Fig 1: TAC suppresses MSC HBEGF-to-HSC signaling.a UMAP plot of Lepr+ MSCs from male mice (n = 3 control; n = 8 TAC) 4 weeks after surgery. b Marker gene expression in each cluster. Additional marker genes are shown in Supplementary Fig. 7. c Heatmap showing expression of known MSC subpopulation markers. d UMAP plot of LSKs from the same male mice used for MSC scRNA-seq. e NicheNet analysis of inferred ligand-receptor interactions between MSCs (ligand producing) and LSKs (receptor-expressing). f Differential gene expression analysis of MSCs. Upregulated and downregulated genes (FDR < 0.05) are shown in red and blue, respectively. Cytokine and growth factor genes are highlighted in dark red, and those with FDR < 0.05 are labeled. Differential expression was assessed using edgeR with two-sided tests, and P values were adjusted for multiple comparisons using the Benjamini–Hochberg method (FDR < 0.05). g PROGENy pathway enrichment analysis. Positive pathway scores indicate increased pathway activity in post-TAC HSCs compared with controls. Top 8 pathways are shown. h Proportions of cells within each MSC cluster in control (CTRL) and TAC groups. Differences in proportions were analyzed using beta regression, and P values were adjusted using the Benjamini–Hochberg method. Adjusted P values are shown only for significantly different comparisons. n = 3 for control male mice, n = 8 for TAC male mice.
Fig 2: Human HBEGF+ MSCs and BM adiposity in HF.a Clustering of human LEPR+ MSCs. b Expression of the marker genes in human LEPR+ MSCs. Violin plot of the full marker genes corresponding to Supplementary Fig. 7b is shown in Supplementary Fig. 20a. c Representative axial CT images of the 7th thoracic vertebra (T7) from a healthy control subject (left) and a patient with HF (right). Bone marrow regions with CT attenuation below 0 HU, corresponding to fat, are shown in red, and regions with attenuation at or above 0 HU are shown in green. d Percentage areas of CT values below 0 within the range from −200 to 1000 in CT values on coronary CT images of the thoracic vertebral BM in HF patients and age- and sex-matched healthy controls. Two-tailed unpaired Student’s t-test. n = 30 for control, n = 15 for heart failure. Box-and-whisker plots show the median, the 25th and 75th percentiles (box bounds), and the minimum and maximum values (whiskers). e Schematic model summarizing the proposed mechanism, comparing the healthy state (left) with heart failure (right). Under homeostasis, Notch–HES1 signaling in Lepr⁺ mesenchymal stromal cells of the bone marrow niche maintains the HBEGF-expressing subpopulation and restrains the adipogenic regulator PPARγ, supporting balanced hematopoiesis and the generation of CCR2ˡᵒ macrophages that help maintain tissue homeostasis and protect multiple organs, including the heart, from stress. In heart failure, cardiac stress suppresses Notch–HES1 signaling, leading to loss of the HBEGF-expressing subpopulation and de-repression of PPARγ with adipogenic conversion of the stromal cells. The resulting decrease in HBEGF–EGFR signaling imprints a lasting memory on hematopoietic stem cells (stress-imprinted HSCs), establishing central innate immune memory that biases hematopoiesis toward CCR2ʰⁱ proinflammatory macrophages and drives long-lasting multimorbidity. A corresponding HBEGF⁺ mesenchymal stromal cell subpopulation is present in human bone marrow, and heart failure is associated with increased bone marrow adiposity in humans, indicating that this mechanism is conserved across species. Created in BioRender. F, K. (2026) https://BioRender.com/g0580kj.
Fig 3: Hbegf downregulation in MSCs promotes myeloid skewing and cardiac dysfunction.a Expression of Hbegf in lentivirally generated NIH-3T3 stable cell lines overexpressing murine Hbegf and Cre recombinase following transfection with the AAV-FLEX–shHbegf transgene plasmid. n = 6 per group. Exact P value: P < 0.0001 (two-tailed unpaired Student’s t-test; t = 10.95, df = 10). b Schematic of the experimental workflow. Lepr-Cre male mice received intravenous AAV administration followed by TAC. c Fractions of primitive HSPC subpopulations. Exact P values: LT-HSC, P = 0.0220; MPP1, P = 0.8770; MPP2, P = 0.7289; MPP3, P = 0.0024; MPP4, P = 0.0006 (two-tailed unpaired Student’s t-tests after logit transformation; n = 7 scramble, n = 6 shHbegf). d Representative flow cytometry plots showing LSK populations. e Numbers of granulocyte/monocyte progenitor (GMP) per femur. f Numbers of CCR2hi and CCR2lo cardiac macrophages normalized to heart weight (cells per milligram of heart tissue). g Representative cardiac sections stained with Picrosirius Red/Fast Green; collagen fibers appear red, and the non-collagen components green. h Fibrosis area in cardiac tissue. i Left ventricular ejection fractions measured by echocardiography. j Ventricular Nppa and Nppb mRNA expression. Data for Fig. 3c, d, g, h, j were obtained from the same experiment (n = 7 scramble, n = 6 shHbegf). Data for Fig. 3e, f, i were obtained from independent experiments (n = 5 scramble, n = 8 shHbegf). Two-tailed unpaired Student’s t-tests were used for Fig. 3a, e, f, i, j and two-tailed unpaired Student’s t-tests after logit transformation were used for Fig. 3c, h. Figure 3b was created in BioRender. F, K. (2026) https://BioRender.com/4ux5x8g.
Supplier Page from BioLegend for Recombinant Mouse HB-EGF (carrier-free)