TGF-Beta Induces Activin A Production in Dermal Fibroblasts Derived from Patients with Fibrodysplasia Ossificans Progressiva

Fibrodysplasia ossificans progressiva (FOP) is a catastrophic, ultra-rare disease of heterotopic ossification caused by genetic defects in the ACVR1 gene. The mutant ACVR1 receptor, when triggered by an inflammatory process, leads to heterotopic ossification of the muscles and ligaments. Activin A has been discovered as the main osteogenic ligand of the FOP ACVR1 receptor. However, the source of Activin A itself and the trigger of its production in FOP individuals have remained elusive. We used primary dermal fibroblasts from five FOP patients to investigate Activin A production and how this is influenced by inflammatory cytokines in FOP. FOP fibroblasts showed elevated Activin A production compared to healthy controls, both in standard culture and osteogenic transdifferentiation conditions. We discovered TGFβ1 to be an FOP-specific stimulant of Activin A, shown by the upregulation of the INHBA gene and protein expression. Activin A and TGFβ1 were both induced by BMP4 in FOP and control fibroblasts. Treatment with TNFα and IL6 produced negligible levels of Activin A and TGFβ1 in both cell groups. We present for the first time TGFβ1 as a triggering factor of Activin A production in FOP. As TGFβ1 can promote the induction of the main driver of FOP, TGFβ1 could also be considered a possible therapeutic target in FOP treatment.


Introduction
Fibrodysplasia ossificans progressiva (FOP) is an ultra-rare genetic disease characterized by progressive ossification of skeletal muscles and associated soft tissues. Heterotopic ossification (HO) in FOP typically transpires through so-called flare-ups: episodes of soft tissue swelling accompanied by other classical features of inflammation such as pain, redness, and warmth [1]. The inflammatory response induces subsequent development of bone through endochondral ossification at the site of injury [2].
Approximately 97% of FOP cases are caused by a single nucleotide substitution mutation (c.617G > A, R206H) in the gene coding for the Activin receptor type 1 (ACVR1)/Activin

SMAD Signalling in Fibroblasts of FOP Patients
SMAD1/5/9 phosphorylation was investigated in cultured primary fibroblasts derived from FOP patients and controls. The classic ACVR1 c.617G>A R206H mutation was confirmed by Sanger sequencing in the patient cell cultures (Supplementary Figure  S1). After overnight serum starvation, control and FOP patient fibroblasts both exhibited SMAD1/5/9 phosphorylation after BMP4 stimulation, and no differences between the FOP and control groups were observed ( Figure 1). Activin A stimulation led to SMAD1/5/9 phosphorylation in FOP but not in control fibroblasts, confirming the reported responsiveness of the mutated ACVR1 receptor to Activin A in this cell model [19,20].
by which inflammatory factors modulate Activin A production. To this end, we investigated the role of TGFβ1, TNFα, and IL-6 on Activin A and TGFβ1 production and expression of downstream TGFβ superfamily target genes in dermal fibroblasts derived from patients with FOP and healthy controls.

SMAD Signalling in Fibroblasts of FOP Patients
SMAD1/5/9 phosphorylation was investigated in cultured primary fibroblasts derived from FOP patients and controls. The classic ACVR1 c.617G>A R206H mutation was confirmed by Sanger sequencing in the patient cell cultures (Supplementary Figure S1). After overnight serum starvation, control and FOP patient fibroblasts both exhibited SMAD1/5/9 phosphorylation after BMP4 stimulation, and no differences between the FOP and control groups were observed ( Figure 1). Activin A stimulation led to SMAD1/5/9 phosphorylation in FOP but not in control fibroblasts, confirming the reported responsiveness of the mutated ACVR1 receptor to Activin A in this cell model [19,20].

Activin A Production by FOP Fibroblasts
In a previous study, we reported the increased potential for osteogenic transdifferentiation in the FOP fibroblasts. Fibroblasts were differentiated towards cells of an osteogenic lineage (osteoblast-like cells) and were validated regarding gene expression and mineralization assays [19]. Interestingly, this potential varied between the patient cell lines; a trend of higher osteogenic transdifferentiation was observed in three of the five FOP fibroblast lines (P2, P4, and P5), as shown by the increased expression of the RUNX2 and ALP osteogenic markers [19]. Considering the role of Activin A as a known driver of osteogenesis in FOP, its production was measured by ELISA in FOP fibroblasts supernatant in standard culture conditions. We found increased Activin A production in FOP fibroblasts compared to control cells ( Figure 2A) (unpaired t-test, p < 0.05), though the production gradually declined over the first three days ( Figure 2B). This difference was also present during the osteogenic transdifferentiation of the fibroblasts.

Activin A Production by FOP Fibroblasts
In a previous study, we reported the increased potential for osteogenic transdifferentiation in the FOP fibroblasts. Fibroblasts were differentiated towards cells of an osteogenic lineage (osteoblast-like cells) and were validated regarding gene expression and mineralization assays [19]. Interestingly, this potential varied between the patient cell lines; a trend of higher osteogenic transdifferentiation was observed in three of the five FOP fibroblast lines (P2, P4, and P5), as shown by the increased expression of the RUNX2 and ALP osteogenic markers [19]. Considering the role of Activin A as a known driver of osteogenesis in FOP, its production was measured by ELISA in FOP fibroblasts supernatant in standard culture conditions. We found increased Activin A production in FOP fibroblasts compared to control cells ( Figure 2A) (unpaired t-test, p < 0.05), though the production gradually declined over the first three days ( Figure 2B). This difference was also present during the osteogenic transdifferentiation of the fibroblasts.

Cytokine Regulation of Activin A Production by FOP Fibroblasts
The higher production of Activin A by FOP fibroblasts prompted us to investigate its regulation. Given the contribution of the inflammatory micro-environment in FOP lesions, Activin-A was quantified by ELISA after 24 and 48 h of stimulation with several disease-relevant recombinant cytokines such as TGFβ1, BMP4, TNFα, and IL6 ( Figure 3A). Stimulation with TGFβ1 resulted in significantly higher production of Activin A in FOP fibroblasts, both at 24 and 48 h of stimulation. Interestingly, the Activin A production was the highest in patients P2, P4, and P5. Stimulation with BMP4 also produced the same effect, although in this case, significant upregulation of Activin A production was also shown at 24 h. Again, P2, P4, and P5 were the most responsive patients to BMP4 stimulation, as shown by the higher Activin A production. Stimulation with TNFα did not lead to significant differences in Activin A production, and most cell lines showed levels close to the sensitivity limit of the ELISA (12 pg/mL). Similarly, low levels of Activin A were produced after stimulation with IL6 in all conditions (Supplementary Figure S2).

Cytokine Regulation of Activin A Production by FOP Fibroblasts
The higher production of Activin A by FOP fibroblasts prompted us to investigate its egulation. Given the contribution of the inflammatory micro-environment in FOP leions, Activin-A was quantified by ELISA after 24 and 48 h of stimulation with several isease-relevant recombinant cytokines such as TGFβ1, BMP4, TNFα, and IL6 ( Figure 3A). timulation with TGFβ1 resulted in significantly higher production of Activin A in FOP broblasts, both at 24 and 48 h of stimulation. Interestingly, the Activin A production was e highest in patients P2, P4, and P5. Stimulation with BMP4 also produced the same ffect, although in this case, significant upregulation of Activin A production was also Considering the clear stimulatory effect of TGFβ1 on Activin A production, we investigated to which extent TGFβ1 secretion could be, in turn, affected by cytokines. To this end, FOP and control cells were stimulated with Activin A, BMP4, and TNFα for 24 h and 48 h ( Figure 3B). Stimulation with Activin A did not reciprocate the effect on TGFβ1, as shown by the lack of statistically significant differences in TGFβ1 production; most cell lines produced TGFβ1 levels close to the sensitivity limit of the ELISA (80 pg/mL). On the contrary, BMP4 treatment led to significantly increased production of TGFβ1 in control cells at 24 h and in both FOP and control cells at 48 h. TNFα stimulation also resulted in low response in TGFβ1 production in all conditions, and vice versa, no detectable production of TGFβ1 was observed after TNFα stimulation (Supplemental Figure S2). As no differences were observed after TNFα or IL-6 stimulation, only the effects of TGFβ and Activin A were validated on a gene expression level.  . Activin A and TGFβ1 production measured by ELISA upon Activin A, TGFβ1, and BMP4 stimulation in FOP and control fibroblasts. Fibroblasts were subjected to overnight serum starvation (-FBS) before being stimulated with TGFβ1 and BMP4 for 24 and 48 h. This was followed by th measurement of (a) Activin A and (b) TGFβ1. Data is shown as the production level per cell line and their mean. * p < 0.05, ** p < 0.005, *** p < 0.0005 as determined by 2-way ANOVA.
Considering the clear stimulatory effect of TGFβ1 on Activin A production, we in vestigated to which extent TGFβ1 secretion could be, in turn, affected by cytokines. To this end, FOP and control cells were stimulated with Activin A, BMP4, and TNFα for 24 h and 48 h ( Figure 3B). Stimulation with Activin A did not reciprocate the effect on TGFβ1 as shown by the lack of statistically significant differences in TGFβ1 production; most cel lines produced TGFβ1 levels close to the sensitivity limit of the ELISA (80 pg/mL). On the contrary, BMP4 treatment led to significantly increased production of TGFβ1 in contro cells at 24 h and in both FOP and control cells at 48 h. TNFα stimulation also resulted in low response in TGFβ1 production in all conditions, and vice versa, no detectable produc tion of TGFβ1 was observed after TNFα stimulation (supplemental Figure S2). As no dif ferences were observed after TNFα or IL-6 stimulation, only the effects of TGFβ and Ac tivin A were validated on a gene expression level.

Expression of BMP and TGFβ Target Genes
Stimulation of the FOP ACVR1 receptor leads to the expression of BMP target genes downstream of pSMAD1/5/9 signaling [21]. Thus, we investigated if that is the case after stimulation with the cytokines Activin A, BMP4, and TGFβ1 for 6 h ( Figure 4A). As ex pected, stimulation with Activin A produced a significant increase in the relative gene expression of ID1 and ID2 (two-way ANOVA, p < 0.05) in FOP compared to control fibro blasts. BMP4 stimulation produced the same effect in both FOP and control cells, in agree ment with the lack of differences in pSMAD1/5/9 expression ( Figure 1A). Interestingly stimulation with TGFβ1 also led to a significant upregulation in ID1 expression in FOP cells (two-way ANOVA, p < 0.05), although this was not observed for ID2. No significan differences were found for BMP target genes DLX1 and MSX2 (Supplemental Figure S3A) Relative gene expression of the TGFβ target genes CTGF, COL1A1, and THBS1 was also measured in the same conditions (supplemental Figures S3B and S4B). Treatment with Activin A did not produce significant changes in CTGF expression, whereas treatmen with BMP4 did lead to a significant CTGF increase (two-way ANOVA, p < 0.05), in agree ment with the increased TGFβ production after BMP4 stimulation ( Figure 3B). As ex . Activin A and TGFβ1 production measured by ELISA upon Activin A, TGFβ1, and BMP4 stimulation in FOP and control fibroblasts. Fibroblasts were subjected to overnight serum starvation (−FBS) before being stimulated with TGFβ1 and BMP4 for 24 and 48 h. This was followed by the measurement of (a) Activin A and (b) TGFβ1. Data is shown as the production level per cell line and their mean. * p < 0.05, ** p < 0.005, *** p < 0.0005 as determined by 2-way ANOVA.

Expression of BMP and TGFβ Target Genes
Stimulation of the FOP ACVR1 receptor leads to the expression of BMP target genes downstream of pSMAD1/5/9 signaling [21]. Thus, we investigated if that is the case after stimulation with the cytokines Activin A, BMP4, and TGFβ1 for 6 h ( Figure 4A). As expected, stimulation with Activin A produced a significant increase in the relative gene expression of ID1 and ID2 (two-way ANOVA, p < 0.05) in FOP compared to control fibroblasts. BMP4 stimulation produced the same effect in both FOP and control cells, in agreement with the lack of differences in pSMAD1/5/9 expression ( Figure 1A). Interestingly, stimulation with TGFβ1 also led to a significant upregulation in ID1 expression in FOP cells (two-way ANOVA, p < 0.05), although this was not observed for ID2. No significant differences were found for BMP target genes DLX1 and MSX2 (Supplemental Figure S3A). Relative gene expression of the TGFβ target genes CTGF, COL1A1, and THBS1 was also measured in the same conditions (Supplemental Figures S3B and S4B). Treatment with Activin A did not produce significant changes in CTGF expression, whereas treatment with BMP4 did lead to a significant CTGF increase (two-way ANOVA, p < 0.05), in agreement with the increased TGFβ production after BMP4 stimulation ( Figure 3B). As expected, TGFβ stimulation led to significantly higher expression of CTGF both in FOP and control cells (two-way ANOVA, p < 0.05); no specific gene expression pattern was observed for COL1A1 and THBS1.

Gene Expression of Cytokines in Fibroblasts
Based on these findings, it is clearly shown that TGFβ1 triggers Activin A production in FOP, and not control, fibroblasts at both 24 and 48 h, whereas after stimulation with BMP4, Activin A production was increased in both FOP and control cells at 24 h and remained high in FOP cells at 48 h. In order to address the possibility of feedback mechanisms, the relative expression of INHBA, TGFBR1, IL6, and IL1b was investigated ( Figure 5). In line with the above, stimulation with TGFβ1 and BMP4 led to significantly increased INHBA expression in FOP cells for TGFβ1 and both cell types for BMP4 (two-way ANOVA, p < 0.05).

Gene Expression of Cytokines in Fibroblasts
Based on these findings, it is clearly shown that TGFβ1 triggers Activin A production in FOP, and not control, fibroblasts at both 24 and 48 h, whereas after stimulation with BMP4, Activin A production was increased in both FOP and control cells at 24 h and remained high in FOP cells at 48 h. In order to address the possibility of feedback mechanisms, the relative expression of INHBA, TGFBR1, IL6, and IL1b was investigated ( Figure  5). In line with the above, stimulation with TGFβ1 and BMP4 led to significantly increased INHBA expression in FOP cells for TGFβ1 and both cell types for BMP4 (two-way ANOVA, p < 0.05).

Discussion
In FOP, inflammation begets ossification when mediators in the inflammatory response set off the osteogenic properties of the mutant ACVR1 receptor. Members of the TGFβ superfamily, such as BMP and Activin A, which directly interact with the FOP ACVR1 receptor, mediate some of these inflammatory responses [12,22]. Our study in a human in vitro model shows that TGFβ1 is a stimulant of Activin A production in FOP. TGFβ1 significantly upregulated INHBA expression and Activin A production in FOP der-

Discussion
In FOP, inflammation begets ossification when mediators in the inflammatory response set off the osteogenic properties of the mutant ACVR1 receptor. Members of the TGFβ superfamily, such as BMP and Activin A, which directly interact with the FOP ACVR1 receptor, mediate some of these inflammatory responses [12,22]. Our study in a human in vitro model shows that TGFβ1 is a stimulant of Activin A production in FOP. TGFβ1 significantly upregulated INHBA expression and Activin A production in FOP dermal fibroblasts. Given that Activin A is considered the major driver of HO in FOP [17], this finding can be of clinical significance. We also found that the FOP fibroblasts produced higher levels of Activin A compared to control cells under basal culture conditions and the initial steps towards osteogenic transdifferentiation.
Given the strong induction of Activin A by TGFβ1, we sought in turn to determine which cytokines may regulate the latter. TGFB1 expression has been demonstrated to be increased in FOP M2 anti-inflammatory macrophages following lipopolysaccharide stimulation; in an unstimulated state, these cells also displayed significantly elevated TGFβ1 production [23]. Treatment of an FOP mouse model expressing a constitutively active mutant form of ACVR1 with an antibody against TGFβ1 was able to significantly decrease HO progression [24]. The same study identified elevated serum TGFβ after trauma induction in a HO mouse model, which coincided with the recruitment of mesenchymal stromal/progenitor cells at the injury site. Furthermore, in mice with no TGFβ1 production in neutrophils and cells of the macrophage/monocyte lineage, injury failed to trigger HO. This is in line with our findings of TGFβ being an important FOP-specific stimulant of Activin A. Interestingly, in mesenchymal stem cells derived from induced pluripotent stem cells (iMSC), treatment with TGFβ1 for 16 h in cells with the ACVR1 mutation did not induce downstream BMP signaling pathways, suggesting that TGFβ1 might not induce Activin expression in all cell types [25]. To summarize, BMP4 appears to be a potent inducer of TGFβ1 in both FOP and control cells, while Activin A production appears more pronounced in FOP cells after TGFβ and BMP4 treatment ( Figure 6).
No induction of Activin A production was observed by cytokines TNFα and IL6 despite previous findings of their upregulation in human-induced FOP macrophages [26] as well as in post-injury lesions and mast cells in the Acvr1 cR206H/+ FOP mouse model [7]. However, the FOP fibroblasts themselves did produce more Activin A than control fibroblasts under basal culture conditions. Given the multipotent plasticity and variability of fibroblasts [27], it can be expected that dermal fibroblasts are able to simulate the nature of the mesenchymal stem progenitors in the FOP lesions. In the early stages of a flare-up, the lesion is known to be infiltrated by B and T lymphocytes and mast cells, likely delivering the Activin A to the site of inflammation [7,[28][29][30][31], which in turn triggers the FAPs to commit to the osteogenic lineage. Indeed, recent single-cell analysis performed in murine FOP-like lesions has revealed how macrophages and fibroblast-like cells are the main sources of local Activin A in damaged muscles, prior to HO [11,32]. Our findings further support the notion that Activin A production may be even further propagated by the mesenchymal stem cells themselves in the HO area.
It is noteworthy that the fibroblasts that showed the highest production of Activin A were derived from patients experiencing a flare-up. Inflammation is known to induce epigenetic changes, which can influence the expression of genes such as INHBA [33]. FOP dermal fibroblasts were shown to react to Activin A based on phosphorylation of SMAD1/5/9 and expression of ID1 and ID2 in combination with higher potential for osteogenic transdifferentiation [19], and no difference between the control and FOP cells were observed in phosphorylated SMAD1/5/9 and BMP target gene expression after BMP stimulation. Similarly, in FAPs, no difference in phosphorylated SMAD1/5/9 was seen between Acvr1 R206H/+ and controls after BMP stimulation. However, canonical SMAD1/5/9 activation does not seem different in FOP monocytes compared to controls [3,23]. Despite this, FOP monocytes and macrophages exhibit a distinct proinflammatory secretome (including TGFβ), suggest-ing that alternative pathways, other than SMAD1/5/9 activation, can mediate FOP cell dysfunction.  Figure 6. Schematic summary of main cytokine interactions with fibroblasts. TGFβ1 and BMP4 stimulation lead to increased Activin A production in FOP fibroblasts. In control fibroblasts Activin A production is increased less following exposure to BMP4 and not increased after TGFβ1 exposure.
In both FOP and control, fibroblasts BMP4 leads to increased TGFβ1 production.

Clinical Characteristics of FOP Patients
Patients experienced their first flare-up during the first 4 years of their life [19]. At the time of the biopsy, patients were questioned and examined to determine if they were currently experiencing a flare-up and if so, which medication was being used. Patients 2, 4, and 5 received medication due to a local flare-up which was administered from at least 1 week prior to biopsy acquisition until at least 1 week thereafter (Table 1). Patients 2 and 4 received nonsteroidal anti-inflammatory drugs (NSAIDs), and although some effect was noted, the flare-up was still active during biopsy acquisition. Patient 5 received pain relief medication, as previously described [19].

Cell Culture
Dermal fibroblasts were isolated and cultured from a 3 mm full-thickness skin biopsy using methods previously described [19]. Fibroblasts were cultured in Ham F10 media (31550-031, Gibco, ThermoFisher, New York, US) supplemented with 10% fetal calf serum Figure 6. Schematic summary of main cytokine interactions with fibroblasts. TGFβ1 and BMP4 stimulation lead to increased Activin A production in FOP fibroblasts. In control fibroblasts Activin A production is increased less following exposure to BMP4 and not increased after TGFβ1 exposure. In both FOP and control, fibroblasts BMP4 leads to increased TGFβ1 production. Latent TGFβ is stored within the bone matrix and can be activated by the resorption of the calcified cartilage during endochondral remodeling [24]. TGFβ exists in three isoforms: TGFβ1, TGFβ2, and TGFβ3. TGFβs are expressed with latency-associated proteins, rendering them often inactive in the extracellular matrix in different tissues. TFGβ1, TFGβ2, and TFGβ3 are key modulators of tissue healing after injury, with one isoform showing overlapping, and also separate effects [34]. TGFβ1, the most prevalent isoform, is often involved in the promotion and differentiation of stem cells. For instance, in bone, TFGβ1 is involved in the proliferation of osteoblastic cells, further inducing bone formation [34]. In an FOP-iPSC model wherein a BMP-specific luciferase reporter construct (BRE-Luc) was transfected, no increase or difference in BMP signaling was found between the different TGFβ isoforms [25]. However, TGFβ3 did enhance chondrogenesis of 3D chondrogenic micro mass formation in both FOP and control cells [25]. Similarly, when investigating the effect of the separate TGFβ isoforms in a C2C12 myoblast line, TGFβ increased proliferation but not in an isoform-dependent manner. It will be intriguing to clarify the role of chondrogenic stimuli by the different TGFβ isoforms. For this new cell, models mirroring the endochondral ossification in FOP will need to be developed. Our model encapsulates the osteogenic transdifferentiation process, but during this process no increased expression of chondrogenic markers was detected, suggesting the endochondral aspect is not well reflected in this model [20].
Given the complex nature of inflammation, it is equally important to determine cross-reaction and synergistic effects with numerous other growth factors in the FOP HO microenvironment and their effect on the HO progenitor cell type(s). It is suggested that flare-ups are characterized by a catabolic stage with tissue destruction and infiltration of immune cells, which is followed by endochondral bone formation [35]. TGFβ can be delivered by immune cells but is thought to be activated during the resorption that takes place in cartilage remodeling. Targeting TGFβ may present a therapeutic window at different stages of HO. There is already experience inhibiting TGFβ as a potential therapy for various diseases. In some forms of cancer, TGFβ signaling contributes to cancer cell proliferation and metastasis. Multiple clinical trials have been performed to inhibit these processes with various TGFβ inhibitors, and although clinical efficacy in these trials has often been disappointing, the treatment itself was generally well tolerated [36]. In the field of bone diseases, Fresolusimab, an antibody neutralizing all TGFβ isoforms, has been tested in a phase 1 trial in Osteogenesis Imperfecta patients, exploring its effects on bone remodeling and bone density. Higher dosages resulted in sustained suppression of bone turnover, and increased bone mineral density and side effects were deemed acceptable [37]. However, long-term data are lacking, and given the pleiotropic signaling of TGFβ, lifetime inhibition of TGFβ may result in toxicity. Systemic suppression of TGFβ may have an impact on wound healing, tissue repair, and inflammatory responses [34]. In a conditional knockout model of TGF-βR2 in renal tubular cells, increased renal inflammation was noted [38]. Though, in soluble TGF-receptor IgGFc chimera transgenic mice, lifetime blocking of TGFβ had no discernable negative side effects [39]. At the time of writing, no specific TGFβ inhibitors have been FDA or EMA approved for clinical use, but given the vast experience with some TGFβ inhibitors, perhaps these could be appropriated for evaluation in a clinical FOP trial after further preclinical evaluation in FOP-specific cell and animal models.
To conclude, the findings of this study revealed TGFβ1 as an upstream FOP-specific inducer of Activin A production in patient-derived fibroblasts; Activin A and TGFβ, in turn, were both upregulated by BMP4 irrespective of mutational ACVR1 status ( Figure 6). Thus, TGFβ may present an important complementary target next to Activin A, blocking of which may (at least partially) prevent the molecular cascade culminating in Activin A-mediated FOP HO. This study was performed in FOP dermal fibroblasts, which can be expected to mirror comparable characteristics of multipotent FOP progenitor cells. We await with excitement future findings about the cell populations in which this mechanism is reflected in FOP in order to provide orientation regarding the development of additional therapeutic modalities for this debilitating disease.

Clinical Characteristics of FOP Patients
Patients experienced their first flare-up during the first 4 years of their life [19]. At the time of the biopsy, patients were questioned and examined to determine if they were currently experiencing a flare-up and if so, which medication was being used. Patients 2, 4, and 5 received medication due to a local flare-up which was administered from at least 1 week prior to biopsy acquisition until at least 1 week thereafter (Table 1). Patients 2 and 4 received nonsteroidal anti-inflammatory drugs (NSAIDs), and although some effect was noted, the flare-up was still active during biopsy acquisition. Patient 5 received pain relief medication, as previously described [19].

DNA Analysis of Fibroblasts
200,000 cells per sample were lysed with a sequencing lysis buffer (QE09050, Biosearch technologies/Lucigen, Hoddeson, UK). Cell lysates were heated at 65 • C for 6 min, followed by heating at 98 • C for 2 min. Extracted DNA was used for PCR amplification by standard methods with a GeneAmp PCR System 9700 (Applied Biosystems, Foster City, USA) and subsequent sequencing analysis.

RNA Isolation, cDNA Synthesis and qPCR
Cell samples (150,000 cells/well) were treated with BMP4, TGFβ, and Activin A, respectively, for 6 h after being serum-starved overnight. Cells were lysed with RNA lysis buffer (R1060-1-50, Zymo Research, Irvine, USA) and RNA was isolated using the Quick RNA kit (R1055, Zymo Research, Irvine, USA) following the manufacturer's instructions. cDNA was synthesized with the SuperScript™ VILO™ cDNA synthesis kit (11754250, Invitrogen™, Waltham, USA). qPCR was performed with 1 pmol/µL forward and reverse primers (IDT) as listed in Table 2, cDNA and SYBR Green master mix (4887352001, Roche) by standard methods with the Lightcycler 480 (Roche, Basel, Switsersland). The relative expression was then calculated with the LightCycler 480 release 1.5.0 SP4 software (Roche, Basel, Switserland) using YWHAZ as a housekeeping gene.

Protein Isolation and Western Blotting Analysis
The fibroblasts were seeded with 150.000 cells/well and treated for 60 min with Activin A or 90 min with BMP4 after serum starvation overnight. Whole-cell lysates were prepared by lysing cells in NuPAGE ® LDS Sample Buffer with 10% NuPAGE ® reducing agent. Proteins were separated in NuPAGE 4-12% BT gels using the XCell SureLock™ system and were subsequently transferred by using the iBlot Dry Blotting system (Invitrogen). Nitrocellulose membranes were blocked in the Odyssey blocking buffer (Westburg). Immunoblotting was performed in Odyssey blocking buffer with 0.1% Triton X-100.

Statistics
Analysis of variance (2-way ANOVA) was used with the factors: genotype and treatments. An adjusted p-value of <0.05 was considered significant for all analyses. All statistical analyses were performed using GraphPad Prism 9 software (Insight Partners, New York, USA). Data are represented as means with error and SD, mean with SD, or summary data.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms24032299/s1, Figure S1: Sanger analysis results of the fibroblasts derived from the FOP patients. Figure S2: Activin A and TGFβ1 production measured by ELISA after TNFα and IL-6 stimulation in FOP and control fibroblasts. Figure   Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.