Abstract
Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening syndrome characterized by an overwhelming activation of diverse immune cells. Chemotherapy based regimen as a first-line therapy for HLH has a substantial risk of mortality. Thus, exploring a less toxic therapy is urgently needed. Accumulating evidence suggests that hypoxia-inducible factors (HIFs) play an important role in the regulation of the immune system. Hypoxia, as well as inflammation, infectious microorganisms, and cancer, triggers HIF expression and stabilization in immune cells. HIFs activation enhances phagocyte capacity, drives T cell differentiation and increases cytotoxic activity. HIFs also regulate the cellular metabolism of immune cells to dictate their fate, development, and function. However, little is known about HIFs' function in HLH pathogenesis.
To determine whether activation of HIF-1α/HIF-1b in hematopoietic cells is sufficient to induce HLH phenotypes, we generated transgenic mice with doxycycline-inducible HIF-1α/HIF-1b expression. Using a Vav1-Cre/Rosa26-LSL-rtTA driver, a doxycycline-inducible expression of both a stable and constitutively active human HIF-1α triple-point-mutation (TPM) and wild-type HIF-1β was achieved. All the inducible HIF-1α/HIF-1b mice developed HLH phenotypes in C57/BL6background and died within three weeks. They quickly developed severe anemia, thrombocytopenia, multi-organ failure, splenomegaly, and hemophagocytosis. Total and type-1 polarized macrophages were significantly increased in the bone marrow (BM) and spleen (SP) of HLH mice compared to the controls. To determine the phagocytic activity of the type-1 polarized macrophages, we generated type-1 and type-2 polarized macrophages from BM mononuclear cells in vitro with M-CSF/IFNγ or M-CSF/IL-4 respectively. We found that type 1 rather than type 2 macrophages engulfed erythroblasts in an in vitro co-culture assay.
IFN-γ signaling is critical for Type-1 macrophage polarization. We generated IFN-γ receptor-/- /Vav1-Cre/LSL/TPM mice and found that knockout of IFN-γ receptor completely blocked the macrophage activation and HLH development. Interestingly, the serum IFN-γ level was only slightly upregulated in Vav1-Cre/LSL/TPM mice, suggesting that IFN-γ locally, but not systemically, exerts its function in our HLH model. However IFN-γ expression in NK cells and CD8+ T cells did not increase. Thus, the source of the IFN-γ for macrophage polarization is still unclear.
In primary HLH, defective cytotoxic function in NK and T cells is important for HLH development. We found that the cell surface CD107a (degranulation) and NK46p (activating receptor) expression didn't change in these HLH mice. Interestingly, the absolute numbers of total NK cells and DX5+ mature cytotoxic NK cells were significantly reduced in the PB, SP, and BM from HLH mice. However, the frequencies of CD8+ T cells, CD4+ T cells, Th17 cells, and total T cells did not change in HLH mice.
In order to dissect the contribution of individual immune cell subpopulations to the HLH pathogenesis, diverse lineage specific Cre transgenic alleles were used. Induction of TPM allele in myeloid cells (LysM-Cre), mature T cells (dLck-Cre), NK cells (NCR-Cre), or multi-lineages (LysM-Cre/dLck-Cre) did not cause HLH. Interestingly, induction of TPM allele in the mononuclear phagocyte system (monocyte, macrophage, and dendritic cells) with Cx3cr1-Cre could give rise to HLH phenotypes, as well as macrophage activation and reduced NK cell numbers, which are similar to the Vav1-Cre model. These results indicate that activation of HIF-1α/HIF-1b in the mononuclear phagocyte system is sufficient to polarize macrophages and induce HLH phenotype. The reduced NK cell numbers may be secondary to macrophage/dendritic cell activation in this HLH model.
In conclusion, we found that; 1) induction of stable and constitutively active form of HIF-1α/HIF-1β expression in hematopoietic cells, especially in the mononuclear phagocyte system, polarizes macrophages and causes HLH, 2) IFN-γ signaling is required for HLH development and macrophages/dendritic cells are critical immune cell populations in this model, 3) Reduced NK cell numbers may be a secondary phenomenon in this model. This new HLH model recapitulates the features of secondary HLH in human, provides a unique model for dissecting the detail mechanisms, and helps in testing new therapies for sHLH.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.