Abstract SY02-01: Inflammatory memory and selective advantage in human clonal hematopoiesis

Abstract A fundamental challenge in cancer biology is to understand why some people develop cancer and others do not despite similar genetic predispositions. Taking the blood system as an exemplar, we seek here to determine if the inflammatory milieu has distinct impact on the clonal fitness and selective advantage of mutant tissue stem cells at the expense of wildtype clones. Clonal hematopoiesis (CH) is an aging-associated phenomenon in which hematopoietic stem cells (HSC) acquire somatic mutations that result in clonal expansion of mutated blood cells and a 12-fold increased risk for myeloid neoplasms, cardiovascular disease, and other aging-associated diseases. Human lineage tracking studies have shown that pre-leukemic CH mutations often arising many decades before disease onset. Although CH has a prevalence as high as 1 in 3 in people over 60 years of age, only some individuals show a readily detectable clone size (≥2% and considered to have clonal hematopoiesis of indeterminate potential). The mechanisms regulating CH clone size are poorly understood, but inflammation is elevated in individuals with CH. Extensive data from murine models also shows that inflammation reprograms HSCs to impair their function. This likely contributes to clonal selection in murine models of CH as acute inflammation also activates murine HSCs, skews them towards myeloid differentiation, and impairs their self-renewal. Some of these inflammatory responses include activation of target genes downstream of TNFα via NFkB, which is known to regulate HSC survival in both murine and human settings. Repeated inflammatory challenges have also been associated with sustained epigenetic changes and accelerated aging of murine HSCs, as well as expansion of clones bearing CH-associated mutations in Dnmt3a and Tet2, the two most frequently mutated genes in CH cases. However, we have shown extensive molecular and functional heterogeneity in the human HSC compartment at steady-state, and it is unclear whether all human HSC respond homogeneously to inflammation. If HSC respond heterogeneously, it is also unclear whether this would impact clonal selection such as during CH. To address these two questions, we developed inflammation-recovery xenograft models to determine the long-term molecular and cellular sequelae of repeated inflammatory challenge on human HSCs in an empirical fashion (Zeng, et al, in revision). No in vivo studies of inflammatory stress on human HSCs have been previously reported. In parallel, we generated a TARGET-seq+ data set of bone marrow hematopoietic stem and progenitor cells from nine DNMT3A or TET2-mutated CH donors and four age-matched controls with no detectable CH mutations (Jakobsen, et al, Cell Stem Cell, 2024). TARGET-seq+ combines simultaneous single cell profiling of transcriptome, genotype, and cell surface immunophenotype, enabling comparisons of mutant (CHMUT) and wild-type (CHWT) HSC obtained from individuals with CH to HSC from age-matched controls without CH. First, we established that human cord blood (CB) HSC in xenotransplantation models recapitulate phenotypes previously observed in the murine system following acute inflammatory activation for 16 hours using human TNFα or lipopolysaccharide (LPS). However, when we mimicked repeated inflammatory challenge through acute treatment with TNFα or LPS at 2w and 10w post-transplantation followed by analysis at 20w, we found a reduced human graft at 20w compared to PBS controls. Thus, repeated challenge resulted in a long-lasting functional HSC impairment long after the treatment. Single cell multiome analysis of human HSCs isolated from this model showed two bona fide HSC subsets: HSC-I showed few molecular changes in response to prior inflammatory stress. However, the second HSC subset, termed HSC inflammatory memory (HSC-iM) showed extensive chromatin accessibility and transcriptional changes 2.5 months after recovery from TNFα or LPS treatment. These changes are primarily in the AP-1 and NF-kB gene regulatory networks and while distinct, also bear striking similarity to epigenetic changes seen in murine epithelial stem cells following wounding. We found HSC-iM share core molecular programs with human memory T cells. Importantly, a transcriptional program reflecting HSC-iM was enriched in HSCs from recovered COVID-19 patients and in HSC from aged individuals and in HSC from individuals with CH from our TARGET-seq+ data set. To gain insight into the impact of the HSC-I and HSC-iM states on selective advantage in CH, we turned to our TARGET-seq+ data set. First, we directly identified both the HSC-I and HSC-iM populations in human CH cases (named HSC1 and HSC2 respectively) validating the relevance of the inflammation-recovery xenograft model established with CB HSC from which they were first identified and confirming that this is not a xenograft artifact. Second, when CHMUT and CHWT within a BM of a subject with CH are compared to each other, gene expression changes occur predominantly in HSC-iM and not HSC-I. Moreover, the presence of the CH mutation results in decreased enrichment of specific inflammatory pathways. Third, we confirmed that CHWT within HSC2-dominant hierarchies were stalled in differentiation, with reduced production of downstream progenitor populations compared to CHWT cells within HSC1-dominant hierarchies. By contrast, mutations in DNMT3A and TET2 were associated with increased progenitor abundance within HSC-iM-dominant hierarchies. Thus, HSC2/HSC-iM is inherently growth restricted, but CH mutations relieve this impairment leading to an increased HSC2/HSC-iM contribution to hematopoiesis and an overall clonal growth advantage. This lends further support to our conclusion that accumulation of HSC-iM underlies age-related hematopoietic dysfunction. We posit that HSC-iM is a “protective adaptation” to inflammation, representing a reserve pool of stem cells that requires a subsequently stronger inflammatory trigger for activation. Additionally, a key hypothesis our work is the notion that selective advantage of CHMUT HSC arises due to a decrease in differentiation output of CHWT HSC-iM. We are currently testing this hypothesis with CRISPR-mediated DNMT3A- and TET2-mutant CB HSC in our xenograft-inflammatory recovery models. We show that HSC-iM progeny monocytes are pro-inflammatory compared to those from HSC-I in our xenograft models, in the COVID patient recovery data, and in steady-state human BM data sets. To gain insight into whether HSC-I or HSC-iM have differing functional capacity or biases to generate downstream progeny, raw RNA sequencing data from the CD34+CD38-CD45RA- (HSC/Progenitor) scMultiome was combined with scRNA-seq data from CD33+ mature myeloid cells isolated from those same xenografts. This allowed us to examine the differentiation output of clonally defined HSC-I and HSC-iM. HSC-iM progeny, including monocytes and dendritic cells, retain transcriptional features of HSC-iM including inflammatory signaling pathways, while HSC-I do not. From a biological viewpoint, this is a step forward to addressing a conundrum in the field: many of the mature myeloid cells that respond to inflammation are extremely short-lived raising the question of how long-term inflammation can be sustained and drive the deleterious phenotypes associated with CH. Our data suggests that HSC-iM continuously produces inflammatory primed myeloid cells nominating HSC-iM as a novel cellular player in exacerbating systemic inflammation. In sum, the discovery of HSC-iM implicates the importance of the inflammatory milieu and provides a framework to address whether HSC-iM emerges as a ‘cost’ of adaptation to inflammation that modifies wildtype HSC fitness to drive selective advantage of CH-bearing HSC clones. Citation Format: Stephanie Xie. Inflammatory memory and selective advantage in human clonal hematopoiesis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_2):Abstract nr SY02-01.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.