CNS allografts have long been considered for treatment of neurological disease and there is extensive evidence that allografts can survive in the CNS without long-term immunosuppression; however, how tissue mismatch might affect outcome in other ways is not clear. We have found that allogeneic transplants elicit immune signaling in the brain which detrimentally affects NPC survival and selectively impairs the production of neurons. These effects appear to occur during the first weeks following transplant and can be prevented by short courses of NSAID administration but not by immunosuppression using CsA.
In humans, allogeneic cell transplants almost invariably include a short course of immunosuppression (most commonly CsA) to minimize chances of graft rejection through a T-cell-mediated adaptive immune response 
. We did not detect any significant changes in the behavior of isografts, allografts, or endogenous progenitors following CsA treatment in vivo
(). In contrast, we have shown in prior work that the NSAIDs can attenuate the negative effects of innate proinflammatory signaling on hippocampal neurogenesis 
and find here that a broad spectrum NSAID, indomethacin, or the more selective PPAR-γ activator rosiglitazone can promote the production and/or survival of neurons produced by transplanted NPCs.
Steroidal anti-inflammatory drugs are commonly used in neural transplantation. Although effective, such drugs were not considered here due to the known suppressive effects of adrenal steroids on adult neurogenesis 
. As a potential substitute, NSAIDs such as indomethacin also have good CNS penetration and act by attenuating innate proinflammatory signaling through inhibition of Cox1/Cox2 and activation of PPAR-γ 
. Unfortunately, like CsA or steroidal anti-inflammatory drugs, indomethacin also has unwanted side effects at doses sufficient to attenuate inflammation in the CNS. In contrast, rosiglitazone, a non-classical NSAID that has shown promise as an immunomodulatory drug in the clinical treatment of Alzheimer's disease 
, is well tolerated over long periods and was equally effective as indomethacin at reversing allograft-specific impairment in neurogenesis.
The effectiveness of rosiglitazone in overcoming an allograft-selective loss of neurons provides several intriguing leads into the mechanisms that may be responsible. Rosiglitazone selectively activates PPAR-γ which consequently alters expression of genes regulated by PPAR-γ There are roughly 100 known targets, among which are numerous proinflammatory cytokine genes 
. Our cytokine expression data confirms that rosiglitazone attenuates proinflammatory cytokine response elicited by allografts. While speculative, it seems reasonable to predict that cytokines that are most prominently attenuated by both indomethacin and rosiglitazone may be relevant to the allo-specific impairment in neuron production or survival. Among these are TNFα and IL-1β, both important early mediators of the acute innate proinflammatory response that triggers subsequent lymphocyte infiltration, antigen presentation, and initiation of the adaptive response. In vitro
, both cytokines have been shown to directly attenuate neurogenesis in NPCs 
. The other four cytokines/chemokines that have been down regulated by both indomethacin and rosiglitazone are IL-12, IL-13, MIP-1α and MCP-3. It is still unknown whether any of these cytokines has direct effect on grafted or endogenous NPCs and future studies are needed to test it. IL-12, MIP-1 and MCP-3 are all pro-inflammatory cytokines which play important role in the regulation of macrophage/microglia activation and migration. It is possible that NSAIDs may work in the present model by simply attenuating proinflammatory cytokine expression following activation of the acute innate immune response.
We have confirmed that purified NPCs in vitro
do not express MHC I or MHC II and it seems unlikely that infiltrating T cells would be strongly activated by presentation of allo-specific MHC. However, the cytokines produced acutely in response to the injection injury (or that may be present due to a disease or injury present in a transplant context) may upregulate NPC MHC expression, as seen in vitro
). This may be sufficient to initiate a recognition response in vivo
. MHC I expression was low in surviving NPCs but also detected in host microglia and macrophages. This is consistent with an upregulation of MHC following exposure to cytokines in the graft with phagocytosis and re-presentation of allo-MHC by host microglia and macrophages (Fig. S3
). This should provide an effective context for the recruitment of allo-specific T cells. Although T cells are present in the grafts, CsA failed to alter outcome and we conclude that classical calcineurin-dependent T cell activation, amplification and lysis of NPCs is unlikely to account for all of the effects. Furthermore, the endogenous neural progenitors would not be recognized by allo-reactive T cells yet they are equally inhibited in the production and survival of neurons in the allograft environment. While this argues against T cell specific lysis of allogeneic neurons, it does not completely exclude a role for T cells. It is possible that allogeneic grafts do trigger a larger number of infiltrating cells during the earliest stages of innate immune activation and antigen presentation. Cytokines acutely released by these cells (including T cells) may reinforce or amplify microglial activation but abundant Fas Ligand 
and other inhibitory factors in the CNS attenuate subsequent steps of the adaptive response and prevent a classical CsA-sensitive T cell amplification and cytolytic graft rejection. The presence of CD4 and CD8 cells at the transplant site (Fig. S4
) confirms that T cells are present and suggests a degree of T cell response from the host, possibly mediated by elevated CD8-positve cells. However, cell survivals at 2 weeks and 4 weeks were not different and inhibition of T cell receptor signaling with CsA did not significantly alter outcome. It seems likely that allograft-induced upregulation of cytokines or chemokines may play a relatively important role in the observed inhibitory effect of allografts on neuron production and survival in the context of the present studies.
It is clear from earlier work that intracerebral allografts show good long-term survival when transplanted into the striatum 
. However, the graft can be recognized and eliminated by a classical allograft response, particularly when the animal has been primed with a peripheral allograft 
. It is possible that hippocampus is a special target site close to the ventricles and therefore more prone to immunological effects than other target areas such as the striatum 
, which may account for the observed reduction in survival of allograft vs. isograft in our work relative to results seen in striatal grafts 
. However, if this response is incrementally larger in the hippocampus, it is still not sufficient to cause systemic priming 
. Splenocytes from transplanted animals show no evidence of prior exposure to allo antigens (Fig. S5
). This does not eliminate a role for T cells but does confirm that the local allo-specific T cell response is not sufficiently strong to elicit systemic immunity. Instead, we find evidence that innate immune signaling and related cytokine signaling may play a larger role.
The mechanism underlying the selective upregulated cytokine production by allo- vs. isografts is unclear but it is possible that these influences may stem from the signaling networks mediated by many immune cell types, including microglia, and infiltrating lymphocytes such as T cells, NK and/or NK-T cells. Reduced cytokine abundance may attenuate many of the subsequent innate and adaptive responses in the host, including T-cell relevant processes of antigen presentation and expression of co-stimulatory molecules. It is also possible that NSAIDs act directly on neural progenitor cells to promote neurogenesis but we have not been able to detect differences in endogenous neurogenesis when these drugs are administered to naïve animals. This suggests that the primary mechanisms are related to cytokine production and/or modulating cytokine response in the host and graft.
CsA was surprisingly ineffective but this result is consistent with earlier studies evaluating allogeneic cell survival in the presence or absence of CsA 
. These studies have led to a general assumption in the field that allogeneic mismatch may not be relevant in neural transplantation because cell survival is similar in the presence or absence of CsA. In our model, even cells carrying the xenotypic jellyfish GFP protein had equivalent survival rates in wild type hosts and in congenic hosts where endogenously expressed GFP would not be a “foreign” protein. This highlights the relative tolerance of the brain to foreign tissues yet strengthens the importance of recognizing and understanding the more subtle immune-brain interaction that revolves around innate immune response and elaboration of proinflammatory cytokines. Although promising, care should be taken in interpreting our results using NSAIDs. The current study focused on the early post-graft period and did not include time points beyond 4 weeks. The chronic effects of both major and minor histocompatibility mismatch have not been fully addressed in the short time frames of the present study and it is possible that drugs such as CsA or adrenal steroids are indeed important in long-term outcome. However, the present study does draw our attention to the unexpected utility of non-steroidal immunomodulatory drugs that attenuate innate immune signaling and cytokine release triggered by transplantation. Together, the results confirm prior work that strongly mismatched allogeneic neural stem cell grafts can survive in the absence of immune suppression. However, it is also clear that histocompatibility does impact cell survival in this model and has a strongly selective effect on the abundance of graft-derived new neurons. Non-traditional immunomodulatory drugs, such as NSAIDs, may be valuable adjunct treatments for improving transplant survival in the CNS and promoting the therapeutic integration and retention of graft-derived neurons.