ATG5 Dependent Autophagy in Graft-Versus-Host Disease Target Organs Mitigates Disease Severity Post Allogeneic Bone Marrow Transplant

In immune mediated diseases such as graft-versus-host disease (GVHD), research has focused almost entirely on hematopoietic cells. It has recently been posited that mechanisms within target tissues also play a significant role in regulating the severity of disease. Autophagy protects cells from a variety of stressors and is known to affect immune cell responses after allogeneic bone marrow transplant (allo-BMT). However, a definitive and exclusive role of target tissue pathways on mitigating GVHD has heretofore not been demonstrated. Because GVHD target organs are under stress from alloreactive T cells, we sought to determine the role of target tissue autophagy in mitigating the severity of GVHD.

We hypothesized that the stress from alloreactive T cells would result in upregulation of target tissue autophagy. To test this in vivo, we utilized LC3-GFP-RFP reporter mice as transplant recipients in the well-characterized MHC disparate GVHD murine model BALB/c→B6. In these reporter mice, autophagosomes are marked with a dual GFP-RFP tag, and can be assessed for in vivo induction of autophagy by confocal microscopy through quantification of red punctal abundance. The recipient B6 reporter mice were lethally irradiated with split doses of 5Gy and transplanted with 5x10⁶ BM and 3x10⁶ splenic T-cells from either syngeneic B6 (syn) or allogeneic BALB/c (allo) donors. We assessed for induction of autophagy in GVHD target organs (gut, liver) and non-target organs (kidney, heart) and observed that autophagy was induced to a greater extent in target organs.

Because the primary driver of morbidity and mortality after allo-BMT is gastrointestinal (GI) GVHD, we next hypothesized that autophagy protects the GI epithelium from alloimmune damage. To determine the effects of autophagy exclusively in the GI tract, we generated mice that lack a macroautophagy protein, ATG5, in the GI epithelium only (Atg5^fl/flVil1-Cre) and utilized them as recipients in BALB/c→B6 allo-BMT. As hypothesized, survival of allo Atg5^fl/flVil1-Cre mice was significantly reduced (P<0.0001) compared to littermate controls. Importantly, all syn Atg5 KO mice survived the entire period of observation, demonstrating that protection from conditioning regimen toxicity did not require ATG5-dependent autophagy in the gut. We harvested both GVHD target and non-target organs and performed histopathological analysis at day 7 post-transplant and found that histological GVHD severity in the gut was significantly higher in allo Atg5^fl/flVil1-Cre mice compared to Cre-null WT littermate controls (P=0.0063). All other organs were unaffected by loss of ATG5 in the GI epithelium, and we observed no difference between WT and Atg5 KO syn recipients in any organ. We harvested splenocytes and tissue-resident lymphocytes from the gut and found similar numbers of donor CD4 and CD8 T cells, activation markers CD69 and CD62L, and IFNγ and TNFα positive cells. Furthermore, similar levels of the proinflammatory serum cytokines IFNγ, TNFα, and IL-6 were observed in the sera of WT and Atg5 KO mice.

We next determined whether autophagy is critical for protection in other GVHD target organs after allo-BMT. To this end, we generated Atg5^fl/flAlb-Cre mice, which lack Atg5 expression in hepatocytes. Using a similar model of acute GVHD as above, we observed dramatically shorter survival in the Atg5^fl/flAlb-Cre allo recipients (P=0.025) along with an increase in the hepatocyte damage markers bilirubin (P=0.0001), alanine aminotransferase (P=0.036), and alkaline phosphatase (P=0.0002) in Atg5 KO allo recipients. As before, there were similar numbers of splenic and liver-resident donor T cells, activation markers, and proinflammatory cytokines.

To determine if autophagy is exclusively protective in GVHD target organs, we next generated Atg5^fl/flMyh6-Cre mice, which lack Atg5 expression in cardiac myocytes, and Atg5^fl/flAlb-Cre mice, which lack Atg5 expression in podocytes, as a control for lack of ATG5 in non-GVHD target organs. Using a similar GVHD model as above, we monitored these mice and littermate controls for survival and severity of GVHD symptoms. In contrast to our findings in the liver and gut, we observed similar survival (P=0.17, Myh6; P=0.49, Nphs2) and GVHD severity between allo recipients.

Collectively, our results demonstrated that autophagy is a tissue-intrinsic protective response that regulates disease severity in GVHD target organs.