NADPH-derived ROS generation drives fibrosis and endothelial-to-mesenchymal transition in systemic sclerosis: Potential cross talk with circulating miRNAs

Patient selection

Serum samples were collected from SSc patients at the Unit of Complex Rheumatology, Sassari. Table 1 summarizes the clinical and serological characteristics of the 16 patients enrolled in this study. All SSc patients met the SSc classification criteria of American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) [39]. The onset of the disease and the classification of patients having limited or diffuse were performed following the ACR/EULAR criteria for SSc classification [39]. All the previous medications were maintained during the study, whereas no other medications were started within the study framework. Posted flyers were used to recruit HD who were enrolled in the study after passing a screening questionnaire to exclude any underlying vascular or autoimmune condition.

1. Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies, and in accordance the tenets of the Helsinki Declaration and has been approved by the Ethics Committee of the Azienda Sanitaria Locale #1 of Sassari with protocol # 779/L 2009.

2. Informed consent: Informed consent has been obtained from all individuals included in this study.

Cell culture

The HPMECs used in this study were obtained from human lungs of HD and were supplied by Innoprot (Innoprot, Bizkaia, Spain). HPMECs were cultured in a specific basal medium supplemented with EC growth supplements as per the supplier’s directions. At confluency, HPMECs were subcultured at a split ratio of 1:2 and used within three/four passages. Unless specified in the text, in all experiments, cells have been grown in 96-well black plates (BD Falcon, Franklin Lakes, New Jersey, United States) and treated with basal medium containing 10% (v/v) of the subjects’ sera. Sera among the different subjects were normalized based on protein content [40,41]. In selected experiments, HPMECs were pretreated with broad NADPH inhibitor diphenyleneiodonium (DPI) chloride to investigate the involvement of NADPH oxidase in the observed phenomena [42].

Measurement of intracellular ROS

The ROS fluorescent molecular probe 2′,7′-dichlorodihydrofluorescein diacetate (H₂-DCFDA) was used to assess the intracellular ROS levels [43,44]. Inside the cell, after removal of the acetate groups by intracellular esterases, the nonfluorescent molecule is converted to a green-fluorescent form on oxidation in the presence of ROS. HPMECs were incubated for 30 min with phosphate-buffered saline (PBS) containing 1 µM H₂-DCFDA, washed with PBS, and treated as described. Fluorescence intensity was measured by using a Tecan GENios Plus microplate reader (Tecan, Switzerland) in a light-protected condition, at the excitation wavelength of 485 nm and an emission wavelength of 535 nm. The treatment-induced variation of fluorescence was kinetically measured over a time course of 5 h. All fluorescence measurements were corrected for the background fluorescence and protein concentration. Results were evaluated by comparing five measurements and expressed as a means  ±  standard deviation (SD) of the relative fluorescence unit (RFU) values.

Determination of cell proliferation

Cell proliferation was assessed by using the BrdU assay (Roche CH), a chemiluminescent immunoassay based on the determination of BrdU incorporation during DNA synthesis [45,46]. When cells are exposed to BrdU, the compound is incorporated into the newly synthesized DNA strands of actively proliferating cells. DNA-incorporated BrdU can be measured using anti-BrdU antibodies, allowing the assessment of the DNA synthesizing cells. Subconfluent HPMECs were treated for 48 h with 10% (v/v) of serum from different subjects. BrdU was added 12 h before the end of the experiments, then the supernatant was removed, and cells were fixed for 30 min with a fixing-denaturating solution (Fix-Denat). At the end of incubation time, the Fix-Denat solution was discarded, and the cells were incubated for 90 min with a horseradish peroxidase-conjugated anti-BrdU antibody. After three washing, the substrate solution was added and allowed to react for 3–10 min at room temperature. Within this time window, the horseradish peroxidase catalyzes the oxidation of diacylhydrazide, and the reaction product, decaying from its excited state, yields light. Finally, a GENios Plus microplate reader (Tecan) was used to read the light emission of treated cells. Results were expressed as a means ± SD of the relative light units/s (RLU/s) values.

Production of lentiviral particles

To assess the human collagen type-I (COL1A1) synthesis, a set of lentiviral particles was created and used to transduce HPMECs exposed to the different experimental treatments [24]. Specifically, the following lentivectors were used to produce the infecting lentiviral particles: pCOL1A1-turbo Green Fluorescent Protein (tGFP), a plasmid containing the green fluorescent protein (GFP) driven by the human COL1A1 promoter; pEFα-LV-FP602, a plasmid containing the red fluorescent protein (FP602) driven by the elongation factor 1-alpha (EFα), which is a constitutive promoter. pCOL1A1-tGFP activation is associated with collagen synthesis, whereas the activation of pEFα-LV-FP602 expresses the red protein FP602 used as a control for normalizing the cell transduction efficiency. Transient transfection of the 293T/17 packaging cells was used to produce the lentiviral particles [25,47]. Briefly, 70% confluence 293T/17 cells were mixed with the transgene expression plasmid (pCOL1A1-tGFP or EFα-FP602), and then the second generation packaging plasmids (pMD2.G and pCMVR8.74) were used to transfect the cells. A calcium-phosphate solution consisting of a 1:1 mixture of 0.25 M of CaCl2:2X BBS (0.28 M NaCl, 0.05 M N,N-bis-(2-hydroxyethyl)-2-aminoethanesulfonic acid, 1.5 mM Na₂HPO₄) was used to transfect the cells. The medium was changed 12–16 h posttransfection, and virus-containing media was harvested twice at 24 and 48 h after changing the medium. The particle-containing medium was centrifuged for 2 h at 19,400 rpm in a L8-80M ultracentrifuge (Beckman, Brea, California, United States), and then, the pellet was washed in 1 mL Hanks’ Balanced Salt Solution (HBSS) and centrifuged again for 2 h at 21,000 rpm in a L8-80M ultracentrifuge (Beckman). The final pellet was resuspended in 200 µL of HBSS, vortexed for 2 h before being stored at −80°C in 20 µL aliquots. Virus titer was assessed by using the p24 ELISA kit (PerkinElmer, Waltham, Massachusetts, United States) according to the manufacturer instruction.

Generation of HPMEC/pCol1GFP-pEFα-FP602 cell lines

Experiments were performed on stable transfected HPMECs containing pCOL1A1-tGFP and EFα-FP602 [48,49]. In brief, the produced lentiviral particles were used to transduce 50–60% confluent HPMECs using an optimized multiplicity of infection to reach an infection efficiency of about 95%. To this end, cells were incubated for 24 h with the lentiviral particles; then, an equal amount of free-virus fresh medium was added for another 24 h. At the end of the transfection process, a flow cytometer analysis was performed to assess the infection efficiency. The obtained stable transduced HPMECs were then used for determining the variations of COL1A1 synthesis under the different experimental conditions.

Collagen determination

COL1A1 synthesis was assessed using the created HPMEC/pCol1GFP-pEFα-FP602 stable cell lines. We used this specific method as it allowed us to per a real-time determination of multiple samples in a 96-well plate using small aliquots of subject sera [24,40,41]. The stable transfected HPMEC/pCol1GFP-pEFα-FP602 cell line was stimulated with basal medium containing 10% sera from HD and 10% (v/v) of sera from SSc patients (SSc). Potential variation of GFP fluorescence induced by the treatments was kinetically assessed over a time-course of 4 h using a Tecan GENios Plus multiplate reader (Tecan, Mannedorf, Switzerland). Fluorescence quantification was performed using 485 and 535 nm as excitation wavelengths and 535 and 590 nm as emission wavelengths, respectively, for pCOL1A1-LV-tGFP and EFα-LV-FP602, and pCOL1A1-LV-tGFP to EFα-LV-FP602 ratio was used to normalize the data for the transduction efficiency. Data were then expressed as a means ± SD of the RFU values.

Protein extraction

HPMECs were grown in T12.5 culture flasks (BD Falcon). Subconfluent cells were treated with 10% (v/v) sera from HD subjects and SSc patients for 72 h. After the treatment, cells were washed with chilled PBS and then incubated in ice-cold lysis buffer (CytoBuster; Novagen, Darmstadt, Germany) containing protease and phosphatase inhibitors (Sigma, St Louis, MO) for 10 min at 4°C. Cells were then scraped, and the lysate was centrifuged at 16,000g for 5 min at 4°C. The supernatant was collected and stored in the same lysis buffer at −80°C. Protein content was measured using the Bradford assay as per the manufacturer’s protocol (Sigma, St Louis, MO).

Determination of alpha-smooth muscle actin (α-SMA) and von Willebrand factor (vWF)

Variation of α-SMA and vWF protein expression in response to sera treatment was quantitatively assessed as previously described by enzyme immunoassay [41]. Briefly, after 48 h of treatment with the sera, HPMECs were washed and collected in the proteins extraction buffer. A total of 100 µL of peroxidase-labeled anti-α-SMA and peroxidase-labeled anti-vWF monoclonal antibody solutions were transferred, respectively, into anti-α-SMA and vWF monoclonal antibody-coated wells. Subsequently, 20 µL of cell extract or standard was added, then the solution was incubated for 3 h at 37°C. Then 100 µL of substrate solution was added to each well. The reaction between the peroxidase-labeled anti-α-SMA and -vWF monoclonal antibodies and substrate results in color development that yield an intensity proportional to the quantity of α-SMA and -vWF present in samples and standards. Finally, 100 µL of stop solution was added to each well after 15 min of incubation at room temperature. The amount of α-SMA and vWF was quantitated by measuring the absorbance (450 nm) using a Tecan GENios Plus microplate reader (Tecan, Switzerland). Then, the concentrations of the two target proteins in the samples were determined by comparing their specific absorbances with those obtained for the standards plotted on a standard curve. All samples were normalized for protein content before proceeding with the ELISA determination. Results were expressed as a means ± SD of ng/mL protein.

RNA isolation and quantitative real-time polymerase chain reaction (PCR)

Total RNA was extracted using the miRNeasy kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions, with minor adjustments. For RNA extraction, 200 µL of the sample (serum) was used with 1 mL of QIAzol. 10 µL (from a 5 fmol/mL stock) of a synthetic analog of the nonhuman Caenorhabditis elegans microRNA-39 (celmiRNA-39; Qiagen) was spiked to normalize RNA extraction efficiency. The TaqMan miRNA reverse transcription kit and the miRNA-specific stem-loop primers system (Thermo Fisher Scientific) were used to perform individual miRNAs reverse transcription. The qPCR was performed in triplicate using 2× Universal PCR Master Mix with No AmpErase UNG (Thermo Fisher Scientific) using the QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific). Cel-miRNA-39 was used to normalize the expression of the studied miRNA [45,50].

Bioinformatic analysis

miRecords was used to determine the potential target genes of selected miRNAs. This database consists of two components: validated targets and predicted targets. The first is a database of experimentally validated miRNA targets resulting from literature. In contrast, the predicted targets is a database of predicted miRNA targets produced by integrating 11 established miRNA target prediction programs [51]. To localize miRNAs targets in signaling pathways, we used miTALOS database [52]. Finally, we used miRwayDB [52] to evaluate miRNAs–pathways association, a database providing comprehensive information of experimentally validated target genes and microRNA–pathway associations in various pathophysiological conditions.

Statistical analysis

Kruskall–Wallis one-way analysis of variance, followed by a post hoc Dunn’s test for multiple comparisons, was used to detect differences among the studied groups. When appropriate, a two-way analysis of variance, followed by a post hoc Šídák’s test for multiple comparisons, was used to detect differences among studied groups. Differences between two groups were detected using the Mann–Whitney–Wilcoxon test. All statistical analyses were performed using GraphPad Prism version 9.00 for Windows (GraphPad Software, San Diego, CA, USA), and p-values <0.05 were considered statistically significant.

SSc sera elicit an NADPH oxidase-dependent increase of ROS production in HPMECs

A growing number of evidence underline the role of oxidative stress in SSc pathogenesis [8,53]. Besides contributing to the disease prolongation, high ROS levels are also the leading cause of endothelial dysfunction [6], which appears to trigger many events underpinning SSc-associated vasculopathy [54]. We earlier reported that the exposition of human pulmonary vascular smooth muscle cells to sera of SSc patients elicited an NADPH oxidase-mediated ROS increase responsible for collagen synthesis overexpression [40]. Here, we hypothesize that a similar mechanism implicating ECs may be involved in SSc-associated vascular damage and fibrosis. As pulmonary fibrosis is the primary cause of mortality in SSc [55], HPMECs have been used in this study. For intracellular ROS levels detection, subconfluent HPMECs were loaded with 10 µM H₂-DCFDA and then cultured in basal medium containing 10% (v/v) of sera from both SSc patients and healthy donor (HD). Variations in intracellular ROS levels were kinetically determined in an 8 h time-course experiment (Figure 1a), and values at 2 h were used for comparison (Figure 1b). Fluorescence data were normalized for protein content and expressed as RFU. As reported in Figure 1a, intracellular ROS levels in SSc-treated HPMECs were significantly higher than those observed in cells exposed to HD sera, suggesting that SSc patients harbor circulating factors able of increasing ECs’ intracellular oxidative status. Moreover, the SSc sera-induced ROS increase appears to be mediated by NADPH oxidases, as cell pretreatment with the broad NADPH inhibitor DPI significantly attenuated the observed surge of ROS (Figure 2b).

Figure 1

Effect of SSc sera on HPMEC intracellular ROS levels: (a) SSc sera increase intracellular ROS production in HPMEC. The real-time assessment of intracellular ROS production in HPMECs exposed to sera of SSc patients and healthy donor. Before sera stimulation, subconfluent HPMECs were loaded with 10 µM of H₂-DCFDA and then cultured in basal medium containing 10% (v/v) of sera from SSc patients (SSc) and healthy donor (HD). Variations in intracellular ROS levels were kinetically determined in an 8 h time-course (a randomly selected representative experiment is reported) and values at 2 h were used for comparisons in B. (b) SSc sera-induced ROS increase in HPMECs is mediated by NADPH oxidases. Before sera stimulation, subconfluent HPMECs were loaded with 10 µM of H₂-DCFDA and then cultured for 8 h in basal medium containing 10% (v/v) of sera from SSc patients and HD, then values at 2 h were used for comparisons. In selected experiments, H₂-DCFDA-loaded subconfluent HPMECs were preincubated for 1 h with 5 µM of NADPH oxidase inhibitor DPI before sera stimulation. Fluorescence data were normalized for protein content and expressed as mean ± SD of the RFUs. (b) Horizontal lines indicate the mean value and standard deviation range. Kruskall–Wallis one-way analysis of variance, followed by a post hoc Dunn’s test for multiple comparisons, was used to detect differences among studied groups. *p ≤ 0.001, significantly different from the control group (HD); #p ≤ 0.0001, significantly different from the SSc group.

Figure 2

Effect of SS sera on HPMECs proliferation: (a) SSc sera induce HPMECs proliferation. Subconfluent HPMECs were cultured for 24 and 48 h in basal medium containing 10% (v/v) of sera from SSc patients and healthy donor. (b) SSc sera-induced proliferation increase in HPMECs is mediated by NADPH oxidases. Subconfluent HPMECs were cultured for 48 h in basal medium containing 10% (v/v) of sera from SSc patients and healthy donor. In selected experiments, cells were incubated for 1 h with 5 µM of the NADPH oxidase inhibitor DPI before exposure to the sera. Fluorescence data were normalized for protein content and expressed as mean ± SD of the RFUs. (a) Horizontal lines indicate the mean value and standard deviation range. Two-way analysis of variance, followed by a post hoc Šídák’s test for multiple comparisons, was used to detect differences among studied groups. (b) Horizontal lines indicate the mean value and standard deviation range. Kruskall–Wallis one-way analysis of variance, followed by a post hoc Dunn’s test for multiple comparisons, was used to detect differences among studied groups. *p ≤ 0.0004, significantly different from the control group (HD); #p ≤ 0.002, significantly different from the SSc group.

NADPH oxidase-derived ROS mediate increased proliferation in HPMECs exposed to SSc sera

In physiological conditions, the endothelium is genetically stable and remains in a quiescent state for a prolonged time [56]. Nonetheless, ECs own the ability to promptly switch to a migratory or proliferative status when activated by proangiogenic or proinflammatory stimuli [57]. In SSc, the overproduction of ROS is responsible for endothelial dysfunction and for promoting a profibrotic and proinflammatory state [8,22,54]. Indeed, based on their concentration, ROS can be important modulators of physiological and pathological cell proliferation. For instance, while persistent ROS production can generally induce apoptosis through JNK pathway activation, a transient increase of ROS levels can elicit cell proliferation through extracellular signal-regulated kinase (ERK) pathway activation [58]. In the attempt to understand whether the observed variations of intracellular ROS could affect HPMECs proliferation, we assessed the rate of DNA synthesis in cells exposed to SSc and HD sera for both 24 and 48 h. As depicted in Figure 2a, a significant increase in cell proliferation was elicited by SSc sera compared to HD sera. Moreover, similarly to data in Figure 1b, NADPH inhibition significantly blunted the increase in cell proliferation elicited by SSc sera (Figure 2b). Besides being consistent with the SSc fibro-proliferative features, this finding confirms the relationship between ROS and cell proliferation, suggesting NADPH-activating prooxidant factors underpinning SSc-induced ROS and proliferation increment observed in HPMECs treatment with SSc sera.

NADPH oxidase-derived ROS mediates increased collagen synthesis in HPMECs exposed to SSc sera

Among the connective macromolecules, a persistent collagen overproduction is primarily responsible for the progressive nature of SSc fibrosis [59]. Although activated myofibroblasts appear to be the primary source of collagen in SSc, it is becoming evident that also ECs play an important role in the disease-associated tissue restructuring by synthesizing collagen and contributing to vessel/microvessel narrowing [60]. In this light, we investigate whether circulating prooxidant factors present in SSc sera could contribute to the SSc-associated fibrotic phenotype by augmenting the synthesis of collagen type l (COL1A1). COL1A1 synthesis, in both the presence of SSc and HD sera, was detected using the HPMEC/pCol1GFP-pEFα-FP602 cell line, obtained by the cotransduction of HPMECs with both pCOL1A1-tGFP and pEFα-FP602 lentiviral particles [24,40]. This technique offers the advantage of kinetically assessing multiple samples simultaneously in a 96-well plate using small amounts of subject sera. As indicated in Figure 3a, the exposure of HPMECs to SSc sera resulted in a progressive time-dependent increase of the COL1A1 promoter activity with values at 2 h significantly higher in cells exposed to SSc sera compared to HD sera (Figure 3a and b). As found for the proliferation (Figure 2a), also the SSc-induced increase of COL1A1 synthesis appeared to be mediated by NADPH-activating prooxidant factors as it was significantly counteracted by DPI (Figure 3b).

Figure 3

Effects of SSc sera on HPMECs collagen (COL1A1) promoter activation. (a) SSc sera induce HPMECs collagen synthesis. Subconfluent HPMECs were transduced with lentiviral particles obtained from the COL1A1-LV-tGFP and EF1α-LV-FP602 lentivectors, and then cultured in basal medium containing 10% (v/v) of sera from SSc patients (SSc) and healthy donor (HD). Variations of COL1A1 promoter activation were kinetically followed for 48 h (a randomly selected representative experiment is reported) and values at 8 h were used for comparison in B. (b) SSc sera-induced COLIA1 increase in HPMECs is mediated by NADPH oxidases. Before sera stimulation, subconfluent HPMECs were transduced with lentiviral particles obtained from the COL1A1-LV-tGFP and EF1α-LV-FP602 lentivectors and then cultured for 48 h in basal medium containing 10% (v/v) of sera from SSc patients and HD; then values at 2 h were used for comparisons. In selected experiments, viral-transduced HPMECs were preincubated for 1 h with 5 µM of the NADPH oxidase inhibitor DPI before sera stimulation. (a) and b) Data are expressed as normalized transduction efficiency by reporting the mean ± SD of the RFUs ratio of COL1A1-LV-tGFP to EF1α-LV-FP602. (b) Horizontal lines indicate the mean value and standard deviation range. Kruskall–Wallis one-way analysis of variance, followed by a post hoc Dunn’s test for multiple comparisons, was used to detect differences among studied groups. *p ≤ 0.0005, significantly different from the control group (HD); #p ≤ 0.005, significantly different from the SSc group (SSc).

NADPH oxidase-derived ROS mediates EndMT in HPMECs exposed to SSc sera

Besides being involved in extracellular matrix (ECM) remodeling, collagen is tightly linked to EMT and EndMT [61,62]. Moreover, as earlier mentioned, activated myofibroblasts appear to be the primary cells involved in SSc-associated fibrosis [21,63], and ECs represent a very likely source of myofibroblasts originated through EndMT. Emerging evidence also indicate a link between oxidative stress and EndM in SSc; in fact, antioxidants reduce EndMT in this disease while oxidative stress mediates its onset and progression [22]. These evidences, along with the increase of ROS and cell proliferation observed following exposure to SSc sera, led us to hypothesize that HPMECs may undergo an EndMT-like mesenchymal transformation, elicited by sera-contained prooxidant factors. During EndMT, ECs progressively lose their typical endothelial markers, acquiring a mesenchymal phenotype [64]. To gain molecular insights on our hypothesis, we checked the protein levels of α-SMA and vWF in HPMECs exposed to both SSc and HD sera. Indeed, α-SMA and vWF are two typical mesenchymal and endothelial markers, respectively. Although α-SMA is essential for the cell contractile activity [14,65], vWF plays a key role in the endothelium pathophysiology by recruiting platelets into the sites of vascular injury [66]. Protein levels of both α-SMA and vWF were assessed by ELISA in HPMECs exposed for 2 days to sera from SSc patients and HD. Consistently with the establishment of EndMT, significant alterations of the two proteins expression were detected. Indeed, α-SMA levels were higher in cells exposed to SSc sera than cells exposed to HD (Figure 4a), whereas a lower expression vWF was noted in cells exposed to SSc sera compared to cells exposed to HD (Figure 4b). Similarly to COL1A1 (Figure 3b), the pretreatment of HPMECs with DPI significantly blunted SSc sera-induced EndMT, indicating the involvement of NADPH-activating prooxidant factors in the observed process.

Figure 4

Effects of SSc sera on EndMT markers expression in HPMECs (a) SSc sera induce α-SMA expression in HPMECs. (b) SSc sera inhibits vWF expression in HPMECs. (a and b) Subconfluent HPMECs were cultured for 48 h in basal medium containing 10% (v/v) of sera from SSc patients (SSc) and healthy donor (HD). Cells were then processed for the determination of both α-SMA and vWF as reported in the material methods. In selected experiments, HPMECs were preincubated for 1 h with 5 µM of the NADPH oxidase inhibitor DPI before sera stimulation. (a and b) Data are expressed as nanogram per milliliter of protein. Horizontal lines indicate the mean value and standard deviation range. Kruskall–Wallis one-way analysis of variance, followed by a post hoc Dunn’s test for multiple comparisons, was used to detect differences among studied groups. (a) *p ≤ 0.006, significantly different from the control group (HD); #p ≤ 0.008, significantly different from the SSc group (SSc). (b) *p ≤ 0.01, significantly different from the control group (HD); #p ≤ 0.02, significantly different from the SSc group (SSc).

SSc sera contain profibrotic and EndMT associated miRNAs

Recent studies reported an association between miRNAs and ROS in vascular diseases [67], indicating miRNAs’ ability to modulate redox-regulated pathways involved in cardiac and pulmonary diseases [68,69]. Indeed, miRNAs have been reported to be essential regulators of NADPH oxidase-mediated angiogenic response and modulators of the expression of genes involved in EndoMT [23,70,71,72]. These evidences suggest miRNAs may represent a mechanistic connection between ROS, endothelial dysfunctions, and SSc-associated fibroproliferative alterations. Therefore, we speculated that SSc patients’ sera contain profibrotic miRNAs, which may drive the currently observed in vitro phenomena and thus be involved in SSc-associated vascular damage and fibrosis in vivo. To test our hypothesis, we investigated the serum content of nine miRNA (miR-21; miR-24; miR-181b; miR-92a; miR-27b; miR-29 a,b,c; and miR-125b) implicated in SSc pathogenesis and altered in sera/tissues of SSc patients [38,73,74]. The relative levels of all detected miRNAs are shown in Figure 5. We found most of the miRNAs significantly increased in the SSc patients’ sera compared to normal sera, except for miR-29b and miR-29a, which were instead, decreased (miR-29b), and not significantly altered (miR-29a; Figure 5). Combining information obtained by both in silico analysis (by miRNA targets/pathway prediction tools) and literature data, these altered miRNAs appeared linked to a signaling cross-talk between ROS and a set of pathways involved in the pathophysiological processes of fibrosis such as TGF-β/SMAD, PI3K/AKT, and ERK/MAP. Moreover, most of these miRNAs have also been validated for being involved in the over-deposition of ECM components, EndMT, and oxidative stress. Table 2 summarizes the known information related to the miRNAs detected in this study. For instance, miR-21 has been reported linked to all the above-mentioned fibrotic pathways [75], and it might thus play a significant role in driving the phenomena we observed following HPMEC treatment with SSc sera. Indeed, the overexpression of miR-21 decreases Smad7, a negative TGF-β regulator, thus activating TGF-β-induced ECM production and consequently increasing fibrosis [76]. However, TGF-β1 and Smad3 upregulate miR-21 expression, which in turn activates TGF-β/Smads signals [77,78,79,80]. Additionally, via modulation of the tissue inhibitor of metalloproteinases 3 (TIMP3) activity, TGF-β1 regulates fibrosis by promoting ECM production and inhibiting ECM degradation [81]. Similarly to TGF-β1, miR-21 might regulate the fibrotic process by the same dual mechanism. Indeed, in SSc, miR-21 regulates the expression of genes responsible for ECM production, such as collagen, fibronectin, and α-SMA, and TIMP-3 is one of the most validated miR-21 target genes [76]. TIMP-3 is also a target gene of miR-181b, which we found increased in SSc sera along with miR-21. Besides the TGF-β/Smads pathway, PI3K/AKT and ERK/MAPK are also two miR-21-regulated signaling pathways involved in fibrosis pathophysiology [75]. In the ERK/MAPK-mediated fibrosis, both miR-21 and miR-29a-b target the sprouty (Spry) genes, which encode potent ERK/MAPK inhibitors. However, they play opposite roles as miR-21 is a Spry inhibitor and hence activates ERK/MAPK-associated fibrosis, whereas miR-29a-b are activators and, therefore, block the fibrotic process mediated by ERK/MAPK [82]. This aspect is consistent with our results reporting the increased expression of miR-21 and downregulation of miR-29a-b in SSc sera. Our current results showing increasing cell proliferation, collagen synthesis, and α-SMA expression on HPMECs treatment with SSc sera (Figures 2–4) appear to be EndMT-associated events ascribable to the ECs differentiation into myofibroblasts. PI3K/AKT signals are also involved in EMT and EndMT [83,84], and miR-21 might regulate EndMT through this signaling pathway. Indeed, the endogenous inhibitor of AKT activation, phosphatase, and tensin homolog (PTEN) is inhibited by miR-21 [85,86]. Moreover, PTEN is also a target of miR-92a (overexpressed in SSc sera), which has been reported to increase cell proliferation and promote EndMT and EMT [87,88,89]. miR-92a can also regulate collagen accumulation by downregulating the metalloproteinase-1, a protein involved in ECM breakdown [74,90].

Figure 5

Content levels of miRNAs present in serum of HD and SSc patients. The expression of the indicated microRNAs (miRNAs) was measured (real-time quantitative polymerase chain reaction) in the serum of healthy donor (HD) and SSc patients (SSc) samples. miRNA levels as expressed as relative expression normalized to levels of cel-miRNA-39. Horizontal lines indicate the mean value and standard deviation range. Mann–Whitney test was used to detect differences among HD and SSc groups.

miR-29a, miR-29b, and miR-29c also play an important role in collagen synthesis regulation. In fact, this miRNAs family, besides being altered in SSc sera/plasma and involved in the fibrotic process [73,91], also targets most collagen encoding genes. In this regard, our current findings show miR-29c upregulated, whereas miR-29b and miR-29a are respectively downregulated and not significantly altered. The interplay between the members of this miRNA family is still unclear, but it has been suggested that miR-29a and miR-29b are collagen synthesis repressors, whereas miR-29c is an activator [92]. We believe that a similar mechanism might be responsible for the increased collagen synthesis observed upon HPMECs exposition to SSc sera.

By activating the TGF-β pathway, miR-24 and miR-27b, two other miRNAs that we found upregulated in SSc sera, are positive regulators of EndMT [93]. In addition, miR-27b is either target or regulator of endothelin-1 (ET-1), a potent vasoconstrictor secreted by ECs and involved in SSc pathophysiology [94,95,96]. A similar regulatory mechanism has been reported between ROS and ET-1, where ROS can increase ET-1 production [97,98] and ET-1 can, in turn, increase ROS levels [54,99]. In SSc, ROS modulate most of the key components involved in the disease pathophysiology, including pathways related to the fibrotic process [8]. For instance, TGF-β and ROS show a reciprocal regulation in fibrosis induction [100], ROS can induce EMT via the TGF-β1/PI3K/Akt/pathway [101], and the ERK/MAPK pathway can be activated by increased ROS [102]. Interestingly, most of the miRNAs we analyzed in SSc sera are implicated in the same pathways, and some of them (miR-21, miR-24, miR-125b, and miR-29b) can increase ROS levels or being modulated by ROS [103,104,105,106]. Considering our experimental results, the in-silico analysis, and the literature, data suggest that miRNAs present in the patients’ sera could drive EndMT in SSc by interacting with the intracellular redox system and regulating fibrosis through a cross talk among the three main fibrosis-associated signaling pathways.