Chronic Infection Depletes Hematopoietic Stem Cells through Stress-Induced Terminal Differentiation

Chronic infections including tuberculosis, hepatitis C, and HIV are estimated to affect over a third of the world's population and are associated with significant health implications including bone marrow suppression and an increased risk for cancer (Jian et al., 2014; Ramos-Casals et al., 2003; Scadden et al., 1989). Pancytopenia, a suppression of blood counts across multiple lineages, can affect as many as 12% of people with miliary tuberculosis and increases risk of death from the infection (Achi et al., 2013). However little is known about the mechanisms by which infections affect hematopoietic stem cells (HSCs). Here we show that chronic infection depletes HSCs and we identify terminal differentiation as the major route of HSC loss. Furthermore, we define Batf2 as a potent new mediator of inflammation-induced differentiation.

Using an established model of chronic infection, we conducted repeated monthly infection of C57Bl/6 WT mice with 2 x 10⁶ cfu Mycobacterium avium, generating a sustained chronic IFNg response. Mice became pancytopenic after 4-6 months of repeated infection. The number of HSCs in the bone marrow was depleted to just 5% of the starting number by 4 months following initial infection, without evidence of extensive myelofibrosis. Poor engraftment upon transplant of whole bone marrow (WBM) from mice repeatedly infected with M. avium confirmed that functional HSCs were lost. When WBM from Ifngr1-deficient mice infected with M. aviumwas transplanted, a similar engraftment defect was not seen, suggesting that HSC loss is IFNg-dependent.

In addition to loss of HSCs, secondary HSC transplants revealed a defect in HSC self-renewal capacity following repeated M. aviuminfection. Mathematical modeling demonstrated that the rate of HSC loss after chronic infection must increase by 57% from steady state to account for the observed decay in total HSCs.

In order to define the mechanism of HSC loss during infection, we conducted mobilization and apoptosis assays after infection or IFNg treatment but found no evidence of increased displacement or death. Next we performed RNAseq profiling of HSCs from infected and control animals. GSEA analysis reflected increased myeloid differentiation rather than apoptosis during infection, consistent with our previous findings. We confirmed that IFNg treatment alone promotes myeloid differentiation of human CD34+ progenitors in vitro, with a reciprocal decrease in the persistence of CD34+CD38- HSPCs. Out of 4 transcription factors among the 151 differentially regulated genes identified by RNAseq, Batf2 stood out as one whose role in HSC biology has not been studied. We confirmed that Batf2 is induced in murine HSCs during infection, and that BATF2 is upregulated in human CD34+ hematopoietic progenitor cells following IFNg treatment. Consistent with a role in differentiation, retroviral overexpression of Batf2 in Sca1+ murine progenitor cells resulted in increased myeloid production upon transplant. Further, knockout of BATF2 in human CD34+ progenitors using CRISPR-Cas9 gene editing resulted in impaired myeloid differentiation in response to IFNg in vitro.

Our studies demonstrate that chronic infection depletes the HSC pool by promoting HSC differentiation via an IFNg-dependent mechanism, and we identify the transcription factor Batf2 as a key player in infection-induced myeloid differentiation. These findings may provide a therapeutic opportunity to protect or restore hematopoiesis in patients with aplastic anemia, hemophagocytic histiocytosis (also associated with high IFNg levels), and chronic infections.