Neuropilin-1 regulates platelet-derived growth factor receptor signalling in mesenchymal stem cells

Abstract

Using human MSCs (mesenchymal stem cells) lacking VEGF (vascular endothelial growth factor) receptors, we show that the pro-angiogenic receptor neuropilin-1 associates with phosphorylated PDGFRs [PDGF (platelet-derived growth factor) receptors], thereby regulating cell signalling, migration, proliferation and network assembly. Neuropilin-1 co-immunoprecipitated and co-localized with phosphorylated PDGFRs in the presence of growth factors. Neuropilin-1 knockdown blocked PDGF-AA-induced PDGFRalpha phosphorylation and migration, reduced PDGF-BB-induced PDGFRbeta activation and migration, blocked VEGF-A activation of both PDGFRs, and attenuated proliferation. Neuropilin-1 prominently co-localized with both PDGFRs within MSC networks assembled in Matrigel and in the chorioallantoic membrane vasculature microenvironment, and its knockdown grossly disrupted network assembly and decreased PDGFR signalling. Thus neuropilin-1 regulates MSCs by forming ligand-specific receptor complexes that direct PDGFR signalling, especially the PDGFRalpha homodimer. This receptor cross-talk may control the mobilization of MSCs in neovascularization and tissue remodelling.

Figure 1. NRP-1 associated with phosphorylated PDGFRs

The association of NRP-1 with PDGFRs was evaluated. (A) Flow cytometry analysis of cell surface (i) IgG₁ used as a control, (ii) NRP-1 and (iii) VEGFR2. A representative example of three independent experiments is shown. (B) The association of NRP-1 with PDGFRs was examined by immunoprecipitation (IP) followed by immunoblot (IB) analysis. MSCs grown on gelatin and cultured for 24 h in serum-free conditions were unstimulated (Con) or stimulated with either 20 ng/ml PDGF-AA, PDGF-BB or VEGF-A₁₆₅ for 10 min at 37 °C, then NRP-1 association with PDGFRs was determined by IP analysis of cell lysates. IP analysis using (i and ii) anti-NRP-1, (iii) anti-PDGFRα or (iv) anti-PDGFRβ, with anti-IgG₁ as a control, then NRP-1 association with PDGFRs detected by IB analysis using (i) anti-PDGFRα, (ii) anti-PDGFRβ or (iii and iv) anti-NRP-1, followed by IB analysis using anti-PDGFRs or NRP-1 as loading controls. A representative of three independent experiments is shown. (C) The percentage of (i and ii) total PDGFRs and (iii and iv) phosphorylated PDGFRs interacting with NRP-1 in a particular cell lysate was estimated by IP and IB analysis. Cell lysates were isolated from MSCs which were either unstimulated (Con) (lysates 1 and 3), or exposed to 20 ng/ml PDGF-AA (lysate 2) or PDGF-BB (lysate 4) for 10 min at 37 °C. Each cell lysate was then split into four separate 100 μg aliquots (i–iv) for IP analysis, using either anti-NRP-1, anti-PDGFRα (Rα) or anti-PDGFRβ (Rβ), then IB analysis using (i and ii) anti-PDGFRα or anti-PDGFRβ and (iii and iv) using anti-PDGFRα-Tyr⁷⁵⁴ or anti-PDGFRβ-Tyr¹⁰²¹. As a loading control blots were re-probed using the corresponding IP antibody. The percentage of total PDGFRα or PDGFRβ interacting with NRP-1 was estimated by quantifying the IB analysis in (i) relative to the corresponding IB analysis in (ii), which was assumed to be 100%. Similarly, the percentage of p-PDGFRα-Tyr⁷⁵⁴ or PDGFRβ-Tyr¹⁰²¹ interacting with NRP-1 was estimated by quantifying the IB analysis in (iii) relative to the corresponding IB analysis in (iv), which was taken to be 100%. This approach gives the proportion of receptor association within the immunoprecipitates, but does not report the total amounts of receptors since the efficiency of each antibody in immunoprecipitating their receptor from a cell lysate may not be 100%. (D) Histograms representing the co-immunoprecipitation data, showing the percentage of (i) total PDGFRα or PDGFRβ and (ii) p-PDGFRα-Tyr⁷⁵⁴ or PDGFRβ-Tyr¹⁰²¹, which interacted with NRP-1. Values are mean percentage values±S.D. determined from two independent experiments. **P<0.001, compared with unstimulated control.

Figure 2. NRP-1 co-localized with phosphorylated PDGFRs

The cellular distribution of NRP-1 and phosphorylated PDGFRs was examined following ligand exposure. MSCs grown on 0.1% gelatin were cultured for 24 h in serum-free conditions, exposed to PDGF ligands, then co-localization of NRP-1 with either phosphorylated PDGFRα at site Tyr⁷⁵⁴ or PDGFRβ at site Tyr¹⁰²¹ was examined by immunofluorescence microscopy. (A) Control unstimulated and (B) exposed to 20 ng/ml PDGF-AA for 10 min, showing PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red). (C) Control unstimulated and (D) exposed to 20 ng/ml PDGF-BB for 10 min, showing PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). For each image, the corresponding red and green channels having similar threshold values and the same particle size range are shown, together with their co-localization represented by the image in yellow. The mean number of co-localized particles±S.D. derived from six different single cell images is denoted in yellow. Nuclei are counter-stained with DAPI (blue). Representative images of at least four independent experiments are shown. Scale bars=20 μm. (E) Histogram showing the ligand-induced increase in co-localization between NRP-1 and PDGFRα-Tyr⁷⁵⁴ or PDGFRβ-Tyr¹⁰²¹, as determined by immunofluorescence analysis. Values are the mean number of co-localized particles±S.D. derived from six different single cell images. **P<0.001, compared with the corresponding unstimulated control.

Figure 3. Comparison of VEGF-A-stimulated MSCs and HUVECs

The cellular distribution of NRP-1 was examined in MSCs and compared with HUVECs following VEGF-A₁₆₅ stimulation. MSCs grown on 0.1% gelatin were cultured for 24 h in serum-free conditions, then co-localization of NRP-1 with either PDGFRα phosphorylated at site Tyr⁷⁵⁴, or PDGFRβ phosphorylated at site Tyr¹⁰²¹, was examined by immunofluorescence microscopy. As a comparison, HUVECs grown on 0.1% gelatin were cultured for 4 h in serum-free conditions, then co-localization of NRP-1 with VEGFR2 phosphorylated at site Tyr¹¹⁷⁵ was similarly determined. (A) Control unstimulated HUVEC and (B) HUVEC exposed to VEGF-A₁₆₅ for 10 min, showing VEGFR2-Tyr¹¹⁷⁵ (red) and NRP-1 (green). (C) Control unstimulated MSC and (D) MSC exposed to VEGF-A₁₆₅ for 10 min, showing PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red). (E) Control unstimulated MSC and (F) MSC exposed to 20 ng/ml VEGF-A₁₆₅ for 10 min, showing PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). Below each image, the corresponding red and green channels which have similar threshold values and the same particle size range are shown, together with their co-localization represented by the image in yellow. The mean number of co-localized particles±S.D. derived from four different single cell images is denoted in yellow. **P<0.001, compared with the corresponding unstimulated control. Nuclei are counter-stained with DAPI (blue). Representative images of at least three independent experiments are shown. Scale bars=20 μm.

Figure 4. NRP-1 enhanced PDGFR phosphorylation

The effects of NRP-1 knockdown on PDGFR phosphorylation were determined. (A) Lysates from MSCs transfected with two different NRP-1 siRNAs, (i) (from Qiagen) and (ii) (from Ambion), or scrambled (Scr) control, were analysed for NRP-1 protein expression by immunoblot (IB) analysis using an anti-NRP-1 antibody. Membranes were reprobed with anti-β-actin as a loading control. (B) The targeting specificity of each siRNA knockdown was evaluated by RT–PCR analysis. Following siRNA knockdown using Scr control or two different NRP-1 siRNAs, (i) (from Qiagen) and (ii) (from Ambion), transcript expression for PDGFRα, PDGFRβ, NRP-1 and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) as a control were determined. A representative of two independent experiments are shown. (C and D) Following siRNA knockdown with either Scr control or target NRP-1 siRNAs, MSCs in serum-free medium were exposed to either 50 ng/ml PDGF-AA, PDGF-BB, PDGF-AB, PDGF-CC or VEGF-A₁₆₅, or no growth factors (basal), for 10 min at 37 °C, then specific PDGFR tyrosine phosphorylation determined by cell-based ELISAs for (C) PDGFRα-Tyr⁷⁴² or (D) PDGFRβ-Tyr⁷⁵¹. Tyrosine phosphorylation is represented by RFUs (relative fluorescent units). Values are mean normalized RFUs±S.D. determined from two independent experiments performed in triplicate, using two different NRP-1 siRNAs. **P<0.001, *P<0.005, compared with the Scr siRNA control.

Figure 5. NRP-1 regulated PDGFR-mediated MSC migration and proliferation

The effects of NRP-1 knockdown on PDGFR-mediated migration and proliferation were investigated. (A) Following siRNA knockdown with either scrambled (Scr) control or target NRP-1 siRNAs, the effects on PDGFR-mediated migration were determined by exposing MSCs to 20 ng/ml PDGF-AA, PDGF-BB or VEGF-A₁₆₅ in the lower half of a Boyden chamber for 5 h. The broken line represents the level of unstimulated MSC migration. Values are the mean number of migratory cells±S.D. determined from ten random fields from each of three independent experiments. **P<0.001, *P< 0.005, compared with the Scr siRNA control. (B) Representative images of migratory cells/field (using a 10× objective lens) on the membrane underside of a Boyden chamber after 5 h. Basal denotes unstimulated MSC migration. Ligand-stimulated MSC migration induced by 20 ng/ml PDGF-AA (AA), PDGF-BB (BB) or VEGF-A₁₆₅ (VEGF) in the absence (−) or presence (+) of 100 nM VEGFR2 tyrosine kinase inhibitor (RTK). (C and D) Following siRNA knockdown with either Scr control or target NRP-1 siRNAs, the effects on PDGFR-mediated proliferation were determined by culturing MSCs for 5 days in either growth medium alone, or growth medium supplemented with (C) 10 ng/ml PDGF-AA or (D) 10 ng/ml PDGF-BB. Values are the mean cell number±S.D. determined from triplicate assays from each of two independent experiments, using two different NRP-1 siRNAs. **P<0.001, *P<0.005, compared with the respective Scr siRNA control MSC proliferation.

Figure 6. NRP-1 co-localization with PDGFRs during MSC network assembly

The co-localization of NRP-1 and PDGFRs was examined during the assembly of MSC networks. MSCs were seeded on to Matrigel™, then co-localization of NRP-1 with either PDGFRα phosphorylated at site Tyr⁷⁵⁴, or PDGFRβ phosphorylated at site Tyr¹⁰²¹, examined by immunofluorescence microscopy. MSCs cultured for (A) 2 h, (C) 6 h or (E) 24 h showing PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red). MSCs cultured for (B) 2 h, (D) 6 h or (F) 24 h showing PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). Below each image, the corresponding red and green channels which have similar threshold values and the same particle size range are shown, together with their co-localization represented by the image in yellow. Nuclei are counter-stained with DAPI (blue). Representative images of at least four independent experiments are shown. Scale bars=20 μm.

Figure 7. NRP-1 regulated MSC network assembly

The role of NRP-1 in regulating MSC network assembly was evaluated following NRP-1 knockdown. Following siRNA knockdown with either scrambled (Scr ↓) control or target NRP-1 (NRP ↓) siRNAs, MSCs were seeded on to Matrigel™ and cultured for 24 h, then co-localization of NRP-1 with either phosphorylated PDGFRα at site Tyr⁷⁵⁴ or PDGFRβ phosphorylated at site Tyr¹⁰²¹ was examined by immunofluorescence microscopy. Control knockdown MSCs, showing (A) PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red), and (C) PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). NRP-1-knockdown MSCs, showing (B) PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red). (D) PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). Nuclei are counter-stained with DAPI (blue). For each image, the corresponding red and green channels having similar threshold values and the same particle size range are shown, together with their co-localization represented by the image in yellow. Wider field images of control knockdown (E) and NRP-1-knockdown (F) MSCs showing PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red). Representative images of at least four independent experiments are shown. (G) Histogram representing the number of branch points per field at 24 h following Scr or target NRP-1 knockdown. Values are the mean number of branch points±S.D. determined from at least six random fields from each of four independent experiments, **P<0.001, compared with the corresponding Scr siRNA control. Scale bars=20 μm.

Figure 8. NRP-1 is essential for in vivo MSC networks

The effects of NRP-1 knockdown on MSC network formation was examined using an in vivo angiogenesis model system, the CAM of the developing chick embryo. Following siRNA knockdown with either Scrambled (Scr ↓) control or target NRP-1 (NRP ↓) siRNAs, MSCs were seeded on to Matrigel™ and implanted in direct contact with a highly vascularized area of CAM for 24 h, then co-localization of NRP-1 with either phosphorylated PDGFRα at site Tyr⁷⁵⁴ or PDGFRβ phosphorylated at site Tyr¹⁰²¹ was examined by immunofluorescence microscopy. Control knockdown MSCs, showing (A) PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red), (C) PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). NRP-1-knockdown MSCs, showing (B) PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red), (D) PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). Control knockdown MSCs, showing (E) PDGFRα-Tyr⁷⁵⁴ (green) and NRP-1 (red), (F) PDGFRβ-Tyr¹⁰²¹ (red) and NRP-1 (green). For each image, the corresponding red and green channels having similar threshold values and the same particle size range are shown, together with their co-localization represented by the image in yellow. Nuclei are counter-stained with DAPI (blue). Representative images of two independent experiments are shown. Scale bars=20 μm.