Allogenic C57B/6->B6D2F1 Transplantation As a Model for Graft Versus Host Disease in the Brain

Bone marrow transplants are a common and reliable treatment for many disorders; however, in roughly 50% of patients, graft versus host disease (GvHD) occurs, causing significant morbidity and mortality. GvHD is an immune mediated disorder in which donor T cells recognize host antigens as foreign, causing donor T cells to proliferate and attack healthy host tissue. Though the pathogenesis of GvHD in organs such as the skin and colon has been well-documented, little is known about the effects of GvHD on the brain. Some clinical case studies have associated extensive neurological issues with onset of GvHD, with symptoms such as seizures, cognitive dysfunction, and cerebellar impairment including hemiparesis. Yet research involving the brain in GvHD lags behind research in other tissues. Our lab, working with a MHCII mismatched GvHD murine model, has studied GvHD in the brain in an effort to better understand the neurological effects of GvHD.

It has been established that an increase in CD4+ and CD8+ T cells in GvHD causes much of the autoimmune reaction associated with the disease. Because of this, we sought to determine if these cell types infiltrate the brain, and if so, where. Using histochemical staining in a C57BL/6->B6D2F1 model (syngenically transplanted B6D2F1 mice were healthy controls), we saw significantly increased populations of CD3+ T cells (p<0.0001) in perivascular, periventricular, and meningeal locations. CD3+ cells are rarely observed in healthy mice and indicate an abnormal immune response in the brain. There was a significant increase in CD8+ (p=0.0316), but not in CD4+, cell counts in brains of GvHD mice as compared to brains of healthy control mice. These T cell populations were not found in parenchymal locations except proximal to micro vessels. The location of infiltrating T cells suggests that the epithelial cells lining neural blood vessels and the ependymal cells lining the ventricles are likely the target of inflammatory attack.

Since an increase in CD8+ T cell populations suggests a pro-inflammatory Th1 immune response, we also investigated B226+ (B-cell) and IBA1+ (macrophage/microglia) cell populations in the brain. B-cells are associated with a Th2 response, while Th1 responses increase IBA1+ cell populations. We found no increase in numbers of B226+ cells in GvHD vs healthy mice, but found a significant increase in the number of IBA1+ cells (p<0.0001) in brains of GvHD mice, indicative of a Th1 immune response. Further, using a luminex multiplex assay, we found decreased serum levels of IL-4 (p = 0.0011) as well as increased levels of VEGF (p=0.0012) and ICAM-1 (p=0.0222) in GvHD mice as compared to healthy controls. The decreased level of IL-4 is consistent with the activation of a Th1 response and suppression of an anti-inflammatory Th2 response. The increase in ICAM-1, a molecule that regulates vascular permeability, indicates that GvHD induces increases in proteins important to vascular integrity, possibly allowing T cells to enter the brain via the blood brain barrier. Past studies have found increased levels of VEGF, and likewise of angiogenesis, in other organs affected by GVHD. Interestingly there is often a spike in VEGF-A levels following central nervous system injury, and the increase observed in our GvHD model may be a result of GvHD injury to the brain.

Finally, since behavioral changes are often the first indicator of neurological involvement in GvHD, we wanted to assess whether our GvHD mouse model showed any notable changes in behavior. We found no motor deficits as measured through inverted grid and pole walk, nor did we see any difference in Morris Water Maze learning and recall between GvHD and control mice. However, we saw slower swimming speeds for GvHD mice (p<0.0001), correlating with signs of generalized somatic illness associated with GvHD, such as weight loss, signs of dehydration, and whisker barbering which may indicate stress. It is likely that the decrease in swimming speed is a result of generalized illness, rather than a neuromotor deficit caused by GvHD.

In conclusion, our data suggest that the epithelial and ependymal cells surrounding the brain are a target of GvHD. Additionally, we found data supporting the theory that acute GvHD is associated with a Th1 response, as well as elevated serum ICAM-1 and VEGF levels. Our data indicate that GvHD may play a role in the brain of GvHD patients, an organ that has not previously been implicated in the pathology of GvHD.