Abstract PR01: Novel anti-leukemia activity of the IRE1-XBP1 signaling pathway in hematopoietic stem and progenitor cells

• Published in: Blood cancer discovery 2023-05, Vol.4 (3_Supplement), p.PR01-PR01
• Main Authors: Barton, Brendan M, Son, Francheska, Verma, Akanksha, Glimcher, Laurie, Adoro, Stanley
• Format: Article
• Language: English
• ISSN: 2643-3249; 2643-3249
• DOI: 10.1158/2643-3249.AML23-PR01

Most patients with acute myeloid leukemia (AML) initially achieve clinical remission after chemotherapy but eventually die from disease relapse as they succumb to chemotherapy-resistant AML cells, including the rare fraction of leukemia stem cells (LSC) that persist and often reinitiate a more aggressive disease. While eliminating LSCs has become a desirable strategy to achieve durable AML remission, much remains unknown about LSC vulnerabilities and the specific cellular programs that enable their disease activity. Spurred by increasing evidence that disruptions in proteome homeostasis or “proteostasis” impair stem cell integrity, we investigated how acute myeloid leukemogenesis is impacted by the inositol-requiring enzyme-1 alpha (IRE1a), the most conserved enzyme of the endoplasmic reticulum unfolded protein response that resolves proteotoxic stress in eukaryotic cells. Activated IRE1a catalyzes an unconventional splicing of the XBP1U mRNA to generate an XBP1S transcript that encodes the transcription factor XBP1 (X-box binding protein-1), the main effector of IRE1a-transduced signals. Using transcriptomics and genomics to profile the activation and function of IRE1a in primary human AML samples and murine AML models, we identified the IRE1a-XBP1 axis as a cell-intrinsic brake against pro-myeloid leukemogenic programs in hematopoietic stem and progenitor cells (HSPC). Notably, compared to blast cells from the same AML patient, LSCs expressed significantly reduced levels of the genes encoding IRE1a (ERN1) and its spliced product XBP1S suggesting that downregulation of IRE1a-XBP1 activity promotes LSC activity and acute myeloid leukemogenesis. Consistent with this hypothesis, genetic induction of XBP1S-encoded XBP1 in patient-derived AML cells suppressed their LSC transcriptional program, impaired their clonogenic capacity, and limited their ability to cause AML in immunodeficient animals. By contrast, genetically engineered IRE1a or XBP1 insufficiency enriched the LSC signature in HSPCs and cooperated with the myeloproliferative oncogene FLT3-ITD to initiate a lethal AML disease in mice. Integrated RNA-Seq and ChIP-Seq of XBP1 target genes in HSPCs revealed a mechanism in which IRE1a-induced XBP1 functions as a potent transcriptional repressor of signature pro-LSC gene programs, exemplified by the Wnt/beta-catenin pathway. Accordingly, we defined an independently prognostic “18-gene” signature of XBP1-repressed beta-catenin target genes that when highly