TGF-β1 signaling and Krüppel-like factor 10 regulate bone marrow-derived proangiogenic cell differentiation, function, and neovascularization

Abstract

Emerging evidence demonstrates that proangiogenic cells (PACs) originate from the BM and are capable of being recruited to sites of ischemic injury where they contribute to neovascularization. We previously determined that among hematopoietic progenitor stem cells, common myeloid progenitors (CMPs) and granulocyte-macrophage progenitor cells (GMPs) differentiate into PACs and possess robust angiogenic activity under ischemic conditions. Herein, we report that a TGF-β1-responsive Krüppel- like factor, KLF10, is strongly expressed in PACs derived from CMPs and GMPs, ∼ 60-fold higher than in progenitors lacking PAC markers. KLF10(-/-) mice present with marked defects in PAC differentiation, function, TGF-β responsiveness, and impaired blood flow recovery after hindlimb ischemia, an effect rescued by wild-type PACs, but not KLF10(-/-) PACs. Overexpression studies revealed that KLF10 could rescue PAC formation from TGF-β1(+/-) CMPs and GMPs. Mechanistically, KLF10 targets the VEGFR2 promoter in PACs which may underlie the observed effects. These findings may be clinically relevant because KLF10 expression was also found to be significantly reduced in PACs from patients with peripheral artery disease. Collectively, these observations identify TGF-β1 signaling and KLF10 as key regulators of functional PACs derived from CMPs and GMPs and may provide a therapeutic target during cardiovascular ischemic states.

Figure 1

TGF-β1 regulates PAC differentiation and function. (A) VEGFR2 expression in BM-derived progenitors isolated from WT mice (A) or TGF-β1⁺/⁺ or TGF-β1^+/− mice (B) and grown in the presence or absence of TGF-β1 (A) for 7 days (n = 3 per group). Percentage of VEGFR2 expression was analyzed by flow cytometry. *P < .01 versus no TGF-β1; **P < .01 versus TGF-β1^+/−. (C) BM-derived progenitors grown in the presence of Ctrl (vehicle) or TGF-β1 and plated in fibronectin-coated wells were assessed for adhesion (n = 6 per group). HPF indicates high power field. *P < .05 versus Ctrl. (D) BM-derived GMPs and MEPs were grown in the presence or absence of TGF-β1 treatment were assessed for VEGFR2 mRNA expression by quantitative PCR (n = 3 per group). *P < .01 versus Ctrl; **P < .05 versus Ctrl. (E) TGF-β1–stimulated expression of the VEGFR2 promoter-luciferase reporter transfected in BM-derived GMPs and MEPs (n = 3 per group). *P < .01 versus Ctrl; **P < .05 versus Ctrl.

Figure 2

Identification of KLF10 expression in PACs and responsiveness to TGF-β1. (A-B) BM-derived progenitors grown in either hematopoietic IMDM medium or EGM-2 were harvested, and expression of Klf10 was examined by quantitative PCR (A) or Western blot analysis (B). *P < .01. β-actin was used as an internal loading control. (C-D) BM-derived progenitors grown in EGM-2 medium in the presence or absence of TGF-β1. Klf10 mRNA expression was examined in GMP-derived PACs by quantitative PCR (C). *P < .01 versus no TGF-β1. (D) VEGFR2 expression was examined by flow cytometry for the indicated WT or KLF10^−/− PACs (n = 3 per group); *P < .01 versus WT. (E-F) BM-derived progenitors were transduced with retrovirus GFP-RV-EV (empty vector; ctrl) or GFP-RV-KLF10. The percentage of GFP⁺ cells that also expressed VEGFR2 in WT CMP-, GMP-, and HSC-derived PACs (E) or TGF-β1⁺/⁺ and TGF-β1^+/− CMP-derived PACs (F) was assessed by FACS. *P < .01 versus empty EV; **P < .05 versus EV TGF-β1⁺/⁺. (G) ChIP analysis of KLF10 binding to the VEGFR2 promoter in GMP-derived PACs. IgG was used as a nonspecific control. Assays were performed in triplicate by real-time quantitative PCR with the use of primers at −294 bp and −30 bp of the VEGFR2 promoter. Values are presented as relative to DNA input. *P < .01 versus without TGF-β1 treatment.

Figure 3

KLF10^−/− CMP- and GMP-derived PACs possess markedly reduced migratory function and release of soluble paracrine factors. (A) The indicated WT or KLF10^−/− BM-derived PACs were assayed for adhesion on fibronectin-coated plates. The number of adherent cells was quantitated after 15 minutes (n = 6 per group). *P < .01 versus WT; **P < .05 versus WT. (B) The indicated WT or KLF10^−/− BM-derived PACs were assayed using a modified Transwell Boyden chamber in response to serum. The number of cells in the lower chamber was quantitated after 4 hours (n = 6 per group). *P < .01 versus WT. (C) Cell surface expression of the chemokine receptors CXCR4, CXCR3, and CCR7 was determined on the indicated WT or KLF10^−/− BM-derived PACs by flow cytometry and expressed as percentage of positivity. (D) The indicated WT or KLF10^−/− BM-derived PACs were assayed with a modified Transwell Boyden chamber in response to 1% BSA control or the chemokines SDF-1α, CXCL10, or CCL21. The number of cells in the lower chamber was quantitated after 4 hours (n = 6 per group). *P < .01 versus WT; **P < .05 versus WT. (E) Culture supernatants were harvested from the indicated WT or KLF10^−/− CMP- and GMP-derived PACs assessed by ELISA for the indicated cytokines, growth factors, or chemokines (n = 3 per group). *P < .01 versus WT.

Figure 4

Effect of KLF10 deficiency on hindlimb ischemia. (A-E) WT or KLF10^−/− mice underwent femoral artery ligation to induce hindlimb ischemia. (A) KLF10^−/− mice (n = 2 of 21) developed autoamputation of the ischemic leg. Mice were photographed with a Nikon Coolpix 4600 digital camera. (B) Images (left) are representative of blood flow recovery for each time point over 28 days. Quantitation (right) of blood flow recovery was calculated as the mean blood flow (right [ischemic] leg)/left [(nonischemic] leg) by laser Doppler imaging (785-nm near-infrared Laser Doppler Imager-2; Moor Instruments).*P < .01 versus WT; **P < .05 versus WT. (C) FACS analyses of circulating PACs (Sca-1⁺/VEGFR2⁺) in WT or KLF10^−/− mice 7 days after femoral artery ligation (n = 9-10 per group). *P < .05 versus WT. (D) WT or KLF10^−/− PACs (1:1 mix of CMP- and GMP-derived PACs) were intramuscularly injected immediately after femoral artery ligation in KLF10^−/− mice. Mean blood flow recovery (ischemic leg/nonischemic leg) was measured by laser Doppler imaging after 3 days. *P < .05 versus WT +PBS. (E-F) Frozen sections of quadriceps muscles harvested 3 days after intramuscular injection of WT or KLF10^−/− PACs were labeled with cell tracker (red) and a FITC-conjugated mAb to CD31 (green). (E) Sections were analyzed for CD31 staining with the use of an AQUA/PM2000 Imaging Platform (HistoRx), and automated quantitative analysis was performed with Software suite Version 2.2 (HistoRx). *P < .01 versus WT; n = 4 mice per group. (F) Sections were examined with an Olympus, Fluoview, Model FV1000 camera at 10× magnification and FV10-ASW Version 02.01 software to determine the percentage of labeled PACs (red) that colocalized (yellow) with CD31⁺ cells (green). *P < .01 versus WT; n = 4 mice per group.

Figure 5

Effect of KLF10^−/−PACs on blood flow recovery in WT mice. (A-D) WT mice underwent femoral artery ligation to induce hindlimb ischemia. WT or KLF10^−/− CMP-, GMP-, MEP-, or HSC-derived PACs were intramuscularly injected immediately after femoral artery ligation (n = 5 per group). Mean blood flow recovery (ischemic leg/nonischemic leg) was measured by tissue Doppler imaging. *P < .05 versus WT PACs.

Figure 6

Reduced angiogenesis in Matrigel plugs implanted in KLF10^−/− mice. (A-B) Matrigel plugs subcutaneously implanted for 8 days in WT or KLF10^−/− mice (n = 10 per group) were stained for CD31. (A) Angiogenesis in whole Matrigel plugs. Matrigel images were photographed with an Olympus, Model SZ61 camera (top). CD31 staining in paraffin sections (5 μm; bottom) was analyzed with an Olympus, Fluoview, Model FV1000 camera at 10× magnification and FV10-ASW Version 02.01 software and quantitated as relative CD31 positivity (B). *P < .01 versus WT.

Figure 7

KLF10 expression in response to hindlimb ischemia in mice and in human patients with PAD. (A) Eight- to 10-week-old, male WT mice underwent femoral artery ligation to induce hindlimb ischemia (ischemia) or sham control operation (sham), and expression of Klf10 was determined by intracellular staining of circulating PACs (Sca1⁺/VEGFR2⁺) after 7 days by flow cytometry. *P < .01 versus sham. (B) Expression of KLF10 was determined in circulating PACs (CD34⁺/VEGFR2⁺) obtained from healthy control subjects or patients with symptomatic PAD. *P < .01 versus healthy controls.