Deletion of AMP-Activated Protein Kinase (AMPK) in Donor T Cells Protects Against Graft-Verus-Host Disease through Control of Regulatory T Cell Expansion and Target Organ Infiltration

Beezhold, Kevin; Moore, Nathan; Chiaranunt, Pailin; Brown, Rebecca; Byersdorfer, Craig A.

doi:10.1182/blood.V128.22.806.806

Abstract

Allogeneic hematopoietic stem cell transplantation (alloHSCT) represents a curative treatment for high-risk leukemia and a number of non-malignant hematologic disorders. However, the therapeutic use of alloHSCT remains limited by acute graft-versus-host disease (GVHD), where activated donor T cells attack and destroy host tissues in the skin, gastrointestinal tract, and liver. We have previously shown that the alloreactive T cells responsible for GVHD increase their dependence on the oxidation of fat relative to either syngeneic or naive T cells. To explore this adaptation mechanistically, we studied the role of AMPK, an intracellular energy sensor and known driver of fat oxidation, in donor T cells during GVHD. Alloreactive T cells increased phosphorylation of AMPK as early as day 3 post-transplant, with up-regulation in pathways both up- and downstream of AMPK. Changes in phosphorylation were up to 8-fold higher in alloreactive T cells compared to naive T cells or syngeneic controls (p=0.0003). We then investigated the role of AMPK during GVHD pathogenesis using donor cells deficient in AMPK (from AMPKα1^fl/flα2^fl/fl x CD4-Cre mice). AMPK^-/- T cells caused significantly less GVHD in both major-MHC and minor-histocompatibility mismatch models of GVHD (Figure 1A), with a coordinated decrease in the number of donor T cells recovered on day 7 post-transplant (3.15 +/- 0.49x10⁶ versus1.87 +/- 0.53x10⁶, p=0.0006, wildtype (wt) versus AMPK^-/- respectively). Importantly, expansion of syngeneic T cells was unaffected by AMPK deficiency (Figure 1B).

We next investigated the ability of AMPK^-/- T cells to mount effective cytotoxic and anti-leukemia responses. AMPK^-/- T cells demonstrated equivalent cytotoxicity against MHC-mismatched targets both in vitro and in vivo and differentiated in similar proportions into cytokine-producing cells (IFN-γ, TNFα, IL-17, and IL-4). We then assessed graft-versus-leukemia (GVL) potential in AMPK^-/- cells using a GVL model with high tumor burden. AMPK^-/- T cells exhibited equivalent clearance of p815 leukemia cells on day 13 post-transplant (Figure 2A), and extended survival of recipient mice similarly to wt T cells (Figure 2B). To elucidate possible mechanisms underlying this separation of GVL and GVHD responses, we evaluated metabolic pathways in wt and AMPK^-/- T cells recovered on day 7. To our surprise, rates of fatty acid oxidation were identical between wt and AMPK^-/- T cells and loss of AMPK did not impact alloreactive T autophagy, nor impair signaling downstream of mammalian target of rapamycin.

To define the mechanism underlying AMPK^-/- benefits, we quantitated levels of regulatory T cells (T_reg) on day 7 post-transplant. In contrast to expectation, both the percentage and total number of T_reg increased in mice receiving AMPK^-/- T cells (0.85 +/- 0.32x10⁴ vs. 1.69 +/- 0.34x10⁴, wt vs. AMPK^-/-, p=0.004). Loss of AMPK facilitated donor T_reg expansion, as elimination of FoxP3⁺ cells prior to transplantation abrogated differences between wt and AMPK^-/- donors on day 7. Importantly, T_reg levels were equal in wt versus AMPK^-/- donors prior to transplantation. Finally, we assessed the ability of AMPK^-/- T cells to infiltrate into GVHD target organs. As shown in Figure 3, peri-portal infiltration of AMPK^-/- cells was significantly reduced compared to wt T cells, and infiltrates in recipients of AMPK^-/- cells contained many fewer CD3⁺ T cells per high-powered field, with CD3⁺ cells representing a lower percentage of cells overall. Decreased hepatic infiltration correlated with a lower percentage of circulating CD4⁺ cells and lower levels of the integrin pair α4β7 (55.2 +/- 1.4% versus 47.2 +/- 2.8% α4β7^Hi cells, p=0.0017, wt vs. AMPK^-/-).

In conclusion, deletion of AMPK in donor T cells decreases GVHD severity but spares anti-leukemia responses and preserves homeostatic immune reconstitution. Mechanistically, this occurs through a decrease in pathogenic T cell numbers, an increase in the number and percentage of T_reg cells, fewer circulating CD4⁺ cells, and decreased infiltration of donor cells into target organs. From these findings, we conclude that AMPK represents a clinically relevant target in donor T cells pre-transplant and are actively exploring ways to translate this exciting therapy into clinical practice.