Inflammageing of Hematopoietic Stem Cells Is Driven By IL-1

Introduction: Lifelong blood production is sustained through a stepwise differentiation program by a limited number of self-renewing hematopoietic stem cells (HSCs) in the bone marrow (BM). Hematopoietic cell development is tightly controlled by both cell-intrinsic and -extrinsic factors and its dysregulation can lead to aplasia or neoplasia. During ageing, HSCs increase in number, reduce self-renewal capacity on a per cell basis, skew towards myeloid differentiation, and show less efficient bone marrow (BM)-homing ability. We here evaluated how and to what extend HSC-extrinsic factors determine HSC behaviour during aging.

Methods, Results and Discussion: To screen for aging-associated extrinsic factors, we performed antibody based protein arrays and transcriptome analysis with total BM of young (6-8 week old) versus aged (2 year old) animals. This demonstrated that RANTES, MIP-2, IL-1α and IL-1β are significantly upregulated in aged BM at both the protein as well as the RNA level. ELISA of peripheral blood (PB) serum and BM lysates indicated that IL-1α and IL-1β are locally increased and produced in BM, but are not significantly increased in PB serum. Further, qPCR of various BM cell types of hematopoietic (myeloid, lymphoid and progenitor cells) and non-hematopoietic/stromal origin indicated that multiple cell types upregulate Il1α and Il1β, with highest increase being derived from myeloid hematopoietic cells. This raised the possibility that elevated IL-1 is a result of an inflammatory response to circulating pathogen-derived compounds, possibly of bacterial origin. Indeed, we previously demonstrated that steady-state levels of granulopoiesis in young steady-state mice depend on heat-resistant microbiota-derived compounds (M.Balmer et al., The Journal of Immunology 2014). To prospectively test the role of IL-1-induced signalling and the microbiome during aging, we investigated the ageing-associated phenotype of HSCs in young and aged IL1RIKO mice and in young and aged germ-free mice. Both IL1RIKO and germ-free aged mice had lower counts of platelets and neutrophils in PB, and lower frequency of LT-HSCs (LKS Flt3-CD34-CD48-CD150+) in BM, compared to aged WT mice. Moreover, aged IL1RIKO LT-HSCs showed improved lymphoid lineage repopulation upon transplantation into lethally irradiated WT mice, compared to LT-HSCs of aged WT mice that demonstrated the known myeloid-biased lineage output. Interestingly, LT-HSCs from aged germ-free mice also demonstrated lymphoid-biased lineage differentiation as observed from young mice. In line with this finding, no difference was observed in IL-1α and IL-1β protein concentrations in BM lysates from young and aged germ-free mice. To test if IL-1 increase in aged steady-state mouse BM is indeed dependent on ligation of pattern recognition receptors and consecutive signalling, we analysed MyD88 and Trif KO mice, respectively. Both aged KO mice showed compared to WT mice reduced BM IL-1 levels and a reduced ageing-phenotype of HSCs, with the most profound difference in Trif KO mice. Interestingly, this correlates with our previous finding on pathogen-Induced TLR4-TRIF innate immune signaling in HSCs, inducing reduced competitive fitness (Takizawa et al., Cell Stem Cell 2017).

Conclusions: Our data demonstrate that ageing associated phenotype and myeloid-biased differentiation of HSCs is a result of signals derived from the microbiome, that act through increased IL-1 signalling, locally in BM.