A potential new approach for treating systemic sclerosis: Dedifferentiation of SSc fibroblasts and change in the microenvironment by blocking store-operated Ca2+ entry

Transforming growth factor-β (TGF-β) is an important target for treating systemic sclerosis (SSc). However, our study revealed three levels of TGF-β1 expression in SSc patients, indicating that inhibiting TGF-β is not sufficient to treat SSc. A previous clinical trial also displayed disappointing results. Thus, our study attempted to search for a potential novel approach. Ingenuity Pathway Analysis (IPA) indicated that the SSc pathological pathways were closely associated with store-operated Ca2+ entry (SOCE)-regulated signals, and SOCE activity was found to be increased in SSc fibroblasts. Further treatment of SSc fibroblasts with SOCE inhibitors, 2APB, and associated calcium channel inhibitors SKF96365, and indomethacin, showed that the SOCE inhibitors selectively decreased fibrosis markers and altered the cell morphology. Consequently, SOCE inhibitors, especially 2APB and indomethacin, caused the dedifferentiation of SSc fibroblasts via cytoskeleton remodeling and altered collagen secretion and restored the cell mobility. We further explained SSc pathogenesis as fibroblast differentiation with SOCE. Treatment with exogenous factors, gelatin-1, FAM20A and human albumin, which were identified from the conditioned medium of SSc fibroblasts, was important for regulating the differentiation of fibroblasts with higher levels of SOCE and α-SMA. Conclusively, to treat SSc, blockage of the increased SOCE activity in SSc induces the dedifferentiation of SSc fibroblasts and simultaneously changes the extracellular matrix (ECM) structure to limit SSc pathogenesis.


Introduction
Systemic sclerosis (SSc), a severe multisystem autoimmune disease, is characterized by progressive fibrosis that can influence all organs in the body [1]. Previous study indicated the majority of SSc deaths involve pulmonary fibrosis, pulmonary arterial hypertension and cardiac causes [2].However, the low efficacy of immunosuppressive treatments suggest a complicated pathogenesis of fibrosis and unknown mechanisms in SSc. For treating fibrotic diseases, especially SSc, several recent studies focused on transforming growth factor-β (TGF-β) as a potential target for anti-fibrotic therapy because TGF-β is a crucial mediator of fibrosis [3].
The excess TGF-β activity in SSc contributes to overproduction of collagen, a component of the extracellular matrix (ECM) [3]. TGF-β or/and TGF-β-induced collagen overproduction can cause the differentiation of fibroblasts to myofibroblasts, and myofibroblasts also accelerate fibrosis by changing ECM collagen fibers [4]. However, blockage of TGF-β to treat fibrosis still faces critical challenges; monoclonal antibodies targeting TGF-β have been used in a clinical trial with disappointing results to date [5]. There is still no efficient approach for treating fibrotic diseases, especially SSc.
To identify an efficient approach for treating fibrotic diseases, we attempted to search for the key factor that is important in regulating SSc development. In addition, in the process of fibrosis in SSc, fibroblasts play a crucial role in interrupting the balance between the ECM and matrix metalloproteinase; in particular, myofibroblasts, which are differentiated from fibroblasts, can dramatically change the ECM structure. Therefore, our study focused on the mechanism underlying the differentiation of fibroblasts in SSc and evaluated whether disrupting myofibroblast activity can inhibit fibrosis.

Cell culture
SSc-fibroblasts were isolated by skin biopsy (5-mm punch) from cutaneous SSc, the area selected were typically tight and firm in clinical condition, and were confirmed by dermatologist, with clinical diagnosis according to the American College of Rheumatology (ACR) criteria as previously described [6]. To avoid inter-individual biologic variation, the control fibroblasts were obtained from paired normal skin tissue of the same patient, which located in symmetrical part of SSc or surrounding of SSc where the texture were relatively normal and not rigid in contrast to the diseased area. The mean age of patients was 45 y (range, 20-68 y), 55% female. Patients had no treatment by any immunosuppressive or steroid therapy. All skin biopsies from SSc patients (n = 20) in this study were received with approval from the Institutional Review Board/Ethics Committee (IRB), number KMUHIRB-F(II)-20160072 in Kaohsiung Medical University Hospital, and approval signed consent were obtained for patients. The isolated procedure of human dermal fibroblasts as described previously [7].

LC/MS E analysis and IPA
Pretreatment was performed with or without 2APB, SKF96365 and indomethacin in fibroblasts or SSc fibroblasts. After 72 h, cell lysates and conditioned media were collected to analyze the components by using a Waters Xevo G2 qTof mass spectrometer (Waters) as previously described [8]. After LC/MS E analysis, we screened for significant differences in the expression of SSc fibroblasts compared with the paired adjacent fibroblasts in each group by Progenesis QI for Proteomics (QIP) (Waters). For IPA, we uploaded the molecules obtained from the first screening by QIP and identified the common molecules (intersection) among the three groups. The molecules from the intersection among the three groups were further searched to determine the corresponding signaling pathways and associated diseases.

Wound healing assay
A wound healing assay was performed using IBIDI Culture-Inserts (IBIDI). In brief, 10 4 cells were plated in each well and were treated with SOCE inhibitors after attachment to the well. After incubating for 24 h, the culture inserts were removed. The movement of cells into the gap area was captured using an Invitrogen EVOS FL Auto 2 Cell Imaging System (Thermo Fisher Scientific) at the indicated time until the gap area had closed. The images of the gap area from the wound healing assay were then measured using ImageJ software (Fiji).

Immunofluorescence staining
Fixed cells were incubated with primary antibodies against vimentin (Abcam), fibronectin (Abcam), and α-SMA (Abcam) at 4˚C overnight. The cells were subsequently incubated with secondary antibody for 1 h and then with 4',6-diamidino-2-phenylindole (DAPI, Thermo Fisher Scientific) for 5 min. Coverslips were inverted and fixed onto glass slides by antifade mounting reagent (Biotium). The fluorescence imaging was performed by a confocal microscope (Olympus FV1000).

Statistical analysis
GraphPad Prism (La Jolla, CA, USA) was used to generate bar charts; error bars indicate the SDs, unless otherwise noted. All statistical analysis data was analyzed by GraphPad-Prism Analysis software (GraphPad-Prism Software Inc., San Diego, CA. USA). Analysis of variance and Student's t-tests were utilized to compare the difference between groups. P-values of less than 0.05 were considered statistically significant for differences between groups.

TGF-β1 in skin was not expressed at an increased level in all SSc patients
Many studies indicated that TGF-β1 is highly expressed in SSc and is a critical target in treating fibrotic diseases [10]. However, our study revealed a completely different result for the expression of TGF-β1 in SSc. We first collected SSc skin tissue and paired normal skin tissue from patients and isolated fibroblasts from the SSc skin and the control skin ( Fig 1A). Additionally, the TGF-β1 expression in SSc fibroblasts and control fibroblasts was measured; not all samples displayed a higher level of TGF-β1 than the paired adjacent fibroblasts. Three distinct groups were found by comparing the TGF-β1 expression in the SSc fibroblasts (S) to that in the control group fibroblasts (F): no difference, a lower level and a higher level ( Fig 1B). We confirmed the TGF-β1 expression in the SSc skin tissue and the paired skin tissue with immunohistochemical (IHC) staining. The result was similar, as shown in Fig 1C and 1D; interestingly, TGF-β1 was expressed not only in fibroblasts but also in keratinocytes. Our results also indicated three levels of TGF-β1 expression were classified in SSc patients, although increased expression of TGF-β1 in SSc-fibroblasts was revealed with an average of total SSc biopsies (S1 Fig). If keratinocytes of SSc aggravate the activation TGF-β independent to fibroblast [11], extracellular factors may play a certain role in regulation of the microenvironment to induce the differentiation in fibroblasts.

Store-operated Ca 2+ entry (SOCE) was activated in SSc fibroblasts and regulated pathological signaling pathways
TGF-β1 levels are not increased in all SSc tissues, so blocking TGF-β may be insufficient to treat SSc or other fibrotic diseases. Therefore, we determined which factor was critically involved in SSc. As shown in S2A and S2B Fig, 16 molecules are common to the three groups, which are firstly screened by Progenesis QI for Proteomics (QIP), as indicated by Ingenuity Pathway Analysis (IPA). These 16 molecules are associated with increased risk of malignancy in various systemic diseases and disorders (S1 Table). We further classified these 16 molecules by their signaling pathways. The majority of these SSc pathological pathways are closely associated with Ca 2+ signals, as shown in S2C Fig; store-operated Ca 2+ entry (SOCE) is the most important Ca 2+ signal associated with inflammation and carcinogenesis [12].SOCE channels include many types of Ca 2+ channels, such as several transient receptor potential (TRP) channels [13], which are crucial in regulating the proliferation, differentiation, and ECM-protein production in fibroblasts [14]. We found that the SOCE activity was higher in SSc fibroblasts than in the control group fibroblasts by applying thapsigargin (TG) (Fig 2A and 2B). This study also applied three SOCE associated inhibitors, 2-aminoethoxydiphenyl borate (2APB), SKF96365 and indomethacin [6,9,15,16]. Indomethacin as a SOCE associated inhibitor [6,15] widely used in clinical treatment as an anti-inflammatory agent that provides analgesic effect to SSc patients, and we applied indomethacin in the study for clinical practical concern. Thus, although pretreatment by SKF96365 and indomethacin restrained TG-induced SOCE, the higher SOCE activity in SSc fibroblasts was significantly inhibited by 2APB in the three groups (Fig 2A and 2B). Our results suggested increased SOCE activity is a promising target for treating SSc.

SOCE inhibitors induced dedifferentiation in SSc fibroblasts
Next, we investigated the effect of SOCE inhibitors on the fibrotic pathogenesis in SSc fibroblasts. As shown in Fig 3A and 3B, fibrosis markers, such as TGF-β1, fibronectin and α-SMA, were not consistently expressed at an increased level in SSc fibroblasts in the three groups, although the cell morphology was changed to become similar to myofibroblasts [17], which are differentiated from fibroblasts ( Fig 3D). However, treatment with SSc fibroblasts with SOCE and associated inhibitors, 2APB, SKF96365 and indomethacin, led to a selective decrease in fibrosis markers in the three groups (Fig 3A and 3B), mainly the TGF-β1, α-SMA and α-tublin. Although there is discrepancy in the fibronectin and β-Actin expression, in general there is a trend of decreased expression. To note, SKF96365 and indomethacin also influenced the percentage of surviving cells (Fig 3C).
Interestingly, SOCE and associated inhibitors altered the cell morphology of SSc fibroblasts. Fibroblasts are typically bipolar or multipolar, have elongated shapes, and a smooth, regular surface with a large nucleus. In contrast, myofibroblasts, which can be induced by adding TGF-β1, is a more differentiated form of fibroblasts that is much larger and has an undulating membrane and multiple processes. The higher ratio of myofibroblast to fibroblast indicates the extent of fibrosis, in other words, the SSc pathogenesis. Fig 3D revealed that after applying 2APB and indomethacin restored the cell morphology of SSc fibroblasts to that of normal fibroblasts. In order to evaluate whether SOCE inhibitors restored the differentiated fibroblasts through dedifferentiation, we utilized TGF-β1-induced myofibroblasts and treated these cells with SOCE inhibitors for three days. As shown in S3 Fig, the TGF-β1-induced fibroblast differentiation was restored due to dedifferentiation by treatment with SOCE and associated inhibitors, especially 2APB and indomethacin. This finding can be observed by both the morphology of cells and also the expression of the cytoskeletons, mainly α-SMA, which is the well-known cytoskeleton that confined the mobility of myofibroblasts. Furthermore, we also calculated the ratio of differentiated fibroblasts in the three groups through identifying cell morphology, and found that the higher ratio of differentiated fibroblasts in the SSc fibroblast group was significantly decreased by treatment with the SOCE inhibitors (Fig 3D and 3E). Consequently, SOCE and associated inhibitors, especially 2APB and indomethacin, resulted in the dedifferentiation of SSc fibroblasts.

SOCE inhibitor-induced dedifferentiation of SSc fibroblasts was due to cytoskeleton remodeling and collagen secretion
Because SOCE inhibitors changed the cell morphology, we supposed that SOCE inhibitors might contribute to dedifferentiation via cytoskeleton remodeling. To prove this hypothesis, we first tracked the motion of cells and found that the SSc fibroblasts migrated in a round circular pattern. However, treatment with 2APB and indomethacin restored the migration pattern of SSc fibroblasts to a pattern similar to that of normal fibroblasts (Fig 4A), and also improved the migratory distance of SSc fibroblasts by calculating the cell migratory trace ( Fig  4B). Although the cell migratory capacity cannot be compared between the normal fibroblasts and the three groups of SSc fibroblasts, we found the migratory ability of SSc fibroblasts was increased by 2APB and decreased by SKF96365 (Fig 4C). In brief, 2APB restored the cell mobility of SSc fibroblasts. In addition, the cytoskeleton distribution of SSc fibroblasts was affected by the three types of SOCE inhibitors; the cytoskeleton, as well as the fibrosis markers vimentin, α-SMA and fibronectin, was rearranged and even decreased in response to treatment with SOCE inhibitors (Fig 4D). For instance, the shape and distribution of vimentin fibers reveal straight and tight arrangement in fibroblasts, but the crossed and inattentive arrangement are showed in SSc fibroblasts; treatment by SOCE inhibitors in SSc fibroblasts induced the rearrangement of vimentin fibers, changed the shape and arrangement of vimentin fibers. The similar results are as shown in α-SMA and fibronectin (Fig 4D). Notably, Furthermore, to consider other changes in addition to cytoskeleton remodeling, we further investigated whether SOCE inhibitors influenced the microenvironment by disrupting collagen secretion. SOCE is important in regulating cytokine release [9].We examined several types of collagen from the medium of SSc fibroblasts: the fibril-forming collagens, collagen types 1,2,3, 5 and 11-α1; the collagen of the basement membrane, collagen type 4; and fibril-associated collagens with interrupted triple helices, type 14 [18]. Our study suggested that SOCE associated inhibitors disrupted the secretion of certain collagens, mainly collagen type 4, and some selective changing in collagen patterns as well as an altered pattern of the ECM microenvironment (Fig 5). Thus, the dedifferentiation of SSc fibroblasts caused by SOCE inhibitors, especially 2APB and indomethacin, was due to cellular cytoskeleton remodeling and an altered ECM composition.

Exogenous factors were important in regulating the differentiation of fibroblasts with SOCE
We further attempted to explain SSc pathogenesis in terms of fibroblast differentiation and SOCE. Previously, we showed that extracellular factors may regulate the microenvironment to induce fibroblast differentiation. Conditioned medium was collected from three groups and used to treat human dermal fibroblasts from healthy people. There was no alteration in cell morphology or in the expression levels of cytoskeletal components and TGF-β1 in response to treatment with the conditional media (Fig 6A and 6C). Notably, the SOCE activity increased, and migratory ability decreased, as shown in Fig 6B and 6D. The condition medium could contain "something" that controlled the SOCE activity, cytoskeletal component expression and migratory ability. We further identified the extracellular factors that were present in the conditioned medium using the IPA system. In the screening of the three groups, gelatin-1, FAM20A and human albumin were the factors identified to be expressed at significantly higher or lower levels among the three groups (Fig 6E). Treatment with gelatin-1, FAM20A and human albumin promoted the expression of the fibrosis marker α-SMA and induced higher SOCE activity, which regulated fibrotic pathogenesis (Fig 6G, 6H and 6I). Treatment with human albumin obviously changed the cell shape with a differentiation from fibroblasts in a TGF-β1-independent process (Fig 6F, 6H and 6I), and restrained the cell migratory ability (S4 Fig). Our results indicated the differentiation of fibroblasts was dependent on extracellular signals, implying that a microenvironmental effect principally contributed to SSc pathogenesis.

Discussion
In this study we attempted to look for a potential novel approach that would be effective for SSc therapy. Although TGF-β1 is a crucial target in treating SSc, skin biopsies from SSc patients revealed that the TGF-β1 level differed from that of the paired normal skin biopsies. Consequently, based on IPA analysis, we suggested that SOCE inhibitors efficiently attenuated SSc pathogenesis, inducing the dedifferentiation of SSc fibroblasts by cytoskeleton remodeling and changes in collagen secretion. The cytoskeleton remodeling in SSc fibroblasts was due to selective inhibition of cytoskeleton components by SOCE inhibitors; the changes in collagen secretion contributed to disruption of the ECM structure and affected the microenvironmental signals applied to the cells. The microenvironment is the major factor resulting in SSc pathogenesis, and treatment with extracellular factors, such as human albumin, caused fibroblast differentiation. Thus, to treat SSc, endogenous blockage of excess Ca 2+ signals and exogenous disruption of the ECM structure have potential as novel approaches. Effect of deficient SOCE activity on SSc pathology Our study indicated that blocking the increased SOCE activity is an efficient method for treating SSc. SOCE is closely related to the regulation of voltage-gated Ca 2+ channels (VGCCs) [19].A previous report pointed out that a VGCC inhibitor, nifedipine, disrupted Ca 2+ signals in fibroblasts and inhibited pulmonary fibrosis in a bleomycin model [20].It is still unclear whether SOCE cooperates with VGCCs, but there is some possibility that inositol triphosphate (IP3; to activate SOCE) and Ca 2+ (entry by VGCCs) cause Ca 2+ oscillations through positive feedback [21]. A cell undergoing Ca 2+ oscillations usually shows increased cellular activity [22]. Nevertheless, VGCCs are known to be an essential players in controlling many physiological functions in excitable cells [23]; the role of VGCCs may be more important in excitable cells than in dermal fibroblasts. The relationship between SOCE and VGCCs in SSc fibroblasts should be further clarified in the future. In addition, may studies mentioned that SOCE is mediated by STIM1-ORAI1 [24], but our results revealed that the expression of STIM1 or ORAI1 did not change significantly (S5 Fig). Therefore we will propose here that the increased SOCE activity in SSc fibroblasts might be due to TRP channels, which are potential targets for treating SSc. We further identified the expression of TRPC channels, which are important in regulating SOCE in three groups of SSc, and found that the increased level of TRPC1, TRPC6 and TRPC7.
Interestingly, 2APB, a SOCE inhibitor used in our study, is a powerful candidate for treating SSc because of its low less toxicity and its efficiency in restoring cell mobility. Treatment with 2APB also significantly attenuated the secretion of multiple types of collagen, as shown in Fig 5. Additionally, a nonsteroidal anti-inflammatory drug (NSAID), indomethacin, is frequently utilized in improving the pain of SSc in clinic, but indomethacin has higher toxicity than 2APB (Fig 3C). Actually, treatment with indomethacin caused SSc fibroblasts to recover the normal morphology of fibroblasts, but we found that several cells had become thinner than normal fibroblasts, indicating a change in cell morphology. It is still unclear whether the toxicity of indomethacin contributes to the change in cell morphology. However, the toxicity of SKF96365 disrupts the cytoskeleton, resulting in a decrease in cell mobility (Fig 4D).
According to our study, the pathogenesis of SSc was dependent on direct exogenous effects, such as changes in ECM structure or extracellular molecules; accordingly, changing the microenvironment might be more efficient than targeting fibrosis markers to treat SSc. The fibrosis markers in SSc fibroblasts display different levels of expression; for instance, although we classified the expression of TGF-β1 into three groups, the expression of downstream molecules in TGF-β cascades did not change in a manner consistent with the TGF-β expression in the three groups. This finding that inhibiting the TGF-β1pathway to treat SSc is not sufficiently effective so far. Overall, our study supports the proposal of a new approach for treating SSc: blockage of the increased SOCE activity in SSc induces the dedifferentiation of SSc fibroblasts to restore the cell motility through cytoskeleton remodeling and simultaneously disrupts the ECM structure to limit SSc pathogenesis. fibroblasts promoted the TG-induced Ca 2+ influx. Right panel: the mean area under the intracellular Ca 2+ response curves after TG was applied (n >120 cells, mean ± SD, � , p<0.05; ��� , p<0.001). (C) Analysis of the expression levels of TGF-β1, α-SMA, and GAPDH after treatment with the conditioned medium for 24 h. (D) The conditioned media blocked the cell migration. (E) Three extracellular molecules, gelatin-1 (Gel-1), FAM20A, and human albumin (Alb), were screened by analysis with the IPA system, which identified the common components in the media from the three groups. Analysis of the expression of gelatin-1, FAM20A, and human albumin in the three groups by LC/MS (mean ± SD, � , p<0.05; �� , p<0.01; ��� , p<0.001). Treatment with gelatin-1 (0.25 ng/ml), FAM20A (1 μg/ml), and human albumin (2 mg/ml) in fibroblasts for 14 days and the subsequent effects on (F) cell morphology, (G) SOCE activity and (H) α-SMA expression. Right panel: the mean area under the intracellular Ca 2+ response curves after the addition of extracellular CaCl 2 (n >120 cells, mean ± SD, � , p<0.05; �� , p<0.01). (I) The quantification of (H) protein expression is shown (data were normalized to the protein expression of the internal control, GAPDH; mean ± SD, �� , p<0.01; ��� , p<0.001). https://doi.org/10.1371/journal.pone.0213400.g006 Effect of deficient SOCE activity on SSc pathology