Allogeneic hematopoietic stem cell transplantation (alloHSCT) is a curative treatment for high-risk leukemia and many non-malignant hematologic disorders. However, the use of alloHSCT as a therapy 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 alloreactive T cells increase phosphorylation of the intracellular energy sensor AMP-activated protein kinase (AMPK) as early as day 3 post-transplant (levels 8-fold higher in alloreactive T cells compared to naive or syngeneic controls (p=0.0003)) and donor T cells lacking AMPK (AMPK KO) cause significantly less GVHD in both major- and minor-histocompatibility mismatch models of GVHD. At the same time, AMPK KO T cells equivalently clear leukemia cells and extend recipient survival in leukemia models with high tumor burden. Transplantation of cells lacking AMPK lowers 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 AMPKKO respectively) but interestingly, increases both the percentage and total number of regulatory T cells (Treg) (0.85 +/- 0.32x104 vs. 1.69 +/- 0.34x104, wt vs. KO, p=0.004).

To understand these findings mechanistically, we sought to study the role of AMPK in donor Treg expansion and stabilization post-alloHSCT. Canonically, interleukin-2 (IL-2) signaling, through phosphorylation of Stat5, promotes Treg generation and stability, while inflammatory signals (e.g. IL-6 signaling through Stat3) are known to decrease Treg stability post-transplant. Studies testing IL-2 sensitivity in KO cells, or the degree of Stat5 phosphorylation post-transplant, were unrevealing. However, AMPK KO Tregs significantly decrease their sensitivity to IL-6, with a 50% reduction in phosphorylated-Stat3 levels following intermediate levels of cytokine stimulation (Figure 1A). To demonstrate the functional consequences of this difference in cytokine sensitivity, we exposed in vitro generated Treg to increasing concentrations of IL-6. As expected, the percentage of wt FoxP3+ cells decreased sharply with increasing IL-6 concentrations (decreasing from 50% to 14%), while AMPK KO Tregs were minimally affected over this dose range (47% versus 42%; Figure 1B, p<0.02).

To determine if increased Treg stability was the predominant reason for improvements in GVHD seen with AMPK KO cells, CD25+ cells were removed from the donor inoculum of AMPK KO or wt mice by magnetic selection (to eliminate >90% of Treg), and then either CD25-replete or CD25-depleted sets of cells were transplanted into B6xDBA2 F1 recipients in a major MHC mismatch model of GVHD. Despite a decrease in AMPK KO Treg numbers on day 7 to levels commiserate with wt donors (Figure 2A), elimination of AMPK KO Treg had only a modest effect on GVHD severity (Figure 2B), implicating a substantial role for AMPK in the non-Treg compartment.

To explore the role of AMPK in non-Tregs further, we investigated the sensitivity of unmanipulated wt and AMPK KO T cells to IL-6 stimulation. Similar to FoxP3+ cells, both naïve CD4 and CD8 T cells lacking AMPK experienced lower levels of Stat3 phosphorylation in response to graded concentrations of IL-6 (Figure 3A). Furthermore, if IL-6 was added to a 4-day culture with anti-CD3 stimulation, the number of wt T cells recovered after 96 hours increased >2.2-fold. In contrast, AMPK KO cells experienced minimal changes in T cell number over a wide range of IL-6 concentrations (Figure 3B).

In conclusion, we define in these studies a novel role for AMPK in controlling donor T cell sensitivity to IL-6, both at the level of Stat3 phosphorylation, and in terms of functional consequences. Not only does this striking result elegantly explain our previous observation of increased Treg percentages, and decreased donor T cell numbers, following transplantation of AMPK KO cells, it continues to make the case that inhibition of AMPK represents an attractive clinical target for the amelioration of GVHD.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.