K';
Genetic Control of Death in Chinese Hamster
Ovary Cultures
by
Joydeep Goswami
M.B.A., MIT-Sloan School of Management, Cambridge, MA, 1998
M.S.C.E.P., Massachusetts Institute of Technology, Cambridge, MA, 1994
B.Tech. Chemical Engineering, Indian Institute of Technology, Bombay, INDIA, 1993
Submitted to the Department of Chemical Engineering
in Partial Fulfillment of the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
September, 1998
©1998 Massachusetts Institute of Technology. All rights reserved.
Signature of Author:
__ __
c~t7
Department of Chemical Engineering
September 1, 1998
v
Certified by:
O-Daniel I. C. Wang
Institute Professor
Thesis Supervisor
Certified by:
SAnthony
J. Sinskey
Professor of Biology
Thesis Supervisor
Accepted by:
n
c-s-s'-. INSr.TU E
OF TECHNOLOGy TEs
LIBRARIES
Robert E. Cohen
St. Laurent Professor of Chemical Engineering
Chairman, Committee for Graduate Students
1
Genetic Control of Death in Chinese Hamster Ovary
Cultures
by
Joydeep Goswami
Submitted to the Department of Chemical Engineering on September 4th, 1998 in Partial
Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Chemical
Engineering
One of the main problems in mammalian cell culture systems, including Chinese Hamster
Ovary (CHO) cultures, is the inability to maintain viability of these cultures over
extended periods of time. This inability translates into lower final protein titers and
higher protein production and recovery costs. This thesis was undertaken to better
understand the processes of death in CHO cells and to find ways to extend the viability
of CHO cultures.
A majority of CHO cells in culture were found to die by apoptosis, a genetically
controlled form of cellular suicide. Protein synthesis inhibition in CHO cells led to rapid
death, indicating that CHO cells were pre-disposed to death and that survival proteins
needed to be continuously synthesized to protect cells from death. Caspases, a class of
proteins found to be universally important in inducing apoptosis, were found to be
activated in apoptotic CHO cells.
Surprisingly, inhibition of caspase activity using z-VAD.fmk, a universal peptide
inhibitor of caspases, failed to significantly extend viability in batch culture although it did
prevent cleavage of known intracellular caspase substrates. In contrast, expression of bcl2, a well-characterized anti-apoptotic gene, was able to significantly increase the life of
CHO batch cultures in response to both glucose limitation and growth factor withdrawal.
Using these results, a pathway for apoptosis in CHO cells, focusing on the caspases and
bcl-2, was suggested.
An experiment was devised to statistically measure the ability of individual cells
to replicate. Replication competence was found to correlate well with viability results
from the acridine orange / ethidium bromide assay, but not with results from the trypan
blue assay. These experiments proved that early apoptotic cells, which have lost
membrane integrity but not chromatin integrity, can be considered dead since they lose
the ability to replicate. In addition, the experiments proved that bcl-2 expression is able
to extend the replication competence of cells under normal culture conditions.
Bcl-2 expression was shown to improve both final product titers and integrated
viable cell densities in CHO fed-batch cultures. It was also able to maintain insulindeprived fed-batch cultures in a viable and productive state for much longer than insulinsupplemented cultures, thus suggesting an easy way to maintain viability and
productivity coupled with slower growth. A slower growth rate has been reported in
literature to yield higher product quality and productivity.
Concentrations of insulin, a growth and survival factor used in CHO culture, were
observed to drop rapidly in fed-batch cultures of CHO cells. The loss of insulin was
concurrent with the accumulation of cells in the GO/G1 state and an increase in expression
levels ofp53, a well-documented growth-inhibiting and apoptosis-inducing gene. Insulin
degrading activity was found to be at least partially caused by release of proteases from
cells into the culture medium. Insulin degradation was sharply reduced by adding sodium
glycocholate, an amino-peptidase inhibitor, suggesting that amino-peptidases play a major
role in insulin degradation in CHO fed-batch cultures.
Based on the above results, possible ways to further reduce death and improve
productivity in CHO cultures are also suggested.
Thesis Supervisor: Dr. Daniel I.C. Wang
Title: Institute Professor
Thesis Supervisor: Dr. Anthony J. Sinskey
Title: Professor of Biology
Acknowledgments
Although the title of my thesis would suggest that I spent most my time at MIT studying
death, my time here was really not that morbid. I feel that my experiences over the last
five years, both inside and outside the lab, have made me a better all-round individual. I
would thus like to take this opportunity to thank the people who my made my stay at
MIT so thoroughly enjoyable.
First, I would like to thank my advisors Prof. Tony Sinskey and Prof. Daniel
Wang. Thank you for pushing me to look harder for answers and teaching me to focus on
the most relevant issues quickly. These are skills I will value a lot in my career. I also
appreciate that both of you were so very understanding and encouraging when the going
was not easy, especially initially when the transfections refused to work. I also
appreciate that you had the confidence to give me the freedom to explore issues that
interested me and the patience to listen when I explained these issues to you. I
understand that I was not the traditional graduate student, and I want to thank you for
your support and recommendations when I applied to the MBA program at Sloan.
Without your encouragement and understanding, working simultaneously on both the
MBA and Ph.D. would never have been possible. I believe that the biggest support that
an advisor can give his students is by expressing confidence in them and their work, and
both of you did a splendid job of letting this confidence show. Dr. Wang I enjoyed our
tennis games a lot, and I am sorry that we did not get a chance to play more often. Of
course, I must confess that your threats of not letting me graduate till you were able to
beat me did not influence my game. Tony, no matter how stressed out I was, your good
humor always cheered me up. I will always remember the day when I walked into your
office and said that I wanted to apply to Sloan, and you said, "Of course, I would be
disappointed if you did not". Your confidence in me has meant and will continue to mean
a lot to me. Thank you also for the great picnics that you and Prof. Rha had for us. It
was nice to have well balanced individuals as advisors.
I would also like to thank the other members of my thesis committee - Dr. Mike
Glacken, Prof. Doug Lauffenburger, Dr. Morris Rosenberg, Prof. Hermann Steller, Prof.
Greg Stephanopoulos. It was great that you spent so much time with me discussing
various parts of my research and life in general. Hermann, despite your protests, I still
think you are a better tennis player than I am. Thanks for all the help with the apoptosis
assays and the p35 experiments. Greg thanks for having those long meetings with me to
discuss my research and for helping me put everything in perspective. Doug thanks for
all the help with the insulin experiments and for making it to every committee meeting.
Mike and Morris, I am grateful to you two for helping me put my work into perspective
from an industrial standpoint.
I would also like to thank two other members of faculty who were not formally a
part of my thesis committee - Prof. Harvey Lodish and Prof. Charlie Cooney. Harvey
you amazed me with your insight and knowledge, and the extent to which you were
involved in the research in your lab. I went to you several times with problems which
were vexing me, and I never came away disappointed. Charlie, I don't think I have ever
met anyone goes more out to his/her way to help others. Thank you not only for your
insightful comments regarding my research, but also for all your help with getting into and
out of Sloan. Your courage, determination and positive attitude toward life have taught
me a lot. Congratulations again, on scaling Mt. McKinley.
Of course, my studies at MIT would not have been possible without generous
financial support. I would like to thank Unilever Ltd. for awarding me the Unilever/
Lever Brothers Practice School Fellowship during my first semester here. In addition, I
would like to thank the National Science Foundation which provided financial support for
most of my graduate study at MIT. I would like to thank the MIT-Merck Fellowship
Foundation which supported me during the last year. Lastly, I would also like to thank
Merck and Co. for their fellowship in 1995 which supported my trip to San Diego to
attend the Cell Culture Engineering Conference V.
I would also like to express my appreciation to the BPEC staff over the years Audrey Childs, Sonia Foster, Lynne Lenker, Joya Gargano, James Leung, John Galvin,
Darlene Ray, Lorraine Cable and Sara Puffer. Thank you for promptly attending to all
my requests and helping me to work out the various administrative details associated with
the life of a graduate student. I would also like to thank the administrative crew in
Chemical Engineering graduate student office - Janet Fischer and Elaine Aufiero. Life in
ChemE would have been unimaginable without your support - starting with helping me
out with lost forms even before I came to MIT, to sorting out late registrations and the
tuition-detail headaches which come with being registered in two departments at the same
time. Business School types professing quality of service should take a page out of your
book.
I have had the opportunity to supervise some wonderful undergraduate and high
school students over the years - Eve Li, Anya Freedman, Gus Blomquist, Brian Roland,
Brock Bienkowski, Sujata Bhatia, Katiuscia Porter and Kevin Sullivan. Working with you
was a lot of fun, and I think I learnt a lot more from you than I taught you.
Most of my time of course was spent with my fellow grad students at BPEC and
Biology, and from you I have learnt the most during my stay at MIT. Along with all the
lab techniques and academic knowledge that I gained through our interactions, I will also
always cherish our discussions on various life-saving topics (read gossip). Of course, it
was fun to feign the part of evil, rampant, heartless capitalist against Populist Steve!
Thanks to Robert Balcarcel, David Chang, Jianxin Chen, John Chung, Grace Colon, Dave
Schaffer, Brian Follstad, Peter Frier, Sherry Gu, Stephane Guillouet, Bryan Harmon,
Bettina Knorr, Araba Lamous-Smith, Dan Lasko, Phil Lessard, Kai-Chee Loh, Margaret
McCormick, Steve Meier, Gautam Nayar, Gregg Nyberg, Chandra Papudesu, Cliff Rutt,
Anna Sanfeliu, Eric Scharin, Marc Shelikoff, Troy Simpson, Rahul Singhvi, Christi Snell,
Margaret Speed, Espe Troyano, Chia-lin Wei, Bruce Woodson, Inn Yuk, Liangzhi Xie and
Jifeng Zhang. Thanks also to Mike Hengartner who helped me start off my career in cell
death. I would also like to thank Glenn Paradis and Mike Connolly for helping me with
flow cytometry experiments.
I would also like to thank some of the many friends I made during my tenure at
MIT, who kept me sane and prevented me from working too hard. Thanks to Ravi Kane,
Arpan Mahorowala, Robert Balcarcel, Suman Banerjee, Aleks Engel, Tim Benish, Fred
Colhoun, Kiko Aumond, Chris Dowd, Radha Nayak, John Konz, Ravi Srinivasan, Steffen
Ernst, Jeff White and Yoky Matsuoka. It was fun hanging out with you guys. Kiko and
Aleks I will always remember your parties as one of the high points of my graduate life.
Vivek Mohindra and Gokaraju Raju, thank you for all the advice and guidance that you
gave me throughout the years. Thanks also to my friends and teammates from Sloan who
helped me juggle both research and classes - Geoff Bloss, Michelle Go, Michelle WilsonClarke, Claudio Ribiero, Eiji Harada, William Dutcher, Paul Evers, Marc Osofsky,
Preetish Nijhawan, Joel Serface, Stephanie Tan, Debjani Deb, Chuck Goeringher, Michelle
LeBlanc, Frank Martelli, Paul White, Julie Suh and Tina Baumgartner. I also want to
thank my tennis partners for getting onto court whenever I requested them to. Thank
you Yoky, Joel, Colin Walden, Manish Bhatia and Jonathan Seelig. Thanks also to my
squash buddies, Ravi Kane, Ravi Srinivasan, Arup Chakraborty, Preetish, Phillip
Hirschon and Ashu Atwal. Lastly, I would like to thank Jennifer Weiloch for making the
incredibly difficult last year so incredibly wonderful. Thanks for all your help and
support.
Finally, I would like to thank my parents. Ever since I was young you have
supported me and encouraged me more than any parents I know. I would have never
made it here if were not for the time and energy you spent in helping me distinguish right
from wrong, and teaching me to have the courage to do what was right. I know it was
incredibly tough for you to let your only child leave to go thousands of miles away, but I
thank you for making that decision with a smile. Thank you for all the love, caring and
dedication throughout the years. I will always be indebted to you and I dedicate this
Ph.D. thesis to both of you.
I
TABLE OF CONTENTS
ABSTRACT.......................................................................................................................
........................................................
ACKNOWLEDGMENTS ...........................................
.......................................................
TABLE OF CONTENTS...........................................
3
5
9
LIST OF FIGURES................................................................................................................
LIST OF TABLES.................................................................................................................
13
1. INTRODUCTION ..............................................................................................................
21
..............................................
1.1 Background...............................................................
1.2 M otivation ...................... .......................................................................................
......................
1.3 Thesis Objectives .............................................................................
.....................
1.4 Thesis Organization ..........................................................................
21
23
24
25
.....................
27
2. LITERATURE REVIEW ..........................................................................
2.1 Cell Death -Necrosis and Apoptosis .....................................
......
.
20
......
27
27
2.......7.................2....
2.1.1 Necrosis...................................................................................
30
..................................
2.1.2 Apoptosis........................................................................
2.2 Genes Regulating Apoptosis........................................................................................34
2.2.1 p53 - a signaling molecule that is involved in both cell cycle control and apoptosis37
.......... .............................. 40
2.2.2 The Bcl-2 family of proteins......................................
2.2.3 Cysteine Proteases or Caspases.................................................49
2.3 Mitochondria - the meeting ground for the key players in apoptosis ...................... 59
2.4 Are caspases essential for apoptosis?............................................ ............................ .. 60
62
...................
2.5 Apoptosis in M amm alian Cell Culture .........................................
3. MATERIALS AND M ETHODS ............................................................................
65
3.1 Cell Culture ..................................................................................................................
65
3.1.1 Cell Line............................................................................................................. 65
3.1.2 Culture M edium......................................................................................................65
3.1.3 Culture M aintenance.............................................................................................. 66
3.1.4 Caspase-InhibitingPeptide Experiments......................................... 67
67
3.1.5 Fed Batch Culture...............................................................................................
68
...... ....... ...........................
3.1.6 Insulin binding studies....................................
3.2 Analytical M ethods ..................................................................................................... 70
.......................... 70
3.2.1 Cell Number and Viability......................................................
3.2.2 Dry Cell Weight...................................................................................................... 73
73
3.2.3 Sugar and Lactate Assays.......................................................
73
..........................
......
Concentration
3.2.4 Determination of IFN-y
3.2.5 Determination of insulin concentration........................................ 74
3.2.6 Western Blot........................................................................................................... 74
77
..................................
3.2.7 Total Protein Assay ..................................
3.2.8 M easurement of Caspase Activity......................................................................... 77
3.2.9 DNA Ladder Technique for Detection of Apoptosis ................................................. 79
..................... 80
...................................
3.2.10 Cell cycle assay......................................
3.3 Transfection and Cloning of cells ....................................................
81
81
3.3.1 Preparationof plasmids ..................................................................
3.3.2 Transfection of Suspension CHO Cells ................................................. 81
3.3.3 M onitoring Transfection Efficiency.........................................................................83
3.3.4 Cloning ..........................................................
.................................................. 85
4. APOPTOSIS AND ITS CONTROL IN CHINESE HAMSTER OVARY BATCH CULTURE..........87
4.1 CHO Cells Die by Apoptosis in Serum-free Batch Culture ..................................... 88
4.2 Protein synthesis inhibition causes rapid, dose-dependent apoptosis in CHO cells...93
4.3 Cysteine protease inhibiting peptides are unable to significantly enhance viability of
CH O cells in culture ...................................................... ............................................... 95
4.4 Bcl-2 extends viability in batch cultures of CHO cells .....................................
100
4.5 Bcl-2 protects better than caspase inhibiting peptides in response to growth and
survival factor withdrawal...............................
104
4.6 Protective effect of bcl-2 is enhanced in clonal batch cultures ................................ 109
4.7 Caspase-independent death pathways which are blocked by bcl-2 expression appear
to exist in CHO cells ........................................
113
4.8 D iscussion and Conclusions.................................................................................... 118
5. CORRELATING VIABILITY AND REPLICATION COMPETENCE ....................................
125
5.1 Experimental Approach ........................................
126
5.2 The ability of cells to replicate correlates much better with AO/EB assay results than
with TB assay results ................................................... ............................................. 128
5.3 Early apoptotic cells lose the ability to replicate...............................
...... 132
5.4 Bcl-2 is able to prolong the replication competence of cells in culture...................132
5.5 Discussion and Conclusions.................................................
........................ 133
6. IMPROVING CHO FED-BATCH CULTURE PERFORMANCE USING BCL-2 EXPRESSION
................................................................................
....................................................
135
6.1 Bcl-2 expression significantly extends viability and enhances product titers in normal
CHO fed-batch cultures ........................................
136
6.2 Bcl-2 expression allows insulin-deprived fed-batch cultures to survive longer than
norm al fed-batch cultures .......................................... ................................................. 141
6.3 CHO cells progressively arrest in the GO/Gi phase during fed-batch culture........ 147
6.4 p53 expression increases with time in CHO fed-batch culture but the expression of
cyclin E does not change.........................................................
................................... 151
6.5 Specific glucose consumption rate drops off sharply just before viable cell density
stops increasing ..........................................................................
................................ 155
6.6 Discussion and Conclusions.................................................
........................ 156
7. THE FATE OF INSULIN IN CHO FED-BATCH CULTURES .............................................
163
7.1 Reduced binding of insulin to its receptor cannot fully explain cessation of growth in
CHO fed-batch cultures ........................................
164
7.2 Insulin rapidly disappears with time in CHO fed-batch cultures ........................... 165
7.3 Insulin degrading activity in CHO fed-batch cultures is concentrated in the
supernatant................. .......................................................................................
168
7.4 Boiling of fed-batch supernatant removes all insulin degrading activity ................. 173
7.5 Aminopeptidase inhibitors are able to substantially reduce the degradation of insulin
in the fed-batch supernatant.......................................................................................175
7.6 Adding large quantities of excess insulin also reduces the rate of degradation of insulin
in the fed-batch supernatant.......................................................................................178
179
7.7 Discussion and Conclusions..............................
8. CONCLUSIONS AND RECOMMENDATIONS .................................................................
185
.......... 185
8.1 C onclusions ................................................................................................
8.2 Recommendations ................................................................................................... 189
9. REFERENCES .......................................................................................................
195
LIST OF FIGURES
Figure 2-1: Events in necrosis of a typical mammalian cell as characterized by changes in
cellular and organelle morphology. The solid arrows indicate the path of progression
29
of the cell in the necrotic process .................................... ................
Figure 2-2: Morphological features identified in the typical apoptotic death of mammalian
cells. The thick solid arrows indicate the path of progression of the apoptotic
process. All stages of the process need not be observed ..................................... 33
Figure 2-3: A map of bcl-2's (and bcl-xL's) interactions with other proteins which help
explain the mechanisms by which it exerts it antiapoptotic effects. See text for
details of specific protective effects. Adapted from Reed, 1997 ........................... 45
Figure 2-4: Examples of receptor, adapter and protease complexes in regulation of caspase
activity. In this illustration, Fas ligand and TNF- (tumor necrosis factor) receptor
are the receptor molecules. FADD (Fas-associated-death-domain protein), TRADD
(TNF-receptor-associated-death-domain protein), RIP (receptor-interacting-protein),
RAIDD (RIP-associated ICH-1/CED-3 -homologous protein) are the adapter
molecules. Pro-caspase-2 and pro-caspase-8 are the end proteases (or caspases)
which are activated through this signaling pathway. Adapted from Cohen, 1997....54
Figure 2-5: The mitochondrial caspase cascade. Release of cytochrome c from the
mitochondria leads to the activation of caspase-9 in the presence of Apaf-1 and
dATP. Caspase-9 subsequently activates other caspase-3 and potentially other
downstream caspases. Adapted from Li, et al., 1997 ............................................... 58
Figure 2-6: A map of apoptosis focusing on caspases and the mitochondrial bcl-2
The diagrams representing receptor-adapter-protease activation of
proteins.
caspases, bcl-2 family regulation and activation of downstream caspases are shown
in greater detail in Figure 2-4, Figure 2-3 and Figure 2-5, respectively. The release of
AIF (apoptosis inducing factor) is in a positive feedback loop with activation of
downstream caspases. Loss of mitochondrial potential and release of reactive
oxygen species (ROS) by themselves can lead to death, but there is an active debate
in literature as to whether this death is apoptotic or necrotic (see text)................. 61
Figure 4-1: Viable (VCD) and total cell densities (TCD) obtained in normal serum-free
....... 89
batch culture of suspension CHO cells .........................................
Figure 4-2: Apoptosis in batch cultures of CHO cells. Percentage of dead, apoptotic
necrotic cells obtained during a normal serum-free batch culture of CHO cells.
percentage of dead cells is obtained by adding the percentages of apoptotic
necrotic cells. Almost all death is seen to occur via apoptosis. .............................
and
The
and
90
Figure 4-3: Apoptosis in batch cultures of CHO cells. Agarose gel photograph of
genomic DNA from CHO cells shows presence of a DNA ladder which coincides
with massive apoptosis on day 4...................................
...............
92
Figure 4-4: Extent of apoptosis induced by varying doses of cycloheximide (CHX).
Cycloheximide induces protein synthesis inhibition in cells. Apoptosis accounted
for almost all the death observed.............................
.......
.................... 94
Figure 4-5: Fluorescence obtained by combining cell-lysate from apoptotic CHO cells
with a fluorescent substrate z-YVAD.AFC, indicating the presence of caspases.....96
Figure 4-6: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures
CHO cells: Total cell density (TCD) and viable cell density (VCD) as a function
tim e.....................................................................................
..................................
of
of
of
98
Figure 4-7: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of
CHO cells: Total cell density and viability of cultures as a function of time.
Viability declines rapidly after 72 hours ................................................ 99
Figure 4-8: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of
CHO cells: Medium glucose concentration as a function of time. Medium glucose
concentration dropped to almost zero after 72 hours in both cultures .................... 101
Figure 4-9: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of
CHO cells: Poly-ADP-Ribose Polymerase (PARP) cleavage is evident in the control
cultures but not the z-VAD.fmk protected cultures ......................................
102
Figure 4-10: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Total cell density (TCD) and viable cell density (VCD) as a
function of tim e .................................................... .............................................. 105
Figure 4-11: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Total cell density (TCD) and viability of cultures as a function
of time. Viability of the control, but not the bcl-2 protected cell line, declines
rapidly after 72 hours............................................... .......................................... 106
Figure 4-12: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Medium glucose concentration as a function of time. Medium
glucose concentration dropped to almost zero after 72 hours in both cultures...... 107
Figure 4-13: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Poly-ADP-Ribose Polymerase (PARP) cleavage is clearly
evident in the control cultures but almost all the PARP is uncleaved in the bcl-2
protected cultures..................................................................................................... 108
Figure 4-14: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells that have been deprived of
insulin and transferrin. To cultures which did not contain z-VAD.fmk we added a
volume of Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver
z-VAD.fmk to the other cultures (also see text). This graph plots total cell density
(TCD) as a function of time ..................................................................................... 10
Figure 4-15: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells which have been deprived of
insulin and transferrin. To cultures which did not contain z-VAD.fmk we added a
volume of Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver
z-VAD.fmk to the other cultures (also see text). This graph plots viability of
............................ 111
cultures as a function of time ...................................................
Figure 4-16: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells which have been deprived of
insulin and transferrin. To cultures which did not contain z-VAD.fmk we added a
volume of Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver
z-VAD.fmk to the other cultures (also see text). Medium glucose concentration is
112
plotted as a function of time. .....................................
Figure 4-17: Bcl-2 CHO clonal populations are able to maintain their viability for a much
longer period of time as compared to the control, pcd-CHO cells in a batch culture.
The main cause of death is glucose depletion at about 72 hours. The total cell
density (TCD) for both cultures is approximately the same. The variation in total
cell density at the end of the culture is due to cells sticking to the walls of the flask. 114
Figure 4-18: Clonal populations of CHO cells expressing
viability for much longer (as compared to the control
to insulin-transferrin deprivation in batch culture.
density of the cultures..............................................
bcl-2 are able to maintain their
pcd-CHO cell line) in response
TCD refers to the total cell
......................................... 115
Figure 4-19: Cleavage of caspase-6 substrate lamin A in CHO cells, in response to insulin
and transferrin deprivation. 'p' refers to the control pcd-CHO cell line. 'b' refers
to the bcl-2 expressing CHO cell line. '+z' and '-z' indicate whether caspase
inhibitor, z-VAD.fmk was or was not added to the culture. The results indicate that
addition of z-VAD.fmk is able to prevent cleavage of lamin A, but is not able to
protect cells from death. Bcl-2 expression prevents both lamin cleavage and death. 117
Figure 4-20: Conventional pathway of death in the mammalian cells, focusing on the bcl-2
and caspase (C) nodes. S1 and S2 are the same or different external stimuli for
apoptosis.................................................................................................................. 12 1
Figure 4-21: Suggested death (including both apoptosis and necrosis) pathway in a CHO
cell. The pathway focuses on interaction between caspases and bcl-2 family
proteins. FADD, AIF and ROS refer to Fas associated death domain, apoptosis
inducing factor and reactive oxygen species, respectively. See text for more details. 123
Figure 5-1: A pictorial representation of the four kinds of cells which can be distinguished
by the acridine orange/ ethidium bromide (AO/EB) assay. Cells which have not lost
their membrane integrity are penetrated by acridine orange only, appear green under
the microscope, and are lightly shaded in the figure. Cells which have lost their
membrane integrity also incorporate ethidium bromide and appear orange under the
microscope. These cells are hatched in the figure. The dark spots in the early and
late apoptotic cells represent condensed and fragmented chromatin material..........129
Figure 5-2: Acridine orange (AO) viability is a good indicator of the ability of a cell to
replicate. AO viability is defined as the percentage of viable cells in the population
of cells with intact membrane integrity. Normalized replication competence is a
normalized measurement of the number of wells in which the cells were able to
replicate .................................................................................................................... 13 1
Figure 6-1: Total (TCD) and viable cell densities (VCD) of bcl-2 expressing and control
(pcd) cell lines in a normal fed-batch culture seeded at 5x105 cells/mL................ 138
Figure 6-2: Total cell density (TCD) and viability of bcl-2 expressing and control (pcd)
cell lines in a normal fed-batch culture seeded at 5x10 5 cells/mL .......................... 139
Figure 6-3: y-interferon production with bcl-2 expressing and control (pcd) cell lines in a
normal fed-batch culture seeded at 5x10 5 cells/mL............................
142
Figure 6-4: Total (TCD) and viable cell densities (VCD) of bcl-2-protected and control
(pcd) cell lines in an insulin-deprived fed-batch with an initial cell density of 5x10 5
cells/mL ................................................................................................................... 144
Figure 6-5: Total cell density (TCD) and viability of bcl-2 protected and control (pcd) cell
lines in an insulin-deprived fed-batch with an initial cell density of 5x10 5 cells/mL.145
Figure 6-6: y-interferon production with bcl-2 protected and control (pcd) cell lines in an
insulin-deprived fed-batch with an initial cell density of 5x10 5 cells/mL............. 146
Figure 6-7: Total (TCD) and viable cell densities (VCD) of bcl-2-protected and control
(pcd) cell lines in an insulin-deprived fed-batch with an initial cell density of 8x10 5
cells/mL . ................................................................................................................. 148
Figure 6-8: Total cell density (TCD) and viability of bcl-2 protected and control (pcd) cell
lines in an insulin-deprived fed-batch with an initial cell density of 8x1 05 cells/mL. 149
Figure 6-9: y-interferon production with bcl-2 protected and control (pcd) cell lines in an
insulin-deprived fed-batch with an initial cell density of 8x1 05 cells/mL ............ 150
Figure 6-10: Cell cycle trends in fed-batch cultures of clonal populations of pcd (a) and
bcl-2 (b) transfected CHO cells. It can be seen that the fraction of cells in the GO/G1
phase increases monotonically with culture time, while the proportion in the S-phase
drops off. VCD refers to viable cell density of the culture. The viable cell density of
the cultures' stops increasing when more than 70% of cells are in the GO/G1 phase. 153
Figure 6-11: Increase in p53 expression in CHO fed-batch cultures coincides with
cessation of an increase in viable cell density and follows the increase in GO/G1
phase population. Cells transfected with bcl-2 are able to maintain their viability for
a longer period of time despite p53 expression. '*' indicates a sample from a batch
culture in log-phase growth, which was used as a control for p53 expression. 'p'
refers to pcd-CHO cells while 'b' refers to bcl2-CHO cells ................................. 154
Figure 6-12: Cyclin E expression remains constant with culture age. The increase in
GO/G1 phase population is therefore not related to cyclin E expression. '*' indicates
a sample from a batch culture in log-phase growth, which was used as a control for
cyclin E expression. 'p' refers to pcd-CHO cells while 'b' refers to bcl2-CHO cells. 154
Figure 6-13: Specific glucose consumption (SGC) and viable cell density (VCD) with time
in a normal fed-batch culture of CHO cells. Specific glucose consumption drops off
before a fall in viable cell density. ................................................... 157
Figure 6-14: The glucose concentration (glc) of the fed batch media increase sharply after
100 hours due to unavoidable overfeeding and inefficient uptake of glucose. The
specific uptake rate of glucose, when measured per mole of glucose in the medium
(SGCG) continues to fall after about 90 hours in the culture. The rise in specific
glucose uptake rate seen in Figure 6-13 may therefore be explained by the rising
158
concentrations of glucose in the culture (see text)...............................
Figure 7-1: Insulin binding to cells taken from different points in a fed-batch culture. Net
radioactivity is a proxy for the quantity of insulin bound in a specific manner to its
receptor (see text). The higher the net radioactivity value the larger to quantity of
insulin bound. 'Old' and 'new' refer to samples taken from day 7 and day 1 of a
fed-batch culture, respectively .....................................
...............
166
Figure 7-2: Insulin concentration as a function of time in a typical fed-batch culture of
CHO cells. The initial concentration of insulin in the culture is 5 mg/L. Insulin
concentrations drop rapidly to almost non-detectable levels by 100 hours. Culture
growth stops soon after. The suffix '-ins' refers to the concentration of insulin in
either the bcl2-CHO or pcd-CHO cell culture. VCD refers to the viable cell density
in either culture...........................................................................................
........... 167
Figure 7-3: Degradation of insulin in fed-batch supernatant (culture medium with cells
spun down and removed) from various days in a CHO fed-batch culture. The
degradation of insulin in samples from each day, under cell-culture conditions, was
followed for 24 hours. 5 mg/L of insulin was added to the supernatant at '0' hours,
and the concentration of insulin measured in the supernatant at this point was
denoted as 100%. Insulin in any particular sample was expressed as a fraction of the
insulin concentration in the same sample at '0' hours ............................................. 170
Figure 7-4: Degradation of insulin in fed-batch medium (with cells) from various days in a
CHO fed-batch culture. The degradation of insulin in samples from each day was
followed for 24 hours. 5 mg/L of insulin was added at '0' hours, and the
concentration of insulin measured in the supernatant at this point was denoted as
100%. Insulin in any particular sample was expressed as a fraction of the insulin
concentration in the same sample at '0' hours .................................................
171
Figure 7-5: Degradation kinetics of insulin in fresh (cell-free) fed-batch medium under
standard cell-culture conditions. 5 mg/L of insulin was added to the medium at 'O'
hours. The concentration of insulin in the medium at various time points was
expressed as a fraction of the concentration at '0' hours ..................................... 172
Figure 7-6: Boiling fed-batch supernatant for three minutes removes its insulin degrading
activity. 5 mg/L of insulin was added to the boiled sample and unboiled control at
'0' hours. Both samples were incubated under standard cell-culture conditions. The
concentration of insulin in the medium at various time points was expressed as a
fraction of the concentration at '0' hours................................
174
Figure 7-7: Effect of various protease inhibitors on the degradation of insulin in the
supernatant from a fed-batch culture. 'All' indicates that all the protease inhibitors
were added to this sample at the concentrations suggested. No protease inhibitors
were added to the 'control' culture. 5 mg/L of insulin was added to the medium at
'0' hours. The concentration of insulin in the medium at various time points was
expressed as a fraction of the concentration at '0' hours. All samples were incubated
under standard cell-culture conditions for the duration of the experiment............1...77
Figure 7-8: Adding excess insulin reduces the rate of insulin degradation, possibly due to
the saturation of proteases or insulin-binding proteins in the fed-batch supernatant.
The control indicated has IX or 5gSg/mL of insulin added to it at time zero. The
concentration of insulin in the medium at various time points was expressed as a
fraction of the concentration at '0' hours. All samples were incubated under
standard cell-culture conditions for the duration of the experiment ...................... 180
LIST OF TABLES
Table 2-I: Changing views in apoptosis ..................................................
Table 2-II: Caspases and their target substrates. .......................................
35
..... 55
Table 3-I: Antibodies used and the companies from which they were sourced..............78
Table 5-I: Normalized replication competence (norm. repl. comp.), acridine orange
viability (AO viab.) and trypan blue viability (TB viab.) data............................130
Table 6-I: A comparison of the results from a normal fed-batch using bcl-2 transfected
and control (pcd) clonal cell lines. IVCD represents integrated viable cell density and
is the area under the viable cell density curve (plotted against time) ....................... 142
Table 6-II: A comparison of the results from the two insulin-deprived fed-batch runs. All
cultures contained clonal population of cells. Run 1 was started with an initial cell
density of 5x10 5 cells/mL while Run 2 had an initial cell density of 8x10 5 cells/mL.
IVCD represents integrated viable cell density and is the area under the viable cell
density curve (plotted against time).........................................................................146
Table 7-I: Concentrations and sources of various protease inhibitors used to study insulin
degradation in CHO fed-batch culture supernatants ............................................... 176
1. Introduction
1.1 Background
Humans had been unknowingly using bacteria and yeast to produce foods and drinks such
as bread, cheese, yogurt, wine, mead and beer long before Pasteur invented the science of
microbiology in the mid-1800s. The first recorded use of microorganisms to fight disease
was by Edward Jenner in 1796, when he used the material from cowpox lesions as the
first effective vaccine against smallpox.
Until the mid-1900s, the initial uses of
microorganisms on an industrial scale remained confined to the food and beverage industry
and the production of simple chemicals such as ethanol, acetone and citric acid. The
discovery of the antibiotic penicillin in 1928 by Alexander Fleming gave a huge boost to
the biological fermentation industry used for pharmaceutical purposes.
The antibiotic
industry remains a multi-billion dollar industry.
The discovery of the double-helix in 1953 by Watson and Crick, and the
subsequent developments in technologies in manipulating DNA led to the development of
the biotechnology industry in the 1970s. Once the gene coding for a particular protein
was identified, recombinant DNA technology allowed the production of this protein in
almost any organism. This meant that large quantities of therapeutic and diagnostic
proteins that were difficult to isolate from natural sources, could now be produced
industrially in optimally designed host organisms. In 1982, human insulin became the
first recombinant pharmaceutical product to be approved by the Food and Drug
Administration.
The choice of host organism for the production of recombinant product is dictated
by several factors, including productivity of host and complexity of the desired protein.
Bacterial production is often the most simple, allowing large quantities of product to be
manufactured in a short time. However, bacteria are not able to form intramolecular
disulfide bonds or glycosylate a protein, both of which are often key to the biological
activity of more complex proteins.
Nevertheless, bacteria remain important for the
production of amino acids, vitamins and simple proteins like insulin. Production in yeast
offers the advantages of rapid production and high productivities obtained in bacterial
culture. Yeast secrete soluble proteins and are able to form intramolecular disulfide
bonds, but are unable to correctly glycosylate most complex proteins. Yeast is still the
organism of choice for certain biologics. Animal cell culture, using transformed cells from
multicellular animals, is still the only way to manufacture most complex proteins.
However, animal cells are slow growing, fragile and offer much lower productivities than
bacterial or yeast systems. In addition to production of complex and therapeutic proteins
such as erythropoietin, tissue plasminogen activator, monoclonal antibodies and
interferons, animal cell culture is also used to produce viruses for vaccines and
bioinsectisides. The recent use of transgenic animals to manufacture proteins in their milk
has the potential to revolutionize high throughput manufacture of some complex proteins.
However, for proteins that may be toxic to the animal in large quantities and for the rapid
manufacture of proteins prior to FDA approval (there is not enough time to develop
transgenic animal factories), animal cell culture will continue to play an important role.
1.2 Motivation
One of the major problems in animal cell culture is the high cost of recovering secreted
product from the medium, primarily due to the low concentrations of product in the
medium. The main cause of low product concentration is that viability of cells in batch or
fed-batch cultures cannot be sustained for sufficiently long periods of time due to their
sensitivity to the culture environment and the accumulation of toxic metabolic products.
Previous studies have indicated the potential of designing medium and feeding
strategies to reduce the production of toxic metabolites such as lactate and ammonia and
increase the viability and productivity of cultures.
Another way to increase the
concentration of the product in the medium would be to improve viability of the cells by
increasing the tolerance of the cells to toxic metabolic end products. Recent studies have
indicated that a substantial number of the cells that die in vitro during culture, do so by
apoptosis, a genetically controlled form of cellular suicide.
However, at the time of
initiation of this study there were no reported studies of apoptosis in Chinese Hamster
Ovary (CHO) cells.
Since apoptosis is a genetically controlled process, a cell's
commitment to apoptosis, can be altered by recombinant expression of genes controlling
the process. To be effective in controlling apoptotic death via genetic means, the genes
controlling the apoptotic process in CHO cells first needed to be identified.
A clearer
understanding of how these genes are ordered in the molecular pathway is also necessary
to determine which genes will be the most effective at preventing death. The effect of
expression of death controlling genes on the productivity of y-interferon, the model
protein expressed by CHO cells studied in our lab, also needed to be studied. Combining
the use of feeding strategies and the use of genetic alteration of cells could provide means
to further extend viability and improve product concentrations. The use of these genetic
techniques could potentially lead to more robust and hardy cell lines and a more costeffective manufacturing process for life-saving drugs and diagnostic proteins.
1.3 Thesis Objectives
The central goal of this thesis was to improve viability in CHO cultures, as a means to
improving final product concentrations. To achieve this goal, the process of death in the
CHO cell was studied at a molecular and genetic level. The specific objectives of these
studies were:
*
To study the process of death in CHO cells and determine whether apoptosis is the
main mode of death
*
If apoptosis is the main mode of death, determine which proteins or genes are the key
players in the apoptotic pathway of CHO cells, how they interact, and the molecular
ordering of this pathway with respect to these players
*
Determine whether the viability of CHO cultures can be improved by overexpressing
or inhibiting some of the key players in the apoptotic pathway
*
Determine whether the viability and productivity of CHO fed-batch cultures can be
extended beyond that obtained by using just medium design and feeding strategies, by
combining these strategies with genetic modulation of the death process
In addition to these objectives, the thesis also sought to develop techniques to
examine viability at a more fundamental level and studied the interaction of the growth
and survival factor insulin with key apoptotic pathway elements in the CHO cell during
the culture process.
1.4 Thesis Organization
This thesis is organized into eight chapters. The first chapter introduces the research
topic and outlines the specific objectives of the thesis. Chapter 2 contains a detailed
review of cell death and the key players in apoptosis.
Chapter 3 provides a complete
description of the techniques and protocols used in the thesis. Chapter 4 describes the
results of experiments conducted to determine the mode of death in CHO cells. It also
compares the results from attempts to control death in CHO cultures using expression of
the antiapoptotic gene bcl-2 and addition of peptide inhibitors of specific apoptotic
proteins (caspases). This chapter concludes with an analysis of the apoptotic pathway
in CHO cells. The results presented in Chapter 5 confirm that the viabilities measured in
Chapter 4 correspond to the ability of cells to replicate (a more stringent test of viability).
Chapter 6 presents the results of experiments conducted to investigate the effect of bcl-2
expression on enhancing performance in CHO fed-batch cultures.
The chapter also
examines some of the underlying causes for cessation in growth of viable cells in CHO
fed-batch cultures. Chapter 7 follows the fate of growth and survival factor insulin in fedbatch cultures and suggest methods to reduce its rapid disappearance in these cultures.
Finally, Chapter 8 summarizes the results presented in this thesis and presents
suggestions for future work in this area.
2. Literature Review
2.1 Cell Death -Necrosis and Apoptosis
The study of cell death has not received due recognition until very recently. This may
partly have been due to the rather restricted notion of cell death as a degenerative
phenomenon caused by injury or age. However, recent studies have clearly shown that
this notion is only partially true.
eukaryotic organisms.
Two distinct forms of cell death exist in most
The first subscribes to the previously held 'degenerative
phenomenon' idea and is called necrosis.
The other process, known as apoptosis,
involves the active self destruction of cells (consuming energy and sometimes involving
protein synthesis) rather than just passive degeneration. The two processes of cell death
can be distinguished by various morphological and biochemical criteria and also by
identifying the circumstances of death. The following discussion elucidates some of the
characteristic features of necrosis and apoptosis.
2.1.1 Necrosis
The occurrence of necrosis is determined, not by factors intrinsic to the cell, but by
violent perturbation in the environment (Wyllie, et al., 1980).
Examples of such
environmental perturbation include severe lack of oxygen (hypoxia) or blood (ischemia),
sudden large fluctuations in environmental temperature (hypo-
or hyperthermia),
disruption of cell membranes by injury or exposure to large doses of toxins. Necrosis is
also the mode of cell death by autolysis, in vitro. There is no evidence of necrosis in
normal embryonic development and metamorphosis (Wyllie, et al., 1980).
Necrosis is seen to affect cells in groups rather than individually.
Early
morphological changes brought on by necrosis include the marginal clumping of loosely
textured chromatin, dilation of the endoplasmic reticulum, mild dispersion of ribosomes
and gross swelling of the mitochondrial matrix (see Figure 2-1).
The late stages of
necrosis are characterized by the rupture of nuclear, organelle and plasma membranes
(leading to a loss of organized structure), the dissolution of ribosomes and lysosomes and
finally to swelling and rupture of the cell. This burst of cells and spewing out of cell
contents causes the familiar exudative inflammation. The debris is eventually ingested and
degraded by specialized phagocytic cells.
The biochemical basis for this observed rupture of membranes arises from a
marked increase in membrane permeability. This increase in permeability is accompanied
by potassium loss, sodium entry and a fall in membrane potential. The rise in membrane
permeability is also reflected in the failure to exclude vital dyes such as nigrosine, trypan
blue, propidium iodide and eosin Y, thus forming a basis for viability tests. The rise in
membrane permeability may be a result of toxins directly attacking membrane function.
Alternatively, toxins may interfere with the energy supply on which the selective control
ER
nucleus _
mitochondrion--
* swollen mitochondrial matrix
* increase in membrane permeabilityn
* cell membrane ruptures
* cell contents spewed out
* inflammation
lysosome
o
H2(
Figure 2-1: Events in necrosis of a typical mammalian cell as characterized by changes in
cellular and organelle morphology. The solid arrows indicate the path of progression of
the cell in the necrotic process.
of fluxes across the plasma membrane depends (Trump and Mergner, 1974).
Hence,
necrosis does not require the active use of cellular energy.
2.1.2 Apoptosis
The word apoptosis is derived from ancient Greek for "falling off of leaves" and was first
used to describe the process of physiological cell death by Kerr, et al., 1972. It occurs
normally in the steady-state kinetics of healthy adult tissues producing a normal turnover
of cells. Salvesen and Dixit, 1997 estimate that as many as 1011 cells die by apoptosis
each day in a normal human being. It is interesting to note that this mechanism for 'cell
suicide' has not developed in single cell organisms (although there have been some reports
to the contrary, see Anderson, 1997), but multicellular organisms have found many uses
for this mechanism. It accounts for the deletion of cells during embryonic development
and metamorphosis, e.g., death of more than 50% of the neurons during the development
of the vertebrate nervous system (Raff, et al., 1993) and the metamorphic degradation of
the cells in the tadpole's tail. It may be brought about by a change in concentration of
trophic hormones, e.g., removal of the endometrial epithelium by apoptosis after estrogen
withdrawal in mice.
In addition to morphogenetic and developmental functions,
apoptosis is also a major defense mechanism. This is aptly demonstrated in plants where
a cell detecting a single bacterium not only kills itself, but also alerts neighboring cells
which then follow suit.
Apoptosis can also be induced by cell-mediated immunity, e.g., cytotoxic T
lymphocytes (CTLs) recognize virus-bearing cells and cause them to undergo apoptosis
by transferring enzyme-bearing granules through pores in their surface formed with
perforin (see Vaux, et al., 1994 for a review). Failure of cells with defective genomic
DNA to die via apoptosis often causes cancer, but the fact that we can induce apoptosis
in cells with radiation, cytotoxic cancer-chemotherapeutic agents and hyperthermia
provides us with an important tool to fight cancer. In addition, excessive or uncontrolled
apoptosis can also lead to pathogenesis. For instance, unregulated apoptosis appears to
be the cause of several neurodegenerative diseases such as Parkinson's disease,
Alzheimer's disease, amyotrophic lateral sclerosis (ALS) etc. (for a review see
Thompson, 1995). As mentioned above, small concentrations of many necrotic agents
have been observed to induce apoptosis (Martin and Cotter, 1994). In fact there appear to
be crossover concentrations for most drugs under which cells die increasingly by
apoptosis than necrosis. This suggests that it is the severity of the insult to cells that
determines the type of death the cell will undergo, in addition to the type of insult itself.
The morphological features of apoptosis are very distinct (see Wyllie, et al., 1980
and Martin and Cotter, 1994 for reviews). In vivo, the cell initially loses contact with its
neighbors. The chromatin condenses into large compact granular masses that lie adjacent
to the nuclear membrane.
Chromosomal DNA is degraded at nucleosomal intervals
(approx. 180 base pairs) leading to the formation of the characteristic 'DNA ladder'. The
nuclear outline is abnormally convoluted and later grossly indented, and the nucleolus
enlarges with its granules becoming coarse and abnormally scattered. There is usually a
marked reduction in cell volume.
Most of the cell organelles, however, retain their
integrity except for the endoplasmic reticulum, which dilates (see Figure 2-2).
There is
also some recent evidence that other organelles such as the mitochondria also dilate during
apoptosis (Vander Heiden, et al., 1997). In the later phases of apoptosis, cell membrane
blebbing leads to cellular fragmentation which in turn leads to the formation of 'apoptotic
bodies'.
These apoptotic bodies contain nuclear fragments as well as cytoplasmic
elements. Finally, neighboring cells and macrophages phagocytose the fragments. Unlike
necrosis, apoptosis does not lead to an inflammatory response or the formation of
residual scars (Fesus, et al., 1991).
In addition, apoptosis is an astonishingly rapid
process (Wyllie, et al., 1980) and cells undergoing apoptosis may disappear in vivo
within 4 hours.
Apoptotic cells do not display increased membrane permeability until well after
the characteristic morphological features have appeared (Wyllie, et al., 1980). This may
explain the initial insensitivity of apoptotic cells to vital dye exclusion-based viability
tests. Furthermore, apoptosis requires the use of energy-rich nucleotides, e.g., death is
accelerated in palatal shelf epithelium by artificially raising the level of cyclic adenine
monophosphate (cAMP).
This energy may be required in the synthesis of enzymes
such as endonucleases which serve to cleave chromosomal DNA, or in the various
apoptotic signaling cascades (Li, et al., 1997). There is evidence to suggest that the
activity of these nucleases is associated with the concentration of divalent cations such as
calcium (Cohen, 1991; Wyllie, et al., 1980). Energy is also used in protein synthesis, as it
is known that many pathways of apoptosis are genetically coded for and require
R
nucleus
mitochondrion
lysosom
Chromatin and cytoplasm condense
Organelles shrink but retain identity
Plasma membrane blebs
Chromosomal DNA degraded
Convoluted membrane
Plasmalemmal sealing - compartments
Apoptotic bodies
fragmented
chromatin
Figure 2-2: Morphological features identified in the typical apoptotic death of mammalian
cells. The thick solid arrows indicate the path of progression of the apoptotic process.
All stages of the process need not be observed.
transcription and translation of the genetic code.
Some observations that protein
inhibitors sometimes induce rather than inhibit apoptosis are explained by the hypothesis
that proteins which induce apoptosis have already been synthesized and are countered by
active synthesis of proteins which inhibit apoptosis. Protein synthesis inhibition in these
cells thus induces apoptosis by preventing the synthesis of apoptosis inhibitor genes
(Cuende, et al., 1993; Gottschalk, et al., 1994; Mercille and Massie, 1994; Mosser and
Massie, 1994).
The field of apoptosis has been in a state of flux, and much of the original
concepts are themselves undergoing apoptosis!
Table 2-I lists some of the early
assumptions in the apoptotic process which are now known to be incorrect (adapted
from Kroemer, 1997). This table should serve as a word of caution to the reader that not
much in the world of apoptosis can be regarded as absolute truth.
2.2 Genes Regulating Apoptosis
Since apoptosis is genetically programmed process, the apoptotic process is controlled
both by genes that induce the cell to die and those that protect the cell from death. The
process of apoptosis can be subdivided into four different phases - initiation, signaling,
effector and
degradation.
Addition
of specific
apoptotic
stimuli,
such
as
chemotherapeutic drugs, or the loss of survival factors such as IGF-1 and interleukins, can
initiate the death process either by turning on the death genes (such as those of the
cysteine protease or caspase family) or by suppressing the activity of the survival genes
(such as those of the bcl-2 family). The diverse stimuli, in almost all cases (except in the
Table 2-I: Changing views in apoptosis
Invalidation
Assumption
Apoptosis involves the de novo synthesis Apoptosis can be induced in all cells in the
of 'killer genes'
presence of cycloheximide
Apoptosis is an abortive cell cycle
Apoptosis can be induced during any phase
of the cell cycle
Apoptosis is a nuclear process
Cytoplasts (enucleated cells) can undergo
receptor-mediated,
bcl-2-regulated
cell
death
Apoptosis is mediated by reactive oxygen Apoptosis can induced in/by the absence of
species
oxygen, in a bcl-2-regulated fashion
Apoptosis involves an elevation of Ca+2 in Ca+2 depletion
can
induce
apoptosis.
Nuclear apoptosis can be induced in Ca+2
some subcellular compartment
free media
Apoptosis is due to cytosolic acidification
Acidification inhibits apoptosis in some
cases
Apoptosis
always
involves
specific Bax and Bak overexpression induce death in
proteases (caspases)
the presence of caspase inhibitors
Apoptosis does not involve mitochondria
Cell-free systems of apoptosis
require
mitochondria or mitochondrial products
case of direct DNA damage to the cell), have to transmit the apoptotic signal through a
series of signaling molecules and signaling events from the outside of the cell, through the
cytosol and to the nucleus to affect gene transcription and translation.
The map of
signaling pathways, which include FADD (Fas-associated death domain, see Nagata, 1996
for a review) and TRADD (TNF-receptor-associated death domain, see Vaux and
Strasser, 1996 for a review), is still incomplete.
There is also evidence, that in some
instances this signal transduction process is short-circuited. For instance, most caspases
already exist in the cytosol as pro-forms of the protein, and are cleaved to the active form
via auto-catalyzed reactions, in response to apoptotic stimuli.
These pro- and anti-
apoptotic genes and their protein products comprise the effector phase of apoptosis. In
turn, the effector phase proteins, use common downstream degradation proteins such as
nucleases, transglutaminases etc. which actually enable the process of death in the cell.
The regulation of apoptosis can occur throughout out the cell - in the nucleus, the cytosol
and also, as we shall see later in cell organelles such the mitochondria and endoplasmic
reticula. Apoptosis is thought to be beyond regulation once the degradation phase is
reached.
Much of the original work in identifying the key genetic players in apoptosis was
done by studying the process of programmed cell death in the nematode, Caenorhabditis
elegans. Three basic genes - ced-3, ced-4 and ced-9 (ced for C. elegans death)- control the
process of apoptosis in these animals through mutual interactions (Ellis, et al., 1991;
Hengartner, 1996). Subsequently, the mammalian homologues of these basic genes have
been discovered, once again pointing to the astonishingly conserved nature of the
pathway of apoptosis in evolution. The following sections review some of the signaling
pathways and the mammalian homologues of the ced genes.
2.2.1 p5 3 - a signaling molecule that is involved in both cell cycle control and
apoptosis
p53 codes for a 53-kilodalton nuclear phosphoprotein that has specific DNA-binding
capability and can positively and negatively regulate transcription. Normal p53 works as
a guardian of the genome (Lane, 1992). It checks for mutations in genomic DNA during
cell replication. If it finds mutations or nicks in genomic DNA, it arrests cell division in
the Gl phase (Lin, et al., 1992; Marx, 1993; Mercer, et al., 1990; Sherr, 1994). The nicks
may be subsequently repaired but it is not clear whether expression of p53 causes it to
proceed with its suicide program (Clarke, et al., 1993) or return to its normal cell cycle
after damages have been repaired (Lowe, et al., 1993). p 5 3 has also been shown to have
3'-5' exonuclease activity which is believed to be important in DNA replication,
recombination and repair (Mummenbrauer, et al., 1996), an observation that further
validates its role as the guardian of the genome. Failure to correct damage to or mutation
of genomic DNA leads to induction of apoptosis by p53 (Clarke, et al., 1993; Han, et al.,
1996; Ko and Prives, 1996; Lowe, et al., 1993; Lowe, et al., 1993; Stasser, et al., 1994).
Small quantities of p53 are always present in the cell, but the protein is rapidly
turned-over (Ko and Prives, 1996). The activation of p53 is post-transcriptional perhaps
through phosphorylation or other post-translational means (Ko and Prives, 1996; Oren
and Prives, 1996). These post-transcriptional and post-translational modifications also
stabilize the p53 protein. The susceptibility to induction of apoptosis by p53 is linked
to the amount of p53 protein that is synthesized.
Thus, p53-heterozygous mice
(containing only one copy of the p53 gene) displayed an apoptotic response to ionizing
radiation and topoisomerase inhibitors (which prevent DNA repair) which was
intermediate between that displayed by wild-type (having 2 copies of the p53 gene) and
p53 homozygous mutant (a p53 knockout with no copies of the p53 gene) (Clarke, et al.,
1993; Lowe, et al., 1993).
However, the observation that treatment with apoptotic
agents like phorbol ester/calcium ionophore, dexamethasone (Lowe, et al., 1993) and
glucocorticoid (Clarke, et al., 1993) lead to similar extents of apoptosis in wild-type,
heterozygous mutants and homozygous mutants, indicates that p53-independent
apoptotic pathways exist.
Some of the above-described experiments suggest that p53 induces apoptosis only
when damage to DNA occurs, suggesting that its role is limited to damage control
pathways.
Another interesting observation made by Clarke, et al., 1993, is that
cycloheximide (a protein synthesis inhibitor) blocks apoptosis caused by both p53dependent and p53-independent pathways. This suggests that the two pathways may
share common elements that require protein synthesis downstream of p53 and that p53
expression may not be required to activate these downstream proteins. Studies have also
indicated that though p53 expression is sometimes necessary for apoptosis, it is not
sufficient for cell death (Lowe, et al., 1993). For instance, irradiation of normal cells
stabilizes p53, but causes cell cycle arrest without apoptosis (Kastan, et al., 1992).
Finally, cells lacking p53 are also known to undergo apoptosis under the influence of
chemotherapeutic agents (causing DNA damage), if these are used at sufficiently high
concentrations (Sen and D'lncalci, 1992). Hence, even DNA damage related death is
possible in the absence of p53, depending on the level of the damage.
Several downstream events by whichp53 effects its control on cellular growth and
apoptosis have been identified. It was observed that expression of p53 lead to the
activation of a gene, alternately called Cipl or WAF1, that produces a 21 kilodalton
protein (El-Deiry, et al., 1993; Harper, et al., 1993; Marx, 1993; Sherr, 1994).
This
protein binds to an enzyme called cyclin-dependent kinase 2 (Cdk2). Cdk2 is responsible
for preparing cells to divide by pushing them out of the GI growth phase of the cell cycle
into the DNA-synthesizing phase (S). The binding of Cipl to Cdk2 blocks Cdk2's
growth-spurring ability.
p53 is also thought to control cell cycle progression by
activating the GADD45 gene (Kastan, et al., 1992) and the cyclin G gene (Okamoto and
Beach, 1994).
p53 also induces cells with a damaged genome to undergo apoptosis via several
mechanisms, although once again the directory of genes activated by p53 is by no means
complete. p53 is known to activate the transcription of an apoptosis inducing member of
the bcl-2 family, Bax (Han, et al., 1996; Matsuyama, et al., 1998; Miyashita and Reed,
1995). p 5 3 is also known to transactivate the insulin-like-growth-factor binding protein 3
(IGF-BP3) gene (Buckbinder, et al., 1995). IGF-BP3 is known to inhibit the action of
IGF-1 and its secretion can adversely affect mitosis and cellular survival. Lastly, p53 has
also been reported to be associated with the Fas apoptotic pathway (Oren and Prives,
1996).
Although p53 is an important signaling molecule in the apoptotic pathway, p53
independent apoptotic pathways clearly exist. In addition, the pathways by which p53
induces death are closely linked to those associated with the bcl-2 family of proteins that
are discussed below.
2.2.2 The Bcl-2 family of proteins
Although the most famous member of this family, bcl-2 (from b-cell lymphoma-2), is one
of the strongest anti-apoptotic genes, the bcl-2 gene family includes several other antiapoptotic as well as pro-apoptotic genes. Bcl-2 itself has strong sequence homology to
the ced-9 anti-death gene in C. elegans (Hengartner and Horvitz, 1994). The other antiapoptotic members of the family include bcl-xL Mcl-1, Al and BI-1, Bcl-w, Bfl-1, Brag-i
(Boise, et al., 1993; Kozopas, et al., 1993; Lin, et al., 1993; Xu and Reed, 1998, while the
pro-apoptotic family members include Bax, Bak, Bcl-xs, Bad and Bik (Boise, et al., 1993;
Boyd, et al., 1995; Chittenden, et al., 1995; Kiefer, et al., 1995; Oltvai, et al., 1993; Yang,
et al., 1995, also see Yang and Korsmeyer, 1996). In addition, genes such as Bag-i,
interact with bcl-2 family of proteins to enhance anti-apoptotic activity (Takayama, et
al., 1995).
Most members of the bcl-2-gene family possess a carboxy-terminal
transmembrane region, although Bid and Bad are notable exceptions. This transmembrane
region influences the subcellular localization of bcl-2-family proteins.
In addition, members of the bcl-2 protein family contain varying amounts of four
so-called bcl-2 homology (BH) domains, BH1 to BH4. The BH1 and BH2 domains are
thought to allow bcl-2 proteins to form homo- or hetero-dimers with other members of
the family. The formation of these dimers allows a level of control in the regulation of
death by this family of proteins. The BH3 domain is required for death agonists, such as
Bax and Bak, to heterodimerize with death-suppressing members of the family and for
their apoptosis promoting function (Zha, et al., 1996). In addition, swapping a 23 aminoacid segment surrounding BH3 from Bax to bcl-2, converted bcl-2 to a death agonist
(Hunter and Parslow, 1996). The BH3 domain thus serves as a potent death-domain
necessary for the apoptosis inducing activity of the death agonist family members
(Kroemer, 1997). The BH4 domain, in contrast, is conserved in all anti-apoptotic bcl-2
family members, and is lacking or ill-conserved in all the proapoptotic members, with the
exception of bcl-xs.
Deletion of the BH4 domain from death-antagonistic family
members, did not inhibit dimerization, but imparted either a loss of function or dominant
negative phenotype (yielding mutants that promote rather than suppress apoptosis, see
Hunter, et al., 1996).
The following sections discuss the activity of anti-apoptotic members of the bcl-2
family in general and some of the emerging mechanisms by which these genes are thought
to exert their effects.
2.2.2.1 Death-antagonisticfamily members
The most well known death antagonistic family members are bcl-2 and bcl-xL. bcl-2 codes
for a 26 kDa integral membrane protein, and was first discovered in patients with B-cell
malignancies. Bcl-2 and bcl-xL observed to prevent apoptosis in response to a large
number of apoptotic stimuli (Lincz, 1998; Reed, 1994; Yang and Korsmeyer, 1996).
Expression of bcl-2 inhibits apoptosis brought about by a withdrawal of growth factors,
IL-3, NGF etc. In all these cases, bcl-2 alone did not affect the proliferation rate of cells,
although there have been some recent reports that bcl-2 delays cell-cycle entry in some
cell lines (Huang, et al., 1997). Bcl-2 overexpression allowed cells to survive in the
absence of growth factors but without further proliferation. In contrast, reductions in bcl2 levels using antisense-mediated techniques, accelerated cell death, showing a clear
function of bcl-2 in reducing apoptosis (Reed, 1994). In all cases examined, expression of
bcl-2 prevented the characteristic morphological changes associated with apoptosis such
as cell shrinkage, chromatin condensation, nuclear fragmentation and DNA laddering,
implying that bcl-2 blocks a relatively early event associated with apoptosis. The extent
of protection offered by bcl-2 was proportional to the level of bcl-2 expression in the
cells (Hockenberry, et al., 1993).
Bcl-2 prevents oxidative injury from resulting in
apoptosis (Hockenberry, et al., 1993). Lipid peroxidation, which is a characteristic event
observed in apoptosis and induced by a large number of apoptotic stimuli, was also
blocked by bcl-2 (Hockenberry, et al., 1993; Kane, et al., 1993).
In addition, bcl-2
prevents induction of apoptosis in response to diverse stimuli such as nutrient depletion,
accumulation of toxins, viral infection, hypoxic
and hyperoxic conditions,
and
hydrodynamic stresses (Singh, et al., 1996; Levine, et al., 1993; Olsen, et al., 1996;
Shimizu, et al., 1996). Bcl-2 and bcl-xL have also been reported to protect against stimuli
which normally kill by necrosis (Kane, et al., 1995; Shimizu, et al., 1996; Vander Heiden,
et al., 1997). It is thought that many of these necrotic agents induce apoptosis by causing
a fall in mitochondrial membrane potential, which bcl-2 by virtue of its sub-cellular
localization is able to prevent (see below).
Bcl-2 has been observed to counter the ability of p53 to induce apoptosis in
response to DNA damage induced by various stimuli (Chiou, et al., 1994; Stasser, et al.,
1994). Bcl-2 did not, however, prevent or reduce DNA damage or enhance DNA repair
rates in any of these cases, suggesting that it acted downstream of these events and
prevented DNA damage from being translated to an apoptotic signal. Bcl-2 was also
unable to prevent cell cycle arrest induced by p53. Growth-arrested cells remained viable
and when the expression of wild-type p53 was turned off, these cells were able to
proliferate even after 10 days in the growth arrested state (Chiou, et al., 1994). Bcl-2 also
appears to be very specific in its action of preventing cell death.
For instance, it
completely inhibits apoptosis induced by c-myc in response to growth factor withdrawal,
but does not affect c-myc's mitotic ability (Bissonnette, et al., 1992; Fanidi, et al., 1992;
Lotem and Sachs, 1993; Vaux, et al., 1988). Bcl-xL's protective effects closely mimic
those of bcl-2 (Boise, et al., 1993; Reed, 1997).
2.2.2.2 Mechanisms of action
Although bcl-2-related proteins have been recognized as key players in apoptosis for a
while now, their mechanism of action is just being clarified. Subcellular localization, levels
of expression, dimerization, post-translational modification and pore-formation sit at the
core of these new theories of bcl-2 action, and are discussed below (also see Figure 2-3 for
a summary of bcl-2 mechanisms of action). In addition, RNA splicing, as in the case of
bcl-xL and bcl-xs, provides another level of control in the manner in which the bcl-2 family
regulates death (Boise, et al., 1993).
2.2.2.2.1 Subcellular localization
Since most members of the bcl-2 family have transmembrane domains, membrane
localization is thought to be necessary for the proteins to perform their proper function in
regulating apoptosis.
Family members lacking transmembrane domains are thought to
exert their function by binding to and affecting the function of the membrane-localized
family members.
Bcl-2 family members have been detected localized to the nuclear
envelope, parts of the endoplasmic reticulum (ER) and the membranes of the
mitochondria (Kroemer, 1997; Reed, 1994). Although there has been some speculation
that the bcl-2 family may regulate apoptosis by controlling Ca +2 release from the ER
(Lam, et al., 1994; Orrenius, et al., 1992) and by being involved in some aspect of nuclear
transport or nuclear envelope assembly and maintenance (Krajewski, et al., 1993), not
much evidence supporting these hypotheses has been found. In addition, studies in cellfree systems have demonstrated apoptosis in the absence of the ER and nuclei, obviating
the need for localization of bcl-2 family proteins to these organelle membranes (Jacobson,
et al., 1994; Liu, et al., 1996). Some other studies have suggested that bcl-2 can prevent
apoptosis by redistributing calcineurin from the cytosol to intracellular membranes. This
prevents calcineurin from dephosphorylating BAD or interacting with substrates such as
phosphorylated NF-AT (Reed, 1997; Shibasaki, et al., 1997).
Mitochondrial PT
(cytochrome c /AIF release)
caspases
p53
I ,
I
I
I
I
NF-AT
Figure 2-3: A map of bcl-2's (and bcl-XL's) interactions with other proteins which help
explain the mechanisms by which it exerts it antiapoptotic effects. See text for details of
specific protective effects. Adapted from Reed, 1997.
However, evidence pointing toward the need for mitochondrial localization is
growing steadily (Bossy-Wetzel, et al., 1998; Kroemer, 1997; Kroemer, et al., 1997;
Newmeyer, et al., 1994; Petit, et al., 1996; Susin, et al., 1996; Vander Heiden, et al., 1997;
Yang, et al., 1997).
Preventing members of the bcl-2 family from localizing to the
mitochondria severely reduced their ability to control apoptosis.
In contrast, once the
ability to localize specifically to the mitochondria was added back to the proteins, they
regained their function (Zha, et al., 1996).
Bcl-2 and Bcl-XL are thought to regulate
apoptosis by maintaining mitochondrial membrane potential and preventing the release of
cytochrome c and apoptosis-inducing-factor (AIF) (Kluck, et al., 1997; Susin, et al., 1996;
Yang, et al., 1997; Zamzami, et al., 1996).
The release of these agents from the
mitochondria activates caspases, which then push the cell further toward the degradation
phase of apoptosis (see below). In addition, it has recently been demonstrated that the
mitochondrial FOF 1-ATPase proton pump is necessary for the apoptotic activity of Bax
(Matsuyama, et al., 1998), demonstrating that Bax activity is also mitochondria
dependent.
2.2.2.2.2 Dimerization and levels of expression
The response of a cell to an apoptotic stimulus is determined by the competitive
dimerization between death-agonisitic and antagonistic members of the bcl-2 family.
These dimerizations, in turn, are affected by the ratios of these proteins to other proteins
that they interact with (Yang and Korsmeyer, 1996). Mutant forms of bcl-XL, which are
unable to bind to Bax or Bak retain substantial portions of their death repressor activity
indicating that family members (at least the death-antagonistic ones) can function without
dimerization (Cheng, et al., 1996). In contrast, binding (formation of heteromers) by the
death agonistic members of the bcl-2 family seems essential for their apoptotic function
(Kroemer, 1997). While it is clear that a dynamic equilibrium between the various homoand heterodimers is instrumental in regulating apoptosis, it is not very clear which of
these dimers is the true regulator of apoptosis.
2.2.2.2.3 Post-translational modification
Post-translational modification of proteins offers another level of control for apoptosis
induction. Serine phosphorylation of bcl-2 related proteins effectively destroys activity
of both death agonistic and antagonistic family members (Kroemer, 1997).
The
disruption of family members often disrupts their ability to bind to other members or the
ability to localize to desired organelles.
The stimulus for this phosphorylation is
sometimes external as with chemotherapeutic agents that act on microtubules.
In
addition, some bcl-2 family proteins can themselves induce phosphorylation of other
family members. For instance, bcl-2 and bcl-XL regulate the phosphorylation of Bad, a
death agonist with no transmembrane domain, via Raf-l kinase (Zha, et al., 1996,
Kroemer, 1997). Once phosphorylated, Bad loses its ability to bind to bcl-2 and hence
its ability to induce apoptosis.
Growth factor receptors and kinase cascades can also
influence phosphorylation of bcl-2 family members (Kroemer, 1997). For instance, in the
presence of growth factors Bad is phosphorylated and bound to the protein 14-3-3, but
once growth factors are removed Bad is de-phosphorylated and is free to bind to Bcl-xL
and promote apoptosis (Lincz, 1998; Zha, et al., 1996).
Bcl-2 and Bcl-XL may also use their mitochondrial localization to interact with and
sequester other apoptosis regulating proteins (e.g., Raf-1, BAG-1 etc.) with which they
interact. This provides yet another level of apoptosis control (Kroemer, 1997; Reed,
1997). Bcl-2 has also been reported to control the translocation of p53 from the cytosol
to the nucleus by binding to a p53 binding protein, p53-BP2 (Naumovski and Cleary,
1996). Interestingly, the death agonistic members of the family are not able to bind with
any of the proteins that the death-antagonistic proteins bind to.
2.2.2.2.4 Pore formation
Since many of the bcl-2 family of proteins are localized to the mitochondria and regulate
mitochondrial membrane potential and release of apoptotic agents from the mitochondria,
it has been hypothesized that bcl-2 proteins may have pore-forming ability. Indeed both
bcl-2 and bcl-XL contain domains which are similar to the pore-forming domains of
bacterial toxins such as colicins and diptheria toxin (Muchmore, et al., 1996).
It is
suggested that differences in charge and structure (due to formation of homo- or heterodimers, or due to binding with other proteins) could determine ion selectivity and
conductance through the pores. Vander Heiden, et al., 1997 suggest that bcl-XL's poreforming ability allows it to regulate volume homeostasis in the mitochondria, and may
explain how cytochrome c release can occur without loss in the mitochondrial membrane
potential. In addition to mitochondrial pore formation, bcl-2's pore-forming ability and
nuclear localization could allow it to catch proteins as they cross over into the nuclear
envelope, thus preventing nuclease activity or transmission of the apoptotic signal to the
nucleus (McConkey, 1996; Reed, 1997).
2.2.3 Cysteine Proteases or Caspases
Though members of the bcl-2 family are important in controlling apoptosis, they are
thought to lie fairly upstream in the apoptotic execution pathway.
Another family of
proteins, which according to some is absolutely essential in apoptosis (Salvesen and
Dixit, 1997), is the cysteine protease family of proteins. A cell is thought to commit to
death at a point at which the rate of cellular damage outstrips the ability of a cell to repair
itself.
Cysteine proteases often affect both these parameters determining a cell's
commitment to death (Martin and Green, 1995).
The cysteine protease family is so called due to the presence of a cysteine residue
in the active site(s) of these proteins. In addition, this family displays a highly conserved
QACXG pentapeptide active site motif. The fourth residue, 'X', can be R, Q or G.
These proteins are also characterized by the absolute requirement of an aspartate residue
in the P1 position of their substrates (the P1 position residue is the last residue on the Nterminal side of the cleaved part of the protein, hence substrate cleavage occurs right after
the P1 position). Another hallmark of this family of enzymes is that they are all initially
produced as inactive zymogens, which are cleaved (often in an autocatalytic fashion) in
the presence of apoptotic stimuli to form the active enzyme.
The family of cysteine
proteases already has at least 10 members, and to reduce the confusion due to multiple
names for the same protease (due to simultaneous discovery of the protein by several
researchers), a trivial name, caspase, was suggested for all family members. The 'c' in
caspase refers to the cysteine protease part and the 'aspase' refers to the ability of these
proteases to cleave after an aspartate residue in their substrates. The members are then
assigned numbers in the order of their discovery.
The first member of the family to be discovered was the interleukin-l-13converting-enzyme or ICE, which is now called caspase-1 (Miura, et al., 1993). Although
not now considered a central piece of the apoptotic machinery, a lot of the initial work on
the structure, function and mechanism of action of caspases was carried out with this
protein. All caspases are now classified as belonging to a family headed by one of the
first three caspases discovered, based on sequence homology and specificity of substrates
(Alnemri, et al., 1996). The caspase-1 (or ICE) subfamily consists of caspases-1, -4 and 5. Caspases-2 and -9 comprise the caspase-2 (or ICH-1/Nedd2) family, while the caspase
3 (or ced-3/CPP32) subfamily consists of caspases-3, -6, -7, -8, and -10. While caspase-1
and possibly caspases-4 and -5 are primarily involved in procytokine activation,
caspases-2, -3, -6, -7, -8, -9 and -10 are considered to promote apoptosis.
The caspase family of proteins has been implicated in apoptosis in several cell
lines and human diseases (Nicholson, 1996; Takahashi and Earnshaw, 1996).
The
structure and the specificity of these proteases have provided several clues as to their
mechanism of action. In addition, it is being becoming clear that caspases comprise a
cascade with upstream members of the family activating downstream members. Thus the
caspase family proteins serve both as signaling molecules and as effector molecules. The
following section discusses some of the proposed mechanisms by which caspases control
the process of apoptosis and their interaction with other families of proteins discussed
above.
2.2.3.1 Structureand activationof caspases
Most caspases are produced as inactive zymogens (or proenzymes) that are cleaved to
the active form. These inactive zymogens consist of an N-terminal peptide prodomain, a
large and a small subunit, and sometimes a linker region between the two subunits.
Crystal structures of caspases-1 and -3 have shown that the active enzyme is a
heterotetramer, consisting of two large and two small subunits (Thomberry
and
Molineaux, 1995; Wilson, et al., 1994). It is believed that the heterotetramer structure of
the active enzyme is maintained in all other caspases. The active site(s) (there are usually
at least two active sites per tetramer) span both subunits, making both of them necessary
for enzyme activity. The cleavage sites in the inactive proenzyme are aspartate sites that
can usually be cleaved by the active form of the enzyme. This supports the hypothesis
that these enzymes can be activated via autocatalysis. Activation by autocatalysis thus
provides a level of control to the cell in regulating apoptosis through proteolysis via
caspases.
The structure of active caspase-1 is often considered a model for other caspases
(see Cohen, 1997 for a review). In caspase-1, Cys-285 and His-237 form a catalytic dyad
in the active site. The conserved QACRG pentapeptide active site is in the larger p20
subunit. However, the Asp pocket (for docking the substrate Asp cleavage site) is
formed by Arg-179, Gln-283, Arg-341 and Ser-347, of which only the first two are in the
p20 subunit. The two Arg residues form hydrogen bonds with the P1 Asp residue of the
substrate, and the presence of these Arg residues is necessary for catalytic activity.
In
contrast, the P2 and P3 residues in the substrate are solvent exposed, explaining why
broad substitution in the P2 position is tolerated in substrates (Thornberry and
Molineaux, 1995). The side chain of the P4 tyrosine binds in a hydrophobic channel.
Different residues in the P2, P3 and P4 positions are required in substrates for different
caspases to ensure that they can bind to the active site and be processed. The structure
of the active site thus determines the specificity of the caspase for its substrates.
The prodomain in the inactive zymogen might provide an additional level of
control and specificity in caspase-regulated apoptosis.
Large prodomains function as
signal integrators that bind adapter molecules involved in the transduction of the
apoptotic signal. The adapter molecules are in turn associated with various cell surface
receptors. For instance, the prodomain of caspase-8 binds to the corresponding motif in
the adapter molecule FADD (Fas-associated-death-domain) allowing for its recruitment to
the CD-95 death receptor signaling complex (Chinnaiyan, et al., 1995; Fraser and Evan,
1996). Figure 2-4 provides an example of some receptor, adapter, protease cascades
pertinent to caspase activity.
Large prodomains often contain a global homophilic
interaction domain referred to as CARD (caspase recruitment domain).
Significantly,
caspase-2 is recruited to the TNFR-1 signaling complex through an interaction involving
the respective CARD domains within the adapter molecule RAIDD and the prodomain of
caspase-2 (Duan and Dixit, 1997). CARD domain sequences are highly specific, ensuring
that only certain caspases are recruited and activated by particular stimuli/receptors. For
instance, only caspase-9, but not other large prodomain caspases, bound to the CARD
region of the apoptotic protease activating factor-i (Apaf-1, a mammalian ced-4
homologue) by virtue of a CARD sequence in its prodomain (Pan, et al., 1998).
2.2.3.2 Targetsubstrates of executionercaspases
The executioner caspases cleave several key proteins within the cell making the
commitment to death an irreversible commitment.
A few of these proteins, their
significance and the primary caspases responsible for their cleavage are shown in Table 2II (also see Cohen, 1997 for a review).
2.2.3.3 Inhibitorsof caspases
There are several genetic and chemical (peptide) inhibitors of caspases.
Amongst the
genetic inhibitors crmA effectively inhibits caspases-1, -4, -6, and -8 but is a poor
inhibitor of caspases-2, -3, -7 and -10. The anti-apoptotic members of the bcl-2 family
are able to prevent caspase activation by preventing release of cytochrome c from the
mitochondria (see below). However, once caspases (especially downstream caspases) are
activated, bcl-2 family proteins are unable to inhibit caspases.
It is interesting to note
that many caspase-related apoptotic pathways that are resistant to crmA are sensitive to
inhibition by bcl-2 family proteins and vice versa (Cohen, 1997).
The baculovirus
Autographa californica p35 gene-product is a surprisingly wide inhibitor of all known
caspases (Bump, et al., 1995; Hershberger, et al., 1994; Rabizadeh, et al., 1993; Xue and
Horvitz, 1995). Another family of baculovirus proteins and their human homologues,
TNF
Fas
TNF
Ligand
RIP
-
AIDD
- FA D D
Death domain
FADD-like
rodomain
prodomain
DED
Pro-caspase-8
..
TRADD
Pro-caspase-2
Figure 2-4: Examples of receptor, adapter and protease complexes in regulation of caspase
activity. In this illustration, Fas ligand and TNF- (tumor necrosis factor) receptor are
the receptor molecules. FADD (Fas-associated-death-domain protein), TRADD (TNFreceptor-associated-death-domain protein), RIP (receptor-interacting-protein), RAIDD
(RIP-associated ICH-1/CED-3-homologous protein) are the adapter molecules. Procaspase-2 and pro-caspase-8 are the end proteases (or caspases) which are activated
through this signaling pathway. Adapted from Cohen, 1997.
Table 2-II: Caspases and their target substrates.
Target Protein
Biological Significance
Cleaving Caspases
Poly(ADP-
Thought to be responsible for DNA
Caspases-3 and -7
ribose)
repair
polymerase
regulation (Lazebnik, et al., 1994).
(PARP)
However, PARP-null
and
activity
endonuclease
mice develop
normally (Cohen, 1997).
DNA-PK
responsible for DNA double-strand-
caspase-3
break repair
Lamins (A, B responsible for structural integrity of caspase-6 cleaves lamin A
(Orth, et al., 1996). Lamin B
the nucleus (McKeon, 1991)
and C)
is
cleaved very
early
in
apoptosis and maybe cleaved
by a more upstream caspase
(Mandal, et al., 1996)
U1-70 kD
essential for splicing
of precursor
caspases-3 and -7
mRNA. Cleavage blocks cellular repair
pathways dependent on new mRNA
synthesis
Fodrin
responsible for cell membrane integrity. possibly caspase-6
Cleavage leads to plasma membrane
blebbing
Protein kinase the 5-fragment of PKC is thought to be PKC-8 is a rare protein
C (PKC) 8 and responsible for chromatin condensation cleaved by caspase-3 but not
Retinoblastoma
and nuclear fragmentation.
protein (Rb)
important
mediator
of
mediation and progression.
Rb is an caspase-7.
cell
cycle
Rb is cleaved by caspase-3
called the IAPs (inhibitors of apoptosis proteins), have been reported to broadly inhibit
caspase activity (Clem, et al., 1996; Orth and Dixit, 1997). The peptide z-VAD.fmk
(benzylocarbonyl-Val-Ala-Asp(OMe).fluoromethyl ketone) is a powerful and irreversible
inhibitor of all tested caspases (Fraser and Evan, 1996; Xiang, et al., 1996).
2.2.3.4 The caspase cascade
Caspases implicated in apoptosis can be divided into initiators and executioners of
apoptosis (Salvesen and Dixit, 1997).
Caspases belonging to the caspase-1 family are
sometimes thought to play an amplifying role in the signaling process from initiator to
executioner caspases, but evidence for this hypothesis is scant (Fraser and Evan, 1996).
The initiator caspases (caspases-8, -10 and possibly caspase-2) are located at the apex of
the caspase apoptotic cascade, and are distinguished by their ability to directly cleave,
and thus activate, zymogens belonging to all other caspases (Cohen, 1997). In contrast,
the executioner caspases can only cleave a limited number of the zymogens of other
family members.
The ordering of some of the executioners of the pathway is
controversial and may differ in different cell lines. Cohen, 1997 suggests that since PARP
is cleaved before lamins in most cell lines undergoing apoptosis, caspases responsible for
cleaving PARP (caspases-3 and -7) may lie upstream of lamin cleaving caspase-6.
The best example of a pure caspase cascade can be found in Fas induced death.
Here the Fas receptor binds pro-caspase-8 via FADD and activates it. Caspase-8 can
then catalytically activate executioner caspases (-9, -3, -6, and -7), either directly or
through caspase-10. The executioner caspases then cleave their substrates to initiate the
degradation process. This caspase cascade does not require the presence of mitochondria
or the release of cytochrome c from the mitochondria (Reed, 1997; Wallach, et al., 1997).
This may explain why bcl-2 is not able to protect against Fas induced apoptosis in many
cell lines. TNFR mediated death may similarly use caspase-2 as an initiator caspase.
Physiologically, proteases such as the cytotoxic cell protease granzyme B can also
directly activate the caspase cascade by cleaving caspase zymogens (Pan, et al., 1998).
Even more remarkable is the recent demonstration that proteases without aspartate
residue specificity can cleave caspase zymogens in the linker regions (regions between
subunits), possibly allowing lysosomes or viruses to engage the apoptotic apparatus
under pathological conditions (Zhou and Salvesen, 1997).
Most of the other caspase cascades however display an absolute requirement for
the presence of mitochondria (Kroemer, 1997). Here the primary event in the activation
of the caspase pathway is the release of cytochrome c from the mitochondria as a result
of as yet incompletely defined upstream events. The release of cytochrome c need not
involve upstream caspases since a universal caspase inhibitor failed to block the release of
mitochondrial cytochrome c (Bossy-Wetzel, et al., 1998; Yang, et al., 1997).
Once
cytochrome c (also called Apaf-2; Apaf is an acronym for apoptotic protease activating
factor) has been released from the mitochondria it binds to Apaf-1 (see Li, et al., 1997 for
a review). The Apaf-1-cytochrome c complex is able, only in the presence of dATP, to
bind and cleave the caspase-9 zymogen (see Figure 2-5).
Activated caspase-9 then
cleaves other downstream executioner caspases such as caspase-3 or caspase-7 (Pan, et
al., 1998). These activated caspases can then further act on the mitochondria inducing
permeability transition.
The permeability transition brings about a reduction in
Apoptotic stimuli
cytochrome c
dATP
Im
(Apaf-2)
mitochondria
pro-caspase-9
(Apaf-3)
- DOLDA
IC
DQLDA
DQLD
Activated caspase-9
pro-caspase-3
activated caspase-3
Q
Figure 2-5: The mitochondrial caspase cascade. Release of cytochrome c from the
mitochondria leads to the activation of caspase-9 in the presence of Apaf-1 and dATP.
Caspase-9 subsequently activates other caspase-3 and potentially other downstream
caspases. Adapted from Li, et al., 1997.
mitochondrial membrane potential, which in turn triggers the liberation of apoptosisinducing-factor (AIF) from the mitochondria and the generation of reactive oxygen species
(Backer, et al., 1990; Bossy-Wetzel, et al., 1998; Susin, et al., 1996; Zamzami, et al.,
1996). AIF alone can provoke nuclear apoptosis. In addition to affecting mitochondrial
potential, the cytochrome c induced activation of the executioner cascade leaves them free
to act on the various biologically critical substrates discussed above and seals the cell's
commitment to death.
2.3 Mitochondria - the meeting ground for the key players in apoptosis
The release of cytochrome c from the mitochondria, which is a key event in the activation
of caspases, is controlled by the bcl-2 family of proteins - bcl-2, bcl-XL and Bax. Pan, et
al., 1998 prove that bcl-XL can interact with Apaf-1 (at a binding site different from that
required for caspase-9 binding). The authors hypothesize that this interaction also allows
bcl-XL another mechanism to prevent caspase-mediated apoptosis.
In support of their
theory, they demonstrate that other members ofbcl-2 family, namely Bax and Bak, which
bind to bcl-XL and disrupt its ability to prevent apoptosis, also prevented its ability to
bind Apaf-1.
Bcl-XL may be controlling Apaf-l-caspase-9 mediated death by also
sequestering Apaf-1 to the mitochondrial matrix and removing it from the cytosol,
although this hypothesis remains to be proved. In addition, bcl-2 family members may
also prevent death more directly by maintaining mitochondrial membrane potential and
preventing the release of AIF and other damaging ROS. The mechanism of cytochrome c
release from the mitochondria is not very clear at this point.
Since a drop in
mitochondrial membrane potential is not essential for cytochrome c release, it is
speculated that the release of cytochrome c occurs due to a transitory opening of the
permeability transition pore (Bossy-Wetzel, et al., 1998). In addition, it is also believed
that bcl-2 proteins are able to mediate the release of cytochrome c via their pore forming
domains. Future research will help to elucidate some of the mechanisms of cytochrome c
release and the ways in which bcl-2 family proteins regulate this.
2.4 Are caspases essential for apoptosis?
Given the central role that caspases play in controlling the executioner phase of
apoptosis, there has been a lot of speculation as to whether they are absolutely necessary
for apoptosis. In an interesting paper, Xiang, et al., 1996 proved that Bax was able to
induce to induce apoptosis-like death in the presence of z-VAD.fmk (considered to be a
universal inhibitor of caspases) in response to treatment of cells with Fas. The hallmarks
of apoptosis - chromatin condensation, membrane blebbing and cytoplasmic vacuolation were observed to lesser degrees than in cells not treated with z-VAD.fmk. Salvesen and
Dixit, 1997, however argue that the observed death was not apoptotic death. Although
the apoptotic nature of the caspase-inhibited death process remains in question, the fact
remains that rapid death can be induced in the absence of caspases.
Bax is thought to
mediate this death by altering mitochondrial function, causing a drop in mitochondrial
potential and allowing the release of reactive oxygen species. Figure 2-6 shows a map of
apoptosis in which both caspase-independent and mitochondria independent forms of
death are indicated.
Apoptotic Stimuli
Direct
activators of
downstream
caspases, e.g.
Granzyme B
Bcl-2 family regulation of
mitochondrial potential and
pore-transition
Activation of
downstream caspases
vLoss of mitochondrial potential
Release of ROS
IMIN1ff
activation of degradation proteins
I
-"
1 Death
"
Figure 2-6: A map of apoptosis focusing on caspases and the mitochondrial bcl-2
proteins. The diagrams representing receptor-adapter-protease activation of caspases,
bcl-2 family regulation and activation of downstream caspases are shown in greater detail
in Figure 2-4, Figure 2-3 and Figure 2-5, respectively. The release of AIF (apoptosis
inducing factor) is in a positive feedback loop with activation of downstream caspases.
Loss of mitochondrial potential and release of reactive oxygen species (ROS) by
themselves can lead to death, but there is an active debate in literature as to whether this
death is apoptotic or necrotic (see text).
2.5 Apoptosisin Mammalian Cell Culture
Several authors have demonstrated apoptosis to be primary cause of death in hybridoma
and myeloma cultures (Franek, 1995; Mastrangelo and Betenbaugh, 1998; Mercille and
Massie, 1994; Singh, et al., 1994; Vomastek and Franek, 1993). Initial reports by Singh,
et al., 1994 suggested that Chinese Hamster Ovary (CHO) cells do not undergo apoptosis
in culture.
However, later reports demonstrated extensive apoptosis in CHO cells
(Moore, et al., 1995; Moore, et al., 1997). Several authors have reported the beneficial
effects of bcl-2 expression in delaying death of cells in culture.
Bcl-2 suppresses
apoptosis in hybridoma cultures under conditions of hyperoxia, hypoxia, glutamine
deprivation, glucose deprivation, serum limitation (Simpson, et al., 1997) and insulin
withdrawal (Chung, et al., 1998). It has even been reported to prevent apoptosis related
to hydrodynamic stress (Singh, et al., 1996). Itoh, et al., 1995 report that bcl-2 is able to
protect hybridoma cells at the end of a batch culture from nutrient limitation-induced
death, and is also able to enhance productivity of the cells. However, no explanation was
offered for the higher productivity observed. Singh, et al., 1997 have also demonstrated
that bcl-2 expression reduces specific nutrient consumption, suggesting that bcl-2
expressing cells consume nutrients more efficiently. Mastrangelo, et al., 1996 report that
bcl-2 improves recombinant protein production by inhibiting apoptosis in response to
viral infection.
There have, however, been instances where bcl-2 has failed to prevent apoptosis
in the context of mammalian cell culture. Murray, et al., 1996 report that bcl-2 expression
failed to prevent the onset of death in an NSO myeloma batch culture. A high indigenous
Bax expression level (negating the effects of bcl-2) or the detection of indigenous bcl-XL
(leading to an already high base protective level) were offered as possible explanations for
these observations. Terada, et al., 1997 also report that bcl-2 is unable to suppress
apoptosis induced by the complete absence of serum in hybridoma cultures.
An
interesting paper by Suzuki, et al., 1997 reports the synergistic effect of expressing both
Bag-i and bcl-2 in preventing apoptosis in hybridoma cell lines.
In addition to genetic approaches to apoptosis reduction, nutrient and growth
factor supplementation approaches have also been suggested (Chung, et al., 1997; Franek,
1995; Franek and Chladkova-Sramkova, 1995; Franek and Dolnikova, 1991). However,
the apoptotic pathways involved in cells undergoing apoptosis in mammalian cell cultures
remains to be investigated, and could provide more clues to extending the life of
mammalian cell cultures.
3. Materials and Methods
3.1 Cell Culture
3.1.1 Cell Line
A Chinese hamster ovary cell line producing recombinant human gamma interferon (yCHO) was obtained from Dr. Walter Fiers at the University of Ghent, Belgium (Scahill, et
al., 1983). The y-CHO cell line was created from a dihydrofolate reductase deficient
(DHFR-) CHO cell line by co-transfecting the cells with genes for DHFR and human
gamma interferon (IFN-y). DHFR- cells require added ribonucleosides to survive, while
transformed cells are able to produce their own ribonucleosides. The cells were selected
for growth in 0.25 gM methotrexate, which is a competitive inhibitor of the DHFR
enzyme. Methotrexate selection leads to amplification of the DHFR gene and adjacent
DNA (Kaufman and Sharp, 1982), which increases the copy number of genes cotransfected with DHFR (IFN-y in our case). To maintain selection pressure, the cells
were grown in ribonucleoside-free medium with 0.25 gM methotrexate.
3.1.2 Culture Medium
The cells are routinely grown in suspension in serum-free RPMI medium (Sigma
Chemicals) with 2.5 g/L Primatone RL (Quest International), 5 mg/L each of insulin and
transferrin (Sigma). Other components added to RPMI to permit serum-free culture
include 0.4 g/L 2-hydroxypropyl-(p-cyclodextrin, 1 g/L (0.1%) Pluronic F-68, 5 mg/L
insulin, 5 mg/L transferrin, 1 mM sodium pyruvate, 1 jtM putrescine, 11 mg/L choline
chloride, 100 jgM ethanolamine, 1.5 gM linoleic acid, 0.25 jtM methotrexate, 10,000
units/L penicillin, 10 mg/L streptomycin, 6.3 mg/L EDTA and trace minerals (10 nM
sodium selenite, 1 nM manganese sulfate, 10 nM molybdic acid, 10 nM ammonium
metavanadate, 10 nM cupric sulfate, 3 gM zinc sulfate and 5 gM ferric citrate).
3.1.3 Culture Maintenance
Cultures were routinely maintained in shake flasks agitated at 70 rpm on an orbital shaker
(Bellco, Vineland, NJ) placed in a 370 C incubator with a 5-10% carbon dioxide overlay.
Every 2-4 days, cells were resuspended in fresh medium at 2x10 5 cells/mL. Frozen stocks
were prepared from cells with viabilities 2 95% by centrifuging the cells at 800 rpm for 7
minutes and resuspending the cells at 7x10 6 cells/mL in freezing medium (7.5% DMSO,
46.3% fresh medium and 46.3% conditioned medium). Vials containing 1 mL of the cell
suspension were placed into a cryogenic 1 C/min freezing container (Cole-Parmer, Niles,
IL) which was placed into a -70 OC freezer over-night. Cells were later transferred to a
liquid nitrogen cell bank for long term storage. New cultures were started by quickly
thawing the frozen cells and placing them directly into 20 mL of fresh medium in a shake
flask. The following day the cells were resuspended at 2x10 5 cells/mL in fresh medium to
remove the DMSO.
Cells were re-suspended every 2-4 days until the viability was
greater than 95%, at which time experiments were initiated. For all experiments mid-log
phase cells were centrifuged and resuspended in fresh medium at an initial density of
4x10 5 cells/mL (unless otherwise noted). All experiments were carried out in duplicate.
3.1.4 Caspase-Inhibiting Peptide Experiments
z-VAD.fmk (N-benzylocarbonyl-Val-Ala-Asp-fluoro-methyl-ketone,
catalog # FK-009,
Enzyme Systems Products) was solubilized in Dimethyl Sulfoxide to yield a 10 mM
stock.
z-VAD.fmk binds irreversibly to several caspases within the cell.
A final
concentration of 60 gLM was used for experiments.
3.1.5 Fed Batch Culture
Fed batch cultures were performed in 250 mL shake flasks (with a maximum of 80 mL of
cell suspension per flask) using the stoichiometric feeding technique developed at MIT
(Xie, et al., 1997; Xie and Wang, 1994a; Xie and Wang, 1994b). Agitation was set at 70
rpm, and the flasks were placed in a 370 C incubator with a 5% carbon dioxide
atmosphere. Cultures were inoculated at 5x10 5 cells/mL in initial medium containing 3
mM glucose and 0.5 mM glutamine (unless otherwise noted). The initial medium was
RPMI-SFM with reduced glucose and glutamine and enhanced buffer capacity.
The
buffer capacity of RPMI-1640 was enhanced by adding 2.09 g/L of MOPS (Sigma) and
6.50 g/L HEPES sodium salt (Sigma). To maintain proper osmolarity when the extra
buffers were added, the sodium chloride concentration was reduced from 6 g/L to 2 g/L.
Feeding was performed with a stoichiometrically designed supplemental medium as
described in Xie, et al., 1997. Cell counts were performed every 24 hours and feeding was
performed every 12 hours. Feeding volume was a function of both cell density and the
expected growth rate of cells. Samples were withdrawn periodically for cell counts and to
collect supernatant and cells for later analysis.
3.1.5.1 Declumping Protocolfor cells in fed-batch culture
CHO cells in fed-batch cultures tended to clump and stick to the walls of
the shake flask on reaching high cell densities. The presence of clumps in
the cell sample withdrawn for cell number and viability determination
introduces high variabilities in these numbers. The following protocol was
developed to solve these problems faced due to clumping and wall growth
of cells.
*
Cells are gently scraped off the wall into the medium with a cell
scraper (Falcon 3089)
*
The cell suspension is pipetted up and down repeatedly with a 5 mL
pipette to reduce the size of the clumps
*
After ensuring that the cells and clumps are uniformly suspended, two
0.6 mL samples are withdrawn and subsequently mixed
*
1.2 mL of 1X trypsin-EDTA (GIBCO-BRL) is added to sample
*
The mixture is then gently pipetted up and down for 2 minutes (visible
clumps should disappear fairly quickly)
*
1.2 mL of 1g/L of soybean trypsin inhibitor (GIBCO-BRL) is added to
the suspension followed by brief up and down pipetting
*
Cells are counted as described below soon after the trypsinization step
is completed.
3.1.6 Insulin binding studies
Insulin binding was measured using radioactive 1125-insulin (Dupont NEN, catalog #
NEX104). Cells from day seven of insulin-free fed-batch culture were trypsinized (to
remove clumps) and counted using a Coulter counter. Non-trypsinized cells were then
resuspended in the culture medium at 1x10 6 cells/mL. Radioactive insulin was added to
the cells at a concentration of 10 ng/mL (0.96 gCi/mL).
In this experiment, it was
important to measure only the radioactivity due to insulin that is bound specifically to
the insulin family receptors on the cell surface, and not the non-specifically bound insulin.
To measure the extent of non-specific binding, cold (non-radioactive) insulin was added at
a concentration of 10 ptg/mL, to half the population of cells prepared above.
This
saturating quantity of insulin effectively dislodges radioactive insulin bound specifically
to insulin receptors at equilibrium, but does not dislodge non-specifically bound
radioactive insulin.
Measuring the radioactivity of this population thus permits a
measurement of the extent of non-specific binding.
Binding measurements were
conducted using the Millipore MultiScreen Assay System (catalog # MADP N6550).
The cells were then plated directly into a 96-well plate with a 0.65 micron filter base at
1.5x1 0' plates per well in, quadruplicate. As a control, to measure the extent of reduction
in insulin binding, cells taken from day one of an insulin-deprived fed-batch culture of
CHO cells was similarly treated.
The cells were agitated in the 96 well plates for two hours at 40 C. They were then
washed with WHIPS (20 mM HEPES, pH 7.4, 130 mM sodium chloride, 5 mM
potassium chloride, 0.5 mM magnesium chloride, 1 mg/mL polyvinylpyrolidone) four
times, using a vacuum manifold to suck the washing medium through the filter at the
bottom of each well.
The washing was performed to remove all unbound insulin
(radioactive and non-radioactive) from the cells. The filters were dried and the associated
radioactivity was measured with a Packard5000 series gamma counter (Packard
Instruments).
3.2 AnalyticalMethods
3.2.1 Cell Number and Viability
Cell number and viability were determined using a hemacytometer and microscope after
staining the sample with trypan blue. Viable cells exclude the dye, while non-viable cells
have lost membrane integrity and are stained blue. Total cell number was also determined
using a Coulter electronic particle counter (Coulter Electronics, Hialeah, FL) after diluting
samples in an isotonic saline solution. Extent of apoptosis was routinely determined by
the acridine orange / ethidium bromide (AO/EB) assay (McGahon, et al., 1995; Mercille
and Massie, 1994) using a hemacytometer (see details below). In each of the assays at
least 200 cells were counted for each sample.
3.2.1.1 Acridine Orange - Ethidium Bromide Apoptosis Assay
AO and EB are both DNA intercalating dyes which fluoresce under u.v.
light. AO enters all cells regardless of their membrane integrity. EB can
only enter cells which have lost their membrane integrity. Using a bluegreen filter and u.v. light, AO stained cells appear green and EB stained
cells orange. EB is a stronger dye than AO and overwhelms AO when
both dyes are present in a cell. Thus cells which have lost their membrane
integrity are stained orange.
Materials
Stock Solutions: 0.5 g/L Acridine Orange (Sigma catalog # A-6014) and 0.5
g/L of Ethidium Bromide (Sigma catalog # E-8751), prepared in water.
Stock solutions should be protected from light (storage in containers
wrapped in aluminum foil is recommended).
Working Dye Mix: 100 pgg/mL of AO and 100 pg/mL of EB in 1X
phosphate buffered saline.
Method
Working dye mix is added to the cell suspension (5x10 5 - 2x10 6 cells/mL)
in the ratio of 1:10-1:30 and mixed well (McGahon, et al., 1995; Mercille
and Massie, 1994). For higher cell concentrations a larger proportion of
dye may be used. Dye uptake is instantaneous. The mixture can then be
counted using a slide or a hemacytometer and a magnification of 20X100X.
There are 4 kinds of cells that may be observed (Mercille and Massie,
1994):
1)
Viable Cells: Cells are stained bright green with non-fragmented
nuclei. Usually the nucleus is very large and the whole cell seems to be
stained green.
2)
Early Apoptotic Cells: Cells are stained green with several bright
green dots or fragments, which correspond to condensed and fragmented
chromatin.
Condensation and fragmentation of chromatin indicates that
the cell is already committed to apoptosis but has not yet lost its
membrane integrity.
3)
Late Apoptotic Cells: Cells are stained orange with several bright
orange dots or fragments, which correspond to condensed and fragmented
chromatin. The orange color and the presence of fragmented chromatin
indicates that the apoptotic cells have lost their membrane integrity and are
in a very late stage of apoptosis.
4)
Necrotic Cells:
Cells are stained orange (indicating loss of
membrane integrity) with randomly degraded chromatin (no condensation
of chromatin is observed). Usually the entire cell will be stained orange.
Notes
1)
AO and EB are both highly mutagenic and hence necessary
precautions should be taken while handling them.
2)
Some cells will appear to be stained both green and orange. In most
cases, these cells will have two distinct bands of chromatin.
These cells
are undergoing mitosis and hence these should be counted as viable. A
good rule of thumb is to count cells as apoptotic only if there are three or
more fragments of chromatin.
3)
If cells have been dead for a long period of time, they may have
begun to lose their chromatin material.
These cells were considered
apoptotic, since counts performed on the same culture (towards the end of
the culture) have indicated that the increase in chromatin-free cells is
proportional to the decrease in late-apoptotic cells.
It is, however,
recommended that this be verified for each cell type used.
4)
It is recommended that the cells not be exposed to the dye too long
prior to actual counting and at least 200 cells be counted to provide a
statistically significant count.
5)
Cells where some orange staining seems to lie atop the green
chromatin material are not dead cells. The faint orange stain is from single
stranded mRNA, which is stained orange by AO.
3.2.2 Dry Cell Weight
Dry cell weight was determined by drying samples in a vacuum drying oven. Sample
volumes of 60-80 mL were collected from the flask and centrifuged at 250 g for 10
minutes, and the supernatant was removed. The cells were washed twice in phosphate
buffered saline and centrifuged. After carefully removing the supernatant, cells were
suspended in purified water and transferred to dried, pre-weighed aluminum weigh boats.
Samples were dried to constant weights in a vacuum drying oven set at 60 0 C.
3.2.3 Sugar and Lactate Assays
Suspension culture samples were centrifuged at 200 g for 8 minutes to remove the cells,
and the supernatant was frozen at -20 0 C until ready for analysis. Three hundred and
forty pl of thawed sample was deproteinated before performing the lactate and sugar
assays by adding 100 pl of 20% rn/v trichloroacetic acid.
Two hundred
1 of
deproteinated sample was then neutralized with 50 l of 25% m/v potassium bicarbonate.
Enzymatic glucose and lactate assays were performed according to the 16-UV and 826UV Sigma assay protocols (Sigma, St. Louis, MO), respectively.
Standards were also
deproteinated according to the above protocol.
3.2.4 Determination of IFN-y Concentration
Gamma interferon (IFN-y) concentrations were measured using a commercially available
ELISA kit (Biosource International, Camarillo, CA). Cells were removed from medium
by centrifuging at 200 g for 10 minutes, and samples were stored at -20 0 C till further
analysis. Just before analysis, samples were thawed at room temperature and diluted to
less than 1 gg/L IFN-y (the maximum standard concentration) in a 10 g/L solution of
bovine serum albumin in phosphate buffered saline (Pierce, Rockford, IL).
It was
important to include BSA in the dilution buffer to avoid loss of IFN-,y due to non-specific
adsorption to tube walls. The assay was performed according to the manufacturer's
protocol.
3.2.5 Determination of insulin concentration
Insulin concentration in the medium was determined using a commercially available
insulin-ELISA kit (American Laboratory Products Company, Ltd. catalog # 10-1113-01).
Cells were removed from medium by centrifuging at 200 g for 10 minutes, and samples
were stored at -20 0 C till further analysis. Just before analysis samples thawed at room
temperature and were diluted to less than 10 gg/L insulin (the maximum standard
concentration) in a 10 g/L solution of bovine serum albumin in phosphate buffered saline
(Pierce, Rockford, IL). It was important to include BSA in the dilution buffer to avoid
loss of insulin due to non-specific adsorption to tube walls. The assay was performed
according to the manufacturer's protocol.
3.2.6 Western Blot
3.2.6.1 Cell Lysis
Intracellular protein expression of various proteins were determined by analyzing cell
lysates with a western blot. Cells were detergent lysed in a buffer containing 0.5% Triton
X-100, 150 mM NaCl, 10 mM Tris-HCl (pH = 8.0), 100 gM EDTA and 0.025% sodium
azide . Immediately prior to use, protease inhibitors were added to the following final
concentrations: 100 mg/L PMSF, and 10 mg/L aprotinin. Protease inhibitors were added
from 100X concentrated stocks prepared in isopropanol (PMSF) or phosphate buffered
saline (aprotinin). PMSF is extremely unstable in aqueous solutions and should be added
last to the ice-cold lysis buffer, and the buffer should be used within 10 minutes of this
addition. All buffer components were from Sigma. Approximately
100 tL of lysis
buffer was used per 2x10 6 cells, which ensured efficient extraction. Cells were centrifuged
at 200 g for 7 minutes and the supernatant was removed. Cells were washed once in PBS,
centrifuged and then suspended in ice-cold lysis buffer.
The samples were then
vigorously vortexed, after which they were further aliquoted in 10 L samples and stored
at -70 0 C until western blot analysis.
3.2.6.2 Electrophoresis
Lysates were prepared for western blot analysis by mixing them 1:1 with 2x SDS-PAGE
sample buffer (125 mM Tris-HCl pH=6.8, 20% v/v glycerol, 4% w/v SDS, 0.0025% w/v
bromophenol blue) before heating the samples in boiling water for 5 minutes.
Two
hundred mM DTT is added to the buffer just before usage. DTT is stored as a 1 M
solution in 0.01 M sodium acetate solution (pH 5.2) at -200 C. DTT is very unstable at a
pH different from 5.2.
Twenty pl of each sample was loaded per lane onto 12%
polyacrylamide gels for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE) in the Bio-Rad Mini-Protean II electrophoresis system (Bio-Rad, Hercules, CA).
The gels were run at a constant voltage of 100 V for 1.25 hours.
3.2.6.3 Transfer to nitrocellulosemembrane
Following electrophoresis, the gels were incubated in transfer buffer (25 mM Tris base
and 14.4 g/L glycine in a 20% methanol solution) at 40 C for 15 minutes. Proteins were
then transferred from the gels onto nitrocellulose membranes (pre-soaked in transfer
buffer for an hour) in the Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad) set at
a constant 100 V for 45 minutes.
3.2.6.4 Probingwith antibodiesand development of blot
Following the transfer, the membranes were rinsed in TBST, a tris buffered saline solution
containing the detergent Tween-20 (10 mM Tris-HC1, 150 mM NaCl, 0.05% Tween-20,
pH=8.0). The rinsed membranes were soaked for at least two hours in a blocking buffer
which contained 5% non-fat dry milk in TBST to saturate the nitrocellulose membrane's
non-specific protein binding capacity. The blocked membranes were then rinsed 3x5
minutes in TBST before adding the primary antibody solution (primary antibody
concentrate dissolved in TBST solution containing 5% non-fat milk). After incubation in
the primary antibody solution, the membranes were again washed 3x5 minutes in TBST.
The next step was an incubation in the secondary antibody solution (horseradish
peroxidase conjugated secondary antibody concentrate dissolved
in TBST solution
containing 5% non-fat milk). In preparation for the development step, the membranes
were washed once in TBST and then rinsed 3x5 minutes in TBS (TBST without Tween20), after completion of incubation with the secondary antibody. The membranes were
developed using the Renaissance Enhanced Chemiluminescence kit (DuPont NEN,
Boston, MA), and the results were recorded by exposing the membranes to X-ray film
(X-OMAT AR, Eastman Kodak Co., Rochester, NY) for exposure times ranging from 5
seconds to 2 minutes. Films were developed in a film processor (Eastman Kodak Co.).
The antibodies used during the course of the thesis and the companies from which they
were sourced are indicated in Table 3-I.
3.2.7 Total Protein Assay
Total protein concentrations in cell lysates were determined using the Bio-Rad DC
Protein Assay (Bio-Rad, Hercules, CA). The DC Protein Assay is compatible with ionic
and non-ionic detergents, and therefore can be used with detergent-lysed cells. Bovine
gamma globulin (Bio-Rad) was used as a standard. The standard was prepared in the
same buffer used for cell lysis at the following concentrations: 0.1, 0.3, 0.5, 0.7 and 0.9
mg/mL.
The assay was performed according to the instructions provided by the
manufacturer, and unknown protein concentrations were determined from a linear
standard curve.
3.2.8 Measurement of Caspase Activity
About 2x10 6 cells were lysed using a single detergent lysis buffer (Sambrook, et al., 1989).
Cells from an apoptotic culture were taken for measurement of caspase activity. The
substrate used to detect caspases, N-benzylocarbonyl-Tyr-Val-Ala-Asp.AFC
(z-
YVAD.AFC, catalog # AFC-120, Enzyme Systems Products), has a fluorescent 7amino4trifluromethyl coumarin (AFC) residue.
z-YVAD.AFC is a specific substrate for
caspases within the cell. Cleavage of this substrate after the aspartate residue, releases
Table 3-I: Antibodies used and the companies from which they were sourced
protein
primary antibody
secondary antibody
15131A (Pharmingen)
catalog # 107-035-142 (Jackson
detected
human bcl-2
ImmunoResearch Laboratories)
Poly-(ADP-
catalog # 1 835 238 (Boehringer
catalog # sc-2004 (Santa Cruz
Ribose)-
Mannheim)
Biotechnology)
catalog # sc-6215 (Santa Cruz
catalog # sc-2020 (Santa Cruz
Biotechnology)
Biotechnology)
catalog # sc-6217 (Santa Cruz
catalog # sc-2020 (Santa Cruz
Biotechnology)
Biotechnology)
Polymerase
Lamin A
Lamin B
p53
cyclin E
catalog # sc-6243 (Santa Cruz catalog # sc-2004 (Santa Cruz
Biotechnology)
Biotechnology)
catalog # sc-481 (Santa Cruz
catalog # sc-2004 (Santa Cruz
Biotechnology)
Biotechnology)
the fluorescent AFC fragment of the protein, which in turn leads to fluorescence. To
measure fluorescence, 120 gL of AFC buffer (50 mM HEPES pH 7.5, 1% sucrose, 0.1%
CHAPS), 50 gL Dithiothreitol (100 mM DTT stock), and 10 gL of AFC-120 (2.5 mM
AFC-120 in Dimethyl Sulfoxide (DMSO)) were added to a cuvette which was then
placed in a fluorimeter. The fluorimeter was then zeroed (excitation frequency = 400 nm,
emission frequency = 505 nm). 20 gL of the cell lysate was then added to the cuvette and
the fluorescence was then monitored with time.
Caspase activity was also measured observing cleavage of caspase 3 (CPP32)
substrate, Poly-ADP-Ribose Polymerase (PARP). PARP cleavage was determined by
western blot analysis of cell lysate using a polyclonal antibody against PARP (Boehringer
Mannheim, catalog # 1 835 238). Uncleaved PARP runs as a 113 kDa band, while
cleaved PARP is detected either as a 89 kDa fragment or a 24 kDa fragment.
3.2.9 DNA Ladder Technique for Detection of Apoptosis
The Apoptotic DNA Ladder Kit from Boehringer Mannheim (catalog # 1 835 246) was
used to observe the characteristic ladder obtained due to internucleosomal cleavage of
DNA in apoptosis. The protocol for this assay is well described in the kit. DNA from
lysed cells bind to microscopic glass fibers in the presence of a chaotropic salt. The salts
are then washed away and the DNA is eluted in Tris-EDTA buffer and run on a 2%
agarose gel (Sambrook, et al., 1989).
3.2.10 Cell cycle assay
Cell cycle analysis of cells from CHO cultures was performed with propidium iodide
with the use of a flourescence activated cell sorter (FACS, Becton Dickinson,
FACSCAN). Propidium iodide binds to DNA and fluoresces when ultraviolet light is
shone on cells (excitation = 488 nm, emission = 640 nm). The extent of fluorescence is
dependent on the quantity of DNA in the cell. The assay estimates phase of the cell
cycle by estimating the amount of fluorescence in each cell, and hence the amount of
DNA in the cell.
Sample containing 1x10 6 cells was centrifuged and the supernatant was removed.
Cells were then resuspended in 0.5 mL of Stain Solution (1.5% Poly-(Ethylene)-Glycol4000 (PEG-4000), 1.5% PEG-8000, 50 gg/mL propidium iodide (PI), 180 units/mL
DNAse-free RNAse, 0.1% Triton-X-100, 4mM sodium citrate pH = 7.8; final pH of
solution adjusted to 7.2). The cell suspension is incubated at 370 C for 20 minutes. Onehalf mL of Salt Solution (1.5% Poly-(Ethylene)-Glycol-4000 (PEG-4000), 1.5% PEG8000, 50 gg/mL propidium iodide (PI), 0.1% Triton-X-100, 0.4 M sodium chloride; final
pH = 7.2) is added to the suspension at the end of incubation. The suspension must be
stored in the dark at 40 C for at least one hour prior to FACS assay, and may be stored for
up to four days before the assay is carried out.
Storage beyond four days is not
recommended as the DNA in the samples starts to degrade.
3.3 Transfectionand Cloning of cells
3.3.1 Preparation of plasmids
Human bcl-2 cDNA (graciously donated by Stanley Korsmeyer, University of
Washington) was excised using EcoR1
sites, inserted into pcDNA3.1+
vector
(Stratagene), and placed under the control of a constitutive CMV promoter.
3.3.2 Transfection of Suspension CHO Cells
A protocol was developed for transfecting CHO cells in suspension, rather than transfect
anchorage dependent cells and then adapt them to suspension culture. CHO cells grown
in suspension were transfected (see transfection protocol below) in standard 6-well plates
using Lipofectamine® (GIBCO-BRL).
Lipofectamine® uses polycationic liposomes to
deliver DNA to the cell. Expression of bcl-2 was verified by western blot analysis using a
purified hamster anti-human bcl-2 monoclonal antibody (catalog # 15131A, Pharmingen).
The control cell line used was one transfected with the null pcDNA3.1+ vector. 600
pg/mL G-418 (Genetecin, GIBCO-BRL) was used to maintain selective pressure on the
mixed culture of cells.
3.3.2.1 Transfection protocol
Ensure that there are enough cells at greater than 90% viability, and
growing in the log phase. About 2x10 6 cells per well of a 6 well plate are
required.
PreparationofDNA
* For each transfection, dilute 2gg of DNA (Maxipreped DNA) into
100 L of Optimem 1 (Life Technologies).
*
For each transfection, dilute 5gL of Lipofectamine" into 100p~L of
Optimem 1".
*
The above is most conveniently done in sterile plastic cuvettes,
especially if many transfections are done using the same set of DNA.
Falcon 2054 12x75 mm tubes were used. Doing at least 3 wells per
type of DNA plasmid is recommended.
*
Combine the DNA and Lipofectamine samples and mix gently (by
pipetting up and down). Incubate at room temperature for about 45
minutes to allow the DNA-liposome complexes to form.
Preparationof cells
* At about 15 minutes before the DNA-lipofectamine mixture is ready:
*
Spin down the requisite number of cells for n+2 wells, where n is the
number of wells that you want to transfect.
*
Remove supernatant and wash cells once in PBS.
*
Resuspend cells in Optimem 1" medium at a density of 2.5x10 6
cells/mL.
*
Pipette 0.8 mL of cell suspension into each well of a 6 well plate.
*
Add 0.2 mL of the DNA-lipofectamine mixture to each well and mix
gently (pipette up and down).
*
Incubate at 370 C for about 6 hours, and then add 1 mL of Optimem 1"
to each well.
*
After about 20 hours after the start of transfection, harvest cells from
plates (this might require a short trypsinization step. Trypsinize for
2-4 minutes, and then quench trypsin by using appropriate amount of
trypsin inhibitor).
Post TransfectionProcedure
* Resuspend cells in normal growth medium and let them grow for 24
hours.
*
Spin down half the cells, and resuspend in selection medium. We use
600 gg/mL of G-418 (Geneticin®, GIBCO BRL) in RPMI medium.
*
Split the culture in half again by spinning down the cells, and
resuspending the cells in fresh selection medium.
* Repeat the above step after 2 days, or sooner if the medium starts
turning yellow.
*
Repeat the above step after 3 days, or sooner if the medium starts
turning yellow.
*
After 1st week, count the number of viable cells and resuspend in fresh
selection medium at 0.5x10 6 cells per mL. After this change medium
every 3 days dilute cells by 1/2 for the next week or till the culture is
about 95% viable.
3.3.3 Monitoring Transfection Efficiency
Transfection efficiency was monitored using green fluorescence protein (GFP). Each time
the cells were transfected with the desired plasmid, a set of transfections was also carried
out using a plasmid containing the GFP gene (pEGFP, catalog # 6077-1, Clontech), using
exactly the same transfection conditions as used with the other plasmids. After the last
step of the transfection procedure described above, the cells were assayed for GFP
expression using a fluorescence activated cell sorter (FACS, Becton Dickinson,
FACSCAN).
The fraction of cells in the transfected population that displayed green
fluorescence (excitation = 488 nm; emission maximum = 507 nm.) was defined as the
transfection efficiency, since it was a measure of the fraction of cells transfected which
took up the transfecting plasmid. Since about 20,000 cells were counted to arrive at this
number, the transfection efficiency measure obtained was statistically very significant.
The transfection efficiency obtained with the GFP plasmid was assumed to be indicative
of the transfection efficiency obtained with our desired plasmid, since the transfection
conditions used for both were identical. The main advantage of using GFP over other
techniques, is that it is very rapid and accurate. Cells do not need to be processed at all,
before the assay is carried out.
Using this technique to monitor transfection efficiency we were able to increase
transfection efficiency from under 1% to over 25%, essentially making three changes to
the transfection process (all of which are incorporated in the above protocol):
a) Optimem 1 (GIBCO-BRL), and not RPMI, was used as the medium during
transfection (no insulin, transferrin, antibiotics or primatone was used as it was suspected
that small molecules might interfere with transfection)
b) Cells were not agitated during transfection (24 hours). Agitation caused many
cells to stick to the sides of the 6 well plates and it is possible that these cells did not take
up any of the DNA, leading to lower transfection efficiencies. The cells were found to be
mostly viable after 24 hours in Optimem 1® without agitation, hence oxygen transport to
the cells was not limiting. Some of the cells did tend to stick to the bottom of the plate,
but these could easily be trypsinized the following day.
c) The ratio of DNA and lipofectamine during transfection was altered. The most
optimum ratio amongst the ones tested was 2gg DNA to 5gg lipofectamine.
The higher rates of transfection meant shorter times to obtain mixed cultures.
This time went down from at least 1.5 months to around two weeks.
The monitoring
procedure ensures that transfection can rapidly be optimized for any future cell line for
which transfection is required.
3.3.4 Cloning
Using a fluorescence activated cell sorter (FACS, Becton Dickinson, FACStarPlus), wells
of a 384-well plate (NUNC) were seeded with live cells (dead cells were gated out using a
propidium iodide stain) from a mixed culture of transfected cells, at one cell per well.
Cells were taken from a culture growing in log-phase. Wells were filled with 100gL of
50% conditioned medium (taken from culture where cells were in the exponential growth
phase) and 50% fresh GN-RPMI. Selective pressure was constantly maintained during
the cloning process. In about 1.5 weeks, wells in which cells had grown were harvested
and the cells were transferred to wells of a 24 well-plate. When cells in these wells
reached a concentration of about 1x10 6 cells/mL, cells were then transferred to a 6-well
plate and then to a T-75 flask.
Selective pressure using 600 gg/mL of G-418 was
maintained at all times. A sample was taken from each clonal population to check for
gene expression using western blotting. The whole process, from the time after harvesting
cells from the 384-well plate to the T-75 flask, took about 1.5 weeks.
4. Apoptosis and its Control in Chinese Hamster Ovary
Batch Culture
The first step in controlling death in cell cultures was to determine the mode of death in
serum-free suspension CHO cultures. If the main mode of death in CHO cells under
serum-free culture conditions were necrosis, the only way to reduce death would be to
make environmental changes in the medium. If on the other hand, cells primarily died due
to apoptosis, death could be reduced by both environmental and genetic means. At the
time this study was initiated, there were conflicting reports in literature as to the principal
mode of death in CHO cells (Moore, et al., 1995; Moore, et al., 1997; Singh, et al., 1994;
Singh, et al., 1997). This chapter details the results of several techniques used to ascertain
the mode of death in CHO cultures. Since we found that apoptosis was the major mode
of death in batch cultures, we were further interested elucidating the apoptotic process in
CHO cells.
In particular, we were interested in determining whether de novo gene
synthesis was required to execute the apoptotic process in our cells. In addition, given
the reported predominance of caspases in apoptosis in most mammalian cell lines we
wanted to determine whether they played a role in apoptosis in CHO cells.
As caspase activity was detected in apoptotic cells, this study attempted to
determine whether specifically inhibiting some of the caspases using peptide-caspaseinhibitors could extend the viability of batch cultures. Since the commitment of a cell to
apoptosis also depends on expression of survival proteins, we studied the effect of
constitutive overexpression of the survival protein, bcl-2, in CHO cells grown in
suspension. Bcl-2 has been shown to protect against apoptosis in a large number of cell
lines (Kroemer, 1997; Mastrangelo and Betenbaugh, 1998; Reed, 1994). This chapter
compares the relative effectiveness of bio-chemical (caspase-inhibiting-peptides) and
genetic means (bcl-2 expression) in extending the life of CHO cells in serum-free batch
culture. In addition, the chapter compares the relative effectiveness of clonal populations
and mixed-culture populations in resisting apoptotic stimuli.
The discussion and
conclusions section of the chapter summarizes the results and presents additional insights
that can be drawn from them.
4.1 CHO Cells Die by Apoptosis in Serum-free Batch Culture
The growth kinetics for a typical serum-free batch culture of CHO cells is shown in
Figure 4-1. After a period of exponential growth cells die very rapidly. There is a small
decline in the number of total cells due to the tendency of dead cells to stick to the walls
of the flask.
The mode of death involved in these cells, as measured by the Acridine Orange /
Ethidium Bromide (AO/EB) assay, is shown in Figure 4-2.
The AO/EB assay is
measures both membrane integrity and DNA integrity of cells (see Materials and
Methods chapter for more assay details).
The apoptotic cells included both early
(chromatin condensation but no loss in membrane integrity) and late (chromatin
condensation and membrane integrity loss) stages of apoptosis.
These results indicate
2.OE+6
..j
E
1.5E+6
*
-1.OE+6
TCD
+-VCD
>
5.OE+5
O.OE+O
0
1
2
3
Time
(day)
4
Figure 4-1: Viable (VCD) and total cell densities (TCD) obtained in normal serum-free
batch culture of suspension CHO cells.
100
80
0
60
cc*
% dead
-*% apoptotic
---
>1
-
- -%
40
necrotic
20
0
0
1
2
Time
3
4
(day)
Figure 4-2: Apoptosis in batch cultures of CHO cells. Percentage of dead, apoptotic and
necrotic cells obtained during a normal serum-free batch culture of CHO cells. The
percentage of dead cells is obtained by adding the percentages of apoptotic and necrotic
cells. Almost all death is seen to occur via apoptosis.
that apoptosis is the main mode of death right from the start of the culture growth
process, which is in agreement with the work of Moore, et al., 1995. Furthermore,
apoptosis accounts for more than 95% of the death occurring during cultivation.
In
contrast, necrosis accounts for less than 5% of the final death observed.
Apoptosis was also confirmed by the presence of a DNA ladder, when genomic
DNA extracted from dead CHO cells was run on an agarose gel (Figure 4-3) (see Materials
and Methods chapter for more assay details). The gel-photograph shows that for the
first two days of the culture when the viability, as detected by the AO/EB assay, is high
the genomic DNA runs as a compact band of high molecular weight. By day 3, some
apoptosis has started to occur and this is reflected in the appearance of the beginnings of
a DNA ladder - bands of lower molecular DNA, approximately in multiples of 180 bp.
This ladder is caused by internucleosomal cleavage of DNA by endonucleases which are
activated during the apoptotic process. By day 4 the apoptotic DNA ladder is much
more prominent, and further cleavage of DNA is apparent from the appearance of lower
molecular bands of DNA. We can conclude that most of the death occurs by apoptosis
and not necrosis, due to the presence of an DNA ladder in lane 4 of the gel. Had necrosis
been the primary mode of death, a smear caused by random cleavage of DNA would have
been observed. Lastly, the complete disappearance of the original band of high molecular
weight DNA indicates that most of the cells have perished, confirming the results of the
AO/EB assay.
viability (%) 99 99 88
day
123
4
Figure 4-3: Apoptosis in batch cultures of CHO cells. Agarose gel photograph of
genomic DNA from CHO cells shows presence of a DNA ladder which coincides with
massive apoptosis on day 4.
4.2 Protein synthesis inhibitioncauses rapid, dose-dependent apoptosisin
CHO cells
Since apoptosis is a genetically controlled form of death where the fate of a cell is
determined by death suppressing and death inducing proteins, we were interested in
determining which class of proteins dominated in CHO cells. One way to determine this
was to inhibit protein synthesis using a protein synthesis inhibitor such as cycloheximide,
and then follow the fate of the cells. Two scenarios are possible. In the first scenario, if
death proteins are present in greater abundance and survival proteins have to be
continuously synthesized to keep the cell alive, prevention of protein synthesis would
cause rapid apoptosis. In the second scenario, if death proteins have to be synthesized de
novo to initiate apoptosis, inhibition of protein synthesis will ultimately lead to death by
necrosis.
Different concentrations of cycloheximide, ranging from 0 RM to 60 gM, were
added to cell suspensions taken from a culture in log-phase growth. Viability of these
cultures were monitored at 13, 16 and 19 hours after the addition of cycloheximide. As
shown in Figure 4-4, we observed extensive and rapid apoptosis, and almost no necrosis
in cultures to which cycloheximide was added. For example, by 19 hours after start of
protein synthesis inhibition, more than 85% of the cells in the culture to which 601M
cycloheximide was added had undergone apoptosis. By contrast, there was less than 8%
apoptosis in the control culture (without cycloheximide). The extent of apoptosis at any
90
80
70
----
06
s60
o 50
.
< S40
40
0 uM CHX
20 uM CHX
30 uM CHX
40 uM CHX
-1- 60 uM CHX
-----
0 30
~20
w 10
13
16
19
Time (hour)
Figure 4-4: Extent of apoptosis induced by varying doses of cycloheximide (CHX).
Cycloheximide induces protein synthesis inhibition in cells. Apoptosis accounted for
almost all the death observed.
point in time was also observed to be dose dependent, since higher levels of
cycloheximide at each time point resulted in higher levels of apoptosis. These results
suggest that the first scenario in our hypothesis is more accurate in describing apoptosis
in CHO cells and survival proteins must be continuously synthesized to protect CHO
cells from death.
4.3 Cysteine protease inhibitingpeptides are unable to significantly
enhance viabilityof CHO cells in culture.
Cysteine proteases (caspases) have been found to be downstream effectors of apoptosis
in a large number of cell-lines, including CHO cells. We were interested in determining
whether our CHO cell line expressed some of these proteins in response to stresses
resulting from normal cultivation conditions. Cell lysate taken from an apoptotic culture
of CHO cells was assayed for caspase activity by adding z-YVAD.AFC, a fluorescent
substrate for caspase-l (see Materials and Methods for more details of assay). Activated
caspase-l in the lysate can cleave z-YVAD.AFC, releasing the fluorescent AFC residue.
This event would be detected by the increase in fluorescence. Figure 4-5 indicates that
the fluorescence was detected when z-YVAD.AFC was added to the lysate, indicating the
presence of activated caspase-1 like caspases in the lysate.
Since cysteine protease activity was detected in our cells, we attempted to delay
apoptosis in our cells by adding specific caspase-inhibiting peptides to our culture. The
peptide we chose was a fluoro-methyl-ketone (fink) with a sequence, benzylocarbonylVal-Ala-Asp(OMe).fluoromethyl ketone (z-VAD.fmk). This is an irreversible inhibitor
ICE-like Protease Activity Detected Using YVADFlourescent Substrate
control flourescence = 0
10
C
MSE
8
6.
4
C
2
0
I
0
---
I
I
6
24
30
Time (min)
Figure 4-5: Fluorescence obtained by combining cell-lysate from apoptotic CHO cells
with a fluorescent substrate z-YVAD.AFC, indicating the presence of caspases.
of cysteine proteases (Enari, et al., 1996; Martin, et al., 1996; Takahashi and Earnshaw,
1996; Thornberry and Molineaux, 1995; Wang, et al., 1995; Wang, et al., 1996). The
methyl group ensures that the peptide is cell membrane permeable. We added 60gM zVAD.fmk (solubilized in Dimethyl Sulfoxide (DMSO)) to batch cultures to examine
whether we could delay the rapid death observed at the end of a batch culture. To the
control culture for each of these experiments, we added the same volume of DMSO as
was used to deliver the z-VAD.fmk.
The results are shown in Figure 4-6 and Figure 4-7. z-VAD.fmk provides only a
small improvement in culture viability 96 hours into the culture. Dead cells protected by
z-VAD.fmk showed loss of membrane integrity, and condensed but not fragmented
chromatin, while those in the control cultures showed a typical apoptotic morphology
with condensed and fragmented chromatin. This indicates that z-VAD.fmk is able to
inhibit, directly or indirectly, the activity of the endonucleases responsible for
fragmenting genomic DNA in the apoptotic process.
Figure 4-8 shows that the
concentration of glucose in both cultures reaches close to zero levels by 72 hours. We
suspect that this depletion of glucose from the medium was the major cause of death in
the culture. The results demonstrate that the viabilities and cell densities of the two
cultures are almost identical for a period of 72 hours. Death starts to occur almost 24
hours after glucose depletion was detected. We verified that z-VAD.fmk does get into the
cells and prevent cleavage of substrates of caspases by monitoring the cleavage of Poly(ADP-Ribose)-Polymerase (PARP) at 96 and 120 hours. PARP is a protein involved in
DNA repair and inhibition of nuclease activity, which has been shown to be cleaved by
2.5E+6
2.OE+6
1.5E+6
- -zVAD
o-e-m
1.0E+6
C \
-
o-
5.OE+5
TCD
DMSOTCD
-- zVAD VCD
--DMSO VCD
\
%
U)
0.0E+0
I
0
20
I I
40
I
60
I I
80 100
120
Time (hour)
Figure 4-6: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of CHO
cells: Total cell density (TCD) and viable cell density (VCD) as a function of time.
100
2.5E+6
2.OE+6
--80
1.5E+6
-60
_
>% -•---
40
I\0
S--e--DMSO
1.0E+6
1C
I
_ 5.OE+5
a)
\
0
I-
.0
1 ----
DMSOTCD
zVAD viability'
viability
20
0
I
0.0E+0
zVAD TCD
20
40
60
80 100 120
Time (hour)
Figure 4-7: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of CHO
cells: Total cell density and viability of cultures as a function of time. Viability declines
rapidly after 72 hours.
caspase-3 during the apoptotic process in a number of different cell lines (Lazebnik, et al.,
1994; Liu, et al., 1996; Mandal, et al., 1996). Figure 4-9 shows that there is no cleavage
of PARP in the z-VAD.fmk protected cultures, while a clear cleavage of PARP is seen in
the cultures not protected by z-VAD.fmk. The cleavage of PARP in the unprotected
culture is further proof of the presence of caspase (caspase-3) activity in apoptotic CHO
cells. The presence of intact PARP in cultures to which z-VAD.fmk was added suggests
that z-VAD.fmk was able to effectively block caspase activity in these cells. Thus the
meager protection in cultures supplemented with z-VAD.fmk cannot be attributed to the
inability of z-VAD.fmk to penetrate the cells.
In addition, since the control and z-
VAD.fmk protected cultures behave almost identically till the cultures start to die, death
in the culture with z-VAD.fmk is not due to any toxic side-effect of the peptide. Higher
concentrations of z-VAD.fmk (upto 120 gM) did not yield improved culture viabilities.
Concentrations higher than 120 jgM could not be used since the DMSO used to solubilize
the z-VAD.fmk was toxic to the cells at those levels.
4.4 Bcl-2 extends viabilityin batch cultures of CHO cells
A mixed culture of CHO cells transfected with the human bcl-2 antiapoptotic gene (see
Kroemer, 1997; Reed, 1997 for a recent review) was used to examine the effect of bcl-2
expression on cell death at the end of a batch culture (see Materials and Methods section
for more details on how the mixed culture was obtained). The performance of the bcl-2
expressing culture was compared to that of the pcd-CHO. All cultures were inoculated
100
14
12
E
10
0
C
8
pcd + zVAD
- - pcd + DMSO
S--o-
6 -
E
.:
4
0
0
20
40
60
80
100
120
Time (hour)
Figure 4-8: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of CHO
cells: Medium glucose concentration as a function of time. Medium glucose concentration
dropped to almost zero after 72 hours in both cultures.
101
cell line
pcd +
zVAD
pcd +
DMSO
32
15
viability (%)
pcd +
DMSO
0
pcd +
zVAD
3
uncleaved
PARP (113 kDa)
hhi
cleaved PARP
(84 kDa)
Figure 4-9: Comparison of the protective effect of z-VAD.fmk (a peptide inhibitor of
caspases) on protected (pcd + zVAD) and control (pcd + DMSO) batch cultures of CHO
cells: Poly-ADP-Ribose Polymerase (PARP) cleavage is evident in the control cultures
but not the z-VAD.fmk protected cultures.
102
with 4x105 cells/mL, and the viabilities and cell densities were measured till the viabilities
of both cultures were below 50% (as measured by the AO/EB assay).
The results are
shown in Figure 4-10, Figure 4-11, Figure 4-12 and Figure 4-13. The fluctuation in total
cell densities at the end of the culture results from a tendency of cells to clump and stick
to the walls of the flask. The bcl-2 culture is more than 40% viable 72 hours after the
control culture is completely dead. As can be seen from Figure 4-10, bcl-2 expression
does not affect the growth rate of the culture. Figure 4-12 indicates that the glucose in
both cultures had been completely depleted by 72 hours. We suspect that this glucose
depletion is the main stimulus for death in the control culture.
We checked whether bcl-2 expression is able to block cleavage of caspase-3
substrate PARP. Previous reports have suggested that bcl-2 lies upstream of caspase-3 in
the apoptotic pathway, hence if it is effective in preventing apoptosis, it should be able
to block caspase-3 activity and hence block cleavage of caspase-3 substrates such as
PARP.
Bcl-2 prevents activation of downstream caspases such as caspase-3, by
preventing the release of cytochrome c from the mitochondrial matrix (Kluck, et al.,
1997). Figure 4-13 shows that this is indeed the case. Even after 168 hours the PARP in
the bcl-2 protected cultures remains essentially uncleaved, suggesting that bcl-2 prevents
caspase-3 activation in CHO cells. We suspect that the small band of cleaved PARP
which appears in the bcl-2 lane, may be caused by apoptosis in the mixed population of
cells which do not express bcl-2. The protective effect of bcl-2 is similar to that observed
by other authors using other cell lines (Itoh, et al., 1995; Simpson, et al., 1997; Singh, et
103
al., 1997; Suzuki, et al., 1997; Terada, et al., 1997). Some hypotheses explaining the
surprisingly different results obtained by preventing caspase activation by bcl-2 and zVAD.fmk are presented in the discussion and conclusions section of this chapter.
4.5 Bcl-2 protects better than caspase inhibitingpeptides in response to
growth and survival factor withdrawal
Growth and survival factor withdrawal is known to be a potent inducer of apoptosis in
mammalian cells, especially in the absence of serum (Bottenstein, et al., 1979; Murakami,
1989; Zhou and Hu, 1995). Insulin and transferrin were used as specific growth and
survival factors in our serum-free medium. We studied the protective effect of bcl-2 and
z-VAD.fmk separately and together in protecting CHO cells in batch culture from the
effects of insulin and transferrin withdrawal. The hypothesis underlying these sets of
experiments was that if pathways which were bcl-2-dependent and caspase-independent
and vice versa, existed, then using both bcl-2 expression and z-VAD.fmk addition would
provide better protection than either alone. A volume of DMSO equal to that used to
deliver the z-VAD.fmk to the other cultures was added to the control (pcd-CHO) culture
and the bcl-2-CHO culture. No insulin or transferrin was added to the medium from the
start of the culture. Cells being grown in medium containing insulin and transferrin were
spun down and washed once with phosphate buffered saline (PBS), before transferring to
medium without insulin and transferrin. All cultures were inoculated at 4x10 5 cells/mL
and the experiments were continued till the viability of all the cultures had dropped below
50%.
104
2.OE+6
-
m
,-E
1.5E+6
0
c=
U
- --
0
O
1.OE+6
---pcd TCD
-o---bcl2 VCD
- .-- pcd VCD
I
C)
I
0l
I
I
5.OE+5
0.0E+0
_: i
50
S
I
I
_
II
Time
bcl2 TCD
_ _ _
100
a11d .1
I
150
i
1
i
I
200
(hours)
Figure 4-10: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Total cell density (TCD) and viable cell density (VCD) as a
function of time.
105
2.OE+6
100
.. I
E
,
80
1.5E+6 -
o
60
o
1.OE+6
0,
U)
0
Q,
-40
Ic
0
5.OE+5 -
-20
i
0.OE+0 ()
m
|
1
i
mii
I
i
I
I
50
I
pi
I.
4
i
II
1iW I-
100
I 1
150
I
0
20 0
bcl2 TCD
--u--pcd TCD
-- o--bcl2 viability
--- -- pcd viability
Time (hour)
Figure 4-11: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Total cell density (TCD) and viability of cultures as a function of
time. Viability of the control, but not the bcl-2 protected cell line, declines rapidly after
72 hours.
106
M 10
E
8
O
-b- bcl 2:
--- pcd
0
0
50
100
Time
150
200
(hour)
Figure 4-12: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Medium glucose concentration as a function of time. Medium
glucose concentration dropped to almost zero after 72 hours in both cultures.
107
pcd
cell line
bcl-2 MC
time (hour)
96
168
viability
6%
49%
uncleaved
U
PARP (113 kDa)
i
cleaved PARP
(84 kDa)
Figure 4-13: Protective effect of bcl-2 expression as compared to a control (pcd) in batch
culture of CHO cells: Poly-ADP-Ribose Polymerase (PARP) cleavage is clearly evident
in the control cultures but almost all the PARP is uncleaved in the bcl-2 protected
cultures.
108
The results of the experiment are shown in Figure 4-14, Figure 4-15 and Figure 416. The potency of insulin and transferrin withdrawal in initiating apoptosis can be
judged from the rapid death in the control culture (pcd + DMSO). It is seen that bcl-2
and z-VAD.fmk acting together do indeed provide better protection than either agent
alone, confirming our hypothesis presented above. Also, bcl-2 expression provides much
better protection than adding z-VAD.fmk to the culture. The sudden decrease in viability
for the pcd + z-VAD.fmk culture at 72 hours is probably due to the exhaustion of glucose
in the culture (Figure 4-16). Dead cells in the z-VAD.fmk protected cultures showed loss
of membrane integrity, and condensed but not fragmented chromatin material. The cells in
the control culture showed both loss of membrane integrity coupled with condensed and
fragmented chromatin. Again, this indicates that z-VAD.fmk is able to inhibit, directly or
indirectly, the activity of the endonucleases responsible for fragmenting genomic DNA in
the apoptotic process. The better growth rates of z-VAD.fmk supplemented cultures
(Figure 4-14) cannot be explained at this time.
4.6 Protective effect of bcl-2 is enhanced in clonal batch cultures
The glucose starvation and insulin-transferrin (growth/survival factor) deprivation
experiments performed with bcl-2 mixed cultures were repeated with a clonal population
of bcl-2 cells (see Materials and Methods chapter for more details on clone selection
procedures). As indicated before, the glucose starvation results naturally at the end of a
batch culture when the cells have consumed all the glucose originally present in the
medium.
In growth factor deprivation experiments, no insulin or transferrin was
109
1.6E+6
E
(
1.2E+6
>
8.-E+5
.. - _
--
,--pcd
---...
----
-4.E+5
--- bcl-2 + DMSO
+ DMSO
bcl-2 + z-VAD
pcd + z-VAD
0.0E+0
0
24
48
Time
72
96
120 144
(hour)
Figure 4-14: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells that have been deprived of insulin
and transferrin. To cultures which did not contain z-VAD.fmk we added a volume of
Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver z-VAD.fmk to
the other cultures (also see text). This graph plots total cell density (TCD) as a function
of time.
110
100
80
60
.
-
60t
-
+- bcl-2 + DMSO
40
-
e- - pcd + DMSO
.----bcl-2 + z-VAD'
S
20
20 --
u
pcd + z-VAD
0
0
24
48
Time
72
96
120
144
(hour)
Figure 4-15: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells which have been deprived of
insulin and transferrin. To cultures which did not contain z-VAD.fmk we added a volume
of Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver z-VAD.fmk
to the other cultures (also see text). This graph plots viability of cultures as a function of
time.
111
5.0
4.0 "
E
i
3.0
%I .-
-o---bcl2 + DMSO
- ---
S2.0
r
----
- --
pcd + DMSO
bcI2
b c + zVAD
pcd + zVAD
1.0
0.0
24
I
48
I
72
Time
I
96
120
(hour)
Figure 4-16: Comparison of the protective effects of bcl-2 expression and z-VAD.fmk,
together and separately, in batch cultures of CHO cells which have been deprived of
insulin and transferrin. To cultures which did not contain z-VAD.fmk we added a volume
of Dimethyl Sulfoxide (DMSO) equal to the volume that was used to deliver z-VAD.fmk
to the other cultures (also see text). Medium glucose concentration is plotted as a
function of time.
112
added to the medium right from the start.
The results from the glucose starvation
experiment are indicated in Figure 4-17. It can be seen that bcl-2 expressing clones are
able to survive more than a 100 hours longer than the control cell line (pcd).
The
protective effect of a clonal population is much larger than that observed in a mixed
culture of bcl-2 CHOs where the additional protective effect over the control was on the
order of 48 hours.
The much stronger protective effect of bcl-2 in a clonal population is evident in
the insulin-transferrin withdrawal experiments as well. Bcl-2 protected cultures were able
to maintain viability around 90% 24 hours after the viability of the control culture had
fallen below 15% (Figure 4-18). As expected, the cells did not grow very well in the
absence of growth factors. The variability in cell density was due to cells sticking to the
side of the flask. Again, the protective effect of bcl-2 in a clonal population is enhanced
as compared to that of a mixed culture population.
For comparison, the mixed culture
viability 24 hours after the control culture had substantially perished was about 50%.
In summary, these experiments confirm the protective effect of bcl-2 in a clonal
population and indicate that the protective effect is stronger in a clonal population than in
a mixed culture.
4.7 Caspase-independent death pathways which are blocked by bcl-2
expression appear to exist in CHO cells
z-VAD.fmk has been reported to be a powerful and universal inhibitor of cysteine
proteases. Yet in our cultures z-VAD.fmk failed to prevent death in response to either
113
2.50E+6
a=~zr
100
~~
-f
-*
90
t A
2.00E+6
80
70
1.50E+6
6o0
1.00E+6
50
.
40
.2
- 30
5.00E+5
-20
-
10
BB
0.OOE+0
50
100
'
150
0
1 1
200
250
--.--
bcl2-TCD
pcd-TCD
bcl2-Viability
1----pcd-Viability
Time (hour)
Figure 4-17: Bcl-2 CHO clonal populations are able to maintain their viability for a much
longer period of time as compared to the control, pcd-CHO cells in a batch culture. The
main cause of death is glucose depletion at about 72 hours. The total cell density (TCD)
for both cultures is approximately the same. The variation in total cell density at the end
of the culture is due to cells sticking to the walls of the flask.
114
100
3.5E+5
90
3.OE+5 -
80
E 2.5E+5 -
70
60
o 2.OE+5 -
50
r
1.5E+5 -
40
a
a
1.OE+5
30
a
o
20
5.OE+4 O.OE+O
10
-----
bcl2-TCD
pcd- TCD
--- o--bcl2-Viability
-- e--pcd-Viability
0
.a
20
.
40
60
80
Time (hour)
Figure 4-18: Clonal populations of CHO cells expressing bcl-2 are able to maintain their
viability for much longer (as compared to the control pcd-CHO cell line) in response to
insulin-transferrin deprivation in batch culture. TCD refers to the total cell density of the
cultures.
115
glucose starvation or growth-factor deprivation. To check whether z-VAD.fmk was able
to block caspases effectively, we checked for cleavage of substrates of at least two
different caspases in lysates from apoptotic CHO cells. Substrate cleavage had to be used
as a proxy for caspase activity, since antibodies which could detect native caspases in
CHO cells could not be found. The first substrate, which has already been discussed
above, is PARP which is cleaved by caspase-3 (CPP32/Yama). The second substrate that
we tested for was lamin A, which is cleaved by caspase-6, but not by caspase-3 (Orth, et
al., 1996). By selecting these different substrates we could be reasonably sure that zVAD.fmk was indeed universally blocking caspases in CHO cells.
As previously shown in Figure 4-9 and Figure 4-13, z-VAD.fmk and bcl-2 were
able to block PARP cleavage in apoptotic CHO cells. In contrast, cells not protected by
either z-VAD.fmk or bcl-2 suffered complete cleavage of PARP. Figure 4-19 shows the
results from the lamin A cleavage study. Cells were exposed to insulin and transferrin
withdrawal as a stimulus for apoptosis for this study. The samples shown were taken on
day 3 of the culture. Extensive apoptosis was observed in the control pcd-CHO culture
to which no z-VAD.fmk was added. The pcd-CHO culture to which 60 gM z-VAD.fmk
was added also experienced extensive death. Dead cells lost their membrane integrity and
had condensed but not fragmented chromatin. The bcl-2 protected cell-lines remained
viable. As shown in Figure 4-19, z-VAD.fmk addition (lane 3) is able to completely
block the lamin A cleavage observed in the control (lane 1).
Therefore, z-VAD.fmk
effectively blocks caspase-6 activity (in addition to blocking caspase-3 activity), but is
not able to protect cells from death. No lamin cleavage is observed in bcl-2 protected cell
116
Sample
p-z
b-z
p+z
b+z
Viability (%)
10
84
11
86
lamin A ---
.il
70 kDa
Figure 4-19: Cleavage of caspase-6 substrate lamin A in CHO cells, in response to insulin
and transferrin deprivation. 'p' refers to the control pcd-CHO cell line. 'b' refers to the
bcl-2 expressing CHO cell line. '+z' and '-z' indicate whether caspase inhibitor, zVAD.fmk was or was not added to the culture. The results indicate that addition of zVAD.fink is able to prevent cleavage of lamin A, but is not able to protect cells from
death. Bcl-2 expression prevents both lamin cleavage and death.
117
lines (with or without z-VAD.fmk). The results confirm that blocking caspase activity is
not sufficient to block death in CHO cells. Secondly, since bcl-2 is able to block death
not blocked by addition of caspase inhibitors, bcl-2-dependent but caspase-independent
apoptotic pathways exist in CHO cells.
4.8 Discussionand Conclusions
There have been conflicting reports in literature as to the mode of death in CHO cells
(Moore, et al., 1995; Moore, et al., 1997; Singh, et al., 1994; Singh, et al., 1997). In this
chapter we have shown that CHO cells grown in batch culture die primarily via
apoptosis, a form of genetically encoded death.
In batch culture, this induction of
apoptosis followed the depletion of glucose in the medium. In addition, we have also
demonstrated, by means of protein synthesis inhibition, that de novo protein synthesis is
not necessary for CHO cells to undergo apoptosis. This implies that death inducing and
death suppressing proteins are always present within the CHO cell, and that death
proteins probably degrade less rapidly than survival proteins (Mercille and Massie, 1994;
Mosser and Massie, 1994). Hence, when protein synthesis is inhibited, death proteins
are able to exert their effect once the survival proteins have been removed.
This
hypothesis also implies that in a surviving CHO cell, survival proteins have to be
continuously synthesized to enable the cell to escape death. This scenario lends support
to our hypothesis that overexpression of the correct survival protein or inhibition of the
death proteins in the CHO cell may be able to delay the onset of apoptosis in response to
certain stress stimuli.
118
ICE-like cysteine protease (caspase) activity was detected in CHO cells. Hence,
our first effort was to determine if countering the effect of cysteine proteases within the
cell would be able to extend the life of a batch culture. Experiments were carried out using
z-VAD.fmk, a cysteine protease inhibiting peptide. This peptide, though not the most
potent inhibitor, was chosen since it was known to have good cell membrane
permeability. Our experiments indicate that inhibition of caspases is able to marginally
delay the onset of death in batch culture of CHO cells. When apoptotic stimuli such as
withdrawal of insulin and transferrin were introduced, cell cultures with z-VAD.fmk
initially exhibited a higher viability, with respect to the control. This protective effect
was observed right from the start of the culture. However, viability rapidly declined as
medium glucose concentrations fell to near zero, suggesting limits to the protective effect
of z-VAD.fmk.
Other experiments were conducted to determine whether overexpression of bcl-2,
a well-studied death-suppressing gene, in CHO cells would be able to extend the life of
batch cultures.
Bcl-2 has been shown to have a protective effect under cell-culture
conditions in several other cell lines (Itoh, et al., 1995; Mastrangelo and Betenbaugh,
1998; Mastrangelo, et al., 1996; Simpson, et al., 1997; Singh, et al., 1996; Singh, et al.,
1997; Suzuki, et al., 1997; Terada, et al., 1997). In the case of both normal and insulin
and transferrin deprived CHO batch cultures, bcl-2 was able to delay the onset of death
substantially. It is interesting to note that both bcl-2 and z-VAD.fmk only start to play a
role once apoptotic stimuli are experienced. The growth and viability of the control
culture is identical to that of the protected culture before the appearance of apoptotic
119
stimuli. In addition, our experiments clearly indicate that bcl-2 overexpression is a much
more effective way of preventing death in CHO cultures than using caspase inhibiting
peptides such as z-VAD.fmk.
The fact bcl-2 overexpression is a better way of extending the life of CHO cultures
than using caspase inhibitors is interesting since the point of action of bcl-2 appears to be
upstream of that of the caspases in the apoptotic pathway (Enari, et al., 1996; Wang, et
al., 1996) (Figure 4-20). This upstream location would suggest that the protective effect
of bcl-2 would at best equal that of caspase inhibitors. We were interested in determining
if having both bcl-2 and z-VAD.fmk together provided any additional protection over just
having z-VAD.fmk alone in the culture. If certain caspase independent pathways exist,
and bcl-2 is also able to block these pathways, then having both bcl-2 and caspase
inhibitors should improve culture viability. Figure 4-15 shows that bcl-2 and z-VAD.fmk
acting together do indeed provide better protection than either agent alone.
This
improved survival could not have been due to z-VAD.fmk protection of the non-bcl-2
expressing cells in the mixed culture, since we know that z-VAD.fmk alone is unable to
protect cells not expressing bcl-2, once glucose is exhausted (Figure 4-15 and Figure 4-16).
Hence, bcl-2-independent and caspase-dependent pathways appear to exist.
Our results are bolstered by reports in literature about two separate apoptotic
pathway factors that bcl-2 is able to block - cytochrome c (Kluck, et al., 1997; Liu, et al.,
1996; Yang, et al., 1997) and Apoptotic Inducing Factor (AIF) (Kroemer, et al., 1997;
Susin, et al., 1996). It is suggested that both these factors reside between the inner and
outer membranes of the mitochondria.
They are released in response to apoptotic
120
S1
S2
bcl-2
I- p35
I
death
I
Figure 4-20: Conventional pathway of death in the mammalian cells, focusing on the bcl-2
and caspase (C) nodes. S and S2 are the same or different external stimuli for apoptosis.
121
stimuli, whereupon they initiate subsequent steps in the apoptotic pathway. Bcl-2, by
virtue of its sub-cellular localization to the mitochondrial membrane, is able to block
apoptotic death by blocking the release of these agents from the mitochondria.
Cytochrome c has been shown to activate cysteine proteases.
The pathway by which
AIF causes apoptosis can be caspase-independent, since AIF has been shown to directly
induce nuclear degradation (Bossy-Wetzel, et al., 1998).
In addition, bcl-2 can also
protect against death caused by Bax activation, which can kill cells in a caspaseindependent manner (Xiang, et al., 1996).
Lastly, bcl-2 can also prevent death by
preventing reactive oxygen species (ROS) release and loss of mitochondrial potential
(which inhibits oxidative phosphorylation). Given this information and the fact that bcl-2
has been shown to act upstream of certain caspases in the apoptotic pathway, we have
attempted to come up with an death pathway (including both apoptotic and necrotic
death) in CHO cells which focuses on bcl-2 and caspases (Figure 4-21). Direct activation
of downstream caspases, such as caspase-3, by Fas or TNFR activated upstream
caspases (caspase-8 and caspase-2) or cytotoxic lymphocyte enzyme, granzyme B, are
examples of bcl-2-independent death pathways. Bax mediated death or death due to loss
of mitochondrial potential and release of ROSs represent caspase-independent death.
The fact that cells still die despite the presence of both z-VAD.fmk and bcl-2, can
be explained by one or both of the following hypotheses. First, as the level of insults gets
progressively greater, it is possible that the level of bcl-2 and/or z-VAD.fmk in the cell is
no longer sufficient to counter the levels of the death protagonist elements of the
122
Direct
activators of
downstream
caspases, e.g.
Granzyme B
Receptor-Adapter-Protease
activation of upstream
caspases, e.g. Fas-FADDcaspase-8
Apoptotic Stimuli I
bcl-2
nproteins
Death
Figure 4-21: Suggested death (including both apoptosis and necrosis) pathway in a CHO
cell. The pathway focuses on interaction between caspases and bcl-2 family proteins.
FADD, AIF and ROS refer to Fas associated death domain, apoptosis inducing factor and
reactive oxygen species, respectively. See text for more details.
123
apoptotic pathway. Second, pathways that are both bcl-2- and caspase-independent may
exist. These pathways may directly trigger events at an as yet unidentified node in the
death pathway, which is downstream of the mitochondrial and caspase checkpoints. At
this point we have no evidence to allow us to choose between these two hypotheses.
The drawback with studies carried out in the batch mode is that cultures
invariably are exposed to a nutrient limitation, which ultimately proves to be too strong a
stimulus to counter using either gene expression or peptide induced inhibition of caspases.
This observation corroborates batch culture results presented by other authors working
with several cell lines (Itoh, et al., 1995; Simpson, et al., 1997; Singh, et al., 1996; Singh,
et al., 1997; Suzuki, et al., 1997). Furthermore, this scenario of nutrient limitation is
unlikely to occur industrially. Fed-batch studies, described in a later chapter, in which
nutrient limitation ceases to be a primary cause of death, will provide an interesting
environment in which to further test the protective effect of death suppressing proteins
such as bcl-2. In addition, these studies will give us an opportunity to study the effect of
bcl-2 expression on total product titers in an environment where nutrient availability is
not a limiting factor.
124
5. Correlating Viability and Replication Competence
One of the best measures of viability for immortalized mammalian cells is their ability to
undergo mitosis when placed in a favorable environment. However, measuring mitosis is
time consuming and other assays have been developed to provide a more rapid
assessment of viability. One type of assay uses the ability of a viable cell, with its
cellular membrane intact, to exclude a vital dye as a measure of viability. Dyes used for
these types of assays include trypan blue (TB) (Patterson, 1979) and propidium iodide
(PI) (Tanke, et al., 1982; Trost and Lemasters, 1994; Vollenweider and Groscurth, 1992).
The underlying assumption in these types of assays is that the commitment of a cell to
death and loss of membrane integrity are virtually simultaneous events in the life of a cell.
Recently, it has been shown that a large number of mammalian cells in culture
undergo death via programmed cell death or apoptosis (Franek and Chladkova-Sramkova,
1995; Goswami, et al., 1998; Itoh, et al., 1995; Mastrangelo and Betenbaugh, 1998;
Mastrangelo, et al., 1996; Mercille and Massie, 1994; Moore, et al., 1995; Moore, et al.,
1997; Perreault and Lemieux, 1994; Singh, et al., 1996; Singh, et al., 1997; Suzuki, et al.,
1997; Terada, et al., 1997; Vomastek and Franek, 1993). However, the loss of cellular
membrane integrity is a very late event in the process of apoptotic death, and cells scored
as viable by the TB assay may have already committed to death. The condensation and
fragmentation of chromatin material has been found to be an event upstream of the loss of
membrane integrity, in a cell's commitment to death by apoptosis (Martin and Cotter,
125
1994; Wyllie, et al., 1980).
The Acridine Orange/ Ethidium Bromide assay, that is
increasingly being used to detect apoptosis in cells (McGahon, et al., 1995; Mercille and
Massie, 1994), overcomes the problems of late detection associated with the membrane
integrity based viability assays, by using both DNA integrity and membrane integrity to
monitor the viability of cells. Our objective was to correlate the viability measurements
provided by the aforementioned types of assays to the ability of cells in culture to
undergo mitosis. Secondly, we were also interested in determining if early apoptotic cells
(cells with intact membrane integrity but fragmented chromatin material) could undergo
mitosis.
These issues were particularly interesting in the light of the experiments
described in the previous chapter where bcl-2 expression was shown to extend viability of
CHO cells in batch culture, as measured by the AO/EB assay.
We were interested in
determining whether this observed extension in 'AO/EB' viability corresponded to an
extension in replication competence of the cells.
These issues were investigated by
following the fate of individual CHO cells.
5.1 Experimental Approach
Serum-free batch cultures of four Chinese Hamster Ovary (CHO) cell lines were grown in
suspension. For the first data-set, one of these cell lines was a control CHO cell line and
the other was a mixed culture of bcl2-CHO cells. Each culture was seeded at 5x10 5
cells/mL and reached a maximum cell density of about 2x10 6 cells/mL on day three.
Cultures reached glucose exhaustion on day three (data not shown) and this is most likely
126
the cause of death of the control culture. For the second data set, clonal populations of
pcd-CHO and bcl2-CHO were used. These cultures were inoculated at 4x10 5 cells/mL
and reached a maximum cell density of about 2x10 6 cells/mL. The control culture started
to die on day 4. No additional glucose or any other carbon source was fed to any of the
cultures. Cellular viability of a sample taken from each culture was evaluated each day (in
duplicate) using a hemacytometer via both the AO/EB and the TB techniques. In each of
the assays at least 200 cells were counted for each sample.
For replication competence assays, wells of a 384 well plate (NUNC, catalog #
242 765) were filled with 100 gL of 50% conditioned medium and 50% fresh medium.
Conditioned medium is medium taken from a culture that is in log phase growth and
where the total cell density of the culture is less than 1x10 6 cells/mL.
Using a
Fluorescence Activated Cell Sorter (FACS; Becton Dickinson, FACStarPlus), the 384
well plates were then seeded at one cell per well with cells from the cultures. Only those
cells that excluded propidium iodide, i.e., cells with intact membrane integrity, were used
to seed the wells. This was done using the FACS to gate only cells that were not stained
by propidium iodide. The 384 well plate was divided into two halves of 192 wells each.
One half of these wells were seeded with cells from the pcd-CHO cultures and the other
half with cells from the bcl2-CHO cultures. After a week, a light microscope was used to
score the number of wells in which the cells had undergone mitosis. The percentage of
wells in which cells grew, was defined as the replication competence of that population of
cells. Since cells do not grow very well when they are isolated from other cells, for each
data-set the percentage of wells in which cells grew in a plate seeded with healthy cells
127
from a log-phase growth culture was used as a base to calculate normalized replication
competence (norm. repl. comp.). To summarize:
replication competence is defined as the percentage of wells in which cells replicated, and,
normalized replication competence is defined as replication competence divided by the
percentage of wells in which healthy cells replicated.
5.2 The abilityof cells to replicate correlates much better with AO/EB
assay results than with TB assay results
The viability of cells seeded according to the trypan blue assay (TB viability) was 100%,
since only cells that were able to exclude vital dye were used to seed wells. The viability
of the seeded cells according to the AO/EB assay, defined as acridine orange (AO)
viability, is the percentage of cells scored viable (%V) by this assay divided by the sum of
the percentages of viable and early apoptotic cells (%EA) (Figure 5-1). Hence,
AO viability = %V / (%V + %EA)
TB viability = 100%
The results of the experiment are shown in Table 5-I. Normalized replication
competence was correlated with AO viability and TB viability. The corresponding plot
is shown in Figure 5-2.
As can be seen from Figure 5-2, normalized replication
competence correlates very well with AO viability (R2 = 0.9364 , t-statistic = 25.18, pvalue = 4.46x10"11). Normalized replication competence does not correlate at all with TB
viability (R2 = 0; graph not shown). This clearly indicates that the AO/EB assay is a
128
AO
EB
Viable
Late
Apoptotic
EB
AO
Early
AOE
Necrotic
EB
Figure 5-1: A pictorial representation of the four kinds of cells which can be distinguished
by the acridine orange/ ethidium bromide (AO/EB) assay. Cells which have not lost their
membrane integrity are penetrated by acridine orange only, appear green under the
microscope, and are lightly shaded in the figure. Cells which have lost their membrane
integrity also incorporate ethidium bromide and appear orange under the microscope.
These cells are hatched in the figure. The dark spots in the early and late apoptotic cells
represent condensed and fragmented chromatin material.
129
Table 5-I: Normalized replication competence (norm. repl. comp.), acridine orange
viability (AO viab.) and trypan blue viability (TB viab.) data
data set
cell type
day
repl. comp.(%)
norm. repl.
AO viab. (%)
comp. (%)
TB viab.
(%)
1
pcd
3
38.69
100.00
93.75
100
1
pcd
4
1.19
3.08
10.59
100
1
pcd
5
0
0.00
0
100
1
bcl-2
3
36.31
93.85
99.22
100
1
bcl-2
4
27.98
72.31
92.30
100
1
bcl-2
5
19.64
50.77
74.50
100
2
pcd-clone
2
57.22
86.55
99.46
100
2
pcd-clone
3
50.00
75.63
98.02
100
2
pcd-clone
5
0.00
0.00
0.00
100
2
bcl-2-clone
2
66.11
100.00
100.00
100
2
bcl-2-clone
3
60.00
90.76
99.66
100
2
bcl-2-clone
5
50.00
75.63
99.28
100
130
100%
o
-
R 2 =0.94
80%
p value = 4.5e-11
cc
= 0_
cc Q
20%
0 o
20%
E
50%
0%
100%
acridine orange viability
Figure 5-2: Acridine orange (AO) viability is a good indicator of the ability of a cell to
replicate. AO viability is defined as the percentage of viable cells in the population of
cells with intact membrane integrity. Normalized replication competence is a normalized
measurement of the number of wells in which the cells were able to replicate.
131
better method than the TB assay to determine the ability of CHO cells in a serum-free
environment to undergo mitosis.
5.3 Early apoptotic cells lose the abilityto replicate
Early apoptotic cells are defined as cells that have condensed or fragmented chromatin
material, but which have not lost their membrane integrity.
However, it is unclear
whether these cells should be classified as viable or dead. From our data, we conclude
that early apoptotic cells may be considered dead if the ability to replicate is considered
to be a true measure of viability. Table 5-I shows that the pcd-CHO cells from days four
and five in data set 1 and day five in data set 2 are mostly in the early apoptotic state and
that most of these cells fail to replicate. In fact, on day five all the pcd-CHO cells are in
the early apoptotic state (both data sets) and none of the cells in these wells grew,
providing strong evidence that early apoptotic cells lose their ability to replicate.
5.4 Bcl-2 is able to prolong the replicationcompetence of cells in culture
Previous experiments have indicated that expression of the bcl-2 gene in CHO cells is able
to delay apoptosis brought about by glucose limitation in batch culture (Goswami, et al.,
1998). Our data proves that these cells, scored as viable the AO/EB assay, also retain the
ability to undergo mitosis. As indicated above, both batch cultures ran out of glucose on
day three. This glucose limitation is the main cause of the apoptosis observed in the
control pcd-CHO cell line.
As seen from Table 5-I, the control pcd-CHO cells
132
completely lose their replication competence by day 5 (in both data sets).
In contrast,
the bcl-2 CHO cells retain their replication competence well into day 5.
The clonal
population (data set 2) performs even better than the mixed culture population (data set
1), perhaps because a larger fraction of the population expresses the bcl-2 protein. Table
5-I indicates that the cells expressing bcl-2 retain their ability to replicate much longer
than cells transfected with the control vector (pcd). Thus transfection with the bcl-2 gene
enables cells to extend the period of time during which they retain the ability to undergo
mitosis under conditions of stress.
More importantly, the viable phenotype that we
observe using the AO/EB assay is not merely an artifact related to the assay.
5.5 Discussionand Conclusions
From Table 5-I, it is evident that normalized replication competence drops off faster than
AO viability. This suggests that some of the cells that are scored viable by the AO/EB
assay, are already committed to death by apoptosis. This is entirely plausible since we
know that the chromatin condensation and fragmentation that is used as an indicator of
death in the AO/EB assay is not the earliest point in the cell's commitment to apoptotic
death. For instance, cysteine proteases are activated before chromatin condensation and
degradation can begin (Lazebnik, et al., 1994; Lazebnik, et al., 1995; Oberhammer, et al.,
1994; Orth, et al., 1996). Hence, even the AO/EB assay, or any other assay that bases its
measurement on a later manifestation of apoptosis, is not 100% accurate in determining
the viability of cells. More accurate assays would have to detect events further upstream
133
in the apoptotic pathway. However, given that apoptosis is a very rapid process (often
taking only a few hours to complete), these assays must be rapid if they are to provide
any benefit over the AO/EB assay.
In summary, the AO/EB assay is a rapid assay, that is able to assess replication
competence of a cell much more accurately than the currently used TB assay. Though we
have compared two specific assays in this chapter, the techniques and results used here
could easily be extended to other assays.
134
6. Improving CHO
Fed-Batch
Culture Performance
Using
Bcl-2 Expression
In Chapter 4 we demonstrated the ability of bcl-2 to protect batch cultures from death in
response to nutrient starvation and growth factor-depletion. Both these stimuli, however,
can easily be avoided by rational feeding strategies. We were, thus, more interested in
investigating the ability of bcl-2 expression to extend the life of cultures where nutrient
limitation was not a factor. Fed-batch culture allows us to study the protective effect of
bcl-2 under conditions in which cells are provided with all the nutrients they require.
This chapter studies the fed-batch culture of CHO cells under two conditions. The first
uses cells grown in a medium maintained complete with all nutrients and with all growth
factors by regular addition of a supplemental medium (Xie, et al., 1997; Xie and Wang,
1994). These cultures are referred to as normal fed-batch cultures.
There are several reports in literature which suggest that slowing the growth rate
of cells significantly improves the length of time for which cells remain viable and also
enhances their productivity (Chung, et al., 1998; Fussenegger, et al., 1997). One way to
reduce the growth rate of cells is to remove growth factors from the medium. In the case
of CHO cells that are used in this lab, the main growth factor is insulin. Unfortunately,
insulin is also a survival factor and cells rapidly die in the absence of insulin. However, as
demonstrated in Chapter 4, bcl-2 expression was able to prevent death resulting from
insulin deprivation but was not able to spur growth in cells.
135
Thus bcl-2 expression
provides us with a tool to reduce growth rate without triggering death. The second kind
of fed-batch cultures, referred to in this chapter as insulin-deprived fed-batch cultures,
were used to investigate whether bcl-2 expression could significantly increase life and
productivity of fed-batch cultures without growth factors, beyond that obtained in
normal fed-batch cultures.
Cell growth in normal cell cultures was observed to cease after a certain point in
time, despite the fact that the cultures were being supplemented with all the necessary
nutrients. The second half of this chapter examines some of the underlying phenomena
that might provide a clue toward the causes of this growth-arrest.
6.1 Bcl-2 expression significantlyextends viabilityand enhances product
titers in normal CHO fed-batch cultures
To investigate the beneficial effect of bcl-2 in fed-batch cultures, growth and viability in
bcl-2-CHO fed-batch cultures were compared with that in control (pcd-CHO) fed-batch
cultures. Both the cultures were clonal populations (see Materials and Methods chapter
for clone development protocols) and were seeded at approximately 5x105 cells/mL. G418 was added to both cultures at a concentration of 600 pg/mL, to maintain selective
pressure throughout the culture process. The initial medium used was modified RPMI
with 3 mM glucose and 0.5 mM L-glutamine (see Materials and Methods section for
complete medium composition). The initial medium glucose and glutamine were reduced
to minimize production of toxic metabolites lactate and ammonia, respectively (Xie, et al.,
136
1997; Xie and Wang, 1994a; Xie and Wang, 1994b). The culture was fed approximately
at 12 hour intervals with a calculated volume of supplemental medium (Xie, et al., 1997).
The volume of supplemental medium added depended on the cell density of the culture
and the expected growth rate of the cells. Cultures were stopped when the viability
dropped below 50%. Below this viability, proteases released by lysis of dead cells could
seriously affect protein quality.
As can be seen from Figure 6-1, the total cell density reached by both cell types
was almost the same (about 3.5x10 6 cells/mL). In addition, the growth rates for both
cultures were similar throughout the life of the culture.
This confirms that bcl-2
expression has no effect on the growth rate of CHO cells in fed-batch culture.
A
comparison of the viabilities and viable cell densities (Figure 6-1 and Figure 6-2),
however, clearly shows the strong protective effect of bcl-2. The bcl-2 expressing cells
reach a higher viable cell density and are able to maintain a viability of greater than 50%
almost 72 hours longer than the control cultures. The protective effect of bcl-2 starts to
become evident after about 72 hours in culture, at which point the control culture begins
to die. The stimulus for death in either culture is not yet clear. The dissolved oxygen
(D.O.) in the both cultures remained above 40% at all times, which is not considered to be
an oxygen limiting condition. However, it is possible that a 40% D.O. is sufficient to
trigger apoptosis in unprotected cells. It may also be possible that certain metabolites
produced by the cells during the culture have an adverse effect on the viability and growth
of cells. In either case, bcl-2 is able to protect against the adverse stimuli and extend
viability for almost 72 hours.
137
4.0E+6
3.5E+6
-
3.0E+6
4'
2.5E+6
4'
.
---bcl2-TCD
----- pcd-TCD
-- 4b -bcl2-VCD
S-- -- pcd-VCD
2.OE+6
CO
4'
1.5E+6
4'
4'
C. 1.OE+6
5.OE+5 -
0.OE+0
0
50
100
150
Time (hour)
200
250
Figure 6-1: Total (TCD) and viable cell densities (VCD) of bcl-2 expressing and control
(pcd) cell lines in a normal fed-batch culture seeded at 5x10 5 cells/mL.
138
4.0E+6 -
100
I1-
9E1
~
3.5E+6 -
~
-q
T
-
q
\\
S90
-80
3.OE+6 -70
tI
2.5E+6 -
-60
1H
-40
1.5E+6 -
-30
bcl2-TCD
S---pcd-TCD
- -e- - bcl2-viability
-- e--- -pcd-viability
--
-50
2.OE+6 -
CU
1.0E+6 -20
5.OE+5 -
-10
0.0E+0 -
- 0
0
50
150
100
Time (hour)
200
250
Figure 6-2: Total cell density (TCD) and viability of bcl-2 expressing and control (pcd)
cell lines in a normal fed-batch culture seeded at 5x10 5 cells/mL.
139
The benefit provided by bcl-2 expression can be quantified by comparing the yinterferon produced in the bcl-2 protected cultures to that produced in the control
cultures. However, since the y-interferon gene in our CHO cell is controlled by a viral
SV40 based S-phase promoter, protein production is growth related and slows down after
the cells have stopped growing (Banik, et al., 1996; Fussenegger, et al., 1997; Gu, et al.,
1993; Mariani, et al., 1981; Subler, et al., 1992, also see cell cycle studies below). In
addition, expression of genes such as p53 (see below) is known to inhibit activity of viral
promoters such as SV40 (Subler, et al., 1992). Since most of the protective effect of bcl-2
manifests itself only after this point of cessation of growth, using y-interferon titers as a
measure of additional benefit provided by bcl-2 expression would tend to underestimate
this protective effect. An alternative way to quantify the protection provided by bcl-2
would be to compare the integrated viable cell densities (IVCD) of the bcl-2 and control
cultures over the life of the cultures. The IVCDs calculated represent the sum of the
average viable cell densities recorded over the life of the culture, till the culture viability
decreased to below 50%. The advantage of using IVCDs to compare performance is that
they dissociate the protective effect of bcl-2 from any growth related effects of the
promoter used to express the desired protein. If a non-growth associated promoter were
used, the final titer of the cultures could easily be calculated by multiplying the IVCD
with the specific protein productivity of the cell. In that case of course, the percentage
improvements calculated based on IVCD and protein titers would be the same.
140
The final difference in the IVCDs of the bcl-2 protected and control cultures is
about 81% of the final IVCD of the control culture. Bcl-2 expression is able to improve
final y-interferon titers by about 76% (see Figure 6-3). A comparison of results is shown
in Table 6-I. Hence, bcl-2 expression is able to substantially improve fed-batch culture
titers by extending culture viabilities and maintaining cells in a productive phase.
6.2 Bcl-2 expression allows insulin-deprivedfed-batch cultures to survive
longer than normal fed-batch cultures
As described above, the purpose of the insulin-deprived fed-batch cultures was to study
whether bcl-2 expression could suppress death in the absence of growth factors in a fedbatch culture mode. Slower growth has been reported to extend viability and enhance
productivity of cells.
Hence, this set of experiments was performed to investigate
whether bcl-2 expression could allow insulin-deprived cultures to survive even longer (and
thus produce even higher titers of y-interferon) than normal fed-batch cultures.
Two different initial cell densities were investigated in our study.
The first set of
experiments compared the performance of a bcl-2 transfected and control pcd-culture,
when the cultures were seeded at 5x10 5 cells/mL (the same initial cell density as the
normal fed-batch cultures).
Of course, as expected the performance of the bcl-2
expressing culture was several fold better than that of the control culture, in terms of both
IVCD and y-interferon titers (see Figure 6-4, Figure 6-5, Figure 6-6 and Table 6-II). As in
batch cultures the difference in viability of the cultures was observed right from the start
141
--- bcl2
--- pcd
50
100
150
200
250
Time (hour)
Figure 6-3: y-interferon production with bcl-2 expressing and control (pcd) cell lines in a
normal fed-batch culture seeded at 5x10 5 cells/mL.
Table 6-I: A comparison of the results from a normal fed-batch using bcl-2 transfected
and control (pcd) clonal cell lines. IVCD represents integrated viable cell density and is
the area under the viable cell density curve (plotted against time).
IVCD
(cell-day/mL)
y-interferon
(mg/L)
bcl-2
pcd
% improvement due to
bcl-2 expression
1.88x10 7
1.04x10 7
81%
9.78
5.55
76%
142
of the cultures, indicating the strong protective effect of bcl-2 expression.
More
interestingly, the bcl-2 expressing insulin-deprived fed-batch culture survived much longer
(-350 hours) than the bcl-2 expressing normal fed-batch culture (-250 hours). However,
as expected the culture did not grow as well due to the lack of insulin, and the IVCD and
y-interferon titer of the bcl-2 culture was lower than that of the bcl-2 expressing normal
fed-batch cultures (Table 6-I and Table 6-II). The y-interferon productivity, which as
described above is very growth related, was particularly affected by the poor growth rate.
As Figure 6-6 shows, bcl-2 cells continuously synthesized y-interferon, albeit at a slower
rate, proving that bcl-2 expression was able to maintain insulin-deprived cells in a viable
and productive state.
Since cells did not grow very well without insulin, the IVCD of insulin-deprived
cultures studied above was low despite a much longer culture life span.
To examine
whether higher initial cell densities in insulin-deprived cultures could lead to substantially
improved IVCDs, another set of fed-batch cultures was initiated at a density of 8x10 5
cells/mL. The bcl-2 transfected culture grew slightly better than the corresponding culture
seeded at a lower cell density, perhaps due to the presence of more autocrine factors
which were partially able to overcome the absence of insulin. Again, the bcl-2 culture
grew more and remained more viable for a much longer time than the control culture
(Figure 6-7, Figure 6-8 and Table 6-II).
In percentage terms the performance-
improvement over the control culture was slightly better than that obtained with the
lower initial cell density, although the higher initial cell density culture did not survive as
143
1.6E+6
1.4E+6
1.2E+6
1.0E+6
---
bcl2-TCD
_---pcd-TCD
-bcl2-VCD
---pcd-VCD
---
8.0E+5
6.0E+5
4.0E+5
2.0E+5
I
0.0E+0
0
50
100
150
200
250
300
350
Time (hour)
Figure 6-4: Total (TCD) and viable cell densities (VCD) of bcl-2-protected and control
(pcd) cell lines in an insulin-deprived fed-batch with an initial cell density of 5x10 5
cells/mL.
144
100
1.6E+6
1.4E+6 -
90
1
A
I
-J 1.2E+6
80
E
70
1.0E+6 -
60
8.0E+5
50 .
o
.
6.OE+5
40
I
30
4.OE+5
4
.'
bc2-TCD
-----pcd-TCD
S-- - bcl2-viability
-- s- -pcd-viability
20
2.0E+5 -
10
O.OE+O I
i
50
100
150
i_
0
200
250
300
350
Time (hour)
Figure 6-5: Total cell density (TCD) and viability of bcl-2 protected and control (pcd) cell
lines in an insulin-deprived fed-batch with an initial cell density of 5x10 5 cells/mL.
145
4.5
4
3.5
3
+bcl-2
-.
---- pcd
2.5
2
1.5
1
0.5
7I
0 0
0
50
50
100
150
200
250
350
300
Time (hour)
Figure 6-6: y-interferon production with bcl-2 protected and control (pcd) cell lines in an
insulin-deprived fed-batch with an initial cell density of 5x1 05 cells/mL.
Table 6-II: A comparison of the results from the two insulin-deprived fed-batch runs. All
cultures contained clonal population of cells. Run 1 was started with an initial cell
density of 5x10 5 cells/mL while Run 2 had an initial cell density of 8x10 5 cells/mL. IVCD
represents integrated viable cell density and is the area under the viable cell density curve
(plotted against time).
Run 1
Run 2
bcl-2
pcd
bcl-2
1.30x10 7
1.74x10 6
1.65x10 7
1.95x10
4.58
1.22
5.51
1.33
Run 1
Run 2
improvement due to
pcd
CI-Z expression
IVCD
(cell-day/mL)
y-interferon
(mg/L)
146
6
646%
746%
274%
315%
long (- 300 hours) as the lower initial density one (- 350 hours). Once again, y-interferon
was produced throughout the life of the bcl-2 protected culture, indicating that bcl-2 is
able to maintain cells in productive state in the absence of insulin (Figure 6-9). More
interestingly, the IVCD of the bcl-2 protected culture was much closer to that obtained
during normal fed-batch culture (Table 6-I and Table 6-11).
This clearly indicates the
potential of insulin-deprived bcl-2 expressing slow growing cultures to achieve high
IVCDs. Higher initial cell densities could perhaps obtain even higher IVCDs.
Cells
maintained in the growth arrested state could conceivably have higher productivities (with
the right promoter; see Fussenegger, et al., 1997) and produce protein of more uniform
quality. There are also added advantages of lower cost and less regulatory issues by not
having to add costly growth factors to the culture medium.
6.3 CHO cells progressively arrest in the GO/G1 phase during fed-batch
culture
The growth arrest observed has interesting implications for the cell cycle in CHO cells. In
particular, we were interested in knowing if there was a particular phase in which cells in
a fed-batch culture of CHO cells arrested. It is known that hybridomas and other myeloid
cell lines arrest in the S-phase of the cell cycle since the c-myc gene in these cell lines is
deregulated (Lotem and Sachs, 1993; Miyazaki, et al., 1995; Sherr, 1994). However, cmyc has not been reported to be deregulated in CHO cells.
147
We were also
2.50E+06
--M 2.00E+06
,_1
-
E
1.50E+06 -
>cv T\
s
bcl2-TCD
----pcd-TCD
-- *--bcl2-VCD
-- o--pcd
a)
0
1.00E+06 -
killE
O 5.00E+05 -
0.00E+00
0
50
100
150
2 00
250
300
Time (hour)
Figure 6-7: Total (TCD) and viable cell densities (VCD) of bcl-2-protected and control
(pcd) cell lines in an insulin-deprived fed-batch with an initial cell density of 8x10 5
cells/mL.
148
100
2.5E+6
_
E
-80
\
!
2.0E+6
\
60
=
1.5E+6
c
U)
1.0E+6
O
5.OE+5 - -
\
40
L--I
O.OE+O
0
.
.0
a
bcl2-TCD
-.-pcd-TCD
- 4 - - bcl2-Viability
--- pcd-Viability
20
I
I
I
I
I
50
100
150
200
250
0
300
Time (hour)
Figure 6-8: Total cell density (TCD) and viability of bcl-2 protected and control (pcd) cell
lines in an insulin-deprived fed-batch with an initial cell density of 8x1 05 cells/mL.
149
5
-1
E
-
4
--
0
3
.---bcl2
--- pcd
0
I
0
0
50
I
100
I
150
I
200
I
250
300
Time (hour)
Figure 6-9: y-interferon production with bcl-2 protected and control (pcd) cell lines in an
insulin-deprived fed-batch with an initial cell density of 8x105 cells/mL.
150
interested in studying how the distribution across the different stages of the cell cycle
changed with culture time in fed-batch cultivation of CHO cells.
Cell cycle analysis, using the propidium iodide technique (see Materials and
Methods section for more details), was performed on CHO cells taken from various time
points in a fed-batch culture. Figure 6-10 shows that cells progressively arrest in the
GO/G1 phase as the culture progresses. There is no appreciable difference in cell cycle
arrest between bcl-2 expressing and control cell lines. At about 60 hours a majority of the
cells are still in the S-phase of the culture, suggesting cell division and growth. But the
proportion of cells in the S-phase drops monotonically with culture time. By the time
the cultures had stopped growing the proportion of cells in the S-phase was below 10%.
In contrast, the proportion of cells in the GO/G1 phase had increased from below 40% to
above 90% during the same time frame. There was no appreciable change in the fraction
of cells in the G2/M phase. Arrest in the GO/G1 phase suggests the expression (or lack
thereof) of specific genes involved in both apoptosis and cell cycle regulation, which
could further provide insight into the death pathway of CHO cells. Expression of some
of these genes is investigated in the following section.
6.4 p53 expression increases with time in CHO fed-batch culture but the
expression of cyclin E does not change
Since cells in CHO fed-batch cultures were observed to accumulate in the GO/Gi, we were
interested in determining whether expression of certain genes could be responsible for this
151
phenomenon. Expression of the p53 gene is known to arrest cells in the GO/G1 phase in
several cell lines (Chiou, et al., 1994; Elledge, 1996; King, et al., 1996; Sherr, 1994; Sherr,
1996). p53 has also been reported to mediate death in several cell lines by transactivating
genes such as Bax, IGF-BP3 and those involved with the Fas-apoptotic pathway (Oren
and Prives, 1996).
The observed GO/G1 arrest in CHO fed-batch culture could be
mediated by increased p53 expression. A western blot analysis of lysates of cells taken
from various points of the fed-batch culture indicated that p53 expression did increase
over the life of the culture (Figure 6-11). The expression of p53 in both bcl-2 and control
pcd cells increased, indicating that increase in p53 expression was not abrogated by bcl-2
expression.
The protective effect of bcl-2 in the presence of overexpressed p53 is
confirmed by reports in literature (Chiou, et al., 1994; Stasser, et al., 1994). The stimulus
for increased p53 expression however is still unknown.
In contrast, the expression of
cyclin E, a protein responsible for helping cells move out of the GO/G1 phase (Draetta,
1994; Koff, et al., 1993; Ohtsubo and Roberts, 1993; Ohtsubo, et al., 1995; Sherr, 1994),
did not change over the life of the culture (Figure 6-12).
This suggests that further
overexpression of cyclin E is unlikely to solve the growth problem in CHO cells (also see
discussion and conclusion section).
152
3.0E+6
100
90-
- 2.5E+6
80
a
E
70-
2.0E+6
601.5E+6
50-
.>
40
20-
- 5.OE+5
GO/G1
G2/M
----
VCD
~-se-VCD
1.OE+6
30-
--i-
O
100
0
50
100
150
- 0.OE+0
200
Time (hour)
3.0E+6
100
90
2.5E+6
80
"
E
70
2.0E+6
60
=
a,
0
1.5E+6
50
.
C
40
1.0E+6
30
20
5.OE+5
GO/G1
-uG2/M
S--VC
VCD
--
)
10
0 I
0
0.OE+0
50
100
150
200
Time (hour)
(b)
Figure 6-10: Cell cycle trends in fed-batch cultures of clonal populations of pcd (a) and
bcl-2 (b) transfected CHO cells. It can be seen that the fraction of cells in the GO/G1
phase increases monotonically with culture time, while the proportion in the S-phase
drops off. VCD refers to viable cell density of the culture. The viable cell density of the
cultures' stops increasing when more than 70% of cells are in the GO/G1 phase.
153
Cell line
*
culture age (hr.)
P
b
p
88
88
112
112
160
b
b
160
208
0 "now-
p5 3
..
Figure 6-11: Increase in p53 expression in CHO fed-batch cultures coincides with
cessation of an increase in viable cell density and follows the increase in GO/G1 phase
population. Cells transfected with bcl-2 are able to maintain their viability for a longer
period of time despite p53 expression. '*' indicates a sample from a batch culture in logphase growth, which was used as a control for p53 expression. 'p' refers to pcd-CHO
cells while 'b' refers to bcl2-CHO cells.
Cell line
culture age (hr.)
b
88
p
rainbow
112 marker 112
-~
cycin E --
:Mi
b
160
160
208
-
**I i
4=1 W.:qwjm-. anow
Figure 6-12: Cyclin E expression remains constant with culture age. The increase in
GO/G1 phase population is therefore not related to cyclin E expression. '*' indicates a
sample from a batch culture in log-phase growth, which was used as a control for cyclin E
expression. 'p' refers to pcd-CHO cells while 'b' refers to bcl2-CHO cells.
154
6.5 Specific glucose consumption rate drops off sharply just before viable
cell densitystops increasing
To obtain a better understanding of the cause of the observed growth arrest, the
consumption of nutrients during the various phases of a fed-batch culture was analyzed.
The specific consumption rate of glucose was found to be particularly interesting. Figure
6-13 shows that the specific glucose consumption rate decreases sharply just before the
viable cell density in the cultures stops increasing. Interestingly, once the viability of the
cultures starts to decrease the glucose consumption rate starts to increase again. This is
most probably due to the accumulation of glucose in the culture, which allows for more
efficient uptake of glucose by cells through the GLUT-1 transporter, the main glucose
transporter in CHO cells (Hausdorff, et al., 1996; Pessin and Bell, 1992). The transport
of glucose via the GLUT-1 transporter is concentration gradient dependent (Burant, et al.,
1991; Carruthers, 1990). The decrease in glucose consumption indicates that cells may be
experiencing a lack of nutrients due to an inability to uptake them, even though we cannot
observe any shortage macroscopically in the cellular environment. Possible causes for
this inability to uptake glucose are investigated in the next chapter.
It is surprising that the specific glucose consumption seems to increase beyond a
certain point in both cultures. To understand why this is so, we have to take into
consideration that the concentration of glucose rises beyond 100 hours in the cultures due
to overfeeding.
Since, the transport of glucose via the GLUT-1 transporters is
concentration dependent (Burant, et al., 1991; Carruthers, 1990), this increase in
concentration in glucose could possibly lead to an increase in glucose uptake rate. To
155
correct for this effect, we measured specific glucose uptake rate on a per cell, per mole of
external glucose present per time basis. As shown in Figure 6-14, there is a steady
decrease in the specific glucose consumption when measured per mole of glucose present
in the medium. The cells therefore are much more inefficient at taking up glucose after
about 100 hours into the culture.
6.6 Discussionand Conclusions
The results presented in this chapter clearly demonstrate the beneficial effect that bcl-2
expression provides in improving the performance of fed-batch cultures.
This
improvement in fed-batch performance is brought about primarily by extending the
viability of cultures. No difference in growth rates of bcl-2-transfected and control cell
cultures was observed. The improvement in the performance of fed-batch cultures was
found to be greater than 75% as measured by both integrated viable cell densities (IVCDs)
and y-interferon titers.
In evaluating these improvements one must remember that
performance of the control cultures, against which improvements are measured, has
already been optimized by stoichiometric feeding strategies.
In addition, the results from insulin-deprived fed-batch cultures suggest that
culture viability can be significantly extended beyond that obtained in normal fed-batch
cultures by using bcl-2 expression. Confirming results from batch culture experiments,
bcl-2 abrogates cell death induced by insulin deprivation, but is unable to induce cell
growth. The IVCDs obtained in insulin-deprived cultures increase with increasing initial
156
5.OE-6
3.0E+6
/
I
I
2.5E+6 -
-4.OE-6
,.
E >
E
'
2.OE+6 3.0E-6
,d/
S1.5E+6 -
/
o 1.OE+6 -
dE
2.0E-6
CD
- 1.OE-6
5.0E+5 -
0.OE+0
00
0
! 0.OE+0
I
SI
50
100
150
200
250
--. -- bcl2-VCD
--- a--pcd-VCD
-- -bcl2-SGC
.- pcd-SGC
Time (hour)
Figure 6-13: Specific glucose consumption (SGC) and viable cell density (VCD) with time
in a normal fed-batch culture of CHO cells. Specific glucose consumption drops off
before a fall in viable cell density.
157
4.0E-7
12
3.5E-7 0
0-
- 3.OE-7
Ec
8
6
2.5E-7
2.OE-7
'i
o
o
"^
4-
S1.5E-7
= 1.OE-7
2
5.OE-8
$-0
0
50
100
Time
4)
dc
i
O
0
150
O.OE+O
200
(hour)
250
.0
0
- - bcl2-glc
----pcd-glc
- --- bd2-SGCG
S.... pcd-SGCG
Figure 6-14: The glucose concentration (glc) of the fed batch media increase sharply after
100 hours due to unavoidable overfeeding and inefficient uptake of glucose. The specific
uptake rate of glucose, when measured per mole of glucose in the medium (SGCG)
continues to fall after about 90 hours in the culture. The rise in specific glucose uptake
rate seen in Figure 6-13 may therefore be explained by the rising concentrations of glucose
in the culture (see text).
158
cell densities. Interestingly, IVCDs obtained in an insulin-deprived fed-batch culture with
an initial cell density of 8x10 5 cells/mL approached those obtained with normal fed-batch
cultures. This suggests that higher initial cell densities could perhaps obtain even higher
IVCDs. The advantage in using slower growing cultures is that cells waste less of their
energy in growth related activities. It has been suggested that they are more productive
and produce less waste products in a slow growing state (Fussenegger, et al., 1997; Lao,
et al., 1996). In addition, the product produced may be of more uniform quality (since
more cells are in the same growth phase). There are also added cost and regulatory
advantages of not having to add costly growth factors to the culture medium. Clearly bcl2 expression has the potential to easily permit slow-growing, but longer surviving, fedbatch culture of CHO cells.
In almost all of our fed-batch cultures, y-interferon titer improvements did not
equal the IVCD improvement.
The trend is particularly acute in insulin-deprived
cultures. The promoter used to express y-interferon is an SV40 based promoter. The
SV40 promoter has been reported by several authors to be a growth-associated promoter,
which is most active in the S-phase of the cell cycle. Since CHO cells were shown to
arrest in the GO/G1 phase during cell culture, it is not surprising that the productivity of
these cultures is lower during the latter part of the cultures. Unfortunately, this is also
the part of the culture in which bcl-2 expression provides maximum benefit. Also, it is
entirely logical that productivity is most seriously affected in insulin-deprived cultures,
where growth is slow throughout the culture. The problem of lower productivity due to
159
slower growth can easily be addressed by using a non-growth-associated promoter.
Fussenegger, et al., 1997 have demonstrated this possibility in CHO cells, by placing the
product gene and a cytostatic gene under the control of the same non-growth associated
promoter. Bcl-2 expression goes a step further than the technique used by Fussenegger,
et al., 1997 to enhance productivity, since it also creates a much more hardy and robust
cell line in the process.
The experiments described in latter part of this chapter focused on identifying the
underlying causes behind the cessation in growth observed in fed-batch cultures.
This
cessation in growth occurs despite ensuring an adequate supply of nutrients. Cell cycle
analysis of CHO cells taken from various points in the culture indicated that cells
progressively arrested in the GO/G1 phase, suggesting a lack of growth stimulants in the
medium. In addition, western blot analysis of cell lysates confirmed that p53 expression
was upregulated in cells taken from later phases of the fed-batch culture.
This
upregulation in p53 expression also indicates loss of growth stimulus or the accumulation
of specific death stimuli in the medium (Elledge, 1996; Ko and Prives, 1996; Levine, 1997;
Sherr, 1994; Sherr, 1996). Another protein frequently associated with G to S phase
transition of cells is cyclin E (Ohtsubo, et al., 1995). Surprisingly, cyclin E expression
was found to be unchanged during the entire culture process. Renner, et al., 1995 have
reported that cyclin E induces proliferation and reduces surface-attachment requirement in
anchorage dependent cell lines. It is possible that during the adaptation of our anchoragedependent CHO cell line to suspension culture, cyclin E expression was deregulated, and
hence we see no reduction in cyclin E expression during cessation of growth in a fed160
batch.
This also suggests that further overexpressing cyclin E in our cell line, via
recombinant techniques, is unlikely to solve the growth problems in CHO fed-batch
culture.
Another surprising result presented above was that of a sudden decrease in the
specific glucose uptake rate just before the viable cell density stopped increasing.
This
would suggest that cells may be experiencing a lack of nutrients due to an inability to
uptake them, even though this nutrient starvation is not apparent through measurement of
nutrient concentrations in the external environment.
Possible reasons underlying this
reduced glucose uptake and stunted growth of cultures are explored in the next chapter.
161
7. The Fate of Insulin in CHO Fed-Batch Cultures
The results from the previous chapter indicated that although the expression of bcl-2 is
able to extend the life of fed-batch cultures of CHO cells, it is unable to affect the growth
of the cultures. It was unclear, however, why the cultures stopped growing after a certain
point in time. In addition, it was unclear what was causing the cells to die even when
they were supplied with the requisite quantities of nutrients. Analysis of the medium for
glucose and the other amino acids indicated that none of them were close to being depleted
during the culture, and hence their depletion could be ruled out as a cause of death.
Moreover, cells from the stationary phase of the fed-batch that were transferred to fresh
RPMI grew normally, indicating that the cell's intrinsic ability to divide and grow was not
in any way impaired.
Results presented in the previous chapter also indicated a progressive arrest of
cells in the GO/G1 phase. This suggested the loss of growth factors from the medium. In
addition, p53, a well known growth-suppressing and apoptosis-inducing gene, was
expressed in larger quantities as the culture progressed.
More interestingly, specific
glucose uptake rate dropped off 24 hours prior to the cessation of growth. This dip in
glucose uptake rate could be a cause of both cessation of growth and of cell death, since
we know from batch experiments that insufficient supply of glucose to cells can trigger
death. To understand why the cellular glucose uptake reduces, elements in the glucose
uptake pathway to the cell were investigated. One obvious step was the binding of
163
insulin to its receptor, which helps transport glucose into the cell. In CHO cells, insulin
binding to its receptor triggers glucose transport mainly by activating the GLUT-i
transporter gene (Hausdorff, et al., 1996; Pessin and Bell, 1992).
The GLUT-i
transporter helps the cell uptake glucose from the surroundings in a concentrationdependent manner (Burant, et al., 1991; Carruthers, 1990). Moreover insulin is a well
documented growth factor in many mammalian cell cultures (Bottenstein, et al., 1979;
Pimentel, 1994), and any disruption in the insulin signal transduction pathway could
hamper growth of cells. Possible events that could hamper the ability of insulin to
facilitate glucose uptake and spur growth in a CHO fed-batch culture were investigated,
and the results are presented in the following sections.
7.1 Reduced bindingof insulin to its receptor cannot fully explain cessation
of growth in CHO fed-batch cultures
An event that could explain the observed phenomena of growth reduction and reduced
glucose uptake was the reduced ability of insulin to bind to its receptor. This could be
caused by a change in the integrity of expressed insulin receptors as the culture
progressed. We compared the insulin-receptor binding efficiency in cells taken from day
1 and day 7 of a fed-batch culture. Samples were incubated with radioactive 1125 insulin
and then washed to remove unbound insulin. The amount of radioactivity measured in a
washed sample provided a measure of the quantity of insulin bound to the sample. The
specifically bound insulin was calculated as the difference between the mean values of
164
radioactivity in samples where only radioactive insulin was added to those in which both
radioactive insulin and a saturating amount of non-radioactive insulin was added (see
Materials and Methods chapter for more assay details). It was found that the amount of
insulin specifically bound to the cells taken from day one of a fed-batch culture was about
1.5 times that bound specifically to cells from day seven of a fed-batch culture (see Figure
7-1). Although this represents a significant reduction in insulin binding, insulin is still
able to bind to cells in the stationary phase of the fed-batch. This would tend to disprove
the hypothesis of failure of insulin binding as the major cause of the stationary phase
observed in CHO fed-batch cultures.
7.2 Insulin rapidly disappears with time in CHO fed-batch cultures
Since the binding of insulin to receptors on the cell surface was not substantially affected
by fed-batch culture, it could be that insulin itself was being affected in the fed-batch
culture. To determine if this was the case, we first decided to measure the concentration
of insulin in the medium supernatant as a function of time.
This was done using a
commercially available insulin-ELISA kit (American Laboratory Products Company, Ltd.
catalog # 10-1113-01). Figure 7-2 shows the concentration of insulin in a typical (normal)
fed-batch as the culture proceeds. The secondary y-axis also shows the corresponding
viable cell densities of the bcl-2- and pcd-CHO cultures. Figure 7-2 indicates that the
concentration of insulin in the supernatant falls rapidly from an initial value of 5 mg/L to
undetectable levels within 96 hours of the start of the culture. The viable densities stop
165
5000
4500
E" 4000o 3500
S3000 0 2500
o 2000
S1500
Z 1000 500
0
old
new
Figure 7-1: Insulin binding to cells taken from different points in a fed-batch culture. Net
radioactivity is a proxy for the quantity of insulin bound in a specific manner to its
receptor (see text). The higher the net radioactivity value the larger to quantity of insulin
bound. 'Old' and 'new' refer to samples taken from day 7 and day 1 of a fed-batch
culture, respectively.
166
3.0
3.0E+6
2.5
2.5E+6
2.0
2.0E+6
u0
1.5
1.5E+6
>
1.0
1.OE+6
Q
0.5
5.OE+5 0
-j
-- 4-- bc2-ins
0.0 !
0
50
100
,,
---
150
200
-a- -pcd-ins
bcl2-VCD1
----- pcd-VCD
0.OE+0
250
Time (hour)
Figure 7-2: Insulin concentration as a function of time in a typical fed-batch culture of
CHO cells. The initial concentration of insulin in the culture is 5 mg/L. Insulin
concentrations drop rapidly to almost non-detectable levels by 100 hours. Culture
growth stops soon after. The suffix '-ins' refers to the concentration of insulin in either
the bcl2-CHO or pcd-CHO cell culture. VCD refers to the viable cell density in either
culture.
167
increasing approximately 24 hours after the insulin disappeared from the medium.
Interestingly, there is no significant difference between the disappearance of the insulin
from the bcl-2 transfected and control (pcd) cultures, indicating that the disappearance of
insulin from the medium is a phenomenon intrinsic to CHOs that is not affected by
transfection with bcl-2.
7.3 Insulin degrading activityin CHO fed-batch cultures is concentrated in
the supernatant
Although the measurement of insulin in the fed-culture medium indicated that it was
disappearing rapidly as the culture progressed, it was not clear what was causing this
disappearance. One possibility was that insulin was binding (perhaps non-specifically)
to the surface of the cells and thus being removed from the medium. Another possibility
was that the cells were using insulin much more inefficiently as the culture progressed,
causing a rapid consumption of insulin. Lastly, it was also possible that the cells were
secreting either a protease or a binding protein that was either degrading the insulin or
preventing it from being used by cells or being detected by ELISA.
To differentiate between these alternatives we added fresh doses of 5 mg/L of
insulin to a 2 mL sample of supernatant (medium from which cells have been removed)
and to cell containing medium on various days of a normal fed-batch culture. The samples
were then incubated under culture conditions, and insulin degradation was then quantified
over a period of 24 hours by measuring the quantity of insulin in the supernatant or the
168
cell-containing medium at 0, 2, 6, 10 and 24 hours.
For ease of comparison, the
concentrations of insulin in the samples were expressed as a fraction of the concentration
at '0' hours. Figure 7-3 and Figure 7-4 show the results from the experiment.
It is
evident that the degradative activity in both the supernatant and the cell-containing
medium increase as a function of culture time. Also, on any particular day, the
concentration of insulin in the sample undergoes a monotonic decrease with time. This is
more evident in the samples taken earlier in the culture where the insulin does not
disappear too rapidly.
In addition, comparing the results from days 6 and 7, the
degradative activity of the supernatant alone is almost identical to that of the cell
containing medium.
Thus the degradative activity is primarily in the supernatant,
favoring the hypothesis of a protease or binding protein being responsible for the
disappearance of insulin. It should also be noted that almost all the insulin added
disappears within 24 hours, especially towards the later part of the culture. This explains
why earlier experiments, in which insulin was added to the culture in an effort to spur
increases in viable cell density, failed to achieve any results.
To guard against the possibility that the conditions of incubation (culture
conditions) are responsible for the disappearance of insulin, a control experiment was
performed to study the degradation kinetics of insulin under these conditions in fresh cellfree fed-batch culture medium.
Figure 7-5 indicates that insulin does not degrade
significantly in a fresh cell-free medium under culture conditions even after almost 200
hours, which is far longer than the time scale of 24 hours in which most of the insulin
disappears from the supernatant of a fed-batch culture. Hence, the insulin degradation
169
100%
'- 90%
80%
*" 70%
60o
r
day 3i
0%
Iaday 4
day 6
7
S40% --__
30% -
_
--_
Dday
_
20%
LL
10% -
0
2
6
10
24
Time (hour)
Figure 7-3: Degradation of insulin in fed-batch supernatant (culture medium with cells
spun down and removed) from various days in a CHO fed-batch culture. The degradation
of insulin in samples from each day, under cell-culture conditions, was followed for 24
hours. 5 mg/L of insulin was added to the supernatant at '0' hours, and the concentration
of insulin measured in the supernatant at this point was denoted as 100%. Insulin in any
particular sample was expressed as a fraction of the insulin concentration in the same
sample at '0' hours.
170
100%
_
80%
w
60% -
o day 5
13 day 6
O day 7
40% -
.0
20%
L
U-
0
6
2
10
24
Time (hour)
Figure 7-4: Degradation of insulin in fed-batch medium (with cells) from various days in a
CHO fed-batch culture. The degradation of insulin in samples from each day was
followed for 24 hours. 5 mg/L of insulin was added at '0' hours, and the concentration of
insulin measured in the supernatant at this point was denoted as 100%. Insulin in any
particular sample was expressed as a fraction of the insulin concentration in the same
sample at '0' hours.
171
120%
100%
C
"
80%
E
S60%
S40%
C
L_
0%
0
50
100
150
200
Time (hour)
Figure 7-5: Degradation kinetics of insulin in fresh (cell-free) fed-batch medium under
standard cell-culture conditions. 5 mg/L of insulin was added to the medium at '0' hours.
The concentration of insulin in the medium at various time points was expressed as a
fraction of the concentration at '0' hours.
172
activity that we observe in the supernatant of a fed-batch culture is due to a substance
present in the supernatant that is not present in the original medium. In addition, insulin
degradation due to incubator conditions is negligible.
7.4 Boiling of fed-batch supernatant removes all insulin degrading activity
To asses the nature of the component in the fed-batch medium which was causing the
insulin to disappear, the supernatant was boiled for three minutes and fresh insulin was
added to it after cooling.
supernatant.
Boiling destroys any protease activity present in the
As Figure 7-6 indicates the boiled supernatant showed almost no
degradation of insulin after 24 hours, while in the unboiled control almost all the insulin
had disappeared after the same period of time. In fact, the boiled control retained almost
complete insulin activity even after 120 hours. Therefore the component responsible for
the disappearance of insulin could be a protease. However, a binding protein, secreted by
the cells during the culture process, which binds to insulin and prevents its detection by
ELISA could also cause the disappearance of insulin.
There is evidence that p53
expression in cells (p53 expression increases during CHO fed-batch cultivation, see
results from previous chapter) leads to secretion of an IGF-1 binding protein
(Buckbinder, et al., 1995). It is possible that CHO cells secrete a similar binding protein,
which directly affects insulin. Boiling of the supernatant would also destroy any such
binding protein activity. The boiling experiment confirmed the presence of degradative
173
100%
0 90%
E 80%
E
*
.
E
S70%
60%
50%
40% 30% -
----
boiled
unboiled control
20%
10%
u.0%
o
o
0
10
20
Time (hour)
30
Figure 7-6: Boiling fed-batch supernatant for three minutes removes its insulin degrading
activity. 5 mg/L of insulin was added to the boiled sample and unboiled control at '0'
hours. Both samples were incubated under standard cell-culture conditions. The
concentration of insulin in the medium at various time points was expressed as a fraction
of the concentration at '0' hours.
174
activity in the supernatant, but, further experiments were needed to distinguish between
protease and binding protein activity.
7.5 Aminopeptidase inhibitors are
able to substantially reduce the
degradation of insulinin the fed-batch supernatant
If the disappearance of insulin in the supematant were due to a protease, addition of
protease inhibitors to the supernatant should prevent its insulin degrading activity. A
literature search on insulin degrading proteases that are normally produced by cells
(Fukuda, et al., 1995; Morishita, et al., 1992; Tozaki, et al., 1997; Yamamoto, et al., 1994)
suggested several protease inhibitors that prevented degradation of insulin in vivo. The
concentrations of various protease inhibitors added to supematant samples are indicated
in Table 7-1. At the start of the experiment (zero hours) 5 mg/L of insulin was added to
each supernatant sample. The concentrations of insulin in the supernatant were measured
at 0, 0.5, 1.5, 3 and 13.5 hours, respectively. Degradation activity in these samples was
compared to the degradation in a control supematant to which no protease inhibitors were
added.
The results from the study are shown in Figure 7-7. For ease of comparison, the
concentrations of insulin in the samples were expressed as a fraction of the concentration
at '0' hours.
The bacitracin obtained from Sigma contained some intrinsic protease
activity, and this explains the rapid disappearance of insulin in the bacitracin and 'All'
samples.
Figure 7-7 indicates that sodium glycocholate was able to reduce insulin
175
Table 7-I: Concentrations and sources of various protease inhibitors used to study insulin
degradation in CHO fed-batch culture supernatants.
Inhibitor Used
Final Concentration
Source
Soybean Trypsin Inhibitor
10 mg/mL
Sigma-Aldrich, T-9128
Bacitracin
20 mM
Sigma-Aldrich, B-0125
Aprotinin
100 gg/mL
Sigma-Aldrich, A-6279
Sodium Glycocholate
10 mM
Sigma-Aldrich, G-7132
All
above concentrations of all
inhibitors added to sample
176
120%
"E 100%
E
a
80%
All
r---- Trypsin Inhibitor
--&--Bacitracin
-=o60%0
40% -
'-v-Aprotinin
20% -
---- Na Glycocholate
---- Control
o
_
0%
0
5
10
15
Time (hour)
Figure 7-7: Effect of various protease inhibitors on the degradation of insulin in the
supernatant from a fed-batch culture. 'All' indicates that all the protease inhibitors were
added to this sample at the concentrations suggested. No protease inhibitors were added
to the 'control' culture. 5 mg/L of insulin was added to the medium at '0' hours. The
concentration of insulin in the medium at various time points was expressed as a fraction
of the concentration at '0' hours. All samples were incubated under standard cell-culture
conditions for the duration of the experiment.
177
degradation in the supernatant to the largest extent, with almost 90% of the insulin
remaining after 13.5 hours. It was confirmed that sodium glycocholate did not cause a
false positive with the insulin assay used. Aprotinin and Soybean Trypsin Inhibitor
(STI) displayed the next best insulin-protease-inhibitor activity.
Supematants treated
with these inhibitors managed to retain 50% of original insulin activity after 13.5 hours.
In contrast, the control, which had no added protease inhibitors, only retained 33%
insulin activity after 13.5 hours.
Sodium Glycocholate is an aminopeptidase inhibitor (Fukuda, et al., 1995).
Hence, the results indicate that the protease(s) in the supernatant of a fed-batch culture
degrades insulin from the N-terminal position.
Furthermore, since serine protease
inhibitors such as aprotinin and STI also reduced the degradation of insulin, it is possible
that there are also some trypsin- and chymotrypsin-like proteases in the supernatant.
7.6 Adding large quantities of excess insulin also reduces the rate of
degradation of insulinin the fed-batch supernatant
If the degradation of insulin in the fed-batch supernatant was caused by binding of insulin
to a binding protein secreted by the cells, the addition of large excesses of insulin should
saturate the binding protein available and prevent disappearance of insulin. In order to
determine the effect of excess insulin on the rate of insulin disappearance from the
supernatant, we added 500 gg/mL and 50 gg/mL insulin to the supernatant, and compared
insulin disappearance in these samples to that in the control (5 gg/mL insulin). Insulin
178
concentrations were measured at 0, 0.5, 1.5, 3 and 13.5 hours after the initial addition of
insulin.
For ease of comparison, the concentrations of insulin in the samples were
expressed as a fraction of the concentration at '0' hours. Figure 7-8 shows that increasing
the concentration of insulin to 100X (500 gg/mL) and to 10X (50 gg/mL), caused only
73% and 66% of the insulin, respectively, to degrade after 13.5 hours. In contrast, only
33% of the insulin at time zero was remaining in the control (IX insulin) after 13.5 hours.
The reduction in the rate of insulin degradation due to the addition of excess insulin could
be due to saturation of binding protein present in the supernatant..
However, if non-
catalytic insulin proteases were present in the medium, the observed reduction in protease
activity could also be due to saturation of the activity of these proteases.
7.7 Discussion and Conclusions
The experiments described in this chapter focused on identifying the underlying causes of
the cessation in growth observed in CHO fed-batch cultures. This cessation in growth
occurs despite ensuring an adequate supply of nutrients and adding extra doses of the
growth factor insulin. The results presented in the previous chapter also indicated that
specific glucose consumption dips just before cessation in viable cell density growth.
Cells in fed-batch cultures progressively arrested in the GO/G1 phase and upregulated p53
expression. This chapter explored some events that might explain these observations.
The ability of insulin to bind its receptor was found to be reduced in later phases
of a fed-batch culture, but the magnitude of this decrease was not sufficient to explain the
179
120%
100%
E
80%
S
-,
100X insulin
I--e-1OX insulin
-- control
60% -
---
)
40%0
o
20%
0%
0
5
10
15
Time (hour)
Figure 7-8: Adding excess insulin reduces the rate of insulin degradation, possibly due to
the saturation of proteases or insulin-binding proteins in the fed-batch supernatant. The
control indicated has 1X or 5gg/mL of insulin added to it at time zero.
The
concentration of insulin in the medium at various time points was expressed as a fraction
of the concentration at '0' hours. All samples were incubated under standard cell-culture
conditions for the duration of the experiment.
180
observed phenomena. However, the increase in p53 expression and the cessation of
growth both appear correlated to the rapid disappearance of insulin from fed-batch
cultures.
Insulin is a known growth and survival factor for the culture of several
mammalian cell lines (Bottenstein, et al., 1979; Murakami, 1989; Pimentel, 1994; Zhou
and Hu, 1995). Insulin concentrations in the culture were observed to fall below
detectable levels at least 24 hours before the viable cell density stopped increasing. The
lag in cessation of growth could be due to insulin receptors remaining phosphorylated for
some time after the insulin has disappeared from the medium. Interestingly, these results
corroborate the work of Drapeau, et al., 1994 and Lao, et al., 1996, who demonstrated
insulin degrading activity in batch-refeed processes and perfusion cultures, respectively.
The authors did not, however, explore the cause of the insulin degrading activity.
Insulin degrading activity in CHO fed-batch culture was shown to be concentrated
in the supernatant, since there was no observed difference in degradation rates of insulin
between culture medium in which cells were still present and culture medium from which
cells were removed. Since the rate of insulin degradation is almost completely removed
by boiling and significantly reduced by specific protease inhibitors, insulin degradation is
definitely protease related. Since the best reduction in protease activity was provided by
sodium glycocholate, an amino-peptidase inhibitor, amino-peptidases appear to play a
major role in insulin degradation in CHO fed-batch cultures. Furthermore, trypsin- and
chymotrypsin-like proteases may also play a role in insulin degradation since serine
protease inhibitors such as aprotinin and soybean trypsin inhibitor were able to reduce
insulin degradation. Interestingly, adding large excesses of insulin to the supernatant also
181
reduces the rate of insulin degradation. This may be due to the saturation of non-catalytic
protease activity or due to the quenching of any insulin-binding-protein activity that may
be present. Binding-protein activity, while not specifically confirmed, cannot be ruled
out at this stage. It may be noted that p53 expression has been shown to cause secretion
of a IGF binding protein, IGF-BP3, in some cell lines. It is not unreasonable, therefore, to
expect that the observed p53 expression in CHO fed-batch cultures can trigger secretion
of similar binding proteins that affect insulin.
No insulin degrading activity was observed in fresh medium.
In addition, the
insulin degrading activity in a fed-batch culture progressively increases with time, as
evidenced by increasingly rapid rates of disappearance of added insulin. Insulin degrading
proteases and binding proteins (if any) must therefore be secreted by the cells. Also the
degradative activity starts to accumulate before extensive death and cell lysis occur,
indicating that the insulin degrading agents are released as a part of the normal metabolic
process. Since the insulin degrading activity increases with culture time, the degrading
agents must accumulate with time indicating that they have a fairly long half-life.
The above experimental outcomes suggest that the degradation of insulin in fedbatch cultures can be reduced in one of two ways. First, very high doses of insulin could
be added to the culture at regular intervals. Of course, cost of adding such large quantities
of insulin is a drawback.
In addition, large quantities of insulin are known to
downregulate the insulin receptor, necessitating even higher quantities of insulin in the
culture. A second strategy to reduce the disappearance of insulin would be to add a
cocktail of protease inhibitors. The above results suggest that amino-peptidase inhibitors
182
should be an essential part of this cocktail. The cost of these protease inhibitors and the
optimum concentrations at which they are most effective without being toxic to cells are
issues that will need to be addressed in future studies.
183
8. Conclusions and Recommendations
8.1 Conclusions
The results presented in this thesis showed that CHO cells in serum free culture die
primarily by apoptosis. A new assay, the acridine orange / ethidium bromide (AO/EB)
assay, was optimized to allow rapid analysis of apoptosis in CHO cells. In addition,
apoptosis was confirmed by the degradation of genomic DNA into multiples of 180 bp
fragments at the onset of apoptosis. Protein synthesis inhibition demonstrated that death
proteins dominate in CHO cells and survival proteins must be continuously synthesized
in order to protect cells from death. Activation of caspases was observed in apoptotic
CHO cells and confirmed by the cleavage of substrates of two different caspases - Poly(ADP-Ribose)-Polymerase (PARP), which is a substrate of caspase-3, and lamin A,
which is a substrate of caspase-6. Caspases are perhaps the most important apoptosis
agonists in eukaryotes and their activation is considered by some researchers to be
essential in the apoptotic process.
Surprisingly, inhibition of caspase activity (as confirmed by non-cleavage of
substrates) by universal and irreversible caspase inhibitor z-VAD.fmk failed to prevent
death in CHO cells. In contrast, expression of the anti-apoptotic gene bcl-2 was able to
significantly extend the viability of CHO cells in batch culture, in response to glucose
starvation at the end of a batch. Bcl-2 expression was able to completely abrogate death
in batch culture response to removal of growth and survival factor, insulin. z-VAD.fmk
again failed to have a significant protective effect. In both these cases, bcl-2 expression
completely prevented caspase activity confirming previous reports that it lay upstream
185
of caspases in the apoptotic pathway. The failure of caspases to protect cells from death
was surprising since previous reports had indicated that caspases were universally
required in apoptosis, and that inhibition of their activity would be a more potent agent
for prevention of death than overexpressing bcl-2. Our results show this not to be the
case, suggesting that bcl-2-dependent and caspase-independent death pathways do exist,
at least in CHO cells.
These results have since been confirmed in other reports in
literature (Xiang, et al., 1996). In conclusion, bcl-2 expression was found to be a much
better way of extending batch culture viability than using peptide inhibition of caspases.
This better protection afforded by bcl-2 is attributed to its ability to not only prevent
activation of downstream caspases but also maintain mitochondrial potential, thus
protecting a cell's ability to carry out oxidative phosphorylation and preventing the
release of reactive oxygen species.
Before claiming significant improvement in cellular viability using bcl-2
expression, it was important to confirm the increased viability with another assay which
went beyond measuring membrane permeability and chromatin integrity.
The more
stringent benchmark chosen for viability was the ability of a cell to undergo mitosis when
placed in a more favorable environment. An experiment that allowed us to follow the fate
of individual cells in a favorable environment was designed.
The results from this
experiment (discussed in Chapter 5) proved that the viabilities observed with bcl-2
correlated with the continuing ability of a cell to undergo mitosis. It also showed that the
AO/EB assay results correlated much better with replication competence than did the
results of the trypan blue assay.
Lastly, this set of experiments proved that early
186
apoptotic cells could be considered dead, if replication competence is considered as a
stringent measure of viability.
Bcl-2 expression combined with stoichiometric medium design and feeding
strategies was able to significantly improve the performance of fed-batch cultures beyond
that obtained using just stoichiometric medium design and feeding strategies without bcl-2
expression. Final gamma interferon (IFN-y) concentrations (the model protein secreted
by the CHO cells used) and integrated viable cell densities (IVCDs) were both improved
by a factor of at least 75% as compared to control cultures, by using bcl-2 clonal cell
lines. Bcl-2 clonal cultures had at least a 600% higher IVCD as compared to the control
culture, in an insulin-deprived environment.
Bcl-2 expression also extended culture
lifespans in these slower growing insulin-deprived fed-batch cultures beyond that
obtained with bcl-2 expressing cells in insulin-supplemented fed-batch medium by as
much as 40%.
For insulin-deprived cultures seeded at higher densities, the IVCDs
obtained with bcl-2 expressing clones approached those obtained with bcl-2 expressing
clones in insulin-supplemented cultures. However, although bcl-2 expression allowed
cells to remain viable and productive in the absence of insulin, the productivities were
lower than those obtained in an insulin-supplemented medium. This was most likely due
to the fact that the SV40-based promoter used to express IFN-y is not as active in the
absence of growth. Nevertheless, these experiments demonstrate the potential of bcl-2
expression to permit slow growing cultures to achieve much higher lifespans and
productivites (provided a non-growth associated promoter is used) than the faster
187
growing cultures. In fact some reports have suggested that slower growing cultures are
more productive than faster growing ones (Fussenegger, et al., 1997). Bcl-2 expression
provides these benefits in addition to advantages of a more robust cell line, lower cost (no
use of costly growth factors) and fewer regulatory issues (related to the use of undefined
growth factors).
The cessation of growth in fed-batch cultures was observed to be concurrent with
accumulation of cells in the GO/G1 phase. The expression of p53, a well-characterized
growth-inhibiting and apoptotic protein, which arrests cells in the GO/G1 phase, was also
observed to increase throughout the culture. The increase in p53 expression could also be
a factor in the increased death observed soon after viable cell density stops increasing.
p53 expression has been shown to activate Bax expression. Bax is an apoptotic member
of the bcl-2 family, which dimerizes with bcl-2. Increasing quantities of Bax can negate
the protective effect of bcl-2 and cause death of cells. This hypothesis may explain why
even bcl-2 expressing fed-batch cultures ultimately perish. The expression of cyclin E, a
protein known to drive G to S phase transition in cells, was observed not to change during
the fed-batch process, indicating that this gene might have been deregulated during the
process of adaptation to suspension culture from anchorage-dependent growth. Further
overexpression of cyclin E is therefore unlikely to increase growth rates.
The specific glucose uptake rate was observed to drop right before the cessation
of viable-density growth in fed-batch cultures.
This suggests that cells might be
experiencing glucose starvation, which has already been shown to be a potent death
stimulus, due to a failure to uptake glucose. A closer look at insulin, known to be a
188
critical factor in cellular glucose uptake and growth, provided some surprising results.
The ability of insulin to bind to its receptor was found to be significantly reduced, but not
enough to explain the complete cessation of growth. More interestingly, insulin was
found to disappear at increasingly rapid rates from the fed-batch medium as the culture
progressed, falling to undetectable levels about 24 hours before the cessation of growth.
The insulin degrading activity was isolated to the medium (and not the cells) and was
shown to be blocked by amino-peptidase inhibitors such as sodium glychocholate and to
a much lesser extent by serine protease inhibitors. Surprisingly, insulin degradation rates
were also reduced by using high concentrations of insulin, suggesting the presence of
irreversible protease activity or insulin binding proteins.
8.2 Recommendations
The fact that bcl-2 is able to significantly extend viabilities in batch and fed-batch culture
of CHO cells does not imply that further improvements cannot be made. There are some
reports to suggest that other homologues of bcl-2, such as bcl-XL, prevent apoptosis in
cases where bcl-2 does not (Boise, et al., 1993; Cheng, et al., 1996; Gottschalk, et al.,
1994). In addition, proteins that interact with bcl-2 to increase its protective effect have
been reported. One such protein, Bag-l, enhances bcl-2 activity presumably by helping it
interact with Raf-1 and inducing phosphorylation and deactivation of apoptotic protein
Bad (Reed, 1997). Co-expression of bcl-xL and/or Bag-1 with bcl-2 in CHO cells could
further improve viabilities.
189
Since caspases are important effectors of apoptosis in CHO cells, further research
needs to be performed at trying to inhibit caspases using genetic expression of caspase
inhibiting proteins.
One such protein, the baculovirus protein p35, has been
demonstrated to be a strong and universal inhibitor of caspases
in the nematode C.
elegans, and some insect and mammalian cell lines (Bump, et al., 1995; Clem, et al., 1991;
Rabizadeh, et al., 1993). However, preliminary attempts to transfect cells stably with
p35 failed in this lab, and reportedly at some other labs (H. Steller, personal
communication). The second set of proteins that have been shown to inhibit caspase
activity are the IAPs (for inhibitor of apoptosis proteins; Bump, et al., 1995; Clem, et al.,
1996; Hawkins, et al., 1996; Liston, et al., 1996; Orth and Dixit, 1997). Several human
homologues of these proteins have recently been identified. Since results presented in
Chapter 4 (and independent reports in literature, see Cuende, et al., 1993; Lincz, 1998;
Reed, 1997) suggest that bcl-2 independent but caspase-dependent death pathways exist
in mammalian cells, co-expression of bcl-2 and specific caspase-inhibiting proteins in
CHO cells has the potential to significantly improve the resistance of cells to death
stimuli.
There have recently been significant developments in mammalian expression
systems that can help us implement some of the above recommendations more efficiently.
Promoters which can provide higher levels of bcl-2 expression than the CMV promoter
are now available commercially. One such promoter is tetracycline-inducible (Clontech),
which would permit the study of different expression levels of bcl-2 and other proteins in
the same cell line. Plasmids with an internal ribosomal entry site (IRES), which can allow
190
the simultaneous expression of two or more genes using the same promoter (translation of
two genes of a single mRNA transcript), are also now commercially available (Clontech;
Fussenegger, et al., 1998). Bicistrionic plasmid expression systems in which ones of the
genes is that responsible for green fluorescence protein expression are also available
(Clontech).
These expression systems permit the easy and rapid selection of high
expression clones using flow-cytometry.
The results presented in Chapter 6 suggest that slower growing fed-batch cultures
survive much longer than faster growing cultures. In addition, previous work in this lab
has suggested that slower growth in cultures leads to higher glycosylation efficiencies
(Nyberg, 1998). Further, reports in literature suggest that cells are more productive in the
stationary phase (Fussenegger, et al., 1997). However, these advantages have not been
translated into higher product titers in our cell line due to the use of a growth phase
dependent promoter (Banik, et al., 1996; Gu, et al., 1993; Gu, et al., 1994; Mariani, et al.,
1981). Expressing IFN-y with a promoter that is active in the S-phase (see Fussenegger,
et al., 1997 for a review) is strongly recommended to be able to better study productivity
and quality issues in a slow-growing cultures.
Insulin disappearance was shown to be significantly reduced in the supernatant
from fed-batch media by using amino-peptidase inhibitor sodium glycocholate and by
adding larger doses of insulin. Further studies need to be performed to investigate the
effects of directly adding protease inhibitors to fed-batch cultures of CHO cells on insulin
degradation, cell growth and cell viability. Protease inhibitors should be chosen based on
191
both effectiveness of protease inhibition and the absence of adverse effects on cells.
Adding large doses of insulin to cultures is not recommended due to cost issues and also
the possibility of downregulation of insulin receptors on the cell surface. Another more
exotic approach would be to design (or select) a clone that can grow in the absence of
insulin (Lao, et al., 1996; Renner, et al., 1995).
It is interesting to observe that even bcl-2 expressing cultures ultimately die
rapidly after a point in fed-batch culture. The level of protection afforded by bcl-2
depends on its level of expression in the cell.
Further increases in levels of bcl-2
expression using a different promoter may increase the life of bcl-2 fed-batch cultures. In
addition, the level of protection offered by bcl-2 expression also depends on the
expression levels of other dimerizing members of the bcl-2 family, most importantly that
of Bax. p53 expression, which was shown to increase during the course of a fed-batch
culture, has been reported to stimulate Bax expression (Han, et al., 1996; Matsuyama, et
al., 1998; Miyashita and Reed, 1995; Oren and Prives, 1996). It would be interesting to
see if Bax expression increases with time in CHO fed-batch cultures. p53 expression has
also been reported to induce the expression of transforming growth factor beta (TGF-3),
another apoptotic and growth inhibiting protein (Fussenegger, et al., 1997).
Very
preliminary studies in our cell line indicated an accumulation of TGF-P in fed-batch
culture of CHOs. Further research into expression of these and other apoptotic stimuli
secreted as a part of the normal metabolic process can provide a better idea of the
192
apoptotic pathway in CHO cells and suggest mechanisms to further extend viability and
productivity in CHO cell cultures.
193
9. References
Alnemri, E. S., Livingston, D. J., Nicholson, D. W., Salvesen, G., Thornberry, N. A.,
Wong, W. W. and Yuan, J. 1996. Human ICE/CED-3 protease nomenclature (letter).
Cell. 87: 171.
Anderson, P. 1997. Kinase Cascades Regulating Entry into Apoptosis. Microbiol. Mol.
Biol. Rev. 61: 33-46.
Backer, J. M., Kahn, C. R. and White, M. F. 1990. The Dissociation and Degradation of
Internalized Insulin Occur in the Endosomes of Rat Hepatoma Cells. J. Biol. Chem.
265: 14828-35.
Banik, G. G., Todd, P. W. and Kompala, D. S. 1996. Foreign protein expression from S
phase specific promoters in continuous cultures of recombinant CHO cells.
Cytotechnology. 22: 179-184.
Bissonnette, R. P., Echeverri, F., Mahboubi, A. and Green, D. R. 1992. Apoptotic cell
death induced by c-myc is inhibited by bcl-2. Nature. 359: 552-554.
Boise, L. H., Gonzalez-GarcIa, M., Postema, C. E., Ding, L., Lindsten, T., Turka, L.
A., Mao, X., NuOez, G. and Thompson, C. B. 1993. Bcl-x, a bcl-2-Related Gene
That Functions as a Dominant Regulator of Apoptotic Cell Death. Cell. 74: 597-608.
Bossy-Wetzel, E., Newmeyer, D. D. and Green, D. R. 1998. Mitochondrial cytochrome c
release in apoptosis occurs upstream of DEVD-specific caspase activation and
independently of mitochondrial transmembrane depolarization. EMBO J. 17: 37-49.
Bottenstein, J., Hayashi, I., Huchings, S., Masui, H., Mather, J., McClure, D., Ohasa, S.,
Rizzino, A., Sato, G., Serrero, G., Wolfe, R. and Wu, R. 1979. The growth of cells in
serum-free hormone supplemented media. In: W. B. Jakoby and I. H. Pastan,
Methods in Enzymology, Academic, New York.
Boyd, J. M., Gallo, G. J., Elangovan, B., Houghton, A. B., Malstrom, S., Avery, B. J.,
Ebb, R. G., Subramanian, T., Chittenden, T., Lutz, R. J. and Chinnadurai, G. 1995.
Bik, a novel death-inducing protein shares a distinct sequence motif with Bcl-2 family
proteins and interacts with viral and cellular survival-promoting proteins. Oncogene.
11: 1921-8.
195
Buckbinder, L., Talbott, R., Velasco-Miguel, S., Takenaka, I., Faha, B., Seizinger, B. R.
and Kley, N. 1995. Induction of the growth inhibitor IGF-binding protein 3 by p53.
Nature. 377: 646-649.
Bump, N. J., Hackett, M., Hugunin, M., Seshagiri, S., Brady, K., Chen, P., Ferenz, C.,
Franklin, S., Ghayur, T., Li, P., Licari, P., Mankovich, J., Shi, L., Greenberg, A. H.,
Miller, L. K. and Wong, W. W. 1995. Inhibition of ICE Family Proteases by
Baculovirus Antiapoptotic Protein p35. Science. 269: 1885-88.
Burant, C. F., Sivitz, W. I., Fukumoto, H., Kayano, T., Nagamatsu, S., Seino, S., Pessin,
J. E. and Bell, G. I. 1991. Mammalian Glucose Transporter: Structure and Molecular
Regulation. Recent Prog. Hormone Res. 47: 349-89.
Carruthers, A. 1990. Facilitated Diffusion of Glucose. Physio. Rev. 70: 1135-76.
Cheng, E. H., Levine, B., Boise , L. H., Thompson, C. B. and Hardwick, J. M. 1996. Baxindependent inhibition of apoptosis by Bcl-xL. Nature. 379: 554-556.
Chinnaiyan, A. M., O'Rourke, K., Tewari, M. and Dixit, V. M. 1995. FADD, a novel
death domain-containing protein, interacts with the death domain of Fas and initiates
apoptosis. Cell. 81: 505-12.
Chiou, S. K., Rao, L. and White, E. 1994. Bcl-2 blocks p53-Dependent Apoptosis. Mol.
Cell. Biol. 14: 2256-2263.
Chittenden, T., Harrington, E. A., O'Connor, R., Flemington, C., Lutz, R. J., Evan, G. I.
and Guild, B. C. 1995. Induction of apoptosis by the Bcl-2 homologue Bak. Nature.
374: 733-736.
Chung, J. D., Sinskey, A. J. and Stephanopoulos, G. N. 1998. Growth Factor and Bcl-2
Mediated Survival During Abortive Proliferation of Hybridoma Cell Line. Biotech.
Bioeng. 57: 164-171.
Chung, J. D., Zabel, C., Sinskey, A. J. and Stephanopoulos, G. N. 1997. Extension of
Sp2/0 Hybridoma Cell Viability Through Interleukin-6 Supplementation. Biotech.
Bioeng. 55: 439-446.
Clarke, A. R., Purdie, C. A., Harrison, D. J., Morris, R. G., Bird, C. C., Hooper, M. L.
and Wyllie, A. H. 1993. Thymocyte apoptosis induced by p53-dependent and
independent pathways. Nature. 362: 849-852.
Clem, R., Fechheimer, M. and Miller, L. K. 1991. Prevention of Apoptosis by a
Baculovirus Gene During Infection of Insect Cells. Science. 254: 1388-90.
196
Clem, R. J., Hardwick, J. M. and Miller, L. K. 1996. Anti-apoptotic genes of
baculoviruses. Cell Death and Differentiaion. 3: 9-16.
Cohen, G. M. 1997. Caspases: the executioners of apoptosis. Biochem. J. 326: 1-16.
Cohen, J. J. 1991. Programmed Cell Death in the Immune System. Adv. Immunol. 50:
55-85.
Cuende, E., AlEs-Martinez, J. E., Ding, L., GUnzalez-GarcIa, M., MartInez, A. C. and
Nunez, G. 1993. Programmed Cell Death by bcl-2-dependent and independent
mechanisms in B Lymphoma Cells. EMBO J. 12: 1555-1560.
Draetta, G. F. 1994. Mammalian GI cyclins. Curr. Op. Cell Bio. 6: 842-46.
Drapeau, D., Luan, Y.-T., Popoloski, J. A., Richards, D. T., Cohen, D. C., Sinacore, M.
S. and Adamson, S. R. 1994. Extracellular insulin degrading activity creates instability
in a CHO-based batch-refeed continuous process. Cytotechnology. 15: 103-109.
Duan, H. and Dixit, V. M. 1997. RAIDD is a new 'death' adaptor molecule. Nature. 385:
86-9.
El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B. and Parsons, R. 1993. WAF1,
a Potential Mediator of p53 Tumor Suppression. Cell. 75: 817.
Elledge, S. J. 1996. Cell Cycle Checkpoints: Preventing an Identity Crisis. Science. 274:
1664-72.
Ellis, R. E., Yuan, J. Y. and Horvitz, H. R. 1991. Mechanisms and functions of cell death.
Annu. Rev. Cell Biol. 7: 663-698.
Enari, M., Talanian, R. V., Wong, W. W. and Nagata, S. 1996. Sequential activation of
ICE-like and CPP32-like proteases during Fas-mediated apoptosis. Nature. 380: 723726.
Fanidi, A., Harrington, E. A. and I., E. G. 1992. Cooperative interaction between c-Myc
and bcl-2 oncogenes. Nature. 359: 554-556.
Fesus, L., Davies, P. J. and Piacentini, M. 1991. Apoptosis: molecular mechanisms in
programmed cell death. Eur. J. Cell Biol. 56: 170-177.
Franek, F. 1995. Starvation-Induced Programmed Death of Hybridoma Cells: Prevention
by Amino Acid Mixtures. Biotech. Bioeng. 45: 86-90.
Franek, F. and Chladkova-Sramkova, K. 1995. Apoptosis and nutrition: Involvement of
amino acid transport system in repression of hybridoma cell death. Cytotechnology.
18: 113-117.
197
Franek, F. and Dolnikova, J. 1991. Hybridoma growth and monoclonal antibody
production in iron-rich protein-free medium: Effect of nutrient concentration.
Cytotechnology. 7: 33-38.
Fraser, A. and Evan, G. 1996. A license to Kill. Cell. 85: 781-.
Fukuda, Y., Tsuji, T., Fujita, T., Yamamoto, A. and Muranishi, S. 1995. Susceptibility of
Insulin to Proteolysis in Rat Lung Homogenate and Its Protection from Proteolysis
by Various Protease Inhibitors. Biol. Pharma. Bull. 18: 891-94.
Fussenegger, M., Mazur, X. and Bailey, J. E. 1997. A Novel Cytostatic Process
Enhances the Productivity of Chinese Hamster Ovary Cells. Biotechnol. Bioeng. 55:
927-39.
Fussenegger, M., Mazur, X. and Bailey, J. E. 1998. pTRIDENT, a Novel Vector Family
for Tricistrionic Gene Expression in Mammalian Cells. Biotechnol. Bioeng. 57: 1-10.
Goswami, J., Sinskey, A. J., Steller, H., Stephanopoulos, G. N. and Wang, D. I. C. 1998.
Apoptosis in Batch Cultures of Chinese Hamster Ovary Cells. Biotech. and Bioeng.
in press:
Gottschalk, A. R., Boise, L. H., Thompson, C. B. and Quintens, J. 1994. Identification
of Immunosuppressant-induced apoptosis in a murine B-cell line and its prevention by
bcl-x but not bcl-2. Proc. Natl. Acad. Sci. USA. 91: 7350-7354.
Gu, M. B., Todd, P. and Kompala, D. S. 1993. Foreign gene expression (betagalactosidase) during the cell-cycle phases in recombinant CHO cells. Biotech. Bioeng.
42: 1113-1123.
Gu, M. B., Todd, P. and Kompala, D. S. 1994. Analysis of Foreign Protein
Overproduction in Recombinant CHO Cells. Annals of N.Y. Acad. Sci. 721: 194-207.
Han, J., Sabbatini, P., Perez, D., Rao, L., Modha, D. and White, E. 1996. The E1B 19K
protein blocks apoptosis by interacting with and inhibiting the p53-inducible and
death-promoting Bax protein. Genes and Development. 10: 461-477.
Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K. and Elledge, S. J. 1993. The p21
CDK-Interacting Protein Cipl Is a Potent Inhibitor of Gl Cyclin-Dependent Kinases.
Cell. 75: 805.
Hausdorff, S. F., Fingar, D. C. and Birnbaum, M. J. 1996. Signalling Pathways Mediating
Insulin-Activated Glucose Transport. Cell & Dev. Biol. 7: 239-47.
198
Hawkins, C. J., Uren, A. G., Hacker, G., Medcalf, R. L. and Vaux, D. L. 1996. Inhibition
of Interleukin 1 beta-converting enzyme-mediated apoptosis of mammalian cells by
baculovirus IAP. Proc. Natl. Acad. Sci. USA. 93: 13786-90.
Hengartner, M. O. 1996. Programmed cell death in invertebrates. Curr. Opin. Genet. Dev.
6: 34-38.
Hengartner, M. 0. and Horvitz, H. R. 1994. Activation of C. elegans cell death protein
CED-9 by an amino-acid substitution in a domain conserved in Bcl-2. Nature. 369:
318-320.
Hershberger, P. A., LaCount, D. J. and Friesen, P. D. 1994. The Apoptotic Suppressor
P35 is Required Early During Baculovirus Replication and is Targeted to the Cytosol
of Infected Cells. J. Virol. 68: 3467-77.
Hockenberry, D. M., Oltvai, Z. N., Yin, X., Milliman, C. L. and Korsmeyer, S. J. 1993.
Bcl-2 Functions in an Antioxidant Pathway to Prevent Apoptosis. Cell. 75: 241-251.
Huang, D. C. S., O'Reilly, L. A., Strasser, A. and Cory, S. 1997. The anti-apoptosis
function of Bcl-2 can be genetically separated from its inhibitory effect on cell cycle
entry. EMBO J. 16: 4628-4638.
Hunter, J. J., Bond, B. L. and Parslow, T. G. 1996. Functional dissection of the human
Bcl2 protein: Sequence requirements for the inhibition of apoptosis. Mol. Cell . Biol.
16: 877-883.
Hunter, J. J. and Parslow, T. G. 1996. A peptide sequence from Bax that converts Bcl-2
into an activator of apoptosis. J. Biol. Chem. 271: 8521-8524.
Itoh, Y., Ueda, H. and Suzuki, E. 1995. Overexpression of bcl-2, Apoptosis Suppressing
Gene: Prolonged Viable Culture Period of Hybridoma and Enhanced Antibody
Production. Biotech. and Bioeng. 48:
Jacobson, M. D., Burne, J. F. and Raff, M. C. 1994. Programmed cell death and Bcl-2
protection in the absence of a nucleus. EMBO J. 13: 1899-1910.
Kane, D. H., Sarafian, T. A., Anton, R., Hahn, H., Gralla, E. B., Valentine, J. S., Ord,
T. and Bredesen, D. E. 1993. Bcl-2 Inhibition of Neural Death: Decreased Generation
of Reactive Oxygen Species. Science. 262: 1274-1276.
Kane, D. J., Ord, T., Anton, R. and Bredesen, D. E. 1995. Expression of bcl-2 Inhibits
Necrotic Neural Cell Death. J. Neurosci. Res. 40: 269-275.
Kastan, M. B., Zhan, Q., El-Deiry, W. S., Carrier, F., Jacks, T., Walsh, W. V.,
Plunkett, B. S., Vogelstein, B. and Fornace, A. J. J. 1992. A mammalian cell cycle
199
checkpoint utilizing p53 and GADD45 is defective in atazia-telangiectasia. Cell. 71:
587-597.
Kaufman, R. J. and Sharp, P. A. 1982. Amplification and Expression of Sequences
Cotransfected with a Modular Dihydrofolate Reductase Complementary DNA Gene.
Journal of Molecular Biology. 159: 601-621.
Kerr, J. F. R., Wyllie, A. H. and Currie, A. H. 1972. Apoptosis, a basic biological
phenomenon with wider implications in tissue kinetics. Br. J. Cancer. 26: 239-45.
Kiefer, M. C., Brauer, M. J., Powers, V. C., Wu, J. J., Umansky, S. R., Tomei, L. D. and
Barr, P. J. 1995. Modulation of apoptosis by the widely distributed Bcl-2 homologue
Bak. Nature. 374: 736-739.
King, R. W., Deshaies, R. J., Peters, J. M. and Kirschner, M. W. 1996. How Proteolysis
Drives the Cell Cycle. Science. 274: 1652-59.
Kluck, R. M., Bossy-Wetzel, E., Green, D. R. and Newmeyer, D. D. 1997. The Release
of Cytochrome c from Mitochondria: A Primary Site for Bcl-2 Regulation of
Apoptosis. Science. 275: 1132-36.
Ko, L. J. and Prives, C. 1996. p53: puzzle and paradigm. Genes Dev. 10: 1054-1072.
Koff, A., Ohtsuki, M., Polyak, K., Roberts, J. M. and Massagu6, J. 1993. Negative
Regulation of G1 in Mammalian Cells: Inhibition of Cyclin E-Dependent Kinase by
TGF-B. Science. 260: 536-9.
Kozopas, K. M., Yang, T., Buchan, H. L., Zhou, P. and Craig, R. W. 1993. MCL1, a gene
expressed in programmed myeloid cell differentiation, has sequence similarity to
BCL2. Proc. Natl. Acad. Sci. 90:
Krajewski, S., Tanaka, S., Takayama, S., Schibler, M. J., Fenton, W. and Reed, J. C.
1993. Investigations of the subcellular distribution of bcl-2 oncoprotein: residence in
the nuclear envelope, endoplasmic reticulum, and outer mitochondrial membranes.
Cancer Res. 53: 4701-4704.
Kroemer, G. 1997. The proto-oncogene Bcl-2 and its role in regulating apoptosis. Nat.
Med. 3: 614-620.
Kroemer, G., Zamzami, N. and Susin, S. A. 1997. Mitochondrial control of apoptosis.
Immunol. Today. 18: 44-51.
Lam, M., Dubyak, G., Chen, L., Nufiez, G., Miesfeld, R. L. and Distelhorst, C. W.
1994. Evidence that bcl-2 represses apoptosis by regulating endoplasmic reticulumassociated Ca + + fluxes. Proc. Natl. Acad. Sci. USA. 91: 6569-6573.
200
Lane, D. P. 1992. p53, guardian of the genome. Nature. 358: 15-16.
Lao, M.-S., Toth, D., Dannel, G. and Schalla, C. 1996. Degradative activities in a
recombinant chinese hamster ovary cell culture. Cytotechnology. 22: 43-52.
Lazebnik, Y. A., Kaufmann, S. H., Desnoyers, S., Poirier, G. G. and Earnshaw, W. C.
1994. Cleavage of poly(ADP-ribose) polymerase by a proteinase with properties like
ICE. Nature. 371: 346-347.
Lazebnik, Y. A., Takahashi, A., Moir, R. D., Goldman, R. D. and Poirier, G. G. 1995.
Studies of the lamin proteinase reveal multiple parallel biochemical pathways during
apoptotic execution. Proc. Natl. Acad. Sci. USA. 92: 9042-46.
Levine, A. J. 1997. p53, the Cellular Gatekeeper for Growth and Division. Cell. 88: 323331.
Levine, B., Huang, Q., Isaacs, J. T., Reed, J. C., Griffin, D. E. and Hardwick, J. M. 1993.
Conversion of lytic to persistent alphavirus infection by the bcl-2 cellular oncogene.
Nature. 361: 739-42.
Li, P., Nijhawan, D., Budihardjo, I., Srinivasula, S. M., Ahmad, M., Alnemri, E. S. and
Wang, X. 1997. Cytochrome c and dATP-Dependent Formation of Apaf-1/Caspase-9
Complex Initiates an Apoptotic Protease Cascade. Cell. 91: 479-489.
Lin, D., Shields, M. T., Ullrich, S. J., Appella, E. and Mercer, W. E. 1992. Growth arrest
induced by wild-type p53 protein blocks cells prior to or near the restriction point in
late G1 phase. Proc. Natl. Acad. Sci. USA. 89: 9210-4.
Lin, E. Y., Orlofsky, A., Berger, M. S. and Prystowsky, M. B. 1993. Characterization of
Al, a novel hemopoietic-specific early-response gene with sequence similarity to bcl2. J. Immunol. (IFB). 151: 1979-88.
Lincz, L. F. 1998. Deciphering the apoptotic pathway: All roads lead to death. Immunol.
Cell Biol. 76: 1-19.
Liston, P., Roy, N., Tamai, K., Lefebvre, C., Baird, S., Cherton-Horvat, G., Farahani, R.,
McLean, M., Ikeda, J. E., MacKenzie, A. and Korneluk, R. G. 1996. Suppression of
apoptosis in mammalian cells by NAIP and a related family of IAP genes. Nature.
379: 349-353.
Liu, X., Kim, C. N., Yang, J., Jemmerson, R. and Wang, X. 1996. Induction of Apoptotic
Program in Cell-Free Extracts: Requriement for dATP and Cytochrome c. Cell. 86:
147-157.
201
Lotem, J. and Sachs, L. 1993. Regulation by bcl-2, c-Myc, and p-53 of Susceptibility to
Induction of Apoptosis by Heat Shock and Cancer Chemotherapy Compounds in
Differentiation-component and -defective Myeloid Leukemic Cells. Cell Growth Diff.
4: 41-47.
Lowe, S. W., Ruley, H. E., Jacks, T. and Housman, D. E. 1993. p53-Dependent
Apoptosis Modulates the Cytotoxicity of Anticancer Agents. Cell. 74: 957-967.
Lowe, S. W., Schmitt, E. M., Smith, S. W., Osborne, B. A. and Jacks, T. 1993. p53 is
required for the radiation induced apoptosis in mouse thymocytes. Nature. 362: 847849.
Mandal, M., Maggirwar, S. B., Sharma, N., Kaufmann, S. H., Sun, S. C. and Kumar, R.
1996. Bcl-2 prevents CD95 (Fas/APO-l)-induced degradation of lamin B and
poly(ADP-ribose) polymerase and restores the NF-kappaB signaling pathway. J.
Biol. Chem. 271: 30354-9.
Mariani, B. D., Slate, D. L. and Schimke, R. T. 1981. S phase -specific synthesis of
dihydrofolate reductase in Chinese Hamster Ovary cells. Proc. Natl. Acad. Sci. USA.
78: 4985-4989.
Martin, S. J., Amarante-Mendes, G. P., Shi, L., Chuang, T., Casiano, C. A., O'Brien, G.
A., Fitzgerald, P., Tan, E. M., Bokoch, G. M., Greenberg, A. H. and Green, D. R.
1996. The cytotoxic cell protease granzyme B initiates apoptosis in a cell-free system
by proteolytic processing and activation of the ICE/CED-3 family protease, CPP32,
via a novel two-step mechanism. EMBO J. 15: 2407-16.
Martin, S. J. and Cotter, T. G. 1994. Apoptosis of Human Leukemia: Induction,
Morphology and Molecular Mechanisms. In: Apoptosis II: The Molecular Basis of
Apoptosis in Disease, Cold Spring Harbor Laboratory Press,
Martin, S. J. and Green, D. R. 1995. Protease Activation during Apoptosis: Death by a
Thousand Cuts? Cell. 82: 349-52.
Marx, J. 1993. Howp53 Suppresses Cell Growth. Science. 262: 1644-1645.
Mastrangelo, A. J. and Betenbaugh, M. J. 1998. Overcoming apoptosis: new methods for
improving protein-expression systems. TIBTECH. 16: 88-95.
Mastrangelo, A. J., Hardwick, J. M. and Betenbaugh, M. J. 1996. Bcl-2 inhibits
apoptosis and extends recombinant protein production in cells infected with Sindbis
viral vectors. Cytotechnology. 22: 169-178.
202
Matsuyama, S., Xu, Q., Velours, J. and Reed, J. C. 1998. The Mitochondrial FoF 1ATPase Proton Pump Is Required for Function of the Proapoptotic Protein Bax in
Yeast and Mammalian cells. Mol. Cell. 1: 327-336.
McConkey, D. J. 1996. Calcium-dependent, interleukin 1-beta-converting enzyme
inhibitor-insensitive degradation of lamin B 1 and DNA fragmentation in isolated
thymocyte nuclei. J. Biol. Chem. 271: 22398-406.
McGahon, A. J., Martin, S. J., Bisonette, R. P., Mahboubi, A., Yufang, S., Mogil, R. J.,
Nishioka, W. K. and Green, D. R. 1995. The End of the (Cell) Line: Methods for the
Study of Apoptosis in Vitro. In: L. M. Schwartz and B. Osborne, Methods in Cell
Biology, Academic Press, Inc., San Diego.
McKeon, F. 1991. Nuclear lamin proteins: Domains required for nuclear targeting,
assembly, and cell-cycle-regulated dynamics. Curr. Opin. Cell Biol. 3: 82-86.
Mercer, W. E., Shields, M. T., Amin, M., Sauve, G. J., Appella, E., Romano, J. W. and
Ullrich, S. J. 1990. Negative growth regulation in a glioblastoma tumor cell line that
conditionally expresses human wild-type p53. Proc. Natl. Acad. Sci. USA. 87: 616670.
Mercille, S. and Massie, B. 1994. Induction of Apoptosis in Nutrient-Deprived
Cultures of Hybridoma and Myeloma Cells. Biotech. and Bioeng. 44: 11401154.
Miura, M., Zhu, H., Rotello, R., Hartwieg, E. A. and Yuan, J. 1993. Induction of
Apoptosis in Fibroblasts by IL- lb-Converting Enzyme, a Mammalian Homolog of the
C. elegans Cell Death Gene ced-3. Cell. 75: 653-660.
Miyashita, T. and Reed, J. C. 1995. Tumor Suppressor p53 Is a Direct Transcriptional
Activator of the Human bax Gene. Cell. 80: 293-299.
Miyazaki, T., Liu, Z. J., Kawahara, A., Minami, Y., Yamada, K., Tsujimoto, Y.,
Barsoumian, E. L., Perlmutter, R. M. and Taniguchi, T. 1995. Three Distinct IL-2
Signaling Pathways Mediated by bcl-2, c-myc, and Ick Cooperate in Hematopoietic
Cell Proliferation. Cell. 81: 223-31.
Moore, A., Donahue, C. J., Hooley, J., Stocks, D. L., Bauer, K. D. and Mather, J. P.
1995. Apoptosis in CHO cell batch cultures: examination by flow cytometry.
Cytotechnology. 17: 1-11.
Moore, A., Mercer, J., Dutina, G., Donahue, C. J., Bauer, K. D., Mather, J. P.,
Etcheverry, T. and Ryll, T. 1997. Effects of temperature shift on cell cycle, apoptosis
and nucleotide pools in CHO cell batch cultures. Cytotechnology. 23: 47-54.
203
Morishita, M., Morishita, I., Takayama, K., Machida, Y. and Nagai, T. 1992. Novel oral
microspheres of insulin with protease inhibitor protecting from enzymatic
degradation. Int. J. of Pharmaceutics. 78: 1-7.
Mosser, D. D. and Massie, B. 1994. Genetically Engineering Mammalian Cell Lines for
Increased Viability and Productivity. Biotech. Adv. 12: 253-277.
Muchmore, S. W., Sattler, M., Liang, H., Meadows, R. P., Harlan, J. E., Yoon, H. S.,
Nettesheim, D., Chang, B. S., Thompson, C. B., Wong, S. L., Ng, S. L. and Fesik, S.
W. 1996. X-ray and NMR structure of human Bcl-xL, an inhibitor of programmed cell
death. Nature. 381: 335-41.
Mummenbrauer, T., Janus, F., Muller, B., Wiesmuller, L., Deppert, W. and Grosse, F.
1996. p53 Protein Exhibits 3'-to-5' Exonuclease Activity. Cell. 85: 1089-1099.
Murakami, H. 1989. Serum-free media used for cultivation of hybridomas. In:
Monoclonal antibodies: Production and applications, Alan R. Liss, New York.
Murray, K., Ang, C., Gull, K., Hickman, J. A. and Dickson, A. J. 1996. NSO Myeloma
Cell Death: Influence ofbcl-2 Overexpression. Biotech. Bioeng. 51: 298-304.
Nagata, S. 1996. Apoptosis: Telling cells their time is up. Curr. Biol. 6: 1241-43.
Naumovski, L. and Cleary, M. L. 1996. The p53-binding protein 53BP2 also interacts
with Bcl2 and impedes cell cycle progression at G2/M. Mol. Cell. Biol. 16: 3884-92.
Newmeyer, D. D., Farschon, D. M. and Reed, J. C. 1994. Cell-Free Apoptosis in
Xenopus Egg Extracts: Inhibition by Bcl-2 and Requirement for an Organelle Fraction
Enriched in Mitochondria. Cell. 79: 353-364.
Nicholson, D. 1996. ICE/CED3-like Proteases as Therapeutic Targets for the Control of
Inappropriate Apoptosis. Nature Biotech. 14: 297-301.
Nyberg, G. B. 1998. Glycosylation Site Occupancy Heterogeneity in Chinese Hamster
Ovary Cell Culture. Massachusetts Institute of Technology.
Oberhammer, F. A., Hochegger, K., Fr6shcl, G., Tiefenbacher, R. and Pavelka, M. 1994.
Chromatin Condensation during Apoptosis Is Accompanied by Degradation of Lamin
A+B, without Enhanced Activation of cdc2 Kinase. J. Biol. Chem. 126: 827-37.
Ohtsubo, M. and Roberts, J. M. 1993. Cyclin-Dependent Regulation of G 1 in
Mammalian Fibroblasts. Science. 259: 1908-11.
204
Ohtsubo, M., Theodoras, A. M., Schumacher, J., Roberts, J. M. and Pagano, M. 1995.
Human Cyclin E, a Nuclear Protein Essential for the G 1 -to-S Phase Transition. Mol.
Cell. Biol. 15: 2612-24.
Okamoto, K. and Beach, D. 1994. Cyclin G is a transcriptional target of the p53 tumor
suppressor protein. EMBO J. 13: 4816-4822.
Olsen, C. W., Kehren, J. C., Dybdahl-Sissoko, N. R. and Hinshaw, V. S. 1996. bcl-2
alters influenza virus yield, spread, and hemagglutinin glycosylation. J. Virol. 70: 663666.
Oltvai, Z. N., Milliman, C. L. and Korsmeyer, S. J. 1993. Bcl-2 Heterodimerizes In Vivo
with a Conserved Homolog, Bax, That Accelerates Programmed Cell Death. Cell. 74:
609-619.
Oren, M. and Prives, C. 1996. p53: upstream, downstream and off stream. Biochim.
Biophys. Acta. 1288: R13-R19.
Orrenius, S., Burkitt, M. J., Kass, G. E. N., Dypbukt, J. M. and Nicotera, P. 1992.
Calcium ions and oxidative cell injury. Ann. Neurol. 32: S33-S42.
Orth, K., Chinnaiyan, A. M., Garg, M., Froelich, C. J. and Dixit, V. M. 1996. The CED3/ICE-like Protease Mch2 Is Activated during Apoptosis and Cleaves the Death
Substrate Lamin A. J. Biol. Chem. 271: 16443-46.
Orth, K. and Dixit, V. M. 1997. Bik and Bak Induce Apoptosis Downstream of CrmA
but Upstream of Inhibitor of Apoptosis. J. Biol. Chem. 272: 8841-8844.
Pan, G., Humke, E. W. and Dixit, V. M. 1998. Activation of caspases triggered by
cytochrome c in vitro. FEBS Lett. 426: 151-154.
Pan, G., O'Rourke, K. and Dixit, V. M. 1998. Caspase-9, Bcl-XL, and Apaf-1 Form a
Ternary Complex. J. Biol. Chem. 273: 5841-5845.
Patterson, M. K. 1979. Measurement of growth and viability of cells in culture. Methods
Enzymol. 58: 141-152.
Perreault, J. and Lemieux, R. 1994. Essential role of optimal protein synthesis in
preventing the apoptotic death of B cell hybridomas. Cytotechnology. 13: 99-105.
Pessin, J. E. and Bell, G. I. 1992. Mammalian Facilitative Glucose Transporter Family:
Structure and Molecular Regulation. Annu. Rev. Physiol. 54: 911-30.
205
Petit, P. X., Susin, S.-A., Zamzami, N., Mignotte, B. and Kroemer, G. 1996.
Mitochondria and programmed cell death: back to the future. FEBS Letters. 396: 713.
Pimentel, E. 1994. Insulin. In: Handbook of Growth Factors, 1. CRC Press,
Rabizadeh, S., LaCount, D. J., Friesen, P. D. and Bredesen, D. E. 1993. Expression of the
Baculovirus p35 Gene Inhibits Mammalian Neural Cell Death. J. Neurochem. 61:
2318-21.
Raff, M. C., Barres, B. A., Burne, J. F., Coles, H. S., Ishizaki, Y. and Jacobson, M. D.
1993. Programmed cell death and the control of cell survival: lessons from the nervous
system. Science. 262: 695-700.
Reed, J. C. 1994. Bcl-2 and the Regulation of Programmed Cell Death. The Journal of Cell
Biology. 124: 1-6.
Reed, J. C. 1997. Cytochrome c: Can't Live with It - Can't Live without It. Cell. 91: 559562.
Reed, J. C. 1997. Double Identity for proteins of the Bcl-2 family. Nature. 387: 773-776.
Renner, W. A., Lee, K. H., Hatzimanikatis, V., Bailey, J. E. and Eppenberger, H. M.
1995. Recombinant Cyclin E Expression Activates Proliferation and Obviates Surface
Attachment of Chinese Hamster Ovary (CHO) Cells in Protein-Free Medium.
Biotech. Bioeng. 47: 476-.
Salvesen, G. S. and Dixit, V. M. 1997. Caspases: Intracellular Signaling by Proteolysis.
Cell. 91: 443-446.
Sambrook, J., Fritsch, E. F. and Maniatis, T. 1989. Molecular Cloning. 2nd. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor.
Scahill, S. J., Devos, R., van der Heyden, J. and Fiers, W. 1983. Expression and
characterization of the product of a human immune interferon cDNA gene in Chines
hamster ovary cells. Proc. Natl. Acad. Sci. USA. 80: 4654-58.
Sen, S. and D'lncalci, M. 1992. Apoptosis. Biochemical events and relevance to cancer
chemotherapy. FEBS Lett. 307: 122-127.
Sherr, C. J. 1994. G1 Phase Progression: Cycling on Cue. Cell. 79: 551-555.
Sherr, C. J. 1996. Cancer Cell Cycles. Science. 274: 1672-1677.
Shibasaki, F., Kondo, E., Akagi, T. and McKeon, F. 1997. Suppression of signalling
through NF-AT by interactions between calcineurin and Bcl-2. Nature. 386: 728-731.
206
Shimizu, S., Eguchi, Y., Kamiike, W., Itoh, Y., Hasegawa, J., Yamabe, K., Otsuki, Y.,
Matsuda, H. and Tsujimoto, Y. 1996. Induction of apoptosis as well as necrosis by
hypoxia and predominant prevention of apoptosis by Bcl-2 and Bcl-XL. Cancer Res.
56: 2161-6.
Simpson, N. J., Milner, A. E. and Al-Rubeai, M. 1997. Prevention of Hybridoma Cell
Death by bcl-2 During Suboptimal Culture Conditions. Biotech. Bioeng. 54: 1-16.
Singh, R. P., Al-Rubeai, M., Gregory, C. D. and Emery, A. N. 1994. Cell Death in
Bioreactors: A role for Apoptosis. Biotech. and Bioeng. 44: 720-726.
Singh, R. P., Emery, A. N. and Al-Rubeai, M. 1996. Enhancement of Survivability of
Mammalian Cells by Overexpression of the Apoptosis-Suppressor Gene bcl-2.
Biotech. Bioeng. 52: 166-175.
Singh, R. P., Finka, G., Emery, A. N. and Al-Rubeai, M. 1997. Apoptosis and its control
in cell culture systems. Cytotechnology. 23: 87-93.
Stasser, A., Harris, A. W., Jacks, T. and Cory, S. 1994. DNA Damage Can Induce
Apoptosis in Proliferating Lymphoid Cells via p53-Independent Mechanisms
Inhibitable by Bcl-2. Cell. 79: 329-339.
Subler, M. A., Martin, D. W. and Deb, S. 1992. Inhibition of viral and cellular promoters
by human wild-type p53. J. Virol. 66: 4757-4762.
Susin, S. A., Zamzami, N., Castedo, M., Hirsch, T., Marchetti, P., Macho, A., Daugas,
E., Geuskens, M. and Kroemer, G. 1996. Bcl-2 Inhibits the Mitochondrial Release of
an Apoptogenic Protease. J. Exp. Med. 184: 1331-41.
Suzuki, E., Terada, S., Ueda, H., Fujita, T., Komatsu, T., Takayama, S. and Reed, J. C.
1997. Establishing apoptosis resistant cell lines for improving protein productivity of
cell culture. Cytotechnology. 23: 55-59.
Takahashi, A. and Earnshaw, W. C. 1996. ICE-related proteases in apoptosis. Curr. Op.
Gen. Dev. 6: 50-55.
Takayama, S., Sato, T., Krajewski, S., Kochel, K., Irie, S., Millan, J. A. and Reed, J. C.
1995. Cloning and Functional Analysis of BAG-1: A Novel Bcl-2-Binding Protein
with Anti-Cell Death Activity. Cell. 80: 279-284.
Tanke, J. J., Van der Linden, P. W. G. and Langerak, J. 1982. Alternative Fluorochromes
to Ethidium Bromide for Automated Read Out of Cytotoxicity Tests. J. Immunol.
Methods. 52: 91-96.
207
Terada, S., Itoh, Y., Ueda, H. and Suzuki, E. 1997. Characterization and fed-batch culture
of hybridoma overexpressing apoptosis suppressing gene bcl-2. Cytotechnology. 24:
135-141.
Thompson, C. B. 1995. Apoptosis in the Pathogenesis and Treatment of Disease.
Science. 267: 1456-1462.
Thomberry, N. A. and Molineaux, S. M. 1995. Interleukin-1-0 converting enzyme: A
novel cysteine protease required for IL-1 3 production and implicated in programmed
cell death. Protein Sci. 4: 3-12.
Tozaki, H., Emi, Y., Horisaka, E., Fujita, T., Yamamoto, A. and Muranishi, S. 1997.
Degradation of Insulin and Calcitonin and Their Protection by Various Protease
Inhibitors in Rat Caecal Contents: Implications in Peptide Delivery to the Colon. J.
Pharm. Pharmacol. 49: 164-168.
Trost, L. C. and Lemasters, J. J. 1994. A Cytotoxicity Assay for Tumor Necrosis Factor
Emplying a Multiwell Flourescence Scanner. Anal. Biochem. 220: 149-153.
Trump, B. F. and Mergner, W. J. 1974. Cell Injury. Second. Academic Press, New York.
Vander Heiden, M. G., Chandel, N. S., Williamson, E. K., Schumacker, P. T. and
Thompson, C. B. 1997. Bcl-xL Regulates the Membrane Potential and Volume
Homeostasis of Mitochondria. Cell. 91: 627-637.
Vaux, D. L., Cory, S. and Adams, J. M. 1988. Bcl-2 gene promotes haemopoietic cell
survival and cooperates with c-myc to immortalize pre-B cells. Nature. 335: 440-442.
Vaux, D. L., Haecker, G. and Strasser, A. S. 1994. An Evolutionary Perspective on
Apoptosis. Cell. 76: 777-779.
Vaux, D. L. and Strasser, A. 1996. The molecular biology of apoptosis. Proc. Natl. Acad.
Sci USA. 93: 2239-2244.
Vollenweider, I. and Groscurth, P. 1992. Comparison of four DNA staining fluorescence
dyes for measuring cell proliferation of lymphokine-activated killer (LAK) cells. J.
Immunol. Methods. 149: 133-135.
Vomastek, T. and Franek, F. 1993. Kinetic of development of spontaneous apoptosis in B
cell hybridoma cultures. Immunol. Letters. 35: 19-24.
Wallach, D., Boldin, M., Varfolomeev, E., Beyaert, R., Vandenabeele, P. and Fiers, W.
1997. Cell death induction by receptors of the TNF family: towards a molecular
understanding. FEBS Lett. 410: 96-106.
208
Wang, X., Pai, J., Wiedenfeld, E. A., Medina, J. C., Slaughter, C. A., Glodstein, J. L. and
Brown, M. S. 1995. Purification of an Interleukin-l-3 Converting enzyme-related
Cysteine Protease that Cleaves Sterol Regulatory Element-binding Proteins between
the Leucine Zipper and Transmembrane Domains. J. Biol. Chem. 270: 18044-50.
Wang, X., Zelenski, N. G., Yang, J., Sakai, J., Brown, M. S. and Goldstein, J. L. 1996.
Cleavage of sterol regulatory element binding proteins (SREBPs) by CPP32 during
apoptosis. EMBO J. 15: 1012-20.
Wilson, K. P., Black, J. A., Thomson, J. A., Kim, E. E., Griffith, J. P., Navia, M. A.,
Murcko, M. A., Chambers, S. P., Aldape, R. A., Raybuck, S. A. and Livingston, D. J.
1994. Structure and mechanism of interleukin-1 beta converting enzyme. Nature. 370:
270-5.
Wyllie, A. H., Kerr, J. F. R. and Currie, A. R. 1980. Cell Death: The Significance of
Apoptosis. Int. Rev. Cytol. 68: 251-306.
Xiang, J., Chao, D. T. and Korsmeyer, S. J. 1996. BAX-induced cell death may not
require interleukin 1-b-converting enzyme-like proteases. Proc. Natl. Acad. Sci. USA.
93: 14559-14563.
Xie, L., Nyberg, G., Gu, X., Li, H., M611born, F. and Wang, D. I. C. 1997. GammaInterferon Production and Quality in Stoichiometric Fed-Batch Cultures of Chinese
Hamster Ovary (CHO) Cells under Serum-Free Conditions. Biotech. and Bioeng. 56:
577-82.
Xie, L. and Wang, D. I. C. 1994a. Fed-batch Cultivation of Animal Cells Using Different
Medium Design and Feeding Strategies. Biotech. Bioeng. 43: 1175.
Xie, L. and Wang, D. I. C. 1994b. Stoichiometric Analysis of Animal Cell Growth and its
Application in Medium Design. Biotech. and Bioeng. 43:
Xu, Q. and Reed, J. C. 1998. Bax Inhibitor-1, a Mammalian Apoptosis Suppressor
Identified by Functional Screening in Yeast. Mol. Cell. 1: 337-346.
Xue, D. and Horvitz, H. R. 1995. Inhibition of Caenorhabditis elegans cell death protease
CED-3 by a CED-3 cleavage site in baculovirus p35 protein. Nature. 377: 248-251.
Yamamoto, A., Taniguchi, T., Rikyuu, K., Tsuji, T., Fujita, T., Murakami, M. and
Muranishi, S. 1994. Effects of Varioius Protease Inhibitors on the Intestinal
Absorption and Degradation of Insulin in Rats. Pharma. Res. 11: 1496-1500.
Yang, E. and Korsmeyer, S. J. 1996. Molecular Thanatopsis: A discourse on the Bcl-2
family and cell death. Blood. 88: 386-401.
209
Yang, E., Zha, J., Jockel, J., Boise, L. H., Thompson, C. B. and Korsmeyer, S. J. 1995.
Bad, a Heterodimeric Partner for Bcl-xL and Bcl-2, Displaces and Promoter Cell
Death. Cell. 80: 285-291.
Yang, J., Xuesong, L., Bhalla, K., Kim, C. N., Ibrado, A. M., Cai, J., Peng, T., Jones, D.
P. and Wang, X. 1997. Prevention of Apoptosis by Bcl-2: Release of Cytochrome c
from Mitochondria Blocked. Science. 275: 1129-32.
Zamzami, N., Susin, S. A., Marchetti, P., Hirsch, T., Gomez-Monterrey, I., Castedo, M.
and Kroemer, G. 1996. Mitochondrial control of nuclear apoptosis. J. Exp. Med. 183:
1533-44.
Zha, H., Aime-Sempre, C., Sato, T. and Reed, J. C. 1996. Proapoptotic protein Bax
heterodimerizes with Bcl-2 and homodimerizes with Bax via a novel domain (BH3)
distinct from BH1 and BH2. J. Biol. Chem. 271: 7440-7444.
Zha, H., Fisk, H. A., Yaffe, M. P., Mahajan, N., Herman, B. and Reed, J. C. 1996.
Structure-function comparisons of the proapoptotic protein Bax in yeast and
mammalian cells. Mol. Cell. Biol. 16: 6494-6508.
Zha, J. P., Harada, H., Yang, E., Jockel, J. and Korsmeyer, S. J. 1996. Serine
phosphorylation of death agonist BAD in response to survival factor results in
binding to 14-3-3 not Bcl-xL. Cell. 87: 619-628.
Zhou, Q. and Salvesen, G. S. 1997. Activation of pro-caspase-7 by serine proteases
includes a non-canonical specificity. Biochem. J. 324: 361-364.
Zhou, W. and Hu, W.-S. 1995. Effect of Insulin on a Serum-Free Hybridoma Culture.
Biotech. Bioeng. 47: 181-185.
210