Inhibition of bone morphogenetic protein 6 receptors ameliorates Sjögren’s syndrome in mice

Primary Sjögren’s syndrome (pSS) is a chronic autoimmune disease, with only palliative treatments available. Recent work has suggested that increased bone morphogenetic protein 6 (BMP6) expression could alter cell signaling in the salivary gland (SG) and result in the associated salivary hypofunction. We examined the prevalence of elevated BMP6 expression in a large cohort of pSS patients and tested the therapeutic efficacy of BMP signaling inhibitors in two pSS animal models. Increased BMP6 expression was found in the SGs of 54% of pSS patients, and this increased expression was correlated with low unstimulated whole saliva flow rate. In mouse models of SS, inhibition of BMP6 signaling reduced phosphorylation of SMAD1/5/8 in the mouse submandibular glands, and led to a recovery of SG function and a decrease in inflammatory markers in the mice. The recovery of SG function after inhibition of BMP6 signaling suggests cellular plasticity within the salivary gland and a possibility for therapeutic intervention that can reverse the loss of function in pSS.

BMP6 signaling is involved in a wide range of biological processes, including iron homeostasis, fat and bone development, and ovulation, as well as stem cell self-renewal and differentiation 11,12 . The central role of BMP signaling has led to the development of specific small-molecule kinase inhibitors. Recent work has identified the BMP signaling inhibitors LDN-212854 and LDN-193189 as highly selective for ALK2 and ALK3 receptors compared with other type I receptors 13 . Some difference in activity exists between these inhibitors with LDN-212854 exhibiting greater selectivity towards ALK2 than ALK3 compared to LDN-193189 [13][14][15] .
In the present study, the ability of these inhibitors to block BMP6 signaling was investigated in two SS mouse models and in HSG cells and evaluated for the ability to block the expression of key proteins involved in SG function and inflammation. Treatment of HSG cells in vitro with LDN-212854 or LDN-193189 resulted in decreased BMP6 signaling and SMAD1/5/8 phosphorylation and led to a recovery of fluid movement across the cell membrane. Daily treatment of BMP6-overexpressing mice with LDN-193189 or C57BL/6.NOD-Aec1Aec2 mice with either LDN-212854 or LDN-193189 restored SG function in mice with established disease. Associated with this functional increase was an increase in expression of AQP5, a protein critical for membrane water permeability in SGs 16 . Treatment with either BMP signaling inhibitor also decreased the infiltration of interferon gamma (IFN-γ) producing CD4 + T cells in the submandibular glands (SMGs) of C57BL/6.NOD-Aec1Aec2 mice. Our findings suggest that the use of small molecule inhibitors of BMP6 signaling is a promising approach for the treatment of pSS.

Methods
Cells. HSG cells were provided by Dr. Indu Ambudkar (National Institute of Dental and Craniofacial Research [NIDCR], National Institutes of Health [NIH]), and cultured in Dulbecco's Modified Eagle Medium (DMEM, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS) in a humidified incubator at 37 °C with 5% CO 2 . HSG cells, which based on short tandem repeat analysis share a common origin with Hela cells, have been used as a model to test regulatory volume decrease (RVD) as a surrogate assay for water movement across a membrane and the molecular mechanisms of secretion from exocrine tissue 4,17,18 . Patient selection criteria. Studies involving healthy subjects were conducted in accordance with approved National Institute of Health (NIH) guidelines. All participants provided informed consent prior to the initiation of any study procedures. Healthy volunteer samples were obtained from NIH Institutional Review Board approved protocols in the Sjögren's Syndrome Clinic at the National Institute of Dental and Craniofacial Research (NIDCR) at the NIH in Bethesda, MD. The protocols utilized in this study are registered at ClinicalTrials.gov (NCT00001390, NCT00001852). In addition, a sequential cohort of seventy-nine deidentified female patients with pSS were selected from the Sjögren's International Collaborative Clinical Alliance (SICCA). All patients fulfilled the 2016 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria for pSS 19 . Their clinical manifestations are summarized in Supplemental Table 1.
Western blot analysis of SMAD signaling. To analyze signaling downstream of BMP6 activation, we studied phosphorylation of SMAD proteins. For the in vitro studies, HSG cells were harvested after treatment and whole-cell lysates were prepared using RIPA Lysis and Extraction Buffer with Halt Protease and Phosphatase Inhibitor Cocktail (Cat# 89900 and 78441, Thermo Fisher Scientific, Waltham, MA, USA). Cells were scraped, sonicated, and then incubated on ice for 20 min. Clarified supernatants were collected from whole-cell lysates that were centrifuged at 14,000 rpm at 4 °C for 20 min. The total protein concentration of the supernatants was measured with the Quick Start Bradford Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA) with bovine serum albumin (BSA) as protein standard. Forty micrograms of protein from each lysate was denatured with 4× Laemmli Sample Buffer and 2-mercaptoethanol (both Bio-Rad Laboratories) at 75 °C for 10 min prior to fractionation on a 4-12% SDS-PAGE gel (Invitrogen, Frederick, MD, USA). The proteins were then transferred to a nitrocellulose membrane using the Trans-Blot Turbo Mini Nitrocellulose Transfer Pack (Bio-Rad Laboratories) at 1.3 A and 25 V for 12 min.
In vivo studies of mice with Sjögren's syndrome-like phenotype. The C57BL/6.NOD-Aec1Aec2 mouse model of Sjögren's syndrome-like phenotype was used to examine exocrine gland dysfunction, lymphocytic infiltration, autoantibody formation, and increased serum proinflammatory cytokines 20,21 . C57BL/6. NOD-Aec1Aec2 mice (male and female) were bred and maintained, as described previously, at the animal facility of the Department of Pathology of the University of Florida 4,22 . To create a second murine model of pSS, a viral vector encoding BMP6 was delivered via retroductal cannulation to the SMGs of 6-8-week-old female C57BL/6J mice (obtained from The Jackson Laboratory; Bar Harbor, ME, USA). BMP6-overexpressing C57BL/6J mice were housed in a pathogen-free animal facility and in compliance with the NIH Guidelines on Use of Animals in Research. All procedures involving animals were performed in an internationally accredited vivarium following institutional guidelines and standard operating procedures. Treatment with ALK2/3 inhibitors, as described herein, was approved by the University of Florida's Institutional Animal Care and Use Committee and Institutional Biosafety Committee.
Regulatory volume decrease measurement. Cell membrane water permeability in the HSG cells was measured using a regulatory volume decrease (RVD) assay, which is a measure of rapid transient osmotic swelling via the movement of water through aquaporin channels in response to addition of hypo-osmotic solutions and hypotonic stress 4 . Briefly, 4 × 10 5 HSG cells per well were cultured with/without 6 ng/mL BMP6. For the indicated cultures, 0.1, 1.0, or 10 nM of LDN-212854 or LDN-193189 was added to the medium, after which the cells were cultured for 4 days. RVD was induced by the addition of 150 mOsm hypotonic solution (HTS), and the change in cell volume was measured before and after HTS stimulation as described previously 3,4,23 . In vivo salivary flow measurements in mice. Pilocarpine-stimulated salivary flow rate (SFR) was determined in C57BL/6.NOD-Aec1Aec2 mice and BMP6-overexpressing mice as described previously 3,22 . Detection of autoantibodies. Serum was tested for the presence of autoantibodies against SSA/Ro or SSB/ La and for antinuclear antibodies (ANA) as described previously 24 . Assessment of salivary gland lymphocytic infiltration. Minor SG focus score (FS) of pSS patients was assessed by the SICCA consortium as described previously 25 . Whole mouse SMGs were surgically removed following euthanasia (see above). Tissue sections were stained with hematoxylin and eosin (H&E) stain and lymphocytic infiltration areas in human and mouse SMG H&E slides were captured and assessed using Aperio digital technology (Leica Biosystems, Buffalo Grove, IL, USA), as described previously 24 . Unstained sections were used for immunofluorescent labeling as described previously 3 .

Immunofluorescent labeling and protein expression quantification. Sections prepared
from formalin-fixed paraffin-embedded minor SG (N = 4) and parotid SG tissues (N = 1) from age-and gender-matched healthy volunteers (HVs) were obtained from the NIDCR Sjögren's Syndrome Clinic or through the cooperative human tissue network. Paraffin-embedded human minor SG slides from N = 79 pSS patients were obtained from the SICCA International Sjögren's Syndrome Biorepository and Data Registry 25 . The level of BMP6 expression in minor SG tissue sections from pSS patients and healthy volunteer (HV) was determined. (Sample size calculations (https://clincalc.com/stats/samplesize.aspx) based on preliminary differences in signal intensity between pSS patient and HV samples with a Student's t test for unequal variances showed that 14 samples would be sufficient to have at least 80% power to find differences with a 5% level of significance). Positive BMP6 expression in SGs was defined as the relative immunofluorescent intensity in the tested slide ≥ average relative immunofluorescent intensity from the HV slide +2SD (142.3 fluorescence units).
To prepare HSG cell slides, cells were immobilized on sterile Nunc Lab-Tek Chamber Slides (Sigma-Aldrich Corp.), and cultured in 1X Minimum Essential Medium (Gibco, Gaithersburg, MD) with 10% FBS containing 1% antibiotics-antimycotic (penicillin, streptomycin, Amphotericin B) (Invitrogen, Frederick, MD) at 37 °C with 5% CO 2 for 48 h. Next, slides were removed from media and washed twice with PBS, followed by fixation in ice cold methanol at −20 °C for 6 min. Chamber slides were washed in PBS five times and before immunofluorescent labeling, with the same procedure of blocking, primary and secondary antibody incubation, and DAPI mounting medium as described above for paraffin-embedded human parotid and minor SG slides.
All images were acquired with the Olympus FluoView 1000 (Center Valley, PA, USA). Analysis and quantification of expression were performed using Volocity (Quorum Technologies, Puslinch, Ontario, Canada). Pixel intensities were analyzed between a range of 1500 and 4095. For human minor SG BMP6 detection, two consecutive sections were labeled with mouse anti-human BMP6 antibody or mouse IgG as control, respectively (see above). For each slide labeled with anti-BMP6 antibody, three to five confocal images were acquired using a 40× objective. The level of BMP6 expression in minor SGs from patients or HVs was determined by averaging pixel intensity from 15-20 representative minor SG epithelial areas using the 40× objective for each slide, with subtraction of corresponding control mean IgG intensity for each tissue section. The difference was used to determine BMP6 expression in minor SGs from patients or HVs. For data acquisition and analysis of all BMPRs, TMEM16A/ANO1 and NKCC1, three to five confocal images were acquired using a 40× objective. Average pixel intensity for each slide was used to determine expression of the corresponding protein. For AQP5 and CFTR in mouse SMGs, Z-stacked images were acquired as follows. Voxel volume (µm 3 ) for each stacked image was used to determine AQP5 expression in the SMG of each animal. For pSMAD1/5/8 and inhibitor of DNA binding protein 3 (ID3), 8 × 8 and 3 × 3 tiled images were acquired for each slide, respectively, and average pixel intensity of each tiled image was used to determine pSMAD1/5/8 or ID3 expression.
All assays were performed according to the instructions provided by the manufacturer. Briefly, mean fluorescent intensities were collected on a Luminex-200 instrument (Bio-Rad Laboratories, Hercules, CA, USA) using Bio-Plex Manager software version 6.2 (Bio-Rad Laboratories, Hercules, CA, USA). Standard curves for each cytokine were generated using the premixed lyophilized standards provided in the kits, and cytokine concentrations were determined from the standard curve using a 5-point regression to transform the median fluorescent intensity values into concentrations. Each sample was run in duplicate, and the average of the duplicates was used as the measured concentration. Any value that was below the detection level was listed as "LOD" (limit of detection) as reported by the Luminex kits. Analyses were performed using Data Pro Manager version 1.02 (Bio-Rad Laboratories, Hercules, CA, USA). Statistical analysis. Unpaired Student's t tests were used to analyze differences in SFR, lymphocytic infiltration, immune cells and protein expression in SMGs between two groups of mice or between minor SGs from pSS patients and HVs. One-way ANOVA followed by Tukey's multiple comparison was used to analyze the RVD

Results
Increased BMP6 expression in minor salivary glands of pSS patients is associated with xerostomia and sialadenitis. To investigate the relationship between BMP6 expression and clinical manifestations of SS, the level of BMP6 was quantified in sections from minor SG biopsies from a large cohort of female pSS patients (N = 79) (Supplemental Table 1). All patients met the ACR/EULAR 2016 criteria for the diagnosis of pSS 19 . A total of 43 (54.4%) patients showed BMP6 expression levels greater than 2X the standard deviation seen in the epithelia of their minor SGs relative to the HVs (fluorescent intensity +/− SEM 242.8 +/−23.2 vs 90 +/−26.6 P < 0.001) (Fig. 1A).
The relationship between minor SG BMP6 expression and two important SG-related manifestations of pSS, SG dysfunction assessed by unstimulated whole saliva (UWS) flow rate and sialadenitis assessed by focus score (FS) or lymphocyte infiltration area was determined. There was a significant negative correlation between minor SG BMP6 expression and UWS flow rate (N = 43, r = −0.33, P < 0.05) (Fig. 1B). A trend of positive correlation www.nature.com/scientificreports www.nature.com/scientificreports/ was found between the minor SG BMP6 expression level and the FS but was not statistically significant (N = 20, r = 0.30, P = 0.20) (Fig. 1C). The lymphocytic infiltration area, another reliable measure of focal infiltration in the SG 3,4,24 , was also positively correlated with the BMP6 expression but not statistically significant (N = 43, r = 0.22, P = 0.15) (Fig. 1D). In addition, the lymphocytic infiltration area was also significantly associated with the FS in a linear regression model (N = 40, r 2 = 0.32 P = 0.0001) (Supplemental Fig. 1). These findings suggest that minor SG BMP6 expression is associated with the loss of SG function (xerostomia) and with increased inflammation (sialadenitis) in pSS patients.

ALK2/3 and BMPR2 are expressed in normal human salivary glands.
To identify the cell types that may be responsive to BMP6 signaling within the SGs, immunofluorescent labeling and confocal imaging were used to detect the expression of BMP type I receptors ALK2 and ALK3 and the type II receptor BMPR2 in histological sections from the parotid gland of a HV. All three receptors were observed on both ductal and acinar cells (Supplemental Fig. 2), suggesting that both types of cells may be responsive to BMP6 signaling.

ALK2/3 inhibition blocks BMP6-induced loss of cell membrane water permeability in vitro.
Previous studies showed that treatment of HSG cells with recombinant BMP6 inhibits AQP5 expression, resulting in decreased cell membrane water permeability, as measured by a reduction in regulatory volume decrease (RVD) 4 . To investigate whether inhibition of BMP6 signaling can block the BMP6-induced loss of RVD, HSG cells were treated with 6 ng/mL of BMP6, with or without LDN-212854 or LDN-193189. BMP6 treatment of HSG cells significantly reduced the RVD (mean ± SEM) compared with PBS-treated control HSG cells (2.93 ± 0.61 vs 92.17 ± 1.83%, respectively; P < 0.01) (Fig. 3). However, addition of LDN-212854 (0.1, 1.0 or 10 nM) inhibited the RVD loss induced by BMP6 in a dose-dependent manner (37.80 ± 5.83, 80.50 ± 6.93 and 90.83 ± 7.17%, respectively vs 2.93 ± 0.61 in BMP6 only-treated cells; all P < 0.05). A dose-dependent increase in the RVD was also measured upon treatment with the same three doses of LDN-193189 (23.40 ± 8.80, 86.50 ± 2.21 and 87.70 ± 8.42%, respectively; all P < 0.05 compared with BMP6 only-treated cells). Moreover, the RVD of HSG cells treated with either LDN-212854 or LDN-193189 at both 1.0 and 10 nM was not significantly different from that of PBS-treated HSG cells (all P > 0.05). These findings suggest that treatment with LDN-212854 or LDN-193189 can block BMP6-induced loss of the RVD in vitro in the HSG cells.
Inhibition of ALK2/3 restores salivary gland function in BMP6-overexpressing mice and C57BL/6.NOD-Aec1Aec2 mice. Based on our finding that BMP signaling inhibitors blocked BMP6-induced RVD loss, we hypothesized that in vivo treatment with these small molecules might restore SG function in animal models of pSS. We tested this possibility in two SS animal models: C57BL/6J mice overexpressing BMP6 specifically in the SMGs 3 , and C57BL/6.NOD-Aec1Aec2 mice 20,21 .

Inhibition of ALK2/3 decreases the number of Th1 cells in the submandibular glands of C57BL/6.
noD-Aec1Aec2 mice. C57BL/6.NOD-Aec1Aec2 mice are similar to pSS patients in that they develop an autoimmune disease with localized inflammation in secretory epithelia especially in salivary and lacrimal glands, and increased production of autoantibodies and proinflammatory cytokines, such as IFN-γ 20,21 . After treatment with ALK2/3 inhibitors, serum, SMGs, and lacrimal glands were collected from C57BL/6.NOD-Aec1Aec2 mice and tested for changes in the salivary gland and systemic immune system. No significant changes in serum levels of antinuclear antibodies or of the SS-associated autoantibodies directed against SSA/Ro and SSB/La were detected in LDN-treated mice compared with the PBS-treated control group (data not shown). No significant changes in the size of the lymphocytic infiltrate were noted in the SMGs and lacrimal glands of LDN-treated C57BL/6.NOD-Aec1Aec2 mice compared with PBS-treated controls (Supplemental Fig. 4). However, there was a greater than two-fold decrease in CD4 + IFN-γ + T cells in the SMGs of LDN-212854-and LDN-193189-treated mice compared with those from PBS-treated control mice, as measured by flow cytometry (N = 6 (3 males and 3 females)/group, P < 0.05) (Table 1, Supplemental Fig. 5). No significant changes in the percentage of CD19 + B cells or other T cell populations were detected (Table 1). These data suggest that although LDN-212854 or LDN-193189 treatment does not change the size of the area of lymphocytic infiltration, ALK2/3 inhibition decreases the relative number of proinflammatory Th1 cells in SMGs.

LDN-212854 and LDN-193189 alter AQP5 protein expression and localization in C57BL/6.
noD-Aec1Aec2 mice. Initiation of fluid secretion from the SG requires movement of water and ions across the membranes of acinar cells. AQP5 is the main water channel for salivary fluid secretion 4,27 . Decreased AQP5 expression has been observed in minor SGs of pSS patients by microarray analysis, and an altered distribution of AQP5 has been described in SS patients and in mice overexpressing BMP6 in their SGs by histological studies 3 . In addition, in pSS patients who respond to treatment with rituximab, AQP5 expression increases 28 . Gene delivery of another water channel, AQP1, to C57BL/6.NOD-Aec1Aec2 mice can rescue the loss of SG function in these mice 4 . These results suggest a central role of AQP-mediated membrane water permeability in SG function.
To determine whether the increase in SG function by ALK2/3 inhibitor treatment is associated with a change in AQP5 expression, SMG sections from LDN-treated and control C57BL/6.NOD-Aec1Aec2 mice were imaged by immunofluorescent labeling and confocal microscopy. A significant increase in AQP5 expression (mean ± SEM) was observed in the SMGs from both LDN-212854-(58.80 ± 3.40 µm 3 ) and LDN-193189-treated mice (45.73 ± 2.63 µm 3 ) compared with the PBS-treated group (30.71 ± 1.67 µm 3 ; both P < 0.0001) (Fig. 6). Confocal imaging analysis of the expression of other proteins that are important for salivary function, including transmembrane member 16A (TMEM16A/ANO1) and Na-K-Cl cotransporter 1 (NKCC1), showed no significant changes in LDN-treated mouse SMGs compared with the control group (Supplemental Fig. 7). LDN-193189 treatment resulted in complete relocalization of the cystic fibrosis transmembrane conductance regulator (CFTR), an apical cell membrane chloride channel, from vesicles to the apical membrane of SMG epithelial cells. Although aggregation of CFTR was still observed following treatment with LDN-212854 (Supplemental Fig. 7) some relocalization was detected, which maybe sufficient for a functional effect. The difference in CFTR response is likely a result of the broader effects of LDN-193189 on both ALK2 and ALK3 compared to LDN-212854. These results support a central role of increased AQP5 levels in salivary gland function in response to ALK2/3 inhibition in models of Sjögren's syndrome.

Discussion
Although increased interferon-alpha signaling is well established in SS 29 , little is known about alterations in other cell signaling pathways. We recently discovered by unbiased microarray mRNA analysis that BMP6, an important cytokine, is up-regulated in the salivary glands of a subset of SS patients and confirmed in mice the association between overexpression of BMP6 and the loss of fluid secretion and increased the inflammatory filtrate within the salivary glands 3 . Here we extended our studies by analyzing BMP6 expression in an independent SS cohort and confirming that increased BMP6 salivary gland protein expression occurs in approximately 54% of the SS subjects compared to case controls. Importantly, the increased expression of BMP6 protein expression seen in SS cases significantly correlated with decreased salivary flow rate. Previous expression of BMP6 in mice showed a statistically significant increase in focus score, but only a trend towards a positive correlation with size of lymphocytic infiltration was detected in patient samples. This may reflect the broader range of expression that is possible to achieve in mice experiments or a threshold of activation effect required for an increase in focus score. In addition, immunostaining confirmed BMP6 responsive receptor expression on both acinar and ductal cell types within human salivary glands laying the groundwork for the rationale targeting of the BMP-6 signaling pathway for the treatment of SS.  Table 1. Flow cytometry analysis of the SMG lymphocytic infiltrated cells from the C57BL/6.NOD-Aec1Aec2 mice (N = 6 mice (3 males and 3 females)/group. *P < 0.05.
Using cell culture and mouse models of SS, we studied whether targeted pharmacological blockade of BMP6 signaling might attenuate autoimmune disease activity. For these experiments, selective ATP competitive small molecule kinase inhibitors of the BMP downstream signaling activity were utilized. Two inhibitors, LDN-212854 and LDN-193189, have IC 50 for BMP responsive ALK receptors in the low nM range when compared to TGF-β/ activin responsive ALK receptors. Treatment of HSG cells with either LDN-212854 or LDN-193189 decreased the BMP6 induced phosphorylation of SMAD1/5/8 and inhibited the BMP6-induced loss of cell membrane water permeability. In vivo treatment with these BMP6 signaling inhibitors in two mouse models of SS restored salivary gland secretion, decreased proinflammatory cells in the gland, decreased cytokines in the serum, and increased levels of TGF-β1, a cytokine often associated with regulatory T cell population. Treatment with LDN-212854 or LDN-193189 also increased AQP5 protein expression and increased salivary secretions in C57BL/6. NOD-Aec1Aec2 mice. Although salivary gland secretion is complex, water channels such as AQP5 play an important role in the final steps of this process. The physiologic response to ALK2/3 inhibitor treatment implies restoration of water movement through inhibition of the SMAD1/5/8 signaling pathway is a fundamental aspect to loss of salivary gland function and supports the possibility of using ALK2/3 inhibition therapeutically.
Sjögren's syndrome is often associated with a strong interferon response and IFN-γ producing Th1 cells play an important role in the initiation and development of SS in patients and animal models 4,22,[30][31][32] . A recent study reported that BMP6 upregulated interferon signaling and downregulated inhibitors of interferon signaling, suggesting a direct interacting between the BMP6 and IFN signaling 30 . The present study supports this and suggests ALK2/3 inhibition can regulate the expression of IFN-γ. Further studies will be necessary to test for BMP6 responsive receptors on IFN-producing immune cells to see if ALK2/3 inhibitors alter their specific cytokine production profile in the context of Sjögren's syndrome. Salivary gland epithelia are reported to be involved in immune regulation as non-professional antigen presenting cells and directly produce pro-inflammatory cytokines such as IFN-γ, IL-17, IL-23 and chemokines (for review, see 31 ). As an important pro-inflammatory T cells, IFN-γ producing Th1 cells play an important role in the initiation and development of the pSS 32 . IFN-γ has been suggested as a key effector point in the inflammation and function in models of pSS 4,22 . In addition,IFN-γ induces the expression of SMAD7, an antagonistic SMAD, which prevents the interaction of SMAD3 with the TGF-β receptor. The results indicate a mechanism of transmodulation between the STAT and SMAD signal-transduction pathways, which has been reported previously 33 . A connection between BMPs and proinflammatory cytokines such as IFN-γ is beginning to emerge 34 . In our study, although the inhibition of ALK2/3 didn't change the infiltrate size or focus score, the observed reduction of IFN-γ producing Th1cells indicated an important decrease in inflammation in the SG and maybe related to the increase in gland secretion.
A key unanswered question remains what is driving the increased mRNA expression of BMP6 in the salivary glands of patients. BMP6 can be produced by many cell types at low levels but its induced overexpression in SS appears to be mainly due to increased transcriptional activity. Additional research will be required to define which cells in the gland are producing the BMP6 and what is the stimulus. Recent work on BMP6 in the liver suggests that in some contexts BMP/SMAD signaling regulates the expression of a broad repertoire of antiviral genes including IRF1, IRF2, IRF7 and IRF9, which in turn regulate many other antiviral genes 30 . Furthermore, it has been shown that BMP6 could enhance an antiviral response to HCV and potently block HCV replication independently of type 1 interferon signaling, suggesting BMP-6 may be interconnected with host antiviral responses. www.nature.com/scientificreports www.nature.com/scientificreports/ Many animal models have been developed for studying Sjögren's syndrome including derivatives of the NOD line, induced models such as Ro or M3R immunized mice, or engineer strains such as STIM1/2 KO, matriptase KO, IL-12 TG, BAFF TG, IL-14α TGPI3K KO, or ID3 KO systemic and T-cell specific models (for review see 35 ). Collectively, these models and interventions that have been tested suggest that many cell types (B, T, mesenchymal stem cells, thyroid, epithelial cells, etc.) can be involved in the development of the disease. Therefore, identifying critical pathways in disease development and then formulating broad pathway specific inhibitors is an important progression in the development of novel interventions for complex autoimmune diseases like Sjögren's syndrome.
Although there are likely alterations in many different signaling pathways in SS, our study focused on BMP6 signaling as one driver of SS inflammation and dysfunction. As pSS is a complex disease, further studies are necessary to identify other pathogenic mechanisms involved in patients with normal levels of BMP6 expression responsible for altered function and inflammation to develop more personalized therapies for this disease. Although our treatment studies were performed in mouse models of SS, our findings suggest the possibility that there might be plasticity within the salivary glands of pSS patients and that treatment altering the BMP6 and other signaling pathways might allow for recovery of the salivary gland function in patients.