Fig 2: EGFL7 immunostaining in HC and dcSSc skin. (A,D) In HC skin, strong EGFL7 expression was detected in ECs and pericytes of vessels (black arrows), in fibroblasts of papillary and reticular dermis (black open arrows), and cells of the basal layer of the epidermis and keratinocytes. (B,E) In EOS dcSSc skin, strong EGFL7 expression was detected in ECs and pericytes of vessels (black arrows), in fibroblasts of papillary and reticular dermis and epidermal (black open arrows), cells of the basal layer of the epidermis and keratinocytes as in the HC skin. ECs of the most vessels were strong immunostained for EGFL7. However, ECs of a few vessels displayed a weaker immunostaining intensity for EGFL7 (blu arrows). Inset shows EGFL7 positive fibroblast (C and F) In LSS dcSSc skin, the remaining microvessels displayed a weak (or absent) immunostaining for EGFL7. Furthermore, fibroblasts of papillary and reticular dermis and cells of the basal layer of the epidermis and keratinocytes displayed a weak (or absent) immunostaining for EGFL7. Inset shows EGFL7 weak positive fibroblast. Semiquantitative scoring was performed independently by 2 blinded observers, based on the observation of immunostained skin sections. Scores are defined as follows: +++ intense staining; +++/− intense staining but not homogeneous positive staining; ++ moderate staining; + weak staining; +/− weak staining but not homogeneous positive staining.

Fig 3: EGFL7 is inducible by TGF-β on HC-FBs and decreases the expression of COL1A1 (A–G). The experiments were performed in triplicate. Bars represent mean values ± SEM before and after stimulation with 10 ng/ml of TGF-β for 48 hrs. The absence of rh proteins from the medium of stimulation was considered as a negative control. HC-FBs, after stimulation by TGF-β, expressed significantly higher levels of EGFL7 RNA, when compared to untreated HC-FBs (***p < 0.0001). On the contrary, TGF-β treated EOS SSc-FBs did not show any difference in the EGFL7-RNA levels, when compared to the untreated EOS SSc-FBs. No effect was observed, on the EGFL7 RNA levels for the TGF-β treated and untreated LSS SSc-FBs (A). The levels of TGF-β in untreated HC-FBs were set to 100% and all the results were normalized to this value. After EGFL7 stimulation, we observed a dose-related fashion decrease of the expression of COL1A1 (*p = 0.018; ***p = 0.0001 for RNA levels), both in RNA and protein levels, in SSc-FBs. On the contrary, when HC-FBs were treated with the same rhEGFL7 concentrations, the expression of COL1A1 protein did not change at any concentration of EGFL7 stimulation (B,C). Full length blot is represented in Supplementary Fig. 2. Densitometric analysis of the bands was performed using ImageJ software (NIH; online at http://rsbweb.nih.gov/ij) (D).

Fig 4: EGFL7 knockdown in HC- and SSc-FBs up-regulates Col1A1 expression after TGF-β stimulation. SSc-FBs were transfected with specific EGFL7-siRNA (siRNA) or non-targeting siRNA (scr), and EGFL7 mRNA expression was evaluated by qRT-PCR analysis. The cells transfected with EGFL7-siRNA showed a decreased expression of EGFL7 mRNA levels when compared with cells transfected with scr siRNA (A). After TGF-β stimulation, the silenced cells show an overexpression of COL1A1, suggesting the anti-fibrotic role of EGFL7, confirming the effects observed in Fig. 3B. Each experimental condition was performed in triplicate. Bars represent mean values ± SEM.

Fig 5: SSc-FBs are responsible for impaired angiogenesis in an organotypic co-culture assay system in vitro but EGFL7 restores angiogenesis. Angiogenesis was not affected when primary HC-FBs were co-cultured with HUVECS (A). Interestingly, when primary SSc-FBs were co-cultured with HUVECs, angiogenesis was impaired when compared with the tubule formation obtained from co-cultured HC-FBs with HUVECs (B–C). EGFL7 increases the tubule formation when added to the cultured medium (100 ng/ml) thus adding EGFL7 to the list of the pro-angiogenic molecules that rescue angiogenesis defects in SSc patients (E–G). VEGF (25 ng/ml) was used as positive control (F–K). Representative microscopic fields at 4X magnification from triplicate wells. Vascular junction and tubule number and total tubule length were analysed by Angiosys System (TCS CellWorks, USA) (D–H–L). Bars represent mean values ± SEM (measurment of 12 microscopic fields from triplicate wells). Statistical analysis was performed by unpaired and paired t test, two tailed. *p < 0.05; **p < 0.001; ***p < 0.0001.