Currently, there are no approved treatments for chronic ischemic stroke patients, having been woefully understudied. Until now, stem cells have been thought to hold potential for improving recovery after stroke by modifying neuroinflammation during the acute phase of ischemia when most cells are lost in the infarct core, or during the subacute phase when more plasticity is observed in the peri-infarct region. Quite remarkably, in this study, however, we found that autologous MSCs transplanted directly into the brain of rats with a chronic infarct steadily enhanced recovery of function over the subsequent months, demonstrating effects long after typical windows of plasticity in rats or stroke patients.
Many studies in rodents and non-human primates have previously shown the benefits of MSC treatment in improving functional recovery in acute ischemic stroke models [14–31, 35, 38, 40, 41, 53–55]. With the exception of a few studies that used autoMSC [35, 40], most investigations have employed alloMSCs delivered systemically [41, 53–55] or via direct intracerebral injection [55–63]. Thus far, very few studies have examined MSC treatment during the chronic phase of stroke, and all of these have used IV delivered alloMSCs [32–34].
Our current study investigated the use of autologous MSCs injected intracerebrally in a chronic stroke rat model. We showed that at all tested cell doses (1x106, 2.5x106, 5x106 autoMSCs), rats gained significant functional recovery measured by behavioral testing scores. Moreover, we found that recovery in sensorimotor function began within one week of autoMSC administration and continued to improve significantly over the next 60 days. The absence of dose dependency seen in these studies has been well documented previously [61] and suggests that there is a minimally effective dose which is exceeded even in our lowest tested cell concentration (1x106 autoMSCs). The striking recovery in behavior following intracerebral autoMSC transplantation indicates the positive and long-lasting effects of direct injection of stem cells into the peri-infarct region, even in a chronic stroke after the window of brain plasticity is presumed to have closed.
There remains inconclusive evidence regarding the effect that MSC administration has on infarct size [35, 56, 61, 64]. In investigating changes to the volume of the stroke after autoMSC administration, it must be noted that MR imaging revealed a much larger infarct on the day following MCAO due to cerebral edema. This acute swelling of the brain subsided by the day of transplantation at which time stroke volume had stabilized in control rats. Nonetheless, in the two lower dose autoMSC treatment groups we found a small but significant reduction in infarct volume when compared with control groups. The higher dose group (MCAO + 5x106 autoMSCs) trended down but did not reach significance, possibly due to the small sample size of the experimental groups or variability in the placement and survival of implanted MSCs. The small graft-associated changes in infarct volume were unsurprising given that cell death in the ischemic core had likely equilibrated by the end of the chronic phase (first 28 days) prior to MSC transplantation. Likewise, we found that the width of the corpus callosum was also unchanged following transplantation in all treatment groups when compared to controls. This differs from previous research in acute stroke models showing that single and repetitive MSC treatment increases the thickness of the corpus callosum, suggestive of regrowth of myelinated fibers and synaptic plasticity [51]. Possibly, an older more chronic stroke does not lend itself to this type of brain recovery.
We further used Q-dot-labeled autoMSCs to track the localization of MSCs in the brain over time. Somewhat surprisingly, we found that labeled implanted cells remained at the locations in the peri-infarct region where that had been originally deposited, without migration elsewhere in the brain, even two months later. It is likely that this long-term survival of autoMSCs and the continual availability of their locally secreted products may have critically contributed, either directly or indirectly, to the observed robust long-term recovery in sensorimotor function seen in this chronic stroke model.
Although the use of stem cells raises a concern of potential tumorigenicity due to their innate ability to self-renew, we found no evidence of cell proliferation after implantation of autoMSCs as evidenced by the lack of Ki-67 in and around the transplanted region at two months. Furthermore, we did not observe abnormal tumor-like anatomy on MRI at any timepoint during the study. Interestingly, while there was no indication that MSCs were dividing in the graft, there was also no evidence that cells had differentiated into other brain phenotypes. Thus, NeuN or GFAP staining was not seen in Q-dot-labeled transplanted MSCs, indicating the absence of differentiation or trans-differentiation of MSCs into neurons and glia.
Of possible further importance is the local cellular and inflammatory landscape into which MSCs were transplanted. Indeed, we demonstrated a sustained increase in the degree of reactivity in astrocytes and in the number of reactive microglia in all rats with a large MCA stroke which was unaltered by MSC therapy, though potential changes in their molecular composition (ie. cytokines, growth factors, ect.) was not studied here. The literature on reactive gliosis following stem cell transplantation in rats with ischemia is conflicted with some studies showing an increase and others a decrease [37, 38, 41, 52, 59]. Regardless, the persistent glial reactivity seen after stroke may reflect critical changes in the local cellular and molecular milieu needed for autoMSCs to produce enhanced functional recovery in this chronic rat stroke model.
Finally, in a separate important study, we directly compared our results using autoMSCs with alloMSCs in the chronic stroke model. Interestingly, in immunosuppressed rats, alloMSCs survived long term in the brain similar to autoMSCs. However, unlike autografts, allografts did not produce functional recovery greater than the spontaneous recovery recorded in control animals. This is in agreement with the observations of others using alloMSCs [33, 34] despite a report of improved blood-brain-barrier (BBB) function in these rats [32].
In our study, the disparity between the efficacy of autoMSCs versus alloMSCs may be due to small differences in cell handling (ie. autoMSCs but not alloMSCs were expanded in the bioreactor) or cell survival in the graft (ie. alloMSCs may be subject to greater immunorejection than autoMSCs). However, more likely, the key difference stems from the fact that autoMSCs were harvested from the bone marrow of a rat with an active stroke while alloMSCs, as in all allografts studied previously in stroke [16–27, 29–34, 37, 38, 41, 53–57, 61, 64–68], were derived from a healthy (non-MCAO) donor rat. This critical difference, which likely impacts the profile of cytokines and growth factors autoMSCs secrete into their environment, may be crucial to treatment efficacy. Possibly alloMSCs, which are known to be most effective when administered soon after stroke [40, 69], are provided this critical activation in the acute stroke model but not in the chronic stroke model unless combined with other potentially activating influences, like rehabilitation therapy [33, 34]. Consistent with this notion, alloMSCs that had been genetically engineered and transplanted as a “modified stem cell product”, and thus potentially activated, proved partially effective in a preliminary clinical trial of chronic stroke patients [15, 28]. Resolving these important underlying mechanisms will require further exploration into the molecular cross talk between local brain cells and implanted MSCs from various sources. Regardless of the mechanisms, the results of the current study in rats have important clinical implications, suggesting that additional recovery in patients with chronic infarcts and long-term disability may be possible with intracerebral autoMSC therapy.