p38MAPK Inhibition Blocks Inflammatory Signaling in Acute Myeloid Leukemia

Background:p38 mitogen-activated protein kinase (p38MAPK) is activated by various pro-inflammatory and stress-related stimuli, and has been an attractive therapeutic target for autoimmune diseases. p38MAPK (hereafter referred to as p38)  signaling is also involved in cell proliferation, differentiation, apoptosis, and invasion, suggesting that it may be a potential therapeutic target for cancer. We found that inflammatory cytokines, including interleukin-1 (IL-1), promote growth and survival of more than half of the acute myeloid leukemia (AML) patient samples we tested. Since p38 is a downstream mediator of inflammatory pathways, we hypothesized that targeting p38 might be an effective therapeutic strategy in AML and other hematologic malignancies. To test this hypothesis, we evaluated the effectiveness of three p38 inhibitors using in vitro studies in primary AML patient samples. We found that targeting p38 blocks IL-1-activated extrinsic signaling and is a critical therapeutic target in a large subset of AML patients.

Methods: We screened ~1000 primary leukemia patient samples for sensitivity to p38 inhibition using varying concentrations of doramapimod (BIRB 796) in a cell growth assay. We compared the sensitivity profile of doramapimod with 2 other small-molecule p38 inhibitors currently in clinical trials: ARRY 614, a dual p38/Tie2 inhibitor, and ralimetinib, which blocks activation of p38 by its substrate MK2. We determined cell viability, survival, and downstream signaling in the presence of 10 ng/ml IL-1α or IL-1β. Patient samples with IC50 < 1000nM were considered drug responsive.

Results: In our patient population, we observed response rates of 31% in AML (109/350), 27% in myelodysplastic syndromes (MDS; 25/93), 19% in myeloproliferative neoplasms (23/123), 13% in mature B-cell neoplasms (30/232), and 10% in precursor lymphoid neoplasms (19/182). Focusing on AML, we compared the sensitivity profile of doramapimod with two other small-molecule p38 inhibitors, ralimetinib and ARRY 614. These inhibitors showed strikingly similar sensitivity profiles to doramapimod when tested in an additional 25 primary AML samples, with ~25% responsive and median IC₅₀ of 11 nM for ARRY 614 (range: 7-650nM), 105 nM for ralimetinib (range: 7-850nM), and 18 nM for BIRB 796 (range: 13-40nM).  Because IL-1 is known to stimulate p38 signaling, we compared the response rates for these three p38 inhibitors with or without IL-1 in a dose-response study. IL-1 increased the percent of AML samples responding to p38 inhibition from 25% to 60%, indicating a potentially important role of extrinsic inflammatory stimuli in p38 inhibitor sensitivity. Consistent with this all three p38 inhibitors were similarly effective in blocking the growth of primary AML CD34+ progenitors, suggesting that targeting p38 might reduce early progenitor AML cells. Further, we compared doramapimod, ralimetinib, and ARRY 614 for their ability to inhibit p38 phosphorylation in primary AML samples using flow cytometry and immunoblot analysis; all three inhibitors blocked p38 pathway activation in AML cells. Notably, in clinical studies of ARRY 614 in MDS patients, preliminary biomarker analyses demonstrated persistent inhibition of phospho-p38 in the bone marrow during the treatment. Also, consistent with functional inhibition of p38, there was a profound decrease in plasma cytokine concentrations, most significantly IL-1, during ARRY 614 treatment. In 250 primary AML samples, we observed no correlation between BIRB 796 sensitivity in vitroand clinical metrics such as white blood cell count, blast percentage in peripheral blood or bone marrow, karyotype, or tumor genotype. This suggests that IL-1 and p38 activation might be independent biomarkers of drug sensitivity.

Conclusions: These data underscore the importance of the p38MAPK pathway in the pathobiology of AML and provide strong preclinical evidence to support p38MAPK as a therapeutic target. Targeting p38MAPK might also block tumor-extrinsic signaling, as indicated by IL-1-activated signaling. That all three p38MAPK inhibitors showed comparable sensitivity profiles holds promise for ARRY614, which showed the lowest median IC50 and is currently in clinical development. In addition, with further study these findings may be extended to hematologic malignancies other than AML.