Relaxin increases human endothelial progenitor cell NO and migration and vasculogenesis in mice

Abstract

The ovarian peptide hormone, relaxin, circulates during pregnancy, contributing to profound maternal vasodilation through endothelial and nitric oxide (NO)-dependent mechanisms. Circulating numbers of bone marrow-derived endothelial cells (BMDECs), which facilitate angiogenesis and contribute to repair of vascular endothelium, increase during pregnancy. Thus, we hypothesized that relaxin enhances BMDEC NO production, circulating numbers, and function. Recombinant human relaxin-2 (rhRLX) stimulated PI3K/Akt B-dependent NO production in human BMDECs within minutes, and activated BMDEC migration that was inhibited by L-N(G)-nitroarginine methyl ester. In BMDECs isolated from relaxin/insulin-like family peptide receptor 2 gene (Rxfp2) knockout and wild-type mice, but not Rxfp1 knockout mice, rhRLX rapidly increased NO production. Similarly, rhRLX increased circulating BMDEC number in Rxfp2 knockout and wild-type mice, but not Rxfp1 knockout mice as assessed by colony formation and flow cytometry. Taken together, these results indicate that relaxin effects BMDEC function through the RXFP1 receptor. Finally, both vascularization and incorporation of GFP-labeled BMDECs were stimulated in rhRLX-impregnated Matrigel pellets implanted in mice. To conclude, relaxin is a novel regulator of BMDECs number and function, which has implications for angiogenesis and vascular remodeling in pregnancy, as well as therapeutic potential in vascular disease.

Figure 1

Relaxin is a chemoattractant for CD34⁺ BMDEC migration. (A) CD34⁺ BMDECs migrate to increasing concentrations of rhRLX. *P < .001 versus 0 ng/mL rhRLX. Bar represents the mean (± SD) infrared (IR) fluorescence of migrating cells from 4 wells per rhRLX concentration. Shown is a representative of 4 experiments. (B) CD34⁺ BMDECs from 10 healthy persons that migrate to rhRLX (50 ng/mL) and SDF-1 (100nM) expressed as mean percentage (± SD) of cells in the lower chamber, above background migration. *P < .029 versus SDF-1. (C) CD34⁺ BMDECs were seeded in the upper Boyden chamber with either vehicle or 60 ng/mL of rhRLX in the top and/or bottom as indicated. Shown is mean percentage (± SD) of 3 wells of CD34⁺ BMDECs loaded to the top chamber that have migrated, after subtracting background. *P < .025 versus black bar.

Figure 2

CD34⁺ BMDECs produce NO in response to rhRLX. Human CD34⁺ BMDECs were isolated in the absence (A-B) and presence (C-D) of 100μM L-NAME and then incubated with DAF-FM before imaging by confocal microscopy. Panels A and C are brightfield images of panels B and D, respectively, the same fields imaged with 495 nm excitation and 515 nm emission. Arrows indicate some of the CD34⁺ BMDECs. Original magnification ×200. (E) CD34⁺ BMDECs were isolated from a healthy volunteer and labeled with DAF-FM for 30 minutes before removing probe and waiting 10 minutes for de-esterification. The cells were monitored for 30 minutes, to confirm a stable baseline of bioavailable NO before vehicle (○; arrow) or 50 ng/mL of rhRLX (●; arrow) was added to the CD34⁺ cells. Shown is the mean fluorescence (± SD) of at least 8 cells in which fluorescence was continuously monitored. *P < .01 versus last baseline value. (F) BMDEC-CFU were isolated from a healthy volunteer and labeled with DAF-FM. After 30 minutes to stabilize NO baseline, the indicated concentration of rhRLX was added and after 30 minutes fluorescence was determined. Data are mean ± SD. *P < .001 versus 0 ng/mL of rhRLX. (G) Cells were incubated with 10μM L-NAME before being placed in the Boyden chamber with 50 ng/mL of rhRLX. The mean percentage (± SD) of fluorescence of cells migrating relative to control is shown. *P < .001 versus all other treatments.

Figure 3

Relaxin stimulates NO production via PI3K/Akt. (A) Human CD34⁺ BMDECs were isolated from a healthy volunteer and pretreated with the PI3K inhibitor LY294002 (10μM) for 30 minutes as indicated before the addition of DAF-FM and continued incubation with LY294002 (LY). After 30 minutes of monitoring to confirm that bioavailable NO was stable, either vehicle (Control) or 100 ng/mL of rhRLX was added to the CD34⁺ BMDECs, and NO was measured again after 30 minutes. Fluorescence was determined for each condition in at least 30 cells within 2 separate wells. Data are mean ± SD. *P < .001 versus all other conditions. (B) CD34⁺ BMDECs were isolated from a healthy volunteer and pretreated with the Akt inhibitor MK2206 (20nM) for 30 minutes as indicated before the addition of DAF-FM and continued incubation with MK2206. After 30 minutes of monitoring, to confirm that bioavailable NO was stable, either vehicle (Control) or 50 ng/mL of rhRLX was added to the CD34⁺ BMDECs, and NO was measured again after 30 minutes. Fluorescence was determined for each condition in 2 wells for at least 30 cells per well. Data are mean ± SD. *P < .001 versus control. There is a significant difference between control and MK2206 (P = .025) but no difference between MK2206 and MK2206 + rhRLX-treated cells.

Figure 4

Relaxin increases circulating BMDECs in mice by colony assay and flow cytometry. (A) BMDEC-CFU stimulated by relaxin have characteristics of late outgrowth colonies. BMDEC-CFU are counted after 5 days of culture, and a brightfield view of colonies is shown (original magnification ×100). (B) True BMDEC-CFUs demonstrate Ulex europaeus 1 staining (original magnification ×200). (C) DiI-AcLDL uptake (original magnification ×200). (D) Dual staining in a merged image of Ulex and DiI-AcLDL staining (original magnification ×200). Instead of staining, the BMDEC-CFU cells can be propagated for months. (E) Brightfield view of cells after 3 months of propagation taking on a cobblestone, endothelial, monolayer appearance. (F) Brightfield view of propagated cells that were trypsinized and plated onto a coverslip for immunostaining with anti–VWF and/or anti–MECA-32. (G) Merged image of cells in panel F stained with anti-VWF (red) and the nuclear stain DAPI (blue; original magnification ×400). (H) Merged image of another coverslip stained with anti-VWF (red), anti–MECA-32 (green), and the nuclear stain DAPI (blue; original magnification ×630). (I-K) Mice were implanted with a osmotic pumps pump containing vehicle (I-J) or rhRLX (K-L), and after 5 days the peripheral blood was collected and stained with fluorochrome-conjugated monoclonal antibody to mouse endothelial cell markers Lin, Sca1, cKit, and Flk1. The gated cells were analyzed for Sca-1 and Lin characteristics (I,K), and the subpopulation of Sca1⁺ cells was analyzed for Flk1 and cKit expression (J,L). Background staining was corrected by use of isotype controls for all markers. Percentages shown are percent of gated cells, not total lymphocytes.

Figure 5

Relaxin activates BMDECs through the RXFP1 relaxin receptor. (A) The peptide B-R13/17K H2 (H2), a relaxin antagonist, was added as indicated to CD34⁺ BMDECs at a concentration of 1μM in the presence or absence of rhRLX (50 ng/mL). Shown is the DAF-FM fluorescence in arbitrary units (± SD) of 2 wells, 20 cells per well. *P < .002 versus all other treatments. Shown is a representative of 3 experiments. (B) BMDECs were isolated from the BM of wild-type littermates and Rxfp1 or Rxfp2 knockout mice and treated with vehicle (black bars) or rhRLX 50 ng/mL (white bars) for 10 minutes before the determining intracellular bioavailable NO by DAF-FM. The NO fluorescence from at least 20 cells was analyzed for each mouse. Shown is the mean DAF-FM fluorescence in arbitrary units (± SD). *P < .009 versus vehicle treated cells. †P < .05 versus untreated cells. n indicates the number of mice studied. (C) Wild-type (^+/+) littermates and Rxfp1 and Rxfp2 knockout (^−/−) mice were implanted with osmotic pumps containing rhRLX, and after 5 days their blood was collected and BMDEC-CFU were determined. Shown is the number of BMDECs (± SD). *P < .01 versus Rxfp1^+/+. n indicates the number of mice studied.

Figure 6

Relaxin recruits BMDECs into areas of neovascularization. (A) Chimeric mice (wild-type mice with GFP BM) were implanted with Matrigel pellets impregnated with vehicle or rhRLX, as indicated, and after 7 days the pellets were isolated, embedded in paraffin, sectioned, and stained with DAPI, a pan-endothelial cell antigen monoclonal antibody MECA-32, anti-GFP, or isotype control antibodies as indicated and imaged using conventional fluorescent microscopy. (B) Same as panel A, except that slides were imaged using confocal microscopy. (C) Percentage of area that fluoresces red (by epifluorescence) indicating MECA-32 staining. Shown is average area of 3 fields from 6 different sections from 9 pellets for each treatment. *P < .001 by paired t test. (D) Percentage of area that expresses dual fluorescence (by epifluorescence) indicating dual MECA-32 and GFP staining. Shown is average area of 3 fields from 6 different sections from 9 pellets. *P < .001 by paired t test.