AN ABSTRACT OF THE THESIS OF
Deryk Thomas Loo
Biochemistry and Biophysics
for the degree of
presented on
Doctor of Philosophy i n
July 19. 1991
Title: Nonsenescent Serum-Free Mouse Proastroblast Cells: Extended Culture. Growth
Responses In Vitro. and Application to the Culture of Human Embryonic Astrocytes.
Abstract
approved:4Redacted for Privacy
David W. Barnes
Mouse embryo cells cultured in vitro in serum-supplemented media undergo growth
crisis, resulting in the loss of genomically normal cells prior to the appearance of
established, aneuploid cell lines. I used the technique of serum-free cell culture to develop
a serum-free mouse embryo (SFME) cell line in which serum was replaced by a set of
defined supplements. SFME cells, cultured in a nutrient medium supplemented with
insulin, transferrin, epidermal growth factor (EGF), high-density lipoprotein (HDL), and
fibronectin, have maintained a diploid karyotype with no detectable chromosomal
abnormalities for more than 200 generations. The cells did not undergo growth crisis and
remain in culture today. SFME cells were dependent on EGF for survival and were
reversibly growth inhibited by serum or platelet-free plasma. Treatment of SFME cells
with serum or transforming growth factor beta led to the appearance of glial fibrillary acid
protein (GFAP), a specific marker for astrocytes, identifying SFME cells as proastroblasts.
Following the derivation of SFME cells my research focussed on (1) defining more
precisely the growth response of SFME cells to medium supplements, (2) investigating the
relationship between the nonsenescent nature of SFME cells and their responses to serum
and EG1., and (3) applying the serum-free cell culture methods to the multipassage culture
of human embryonic astrocytes.
SFME cells in serum-containing medium arrested in the G1 phase of the cell cycle
with greatly reduced DNA replication activity. A portion of the inhibitory activity of serum
was extracted by charcoal, a procedure that removed steroid and thyroid hormones.
However, the effect of serum on untransformed SFME cells could not be prevented by
addition of antiglucocorticoid, and ras-transformed clones of SFME cells, which are not
growth inhibited by serum, retained inhibitory responses to glucocorticoid and thyroid
hormone T3. These results suggest that glucocorticoid or thyroid hormones may contribute
to the inhibitory activity of serum on SFME cells, but additional factors are involved.
SFME cell death resulting from EGF deprivation exhibited characteristics associated
with apoptosis or programmed cell death. Ultrastructural analysis showed cells became
small and vacuolated, with pyknotic nuclei. The cultures contained almost exclusively G1-
phase cells. Chromatin exhibited a pattern of degradation into oligonucleosome-length
fragments generating a regularly spaced ladder.
I applied the serum-free approach used to derive SFME cells to the multipassage
culture of human embryonic astrocytes.
Cells were cultured in nutrient medium
supplemented with insulin, transferrin, EGF, HDL, fibronectin, basic fibroblast growth
factor and heparin. Cultures were maintained for a maximum of 70 population doublings
before proliferation ceased. The cells synthesized GFAP.
Nonsenescent Serum-Free Mouse Proastroblasts: Extended Culture,
Growth Responses In Vitro, and Application to the Culture
of Human Embryonic Astrocytes
by
Deryk Thomas Loo
A THESIS
Submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Doctor of Philosophy
Completed July 19, 1991
Commencement June 1992
APPROVED:
Redacted for Privacy
Professor of Biochemistry and Biophysics in charge of major
Redacted for Privacy
Head of Department of Biochemistry and Biophysics
Redacted for Privacy
Dean of GraduKte School
Date thesis presented
Typed by
July. 19, 1991
Deryk Thomas Loo
DEDICATION AND ACKNOWLEDGMENTS
The path I followed to the degree of doctor of philosophy has been an enormous
learning experience, one I will never forget. I am thankful for all the lessons I have been
taught, both inside and outside of the laboratory. Without the love and support of my
parents, Eddie and Diana, and my wife, Jenefer, this thesis would not have been possible.
I dedicate my thesis to my parents and my wife. Thank you. I would also like to thank my
sisters, Megan and Kelly, for tolerating their older brother, and my Godmother, Elaine
Currie, for her support throughout my life.
The research presented in my thesis resulted from a collaboration among the many
members of the laboratory. I would like to thank everyone in the lab, especially Cathy
Rawson, for their help and friendship. Finally, I would like to thank David Barnes for his
financial support and for his unique six year course in molecular and cellular biology.
TABLE OF CONTENTS
CHAPTER 1
1
Introduction and Review of Literature
INTRODUCTION
1
REVIEW OF LITERATURE
3
Animal cell culture
3
Brain cell diversity and mammalian cell culture
10
Cell death
13
Orientation to format
17
CHAPTER 2
18
Extended Culture of Mouse Embryo Cells Without Senescence:
Inhibition by Serum
SUMMARY
19
INTRODUCTION
20
METHODS AND RESULTS
21
DISCUSSION
25
ACKNOWLEDGMENTS
26
CHAPTER 3
35
Serum-Free Mouse Embryo Cells: Growth Responses In Vitro
SUMMARY
36
INTRODUCTION
37
MATERIALS AND METHODS
38
Materials
38
Serum-free cell culture procedures
38
Serum-free culture medium
39
Experimental procedures
41
RESULTS
42
Effects of insulin and transferrin on SFME cells
42
Effects of EGF and FGFIHBGF on SFME cells
42
Effects of attachment proteins and HDL on SFME cells
44
DISCUSSION
46
ACKNOWLEDGMENTS
50
CHAPTER 4
60
Glucocorticoid and Thyroid Hormones Inhibit Proliferation
of Serum-Free Mouse Embryo (SFME) Cells
SUMMARY
61
INTRODUCTION
62
MATERIALS AND METHODS
64
Materials
64
SFME cell culture
64
Assay of cell growth
65
Seru-n treatments
66
RESULTS
67
Inhibition of SFME cells by glucocorticoid and thyroid hormone
67
Contribution of glucocorticoid and thyroid hormone to the
68
inhibitory activity of serum on SFME cells
DISCUSSION
70
ACKNOWLEDGMENTS
73
CHAPTER 5
84
Serial Passage of Embryonic Human Astrocytes In Serum-Free,
Hormone-Supplemented Medium
SUMMARY
85
INTRODUCTION
86
MATERIALS AND METHODS
88
Cell culture
88
Western blots
89
GFAP culture immunoassay
90
RESULTS
92
Serial passage culture of embryonic human astrocytes
92
Growth characteristics of embryonic human astrocytes in vitro
93
Expression of GFAP by embryonic human astrocytes in vitro
94
DISCUSSION
95
ACKNOWLEDGMENTS
99
CHAPTER 6
107
Conclusion and Future Prospectives
CONCLUSION
107
FUTURE PERSPECTIVES
112
BIBLIOGRAPHY
115
APPENDIX
1
129
Serum Inhibition of Proliferation of Serum-Free
Mouse Embryo Cells
SUMMARY
130
INTRODUCTION
131
MATERIALS AND METHODS
133
Cell culture
133
Flow cytometry
133
Assays
134
Karyotyping
134
RESULTS
135
Effects of serum on SFME cells
135
Variant SFME cells capable of growing in serum-containing medium
136
DISCUSSION
138
ACKNOWLEDGMENTS
140
APPENDIX 2
151
Death of Serum-Free Mouse Embryo Cells Caused
by Epidermal Growth Factor Deprivation
SUMMARY
152
INTRODUCTION
153
MATERIALS AND METHODS
155
Cell culture
155
Electron microscopy
155
DNA fragmentation analysis
156
Enzyme assay
157
Flow cytometry
157
Polysome profiles
158
Northern blot hybridization analysis
159
RESULTS
161
Ultrastructural changes associated with EGF deprivation
161
Biochemical and physiological changes associated with EGF deprivation
163
DISCUSSION
165
ACKNOWLEDGMENTS
168
LIST OF FIGURES
Eau
Figure
2.1
Growth of BALB/c mouse embryo cells upon successive
27
transfer in serum-containing or serum-free media
2.2
Karyotype of mouse embryo cell cultures
29
2.3
Effect of medium supplements on growth of mouse embryo
31
cells that were derived in serum-free medium
2.4
Growth inhibition of serum-free-derived mouse embryo
33
cells by plasma or serum
3.1
Growth response of SFME cells to insulin and IGF I
52
3.2
Growth response of SFME cells to transferrin and lactoferrin
53
3.3
Growth response of SFME cells to EGF and TGF alpha
54
3.4
Photomicrographs of SFME cells
55
3.5
Recovery of SFME cells from plating in growth-factor-deficient
56
medium
3.6
Growth response of SFME cells to acidic and basic FGF
57
in the absence of EGF
3.7
Growth response of SFME cells to FGF in the presence
58
and absence of EGF and heparin
3.8
Growth response of SFME cells to HDL
59
4.1
Partial removal of inhibitory activity of calf serum for
76
SFME cells by charcoal treatment
4.2
Steroid inhibition of SFME cell growth
77
4.3
Concentration dependence of hydrocortisone inhibition
78
of SFME cell growth
4.4
Concentration dependence of triiodothyronine inhibition
79
of SFME cell growth
4.5
Specificity of thyroid hormone inhibition of SFME cells
80
4.6
Concentration dependence of charcoal-treated calf serum for
81
glucocorticoid and T3 inhibition of SFME cell growth
4.7
Time course and reversibility of triiodothyronine and hydrocortisone
82
inhibition of SFME cell growth
4.8
Partial removal of inhibitory activity of calf serum by
83
anion exchange resin
5.1
Long-term culture of embryonic human astrocytes in serum-free
100
medium
5.2
Photomicrograph of embryonic human astrocytes in serum-free
101
medium.
5.3
Long-term culture of embryonic human astrocytes in medium
102
supplemented with low concentrations of calf serum
5.4
Growth of embryonic human astrocytes in serum-free and
103
serum-supplemented media
5.5
Response of embryonic human astrocytes to peptide growth factors
104
in serum-free medium
5.6
Expression of GFAP in embryonic human astrocytes
105
5.7
Western immunoblot of GFAP in extracts of embryonic human
106
astrocytes in serum-free medium
LIST OF TABLES
Table
ragg
3.1
Effect of attachment factors on SFME cell growth
51
4.1
Calf serum inhibition of SFME cells in the presence of
74
antiglucocorticoid RU38486
4.2
Glucocorticoid and triiodothyronine inhibition of rastransformed SFME cells
75
LIST OF APPENDIX FIGURES
Figure
Page
A1.1
Culture of SFME cells in serum-containing mediu
143
A1.2
Recovery of SFME cells from plating in serum-containing medium
144
A1.3 Flow cytometric analysis of SFME cells in serum-containing medium
A1.4
Selection of variant SFME cells capable of growing in serum-
145
146
containing medium
A1.5
Karyotype of variant SFME cells selected by culture in serum-
147
containing medium
A1.6
Karyotype of variant SFME clones capable of growing in serum-
148
containing medium
A1.7
Growth of variant SFME cells in serum-containing and serum-free
149
medium
A1.8
Reduced inhibition of variant SFME cells by thyroid hormine,
150
glucocorticoid, and gamma-interferon
A2.1
SFME cells cultured with EGF or without EGF for 4 hours
169
A2.2
Cell death and degeneration in SFME cells cultured for 8 hours
172
in the absence of EGF
A2.3 SFME cell death and severe degeneration after 24 hours of EGF
175
deprivation.
A2.4 DNA fragmentation in EGF-deprived SFME cells
177
A2.5
178
Release of cytoplasmic enzyme in culture medium by SFME cells
in the absence of EGF
A2.6
Flow cytometric size analysis of SFME cells
179
A2.7
Flow cytometric DNA analysis of SFME cells
180
A2.8
Polysome profiles of SFME cells
181
A2.9
Northern blot hybridization analysis of actin mRNA from SFME cells
182
LIST OF APPENDIX TABLES
Ent
Table
A1.1
Synthetic and enzymatic activities of SFME cells in serum-free
141
serum-containing media
A1.2 Karyotypic abnormalities of variant SFME clones capable of growing
in serum-containing media
142
Nonsenescent Serum-Free Mouse Proastroblasts:
Extended Culture,
Growth Responses In Vitro, and Application to the Culture
of Human Embryonic Astrocytes
CHAPTER 1
Introduction and Review of Literature
INTRODUCTION
One of the fundamental problems pursued by cell and molecular biologists is the
mechanism underlying the developmental regulation of cells in an organism. This is a great
undertaking. The sheer number and diversity of cells, their complex organization and
interactions, and the difficulty of controlling experimental variables in intact organisms and
interpreting observations from these experiments increase the complexity of the whole
animal system. One technique scientists have taken in order to study questions directed at
complex systems is to break the system down into smaller, more definable pieces. The
elucidation of small pieces of the large system may then be taken together to aid in
describing the system as a whole. Biologists have used this approach in order to study the
cellular and molecular mecnanisms involved in growth and development. This approach
necessarily adopts the "cell theory" concepts proposed first by Hooke then expanded upon
by others. First, the cell theory recognizes that compartmentalization is a universal
phenomenon of all or nearly all parts of both plants and animals, and that therefore the cell
is a universal structural unit of all living creatures. Second, the cell theory recognizes that
2
the ubiquitous structural units are also distinct physiological and developmental units,
capable of a separate and individual existence under certain conditions.
The level of resolution employed by cell biologists is the organ, tissue or cell.
Organ or tissue culture involves the explanting of a piece of tissue from an animal directly
into an in vitro environment. The tissue may be a thin slice, a larger chunk, or a complete
organ. Cell culture involves the culturing of dissociated cells obtained from tissue
explants. The dissociated cells may be primary cultures derived directly from tissue, they
may be secondary or subsequent passages of cells from primary cultures, or they may be
continuous established cell lines.
Studies at in vitro levels form the basis for the construction of models which may
be used to design and implement experiments in vivo on whole organs or organisms to test
the validity of a model. Similarly, models based on results obtained from in vivo studies
may be tested and expanded by using the tools of molecular and cellular biology. The
combined use of in vivo and in vitro studies provides a powerful approach to study
mechanisms of growth control, replication, and differentiation.
The research presented in my thesis describes two in vitro cell culture systems. My
thesis research focusses on (1) the derivation and long-term culture of SFME cells, (2)
defining the growth response of SFME cells to medium supplements, (3) investigating the
relationship between the nonsenescent nature of SFME cells and their responses to serum
and EGF, and (4) applying the serum-free cell culture methods that allowed the extended
culture of SFME cells to the multipassage culture of human embryonic astrocytes.
3
REVIEW OF LITERATURE
Animal cell culture. In order for cell culture to be carried out successfully several
criteria need to be met. The populations of cells have to be grown from single cells, the cell
population has to grow continuously for an extended period of time, the culture has to
consist of normal cells, and finally the cells must represent and display the properties of
cells from their tissue of origin.
Ross Grenville Harrison has been credited with the founding of modern cell
culture. In 1907 he performed an experiment designed to elucidate the process that takes
place in vivo resulting in extension of nerve fibers between nerve cell bodies during
embryonic development (Harrison, 1907). He isolated fragments from frog embryos
known to give rise to nerve fibers and cultured them in vitro in a lymph clot on a cover slip.
The cover slip was then inverted over a hollow slide and sealed with paraffin. Using this
approach Harrison was able to culture the tissue for a period of several weeks and
continually observe its development under a light microscope. He observed that in a large
number of cases fibers extended from the mass of nerve tissue and extended out into the
surrounding lymph clot. The most remarkable feature of the fiber was its enlarged,
ameboid end, from which numerous filaments extended. These projections have since
been named growth cones. This experiment showed that nerve fibers develop by
elongation from the nerve cell body rather than from protoplasmic bridges formed between
cells. More important, this work demonstrated that cells could survive and develop in vitro
in a manner closely resembling their normal embryonic development in vivo.
Having found that tissues from chick embryos grow readily in artificial media,
nutrient agar and bouillon, Lewis and Lewis (1911) proposed the great advantages of
culturing such tissues in media composed entirely of defined components. In their work
they studied the growth promoting effects of sugars and amino acids on chick embryo
4
explants in solutions of various simple salts. However, long-term growth was not
achieved.
Carrel (1912) proposed that death of cultures may result from preventable
occurrences; such as accumulation of catabolic substances and exhaustion of the medium.
By passaging chick embryo tissue explants into fresh medium and including plasma and
embryo extract in the medium, Carrel found the tissue could be passaged more than 20
times over a period of many months. Carrel's view was generally accepted that the entire
growth-promoting activity of his medium for fibroblasts resided in the embryo extract.
Plasma was regarded as an almost inert substratum. Because of his success, over the next
several decades most investigators adopted modifications of his approach, and little attempt
was made to develop more defined media.
Based on his earlier work developing media for plant tissue culture, White (1946)
developed the first defined medium able to support the prolonged survival but not active
growth of heart muscle and skeletal muscle from chick embryos. This medium was based
on the current knowledge of the nutrient requirements of animals and included glucose,
salts, amino acids, and vitamins. While this medium did not allow extended culture, it
demonstrated the potential of fully synthetic media and its use in defining the nutritional
requirement of cells. Soon after, Morgan et al. (1950) developed a defined medium
designated Mixture 199. This medium contained 68 defined compounds and supported
survival of chick embryo tissues for peroids averaging 4-6 weeks; however, no substantial
proliferation of cells was obtained.
Around this time, Sanford et al. (1948) developed the first permanent cell line.
Sanford cloned a strain of transformed, continuously passaged mouse fibroblast cells (L
strain) isolated originally by Earle et al. (1943). This development led to an era of
advances in the cultivation of animal cells in chemically defined media. Work by several
investigators, based on the earlier medium formulations of White (1946) and Morgan
(1950), led to the development of the first defined media that would support the active
5
growth of L cells in the absence of serum (Healy et al., 1954; Evans et al., 1955). Healy's
medium formulation, which contained 58 defined components, yielded 5- to 6-fold
increases of L cells per week. Evan's medium formulation, designated NCTC 109,
contained 71 defined components including 25 amino acids, 18 vitamins, 6 coenzymes, 5
nucleic acid precursors, 3 fatty acids, glucose and glutathione. This medium yielded 10- to
13-fold increases of L cells per week. However, the inclusion of horse serum gross
globulin greatly increased the mitogenic activity of NCTC 109, indicating the need to
identify additional medium supplements. Nearly all the early successes in animal cell
cultivation achieved in chemically defined media were obtained with strain L cells.
Although investigators developed numerous basal nutrient media and began to
optimize the concentration of sugars, salts, amino acids, nucleic acid precursors, vitamins
and other components for cells in culture, investigators over the next 20 years still relied on
undefined medium supplements in order to achieve acceptable growth of animal cells in
culture. The supplement most often used was bovine, calf or fetal calf serum. This
practice of supplementation of media with serum is still widely used today.
Using dialyzed serum as a medium supplement, Eagle systematically studied the
nutritional requirements of mammalian cells in an effort to establish their minimal essential
nutritional requirement (1955). These studies were conducted with He La cells, a human
cervical carcinoma cell line isolated by Gey et al. (1953), and with the transformed mouse
fibroblast L cell line (Sanford et al., 1948). Eagle showed that in addition to a small
amount of dialyzed serum, 13 amino acids, 7 vitamins, 6 inorganic ions and glucose or
another carbohydrate were required for cell growth. While limited by the inclusion of
serum, Eagle's work contributed greatly to the understanding of the nutritional requirement
of animal cells and resulted in the formulation of Eagle's minimal essential medium (MEM)
(Eagle et al., 1959). Important advances in enzymology also resulted from this work. For
instance, it was observed that while glutamine is essential for most serially propagated cell
lines, high concentrations of glutamic acid can substitute for glutamine (Eagle et al., 1956).
6
De Mars (1958) observed that at high concentrations of glutamic acid, cells formed greatly
elevated levels of glutamine synthetase. This increased enzyme formation was productinhibited; the presence of glutamine inhibited the formation of glutamine synthetase. This
was the first reported case of enzyme induction in serially propagated mammalian cell
cultures.
Hayflick and Moorhead (1961, 1965) described and characterized the development
of numerous cell strains of human origin (WI series), derived from fetal tissue and cultured
in Eagle's minimal essential medium supplemented with 10% calf serum. One of the goals
of this work was to create from normal tissue human cell strains which could be used as
models to develop viral vaccines. Cells were passaged twice weekly (2:1 split ratio) for as
many as 55 passages over a period of 8 months and retained a diploid karyotype. Without
exception, all of the strains were fibroblasts, and all cultures eventually entered a period of
growth crisis or senescence and stopped growing. Based on their work Hayjlick and
Moorhead advanced the hypothesis that the finite lifetime of diploid cell strains in vitro may
be an expression of aging or senescence at the cellular level.
Todaro and Green (1963) used a quantitative approach to study the growth of
mouse embryo cells in culture. They examined in detail the growth properties of mouse
embryo cells from the time they were placed in culture until an established cell line was
developed. Particular attention was paid to the cells as they entered growth crisis, or
senescence, and as they "converted" to become an immortalized cell line. Using various
passaging schemes based on time between passaging and plating density, they observed
that all cultures showed at first a progressively declining growth rate, followed by a period
of little or no growth, then the emergence, in most cultures, of an established cell line
within 3 months of culture. The established cell lines acquired several growth properties
distinct from their cells of origin. The established cells had a much greater ability to grow
at low cell density and in most cases a greater ability to form high density multilayered
cultures. However 3T3 cells (3 day passage, 3 x 105/plate), which unlike most cultures
7
did not reach confluent densities during establishment, remained contact inhibited under
normal culture conditions. These cells also remained nontumorigenic, while cells which
were not contact inhibited were tumorigenic. Most important, all of the established cells
were found to have near-tetraploid karyotypes and other gross chromosomal abnormalities.
These chromosomal alterations were observed to occur very rapidly following the
emergence from growth crisis, suggesting a chromosomal shift in a large fraction of the
diploid or quasidiploid population. Swiss, BALB, and NIH lines (3T3, 3T6, and 3T12),
which were derived using this approach, continue to be used extensively in studies of
growth regulation, differentiation, and carcinogenesis (Todaro and Green, 1963; Aaronson
and Todaro, 1968; Jainchill et al., 1969; Reznikoff et al., 1973).
Studies concerning the nutritional requirements of mammalian cells in vitro were
continued by Ham and coworkers. They developed media which were optimized for the
clonal growth of particular cell types using a systematic approach in which the components
of the defined medium were titrated against one another, in the presence of decreasing
amounts of serum, until little or no supplementation with dialysed serum was required to
obtain clonal cell growth (Ham,1963,1965; McKeehan and Ham,1978; Bettger and Ham,
1981). Using this approach they developed the well-known media F10 and F12 for the
clonal growth of Chinese hamster ovary (CHO) cells in the absence of serum, MCDB 105
for the growth of normal diploid fibroblasts in 1% dialyzed serum, and hormonesupplemented MCDB 108 and 110 for the clonal growth of normal diploid fibroblasts in
the absence of serum. These medium formulations are still widely used today. Serum-free
mouse embryo (SFME) cells were derived in and are routinely cultured in a basal nutrient
medium consisting of a one-to-one mixture of Ham's F12 and Dolbecco's modified Eagle's
medium (Loo et al., 1987).
Another approach taken to understand the role of serum in cell culture was to isolate
and characterize growth promoting activities from serum (Temin et al., 1972; Brooks,
1975). In general, these studies were unsuccessful. Purification of growth stimulatory
8
factors from serum has proved to be a very difficult task for numerous reasons. Large
amounts of starting material are required, many of the active factors found in serum are
bound to carrier proteins, active factors are often synergistically associated with other
factors, and finally many of the protein components of serum are physically similar (Barnes
and Sato, 1980).
Sato and coworkers, recognizing that a primary role of serum was to provide
hormones, developed a unique approach which has eliminated the requirement of serum in
cell culture medium for many cell types (Sato, 1975, Hayashi and Sato, 1976; Rizzino et
al., 1979; Barnes and Sato, 1980a,b; Barnes, 1987). These studies were based on the
following reasoning. Cells in vivo exist in and react to a complex set of nutritional,
hormonal, and stromal influences. If a cell is to survive or proliferate when removed from
an animal and placed in culture, the medium must carry out the functions previously served
by the complicated external in vivo environment of the cell. The approach Sato and
coworkers took was to begin with culture medium whose serum concentration had been
reduced and could not support significant cell proliferation. Purified hormones and serum
factors were then added to this medium individually or in various combinations. Factors
which restore cell growth were included in the medium and the serum concentration was
reduced until cell growth ceased. The process of hormone addition and serum reduction
was repeated until the serum was completely eliminated or no further improvements could
be made by addition of known factors (Rizzino et al., 1979). The serum factors identified
using this approach fall into several categories: hormones or growth factors such as insulin
and epidermal growth factor (EGF), transport proteins such as transferrin and albumin,
attachment factors such as fibronectin and serum spreading factor, and nutrients such as
fatty acids, polyamines and trace elements.
Using this serum-free approach, serum-free media have been developed which will
support the growth of many cell types, including cells of endocrine origin, epithelial origin,
neural and related origins, and mesodermal origin (Barnes, 1987). While most of the cells
9
capable of long-term proliferation in serum free medium are either immortalized cell lines,
transformed cell lines or cell lines derived from tumors, several normal cell lines have been
derived and maintained long-term in culture under serum-free conditions.
Orly et al. (1980) cultured primary rat granulosa cells in a 1:1 mixture of Ham's
F12 and Dulbecco's modified Eagle's medium (F12:DME) supplemented with insulin,
hydrocortisone, transferrin and fibronectin. The cells remained healthy and responsive to
follitropin (FSH) for at least 60 days. Cell response to FSH was suppressed by serum.
Using Eagle's MEM supplemented with nonessential amino acids, various trace
elements, insulin, transferrin, EGF, and 3, 5, 3'-tri-iodothyronine (T3), Kan and Yamane
(1982) cultured human diploid fibroblasts for up to 80 population doublings over a peroid
of more than 80 days before growth ceased. The addition of bovine serum albumin (BSA)
to the serum-free medium enhanced growth while delipidated BSA was ineffective.
Reconstitution of delipidated BSA with certain lipids restored the ability of BSA to enhance
growth, suggesting that BSA provided nutritional lipids.
Using Ham's F12 supplemented with 0.1-0.5% calf serum and six hormones,
Ambesi-Impiombato et al. (1980) cultured primary rat thyroid cells (FRTL strain) for more
than three years, with maintenance of highly differentiated functions. Hammond et al.
(1984) reported the long-term growth of normal human mammary epithelial cells for more
than 70 population doublings by using basal nutrient medium MCDB 170 supplemented
with insulin, hydrocortisone, EGF, ethanolamine and bovine pituitary extract.
Brandi et al. (1986), using Coon's modified Ham's F12 supplemented with bovine
hypothalamic extract, bovine pituitary extract, EGF, insulin, transferrin, selenous acid,
hydrocortisone, T3, retinoic acid, and galactose, cultured bovine parathyroid cells for more
than 140 population doublings before the culture began to deteriorate. The cells were
epithelioid and maintained highly differentiated functions.
Finally, in collaboration with my coworkers, I succeeded in the long-term culture of
mouse and human astrocyte precursor cells. Serum-free mouse embryo (SFME) cells,
10
cultured in a nutrient medium (F12:DME) supplemented with insulin, transferrin, EGF,
high-density lipoprotein (HDL), and fibronectin, have maintained a diploid karyotype with
no detectable chromosomal abnormalities for more than 200 generations. The cells did not
undergo growth crisis and remain in culture today. SFME cells are absolutely dependent
on EGF for survival, growth inhibited by serum or plasma, and nontumorigenic in
syngeneic hosts (Loo et al., 1987, 1989a,b, 1990; Rawson et al., 1990, 1991a,b).
Treatment of SFME cells with serum or transforming growth factor beta (TGF beta) leads
to the appearance of glial fibrillary acidic protein, a specific marker for astrocytes (Sakai et
al., 1990; Lazarides, 1982). Human embryonic brain cells, cultured in a nutrient medium
(F12:DME) supplemented with insulin, transferrin, EGF, fibroblast growth factor (FGF),
heparin, HDL, and fibronectin, were maintained for a maximum of 70 generations before
entry into growth crisis (Loo et al., 1991). Human embryonic brain cells expressed glial
fibrillary acidic protein (GFAP) in response to serum or TGF beta similar to SFME cells.
However, unlike SFME cells the human embryonic brain cells were more responsive to
FGF than EGF, they did not acutely require EGF or FGF for survival, and they were not
growth inhibited by serum. Most important, the human embryonic brain cells, unlike
SFME cells, could not be maintained in vitro for periods beyond that observed with other
human cell types in conventional serum-containing medium.
Brain cell diversity and mammalian cell culture. In the earliest stages of mammalian
development an invasion of mesodermal cells triggers the induction of overlying ectodermal
cells to become neuroectoderm (Spemann, 1938). The neuroectoderm then rolls up along
the length of the embryo and pinches off to form the neural tube. The cells of the neural
tube are precursors for the major differentiated cell types in the central nervous system
(CNS):
astrocytes, oligodendrocytes, neurons and ependymal cells (Sauer, 1935).
Recently, much research has focussed on identifying mechanisms which regulate the
production of differentiated cells in the CNS (McKay, 1989). Specifically, investigators
11
are looking at the lineal relationship between neurons, type-1 and type-2 astrocytes, and
oligodendrocytes.
Neurobiologists are using cell culture methods to study the development of the
CNS. Both cell lines and primary culture of brain cells are used. Many cell lines have
been established from brain-derived tumors, including the mouse C1300 neuroblastoma
cell line (Sato et al., 1969; Schubert et al., 1969), the rat C6 glioma cell line (Benda, et al.,
1969), the rat RN22 schwannoma cell line (Pfeifer, 1973), and the rat pheochromocytoma
cell line PC12 (Greene, 1976). The cell lines retain differentiated functions which are
associated with their cells of origin in vivo. They serve as valuable models for many
research interests, including membrane electrophysiology and ion transport,
neurotransmitter metabolism, cellular endocrinology, neurotoxicity and regulation of
growth and differentiation.
The isolation of stem cell lines are of particular interest. One clonal stem cell line,
RT4-AC, derived from a rat peripheral neurotumor, spontaneously differentiated in culture
into two morphologically distinct cell types: a glial-type cell, which expressed the glial
marker S-100 protein, and a neuronal-type cell, which possessed veratridine-dependent
sodium influx (Tomozawa and Sueoka, 1978).
Established cell lines have also been developed by transformation of embryonic
cells with various oncogenes, including c-myc and SV40 large T antigen. Ten-day
embryonic mouse neuroepithelial cells transformed with c-myc differentiate in response to
FGF into astrocytes, containing GFAP, and neurons, expressing the A2B5 marker and
neurofilaments (Bartlett et al., 1988). This finding indicated that some cells in the
neuroepithelium at embryonic day 10 are multipotent and are not restricted to either the
glial or neuronal cell lineage. Recently, Anderson and coworkers used a retrovirallytransformed adrenal chromaffin cell line to study neuronal development. The cell line they
used differentiates to contribute to neurons of the sympathetic nervous system and
secretory cells in the adrenal gland (Stemple et al., 1988; Barinaga et al., 1990). Anderson
12
found that FGF induced the cells to synthesize NGF (nerve growth factor) receptors.
Following priming by FGF the cells became responsive to NGF. Exposure to NGF
caused the cells to differentiate and become sympathetic neurons. In vivo, cells which
escape priming by FGF are thought to move on to the adrenal gland where they become
secretory cells under the influence of cortisol.
A recent approach to develop established cell lines was to derive cultures from the
cerebral cortex of transgenic mice that carried the polyoma virus large T antigen gene. Cell
lines established by this approach synthesize laminin and neural cell adhesion molecules
and induce primary neurons to develop neuritic processes. Upon reaching confluence the
cells expressed GFAP, and expression of GFAP is reversibly correlated with cell division
arrest (Galiana, et al., 1990).
Tumor-derived and oncogene transformed cell-lines have the advantage of
providing large quantities of homogeneous populations of cells. It must be remembered
however, that these cell lines have acquired properties associated with transformed and
malignant cells, and should not be considered to represent normal cells. They may be
tumorigenic in vivo and often have reduced requirements for growth factors (Scher et al.,
1978; Paul et al., 1978; Cherington et al., 1979; Chiang et al., 1985). Studies using
transformed cell lines can be extremely valuable when coupled with complementary in vivo
and primary culture studies.
Primary cultures have been used extensively to study the in vitro differentiation and
interactions of glial and nerve cells. Refinements in chemically defined media have
facilitated the growth of purified astrocytes, oligodendrocytes and neurons in culture
(Brunner et al., 1982; Bottenstein, 1985, 1986; Morrison and de V ellis, 1984; Conn,
1990). Raff, Noble and coworkers have used primary culture of the rat optic nerve as a
model to study glial cell development. The rat optic nerve contains the axons of retinal
ganglion neurons that project from the eye to the brain. Oligodendrocytes and astrocytes
structurally and functionally support the axons in the nerve. Raff et al. (1983) showed that
13
cultured rat optic nerve contained two types of astrocytes which could be distinguished by
morphology, cell-surface markers, and response to growth factors. They have been called
type-1 and type-2 astrocytes. Type-1 astrocytes were first detected at embryonic day 16,
oligodendrocytes on the day of birth, and type-2 astrocytes at the beginning of the second
week following birth (Miller et al., 1985). Using this system, Raff and coworkers showed
the rat optic nerve contains a bipotential glial progenitor cell (O -2A progenitor cell) found in
the rat CNS whose development was dependent on cell-cell interactions. The 0-2A
progenitor cell developed into a type-2 astrocyte or an oligodendrocyte in vitro. Raff
suggests that a collaboration between ciliary neurotrophic factor (CNTF) and the
extracellular matrix (ECM) influenced the differentiation of 0-2A progenitor cells. CNTF,
a protein produced by type-1 astrocytes in culture, induced 0-2A cells to express GFAP in
serum-free cultures of perinatal optic nerve cells (Hughes et al., 1988; Lillien et al., 1988,
1990; Raff, 1989). GFAP was expressed only transiently, however, and later the culture
reverted to become differentiated oligodendrocytes. The addition of ECM was required to
achieve stable differentiation to type-2 astrocytes. Type-1 astrocytes derive from a different
precursor cell (Raff et al., 1984). The mechanisms regulating the differentiation pattern of
0-2A progenitor cells are actively being studied.
Cell death. The concept of cell death evolved early in the history of cellular
pathology and has long been regarded as a degenerative phenomenon resulting from injury.
Only recently has cell death been investigated as a mechanism necessary for normal growth
and development of multicellular organisms. Based on morphological and biochemical
criteria, cell death has been divided into two classifications: necrosis and apoptosis (Wyllie
et al., 1980).
The occurrence of necrosis has been associated with nonphysiological conditions
such as severe hypoxia and ischemia, major changes in environmental temperature,
complement-mediated cell membrane disruption, and exposure to toxins. Necrotic insults
14
affect cells in groups rather than singly. No evidence has shown a connection between
necrosis and normal embryonic development. Morphological features of necrosis include
gross swelling of the matrix of mitochondria, swelling and eventural rupture of nuclear,
organelle, and plasma membranes and dissolution of ribosomes and lysosomes.
Chromatin initially appears as marginated, loose clumps, but with swelling of the nucleus
and rupture of its membrane, the marginated chromatin masses become dispersed and often
disappear (Wyllie et al., 1980).
In contrast to necrosis, apoptosis or programmed cell death appears to be
intrinsically programmed and mediates several important physiological and pathological
processes (Wyllie et al., 1980; Arends et al., 1990). It has been shown to occur in
developmentally regulated cell death in the embryo (Wyllie, 1981; Searle et al., 1982;
Hinchliffe, 1981) and require protein and RNA synthesis (Oppenheim et al., 1990;
Schwartz et al., 1990). It also has been observed with the deletion of autoreactive T-cells
during thymus development (Smith et al., 1989; S hi et al., 1989) and following the
addition of specific regulatory hormones (Cohen and Duke, 1984) or the deprivation of
specific growth factors (Kerr et al., 1973; Koury and Bondurant, 1990; Martin et al., 1988;
Rodriguez-Tarduchy and Lopez-Rivas, 1989; Kyprianou and Isaacs, 1989). Apoptosis
affects single cells rather than groups and has been observed to occur in three
morphologically distinct stage (Wyllie et al., 1980; Arends et al., 1990). During the first
stage chromatin condenses into a crescent-shaped cap that abuts the nuclear membrane and
the nucleus becomes convoluted and shrinks, forming a pyknotic nucleus. The total cell
volume decreases, the cytoplasm condenses, and the endoplasmic reticulum swells.
Mitochondria and other organelles remain normal. In the second stage, which may overlap
with the first stage, the nuclear and cellular membranes bud and separate into multiple
membrane-bounded apoptotic bodies containing condensed cytoplasm and in some cases
nuclear material. The apoptotic bodies are dispersed in the intercellular spaces and are
either transported to an adjacent lumen or phagocytosed by neighboring cells or
15
macrophages. In the final stage degeneration of the nucleus and cytoplasmic structures
proceeds, while the cell body remains small.
In addition to the morphological uniqueness of apoptosis, a biochemical hallmark of
apoptosis has been cleavage of chromatin into oligonucleosome- sized fragments (Arends et
al., 1990). Evidence suggests that cleavage results from activation of a calcium- and
magnesium-dependent endonuclease activity capable of cleaving chromatin at
internucleosomal sites. In many cases, apoptosis has been prevented by inhibitors of RNA
and protein synthesis, suggesting that macromolecular synthesis may be required for
programmed cell death.
Apoptosis has been studied both in vivo and in vitro by using several model
systems. In the intact adult male, the supply of androgen regulates a balance between cell
death and proliferation of the prostate gland. Castration-induced androgen deprivation in
the adult male rat led to loss of nuclear androgen receptors and involution of the ventral
prostate. Calcium- and magnesium-dependent DNA fragmentation and accumulation of
apoptotic bodies was observed (Kyprianou and Isaacs, 1988). Actinomycin D and
cycloheximide, inhibitors of cellular RNA and protein synthesis, partially inhibited
regression of the ventral prostate.
Primary sympathetic neurons in culture require nerve growth factor (NGF) for
survival. Neurons deprived of NGF for 48 hours died, and exhibited ultrastructural and
physiological properties consistent with apoptosis (Martin et al., 1988). Agents capable of
elevating cAMP prevented neuronal cell death, while activators of protein kinase C did not.
As with the rat ventral prostate, inhibitors of RNA and protein synthesis prevented
neuronal cell death following NGF withdrawal.
Erythropoietin, a glycoprotein produced in the kidney and liver, controls
erythrocyte production by regulation of erythroid progenitor cells. In the presence of
erythropoietin, progenitor cells survived and differentiated into reticulocytes. In the
absence of erythropoietin however, DNA fragmentation and death of erythroid progenitor
16
cells occurred within 16 hours. Cycloheximide inhibits DNA fragmentation and cell death
(Koury and Bondurant, 1990).
In the interleukin -2 (IL-2) dependent T-lymphocyte cell line CTLL -2, deprivation of
IL-2 led to DNA fragmentation and cell death within 10 hours (Rodriguez-Tarduchy and
Lopez-Rivas, 1989). Phorbol 12,13-dibutyrate, an activator of protein kinase C, inhibited
DNA fragmentation and blocked cell death resulting from IL-2 deprivation.
In another system, glucocorticoid hormones have been shown to kill immature
thymocytes through activation of programmed cell death (McConkey et al., 1989; Arends et
al., 1990). Glucocorticoid treatment of immature thymocytes resulted in an increase in
cytosolic calcium, calcium-dependent DNA fragmentation and cell death. Quin-2, which
buffers cytosolic calcium, and inhibitors of RNA and protein synthesis limited the increase
in cytosolic calcium and prevented DNA fragmentation and cell death. Inhibitors of
calmodulin also prevented cell death.
Studies which showed that inhibitors of RNA and protein synthesis blocked
apoptosis suggests that synthesis of new RNA and protein may be required for activation
of the cell death program. In the case of growth factor-deprivation induced apoptosis,
presence of the growth factor may suppress the expression of proteins which directly or
indirectly participate in programmed cell death. Much attention has focussed on isolating
and characterizing genes and proteins whose expression is regulated during programmed
cell death. Using the technique of differential cloning, several labs have cloned genes
whose expression increased.
By using the in vivo rat ventral prostate system, several androgen-repressed genes
have been cloned and identified. Expression If sulfated glycoprotein-2 (SGP-2) and the
related protein testosterone-repressed prostate message-2 (TRPM-2) were highly increased
in the ventral prostate following castration (Montpetit et al., 1986; Leger et al., 1987;
Bettuzzi et al., 1989). SGP-2 is a major glycoprotein secreted by Sertoli cells. Other
androgen repressed genes include glutathione S-transferase (Bettuzzi et al., 1989), c-myc,
17
c-fos, heat shock protein 70K and alpha tubulin (Ruttyan et al., 1988), as well as other
genes yet to be identified. The complexity of this in vivo system has made it difficult to
determine the relationship between androgen deprivation and expression of the identified
genes. What role any of these genes may play in apoptosis remains to be determined.
Genes regulated during programmed cell death in the in vitro systems have not been
identified.
Orientation to format. The following four chapters and two appendicies constitute
results of my thesis research. Each chapter and appendix represents work related to the
long-term growth and characterization of mouse and human astrocyte precursor cell lines.
Each of the chapters and appendices has been submitted and published in a peer review
journal. They are presented in publication format, including a brief SUMMARY, an
INTRODUCTION to provide the reader with background information and an introduction
to the scientific approach taken, a METHODS section, a RESULTS section, an
ILLUSTRATIONS section containing the data of the manuscript, and a DISCUSSION
section.
I was the principal investigator of the research presented in the first four
manuscripts, thererore they represent chapters of my thesis. I was co-principal investigator
of the last two manuscripts. Since I was a major contributor to these manuscripts, and the
work contributes significantly to my thesis research, they are included as appendices. All
references are compiled at the end of the thesis.
18
CHAPTER 2
Extended Culture of Mouse Embryo Cells Without Senescence:
Inhibition by Serum
Deryk Loo, Janice Fuquay, Cathleen Rawson and David Barnes
Published in: Science. Volume 236, pp. 200-202 (1987).
Coauthor contribution: D.L., primary investigator; J.F., technical assistant;
C.R., post-doctoral scientist; D.B., research director
19
SUMMARY
Mouse embryo cells cultured in vitro in serum-supplemented media undergo growth
crisis, resulting in the loss of genomically normal cells prior to the appearance of
established, aneuploid cell lines. Mouse embryo cells established and maintained for
multiple passages in the absence of serum did not exhibit growth crisis or gross
chromosomal aberration. Cells cultured under these conditions were dependent on
epidermal growth factor for survival. Proliferation was reversibly inhibited by serum or
platelet-free plasma, suggesting that mouse embryo cultures maintained by conventional
procedures are under the influence of inhibitory factors.
20
INTRODUCTION
Mouse embryo cells cultured in conventional serum-supplemented media rapidly
lose proliferative potential, leading to growth crisis or senescence followed by the
appearance of genomically altered immortalized (established) cell lines. Swiss, Balb, and
NIH lines (3T3, 3T6, and 3T12) as well as C3H 10T1/2, which are commonly used in
studies of carcinogenesis and growth control, were derived in this way and exhibit near-
tetraploid karyotypes with major chromosomal abnormalities (Todaro and Green, 1963;
Aaronson and Todaro, 1968; Jainchill et al., 1969; Reznikoff et al., 1973). Because
extended culture of some cell types has been achieved by replacing serum in the culture
medium (Orly et al., 1980; Barnes and Sato, 1980b; Ambesi-Impiombato et al., 1980;
Brandi et al., 1986; Hammond et al., 1984; Masui et al., 1986; Giguere et al., 1982), we
developed a serum-free approach to the culture of mouse embryo cells. These procedures
allowed extended cell proliferation in the absence of detectable crisis or gross genomic
alterations. Growth of cultures derived under these conditions was markedly inhibited by
both serum and plasma, suggesting the presence of circulating inhibitors distinct from
platelet-associated factors.
21
METHODS AND RESULTS
Cells from 16-day-old mouse embryos were cultured in nutrient medium
supplemented with 10% calf serum (Todaro and Green, 1963; Aaronson and Todaro,
1968; Jainchill et al., 1969; Reznikoff et al., 1973) or in medium in which serum was
replaced with insulin, transferrin, epidermal growth factor (EGF), high-density lipoprotein
(HDL), and fibronectin. This medium was developed originally for the growth of the
serum-derived, established C3H 10T1/2 and Swiss NIH 3T3 mouse embryo cell lines
(Pipas et al., 1983; Chiang et al., 1985) and will also support the growth, at least in the
short term, of early passage mouse embryo cells cultured initially in serum. Growth of
serum-derived cultures in this medium is enhanced by the addition of platelet-derived
growth factor (PDGF). The medium formulation is related in nutritional and hormonal
composition to other media developed for the growth of established lines of mouse embryo
cells (Serrero et al., 1979; Shipley and Ham, 1981).
Cultures carried in serum-containing media, in the presence or absence of additional
supplementation with the above factors, underwent a well-defined crisis (Fig. 2.1). A
single batch of unheated serum was used for the serum-containing cultures for the duration
of the experiment. Similar results were obtained when the experiment was repeated with a
different batch of serum or with heat-inactivated serum (56°C, 30 minutes). Cultures
maintained in serum-free media grew exponentially without a significant time lag and
represented a major portion of the population of embryo cells initially plated. Cell number
in primary cultures 6 days after plating was three times higher in serum-free medium than
in serum-containing medium. Cells of serum -free, cultures appeared morphologically
homogeneous upon light microscopic examination by the second or third passage and
remained so upon long-term culture.
Cultures were derived in serum-free medium from both Balb/c and Swiss embryos.
Initiation and long-term growth in the absence of serum was carried out with three
22
independent Swiss mouse embryo cultures; similar results were observed in all cases.
Cells spread poorly in low-density serum-free cultures, occasionally growing as detached
clumps. At high density the cells exhibited elongated, bipolar, spindle-shaped morphology
and organized into characteristic fibroblastic swirls. Swiss mouse embryo cells cultured
for 200 generations in serum-free medium were nontumorigenic as measured by injection
of 107 cells subcutaneously into 4- and 7-week-old male and female athymic Swiss mice.
Mice were observed for 6 months. Post-crisis mouse embryo cells derived in serumcontaining medium exhibited aneuploid karyotypes as expected, but cells cultured under
serum-free conditions retained chromosome numbers in the diploid range (Fig. 2.2).
Karyotype analysis by Giemsa banding was carried out at the same time as
chromosome counts. All of the metaphases from Swiss mouse embryo cultures in serum-
containing medium contained one or more translocated marker chromosomes after 20
population doublings had occurred. Giemsa banding indicated that one marker was the
result of a Robertsonian fusion involving chromosome 5 and an unidentifiable
chromosome; the other consistently observed marker was a minute chromosome. No
chromosomal abnormalities were identified in the serum-free derived cultures.
Trans locations were detected in some clones derived from the serum-free Swiss
mouse cultures, suggesting that cloning or freezing procedures may affect karyotype. One
clone containing a translocation was examined further.
This clone remained
nontumorigenic and had a modal chromosome number of 40. Cultures derived in serumfree medium contained a small fraction of cells (4 to 7%) with karyotypes in the tetraploid
range, similar to that reported for pre-crisis Swiss mouse embryo cells maintained in
serum-containing media (Todaro and Green, 1963).
Mouse embryo cells derived in serum-free medium required EGF for survival (Fig.
2.3). The EGF requirement was also observed if MCDB 402, a formulation developed for
Swiss 3T3 cells (Shipley and Ham, 1981), was used as the basal nutrient medium.
Individual omission of the other supplements resulted in decreased cell growth, but did not
23
result in immediate cell death. Flow cytometric analysis indicated that cells maintained in
the absence of EGF accumulated in the G1 phase of the cell cycle prior to loss of viability
(Rawson and Barnes, unpublished data). The EGF requirement of mouse embryo cells in
vitro suggests that this molecule or EGF-like factors (Marquardt et al., 1983) may play a
critical role in fetal development. A comparison of biochemical properties suggest that
EGF is not the survival factor identified previously for 3T3 cells (Lipton et al., 1972).
Mouse embryo cells derived in serum-free medium were capable of growing to very
high densities (Fig. 2.2 ). Under these conditions the cells were smaller than at lower
densities, forming multilayered piles and aggregates. It has been shown previously that
fibroblastic cultures can be grown to high cell densities if adequate nutritional conditions
are maintained (Dulbecco and Elkington, 1973). The apparent decrease in growth rate of
serum-free mouse embryo cultures at high density was the result of both increased cell loss
due to depletion of nutrients and accumulation of cells in G1 due to rapid depletion of EGF.
Survival of the cultures at very high cell densities was precarious because of rapid depletion
of EGF.
Addition of serum to the culture medium led to inhibition of growth;
supplementation of serum-containing medium with the serum-free medium marginally
improved cell growth (Fig. 2.3). Inhibition of cell proliferation was reversible upon
replacement of serum-containing with the serum-free medium. Flow cytometric analysis
indicated that cells in serum-containing medium accumulated in the G1 phase of the cell
cycle.
Both calf serum and platelet-free plasma were effective at inhibiting cell growth
(Fig. 2.4). In addition to the inhibitory activity of plasma factors, platelet-derived growth
factors (Masui et al., 1986; Tucker et al., 1984) also may be inhibitory for serum-freederived mouse embryo cells, since the inhibitory activity per milligram of protein was
somewhat higher for serum than for plasma. Inhibitory activity was detected in six
24
commercial lots of calf serum, all of which were capable of supporting the growth of 3T3
mouse embryo cell lines established by conventional procedures.
25
DISCUSSION
Our work indicates that mouse embryo cells exhibiting the characteristics of precrisis cultures can be maintained on a long-term basis if serum is replaced by hormonal and
nutritional supplements to the basal nutrient medium. Although the precise relationship of
the cell type in the cultures derived in serum-free medium to those derived in serumcontaining media remains to be resolved, several lines of evidence suggest that the serum-
free-derived cells may be of mesodermal origin, like serum-derived lines. Under
appropriate conditions the cells derived in serum-free medium were capable of expressing
classical fibroblastic morphology. In addition, malignant cells derived from serum-free
cultures by calcium phosphate-mediated oncogene transfection formed undifferentiated
sarcomas in athymic mice (Fuquay et al., unpublished data).
The phenomena of crisis (or senescence) of diploid rodent embryo cells that have
been maintained in serum-supplemented media and the transfection-induced immortalization
or establishment of genomically altered lines are being actively studied. Inhibition of
growth by serum or plasma factors and selection of cells unresponsive to these factors may
contribute to the processes of crisis and immortalization under conventional culture
conditions. Although we found it possible in serum-free medium to carry out extended
culture of mouse embryo cells without an obvious crisis, human fibroblasts have been
reported to undergo senescence in serum-free medium at approximately the same number of
population doublings as serum-grown cultures (Kan and Yamane, 1972). These results
may reflect differences in species or in serum-free medium formulations. In any case,
mouse embryo cells in serum-free media may provide a stable model system for the
identification and study of growth inhibitors and examination of the mechanisms by which
cells become hyperploid, malignantly transformed, or otherwise genomically aberrant.
26
ACKNOWLEDGMENTS
We thank A. Rizzino for calf serum and plasma and information on TGF beta
concentrations, K. Nusser and A. Helmrich for assistance and suggestions, T. Ernst and
W. Peterson for karyotype analysis, and J. Willard for flow cytometry assistance. D.B. is
a recipient of a Faculty Research Award from the American Cancer Society.
Janice Fuquay prepared mouse embryo cells for karyotyping. Cathleen Rawson
prepared culture medium and assisted with cell culture. David Barnes provided financial
and scientific support.
27
Fig. 2.1. Growth of Balb/c mouse embryo cells upon successive transfer in
serum-containing or serum-free media. Cultures were grown in (0), medium
supplemented with 10% calf serum (Gibco); (0), serum-free medium supplemented with
bovine insulin (10 µg/ml; Sigma), human transferrin (20 pg/m1; Sigma), EGF (50 ng/ml;
Collaborative Research), human HDL (20 µg/ml; Meloy) on flasks precoated with human
fibronectin (20 µg/m1; Meloy) as described (Barnes et al., 1983), except without the use of
bovine serum albumin; (II), cultures containing both the serum-free supplements and 10%
calf serum. Mouse embryo cells were prepared as described (Todaro and Green, 1963).
Cultures were initiated at 105 cells per 25-cm2 flask, medium was changed 3 or 4 days
after plating, and cells were passaged 6 or 7 days after plating. Cells remained
subconfluent under these circumstances. Serum-free cultures were passaged with the use
of trypsin and trypsin inhibitor (Barnes and Sato, 1980b). Cell number per flask was
determined on the first day after plating and at passage by counting suspensions of
trypsinized cells in a Coulter particle counter and by hemocytometer. Three or more
replicate flasks were initiated at each passage; cell number per flask for replicates varied
less than 10% from the average. Nutrient medium was a 1:1 mixture of Dulbecco's
modified Eagle's medium and Ham's F12 (Gibco) containing 10 nM sodium selenite,
sodium bicarbonate (1.2 g/liter), 15 mM Hepes buffer, penicillin (200 IU/ml),
streptomycin (200 gg/m1), and ampicillin (25 p.g/ml) (Barnes and Sato, 1980b).
28
Figure 2.1
1024
1020
1016
1012
108
104
100
Days
29
Fig. 2.2. Karyotype of mouse embryo cell cultures. (A) Swiss mouse embryo
cells cultured in medium containing 10% calf serum for two population doublings in vitro;
25-hour doubling time (primary culture). (B) Swiss mouse embryo cells cultured in
medium containing 10% calf serum for 20 population doublings; 23-hour doubling time
(post-crisis culture). (C) Swiss mouse embryo cells cultured in serum-free medium for
200 population doublings; 20-hour doubling time. (D) Balb/c mouse embryo cells cultured
in medium containing 10% calf serum, seven population doublings; 30-hour doubling time.
(E) Balb/c mouse embryo cells cultured in serum-free medium for 70 population doublings;
24-hour doubling time. Balb/c cultures and Swiss cultures in serum-free medium were
initiated and passaged as described previously for the derivation of Swiss 3T3 lines
(Todaro and Green, 1963). Swiss cells in serum-free medium were maintained initially in
the presence of human PDGF (1 unit/m1; Collaborative Research). PDGF stimulation of
cell growth in the presence of the other supplements was minimal, and addition of PDGF
was discontinued at 60 population doublings.
30
Figure 2.2
60
A
1
20
15
5
40
0
5
D
1
,
50
I
I,
1111a
.
10
20
60
100
Chromosomes per metaphase
31
Fig. 2.3. Effect of medium supplements on growth of mouse embryo cells
that were derived in serum-free medium. Cultures were grown in (0), serum-free
medium with supplements as described in Fig. 2.1; (), medium supplemented with 10%
calf serum; (), medium containing both the serum-free supplements and 10% calf serum;
(0), serum-free medium with all supplements as described in Fig 2.1 except EGF; (A),
medium supplemented with 10% calf serum for the first 3 days followed by a change to
serum-free medium with the supplements. Swiss mouse embryo cells initiated and grown
for approximately 200 generations in serum-free medium were used for the experiment.
Similar results were obtained with Balb/c mouse embryo cells that had been derived in
serum-free medium. Cells were plated at 105 per 35-mm dish. Cell number was
determined at the indicated times after plating by counting cell suspensions in a Coulter
particle counter, individual values varied less than 10% from the average. All cells in the
serum-free plates from which EGF was omitted were dead by day 5.
32
Figure 2.3
107
3 x 106
.c
i
0)
106
a.
,A
N
15
c
3 x 105
a
105
a
0-1...sa_.- .4`-=-M.."--.0-.
1-
WEI
'0
3 x 104
1
1
3
%
i
I
I
I
5
7
9
11
Days after plating
33
Fig. 2.4. Growth inhibition of serum-free-derived mouse embryo cells by
plasma or serum. Cells were plated in serum-free medium supplemented as described
in Fig. 2.1 in the presence of the indicated concentrations of calf serum (I) or plasma (0).
Cell number was determined 6 days later by counting cell suspensions in a Coulter particle
counter, individual values varied less than 10% from the average. Plasma was defibrinated
before use as described (Pledger et al., 1978) and serum was similarly treated. Cells were
plated at 105 per 35-mm dish. Swiss mouse embryo cells initiated and grown for
approximately 200 generations in serum-free medium were used for the experiment.
Similar results were obtained with serum-free-derived Balb/c mouse embryo cells. Assay
of calf plasma indicated less than 2 pg of transforming growth factor beta per milligram of
plasma protein.
34
Figure 2.4
4
0
31
omm,
0
0
0.1
1.0
Plasma or serum protein (mg/ml)
10
35
CHAPTER 3
Serum-Free Mouse Embryo Cells:
Growth Responses In Vitro
Deryk Loo, Cathleen Rawson, Angela Helmrich, and David Barnes
Published in: Journal of Cellular Physiology. Volume 139, pp. 484-491 (1989).
Coauthor contribution: D.L., primary investigator; C.R., post-doctoral scientist;
A.H., technical assistant; D.B., research director
36
SUMMARY
We have derived serum-free mouse embryo (SFME) cultures in a basal nutrient
medium supplemented with insulin, transferrin, epidermal growth factor (EGF), highdensity lipoprotein (HDL), and fibronectin. These cells are nontumorigenic, lack gross
chromosomal aberrations, and exhibit several other unique properties, including
dependence on EGF for survival and growth inhibition by serum. We have examined the
concentration dependence of the growth stimulatory effects of protein supplements used in
the SFME medium formulation and surveyed other supplements that might act as alternative
or complementary additions to the culture medium. Insulin could be replaced by insulinlike growth factor I and EGF could be replaced by transforming growth factor alpha in the
same concentration range. Transferrin could be replaced by higher concentrations of
lactoferrin. Deterioration of cultures in the absence of EGF began within 8 hours of the
removal of the growth factor, and could be prevented by the addition of fibroblast growth
factor/heparin-binding growth factor. Attachment proteins other than fibronectin were
effective on SFME cells, but limited success was obtained when substituting other lipid
preparations for HDL. These data introduce a precise system for exploring the unusual
characteristics of SFME cells and contribute additional information that may be useful in the
extension of these approaches to other cell types and species.
37
INTRODUCTION
Mouse embryo cells cultured multipassage in basal nutrient medium supplemented
with serum eventially undergo growth crisis or senescence, often followed by the
appearance of genomically altered, immortalized and sometimes tumorigenic cell lines
(Todaro and Green, 1963). Replacing serum in the culture medium by growth factors and
other supplements allows extended culture of some cell types that cannot be maintained on
a long-term basis in conventional serum-containing media (Ambesi-Impiombato et al.,
1980; Orly et al., 1984; Gospodarowicz et al., 1981; Hammond et al., 1984; Brandi et al.,
1986). We have applied a similar serum-free approach to the culture of mouse embryo
cells. We have reported that mouse embryo cultures can be initiated and maintained on a
multi-passage basis in a rich basal nutrient medium supplemented with insulin, transferrin,
epidermal growth factor (EGF), high-density lipoprotein (HDL), and fibronectin (Loo et
al., 1987).
Serum-free mouse embryo (SFME) cells derived in this manner display several
unique properties. SFME cells do not exhibit growth crisis or gross chromosomal
aberration and are nontumorigenic in syngeneic or athymic mice; these cells are dependent
on EGF for survival and are growth inhibited by serum or platelet-free plasma (Loo et al.,
1987). Because an appreciation of the behavior of SFME cells in serum-free culture is
necessary to further pursue an understanding of the unusual properties of the cells, as well
as to allow the further application and modification of this approach to cells derived from
embryos, neonates, or adults of mice and other species, we have examined in detail the
behavior of SFME cells in serum-free culture. In this report we characterize the effects on
SFME cells of growth stimulatory protein supplements to the basal serum-free medium.
38
MATERIALS AND METHODS
Materials. Bovine insulin, human transferrin, human lactoferrin, trypsin, bovine
serum albumin (BSA), heparin, polylysine, and soybean trypsin inhibitor were obtained
from Sigma Chemical Corp. (St. Louis, MO). Nerve growth factor (NGF) was obtained
from Bioproducts for Science, Inc. (Indianapolis, IN). Recombinant human insulin-like
growth factor (IGF I) was obtained from Amgen (Thousand Oaks, CA). Synthetic human
transforming growth factor alpha (TGF alpha) was obtained from Peninsula Laboratories
(Palo Alto, CA). Mouse laminin and type IV collagen was obtained from BRL (Bethesda,
MD). Type I collagen was obtained from Collagen Corp. (Palo Alto, CA). Bovine acidic
fibroblast growth factor/heparin binding growth factor-1 (endothelial cell growth factor;
Schreiber et al., 1985) and basic fibroblast growth factor/heparin binding growth factor-2
(FGF/HBGF), human transforming growth factor beta (TGF beta), and human platelet-
derived growth factor (PDGF) was obtained from R and D Inc. (Minneapolis, MN).
Mouse EGF was obtained from Upstate Biotechnologies (Lake Placid, NY).
Powdered medium formulations and antibiotics were obtained from Grand Island
Biological Company (Grand Island, NY). Sodium selenite and ethylenediaminetetraacetate (EDTA) were obtained from Fisher Scientific (Pittsburgh, PA). 4-(2-hydroxy-
ethyl)-1-piperazine-ethanesulfonic acid (HEPES) was obtained from Research Organics,
Inc. (Cleveland, OH). Gelatin-Sepharose was obtained from Pharmacia Chemicals
(Piscataway, NJ). Plasticware for cell culture was obtained from Falcon. Human serum
spreading factor was prepared as described (Barnes and Silnutzer, 1983). Human plasma
was obtained from volunteers.
Serum-free cell culture procedures. Cultures of Balb/c SFME cells giving rise to
the continuous cultures used for the experiments described in this report were initiated from
16-day-old mouse embryos (Loo et al., 1987). The embryos were minced and dispersed in
0.25% crude trypsin with 1 mM EDTA in phosphate-buffered saline without calcium or
39
magnesium (PBS) and incubated for 15 minutes at 37°C. Large chunks of tissue were
discarded, and the cell suspension was centrifuged at low speed from the trypsin solution
and resuspended in 10 ml of serum-free basal nutrient medium containing 1 mg/ml soybean
trypsin inhibitor. Cells were then recentrifuged, resuspended in fresh serum-free basal
nutrient medium, and plated in serum-free basal nutrient medium with growth stimulatory
supplements as described below.
Passaging of SFME cells was accomplished by trypsinization followed by dilution
of the cells into an equal volume of soybean trypsin inhibitor solution and centrifugation
from solution. Cells were routinely grown in a 5% CO2-95% air atmosphere at 37°C and
passaged weekly, plating at a density of 105 viable cells/25 cm2 flask. SFME cells are
extremely sensitive to trypsin; only a few seconds of exposure to the trypsin solution at
room temperature was necessary for removal from the flask and small volumes of trypsin
and trypsin inhibitor solution are needed (0.5 ml per 25 cm2 flask). Loss of cells
sometimes occurred when passaging under serum-free conditions if dilute cell suspensions
were manipulated. This is presumably due to adherence of the cells to pipets and tubes.
The problem could be reduced somewhat by maintaining high cell densities in the
suspensions during passaging and by the addition of 1 mg/ml BSA (Cohn Fraction V) to
the trypsin inhibitor solution. SFME cultures have been passaged continuously without
evidence of growth crisis or senescence for 200 doublings (Loo et al., 1987).
Serum-free culture medium. The basal nutrient medium was a one-to-one mixture
of Dulbecco-modified Eagle's medium containing 4.5 g/L glucose and Ham's F12 (Ham
and McKeehan, 1979) supplemented with 15 mM HEPES, pH 7.4, 1.2 g/L sodium
bicarbonate, sodium selenite (10 nM), penicillin (200 U/ml), streptomycin (200 µg/ml),
and ampicillin (25 tg/m1) (F12:DME) (Mather and Sato, 1979). Details and general
approaches of serum-free culture techniques and medium preparations are published
elsewhere (Ham and McKeehan, 1979; Sato et al., 1982; Bettger and Ham, 1982; Barnes,
1987). Cells were cultured in F12:DME supplemented with insulin (10 gg/m1), transferrin
40
(25 14/m1), human high-density lipoprotein (HDL) (10 p.g/ml), and EGF (50 ng/ml), in
dishes or flasks precoated with bovine fibronectin (10 gg/m1). Plating was accomplished
by adding the appropriate cell number in 1 ml of prewarmed medium to culture vessels
previously precoated with fibronectin and preincubated in the incubator for approximately
15 minutes with the remaining portion of medium (e.g., 1 ml of cell suspension added to 1
ml of medium in a 35 mm-diameter plated or 4 ml of medium in a 25 cm2 flask). Insulin,
transferrin, EGF, and HDL were then added directly to the individual plates or flasks as
small aliquots from concentrated stocks.
Insulin was prepared at 1 mg/ml in 20 mM HC1 and transferrin at 5 mg/ml in
F12:DME; both were filter sterilized. EGF was obtained as sterile, lyophilized powder and
reconstituted at 50 1.4/m1 with sterile PBS. FGF/HBGF was obtained as sterile,
lyophilized powder and reconstituted at 10 pg/ml with sterile PBS containing 1 mg/ml
BSA. Stocks of insulin, transferrin, EGF and FGF/HBGF were stored at -86°C in aliquots
until needed, then stored at 4°C after thawing. HDL (density=1.068-1.121 g/cc) was
prepared by KBr ultracentrifugation as described (Gospodarowicz, 1984) and filter
sterilized. Protein concentration of HDL stocks were 10-20 mg/mi. Fatty-acid-free BSA
was added to the HDL solutions at a final concentration of 1 mg/ml to act as a trap for toxic
lipid liberated by decomposition of the HDL with time. Stocks of HDL were stored at 4°C.
Fibronectin was prepared by gelatin-affinity chromatography (Ruoslahti et al.,
1982) and filter sterilized. Stocks were maintained at approximately 1 mg/ml in 50 mM
Tris, pH 7.5, containing 4 M urea. Fibronectin was provided to cells by precoating plates
with a solution of fibronectin at the indicated concentration made by adding a small volume
of concentrated stock to F12:DME (1 ml per 35 min-diameter plate) and incubating for 30
minutes at 37°C in a CO2 incubator, followed by removal of the precoating solution and
one wash with 2 or 5 ml of F12:DME. Stocks of fibronectin were stored at -86°C in
aliquots until needed, then stored at 4°C after thawing.
41
Experimental procedures. Cells cultured continuously in the complete serum-free
medium were detached from stock flasks by trypsinization, diluted into trypsin inhibitor
solution, centrifuged, resuspended, counted, and plated at 105/35 mm-diameter dish (or,
otherwise, at the density indicated) in media of the indicated composition. Cell number
was determined 5 days after plating, or otherwise at the indicated times after plating, by
counting trypsinized cell suspensions in a Coulter particle counter. Data shown are the
average of values from duplicate plates. Individual values among replicate plates varied
less than twelve percent from the mean. Stock SFME cultures used for the experiments
described in this report had undergone between 70 and 100 population doublings.
42
RESULTS
Effects of insulin and transferrin on SFME cells. Virtually all cell types in serum-
free culture are responsive to insulin (Sato et al., 1982; Barnes, 1987). The lowest
concentration of insulin to which SFME cells responded maximally was 3 .tg/m1 (Fig.
3.1). Response of cells to nonphysiological concentrations of insulin often is taken to
indicate that insulin is mimicking the action of IGF's, and these peptide growth factors
frequently are found to stimulate growth at concentrations that are lower than the effective
concentrations of insulin (Rosenfeld and Hintz, 1986). However, we found for SFME
cells that IGF I was effective at stimulating cell growth in approximately the same
concentration range as insulin (Fig. 3.1). No additional effect of IGF I (1 tg/ml) was
observed if it was added to SFME cells in the presence of a maximally effective
concentration of insulin (10 1.4/m1). The IGF used in these experiments was maximally
active at 1-5 ng/M1 in the stimulation of thymidine incorporation in chick embryo
fibroblasts, a standard assay of insulin-like growth factor activity, and similar results on
SFME cells were obtained with two different IGF lots (not shown).
As with insulin, almost all cell types in serum-free culture are responsive to
transferrin (Sato et al., 1982; Barnes, 1987). The lowest concentration of transferrin to
which SFME cells responded maximally was 10 p.g/m1 (Fig. 3.2). Lactoferrin has been
reported to stimulate cell growth and replace transferrin for some cell systems (Hashizume
et al., 1987), and we found that lactoferrin also would stimulate SFME cell growth,
although effective concentrations of lactoferrin were approximately tenfold higher than
those of transferrin (Fig. 3.2). No additional effect of lactoferrin (30 pg/m1) was observed
if it was added to SFME cells in the presence of a maximally effective concentration of
transferrin (30 µg/ml).
Effects of EGF and FGF /HBGF on SFME cells. Many cell types in serum-free
culture are mitogenically responsive to nanomolar concentrations of EGF (Sato et al., 1982;
43
Barnes, 1987). Maximal stimulation of SFME cell growth by EGF was observed at 30
ng/ml (Fig. 3.3). TGF alpha, which binds to the EGF receptor and mimics the action of
EGF (Marquardt et al., 1984), was also effective in the same concentration range as EGF,
although EGF was of somewhat higher specific activity than TGF alpha. Addition of both
EGF and TGF alpha together at 100 ng/ml produced an effect no larger than the addition of
either separately.
SFME cells were strikingly dependent on EGF for survival in serum-free medium;
when EGF was omitted from the medium, greater than 90% of the cells were dead within
24-48 hours (Fig. 3.4). This result was qualitatively different from those obtained when
any of the other medium supplements were singly omitted from the medium; SFME cells in
the presence of EGF and the individual absence of any of the other components were
capable of survival and growth, although cell number under these conditions was lower
than that of control plates in the complete serum-free medium (e.g., Figs. 3.1, 3.2, 3.4).
In experiments in which EGF was omitted and then restored to the medium and the cultures
allowed to recover for 4 days in the complete serum-free medium, we found that
incubations without EGF for as little as 8 hours led to significant reduction in the ability of
the cultures to recover, suggesting that cell death begins in the cultures very soon after the
omission of EGF (Fig. 3.5). By comparison, little time-dependent effect of the omission
and readdition of insulin was observed in this protocol.
Neither PDGF (1-10 ng/ml), NGF (1-10 ng/ml), nor TGF beta (0.1-10 ng/ml)
could replace EGF for survival or growth of SFME cells, but cells did survive and
proliferate slowly if EGF was replaced with FGF/HBGF (Figs. 3.4, 3.6). TGF beta
added in addition to EGF resulted in no effect or variable, small growth inhibition,
although this factor markedly improved cell adhesion and spreading. PDGF added in
addition to EGF was marginally stimulatory: about a 10% increase in cell number over
controls without PDGF was observed.
44
Both acidic FGF/HBGF-1 (endothelial cell growth factor) and basic FGF/HBGF-2
were effective and proliferation-promoting effects of FGF also were observed when the
growth factor was added to the complete serum-free medium containing EGF (Figs. 3.6,
3.7). Heparin potentiated the effect of both basic and acidic FGF/HBGF (Figs. 3.6, 3.7).
Potentiation of the FGF effect is thought to occur because the growth factor is more stable
when bound to heparin (Schreiber et al., 1985; Gospodarowicz and Cheng, 1986;
Klagsbrun et al., 1987).
Effects of attachment proteins and HDL on SFME cells. In addition to hormones
like insulin, IGF I, EGF, FGF, or TGF alpha and binding proteins like transferrin or
lactoferrin, attachment factors are also required for the proper anchorage of cells in serum-
free culture (Barnes, 1987). SFME cells were derived from primary culture using
fibronectin as an attachment factor, and responded maximally to about 10 14/m1 of this
protein (Table 3.1). Laminin and serum spreading factor, other attachment factors used in
serum-free medium formulations for various cell types (Grotendorst et al., 1982; Barnes
and Silnutzer, 1983; Barnes, 1987), also supported attachment and stimulated growth of
SFME cells (Table 3.1). Combinations of these glycoproteins with fibronectin did not
produce effects on cell growth larger than the effects of each alone. Polylysine precoating
of culture dishes as described by McKeehan and Ham (McKeehan and Ham, 1976) was
effective at promoting attachment, but was inhibitory for cell growth unless fibronectin was
also present. Type I and type IV collagen and gelatin were ineffective at supporting SFME
cell attachment or stimulating cell growth (not shown). Interestingly, SFME cells in the
absence of fibronectin were capable of extended growth as unattached multi-cellular
aggregates. Cells in plates without fibronectin underwent about five population doublings
in 6 days (Table 3.1).
In many cases, additional nutritional lipid supplements beyond those provided by
the basal medium are required for maximal growth of cells in serum-free culture (Ham and
McKeehan, 1979; Sato et al., 1982; Bettger and Ham, 1982; Barnes, 1987). HDL is
45
growth stimulatory for many cell types in serum-free medium (Gospodarowicz, 1984;
Rizzino, 1984; Hoshi and McKeehan, 1984), and is also effective on SFME cells (Fig.
3.8). Delipidated (apo) HDL did not exhibit the stimulatory activities of HDL on SFME
cells. Low-density lipoprotein (LDL) at similar concentrations was growth stimulatory, but
considerable variability in activity was observed among LDL preparations, and some
preparations were toxic at doses higher than 1 pg/ml. Very low-density lipoprotein
(VLDL) was inhibitory even at 1 µg/ml (not shown). These effects are presumably due to
the toxic nature of the particular lipid components or lipid decomposition products of these
lipoprotein fractions.
The lipid composition of HDL is primarily phospholipid and fatty acid cholesterol
esters (Gospodarowicz, 1984). Replacement of HDL with linoleic acid, oleic acid,
arachadonic acid, dioleoyl phosphatidyl choline, dilinoleoyl phosphatidyl choline,
diarachadonyl phosphatidyl choline, or cholesterol, with or without fatty-acid-free BSA
present as a lipid carrier, produced modest stimulatory effects in some cases, but the effect
of HDL could not be entirely replaced by these supplements and toxicity was observed at
higher doses of some lipids (e.g., greater than 500 nM linoleic acid) due to detergent
effects (not shown).
46
DISCUSSION
SFME cells exhibit a number of unusual characteristics for cells that can be grown
indefinitely in vitro. These cells are growth inhibited by serum or platelet-free plasma and
dependent on EGF (or FGF) for survival (Loo, et al., 1987). Here we have described the
responses of SFME cells to growth stimulatory supplements included in the serum-free
medium formulation in which these cells were derived, and also characterized the effects of
FGF on these cells. The formulation used for SFME cells was developed originally for
serum-free culture of C3H 10T1/2 and Swiss NIH 3T3 mouse embryo cell lines (Pipas et
al., 1984; Chaing et al., 1985) and is related in nutritional and hormonal composition to
other media developed for the growth of established lines of mouse embryo cells and
related cell types (Serrero et al., 1979; Darmon et al., 1981; Shipley and Ham, 1983).
The serum-free medium employed embodies a highly specific formulation that
imparts an intense selective pressure for the SFME cell type and the distinctive survival and
growth characteristics described. Not only are these cultures selected in the absence of
serum proteins such as albumin and platelet factors such as platelet-derived growth factor
and transforming growth factor beta, but also are cultured in lower concentrations of
transferrin, HDL, and fibronectin than one might expect to find in conventional serumsupplemented medium. Thus, the extracellular environment governing the survival and
growth of our serum-free cultures is significantly different than that of cells cultured in
medium supplemented with 10% serum.
Although it is not surprising that SFME cells respond to insulin and transferrin, it is
unusual to find that insulin and IGF I are effective at similar concentrations. Although the
optimal concentration of insulin or IGF I for growth stimulation of SFME cells is higher
than might be expected to exist under physiological conditions, suboptimal growth-
promoting effects of concentrations in the nanomolar range were observed for both
47
peptides. The relevant receptor presumably is of low affinity for these ligands; perhaps
both insulin and IGF I are interacting with the IGF II receptor.
The ubiquitous requirement of serum-free cultures for transferrin (Sato et al., 1982;
Barnes, 1987) is presumed to be due to the function of this protein as a carrier of iron, and
the iron-binding lactoferrin molecule could replace transferrin for SFME cells, as has been
observed for some other cell types (Hashizume et al., 1987). The optimal concentration of
transferrin for growth stimulation of SFME cells was within the physiological range; the
concentration of transferrin in serum is greater than 1 mg/nil (Beck, 1982).
Both EGF and TGF alpha at physiologically relevant concentrations were capable
of supporting survival and growth of SFME cells. These peptides in the range of 10-100
ng/ml were strongly growth stimulatory for SFME cells, but even 1 ng/ml was capable of
preventing cell death and supporting some proliferation beyond the initial plating density
(Fig. 3.3). TGF alpha levels can reach 1 ng/ml in culture medium from cells producing
this peptide and are present at about 1 ng/mg protein in mouse embryos (Twardzik et al.,
1982; Salomon et al., 1987). The level of circulating EGF in the mouse is about 1 ng/ml,
but can be boosted to over 100 ng/ml by appropriate stimulation of the submaxillary gland
(Carpenter and Cohen, 1979). Circulating levels of EGF in vivo should be sufficient to
allow survival of SFME cells injected into mice in tumorigenicity studies, especially when
one considers that the concentration in vivo is constant, while EGF levels in the cell culture
experiments described were diminished over the course of the experiment.
Replacement of EGF with FGF allowed survival and slow growth of the cells, but
the rate of proliferation was reduced compared to that in the presence of EGF, and the
gross morphology of cells was different in medium containing EGF compared with
medium in which EGF was replaced with FGF. The stimulation of SFME cell growth by
FGF in medium containing EGF suggests that serum-free medium containing FGF in
addition to the other components might be useful in the initiation and multipassage culture
of rodent embryo cells. We have utilized medium containing 10 ng/ml FGF in addition to
48
the other supplements used routinely with SFME cells to initiate cultures from C3H mouse
and Fischer rat embryos, as well as newborn Balb/c mice, and have obtained cells with
properties similar to those of Balb/c SFME cells.
Cell types derived in serum-containing medium, including mouse embryo lines
such as the various 3T3 and 10T1/2 strains as well as human fibroblasts, are generally
mitogenically responsive to EGF or FGF (Serrero et al., 1979; Darmon et al., 1981;
Shipley and Ham, 1983; Pipas at al., 1984; Chiang et al., 1985; Phillips and Cristofalo,
1988), but are not dependent on these peptides for survival. The cellular function required
for survival of SFME cells that is presumably met by activation of the EGF/TGF alpha
receptor is unknown, as is the mechanism by which FGF supplants the requirement. The
dramatic requirement of the peptide growth factors for survival of these mouse embryo
cells in culture suggests that an EGF-like or FGF-like factor may be critical for proper fetal
development.
SFME cells responded to fibronectin and other glycoprotein attachment factors in
serum-free culture. Even in the absence of any attachment factor, SFME cells were capable
of sustained growth. The ability to grow under anchorage-independent conditions is a
phenomenon not usually observed in nontumorigenic cells that also are capable of adherent
growth. SFME cells grow to very high densities in adherent culture (Fig. 3.6), also an
unusual property for nontumorigenic cells. Extended maintenance of high-density cultures
leads to an apparent plateau in cell number per plate that is the result of increased cell loss
due to medium depletion of nutrients and growth arrest prior to cell death due to depletion
of growth factors (Loo et al., 1987). SFME cells in high-density culture grow in
multilayered arrangements throughout the dishes, and no formation of foci is observed.
Nutritional lipids were supplied to the basal nutrient medium by the addition of
HDL. The concentration of human HDL used in serum-free medium for SFME cells (10
14/m1) is considerably less than that found in human plasma (about 1 mg/ml). Limited
49
success was obtained when substituting other lipoprotein fractions or lipid preparations for
HDL.
Although most of the growth stimulatory supplements to the basal nutrient medium
used for SFME cells are present in commercial sera, SFME cells do not grow in
conventional serum-containing media (Loo et al., 1987). The serum-free cell culture
techniques described here allow the elimination of undefined inhibitory serum factors from
the medium and also provide an assay system for the identification and study of these
inhibitory activities. In addition to usefulness in the study of growth-inhibitory factors,
SFME cells also provide a novel model for the study of other phenomena related to the
regulation of proliferation of animal cells, including the exceptional activities of peptide
growth factors and actions of oncogenes to block serum inhibition of growth or mimic
peptide growth factor action (Barnes et al., 1989). The explicit control of the culture
environment afforded by the SFME system confers advantages in approaching these topics
experimentally, as well as providing the only means by which cells exhibiting these
properties have been isolated. Application of these techniques to the direct serum-free
culture of untransformed cells from other species, including humans, should provide
interesting results in the future.
50
ACKNOWLEDGMENTS
We thank Kevin Nusser and Wayne Englander for help with this work. Support
was provided by NM-NCI 40475 and NIH-AG 07560. D.B. is the recipient of Research
Career Development Award 01226 from the National Cancer Institute.
Cathleen Rawson prepared HDL and performed insulin/IGF I dose response
experiment. Angela Heinrich prepared medium and performed experiments involving
attachment factors. David Barnes provided financial and scientific support.
51
Table 3.1. Effect of attachment factors on SFME cell growth. Cells were plated
at 105/dish in F12:DME medium supplemented with insulin (10 gg/m1), transferrin
(2514/m1), HDL (10 µg/ml), and EGF (50 ng/ml). Dishes were precoated with attachment
factors as indicated (1 ml precoating solution/plate) and washed prior to plating as
described in Materials and Methods. Cell number was determined 5 days after plating.
Attachment factor
Experiment 1
None
Fibronectin (1 µg /ml)
Fibronectin (2.5 µg /ml)
Fibronectin (5 14/m1)
Fibronectin (10 µg /ml)
Fibronectin (25 µg /ml)
Experiment 2
None
Fibronectin (10 µg /ml)
Serum spreading factor (10 µg /ml)
Laminin (10 µg /ml)
Polylysine (100 µg /ml)
Fibronectin (10 p.g/m1)
+ Serum spreading factor (10 i.g /ml)
Cells/dish
4.1
4.7
5.1
5.3
5.9
6.1
(±.13) x 106
(±.19) x 106
(-±.17)x 106
(±.17) x 106
(-±.36)x 106
(±.72) x 106
4.1
6.6
5.7
5.3
2.2
(±.03) x 106
(±.11)x 106
(±.36) x 106
(±.08) x 106
(±.19)x 106
6.4 ( ±.44) x 106
Fibronectin (10 1.1.g/m1)
+ Laminin (10 µg /ml)
Fibronectin (10 p.g /ml)
+ Polylysine (100 µg /ml)
6.0 (±.04) x 106
6.0 (±.47) x 106
52
Fig. 3.1. Growth response of SFME cells to insulin and IGF I. Cells were
plated at 105/dish in F12:DMA basal nutrient medium supplemented with transferrin (25
p.g/ml), HDL (10 µg/ml), and the indicated concentration of insulin or IGF I. Dishes were
precoated with fibronectin (10 j.t.g/m1). Cell number was determined 5 days after plating.
0, insulin; S. IGF 1.
6
5
4
GI
3
2
1
fr.-4r
0
0
//
0.001
61-
0.01
0.1
Insulin or IGF (µg/m1)
I.0
10.0
53
Fig. 3.2. Growth response of SFME cells to transferrin and lactoferrin. Cells
were plated at 105/dish in F12:DME basal nutrient medium supplemented with insulin (10
mg/m1), EGF (50 ng/ml), and the indicated concentration of transferrin or lactoferrin.
Dishes were precoated with fibronectin (10 µg/ml). Cell number was determined 5 days
after plating. 0, transferrin; 0, lactoferrin.
2
o
1
kJ
...If' 0
. 0'
,w
.
.. .
o'
,
0 - y ,i -.
0
'
0
II
L
0.1
1.0
10.0
Transferrin or Lactoferrin (µg/m1)
100.0
54
Fig. 3.3. Growth response of SFME cells to EGF and TGF alpha. Cells were
plated at 105/dish in F12:DME medium supplemented with insulin (1014/m1), transferrin
(2514/m1), HDL (10 µg/ml), and the indicated concentration of EGF or TGF alpha.
Dishes were precoated with fibronectin (10 µg/m1). Cell number was determined 5 days
after plating. 0, EGF; , TGF alpha.
6
5
3
2
1
0
1
I0
EGF or TGF alpha (ng/ml)
100
55
Fig. 3.4. Photomicrographs of SFME cells. A: Complete serum-free medium. B:
Medium from which insulin was omitted at the time of plating. C: Medium from which
EGF was omitted at the time of plating. D: Medium in which EGF was replaced by basic
FGF (10 ng/ml) and heparin (10 mg/nil). Photographs were taken 2 days after plating.
Composition of complete serum-free medium was as indicated in Materials and Methods.
B
6?
0
a
0a
a 2,
O
0
a
J
3
O
3
r,
56
Figure 3.5. Recovery of SFME cells from plating in growth-factor-deficient
medium. Cells were plated at 5 x 104/dish in F12:DME basal nutrient medium
supplemented with insulin (10 µg/m1), transferrin (25 tg/m1), and HDL (10 .Lg/m1)
(without EGF) or in medium supplemented with EGF (50 ng/ml), transferrin (25 1./g/m1),
and HDL (10 .tg/m1). At the indicated times, insulin (0) or EGF () was added to the
medium in plates deficient in that factor and cell number was determined four days later.
Medium was changed 48 hours before cell counts. Cells in plates maintained without EGF
i
for more than 24 hours were dead (see Fig. 3.4)
5
0
A
O
O
..
0
I
0
I
48
24
Hours Without EGF or Insulin
72
57
Fig. 3.6. Growth response of SFME cells to acidic and basic FGF in the
absence of EGF. Cells were plated at 105/dish in F12:DME basal nutrient medium
supplemented with insulin (10 µg/ml), transferrin (25 tg/ml), and HDL (10 µg/ml) (0),
and 10 ng/ml basic FGF with 10 µg/ml heparin (0), 10 ng/ml acidic FGF (A), or 10 ng/ml
acidic FGF with 10 µg/m1 heparin (A). Control cultures (I) contained insulin, transferrin,
HDL, and EGF (50 ng/ml). Dishes were precoated with fibronectin (10 .tg/ml). Cell
number was determined on the indicated days after plating. Medium was changed 4 days
after plating.
IC
3 x 106
106
3 x I05
105
0
2
4
Days After Plating
6
8
58
Fig. 3.7. Growth response of SFME cells to FGF in the presence and
absence of EGF and heparin. Cells were plated at 5 x 104/dish in F12:DME basal
nutrient medium supplemented with insulin (10.tg/ml), transferrin (25 1.4/m1), HDL (10
14/m1), and the indicated concentrations of basic FGF, with or without heparin (10 p.g/ml)
and with or without EGF (50 ng/ml). Dishes were precoated with fibronectin (10 p.g/ml).
Cell number was determined 5 days after plating. 0, without EGF, without heparin; I,
without EGF, with heparin; 0, with EGF, without heparin; I, with EGF, with heparin.
3 x 106
106
I
oif (
-- ---' 0
T
-
--a
- - ...
0
0
0
oil.°
.,
.
0
111,,,........../
105
lil -sr ------7". 111
3 x 104
...
.
...-,0
/
T/
0 0.1
1.0
FGF
( ng/ml)
10.0
59
Fig. 3.8. Growth response of SFME cells to HDL. Cells were plated at 105/dish
in F12:DME basal nutrient medium supplemented with insulin (1014/m1), transferrin (25
µg/ml), EGF (50 ng/ml), and the indicated concentration of HDL (A) or delipidated (apo)
HDL (). Dishes were precoated with fibronectin (10 µg/m1). Cell number was
determined 5 days after plating.
8
0
7
6
6.
0
x
5
4E11E0
.111.
Mln.
Mln.
4
3
2
1
0
5
10
15
0
20
High Density Lipoprotein (i.s.g/m1)
60
CHAPTER 4
Glucocorticoid and Thyroid Hormones Inhibit Proliferation
of Serum-Free Mouse Embryo (SFME) Cells
Deryk Loo, Cathleen Rawson, Molly Schmitt, Katherine Lindburg, and David Barnes
Published in: Journal of Cellular Physiology (1990). Volume 142, pp. 210-217.
Coauthor contribution: D.L., primary investigator, C.R., post-doctoral scientist;
M.S. and K.L., technical assistants; D.B., research director
61
SUMMARY
Mouse embryo cells derived in a serum-free medium formulation (SFME cells) do
not exhibit growth crisis or chromosomal abnormalities and are nontumorigenic in vivo;
these cells are also reversibly growth inhibited by serum or platelet-free plasma (Loo et al.,
1987). A portion of the inhibitory activity of serum could be extracted by charcoal, a
procedure that removes steroid and thyroid hormones. Both L-3,5,3'-triiodothyronine
(T3) and hydrocortisone inhibited growth of SFME cells in a reversible manner. The
inhibitory activity of serum also was partially removed by treatment with anion exchange
resin in a procedure designed to deplete serum of thyroid hormone. However, the effect of
serum on untransformed SFME cells could not be prevented by addition of the
antiglucocorticoid RU38486, and ras-transformed clones of SFME cells, which are capable
of growing in serum-containing medium, retained inhibitory responses to glucocorticoid
and, with some clonal variability, to T3. These results suggest that glucocorticoid or
thyroid hormones may contribute to the inhibitory activity of serum on SFME cells, but
additional factors are also involved.
62
INTRODUCTION
Mouse embryo cells derived in serum-free basal nutrient medium supplemented
with insulin, transferrin, epidermal growth factor (EGF), high-density lipoprotein (HDL),
and fibronectin exhibit a number of unusual properties. These serum-free mouse embryo
(SFME) cells do not exhibit an obvious growth crisis or chromosomal abnormalities and
are nontumorigenic in vivo (Loo et al., 1987,1989a). These cells are also dependent on
EGF for survival and are markedly inhibited by serum added to the medium at levels that
are stimulatory for cell lines derived in serum-containing medium (Loo et al., 1987). Sera
from a variety of species, including calf, adult, and fetal bovine sera, human serum and
mouse serum, as well as platelet-free plasma, exert similar inhibitory activity on SFME
cells. Fibroblast growth factor (FGF) can replace EGF in the serum-free formulation,
although growth with FGF in place of EGF is suboptimal (Loo et al., 1989b).
The influence of serum on untransformed SFME cells is likely to represent a
meaningful physiological response because the effect occurs at low serum concentrations,
is a reversible growth inhibition, rather than a toxicity, and results in the accumulation of
cells in the G1 phase of the cell cycle, rather than random inhibition at all points in the cycle
(Loo et al., 1987). Furthermore, SFME cells transformed by the human Harvey ras
oncogene no longer require EGF for survival, are capable of sustained proliferation in
serum-containing medium, and are tumorigenic in syngeneic or athymic mice (Barnes et
al., 1989). The ability of oncogene-transformed SFME cells to overcome growth
inhibition by serum raises the possibility that the loss of response to the inhibitory activity
may reflect a component of the malignant phenotype.
In this paper we report experiments designed to examine the nature of the inhibitory
activity of serum on SFME cells. We found that a portion of the activity was extracted by
charcoal, a classical procedure for removal of hydrophobic hormones (Westphal et al.,
1975).
Both L-3,5,3'-triiodothyronine (T3) and glucocorticoid at concentrations
63
comparable to those found in serum were capable of inhibiting growth of SFME cells.
These inhibitory activities were specific, reversible, and synergistic with charcoal-extracted
serum.
The combination of T3 and hydrocortisone at optimal concentrations produced an
inhibitory effect greater than the effect of each separately, and this effect was of the same
magnitude as that seen with an optimal amount of serum. In addition, the inhibitory
activity of serum was partially removed by treatment with anion exchange resin, which
removes thyroid hormone (Samuels et al., 1979), as well as other molecules, from serum.
However, the effect of serum on untransformed SFME cells could not be prevented by
addition of the antiglucocorticoid RU38486 (Moguilewsky and Philibert, 1984); rastransformed clones of SFME cells retained inhibitory responses to glucocorticoid; and
some clones retained inhibitory responses to T3.
64
MATERIALS AND METHODS
Materials. Bovine insulin, human transferrin, steroids and thyroid hormone and
derivatives, Norit A charcoal, trypsin, and soybean trypsin inhibitor were obtained from
Sigma Chemical Corp., St. Louis, MO. AG 1-X8 resin was obtained from Bio-Rad,
Richmond, CA. Mouse EGF was obtained from Upstate Biotechnologies, Lake Placid
NY. HDL (density = 1.068-1.121 g/cc) was prepared by KBr ultracentrifugation as
described (Gospodarowicz, 1984) and filter sterilized after dialysis. Powdered medium
formulations and antibiotics were obtained from Grand Island Biological Company, Grand
Island, NY. Sodium selenite and ethylenediaminetetraacetate (EDTA) were obtained from
Fisher Scientific, Pittsburgh, PA; 4-(2-hydroxy-ethyl)-1-piperazine-ethanesulfonic acid
(HEPES) was obtained from Research Organics, Inc., Cleveland, OH. Plasticware for cell
culture was obtained from Falcon. RU38486 was obtained from Roussel-Uclaf, 111
Route de Noisy, 93230 Romainville, France.
SFME cell culture. Details of derivation and behavior of SFME cells and serumfree culture methods have been published (Barnes, 1987; Loo et al., 1987, 1989a,b). The
basal nutrient medium was a one-to-one mixture of Dulbecco-modified Eagle's medium
containing 4.5 g/L glucose and Ham's F12 (Ham and McKeehan, 1979; Mather and Sato,
1979) supplemented with 15 mM HEPES, pH 7.4, 1.2 g/L sodium bicarbonate, sodium
selenite (10 nM), penicillin (200 U/ml), streptomycin (200 p.g/m1), and ampicillin (25
.tg/ml) (F12:DME).
Cells were cultured in F12:DME supplemented with insulin (10 .tg/ml), transferrin
(25 .tg/ml), human high-density lipoprotein (HDL) (10 pg/m1), and EGF (50 ng/ml), in
dishes or flasks precoated with bovine fibronectin (10 .tg/ml). Passaging of SFME cells
was accomplished by rapid trypsinization at room temperature (0.25% crude trypsin with 1
mM EDTA in phosphate-buffered saline without calcium or magnesium) followed by
65
dilution of the cells into an equal volume of soybean trypsin inhibitor solution (1 mg/ml in
F12:DME) and centrifugation from suspension.
Plating was accomplished by adding the appropriate cell number in 1 ml of
prewarmed medium to culture vessels previously precoated with fibronectin and
preincubated in the incubator (5% CO2-95% air atmosphere at 37°C) for approximately 15
minutes with the remaining portion of medium. Insulin, transferrin, EGF, and HDL were
added as small aliquots from concentrated stocks directly to the individual culture vessels
immediately after plating cells. Fibronectin was provided to cells by precoating plates with
a solution of fibronectin made by adding 5 gl of a 1 mg/ml stock to F12:DME (1 mg per 35
mm-diameter plate) and incubating for 60 minutes at 37°C in a CO2 incubator, followed by
removal of the precoating solution and one wash with 1 ml of F12:DME.
Insulin was prepared at 1 mg/ml in 20 mM HC1 and transferrin at 5 mg/ml in
F12:DME; both were filter sterilized. EGF was obtained as sterile, lyophilized powder and
reconstituted at 50 µg/ml with sterile PBS. Fibronectin stocks were filter sterilized and
adjusted to approximately 1 mg/ml in 50 mM Tris, pH 7.5 containing 4 M urea. Stocks of
insulin, transferrin, EGF, and fibronectin were stored at -86°C in aliquots until needed and
then stored at 4°C after thawing. Fatty-acid-free bovine serum albumin was added to HDL
solutions to a final concentration of 1 mg/nil and aliquots were stored at 4°C. Steroids were
prepared at 10-3 M in absolute ethanol and stored at -20°C. Thyroid hormone and
derivatives were prepared at 10-5 M in 0.2 N NaOH and stored at -20°C.
Human Ha-ras-transformed SFME cells were derived by calcium phosphate-
mediated transfection (Loo et al., 1989a) with a plasmid containing the oncogene
(pUCEJ6.6) (Shih and Weinberg, 1982; Land et al., 1983) obtained from the American
Type Culture Collection, Rockville, MD. Transformed cells were obtained by selection in
medium without EGF and isolated with cloning cylinders.
Assay of cell growth. Cells cultured continuously in the complete serum-free
medium were detached from stock flasks by trypsinization, diluted into trypsin inhibitor
66
solution, centrifuged, resuspended, counted, and plated at 105/35 mm-diameter dish in
media of the indicated composition. In experiments employing antiglucocorticoid,
RU38486 was added as small volumes of dilutions in F12:DME medium from a stock (1
mM) in absolute ethanol. Cell number was determined at the indicated times after plating
by counting trypsinized cell suspensions in a Coulter particle counter. Data shown are the
average of values from duplicate plates. Stock SFME cultures used for the experiments
described in this report had undergone between 60 and 120 population doublings.
Transformed cells had undergone approximately 30 to 60 doublings after cloning.
Measurements of T3 receptor levels in untransformed and transformed SFME cells were
carried out according to the procedure of Samuels et al. (Samuels et al., 1975).
Serum treatments. Serum was extracted twice with 50 mg/ml charcoal (Norit A) for
30 minutes at 56°C, followed by centrifugation to remove charcoal particles and filtration
through a 0.2 1.t.m filter (Westphal et al., 1975).
Charcoal treatment reduced the
radiolabeled tracer dexamethasone and T3 added prior to treatment by 99.8%. Control
serum was incubated without charcoal for 30 minutes at 56°C, followed by centrifugation.
Charcoal treatment reduced the protein concentration of the serum by 11%.
For resin treatment, serum was incubated with AG 1-X8 (chloride) resin at room
temperature (50 mg resin/ml serum, three changes at intervals of 5, 18, and 24 hours
(Samuels et al., 1976). Resin was autoclaved before use and sterility of serum was
confirmed after resin treatment. Control serum was incubated in the same protocol without
resin. Resin treatment reduced radiolabeled tracer T3 and dexamethasone added prior to
treatment by greater than 99.9%. Resin treatment reduced protein concentration of the
serum by 6%.
67
RESULTS
Inhibition of SFME cells by glucocorticoid and thyroid hormone. Serum is
strongly inhibitory for SFME cell growth, even when added to medium supplemented with
all of the growth - stimulatory supplements used in the serum-free medium formulation (Loo
et al., 1987). In attempts to characterize the inhibitory activity of calf serum on SFME
cells, we found that incubation of serum with activated charcoal at 56°C diminished or
eliminated the inhibitory activity of calf serum at low concentrations, although residual
inhibitory activity remained at higher serum concentration (Fig. 4.1). Serum was extracted
twice with charcoal; further extractions did not lead to additional reduction in inhibitory
activity. Incubation of serum at 56°C in the absence of charcoal did not affect the inhibitory
activity (Fig. 4.1). The inhibitory activity of calf plasma and human serum and plasma also
was partially removed by charcoal treatment. Lipoproteins are inhibitory for some cell lines
(Leffert and Weinstein, 1976; Ito et al., 1982; Barnes and Reed, 1989), and might be
expected to be removed to some extent by charcoal treatment, but delipidation of calf serum
or human plasma did not remove the inhibitory activity on SFME cells (not shown).
Because charcoal treatment is an established procedure for removal of steroids from
serum, we tested steroid hormones for inhibitory activity on SFME cells in medium
containing 2.5% charcoal-treated serum. Glucocorticoids were inhibitory at concentrations
that may be obtained in serum (Figs. 4.2,4.3). Other classes of steroids had no effect on
SFME cell proliferation.
Similarly, charcoal treatment removes thyroid hormones, and we found that T3 also
was inhibitory for SFME cell growth at physiological concentrations (Fig. 4.4). Thyroxine
(T4) was also active in this assay, while related compounds that do not possess thyroid
hormone activity in vivo did not inhibit SFME cell growth (Fig. 4.5). Both hydrocortisone
and T3 acted synergistically with charcoal-treated serum (Fig. 4.6). The synergistic effect
of charcoal-treated serum with thyroid hormone was not mimicked by T4 binding protein
68
added with T4 under serum-free conditions. The synergistic effect of charcoal-treated
serum with either hydrocortisone or T3 also was not mimicked with bovine serum albumin
at 1 mg/m1 (not shown).
Contribution of glucocorticoid and thyroid hormone to the inhibitory activity of
serum on SFME cells. SFME cells plated in F12:DME medium with growth-stimulatory
supplements, charcoal-treated calf serum, and either hydrocortisone or T3 grew more
slowly than controls throughout the culture period, and growth rate was further reduced if
both hydrocortisone and T3 were present (Fig. 4.7A). This inhibition was of similar
magnitude to that observed with 10% calf serum, and the effect, like that of calf serum,
was reversible upon removal of the inhibitory supplement (Fig. 4.7B). Treatment of serum
with AG 1-X8 anion exchange resin, which removed thyroid hormone, significantly
reduced the growth inhibitory activity of the serum, although resin-treated serum was still
strongly inhibitory at a concentration of 10% (Fig. 4.8).
This treatment, like charcoal treatment, also removed glucocorticoid from the
serum. However, addition of antiglucocorticoid RU38486 at 1 tM did not reduce the
inhibitory activity of serum, although this compound was capable of reversing the
inhibitory effect of added hydrocortisone on SFME cells (Table 4.1). RU38486 at 11.1M
also was ineffective at further reducing the residual inhibitory activity of resin-treated sreum
at a concentration of 10% (Fig. 4.8), and resin treatment of charcoal-treated serum did not
reduce inhibitory activity beyond that of treatment with resin or charcoal only. SFME cells
cultured for 5 days in medium supplemented with the growth-stimulatory factors of the
serum-free formulation and additionally with 10% resun-treated, charcoal-treated serum
were growth inhibited by 85%, compared with controls containing no serum.
SFME cells transformed with the human Ha-ras oncogene grow in the presence of
serum concentrations that are strongly inhibitory for untransformed SFME cells (Barnes et
al., 1989). We tested several clones of ras-transformed SFME cells for inhibitory activity
of triiodothyronine and hydrocortisone; results with two clones am shown in Table 4.2.
69
Both clones were inhibited by hydrocortisone or hydrocortisone and T3 together to a
degree similar to that observed with untransformed SFME cells. Inhibition of rastransformed cells by T3 was reduced in one clone, but was not reduced in another (Table
4.2). The concentration dependence of the inhibitory effect of T3 on the latter clone was
similar to that of untransformed SFME cells. T3 receptor level in the clone that was less
inhibited by T3 was not reduced, relative to untransformed SFME cells (not shown).
70
DISCUSSION
The unusual inhibitory response of SFME cells to serum and the ability of
tumorigenic derivatives to overcome the serum inhibition prompted an examination of the
molecules in serum that might be responsible. Although the platelet-derived peptide
transforming growth factor (TGF) beta is inhibitory for some cell types in culture (Tucker
et al., 1984; Sporn and Roberts, 1985), platelet-free plasma containing no detectable TGF
beta is strongly inhibitory for SFME cell growth (Loo et al., 1987), and addition of
purified TGF beta over a wide range of concentrations produced little or no SFME cell
growth inhibition under a variety of conditions.
Other potentially inhibitory peptides-tumor necrosis factor, interleukins, and
interferon (Clemens and McNurlan, 1985; Tsai and Gaffney, 1986; Saneto et al., 1986;
Kohase et al., 1987; Barnes and Reed, 1989) are probably not responsible for the
inhibitory effect of serum on SFME cells because tumor necrosis factor and interleukins I
and II had no effect when tested on SFME cells, and serum containing no detectable
interferon was strongly inhibitory for SFME cell growth (unpublished results). We have
presented evidence in this report that nonpeptide hormones are capable of inhibiting SFME
cell growth.
Removal of steroid and thyroid hormones by treatment with charcoal reduced the
inhibitory activity of serum on SFME cells. Of the steroids tested, only glucocorticoids
were inhibitory for SFME cell growth, and thyroid hormones (T3 or T4) also were
inhibitory. The effect of T3 and hydrocortisone together in the presence of 2.5% charcoal-
treated calf serum was greater than either separately, and was similar in magnitude to that
observed with 10% calf serum. It is possible that hydrocortisone or thyroid hormones
inhibit cell growth coordinately with the induction of a differentiated state in SFME cells,
but if this is the case, the differentiation is not terminal, because the inhibitory effect of
these hormones, like that of serum, was reversible.
71
Steroid and thyroid hormones induce multiple effects on cells, depending on the cell
type, but display similarities in receptors and mechanisms of action at the molecular level
(Evans, 1988; Beato, 1989). The action of hydrocortisone on SFME cells led to reduced
growth, but did not cause cell death, and thus differs from the irreversible toxicity of this
steroid on thymocytes (Kaiser and Edelman, 1977; Bourgeois and Gasson, 1985). The
effect of glucocorticoid on SFME cells resembles the inhibitory activity of steroids on
chicken fibroblasts and human mammary cells in culture (Lippman et al., 1976; Rubin,
1977).
Both hydrocortisone and T3 were active at concentrations found in serum, and both
were maximally effective when added in the presence of a low concentration of charcoal-
treated calf serum. This effect may be due to binding proteins of serum that modulate the
action of steroid and thyroid hormones. These proteins may not be extracted with charcoal,
or may be present in excess of the ligands. We were not able to mimic the effect of
charcoal-treated serum with T4-binding protein or albumin, another steroid and thyroid
hormone-binding serum protein, and the molecular nature of the factor or factors
responsible for the synergism of hydrocortisone and T3 with charcoal-treated serum is
unknown.
We were unable to alter the inhibitory activity of serum by addition of
antiglucocorticoid, although the antihormone clearly was capable of preventing the effect of
added hydrocortisone in the presence of charcoal-treated serum. Furthermore, rastransformed cells capable of growing in serum-containing medium remained inhibited by
hydrocortisone. Although the degree to which endogenous glucocorticoid in serum may
contribute to the inhibitory activity of serum on SFTAE cell growth is unclear, these cells
may provide a useful model for the study of glucocorticoid inhibition of proliferation under
well-controlled culture conditions.
We obtained a partial removal of the inhibitory activity of serum by treatment with
anion exchange resin that was designed to remove thyroid hormone, suggesting that
72
endogenous thyroid hormone may play a role in serum inhibition of SFME cells.
However, the resin treatment also removed glucocorticoid; partial removal of steroid from
serum by this treatment has been reported previously, and has been attributed to adsorption
to the hydrophobic support material of the resin (Samuels et al., 1979). In addition, we
found that some clones of ras-transformed SFME cells capable of growing in serumcontaining medium remained inhibited by T3, indicating that escape from serum inhibition
of SFME cell growth is not solely the result of loss of inhibitory response to thyroid
hormone.
Recent evidence that the erbA oncogene is related to the T3 receptor (Weinberger et
al., 1986; Sap et a., 1986) has focused speculation that this oncogene may represent a
mutationally activated receptor capable of functioning in the absence of ligand in a growthpromoting manner analogous to the mutationally activated EGF receptor, or erbB oncogene
(Downward et al., 19484; Samuels et al., 1988; Evans, 1988). An alternative possibility
raised by observations that thyroid hormone can be growth inhibitory is that a thyroid
hormone receptor might become altered in a manner that leads to interference with the
inhibitory function, thus releasing a cell from T3-mediated negative growth control.
Recently a non-hormone-binding erbA protein capable of inhibiting thyroid hormone action
has been reported (Koinig et al., 1989).
Although the phenomena we have reported here may explain partially the effect of
serum on SFME cell growth, significant growth-inhibitory activity existed in serum that
was exhaustively extracted with charcoal or anion exchange resin, or with a combination of
both treatments. Charcoal-treated or resin-treated serum at a concentration of 10%
remained strongly inhibitory for SFME cells. In addition, both treatments are capable of
extracting from serum factors other than steroid and thyroid hormones, such as peptides
(Titani et al., 1978). It appears that the effect of serum on SFME cells involves at least one
additional factor that is not removed by charcoal extraction. The nature of this factor
remains to be explored.
73
ACKNOWLEDGMENTS
We thank Carrie Smith for technical assistance. Supported by NIH -NAI- 07560. D.
Barnes is the recipient of Research Career Development Award NEI-NCI-01226.
Cathleen Rawson assisted with all growth curve experiments. Molly Schmitt
prepared charcoal-treated calf serum and assisted with SFME stock culturing. Katherine
Lindburg prepared medium and assisted with experiments using transformed-SFME cells.
David Barnes provided financial and scientific support.
74
Table 4.1.
Calf serum inhibition of SFME cells
in
the presence of
antiglucocorticoid RU38486. Experiment 1 (control): SFME cells were plated at
105/35 mm-diameter dish in F12:DME medium supplemented as described with insulin,
transferrin, EGF, HDL, fibronectin, and 2.5% charcoal-treated calf serum, with or without
hydrocortisone (100 nM) and RU38486 (1 gM). Cell number was determined 5 days after
seeding. Experiment 2: Plates were supplemented with insulin, transferrin, EGF, HDL,
fibronectin, and the indicated concentrations of serum, with or without RU38486 (1 pM).
Cell number was determined 6 days after seeding.
Cells x 10-6
Experiment 1
hydrocortisone
+ hydrocortisone
Experiment 2
0% calf serum
1.0% calf serum
2.5% calf serum
5.0% calf serum
10% calf serum
RU38386
+ RU38386
1.56 ± .19
0.43 ± .07
1.91 ± .07
1.58 ± .13
4.66
2.19
.288
.145
.159
± .32
± .47
± .04
± .01
± .01
4.39
1.98
.258
.144
.152
± .11
± .48
.02
± .01
± .02
75
Table 4.2. Glucocorticoid and triiodothyronine inhibition of ras-transformed
SFME cells. Cells were plated at 105/35 mm-diameter dish in F12:DME medium
supplemented as described with insulin, transferrin, EGF, HDL, fibronectin, and 2.5%
charcoal-treated calf serum, with or without hydrocortisone (100 nM) and, T3 (10 nM), or
both. Plates containing calf serum (10%) also received insulin, transferrin, EGF, HDL,
and fibronectin. Cell number was determined 6 days after plating.
SFME
Control
Hydrocortisone
Triiodothyronine
Hydrocortisone +
triiodothyronine
Calf serum
2.38
0.73
0.91
0.28
± 0.20
± 0.01
± 0.01
± 0.02
0.21 ± 0.03
Cells x 10m86
ras-SFME
clone A
2.75 ± 0.19
0.92 ± 0.05
1.00 ± 0.04
0.40 ± 0.03
ras-SFME
clone B
3.65 ± 0.37
0.93 ± 0.01
2.41 ± 0.09
0.47 ± 0.05
1.19 ± 0.01
2.06 ± 0.18
76
Fig. 4.1. Partial removal of inhibitory activity of calf serum for SFME cells
by charcoal treatment. Cells were plated at 105/35 mm-diameter dish in F12:DME
medium supplemented with insulin (10 1.4/m1), transferrin (25 1.4/m1), EGF (50 ng/ml),
and HDL (1014/m1) and the indicated concentrations of calf serum. Dishes were precoated
with bovine fibronectin (10 p.g/m1). Serum was extracted with charcoal at 56°C as
described. Control serum was incubated at 56°C without charcoal. Cell number was
determined 6 days after plating by counting trypsinized cell suspensions in a Coulter
particle counter. (II), untreated serum; (0), 56°C-treated serum; (0), charcoal-treated
serum. Error bars indicate variation of individual samples from the mean.
7
b6
45-
x
.c
5
60 4
, 3
a°
N
T
2
U
-)
I
0
0.0
2.5
5.0
Serum ( %)
7.5
10.0
77
Fig. 4.2. Steroid inhibition of SFME cell growth. Cells were plated at 105/35
mm-diameter dish in F12:DME medium supplemented as described with insulin,
transferrin, EGF, HDL and fibronectin as described. Dishes also contained 2.5% charcoal-
treated calf serum and the indicated steroids (10 nM). Cell number was determined 6 days
after plating. Error bars indicate variation of individual samples from the mean.
7
6
5
4
3
2
0
0
V
0
SI
SI
C
0
CU
it
UJ
O
78
Fig. 4.3. Concentration dependence of hydrocortisone inhibition of SFME
cell growth. Cells were plated at 105/35 mm-diameter dish in F12:DME medium
supplemented as described with insulin, transferrin, EGF, HDL and fibronectin as
described. Dishes also contained 2.5% charcoal-treated calf serum and the indicated
concentration of hydrocortisone. Cell number was determined 6 days after plating. Error
bars indicate variation of individual samples from the mean.
I
4-
i
/iNr---I
EN
I
0
I
0
//
I
10-10
I
i
1
10-9
10-8
10-7
Hydrocortisone ( M )
79
Fig. 4.4. Concentration dependence of triiodothyronine inhibition of SFME
cell growth. Cells were plated at 105/35 mm-diameter dish in F12:DME medium
supplemented as described with insulin, transferrin, EGF, HDL and fibronectin as
described. Dishes also contained 2.5% charcoal-treated calf serum and the indicated
concentration of T3. Cell number was determined 6 days after plating. Error bars indicate
variation of individual samples from the mean.
0
,
1J-711
'
0
10-12
I
4
3
o
2
I
0
IO-II
i0-10
10-9
Triiodothyronine (M)
10-8
80
Fig. 4.5. Specificity of thyroid hormone inhibition of SFME cells. Cells were
plated at 105/35 mm-diameter dish in F12:DME medium supplemented as described with
insulin, transferrin, EGF, HDL and fibronectin as described. Dishes also contained 2.5%
charcoal-treated calf serum and 1 nM triiodothyronine (T3), thyroxine (T4),
diiodothyronine (T2), iodotyrosine (I-tyr) or diiodotyrosine (I2-tyr). Cell number was
determined 6 days after plating. Error bars indicate variation of individual samples from
the mean.
a
7
6
5
4
3
2
1
0
Control
T3
T4
T2
I-Tyr I2-Tyr
81
Fig. 4.6. Concentration dependence of charcoal-treated calf serum for
glucocorticoid and T3 inhibition of SFME cell growth. Cells were plated at
105/35 mm-diameter dish in F12:DME medium supplemented as described with insulin,
transferrin, EGF, HDL and fibronectin as described. Dishes also contained 10 nM
triiodothyronine (0) or 100 nM hydrocortisone (0) and the indicated concentrations of
charcoal-treated serum. Data are expressed as percent of cell number in control plates
containing the indicated concentration of charcoal-treated serum but without
triiodothyronine or hydrocortisone. Cell number was determined 6 days after plating.
Error bars indicate variation of individual samples from the mean.
2.5
5.0
Serum (%)
7.5
10.0
82
Fig. 4.7.
Time
course and reversibility of triiodothyronine and
hydrocortisone inhibition of SFME cell growth. Cells were plated at 105/35 mmdiameter dish in F12:DME medium supplemented as described with insulin, transferrin,
EGF, HDL, fibronectin, and 2.5% charcoal-treated calf serum. Some dishes also were
supplemented with insulin, transferrin, EGF, HDL, fibronectin, and 10% calf serum. A:
(0), serum-free medium without T3 or hydrocortisone; (A), with hydrocortisone; (V), with
T3. B: (0) serum-free medium without T3 or hydrocortisone, as in A; (), with both T3
and hydrocortisone; (), 10% calf serum; (15) T3 and hydrocortisone added at plating,
changed on day 3 to medium without T3 and hydrocortisone; (0), 10% calf serum added at
plating, changed on day 3 to medium without calf serum. Error bars indicate variation of
individual samples from the mean.
I07
10 6
0
103
3
5
Days After Plating
7
9
3
5
7
Days After Plating
9
11
83
Fig. 4.8. Partial removal of inhibitory activity of calf serum by anion
exchange resin. Serum was extracted with resin as described. SFME cells were plated
at 105/35 mm-diameter dish in F12:DME medium supplemented as described with insulin,
transferrin, EGF, HDL, fibronectin, and the indicated concentration of serum. (0),
unextracted calf serum; (0), resin-extracted serum; (0), resin-extracted serum plus
antiglucocorticoid RU38486 (1 iiM). Error bars indicate variation of individual samples
from the mean.
Serum ( %)
84
CHAPTER 5
Serial Passage of Embryonic Human Astrocytes in
Serum-Free, Hormone-Supplemented Medium
Deryk Loo, Yoshio Sakai, Cathleen Rawson, and David Barnes
Published in: Journal of Neuroscience Research. Volume 28, pp. 101-109 (1991).
Coauthor contribution: D.L., primary investigator, Y.S., visiting professor;
C.R., research associate; D.B., research director
85
SUMMARY
We applied serum-free cell culture methods that allow extended proliferation of
mouse astrocyte precursor cells to the multipassage culture of embryonic human brain cells.
Cells were cultured in nutrient medium supplemented with insulin, transferrin, epidermal
growth factor, fibroblast growth factor, heparin, high-density lipoprotein, and fibronectin.
Cultures were maintained for a maximum of 70 population doublings before proliferation
ceased. The cells synthesized glial fibrillary acidic protein, an astrocyte marker, and
expression of this protein was increased by incubation of the cells with transforming
growth factor beta or serum. These results identify extracellular factors important for
proliferation and differentiation of embryonic human astrocytes and provide a controlled
system for multipassage culture.
86
INTRODUCTION
Our laboratory has derived and characterized cell cultures from 16 day mouse
embryos in basal nutrient medium supplemented with insulin, transferrin, epidermal
growth factor (EGF), high-density lipoprotein (HDL), and fibronectin (Loo et al., 1987,
1989a, b, 1990; Rawson et al., 1990; Shirahata et al., 1990; Sakai et al., 1990). These
serum-free mouse embryo (SFME) cells exhibit a number of unusual properties. These
cells do not lose proliferation potential or develop gross chromosomal aberration when
cultured for more than ten times the number of population doublings that can be achieved
with mouse embryo cells in conventional, serum-containing medium (Loo et al., 1987),
and proliferation of SFME cells is reversibly inhibited by serum (Loo et al., 1987, 1990;
Shirahata et al., 1990).
Recently we reported that treatment of SFME cells with serum or transforming
growth factor beta (TGF beta) leads to the appearance of glial fibrillary acidic protein
(GFAP), an intermediate filament protein of about 50,000 daltons that is a specific marker
for astrocytes (Sakai et al., 1990; Lazarides, 1982; De Armond et al., 1983; de Vellis et al.,
1986; Morrison et al., 1985). TGF beta is a platelet-derived growth factor that influences
growth and differentiation of a number of cell types (Sporn and Roberts, 1988; Massague,
1987; Rizzino, 1988).
Cells with properties like those of SFME cells can be isolated directly from mouse
brain (Sakai et al., 1990), and formulations of serum-free media developed specifically for
astrocytes (Morrison and de Vellis, 1981; Raff et al., 1983; Morrison et al., 1985; Raff et
al., 1984) are similar to the medium formulation used for SFME cells. Fibroblast growth
factor (FGF), pituitary extract (a source of FGF), or EGF are used in these astrocyte
media, and FGF can partially replace EGF for SFME cell survival and growth (Loo et al.,
1989b).
87
Results with mouse cells led us to ask if similar cultures could be developed from
embryonic human brain. Here we report the multipassage growth of human astrocytes
from 20- to 24-week-old embryos in media similar to that used for SFME cells. Several of
these cultures were maintained for 60 to 70 population doublings using this approach, but,
unlike SFME cells, the human cells eventually ceased proliferation. These cultures were
different in several other ways from SFME cultures. For instance, human cells were
dependent on FGF for proliferation and were less responsive to EGF than to FGF, while
SFME cells were dependent on EGF for survival and less responsive to FGF (Loo et al.,
1989b; Rawson et al., 1990). Response of the human cells to mitogens, in general,
decreased with population doubling level.
Medium supplemented with 10% calf serum did not support growth of the human
cells well unless supplemented additionally with the growth factors of the serum-free
medium. GFAP expression was increased in the cells by TGF beta or serum, but human
cells, unlike SFME cells, expressed GFAP at easily detectable levels in the absence of TGF
beta or serum. These results show that a cell type related to SFME cells is present in
embryonic human brain, identify extracellular factors important for proliferation and
differentiation of these cells, and provide a well-controlled system of multipassage culture
for further study.
88
MATERIALS AND METHODS
Cell culture. Embryonic tissue (brains from 20- to 24-week-old embryos from
saline-induced abortion) was obtained from the International Institute for the Advancement
of Medicine, Essington, Pennsylvania. For initiation of cultures, brains were dissected
under sterile conditions and dissociated by the procedure of Brunner at al. (Brunner et al.,
1982). Brains were minced and fragments were washed twice in 10 ml sterile basal
nutrient medium, then transferred to a sterile 10 cm-diameter tissue culture dish and cut into
1-3 mm-sized pieces. Trypsin was added (10 ml of 0.5 mg/ml crude porcine trypsin in
medium) and incubated at 37°C for 10 minutes. Tissue fragments were resuspended by
pipetting and 3 ml of a RNase/DNase solution (1 mg/ml each in phosphate-buffered saline,
PBS) was added and incubated at 37°C for 5 minutes.
Five milliliters of a collagenase/hyaluronidase solution (4 mg/ml each in PBS) was
added and incubated at 37°C for 30 minutes, resuspending tissue every 5 minutes. This
was followed by the addition of 3 ml RNase/DNase, 5 ml trypsin inhibitor (1 mg/ml in
medium), and an additional 3 ml of RNase/DNase with pipetting between each addition.
The tissue digest was filtered through a 30 gm pore nylon filter into a conical
polypropylene tissue culture tube and centrifuged at 600g for 5 minutes. The cell pellet
was washed once with medium and cells were plated in tissue culture flasks and cultured in
serum-free medium with the supplements as described below.
Detailed procedures for the culture of Baib /c SFME cells have been published (Loo
et al., 1989a,b), and similar procedures were employed for embryonic human brain
cultures The basal nutrient medium was a 1:1 mixture of Dulbecco-modified Eagle's
medium containing 4.5 g/L glucose and Ham's F12 (Ham and McKeehan, 1979; Mather
and Sato, 1979) supplemented with 15 mM HEPES, pH 7.4, 1.2 g/L sodium bicarbonate,
sodium selenite (10 nM), penicillin (200 U/ml), streptomycin (200 p.g/m1), and ampicillin
(2514/m1) (F12:DME). SFME cells were cultured in F12:DME supplemented with insulin
89
(lOgg/m1), transferrin (10 gg/m1), HDL (10 µg/ml), basic FGF (10 ng/ml), heparin (1
p.g/ml), and EGF (50 ng/ml) in dishes or flasks precoated with bovine fibronectin (10
gg/1111)
Bovine insulin and human transferrin were obtained from Sigma Chemical Corp.
(St. Louis, MO). Bovine basic FGF and human TGF beta were obtained from R and D
Inc. (Minneapolis, MN). Mouse EGF was obtained from Upstate Biotechnologies (Lake
Placid, NY). HDL and fibronectin were prepared as described (Loo et al., 1989a).
Passaging of human embryonic brain cells was accomplished by trypsinization followed by
dilution of the cells into an equal volume of soybean trypsin inhibitor solution and
centrifugation from solution.
Cells were routinely grown in a 5% CO2-95% air
atmosphere at 37°C and passaged at a 1:10 dilution. Medium was changed or additional
growth factors were added 4 days after plating, and cultures were passaged or medium was
changed 7 days after plating. Population doublings were calculated from passage dilution
at each trypsinization.
Western blots. For the preparation of extracts, cells were scraped from 10 cm-
diameter plates into 1 ml ice-cold PBS, pH 7.2, with 0.02% PMSF. Cell suspensions
were centrifuged and pellets were stored at -86°C until extracts were prepared. To prepare
extracts, one to two volumes of extract buffer (0.1% Triton X-100, 0.02% PMSF, 0.125
M Trizma base, pH 6.8, 250 units/ml DNase) was added to cell pellets and thoroughly
mixed using constant pipetting. Then 0.5% SDS and 0.5% 2-mercaptoethanol were added
to each lysate followed by extensive mixing. Protein assays were performed on the lysates
using the method of Bradford (Bradford et al., 1976). For preparation of tissue extract
from uncultured brain, minced tissue fragments were treated as described for cell
suspensions.
Cell extracts (40 .tg protein per lane) were reduced with 2-mercaptoethanol and
fractionated by electrophoresis on a 10% polyacrylamide slab gel (80 V/29 mA for 5
minutes followed by 100 V/35 mA for 15 minutes, then 25 V/6 mA for 13.5 hours). The
90
polyacrylamide gel was removed from between glass plates and encased in an
electroblotting apparatus (Transblot, BioRad). Proteins were electrophoretically transferred
from the gel to nitrocellulose filters (0.45 gm, Schleicher and Schuell) in transfer buffer
(150 mM glycine, 20 mM Trizma base, pH 8.3, 20% methanol, at 4°C) at 75 V for 3
hours. The nitrocellulose filters were air dried at room temperature for 16 hours.
The nitrocellulose blots were immersed in TBS (20 mM Tris, pH 7.5, 500 mM
NaC1) and incubated with gentle agitation at room temperature for 2 minutes. Then filters
were incubated for 1 hour in blocking solution (3% gelatin (w/v) in TBS) and washed for
10 minutes in TTBS (20 mM Tris, pH 7.5, 500 mM NaC1, 0.05% Tween-20). In order to
detect GFAP, filters were incubated for 1.5 hours in a 1:1,000 dilution of anti-GFAP
(rabbit anti-bovine GFAP antisera, Biomedical Technologies, Inc.) in antibody buffer
(1.0% gelatin (W/V) in TTBS) then washed twice, 5 minutes each in TTBS.
Filters were then incubated for 1.5 hours in a 1:3,000 dilution of affinity-purified
goat anti-rabbit IgG conjugated with horseradish peroxidase (BioRad, blotting grade) in
antibody buffer, washed twice, 5 minutes each, in TTBS and once for 5 minutes in TBS.
Filters were incubated with substrate (.067 mg/ml 3-amino-9-ethylcarbazole, 50 mM Tris-
HC1, pH 7.3) for 10 minutes to 1 hour with gentle agitation in the dark at room
temperature. When bands were visible development was stopped by decanting substrate
solution and washing the filter twice with water, l0 minutes each time, at room temperature
in the dark. Experiments with rabbit antibody were confirmed with mouse monoclonal
anti-GFAP (Boehringer-Mannheim) and anti-mouse IgG conjugated with horseradish
peroxidase (BioRad).
GFAP culture immunoassay. Cells were plated at 5 x 105 cells per 35 mmdiameter dish. At the indicated time, cells were fixed by aspirating medium and slowly
adding 1 ml 70% ethanol followed by incubation for 5 minutes at room temperature,
followed by an identical treatment with 95% ethanol. The ethanol was then aspirated and
plates were air dried. Fixed cells were incubated with a solution containing 15 mM
91
NaPO4, pH 7.1, 1% Triton X-100, 0.6 M KC1, 10 mM MgC12, 2 mM EDTA, and 1 mM
EGTA for 5 minutes at room temperature and then washed with 1 ml PBS for 5 minutes at
room temperature. One milliliter of PBS with 10% calf serum (CS) was added and
incubated for 1 hour at room temperature. The 10% CS solution was aspirated and 1 ml of
anti-GFAP antiserum (1:1,000 in PBS with 10% CS) was added and incubated at room
temperature for 1 hour followed by two washes with 2 ml PBS, each 5 minutes at room
temperature. Controls received an equivalent amount of normal rabbit serum in place of
anti-GFAP.
Affinity-purified goat anti-rabbit IgG (H & L) conjugated with horseradish
peroxidase (BioRad, blotting grade) (1 ml of a 1:3,000 dilution in PBS with 10% CS) was
added and incubated for 1 hour at room temperature. This was aspirated and plates were
washed twice, 5 minutes each, in 2 ml PBS at room temperature. One milliliter of 2 mM
2,2'-azino-di-(3-ethyl-benzothiazolin-sulfonate) in 0.1 M sodium acetate, 0.05 M Naphosphate and 2.5 mM H202 was added and the plates were incubated for 30 minutes in
the dark at room temperature. Absorbance at 405 nm was determined for each sample.
Cell number per dish for each condition was determined by trypsinizing replicate dishes
cultured identically to those used for the GFAP assay and counting in a Coulter particle
counter. Control values from plates receiving normal rabbit serum in place of anti-GFAP
were less than 20% of those with anti-GFAP and were subtracted from the corresponding
value with anti-GFAP in each case. Variation in individual plates in GFAP per cell,
expressed as (A405)/106 cells, was less than 10% from the average of duplicate dishes.
Immunocytochemical assay was carried out in the same manner except substrate was 3amino-9-ethylcarbazole (.067 mg/ml) in 50 mM Tris-HC1, pH 7.3, and incubation was 3 to
24 hours in the dark at room temperature.
92
RESULTS
Serial passage culture of embryonic human astrocytes. Human brain cultures (22
week-old embryo) initiated in the serum-free medium developed for SFME cells and further
supplemented with FGF and heparin initially grew slowly (Fig. 5.1). Ultimately a
population of homogeneous appearance emerged (Fig. 5.2) and, in the example shown in
Figure 5.1, proliferation rate increased. Similar results were obtained with two cultures
initiated independently from the same specimen and carried independently (Fig. 5.1).
Human foreskin fibroblasts initiated and carried in the serum-free medium initially
proliferated rapidly, but growth halted abruptly after about 30 population doublings (Fig.
5.1). Companion cultures from the same preparations initiated in medium supplemented
only with 10% serum could be maintained for only a few generations in this medium before
proliferation ceased (not shown). Ultimately proliferation rate of the brain-derived cells
declined and the cultures reached a point at which cell proliferation ceased.
Figure 5.3 is an example of data from a culture (24 week-old embryo) that did not
recover from the initial slow growth rate in serum-free medium, but exhibited a behavior
similar to that of the cells of Figure 5.1 when 0.5% serum was added at the time of
passaging in addition to the other medium supplements. The medium was changed to
serum-free medium the following day, so the cells were exposed to medium containing the
small amount of serum for less than 24 hours at each passage.
Continuous
supplementation of cultures with 0.5% or 1.0% calf serum did not prevent the ultimate loss
of proliferative potential after 60 to 70 population doublings. Similarly, addition of 2% or
10% serum to the medium of cells carried serum-free to population doubling levels of 60 to
70 as shown in Figure 5.1 did not restore proliferative potential.
Another independent culture (23 week-old embryo) derived in serum-free medium
underwent the initial growth and loss of proliferative potential, but did not resume growth
when 1% serum was added to the medium (not shown). Attempts were made to culture
93
cells from three other embryos (7 to 11 weeks old), and all were successfully maintained in
primary culture and passaged, but long-term culture was not undertaken. Clonal growth of
all of the cultures was poor in serum-free medium but was substantially improved by
supplementation of the medium with serum in addition to the other factors.
Growth characteristics of embryonic human astrocytes in vitro. Cells passaged in
serum-free medium supplemented with insulin, transferrin, EGF, FGF, heparin, HDL, and
fibronectin initially underwent a lengthy lag period after plating before proliferation was
observed. Additional hormonal supplements such as steroids and thyroid hormones and
other peptide growth factors did not improve growth. Little proliferation was observed in
plates containing medium supplemented only with 10% calf serum, while cells proliferated
with no lag in medium supplemented with both the factors of the serum-free medium
formulation and 10% calf serum. The pattern of cell growth in these media was the same in
both low-passage and high-passage cultures, but the overall level of growth was reduced in
the high-passage cultures (Fig. 5.4).
Individual omission of each of the peptide growth factors from the serum-free
medium reduced the cell number per plate after 7 days, and similar results were seen in
both low- and high-passage cells, although growth was generally reduced in the highpassage culture. Cell number was reduced more by omission of FGF/heparin or insulin
than by omission of EGF (Fig. 5.5). The response of the human cells to removal of both
FGF and EGF from the medium was qualitatively different from the response of SFME
cells. Greater than 80% of the human cells remained alive in the absence of these growth
factors through the 7-day period of the experiment, while all SFMF. cells under identical
conditions die within 48 hours (Loo et al., 1989a; Rawson et a1., 1990). Results shown in
Figures 5.4 and 5.5 were obtained with the culture for which multipassage growth
characteristics are shown in Figure 5.1. Similar results were obtained with the culture for
which growth characteristics are shown in Figure 5.3.
94
Expression of GFAP by embryonic human astrocytes in vitro. GFAP was detected
in a quantitative assay using both early and late passage cultures of human brain cells
cultured in serum-free medium (Fig. 5.6). Although the human brain cells, like SFME
cells, increased expression of GFAP in response to TGF beta and calf serum, a significant
amount of GFAP was evident in these cultures in the immunoassay and in Western blots
(Fig. 5.7) in the absence of serum or TGF beta. SFME cells do not produce easily
detectable amounts of GFAP in the absence of the inducers, but markedly express GFAP in
response to serum or TGF beta in immunoassays and Western blots identical with those
described here for human cells, while mouse 3T3 cells are negative in these assays (Sakai
et al., 1990).
Western blots of extracts of the serum-free human brain cells probed with antiserum
to neurofilaments or neuron-specific enolase were negative and no induction of these
markers was observed when the cells were incubated with 10% calf serum (not shown).
Western blot controls with normal rabbit serum in place of anti-GFAP were negative.
Results identical to those shown in Figure 5.7 were obtained with monoclonal antibody to
GFAP in place of rabbit anti-GFAP and negative results were obtained with an equivalent
concentration of nonspecific mouse IgG in place of monoclonal anti-GFAP.
Immunohistochemical staining detected expression of GFAP in greater than 80% of the
cells of cultures incubated 48 hours in medium with 10% calf serum, with over half of the
cells staining very strongly. Results were similar with cultures at 20 population doublings
and 60 population doublings, indicating that the medium is supportive of astrocyte
precursors over multipassage culture. Controls with normal rabbit serum in place of antiGFAP showed no stained cells.
95
DISCUSSION
We have applied the serum-free approach developed for mouse embryo cells to
human embryonic brain cells and achieved extended multipassage culture of human
astrocytes.
Although rodent astrocytes are routinely cultured short-term under
conventional, serum-containing or serum-free conditions (de Vellis et al., 1986; Morrison
and de Vellis, 1981; Morrison et al., 1985; Raff et al., 1983, 1984), extended culture of
these cells in serum containing medium results in spontaneous transformation and
alterations in expression of differentiated properties (Bressler and de Vellis, 1985).
Previously cell culture of human embryonic astrocytes under conventional conditions has
been limited to a few passages (Grenier et al., 1989; Rutka et al., 1987; Milsted et al.,
1990).
Several cultures were maintained for 60 to 70 population doublings before
proliferation ceased, providing a useful new culture model. This is sufficient proliferative
potential to allow long-term experiments such as transfection and selection of cells
expressing exogenous genes integrated into the genome or chronic exposure to chemical
agents and evaluation of toxic or carcinogenic effects. The population doubling levels
observed in this system are similar to those obtained under conventional conditions for
human fibroblasts and in serum-free medium for mammary epithelia: cells (Hayflick, 1965;
Martin et al., 1970; Hammond et al., 1984; Cristofalo and Stanuli.;-Praeger, 1982). For
long-term growth one of the cultures required at each passage the addition of a low
concentration of serum for a short period. The serum may provide undefined growth
factors, but may also contribute nutritional factors, adhesion proteins, or other factors that
may help the cells recover from the trypsinization (Barnes, 1987; Ham and McKeehan,
1979).
The medium employed for serum-free culture of human astrocytes was a
modification of the medium used for SFME cells; FGF was added in addition to the other
96
components and heparin also was added to stabilize FGF in the culture medium
(Gospodarowicz and Cheng, 1986). FGF is growth stimulatory for SFME cells in the
presence of the other factors (Loo et al., 1989b), and is a component of a serum-free
medium formulated specifically for astrocytes (Morrison et al., 1985; Morrison and de
Vellis, 1982). Of the three peptide growth factors in the serum-free medium--FGF, EGF,
and insulin--the human astrocyte cultures were most strongly stimulated to proliferate by
FGF. SFME cells are most responsive to EGF (Loo et al., 1987, 1989b).
SFME cells are markedly inhibited by serum in the culture medium (Loo et al.,
1987; Shirahata et al., 1990); multiple factors in serum may be responsible for inhibition of
SFME cell proliferation, including glucocorticoid and thyroid hormones (Loo et al., 1990).
Human embryonic astrocytes derived in the serum-free medium were not inhibited by
serum when it was added to medium containing the appropriate supplements used in the
serum-free formulation. We were not able to maintain multipassage cultures of human
embryonic astrocytes in medium supplemented only with serum, and it is unlikely that
initiation and continuous maintenance of the cultures in medium containing both calf serum
and the supplements of the serum-free medium would be successful. because this might be
expected to lead to growth of multiple cell types and possibly to overgrowth of fibroblasts.
Multiple serum factors have been shown to regulate expression of GFAP in rodent
astrocyte cultures (Bologa et al., 1988). Based on results with SFME cells (Sakai et al.,
1990), we anticipated that serum and TGF beta could influence expression of
differentiation on the human cells derived in the serum-free medium. Immunoassay of
fixed cells demonstrated an increase in GFAP in the human cells after treatment with TGF
beta or serum, and immunoblots with the same antiserum confirmed that the protein
recognized was of the appropriate molecular weight for GFAP. TGF beta may play a role
in control of differentiation of astrocytes in the developing human brain, and the growth
factor released from platelets at the site of a wound or other lesion also may contribute to
GFAP expression in the brain in response to trauma. The quantitatively greater response of
97
the cells to calf serum than to TGF beta in this assay is consistent with the view that other
factors in serum contribute to increased GFAP expression (Bologa et al., 1988).
Serum-free cultures from human embryonic brain are similar to serum-free mouse
astrocyte precursor cultures in TGF beta response and GFAP expression, but the human
cells differ from the mouse cells in ability to survive for longer periods in the absence of
EGF and FGF, in growth factor requirements for optimal proliferation, and in growth
response to serum. Most important, the serum-free human cultures, unlike SFME cells,
could not be maintained in vitro for periods beyond that observed with other human cell
types in conventional serum-containing medium. High-passage cells did not show a loss
of response to a specific mitogen, but instead were generally less responsive under all
growth conditions tested.
One hypothesis regarding "senescence" in vitro is that loss of proliferative potential
is related to increased differentiation of the cells with increased population doubling level
(Cristofalo and Stanulis-Praeger, 1982; Martin et al., 1970). In our cultures low- and
high-passage cells expressed GFAP at approximately equal levels and remained responsive
to TGF beta and serum in the expression of the differentiation marker, indicating that the
loss of proliferative potential with increasing population doublings was not correlated with
increased expression of differentiated function. The higher-passage cells were somewhat
less responsive to serum and TGF beta for expression of the differentiation marker than
were the lower-passage cells.
Several explanations exist for the difference in potential culture lifetime between
human and mouse serum-free cultures. Possibly the mouse cells have undergone an
undetected genomic alteration leading to "immortalization" and the human cells did not
undergo this alteration. It is well established that mouse cells in culture are much more
likely to experience spontaneous genetic changes than are human cells (Cristofalo and
Stanulis-Praeger, 1982; Martin et al., 1970; Todaro and Green, 1963). However, such a
change must be considerably more subtle than the change accompanying "immortalization"
98
in mouse embryo cells in conventional culture, because the conventional alterations lead to
gross karyotypic abnormalities (Todaro and Green, 1963; Loo et al., 1989a). In addition,
if this explanation is correct, the spontaneous genetic change leading to immortalization
occurs reproducibly in multiple initiations of cultures from several mouse strains (Loo et
al., 1987,1989a).
Another explanation, suggested by requirement of the human cells for additional
medium supplementation beyond that required for mouse cells, is that the medium used for
the human cells remains suboptimal, and additional medium formulations would allow
further growth of the human cultures. Finally, the differences in the behavior of the human
and mouse cultures with regard to serum inhibition of growth and growth factor responses
suggest that the cell type represented by the human cultures is related to, but not identical
with, SFME cells. It is possible that a human homologue of SFME cells remains to be
cultured and this cell type is capable of population doublings exceeding that reported in this
paper. It is also possible that the precise SFME cell type simply does not exist in humans;
significant differences exist in the developmental neurobiology of glial cells in these two
species (Choi, 1981; Levitt et al., 1983).
99
ACKNOWLEDGMENTS
This work was supported by NM-NAT-07560. D. Barnes is the recipient of
Research Career Development Award NIH-NCI-01226.
Yoshio Sakai performed the GFAP western immunoblot. Cathleen Rawson
assayed the response of embryonic human astrocytes to peptide growth factors in serumfree medium. David Barnes provided financial and scientific suppon
100
Fig. 5.1. Long-term culture of embryonic human astrocytes in serum-free
medium. Procedures for initiation and serial culture are described in the text. Population
doublings were calculated from passage dilution at each trypsinization. x x x and xx ,
replicate cultures from human brain (22 week-old embryo) initiated and carried
independently; x x -x
,
culture of human foreskin fibroblasts initiated and carried by
identical procedures.
60
40
POP
DBL 20
0
0
100
300
200
Days
400
500
101
Fig. 5.2. Photomicrograph of embryonic human astrocytes in serum-free
medium. Multipassage growth characteristics of the cells are shown in Figure 5.1.
Photograph was taken at 40 population doublings.
102
Fig. 5.3. Long-term culture of embryonic human astrocytes in medium
supplemented with low concentrations of calf serum. The culture was initiated
from human brain (24 week-old embryo) and carried for 20 population doublings in serum-
free medium identical with that used for the experiment of Figure 5.1
xxx
.
At this
point cultures were divided into three groups and serially passaged with the supplements of
the serum-free medium and the following: 0.5% calf serum added at the time of plating and
removed upon medium change the following day,
continuously in the medium,
medium,
.
1E51
.
++.
0.5% calf serum present
1.0% calf serum present continuously in the
Population doublings were calculated from passage dilution at each
trypsinization.
80
60
POP
40
D131.
20
0
0
100
200
Days
300
400
103
Fig. 5.4. Growth of embryonic human astrocytes in serum-free and serumsupplemented media. Cells were plated in complete serum-free medium with all
supplementary factors (SF), basal nutrient medium (F12:DME) supplemented only with
10% calf serum (10%CS), or complete serum-free medium supplemented additionally with
10% calf serum (CS + factors). Cells were plated at 105/35 mm-diameter plate and cell
number was determined by counting trypsinized cell suspensions 7 days after plating.
Medium was exchanged for fresh medium of the same composition 4 days after plating.
Open bars are date from cultures at 10 population doublings. Solid bars are data from same
cultures at 50 population doublings. Multipassage growth characteristics of the cells are
shown in Figure 5.1. Data are the average of cell counts from two replicate dishes.
Individual values for replicates did not vary more than 5% from the average.
30
SF
10%CS
CS+f actors
104
Fig. 5.5. Response of embryonic human astrocytes to peptide growth factors
in serum-free medium. Cells were plated in complete serum-free medium containing
all supplementary factors (SF) or in medium individually lacking EGF, FGF, or insulin (I)
or lacking both EGF and FGF. Cells were plated at 105/35 mm-diameter plate and cell
number was determined by counting trypsinized cell suspensions 7 days after plating.
Medium was exchanged for fresh medium of the same composition 4 days after plating.
Open bars are data from cultures at 10 population doublings. Solid bars are data from
cultures at 50 population doublings. Multipassage growth characteristics of the cells are
shown in Figure 5.1. Data are the average of cell counts from two replicate dishes.
Individual values for replicates did not vary more than 5% from the average.
SF
-EGF
-FGF
-EGF&FGF
-I
105
Fig. 5.6. Expression of GFAP in embryonic human astrocytes.
Serum-free
human astrocytes at 10 and 40 population doublings were incubated in the complete serum-
free medium as described (solid bars), in this medium supplemented with 10 ng/ml TGF
beta (crosshatched bars) or in basal nutrient medium (F12:DME) supplemented with 10%
calf serum (open bars). Cells were fixed after 2 days and assayed for GFAP as described
in the text. Control values from plates receiving normal rabbit serum in place of antiGFAP were less than 20% of that obtained with anti-GFAP and were subtracted from the
corresponding value with anti-GFAP in each case. Multipassage growth characteristics of
the cells are shown in Figure 5.1.
GFAP
(A405)
per
1
6
10 cells
10 POP DBLS
40 POP DBLS
106
Fig 5.7. Western immunoblot of GFAP in extracts of embryonic human
astrocytes in serum-free medium. Extracts were made from cultures after 10
population doublings in vitro. A, B: Extracts from two independent cultures probed with
anti-GFAP (multipassage growth characteristics of the cells are shown in Figs. 5.1, 5.3).
C: Blank. D: Human brain extract probed with anti-GFAP. E: Cell culture extract as in
A, probed with normal rabbit serum in place of anti-GFAP. F: Human brain extract as in
D, probed with normal rabbit serum in place of anti-GFAP. GFAP is sensitive to
degradation by intracellular proteinases, resulting in multiple bands detected in Western
blots for GFAP in control brain extracts (De Armond et al., 1983) and occasionally in cell
extracts.
68-_
'
.........,
4.1111or
43-
A
B
C
D
E
F
107
CHAPTER 6
Conclusion and Future Perspectives
CONCLUSION
The common thread binding this thesis together is the use of a serum-free approach
to the long-term culture of mammalian embryonic astrocytes. A serum-free approach to the
long-term culture of mouse embryo cells was undertaken based in part on the following
reasoning. Serum is not a physiologically relevent substance for most cell types since most
cells in vivo are not continually exposed to blood serum. Furthermore, serum is a very
complex material, and contains substances which may effect both cell function and cell
growth in vitro. The entry into and subsequent exit from growth crisis of mouse embryo
fibroblasts cultured in serum-containing medium may reflect in part an adaptation of the
cells to long-term culture in the presence of antagonistic serum factors. We reasoned that
by replacing serum with a set of def..:ed growth supplements we could avoid some of the
drawbacks of serum, and perhaps culture cells that were genetically and physiologically
more normal than cell lines previously derived using serum-containing media.
The derivation of SFME cells, detailed in chapter 2, forms the cornerstone of my
thesis. This research demonstrated for the first time that karyotypically normal mouse
108
embryo cells could be cultured long-term without growth crisis using a serum-free
approach. SFME cells derived using this approach displayed several unique properties.
The cells were dependent on EGF for survival and were reversibly growth inhibited by
serum. Later, research by Yoshio Sakai (1990) showed that SFME cells expressed glial
fibrillary acidic protein (GFAP), an intermediate filament protein expressed uniquely by
astrocytes, when grown in the presence of calf serum or transforming growth factor beta
(TGF beta). GFAP expression was reversible following removal of calf serum or TGF
beta.
Sakai's work suggests SFME cells represent astrocyte precursor cells or
proastroblasts. SFME cells may provide a unique cell culture model to study the biology
of astrocyte precursor cells. In addition, the long-term culture of SFME cells provides
evidence that normal cells can be cultured long-term under the proper environmental
conditions.
Following the derivation of SFME cells my thesis research branched into several
projects based on observations made during the long-term culture of SFME cells. The
experiments I conducted were aimed at answering the following questions. First, what
type of cell do SFME cells represent in vivo, and what are the optimal growth conditions
for the cells? Second, can the serum-free approach used for the long-term culture of SFME
cells be extended to the long-term culture of human embryonic astrocytes? Third, by what
mechanism are SFME cells growth inhibited by serum? Fourth, by what mechanism does
EGF deprivation cause SFME cell death?
Chapter 3 detailed the growth responses of SFME cells in vitro. This work
examined precisely the growth requirements of SFME cells in vitro and detailed the cell
culture methods used to derive SFME cells. The concentration dependence of the growth
stimulatory effects of protein supplements used in the serum-free medium formulation was
determined. A survey of other supplements that might act as alternative or complementary
additions to the culture medium was also done. Three main conclusions of this work were
(1) transforming growth factor alpha (TGF alpha), at equivalent concentrations, replaced
109
EGF for SFME cell survival and growth, (2) fibroblast growth factor partially replaced
EGF and allowed for survival and slow growth of SFME cells, and (3) the viability of
SFME cells was reduced when cells were deprived of EGF for as little as 8 hours.
In chapter 4 and appendix 1 the inhibition of SFME cell growth by calf serum was
explored. Research detailed in chapter 4 showed that SFME cells were reversibly growth
inhibited by low concentrations of calf serum or plasma, and cells arrested in the G1 phase
of the cell cycle. Additionally, charcoal extraction, a procedure that removes hydrophobic
molecules, removed a portion of the inhibitory activity from serum. Both glucocorticoid
hormones and thyroid hormones (T3 and T4), molecules that would be removed by
charcoal, were growth inhibitory for SFME cells. The inhibitory activity was specific,
reversible, and synergistic with charcoal-extracted serum. The primary conclusion from
these results was that glucocorticoid or thyroid hormones may contribute to the inhibitory
activity of serum on SFME cells, but additional unidentified inhibitory factors were also
involved and remain to be identified.
Appendix 1 reported that SFME cells cultured in serum-containing medium arrested
in the G1 phase of the cell cycle, with a greatly reduced rate of incorporation of precursors
into DNA and thymidine kinase activity. A reduction in rate of incorporation of amino
acids into protein was not observed.
These observations are associated with
physiologically relevant inhibition of cell growth, and are consistent with growth inhibition
associated with induction of cell differentiation. In addition, appendix 1 revealed that
variant SFME cells capable of growing in serum-containing medium were selected in
medium containing 10% calf serum using a 3T3 protocol. Proliferation of variant cells was
observed with a frequency of about 1 in 106 SFME cells following a one-month lag of no
cell growth, similar to the growth crisis period observed with serum-derived mouse
fibroblasts. Karyotypic analysis revealed that variant SFME cells capable of growing in
serum displayed abnormal karyotypes with a strong tendency toward hyperploidy,
suggesting that serum-containing medium selected for, or induced, an abnormal karyotype
110
in SFME cells. While gene dosage may play a role in release from serum inhibition, SFME
cells made polyploid by cell fusion remained inhibited by serum. In addition, close
examination of the relationship between exit from growth crisis and karyotype shift from a
diploid to a tetraploid karyotype of the population of serum-derived mouse embryo cells
suggested that cells exited from growth crisis one or two generations before the karyotype
switch occurred. These observations suggest that exit from growth crisis of serum-grown
mouse embryo cells resulted from one or more changes within the cell, either genetic or
epigenetic, which allow the cells to be refractile to inhibition by serum. In addition, the
change appeared to reduce the cells ability to properly maintain a normal karyotype.
Chapter 5 detailed research addressing whether SFME-like cells could be cultured
from human embryonic brain. This chapter described the long-term growth of astrocyte
cells from human embryonic brain with properties common to SFME cells. The human
astrocytes displayed a morphology similar to SFME cells in culture and expressed GFAP
when grown in the presence of calf serum or TGF beta. However, the human astrocytes
differed from SFME cells in ability to survive for longer periods in the absence of EGF and
FGF, in growth factor requirements for optimal proliferation, and in growth response to
serum. The human astrocytes also expressed detectable amounts of GFAP when grown
serum-free. Most important, the serum-free human cultures, unlike SFME cells, could not
be maintained in vitro for periods beyond that observed with other human cell types in
conventional serum-containing medium.
Several explanations were proposed in chapter 5 for the difference in potential
culture lifetime between the human and mouse serum-free cultures. The mouse cells may
have undergone an undetectable genetic change which resulted in the ability of the cells to
avoid growth crisis. The genome of human cells is more stable than rodent cells in vitro,
so genetic changes required for human cells to escape growth crisis may occur less
frequently. The medium formulation used to culture the human astrocytes may have been
suboptimal, and additional formulations might have allowed for further growth. The
111
growth stimulatory effect of calf serum added to cells growing in the serum-free medium
suggests other growth stimulatory factors exist. The human astrocytes may be a related but
not identical cell type to SFME cells. It is possible that SFME-type cells remain to be
cultured from embryonic human brain, or perhaps SFME-type cells do not exist in
embryonic human brain.
Finally, appendix 2 described research directed at understanding the mechanism by
which EGF deprivation led to death of SFME cells. Electron microscopic examination of
SFME cells showed that cells became small and severely degenerate. growth arrested in the
G1 -phase of the cell cycle, and displayed characteristic pyknotic nuclei following removal
of EGF. Chromatin isolated from cultures undergoing EGF deprivation-dependent cell
death exhibited a pattern of fragmentation suggesting that DNA was cleaved by activated
nucleases at regions unprotected by histones. At the same time mRNA from these cells
remained intact. Many of these characteristics have been associated with cell death in
several growth factor-dependent systems, both in vitro and in vivo (refer to introduction).
It is well established that apoptosis, or programmed cell death, is required for the normal
development and survival of all multicellular organisms.
It has been suggested that cell
death, following growth factor withdrawal in growth factor-dependent systems, may occur
by apoptosis.
It has been further suggested, based on results from experiments using
RNA and protein synthesis inhibitors, that apoptosis, or programmed cell death in these
growth factor-dependent systems, required the synthesis of RNA and protein. SFME cell
death following EGF withdrawal is blocked by inhibitors of RNA and protein synthesis.
The general conclusion of appendix 2 is that SFME cells, when deprived of EGF, rapidly
entered a cell death program, termed apoptosis, that is morphologically, physiologically,
and biochemically distinct from necrosis.
112
FUTURE PERSPECTIVES
Research using SFME cells falls into two broad categories. One, experiments may
be designed to answer questions about SFME cells, in order to expand our knowledge of
the biology of SFME cells. Two, experiments may be designed where SFME cells are
used as an in vitro model to study a particular phenomenon. These two categories of
experiments often overlap. I believe it is important that experiments involving SFME cells
continue to pursue both lines of research. In the next several paragraphs I will briefly pose
some broad questions that I feel are important to follow. Some fall into category one,
others into category two, and still others somewhere in between. I will then briefly
propose some ways in which these questions might be pursued.
First, how do SFME contribute to the adult mouse brain? It would be valuable to
be able to show that SFME cells contribute to the normal development of the mouse brain,
and would open up numerous research avenues. One direct approach would be to create
SFME cells that express a specific, detectable markers, such as beta galactosidase or
luciferase. Once available, the cells could be transplanted into the brains of mice at various
stages of development. The mouse brain is not fully developmed at birth, and newborn
mice may prove to be the best candidate. Progress of the transplanted SFME cells may be
followed by sacrificing mice later in development and assaying thin tissue slices for the
specific marker expressed by the SFME cells. At the same time, geneticin (G -418)resistant SFME cells could be transplanted into similar mice, and later in development the
brain could be recovered and the presence of G-418-resistant cells assayed for in culture
using G-418-containing serum-free medium.
If transplanted SFME cells contribute to normal brain development then the role of
SFME cells in vivo may be addressed. For instance, how do SFME cells contribute to
reactive gliosis? Do SFME cells participate in the repair of brain lesions? Evidence
suggests that lesioning of the rat brain results in an increase in membrane EGF receptor
113
content of astrocytes adjacent to the lesion, and a decrease in an EGF receptor-like astrocyte
mitogen inhibitor (Nieto-Sampedro, 1988). In addition, macrophages and microglia are
attracted to lesioned areas of the brain, and tissue plasminogen activator (TPA) protein has
been found associated with lesioned areas of the brain. Both low density lipoprotein (LDL)
receptors and TPA contain EGF-like domains and may promote SFME cell growth.
Finally, TGF beta may be released from platelets at the site of the wound and promote
astrocyte differentiation. An interplay between growth inhibitory and growth stimulatory
agents may modulate, perhaps through the EGF receptor, the response of astrocytes to
injuries to the brain.
Second, what are the additional growth inhibitory factors in serum, and by what
mechanism do they cause growth inhibition of SFME cells? Purifying the inhibitory
components from serum using biochemical techniques has not been very successful. One
approach that may yield information about the mechanism of serum inhibition is to use the
technique of differential cloning to identify genes that are regulated by serum. This
approach may lead to the identification of genes that regulate SFME cell growth and may
also identify cell-specific markers of astrocyte precursor cells. Another approach might be
to examine the signal transduction pathway of the EGF receptor in SFME cells following
exposure to serum. Although the level of membrane-associated EGF receptors is similar
between SFME cells growth serum-free or in serum-containing medium, the
phosphorylation state of the receptor may be altered, or the EGF receptor complex as a
whole may be altered. Finally, gene expression between normal SFME cells and serum-
selected SFME cells may be compared in order to examine whether gene expression of
serum-selected SFME cells differs from normal SFME cells.
Third, do genes repressed by EGF exist in SFME cells, and do they participate in
the EGF-deprivation dependent cell death program? One approach would be to continue
using differential cloning techniques to identify genes that are specifically expressed
following EGF withdrawal. Although I was unable to identify EGF-repressed genes 15
114
hours following EGF withdrawal, examining gene expression at earlier time points
following growth factor removal may prove more successful. Earlier time points may
avoid artifacts, such as the differential accumulation of fragmented DNA, resulting from
cellular degradation as the cells undergo apoptosis. Another approach might be to examine
the EGF receptor signal transduction pathway of SFME cells grown in the presence and
absence of EGF. For instance, blocking points in the signal transduction may lead to
SFME cell death and identify points of the pathway important for EGF-repression of
programmed cell death.
Finally, the serum-free approach used for the long-term culture of both mouse and
human embryonic astrocytes may be extended to the long-term culture of astrocytes
recovered from individuals afflicted with degenerative brain disorders, including
Alzheimer's disease and Parkinson's disease. Comparisons between cultures of astrocyte
cells from normal and diseased tissues may provide important information about the role
astrocytes may play in the disease state.
115
BIBLIOGRAPHY
Aaronson, S.A. and Todaro, G.J. 1968. Development of 3T3-like lines from Balb/c
mouse embryo cultures: transformation susceptibility to SV40. J. Cell Phys. 72:141-148.
Ambesi-Impiombato,F.S., Parks, L.A.M. and Coon H.G. 1980. Culture of hormonedependent functional epithelial cells from rat thyroids. Proc. Natl. Acad. Sci. USA.
77:3455-3459.
Arends, M.J., Morris, R.G. and Wyllie, A.H.
1990.
Apoptosis:
the role of the
endonuclease. Amer. J. Path. 136:593-608.
Augusti-Tocco, G. and Sato, G. 1969. Establishment of functional clonal lines of
neurons from mouse neuroblastoma. Proc. Natl. Acad. Sci. USA. 64:311-315.
Barinaga, M. 1990. Technical advances power neuroscience. Science. 250:908-909.
Barnes, D. 1987. Serum-free animal cell culture. BioTechniques. 5:1-9.
Barnes, D. and Sato, G. 1980a. Methods for growth of cultured cells in serum-free
medium. Anal. Biochem. 102:255-270.
Barnes, D. and Sato, G. 1980b. Serum-free cell culture: a unifying approach. Cell.
22:649-655.
Barnes, D. and Silnutzer, J. 1983. Isolation of human serum spreading factor. J. Biol.
Chem. 258:12548-12552.
1989. Effects of
oncogenes on EGF requirement and serum inhibition of serum-free derived mouse embryo
(SFME) cells. J. Cell Biol. 107:4452 (abstract).
Barnes, D., Loo, D., Shirahata, S., Rawson, C. and Chang, Y.
Barnes, D.W. and Reed, D.J. 1989. Cytotoxicity and cell culture conditions. In: Milo,
G.E. (ed): "Transformation of Human Diploid Fibroblasts."
Barnes, D.W., Silnutzer, J., See. C. and Shaffer, M. 1983. Characterization of human
serum spreading factor with monoclonal antibodies. Proc. Natl. Acad. Sci. USA.
80:1362-1366.
Bartlett, P.F., Reid, H.H., Bailey, K.A. and Bernard, 0. 1988. Immortalization of
mouse neural precursor cells by the c-myc oncogene. Proc. Natl. Acad. Sci. USA.
85:3255-3259.
Baserga, R., Waechter, D.E., Soprano, K.J. and Galanti, N. 1982. Molecular biology of
cell division. Ann. N.Y. Acad. Sci. 397:110-120.
Beato, M. 1989. Gene regulation by steroid hormones. Cell. 56:335-344.
Beck, W.S. 1982. Hypochromic anemias I. Iron deficiency and excess. In: Beck, W.S.
(ed): "Hematology." MIT Press, Cambridge, MA. 102-103.
116
Benda, P., Lightbody, J., Sato, G., Levine. L. and Sweet, W. 1969. Differentiated rat
glial cell strain in tissue culture. Science. 161:370-371.
Bettger, W.J. and Ham, R.G. 1982. The nutrient requirements of mammalian cells grown
in culture. Adv. Nutr. Res. 4:249-286.
Bettger, W.J., Boyce, S.T., Walthall, B.J. and Ham, R.G. 1981. Rapid clonal growth
and serial passage of human diploid fibroblasts in a lipid-enriched synthetic medium
supplemented with epidermal growth factor, insulin, and dexamethasone. Proc. Natl.
Acad. Sci. USA. 78:5588-5592.
Bettuzzi, S., Hiipakka, R.A., Gilna, P. and Liao, S. 1989. Identification of an androgenrepressed mRNA in rat ventral prostate as coding for sulphated glycoprotein 2 by cDNA
cloning and sequence analysis. Biochem. J. 257:293-296.
Bishop, J.M. 1987. The molecular genetics of cancer. Science. 235:305-311.
Bologa, L., Cole, R., Chiappelli, F., Saneto, R.P. and de Vellis, J. 1988. Expressoin of
glial fibrillary acidic protein by differentiated astrocytes is regulated by serum antagonistic
factors. Brain Res. 457:295-302.
Bottenstein, J.E. 1985. Growth and differentiation of neural cells in defined media. In:
Bottenstein, J.E. and Sato, G. (eds): "Cell Culture in the Neurosciences." Plenum Press,
NY.
Bottenstein, J.E. 1986. Growth requirements in vitro of oligodendrocyte cell lines and
neonatal rat brain oligodendrocytes. Proc. Natl. Acad. Sci. USA. 83:1955-1959.
Bourgeois, S. and Gasson, J.C. 1985. Genetic and epigenetic bases of glucocorticoid
resistance in lymphoid cell lines. Biochem. Actions Horm. 12:310-351.
Bradford, M.M. 1976. A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248254.
Brandi, M.L., Fitzpatrick, L.A., Coon, H.G. and Aurbach, G.D.
1986. Bovine
parathyroid cells: cultures maintained for more than 140 population doublings. Proc. Natl.
Acad. Sci. USA. 83:1709-1713.
Bressler, J.P. and de Vellis, J. 1985. Neoplastic transformation of newborn rat astrocytes
in culture. Brain Res. 348:21-27.
Brooks, R.F. 1975. Growth regulation in vitro and the role of serum. In: Allison, A.C.
(ed): "Structure and Function of Plasma Proteins." Plenum, NY, 1. 112.
Brunner, G., Lang, K., Wolfe, R.A., McClure, D.B. and Sato, G.H. 1982. Selective
cell culture of brain cells by serum-free, hormone-supplemented media: a comparative
morphological study. Dev. Brain Res. 2:563-575.
Buttyan, R., Zakeri, Z., Lockshin, R. and Wolgemuth, D. 1988. Cascade induction of cfos, c-myc, and heat shock 70K transcripts during regression of the rat ventral prostate
gland. Mol. Endocrin. 2:650-657.
117
Carpenter, G. and Cohen, S. 1979. Epidermal growth factor. Annu. Rev. Biochem.
48:193-216.
Carrel, A. 1912. On the permanent life of tissues outside of the organism. J. Exp. Med.
15:516-528.
Cherington, P.V., Smith, B.L. and Pardee, A.B. 1979. Loss of epidermal growth factor
requirement and malignant transformation. Proc. Natl. Acad. Sci. USA. 76:3937-3941.
Chiang, L.-C., Silnutzer, J., Pipas, J.M. and Barnes, D.B. 1985.
Selection of
transformed cells in serum-free media. In Vitro Cell. Dev. Biol. 21:707-712.
1981. Radial glia of developing human fetal spinal cord: golgi,
immunohistochemical and electron microscope study. Dev. Brain Res. 1:249-267.
Choi, B.H.
Clemens, M.J. and mcNurlan, M.A.
1985.
Regulation of cell proliferation and
differentiation by interferons. Biochem. J. 226:345-350.
Cohen, J.J. and Duke, R.C. 1984. Glucocorticoid activation of a calcium-depenent
endonuclease in thymocyte nuclei leads to cell death. J. Immunology. 132:38-42.
Conn, P.M. 1990. Methods in neurosciences. Volume 2.
Cristofalo, V.J. and Stanulis-Praeger, B.M. 1982. Cellular senescence in vitro. Adv.
Cell Culture. 2:1-53.
Darmon, M., Serrero, G., Rizzino, A. and Sato, G. 1981. Isolation of myoblastic,
fibroadipogenic and fibroblastic clonal cell lines from a common precursor and study of
their requirement for growth and differentiation. Exp. Cell Res. 132:313-327.
de Vellis, Wu, D. and Kumar, S. 1986. Enzyme induction and regulation of protein
synthesis. In: Fedoroff, S. and Vernadakis, A. (eds): "Biochemistry, Physiology, and
Pharmacology of Astrocytes." Academic Press, NY. 2:209-237.
DeArmond, S.J., Fajardo, M. and Eng, L.F. 1983. Degradation of glial fibrillary acidic
protein by a calciium dependent proteinase: an electoblot study. Brain Res. 262:275-282.
Demars, R. 1958. The inhibition by glutamine of glutamyl transferase formation in
cultures of human cells. Biochim. et Biophys. Acta. 27:435-436.
Downward, J., Yarden, Y., Mayer., E., Scrace, G., Totty, N., Stockwell, P., Ullrich,
A., Schlessinger, J. and Waterfield, M.D. 1984. Close similarity of epidermal growth
factor receptor and v-erb-B oncogene protein sequences. Nature. 307:521-527.
Duke, R.C. and Cohen, J.J. 1986. IL-2 addiction: withdrawl of growth factor activates a
suicide program in dependent T cells. Lymphokim Res. 5:289-299.
Dulbecco, R. and Elkington, J. 1973. Conditions limiting multiplication of fibroblastic
and epithelial cells in dense cultures. Nature. 246:197-199.
Eagle, H. 1955a. Nutrition needs of mammalian cells in tissue culture. Science. 122:501504.
118
Eagle, H. 1955b. The minimum vitamin requirements of the L and He La cells in tissue
culture, the production of specific vitamin deficiencies, and their cure. J. Exp. Med.
102:595-600.
Eagle, H. 1955c. The specific amino acid requirements of a human carcinoma cell (strain
He La) in tissue culture. J. Exp. Med. 102:37-48.
Eagle, H. 1955d. The specific amino acid requirements of a mammalian cell (strain L) in
tissue culure. J. Biol. Chem. 214:839-853.
Eagle, H. 1959. Amino acid metabolism in mammalian cell cultures. Science. 130:432437.
Eagle, H., Oyama, V.I., Levy, M., Horton, C.L. and Fleischman, R. 1956. The growth
response of mammalian cells in tissue culture to L-glutamate and L-glutamic acid. J. Biol.
Chem. 218:607-617.
Earle, W.R., Schilling, E.L., Stark, T.H., Straus, N.P., Brown, M.F. and Shelton, E.
1943. Production of malignancy in vitro. IV. the mouse fibroblast cultures and changes
seen in the living cells. J. Natl. Cancer Inst. 4:165-212.
Ernst, T., Jackson, C. and Barnes, D. 1990. Factors influencing karyotypic stability of
serum-free mouse embryo (SFME) cells. Cytotech. in press.
Evans, R.M. 1988. The steroid and thyroid hormone receptor superfamily. Science.
240:889-895.
Evans, V.J., Bryant, J.C., Fioramonti, M.C., McQuilkin, W.T., Sanford, K.K., and
Earle, W.R. 1955. Studies of nutrient media for tissue cells in vitro. Cancer Research.
16:77-86.
Farber, R.A. and Liskay, R.M. 1974. Karyotypic analysis of a near-diploid established
mouse cell line. Cytogenet. Cell Genet. 13:384-396.
Galiana, E., Borde, I., Marin, P., Rassoulzadegan, M., Cuzin, F., Gros, F., Rouget, P.
and Evrard, C. 1990. Establishment of permanent astroglial cell lines, able to differentiate
in vitro, from transgenic mice carrying the polyoma virus large T gene: an alternative
approach to brain cell immortalization. J. Neurosci. Res. 26:269-277.
Gey, G.O., Coffman, W.D. and Kubicek, M.T. 1952. Tissue culture studies of the
proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res. 12:264265.
Giguere, L., Cheng, J. and Gospodarowicz, D. Factors involved in the control of
proliferation of bovine corneal endothelial cells maintained in serum-free medium. J. Cell
Phys. 110:72-80.
Gospodarowicz, D. 1984. Preparation and uses of lipoproteins to culture nortnal diploid
and tumor cells under serum-free conditions. In: Barnes, D.W., Sirbasku, D.A. and Sato,
G.H. (eds): "Cell Culture Methods for Molecular and Cell Biology." Alan R. Liss, NY.
1:69-88.
Gospodarowicz, D. and Cheng, J. 1986. Heparin protects basic and acidic FGF from
inactivation. J. Cell Phys. 28:475-484.
119
1981. Factors
controlling the proliferative rate, final cell density and life span of bovine vascular smooth
muscle cells in culture. J. Cell Biol. 89:568-578.
Gospodarowicz, D., Hirabayashi, K., Giguere, L. and Tauber, H.
Gray, J.W. and Coffin, P. 1979. Flow cytometry. Methods Enzymol. 58:233-238.
Greene, L.A. and Tisch ler, A.S. 1976. Establishment of a noradrenergic clonal line of rat
adrenal pheochromocytoma cells which respond to nerve growth factor. Proc. Natl. Acad.
Sci. USA. 73:2424-2428.
Grenier, Y., Ruijs, T.C.J., Olivier, A., Robitaille, Y. and Antel, J.P. 1989. DNA
synthesis by young adult human-derived astrocytes in vitro. Brain Res. 480:87-91.
1989. Thymidine kinase synthesis is repressed in
nonreplicating muscle cells by a translational mechanism that does not affect the polysome
distribution of thymidine kinase mRNA. Proc. Natl. Acad. Sci. USA. 86:4987-4991.
Gross, M.K. and Merrill, G.F.
Grotendorst, G., Kleinman, H., Rohrbach, D., Hewitt, A., Varner, H., Horigan, E.,
Hassel, J., Terranova, V. and Martin, G. 1982. Role of attachment factors in mediating
the attachment, distribution, and differentiation of cells. In: Sato,G. and Sirbasku, D.
(eds): "Cold Spring Harbor Conference on Cell Proliferation." Cold Spring Harbor Press,
Cold Spring Harbor, NY, 9:403-413.
Ham, R.G. 1963. An improved nutrient solution for diploid chinese hamster and human
cell lines. Exp. Cell Research. 29:515-526.
Ham, R.G. 1965. Clonal growth of mammalian cells in a chemically defined, synthetic
medium. Proc. Natl. Acad. Sci. USA. 53:288-293.
Ham, R.G. and McKeehan, W.L. 1979. Media and growth requirements. Methods
Enzymol. 58:44-93.
Hamburger, V., Brunso-Bechtold, J.K. and Yip, J.W. 1981. Neuronal death in the
spinal ganglia of the chick embryo and its reduction by nerve growth factor. J. Neurosci.
1:60-71.
Hammond, S.L., Ham, R.G. and Stampfer, M.R. 1984. Serum-free growth of human
mammary epithelial cells: rapid clonal growth in defined medium and extended serial
passage with pituitary extract. Proc. Natl. Acad. Sci. USA. 81:5435-5439.
Harrison, R.G. 1907. Observations on the living developing nerve fiber. Proc. Soc.
Exp. Biol. Med. 4:140-143.
Hashizume, S., Kuroda, K. and Murakami, H. 1984. Cell culture assay of biological
activity of lactoferrin and transferrin. Methods Enzymol. 147:302-314.
Hayashi, I. and Sato, G. 1976. Replacement of serum by hormones permits growth of
cells in defined medium. Nature. 259:132-134.
Hayat, M.A. 1986. Basic techniques for transmission electron microscopy. Academic
Press, Orlando, FL. 226.
120
Hayflick, L. 1965. The limited in vitro lifetime of human diploid cell strains. Exp. Cell
Rsearch. 37:614-636.
Hayflick, L. and Moorhead P.S. 1961. The serial cultivation of human diploid cell
strains. Exp. Cell Research. 25:585-621.
Healy, G.M., Fisher, D.C. and Parker, R.C. 1954. Nutrition of animal cells in tissue
culture. IX. synthetic medium No. 703. Can. J. Biochem. Physiol. 32:327-337.
Hinchliffe, J.R. 1981. Cell death in embryogenesis. In: Bowen, I.D. and Lockshin,
R.A. (eds): "Cell Death in Biology and Pathology." Chapman and Hall Press, London,
NY, 35-78.
Hoshi, H. and McKeehan, W.L. 1984. Brain cell derived and liver cell derived factors are
required for growth of human endothelial cells in serum-free culture. Proc. Natl. Acad.
Sci. USA. 81:6413-6417.
Hughes, S.M., Lillien, L.E., Raff, M.C., Rohrer, H. and Sendtner, M. 1988. Ciliary
neurotrophic factor induces type-2 astrocyte differentiation in culture. Nature 335:70-73.
Isofort, R.J. 1990. Frequency and mechanisms of factor independence in IL-3-dependent
cell lines. Somatic Cell Mol. Genet. 16:109-121.
Ito, F., Takii, Y., Sizuki, J. and Masamune, Y. 1982. Reversible inhibition by human
serum lipoproteins of cell proliferation. J. Cell Phys. 13:1-13.
Jainchill, J.L., Aaronson, S.A. and Todaro, G.J. 1969. Murine sarcoma and leukemia
viruses: assay using clonal lines of contact-inhibited mouse cells. J. Virol. 4:549-553.
Kaiser, N. and Edelman, I.S. 1977. Calcium dependence of glucocorticoid-induced
lymphocytolysis. Proc. Natl. Acad. Sci. USA. 74:638-642.
Kan, M. and Yamane, I. 1982. In vitro proliferation and lifespan of human diploid
fibroblasts in serum-free BSA-containing medium. J. Cell. Phys. 111:155-162.
Kerr, J.F.R. and Searle, J. 1973. Deletion of cells by apoptosis during castration-induced
involution of the rat prostate. Virchows Arch. 13:87-102.
Klagsbrun, M., Sullivan, R., Smith, S., Rybka, R. and Shing, Y. 1987. Purification of
endothelial cell growth factors by heparin affinity chromatography. Methods Enzymol.
147:95-106.
Koenig, R.J., Lazar, M.A., Hodin, R.A., Brent, G.A., Larsen, P.R., Chin, W.W. and
Moore, D.D. 1989. Inhibition of thyroid hormone action by a non-hormone binding cerbA protein generated by alternative mRNA splicing. Nature. 337:659-661.
Kohase, M., May, L., Tamm, I., Vilcek, J. and Sehgal, P. 1987. A cytokine network in
human diploid fibroblasts: interactions of beta interferons, tumor necrosis fator, plateletderived growth factor and interleukin 1. Mol. Cell. Biol. 7:273-280
Koury, M.J. and Bondurant, M.C. 1990. Erythropoietin retards DNA breakdown and
prevents programmed death in erythroid progenitor cells. Science. 248:378-381.
121
Kyprianou, N. and Isaacs, J.T. 1988. Activation of programmed cell death in the rat
ventral prostate after castration. Endocrinology. 122:552-562.
Land, H., Parada, L.F. and Weinberg, R.A. 1983. Tumorigenic conversion of primary
embryo fibroblasts requires at least two cooperating oncogenes. Nature. 304:596-602.
Lazarides, E. 1982. Intermediate filaments: a chemically heterogeneous, developmentally
regulated class of proteins. Annu. Rev. Biochem. 51:219-250.
Leffert, H.L. and Weinstein, D.B. 1976. Growth control of differentiated fetal rat
hepatocytes in primary monolayer culture IX. Specific inhibition of DNA synthesis
initiation by very low density lipoprotein and possible significance to the problem of liver
regeneration. J. Cell. Biol. 70:20-27.
Leger, J.G., Montpetit, M.L. and Tenniswood, M.P. 1987. Characterization and cloning
of androgen-repressed mRNAs from the rat ventral prostate. Biochem. Biophys. Res.
Comm. 147:196-203.
Levitt, P., Cooper, M.L. and Rakic, P. 1983. Early divergence and changing proportions
of neuronal and glial precursor cells in primate cerebral ventricular zone. Dev. Biol.
96:472-484.
Lewis, R.L. and Lewis W.H. 1911. The cultivation of tissues from chick embryos in
solutions of NaCl, CaC12, KCl and NaHCO3. The Anatomical Record. 5:277-285.
Lil lien, L.E. and Raff, M.C. 1990a. Analysis of the cell-cell interactions that control type2 astrocyte development in vitro. Neuron. 4:525-534.
Li llien, L.E. and Raff, M.C. 1990b. Differentiation signals in the CNS: type-2 astrocyte
development in vitro as a model system. Neuron. 5:111-119.
Lillien, L.E., Sendtner, M., Rohrer, H., Hughes, S.M. and Raff, M.C. 1988. Type-2
astrocyte development in rat brain cultures is initiated by a CNTF-like protein produced by
type-1 astrocytes. Neuron. 1:485-494.
Lippman, M., Bolan, G. and Huff, K.
1976.
The effects of glucocorticoids and
progesterone on hormone-responsive human breast cancer in long-term tissue culture.
Cancer Res. 36:4602-4609.
Loo, D., Rawson, C., Helmrich, A. and Barnes, D. 1989b. Serum-fee mouse embryo
(SFME) cells: growth responses in vitro. J. Cell Phys. 139:484-491.
Loo, D., Rawson, C., Schmitt, M., Lindburg, K. and Barnes, D. 1990. Glucocorticoid
and thyroid hormones inhibit proliferation of serum-free mouse embryo (SFME) cells. J.
Cell Phys. 142:210-217
Rawson, C.L. and Barnes, D.W. 1987. Extended culture of
Loo, D.T., Fuquay,
mouse embryo cells without senescence: inhibition by serum. Science. 236:200-202.
Loo, D.T., Rawson, C.L., Ernst, T., Shirahata, S. and Barnes, D.W. 1989a. Serum-free
primary culture of mouse embryo cells. In Baserga, R. (ed): "Cell Growth and Cell
Division: A Practical Approach." London: IRL Press, pp17-34.
122
Loo, D.T., Sakai, Y., Rawson, C.L. and Barnes, D.W. 1991. Serial passage of
embryonic human astrocytes in serum-free, hormone-supplemented medium. J. Neurosci.
Res. 28:101-109.
Maniatis, T., Fritsch, E.F. and Sambrook, J. 1982. Molecular cloning, a laboratory
manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.
Marquardt, H., Hunkapiller, M.W., Hood, L.E. and Todaro, G. 1984. Rat transforming
growth factor type 1: structure and relation to epidermal growth factor. Science.
223:1079-1082.
Martin, D.P., Schmidt, R.E., Di Stefano, P.S., Lowry, 0.H., Carter, J.G. and Johnson,
E.M. 1988. Inhibitors of protein synthesis and RNA synthesis prevent neuronal death
caused by nerve growth factor deprivation. J. Cell Biol. 106:829-844.
Martin, G.M., Sprague, C.A. and Epstein, C.J. 1970. Replicative life-span of cultivated
human cells. Lab. Invest. 23:86-92.
Massague, J. 1987. The TGF-beta family of growth and differentiation factors. Cell.
49:437-438.
Masui, T., Wakefield, L.M., Lechner, J.F., LaVeck, M.A., Sporn, M.B. and Harris,
C.C. 1986. Type beta transforming growth factor is the primary differentiation-inducing
serum factor for normal human bronchial epithelial cells. Proc. Natl. Acad. Sci. USA.
83:2438-2442.
Mather, J.P. and Sato, G.H. 1979. The use of hormone-supplemented serum-free media
in primary cultures. Exp. Cell Res. 124:215-221.
McConkey, D.J., Nicotera, P., Hartzell, P., Bellomo, G., Wyllie, A.H. and Orrenius, S.
1989. Glucocorticoids activate a suicide process in thymocytes through an elevation of
cytosolic calcium concentration. Arch. Biochem. Biophys. 269:365-370.
McKay, R.D.G. 1989. The origins of cellular diversity in the mammalian central nervous
system. Cell. 58:815-821.
McKeehan, W. and Ham, R. 1976. Stimulation of clonal growth of normal fibroblasts
with substrata coated with basic polymers. J. Cell Biol. 71:727-734.
McKeehan, W.L., Genereux, D.P. and Ham, R.G. 1978. Assay and partial purification
of factors from serum that control multiplication of human diploid fibroblasts. Biochem.
Biophys. Res. Comm. 80:1013-1021.
Merrill, G.F., Hauschka, S.D. and McKnight, S.L. 1984. tk enzyme expression in
differentiating muscle cells is regulated through an int,mal segment of the cellular tic gene.
Mol. Cell. Biol. 4:1777-1784.
Miller, R.H., David, S., Patel., R., Abney, E.R. and Raff, M.C. 1985. A quantitative
immunohistochemical study of macroglial cell development in the rat optic nerve: in vivo
evidence for two distinct astrocyte lineages. Dev. Biol. 111:35-41.
Milsted, A., Barna, B.P. Ronsohoff, R.M., Brosnihan, K.B. and Ferrario, C.M. 1990.
Astrocyte cultures derived from human brain tissue express angoitensinogen mRNA.
Proc. Natl. Acad. Sci. USA. 87:5720-5723.
123
Moguilewsky, M. and Philibert, D. 1984. RU38486: potent antiglucocorticoid activity
correlated with strong binding to the cytosolic glucocorticoid receptor followed by an
impaired activation. J. Steroid Biochem. 20:271-276.
Montpetit, M.L., Lawless, K.R. and Tenniswood, M.
1986.
Androgen-repressed
messages in the rat ventral prostate. The Prostate. 8:25-36.
Morgan. J.F., Morton, H.J. and Parker, R.C. 1950. Nutrition of animal cells in tissue
culture. I. inital studies on a synthetic medium. Proc. Soc. Exp. Biol. Med. 73:1-8.
Morrison, R.S. and de Vellis, J. 1981. Growth of purified astrocytes in a chemically
defined medium. Proc. Natl. Acad. Sci. USA. 78:7205-7209.
Morrison, R.S. and de Vellis, J. 1984. Preparation of a chemically defined medium for
purified astrocytes. In: Barnes, D.W., Sirbasku, D.A. and Sato, G.H. (eds): "Cell
Culture Methods for Molecular and Cell Biology." Alan R. Liss, NY. 4:15-22.
Morrison, R.S., de Vellis, J., Lee, Y.L., Bradshaw, R.A. and Eng, L.F.
1985.
Hormones and growth factors induce the synthesis of glial fibrillary acidic protein in rat
brain astrocytes. J. Neurosci. Res. 14:167-176.
Nieto-Sampedro, M. 1988. Astrocyte mitogen inhibitor related to epidermal growth factor
receptor. Science. 240:1784-1786.
Nicola, N.A. 1989. Hemopoietic cell growth factors and their receptors. Annu. Rev.
Biochem. 58:45-77.
Olivier, A.R., Ballou, L.M. and Thomas, G. 1988. Differential regulation of S6
phosphorylation by insulin and epidermal growth factor in Swiss mouse 3T3 cells: insulin
activation of type 1 phosphatase. Proc. Natl. Acad. Sci. USA. 85:4720-4724.
Oppenheim, R.W., Prevette, D., Tytell, M. and Homma, S. 1990. Naturally occurring
and induced neuronal death in the chick embryo in vivo requires protein and RNA
sytthesis: evidence for the role of cell death genes. Dev. Biol. 138:104-113.
Orly, J., Sato, G. and Erickson, G. 1980.
Serum suppresses the expression of
hormonally-induced functions in cultured granulosa cells. Cell. 20:817-827.
Pardee, A.B., Dubrow, R., Hamlin, J.L. and Kletzien, R.F. 1978. Animal cell cycle.
Annu. Rev. Biochem. 47:715-750.
Paul, D., Brown, K.H., Rupniak, H.T. and Ristow, H.J. 1978. Cell cycle regulation by
growth factors and nutrients in normal and transformed cells. In Vitro Cell. Dev. Biol.
14:76-84.
Pfeiffer, S.E. 1973. In: Sato, G. (ed): "Tissue Culture of the Nervous System."
Plenum Press, NY.
Phillips, P.D. and Cristofalo, V.J. 1988. Classification system based on the functional
equivalency of mitogens that regulate WI-38 cell proliferation. Exp. Cell Res. 175:396403.
124
Pipas, J.M., Chaing, L.C. and Barnes, D.W. 1984. Effect of cell type, hormonal
conditions and viral genetic background on transforming infections of S V40. Cancer
Cells. 2:355-363.
Pledger, W.J., Stiles, C.D., Antoniades, H.N. and Scher, C.D. 1978. An ordered
sequence of events is required before BALB/c-3T3 cells become committed to DNA
synthesis. Proc. Natl. Acad. Sci. USA. 75:2839-2843.
Raff, M.C. 1989. Glial cell diversification in the rat optic nerve. Science. 243:14501455.
Raff, M.C., Abney, E.R. and Miller, R.H. 1984. Two glial cell lineages diverge
prenatally in rat optic nerve. Dev. Biol. 106:53-60.
Raff, M.C., Abney, E.R., Cohen, J., Lindsay, R. and Noble, M. 1983. Two types of
astrocytes in cultures of developing rat white matter: differences in morphology, surface
gangliosides, and growth characteristics. J. Neurosci. 3:1289-1300.
Raff, M.C., Miller, R.H. and Noble, M. 1983. A glial progenitor cell that develops in
vitro into an astrocyte or an oligodendrocyte depending on culture medium. Nature.
303:390-396.
Raff, M.C., Williams, B.P. and Miller, R.H. 1984. The in vitro differentiation of a
bipotential glial progenitor cell. EMBO J. 3:1857-1864.
Rawson, C., Cosola-Smith, C. and Barnes, D. 1990. Death of serum-free mouse embryo
cells caused by epidermal growth factor deprivation is prevented by cycloheximide, 12-0tetradecanoylphorbol-13-acetate, or vanadate. Exp. Cell Res. 186;177-181.
Rawson, C., Loo, D., Helmrich, A., Ernst, T., Natsuno, T., Merrill, G. and Barnes, D.
1991b. Serum inhibition of proliferation of serum-free mouse embryo cells. Exp. Cell
Res. 192:271-277.
Rawson, C.L., Loo, D.T., Duimstra, J.R., Hedstrom, O.R., Schmidt, E.E. and Barnes,
D. 1991a. Death of serum-free mouse embryo cells caused by epidermal growth factor
deprivation. J. Cell Biol. 113:671-680.
Reznikoff, C.A., Brankow, D.W. and Heidelberger, C. 1973. Establishment and
characterization of a cloned line of C3H mouse embryo cells sensitive to postconfluence
inhibition of division. Cancer Research. 33:3231-3238.
Rizzino, A. 1984. Behavior of transforming growth factors in serum-free media--an
improved assay for gransforming growth factors. In Vitro Cell Dev. Biol. 20:815-822.
Rizzino, A. 1988. Transforming growth factor beta: multiple effects on cell differentiation
and extracellular matrices. Dev. Biol. 130:411-422.
Rizzino, A., Rizzino, H. and Sato, G. 1979. Defined media and the determination of
nutritional and hormonal requirements of mammalian cells in culture. Nutrition Reviews.
37:369-378.
Rodriguez-Tarduchy, G. and Lopez-Rivas, A. 1989. Phorbol esters inhibit apoptosis in
IL-2 dependent T lymphocytes. Biochem. Biophys. Res. Comm. 164:1069-1075.
125
Rosenfeld, R.G. and Hintz, R.L. 1986. Somatomedin receptors: structure, function and
regulation. In: Conn, P.M. (ed): "Receptors." Academic Press, Orlando, FL. 3:281329.
Rubin, H. 1977. Antagonistic effects of insulin and cortisol on coordinate control of
metabolism and growth in culture fibroblasts. J. Cell. Phys. 91:249-260.
Ruoslahti, E., Hayman, E., Peirschbacher, M. and Engvall, E. 1982. Fibronectin;
purification, immunochemical properties and biological activities. Methods Enzymol.
82:803-830.
Rutka, J.T., Gib lin, J.R. Balkissoon, R., Wen, D., Myatt, C.A., McCulloch, J.R. and
Rosenblum, M.L. 1987. Characterization of fetal human brain cultures. Dev. Neurosci.
5:154-173.
Sakai, Y., Rawson, C., Lindburg, K. and Barnes, D. 1990. Serum and transforming
growth factor beta regulate glial fibrillary acidic protein in serum-free-derived mouse
embryo cells. Proc. Natl. Acad. Sci. USA. 87:8378-8382.
Salomon, D.S., Perroteau, I., Kidwell, W.R., Tam, J. and Derynck, R. 1987. Loss of
growth responsiveness to epidermal growth factor and enhanced production of alphatransforming growth factors in ras-transformed mouse mammary epithelial cells. J. Cell
Phys. 130:397-409.
Samuals, H.H., Stanley, F. and Shapiro, L.E. 1976. Dose-dependent depletion of
nuclear receptors by L-triiodothyronine: evidence for a role in induction of growth
hormone synthesis in cultured GH1 cells. Proc. Natl. Acad. Sci. USA. 73:3877-3881.
Samuels, H.H., Forman, B.M., Horowitz, Z.D. and Ye,Z.S. 1988. Regulation of gene
expression by thyroid hormone. J. Clin. Invest. 84:957-967.
Samuels, H.H., Stanley, F. and Casanova, J.
1979. Depletion of L-3,5,3'triiodothyronine and L-thyroxine in euthyroid calf serum for use in cell culture studies of
the action of thyroid hormone. Endocrinology. 105:80-85.
Saneto, R.P., Altman, A., Knob ler, R.L. Johnson, H.M. and de Vellis, J.
1986.
Interleukin -2 mediates the inhibition of oligodendrocyte progenitor cell proliferation in
vitro. Proc. Natl. Acad. Sci. USA. 83:9221-9225.
Sanford, K.K., Earle, W.R., and Likely, G.D. 1948. The growth in vitro of single
isolated tissue cells. J. Natl. Cancer Inst. 9:229-246.
Sap, J., Munoz, A., Damm, k., Goldberg, Y., Ghysdael, J., Leutz, A., Beug, H. and
Vennstrom, B. 1986. The c-erb-A protein is a high-affinity receptor for thyroid hormone.
Nature. 324:635-640.
Sasaki, M.S. and Kodama, S. 1987. Establishment and some mutational characteristics of
3T3-like dear-diploid mouse cell line. J. Cell. Physiol. 131:114-122.
Sato, G., Pardee, A. and Sirbasku, D (eds). 1982. In: "Growth of Cells in Hormonally
Defined Media." Cold Spring Harbor Press, Cold Spring Harbor, NY.
Sato, G.H. 1975. The role of serum in cell culture. In: Litwack, G. (ed): "Biochemical
Action of Hormones." Academic Press, 391-396.
126
Scher, C.D., Pledger, W.J., Martin, P., Antoniades, H. and Stiles, C.D.
1978.
Transforming viruses directly reduce the cellular growth requirement for a platelet derived
growth factor. J. Cell Phys. 97:371-380.
Scherer, W.F., Syverton, J.T. and Gey, G.O. 1953. Studies on the propagation in vitro
of poliomyelitis viruses. J. Exp. Med. 97:695-709.
Schreiber, A., Kenney, J,. Lowalski, W., Firesel, B., Mehlman, T. and Maciag, T.
1985. The interaction of endothelial cell growth factor with heparin: characterization by
receptor and antibody recognition. Proc. Natl. Acad. Sci. USA. 82:6138-6142.
Schubert, D., Humphreys, S., Baroni, C. and Cohn, M. 1969. In vitro differentiation of
a mouse meuroblastoma. Proc. Natl. Acad. Sci. USA. 64:316-323.
Schwartz, L.M., Kosz, L. and Kay, B.K. 1990. Gene activation is required for
developmentally programmed cell death. Proc. Natl. Acad. Sci. USA. 87:6594-6598.
Searle, J., Kerr, J.F.R. and Bishop, C.J. 1982. Necrosis and apoptosis: distinct modes
of cell death with fundamentally different significance. Pathol. Annu. 17:229-259.
Serrero, G., McClure, D. and Sato, G. 1979. Growth of mouse 3T3 fibroblasts in
serum-free, hormone-supplemented media. Cold Spring Harbor Conf. Cell Prolif. 6:523530.
Shi, Y., Sahai, B.M. and Green, D.R. 1989. Cyclosporin A inhibits activation-induced
cell death in T-cell hybridomas and thymocytes. Nature. 339:625-626.
Shih, C. and Weinberg, R.A. 1982. Isolation of a transforming sequence from a human
bladder carcinoma cell line. Cell. 29:161-169.
Shipley, G.D. and Ham, R.G. 1983. Multiplication of swiss 3T3 cells in a serum-free
medium. Exp. Cell Res. 146:249-260.
Shirahata, S., Rawson, C., Loo, D., Chang, Y. and Barnes, D. 1990. Ras and neu
oncogenes reverse serum inhibition and epidermal growth factor dependence of serum-free
mouse embryo (SFME) cells. J. Cell. Physiol. 144:69-76.
Smith, C.A., Williams, G.T., Kingston,R., Jenkinson, E.J. and Owen, J.T. 1989.
Antibodies to CD3a-cell receptor complex induce death by apoptosis in immature T cells in
thymic cultures. Nature. 337: 181-184.
Spemann, H. 1938. In "Embryonic Development and Induction."
New Haven, CT,
Yale Univ. Press.
Sporn, M.B. and Roberts, A.B. 1985. Autocrine growth factors and cancer. Nature.
313:745-747.
Sporn, M.B. and Roberts, A.B. 1988. Peptide growth factors are multifunctional.
Nature. 332:217-219.
Stemple, D.L., Mahanthappa, N.K. and Anderson, D.J. 1988. Basic FGF induces
neuronal differentiation, cell division, and NGF dependence in chromaffin cells: a
sequence of events in sympathetic development. Neuron. 1:517-525.
127
Temin, H.M., Pierson, R.W. and Dulak, N.C. 1972. The role of serum in the control of
multiplication of avian and mammalian cells in culture. In: Rothblat, G. and Cristofalo,
V.J. (eds): "Growth, Nutrition and Metabolism of Cells in Culture." Academic Press,
NY, 1:50-81.
Thoenen, H. and Barde, Y. -A. 1980. Physiology of nerve growth factor. Physiol. Rev.
60:1284-1335.
Titani, K., Koide, A., Ericsson, L., Kumar, S., Hermann, J., Wade, R.D., Walsh, K.A.,
Neurath, H. and Fischer, E.H. 1978. Sequence of the carboxyl-terminal 492 residues of
rabbit muscle glycogen phosphorylase including the pyrodoxal 5'-phosphate binding site.
Biochemistry. 17:5680-5695.
Todaro, G.J. and Green, H. 1963. Quantitative studies of the growth of mouse embryo
cells in culture and their development into established cell lines. J. Cell Biol. 17:299-313.
Tomozawa, Y. and Sueoka, N. 1978. In vitro segregation of different cell lines with
neuronal and glial properties from a stem cell line of rat neurotumor RT4. Proc. Natl.
Acad. Sci. USA. 75:6305-6309.
Tsai, S. and Gaffney, E.V. 1986. Inhibition of cell proliferation by interleukin 1 derived
from monocytic leukemia cells. Cancer Res. 46:1271-1278.
Tucker, R., Shipley, G., Moses, H. and Holley, R. 1984. Growth inhibitor for BSC-1
cells closely related to platelet type beta transforming growth factor. Science. 226:705707.
Twardzik, D.R., Ranchalis, J.E. and Todaro, G.J. 1982. Mouse embryonic transforming
growth factors related to those isolated from tumor cells. Cancer Res. 42:590-593.
Weinberger, C., Thompson, C.C., Ong, E.S., Lebo, R. Gruol, D.J. and Evans, R.M.
1986. The c-erb-A gene encodes a thyroid hormone receptor. Nature. 374:641-646.
Westphal, U., Burton, R.M. and Harding, G.B. 1975. Characterization of steroidbinding glycoproteins: methodological comments. Methods Enzymol. 36:91-104.
White, P.R. 1946. Cultivation of animal tissues in vitro in nutrients of precisely known
constitution. Growth. 10:231-289.
Worton, R.G. and Duff, C. 1979. Karyotyping. Methods Enzymol. 58:322-344.
Wyllie, A.H. 1981. Cell death: a new classification separating apoptosis from necrosis.
In: Bowen, I.D. and Lockshin, R.A. (eds): "Cell Death in Biology and Pathology."
Chapman and Hall Press, London, NY, 9-34.
Wyllie, A.H., Kerr, J.F.R. and Currie, A.R. 1980. Cell death: the significance of
apoptosis. In: "International Review of Cytology." Academic Press, NY. 68;251-306.
Yanishevsky, R.M. and Stein, G.H. 1981. Regulation of the cell cycle in eukaryotic
cells. Int. Rev. Cytol. 69:223-259.
Yanker, B.A. and Shooter, E.M. 1982. The biology and mechanism of action of nerve
growth factor. Annu. Rev. Biochem. 51:845-868.
128
Yip, H.K. and Johnson, E.M. 1984. Developing dorsal root ganglion neurons require
trophic support from their central processes: evidence for a role of retrogradely transported
nerve growth factor from the central nervous system to the periphery. Proc. Natl. Acad.
Sci. USA. 81:6245-6249.
APPENDICES
129
APPENDIX
1
Serum Inhibition of Proliferation of Serum-Free Mouse Embryo Cells
Cathleen Rawson, Deryk Loo, Angela Helmrich, Ted Ernst,
Tetsuya Natsuno, Gary Merrill, and David Barnes
Published in: Experimental Cell Research. Volume 192, pp. 271-277 (1991).
Coauthor contribution: C.R., primary investigator, D.L., co-primary
investigator, A.H., contributing investigator, T.E., technical assistant; T.N.,
visiting scientist; G.M., collaborating scientist; D.B., research director
130
SUMMARY
Serum-free mouse embryo (SFME) cells, derived in medium supplemented with
insulin, transferrin, high density lipoprotein, epidermal growth factor, and fibronectin, do
not undergo crisis, maintain a predominately diploid karyotype with no detectable
chromosomal abnormalities for well over 100 population doublings in vitro, and are
growth inhibited by concentrations of serum that are growth-stimulatory for most cell lines
in culture. Serum inhibition of SFME cell proliferation was reversible and was not
prevented by addition of the supplements of the serum-free medium, even when added
repeatedly during the culture period. The serum effect on SFME cell proliferation could be
detected after incubation in serum-containing medium for as little as 8 hours. SFME cells
in serum-containing medium were arrested in the G1 phase of the cell cycle with a greatly
reduced rate of incorporation of precursors into DNA and thymidine kinase activity, while a
reduction in rate of incorporation of amino acids into protein was not observed. SFME
cultures maintained for extended periods in serum-containing medium underwent a crisis-
like period followed by the appearance of variant cells capable of growing in serumsupplemented medium. These cells exhibited abnormal karyotype and were resistant to
several inhibitors of proliferation active on the parent SFME cell type.
131
INTRODUCTION
Mouse embryo cells initiated in the absence of serum in a rich basal nutrient
medium supplemented with insulin, transferrin, epidermal growth factor (EGF), high
density lipoprotein, and fibronectin display several unusual properties (Loo et al.,
1987,1989a, 1989b, 1990; Rawson et al., 1990; Shirahata et al., 1990) that distinguish
them from mouse embryo cells established in conventional protocols utilizing serum as a
source of growth factors (e.g., 3T3 cells) (Todaro and Green, 1963; Aaronson and
Todaro, 1968; Jainchill et al., 1969; Reznikoff et al., 1973). Serum-free mouse embryo
(SFME) cells do not exhibit growth crisis or chromosomal aberration when carried in vitro
for more than 10 times the number of population doublings that can be obtained before
crisis with mouse embryo cells in serum-containing medium (Loo et al., 1987, 1989b;
Rawson et al., 1990). In addition, SFME cells are dependent on EGF for survival and are
growth inhibited by serum. Transformation of SFME cells with the ras or neu oncogenes
releases SFME cells from the growth inhibitory effect of serum as well as the EGF
requirement for survival (Loo et al., 1987; Shirahata et al., 1990). Part of the growth
inhibitory activity of serum may be due to glucocorticoids and thyroid hormones, both of
which can inhibit growth of SFME cells. However, additional factors that contribute to the
inhibitory activity on SFME cells are also present in serum (Loo et al., 1990).
In order to better understand mechanisms by which serum inhibits proliferation of
SFME cells and oncogenic activity abrogates this response, we examined the characteristics
of SFME cells under the influence of serum inhibition and also isolated and characterized
rare SFME variants that did not undergo growth inhibition in response to serum. Exposure
of SFME cells to medium with 10% calf serum caused a reversible inhibition of cell growth
with accumulation of cells in the G1 phase of the cell cycle and depletion of S phase cells.
Serum caused an inhibition of incorporation of thymidine into SFME cells which was
132
associated with a decrease in thymidine kinase activity but was not accompanied by a
decrease in amino acid incorporation.
SFME cells cultured in serum-containing medium underwent a crisis period of
approximately 1 month similar to that observed with mouse embryo cells derived in serum-
containing media (Todaro and Green, 1963; Aaronson and Todaro, 1968; Jainchill et al.,
1969; Reznikoff et al., 1973), followed by the appearance of variant cells capable of
growing in serum-containing medium. The variant cells exhibited an abnormal, hyperploid
karyotype and were resistant to growth inhibition by hydrocortisone, thyroid hormone or
gamma-interferon, suggesting an alteration had occurred in a basic mechanism by which
SFME cells are growth inhibited.
133
MATERIALS AND METHODS
Cell culture. Detailed procedures for the initiation and continuous culture of Balb/c
SFME cells have been published (Loo et al., 1989a, 1989b). The basal nutrient medium
was a one-to-one mixture of Dulbecco's modified Eagle's medium containing 4.5 g/liter
glucose and Ham's F12 (Mather and Sato, 1979; Ham and McKeehan, 1979)
supplemented with 15 mM Hepes, pH 7.4, 1.2 g/liter sodium bicarbonate, sodium selenite
(10 nM), penicillin (200 U/ml), streptomycin (200 µg/ml), and ampicillin (25 .tg/ml)
(F12:DME). SFME cells were cultured in F12:DME supplemented with insulin (10
µg/ml), transferrin (10 p.g/ml), human high density lipoprotein (HDL) (10 µg/m1), and
EGF (50 ng/ml) in dishes or flasks precoated with bovine fibronectin (1014/m1). Bovine
insulin and human transferrin were obtained from Sigma Chemical Co. (St. Louis, MO).
Mouse EGF was obtained from Upstate Biotechnologies (Lake Placid, NY). Fibronectin
and HDL were prepared as described (Loo et al., 1989a; Gospodarowicz, 1984). Calf
serum (CS) was obtained from Grand Island Biological Co. (Grand Island, NY). Gamma-
interferon was obtained from Amgen (Thousand Oaks, CA). Stock SFME cultures used
for the experiments described in this report had undergone between 60 and 120 population
doublings.
Flow cytometry. Cells cultured as indicated were harvested by trypsinizing in
0.25% crude trypsin with 1 mM EDTA in phosphate-buffered saline without calcium or
magnesium (PBS) and diluted into an equal volume of F12:DME containing 1 mg/m1
soybean trypsin inhibitor. Cells were washed in F12:DME and resuspended in cold PBS.
Cold 70% ethanol was added dropwise with gentle mixing. Fixed cells were centrifuged
and resuspended in a solution containing chromomycin A3 (20 ii.g/m1) and 15 mM MgC12
to give a concentration of 1 to 5 x 106 cells/ml. Suspensions were incubated 30 minutes at
room temperature in the dark. Cells were centrifuged from suspension and resuspended in
the same volume of PBS.
134
Basic procedures for flow cytometry were described by Gray and Coffin (1979).
Flow cytometry was carried out with an Epics flow cytometer-cell sorter (Coulter Epics
Division, Hialeah, FL). Chromomycin A3-stained cells were excited with an output of 200
mW of 457 nm light emitted from an argon laser. The resulting cell-associated
fluorescence was captured through a 495-nm longpass absorbance filter (Coulter) and a
blocking interference filter (Corion).
Assays. Cells were plated in F12:DME with serum-free supplements or with 10%
dialyzed CS. After 24 hours medium was changed to fresh medium of the same
composition and isotopes (4 liCi/m1 [3H]-thymidine, New England Nuclear, NET042H or
14C-amino acid mixture, New England Nuclear, NEC445) were added three hours after
the medium change. Four hours later medium was removed, cells were washed three times
with cold PBS, and macromolecules were precipitated with 10% trichioroacetic acid and
counted in a liquid scintillation counter. Cell number per plate was determined by counting
suspensions of trypsinized cells from duplicate plates in a Coulter particle counter.
Thymidine kinase assay was carried out as described (Merrill, et al., 1984).
Karyotyping. Procedures for the analysis of mitotic chromosomes from SFME cell
cultures have been previously published (Loo et al., 1989a; Worton and Duff, 1979). Fifty
to two-hundred metaphases from each culture were stained with a 3% Giemsa stain and
analyzed at 1000X with an oil immersion objective. Cells containing fewer than 35
chromosomes were considered artifacts of the procedure and excluded from the analyses.
135
RESULTS
Effects of serum on SFME cells. SFME cells plated in basal nutrient medium
supplemented with 10% CS do not proliferate and addition of the supplements of the
serum-free medium with 10% CS allows approximately one population doubling over a
period of about 48 hours before growth arrest (Loo et al., 1987) (Fig. A1.1). No further
growth is observed if the supplements are added repeatedly over the course of the
experiment, and initial exposure of SFME cells to medium containing 10% CS for 2 days
followed by addition of the supplements to the serum-containing medium resulted in no
growth response to the supplements (Fig. A1.1).
In experiments in which SFME cells were incubated in medium containing 10% CS
for various times and then switched to serum-free medium with the growth-stimulatory
supplements, exposure of cells to 10% CS for as little as 8 hours produced a measurable
reduction in cell number, and the effect of serum did not increase markedly upon exposure
of SFME cells to serum-containing medium for longer periods (Fig. A1.2). Because EGF
is critical for the growth of SFME cells (Loo et al., 1987, 1989b; Rawson et al., 1990), we
examined levels of the EGF receptor on SFME cells in serum-free and serum-containing
media. No differences were observed, and SFME cells in serum-containing medium also
exhibited the ability to reexpress EGF receptors on the cell surface after down regulation by
a preincubation with the ligand (not shown).
Flow cytometric analysis indicated that SFME cells cultured in serum-containing
medium for 2 days accumulated in the G1 phase of the cell cycle, with S phase cells
disappearing from the cultures (Fig. A1.3). The cytometric pattern was similar to that
observed in serum-derived Balb/c 3T3 cells growth arrested by incubation on 0.5% CS,
except that the amount of DNA per SFME cell was less than that observed with 3T3 cells
because SFME cells remain predominately diploid while 3T3 cells are approximately
tetraploid (Todaro and Green, 1963; Aaronson and Todaro, 1969; Jainchill et al., 1969).
136
The serum-induced arrest of SFME cell proliferation was reflected in a reduced
incorporation of thymidine into DNA and reduced levels of thymidine kinase, an S phase-
dependent enzyme (Table A1.1). A decrease in thymidine kinase activity was associated
with a reduction in mRNA for this enzyme detectable by Northern hybridization analysis
(not shown). Incorporation of amino acid into protein was not reduced under conditions in
which DNA synthesis was inhibited by CS.
Variant SFME cells capable of growing in serum-containing medium. SFME cells
cultured for 96 population doublings in serum-free medium were carried for an extended
period in a 3T3 protocol (plating at 3 x 105 cells/25-cm2 flask, passaging every 3 days
(Todaro and Green, 1963)) in medium containing 10% CS. Cell proliferation was
observed in serum-containing medium after a period of about 1 month (Fig. A1.4).
Karyotyping of cells capable of growth in serum-containing medium showed a strong
tendency toward hyperploidy; 44 of 50 metaphases examined contained greater than the
diploid number of 40 chromosomes (Fig. A1.5). Chromosomal abnormalities were also
evident in these metaphases.
The SFME culture selected in serum-containing medium in the 3T3 protocol may
have been derived from several variant cells present in the initial population or generated
during the selection in serum-containing medium. Each variant might confer a
characteristic diploid, hyperploid, or polyploid contribution to the total karyotype of the
population. Because of this we selected directly from SFME cultures clonally derived
variants capable of growing in serum-containing medium to determine if all cells capable of
growing in serum-containing medium exhibited abnormal karyotype and to determine if a
common structural abnormality could be detected in all clones.
SFME cells were plated in medium containing 10% CS at low density (105
cells/80-cm2 dish) and cultured with weekly medium changes for 5 weeks. Colonies
capable of growing in serum-containing medium developed with a frequency of about 1 in
106 SFME cells. Five of these were propagated and karyotyped (Fig. A1.6). All clones
137
exhibited a degree of hyperploidy fivefold or greater than that of the parent cells cultured in
serum-free medium (Table A1.2).
Three clones showed two subpopulations: one with chromosome numbers
hyperploid but near-diploid and one with chromosome numbers approximately tetraploid.
One clone was largely near-tetraploid and one clone was moderately hyperploid with less
polyploidy (Fig. A1.6, Table A1.2). Structural alterations in chromosomes were also
evident as metacentric and submetacentric chromosomes (all normal mouse chromosomes
are acrocentric) and as double minute chromosomes (Table A1.2), but no structural
alteration was common to all five clones. Although hyperploidy was a common feature of
all clones, it was not sufficient for release of SFME cells from serum inhibition of
proliferation, because SFME cells made hyperploid by polyethylene glycol-induced fusion
did not grow in serum-containing medium.
The proliferative behavior of one of the clones was examined further. This clone
grew slowly in medium supplemented only with 10% CS, and grew well in serum-free
medium or in medium containing both the supplements of the serum-free medium and 10%
CS, indicating that the inhibitory proliferative response was not elicited by CS on these
cells (Fig. A1.7). We have reported previously that glucocorticoid and thyroid hormones
are inhibitory for growth of SFME cells in the presence of low concentrations of charcoal-
treated CS (Loo et al., 1990). Gamma-interferon was also markedly inhibitory for SFME
cells (Fig. A1.8), although none of these factors are solely responsible for the effect of
serum on SFME cells (Loo et al., 1990). The variant clone that was not inhibited by serum
also showed reduced inhibitory response to all three of these factors (Fig. A1.8).
138
DISCUSSION
SFME cells are inhibited by CS, but continue to grow for about one population
doubling if plated in 10% CS with insulin, transferrin, EGF, HDL, and fibronectin.
Further addition of these factors upon extended culture did not lead to any additional
proliferation in the presence of CS and cells plated for 48 hours in 10% CS before the
addition of the factors did not undergo cell division. The growth inhibitory effect is
reversible upon removal of serum from the culture medium, suggesting that the
phenomenon is not simply due to toxicity of serum for SFME cells. Furthermore, serum
inhibition of proliferation was associated with buildup of the cells in the 01 phase of the
cell cycle and reduction of the fraction of cells in S phase, a decrease in thymidine kinase
activity, and a reduction in DNA synthesis.
All of these phenomena are associated with physiologically relevant inhibition of
cell division such as that observed when cells become growth arrested upon growth factor
depletion or upon induction of differentiation (Pardee et al., 1978; Yanishevsky and Stein,
1981; Baserga et al., 1982). Growth factor depletion is not applicable in this circumstance,
but coordinate inhibition of proliferation and induction of differentiation of SFME cells by
serum is consistent with the observed lack of inhibition of incorporation of amino acid by
CS treatment of SFME cells.
Almost without exception mouse embryo cell lines derived in serum-containing
media are hyperploid with one or more chromosomal abnormality; even lines in which
karyotypic abnormalities are minimal are still aneuploid or karyotypically unstable (Todaro
and Green, 1963; Aaronson and Todaro, 1968; Jainchill et al., 1969; Reznikoff et al.,
1973; Farber and Liskay, 1974; Sasaki and Kodama, 1987). Exposure of SFME cells that
had been grown long-term in serum-free medium to a 3T3 passage regimen in serum-
containing medium led to a peroid of no proliferation, followed by the appearance of
variant cells capable of growing in serum-containing medium and expressing abnormal,
139
hyperploid karyotype with structural chromosomal abnormalities. This behavior is similar
to the phenomena of crisis and immortalization observed in mouse embryo cells derived
initially in serum-containing medium and suggests that serum-containing medium selected
for, or induced, an abnormal karyotype in SFME cells.
Variant SFME cells selected for the ability to grow in serum no longer were
inhibited by serum factors and exhibited reduced responses to several hormones that were
growth inhibitory for SFME cells. An alteration apparently occurred in these cells that led
to reduced inhibitory growth responses in general, unrelated to a specific response to a
particular factor that might be present in serum. Although no consistent chromosomal
abnormality was detected in the clones capable of growing in serum, the tendency toward
hyperploidy in these cells suggests that gene dosage effects may provide selective growth
advantage for variant clones under the selective conditions employed, as is probably the
case with postcrisis, hyperploid mouse embryo cell lines arising from cultures initiated in
serum-containing medium. This explanation is not completely simple, however, because
SFME cells made polyploid by cell fusion remained inhibited by serum. Further work is
necessary to determine the relationship between abnormal karyotype, immortalization or
escape from crisis, and escape from serum inhibition observed in variant SFME cells.
140
ACKNOWLEDGMENTS
This work was supported by NIH-NAI-07560. D. Barnes is the recipient of
RCDA NIH-NCI01226. G. Merrill is the recipient of RCDA NIH-AG00334. Flow
cytometry was carried out through the Oregon State University Environmental Health
Sciences Center Flow Cytometry Core Unit (NIEHS ES00210).
Cathleen Rawson assisted with cell culture experiments and performed flow
cytometry. Angela Heinrich provided cell culture support. Ted Ernst prepared karyotypes
and statistical data. Tetsuya Natsuno performed the amino acid incorporation experiment.
Gary Merrill performed the thymidine incorporation and thymidine kinase activity
experiments. David Barnes provided financial and scientific support.
141
Table A1.1. Synthetic and Enzymatic Activities of SFME Cells in Serum-
Free and Serum-Containing Media.
SFME cells were plated in medium
supplemented with insulin, transferrin, EGF, HDL, and fibronectin (serum-free) or 10%
dialyzed CS.
24 hours later, plates were assayed for DNA synthesis (thymidine
incorporation, 4 hour pulse), protein synthesis (amino acid incorporation, 4 hour pulse),
and thymidine kinase activity as described under Materials and Methods.
Serum-free
Calf serum
56.6 ± 4.5
4.05 ± 0.40
(cpm X 10-4/106 cells)
13.3 ± 0.40
17.7 ± 2.0
Thymidine kinase activity
(pmol/min/106 cells)
91.0 ± 5.55
1.30 ± 0.18
Thymidine incorporation
(cpm X 10-4/106 cells)
Amino acid incorporation
142
Table A1.2. Karyotypic Abnormalities of Variant SFME Clones Capable of
Growing in Serum-Containing Media. Parent cultures were initiated in serum-free
medium and had undergone 96 population doublings in this medium at the time of the
experiment. The normal mouse contains 40 acrocentric chromosomes; metacentric and
submetacentric chromosomes are indications of structural abnormalities. Hyperploid cells
were those with more than 40 chromosomes. Polyploid cells were considered to be cells
containing more than 70 chromosomes.
Parent
% Hyperdiploid
% Polyploid
% with metacentric chromosome
% with submetacentric chromosome
% with double minute chromosomes
16
6
1
0
1
Clone A Clone B Clone C Clone D Clone F
78
64
56
0
6
82
30
49
96
52
4
98
82
80
23
0
100
100
18
81
25
0
0
60
5
143
Fig. A1.1. Culture of SFME cells in serum-containing medium. Cells were
seeded at 105/35 mm-diameter plate in the media listed below and cell number was
determined on the indicated days after plating. (0), medium containing 10% CS; (I),
medium containing 10% CS with addition of 10 µg/ml insulin, 10 µg/ml transferrin, 10
p.g/m1HDL, and 50 ng/ml EGF on Day 3 after plating; (), medium containing 10% CS,
insulin, transferrin, HDL, and EGF; (0), medium containing 10% CS, insulin, transferrin,
HDL, and EGF with additional insulin, transferrin, HDL, and EGF added on Days 2, 4,
and 6 after plating. All plates were precoated with fibronectin.
m
t
0
..-1
X
2
.1mm,
i
0
1
2
3
4
5
Days After Plating
6
7
144
Recovery of SFME cells from plating in serum-containing
Fig. A1.2.
medium. Cells were plated at 5 x 104/35 mm-diameter dish in medium containing 10%
CS and precoated with fibronectin. At the indicated times, the medium was removed,
plates were washed once with serum-free medium, and medium supplemented with insulin,
transferrin, FIDL, and EGF was added. Cell number was determined 4 days later.
5
O
in
1
0
-4
x
4
3
2
0
0---°"-----...........0
O
1
0
0
24 hrs
48 hrs
72 hrs
Preincubation With 10% Calf Serum
145
Fig. A1.3.
Flow cytometric analysis of SFME cells in serum-containing
medium. Tracings show relative cell number (Y axis) versus relative amount of DNA per
cell (X axis). (A), 3T3 cells growing in medium containing 10% CS; (B), 3T3 cells
growth arrested by incubation in medium containing 0.5% CS for 2 days; (C), SFME cells
growing in serum-free medium supplemented with insulin, transferrin, EGF, HDL, and
fibronectin; (D), SFME cells growth arrested in medium containing 10% CS for 2 days.
A
B
.k-
C
Jar
D
146
Fig. A1.4. Selection of variant SFME cells capable of growing in serumcontaining medium. SFME cells initiated and cultured for 96 population doublings in
serum-free medium as described were cultured in medium containing 10% CS, plating at 3
x 105/25-cm2 flask and passaging every 3 days (3T3 protocol). Cell number was
determined at each passage. Total cell number decreased during the first 5 pasages due to
loss of cells upon transfer. An increase in cell number was first detected 27 days after
beginning the experiment.
7
102
6
5
4
101
3
2
1
100
0
10
20
30
Days
40
50
147
Fig. A1.5. Karyotype of variant SFME cells selected by culture in serumcontaining medium. SFME cells initiated and cultured for 96 population doublings in
serum-free medium as described were cultured in medium containing 10% CS, plating at 3
x 105 cells/25-cm2 flask and passaging every 3 days. Karyotyping was carried out as
described in the text. (A), parent culture at 96 population doublings; (B), parent culture
after an additional 6 population doublings in serum-free medium; (C), variants after
approximately seven population doublings in serum-containing medium.
A
80
80
70
70
60
6
50
80
70
In
50
N
a,
a,
co
g
.c 40
50
co
Col
40
40
-a;
co
c8
;0
2
O
30
o 30
30
a,
a,
cn
a
pl
; 20
ccci
C
,_,J
20
20
a,
a-
10
0
10
10
_.A11L
35
38
41
I
44 >80
Number of Chromosomes
35
38
41
44 >80
Number of Chromosomes
35 38 41 44 66 69 72 75 78 81 84 87 90,160
Number of Chromosomes
148
Fig. A1.6. Karyotype of variant SFME clones capable of growing in serumcontaining medium. Clones were derived and karyotyped as described in the text.
C
B
A
50
30
50
e t0
a
a
.0
in
04.
e
a
.=
a
230
a
a
I; 30
m 30
2
2
o
2
"O
2.
0
e
0 so
o
a
a
Z
.zi 10
m
a.
0.
iii Iii I
Hill
I
I 35 36 41 44 611 69 72 75 75 SI
0
0
33 36 41 44 86 55 72 75 75 et 64 87 90.136
Number of Chromosomes
36 36
54
41 44 68
,
E
50
40
a
a
z
30
00
S
2
0
0
1i
20
o
a
20
0
u 10
10
oll
35
ss
III
4.
50
0
0.
7 SO
5.1 a .5 r. II 7. $35
II
Cl
so
Number of Chromosomes
59
72
75
75 65 64 87
Number o Chromosomes
Number of Chromosomes
0
I1
35 36
44
I
.
76
11111 1
71 52 65
Number of Chromosomes
149
Fig. A1.7. Growth of variant SFME cells in serum-containing and serumfree medium. Variant SFME cells (clone shown in Fig. A1.6E) were plated at 105/35
mm-diameter plate and cell number was determined on the indicated days after plating.
(), medium containing 10% CS; (0), medium containing 10% CS with addition of 10
p.g/ml HDL and 50 ng/ml EGF; (0), medium containing insulin, transferrin, HDL, and
EGF. All plates were precoated with fibronectin.
3 X 106
U)
.1.)
.-4
0.
10
6
C_
a)
0
al
-)
3 X 105
v-1
a)
U
105
0
2
4
6
Days After Plating
a
150
Fig. A1.8. Reduced inhibition of variant SFME cells by thyroid hormone,
glucocorticoid, and gamma-interferon. Parent or variant SFME cells capable of
growing in serum-containing medium (clone shown in Figs. A1.6E and A1.7). were plated
at 105/35 mm-diameter dish in medium supplemented with insulin, transferrin, EGF,
frDL, fibronectin, and 2.5% charcoal-treated CS (Shirahata, et al., 1990), with or without
hydrocortisone (100 nM), T3 (10 nM), gamma-interferon (3 ng/ml) or CS (10%). Cell
number was determined 6 days after plating. Data for parent SFME cultures are on the top;
data for variant clone are on the bottom.
rn
Cu
0
.-4
0
L
41
C
0
U
Cr)
F-
+
U
1
+
UI-1
+
U)
L)
+
r-i
0
C-
4.)
C
0
0
I uI
u_
cr)
+
+
+
rn
+
u
151
APPENDIX 2
Death of Serum-Free Mouse Embryo Cells Caused
by Epidermal Growth Factor Deprivation
Cathleen Rawson, Deryk Loo, Julie Duimstra,
Olaf Hedstrom, Edward Schmidt, and David Barnes
Published in; The Journal of Cell Biology. Volume 113, Number 3, pp. 671-680 (1991).
Coauthor contributions: C.R., primary investigator; D.L., co-primary investigator; J.D.,
technical assistant; O.H. and E.S., collaborating scientists; D.B., research director
152
SUMMARY
Serum-free mouse embryo (SFME) cells, derived in medium in which serum is
replaced with growth factors and other supplements, are proastroblasts that are acutely
dependent on epidermal growth factor (EGF) for survival. Ultrastructurally, an early
change found in SFME cells deprived of EGF was a loss of polysomes which
sedimentation analysis confirmed to be a shift from polysomes to monosomes. The
ribosomal shift was not accompanied by decreased steady-state level of cytoplasmic actin
mRNA examined as an indicator of cellular mRNA level. With time the cells became small
and severely degenerate and exhibited nuclear morphology characteristic of apoptosis.
Genomic DNA isolated from cultures undergoing EGF deprivation-dependent cell death
exhibited a pattern of fragmentation resulting from endonuclease activation characteristic of
cells undergoing apoptosis or programmed cell death. Flow cytometric analysis indicated
that cultures in the absence of EGF contained almost exclusively G1 -phase cells. Some of
the phenomema associated with EGF deprivation of SFME cells are similar to those
observed upon NGF deprivation of nerve cells in culture, suggesting that these
neuroectodermal-derived cell types share common mechanisms of proliferation control
involving peptide growth factor-dependent survival.
153
INTRODUCTION
We have described the long-term culture of mouse embryo cells derived and
passaged under conditions in which the serum supplement to the basal nutrient medium is
replaced by growth factors and other supplements: insulin, transferrin, EGF, high-density
lipoprotein, and fibronectin (Loo et al., 1987, 1989a, b). These serum-free mouse embryo
(SFME) cells exhibit a number of unusual properties. Unlike mouse embryo cells derived
in conventional, serum-supplemented media which undergo growth crisis and
immortalization, SFME cells do not lose proliferative potential or develop gross
chromosomal aberration when cultured for more than 10 times the number of population
doublings that can be achieved with mouse embryo cells in conventional, serum-containing
medium (Todaro and Green, 1963; Loo et al., 1987, 1989a; Ernst et al., 1990).
Proliferation of SFME cells is reversibly inhibited by serum (Loo et al., 1987,
1990; Rawson et al., 1991), and treatment of SFME cells with serum or transforming
growth factor beta leads to the appearance of glial fibrillary acidic protein (GFAP), an
intermediate filament protein that is a specific marker for astrocytes (Sakai et al., 1990).
Cells with properties like those of SFME cells also can be isolated directly from mouse
brain (Sakai et al., 1990), and a modification of the SFME medium formulation allows the
multipassage serum-free culture of human brain cells that express GFAP (Loo et al.,
1991). These results suggest that SFME cells are proastroblasts.
SFME cells are acutely dependent on EGF for survival, and cycloheximide or
actinomycin D prevents death caused by EGF deprivation (Raawson et al., 1990),
suggesting that EGF-dependent survival may depend on suppression of synthesis of
proteins that cause cell death, a phenomenon similar to that reported for neuronal
"programmed cell death" in the absence of nerve growth factor (NGF) (Martin et al.,
1988).
Thus, multiple cell types of neuroectodermal origin may share common
mechanisms of proliferative control involving peptide growth factor-dependent survival.
154
SFME cell death in the absence of EGF also is delayed by orthovanadate, an inhibitor of
phosphotyrosine phosphatases, and 12- O- tetradecanoylphorbol 13-acetate, an activator of
protein kinase C (Rawson et al., 1990).
Greater than 90% of SFME cells are dead within 48 hours after removal of EGF
from the culture medium, and decreased cell survival is observed after EGF deprivation for
as little as 8 hours (Loo et al., 1987, 1989b). Individual omission of any of the other
medium supplements leads to reduced SFME cell growth, but does not induce cell death
(Loo et al., 1989b). NGF, platelet-derived growth factor, and transforming growth factor
beta will not substitute for EGF in this system, while fibroblast growth factor will allow
survival and slow growth of SFME cells in the absence of EGF (Loo et al., 1989b). We
have suggested that at least two alternative mechanisms for the control of cell proliferation
may exist: one related to intrinsic limitation of proliferative potential observed in
conventional cell cultures, and the other governed by growth factor-dependent survival
observed in SFME cell cultures (Sakai et al., 1990).
We examined ultrastructural, biochemical, and physiological responses of SFME
cells to EGF deprivation as a step toward defining the primary mechanisms leading to death
of these cells. Early changes included a shift of ribosomes from a polysomal organization
to monosomes. The ribosomal shift was not accompanied by decreased steady-state level
of actin messenger RNA in the cytoplasm. With time the cells became small and
degenerate, and pyknotic nuclei appeared. Flow cytometric analysis indicated that cultures
in the absence of EGF contrained almost exclusively a diploid amount of DNA (G1- phase).
Genomic DNA isolated from cultures undergoing EGF deprivation-dependent cell death
exhibited a pattern of fragmentation associated with programmed cell death in other systems
(Arends et al., 1990).
155
MATERIALS AND METHODS
Cell culture. Detailed procedures for the initiation and continuous culture of Balb/c
SFME cells have been published (Loo et al., 1989a,b). The basal nutrient medium was a
one-to-one mixture of DME containing 4.5 g/1 glucose and Ham's F12 (Mather and Sato,
1979; Ham and McKeehan, 1979) supplemented with 15 mM Hepes, pH 7.4, 1.2 g/1
sodium bicarbonate,sodium selenite (10 nM), penicillin (200 U/ml), streptomycin (200
tg/m1) and ampicillin (25 p.g/ml) (F12:DME). Stock cultures of SFME cells were cultured
in F12:DME supplemented with insulin (10 µg/ml), transferrin (1014/m1), human high
density lipoprotein (HDL) (10 µg/ml) and EGF (50 ng/ml) in dishes or flasks precoated
with bovine fibronectin (10 µg/ml). Methods for preparation, storage, and use of medium
supplements and fibronectin precoating have been published (Loo et al., 1989b).
Plating was accomplished by adding cells in prewarmed medium to plates or flasks
previously precoated with fibronectin and preincubated in the incubator (5% CO2-95% air
atmosphere at 37°C) for 15 minutes with the remaining portion of medium. Insulin,
transferrin, EGF, and HDL were added directly to the individual culture vessels as small
aliquots from concentrated stocks immediately after plating the cells. Bovine insulin,
human transferrin, trypsin, and soybean trypsin inhibitor were obtained from Sigma
Chemical Co. (St. Louis, MO). Mouse EGF was obtained from Upstate Biotechnologies
(Lake Placid, NY).
HDL (density = 1.068-1.121 g/cc) was prepared by KBr
ultracentrifugation as described (Gospodarowicz, 1984; Loo et al., 1989b) and filter
sterilized after dialysis. Preparation of fibronectin has been described (Loo et al., 1989b).
Electron microscopy. SFME cells were plated in 10-cm-diameter plates (2 x 107
cells/plate) in serum-free medium with 5 ng/ml EGF. The following day some plates were
changed to serum-free medium without EGF (10m1/plate). Control plates were changed to
serum-free medium with 50 ng/ml EGF. At the indicated times 8 ml of the culture medium
were carefully removed from a plate, 3 ml of fixative (1.75% glutaraldehyde in 0.1 M
156
sodium cacodylate buffer, pH 7.40) were added, and plates were then incubated at room
temperature for 30 minutes. After fixation, the cells were removed from the dishes with a
rubber policeman and were resuspended in fixative, pelleted, and the pellet was minced into
pieces 1 mm3. The samples were washed through two changes of 0.2 M sodium
cacodylate buffer, pH 7.4, and were postfixed for 1 hour in 1.0% 0s04 in 0.15 M sodium
cacydylate buffer.
Samples were washed through two changes of 0.2 M sodium cacodylate buffer and
dehydrated through a graded acetone series (X3), infiltrated through several mixtures of
acetone/resin, and embedded in freshly prepared Medcast-Araldite 502 resin (Ted Pella,
Inc., Irvine, CA) and polymerized at 60°C for 24 hours (hayat, 1986). Ultrathin sections
were cut on an ultramicrotome (model MT-500; Sorvall Instruments Div., DuPont Co.,
Newton, CT) with a sapphire knife. Sections 60-70 nm in thickness were collected on 300
hexagonal mesh copper grids and stained with saturated ethanolic uranyl acetate followed
by bismuth subnitrate. The sections were examined with a transmission electron
microscope (EM 10/A; Zeiss, Oberkochen, Germany) at an accelerating voltage of 60 kV.
DNA fragmentation analysis. SFME cells (2 x 107/plate) were seeded in 10-cm-
diam. plates in serum-free medium with 5 ng/ml EGF. The following day some plates
were changed to serum-free medium without EGF (10 ml/plate). Control plates were
changed to serum-free medium with 50 ng/ml EGF. At the indicated times duplicate cells
were harvested by scraping into cold phosphate-buffered saline with calcium or magnesium
(PBS) followed by centrifugation at 4°C in 1,000 g for 5 minutes. The supernatant was
aspirated and the cell pellets were immediately stored at -86°C.
Cell pellets were thawed on ice and incubated at 37°C for 4 hours in 4 ml of lysis
buffer (200 mM Tris, pH 8.5, 100 mM EDTA, 50 j.ig/m1 proteinase K, 1% SDS). The
DNA solution was extracted twice with phenol. The aqueous phase was dialyzed overnight
against 10 mM Tris HC1, pH 7.5, 1 mM EDTA. After dialysis the DNA solution was
incubated at 37°C for 5 hours with 50 p.g/ml of DNase-free RNase A. Proteinase K (50
157
µg/ml) was added and the DNA solution was further incubated at 37°C for 5 hours. The
DNA solution was extracted once with phenol and once with phenol:chloroform:isoamyl
alcohol (25/24/1). DNA was precipitated in ethanol and resuspended in distilled water.
DNA concentration was determined from the absorbance at 260 nm.
For analysis of fragmentation, DNA (20 µg/lane) was fractionated by
electrophoresis for 16 hours at 40 V on a 1.0% agarose gel with 90 mM Tris HC1, 90 mM
boric acid, 2 mN EDTA, pH 8.0 as running buffer. DNA was visualized by shortwave
UV illumination after ethidium bromide staining (Maniatis et al., 1982). A 123 by DNA
ladder (Bethesda Research Laboratories, Gaithersburg, MD) was used as a standard.
Enzyme assay. Culture medium (0.5 ml) from Balb SFME or Balb 3T3 cultures
maintained for the indicated times under serum-free conditions in the presence or absence
of EGF was added to an equal volume of lactate dehydrogenase assay reagent (100 mM
lactate, 14 mM nicotinamide adenine dinucleotide, pH 8.9; #228-10; Sigma Diagnostics,
St. Louis, MO). Enzyme units were quantified by determining the rate of absorbance
change at 340 nm over a 60 second time interval.
Flow cytometry. SFME cells were seeded at 2 x 107 cells per 10-cm-diam. dish
and cultured overnight in serum-free medium with 5 ng/ml EGF. The following day some
plates were changed to serum-free medium without EGF (10 ml/plate). Control plates were
changed to serum-free medium with 50 ng/ml EGF. At the indicated times cells were
analyzed for cell size by measuring forward angle light scatter (FALS) or cell cycle by
measurement of DNA staining with chromomycin A3 (IGFL).
For size analysis cells were removed from plates with 0.25% crude trypsin/1 mM
EDTA in PBS and diluted into an equal volume of F12:DME containing 1 mg/ml soybean
trypsin inhibitor. Cells were washed with ice-cold PBS by centrifugation and resuspension
in 5 ml PBS. Forma lin (550 gl, 37%) was added while vortexing and fixed cells were
stored at 4°C.
158
For chromomycin staining, cells were removed from plates, centrifuged,
resuspended in 5.0 ml ice-cold PBS as described above, and then pelleted by
centrifugation. The supernatant was aspirated, leaving 0.5 ml in the bottom of the tube,
and 5.0 ml of ice-cold 70% ethanol was added dropwise while vortexing the cell pellet.
Cells were stored in ethanol at 4°C until all samples could be processed together. Fixed
cells were centrifuged, resuspended in chromomycin A3 (20 p.g/m1) with 15 mM MgC12,
and incubated for 0.5 hours at room temperature in the dark. Cells were centrifuged from
solution, resuspended (5 x 105 to 5 x 106 cells/m1) in ice-cold PBS, and filtered through a
40 pm nylon filter.
Basic procedures for flow cytometry were as described by Gray and Coffino
(1979). Flow cytometry was carried out with an Epics flow cytometer-cell sorter (Coulter
Electronics Inc, Hialeah, FL). Chromomycin A3-stained cells were excited with an output
of 100 mW of 457-nm light emitted from an argon laser. The resulting cell-asociated
fluorescence was captured through a 495-nm-long pass absorbance filter (Coulter
Electronics Inc.) and blocking interference filter (Corion Corp., Holliston, MA). Cell
sizing was analyzed with a ND-1 filter for FALS.
Polysome profiles. SFME cells were plated at 2 x 10 cells/10-cm-diam. dish and
cultured for 24 hours in serum-free medium with 5 ng/ml EGF. The medium was changed
on one plate to serum-free conditions with 50 ng/ml EGF and on the other plate to serumfree conditions without EGF. After 6 hours the medium was aspirated, plates were washed
twice with PBS containing 10 gg/m1 cycloheximide, and then stored frozen at -86°C. All
subsequent manipulations were performed on ice or at 4°C. Plates were scraped in 2x lysis
buffer (500 mM NaCl, 50 mM MgC12, 100 mM Tris, pH 7.5, 1% Triton X-100, 400 U/ml
RNasin, and 40 gg/m1cyclohexirnide) and if necessary sterile water was added to achieve a
lx lysis buffer concentration.
Lysates were centrifuged for 10 minutes at 13,000 rpm in a precooled (model
SS34; Sorvall Instruments Div., DuPont Co.) rotor. The supernatant was carefully
159
transferred to the top of a freshly prepared 15-50% sucrose gradient (250 mM NaC1, 25
mM MgC12, 50 mM Tris, pH 7.5, 20 µg/ml cycloheximide, and 15-50% sucrose).
Gradients were centrifuged for 130 minutes in a precooled rotor (model SW40; Beckman
Instruments Inc., Palo Alto, CA) at 32,000 rpm and scanned from top to bottom at 254 nm
using a model 185 density gradient fractionator (ISCO, Lincoln, NE) at a flow rate of
0.375 ml/minute (Gross and Merrill, 1989). Absorbance was monitored by a UA-5
absorbance/fluorescence detector and type 6 optimal unit (ISCO), which was calibrated to
an internal standard immediately before use.
Northern blot hybridization analysis. For isolation of cytoplasmic RNA, plates
containing SFME cells were washed with ice-cold PBS, cells were scraped into cold PBS
and centrifuged (4°C, 1,000 g, 5 minutes). Cell pellets were resuspended in 10 mM Tris
HC1, pH 7.6, 140 mM NaC1, 1.5 mM MgC12, 200 p.g/ml heparin, lysed by added 0.5%
NP-40, and vortexed for 20 seconds, and then centrifuged at 4,000 g for 10 minutes to
pellet nuclei. 1% SDS and 10 mM EDTA were added to the supernatant followed by
extraction twice each with phenol, phenol:chloroform:isoamyl alcohol (25/24/1), and
chloroform:isoamyl alcohol (24/1). RNA was precipitated in ethanol and stored in
diethylpyrocarbonate-treated water at -86°C. Poly(A)+ RNA was prepared from
cytoplasmic RNA using standard methods (Maniatis et al., 1982).
Poly(A)+ RNA (1.0 lig/lane) was fractionated by electrophoresis on a 1.2%
agarose gel containing 6.6% formaldehyde and transferred for 24 hours to a nitrocellulose
membrane (Maniatis et al., 1982). The plasmid pCD-beta actin was linearized with HindIII
and labeled (5 x 108 cpm/p.g DNA) by oligonucleotide-primed extension (kit from
Boehringer Mannheim Diagnostics, Indianapolis, IN). The membrane was prehybridized
for 4 hours, and then hybridized with probe (2 x 106 cpm/ml) for 20 hours in 50%
formamide, 50 mM sodium phosphate, pH 6.5, 0.8 M NaC1, 1 mM EDTA, 0.1% SDS,
0.05% BSA, 0.5% Ficoll, 0.05% polyvinylpyrrolidone, 250 Mg/ml sonicated salmon
sperm DNA, and 500 µg/ml yeast RNA at 57°C. Filters were washed twice with lx SSC
160
and 0.1% SDS at room temperature for 30 minutes each, and then twice with 0.5x SSC
and 0.1% SDS at 65°C for 30 minutes each. Autoradiography was carried out on film (XOMAT AR; Kodak Co., Rochester, NY) for 12-36 hours at -86°C.
161
RESULTS
Ultrastructural changes associated with EGF deprivation. SFME cells were
cultured in the absence of EGF and fixed for EM 4, 8, 16, and 24 hours later. Control
cells grown in the presence of EGF were fixed at the beginning and end of the experiment.
Cells grown in the presence of EGF had a high nuclear to cytoplasmic ratio relative to most
cultured cells and contained abundant aggregates of polyribosomes that were generally
uniformly distributed (Fig A2.la and b). An additional prominent cytoplasmic feature of
control cells was frequent large vacuoles and multivesicular bodies (MVB) that tended to be
regionally located. A well-developed Golgi complex and abundant coated vesicles were
present and intermediate filaments and microtubules were prominent. Cytoplasmic lipid
droplets were rarely seen. Nuclear chromatin appeared as a thin peripheral rim of
heterochromatin with small aggregates of heterochromatin also randomly dispersed
throughout the nuclear matrix. In most nuclei, single or multiple compact nucleoli were
seen either centrally or marginally located adjacent to the nuclear envelope.
SFME cells cultured in the absence of EGF for 4 hours exhibited a loss of
polyribosomes (Fig A2. 1c) and predominance of monosomes. When SFME cells were
cultured for 8 hours without EGF, cellular degeneration, cell death, and other
ultrastructural pathologic alterations were visible. Cellular morphology was pleomorphic;
some cells were small and round, while other cells were larger and elongated with distinct
ameboid-like cytoplasmic projections (Fig A2.2a and b).
Some of the small cells were severely degenerate and contained electron-dense
nuclei with a thick band of marginated chromatin in a circumscribed or crescent form with
nucleoli condensed or absent. Cytoplasm in these degenerate cells was condensed and
cytoplasmic organelles were scant but recognizable; occasional cytoplasmic blebs or
projections were found. A portion of cytoplasmic and nuclear matrix of most degenerate
cells displayed a light-grey granular matrix. The few mitochondria found in degenerate
162
cells with light grey cytoplasmic matrix were normal, whereas the few profiles of rough ER
observed were slightly dilated or vesicular with a loss of polyribosomes.
Dead cells were distinguished by prominent electron-dense cytoplasm containing
lucent cytoplasmic vacuoles and pyknotic nuclei. The nuclei had either a wide, electrondense marginated band of chromatin with the remaining nuclear matrix of similar electron
density to that of the cytoplasm, or the nucleus was round and filled with electron-dense
chromatin. Nucleoli were not found in these cells.
The predominant population of cells in the absence of EGF appeared smaller than
control cells cultured in the presence of EGF because of a loss of cytoplasm. Cytoplasmic
projections with the remaining portion of the cell out of the plane of section and
cytoplasmic debris were scattered throughout the fields, suggesting that cells may have
been in an early phase of disintegration. Frequently, nuclei were found that contained large
nucleoli composed primarily of a light granular component (Fig. A2.2c).
At 16 hours, a greater proportion of SFME cells grown without EGF were
ultrastructurally altered than at 8 hours; however, the extent of the ultrastructural changes,
except for minor differences, were similar to those changes described at 8 hours.
In the remaining cells the nuclear envelope was frequently deeply indented or
invaginated resulting in segmented nuclei. Nucleoli, if found, were condensed and small.
Cytoplasmic vacuoles appeared expanded and occupied a larger proportion of the
cytoplasm. Lipid droplets were commonly found, Golgi were atrophied, intermediate
filaments, and coated vesicles were difficult to find, and aggregates of polyribosomes were
sparse. Nuclear karyorrhexis, and apparent segmental loss of the nuclear envelope was
occasionally seen in some early degenerate cells, but cytoplasmic organelles in these cells
were recognizable. This finding suggests that nuclear condensation and eventual pyknosis
occurred before cytoplasmic condensation and increased electron density.
The majority of SFME cells cultured without EGF for 24 hours were either dead or
degenerate (Fig. A2.3a). Dead cells, severely degenerate cells, cytoplasmic projections,
163
and cellular debris were similar in appearance to that described above, but were a greater
fraction of the total population relative to 8 and 16 hour samples. The remaining cells
appeared to be degenerate because the cytoplasm was filled with large vacuoles and
abundant lipid droplets. Vacuoles often contained abundant membranous material in
circular or concentric myelin figure profiles (Fig. A2.3b). Vacuoles also contained variable
amounts of electron-dense amorphous debris and vesicles.
Biochemical and physiological changes associated with EGF deprivation.
Fragmentation of DNA isolated from SFME cells cultured without EGF was detected 8
hours after removal of the growth factor (Fig. A2.4). The pattern of degradation into
oligonucleosome-length fragments, generating a regularly spaced "ladder," is characteristic
of programmed cell death (Arends et al., 1990) resulting from activation of endonuclease
cleaving chromatin into polynucleosomes. Lactate dehydrogenase released by cells into the
culture medium was assayed as an indication of cell lysis. A large increase in enzyme
activity was only observed in SFME cell medium at relatively late times after removal of
EGF from the cultures (Fig. A2.5). No major release of lactate dehydrogenase over the
same time course was detected in identical experiments with 3T3 mouse embryo cells,
which are not acutely dependent on EGF for survival (Shipley and Ham, 1983).
Nuclease assay of medium from SFME cells cultured in the absence of EGF did not
detect increased activity before the time in which a general release of cellular enzymes was
detected by assay of lactate dehydrogenase, indicating that nuclease causing the cleavage
pattern of Fig. A2.4 probably was not released as cells underwent the progressive changes
in the early phases leading to cell death resulting from EGF deprivation. Incubation of
SFME cells in the presence of EGF with extracts of SFME cells cultured without EGF for
24 hours did not lead to cell death.
Light and electron microscopic examination suggested that SFME cells became
smaller after removal of EGF. This was demonstrated quantitatively by flow cytometric
analysis of forward angle light scatter. The average size of the population of cells became
164
smaller with time after removal of EGF, while SFME cells from parallel cultures incubated
with EGF did not show this change (Fig. A2.6). Analysis of DNA content of SFME cells
incubated without EGF showed that the percentage of cells in G2 and S phase of the cell
cycle markedly decreased with time after the removal of EGF (Fig. A2.7). Most of the
cells 24 hours after removal of EGF were in the 01 stage of the cell cycle. No timedependent change in the relative percentages of cells in cell cycle stages was seen over the
course of the experiment when EGF was included in the medium.
To confirm the electron microscopic observations that polysomal aggregation
decreased in SFME cells upon removal of EGF, we generated polysome profiles by
gradient centrifugation and spectrophotometrically quantified the relative polysome and
monosome content of SFME cells cultured for 6 hours with or without EGF (Fig. A2.8).
Integration of the spectrographic tracings revealed that polysome-associated ribosomes
were reduced 50% in EGF-deprived cells, relative to controls, with a corresponding
increase in monosomes and ribosomal subunits. It is unlikely that the shift from
polysomes to monosomes was because of an overall reduction in steady state level of
mRNA in the cultures incubated without EGF, because Northern blot hybridization
analysis did not detect a change in actin mRNA in SFME cells after 16 hours without
EGF(Fig. A2.9).
165
DISCUSSION
NGF-dependent neurons in vitro represent a well-studied growth factor-dependent
cell culture system. NGF is necessary to establish cultures of these cells and neurons die
within 48 hours in vitro if the growth factor is omitted from the medium (Yanker and
Shooter, 1982; Thoenen and Barde, 1980; Martin et al., 1988). The behavior of these cells
in vitro has been compared to programmed neuronal cell death in vivo, and the ability of
NGF to promote neuronal survival in vivo has been demonstrated (Hamburger et al., 1981;
Yip and Johnson, 1984). EGF-dependent SFME cells represent another growth factor-
dependent culture system of neural tissue origin (Sakai et al., 1990). Growth factordependent systems also exist among hematopoietic stem cells: interleukin-dependent
lymphocytes and erythropoietin-dependent reticulocyte progenitors (Koury and Bondurant,
1990; Rodriguez-Tarduchy and Lopez-Rivas, 1989; Nicola, 1989).
Identification of SFME cells as astrocyte precursors suggests mechanistic and
developmental relationships may exist between the NGF dependence of neurons and the
EGF dependence of SFME cells. In one respect, however, the SFME culture system more
closely resembles the hematopoietic cells. Like lymphocytes and reticulocyte precursors,
and unlike neurons, the cells are capable of proliferation in vitro, and the growth factor
promotes both proliferation and survival. A common feature among all the growth factor-
dependent cell types is the loss of the growth factor requirement after oncogenic
transformation; in some instances malignant transformation is accompanied by autocrine
production of growth factor or activation of a receptor (Shirahata et al., 1990; Isfort, 1990;
Bishop, 1987).
Analysis of DNA from SFME cells cultured without EGF revealed a pattern
consistent with the notion that the cells underwent degradation of chromatin in the regions
unprotected by nucleosomes. This pattern is characteristic of cells undergoing apoptosis or
programmed cell death, through which an endogenous endonuclease is activated (Arends,
166
1990). The time course of appearance of fragmented DNA in cultures incubated without
EGF indicates that increased levels of fragmented DNA occurred coordinately with
increased frequency of appearance of dead cells with pyknotic nuclei. Degenerate cells in
our cultures show ultrastructural similarities to "phase 1" apoptotic cells, identified by
Arends et al. (1990) as undergoing early phases of programmed cell death, and may not yet
have undergone extensive DNA fragmentation associated with pyknotic nuclei. Cells
identified as degenerate appeared before the large scale manifestation of dead cells with
pyknotic nuclei, also suggesting that degenerate cells are precursors of apoptotic cells.
In experiments in which EGF was removed for 8 to 24 hours and then added back,
we found that incubation for only 8 hours in the absence of EGF reduced by 50% the
number of cells that ultimately were recovered after EGF deprivation (Loo et al., 1989b),
suggesting that the degenerate cells traverse an irreversible pathway leading to death.
Incubation for 24 hours in the absence of EGF reduced by 80% the number of cells that
recovered after EGF deprivation (Loo et al., 1989b; Rawson et al., 1990). At this time
most of the cells in the cultures appeared either dead or degenerate by ultrastructural
analysis and significant leakage of cytoplasmic enzyme into the culture medium was
detected.
Assay of cytoplasmic enzyme release into the medium did not detect a large release
at earlier time points when ultrastructural and some biochemical changes were obvious,
indicating that plasma membrane integrity was maintained through this early period and
implying that degenerate cells were not highly permeable. Nuclease activity was not
released in a specific manner before the general breakdown of membrane integrity and
extracts of dying cells did not kill SFME cells in the presence of EGF, suggesting that cell
death resulting from EGF deprivation is not recruited in some cells by release of enzyme
from other cells. Flow cytometry showed that the cultures in the absence of EGF for 24
hours were entirely in the G1 phase of the cell cycle. This probably reflects an inability of
167
cells to progress into S phase in the absence of EGF, and may also indicate an initial
preferential loss of cells in S or G2 in the absence of EGF.
Ultrastructural analysis of SFME cells cultured in the absence of EGF identified a
time-dependent reduction in polysomes, alterations in nuclear morphology, a reduction in
cell size and cytoplasmic volume and the appearance of lipid or membrane-containing
vacuoles and apoptotic nuclei. The appearance of lipid-containing vacuoles and
multivesicular bodies and a progression to small cell size has been noted in neurons
deprived of NGF (Martin et al., 1988). In both culture systems apoptotic nuclei appear
with little evidence of a transition state, suggesting that this change occurs rapidly (Martin
et al., 1988). Primary disturbances in ER are observed in the neuronal system, but were
not prominent in SFME cells without EGF, and the shift from polysomes that we observed
with SFME cells is not seen in neurons cultured without NGF (Martin et al., 1988).
The reduction in SFME cell size perceived microscopically was confirmed by flow
cytometric analysis. This effect appears to be the result of fragmentation and loss of
cytoplasm from the cells; the small cells are essentially nuclei surrounded by a minimal
amount of cytoplasm and plasma membrane. A loss of polysomes was the earliest change
in SFME cells observed in the electron micrographs after removal of EGF, and the EGFdependent polysome shift was confirmed in sedimentation profiles. mRNA remained intact
in these cells, suggesting that an EGF-dependent regulation of translation may be
responsible.
EGF and other peptide-growth factors stimulate phosphorylation of ribosomal
protein S6 in other cell culture systems (Oliver et al., 1988), but the functional significance
of this effect has not been established. The relationship of the ultrastructural, biochemical,
and physiological alterations we have observed in EGF-deprived SFME cells to the primary
mechanism of cell death in this system, as well as the means by which inhibitors of protein
and RNA synthesis prevent cell death, remains to be determined.
168
ACKNOWLEDGMENTS
Supported by National Institutes of Health National Aging Institute-07560 and
Council for Tobacco Research 1813. D. Barnes is the recipient of Research Career
Development Award, NIH-National Cancer Institute-01226. Flow cytometry was carried
out through the Oregon State University Environmental Health Sciences Center Flow
Cytometry Core Unit (NIEHS ES00210).
Cathleen Rawson performed flow cytometry experiments and assisted with
polysome profiles. Julie Duimstra and Olaf Hedstrom provided electron microscopy data.
Edward Schmidt performed polysome profile experiments. David Barnes provided
financial and scientific support.
169
Fig. A2.1. SFME cells cultured with EGF or without EGF for 4 hours. (A),
with EGF, low power. Polygonal or rounded cells, found in sheets, contain large, round
to oblong, slightly indented nuclei. Nuclei usually have a thin peripheral rim of
heterochromatin, abundant finely dispersed chromatin, and prominent nucleoli. (B), with
EGF, at higher magnification, the cytoplasm is rich in organelles: abundant polyribosomes
(P), prominent clear vacuoles (V) containing delicate membranous-like material,
multivesicular bodies (X), well-defined Golgi complex (G), coated vesicles (arrows),
short, slightly dilated profiles of RER (arrowhead), and mitochondria (M). (C), without
EGF for 4 hours. The first significant alteration detected was decreased polyribosomic
aggregates giving the cytoplasm an increased electron lucency. A few cytoplasmic lipid
droplets (L) were also found. A well-developed Golgi complex (G), clear vacuoles (V),
multivesicular bodies (X), intermediate filaments (F), and mitochondria (M) were
ultrastructurally similar to corresponding control organelles. Bars: (A) 10 gm; (B) 1 p.m;
(C) 1 pm.
170
Figure A2.1. A, B
14+
'
:.41.
Ira
.
f,...
4
171
Figure A2.1. C
im:aith14.-r
}arv
,
172
Fig. A2.2. Cell death and degeneration in SFME cells cultured for 8 hours in
the absence of EGF. (A and B), low power magnification of SFME cells depicting
pleomorphic cells, severely degenerate cells (arrows), dead cells (D), cellular debris
(arrowhead), and elongated cytoplasmic projections with the remaining portion of the cell
out of the plane of section (P). (C), higher magnification of SFME cells showed elongated
cytoplasmic projections (P), cytoplasmic blebs (B), atrophied Golgi (G), and lack of
aggregates of poyribosomes.
peripherally located.
Note the prominent light granular nucleolus (Nu)
173
Figure A2.2. A, B
174
Figure A2.2. C
175
Fig. A2.3. SFME cell death and severe degeneration after 24 hours of EGF
deprivation. (A), dead (D) or severely degenerate cells (arrows), and dispersed cellular
debris (arrowheads) were a major feature at this time period. The cytoplasm of remaining
cells contained abundant large vacuoles, some of which compressed adjacent structures
(i.e.., nucleus). Nuclei of these cells were usually indented or partially segmented. (B),
the presence of abundant cytoplasmic lipid droplets was an additional feature observed
during this time period. Note the numerous lipid droplets (L), peripheral collapsed
vacuoles, and condensed granular cytoplasm of a severely degenerate cell. In adjacent
cells, cytoplasmic vacuoles contained a predominance or a mixture of the following
structures: circular membranous material, vesicles, or cellular debris. Multivesicular
bodies (X) or in some vacuoles, membranous whorls (myelin figures) (M) were also seen.
Bars: (A) 1011m; (B) 1 p.m.
176
Figure A2.3.
ff
7,
0
,
.
-.:4
%,,e
.', .- "
Z
4
4
e''
;
',. 4
-%,
....
..--:.,
:4.: ,?1' v. -.!:.71.--1.44c:.
, ;,
, :. /5 . II, °
.> ,
...
,;
4.111 -5'''.-
.44471:11,s ;,i. '.) -L..,
,,,
:,
"
..1,...,...-.
.
,", tit:* r
..' .2.:4;',..--r,
,
T.-
:
,
to)
.
,
.1.
-
177
Fig A2.4. DNA fragmentation in EGF-deprived SFME cells. Cultures were
changed to medium without EGF and incubated for up to 24 hours. DNA was isolated
from samples during this time course, fractionated by electrophoresis, and visualized by
ethidium bromide staining. (A), 0 h without EGF; (B), 4 h without EGF; (C), 8 h without
EGF; (D) 16 h without EGF; (E), 24 h without EGF; (F), 24 h with EGF; (G), 123-bp
DNA ladder standard. Arrowheads mark bands comprising oligonucleosome ladders
indicative of DNA fragmentation by endonuclease (Arends et al., 1990).
ABCDEFG
178
Fig. A2.5. Release of cytoplasmic enzyme in culture medium by SFME cells
in the absence of EGF. SFME cells or 3T3 cells were incubated the indicated times in
serum-free medium with or without EGF and culture medium assayed for lactate
dehydrogenase as described. 3T3 cells are mouse embryo control cells that are not acutely
dependent on EGF for survival. (), with EGF; (), without EGF.
50
40
LDH 30
Iml. IND
20
10
0
Fil
1
4 HR
8 HR
SFME
i
24 HR
now
f-M
4 HR
8 HR
3T3
1
24 HR
179
Fig. A2.6.
Flow cytometric size analysis of SFME cells.
Cultures were
incubated with or without EGF and processed as described for flow cytometry size analysis
by measurement of forward angle light scattering. Tracings show relative cell number (Y
axis) versus relative volume per cell (X axis). (A), no EGF, 0 h; (B), no EGF, 8 h; (C),
no EGF, 24 h; (D), no EGF, 48 h; (E-H), EGF present, 0, 8, 24, 48 h.
180
Fig. A2.7.
Flow cytometric DNA analysis of SFME cells. Cultures were
incubated with or without EGF and processed as described for analysis of DNA content per
cell. Tracings show relative cell number (Y axis) versus relative amount of DNA per cell
(X axis). (A), no EGF, 0 h; (B), no EGF, 8 h; (C), no EGF, 24 h; (D-F), EGF present,
0, 8, 24 h.
A
181
Fig. A2.8 Polysome profiles of SFME cells. Cells were incubated with or without
EGF, lysed in the presence of cycloheximide, and polysome distributions were determined
by sedimentation in a sucrose gradient and spectroscopic measurement as described. (A),
without EGF, 6 h; (B), control, with EGF. Arrows show direction of sedimentation. (T),
absorptive material at the top of the gradient; (R), monosomes and ribosomal subunits; (P),
polysomes.
182
Fig. A2.9. Northern blot hybridization analysis of actin mRNA from SFME
cells. SFME cells were cultured as described in serum-free medium with (A) or without
(B) EGF for 16 hours followed by isolation of poly-A-containing RNA and blot
hybridization as described. The blot was probed with labeled plasmid containing the gene
for beta actin.
-4 28 S
110 le
AB
-418s