Inflammatory Niche Signalling Drives Genotoxic Stress in Hematopoietic Stem Cells and Predicts Leukemic Evolution in Human Leukemia Predisposition Syndromes

Primary alterations of the mesenchymal niche can induce myelodysplasia and acute myeloid leukemia in mouse models, introducing a concept of niche-driven leukemogenesis (Raaijmakers et al, Nature 2010). The molecular mechanisms and human relevance of this concept, however, have remained elusive.

We addressed these key questions by modelling Shwachman-Diamond-Syndrome (SDS), a human monogenic congenital disorder caused by loss-of function mutation in the SBDS gene and characterized by skeletal defects, bone marrow failure and a striking propensity for leukemic evolution.

Targeted Sbds deletion from mesenchymal progenitor cells (MPCs) in mice (Osx^Cre/+Sbds^f/f; OCS^f/f) resulted in bone abnormalities faithfully recapitulating human disease, including short stature and early-onset osteoporosis. Skeletal defects were associated with genotoxic stress in hematopoietic stem and progenitor cells (HSPCs) as demonstrated by mitochondrial membrane hyperpolarization, oxidative stress, DNA damage and cell cycle checkpoint activation (transcriptional modulation of DNA damage response/repair pathways and G0-G1 cell cycle arrest). DNA damage could be partially rescued by in vivo administration of the ROS scavenger N-acetylcysteine supporting the notion of niche induced DNA damage in HSPCs induced by mitochondria-derived superoxide radicals.

Mechanistically, Sbds deficiency caused activation of the p53 tumor suppressorpathway in MPCs (upregulation of P53 and transcriptional activation of downstream targets (GSEA). Genetic deletion of Trp53 from MPCs (Osx^cre/+Sbds^f/fTrp53^f/f mice) rescued the skeletal phenotype and genotoxic stress in HSPCs. Comparison of the transcriptome of MPCs from OCS^f/f mice to their highly FACS-purified mesenchymal (CD45^-CD235^-7AAD^-CD31^-CD271⁺CD105⁺) human equivalents from SDS patients (RNAseq; n=5) demonstrated a striking overlap in disrupted gene programs (GSEA), including ribosome biogenesis and significant overexpression of the proinflammatory molecules such as S100A8 and S100A9, bona fide p53 downstream targets. Activation of p53 and inflammatory molecules was an MPC-autonomous consequence of Sbds depletion as demonstrated by ex vivo knockdown of the gene in OP9 cells. S100A8/A9 overexpression and secretion from MPCs from OCS^f/f mice was confirmed by FCM and serum ELISA. Exposure of HSPCs to recombinant murine S100A8/9 resulted in increased DNA damage and apoptosis associated with transcriptional activation of TLR4 downstream signaling, a bona fide S100A8A9 receptor. In vivo TLR4 blockade by neutralizing antibodies resulted in reduced γH2AX foci in HSPCs from OCS^f/f mice, in support of the existence of a Tpr53-S100A8/A9-TLR4 axis driving genotoxic stress. Formal demonstration that niche-derived S100A8/9 is sufficient to drive genotoxic stress in HSPCs was provided by transplantation of wild-type hematopoietic cells into recipient S100A8/A9 transgenic mice (Cheng et al., 2008) resulting in accumulation of mitochondrial superoxide radicals and DNA-damage in wild-type HSPCs.

Finally, to further define the clinical relevance of this inflammatory MPC-HSPC axis to human disease, we performed massive parallel RNA-sequencing of FACS purified mesenchymal cells from homogeneously treated low-risk MDS patients (n=45). Overexpression of S100A8 and S100A9 in MPCs(confirmed by IHC) was found in a considerable subset of patients (17/45; 38%). S100A8/9⁺ mesenchymal cells displayed transcriptional activation of p53 and TLR programs, in line with findings in the mouse model. Strikingly, patients in the niche-S100A8/9⁺ group displayed a higher frequency of leukemia evolution (29.4% vs. 14.2%) with significantly shorter evolution time (average 3.4 (1-7.5) vs 18.5 (7-40); p=.03) and progression-free survival (median 11.5 vs. 53 months, p=.03), independent of established prognostic factors and risk classification systems.

Collectively, the data define niche-HSPC inflammatory signaling through the p53-S100A8/A9-TLR axis as an actionable determinant of genotoxic stress and disease outcome in human preleukemia, opening the way to niche-instructed, therapeutic targeting to attenuate leukemic evolution.