Neovascularization capacity of mesenchymal stromal cells from critical limb ischemia patients is equivalent to healthy controls

Abstract

Critical limb ischemia (CLI) is often poorly treatable by conventional management and alternatives such as autologous cell therapy are increasingly investigated. Whereas previous studies showed a substantial impairment of neovascularization capacity in primary bone-marrow (BM) isolates from patients, little is known about dysfunction in patient-derived BM mesenchymal stromal cells (MSCs). In this study, we have compared CLI-MSCs to healthy controls using gene expression profiling and functional assays for differentiation, senescence and in vitro and in vivo pro-angiogenic ability. Whereas no differentially expressed genes were found and adipogenic and osteogenic differentiation did not significantly differ between groups, chondrogenic differentiation was impaired in CLI-MSCs, potentially as a consequence of increased senescence. Migration experiments showed no differences in growth factor sensitivity and secretion between CLI- and control MSCs. In a murine hind-limb ischemia model, recovery of perfusion was enhanced in MSC-treated mice compared to vehicle controls (71 ± 24% versus 44 ± 11%; P < 1 × 10(-6)). CLI-MSC- and control-MSC-treated animals showed nearly identical amounts of reperfusion (ratio CLI:Control = 0.98, 95% CI = 0.82-1.14), meeting our criteria for statistical equivalence. The neovascularization capacity of MSCs derived from CLI-patients is not compromised and equivalent to that of control MSCs, suggesting that autologous MSCs are suitable for cell therapy in CLI patients.

Figures

Figure 1

Gene expression. (a) Relative expression of genes in CLI-MSCs/Control MSCs. On this M/A plot, the x-axis denotes the average expression (in units of fluorescence intensity), the y-axis denotes the ratio of expression in CLI- MSCs/Control MSCs. Expression of genes shown in green is with 95% confidence intervals less than a factor differentially expressed. The remaining genes are inconclusive, genes with an average fold change of >2 are indicated in red and 95% confidence intervals are given. (b) Heatmap showing genes associated with angiogenesis, clustering on columns shows no clear separation between CLI- and Control MSCs.

Figure 2

Differentiation. (a) Quantification of alizarin red retention in osteogenically differentiated MSCs was found to be equivalent; CLI:Control = 0.99 (95% CI = 0.81–1.17), the open circle indicates and outlier that has been left out of analysis. (b) ALP activity in MSCs after osteogenic differentiation; CLI:Control = 1.07, 95% CI = 0.75–1.39. (c) Adipogenic Differentiation measured by Lipidtox Green staining, CLI:Control = 1.2, 95% CI = 0.71–1.49. (d) Chondrogenic differentiation measured by sGAG production. sGAGs levels were significantly lower in CLI-MSCs.

Figure 3

Senescence. (a) Representative example of SA-β-galactosidase activity as measured by flow cytometry. (b) Senescence is increased in CLI MSCs compared to Healthy controls (P = 0.01). (c) MSC senescence is significantly associated with donor age (P = 0.02), shaded band indicates 95% confidence interval.

Figure 4

Migration. (a) Real-time migration of MSCs towards different concentrations of PDGF-BB (lines for higher concentrations have been omitted for clarity). (b) Dose-response curve of MSCs to PDGF-BB. (c) EC50 values of dose-response curves of CLI- and control MSCs (CLI:Control = 1.05, 95% CI = 0.66–1.62). (d) Scratch wound closure of microvascular endothelial cells after exposure to MSC-conditioned medium. N and H indicate CM collected under normoxic and hypoxic culturing conditions, respectively. There is a significant effect of hypoxia (P = 0.02), but not of donor group (P = 0.33) The Negative control consists of empty medium, the positive control of medium containing 10% FCS. Error bars indicate SEM. (e) Population doubling times of HMEC-1 endothelial cells grown in MSC-CM (CLI: Control = 0.96, 95% CI = 0.47–2.12). (f) Number of junctions per image in tubule formation assay of HMEC-1 cells on matrigel, in the presence of MSC CM. No differences was observed between CLI and Control MSC CM or between CM collected under hypoxic conditions. N stands for normoxic, H for hypoxic CM.

Figure 5

Hind-limb ischemia model. (a) Hind-limb ischemia model and representative graphical readout. (b) Dose-response curve in of intramuscularly injected MSCs in the HLI model. (c) Relative perfusion over time of ischemic hind limbs over time in vehicle- and MSC-treated animals, MSC-treated animals show significantly higher perfusion than vehicle-treated animals (P < 1·10−6). Error lines indicate SEM. (d) AUCs of relative perfusion over time as a summary estimate for equivalence (CLI:Control 0.98, 95% CI = 0.82–1.14).

Figure 6

Histology. CD31 staining (red) in (a) CLI-MSC–treated and (b) vehicle-treated animals, perimysia are stained in green (scale bar = 100 µmol/l). (c) Number of vessels as ratio vessels:myofibrils; vehicle-treated animals show a significant reduction in the number of vessels compared to MSC-treated animals (P < 0.001). The number of vessels in CLI- and healthy MSC–treated animals is equivalent (CLI:Control 0.96, 95% CI = 0.83–1.09). (d) Staining with human specific anti A/C Lamin shows that injected MSCs are predominantly retained in localized clusters, presumably the injection site, after 14 days (scale bar = 100 µmol/l). (e) Representative image of aSMA (red) staining, CD31 staining is shown in green (scale bar = 100 µmol/l). (f) The number of aSMA+ vessels did not differ per group.