Exp Brain Res (1990) 81:25-34
Experimental BrainResearch 9 Springer-Verlag1990
Spinal cord transplants enhance the recovery of locomotor function after spinal cord injury at birth E. Kunkel-Bagden and B.S. Bregman Department of Anatomy and Cell Biology, Georgetown University School of Medicine, 3900 Reservoir Road NW, Washington, DC 20007, USA ReceivedJuly 4, 1989 / Accepted November 16, 1989
Summary. Fetal spinal cord transplants placed into the site of a neonatal spinal cord lesion alter the response of immature CNS neurons to injury. The transplants prevent the retrograde cell death of immature axotomized neurons and support the growth of axons into and through the site of injury. In the present experiments we used a battery of locomotor tasks to determine if these transplants are also capable of promoting the recovery of motor function after spinal cord injury at birth. Embryonic (El4) spinal cord transplants were placed into the site of a spinal cord "over-hemisection" in rat pups. Three groups of animals were used: 1) normal control animals, 2) animals with a spinal cord hemisection only, and 3) animals with a spinal cord transplant at the site of the hemisection. Eight to twelve weeks later, the animals were trained and videotaped while crossing runways requiring accurate foot placement and footprinted while walking on a treadmill. The videotapes and footprints were analyzed to obtain quantitative measures of locomotor function. Footprint analysis revealed that the animals' base of support during locomotion was increased by a neonatal hemisection. The base of support in animals with transplants was similar to control values. Animals with a hemisection rotated their hindlimbs further laterally than did control animals during locomotion. A transplant at the site of injury modified this response. Normal animals were able to cross a grid runway quickly with only a few errors. In contrast, animals with a hemisection took a longer time and made more errors while crossing. The presence of a transplant at the site of injury enabled the animals to cross the grid more quickly and to make fewer errors than the animals with a hemisection only. Animals that received the transplants demonstrated qualitative and quantitative improvements in several parameters of locomotion. Spinal cord transplants at the site of neonatal spinal cord injury result in enhanced sparing or recovery of motor function. We suggest that this transplant induced recovery of function is a consequence of the anatomical plasticity elicited by the transplants. Offprint requests to: E. Kunkel-Bagden (address see above)
Key words: Spinal cord - Transplants - Recovery of function - Development - Locomotion
Rat
Introduction A number of studies have shown that transplants can ameliorate some deficits or mediate recovery of function following CNS damage. Transplants may influence recovery of host brain function by a variety of mechanisms (see Bj6rklund et al. 1987 for review) including nonspecific or negative consequences of the implantation surgery (Dunnett et al. 1987a), trophic actions (Haun and Cunningham 1984; Haun et al. 1989; Kesslak et al. 1986), diffuse release of hormones (Gash et al. 1980; Gibson et al. 1984; Arendash and Gorski 1982) or transmitters (Freed et al. 1981 and Dunnett et al. 1985), reinnervation of elements of the host by the transplant (McLoon et al. 1985; Sotelo and Alvardo-Mallart 1987; and Bj6rklund and Stenevi 1984), and formation of reciprocal connections between the transplant and the host (Sotelo and Alvardo-Mallart 1987). Peripheral nerve segments (Richardson et al. 1980, 1982), fetal brainstem containing either locus coeruleus (Nygren et al. 1977; Buchanan and Nornes 1986; Nornes et al. 1983) or raphe (Privat et al. 1986, and Foster et al. 1985), and embryonic spinal cord (Reier et al. 1986; Houle and Reier 1988) have been transplanted into mature spinal cord after injury. Embryonic spinal cord transplants fill the lesion cavity and become apposed to the host cord but promote only a limited growth of the host fibers into the transplant (Reier et al. 1986; Houle and Reier 1988). The locus coeruleus and raphe transplants into denervated cord survive and produce a network of nerve terminals in the host gray matter surrounding the transplanted cells (Nygren et al. 1977; Foster et al. 1985; Privat et al. 1986). Despite modest axonal outgrowth, some recovery of function, indicated by changes in sexual reflexes and hindlimb flexion reflexes, has been associated with the fetal brainstem transplants (Privat et al. 1988; Buchanan and Nornes 1986).
26 Lesions in the developing central nervous system often have effects dramatically different f r o m the same lesion in the adult. This is referred to as the infant lesion effect (Bregman and Goldberger 1983 a-c), and reflects the greater anatomical plasticity and greater sparing of function following a CNS lesion m a d e at birth, c o m p a r e d to the same lesion m a d e in the adult. The response of the immature nervous system is not uniform, however, and in m a n y cases, effects of injury to the developing nervous system are m o r e severe than those in the adult (Bregman and Goldberger 1982, 1983b, c). F o r example, axotomized immature neurons are m o r e likely to die than axotomized adult neurons (Bregman and Goldberger 1982, 1983 a-c). We have shown that embryonic spinal cord transplants placed into the d a m a g e d spinal cord of a neonate enhance the anatomical plasticity o f the immature CNS. Thus, transplants prevent the retrograde cell death o f immature axotomized rubrospinal neurons (Bregman and Reier 1986) and allow the growth of descending axons b o t h into and through the site of a spinal cord injury (Bregman 1987a, b; Bregman et al. 1989). Some o f these axons reach normal targets within the host spinal cord caudal to the lesion and transplant (Bregman 1987a, b; Bregman et al. 1989). We now want to determine if these anatomical changes induced by the fetal spinal cord transplants are associated with sparing and recovery of l o c o m o t o r function. The contribution of transplants to functional recovery can be evaluated in the spinal cord. The spinal cord has the intrinsic capacity to mediate basic l o c o m o t o r patterns (spinal pattern generator for locomotion, S P G L ) which are controlled and modified by b o t h descending and segmental afferent input (Grillner 1975). The serotonergic, rubrospinal, corticospinal and dorsal root systems contribute to the descending and ascending influences on the generation o f locomotion (Donatelle 1977; Nygren and Olson 1977; Westlund et al. 1983; Martin and Ghez 1988). Each of these systems exhibits anatomical plasticity elicited by the presence of a transplant (Bregman and Reier 1986; Bregman 1987a, b; Bregman et al. 1989 and unpublished results). We suggest that through such neuronal projections the transplants influence the generation of locomotion and thus permit functional recovery. Since there is considerable recovery of function after spinal cord injury, mediated by intact descending and dorsal root afferent systems (Bregman and Goldberger 1982, 1983b; Goldberger 1988a, b) as well as by intrinsic spinal systems (Barbeau and Rossignol 1987), it is essential to use quantitative methods for analysis of reflex and l o c o m o t o r function to distinguish intrinsic recovery mechanisms f r o m transplant mediated alterations in m o t o r function. In the present study we have developed behavioral tests to assess several specific m o t o r functions quantitatively in the rat. This quantitative a p p r o a c h has allowed us to distinguish between the contribution of recovery of function due to reorganization o f the host CNS and recovery o f function due to the presence of a transplant. In order to test the hypothesis that transplants of fetal spinal cord tissue placed into the site of a neonatal spinal cord lesion p r o m o t e sparing and recovery of lo-
c o m o t o r function, we assessed the l o c o m o t o r function of three groups of animals: 1) normal control animals, 2) animals that received a spinal cord "over-hemisection" as neonates, and 3) animals with spinal cord transplants within the site of the neonatal hemisection. As adults, animals were evaluated using quantitative and qualitative tests of l o c o m o t o r function to determine the accuracy and pattern of locomotion o f the animals. Our results indicate that the presence of a transplant enhances recovery or sparing of some l o c o m o t o r functions, but not others, after spinal cord injury at birth.
Methods
Spinal cord lesions and transplantation procedures A total of 47 animals were used in this study. Three groups of animals were used: 1) normal control (CON) animals, 2) animals with a spinal cord hemisection only (HX), and 3) animals with a spinal cord transplant at the site of the hemisection (HX + TP). Sprague--Dawley rat pups (48 h postnatal, N= 36) were anesthetized by hypothermia. Under a dissecting microscope, the spinal cord was exposed by partial laminectomy at T5-T6. The dura was opened and the dorsolateral sulci, dorsal rootlets, and dorsal columns identified bilaterally. Iridectomy scissors were used to sever both dorsal columns plus the right lateral and ventral funiculus (over-hemisection). A lesion of this type is very reproducible and results in good apposition of the transplant at the rostral and caudal ends of the injury. In contrast to a transection the survival of our animals with hemisections is excellent (nearly 100 %). Lesions of this type also allow us to compare one limb to another. In 18 of the animals a transplant of fetal spinal cord tissue was placed into the lesion site and covered with an artificial dural material (durafilm, Codman-Shurtleff, Inc.). Transplantation techniques were identical to those described previously (Bregman and Reier 1986; Reier et al. 1986; Bregman 1987a, b). Briefly, pregnant Sprague-Dawley rats, 14 days gestation, were anesthetized with chloral hydrate and laparatomized. Fetuses were removed individually as required and placed in sterile culture medium (Dulbecco's Modified Eagle's Medium, Gibco). The spinal cord was removed from the vertebral canal and stripped of meninges and dorsal root ganglia. The thoracic spinal cord was blocked into transverse sections 2-3 mm in length. The donor tissue was floated onto sterile forceps and placed into the lesion site. In the lesion control animals gelfoam was placed into the site of the hemisection and covered with durafilm. The muscle and skin were sutured in layers and the wound covered with celloidin. Animals were warmed, received prophylactic injections of antibiotics (bicillin, Wyeth Laboratories) and returned to their mothers. Unoperated control littermates (N = 11) were used and housed with the lesioned animals (10 pups per litter). At 4 weeks of age the animals were weaned, males and females separated, and housed 2 or 3 per cage in the animal care facilities under a 10/14 hr light/dark cycle.
Behavioral training At 8-12 weeks of age behavioral testing began. Water was removed 48 h before training was to begin, Animals were trained to cross the runways or walk on the treadmill. During the training period of five days, animals received water only upon crossing the runways or while walking on the treadmill (12 ml/day). Once the rats were trained, they were allowed free access to additional water at the end of the testing session for 15 min daily. Animal weights were monitored and all animals received supplemental water if weight loss in any animal was noted. After training for 5 days all animals were able to walk at the speed (6.6 m/rain) of the treadmill.
27
Fig. 1. A treadmill was used to produce continuous locomotion by the animals. Animals were trained for five days to walk at the speed of the treadmill. The animal's feet were inked to obtain footprints while on the treadmill
Locomotor function tests
of animals reported for the different behavioral analyses in the results differs.
All the animals were tested by observers unaware of the animals' surgical treatment. During the testing sessions qualitative observations were made on the animals. In an early group of animals in this study, only some of the behavioral tests were examined. Since footprints were not consistently legible in this group, the number
m m m m
RS~ 9
9
9 9
m 9
9 9
m 9
9
m
LsLm 9
Footprint analysis. The treadmill used in our locomotor analysis is illustrated in Fig. 1. Footprint analysis was modified from that described by de Medinacelli (de Medinacelli et al 1982). The animals fore- and hindpaws were inked and the footprints were recorded on paper covering the surface of the treadmill. Measurements of different parameters of the locomotor pattern were calculated from the hindlimb footprints. A reconstruction of the footprints is illustrated in Fig 2. Measurements were made only on serial prints (a minimum of 3 consecutive prints) and only when the animal was walking continuously. Mean values ( N = 6 - 1 0 ) for each animal were calculated from prints obtained on two days of testing. The base of support was determined by measuring the distance between the central pads of the hindfeet (DBF in Fig. 2). The angle formed by the intersection of lines from the left and right prints was used as an indication of the limb rotation (R in Fig. 2). Stride length (RSL, LSL in Fig. 2) was measured between two consecutive prints on each side.
Grid. The animals were trained to cross a grid runway 180 cm long with 50 mm x 50 mm holes (Fig. 3). The animals crossed in both
9 9
m m m
Fig. 2. An illustration of hindlimb footprints reconstructed from a normal animal's prints indicating the measurements used in this study. Base of support was measured as the distance between the left and right prints (DBF). Limb rotation (R) is the angle produced by the left and right prints. Stride length (RSL, LSL) is the distance the right or left limb traverses with each step cycle
Fig. 3. Accuracy of locomotion was assessed on a grid runway. Animals were trained to cross a grid of 180 cm length with 50 m i n x 50 mm holes. The time to cross the runway and the number of errors in foot placement (footfalls below the grid) of the ipsilateral (right) and contralateral (left) hindlimb were counted. The arrow indicates an error with the right hindlimb
28
Fig. 4. Photomicrograph of a cross-section through a typical transplant within the host spinal cord. The transplant (TP) is in good apposition with the host spinal cord; the border between host and
directions (4 crossings/day) for the water reward at the end of the runway. The animals were trained for 5 days then videotaped for 5 days. The time to cross the runway and the number of errors made by either hindlimb was determined from the slow motion analysis of the videotapes. The errors measured were foot falls when the animal's foot was misplaced so that it fell into the grid hole, rather than being placed onto the rungs. The mean time to cross and number of errors were determined for each animal from the 20 runs.
Statistical analysis. Single factor ANOVAs were carried out using the measurements for each parameter calculated from the footprints and the data obtained crossing the grid. Between-group differences were examined with 2-sample T-tests. The Dynastat Professional (Dynamic Microsystems, Incorporated, Washington, D.C.) statistical software package was used for the statistical calculations. Lesion analysis At the completion of the behavioral testing the animals were sacrificed and perfused intracardially with 0.9% saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. Serial transverse 18 gm cryostat sections were stained with cresyl violet. A microprojector and a Zeiss microscope were used to reconstruct the lesion site and to determine the presence of a transplant. The histological characteristics of the embryonic tissue enabled us to
transplant is indicated by arrows. The host ventral horn (VH), lateral funiculus (LF) and ventral funiculus (VF) are indicated. Cresyl violet stain, scale bar is 500 gm
distinguish host tissue from transplant and allowed the areas of apposition between host and transplant to be determined.
Results
Spinal cord lesions and transplants T h e a p p e a r a n c e o f the s p i n a l c o r d t r a n s p l a n t s (Fig. 4) a n d the p e r c e n t survival were similar to t h a t d e s c r i b e d in o u r p r e v i o u s studies ( B r e g m a n 1987a, b ; B r e g m a n et al. 1989). I n m o s t cases there were extensive a r e a s o f a p p o s i t i o n b e t w e e n the t r a n s p l a n t s a n d the h o s t spinal c o r d . A n a l y s e s o f the b e h a v i o r a l tests were c a r r i e d o u t o n a n i m a l s w h o s e lesions were w i t h i n the m a x i m u m a n d m i n i m u m r a n g e o f d a m a g e (Fig. 5). T h e m a x i m u m lesion i n c l u d e d was the entire right side o f the c o r d p l u s on the left the d o r s a l c o l u m n , p o r t i o n s o f d o r s o l a t e r a l funiculus, i n t e r v e n i n g g r a y m a t t e r a n d a s m a l l p o r t i o n o f the vent r a l funiculus. T h e m i n i m u m a m o u n t o f d a m a g e i n c l u d e d was loss o f the d o r s a l c o l u m n o n the left a n d the entire r i g h t side except a v e r y small p o r t i o n o f the v e n t r a l
29
Maximum
Minimum
HX
a
Maximum
b
Minimum
HX+TP
Fig. 5a, b. Lesion reconstruction from serial sections are plotted onto standard diagrams of the cord for comparison. Maximum and minimum transverse extent of the lesion included in the study are illustrated in a. Only animals with transplants apposed to the host spinal cord were included within the study. Representative lesions plus transplants are drawn in b funiculus. We used the same lesion criteria for both the lesion group and the lesion plus transplant group; thus the transverse extent of damage in both groups was similar. Animals in the transplant group had the additional requirement for inclusion in these studies, that the transplant had to be present and apposed to the host spinal cord. Reconstruction of two representative transplants are shown in Fig. 5b. O f the 36 operated animals, 25 (13 H X and 12 H X + T P ) fulfilled all o f our lesion criteria for inclusion in the behavioral analysis.
Qualitative observations of locomotion Both hemisection only and hemisection plus transplant animals were capable of considerable locomotor function. All lesioned animals were able to walk easily on wide runways (requiring little accuracy of foot placement in order to cross) using all four limbs. Neonatal spinal cord lesions led to an apparent increase in the base of support and in the amount o f rotation between the hindlimbs (Fig. 6). In the animals with a neonatal lesion and transplant, both the amount o f rotation and the base o f support were similar to that observed in control animals. Control animals had little difficulty crossing runways requiring accurate foot placement. They crossed quickly and rarely made errors in foot placement. Both lesioned groups made many errors in foot placement while crossing the grid runway. The animals with transplants, however, recovered rapidly from the error, while the lesion
Fig. 6a-e. Photographs taken from videotapes of the rats walking on the treadmill, a Control (CON); b lesion plus transplant (HX + TP); e lesion only (HX). Neonatal hemisection results in permanent deficits in locomotion (see text). The presence of a transplant at the lesion site results in a pattern of locomotion that is similar to that in control animals
only group repeatedly flexed and extended the hindlimb in the grid hole, prior to regaining support. With all qualitative observations of locomotion, the animals which no longer had transplants were indistinguishable from lesion only animals.
Footprint analysis A neonatal hemisection resulted in a significant increase in the animals' base of support (HX = 50 ram) when
30
BASE OF SUPPORT (DBF)
,11 I
I i1#
60. I ~
* W i,
con HX HX+TP
~
50.
Z LLI W
40 bJ if) W (.3
z
\
30
CO
20
CON
HX
HX
HX+TP
Fig. 7. Base of support (see diagram Fig. 2). The mean ( + SEM) distance between the feet for each group. This distance was significantly (indicated by an asterisk) greater in the hemisection group (HX, N = 7, p
Experimental BrainResearch 9 Springer-Verlag1990
Spinal cord transplants enhance the recovery of locomotor function after spinal cord injury at birth E. Kunkel-Bagden and B.S. Bregman Department of Anatomy and Cell Biology, Georgetown University School of Medicine, 3900 Reservoir Road NW, Washington, DC 20007, USA ReceivedJuly 4, 1989 / Accepted November 16, 1989
Summary. Fetal spinal cord transplants placed into the site of a neonatal spinal cord lesion alter the response of immature CNS neurons to injury. The transplants prevent the retrograde cell death of immature axotomized neurons and support the growth of axons into and through the site of injury. In the present experiments we used a battery of locomotor tasks to determine if these transplants are also capable of promoting the recovery of motor function after spinal cord injury at birth. Embryonic (El4) spinal cord transplants were placed into the site of a spinal cord "over-hemisection" in rat pups. Three groups of animals were used: 1) normal control animals, 2) animals with a spinal cord hemisection only, and 3) animals with a spinal cord transplant at the site of the hemisection. Eight to twelve weeks later, the animals were trained and videotaped while crossing runways requiring accurate foot placement and footprinted while walking on a treadmill. The videotapes and footprints were analyzed to obtain quantitative measures of locomotor function. Footprint analysis revealed that the animals' base of support during locomotion was increased by a neonatal hemisection. The base of support in animals with transplants was similar to control values. Animals with a hemisection rotated their hindlimbs further laterally than did control animals during locomotion. A transplant at the site of injury modified this response. Normal animals were able to cross a grid runway quickly with only a few errors. In contrast, animals with a hemisection took a longer time and made more errors while crossing. The presence of a transplant at the site of injury enabled the animals to cross the grid more quickly and to make fewer errors than the animals with a hemisection only. Animals that received the transplants demonstrated qualitative and quantitative improvements in several parameters of locomotion. Spinal cord transplants at the site of neonatal spinal cord injury result in enhanced sparing or recovery of motor function. We suggest that this transplant induced recovery of function is a consequence of the anatomical plasticity elicited by the transplants. Offprint requests to: E. Kunkel-Bagden (address see above)
Key words: Spinal cord - Transplants - Recovery of function - Development - Locomotion
Rat
Introduction A number of studies have shown that transplants can ameliorate some deficits or mediate recovery of function following CNS damage. Transplants may influence recovery of host brain function by a variety of mechanisms (see Bj6rklund et al. 1987 for review) including nonspecific or negative consequences of the implantation surgery (Dunnett et al. 1987a), trophic actions (Haun and Cunningham 1984; Haun et al. 1989; Kesslak et al. 1986), diffuse release of hormones (Gash et al. 1980; Gibson et al. 1984; Arendash and Gorski 1982) or transmitters (Freed et al. 1981 and Dunnett et al. 1985), reinnervation of elements of the host by the transplant (McLoon et al. 1985; Sotelo and Alvardo-Mallart 1987; and Bj6rklund and Stenevi 1984), and formation of reciprocal connections between the transplant and the host (Sotelo and Alvardo-Mallart 1987). Peripheral nerve segments (Richardson et al. 1980, 1982), fetal brainstem containing either locus coeruleus (Nygren et al. 1977; Buchanan and Nornes 1986; Nornes et al. 1983) or raphe (Privat et al. 1986, and Foster et al. 1985), and embryonic spinal cord (Reier et al. 1986; Houle and Reier 1988) have been transplanted into mature spinal cord after injury. Embryonic spinal cord transplants fill the lesion cavity and become apposed to the host cord but promote only a limited growth of the host fibers into the transplant (Reier et al. 1986; Houle and Reier 1988). The locus coeruleus and raphe transplants into denervated cord survive and produce a network of nerve terminals in the host gray matter surrounding the transplanted cells (Nygren et al. 1977; Foster et al. 1985; Privat et al. 1986). Despite modest axonal outgrowth, some recovery of function, indicated by changes in sexual reflexes and hindlimb flexion reflexes, has been associated with the fetal brainstem transplants (Privat et al. 1988; Buchanan and Nornes 1986).
26 Lesions in the developing central nervous system often have effects dramatically different f r o m the same lesion in the adult. This is referred to as the infant lesion effect (Bregman and Goldberger 1983 a-c), and reflects the greater anatomical plasticity and greater sparing of function following a CNS lesion m a d e at birth, c o m p a r e d to the same lesion m a d e in the adult. The response of the immature nervous system is not uniform, however, and in m a n y cases, effects of injury to the developing nervous system are m o r e severe than those in the adult (Bregman and Goldberger 1982, 1983b, c). F o r example, axotomized immature neurons are m o r e likely to die than axotomized adult neurons (Bregman and Goldberger 1982, 1983 a-c). We have shown that embryonic spinal cord transplants placed into the d a m a g e d spinal cord of a neonate enhance the anatomical plasticity o f the immature CNS. Thus, transplants prevent the retrograde cell death o f immature axotomized rubrospinal neurons (Bregman and Reier 1986) and allow the growth of descending axons b o t h into and through the site of a spinal cord injury (Bregman 1987a, b; Bregman et al. 1989). Some o f these axons reach normal targets within the host spinal cord caudal to the lesion and transplant (Bregman 1987a, b; Bregman et al. 1989). We now want to determine if these anatomical changes induced by the fetal spinal cord transplants are associated with sparing and recovery of l o c o m o t o r function. The contribution of transplants to functional recovery can be evaluated in the spinal cord. The spinal cord has the intrinsic capacity to mediate basic l o c o m o t o r patterns (spinal pattern generator for locomotion, S P G L ) which are controlled and modified by b o t h descending and segmental afferent input (Grillner 1975). The serotonergic, rubrospinal, corticospinal and dorsal root systems contribute to the descending and ascending influences on the generation o f locomotion (Donatelle 1977; Nygren and Olson 1977; Westlund et al. 1983; Martin and Ghez 1988). Each of these systems exhibits anatomical plasticity elicited by the presence of a transplant (Bregman and Reier 1986; Bregman 1987a, b; Bregman et al. 1989 and unpublished results). We suggest that through such neuronal projections the transplants influence the generation of locomotion and thus permit functional recovery. Since there is considerable recovery of function after spinal cord injury, mediated by intact descending and dorsal root afferent systems (Bregman and Goldberger 1982, 1983b; Goldberger 1988a, b) as well as by intrinsic spinal systems (Barbeau and Rossignol 1987), it is essential to use quantitative methods for analysis of reflex and l o c o m o t o r function to distinguish intrinsic recovery mechanisms f r o m transplant mediated alterations in m o t o r function. In the present study we have developed behavioral tests to assess several specific m o t o r functions quantitatively in the rat. This quantitative a p p r o a c h has allowed us to distinguish between the contribution of recovery of function due to reorganization o f the host CNS and recovery o f function due to the presence of a transplant. In order to test the hypothesis that transplants of fetal spinal cord tissue placed into the site of a neonatal spinal cord lesion p r o m o t e sparing and recovery of lo-
c o m o t o r function, we assessed the l o c o m o t o r function of three groups of animals: 1) normal control animals, 2) animals that received a spinal cord "over-hemisection" as neonates, and 3) animals with spinal cord transplants within the site of the neonatal hemisection. As adults, animals were evaluated using quantitative and qualitative tests of l o c o m o t o r function to determine the accuracy and pattern of locomotion o f the animals. Our results indicate that the presence of a transplant enhances recovery or sparing of some l o c o m o t o r functions, but not others, after spinal cord injury at birth.
Methods
Spinal cord lesions and transplantation procedures A total of 47 animals were used in this study. Three groups of animals were used: 1) normal control (CON) animals, 2) animals with a spinal cord hemisection only (HX), and 3) animals with a spinal cord transplant at the site of the hemisection (HX + TP). Sprague--Dawley rat pups (48 h postnatal, N= 36) were anesthetized by hypothermia. Under a dissecting microscope, the spinal cord was exposed by partial laminectomy at T5-T6. The dura was opened and the dorsolateral sulci, dorsal rootlets, and dorsal columns identified bilaterally. Iridectomy scissors were used to sever both dorsal columns plus the right lateral and ventral funiculus (over-hemisection). A lesion of this type is very reproducible and results in good apposition of the transplant at the rostral and caudal ends of the injury. In contrast to a transection the survival of our animals with hemisections is excellent (nearly 100 %). Lesions of this type also allow us to compare one limb to another. In 18 of the animals a transplant of fetal spinal cord tissue was placed into the lesion site and covered with an artificial dural material (durafilm, Codman-Shurtleff, Inc.). Transplantation techniques were identical to those described previously (Bregman and Reier 1986; Reier et al. 1986; Bregman 1987a, b). Briefly, pregnant Sprague-Dawley rats, 14 days gestation, were anesthetized with chloral hydrate and laparatomized. Fetuses were removed individually as required and placed in sterile culture medium (Dulbecco's Modified Eagle's Medium, Gibco). The spinal cord was removed from the vertebral canal and stripped of meninges and dorsal root ganglia. The thoracic spinal cord was blocked into transverse sections 2-3 mm in length. The donor tissue was floated onto sterile forceps and placed into the lesion site. In the lesion control animals gelfoam was placed into the site of the hemisection and covered with durafilm. The muscle and skin were sutured in layers and the wound covered with celloidin. Animals were warmed, received prophylactic injections of antibiotics (bicillin, Wyeth Laboratories) and returned to their mothers. Unoperated control littermates (N = 11) were used and housed with the lesioned animals (10 pups per litter). At 4 weeks of age the animals were weaned, males and females separated, and housed 2 or 3 per cage in the animal care facilities under a 10/14 hr light/dark cycle.
Behavioral training At 8-12 weeks of age behavioral testing began. Water was removed 48 h before training was to begin, Animals were trained to cross the runways or walk on the treadmill. During the training period of five days, animals received water only upon crossing the runways or while walking on the treadmill (12 ml/day). Once the rats were trained, they were allowed free access to additional water at the end of the testing session for 15 min daily. Animal weights were monitored and all animals received supplemental water if weight loss in any animal was noted. After training for 5 days all animals were able to walk at the speed (6.6 m/rain) of the treadmill.
27
Fig. 1. A treadmill was used to produce continuous locomotion by the animals. Animals were trained for five days to walk at the speed of the treadmill. The animal's feet were inked to obtain footprints while on the treadmill
Locomotor function tests
of animals reported for the different behavioral analyses in the results differs.
All the animals were tested by observers unaware of the animals' surgical treatment. During the testing sessions qualitative observations were made on the animals. In an early group of animals in this study, only some of the behavioral tests were examined. Since footprints were not consistently legible in this group, the number
m m m m
RS~ 9
9
9 9
m 9
9 9
m 9
9
m
LsLm 9
Footprint analysis. The treadmill used in our locomotor analysis is illustrated in Fig. 1. Footprint analysis was modified from that described by de Medinacelli (de Medinacelli et al 1982). The animals fore- and hindpaws were inked and the footprints were recorded on paper covering the surface of the treadmill. Measurements of different parameters of the locomotor pattern were calculated from the hindlimb footprints. A reconstruction of the footprints is illustrated in Fig 2. Measurements were made only on serial prints (a minimum of 3 consecutive prints) and only when the animal was walking continuously. Mean values ( N = 6 - 1 0 ) for each animal were calculated from prints obtained on two days of testing. The base of support was determined by measuring the distance between the central pads of the hindfeet (DBF in Fig. 2). The angle formed by the intersection of lines from the left and right prints was used as an indication of the limb rotation (R in Fig. 2). Stride length (RSL, LSL in Fig. 2) was measured between two consecutive prints on each side.
Grid. The animals were trained to cross a grid runway 180 cm long with 50 mm x 50 mm holes (Fig. 3). The animals crossed in both
9 9
m m m
Fig. 2. An illustration of hindlimb footprints reconstructed from a normal animal's prints indicating the measurements used in this study. Base of support was measured as the distance between the left and right prints (DBF). Limb rotation (R) is the angle produced by the left and right prints. Stride length (RSL, LSL) is the distance the right or left limb traverses with each step cycle
Fig. 3. Accuracy of locomotion was assessed on a grid runway. Animals were trained to cross a grid of 180 cm length with 50 m i n x 50 mm holes. The time to cross the runway and the number of errors in foot placement (footfalls below the grid) of the ipsilateral (right) and contralateral (left) hindlimb were counted. The arrow indicates an error with the right hindlimb
28
Fig. 4. Photomicrograph of a cross-section through a typical transplant within the host spinal cord. The transplant (TP) is in good apposition with the host spinal cord; the border between host and
directions (4 crossings/day) for the water reward at the end of the runway. The animals were trained for 5 days then videotaped for 5 days. The time to cross the runway and the number of errors made by either hindlimb was determined from the slow motion analysis of the videotapes. The errors measured were foot falls when the animal's foot was misplaced so that it fell into the grid hole, rather than being placed onto the rungs. The mean time to cross and number of errors were determined for each animal from the 20 runs.
Statistical analysis. Single factor ANOVAs were carried out using the measurements for each parameter calculated from the footprints and the data obtained crossing the grid. Between-group differences were examined with 2-sample T-tests. The Dynastat Professional (Dynamic Microsystems, Incorporated, Washington, D.C.) statistical software package was used for the statistical calculations. Lesion analysis At the completion of the behavioral testing the animals were sacrificed and perfused intracardially with 0.9% saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. Serial transverse 18 gm cryostat sections were stained with cresyl violet. A microprojector and a Zeiss microscope were used to reconstruct the lesion site and to determine the presence of a transplant. The histological characteristics of the embryonic tissue enabled us to
transplant is indicated by arrows. The host ventral horn (VH), lateral funiculus (LF) and ventral funiculus (VF) are indicated. Cresyl violet stain, scale bar is 500 gm
distinguish host tissue from transplant and allowed the areas of apposition between host and transplant to be determined.
Results
Spinal cord lesions and transplants T h e a p p e a r a n c e o f the s p i n a l c o r d t r a n s p l a n t s (Fig. 4) a n d the p e r c e n t survival were similar to t h a t d e s c r i b e d in o u r p r e v i o u s studies ( B r e g m a n 1987a, b ; B r e g m a n et al. 1989). I n m o s t cases there were extensive a r e a s o f a p p o s i t i o n b e t w e e n the t r a n s p l a n t s a n d the h o s t spinal c o r d . A n a l y s e s o f the b e h a v i o r a l tests were c a r r i e d o u t o n a n i m a l s w h o s e lesions were w i t h i n the m a x i m u m a n d m i n i m u m r a n g e o f d a m a g e (Fig. 5). T h e m a x i m u m lesion i n c l u d e d was the entire right side o f the c o r d p l u s on the left the d o r s a l c o l u m n , p o r t i o n s o f d o r s o l a t e r a l funiculus, i n t e r v e n i n g g r a y m a t t e r a n d a s m a l l p o r t i o n o f the vent r a l funiculus. T h e m i n i m u m a m o u n t o f d a m a g e i n c l u d e d was loss o f the d o r s a l c o l u m n o n the left a n d the entire r i g h t side except a v e r y small p o r t i o n o f the v e n t r a l
29
Maximum
Minimum
HX
a
Maximum
b
Minimum
HX+TP
Fig. 5a, b. Lesion reconstruction from serial sections are plotted onto standard diagrams of the cord for comparison. Maximum and minimum transverse extent of the lesion included in the study are illustrated in a. Only animals with transplants apposed to the host spinal cord were included within the study. Representative lesions plus transplants are drawn in b funiculus. We used the same lesion criteria for both the lesion group and the lesion plus transplant group; thus the transverse extent of damage in both groups was similar. Animals in the transplant group had the additional requirement for inclusion in these studies, that the transplant had to be present and apposed to the host spinal cord. Reconstruction of two representative transplants are shown in Fig. 5b. O f the 36 operated animals, 25 (13 H X and 12 H X + T P ) fulfilled all o f our lesion criteria for inclusion in the behavioral analysis.
Qualitative observations of locomotion Both hemisection only and hemisection plus transplant animals were capable of considerable locomotor function. All lesioned animals were able to walk easily on wide runways (requiring little accuracy of foot placement in order to cross) using all four limbs. Neonatal spinal cord lesions led to an apparent increase in the base of support and in the amount o f rotation between the hindlimbs (Fig. 6). In the animals with a neonatal lesion and transplant, both the amount o f rotation and the base o f support were similar to that observed in control animals. Control animals had little difficulty crossing runways requiring accurate foot placement. They crossed quickly and rarely made errors in foot placement. Both lesioned groups made many errors in foot placement while crossing the grid runway. The animals with transplants, however, recovered rapidly from the error, while the lesion
Fig. 6a-e. Photographs taken from videotapes of the rats walking on the treadmill, a Control (CON); b lesion plus transplant (HX + TP); e lesion only (HX). Neonatal hemisection results in permanent deficits in locomotion (see text). The presence of a transplant at the lesion site results in a pattern of locomotion that is similar to that in control animals
only group repeatedly flexed and extended the hindlimb in the grid hole, prior to regaining support. With all qualitative observations of locomotion, the animals which no longer had transplants were indistinguishable from lesion only animals.
Footprint analysis A neonatal hemisection resulted in a significant increase in the animals' base of support (HX = 50 ram) when
30
BASE OF SUPPORT (DBF)
,11 I
I i1#
60. I ~
* W i,
con HX HX+TP
~
50.
Z LLI W
40 bJ if) W (.3
z
\
30
CO
20
CON
HX
HX
HX+TP
Fig. 7. Base of support (see diagram Fig. 2). The mean ( + SEM) distance between the feet for each group. This distance was significantly (indicated by an asterisk) greater in the hemisection group (HX, N = 7, p
