Plasma Hsp90 levels in patients with systemic sclerosis and relation to lung and skin involvement: a cross-sectional and longitudinal study

Our previous study demonstrated increased expression of Heat shock protein (Hsp) 90 in the skin of patients with systemic sclerosis (SSc). We aimed to evaluate plasma Hsp90 in SSc and characterize its association with SSc-related features. Ninety-two SSc patients and 92 age-/sex-matched healthy controls were recruited for the cross-sectional analysis. The longitudinal analysis comprised 30 patients with SSc associated interstitial lung disease (ILD) routinely treated with cyclophosphamide. Hsp90 was increased in SSc compared to healthy controls. Hsp90 correlated positively with C-reactive protein and negatively with pulmonary function tests: forced vital capacity and diffusing capacity for carbon monoxide (DLCO). In patients with diffuse cutaneous (dc) SSc, Hsp90 positively correlated with the modified Rodnan skin score. In SSc-ILD patients treated with cyclophosphamide, no differences in Hsp90 were found between baseline and after 1, 6, or 12 months of therapy. However, baseline Hsp90 predicts the 12-month change in DLCO. This study shows that Hsp90 plasma levels are increased in SSc patients compared to age-/sex-matched healthy controls. Elevated Hsp90 in SSc is associated with increased inflammatory activity, worse lung functions, and in dcSSc, with the extent of skin involvement. Baseline plasma Hsp90 predicts the 12-month change in DLCO in SSc-ILD patients treated with cyclophosphamide.

Systemic sclerosis (scleroderma, SSc) is an autoimmune connective tissue disease characterized by vasculopathy and fibrotic changes in multiple organs, particularly the skin and the lungs. It is a rare chronic rheumatic disorder of a complex aetiopathogenesis 1 . Genetic and environmental factors and epigenetics most probably underlie susceptibility to this disease. Scleroderma is traditionally seen as a prototypical multi-system fibrotic disorder mediated by transforming growth factor (TGF)-β and fibroblasts, which produce excessive extracellular matrix proteins such as collagen. However, another hypothesis attributes this disease to the malfunction of the connective tissue repair mechanism in response to injury 1,2 . Patients with SSc can be categorized into two subgroups based on the affected regions: diffuse cutaneous (dc) SSc with proximal involvement, and limited cutaneous (lc) SSc with fibrosis of the skin distal to the elbows or knees, and occasionally of the face and neck 1 . Internal organ complications associated with SSc contribute to its highest morbidity and mortality amongst inflammatory rheumatic disorders 3 . Up to 80% of SSc patients develop interstitial lung disease (ILD), and 25-30% of these patients present with the progressive phenotype. The underlying mechanisms of fibrotic changes in ILD associated with SSc (SSc-ILD) include mesenchymal cell activation, structural changes to epithelial and endothelial cells, and cellular lesions 4 . Currently, ILD is the leading cause of death for SSc patients, accounts for up to 30% of deaths, and has a 10-year mortality of up to 40% 5 . Clinical management of SSc-ILD is challenging and relies on the identification of clinically meaningful disease, assessment of response to pharmacological therapy, and implementation of routine monitoring 3 . Circulating biomarkers reflect relevant physiological or pathological processes, thus may serve as tools for diagnosis, risk evaluation, and prediction of treatment response or safety monitoring. However, disease-specific biomarkers that could be used in routine clinical practice are lacking 6 . Recently, several new promising or effective therapies for SSc-ILD have become available. Therefore, there is an unmet need for disease-or organ-specific biomarkers to assess and to predict treatment response 6,7 . Heat shock proteins (Hsp) are chaperone proteins first described in 1962 in Drosophila based on their increased expression at elevated temperatures 8 . More recent studies found that Hsp are also induced by other forms of cellular stress (e.g. increased pH, disruption of homeostasis, viral infection, reparative healing, and tissue remodeling, etc.) 9,10 . Under physiological conditions as well as during stress response, Hsp recognize and adhere to misfolded proteins in non-native conformations and protect these client proteins against irreversible degradation 11 . In this study, we focused on Hsp90, which is highly conserved and ubiquitously expressed in eukaryotic organisms and in bacteria. It is crucial for many physiological processes, e.g., the cell cycle regulation, cellular apoptosis, protein degradation, and other signalling pathways [12][13][14] . The abundant Hsp90 activates and stabilizes its various client proteins, such as transcription factors, kinases, and ubiquitin ligases 15 . Some client proteins of Hsp90 are involved in the progression of a diverse group of disorders, e.g., malignancies, neurodegenerative diseases, and bacterial or viral infections 16 . Moreover, Hsp90 is also believed to function as a cytokine in the immune response, and participates in innate and adaptive immunity. Nevertheless, the cytokine-like properties of Hsp90 are yet to be elucidated [17][18][19] . Of particular interest is the significance of Hsp90 in stabilizing and folding TGF-β receptors and Src kinases that promote fibrosis in SSc [20][21][22] . Consequently, inhibition of Hsp90 results in proteasomal degradation of TGF-β receptors and Src 23 . We have previously shown that the Hsp90 expression is increased in SSc skin and dermal fibroblasts and is critical for TGF-β signalling. Pharmacological inhibition of Hsp90 effectively blocked the pro-fibrotic effects of TGF-β in cultured fibroblasts and three different preclinical models of SSc 24 . These results have translational implications because several Hsp90 inhibitors have already been used in clinical trials for other indications, mostly tumorous conditions 16,25 .
Therefore, we aimed to assess the systemic levels of Hsp90 in patients with SSc and age-/sex-matched healthy individuals, and analyze its association with SSc-related clinical features with a focus on fibrotic manifestations. www.nature.com/scientificreports/ Our next aim was to evaluate the potential of systemic Hsp90 as a predictor of response to cyclophosphamide treatment in a group of SSc-ILD patients.

Results
Patients and healthy controls: clinical and demographic characteristics. In the cross-sectional analysis, a total of 92 patients and 92 healthy controls were included. Clinical and demographic characteristics of the patients and healthy controls are stated in Table 1.
In the longitudinal analysis, a total of 30 patients with SSc-ILD with active alveolitis without pulmonary arterial hypertension who underwent a routine 6-month (n = 16) or 12-month (n = 14) treatment with i.v. cyclophosphamide were included. Clinical and demographic attributes of SSc-ILD patients treated with cyclophosphamide are presented in Table 2. The lung function tests FVC, and FEV1 in these SSc-ILD patients remained stable for both 6 and 12 months upon treatment with cyclophosphamide; no statistical differences from baseline were found. However, a trend towards improvement in DLCO from baseline was found after 12 months (Table 2). Plasma Hsp90 in patients with systemic sclerosis and healthy controls: a cross-sectional analysis. The plasma levels of Hsp90 were elevated in all SSc patients compared to healthy individuals [12.5 (9.6-17.9) vs 9.8 (7.7-12.4) ng/mL, p < 0.001]. After dividing the patients into two basic SSc subtypes, patients with lcSSc [13.4 (9.6-17.9) ng/mL, p < 0.001], as well as patients with dcSSc [11.4 (9.6-17.3) ng/mL, p = 0.012] showed a significant increase in Hsp90 compared to healthy controls. However, no significant difference between lcSSc and dcSSc was detected (p = 0.311) (Fig. 1).
Plasma levels of Hsp90 in patients with SSc-ILD treated with cyclophosphamide: a longitudinal analysis. In the prospective analysis of patients with SSc-ILD treated with 6 or 12 monthly i.v. pulses of cyclophosphamide we did not observe any significant differences between baseline Hsp90 and after 1, 6, and 12 months (Fig. 3a). However, baseline Hsp90 was able to predict the 12-month absolute change in DLCO% predicted upon cyclophosphamide treatment (DLCO month12-month0 ; r = -0.494, p = 0.037) (Fig. 3b). Moreover, plasma Hsp90 levels at month 1 predicted the deterioration (absolute decrease by ≥ 5%), or stabilization (absolute increase or decrease by < 5%) and improvement (absolute increase by ≥ 5%) of DLCO % predicted at month 12 with respect to baseline (Fig. 3c).

Discussion
In this study, we determined the concentration of plasma Hsp90 in a cross-sectional cohort of patients with SSc compared to healthy controls and analyzed its possible association with disease-specific features. Furthermore, we evaluated the association between plasma Hsp90 and treatment response to cyclophosphamide in a longitudinal cohort of SSc-ILD patients. For the first time, we report a modest increase in plasma levels of Hsp90 in SSc patients compared to healthy controls, which was statistically significant. However, no significant differences between diffuse and limited cutaneous SSc were observed. To date, systemic levels of Hsp90 have been demonstrated to be elevated in some rheumatic conditions (e.g., systemic lupus erythematosus, axSpA and inflammatory myopathies), as well as in type 1 diabetes and several tumorous conditions (e.g., melanoma, liver and lung cancer) [26][27][28][29][30][31][32] .
Our analysis demonstrated a positive correlation between Hsp90 and C-reactive protein (CRP). IL-6 and CRP play an important role in both the pathogenesis and clinical manifestations of SSc and may be useful indicators of disease activity, severity, and poor prognosis 33 . Hsp90 mediates a nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) dependent inflammatory pathway associated with increased expression and secretion of IL-6 18,19 . Thus, the positive correlation between Hsp90 and CRP could reflect the possible interaction between Hsp90 and IL-6, which induces CRP production.
Furthermore, increased Hsp90 was associated with impaired lung function, indicated by a decrease in FVC, FEV1, DLCO, and SpO 2 . All parameters, except for FEV1, remained statistically significantly associated with Hsp90 in a multivariate regression model adjusted for CRP. Recently, Sontake et al. demonstrated increased Hsp90 expression in lung fibrosis and elevated Hsp90 ATPase activity in fibroblasts isolated from fibrotic lung lesions. Moreover, inhibition of Hsp90 exerted potent anti-fibrotic effects both in vitro and in murine models of pulmonary fibrosis 34,35 . In addition, Hsp90 could also contribute to the development of lung fibrosis through IL-6. Interleukin-6 signaling is necessary for collagen production and substantially depends on the proper function of the transcription factor Signal Transducer and Activator of Transcription (STAT)3 [36][37][38][39] . Hsp90 stabilizes and activates STAT3, thus may play a crucial role in the progression of lung fibrosis mediated by IL-6 40 . In our longitudinal analysis of SSc-ILD patients treated with i.v. cyclophosphamide, we did not find any statistically significant changes in the lung function tests or in plasma Hsp90 levels. We observed stabilization of FVC, and a trend towards improvement in DLCO at month 12 from baseline, which are consistent with other observational studies on cyclophosphamide 41 . Nevertheless, baseline plasma Hsp90 was able to predict a 12-month absolute change in DLCO% predicted. In line with this finding, plasma Hsp90 measured after one month of cyclophosphamide treatment predicted the deterioration, stabilization, or improvement of DLCO at month 12. Since there are a few treatment options currently available for SSc-ILD, and novel therapeutic targets are currently under evaluation in ongoing clinical trials, these results might have practical implications in the stratification of SSc-ILD patients according to potential therapeutic response.
In our cross-sectional analysis, we also established a positive correlation between Hsp90 and the extent and severity of skin involvement. However, this was only found in patients with dcSSc, possibly due to more extensive and severe skin involvement compared to lcSSc. This is in line with our previous findings demonstrating www.nature.com/scientificreports/ Table 3. Univariate and multivariate regression analysis predicting plasma Hsp90 levels based on lung function parameters, unadjusted and adjusted for CRP levels. FVC forced vital capacity, FEV1 forced expiratory volume in one second, DLCO diffusing capacity for carbon monoxide, SpO 2 oxygen saturation assessed by pulse oximetry, CRP C-reactive protein, b y intercept of a line, CI confidence interval. Statistically significant differences (p < 0.05) are marked in bold. www.nature.com/scientificreports/ increased local expression of Hsp90 in the involved skin and cultured fibroblasts of patients with dcSSc, as well as in the fibrotic skin of three preclinical models of SSc. In that study, inhibition of Hsp90 indicated potent anti-fibrotic properties in three murine models of dermal fibrosis, which mimic different clinical stages of SSc 24 . Moreover, a recent study showed that Hsp90 interacts very closely with fibronectin and may regulate matrix assembly, which is upregulated in SSc 2,42 . Our study is limited by small numbers, especially in the longitudinal analysis, and the lack of other longitudinal SSc-ILD cohorts treated with other currently available pharmacological agents. Therefore, further studies are needed to confirm and validate our data. However, based on our results, extracellular Hsp90 could become an attractive soluble biomarker of skin and lung involvement in SSc with the potential to predict treatment response in SSc-ILD patients.

Conclusion
Our study demonstrates that Hsp90 is elevated in the peripheral blood of patients with systemic sclerosis compared to healthy individuals. Plasma Hsp90 levels in SSc patients rise with decreased pulmonary function and increased systemic inflammation. In diffuse cutaneous SSc, Hsp90 is associated with more extensive skin involvement. Of particular interest, in SSc-ILD patients, baseline Hsp90 predicts the 12-month change in DLCO upon cyclophosphamide treatment. These data further support the important role of Hsp90 as a regulator of fibroblast activation and tissue fibrosis in SSc and might have a direct practical application in the real-life stratification of patients with SSc-ILD according to potential therapeutic response.

Methods
Patients and healthy controls. All patients involved in this study fulfilled the ACR/EULAR classification criteria for SSc 43 . Plasma samples from all patients and healthy individuals were prepared from the whole blood collected into commercially available EDTA-treated tubes. For the cross-sectional analysis, plasma samples were obtained from 92 Caucasian patients with systemic sclerosis (SSc) and 92 age-and sex-matched Caucasian healthy controls. For the longitudinal analysis, plasma samples were obtained prospectively at baseline, and 1, 6, and 12 months thereafter from 30 Caucasian patients with SSc-ILD with active alveolitis without pulmonary arterial hypertension who underwent a routine 6-month (n = 16) or 12-month (n = 14) treatment with i.v. cyclophosphamide (CPA, 500 mg/m 2 monthly). Active alveolitis was defined as the presence of areas of ground-glass attenuation on high-resolution computed tomography (HRCT) and reduced levels of diffusing lung capacity for carbon monoxide (DLCO) and/or forced vital capacity (FVC), as described elsewhere 44 . Other SSc-related clinical features were assessed according to generally accepted definitions and recorded, such as the presence of pulmonary arterial hypertension, renal, cardiac and gastrointestinal involvement, Raynaud's phenomenon, and digital ulcers 45 . Skin involvement was evaluated using the modified Rodnan skin score (mRSS) 46 . Disease activity was determined by the European Scleroderma Study Group (ESSG) SSc activity score 47 . Pulmonary function tests (PFT) were routinely performed using standard methods, in accordance with the ATS recommendations 48 . The DLCO was measured by a single-breath method using a gas mixture of 0.2% CO and 8% helium, with correction for hemoglobin. Peripheral oxygen saturation (SpO 2 ) was measured by a handheld pulse oximeter (CR-100, Noramedica, Czech Republic). In the longitudinal analysis PFTs were performed at baseline, and 6 and 12 months thereafter, and the results are expressed as a percentage of the normal predicted values based on the patient's sex, age, and height. The research was confirmed by the local ethics committee at the Institute of Rheumatology in Prague, and each patient signed an informed consent form. All methods were performed in accordance with the relevant guidelines and regulations. Laboratory measurements. All samples were stored at − 80 °C until use. Serum levels of C-reactive protein (CRP) were determined by an immuno-turbidimetric technique using an Olympus AU 400 biochemical analyzer (Olympus Optical, Tokyo, Japan), and erythrocyte sedimentation rate (ESR) was measured according to the Fahreus and Westergren method. ANAs were detected using indirect immunofluorescence on HEP2 cells, and the autoantibodies of the ENA complex (anti-U1RNP, anti-Ro, anti-La, anti-DNA-topoisomerase I, anti-Jo-1, anti-P protein, anti-Sm, and anti-centromere) were assayed by immunoblot. Plasma levels of Hsp90 were assessed by a high-sensitivity ELISA kit (eBioscience, Vienna, Austria) according to the manufacturer's protocol. The assay recognizes human Hsp90 alpha. The calculated sensitivity is 0.03 ng/mL. The absorbance value was established at 450 nm by an ELISA reader (SUNRISE; Tecan, Grödig, Austria). Statistical analysis. Basic descriptive statistics (mean, median, standard deviation, interquartile range, skewness, and kurtosis) were computed for all variables, which were subsequently tested for normality using the Kolmogorov-Smirnov and Shapiro-Wilk tests. Differences in interval variables (e.g. Hsp90, age, etc.) were tested using the Mann-Whitney U test, while the chi-square test was used to compare frequency counts of categorical variables (e.g. gender). The bivariate relationships between variables under study were assessed using the Spearman correlation coefficient. Linear regression analysis was used to predict patients' Hsp90 levels by a set of predictors (FEV1, FVC, DLCO, SpO 2 ) while adjusting for a confounder (CRP). Friedman's test was used to analyze repeated longitudinal measurements taken: (a) at baseline, (b) after 1 month, (c) after 6 months, and (d) after 12 months of therapy with cyclophosphamide. Data are presented as median (IQR) unless stated otherwise. Statistical significance was set at p < 0.05. All analyses were conducted using SPSS version 25 (SPSS, Inc., Chicago, IL, USA). Graphs were created using GraphPad Prism 5 (version 5.02; GraphPad Software, La Jolla, CA, USA).