Deferoxamine preconditioning to restore impaired HIF-1α-mediated angiogenic mechanisms in adipose-derived stem cells from STZ-induced type 1 diabetic rats

Abstract

Objectives: Both excessive and insufficient angiogenesis are associated with progression of diabetic complications, of which poor angiogenesis is an important feature. Currently, adipose-derived stem cells (ADSCs) are considered to be a promising source to aid therapeutic neovascularization. However, functionality of these cells is impaired by diabetes which can result from a defect in hypoxia-inducible factor-1 (HIF-1), a key mediator involved in neovascularization. In the current study, we sought to explore effectiveness of pharmacological priming with deferoxamine (DFO) as a hypoxia mimetic agent, to restore the compromised angiogenic pathway, with the aid of ADSCs derived from streptozotocin (STZ)-induced type 1 diabetic rats ('diabetic ADSCs').

Materials and methods: Diabetic ADSCs were treated with DFO and compared to normal and non-treated diabetic ADSCs for expression of HIF-1α, VEGF, FGF-2 and SDF-1, at mRNA and protein levels, using qRT-PCR, western blotting and ELISA assay. Activity of matrix metalloproteinases -2 and -9 were measured using a gelatin zymography assay. Angiogenic potential of conditioned media derived from normal, DFO-treated and non-treated diabetic ADSCs were determined by in vitro (in HUVECs) and in vivo experiments including scratch assay, three-dimensional tube formation testing and surgical wound healing models.

Results: DFO remarkably enhanced expression of noted genes by mRNA and protein levels and restored activity of matrix metalloproteinases -2 and -9. Compromised angiogenic potential of conditioned medium derived from diabetic ADSCs was restored by DFO both in vitro and in vivo experiments.

Conclusion: DFO preconditioning restored neovascularization potential of ADSCs derived from diabetic rats by affecting the HIF-1α pathway.

Figure 1

Characterization and morphological assessment of ADSC s. (a) Flow cytometry showed that the ADSCs were negative for CD45, CD11b and CD31 but positive for CD44, CD73 and CD90. Data are shown as a histogram plot, with black representing isotype control and white demonstrating experiments. (b) Assessment of morphology of ADSCs demonstrated that (I) diabetic ADSCs had more flattened and rounder morphology than (II) normal ADSCs. Dashed lines indicate cells.

Figure 2

Colony formation and iron assessment of ADSC s. (a) Representative images of colonies of diabetic ADSCs (I) and normal ADSCs (II) by light microscopy. (b) Assessment of size (I) and number (II) of colonies derived from diabetic and normal ADSCs, by NIH imageing. (c) Assessment of iron content of normal and diabetic ADSCs. Data are represented as mean ± SEM (n = 3, **P < 0.01, ***P < 0.001 versus normal cells). Statistical significance was measured by unpaired Student's t‐test.

Figure 3

Effect of DFO on proliferation of ADSC s. (a) Viability of diabetic cells significantly was lower than viability of normal ADSCs after 72 h measured by MTT assay. Diabetic ADSCs were treated with 75 μm (b), 150 μm (c) and 300 μm (d) of DFO for 24, 48 and 72 h and their viability was measured by MTT assay. Data are represented as mean ± SEM (n = 5; *P < 0.05, **P < 0.01, ***P < 0.001 versus normal cells). Statistical significance was measured by repeated measures ANOVA.

Figure 4

Effect of DFO on HIF ‐1α expression at protein level. (a) Representative blot (I) and semi‐quantitative data (II) of HIF‐1α expression at protein level in normal and diabetic ADSCs and treated dADSCs at different concentrations of DFO after 24 h. (b) Representative blot (I) and semi‐quantitative data (II) of HIF‐1α expression at protein levels in treated dADSCs with DFO (150 μm) after different times. Data are represented as mean ± SEM (n = 3; *P < 0.05, ***P < 0.001 versus normal cells and ^## P < 0.01, ^### P < 0.001 versus diabetic cells and ^xxx P < 0.001 versus 300 μm DFO‐treated dADSCs). Statistical significance measured by one‐way ANOVA.

Figure 5

Effect of DFO on angiogenic gene expression. (a) Diabetic ADSCs were exposed to DFO (150 and 300 μm) for 24 h and mRNA expression levels of VEGF (a), SDF‐1(b) and FGF‐2 (c) genes determined by qRT‐PCR. (b) Secretion measurement of VEGF in CM of ADSCs after treatment with DFO by ELISA assay. (c) Representative blots and semi‐quantitative data of SDF‐1 and FGF‐2 expression at protein levels in non‐treated and treated ADSCs with different concentrations of DFO after 24 h. Data are represented as mean ± SEM (n = 3; *P < 0.05, **P < 0.01, ***P < 0.001 versus normal cells and ^# P < 0.05, ^## P < 0.01, ^### P < 0.001 versus diabetic cells and ^xxx P < 0.001 versus 300 μm DFO‐treated dADSCs). Statistical significance measured by one‐way ANOVA.

Figure 6

Assessment of gelatinases by gelatin zymography in ADSC s. Representative zymogram (a) and semi‐quantitative data of MMP‐2 (b) and MMP‐9 (c) activity in conditioned media derived from nADSCs, DFO‐treated and non‐treated dADSCs. The areas of protease activity appeared as white bands. Data are represented as mean ± SEM. Data are represented as mean ± SEM (n = 3; **P < 0.01, ***P < 0.001 versus normal cells and ^### P < 0.001 versus diabetic cells and ^xxx P < 0.001 versus 300 μm DFO‐treated dADSCs). Statistical significant differences by one‐way ANOVA.

Figure 7

Assessment of angiogenic potential of conditioned media derived from ADSC s. (a) Representative image of migration of HUVECs cultured in different media (I) medium without FBS, (II) CM derived from normal ADSCs, (III) diabetic ADSCs, (IV) 150 μm and (V) 300 μm DFO‐treated ADSCs by inverted microscopy after 24 h. (b) Assessment of areas free of HUVECs by NIH image software, in five separated fields. (c) Sprout formation of HUVECs in (I) serum‐free medium, (II) CM derived from nADSCs, (III) dADSCs, (IV) 150 μm and (V) 300 μm DFO‐treated dADSCs were monitored by inverted microscopy after 48 h. (d) Change in sprout formation of HUVECs determined by NIH image. Data represented as mean ± SEM (n = 3; **P < 0.01, ***P < 0.001 versus normal cells and ^## P < 0.01, ^### P < 0.001 versus diabetic cells). Statistical significance measured by one‐way ANOVA.

Figure 8

Assessment of rat wound closure rate after treatment with different conditioned media derived from ADSC s. Following creating 8 mm full‐thickness excisional skin wounds, wounds were injected with 100 μl concentrated medium and concentrated CM derived from ADSCs. Yellow dash lines indicate margins of primary wounds. (a) Representative gross picture of healing process of wounds injected with conditioned media derived from non‐treated diabetic ADSCs (control) and 300 μm DFO‐treated dADSCs (DFO‐treated) after 0, 7, 10, 12 and 14 days. (b) Quantification of wound closure rate in all groups using NIH ImageJ software. CM derived from normal and 300 μm DFO‐treated ADSCs significantly increased wound healing rates compared to CM derived from dADSC groups over 10 days. Data are represented as mean ± SEM. (n = 3; *P < 0.05, **P < 0.01 versus normal cells, ^# P < 0.05 versus diabetic cells). Statistical significance by two‐way ANOVA.

Figure 9

Assessment of rat wound angiogenesis after treatment with different conditioned media derived from ADSC s. (a) Haematoxylin and eosin‐stained histological sections for assessment of angiogenesis in wounds treated with serum‐free medium (I&V), CM derived from (II&VI) nADSCs, (III&VII) dADSCs and (IV&VIII) 300 μm DFO‐treated dADSCs are shown after 3 and 7 days respectively. Arrows denote blood vessels. (b) Comparison between angiogenic scores in groups after 3 and 7 days. Data are represented as mean ± SEM. (n = 3; ***P < 0.001 versus normal cells). Statistical significance measured by two‐way ANOVA.

Figure 10

Effect of CM derived from DFO pre‐treated cells on epitheliarization. Epitheliarization was evaluated using histological analysis after H&E staining at x100 magnification by light microscopy. Arrows denote epithelium. Wounds injected with conditioned media derived from normal ADSCs (b) and 300 μm DFO‐treated diabetic ADSCs (d) had higher tissue epitheliarization compared to those injected with CM derived from serum‐free medium (a) and non‐treated diabetic ADSCs (c).

Figure 11

Effect of CM derived from DFO pre‐treated cells on collagenization. Collagenization was evaluated using Masson's trichrome staining at x400 magnification, by light microscopy. Collagen fibres are stained blue and arrows denote collagen fibres. Wounds injected with conditioned media derived from normal ADSCs (b) and 300 μm DFO‐treated diabetic ADSCs (d) had higher tissue collagenization compared to those injected with CM derived from non‐treated diabetic ADSCs (c) and serum‐free medium (a).