Impaired therapeutic capacity of autologous stem cells in a model of type 2 diabetes

Abstract

Endogenous stem cells in the bone marrow respond to environmental cues and contribute to tissue maintenance and repair. In type 2 diabetes, a multifaceted metabolic disease characterized by insulin resistance and hyperglycemia, major complications are seen in multiple organ systems. To evaluate the effects of this disease on the endogenous stem cell population, we used a type 2 diabetic mouse model (db/db), which recapitulates these diabetic phenotypes. Bone marrow-derived mesenchymal stem cells (MSCs) from db/db mice were characterized in vitro using flow cytometric cell population analysis, differentiation, gene expression, and proliferation assays. Diabetic MSCs were evaluated for their therapeutic potential in vivo using an excisional splint wound model in both nondiabetic wild-type and diabetic mice. Diabetic animals possessed fewer MSCs, which were proliferation and survival impaired in vitro. Examination of the recruitment response of stem and progenitor cells after wounding revealed that significantly fewer endogenous MSCs homed to the site of injury in diabetic subjects. Although direct engraftment of healthy MSCs accelerated wound closure in both healthy and diabetic subjects, diabetic MSC engraftment produced limited improvement in the diabetic subjects and could not produce the same therapeutic outcomes as in their nondiabetic counterparts in vivo. Our data reveal stem cell impairment as a major complication of type 2 diabetes in mice and suggest that the disease may stably alter endogenous MSCs. These results have implications for the efficiency of autologous therapies in diabetic patients and identify endogenous MSCs as a potential therapeutic target.

Figure 1.

Phenotypic characterization of diabetic and nondiabetic mice. (A): Age-matched 8-week-old nondiabetic (left; wild-type C57BLKS) and db/db (right; BKS.Cg-Dock7m +/+ Leprdb/J) mice are shown. The db/db mice exhibited hyperglycemia and obesity. Evaluation of the phenotype was completed using weight recordings (B) and plasma glucose levels (C) (both taken weekly at the same time), from wild-type and db/db mice using the tail vein blood. Data are expressed as the mean ± SEM (n = 5 per group; **, p < .001).

Figure 2.

Population analysis of endogenous MSCs in vitro. (A): MSCs were evaluated using flow cytometric analysis of the total number of diabetic (db) MSCs and wild-type (wt) MSCs with positive MSC markers CD29, CD44, CD90, and CD166 from passage 0 to passage 8. Each data point represents the pooled mean of the percentage of positively labeled MSCs markers. All positively labeled MSCs were negative for hematopoietic lineage-restricted markers CD14, CD34, and CD45. All mouse MSCs were derived in-house from db/db (n = 10) and WT (n = 10) mice. (B): The total number of cells and the time to reach a level of confluence appropriate for passaging (80% confluence) after each passage were recorded. A trypan blue exclusion assay was implemented to quantify viable cells. Each data point represents the total number of cells (solid lines, open symbols) and the number of weeks between passages (dashed lines, filled symbols). At early passages, dbMSCs had a lower proportion of viable cells and required longer to reach confluence. However, with time in standard culture conditions, the performance of the remaining dbMSCs largely matched that of wtMSCs. Assessment of MSC proliferation and survival revealed decreased efficiency of dbMSCs. (C): Proliferation and survival were assessed at passages 2 and 8 with BrdU. Proliferation of dbMSCs was significantly impaired at passage 2 compared with wtMSCs. However, by passage 8, dbMSCs increased to levels of WT proliferation (*, p < .05; **, p < .001). (D): Cultures at passages 2 and 8 were incubated with the nucleic acid marker Sytox Green, and the total number of cells in the same well was quantified. At both passage 2 and passage 8, dbMSCs had significantly fewer surviving cells despite an improvement in cell survival over time. Each column represents the total number of counted cells (error bars indicate ±SEM); n = 6 wells per group (**, p < .001). Abbreviations: BrdU, 5-bromo-2′-deoxyuridine; MSC, mesenchymal stem cell; WT, wild-type.

Figure 3.

Reverse transcription-polymerase chain reaction analysis of endogenous MSCs in vitro. WT and dbMSCs were analyzed at passage 4 for gene expression of VEGF (white) and Wnt3a (black). Although there was no difference between VEGF levels, there was a significant difference in the amount of Wnt3a between the two groups (error bars indicate ±SEM; *, p < .05). Abbreviations: MSC, mesenchymal stem cell; VEGF, vascular endothelial growth factor; WT, wild-type.

Figure 4.

Quantification of endogenous stem cell populations after wounding. (A): An excisional wound model using a donut-shaped splint kept the wounds open, and phosphate-buffered saline (vehicle-only) or cells were delivered to db/db or WT mice. We recorded the baseline levels of healing with no therapeutic interference. (B–D): Population analysis of diabetic and WT MSCs from dissociated wound beds from db/db and WT mice with no grafted MSCs using flow cytometric analysis using the MSC markers CD29 and CD90 and the endothelial progenitor marker CD34, at 1 (B), 5 (C), and 10 (D) days postwounding (n = 5 per time point; *, p < .05; **, p < .001). (E, F): Reverse transcription-polymerase chain reaction analysis was performed on naïve wound bed biopsies, and the levels of VEGF (E) and Wnt3a (F) were quantified at days 1, 5, and 10 from WT and db/db wound beds. There was a significant decrease in the level of Wnt3a expression in the db/db wound at days 1, 5, and 10, and by day 10 there was a significant decrease in the level of VEGF expression in the db/db wound (*, p < .05). Abbreviations: VEGF, vascular endothelial growth factor; WT, wild-type.

Figure 5.

SDF-1 expression in WT and diabetic (db) MSCs in vitro. WT (A–C) and dbMSCs (D–F) at passage 4 were also stained for positive MSC marker CD44 (Cy2, green) and SDF-1 (Cy3, red). A greater percentage of wtMSCs were SDF-1 positive compared with the dbMSCs in culture. Abbreviations: SDF-1, stromal-derived factor-1; WT, wild-type.

Figure 6.

Wound healing was impaired in diabetic animals, and dbMSCs had reduced therapeutic capacity. The progress of wound closure of dorsal, full-thickness excisional wounds held open by sutured (black threads) silicone splints (orange) following delivery of vehicle (PBS) to DB or nondiabetic (WT) subjects was recorded. Control WT mice (A–D) and db/db mice (E–H) received PBS (vehicle only). (I): db/db mice receiving only PBS exhibited significantly delayed wound closure compared with WT mice. WT (J–M) and db/db (S–V) mice received 1 × 10⁶ wtMSCs derived from WT mice or WT (N–Q) and db/db (W–Z) mice received 1 × 10⁶ dbMSCs derived from db/db mice. (R): In the WT host, dbMSC engraftment did not improve WT closure compared with PBS, but wtMSC engraftment indicated improved closure in the WT mouse. (AA): In the db/db host, wtMSCs improved closure in db/db mouse, but dbMSC engraftment alone improved closure in db/db mice but not in WT mice. The extent of reepithelialization and granular tissue formation was monitored daily; representative images are shown at initial treatment (day 0) and at 5, 10, and 15 days postengraftment (n = 10–13 per condition). Each data point represents the mean of the percentage of the area closed (I, R, AA). Closure was calculated from stereological analysis of micrographs using the Cavalieri point probe estimator (error bars indicate ±SEM; n = 10–13 per group; *, p < .05; **, p < .001). Abbreviations: DB, diabetic; dbMSC, diabetic mesenchymal stem cell; PBS, phosphate-buffered saline; WT, wild-type; wtMSC, wild-type mesenchymal stem cell.