Effects of Implantation Site of Stem Cell Grafts on Behavioral Recovery From Stroke Damage

Animals

Forty-two Sprague-Dawley rats (Charles River, UK), (weight, 270 to 330 g at surgery) were maintained on a 12-hour light/dark schedule (lights on at 8 am), with food available ad libitum until recovery from surgery and then restricted (10 g/100 g) to standardize weight. Procedures complied with the UK Animals (Scientific) Procedures Act (1986) and the Ethical Review Process of the Institute of Psychiatry.

MCA Occlusion and Transplant Surgery

The right MCA was occluded for 60 minutes by insertion of a round-tipped polypropylene (Prolene) filament via the common carotid artery to the ostium of the MCA in the circle of Willis. Halothane (in 70% NO₂/30% O₂) anesthesia was used for insertion and removal of the filament, with temperature held at 37±1°C by rectal probe and heating pad. During occlusion rats were assessed for forelimb flexion and contralateral circling to confirm ischemia. During the recovery period, rats were tested for neurological deficits and lesion asymmetry baselines.⁹

Rats were assigned to groups showing equivalent forelimb flexion⁹ to receive cells grafted into the lesioned (ipsilateral; n=8) or intact (contralateral; n=9) striatum or the right ventricle (intraventricular graft; n=7) 2 to 3 weeks after MCA occlusion (MCAO). Sham-occluded (n=11) and stroke (n=7) animals served as controls.

MHP36 cells from frozen stock (passage 42) were prelabeled by incubation for 4 minutes with the fluorescent cell membrane marker PKH26 (Sigma)¹⁰ and suspended (25 000 cells per microliter) with 1 mmol/L N-acetyl-l-cysteine (Sigma) in Hanks’ balanced salt solution (Gibco). Cell viability, determined by trypan blue (Sigma) exclusion, ranged from 83% to 92% before grafting to 73% to 85% after grafting. Rats were anesthetized by midazolam (Hypnovel; 0.03 mL/100 g; Roche) and etorphine hydrochloride and methotrime- prazine (Immobilon; 0.01 mL/100 g; C-Vet Products) to deliver a total of 8 μL (1 μL/min) of cell suspension. For parenchymal grafts, 2 deposits were made at 2 sites either ipsilateral (+) or contralateral (−) to the right MCAO at the following coordinates: (1) anteroposterior=+0.7, lateral=±2, ventral=−5.5/−2.0 and (2) anteroposterior=−0.3, lateral=±3, ventral=−5.5/−2.5. A single deposit of 8 μL was made into the right ventricle at anteroposterior=+0.8, lateral=−1.5, ventral=−4.5. After surgery animals were injected with cyclosporin A (10 mg/kg SC; Sandoz) in cremphor EL (Sigma), and anesthesia was reversed by diprenorphine (Revivon; 0.01 mg/100 g IM; C-Vet Products). All animals received cyclosporin A 3 times per week for 2 weeks.

Behavioral Assessment

The test battery was identical to that used by Veizovic et al.⁵ Time courses for procedures are provided in Table 1.

The bilateral asymmetry test of sensorimotor dysfunction measured the disparity in time taken to remove sticky tape strips (0.8×6 cm) from the affected and unaffected forepaws for 180 seconds.⁹

The rotameter (TSE GmbH) measured motor rotation asymmetry in response to amphetamine (2.5 mg/kg) and apomorphine (1.0 mg/kg) 30 minutes before testing, which stimulate presynaptic and postsynaptic dopamine receptors, respectively, on the intact left side to induce marked rightward circling in stroke animals. Rotation bias was calculated as the total number of anticlockwise turns divided by the total number of clockwise turns, and total activity was calculated by summing turns in both directions.

Learning to find the safe platform in the water maze (4 trials of 60 seconds each per block of testing, 200-cm-diameter pool, 10-cm-diameter platform, water at 24±2°C) provided a measure of spatial navigation deficits (in time taken to find the platform, distance, and swimming speed) as recorded by an image analysis system (HVS Image Ltd).⁹

Histology

Animals were overdosed with pentobarbitone sodium (Animal Care Ltd), flushed transcardially with heparinized saline, and perfused with 4% paraformaldehyde in 0.2 mol/L PBS. Brains were cryoprotected by 30% sucrose and cut by cryostat into 50-μm sections, a thickness chosen to avoid distortion or breakage that might occur with thinner sections of cavitated brain.

Lesion volume was calculated in every 20th section by Simpson’s rule with the use of digital microscope images analyzed by Image Pro Plus (Media Cybernetics) and expressed as a percentage of total brain volume. Transplanted cells were identified by the fluorescent PKH26 label under confocal microscopy (Leica). To determine graft survival semiquantitatively, transplanted cells in the same focal plane were counted (×400 magnification) bilaterally in 3 regions of interest (ROIs): somatosensory cortex, striatum, and septum at the level of implantation (+0.7 mm from bregma) in a rectangular field (0.202×0.252 mm) selected to provide an appreciation of graft survival within areas equivalent relative to implantation tracts in each ROI for each brain (Figure 5A). Since it is difficult to measure thinly scattered grafted cells, the purpose of standardized ROI cell counts was not to estimate the total number of grafted cells but to obtain an accurate count within selected ROIs across groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. The highest concentration of transplanted cells was close to the injection site (A), but a substantial number of cells migrated to more distant regions with the same and the opposite hemisphere irrespective of their injection site. Cell counting was performed within the red squares (on the template MCAO brain) to provide a comparison between implantation sites within these ROIs. Comparison of ROIs indicated that contralateral (Contra) grafts survived best in the contralateral striatum (Str, Stri) and somatosensory cortex (SMC) (B). In all sites, approximately 30% to 35% of grafted cells differentiated into neurons (C). Overall, after contralateral implantation more cells were found in ROIs compared with ipsilateral (Ipsi) and intraventricular (ICV) grafts (D). Absolute number of MHP36 cells differentiating into neurons (E) showed a pattern similar to the number of transplanted cells within each ROI (B).

Differentiation of transplanted cells was determined by colocalizing phenotypic markers for different cell types (Table 2) with PKH26 as a marker for transplanted cells. Primary antibodies were incubated overnight at room temperature before either the fluorescent anti-mouse or anti-rabbit Alexa 488 (Molecular Probes; 1:500) secondary antibody was applied for 1 hour. To determine neuronal differentiation, cells colocalizing with PKH26 were counted after counting PKH26-positive cells within the same ROI.

Statistical Analyses

Results for the water maze and bilateral asymmetry tests were analyzed by repeated-measures ANOVA (SPSS), with groups as the between-subjects factor and trials/days as the within-subjects factor. We used t tests to analyze postlesion bilateral asymmetry deficits. One-way ANOVAs were used for group differences on the rotameter, the water maze probe trial, lesion volume, and differences in cell counts in ROI. The Bonferroni post hoc test was used to compare groups.

Behavioral Assessment

Bilateral Asymmetry Test

After stroke surgery, animals removed the left tape significantly more slowly than the right tape (t=4.162, df=64, P<0.001), whereas there was no difference between removal times for controls. After grafting, the difference between paws remained substantial in stroke controls and in rats with intraventricular grafts but gradually decreased in rats with parenchymal grafts. Over the last 4 weeks, rats with parenchymal grafts no longer showed a significant difference in removal times between paws (Table 3). Stroke controls and rats in the intraventricular graft group removed tape from the unaffected right paw more rapidly than control and parenchymal graft groups, emphasizing the lack of competition between paws in rats with asymmetrical bias that adopt a serial strategy to remove tape from the unaffected before the affected paw (Figure 1).

View this table:

• View inline
• View popup

Statistical Summary of Paw Differences in the Bilateral Asymmetry Test

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Bilateral asymmetry test. Controls (A) approached right and left paws first equally often and took the same time to remove the tape from each paw. Stroke controls (B) and rats with intraventricular grafts removed tape from the right paw, which they always approached first, 40 to 60 seconds faster than from the left paw. Animals with intraventricular grafts (D) appeared to show an improvement of paw asymmetry by 5 and 6 weeks after grafting, but this transient improvement did not result in sustained recovery. In contrast, rats with parenchymal grafts (data from the ipsilateral grafted group shown here) (C) showed reduced bias over time so that there was no difference between paws from 5 and 6 weeks after grafting until the end of testing.

Rotameter

There was no difference between groups in baseline (saline) rotation bias (Figure 2), although a modest overall effect of group (F_4,38=2.722, P<0.05) suggested a tendency to bias in stroke controls. In contrast, injection of amphetamine (F_4,37=5.283, P<0.01) and apomorphine (F_4,38=6.378, P<0.001) induced a significant bias in all stroke groups compared with controls (P<0.01), which was not corrected in any grafted group. There were no differences in total activity between groups under saline (F_4,38=1.199, P=NS), amphetamine (F_4,37=1.888, P=NS), or apomorphine (F_4,38=1.768, P=NS) treatment, indicating that locomotor activity did not significantly differ between groups.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Rotameter. After systemic injection of amphetamine or apomorphine, stroke animals showed a rotational bias (*P<0.05, **P<0.01) that was not present after saline treatment (spontaneous rotation). Unbiased rotation was indicated by a rotation ratio close to 1. The total number of turns did not show any difference between groups, so that all groups responded similarly to the drugs in terms of activity. Ipsi indicates ipsilateral; Contra, contralateral; and ICV, intraventricular.

Water Maze

Controls found the platform significantly faster (F_4,5=23.819, P<0.001) than rats with ipsilateral grafts (P<0.001), rats with contralateral grafts (P<0.001), or stroke controls (P<0.001). In contrast, rats with intraventricular grafts found the platform as rapidly as controls (Figure 3) and were significantly superior to rats with contralateral grafts (P<0.001), rats with ipsilateral grafts (P<0.01), and MCAO controls (P<0.001). Rats with intraparenchymal grafts and stroke controls showed comparable impairments (Figure 3). Similar results were seen for path length (F_4,5=24.267, P<0.001), although rats with intraventricular grafts were marginally impaired (P<0.05) relative to controls while superior to all other stroke groups (P<0.01). Groups differed in speed of swimming (F_4,5=13.504, P<0.001), since controls swam more slowly than all other groups (P<0.001, except for the intraventricular group, in which P<0.05), suggesting that accurate searching reduced speed. Rats with intraventricular grafts swam more slowly than stroke controls (P<0.05) but not grafted striatal groups, indicating an intermediate speed.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Water maze. Mean (±SEM) time to find the platform during training showed that the performance of rats with intraventricular grafts was similar to controls and superior to all other groups. Blocks of testing consisted of 4 trials (2 per day) of 60 seconds each. Abbreviations are as defined in Figure 2.

Histology

Distribution of Transplanted Cells

In all regions of interest, grafted cells differentiated into both neurons and astrocytes (Figure 4). Figure 5A shows the distribution of grafted cells mapped on a template reflecting typical 60-minute MCAO damage. Pooled data from all ROIs (F_2,98=7.818, P<0.001) showed that a greater number of cells survived in ROIs after contralateral than ipsilateral (P<0.001) or intraventricular (P<0.05) implantation; ipsilateral and intraventricular grafts did not differ in cell counts within these sampled regions (Figure 5B). Greatest cell numbers were seen in the contralateral striatum (F_2,18= 14.762, P<0.001), which harbored more cells from contralateral than from ipsilateral (P<0.01) or intraventricular (P<0.05) grafts. In both groups of rats with parenchymal grafts, approximately a third of cells migrated to the other side of the brain, while two thirds remained within the grafted side. Hence, bilateral comparisons of ROI differed in rats with ipsilateral (F_4,34=5.04, P<0.01) and contralateral (F_4,34=4.163, P<0.01) grafts. However, ROI did not differ in rats with intraventricular grafts (F_4,34=1.769, P=NS) because cells spread evenly from the midline infusion site.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Confocal images show the differentiation of transplanted cells (PKH26 in red) into neurons (NeuN in green; A to C) and astrocytes (GFAP in green; D to F) at low and high magnification. Low magnification (A to F) shows the distribution of transplanted cells within the tissue. Higher magnification of inserts (G to L) confirms the colabeling of both markers in the same cell.

Phenotypes of Grafted Cells

Transplanted cells differentiated into site-appropriate phenotypes. In the striatum, transplanted cells differentiated to a large degree into interneurons (Figure 6A) and also into GABAergic output neurons that are DARPP-32 positive (Figure 6B). Grafted cells were also surrounded by fibers innervating the striatum, such as tyrosine hydroxylase–positive dopaminergic fibers. Nesting of PKH26-positive cells within the matrix of input fibers suggests that innervating host neurons formed synaptic connections with grafted cells. The compartmentalized nature of the striatum helped to determine whether grafted cells expressed uncharacteristic phenotypes (as determined by the markers used in this study). No atypical phenotypes were seen, and differentiation of grafted cells always paralleled the host phenotypes within a particular region of the striatum

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Confocal images of neuronal phenotypes of transplanted cells in the striatum. Transplanted cells (PKH26 in red) appropriately differentiated into interneuronal phenotypes expressing choline acetyl transferase (ChAT; A to C), and the Ca²⁺-binding proteins parvalbumin (Parv; D to F), somatostatin (Som; G to I), and calretinin (Cal; J to L). Transplanted cells also differentiated into striatal GABAergic output neurons (DARPP-32; M to O) and were found to be enclosed by markers for substance P (Sub P; V to X), tyrosine hydroxylase (TH; P to R), and enkephalin (Enk; S to U), indicating that transplanted cells differentiated into site-appropriate phenotypes and were integrated into the host neuropil.

Cell Distribution and Number

A large proportion of intraparenchymal grafted cells (30% to 35%) migrated across the midline not only from the intact to the lesioned hemisphere, as found by Veizovic et al,⁵ but also from the lesioned to the intact side, contrary to our previous belief that stem cells only migrate to sites of injury.⁴ This novel finding indicates that denervation and reorganization in the so-called intact hemisphere generate signals that are as attractive to stem cells as ischemic injury. The finding that cell counts within ROIs were significantly greater with contralateral implantation supported our view⁵ that grafting into the intact striatum might foster cell survival. However, ipsilateral grafts were as effective as contralateral grafts in reducing bilateral asymmetry, so that increased cell numbers did not translate into greater functional recovery, possibly because cells in ipsilateral striatum differentiated more readily into neurons. Cell number per se also cannot account for differences in effects of parenchymal versus intraventricular grafts, since numbers in intraventricular grafts were similar to those in ipsilateral grafts. It should be noted, however, that the present study only used semiquantitative measures that could underestimate or overestimate graft survival or differentiation. The extensive migration and the extent of structures affected by MCAO damage render a quantitative or stereological approach technically difficult, and use of 2-dimensinal analyses as a method of providing reasonable “snapshot” representations of cell distribution has been advocated.¹¹

The reduction in bilateral asymmetry possibly depended on sufficient engraftment of the striatum that was provided by parenchymal grafts on either side. This possibility is supported by our findings¹² that apolipoprotein E, a lipid transporter associated with both clearance of dead cells and neuronal reorganization after stroke,¹³ was upregulated in the contralateral striatum of rats with parenchymal but not intraventricular grafts and seen in both host and grafted neurons and astrocytes. These results are consistent with our previous suggestion that correction of motor asymmetry involves interactions with striatal regions undergoing both degeneration and reorganization after stroke, to foster repair of local pathways and to stimulate host reorganization of normally silent contralateral corticospinal links. Conversely, concentration of intraventricular grafts in the septum, an area that governs cholinergic and GABAergic projections to the hippocampus, may have improved communications with this important structure for spatial learning, whose projections are disrupted by MCAO, even though gross damage may not be apparent.

Cell Differentiation and Phenotype

Findings that approximately 30% of grafted cells differentiated into neurons in ROIs after ipsilateral, contralateral, or intraventricular grafting indicate that the overall proportion of neurons is not a correlate of differential graft efficacy. However, attempts to determine whether specific phenotypes of grafted cells relate to their functional effects were also not successful in discriminating between grafts, since similar phenotypes were seen in all groups. Transplanted MHP36 cells differentiated to a large extent into Ca²⁺-binding protein–positive interneurons and GABAergic output neurons, reflecting the general proportion of these cells in the striatum. Double labeling clearly identified grafted cell types at high magnification, with confocal slices and rotation to clarify markers within cells (Figure 6). However, with thinly scattered cells, confocal fluorescence may not be appropriate to quantify cell types at this level of detail throughout ROIs. Thus, although double-labeled cells were identified, accurate comparisons of numbers were not feasible, and it remains to be seen whether implantation site differentially influenced mature phenotype expression.

Lesion Volume

Veizovic et al⁵ reported that contralateral MHP36 grafts significantly reduced infarct volume by 36% at 11 months after grafting. However, the present results showed no evidence for a significant reduction in lesion volume in any of the grafted groups at 3 months after grafting, a time at which lesion volume itself was smaller than at 11 months (18% as opposed to 26% of total brain volume). Taken together, these results suggest that the effects of MHP36 grafts on stroke volume do not involve an early cerebroprotective mechanism to prevent the development of ischemic cell death but might limit the progression of cavitation and secondary damage distant from the infarct in the long run. Veizovic et al⁵ also found that dopamine agonist–induced rotation bias was normalized in rats with contralateral grafts at 38 to 45 weeks after grafting, whereas no reduction in bias was seen at 4 to 7 weeks in the present study. Thus, reduced rotation may be a late effect and, if so, may be related to delayed graft effects on lesion volume. The bias on the bilateral asymmetry test also appeared to include a motor deficit,⁵ whereas the present study only detected a sensorimotor bias. In stroke animals the initial lack of response to the affected paw induces a serial strategy to remove the adhesive tapes, whereas controls have to resolve response competition between paws. Motor impairments indicated by increases in removal times were not apparent in the present studies, but the results clearly show a difference between paws indicative of a sensorimotor bias that was reduced after intraparenchymal grafting but not intraventricular grafting. A time course study is therefore needed to evaluate the contribution of the evolution of stroke damage to effects of grafted cells on lesion volume and behavioral outcome, in interaction with site of transplantation.

Conclusions

This study has shown that within a 3-month period, intraparenchymal stem cell grafts improved sensorimotor dysfunction, whereas intraventricular grafts promoted normal spatial learning in rats after MCAO, indicating that graft placement critically affects the type of functional recovery that occurs. Histological measures such as grafted cell number, distribution, neuronal differentiation, and lesion volumes were not obviously related to differences in graft effects. However, a substantial proportion of grafted cells migrated to the opposite side of the brain, whether this was lesioned or intact, indicating that stem cells are attracted by and interact with regions of both degeneration and reorganization. Differential density of grafted cells in the striatum and septum with intraparenchymal and intraventricular grafts may offer clues to their different functional effects, which should be confirmed by more detailed quantification of possible differences in phenotype. While these results raise many questions, they indicate that several sites of implantation may be optimal for different impairments and that several mechanisms may be involved in stem cell repair of widespread stroke damage.