HUC blood is a promising source of cells for the treatment of stroke, either by using a heterogeneous mix of unprocessed cells or by preselecting a specific cell population from it . Although these two therapeutic avenues have been tested in various preclinical stroke studies with beneficial effects, it is not yet known whether one is more effective than the other. Thus, we compared the cbMNC population, which is a heterogeneous mix of hematopoietic, mesenchymal, and endothelial stem/progenitor cells, along with immature immunological cells , with MSCs derived from cbMNCs and cord matrix.
Although, in principle, transplantation of a xenogenic graft might induce a local immune response in the host, preclinical and clinical reports have raised the possibility of using HUC-derived cells without immunosuppression [12, 17]. Reports have also suggested that immunosuppression is neuroprotective [73, 74]. Therefore, to assess the safety and unbiased efficacy of these cells, we omitted immunosuppression in our study.
Cord-blood-derived cells and other stem cells can be chemotactically attracted to the site of injury . It is thus reasonable to opt for systemic rather than intracerebral delivery to avoid tissue damage at the injection site . Further, selection of intraarterial over intravenous injection was based on study results that showed that more cells reach the MCA territory when transplanted intraarterially [76, 77] and another study that showed approximately 96% of cells transplanted intravenously were trapped in the lungs and did not reach the arterial circulation in a traumatic brain injury model .
As transplanted cell survival in the host brain of immunocompetent rats was one of our outcome measures, we opted to transplant these cells intraarterially.
A dose-response study of cord-blood-derived cells suggested that delivery of 106 cells was sufficient to result in significant functional recovery and that doses of >106 cells resulted in functional as well as histopathologic recovery . We thus opted for a dose of 107 cells to assess the comparative effects of transplanted cbMNCs and cb/cmMSCs. However, we decreased the dose to 5 × 106 for cb/cmMSCs, as intraarterial injection of 107 cmMSCs resulted in severe inflammation of the ipsilateral eye, followed by acute mortality in our study animals. Thus, 5 × 106 cells was the maximum number of cb/cmMSCs we could safely transplant (Table 2). MSCs have a tendency to aggregate into multicellular globules , thus at higher concentrations, they can result in vascular embolization.
Our data suggest that cbMNCs are more effective than cb/cmMSCs in promoting functional recovery and reducing infarct. Although we would have preferred to inject the same doses of cbMNCs and cb/cmMSCs, we think it is unlikely that a twofold difference in transplanted-cell numbers could account for increased beneficial effects of cbMNCs. Furthermore, cbMSCs comprise a very small subset (0.001% to 0.000001%) of cbMNCs . Therefore, it is reasonable to assume that an enriched preselected population of MSCs would have been equally or more effective at a lesser dose, if they were the critical cbMNC subpopulation that mediates recovery. Our flow-cytometry data showed that the cultured fraction of cbMNCs was enriched in CD73+CD166+CD90+CD45-HLA-DR- cells (cbMSCs). These cultured cbMSCs showed an approximately 20-fold increase in expression of MSC markers, resulting in a very effective enrichment of cbMSCs from the cbMNCs. Thus, 5 × 106 transplanted cbMSCs contained 10-fold more cbMSCs than in 10 × 106 cbMNCs. The enhanced recovery in the cbMNC group compared with the enriched cb/cmMSC groups could be due to additional cell fractions in the cbMNCs that might secrete factors important for recovery. Thus, the as-yet-uncharacterized functional properties of cbMNC subpopulations and their possible interactions, rather than total cell numbers, could be responsible for the beneficial effects of transplanted cbMNCs. The enhanced recovery with cbMNCs thus may be partially attributable to the integrative effects of various stem/progenitor cell fractions present in this cell population.
A recent study  compared the effects of cbMNCs with CD34-enriched and CD34-depleted cbMNC fractions in spontaneously hypertensive stroke rats. The study reported superior effects of cbMNCs relative to the CD34+ and CD34- fractions, suggesting the possibility that the combined effect of other cell fractions was necessary for the overall neuroprotective effect.
Although rats in both the cbMNC and cbMSC groups showed improvement in sensory motor functions, only cbMNC-treated rats showed early (within 7 days) recovery in the grid-walk test. This was likely attributable to the significant improvement seen in both contralateral forelimb and hindlimb deficits in this group. Infarct size was also most reduced in the cbMNC group and was significantly smaller than that in the cmMSC group. Although cbMNCs have been shown to produce various growth factors, such as VEGF and BDNF, it is probably their ability to induce enhanced expression of endogenous BDNF that partially resulted in enhanced recovery in this group. BDNF is known to mediate proliferation of existing vascular endothelial cells [83, 84], survival and migration of neuronal cells, along with modulation of their synaptic functions [51, 52], and to exert neuroprotective effects via downregulation of neuronal NOS (nNOS) activity . Intravenous infusion of BDNF has been shown to reduce infarct volume as early as 5 hours after stroke .
Further, the increased mRNA expression of rat-specific BDNF seen in sham versus PBS animals implies a reduction in the BDNF endogenous levels after stroke. Thus, cbMNCs and, to a lesser extent, cb/cmMSCs, possibly restore the stroke-induced depletion of endogenous BDNF. In addition, the trend of increased expression of rat GPx-4 mRNA in cbMNC-treated animals indicates a possible role of MNCs in abating the effects of oxidative stress on GPx-4 levels. The restored GPx-4 might in turn downregulate lipid peroxidation, resulting in decreased neuronal cell death and enhanced overall recovery.
Similar to earlier reports [16, 24], transplanted cbMNCs, cbMSCs, and cmMSCs were seen predominantly in the ischemic hemisphere where homing is likely facilitated by chemokine receptor type-4 (CXCR4)-CXCL12 or CD117-stem cell factor (SCF) interactions. Upregulation of CXCL12/stromal-derived factor 1(SDF-1) has been reported in the ischemic penumbra [87, 88], and its interaction with (CXCR-4)/CD184 (expressed by all three cell populations used in our study) is known to promote migration of cbMNCs . Also, enhanced expression of SCF has been reported in neurons within the injured hemisphere  and might have played a role in directed migration of transplanted cb/cmMSCs, both of which had significant expression of the SCF receptor CD117, as shown in our comparative flow-cytometry experiment.
We observed decreased numbers of CD68- and amoeboidal Iba-1-positive cells in our cell-treated versus PBS animals, indicating a decrease in activated microglia. These resident immune cells are known to acquire a phagocytic phenotype after stroke that disrupts the blood–brain barrier and increases inflammation through release of proinflammatory cytokines, free radicals, and recruitment of leukocytes from the circulatory system [91, 92]. Thus, a decrease in activated microglial cells in the injured brain after HUC-derived cell treatment represents an immunomodulatory effect that could reduce neuroinflammation and increase recovery. Although how HUC-derived cells suppress activated microglia is unknown, a recent study implicates the role of CD11b+ and CD19+ cbMNCs in reducing microglial survival and CD4+ in sustaining microglia in vitro after hypoxic conditions . In our study, cbMNCs were CD19+CD4+, cmMSCs were CD19-, but had enriched CD4+ expression, whereas cbMSCs had negligible expression of both CD19 and CD4. Thus, markedly reduced activated microglial cells in each of these three cell groups indicates that additional factors could be involved in mediating microglial suppression as well as the possibility that CD19+ cells override the protective effects of CD4+ cells.
In stroke, various pro-inflammatory cytokines are secreted from infiltrated leukocytes and macrophages through activation of resident microglia. The insignificant differences in IL-6, IL-β, and TNF-alpha mRNA expression within the cbMNCs and cmMSCs groups compared with control animals may be due to time-dependent expression patterns. These cytokines have been shown to peak in expression in the ischemic hemisphere at day 7 after stroke, which was decreased by day 14 . The significant decrease in IL-2, IL-6, and IL-1beta mRNA in the cbMSC group could indicate a mechanistic shift of action of these cells from their parental cbMNC population. Whereas suppression of proinflammatory factors might mediate the recovery observed in cbMSC-treated animals, in the cbMNC group, recovery is more likely associated with the release of growth factors and the ability to attenuate oxidative stress. The negligible antiinflammatory cytokine IL-10 mRNA seen in our study is similar to that of previous reports  in which no expression was seen at day 14 after stroke.
The absence of any human cytokine expression in the ipsilateral hemisphere of the transplanted animals is possibly due to the moderate number of surviving human cells, low levels of cytokine secretion by the transplanted cells, or reduced overall inflammation by day 14 after stroke. Future studies assessing the treatment and time-dependent profile of these cytokines will enhance our understanding of inflammation and the effects of umbilical cord cell subtypes on the same. At 2 weeks after stroke, insignificant differences in the expression of reelin and EGF mRNA between the PBS, sham, and cell groups could indicate that reelin-mediated migration of progenitors and EGF-mediated recovery after stroke is time dependent.
Cell survival in immunocompetent stroke rat brains and the absence of adverse events in cbMNCs and cbMSCs might be due to their immunomodulatory function in addition to the presence of phenotypically and functionally immature T-lymphocytes . Cord blood has a higher percentage of homogeneous regulatory T-lymphocytes (Tregs) compared with heterogeneous Tregs present in peripheral blood and a smaller percentage in bone marrow . Tregs are known to dampen the immune response  and have been shown to exert potent antiinflammatory neuroprotective effects after stroke . HUC blood Tregs have also been shown clinically to prevent allogeneic acute graft-versus-host disease (GVHD) . It is therefore important to understand the role of this subpopulation in mediating the survival of transplanted HUC-derived cells without adverse effects and whether these cells play a role in poststroke recovery. Although many studies with Treg-depleted donor allografts have shown enhanced GVHD [99–101], it is worth exploring whether Treg-depleted cord-blood-derived cells home and exert similar neuroprotective effects in stroke, as seen in this study. Further, it would be interesting to determine whether cord-blood and matrix cells have phenotypically and functionally similar Treg populations.
Last, although both cbMSCs and cmMSCs showed similar expression of MSC-specific markers, they differed in their expression of Lin1, CD56, CD4, and CD14, which could account for the differences in functional and histopathologic recovery seen with these cells. It would be interesting to determine whether these differences are preserved with cbMSCs and cmMSCs isolated from the same donor. Further, because of phenotypical and functional heterogeneity within MSC preparations and variations between MSC donors, it would be informative to include multiple MSC preparations from multiple donors in MSC studies [102, 103].
We analyzed cbMSCs isolated from a single donor, which is a limitation of our study. Future studies comparing cbMSCs isolated from multiple donors would therefore provide valuable information on donor-dependent and/or MSC-subpopulation variations in the context of cbMSC transplantation after stroke.