Intestinal Epithelial HIF-2 Is Protective in Gut Graft-Versus-Host-Disease

The failure to control both innate and adaptive immune responses in the gut has recently been implicated as a major pathogenic mechanism in the development of graft-versus-host disease (GVHD). Reduced oxygen availability in the intestine has been causally linked to gastrointestinal disease. During intestinal inflammation, increased metabolic activity of resident and infiltrating immune cells, bacteria and reduced blood flow may lead to a sharp decrease of oxygen, resulting in "inflammatory" hypoxia. The transcription factor family, hypoxia-inducible factor (HIF) originally discovered as a master regulator of the adaptive response to hypoxia, has recently emerged as a key regulator of the innate and adaptive immune responses. The HIF heterodimer consists of an oxygen-labile α subunit (HIFα) and a constitutively expressed HIF-1β subunit. Both HIF-1α and HIF-2α expression are markedly elevated in intestinal epithelial cells of patients with inflammatory bowel disease (IBD) and intestinal epithelial HIF-1 attenuates colitis in preclinical mouse models. Although HIF-2 has not been studied extensively in intestinal inflammation, it has emerged as a key regulator in intestinal iron homeostasis. Given that IBD and GVHD share many pathogenic mechanisms, we hypothesized that a sustained HIF response will protect the host intestinal epithelium from conditioning- and alloreactive T cell-induced gut damage.

To determine the functional significance of intestinal epithelial HIF-1 and HIF-2 in gut GVHD, we generated conditional intestinal epithelial HIF-1α (HIF-1α^ΔIE) and HIF-2α^ΔIE vil-cre knockout (KO) mice on a C57BL/6 (B6) background lacking HIF-1α or HIF-2α in the host intestinal epithelium. Using a fully MHC mismatched B10.BR (H2^k)→B6 (H2^b) bone marrow transplant (BMT) model, loss of intestinal epithelial HIF-2 reduced the median survival time (43d) compared to wild-type (WT) recipients (58d, log-rank test, P < 0.005). Although intestinal epithelial HIF-1 deficiency shortened the median survival time (48.5d), it did not reach statistical difference. Loss of intestinal epithelial HIF-1 or HIF-2 worsened GVHD-induced histopathologic crypt damage compared to WT mice transplanted with T cell depleted bone marrow (BM) and enriched T cells (BM+T), 8d post- BMT. Pronounced subepithelial lifting, mucosal edema and sloughing were more evident in the villus tips of HIF-2α^ΔIE mice than HIF-1α^ΔIE mice. Hyperplastic crypts that are characteristic of regenerating crypts after radiation-induced damage were observed in Ki67-stained ileal/jejunal sections of WT mice post-BMT whereas fewer regenerating Ki67-labeled crypts were found in both HIF-1α^ΔIE and HIF-2α^ΔIE mice. In control T cell depleted BM groups (WT, HIF-1α^ΔIE and HIF-2α^ΔIE), Ki67⁺ -proliferating cells resided at the crypt base. Using quantitative real-time PCR analysis, we determined whether intestinal epithelial HIF-1 and HIF-2 differentially regulated the expression of Paneth cells and intestinal stem cell markers in the jejunum, 8d post-BMT. A 5-fold and 2-fold decrease in lysozyme (Lyz) mRNA levels occurred in WT (p<0.001) and HIF-1α^ΔIE BM+T mice (p<0.001) compared to their respective BM groups, 8d post-BMT. However, due to a 2-fold decrease in endogenous Lyz expression in HIF-1α^ΔIE (p<0.01) and HIF-2α^ΔIE (p<0.001) BM mice compared to WT BM mice, Lyz levels were not differentially changed among BM+T groups. Loss of either intestinal epithelial HIF-1 or HIF-2 reduced Reg3γ (p<0.05) and Sox9 (p<0.01) levels whereas only epithelial HIF-1 deficiency reduced Hes-1 by 2-fold (p<0.001) and prevented the recovery of the Lgr5 levels (p<0.01) compared to WT BM+T mice. In summary, we found that both intestinal epithelial HIF-1 and HIF-2 may protect the intestinal stem cell niche from GVHD-induced injury. Importantly, our results suggest that intestinal epithelial HIF-2 may be necessary in preventing GVHD-induced mortality.