Infection Is a Driver of Dnmt3a-Mutant Clonal Hematopoiesis

Clonal hematopoiesis of indeterminate potential (CHIP) occurs increasingly with age and is associated with an increase in all-cause mortality. Mutations in DNMT3A, the gene most commonly altered in CHIP, are nearly universal in individuals by the age of 50 but not everyone develops CHIP. This dichotomy suggests that environmental factors likely play a major role in CHIP emergence. However, the environmental drivers of DNMT3A-mutant CHIP have not been determined.

Infection has been reported to be epidemiologically linked to a variety of cancers, including acute myelogenous leukemia. In prior work, we demonstrated that systemic chronic infection depletes the hematopoietic stem cell (HSC) pool as a result of excessive differentiation. Since prior studies showed that Dnmt3a-mutant HSCs demonstrate defective differentiation and enhanced self-renewal, we hypothesized that infection may provide an environmental selection pressure to favor the expansion of Dnmt3a-mutant HSCs over their wild-type (WT) counterparts.

First, to confirm whether Dnmt3a mutations affect hematologic responses to infection, we infected WT and Mx1-Cre Dnmt3a^f/f mice, in which Dnmt3a-deletion was induced by PIPC treatment, with M. avium. After four weeks of infection, we found few differences in the number of HSCs, Ki67+ levels, or apoptosis in the blood and bone marrow of WT and Dnmt3a^-/- mice. However, Dnmt3a-deficient mice became anemic upon M. avium infection, suggesting a defect in differentiation. To examine the differentiation capacity of DNMT3A-mutant HSCs in response to inflammatory stress, we conducted in vitro differentiation assays with human CD34+ hematopoietic progenitors. Human CD34+ cells in which DNMT3A was disrupted by CRISPR gene editing did not differentiate normally in response to treatment with interferon gamma (IFNy), showing little change in either their HSCs or neutrophil numbers after culture with IFNy compared to control CD34+ cells. These data demonstrate a striking defect in inflammation-induced differentiation in DNMT3A-mutant murine and human HSCs.

Next, we utilized a mathematical model to predict the relative prevalence of different populations of HSCs when exposed to the same chronic infectious stress. Specifically, we modeled the behavior of a "minor population" of Dnmt3a-mutant HSCs, compared to a "major population" of WT HSCs. This modeling exercise demonstrated that impaired HSC differentiation in Dnmt3a-mutants during inflammatory stress is sufficient to enable the minor Dnmt3a-mutant HSCs to quickly overtake the major WT population. To test this prediction, we created mosaic mice by transplanting a 1:6 mixture of Dnmt3a^-/- and WT competitive WBM into lethally irradiated WT mice. Two months after transplantation, mosaic mice were systemically infected with M. avium, and the infection was allowed to continue for 2 months prior to examining the bone marrow by flow cytometry. In line with the modeling prediction, there was a significant expansion of HSCs only among the infected mosaics containing Dnmt3a-mutant marrow. Expansion of Dnmt3a-mutant cells was also detectible in downstream populations (i.e., multipotent progenitors (MPPs)). Dnmt3a-mutant HSCs expansion was similar upon infection of mosaics created with Vav-Cre Dnmt3a^f/f donors, where no PIPC-injections are required.

To assess the role of IFNy signaling in Dnmt3a-mutant clonal expansion, we infected mosaic mice containing a 1:6 ratio of Dnmt3a^-/- IFNyR1^-/- DKO to WT HSCs. Strikingly, no clonal expansion was evident among infected mosaics with the DKO minor population, indicating a requirement for IFNy signaling for Dnmt3a-mutant clonal expansion. Next, we conducted RNA-seq to compare the expanded Dnmt3a-mutant population versus an unexpanded Dnmt3a-mutant population. These data showed that, in stark contrast to WT HSCs, the differentiation factor Batf2 was not induced in Dnmt3a-mutant HSCs upon M. avium infection. By using Batf2^-/- mice, we found that Batf2-mutant HSCs are resistant to depletion during chronic infection, thus phenocopying Dnmt3a-mutant HSCs

In summary, we present the first controlled biological experiment to demonstrate that an environmental factor - infection - can promote Dnmt3a-mutant CHIP. Our studies indicate that a defect in differentiation is sufficient to confer a survival advantage to Dnmt3a-mutant clones during infectious conditions.