Objectives In systemic sclerosis (SSc), vascular involvement is characterised by vascular endothelial growth factor (VEGF)-A/VEGF receptor (VEGFR) system disturbances. Neuropilin-1 (NRP1), a receptor for both class-3 semaphorins (Sema3s) and VEGF-A, is required for optimal VEGF-A/VEGFR-2 signalling. Here, we investigated the possible involvement of Sema3A/NRP1 axis in SSc.
Methods Circulating Sema3A and soluble NRP1 (sNRP1) were measured in patients with SSc and controls. NRP1 and Sema3A expression in skin biopsies was evaluated by immunofluorescence and western blotting. NRP1 expression was assessed in SSc and healthy dermal microvascular endothelial cells (SSc-MVECs and H-MVECs), and in SSc and control endothelial progenitor cell (EPC)-derived endothelial cells (ECs). The possible impact of transcription factor Friend leukaemia integration 1 (Fli1) deficiency on endothelial NRP1 expression was investigated by gene silencing. The binding of Fli1 to NRP1 gene promoter was evaluated using chromatin immunoprecipitation. Capillary morphogenesis was performed on Matrigel.
Results Decreased sNRP1 levels in SSc were associated with active and late nailfold videocapillaroscopy patterns and digital ulcers. No difference in Sema3A was found between patients and controls. NRP1 was significantly decreased in SSc-MVECs both ex vivo and in vitro. NRP1 and Fli1 significantly decreased in H-MVECs challenged with SSc sera, while they were not different in SSc and control EPC-derived ECs. Fli1 occupied the NRP1 gene promoter and Fli1 gene silencing reduced NRP1 expression in H-MVECs. NRP1 gene silencing in H-MVECs resulted in a significantly impaired angiogenic capacity comparable to that of cells treated with SSc sera.
Conclusion In SSc, NRP1 deficiency may be an additional factor in the perturbed VEGF-A/VEGFR-2 system contributing to peripheral microvasculopathy and defective angiogenesis.
- Systemic Sclerosis
- Autoimmune Diseases
- Qualitative research
Statistics from Altmetric.com
Systemic sclerosis (SSc, scleroderma) is a life-threatening connective tissue disorder of unknown aetiology, characterised by widespread vascular injury and dysfunction, impaired angiogenesis, immune dysregulation and progressive fibrosis of the skin and internal organs.1 ,2 The dysregulation of vascular tone control, clinically evident as Raynaud's phenomenon, and microcirculatory abnormalities paralleled by nailfold capillaroscopic changes are the earliest clinical manifestations of SSc and may precede skin and visceral involvement by months or years.1–3 The whole process is characterised by an uncontrolled regeneration of the microvasculature and subsequent loss of microvessels, due to defects in both vascular repair and expected increase in new vessel growth (angiogenesis), leading to severe peripheral ischaemic manifestations, such as digital ulcers and gangrene.2 ,3
Recent studies have highlighted the anatomical and structural similarities between blood vessels and nerves.4 The two networks are often aligned, with nerve fibres and blood vessels following parallel routes. Furthermore, both systems require precise control over their guidance and growth. Several molecules with attractive and repulsive properties have been found to modulate the guidance both of nerves and blood vessels.4 These include the neuropilin (NRP) receptors and their semaphorin (Sema) ligands, as well as netrins, slits and their receptors.4 Among these, NRP1 was initially described as an axonally expressed receptor for secreted class-3 Semas (Sema3s), a family of soluble molecules which modulate the development of the nervous and vascular systems.5 ,6 NRP1 also serves as specific vascular endothelial growth factor-A (VEGF-A) co-receptor on endothelial cells (ECs) and regulates VEGF receptor (VEGFR) signalling, leading to enhanced migration7 and survival of ECs in vitro.8 ,9 Furthermore, NRP1 has been implicated in VEGFR-2-mediated endothelial permeability10 and in VEGF-A-induced three-dimensional EC biology, such as vessel sprouting and branching.11 The absence of functional NRP1 in mice results in embryonic death due to impaired heart and blood vessel development, thus suggesting that this receptor plays a central regulatory role during developmental angiogenesis.4 The exact molecular mechanisms by which NRP1 modulates VEGF-A biology remain to be elucidated. It has, however, been shown that NRP1 potentiates the VEGF-A/VEGFR-2 signalling pathways implicated in the migratory response of ECs.12
The evidence that NRP1 functions as a receptor for both VEGF-A and Sema3s suggests that the latter may also play a role in the modulation of angiogenesis. In particular, it has been reported that Sema3A acts as an antiangiogenic molecule impairing EC adhesion, migration and survival in vitro13–15 and regulates tumour-induced angiogenesis in vivo.16 Moreover, Sema3A null mice exhibit defects in blood vessel reshaping.15 ,17 ,18 The molecular mechanisms underlying the antiangiogenic effects of Sema3A are complex. Hence, it was initially suggested that Sema3A competes with VEGF-A for NRP1 binding, thus inhibiting VEGF-A-induced angiogenesis. However, recent reports have also shown that Sema3A increases vascular permeability, inhibits EC proliferation and induces apoptosis even in the absence of VEGF-A, suggesting that Sema3A may activate its own signalling pathways.19 ,20
On these bases, we hypothesised that the Sema3A/NRP1 axis might play a role in the pathogenesis of SSc-related microvascular abnormalities. Therefore, the aim of the present study was to investigate whether the levels of Sema3A and NRP1 could be altered in the circulation, skin and ECs of patients with SSc, as well as the mechanism explaining the dysregulated expression of these molecules and their possible contribution to the disturbed angiogenesis of SSc.
An extended methods section is provided in the online supplementary material.
Patients, controls, serum samples and skin biopsies
Serum samples were obtained from 49 patients with SSc1 (45 women and 4 men; median age 64 years, range 37–80 years, and median disease duration 10 years, range 2–31 years) classified as limited cutaneous SSc (lcSSc; n=32) or diffuse cutaneous SSc (dcSSc; n=17),21 and from 39 age-matched and sex-matched healthy individuals. All patients were clinically assessed as described elsewhere.3 ,22 ,23 Clinicodemographical characteristics of patients with SSc used for collection of serum samples are shown in online supplementary table S1. Full-thickness skin biopsies were obtained from the clinically involved skin of one-third of the distal forearm of 18 patients with SSc (15 women, 3 men; median age 48.5 years, range 29–73 years, and median disease duration 7.2 years, range 1–18 years). Skin samples from the same forearm region of 11 age-matched and sex-matched healthy donors were used as controls. Each skin biopsy was divided into two specimens and processed for immunohistochemistry and biomolecular analysis as described elsewhere.23 The study was approved by the local institutional review board at the Azienda Ospedaliero-Universitaria Careggi (AOUC), Florence, Italy, and all subjects provided written informed consent.
Isolation, culture and stimulation of dermal microvascular ECs
Dermal microvascular ECs (MVECs) were isolated from biopsies of the involved forearm skin from five patients with dcSSc and from five healthy subjects, as described elsewhere.3 ,23 MVECs from healthy subjects (H-MVECs) and patients with SSc (SSc-MVECs) were used between the third and seventh passages in culture. For stimulation experiments, H-MVECs were grown to 70% confluence, and then were washed three times with serum-free medium and serum-starved overnight in MCDB 131 medium (Sigma-Aldrich, St. Louis, Missouri, USA) supplemented with 2% fetal bovine serum (FBS). Medium was removed and cells were incubated with 2% FBS-MCDB medium containing recombinant human VEGF-A165 (10 ng/mL; R&D Systems, Minneapolis, Minnesota, USA), or 10% serum from patients with SSc (n=5) and healthy subjects (n=5) for 24 h.
Late-outgrowth peripheral blood endothelial progenitor cell-derived ECs
Late-outgrowth endothelial progenitor cell (EPC)-derived ECs were obtained from the peripheral blood of 15 patients with SSc (13 women and 2 men; n=9 with lcSSc and n=6 with dcSSc; median age 60 years, range 42–78 years) and eight healthy individuals (all women; median age 55 years, range 30–65 years), as described elsewhere.3 ,24 ,25
ELISA for serum Sema3A and soluble NRP1
The levels of Sema3A and soluble NRP1 (sNRP1) in serum samples were measured by commercial quantitative colorimetric sandwich ELISA (catalogue number ABIN481720 and ABIN415191, respectively; Antibodies-Online, Atlanta, Georgia, USA), according to the manufacturer's protocol. Each sample was measured in duplicate.
Proteins were extracted from skin biopsies, dermal MVECs and late-outgrowth peripheral blood EPC-derived ECs as described elsewhere.3 ,25 ,26 Western blotting was carried out according to previously published protocols.3 For primary antibodies, refer to the online supplementary material.
Gene silencing of Friend leukaemia integration 1 and NRP1
MVECs were seeded shortly before transfection. The cells were transfected with 10 nM of Friend leukaemia integration 1 (Fli1) small interfering RNA (siRNA), 10 nM of NRP1 siRNA or non-silencing scrambled RNA (SCR) (Santa Cruz Biotechnology, Dallas, Texas, USA) using HiPerfect transfection reagent (Qiagen, Milan, Italy) for 72 h.
RNA purification, cDNA synthesis and quantitative real-time PCR
Total RNA isolation from MVECs, first strand cDNA synthesis and mRNA quantification by SYBR Green real-time PCR were performed as reported elsewhere.27 For predesigned oligonucleotide primer pairs obtained from Qiagen, refer to the online supplementary material.
Chromatin immunoprecipitation assay
The chromatin immunoprecipitation (ChIP) assay was carried out as previously described28 using a rabbit polyclonal anti-Fli1 antibody (catalogue number ab15289, Abcam, Cambridge, UK). Putative Fli1 transcription factor binding site was predicted by Tfsitescan. The primers were as follows: NRP1 Forward, 5′-CTAGGGGTGCAGAGCGAG-3′; NRP1 Reverse, 5′-GAAGGAAGGCGCTGGGAG-3′.
In vitro capillary morphogenesis assay
Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) software for Windows, V.20.0 (SPSS, Chicago, Illinois, USA). Data are expressed as mean±SD or median and IQR. The Student's t test and non-parametric Mann–Whitney U test were used where appropriate for statistical evaluation of the differences between two independent groups. A p value of <0.05 was considered statistically significant.
Serum Sema3A and sNRP1 levels in SSc
No significant differences in serum levels of Sema3A were detected between patients with SSc (median 2.22 ng/mL, IQR 1.84–3.17 ng/mL) and healthy controls (median 3.86 ng/mL, IQR 1.64–4.73 ng/mL) (figure 1A). Circulating sNRP1 levels were significantly reduced in patients with SSc (median 0.22 ng/mL, IQR 0.0–0.6 ng/mL) compared with healthy individuals (median 0.69 ng/mL, IQR 0.0–2.5 ng/mL; p=0.001) (figure 1B). Next, we evaluated the possible correlation of serum sNRP1 levels with the nailfold videocapillaroscopy (NVC) pattern as a measure of peripheral microvascular involvement. sNRP1 levels were significantly decreased in patients with SSc having active (median 0.14 ng/mL, IQR 0.0–0.4 ng/mL) or late (median 0.09 ng/mL, IQR 0.0–0.72 ng/mL) NVC patterns than in controls (p=0.003 and p=0.01, respectively) (figure 1C). Conversely, no difference in serum sNRP1 was found between patients with SSc having early NVC pattern (median 0.45 ng/mL, IQR 0.22–0.66 ng/mL) and healthy controls (figure 1C). Moreover, sNRP1 levels were significantly decreased in patients with SSc having digital ulcers (median 0.06 ng/mL, IQR 0.0–0.27 ng/mL) compared both with patients without digital ulcers (median 0.43 ng/mL, IQR 0.17–0.71 ng/mL; p=0.009) and controls (p=0.001) (figure 1D). No significant association was found with other clinicodemographical and laboratory parameters or with clinical SSc subset.
Decreased expression of NRP1 in SSc dermal ECs ex vivo and in vitro
The expression of Sema3A and NRP1 protein in forearm skin biopsies from patients with SSc and controls was investigated by immunofluorescence and western blot. No significant differences in Sema3A expression were detected between SSc and control skin (figure 2A–C). On the contrary, NRP1 expression was decreased in clinically affected skin biopsies from patients with SSc compared with healthy skin, in particular in dermal ECs and perivascular stromal cells (figure 3A–F). The localisation of NRP1 staining in vascular ECs was confirmed by NRP1/CD31 double immunofluorescence staining (figure 3C, F). Moreover, western blot analysis confirmed that NRP1 protein expression levels were significantly reduced in SSc skin with respect to control skin (p<0.001) (figure 3G, H).
Western blot analysis on cultured dermal MVECs revealed that NRP1 protein expression levels were significantly reduced in SSc-MVECs compared with H-MVECs (p<0.005) (figure 4A). Moreover, NRP1 expression in H-MVECs significantly increased after treatment with healthy sera compared with basal condition, while it decreased after challenging with SSc sera (both p<0.005 vs basal H-MVECs). As expected, stimulation with recombinant human VEGF-A165 strongly upregulated NRP1 expression in H-MVECs (p<0.005 vs basal H-MVECs) (figure 4A). On the contrary, no obvious differences in NRP1 protein levels could be found between late-outgrowth EPC-derived ECs from patients with SSc and healthy controls (figure 4B).
Fli1 deficiency contributes to the downregulation of NRP1 gene in SSc-MVECs
We next examined the potential mechanism by which NRP1 expression is downregulated in dermal SSc microvessels. As the expression of the transcription factor Fli1 is markedly downregulated at least partially via an epigenetic mechanism in SSc dermal ECs, and experimental endothelial Fli1 deficiency reproduces the histopathological and functional abnormalities characteristic of SSc vasculopathy,28–30 we hypothesised that endothelial Fli1 deficiency could inhibit the expression of NRP1 in SSc-MVECs.
First, we analysed Fli1 protein expression in cultured dermal MVECs and late-outgrowth EPC-derived ECs from patients with SSc and controls (figure 4C, D). As shown in figure 4C, Fli1 protein expression in H-MVECs closely paralleled that of NRP1 in the different experimental conditions assayed and was strongly downregulated in SSc-MVECs. Similarly to what was observed for NRP1 expression, Fli1 protein levels did not differ between SSc and control late-outgrowth EPC-derived ECs (figure 4D).
Moreover, we examined the effect of Fli1 gene silencing on mRNA levels of the NRP1 gene in H-MVECs. As displayed in figure 4E, gene silencing of Fli1 significantly suppressed the mRNA expression levels of the NRP1 gene in H-MVECs (p<0.01). In addition, ChIP analysis revealed that Fli1 occupied the promoter region of the NRP1 gene in H-MVECs (figure 4F). These results indicate that Fli1 directly targets the NRP1 gene promoter and is required for homeostatic NRP1 expression in ECs.
NRP1 deficiency contributes to the impaired angiogenesis of SSc-MVECs
To verify whether endothelial NRP1 deficiency has a role in the modulation of angiogenesis, we carried out in vitro capillary morphogenesis on Matrigel matrix. Consistent with previous findings,26 capillary morphogenesis was significantly impaired in SSc-MVECs compared with H-MVECs (p<0.01) (figure 5). H-MVECs stimulated with healthy sera produced an abundant network of branching cords (figure 5). On the contrary, as previously reported,31 ,32 angiogenesis was significantly reduced on challenge with SSc sera (p<0.01 vs basal H-MVECs) (figure 5). The addition of recombinant human VEGF-A165 or anti-VEGF-A165b blocking antibodies to SSc sera significantly increased H-MVEC angiogenesis compared with cells treated with SSc sera alone (both p<0.05) (figure 5). NRP1 gene silencing in H-MVECs resulted in a significant impairment of angiogenic capacity comparable with that of cells treated with SSc sera (p<0.01 vs basal H-MVECs) (figure 5). Stimulation of NRP1-silenced H-MVECs with recombinant human proangiogenic VEGF-A165 or antiangiogenic VEGF-A165b could only slightly increase or decrease angiogenesis, respectively (figure 5).
Here, we investigated for the first time the possible involvement of the Sema3A/NRP1 axis in the pathogenesis of SSc. Our present findings clearly demonstrate that serum levels and dermal expression of NRP1 are significantly decreased in patients with SSc and that lower circulating sNRP1 levels correlate with the severity of NVC abnormalities and the presence of digital ulcers. In contrast to constitutive endothelial expression of NRP1 in healthy skin, NRP1 was found to be strongly reduced ex vivo in SSc dermal microvessels, and NRP1 downregulation was maintained in vitro in MVECs obtained from SSc dermis. On the contrary, we could not find any difference in NRP1 protein levels between peripheral blood EPC-derived ECs from patients with SSc and healthy controls. This evidence suggests that the dysregulated expression of this receptor is restricted to locally injured microvasculature in an overt disease without affecting bone marrow-derived circulating endothelial progenitors. As far as Sema3A is concerned, no difference in its expression was observed between SSc and controls either in the circulation or in the cutaneous tissue.
The importance of NRP1 for vascular development is well established and shown by the generation of knockout mice, which display an embryonic lethal phenotype characterised by severe vascular defects due to impaired angiogenic sprouting and branching very much resembling the disturbed vessel morphology seen in patients with SSc.4 ,33 ,34 Moreover, conditional NRP1 knockout in ECs is associated with important cardiac and vascular defects, thus suggesting a crucial role of NRP1 in EC functions.35 Even though NRP1 was originally identified as an adhesion molecule in the nervous system, it is more commonly studied as receptor for the neuronal guidance molecule Sema3A and as co-receptor for the VEGF-A165/VEGFR-2 complex with key roles in neuronal and vascular development.12 ,36 Of note, several studies have implicated a dysfunctional VEGF-A/VEGFR-2 system in the impaired angiogenic process characteristic of SSc.2 ,26 ,37–39 Moreover, besides VEGF-A, it is well known that a dysregulated expression of a large array of proangiogenic and antiangiogenic (angiostatic) factors present in the circulation of patients with SSc may be mostly responsible for such angiogenic deficit.2 ,37–40
In our study, we observed that the proangiogenic NRP1 receptor was constitutively downregulated in dermal SSc-MVECs and that treatment with SSc sera could significantly reduce NRP1 expression in H-MVECs, which is in line with the reported antiangiogenic properties of SSc sera.2 ,31 ,32 ,37 ,41 Strikingly, we also found that NRP1 gene silencing in H-MVECs resulted in a significantly impaired angiogenic process comparable to that of cells treated with SSc sera, further supporting the implication of NRP1 deficiency in the disturbed angiogenesis of SSc.
Consistent with previous studies, stimulation with recombinant proangiogenic VEGF-A165 strongly upregulated NRP1 expression in H-MVECs, suggesting that this growth factor not only can activate ECs directly, but can also contribute to angiogenesis by a mechanism that involves upregulation of its homologous receptor NRP1.36 The findings of NRP1 downregulation in H-MVECs on challenge with SSc sera are in agreement with the evidence that the majority of VEGF-A detected in SSc circulation is not the proangiogenic VEGF-A165, but rather the antiangiogenic VEGF-A165b isoform.26 ,42 Further, it has been reported that VEGF-A165b is unable to bind the co-receptor NRP1 because the basic C-terminal amino acids essential for NRP1 binding are absent in this splice variant.43 ,44 Interestingly, we observed that stimulation of NRP1-silenced H-MVECs with proangiogenic VEGF-A165 slightly increased angiogenesis, while stimulation with VEGF-A165b isoform resulted only in a slight decrease in their angiogenic capacity. These data are consistent with the antiangiogenic action of VEGF-A165b being mainly dependent on its inability to recruit VEGFR-2/NRP1 co-receptor complex and activate downstream signalling. In a recent study, it was demonstrated that VEGF-A165 and VEGF-A165b may control the balance between VEGFR-2 recycling, degradation and signalling. In particular, due to the lack of NRP1 co-receptor binding, VEGF-A165b may induce differential intracellular vesicular trafficking of VEGFR-2 towards the degradative pathway.45 Thus, both a switch from the proangiogenic to the antiangiogenic VEGF-A isoform and the concomitant NRP1 co-receptor downregulation may have a crucial role in the insufficient angiogenic response found in SSc (figure 6). Indeed, here we also demonstrated that the addition of recombinant human VEGF-A165 or anti-VEGF-A165b blocking antibodies could significantly dampen the antiangiogenic effects of SSc sera on H-MVECs.
The clinical correlation of serum sNRP1 levels with the severity of SSc-related peripheral microvasculopathy also deserves discussion. Indeed, circulating levels of sNRP1 progressively decreased reaching the lowest values in patients with SSc having the active and late NVC patterns, which are characterised by severe architectural changes of microvessels and progressive capillary loss with formation of avascular areas.46 In addition, patients with active/late NVC patterns and digital ulcers showed serum sNRP1 levels significantly lower than healthy controls, whereas sNRP1 levels did not differ significantly between controls and patients with SSc with early NVC pattern and lack of digital ulcers. However, since sNRP1 may be largely released by ECs, we should also consider that the reduction in circulating levels of sNRP1 might be either a cause or a consequence of the disease, which is characterised by progressive loss of the peripheral microcirculation.2 ,3 ,46 Circulating levels of sNRP1 could even serve as a biomarker reflecting the severity and progression of SSc microvasculopathy. Accordingly, further prospective studies on larger cohorts of patients with SSc are warranted.
Finally, our mechanistic findings indicate that in SSc, endothelial NRP1 expression is suppressed at least partially due to Fli1 transcription factor deficiency. In fact, here we provide the first evidence that NRP1 is a member of the angiogenesis-related gene programme regulated by Fli1 in dermal MVECs. In this context, the impact of Fli1 deficiency in the loss of EC integrity and the development of peripheral microvasculopathy during SSc has been well established.28–30 Of note, it has been demonstrated that in SSc Fli1 expression is markedly suppressed at least partially through an epigenetic mechanism.29 ,30 Thus, such an epigenetic modification might partly explain the persistence of the multiple downstream effects of Fli1 deficiency in an in vitro culture system, as supported by the downregulation of endothelial NRP1 observed ex vivo in SSc dermal microvessels and maintained in cultured dermal SSc-MVECs. Interestingly, NRP1 was also found to be a target of the antiangiogenic microRNA miR-320, and a dysregulated microRNA profile is being increasingly reported in SSc.47 ,48
In conclusion, we shed light on NRP1 deficiency as a novel key factor contributing to peripheral microvasculopathy and defective angiogenesis in SSc. Further studies are warranted to decipher whether therapeutic modulation of VEGF-A/VEGFR-2/NRP1 co-receptor signalling might pave the way for boosting angiogenesis and blocking the progression of peripheral microvasculopathy in SSc.
IC was supported by a research grant from the Foundation for the Development of Internal Medicine in Europe (FDIME).
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
Handling editor Tore K Kvien
ER, IC and MM contributed equally.
Contributors Study conception and design: ER, MM, LI-M, MM-C and SG. Acquisition of data: ER, IC, MM, CM, IR, SB-R, JB, JA and YA. Interpretation of data: ER, IC, MM, RS, LI-M, MM-C and SG. Manuscript preparation: ER, IC, MM, IR, LI-M, MM-C and SG.
Funding The study was supported in part by grants from the University of Florence (Progetti di Ricerca di Ateneo to LI-M and MM-C).
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study was approved by the local institutional review board at the Azienda Ospedaliero-Universitaria Careggi (AOUC), Florence, Italy, and all subjects provided written informed consent.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.