Vanadate Impedes Adipogenesis in Mesenchymal Stem Cells Derived from Different Depots within Bone

Abstract

Glucocorticoid-induced osteoporosis (GIO) is associated with an increase in bone marrow adiposity, which skews the differentiation of mesenchymal stem cell (MSC) progenitors away from osteoblastogenesis and toward adipogenesis. We have previously found that vanadate, a non-specific protein tyrosine phosphatase inhibitor, prevents GIO in rats, but it was unclear whether vanadate directly influenced adipogenesis in bone-derived MSCs. For the present study, we investigated the effect of vanadate on adipogenesis in primary rat MSCs derived from bone marrow (bmMSCs) and from the proximal end of the femur (pfMSCs). By passage 3 after isolation, both cell populations expressed the MSC cell surface markers CD90 and CD106, but not the hematopoietic marker CD45. However, although variable, expression of the fibroblast marker CD26 was higher in pfMSCs than in bmMSCs. Differentiation studies using osteogenic and adipogenic induction media (OM and AM, respectively) demonstrated that pfMSCs rapidly accumulated lipid droplets within 1 week of exposure to AM, while bmMSCs isolated from the same femur only formed lipid droplets after 3 weeks of AM treatment. Conversely, pfMSCs exposed to OM produced mineralized extracellular matrix (ECM) after 3 weeks, compared to 1 week for OM-treated bmMSCs. Vanadate (10 μM) added to AM resulted in a significant reduction in AM-induced intracellular lipid accumulation and expression of adipogenic gene markers (PPARγ2, aP2, adipsin) in both pfMSCs and bmMSCs. Pharmacological concentrations of glucocorticoids (1 μM) alone did not induce lipid accumulation in either bmMSCs or pfMSCs, but resulted in significant cell death in pfMSCs. Our findings demonstrate the existence of at least two fundamentally different MSC depots within the femur and highlights the presence of MSCs capable of rapid adipogenesis within the proximal femur, an area prone to osteoporotic fractures. In addition, our results suggest that the increased bone marrow adiposity observed in GIO may not be solely due to direct effect of glucocorticoids on bone-derived MSCs, and that an increase in femur lipid content may also arise from increased adipogenesis in MSCs residing outside of the bone marrow niche.

Keywords: adipogenesis; bone; glucocorticoids; mesenchymal stem cells; vanadate.

Figure 1

Characterization of bone cell isolates. (A) A scatter plot showing homogeneous populations of bone marrow- and proximal femur-derived cells. (B) Cell surface expression of the mesenchymal progenitor cell marker CD90 (column I), bmMSC marker CD106 (column II), fibroblast marker CD26 (column III), and the hematopoietic marker CD45 (column IV) was examined using flow cytometry. The first row (1) represents the negative (unstained) control samples, whereas the results for surface marker expression in bmMSCs and pfMSCs are shown in rows 2 and 3, respectively. Regions of positive staining are indicated to the right-hand side of the vertical line. Experiments were repeated in duplicate on MSCs derived from nine animals (n = 9), and the graphs show the results for one representative animal. For each experiment 15,000 events were recorded of which 10,000 were gated and analyzed using FlowJo Vx (Treestar) software.

Figure 2

Characterization of the osteoblastic differentiation potential of bmMSCs and pfMSCs. Cells were treated with osteogenic media (OM) as described under Section “Materials and Methods,” for the number of days as indicated on the figure, and subsequently stained with Alizarin Red S. Images show representative results of three independent experiments, and were taken at 10× magnification, with the size bar = 1 mm.

Figure 3

Characterization of the adipocytic differentiation potential of bmMSCs and pfMSCs. Cells were treated with adipogenic media (AM) as described under Section “Materials and Methods,” for the number of days as indicated on the figure, and subsequently stained with Oil Red O. Images show representative results of three independent experiments, and were taken at 20× magnification, with the size bar = 1 mm.

Figure 4

The effect of vanadate on lipid accumulation in bmMSCs and pfMSCs. Cells were treated with adipogenic media (AM) or AM plus 10 μM vanadate (AMV) for the number of days as indicated, before being stained with Oil Red O (ORO) and counterstained with crystal violet (CV). The ORO/CV ratio for control cells was set as 1. Different lower-case letters (a vs. b) indicates statistically significant differences (P < 0.05, with n = 4 for bmMSCs and n = 3 for pfMSCs).

Figure 5

The effect of vanadate on the expression of early markers of adipogenesis. (A,B) Cells were treated with adipogenic media (AM) or AM plus 10 μM vanadate (AMV) for the number of days as indicated, and the expression of PPARγ2 (A) and C/EBPα (B) were measured by qRT-PCR. All target gene measurements were normalized to the housekeeping gene acidic ribosomal phosphoprotein (ARBP). (A) As PPARγ2 expression was undetectable in Naïve cells, PPARγ2 expression in day 3 AM-treated pfMSCs or day 21 AM-treated bmMSCs was set as 1. The graph presents the combined data of n = 3 for bmMSCs and n = 3 for pfMSCs. (B) C/EBPα expression in control cells was set as 1. The graph presents representative data from one experimental repeat out of three repeats for bmMSCs and one out of four repeats for pfMSCs.

Figure 6

The effect of vanadate on the expression of late markers of adipogenesis. Cells were treated with adipogenic media (AM) or AM plus 10 μM vanadate (AMV) for the number of days as indicated, and the expression of adipocyte protein-2/fatty acid binding protein-4 (aP2/FABP4), adipsin, and fatty acid synthase (Fasn) was measured with qRT-PCR. All target gene measurements were normalized to the housekeeping gene acidic ribosomal phosphoprotein (ARBP). The expression of aP2/FABP4, adipsin, and Fasn in AM-treated cells was set as 1. The graph presents the combined data of n = 3 for bmMSCs and n = 4 for pfMSCs. Statistically significant differences are indicated as **P < 0.01 or ***P < 0.001.

Figure 7

Comparison of gene expression between bmMSCs and pfMSCs. The relative expression levels of adipogenic genes (A) and osteogenic genes (B) in Naïve bmMSCs and pfMSCs were measured by qRT-PCR. All target gene measurements were normalized to the housekeeping gene acidic ribosomal phosphoprotein (ARBP). For each target gene, the average expression in bmMSCs (n = 3) was set as 1, and the expression in pfMSCs (n = 4) calculated relative to expression in bmMSCs. (C) bmMSCs were treated with adipogenic media (AM) or AM plus 10 μM vanadate (AMV) as indicated, and the expression of Wnt10b was measured by qRT-PCR after 14 and 21 days. Expression in control cells was set as 1. Different lower-case letters (a vs. b vs. c) indicates statistically significant differences (P < 0.05, n = 3).

Figure 8

The effect of GC treatment on bmMSCs and pfMSCs. BmMSCs and pfMSCs were treated with 1 μM dexamethasone (Dex) for 21 days, and compared to vehicle-treated cells. Images show unstained cells at 20× magnification and are representative of three independent experiments. The size bar = 1 mm.

Figure 9

The effects of GCs and vanadate on cell viability and apoptosis in bmMSCs and pfMSCs. Naïve cells were treated with 1 μM dexamethasone (Dex) in the absence or presence of 10 μM vanadate. (A,B) Cells were stained with crystal violet (A) or MTT (B), and the staining was quantified spectrophotometrically. Absorbance measurements of control cells were set as 1. (C) The percentage of apoptotic cells in treated pfMSC cultures was determined with flow cytometry, using annexin V as an apoptotic marker. Statistically significant differences are indicated as *P < 0.05, **P < 0.01, or ***P < 0.001 (n = 3 for bmMSCs, n = 3 for pfMSCs).