Combination therapies of robot rehabilitation and viral delivery of brain-derived neurotrophic factor (BDNF)
to lumbar cord promote large functional gains after a complete transection SCI in adult rats
John K. Lee1,2, Veronica J. Tom2, and Simon F. Giszter1,2
1School
3. PRELIMINARY DATA: Recovery of
Locomotion in Two SCI Models with
Robot-Assisted Rehabilitation
Changes in AOB score with Robot Training
20
B
Changes in % BWS with Robot Training
22
Adult Transection (ATX)
**
50
*
40
% BWS
14
15
*
*
12
10
8
30
6
0
10
0
Start of Robot Training
End of Robot Training
0
0
Start of Robot Training
TRUNK ROBOT REHABILITATION
Significant improvement in
locomotor function ON robot1,2
WITHOUT
any additional
intervention
WITH additional
intervention
•  ATX rats without any additional stimulation (perineal or
epidural) fail to improve function with robot rehablitation.
•  NTX (non-WS) rats and ATX Rats with robot-driven epidural
stimulation show significant improvement in locomotor
function on robot as seen through an increase in functional
kinematic recovery scores (A) and decrease in vertical %
BWS (body weight support) provided by the robot (B).
** Data from Hsieh / Giszter 2011. Reproduced with permission.
** NTX data from Oza CS, Giszter SF submitted for publication & Udoewkere UI, Oza CS, Giszter SF submitted for publication
NO significant
Significant
improvement in
improvement in
locomotor
locomotor function
function ON robot
ON robot3
•  Recent work by Boyce et al. where ATX rats were treated with
Brain-Derived Neurotrophic Factor (BDNF) administered at the
injury site using Adeno-associated virus-5 (AAV5) showed recovery
of autonomous stepping, post-injury.
•  We hypothesize adult spinalized rats given robot
rehabilitation and AAV5-BDNF treatment may learn to step and
weight support better than those that do not receive viral
treatment.
•  We thus propose to test a combined treatment approach,
using viral delivery of BDNF in addition to robot training to
induce autonomous stepping and potentially achieve greater
functional recovery, through synergistic interventions
10
15
20
25
Persistent Improvement
30
Days of Training
Start and End of Training
20
We observed
significant
improvement in
locomotor behavioral
scores in ATX rats
treated with combined
robot rehabilitation and
AAV5-BDNF after
training.
§
*
AAV5-GFP
15
AAV5-BDNF
10
n=5
n=8
5
0
End of Training
Start of
Training
20
75
50
25
0
15
10
5
0
0
5
10
15
20
25
30
0
5
10
Days of Training
15
20
25
30
35
25
30
35
Days of Training
Improvement Followed By Collapse
JLTX019 %WSS
JLTX019 AOB
Control
Complete adult SCI
(T9/10)
(n=8)
(n=5)
75
AOB Scores of BDNF Rats
20
50
25
0
15
10
5
0
0
5
10
15
20
25
30
35
0
5
10
Days of Training
15
15
20
Days of Training
AOB Score
JLTX011
AAV5-BDNF injection
1 mm caudal to injury
AAV5-GFP injection 1
mm caudal to injury
JLTX010
JLTX007
JLTX012
10
JLTX017
JLTX019
JLTX022
JLTX026
5
0
0
5
10
15
20
25
30
35
Days of Training
Robot Training
2. ROBOT REHABILITATION
(20 mins/day; 5 days/wk;
4-6 weeks)
Mean Vertical Interactive Force (zForce)
Robot Training
0.5
(20 mins/day; 5 days/wk;
4-6 weeks)
y = -0.004x + 0.4605
R² = 0.70865
0.45
5. METHODS
0.4
zForce (N)
AAV5-GFP is an Adeno-associated virus-5 that produces
Green Fluorescent Protein (GFP) when expressed
AAV5-BDNF
0.25
Linear (AAV5-GFP)
Linear (AAV5-BDNF)
0.2
•
Four microinjections of 2 µl were injected following the pattern
above
Total 8 µl of virus: 2.5 × 1010 viral particles
Saline (Control)
AAV5-GFP (Experimental)
0
1
2
3
4
5
6
7
8
Weeks of Training
We observed
significantly less
usage of robot
forces in the z-axis
in ATX rats treated
with combined robot
rehabilitation and
AAV5-BDNF after
training.
AAV5-GFP
AAV5-BDNF
zForce at Start and End of
Training
§
0.6
*
0.5
zForce (N)
•
C
ATX animals treated with AAV5-BDNF demonstrated significant
locomotor recovery compared to those treated with AAV5-GFP.
We found significant improvement in animals treated with BDNF
in these three outcome measures:
1.  AOB: BDNF animals showed significant improvement in overall
score, with score level indicating increased weight-support, rightleft hind limb alternation, and plantar foot placement.
2.  Robot Force: BDNF animals showed a significant decrease in use of
robot forces, indicating increased body weight support.
3.  %WSS: BDNF animals exhibited a significant increase in the
percentage of the number of steps they took that were weightsupporting. The pattern of recovery also closely resembled patterns
of AOB recovery.
The increased excitability of the spinal cord from viral
delivery of BDNF coupled with robot training has
potential to create new avenues of combined therapy
to treat SCI.
y = -0.0256x + 0.4662
R² = 0.90818
AAV5-GFP
R
7. CONCLUSIONS
Combining varied methods to improve locomotor recovery in
treating SCI may result in synergistic effects that may provide
more robust recovery.
0.35
0.3
F = force required to maintain height
k = stiffness (45 N/m)
x = height of the pelvis
x0 = desired equilibrium height
100
20
Complete adult SCI
(T9/10)
Udoekwere UI, Oza CS, Giszter SF submitted for publication
2 Oza CS, Giszter SF submitted for publication
3 Hsieh FH, Giszter SF. “Robot-driven spinal epidural stimulation compared with conventional stimulation in adult spinalized rats.” Conf Proc
IEEE Eng Med Biol Soc. 2011;2011:5807-10.
F = k ( x − xo )
JLTX012 AOB
100
1
Cantilevered Phantom haptic robot (SensAble Devices Inc.)
attaches via surgically implanted pelvic orthosis and
applies isotropic elastic force fields (stiffness k = 45 N / m)
at the pelvis during treadmill locomotion for 20 mins/day, 5
days/week, 4-5 weeks.
JLTX012 %WSS
* t-test (assuming unequal variances), p = 0.000013
§ paired t-test, p = 0.000009
4. EXPERIMENT
Experimental
5
End of Robot Training
AOB score
80% of rats have less
than 50% WS stepping
TRUNK ROBOT REHABILITATION
* p = 0.006, t-test; ** p = 0.006, paired t-test
n=5
* p<0.5, t-test
20% of rats greater than
50% WS stepping
AAV5-BDNF
5
2
Weight Support (WS)
10
AAV5-GFP
20
4
Non-weight Support
(NWS)
n=8
%WSS
NO Autonomous Stepping
Autonomous Stepping
*
16
AOB Score
Neonatal Transection (NTX)
(Post-natal day 5/6)
18
ATX-Robot
ATX-Robot + Epidural
NTX-Robot
AOB Score
20
Percent Weight-Supported Stepping
•  Control group (n=5): No animals developed any weightsupported stepping (WSS).
•  Experimental Group (n=8): All animals began with 0% WSS
•  5 of 7 were able to provide > 5% WSS
•  Average final %WSS: 43.73 ± 13.07 %WSS (three animals
were above 75%)
•  Significant improvement in %WSS between control and
experimental* and within the experimental group **.
Comparison of AOB Scores
AOB Score
A
6. RESULTS, cont.
%WSS
Complete Spinal Cord Injury (SCI) Model at T10
6. RESULTS
AOB Score
1. INTRODUCTION
of Biomedical Engineering, Science and Health Systems, 2Department of Neurobiology & Anatomy, Drexel University
0.4
n=5
n=8
0.3
0.2
0.1
0
Start of Training
End of Training
* t-test (assuming unequal variances), p = 0.0015
§ paired t-test, p = 0.0009
ACKNOWLEDGEMENTS
Supported by the Craig H. Neilsen Foundation and the National Institutes of Health (NINDS). We
would like to thank Michelle Klaw, from the Tom lab in the Department of Neurobiology & Anatomy
at the Drexel University College of Medicine, for her help in histology. We would also like to thank
Chintan S. Oza, from the Department of Biomedical Engineering at Duke University Pratt School of
Engineering, for his help while a graduate student in Giszter lab. Also, we would like to thank Kavon
Noorbehesht and Kristina Carter for their help in animal care.