Global remodeling of the vascular stem cell niche in bone marrow of diabetic patients: implication of the microRNA-155/FOXO3a signaling pathway

Abstract

Rationale: The impact of diabetes mellitus on bone marrow (BM) structure is incompletely understood.

Objective: Investigate the effect of type-2 diabetes mellitus (T2DM) on BM microvascular and hematopoietic cell composition in patients without vascular complications.

Methods and results: Bone samples were obtained from T2DM patients and nondiabetic controls (C) during hip replacement surgery and from T2DM patients undergoing amputation for critical limb ischemia. BM composition was assessed by histomorphometry, immunostaining, and flow cytometry. Expressional studies were performed on CD34(pos) immunosorted BM progenitor cells (PCs). Diabetes mellitus causes a reduction of hematopoietic tissue, fat deposition, and microvascular rarefaction, especially when associated with critical limb ischemia. Immunohistochemistry documented increased apoptosis and reduced abundance of CD34(pos)-PCs in diabetic groups. Likewise, flow cytometry showed scarcity of BM PCs in T2DM and T2DM+critical limb ischemia compared with C, but similar levels of mature hematopoietic cells. Activation of apoptosis in CD34(pos)-PCs was associated with upregulation and nuclear localization of the proapoptotic factor FOXO3a and induction of FOXO3a targets, p21 and p27(kip1). Moreover, microRNA-155, which regulates cell survival through inhibition of FOXO3a, was downregulated in diabetic CD34(pos)-PCs and inversely correlated with FOXO3a levels. The effect of diabetes mellitus on anatomic and molecular end points was confirmed when considering background covariates. Furthermore, exposure of healthy CD34(pos)-PCs to high glucose reproduced the transcriptional changes induced by diabetes mellitus, with this effect being reversed by forced expression of microRNA-155.

Conclusions: We provide new anatomic and molecular evidence for the damaging effect of diabetes mellitus on human BM, comprising microvascular rarefaction and shortage of PCs attributable to activation of proapoptotic pathway.

Figure 1. Diabetes mellitus induces bone marrow (BM) remodeling and vascular rarefaction

A and B, Histomorphometric analysis shows replacement of marrow with fat and bone rarefaction. A, Representative microphotograph of hematoxylin and eosin-stained BM sections: (i) Control, (ii) type-2 diabetes mellitus (T2DM) patient, and (iii) T2DM patient with critical limb ischemia (CLI). B, Bar graph showing average data of marrow fractions. C, Representative microphotograph of human BM showing CD31-positive vascular structures. Arrows indicate the type of vessel. D, Confocal microscopy photographs of human BM showing vascular structures stained with the endothelial marker von Willebrand factor (VWF) and the vascular smooth muscle marker α-smooth muscle actin (αSMA). Nuclei are stained blue by 4′,6-diamidino-2-phenylindole (DAPI). E, Bar graph showing average data of microvascular density. *P<0.05 and **P<0.01 versus Controls, §P<0.05 versus T2DM. Controls, n=10; T2DM, n=7; T2DM+CLI, n=10.

Figure 2. Reduced abundance of CD34^pos cells in bone marrow (BM) of type-2 diabetes mellitus (T2DM) patients

A, Representative confocal microscopy photographs of human BM showing the presence of CD45^pos (green arrow), CD45^posCD34^pos (pink arrowhead) (i), and CD45^negCD34^pos cells (red arrow) (ii). Nuclei are stained blue with 4′,6-diamidino-2-phenylindole (DAPI). Bar graphs showing the average density of CD45^posCD34^pos cells (iii, median and 5%–95% distribution) and CD34^posCD45^neg cells (iv, mean±SEM). B, Increased abundance of apoptotic mononuclear cells and CD34^pos cells in BM from diabetic patients. Representative microphotographs of fluorescent terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) positive cells (i) and bar graphs showing average data (ii and iii). *P<0.05, **P<0.01 and ***P<0.001 versus controls, §P<0.05 and §§P<0.01 versus T2DM. Controls, n=10; T2DM, n=7; T2D+critical limb ischemia (CLI), n=8.

Figure 3. Flow cytometry characterization of hematopoietic cells in bone marrow (BM) and peripheral blood (PB) of type-2 diabetes mellitus (T2DM) patients

A, Gating strategy of multicolor flow cytometry. B to D, Bar graphs showing the abundance of CD45^dimCD34^pos (Bi and Bii), CD45^dimCD133^posCD34^pos cells (Ci and Cii) and CD45^posCD133^posCD34^neg cells (Di and Dii). *P<0.05, **P<0.01 and ***P<0.001 versus controls. Controls, n=8 to 14; T2DM, n=7 to 9; T2DM+critical limb ischemia (CLI), n=7 to 11.

Figure 4. Flow cytometry characterization of bone marrow (BM) and peripheral blood (PB) endothelial progenitors

A, Gating strategy of CD45^dimCD34^posKDR^pos mononuclear cells (MNCs) (i), and bar graphs showing the abundance of this population in BM (ii) and PB (iii). B, Gating strategy of CD34^posCD14^posCD45^dimKDR^posCXCR4^pos cells (i), and bar graphs showing the abundance of this population in BM (ii) and PB (iii). *P<0.05 versus controls. Controls, n=8 to 11; type-2 diabetes mellitus (T2DM), n=6 to 9; T2DM+critical limb ischemia (CLI), n=7 to 10.

Figure 5. Flow cytometry characterization of lineage committed hematopoietic cells and endothelial cells (ECs)

A, Gating strategy for identification of B-lymphocytes, T-lymphocytes, and natural killer (NK) cells. B to D, Bar graphs showing the abundance of B-lymphocytes (B), T-lymphocytes (C) and NK cells (D) in bone marrow (BM) (i) and peripheral blood (PB) (ii). E, Gating strategy for identification of BM ECs (i) and bar graph showing average values (ii) *P<0.05 and **P<0.01 versus controls, §P<0.05 versus type-2 diabetes mellitus (T2DM). Controls, n=9 to 16; T2DM, n=7 to 9; T2DM+critical limb ischemia (CLI), n=7 to 8.

Figure 6. Diabetes mellitus-induced expressional changes in bone marrow (BM) CD34^pos cells

A and B, Bar graphs showing mRNA levels of microRNA (miR)-155 (A) and its target FOXO3a (B). Controls, n=10; type-2 diabetes mellitus (T2DM), n=7; T2DM+critical limb ischemia (CLI), n=6. C, Graph showing the inverse correlation between miR-155 and FOXO3a mRNA levels in CD34^pos cells. D, Representative microphotographs (i) and bar graph (ii) showing the in situ expression of FOXO3a in BM cells. Confocal microphotographs showing FOXO3a (red) localization in the cytoplasm (iii) and nucleus (iv) of CD34^pos cells (green). Nuclei are stained blue with 4′,6-diamidino-2-phenylindole (DAPI). n=5 per group. E and F, Bar graphs showing mRNA levels of CDKN1A/p21 (E) and CDKN1B/p27^kip1 (F). Controls, n=10; T2DM, n=7; T2DM+CLI, n=6. *P<0.05 and **P<0.01 versus controls.

Figure 7. MicroRNA (miR)-155 overexpression in bone marrow (BM) CD34^pos cells reverts high glucose (HG)-induced expressional changes and reduces the generation of myeloid and erythroid colonies

A, Bar graphs showing the effect of HG and pre–miR-155 or control scramble (SCR) transfection on CD34^pos cell number. *P<0.05 versus SCR HG. Cells were cultured either in normal glucose (NG) (5 mmol/L D-glucose, NG) or HG (25 mmol/L D-glucose, HG) for 48 hours. B, Bar graphs showing mRNA levels of (i) FOXO3a, (ii) CDKN1A/p21, and (iii) CDKN1B/p27^kip1 after pre–miR-155 or control SCR transfection. N=5 healthy donors per group assayed in duplicate. C, Colony forming unit (CFU) assay of CD34^pos cells. (i) Representative images of CFU-granulocytes, erythroid, macrophage, megakaryocyte (CFU-GEMM), CFU-granulocyte, macrophage (CFU-GM), and CFU-erythroid (CFU-E) colonies. ii, Immunocytochemical characterization of CFU-isolated cells by staining for myeloperoxidase (MPO), a marker for granulocytes, CD68, a marker of macrophages, and glyphorin, a marker of erythrocytes. Arrows in CFU-GEMM point at MPO^pos granulocytes. Nuclei are stained blue with hematoxylin. iii, Bar graph showing the effect of miR-155 overexpression on number (top) and area (bottom) covered by colonies. N=5 healthy donors per group. *P<0.05 and **P<0.01 versus SCR HG, §P<0.05 and §§P<0.01 versus SCR NG. Data represent means±SEM.