Background: While inflammation is necessary to fight infection and repair damaged tissue, excessive inflammation can cause bone marrow suppression and promote cancer. In an extreme example, high levels of the inflammatory cytokine interferon gamma (IFNg) deplete hematopoietic stem cells (HSCs), resulting in aplastic anemia. Patients with this dangerous disease are pancytopenic and therefore at high risk of death from infection. Pancytopenia also occurs to a lesser extent in other inflammatory conditions such as chronic infections (tuberculosis, HIV), and autoimmune diseases (hemophagocytic histiocytosis). However, the mechanism by which HSCs are damaged by IFNg remains poorly understood. We used a mouse model of Mycobacterium avium infection to study the effects of sustained IFNg exposure on primitive hematopoiesis. In prior work, we found, surprisingly, that IFNg promotes division of quiescent HSCs. We hypothesized that cell division might lead to loss of HSCs through terminal differentiation, displacement, or activation of p53-dependent apoptosis pathways.

Objective: We sought to determine whether prolonged IFNg stimulation would lead experimentally to exhaustion of the HSC compartment, and to determine the mechanism of inflammation-mediated HSC loss.

Methods: We conducted repeated monthly infection of C57Bl/6 WT mice with 2 x 106 cfu M. avium, thereby generating a sustained chronic IFNg response. We characterized the blood and bone marrow of treated mice by histology, flow cytometry, colony forming assays, and bone marrow transplant.

Results: Mice infected with M. avium became anemic and leukopenic after 6 months of repeated infection. High IFNg levels were sustained in the mice, with evidence of IFNg production by T cells and NK cells in the bone marrow. The number of committed hematopoietic progenitors gradually decreased and HSCs were depleted in the bone marrow by four months following initial infection, without evidence of extensive myelofibrosis. The marrow was hypercellular with a significant increase in granulocytes. Meanwhile, the myeloid differentiation capacity of the marrow was reduced, consistent with terminal differentiation of myeloid-biased HSCs, as we have previously described. Despite an overall reduction in HSC number, the HSCs that remained in chronically infected animals mostly retained their self-renewal potential, with subtle self-renewal defects evident only after two rounds of transplantation. Homing of HSCs from infected animals was not impaired, but ex vivo culture and apoptosis assays indicated that HSCs from chronically infected animals had reduced colony forming ability and were more prone to cell death upon secondary stress. These findings were recapitulated by introduction of recombinant IFNg alone. RNAseq profiling of HSCs from infected and control animals reflected increased proliferation and differentiation during infection, consistent with the above findings.

Conclusions: We have established a novel mouse model of bone marrow failure related to chronic IFNg stimulation. We demonstrate that chronic infection can deplete the HSC pool by promoting HSC differentiation and lowering the threshold for apoptosis. These mechanisms may drive marrow suppression in patients with aplastic anemia, hemophagocytic histiocytosis (also associated with high IFNg levels), and patients with marrow failure associated with chronic infection. Furthermore, since a reduction in HSC number results in depletion of clonal heterogeneity, our findings have significant implications regarding the mechanism by which chronic inflammation can contribute to the emergence of clonal hematopoiesis and hematologic malignancies with age.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.