Hematopoietic Stem and Progenitor Cells Acquire an Inflammatory Memory of the Acute Myeloid Leukemia Niche

Recent studies suggest that acute myeloid leukemia (AML) generates inflammation in the bone marrow (BM) niche to an extent that compromises the function of healthy hematopoietic stem and progenitor cells (HSPC) and impacts patient prognosis. AML xenograft studies have identified several pro-inflammatory mediators (IL-6, IL-1b) that are secreted into the BM niche by leukemic blasts. Prior work by our laboratory also showed contributions involving other AML-secreted factors, in regulating HSPCs in the leukemic niche. HSPC are cellular components of the innate immune system, but the full extent of their inflammatory adaptation by AML is unknown. To overcome some of the experimental constraints inherent in the xenograft approach, we first established non-conditioned congenic grafts of C1498 murine AML cells in C57B/6 recipients. Assessment of the BM revealed that, even at low AML burden, these animals contained a range of elevated pro-inflammatory factors. Single cell RNA-Seq studies of healthy HSPCs, obtained concurrently at sacrifice confirmed an inflammatory transcriptional signature in HSC and MPP populations, with GSEA enrichment for interferon-alpha and -gamma signaling pathways. These observations led us to hypothesize that AML-experienced (inflammation-primed) healthy HSPCs may acquire a durable long-term inflammatory memory that shapes their response to subsequent stimulation during relapse or infection. To test this hypothesis, we adoptively transferred FACS-sorted healthy AML-experienced HSPCs from C1498-engrafted mice (HSPC^AML) into secondary recipients. HSPCs transplanted from PBS-injected recipients (HSPC^PBS) serve as controls. To assess whether HSPC^AML would respond to immune challenge more aggressively compared to HSPC^PBS, we exposed these animals with a single-injection of lipopolysaccharide (LPS), followed by tissue harvest and analysis after 24 hours. Bulk RNA-Seq analysis of HSPC^AML from the LPS-challenged recipients showed amplified GSEA signatures for tumor necrosis factor- and NF-kB signaling pathways compared to HSPC^PBS. In addition, LPS-challenged HSPC^AML showed amplified expression by several krüppel-like factors (KLF) genes. Analysis for transposase-accessible chromatin loci (ATAC-Seq), concurrently obtained from LPS-challenged HSPC^AML echoed the enrichment of KLF motifs as well. Collectively, the C1498 model suggests that initial AML exposure programs HSPC^AML for an amplified inflammatory and differentiation program. To further ascertain disease relevance and reproducibility we turned to a doxycycline-inducible (i)MLL-AF9 AML model. Given the low leukemic burden in the BM of C1498-grafted animals, we wanted to test the hypothesis that secreted factors accounted, at least in part, for the inflammatory memory. Indeed, injection of AML secreted factors from iMLL-AF9 models revealed a strong inflammatory signature in bulk RNA-Seq studies. To determine if innate immune priming in HSPC can be directly mediated by AML crosstalk, we exposed ex vivo-expanded HSPCs to purified secreted factors (SF^AML) from (i)MLL-AF9 AML blasts, followed by a secondary inflammatory challenge with LPS seven days later. Gene expression analysis of SF^AML-experienced LPS-challenged HSPCs confirmed myeloid-skewed progenitor differentiation and increased expression of inflammatory targets. Together, our data show for the first time that phenotypically normal, AML-experienced HSPCs carry a durable inflammatory legacy whereby primary exposure to AML secreted factors alters their response to secondary stimuli. Our experiments phenocopy observations of trained immunity in other experimental systems, and suggest the acquisition of an innate immune memory in the AML niche. Immune priming may impact the role of these HSPC in the context of inflammation associated with aging, or clones with predisposing molecular lesions such as TET2 or DNMT3A.