Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.
Advertisement
Communications Biology volume 5, Article number: 1145 (2022)
110
2
Metrics details
Diabetes affects select organs such as the eyes, kidney, heart, and brain. Our recent studies show that diabetes also enhances adipogenesis in the bone marrow and reduces the number of marrow-resident vascular regenerative stem cells. In the current study, we have performed a detailed spatio-temporal examination to identify the early changes that are induced by diabetes in the bone marrow. Here we show that short-term diabetes causes structural and molecular changes in the marrow, including enhanced adipogenesis in tibiae of mice, prior to stem cell depletion. This enhanced adipogenesis was associated with suppressed transforming growth factor-beta (TGFB) signaling. Using human bone marrow-derived mesenchymal progenitor cells, we show that TGFB pathway suppresses adipogenic differentiation through TGFB-activated kinase 1 (TAK1). These findings may inform the development of novel therapeutic targets for patients with diabetes to restore regenerative stem cell function.
The global prevalence of diabetes has been steadily increasing worldwide and is projected to affect 10.4% of all adults by 20401. Diabetes poses a substantial burden on the health care system due to the challenges associated with managing various secondary complications. These complications include retinopathy, cardiomyopathy, nephropathy, and atherosclerosis. More recently, human diabetes and experimental models of the disease have shown that the bone marrow may also be a target organ of diabetic complications2,3,4. Both enhanced bone marrow adiposity and skeletal fragility have been well documented5,6,7. Diabetes has been shown to cause numerous other abnormalities8 in the bone marrow, including small vessel rarefaction5, reduced hematopoietic stem cell retention and quiescence factor expression9,10, myelopoiesis9,11, and neuropathy12. Some of these abnormalities, such as enhanced myelopoiesis/increased monocyte generation13, blunted hematopoietic stem cell mobilization14, and lower bone marrow-derived endothelial colony forming cells15,16,17,18 may increase the risk of cardiovascular disease in diabetes. Although the pathogenesis of these bone marrow abnormalities is slowly being uncovered, evidence exists that suggests enhanced bone marrow adipogenesis leads to hematopoietic and non-hematopoietic stem cell deficits. For example, we and others have observed depletion of non-hematopoietic bone marrow stem cells coupled to enhanced marrow adiposity6,19. In addition, examination of rodent bones with varying levels of adipocytes, and bones from A-ZIP/F1 fatless mice show that adipocytes in the bone marrow negatively regulate hematopoietic stem cells20. Importantly, A-ZIP/F1 mice—which show polyuria, polydipsia, and hyperglycemia21—illustrate that it is not simply the diabetic milieu but the presence of adipocytes that reduces hematopoietic stem cells in bones. Collectively, depletion of bone marrow regenerative cells, likely driven by enhanced adipogenesis, may underlie the development of other secondary vascular complications of diabetes. Therefore, it is imperative that we understand how diabetes enhances adipogenesis and leads to the depletion of marrow-resident stem cells.
Identifying mechanisms that cause enhanced bone marrow adipogenesis in diabetes and when these alterations take place, in relation to other complications, requires an experimental model that recapitulates the human disease. Streptozotocin (STZ)-induced diabetic mice and rats are the most established and recommended by the Animal Models of Diabetic Complications Consortium (http://www.diacomp.org). Using this model, researchers have produced a precise timeline of pathological changes in target organs. For example, for diabetic retinopathy, retinal ganglion cells are reduced at 6 weeks22,23, and capillary basement membrane thickening is evident at 17 weeks of hyperglycemia24. Similarly, for diabetic nephropathy, glomerular hypertrophy, glomerular basement membrane thickening, and mesangial expansion take more than 6 months of diabetes25,26. For cardiac dysfunction, increased left ventricle fibrosis is evident at 8 weeks following STZ-induced diabetes27. Collectively, these studies provide a benchmark to identify bone marrow dysfunction in diabetes.
Since bone marrow adipocytes are considered negative regulators of marrow-resident stem cells, we hypothesized that enhanced marrow adiposity in diabetes will precede stem cell depletion. To test our hypothesis, we examined the effect of STZ-induced diabetes in bone tissues of mice. Our results show that 1 month of diabetes is sufficient to enhance marrow adiposity in tibiae of mice. This alteration was observed before deficits in bone hematopoietic fraction and the depletion of bone marrow-resident stem cells. Additionally, we show that STZ-induced hyperglycemia suppresses transforming growth factor-beta (TGFB) pathway in the bone marrow. In cultured primary bone marrow cells, we demonstrate that high levels of glucose suppress TGFB, and the addition of exogenous TGFB inhibits adipogenesis through the non-canonical TGFB-activated kinase 1 (TAK1) pathway. Our studies have provided fundamental insight into diabetic marrow disease and identified the potential underlying mechanisms.
With the goal of identifying early changes in the bone marrow of diabetic mice, we induced diabetes in 5-week-old C57BL/6 mice (corresponding to 15–20 human years). We selected this age to identify diabetes-specific adipogenesis. We know that bone marrow adiposity increases with age28. In C57BL/6 mice, age-related bone marrow adiposity is seen from 12 weeks to 56 weeks of age28. Therefore, we induced diabetes in 5-week-old mice with multiple low doses of streptozotocin (STZ) on 5 consecutive days (Fig. 1a). STZ-injected mice showed significantly lower body weights and higher blood glucose levels, 1 week after the last STZ injection (d0; Supplementary Fig. S1a, b) and after 1 month (Supplementary Fig. S1c, d). As expected, the cumulative body weight difference in the STZ-induced diabetic mice was significantly less compared with the control mice (Supplementary Fig. S1e). Histological analysis of H&E and insulin antibody-stained pancreata displayed significantly reduced size of the islets, and the number of insulin-immunoreactive areas in STZ-induced mice compared with the control mice (Supplementary Fig. S2).
a Experimental scheme for diabetic mouse study. Diabetes was induced in C57BL/6 mice with daily intraperitoneal injections of streptozotocin (STZ; 50 mg/kg) for 5 consecutive days. Non-diabetic control mice received an equal volume of citrate buffer. Blood glucose levels were checked 1 week after the last STZ injection to confirm hyperglycemia (d0). b Representative H&E-stained images of the mouse femur [scale bar = 100 μm]. Inserts showing higher magnification. Quantification of adipocyte area per bone area (c) and adipocyte number per bone area (d) in the femurs. Parameters were measured using MarrowQuant [Mean ± SD; n = 4 (5 for control in panel d); each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. e Representative images of the H&E-stained mouse tibia [scale bar = 100 μm]. Inserts showing higher magnification. Quantification of adipocyte area per bone area (f) and adipocyte number per bone area (g) in the tibiae [Mean ± SD; n = 7 control and 4 STZ in panels f and g; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. Immunofluorescence staining of mouse tibia for perilipin-1 (PLIN1; green). Sections were counterstained with DAPI (blue) [scale bar = 50 μm]. Figure showing shaft (h) and distal regions (i) of the tibia. j Quantification of PLIN1 intensity per area, as determined by ImageJ [Mean ± SD; n = 10 control and 9 STZ; two-tailed student’s t-test: *p < 0.05]. mRNA levels of osteogenesis-associated genes were measured in the tibiae (k, l) and femurs (m, n) of control and diabetic (STZ) mice [Data normalized to Actb and Gapdh; Mean ± SD; n = 5 control and 6 STZ in panel k, 4 control and 6 STZ in panel l, 5 control and 6 STZ in panel m; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05].
We know that, histologically, diabetic complications manifest as increased extracellular matrix deposition and basement membrane thickening29, cell loss30, infiltration of myeloid cells and inflammation in target organs31. These structural features become obvious in STZ-induced models at different times, but typically require 2-plus months of diabetes duration. Examination of H&E-stained sections of the heart, liver, kidney, and lung did not reveal any significant changes in STZ-induced diabetic mice compared to the non-diabetic mice at 1 month of disease duration (Supplementary Fig. S3a). Retina, heart, and kidney sections stained with Periodic Acid-Schiff (PAS) to detect changes in polysaccharide deposition also showed no alterations (Supplementary Fig. S3b). Next, we stained the tissues with Picro-Sirius Red to highlight collagen deposition and observed no significant fibrosis in the tissues (Supplementary Fig. S3c–f). These results show that 1 month of hyperglycemia is not sufficient to induce hallmark structural changes in well-known target organs. Analysis of retina tissues revealed a significant decrease in the thickness of the inner nuclear layer in the diabetic mice (Supplementary Fig. S4), whereas the other retinal layers did not change. This observation is consistent with a recent study in diabetic rats highlighting early structural changes in the retina at 1 month32.
Epididymal fat of mice showed a difference between the STZ-induced diabetic mice and the non-diabetic controls. Diabetes was associated with increased nuclei per tissue area (Supplementary Fig. S5a–c) and suppressed levels of adipogenesis-associated genes, including Pparg, Cepba, Fabp4, Lpl, and Adipoq (Supplementary Fig. S5d–h). Increased lipolysis33,34 in this model of type 1 diabetes may explain suppressed adipogenesis-related genes and increased nuclei.
Detailed examination of liver tissues harvested from the STZ-diabetic mice showed a marked increase in vacuole accumulation, as highlighted by PAS-diastase staining (Supplementary Fig. S6a). To determine whether these vacuoles are lipid droplets, we stained the liver sections with perilipin-1 (PLIN1), a selective lipid droplet-associated protein35. Interestingly, liver tissues isolated from STZ-induced diabetic mice displayed robust PLIN1 immunoreactivity indicating that the vacuoles are indeed lipid droplets (Supplementary Fig. S6b). Collectively, these results suggest that the STZ-induced model of diabetes is effective and causes only early changes in mice after 1 month.
Tibiae and femurs were isolated for the discovery of early pathogenetic changes in the bone tissues that may entail disrupted osteo- and adipogenesis. No changes in the lengths of the bones were observed between the two experimental groups (Supplementary Fig. S7). Using MarrowQuant, a semi-automated image analysis plug-in for QuPath36, we observed a significantly increased area and number of adipocytes in the tibia of diabetic mice but not in the femur (Figs. 1b–g). These findings suggest that diabetes, at least early on, specifically disrupts the tibia. To confirm this finding, we stained the tibia sections with PLIN1 and observed a significant increase in PLIN1 intensity in the marrows of diabetic mice (Figs. 1h–j). In the control mice, PLIN1-positive adipocytes were only found in the distal tibia (Fig. 1i). It is known that distal tibia contains constitutive adipose tissue that develops soon after birth and is relatively stable. Regulated adipose, on the other hand, is found in the shaft and proximal regions of tibia and responds to a wide range of changes including nutritional, physiological, and genetic37. Consistent with this classification, we found diabetes to enhance regulated adipose in shaft regions of the tibia. To supplement these results, we prepared a marrow flush from tibiae and femurs of the mice (Supplementary Fig. S8) to measure gene expression levels. We first examined adipogenesis-associated genes and show that transcriptional factors Cebpa and Pparg, and adipocyte-related adiponectin (Adipoq) are not increased in the tibia (nor femur) after 1 month of diabetes (Supplementary Fig. S9). However, analysis of osteogenesis-related genes showed a significant reduction in Runx2 in both the tibia and femur of the diabetic mice (Figs. 1k–n). Another osteoblast transcription factor, Sp7, was only downregulated in the femur but not the tibia at 1 month. Previous studies have shown that RUNX2 is the first transcription factor required for determination of the osteoblast lineage, while SP7 further directs the fate of mesenchymal cells to osteoblasts38. These findings show that diabetes may cause suppressed osteoblastogenesis in mice, and the effects may be more pronounced in the femur compared to the tibia.
Bone marrow adipogenesis in an STZ-induced type 1 model of diabetes in our current study and other previous studies39,40,41 suggests that the process is insulin-independent. However, insulin may enhance the process through the regulation of fatty acid transporters42,43 and induction of Pparg44. Acyl-CoA synthetase long chain family member 145 (ACSL1), CD3646, and solute carrier family 27 member 143 (also known as fatty acid transport protein 1, FATP1) have been shown to mediate lipid accumulation in various cell types. We first measured the expression of these fatty acid transporters in human bone marrow-derived mesenchymal progenitor cells (bm-MPCs; precursors of adipocytes) and show that the three fatty acid transporters are expressed robustly (housekeeping ACTB, threshold cycle (CT) 28.4 ± 0.15 (Mean ± SD); ACSL1, CT 28.31 ± 0.02; CD36, CT 30.19 ± 0.05; SLC27A1, CT 30.47 ± 0.06). We then tested whether bm-MPCs are capable of adipogenic differentiation in an insulin-independent manner. We produced an insulin-deficient adipogenesis induction media (idADP) containing serum, dexamethasone, isobutylmethylxanthine, and indomethacin (Supplementary Table S1). We then supplemented this base idADP with various test agents including high glucose (HG; 25 mmol/L), linoleic acid-oleic acid (LIN/OL; a source of fatty acids), or a combination. Lipid-laden cells were seen in idADP and were increased in the presence of HG and/or LIN/OL (Supplementary Fig. S10). We know that STZ induction of hyperglycemia in mice increases triglycerides and free fatty acids as early as 2 weeks following onset47. Therefore, our results suggest that adipogenesis in the bone marrow is insulin-independent and is enhanced by hyperglycemia and the availability of fatty acids.
Enhanced adipogenesis in the tibia of STZ-induced diabetic mice at 1 month of diabetes did not appear to be associated with any alteration in the marrow cellularity/nucleated area (majority of the cells being of hematopoietic origin) and Cd45 (Ptprc) mRNA levels (Supplementary Fig. S11). We and others have previously shown that diabetes does not alter the number of CD45-positive hematopoietic cells in experimental diabetes6. Other researchers have also shown that it takes 27–30 weeks of diabetes onset in rodents to decrease bone marrow cellularity5. In human type 2 diabetic patients, no changes in CD45-positive cells were noted6,19. Therefore, our results confirm previous reports of unaltered CD45-positive hematopoietic fraction at this early time point.
Since we and others have shown enhanced bone marrow adipogenesis to be associated with a reduction in non-hematopoietic stem cells in human diabetes6,19, we stained tibiae of mice with stem cell antigen-1 (SCA1) and sex determining region Y-box 2 (SOX2). Functional differentiation assays and single-cell RNA sequencing have confirmed the identity of non-hematopoietic marrow stem cells as SCA1-positive48,49,50. Our data shows no changes in SCA1 or SOX2 immunoreactive cells (Fig. 2a–d). Since SCA1 has been reported on different marrow cell subpopulations including stromal cells51, we further measured transcript levels of other stem cell genes Sca1 (Ly6a), Sox2, Oct4 (Pou5f1), and Nanog. Our results show no significant alteration in stem cell markers and transcripts in the bones of diabetic mice (Fig. 2e–l). These results show that bone marrow adiposity precedes tissue-resident stem cell depletion.
a Immunofluorescence staining of the tibiae of control and diabetic (STZ; 1 month) mice for SCA1 (green). Sections were counterstained with DAPI (blue) [scale bar = 50 μm]. b Quantification of SCA1 intensity per area, as determined by ImageJ [Mean ± SD; n = 4; two-tailed student’s t-test: *p < 0.05]. c Immunofluorescence staining of the tibiae of control and diabetic (STZ; 1 month) mice for SOX2 (green). Sections were counterstained with DAPI (blue) [scale bar = 50 μm]. d Quantification of SOX2 intensity per area, as determined by ImageJ [Mean ± SD; n = 3; two-tailed student’s t-test: *p < 0.05]. e–h mRNA levels of stem cell markers in tibia flush samples, showing Sca1 (Ly6a), Sox2, Oct4 (Pou5f1), and Nanog [Data normalized to Actb and Gapdh; Mean ± SD; n = 5 control and 7 STZ in panel e, 6 control and 7 STZ in panel f, 5 control and 7 STZ in panel g, 6 control and 7 STZ in panel h; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. i–l mRNA levels of stem cell markers in femur flush samples, showing Sca1 (Ly6a), Sox2, Oct4 (Pou5f1), and Nanog [Data normalized to Actb and Gapdh; Mean ± SD; n = 6 control and 7 STZ; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05].
Very recently, TGFB pathway was shown to regulate lineage specification of mesenchymal progenitor cells in the marrow during development52. Therefore, we examined the influence of hyperglycemia on TGFB signaling in diabetic mice. In the tibia, Tgfbr1, Tgfb1, Smad3, Smad6, and Bmp4 levels were significantly lower in STZ-induced diabetic mice compared with the non-diabetic controls (Fig. 3a–e). Similarly, in the femur, diminished levels of Smad2, Smad3, Smad6, and Bmp4 were observed in the diabetic mice (Fig. 3f–i). Immunostaining of tibiae for TGFB1 largely confirmed gene expression results (Fig. 3j, k). Suppressed expression of the TGFB pathway genes in bone tissues of mice suggests that the TGFB pathway may be involved in early diabetes-induced changes in the bone tissues.
Tibia and femur samples were harvested from control and streptozotocin (STZ)-induced diabetic mice after 1 month. a–e mRNA levels of the TGFB pathway genes in the tibiae of mice [For panels a, b, data normalized to Actb, Atp5f1, and Pgk1; for panels c–e, data normalized to Actb and Gapdh; Mean ± SD; n = 6 control and 7 STZ in panel a, 5 control and 7 STZ in panel b, 6 control and 6 STZ in panel c, 4 control and 7 STZ in panel d, 5 control and 7 STZ in panel e; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. f–i mRNA levels of the TGFB pathway genes in the femurs of mice [For panels f–i, data normalized to Actb and Gapdh; Mean ± SD; n = 6 control and 7 STZ in panel f, 6 control and 6 STZ in panel g, 6 control and 7 STZ in panel h; 5 control and 7 STZ in panel i; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. j Immunostaining of tibia marrow for TGFB1 (green). Sections were counterstained with DAPI (blue) [scale bar = 100 μm]. Inserts showing higher magnification. k Quantification of TGFB1 intensity per area, as determined by ImageJ [Mean ± SD; n = 4 control and 3 STZ; two-tailed student’s t-test: *p < 0.05].
In an attempt to identify which cell types express TGFB1 and downregulate its expression in diabetes, we carried out in situ hybridization (ISH) in the bone tissues of mice. Previous studies have shown that TGFB1 is detectable in bone marrow cells, chondrocytes, and cartilaginous matrix in both neonatal and adult mice53. Our Tgfb1 ISH showed a robust signal in megakaryocytes (Supplementary Fig. S12). This finding is consistent with a recent study showing that megakaryocytes express higher levels of Tgfb1 compared to stromal cells54. Other bone marrow cells also expressed Tgfb1, but the signal was obscured by the compact morphology of the marrow cells and hematoxylin stain. Examination of these stained sections using mercury arc illumination showed that most bone marrow cells express Tgfb1 and the likely reduction in Tgfb1 in diabetic mice is attributable to most cells, not specific cell subsets (Supplementary Fig. S13). Indeed, our immunohistochemical staining of the mouse tibia shows widespread presence of TGFB1 protein and overall reduction around all marrow cells (Fig. 3j).
We next examined bone tissues of longer-duration diabetes to determine whether TGFB suppression is maintained in diabetes and whether this associates with enhanced tissue adipogenesis. For this, we analyzed the bone marrows of mice at 2 months of diabetes duration. As expected, diabetic mice weighed significantly less and had higher blood glucose levels compared with the non-diabetic control mice at 2 months (Supplementary Fig. S14). Other structural changes such as an increase in collagen deposition also appear in organs (Supplementary Fig. S15). Moreover, we observed enhanced adiposity in the tibia of diabetic mice, coupled with a significant increase in Cebpa mRNA (Fig. 4a–c). Similar to the 1-month diabetic mouse model, the femur showed no significant increase in adiposity nor upregulated levels of genes associated with adipogenesis even after 2 months (Fig. 4i–k). Interestingly, Sp7 was downregulated in the tibia of diabetic mice at 2 months (Fig. 4e), unlike the results at 1 month. Furthermore, Adipoq, an adipocyte-related protein, and insulin-sensitizing hormone55, levels were significantly reduced in the bone marrow of diabetic mice (Fig. 4d, l). Importantly, analysis of bone tissues from diabetic mice at 2 months revealed a significant reduction in TGFB signaling in the tibia and femur (Fig. 4g–h and 4n–p), suggesting that these changes are sustained over time. Collectively, these findings show that hyperglycemia downregulates the TGFB signaling pathway in the bone marrow, which may contribute to the disruption of the marrow cellular composition.
C57BL/6 male mice received streptozotocin (STZ; 50 mg/kg) or citrate buffer (non-diabetic controls). Tibiae were harvested 2 months after the onset of diabetes. a Representative H&E-stained sections of the tibia showing proximal, shaft, and distal regions [scale bar = 200 μm]. mRNA levels of adipogenesis-associated (b–d) and osteogenesis-associated (e) genes in marrow flush samples of mouse tibia [Data normalized to Actb, Atp5f1, and Pgk1; Mean ± SD; n = 4 control and 3 STZ in panel b, 4 control and 6 STZ in panel c, 4 control and 3 STZ in panel d, 6 control and 4 STZ in panel e; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. mRNA levels of TGFB pathway genes in the tibiae of control or diabetic (STZ) mice after 2 months of diabetes onset [Data normalized to Actb and Gapdh; Mean ± SD; n = 4 control and 5 STZ in panels f and g, 3 control and 5 STZ in panel h; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. i Representative H&E-stained sections of the femur showing shaft and distal regions [scale bar = 200 μm]. mRNA levels of adipogenesis-associated (j–l) and osteogenesis-associated (m) genes in marrow flush samples of mouse femur [Data normalized to Actb, Atp5f1, and Pgk1; Mean ± SD; n = 4 control and 3 STZ in panel j, 5 control and 6 STZ in panel k, 5 control and 4 STZ in panel l, 3 control and 4 STZ in panel m; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05]. mRNA levels of TGFB pathway genes in the femurs of control or diabetic (STZ) mice after 2 months of diabetes onset [Data normalized to Actb and Gapdh; Mean ± SD; n = 4 control and 6 STZ in panel n, 5 control and 5 STZ in panel o, 6 control and 6 STZ in panel p; each data point represents a mouse; two-tailed student’s t-test: *p < 0.05].
A recent study found that the deletion of Tgfbr2 in Osx-Cre-targeted mesenchymal cells led to changes in Cxcl12 mRNA expression in the bone marrow52. Since our studies identified suppressed TGFB signaling in diabetes, we examined if stem cell niche factors, including CXCL12, were affected. As anticipated, there was a significant reduction in Cxcl12 levels in the tibia and femur of diabetic mice (Supplementary Fig. S16a, e). Unlike Cxcl12, however, no significant changes were seen in Kit and Kitl levels in the tibia or femur (Supplementary Fig. S16c, d, g, h), indicating specificity in the changes. Furthermore, levels of other known niche factors Angpt1, Icam1, and Vcam1 were not significantly altered in the bone marrow of diabetic mice (Supplementary Fig. S17). In support of a selective effect on Cxcl12 is the observation that the CXCL12 receptor, Cxcr4, which is expressed on stem cells was not altered in the tibia nor femur (Supplementary Fig. S16b, f). We also found that the CXCL12 staining intensity in the bone marrow was reduced in bone tissues of diabetic mice (Supplementary Fig. S16i, j). Analysis of tissues from mice at 2 months of diabetes duration revealed a sustained reduction in Cxcl12 expression in the tibia and femur (Supplementary Fig. S16k, l), suggesting that these changes are also sustained over time, much like TGFB suppression. Such changes did not appear to be mediated through a reduction in CXCL12-expressing cells56,57 (known as CAR cells), as leptin receptor (LEPR), a marker of CAR cells58,59, did not show significant alteration in the tibia of 1-month diabetic mice (Supplementary Fig. S18).
Our laboratory has previously found that non-canonical WNT11 mediates high glucose-induced adipogenic differentiation of MPCs in culture6. Additionally, previous studies have reported a potential crosstalk between the TGFB and Wnt signaling pathways60. Therefore, we examined whether canonical and non-canonical Wnt signaling pathways may be altered in STZ-induced model of early hyperglycemic alterations. We used Ctnnb1 and Wnt11 levels as a proxy for the canonical and non-canonical Wnt signaling, respectively. Analysis of bone tissues, however, showed no changes in the Wnt signaling pathway (Supplementary Fig. S19). These findings suggest that TGFB alteration may precede Wnt alteration in diabetes-induced bone marrow deficits.
We have previously shown that culture conditions that mimic diabetes cause enhanced adipogenic differentiation in human bm-MPCs6. Furthermore, bm-MPCs from type 2 diabetic patients express high levels of Cebpa, favouring adipogenic differentiation61. Therefore, we used bm-MPCs to examine the signaling mechanisms in detail. We first investigated the effect of high levels of glucose on the TGFB pathway by culturing cells in normal growth media containing elevated levels of glucose for 21 days (Supplementary Fig. S20). We screened for genes associated with adipogenic differentiation and show that cells cultured in media containing 25 mmol/L glucose (high glucose; HG) exhibited suppressed mRNA levels of the TGFB family, including the downstream signaling mediators, SMAD2, SMAD3, and SMAD4, compared with cells cultured in media with 5 mmol/L glucose (Supplementary Fig. S20). In contrast to the TGFB pathway, no changes were observed in the fibroblast growth factor and insulin-like growth factor signaling pathways, other known adipogenesis-inducing pathways. This suggests that hyperglycemia potentially suppresses TGFB signaling in bm-MPCs to regulate cell fate changes.
We next explored the effect of exogenously supplied TGFB1 on adipogenic differentiation of bm-MPCs. We exposed bm-MPCs to adipogenic differentiation media with or without 10 ng/mL TGFB1 for 72 h. This time point has been previously reported to be sufficient for cultured cells to initiate differentiation62,63. TGFB1 exposure inhibited adipogenic differentiation of cells as illustrated by a significant reduction in PPARG2 (Fig. 5a), which supports the findings from a mouse pre-adipocyte cell line64. Moreover, TGFB1 reduced the number of adipocytes in the culture and the level of lipid accumulation in the cells (Figs. 5b–d). We next tested whether elevated glucose levels further increase adipogenic programming. The addition of 25 mmol/L glucose to adipogenic differentiation media significantly increased PPARG2 levels compared with adipogenic differentiation media with only 5 mmol/L glucose (Figs. 5e, f). These results are in alignment with our previous studies65 and support the hypothesis that hyperglycemia in diabetes enhances bm-MPC adipogenesis. Interestingly, the addition of exogenous TGFB1 in high-glucose adipogenic media decreased PPARG2 to levels comparable to adipogenic media with 5 mmol/L glucose, and reduced lipid content within the cells (Figs. 5e, f).
Human bone marrow-derived progenitor cells (bm-MPCs) were induced to differentiate in an adipogenesis-inducing media (ADP), with or without TGFB1 (10 ng/mL) for 72 h. a mRNA levels of PPARG2 in bm-MPCs [Data normalized to ACTB, GAPDH, and RPLP0; Mean ± SD; n = 4; each data point represents an independent sample; ANOVA followed by Bonferroni post hoc analysis: *p < 0.05]. b bm-MPCs, treated as indicated in panel a, were stained with LipidTOX (green) to detect intracellular lipid accumulation. Cells were counterstained with DAPI (blue) [scale bar = 50 μm]. Inserts showing higher magnification. The number of adipocytes and frequency of lipid droplets were measured by CellProfiler [Mean ± SD; n = 3 for panel c, 3 images per replicate were measured for panel d; ANOVA followed by Bonferroni post hoc analysis: *p < 0.05]. e bm-MPCs were cultured in control, adipogenic media (ADP), or ADP supplemented with TGFB1 (10 ng/mL) or high glucose (HG; 25 mmol/L) for 7 days. mRNA levels of PPARG2 were measured [Data normalized to ACTB; Mean ± SD; n = 4; each data point represents an independent sample; ANOVA followed by Bonferroni post hoc analysis: *p < 0.05 compared with control, †p < 0.05 compared with ADP]. f Detection of intracellular lipid accumulation in bm-MPCs by LipidTOX (green) staining. Cells were counterstained with DAPI (blue) [scale bar = 50 μm]. Cells were treated as indicated in Panel e.
Next, we used a variety of inhibitors of the TGFB signaling pathway to identify which arm of the TGFB pathway may be inhibiting adipogenesis in bm-MPCs. The inhibition of ALK5, TGFB-associated kinase 1 (TAK1), and JNK significantly rescued PPARG2 and lipid accumulation in cells, negating the effects of TGFB1 (Figs. 6a, b). This triggered the question of whether cells would enhance adipogenic differentiation potential with these inhibitors under normal growth media. As mentioned earlier, bm-MPCs from type 2 diabetic patients show high expression levels of Cebpa, which intrinsically favours adipogenic differentiation61. Although, ALK5 inhibition did not change the PPARG2 levels, inhibiting JNK significantly induced PPARG2 in the cells (Fig. 6c). A dampened response was also seen with TAK1 inhibitor. These results suggest that TGFB restricts adipogenic differentiation in bm-MPCs by maintaining non-canonical TAK1 activity.
Bone marrow-derived mesenchymal progenitor cells (bm-MPCs) were cultured in adipogenic media (ADP) with TGFB1 (10 ng/mL), and various inhibitors of the TGFB signaling pathway. All inhibitors were tested at 10 µmol/L concentration and for 72-hour exposure. a Levels of PPARG2 mRNA in bm-MPCs [Data normalized to ACTB, GAPDH, and RPLP0; Mean ± SD; n = 4 for all treatments except ADP + TGFB1 + JNKi and ADP + TGFB1 + p38i which were 3; each data point represents an independent sample; two-tailed student’s t-test: *p < 0.05 compared with adipogenic media and TGFB1 (ADP + TGFB1)]. b bm-MPCs, treated as indicated in panel a, were stained for intracellular lipid accumulation by LipidTOX (green). Cells were counterstained with DAPI (blue) [scale bar = 50 μm]. Inserts showing higher magnification. c bm-MPCs cultured in control (DMEM) media supplemented with TAK1 inhibitor (TAK1i; 10 µmol/L) or JNK inhibitor (JNKi; 10 µmol/L) for 72 h were analyzed for PPARG2 mRNA levels [Data normalized to ACTB, GAPDH, and RPLP0; Mean ± SD; n = 3 control, 3 ALK5i, 3 JNKi, and 2 TAK1i; each data point represents an independent sample; two-tailed student’s t-test: *p < 0.05 compared with control].
We performed a microarray analysis to understand the mechanisms of TGFB1-mediated inhibition of adipogenesis and to identify targetable pathways. We cultured bm-MPCs in normal growth media (Control), adipogenic differentiation media (ADP), adipogenic differentiation media with TGFB1 (TGFB1), and adipogenic differentiation media with TGFB1 + TAK1 inhibitor (TAK1i). With a threshold of p = 0.05 (Fig. 7a), we found a total of 193 genes that were significantly higher in ADP, suppressed in TGFB1, and rescued with TAK1i. On the other hand, 154 genes were downregulated in ADP, increased in TGFB1, and normalized by TAK1i. From these 347 identified genes, Control and TGFB1 groups, and ADP and TAK1i groups displayed similar gene expression signatures and were clustered together (Fig. 7b).
a A threshold of p = 0.05 was used to identify differentially regulated genes between the groups. The patterns sought, either (1) genes upregulated with adipogenic media with/without TAK1 inhibitor (TAK1i), which is normalized with TGFB1; or (2) genes downregulated with adipogenic media with/without TAK1i, which is normalized with TGFB1. The identified targets were further analyzed with the DAVID informatics tool to identify significant biological processes. b A heatmap of the differentially expressed genes (347 genes) from the Clariom S Assay (n = 2 for each group) that clusters Control and TGFB1; and ADP and TAK1i. The gene set enrichment analysis for groups: c Control versus adipogenic differentiation; d adipogenic differentiation versus adipogenic differentiation with TGFB1; and e adipogenic differentiation with TGFB1 versus adipogenic differentiation with TGFB1 and TAK1 inhibitor. A positive normalized enrichment score indicates gene sets that are enriched in bm-MPCs exposed to adipogenic differentiation media (c, d) and adipogenic differentiation media with TGFB1 and TAK1 inhibitor (e). A Venn diagram illustrating the enriched pathways found in control and TGFB1 (f) and adipogenic differentiation and TAK1 inhibitor (g). The list of pathways is the common pathways found in all three comparisons.
To identify biological process targets, we performed a GO analysis of the identified genes. We found that the upregulated genes in ADP (and rescued TAK1i in TGFB1-treated cells) are largely involved in fatty acid and lipid metabolism and regulation, with 38 GO biological processes identified with a p < 0.05. This is not unexpected in stem cell differentiation studies where most of the genes and pathways would be related to the phenotype of cells being derived. More interestingly, the downregulated genes are involved in various biological processes, including osteoblast differentiation and the Wnt signaling pathway, with 19 GO biological processes identified with a p < 0.05 (Supplementary Table S2, S3).
To further extract biological insight from the microarray dataset, we performed a Gene Set Enrichment Analysis (GSEA) for hallmark gene sets (v7.4; 50 gene sets) comparing two groups at a time. The groups that were compared were (1) Control vs ADP (Fig. 7c, Supplementary Table S4), (2) ADP vs TGFB1 (Fig. 7d, Supplementary Table S5), and (3) TGFB1 vs TAK1i (Fig. 7e, Supplementary Table S6). Control and TGFB1 groups were indicated with negative normalized enrichment scores, whereas ADP and TAK1i groups were indicated with positive normalized enrichment scores. Next, common enriched gene sets were identified within the groups that displayed similar gene expression signatures. A total of 12 gene sets were enriched in the control and TGFB1 groups, and 10 gene sets were enriched in the ADP and TAK1i groups (Figs. 7c–e). Of the 12 gene sets that were identified in the control and TGFB1 groups, the Wnt Beta Catenin Signaling was of great interest, as our laboratory has previously identified that high-glucose induced adipogenic differentiation of bm-MPCs involves the switch from canonical to noncanonical Wnt signaling6.
To investigate the potential Wnt signaling mechanism, downstream of the TGFB pathway, we exposed bm-MPCs to TGFB1 in adipogenic media and examined mediators involved in the canonical Wnt pathway. Apart from TCF7L1, other canonical Wnt mediators, including CTNNB1 and LEF1, were significantly induced with TGFB1 exposure (Supplementary Fig. S21a). Moreover, TLE1 protein, a co-repressor in the Wnt pathway, localization in nuclei of the cells was lost with TGFB1 exposure (Supplementary Fig. S21b), supporting activation of the canonical Wnt pathway. Exposure to TGFB1 in normal growth media was able to significantly increase CTNNB1 while decreasing CCND1 mRNA (Supplementary Fig. S21c). Furthermore, TAK1 inhibition in the presence of TGFB1 and adipogenic differentiation signals significantly increased PPARG2 while significantly decreasing WISP1 (canonical Wnt target gene) levels (Supplementary Fig. S21d, e). This suggests that TGFB1 may activate the canonical Wnt signaling pathway in bm-MPCs to regulate cell differentiation.
As TGFB1 inhibits adipogenic differentiation, it is possible that one underlying mechanism involves skewing the differentiation of bm-MPCs towards the osteogenic lineage. In addition, the osteoblast pathway was identified in the microarray analysis in cells exposed to TGFB1. Therefore, we examined the effect of exogenous TGFB1 exposure on osteogenic differentiation of bm-MPCs. Previous studies have shown that TGFB1 can regulate mesenchymal cell differentiation66. No changes were seen in the osteogenesis-associated RUNX2 transcription factor (Supplementary Fig. S22a). However, SP7 was induced by TGFB1. In addition, the late marker of osteoblast differentiation, BGLAP (also known as Osteocalcin) was significantly downregulated. Staining of cells with Alizarin Red S, which marks mineralization, showed reduced levels in cells exposed to TGFB1 (Supplementary Fig. S22b, c). Hence, these results suggest that TGFB1 may prime bm-MPCs to differentiate into osteoblasts, but cells do not reach maturation, possibly due to continued exposure to TGFB1.
Our studies have discovered that diabetes mediates early structural alterations in the bone marrow. The key findings of our study include: (1) 1 month of diabetes is sufficient to observe enhanced tibia marrow adipogenesis in mice, (2) femurs may be more resistant to diabetes-mediated adipogenesis compared with tibiae, while both are susceptible to reduced osteogenesis, and (3) TGFB signaling pathway is suppressed in bone tissues of diabetic mice. We further show that enhanced marrow adipogenesis precedes the depletion of marrow-resident stem cells. The main findings of the study are summarized in Fig. 8.
Transforming growth factor-beta (TGFB) signaling in healthy bone tissues maintains a balance between osteoblastogenesis and adipogenesis, in collaboration through canonical Wnt signaling. At least for adipogenesis, TGFB restricts differentiation of marrow progenitor cells through the non-canonical TGFB-activated kinase 1 (TAK1) mechanism. Elevated glucose levels in diabetes suppress the TGFB signaling pathway, which, through alleviation of TAK1, leads to Pparg induction and expression of fatty acid- and lipid metabolism-regulating genes to favour adipogenesis. Dashed black arrows indicate linkages that require further investigation.
STZ-induced hyperglycemia in rodents generates readouts that mimic human diabetic complications, including skeletal fragility and enhanced adiposity in the bone marrow67. Moreover, microangiopathy, neuropathy, and stem cell rarefaction are observed in the bone marrow of diabetic patients and experimental models6,8,19. Our studies show increased adipocytes in the tibia to be associated with suppressed TGFB pathway. A recent study investigated the effect of TGFB signaling pathway in bone tissues during development by deleting Tgfbr2 using a doxycycline-repressible Sp7 (osterix)-Cre transgene (Osx-Cre) in mice 3 weeks of age. These mice show a significant reduction in growth, body weight, bone mass and quality52. Moreover, marrow adiposity was significantly enhanced in Osx-Cre, Tgfbr2fl/fl mice, where osmium staining and Oil Red O-positive cells in the bone marrow increased almost 80-fold. Therefore, Tgfbr2 deletion in the bone marrow mirrors diabetes-induced changes such that reduced expression of the TGFB pathway, including the downstream mediators, may remove the suppression and allow bm-MPCs to readily differentiate into adipocytes.
Although our studies focused on the type 1 model of diabetes, extensive investigations by researchers show expansive bone marrow adiposity in type 2 diabetic patients19,61, and experimental models of type 2 diabetes61. It is unknown whether suppressed Tgfb1 mediates enhanced adipogenesis in the bone marrow in type 2 diabetes. No changes in Tgfb1 were seen in adipocytes isolated from the marrow of type 2 diabetic patients compared to cells from non-diabetic patients61. It is quite possible that hyperinsulinemia and adipokine-mediated processes may be involved in type 2 diabetes. In support of this notion, bm-MPCs from type 2 diabetic patients express high levels of adipogenesis-associated Cebpa61. Furthermore, bone marrow adipocytes from type 2 diabetic patients, but not healthy subjects, produce paracrine factors that enhance adipogenesis in MPCs61. Whether this represents an enhanced self-propagating68,69 property of bone marrow adipocytes from diabetic patients or a TGFB1-independent mechanism is an exciting avenue to explore in future studies.
Another interesting finding of our study is the reduced expression of Cxcl12 in the marrow of diabetic mice. A recent study has shown that purified bone marrow adipocytes in mice express high levels of Cxcl1270. It is possible that the reduced Cxcl12 expression seen in marrow flush samples in our study is partly due to the loss of adipocytes, which express Cxcl12, during centrifugation and lysis. However, when examined with other experimental evidence presented, there does appear to be an overall reduction in marrow Cxcl12 expression. For example, femur tissues contained almost no adipocytes in the control and STZ-induced diabetic mice at the 1-month time point but showed a reduced level of Cxcl12 in the diabetic mice. Furthermore, staining of intact bone tissues also showed reduced CXCL12 levels in tissues harvested from diabetic mice. It is possible that CXCL12 suppression is related to disrupted TGFB signaling in diabetes. In support of this potential link to TGFB, exposure of hepatocytes to recombinant TGFB1 has been shown to upregulate CXCL1271. However, empirical evidence is needed in bone tissues and bone marrow-derived cells. Nonetheless, reduced CXCL12 level in the marrow is of great significance based on our current knowledge of the potential role of CXCL1256. In vitro, CXCL12 acts as a potent chemoattractant for primitive bone marrow CD34+CD38– cells that include hematopoietic stem cells72,73,74. In vivo, CXCL12 has been shown to promote the engraftment of transplanted hematopoietic cells in the bone marrow and subsequent hematopoietic reconstitution75. In diabetic mice, granulocyte-colony stimulating factor fails to reduce Cxcl12 in the bone marrow, while the levels were efficiently downregulated in non-diabetic mice76. However, it should be noted that control mice had significantly higher starting levels of Cxcl12 compared with diabetic mice. Therefore, reduced basal expression of CXCL12 may potentially participate in hematopoietic stem cell mobilization seen in diabetic patients.
Our studies temporally place TGFB disruption before Wnt alteration in diabetic mice and cultured bm-MPCs. We show that the TGFB1 pathway regulates canonical Wnt signaling in bm-MPCs. Furthermore, TGFB1 exposure enriched the Wnt Beta Catenin Signaling gene sets in our profiling. In pre-adipocytes, Wnt signaling decreases as cells differentiate into mature adipocytes, and the expression of CTNNB1 shows an inverse relationship with levels of CEBPA, a key regulator of adipogenesis77. In our study, the TGFB1-mediated activation of the canonical Wnt pathway was evident with the loss of nuclear expression of TLE1 along with the induction of CTNNB1.
In summary, our studies identified bone marrow alterations as one of the early complications of diabetes. We show enhanced bone marrow adiposity in tibiae of mice to be associated with suppressed TGFB signaling, a known regulator of cell differentiation. We explored the functional significance of disrupted TGFB signaling using bm-MPCs and showed that TGFB1-mediated TAK1-JNK axis activation plays an important role in preventing adipogenic differentiation. Identifying avenues to activate the noncanonical TGFB signaling in the marrow may be important to preserving bone integrity in diabetes and the marrow regenerative stem cells.
All mouse studies were initiated after receiving approval from the Animal Care and Veterinary Services at Western University (Animal Use Protocol: #2019-125).
Four-week-old male C57BL/6 mice (Charles River Canada) were obtained and allowed to acclimate for 1 week. Mice were kept on a light/dark (12 h/12 h) cycle at 23 °C and received standard rodent chow and water ad libitum. Mice were divided into two groups, diabetic and non-diabetic, by ensuring the body weights were similar between the two groups. Diabetes was induced by multiple low-dose intraperitoneal injections of freshly prepared streptozotocin (STZ; 50 mg/kg in 0.1 M citrate buffer, pH 4.5; Sigma-Aldrich; Cat# S0130) using 25 G needles, for 5 consecutive days78. Non-diabetic control mice received an equivalent volume of freshly prepared citrate buffer alone. Body weights were measured daily for the duration of the study. Mice were also monitored daily for body condition scores and overall health. Hyperglycemia was confirmed 1 week after the last STZ injection by measuring non-fasting blood glucose levels from the tail vein, using a glucometer (CONTOUR Next One, Ascensia Diabetes Care). This day was considered day zero (d0). Mice with non-fasting blood glucose levels greater than 11.1 mmol/L (>200 mg/dL) were considered hyperglycemic/diabetic. Mice that did not meet the threshold were measured again after 2 days to confirm hyperglycemia. All mice injected with STZ exhibited hyperglycemia and were included in the study. Mice in non-diabetic and diabetic groups remained at a body condition score of 3 (well-conditioned). Each group contained seven mice.
One day prior to euthanasia, blood glucose levels were measured again. Following CO2 euthanasia, mice were monitored for cessation of breathing/heartbeat. An incision in the chest wall was made immediately to reach the heart. A small slit was made into the right atrium. Mice were then perfused with 5 mL of cold phosphate-buffered saline (PBS; Fisher Scientific; Cat# BP399-1) by inserting a 28 G needle into the left ventricle. Femurs and tibiae were then removed and freed from soft tissue. Limbs from one side were used for histology, and the other side was used for preparing flush samples. For histological analyses, the bones were fixed in 10% neutral buffered formalin (Sigma-Aldrich; Cat# HT501128) overnight at room temperature, decalcified in 0.5 mol/L ethylenediaminetetraacetic acid (EDTA; Fisher Scientific; Cat# S311) for 7 days at 4 °C, and rehydrated in 30% sucrose solution (Sigma-Aldrich; Cat# 84100) overnight at 4 °C. Samples were moved to 70% ethanol and processed and embedded in paraffin. All bone tissues were sectioned at a 4-μm thickness to ensure tissues remain affixed to the glass slides (Fisher Scientific; Cat# 1255017). For gene expression analyses, the epiphyses of the femurs and tibiae were removed, and tissues were placed in a 0.5 mL centrifuge tube nested in a 1.5 mL centrifuge tube. A small hole was created at the bottom of the 0.5 mL tubes with an 18G needle, prior to adding the tissues. Tubes containing bone tissues were then centrifuged at 10,000 xg for 15 s to flush the bone marrow. This method of preparing cell suspensions removed all cellular content from the marrows, as can be detected by histological analysis before and after the procedure. The resulting flush was stored in a cell freezing media solution (90% fetal bovine serum and 10% dimethyl sulfoxide; Millipore; Cat# S-002-5F) at −20 °C until further analysis. For RNA isolation from the marrow cells, the flush samples were quickly thawed to 37 °C in a water bath and centrifuged at 300 xg to pellet the cells. Cells were then suspended in RLT cell lysis buffer (RNeasy Plus Mini Kit; Qiagen, Cat# 74134).
In addition to the bone tissues, other tissues from mice were collected. Specifically, eyes, heart, lung, kidney, liver, pancreas, and epididymal adipose tissues were removed for histology. The tissues were fixed in 10% neutral buffered formalin (Sigma-Aldrich; Cat# HT501128) overnight at room temperature prior to paraffin embedding. Non-bone tissues were sectioned at 5-μm thickness for immunostaining and other histological studies.
A longer-term diabetes study was also performed. Following the same procedure as indicated above, non-diabetic and STZ-induced diabetic mice were followed up at 2 months. Femurs and tibiae were removed and processed as indicated above. Bone marrow flush samples were also prepared for gene expression studies. Select target organs of diabetic complications such as the heart, kidney, and eyes were also harvested for histological studies.
Total RNA from cultured cells and marrow flush samples was isolated using RNeasy Mini Plus Kit (Qiagen; Cat# 74134). RNA from formalin-fixed paraffin-embedded (FFPE) epididymal adipose tissue sections (total of 20-μm thickness) was collected using the PureLink FFPE RNA Isolation Kit (Thermo Fisher; Cat# K156002). Total RNA was measured using Qubit RNA Broad Range Assay (Thermo Fisher; Cat# Q10210) or the Qubit RNA High Sensitivity Assay (Thermo Fisher; Cat# Q32852) in a Qubit Fluorometer (Thermo Fisher; Cat# Q32857). cDNA was synthesized using iScript cDNA Synthesis Kit (Bio-Rad Laboratories; Cat# 1708841). qPCR reactions for SYBR-based amplification and detection consisted of 10 µL RT2 SYBR Green qPCR Mastermix (Qiagen; Cat# 330502), 2 µL of both forward and reverse primers (Supplementary Table S7 and S8) at 10 µmol/L concentration, 2 µL cDNA, and 6 µL nuclease-free H2O (Ambion; Cat# AM9937). For TaqMan-based chemistry, the reactions consisted of 5 µL TaqMan Fast Advanced Master Mix (Thermo Fisher; Cat# 4444963), 0.5 µL of the gene probe (Supplementary Table S7 and S8), 1 µL cDNA, and 3.5 µL nuclease-free H2O. All reactions were run on CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories). Target gene mRNA data were normalized to housekeeping genes identified in figure legends. For most studies, the geometric mean of multiple housekeeping genes was used for normalization. SYBR Green-based reactions were performed for 40 cycles using the following temperature profiles: 95 °C for 15 s (initial denaturation), and 60 °C for 60 s (annealing and extension). TaqMan-based reactions were performed for 40 cycles using the following temperature profiles: 95 °C for 3 s (initial denaturation), and 60 °C for 30 s (annealing and extension). Data were analyzed by CFX Manager software (Bio-Rad; Cat# 1845000) using the normalized (ΔΔCT) method.
Histomorphometric analyses were performed on mouse retina, epididymal fat, and bone tissues. The thickness of retinal layers and bone lengths were quantified using QuPath79. The frequency of the area of adipocytes found in the white epididymal adipose tissue was analyzed using Adiposoft, an automated software80. The number of nuclei per area in tissues was determined by QuPath79. Bone marrow analysis was conducted while considering the guidelines and standards previously reported81. The bone marrow histological sections were analyzed using MarrowQuant36 and QuPath79.
Deparaffinized tissue sections were stained using hematoxylin (Leica Biosystems; Cat# 3801561) and eosin (Leica Biosystems; Cat# 3801601) for routine histological analysis. Picro-Sirius Red (Abcam; Cat# ab246832) was used to mark collagen deposition. Periodic Acid-Schiff (PAS) stain kit (Abcam; Cat# ab150680) was used to detect the presence of carbohydrates and carbohydrate compounds such as polysaccharides, mucin, and glycogen. Liver sections were also subjected to α-amylase (Sigma-Aldrich; Cat# A3176) treatment prior to PAS (PAS-diastase Staining) to remove glycogen. The stained slides were digitalized using Aperio slide scanning system (Aperio Technologies, Inc., Vista, CA) with a 20X and 40X objective and saved in Tagged Image File Format (TIFF). Image analyses were performed with the ImageJ software package (National Institutes of Health, Bethesda, MD, USA). Picro-Sirius red-stained area was automatically assessed by means of a macro (https://github.com/northcottj/picrosirius-red). The Picro-Sirius Red-positive area of the specimens was quantified using ImageJ software and shown as a percentage of the total section area.
Tissue sections (all except bone tissues) were deparaffinized in xylenes, hydrated in an ethanol gradient, and subjected to a heat-induced antigen retrieval using Citrate/EDTA buffer (10 mmol/L sodium citrate, 1 mmol/L EDTA, 0.05% Tween-20, pH 6.0) or Tris/EDTA buffer (10 mmol/L Trizma base, 1 mmol/L EDTA, 0.05% Tween-20, pH 9.0) in 2100 Retriever (Electron Microscopy Science, Hatfield, PA). Sections were incubated with rabbit anti-mouse perilipin-1 (PLIN1; Abcam; Cat# ab3526; 1:200 dilution) or insulin (INS; Thermo Fisher; Cat# 15848-1-AP; 1:1000 dilution) at room temperature for 1 h. FITC- or peroxidase-conjugated secondary antibodies (Vector Laboratories, Burlington, ON) were used for detection. Where peroxidase-conjugated secondary antibody was used, AEC (3-amino-9-ethylcarbazole) substrate (Vector Laboratories; Cat# SK-4200) was utilized for detection, followed by hematoxylin counterstaining. For immunofluorescence staining, sections were exposed to Vector TrueVIEW Autofluorescence Quenching Kit (Vector Laboratories; Cat# SP-8400-15) for 5 min, followed by counterstaining with DAPI (Sigma Aldrich; Cat# F6057). Fluorescence images were captured with an Olympus BX-51 microscope equipped with Infinity 3-1 Color CCD camera (Teledyne Lumenera, Ottawa, Canada) and Infinity Analyze software (Teledyne Lumenera). AEC-stained slides were digitalized using an Aperio slide scanning system (Aperio Technologies) with a 20X objective and saved in TIFF.
Bone tissue sections were deparaffinized in xylenes and rehydrated through an ethanol gradient. Tissue sections were then subjected to a heat-induced antigen-retrieval process using a microwave. Briefly, jars containing the antigen retrieval buffer (10 mmol/L sodium citrate, 1 mmol/L EDTA, 0.05% Tween-20, pH 6.0 or 10 mmol/L Tris, 1 mmol/L EDTA, 0.05% Tween-20, pH 9.0) were placed in the microwave and the buffer was boiled. Tissue sections were then immersed in the jar when the buffer had cooled and was not boiling. This process was repeated until tissues had been subjected to a total of 15 min in a heated buffer. Sections were then blocked in 1% bovine serum albumin (Sigma-Aldrich; Cat# A7906) solution in PBS containing 0.1% Tween-20 (Sigma-Aldrich; Cat# P9416) for 1 h, and then incubated with primary antibody (Supplementary Table S9) for 1 h at room temperature. Fluorophore-conjugated secondary antibodies were applied for 1 h at room temperature. Secondary antibodies included Alexa Fluor 488-conjugated anti-goat (Thermo Fisher; Cat# A-11079), FITC-conjugated anti-rabbit (Thermo Fisher; Cat# F-2765), or FITC-conjugated anti-mouse (Vector Laboratories; Cat# FL-2000). All sections were then exposed to Vector TrueVIEW Autofluorescence Quenching Kit (Vector Laboratories; Cat# SP-8400-15) for 5 min, followed by counterstaining with DAPI (Sigma Aldrich; Cat# F6057). Negative control sections were included for all staining studies and consisted of the same tissue and procedure but with the omission of the primary antibody. Fluorescence images were captured using an Olympus BX-51 microscope. Where fluorescence intensity was measured, sections from all experimental conditions were subjected to the same exposure time when capturing images. Fluorescence staining intensity per area of the sections was performed using QuPath and ImageJ.
Tgfb1 mRNA was detected in mouse bone tissues using RNAscope in situ hybridization probes and RNAscope 2.5 Red Reagent Kit (Advanced Cell Diagnostics; Cat # 322350). Paraffin-embedded tissue slides were baked at 60 °C for 1 h in Lab-Line TempCon Oven (American Scientific Products, Livonia, MI). Slides were deparaffinized in xylenes for 10 min and then incubated in absolute ethanol for 2 min. Endogenous peroxidase activity was blocked by incubating slides in RNAscope Hydrogen Peroxide (provided with Red Reagent Kit) for 10 min. Slides were washed in dH2O and immersed in Custom Pretreatment reagent for bone samples (Advanced Cell Diagnostics; Cat # 300040) at 40 °C for 30 min in HybEZ Oven (Advanced Cell Diagnostics; Cat # 310010). Probes were then added, and slides were incubated at 40 °C for 2 h in HybEZ Oven. Probes included Mm-Tgfb1 (Cat# 407751) and Bacillus subtilis dihydrodipicolinate reductase (Dapb) negative control (Cat # 10043). Signal amplifiers were added in sequence, as recommended by Advanced Cell Diagnostics. Detection was performed with the RED reagent (Advanced Cell Diagnostics; Cat # 322350). Slides were counterstained with hematoxylin and mounted. Brightfield images were taken with an Olympus BX51 microscope. The same slides were imaged using mercury arc illumination and TRITC filter. RNAscope 2.5 Red Reagent is naturally fluorescent around 570 nm (TRITC/CY3).
Human bone marrow-derived mesenchymal progenitor cells (bm-MPCs) were prepared from freshly isolated human bone marrow mononuclear cells (Lonza Inc., Walkersville, MD). Multiple donors were used in this study (Lot# 081032A, 43-year-old Caucasian female; Lot# 081109 A, 43-year-old Hispanic male; Lot# 080511A, 20-year-old Black female; Lot# 080500B, 19-year-old Black male; Lot# 081362 A, 27-year-old Caucasian female; Lot# 080455B, 19-year-old Black male; Lot# 081393A, 19-year-old Black female). Bone marrow mononuclear cells were seeded on tissue culture plates in low glucose Dulbecco’s Modified Eagle Medium (DMEM; Thermo Fisher; Cat# 11885084), supplemented with 10% fetal bovine serum (FBS; Lonza, Cat# 14-507 or Thermo Fisher, Cat# 12484028) and 1X PSF (antibiotic-antimycotic solution; Corning, Cat# 30-004-CI).
All in vitro studies were performed with at least four different donor preparations (N) conducted at different times, and more than two experimental replicates (n) each time. All bm-MPC preparations were used before sub-passage 7. To induce adipogenic differentiation, bm-MPCs were plated at a density of 40,000 cells/cm2. The next day, cells were exposed to adipogenesis-inducing media for various time points, ranging from 48 h to 7 days (Supplementary Table S10). Adipogenesis-inducing media comprised of DMEM low glucose media supplemented with 1X StemPro Adipogenesis Supplement from StemPro Adipogenesis Differentiation Kit (Thermo Fisher; Cat# A1007001). We did not use the base media provided in the Adipogenesis Differentiation kit as it contained high levels of glucose. Seeding cells at a high density prior to the adipogenesis media exposure was performed to ensure we capture the differentiation phase only and not the early mitotic burst-related changes. Although the formulation of commercial adipogenesis media is proprietary, adipogenesis in human mesenchymal precursors is easily induced in media containing 10% FBS, 10 μg/mL insulin, 1 μM dexamethasone, 0.5 mM isobutylmethylxanthine, and 60 μM indomethacin. However, for consistency in experiments conducted with different donor cells and at different times, the pre-formulated commercially available media was used. We did confirm with the vendor that the media formulation does not contain peroxisome proliferator activated receptor gamma (PPARG) agonists, such as pioglitazone, which would confound the results. Control media consisted of DMEM low glucose media supplemented with 10% FBS. Media was changed every other day.
For some studies, we tested adipogenic differentiation of bm-MPCs in the absence of insulin. For these experiments, we produced an insulin-deficient adipogenesis-inducing media (idADP) containing serum, dexamethasone, isobutylmethylxanthine, and indomethacin (Supplementary Table S1). We then supplemented this base idADP with various test agents including high glucose (25 mM), linoleic acid-oleic acid (LIN/OL; a source of fatty acids), or a combination. Cells were cultured in the test media for 7 days with media change every other day. Phase images were captured on day 7 using Olympus BX-51. Cells were then fixed with paraformaldehyde (Electron Microscopy Sciences; Cat# 15710) for 10 min, and then stained with HCS LipidTOX Green neutral lipid stain (Thermo Fisher; Cat# H34475) for 30 min, as detailed below. Fluorescence images were captured.
To determine the effect of TGFB1 on bm-MPC differentiation, exogenous human TGFB1 (R&D Systems, Cat# 100-B-001) was added to the adipogenesis induction media (DMEM low glucose media supplemented with 1X StemPro Adipogenesis Supplement) at 10 ng/mL concentration. To determine how TGFB1 mediates its effect in bm-MPC differentiation, downstream signaling proteins that are responsive to TGFB1 were inhibited using commercially available inhibitors (Supplementary Table S11).
To induce osteoblast differentiation in bm-MPCs, cells were plated at a density of 40,000 cells/cm2. After 24 h, cells were exposed to osteogenesis-inducing media for various time points, ranging from 9 days to 21 days (Supplementary Table S10). To prepare osteogenesis-inducing media, DMEM low glucose was supplemented with 1X StemPro Osteogenesis Supplement provided in the StemPro Osteogenesis Differentiation Kit (Thermo Fisher; Cat# A1007201). At the end of the experiment, cells were fixed and stained with Alizarin Red S (Sigma-Aldrich; Cat# A5533), as detailed below.
Total RNA from bm-MPCs was isolated using RNeasy Mini Plus Kit (Qiagen; Cat#74134) as detailed below. RNA samples were sent to the Genetic and Molecular Epidemiology Laboratory, David Braley Research Institute (Hamilton, Canada) for human Clariom S Assay (Thermo Fisher; Cat# 902926) with standard input (“Plus” assay). Raw data files were obtained to be viewed using Transcriptome Analysis Console Software (Thermo Fisher).
Microarray data discussed in this publication is deposited in NCBI’s Gene Expression Omnibus82 and are accessible through GEO Series accession number GSE18461283.
Gene ontology (GO) analysis was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) online bioinformatics program (https://david.abcc.ncifcrf.gov84,85). Gene set enrichment analysis (GSEA) was performed with the GSEA software (http://www.broadinstitute.org/gsea/86,87). Enrichment score (ES) is the degree to which a gene set is over-represented in the expression dataset. Normalized enrichment score (NES) is the enrichment score that has been normalized across analyzed gene sets. Venn diagrams were created with InteractiVenn (http://www.interactivenn.net/index.html88) and are not plotted to scale.
Cells were plated at a density of 40,000 cells/cm2 on 8-chambered Nunc Lab-Tek Chamber Slide System slides (Thermo Fisher; Cat# 177402). Cells were allowed to attach for 24 h. After various treatments, cells were fixed in cold methanol for 10 min on ice. Cells were then permeabilized with 0.25% Triton X-100 (VWR; Cat# VWRVM143) in PBS before incubating with primary antibody against TLE1 (Abcam; ab183742; 1:200 dilution). Following 1 h of primary antibody incubation, FITC-conjugated secondary antibody (Vector Laboratories; Cat# FI-1000) was applied for 1 h at room temperature. Slides were mounted using a medium containing DAPI (Sigma-Aldrich; Cat# F6057). Images were taken using an Olympus BX-51.
To detect intracellular lipid accumulation in cells, HCS LipidTOX Green neutral lipid stain (Thermo Fisher; Cat# H34475) was used. After various treatments, cells were fixed with 3–4% paraformaldehyde (Electron Microscopy Sciences; Cat# 15710) for 10 min. LipidTOX staining solution was then added for 30 min. Cells were counterstained with DAPI. The number of adipocytes and lipid droplets was measured by CellProfiler software89.
To detect osteoblast differentiation and mineralization, Alizarin Red S staining was performed. Cells were fixed in 10% neutral buffered formalin for 30 min. A 2% Alizarin Red S solution (Sigma-Aldrich; Cat# A5533) was added for 5 min. Slides were then cleared in 50% xylene and 50% acetone. Slides were mounted for imaging using a water-based mounting media. To quantify the staining, a 10% acetic acid solution was added to each well and heated at 85 °C for 10 min. The solution was allowed to cool, and samples were centrifuged at 20,000 x g for 15 min. The supernatant was then transferred to a new tube and sodium hydroxide solution was added to neutralize the acid to a pH between 4.1 and 4.5. Absorbance was measured at 405 nm using Thermo Scientific Multiskan FC Microplate Photometer (Thermo Scientific).
All data were expressed as Mean ± standard deviation (SD). A single sex and starting age of the animals was studied. All tissues were analyzed from mice, without exclusion. For cell-based studies, all experiments were repeated in at least 4 biological replicates (donors). Small ‘n’ represents an independent sample, not a technical replicate or an additional reading of the same sample. An a priori analysis was not performed. Sample size for cell culture studies was based on our previous publications6,65. Sample size for in vivo studies was based on previous studies showing gene expression changes90 and target organ cell composition changes91 in diabetes. Statistical and graphical analyses were performed using GraphPad Prism 7 and Microsoft Excel (basic statistical functions for descriptive statistics only). Data were tested for normality. When comparing parametric data from two groups, a two-tailed Student’s unpaired t-test was used. For multiple comparisons, an analysis of variance (ANOVA) followed by Bonferroni post hoc analysis was performed. P values <0.05 were considered statistically significant.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
All data figures show raw/individual data points. Source data for the main figures is provided as Supplemental Data 1. Gene profiling data is deposited in public repository and available to researchers (GSE184612). Data for Supplementary Figures are available from the corresponding authors on reasonable request. Reagents and other research material used in this study, if not depleted, will also be provided to other researchers upon request.
Ogurtsova, K. et al. IDF Diabetes Atlas: Global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin. Pr. 128, 40–50 (2017).
CAS Google Scholar
Yellowlees Douglas, J. et al. Bone marrow-CNS connections: implications in the pathogenesis of diabetic retinopathy. Prog. Retin Eye Res 31, 481–494 (2012).
CAS PubMed PubMed Central Google Scholar
Bhatwadekar, A. D. et al. Hematopoietic stem/progenitor involvement in retinal microvascular repair during diabetes: Implications for bone marrow rejuvenation. Vis. Res 139, 211–220 (2017).
PubMed Google Scholar
Santopaolo, M., Sambataro, M., Spinetti, G. & Madeddu, P. Bone marrow as a target and accomplice of vascular complications in diabetes. Diabetes Metab. Res Rev. 36(Suppl 1), e3240 (2020).
PubMed Google Scholar
Oikawa, A. et al. Diabetes mellitus induces bone marrow microangiopathy. Arterioscler Thromb. Vasc. Biol. 30, 498–508 (2010).
CAS PubMed Google Scholar
Keats, E. C., Dominguez, J. M. 2nd, Grant, M. B. & Khan, Z. A. Switch from canonical to noncanonical Wnt signaling mediates high glucose-induced adipogenesis. Stem Cells 32, 1649–1660 (2014).
CAS PubMed Google Scholar
Rubin, M. R. Skeletal fragility in diabetes. Ann. N. Y Acad. Sci. 1402, 18–30 (2017).
CAS PubMed Google Scholar
Fadini, G. P., Ferraro, F., Quaini, F., Asahara, T. & Madeddu, P. Concise review: diabetes, the bone marrow niche, and impaired vascular regeneration. Stem Cells Transl. Med 3, 949–957 (2014).
CAS PubMed PubMed Central Google Scholar
Hoyer, F. F. et al. Bone marrow endothelial cells regulate myelopoiesis in diabetes mellitus. Circulation 142, 244–258 (2020).
CAS PubMed PubMed Central Google Scholar
Tepper, O. M. et al. Decreased circulating progenitor cell number and failed mechanisms of stromal cell-derived factor-1alpha mediated bone marrow mobilization impair diabetic tissue repair. Diabetes 59, 1974–1983 (2010).
CAS PubMed PubMed Central Google Scholar
Nagareddy, P. R. et al. Hyperglycemia promotes myelopoiesis and impairs the resolution of atherosclerosis. Cell Metab. 17, 695–708 (2013).
CAS PubMed PubMed Central Google Scholar
Albiero, M. et al. Diabetes causes bone marrow autonomic neuropathy and impairs stem cell mobilization via dysregulated p66Shc and Sirt1. Diabetes 63, 1353–1365 (2014).
CAS PubMed Google Scholar
Barrett, T. J., Murphy, A. J., Goldberg, I. J. & Fisher, E. A. Diabetes-mediated myelopoiesis and the relationship to cardiovascular risk. Ann. N. Y Acad. Sci. 1402, 31–42 (2017).
PubMed PubMed Central Google Scholar
Albiero, M. & Fadini, G. P. Pharmacologic targeting of the diabetic stem cell mobilopathy. Pharm. Res 135, 18–24 (2018).
CAS Google Scholar
Fadini, G. P., Agostini, C. & Avogaro, A. Characterization of endothelial progenitor cells. Biochem Biophys. Res Commun. 336, 1–2 (2005).
CAS PubMed Google Scholar
Fadini, G. P. et al. Number and function of endothelial progenitor cells as a marker of severity for diabetic vasculopathy. Arterioscler. Thromb. Vasc. Biol. 26, 2140–2146 (2006).
CAS PubMed Google Scholar
Loomans, C. J. et al. Endothelial progenitor cell dysfunction: a novel concept in the pathogenesis of vascular complications of type 1 diabetes. Diabetes 53, 195–199 (2004).
CAS PubMed Google Scholar
Vasa, M. et al. Number and migratory activity of circulating endothelial progenitor cells inversely correlate with risk factors for coronary artery disease. Circ. Res 89, E1–E7 (2001).
CAS PubMed Google Scholar
Spinetti, G. et al. Global remodeling of the vascular stem cell niche in bone marrow of diabetic patients: implication of the microRNA-155/FOXO3a signaling pathway. Circ. Res 112, 510–522 (2013).
CAS PubMed Google Scholar
Naveiras, O. et al. Bone-marrow adipocytes as negative regulators of the haematopoietic microenvironment. Nature 460, 259–263 (2009).
CAS PubMed PubMed Central Google Scholar
Moitra, J. et al. Life without white fat: a transgenic mouse. Genes Dev. 12, 3168–3181 (1998).
CAS PubMed PubMed Central Google Scholar
Yang, Y. et al. Decrease in retinal neuronal cells in streptozotocin-induced diabetic mice. Mol. Vis. 18, 1411–1420 (2012).
PubMed PubMed Central Google Scholar
Martin, P. M., Roon, P., Van Ells, T. K., Ganapathy, V. & Smith, S. B. Death of retinal neurons in streptozotocin-induced diabetic mice. Invest Ophthalmol. Vis. Sci. 45, 3330–3336 (2004).
PubMed Google Scholar
Kuiper, E. J. et al. Connective tissue growth factor is necessary for retinal capillary basal lamina thickening in diabetic mice. J. Histochem Cytochem. 56, 785–792 (2008).
CAS PubMed PubMed Central Google Scholar
Okada, S. et al. Intercellular adhesion molecule-1-deficient mice are resistant against renal injury after induction of diabetes. Diabetes 52, 2586–2593 (2003).
CAS PubMed Google Scholar
Sugimoto, H., Grahovac, G., Zeisberg, M. & Kalluri, R. Renal fibrosis and glomerulosclerosis in a new mouse model of diabetic nephropathy and its regression by bone morphogenic protein-7 and advanced glycation end product inhibitors. Diabetes 56, 1825–1833 (2007).
CAS PubMed Google Scholar
Chandramouli, C. et al. Diastolic dysfunction is more apparent in STZ-induced diabetic female mice, despite less pronounced hyperglycemia. Sci. Rep. 8, 2346 (2018).
PubMed PubMed Central Google Scholar
Scheller, E. L. et al. Region-specific variation in the properties of skeletal adipocytes reveals regulated and constitutive marrow adipose tissues. Nat. Commun. 6, 7808 (2015).
CAS PubMed PubMed Central Google Scholar
Ban, C. R. & Twigg, S. M. Fibrosis in diabetes complications: pathogenic mechanisms and circulating and urinary markers. Vasc. Health Risk Manag 4, 575–596 (2008).
CAS PubMed PubMed Central Google Scholar
Khan, Z. A. & Chakrabarti, S. Therapeutic targeting of endothelial dysfunction in chronic diabetic complications. Recent Pat. Cardiovasc Drug Discov. 1, 167–175 (2006).
CAS PubMed Google Scholar
Tsalamandris, S. et al. The role of inflammation in diabetes: current concepts and future perspectives. Eur. Cardiol. 14, 50–59 (2019).
PubMed PubMed Central Google Scholar
Aung, M. H., Kim, M. K., Olson, D. E., Thule, P. M. & Pardue, M. T. Early visual deficits in streptozotocin-induced diabetic long evans rats. Invest Ophthalmol. Vis. Sci. 54, 1370–1377 (2013).
PubMed PubMed Central Google Scholar
Ferrannini, E., Barrett, E. J., Bevilacqua, S. & DeFronzo, R. A. Effect of fatty acids on glucose production and utilization in man. J. Clin. Invest 72, 1737–1747 (1983).
CAS PubMed PubMed Central Google Scholar
Foley, J. E. Mechanisms of impaired insulin action in isolated adipocytes from obese and diabetic subjects. Diabetes Metab. Rev. 4, 487–505 (1988).
CAS PubMed Google Scholar
Greenberg, A. S. et al. Perilipin, a major hormonally regulated adipocyte-specific phosphoprotein associated with the periphery of lipid storage droplets. J. Biol. Chem. 266, 11341–11346 (1991).
CAS PubMed Google Scholar
Tratwal, J. et al. MarrowQuant across aging and Aplasia: A digital pathology workflow for quantification of bone marrow compartments in histological sections. Front Endocrinol. 11, 480 (2020).
Google Scholar
Li, Z., Hardij, J., Bagchi, D. P., Scheller, E. L. & MacDougald, O. A. Development, regulation, metabolism and function of bone marrow adipose tissues. Bone 110, 134–140 (2018).
CAS PubMed PubMed Central Google Scholar
Komori, T. Regulation of osteoblast differentiation by Runx2. Adv. Exp. Med Biol. 658, 43–49 (2010).
CAS PubMed Google Scholar
Martin, L. M. & McCabe, L. R. Type I diabetic bone phenotype is location but not gender-dependent. Histochem. Cell Biol. 128, 125–133 (2007).
CAS PubMed Google Scholar
Botolin, S. & McCabe, L. R. Inhibition of PPARgamma prevents type I diabetic bone marrow adiposity but not bone loss. J. Cell Physiol. 209, 967–976 (2006).
CAS PubMed Google Scholar
Botolin, S. & McCabe, L. R. Bone loss and increased bone adiposity in spontaneous and pharmacologically induced diabetic mice. Endocrinology 148, 198–205 (2007).
CAS PubMed Google Scholar
Stahl, A., Evans, J. G., Pattel, S., Hirsch, D. & Lodish, H. F. Insulin causes fatty acid transport protein translocation and enhanced fatty acid uptake in adipocytes. Dev. Cell 2, 477–488 (2002).
CAS PubMed Google Scholar
Wu, Q. et al. FATP1 is an insulin-sensitive fatty acid transporter involved in diet-induced obesity. Mol. Cell Biol. 26, 3455–3467 (2006).
CAS PubMed PubMed Central Google Scholar
Laviola, L., Perrini, S., Cignarelli, A. & Giorgino, F. Insulin signalling in human adipose tissue. Arch. Physiol. Biochem 112, 82–88 (2006).
CAS PubMed Google Scholar
Oikawa, E. et al. A novel acyl-CoA synthetase, ACS5, expressed in intestinal epithelial cells and proliferating preadipocytes. J. Biochem 124, 679–685 (1998).
CAS PubMed Google Scholar
Gao, H. et al. CD36 is a marker of human adipocyte progenitors with pronounced adipogenic and triglyceride accumulation potential. Stem Cells 35, 1799–1814 (2017).
CAS PubMed Google Scholar
Willecke, F. et al. Lipolysis, and not hepatic lipogenesis, is the primary modulator of triglyceride levels in streptozotocin-induced diabetic mice. Arterioscler Thromb. Vasc. Biol. 35, 102–110 (2015).
CAS PubMed Google Scholar
Morikawa, S. et al. Prospective identification, isolation, and systemic transplantation of multipotent mesenchymal stem cells in murine bone marrow. J. Exp. Med 206, 2483–2496 (2009).
CAS PubMed PubMed Central Google Scholar
Ambrosi, T. H. et al. Adipocyte Accumulation in the Bone Marrow during Obesity and Aging Impairs Stem Cell-Based Hematopoietic and Bone Regeneration. Cell Stem Cell 20, 771–784 e776 (2017).
CAS PubMed PubMed Central Google Scholar
Zhong, L., et al. Single cell transcriptomics identifies a unique adipose lineage cell population that regulates bone marrow environment. Elife 9, e54695 (2020).
Ortega, G., Korty, P. E., Shevach, E. M. & Malek, T. R. Role of Ly-6 in lymphocyte activation. I. Characterization of a monoclonal antibody to a nonpolymorphic Ly-6 specificity. J. Immunol. 137, 3240–3246 (1986).
CAS PubMed Google Scholar
Abou-Ezzi, G. et al. TGF-beta signaling plays an essential role in the lineage specification of mesenchymal stem/progenitor cells in fetal bone marrow. Stem Cell Rep. 9, 48–60 (2019).
Thompson, N. L. et al. Expression of transforming growth factor-beta 1 in specific cells and tissues of adult and neonatal mice. J. Cell Biol. 108, 661–669 (1989).
CAS PubMed Google Scholar
Zhao, M. et al. Megakaryocytes maintain homeostatic quiescence and promote post-injury regeneration of hematopoietic stem cells. Nat. Med 20, 1321–1326 (2014).
CAS PubMed Google Scholar
Kubota, N. et al. Disruption of adiponectin causes insulin resistance and neointimal formation. J. Biol. Chem. 277, 25863–25866 (2002).
CAS PubMed Google Scholar
Sugiyama, T., Kohara, H., Noda, M. & Nagasawa, T. Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity 25, 977–988 (2006).
CAS PubMed Google Scholar
Omatsu, Y. et al. The essential functions of adipo-osteogenic progenitors as the hematopoietic stem and progenitor cell niche. Immunity 33, 387–399 (2010).
CAS PubMed Google Scholar
Ding, L., Saunders, T. L., Enikolopov, G. & Morrison, S. J. Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 481, 457–462 (2012).
CAS PubMed PubMed Central Google Scholar
Ding, L. & Morrison, S. J. Haematopoietic stem cells and early lymphoid progenitors occupy distinct bone marrow niches. Nature 495, 231–235 (2013).
CAS PubMed PubMed Central Google Scholar
Luo, K. Signaling cross talk between TGF-beta/Smad and other signaling pathways. Cold Spring Harb. Perspect. Biol. 9, a022137 (2017).
Ferland-McCollough, D. et al. MCP-1 feedback loop between adipocytes and mesenchymal stromal cells causes fat accumulation and contributes to hematopoietic stem cell rarefaction in the bone marrow of patients with diabetes. Diabetes 67, 1380–1394 (2018).
CAS PubMed Google Scholar
Ignotz, R. A. & Massague, J. Type beta transforming growth factor controls the adipogenic differentiation of 3T3 fibroblasts. Proc. Natl Acad. Sci. USA 82, 8530–8534 (1985).
CAS PubMed PubMed Central Google Scholar
Marquez, M. P. et al. The role of cellular proliferation in adipogenic differentiation of human adipose tissue-derived mesenchymal stem cells. Stem Cells Dev. 26, 1578–1595 (2017).
CAS PubMed PubMed Central Google Scholar
Choy, L. & Derynck, R. Transforming growth factor-beta inhibits adipocyte differentiation by Smad3 interacting with CCAAT/enhancer-binding protein (C/EBP) and repressing C/EBP transactivation function. J. Biol. Chem. 278, 9609–9619 (2003).
CAS PubMed Google Scholar
Keats, E. & Khan, Z. A. Unique responses of stem cell-derived vascular endothelial and mesenchymal cells to high levels of glucose. PLoS One 7, e38752 (2012).
CAS PubMed PubMed Central Google Scholar
Zhao, L. & Hantash, B. M. TGF-beta1 regulates differentiation of bone marrow mesenchymal stem cells. Vitam. Horm. 87, 127–141 (2011).
CAS PubMed Google Scholar
Motyl, K. & McCabe, L. R. Streptozotocin, type I diabetes severity and bone. Biol. Proced. Online 11, 296–315 (2009).
CAS PubMed PubMed Central Google Scholar
Fazeli, P. K. et al. Marrow fat and bone–new perspectives. J. Clin. Endocrinol. Metab. 98, 935–945 (2013).
CAS PubMed PubMed Central Google Scholar
Lecka-Czernik, B. et al. Divergent effects of selective peroxisome proliferator-activated receptor-gamma 2 ligands on adipocyte versus osteoblast differentiation. Endocrinology 143, 2376–2384 (2002).
CAS PubMed Google Scholar
Tedesco, S. et al. Pharmacologic PPAR-gamma activation reprograms bone marrow macrophages and partially rescues HSPC mobilization in human and murine diabetes. Diabetes 69, 1562–1572 (2020).
PubMed Google Scholar
Strickland, J., Garrison, D. & Copple, B. L. Hypoxia upregulates Cxcl12 in hepatocytes by a complex mechanism involving hypoxia-inducible factors and transforming growth factor-beta. Cytokine 127, 154986 (2020).
CAS PubMed PubMed Central Google Scholar
Aiuti, A., Webb, I. J., Bleul, C., Springer, T. & Gutierrez-Ramos, J. C. The chemokine SDF-1 is a chemoattractant for human CD34+ hematopoietic progenitor cells and provides a new mechanism to explain the mobilization of CD34+ progenitors to peripheral blood. J. Exp. Med 185, 111–120 (1997).
CAS PubMed PubMed Central Google Scholar
Kim, C. H. & Broxmeyer, H. E. In vitro behavior of hematopoietic progenitor cells under the influence of chemoattractants: stromal cell-derived factor-1, steel factor, and the bone marrow environment. Blood 91, 100–110 (1998).
CAS PubMed Google Scholar
Peled, A. et al. The chemokine SDF-1 activates the integrins LFA-1, VLA-4, and VLA-5 on immature human CD34(+) cells: role in transendothelial/stromal migration and engraftment of NOD/SCID mice. Blood 95, 3289–3296 (2000).
CAS PubMed Google Scholar
Peled, A. et al. Dependence of human stem cell engraftment and repopulation of NOD/SCID mice on CXCR4. Science 283, 845–848 (1999).
CAS PubMed Google Scholar
Ferraro, F. et al. Diabetes impairs hematopoietic stem cell mobilization by altering niche function. Sci. Transl. Med 3, 104ra101 (2011).
PubMed PubMed Central Google Scholar
Bennett, C. N. et al. Regulation of Wnt signaling during adipogenesis. J. Biol. Chem. 277, 30998–31004 (2002).
CAS PubMed Google Scholar
Like, A. A. & Rossini, A. A. Streptozotocin-induced pancreatic insulitis: new model of diabetes mellitus. Science 193, 415–417 (1976).
CAS PubMed Google Scholar
Bankhead, P. et al. QuPath: Open source software for digital pathology image analysis. Sci. Rep. 7, 16878 (2017).
PubMed PubMed Central Google Scholar
Galarraga, M. et al. Adiposoft: automated software for the analysis of white adipose tissue cellularity in histological sections. J. Lipid Res 53, 2791–2796 (2012).
CAS PubMed PubMed Central Google Scholar
Tratwal, J. et al. Reporting Guidelines, Review of Methodological Standards, and Challenges Toward Harmonization in Bone Marrow Adiposity Research. Report of the Methodologies Working Group of the International Bone Marrow Adiposity Society. Front Endocrinol. (Lausanne) 11, 65 (2020).
Google Scholar
Edgar, R., Domrachev, M. & Lash, A. E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30, 207–210 (2002).
CAS PubMed PubMed Central Google Scholar
Kum, J. J., Howlett, C. J. & Khan, Z. A. TGFB in bone marrow dysfunction in diabetes. Gene Expression Omnibus GSE184612 (2022).
Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44–57 (2009).
PubMed Google Scholar
Huang da, W., Sherman, B. T. & Lempicki, R. A. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37, 1–13 (2009).
PubMed Google Scholar
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).
CAS PubMed PubMed Central Google Scholar
Mootha, V. K. et al. PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat. Genet 34, 267–273 (2003).
CAS PubMed Google Scholar
Heberle, H., Meirelles, G. V., da Silva, F. R., Telles, G. P. & Minghim, R. InteractiVenn: a web-based tool for the analysis of sets through Venn diagrams. BMC Bioinforma. 16, 169 (2015).
Google Scholar
Adomshick, V., Pu, Y. & Veiga-Lopez, A. Automated lipid droplet quantification system for phenotypic analysis of adipocytes using CellProfiler. Toxicol. Mech. Methods 30, 378–387 (2020).
CAS PubMed PubMed Central Google Scholar
Feit-Leichman, R. A. et al. Vascular damage in a mouse model of diabetic retinopathy: relation to neuronal and glial changes. Invest Ophthalmol. Vis. Sci. 46, 4281–4287 (2005).
PubMed Google Scholar
Kumar, S. & Zhuo, L. Longitudinal in vivo imaging of retinal gliosis in a diabetic mouse model. Exp. Eye Res 91, 530–536 (2010).
CAS PubMed Google Scholar
Download references
This work was supported by grants from the Canadian Diabetes Association (ZAK; OG-3-13-4034-ZK), Western Strategic Support (ZK), and Lawson Health Research Institute (ZAK). JJYK was funded, in part, through support from the Ontario Graduate Scholarship, Frederick Winnett Luney Graduate Scholarship, and Western’s Interdisciplinary Development Initiative in Stem Cells and Regenerative Medicine Graduate Student Training Award. Authors thank Ms. Linda Jackson-Boeters in the Department of Pathology and Laboratory Medicine at Western University for tissue embedding and sectioning. The authors also thank Dr. Biao Feng in the Department of Pathology and Laboratory Medicine at Western University for initiating diabetes and maintaining mice in the 2-month follow-up study.
Pathology and Laboratory Medicine, Schulich School of Medicine & Dentistry, Western University, London, ON, Canada
Jina J. Y. Kum, Christopher J. Howlett & Zia A. Khan
Pathology and Laboratory Medicine, London Health Sciences Centre, London, ON, Canada
Christopher J. Howlett
Lawson Research Institute, London, ON, Canada
Christopher J. Howlett & Zia A. Khan
Division of Genetics & Development, Children’s Health Research Institute, London, ON, Canada
Zia A. Khan
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
You can also search for this author in PubMed Google Scholar
All experiments included in this study were performed by J.J.Y.K. J.J.Y.K. also designed the experiments and analyzed the results. Z.A.K. and C.J.H. were involved in the conceptual design of the study, interpretation of the data, and editorial review. C.J.H. read all staining images for histopathological assessment. All authors were involved in writing of the manuscript and approve the final submission.
Correspondence to Christopher J. Howlett or Zia A. Khan.
The authors declare no competing interests.
All researchers involved in the study are listed as authors or acknowledged, as appropriate.
Communications Biology thanks Paolo Madeddu and Mattia Albiero for their contribution to the peer review of this work. Primary Handling Editor: Christina Karlsson Rosenthal. Peer reviewer reports are available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and Permissions
Kum, J.J.Y., Howlett, C.J. & Khan, Z.A. Dysregulated transforming growth factor-beta mediates early bone marrow dysfunction in diabetes. Commun Biol 5, 1145 (2022). https://doi.org/10.1038/s42003-022-04112-2
Download citation
Received: 23 March 2022
Accepted: 14 October 2022
Published: 28 October 2022
DOI: https://doi.org/10.1038/s42003-022-04112-2
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
Advertisement
© 2022 Springer Nature Limited
Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.