Comprehensive analysis of hematopoietic alternations with age shows a discrepancy of age-associated changes between peripheral blood and bone marrow

(A) Average frequency of myeloid cells (neutrophils, monocytes) and lymphoid cells (B cells, T cells, and NK cells) in PB at the age of 2-3 months (n = 6), 6 months (n = 6), 12-13 months (n = 6), 18 months (n = 6), 20-21 months (n = 5), ≥ 23 months (n = 16). Abbreviation: PB = Peripheral blood.

(B) Average frequency of immunophenotypically defined HSC and progenitor cells in BM of 2-3-month mice (n = 6), 6-month mice (n = 6), 12-13-month mice (n = 6), ≥ 23-month mice (n = 7).

(C) Age-associated changes of immunophenotypically defined HSC and myeloid differentiation components (CMP and myeloid cells in the PB). The ratio of aged to young frequency was calculated as (the fraction frequency at each aged mice (%)) / (the average fraction frequency at 2-3-month mice (%)). *P < 0.05. **P < 0.01. ***P < 0.001. Data and error bars represent means ± standard deviation.

The expansion of myeloid-biased clones was not observed in 2-year-old LT-HSCs after their transplantation.

(A) Hoxb5 reporter expression in bulk-HSC, MPP, Flk2+, and Lin-Sca1-c-Kit+ populations in the 2-year-old Hoxb5-tri-mCherry mice. Values indicate the percentage of mCherry+ cells ± standard deviation in each fraction (n = 3).

(B) Experimental design to assess the long-term reconstitution ability of Hoxb5+ HSCs or Hoxb5- HSCs. Hoxb5+ HSCs and Hoxb5- HSCs were isolated from 2-year-old CD45.2 Hoxb5-tri-mCherry mice and were transplanted into lethally irradiated CD45.1 recipient mice with 2 × 105 supporting cells (Hoxb5+ HSCs, n = 13; Hoxb5- HSCs, n = 10). For secondary transplants, 1 × 107 whole BM cells were transferred from primary recipient mice. Abbreviations: PB = Peripheral blood, RT = Radiation therapy.

(C) Percentage chimerism at 16 weeks after receiving ten aged Hoxb5- HSCs or ten aged Hoxb5+ HSCs. Each column represents an individual mouse.

(D) Percentage chimerism at 16 weeks after whole BM secondary transplantation. Donor whole BM cells for secondary transplantation were taken from mice denoted by * in Figure 2C.

(E) Kinetics of average donor chimerism in each PB fraction after primary transplantation.

(F) Kinetics of average donor chimerism after secondary transplantation.

(G) Kinetics of average donor myeloid output (myeloid proportion in donor cells) in LT-HSC recipient mice after primary and secondary transplantation. *P < 0.05. **P < 0.01.

***P < 0.001. Data and error bars represent means ± standard deviation.

Aged LT-HSCs show balanced hematopoiesis throughout life

(A) Experimental design for competitive co-transplantation assay using young LT-HSCs sorted from Hoxb5-tri-mCherry GFP mice and aged LT-HSCs sorted from Hoxb5-tri-mCherry mice. Ten CD45.2+ young LT-HSCs and ten CD45.2+ aged LT-HSCs were transplanted with 2 × 105 CD45.1+/CD45.2+ supporting cells into lethally irradiated CD45.1+ recipient mice (n = 8).

(B) Lineage output of young or aged LT-HSCs at 4, 8, 12, 16 weeks after transplantation. Each bar represents an individual mouse.

(C) Lineage output kinetics of young LT-HSCs or aged LT-HSCs at 4, 8, 12, 16 weeks post-transplant.

(D) Competitive analysis of young LT-HSCs vs. aged LT-HSCs lineage output at 4, 8, 12, 16 weeks post-transplant. Competitive ratio was calculated as the proportion of young LT-HSCs derived cells vs. aged LT-HSCs derived cells in each fraction. Abbreviations: Neut = Neutrophils, Mo = Monocytes, B = B cells, T = T cells, and NK = NK cells. Data and error bars represent means ± standard deviation. “n.d.” stands for “not detected.”

Myeloid-associated genes were not enriched in aged LT-HSCs compared to their young counterparts

(A) Experimental schematic for transcriptome analysis. LT-HSCs (n = 3), ST-HSCs (n = 3), and bulk-HSCs (n = 3) were sorted from young (2-3 months) or aged (23-25 months) Hoxb5-tri-mCherry mice, after which each RNA was harvested for RNA sequencing.

(B) Hierarchical clustering dendrogram of whole transcriptomes using Spearman distance and the Ward clustering algorithm.

(C) Violin plots showing normalized gene expression levels for each gene set in young and aged LT-HSCs, ST-HSCs, and bulk-HSCs. Expression values for each gene were standardized independently by applying Z score transformation.

(D, E) Venn diagram showing the overlap of genes between three myeloid signature gene sets, and lymphoid signature gene sets (A Sanjuan-Pla et al., 2013, Pronk et al., 2007, SM Chambers et al., 2007).

(F, G) Signature enrichment plots from GSEA analysis using defined myeloid and lymphoid signature gene sets that overlapped in the three gene sets. Values indicated on individual plots are the normalized Enrichment Score (NES) and q-value of enrichment.

The memory-type lymphocytes in the peripheral blood make it look as if ST-HSCs are lymphoid-biased HSCs

(A) Experimental design for assessing the lineage output of young LT-HSCs or ST-HSCs. Ten LT-HSCs, or ten ST-HSCs were isolated from 2-month-old CD45.2 Hoxb5-tri-mCherry GFP mice and were transplanted into lethally irradiated CD45.1 recipient mice with 2 × 105 supporting cells (LT-HSCs, n = 3; ST-HSCs, n = 4).

(B) Kinetics of average frequency of lymphoid cells (B cells, T cells, and NK cells) in donor fraction after LT-HSCs or ST-HSCs transplantation.

(C) Gating scheme to identify memory (Central and effector) T cells and naive T cells in the PB after excluding doublets, dead cells, and non-donor cells.

(D) Percentage of memory (Central and effector) T cells and naive T cells in donor CD4+ fraction 10 months after LT-HSC or ST-HSC transplantation.

(E) Gating scheme to identify donor cells in bulk-HSC fraction in bone marrow analysis.

(F) Donor chimerism in bulk-HSC fraction 12 months after LT-HSCs or ST-HSCs transplantation. *P < 0.05. ***P < 0.001. Data and error bars represent means ± standard deviation.

Hematopoiesis after transplantation inclined either toward myeloid or lymphoid cell production by artificially changing the ratio of LT-HSC/ST-HSC

(A) Experimental design for the transplantation of 2-3-month-old LT-HSCs and ST-HSCs in a 2:8 ratio (the same ratio as in young mice BM) or 5:5 ratio (the same ratio as in aged mice BM). Donor cells were transplanted with 2 × 105 CD45.1+/CD45.2+ supporting cells into lethally irradiated CD45.1+ recipient mice (2:8 ratio, n = 18; 5:5 ratio, n = 23).

(B) Donor lineage output of young LT-HSCs and ST-HSCs transplanted either in a 2:8 ratio or a 5:5 ratio at 4, 8, 12, 16 weeks post-transplant. Each bar represents an individual mouse.

(C) Kinetics of average lineage output of young LT-HSCs and ST-HSCs in a 2:8 ratio or a 5:5 ratio at 4, 8, 12, 16 weeks post-transplant.

(D) Frequency of myeloid cells in donor cell fraction. *P < 0.05. **P < 0.01. Error bars represent standard deviation. Data represent two independent experiments.

Age-associated physiological changes drive differentiation of LT-HSCs toward myeloid cells

(A) Experimental design for assessing the impact of age-associated physiological changes on differentiation of LT-HSCs. Ten GFP+ LT-HSCs sorted from young (2-3 months) Hoxb5-tri-mCherry GFP mice, were transplanted with 2 × 105 CD45.1+/CD45.2+ supporting cells into lethally irradiated young or aged recipient mice. We defined donor cells as GFP+ cells and supporting cells as CD45.1+/CD45.2+ cells.

(B) Survival rate of recipient mice in each group.

(C) Donor lineage output in young or aged recipient mice 11-12 weeks after transplanting young LT-HSCs (young recipient, n = 17; aged recipient, n = 10).

(D) Myeloid output (Frequency of donor myeloid cells in donor fraction) in young or aged recipient mice 11-12 weeks after transplantation.

(E) Kinetics of lineage output from donor LT-HSCs in young or aged recipient mice 4, 8, 11-12 weeks after transplantation.

(F) Average frequency of donor bulk-HSC and progenitor cells in donor c-Kit+ cells in BM (young recipient, n = 8; aged recipient, n = 8). BM samples were taken from mice denoted by * in Figure 7C.

(G) Representative immunofluorescence images of frozen spleen sections derived from young or aged recipient mice. Green: donor cells (GFP fluorescence); blue: DNA (DAPI); Scale bar: 200 μm.

(H) Frequency of donor cells in spleen B cells of young or aged recipient mice (young recipient, n = 8; aged recipient, n = 8). Spleens are taken from mice denoted by * in Figure 7C.

(I) Representative immunofluorescence images of frozen thymus sections derived from young or aged recipient mice. Green: donor cells (GFP fluorescence); blue: DNA (DAPI); Scale bar: 200 μm.

(J) Frequency of donor cells in thymus T cells of young or aged recipient mice (young recipient, n = 8; aged recipient, n = 8). Thymi are taken from mice denoted by * in Figure 7C. *P < 0.05. ***P < 0.001. Error bars represent standard deviation. Data represent two independent experiments.

Our new model: Self-renewal heterogeneity model. It has been thought that there were myeloid (My-) or lymphoid biased (Ly-) HSCs, and that clonal selection of My-HSCs caused age-associated myeloid biased hematopoiesis. However, in our model, LT-HSCs represent unbiased hematopoiesis throughout life. ST-HSCs lose their hematopoietic ability within a short period and memory-type lymphocytes remains in the PB after ST-HSC transplantation. These remaining memory-type lymphocytes make it look as if ST-HSCs are lymphoid-biased (The upper section). As a result, the age-associated relative decrease of ST-HSCs in bulk-HSC fraction causes myeloid biased hematopoiesis with age. Additionally, the blockage of lymphoid differentiation at spleen and thymus accelerates further myeloid biased hematopoiesis in aged mice (The lower section).