Motor Deficits Are Produced By Removing Some Cortical Transplants Grafted Into Injured Sensorimotor Cortex of Neonatal Rats

Fetal frontal cortex was transplanted into cavities formed in the right, motor cortex of neonatal rats. As adults, the animals were trained to press two levers in rapid succession with their left forelimb to receive food rewards. Once they had reached an optimal level of performance, the effect of removing their transplants was assessed. Surgical removal of transplants significantly impaired the performance of 2 of 4 subjects. Placing a crossstrain skin graft to induce the immunological rejection of the transplants produced a behavioral deficit in 1 of 2 subjects with complete transplant removal. Skin grafts produced no behavioral effects in four subjects that had surviving transplants. Since the motor deficit produced by transplant removal resembled those observed following the removal of normal motor cortex, we propose that these three transplants functioned within the host brain. Histology Showed that the procedures used to remove cortical grafts did not injure any host brains. Therefore, host brain damage is unlikely to account for the behavioral deterioration that followed transplant removals.


INTRODUCTION
While fetal cortical grafts survive and establish connections with host brains /5-8,11-13,18/, many questions regarding their function remain unanswered. The cognitive deficits and behavioral impairments produced by bifrontal cortical lesions /23/ or basal forebrain lesions /1,36/in mature rats are partially reversed by fetal grafts. Also, though the motor deficits induced by cortical lesions in mature subjects may not be reversed by cortical grafts/10,22,23/, motor deficits produced by frontal cortex injury of neonatal rats are ameliorated by transplanting dissociated fetal cortical cells/27/.
In this study we transplanted fetal cortex into injured fight motor cortex of neonatal rats. Once hosts matured, we examined whether removing the grafts affected motor performance of their left forelimbs. To accomplish this, rats were trained to press two levers in fast succession/16/. The time it took them to press the two levers, called interresponse time (IRT), was quantified.
In normal rats, a lesion in contralateral forelimb motor cortex acutely and chronically increases IRTs/16[. Once the rats in this study reached optimal performance, their transplants were removed surgically or they received cross-strain skin grafts to induce immunological rejection of their transplants /2,9,14,26,35L Transplant removal produced a significant deterioration in the IRTs in three subjects. This suggests that these transplants functioned during the performance of the present motor task in a way analogous to normal motor cortical tissue.

Animals
Thirty six neonatal Long-Evans rats served as hosts. Animals were weaned at 3 weeks of age and given standard Purina rat chow and water ad libitum. Training commenced at 6 8 weeks of age. At this time, subjects were given enough food to maintain them at 80% of their freefeeding body weight. Though on a diet, host rats steadily gained weight and appeared healthy throughout the experiment.

Transplantation surgery
One day old host rats were anesthetized with hypothermia. The scalp was incised, and a 1.5 mm per side square bone flap was made over the right frontal cortex, centered 0.3 mm anterior to bregma and 2 mm lateral to the midline. The underlying cortex, which corresponds in location to the forelimb motor cortex in the adult rat, was removed down to the white matter with a syringe and blunt needle. After bleeding was controlled, the cavity was filled with Gelfoam, the overlying skull was -folded back into place and the skin was sutured.
At 7 days of age, host subjects were reanesthetized with hypothermia. The cavities were re-opened and the Gelfoam was removed. E17-E18 day fetuses were removed individually from anesthetized, pregnant mothers. The fetal brain was exposed, and a slab of frontal cortex approximately 2 mm 2 was removed with a scalpel and placed into the cavities of the host animals /17,18,28,29,31 Four of the 12 subjects were re-anesthetized and Magnetic Resonance Imaging (MRI) scans of their heads were obtained. MRIs were performed on a General Electric CSI imaging spectrometer operating at 2.0 Tesla (85.6 Mhz proton resonance frequency). A standard spinecho 2DFT spin-warp proton imaging sequence yielded single slices from 2 or 4 phase-cycled image excitations. Slice selection in each image was 2 mm with 128 or 256 phase-encoding eyele x 512 point complex readout raw data matrices.
For T1 weighted images, the repetition tim (TR) was 400 msee and the echo time (TE) was 16 msee. For T2 weighted images the TR was 2000 msee and the TE was 60 msee. During imaging rats were positioned supine in an imaging coil of the low-pass birdcage resonator design /19/ with an inner diameter of 5 em.
Because of limited access to the MRI scanneri scans were limited to subjects that were to receive skin grafts to produce immunological rejection of transplants.

Apparatus
Rats were trained in modified rodent operant chambers equipped with an automatic feeder (Lafayette Instruments, Lafayette, IN). To the left of the feeder there were two levers with smooth steel balls, 8 mm in diameter, attached to their ends/16/. A metal partition between the levers prevented rats from pressing both leven at once. Operant chambers were interfaced to a Compaq computer via a LVB interface (Med Associates, Fairfield, VT). OPN operant software was used to monitor response times and number of lever presses, to deliver food pellet reinforcements, and to store data/16/. Training Rats were trained to use their left forelimb to press the fight lever and then the left lever using successive approximation operant techniques /16/. At first, rats had u_!imited time to perform this task and receive a reward. Later, responses were rewarded only if the time between lever presses, called the interresponse time (IRT), was less than a fixed interval. These time intervals were slowly decreased to force the rats to perform the task as quickly as possible. When a rat's daily average IRT was 2 see or less, unwanted behaviors, such as the use of the right paw or snout, were shaped out by withholding reinforcements or by administering low amperage shocks (Master Shocker, Lafayette Inst.). Rats were deemed trained once, in a typical lever session, their IRTs were _< 600 msec and received between 50 and 80 rewards.

Transplant removal
Rats were judged to have reached an optimal level of performance when two criteria were met. First, for a month's period a subject's daily IRTs were within 20% of its fastest score. Second, during this period the slope of the regression line of its daily IRTs was not significantly different from zero. These criteria were reached when rats were between 6 and 8 months old.
Once optimal performance was attained, sham operations were performed in all subjects.
All subjects were anesthetized with ketamine (80 mg/kg) and xylazine (20 mg/kg) i.p. Subjects that were to have their transplants removed surgically had their scalp incised, their skulls scraped, and their skin sutured. Subjects that were to receive skin grafts were given a 20 mm long skin incision on their mid-back and sutured. The day after sham surgery, subjects were returned to their usual testing schedule. After a delay of at least one week, transplant removal was attempted using two methods. Subjects #5, 7, and 13 had their grafts removed surgically. They were re-anesthetized and placed in a stereotaxic frame. Their scalps were incised and the skull overlying the transplant was thinned with a drill and opened VOLUME 2, NO. [3][4]1991 using a scalpel blade. After the transplants' edges were identified, an incision was made on the transplant side of this border to avoid host cortical injury. The inferior surface of the transplant was cut and the transplant removed with a spatula. The excised tissue was frozen in 2-methyl butane at-20"C. Folio.wing control of bleeding, the bone flap was dosed and the incision sutured. Testing sessions were resumed the next day.
Subjects #19, 22, 23, 27, 30, and 33, received cross strain skin grafts in an attempt to produce immunological rejection of their transplant.s /2,9,14,35L Hosts were re-anesthetized, and a 200 mm 2 (1 x 2 cm) portion of skin was removed next to the sham incision site. An analogous piece of skin from donor adult, anesthetized Sprague-Dawley rats (250 gin) was sutured in this place. Testing resumed the next day. Rat #31, known not to have a transplant following MRI imaging (Fig. 1C,D), also received a skin graft to serve as a control for the nonspecifie effects of skin graft rejection.
An MRI scan performed in rat #19 six weeks after receiving a skin graft revealed the survival of a large cortical transplant (Fig. 1A,B). This subject was then re-anesthetized and its transplant removed surgically as described above.

Histology
At the conclusion of behavioral testing all host rats were deeply anesthetized and perfused through the heart with 4% paraformaldehyde in 0.1M phosphate buffer (pH 7.4). The brains were removed, placed in 10% sucrose phosphate buffer overnight, and 20/xm sections cut on a cryostat at -20* C. Excised transplants were also cut in a-cryostat in 20 tzm sections. All sections were picked up onto cover slips and air dried. Alternate sections were stained for cresyl echt violet (Nissl) and hematoxylin and eosin (H&E).

Data analysis
Analysis of variance with repeated measurements tests were performed to assess the significance of differences between daily IRTs of the baseline, sham lesion, or transplant removal conditions. Separate tests were done for animals whose transplants were removed surgically or received skin grafts. Also, the results 224 R. SANDOR, M. F. GONZALEZ, M. MOSELEY & F.R. SHARP of individual subjects were analyzed using Kruskal-Wallis analyses of variance by ranks, and Kruskal-Wallis tests for multiple comparisons between treatments.

RESULTS
Rats vaned in their ability to learn the bar pressing task. It took them 8 to 12 weeks to learn to press alternate bars, and another 8 to 16 weeks to reach optimal performance. Twelve large transplants survived as documented by direct visual observation, MRI imaging (Fig. 1), and histology. Three rats with surviving transplants failed to reach the performance criteria and were eliminated from the study.
Consequently, 9 rats, plus an additional rat (#31), known not to have a graft, remained in the study.
Magnetic Resonance Imaging (MRI) scans of rat brains with (A,B) and without (C,D) transplants. MRI coronal (A,C) and saggital (B,D) scans showed the size and location of surviving grafts or cavities. (A and B) MRI scan of subject #19 showing the presence of a transplant (arrows) 7 weeks after a skin graft was performed to promote its immunological rejection. This transplant was histologically verified following surgical removal (Fig. 5C,D). (C and D) MRI scans of subject #33 showing a region of CSF density (white arrows) representing a cavity in the brain without evidence of a surviving transplant 8 weeks following a skin graft.  Therefore, none of the subjects showed evidence of behavioral deterioration during the first month following skin grafting ( Fig. 2; Table 1).
The basis for tissue identification from MRI images is believed to be due to inherent differences in the spin-lattice (T1) and spin-spin (T2) relaxation times of gray matter, white matter, and CSF /19/. The images obtained dearly depicted surviving grafts or cavities (Fig.   1), and were used to follow transplants' condition following skin grafts. Five or six weeks following skin grafts, cortical transplants were dearly visible within some subjects (Rat #19, Fig.  1A, B), but had virtually disappeared in others (Rat #30, Fig. 1C,D). Three subjects (#23, 27, 30) were sacrificed 5-6 weeks following skin grafts. Their motor performance during the experiment did not correlate with either their MRI images, or the histology of their brains. Two of the remaining subjects, rats #22, which two weeks before the skin graft. All subjects had large transplants at 6 weeks of age, except #31 (plotted using large squares) which did not have a transplant. Several subjects exhibited reductions in IRT values following skin grafts, but only #33 (plotted using large circles) exhibited a deterioration in performance that started 4 weeks after the skin graft.
had exhibited very steady motor performance, and #33, which gave signs of motor performance deterioration, continued to be tested. The remaining rat (#19) had a large cortical graft revealed by a T1 weighted MRI scan 6 weeks after receiving a skin graft (Fig. 1A,B). This subject was anesthetized, and its transplant surgically removed. This result is desen"oed in a section below.
The IRTs of subject #22 remained unchanged during the first and second months following skin grafting ( Fig. 2; Table 1). However, the IRTs of subject #33 (Fig. 2; Table  1) signifieamly increased starting on week 4 following the skin graft. This impairment of motor behavior persisted during the following eight weeks of testing ( Fig. 2; Table 1). An analysis of variance on the combined data of all six subjects failed to show that skin grafts produced significant changes in performance.
However, nonparametrie test showed that the improvements in performance of rats #19 and 30 of the first month following skin grafts, and the increases in IRTs seen in rat #33 during the second and third months following skin grafting were statistically significant (Table 1).
Rat #31, the control rat without a surviving graft, showed a small but significant decrease in IRTs following skin grafting ( Fig. 3; Table 1), suggesting that a skin graft by itself does not adversely affect forelimb motor function.

Surgical lesions
The IRTs of 2 of 3 rats that had their transplants removed surgically (#5 and 13) did not increase following surgery. Rat #5's performance remained stable ( Fig. 3 while rat #13 exhibited a small but statistically significant reduction in its IRTs following surgical lesion (Fig. 3; Table 1) which resembled the reductions seen in some subjects following skin grafts ( Fig. 3; Table 1).
Conversely, rat #7 had a marked deterioration in motor performance, manifested by an increase in IRTs, during the first two weeks that followed surgical removal of the transplant ( Fig.   3; Table 1). The performance of this subject oscillated over the next six weeks but its overall performance was worse than pre-operative levels ( Fig. 3; Table 1). Rat #19 had its transplant surgically removed six weeks following skin grafting when a MRI scan showed what appeared to be an intact transplant (Fig. 1A).
Following the surgical removal of its transplant, subject #19 had a marked increase in its IRTs ( Fig. 4) similar to that of subject #7. This deterioration in motor performance persisted for over a month (Fig. 4; Table 1).  (Figs. 3,4). This contrasts to subject #33, whose performance was normal immediately following skin grafting and deteriorated many weeks later, presumably following the immunological rejection and removal of the transplant (Fig. 2).

Weeks
Plot of motor performance of rat #19 during baseline condition and following a sham skin lesion, a skin graft, and surgical removal ol its transplant. Since a MRI scan (Fig. 1A) showed that the skin graft did not produce immunological rejection of the transplant and since the skin graft did not produce behavioral effects for 6 weeks, the transplant was removed surgically. Note the immediate increase in IRTs following this procedure. Error bars indicate S.E.M. values.

229
Histological analysis Histology showed that 2 of the 6 animals that received skin grafts, subjects #23 (Fig. 5A) and #33 (Fig. 1C,D) had their transplants entirely removed. Subjects #19 (Fig. 5C,D), #27, and #30 (not shown), had large transplants, and #22 (not shown) had remnants of transplanted tissue still present in their brains. It is noteworthy that subject #33, the only subject that showed behavioral deterioration following skin grafting, had its transplant successfully removed.
Residual transplants following attempted immunological rejections contained cellular infiltrates at the host-transplant margins (Fig.   5B). H&E stained sections showed that these cellular infiltrates were small blue staining lymphocytes (Fig. 5D, solid arrow). These cells were confined entirely to the transplants (Fig.   5B). In one subject (#27) the lymphocytes infiltrated the entire transplant (Fig. 5B), and in another (#19)the lymphocytes clustered around vessels within the transplant (Fig. 5C, solid arrow, D). Some transplants also had large, multinucleated cells that appeared to be maerophages (not shown).
All four rats that had their transplants removed surgically exhibited residues of transplanted tissue at the bottoms and sides of their brain cavities. Neurons and glia could be discerned within all residual transplants, but no dear histological differences were found between the host brains or transplants of subjects that exhibited behavioral deterioration compared to those that did not.
Histology of the brains of subjects that had transplants removed surgically demonstrated that normal cortex was not removed with the transplants and that host cortex was not injured in any subject. In addition, no evidence of cellular infiltration or other type of histological injury was found in the host brains of subjects in which transplant removals were attempted using immunological rejection.

DISCUSSION
The removal of transplants in 3 subjects led to significant deterioration in motor performance. We have shown/16/ that removal of forelimb motor/sensory cortex produces a similar effect in normal rats. Therefore we suggest that these VOLUME 2, NO. 3-4, 1991 transplants functioned in a way analogous to normal cortex.
The explanation why the removal of these 3 transplants produced motor deficits while removal of transplants in the other subjects did not is uncertain. Transplant function could depend on the formation of neural interconnections between transplants and host brain. Pyramidal neurons in fetal cortex transplants project to newborn host thalamus, cortex, spinal cord /5,8,12,13,20,34/, and fibers from host thalamic and neocortical neurons synapse inside transplants/5,8,11,17,18,20,24/. It could be hypothesized that the neural connections between these 3 transplants and the host brains were more profuse. These connections, in turn, were either functional, or exerted atrophic effect on other structures of the host brain that participate in the lever pressing behavior.
The failure to produce behavioral deficits by surgically removing 2 grafts could be due to several factors. The residual transplant tissue that remained in the host brain cavities could explain this result. Alternatively, these transplants may have had very few interconneetions with host brain, and their removal did not produce changes in the function of host brain. We have previously observed that degrees of transplant/host connectivity/18/and the number of surviving neurons specifically stained for NADPH-d/17/, cholinesterase /24/ and neuropeptides /32/ can vary from animal to animal. Furthermore, some transplants exhibit decreased glucose metabolic rate /29/ and lower oxidative enzyme activity /28/ suggesting that their connections and cellular composition are abnormal. The fact that the removal of these transplants from the mature animals did not produce behavioral deficits, however, does not imply that they did not function. Other investigators have shown that while cortical grafts may have a beneficial effect on the performance of cognitive tasks, removing these grafts, once learning has occurred, does not produce behavioral deterioration/21L While immunological transplant removal was attempted in 6 subjects, only 2 grafts were successfully removed. It is noteworthy that behavioral deterioration occurred in one of these two subjects (Rat #33, Fig. 2  (C and D) Sections of the transplant of subject #19 which was surgically removed after a MRI scan showed that a skin graft did not induce its rejection.
Nissl staining (C) shows cellular infiltrates around blood vessels (solid arrow) which appear to be lymphocytes using H&E staining (D). deterioration observed in this subject could be due to the delay of the actual removal of the graft in situ. No evidence of behavioral deterioration occurred in five subjects that had surviving transplants following attempted removal using immunological rejection. If any of these grafts were actually functioning, the presence of inflammatory cells within the grafts apparently did not affect the motor function assessed in this study.
The reasons why we failed to immunologically remove most of the transplants are uncertain. Freed/14/used similar methods to induce the immunological rejection of fetal substantia nigra grafts and observed behavioral deterioration in his subjects. Carder et al. /2/ recently used immunological rejection between species (rats and mice) to accomplish high rejection rates of transplanted dissociated dopaminergic neurons. It is possible that the success in the last case was related to a better immune attack taking place against groups of individual cells. Recent results indicate that transplants between species are more likely to be rejected, whereas transplants between rat strains are less likely to be rejected unless there are major immunological differences between the strains/9/.
In order to conclude that 3 subjects in this study had transplants that functioned by neural interactions with their host brains, it is necessary to establish that measuring IRTs is a valid way to assess cortical transplants' function. Even though this behavioral test utilizes quantified data unaffected by observer bias, it could not be argued that it is specific to motor cortex injury. Lesions of sensory cortex, substantia nigra, cerebellum and other structures are likely to increase IRTs as well. However, our previous work has shown that increases of IRTs are a sensitive index of injury confined to motor/ sensory cortex in adult rats/16/. We suggest that the use of IRTs in this study was valid because (a) newborn rats sustained injury only to motor/sensory cortex, and (b) we took precautions to ensure that the removal of transplants never affected the host brain. Histological evidence showed that surgical removal of transplants never injured the surrounding host brain. In addition, lymphocytes or macrophages were never found in the host brains of subjects that received skin grafts, but were confined to the transplants. Therefore, the behavioral deterioration observed in the three subjects could not be ascribed to host brain injury.
Neonatal rather than mature hosts were used in this study for several reasons. First, many pyramidal neurons in fetal cortical transplants project to neonatal host cortex, thalamus, ports, and spinal cord/6,7,12,13,20,34/in comparison to the very few pyramidal neurons in fetal cortical transplants that project to adult host thalamus /18/ and by inference to other subcortical structures. Secondly, newborn motor/sensory cortex is more "plastic" than juvenile or adult cortex. Newborn cortical injury results in greater bilateral subeortieal connections than adult cortical injury /3,4,15,25,30/. This greater plasticity might promote the formation of host-transplant interconnections in the newborn host. Thirdly, there is greater gliosis in transplants placed in juvenile compared to newborn hosts, suggesting that normal projections of cells are more likely to occur in transplants in newborn hosts /28/. Fourthly, fetal cortical transplants ameliorate the thalamic atrophy which occurs following frontal cortical lesions in the newborn host/31L Finally, while grafts of cortical cells in injured frontal cortex of neonatal rats reduce some motor deficits /27/, cortical grafts in adult brains have produced deleterious/22/or non-beneficial/10[ effects.