Soluble CD90 as a potential marker of pulmonary involvement in systemic sclerosis

Authors


Abstract

Objective

Vascular injury and endothelial cell (EC) activation are pathogenic hallmarks of systemic sclerosis (SSc; scleroderma). Human CD90 is highly expressed on activated ECs and can be shed from the cell surface. This study was conducted to examine whether soluble CD90 (sCD90) is elevated in the sera of patients with SSc and linked to pulmonary involvement and in particular, pulmonary arterial hypertension (PAH).

Methods

sCD90 serum concentrations were assessed in 76 patients with SSc and related to clinical data, lung function, 6-minute walk distance, echocardiography, bronchoalveolar lavage fluid, and laboratory parameters. Thirty-one healthy volunteers and 29 patients with idiopathic retroperitoneal fibrosis (IRF) served as controls.

Results

sCD90 serum concentrations were elevated in patients with SSc compared to healthy volunteers (P = 0.001) and patients with IRF (P = 0.01). SSc patients with pulmonary fibrosis (P = 0.006) and patients with PAH (P < 0.001) had increased sCD90 serum concentrations compared to patients without the respective pulmonary manifestation of SSc. sCD90 levels correlated with diffusing capacity for carbon monoxide (n = 65; r = −0.348, P = 0.005) and systolic pulmonary artery pressure (n = 53; r = 0.469, P < 0.001). Receiver operating characteristic curve testing determined an optimal cutoff value of ≥626 ng/ml with a sensitivity of 68% and a specificity of 83% for PAH (area under the curve 0.773, 95% confidence interval 0.648–0.898; P < 0.001).

Conclusion

sCD90 concentrations were increased in the sera of SSc patients, particularly in patients with vascular involvement of the lungs. These data suggest that sCD90 should be further evaluated as a marker for diagnosis of PAH in SSc.

INTRODUCTION

Systemic sclerosis (SSc; scleroderma) is a multisystem fibrotic disease of unknown etiology. The mortality is high and mostly due to pulmonary involvement (1). Vascular manifestations in terms of digital ulcers are recurrent and considerably disabling (2). Pathogenetic concepts are dealing with endothelial dysfunction and activation, fibroblast abnormalities, and alternatively activated macrophages (3–6). There are several efforts in establishing markers of pulmonary involvement in SSc, since mortality is mainly determined by pulmonary arterial hypertension (PAH) and pulmonary fibrosis.

Human CD90 is a glycosylphosphatidylinositol-anchored adhesion molecule (7, 8) that is expressed on activated endothelial cells (ECs), fibroblasts, neurons, and a subpopulation of hematopoietic stem cells (9–11). Soluble adhesion molecules such as soluble vascular cell adhesion molecule 1 (sVCAM-1), sE-selectin, and soluble intercellular adhesion molecule 1 (sICAM-1) are known to be elevated in the sera of patients with SSc and are related to active disease (12, 13). Shedding of CD90 from the cell surface of fibroblasts is associated with differentiation toward a profibrogenic myofibroblast phenotype (13, 14). We have previously described an enzyme-linked immunosorbent assay (ELISA) for detection of soluble CD90 (sCD90) (15). Up to now, it is not known if sCD90 serum concentrations are elevated in patients with SSc and whether there is an association with vascular and parenchymal involvement of the lungs.

Therefore, we measured sCD90 serum concentrations in a large cohort of SSc patients. We assessed anthropometric, clinical, laboratory, and functional data in correlation with sCD90 levels in these patients. As controls, sCD90 serum concentrations were measured in healthy volunteers and patients with idiopathic retroperitoneal fibrosis (IRF), a fibroinflammatory disease that is not affecting the lungs and the peripheral vasculature (16).

Significance & Innovations

  • Serum levels of soluble CD90 (sCD90) were elevated in a large cohort of patients with systemic sclerosis (SSc) in comparison to healthy volunteers and patients with idiopathic retroperitoneal fibrosis.

  • In particular, patients with pulmonary arterial hypertension and pulmonary fibrosis revealed highly increased concentrations of sCD90. Moreover, sCD90 levels correlate with diffusing capacity for carbon monoxide and systolic pulmonary artery pressure in echocardiography.

  • These results suggest a potential value of sCD90 as a marker for the presence of pulmonary arterial hypertension in SSc.

SUBJECTS AND METHODS

Patients and healthy volunteers.

Patients were recruited at the Department of Rheumatology and Clinical Immunology, University Medical Center Freiburg, Freiburg, Germany, and the Department of Rheumatology, University Hospital Basel, Basel, Switzerland. SSc patients fulfilled the American College of Rheumatology criteria for SSc (17). Patients with clinical signs of acute infection were excluded. The diagnosis in patients with IRF was based on imaging and clinical findings. Patients with IRF due to known secondary causes were excluded (18). Next to anthropometric and clinical data, bronchoalveolar lavage (BAL) fluid differentials and laboratory parameters were obtained. Skin score was measured by the modified Rodnan skin thickness score (MRSS) (19, 20). Gas exchange at rest was assessed by the determination of hemoglobin-corrected single-breath diffusing capacity for carbon monoxide (DLCO) (21). Spirometry, body plethysmography, and 6-minute walk distance (6MWD) were measured according to established standards (22, 23). PAH (systolic pulmonary artery pressure [PAP] ≥40 mm Hg on echocardiography) and pulmonary fibrosis (fibrotic changes on chest radiograph) were defined as proposed by the European League Against Rheumatism Scleroderma Trial and Research Group (24). We analyzed the sera of healthy volunteers and patients with IRF as controls. The study was carried out in accordance with the Declaration of Helsinki and was approved by our local ethics committee. Informed consent was obtained from all subjects.

ELISA.

Microtiter plates were coated with 0.25 μg/well of anti–Thy-1 antibody in 0.1 M NaHPO4/0.1 M NaH2PO4, pH 9.0 (clone AS02; Dianova), overnight at 4°C. Plates were washed 3 times with phosphate buffered saline (PBS) and blocked with PBS/10% fetal calf serum (FCS) for 1 hour at 4°C. After several washes with PBS, samples and standard (recombinant human Thy-1 expressed in Chinese hamster ovary cells) were diluted in PBS/10% FCS and incubated overnight at 4°C. The plates were washed 5 times with PBS. The biotinylated anti-CD90 monoclonal antibody (clone 5E10; Pharmingen) was added for 90 minutes at 37°C. After 3 washes with PBS, avidin-conjugated peroxidase (eBioscience) diluted at 1:10,000 in assay diluent (eBioscience) was added for 1 hour at room temperature. Plates were rinsed 5 times with PBS. Subsequently, tetramethylbenzidine was used to generate the color reaction that was measured at 405 nm. The serum concentrations of sVCAM-1 and sE-selectin were quantified by a commercially available ELISA kit (eBioscience), according to the manufacturer's instructions.

Statistical analysis.

All values are shown as medians and interquartile ranges (IQRs). Subgroups were compared using the nonparametric Mann-Whitney U test and Fisher's exact test. For correlations, Spearman's rank correlation coefficient was calculated. Multivariate logistic regression analysis was conducted to examine the influence of age, sex, and sCD90 serum levels on PAH. Receiver operating characteristic (ROC) curves were calculated to analyze the value of sCD90 for diagnosing PAH and pulmonary fibrosis in SSc. P values less than 0.05 were considered significant. SPSS, version 17 (IBM), and GraphPad Prism 5.0 were used for database management and statistical analysis.

RESULTS

Study subjects.

A total of 76 sera samples of patients with SSc (58 women, 18 men) referred between 2000 and 2011 were analyzed. Since December 2010, data entry was performed prospectively. The clinical characteristics of the study population are shown in Table 1. The sera of 31 healthy volunteers (19 women, 12 men) and 29 patients with IRF (16 women, 13 men) served as controls. SSc patients did not differ from healthy controls with respect to age (P = 0.17) and sex (P = 0.16). There were more male patients in the cohort of patients with IRF (P = 0.004) in comparison to patients with SSc, whereas differences of age were not observed (P = 0.76).

Table 1. Clinical characteristics of patients with systemic sclerosis*
 Value
  • *

    Values are the median (interquartile range) unless otherwise indicated. MRSS = modified Rodnan skin thickness score; PAH = pulmonary arterial hypertension; PAP = pulmonary artery pressure; NT-proBNP = N-terminal pro–brain natriuretic peptide; TLC = total lung capacity; FVC = forced vital capacity; DLCO = hemoglobin-corrected single-breath diffusing capacity for carbon monoxide; 6MWD = 6-minute walk distance; BAL = bronchoalveolar lavage.

  • By echocardiography.

Subjects, no.76
Disease duration, months66 (107)
Active digital ulcers, no. yes/no8/68
MRSS6 (15)
Pulmonary fibrosis, no. yes/no31/45
PAH, no. yes/no19/57
Systolic PAP, mm Hg28 (13)
NT-proBNP, pg/ml103 (321)
TLC, % predicted90 (27)
FVC, % predicted89 (34)
DLCO, % predicted65 (31)
6MWD, meters433 (107)
BAL fluid macrophages, %79.0 (27.5)
BAL fluid neutrophils, %4.5 (9.0)
BAL fluid eosinophils, %1.5 (4.3)
BAL fluid lymphocytes, %7.0 (15.7)

Elevation of sCD90 serum concentrations in SSc.

Median serum sCD90 concentrations were elevated in patients with SSc (478 ng/ml, IQR 266–662) compared to healthy controls (258 ng/ml, IQR 168–445; P = 0.001) and patients with IRF (310 ng/ml, IQR 187–441; P = 0.01). There was no difference between sCD90 levels of healthy volunteers and IRF patients (P = 0.53) (Figure 1). The median disease duration of patients with SSc was 66 months (IQR 14–122 months). There was no correlation between serum concentrations of sCD90 and disease duration (n = 76; r = −0.026, P = 0.836). There were 29 patients (38.2%) with diffuse cutaneous SSc (dcSSc), 44 patients (57.9%) with limited cutaneous SSc (lcSSc), and 3 patients (3.9%) with SSc sine scleroderma. The medium serum concentration of sCD90 did not differ between patients with lcSSc and dcSSc (398 ng/ml, IQR 244–652 versus 539 ng/ml, IQR 319–780; P = 0.201). Six SSc patients (7.9%) were receiving immunosuppressive treatment regimens (azathioprine, n = 2; cyclophosphamide, n = 1; mycophenolic acid, n = 3) for fibrotic disease, whereas 53 patients (69.7%) were not. There were no differences between sCD90 serum concentrations between these subgroups (P = 0.68). Four SSc patients were receiving endothelin receptor antagonists to treat PAH. There was no difference in sCD90 levels between these patients and PAH patients without treatment (P = 0.26). Twenty-one (72.4%) of the total subgroup of IRF patients received immunosuppressive treatment (methotrexate, n = 2; azathioprine, n = 9; cyclophosphamide, n = 7; mycophenolic acid, n = 2; cyclosporin, n = 1). There were no differences between sCD90 serum levels in these subgroups (P = 0.46).

Figure 1.

Soluble CD90 (sCD90) serum concentrations of patients with systemic sclerosis (SSc) compared to healthy control subjects (HC) and patients with idiopathic retroperitoneal fibrosis (IRF). ** = P < 0.005; * = P < 0.05.

sCD90 serum concentrations and clinical/laboratory data.

SSc patients with pulmonary fibrosis (n = 31) revealed higher median serum concentrations of sCD90 than SSc patients without pulmonary fibrosis (576 ng/ml, IQR 354–780 versus 335 ng/ml, IQR 232–615; P = 0.006). There was a difference of median sCD90 serum concentrations between patients with PAH (n = 19) and those without PAH (694 ng/ml, IQR 388–882 versus 350 ng/ml, IQR 238–578; P < 0.001) (Figure 2A). Patients with PAH and pulmonary fibrosis (n = 9) showed increased median sCD90 levels compared to patients with neither PAH nor pulmonary fibrosis (787 ng/ml, IQR 601–1,000 versus 317 ng/ml, IQR 222–528; P < 0.001). Median sCD90 levels in patients with isolated PAH (n = 10) were elevated in comparison to patients without PAH and pulmonary fibrosis (647 ng/ml, IQR 328–792 versus 317 ng/ml, IQR 222–528; P = 0.02). Patients with isolated pulmonary fibrosis (n = 22) revealed increased median sCD90 levels compared to patients without PAH and pulmonary fibrosis (502 ng/ml, IQR 329–660 versus 317 ng/ml, IQR 222–528; P = 0.03) (Figure 2B). Between SSc patients with and those without pulmonary fibrosis, there was no difference regarding age (P = 0.19), whereas patients with PAH were older (P < 0.001). Multivariate logistic regression analysis of age, sex, sCD90, and pulmonary fibrosis revealed that age and sCD90 serum concentrations were associated with PAH, whereas sex and pulmonary fibrosis were not (Table 2). Moreover, systolic PAP correlated with sCD90 serum concentrations (n = 53; r = 0.469, P < 0.001). There was no correlation between serum concentrations of N-terminal pro–brain natriuretic peptide (NT-proBNP) and sCD90 (n = 27; r = 0.338, P = 0.08). Forced vital capacity (n = 63; r = −0.145, P = 0.26), total lung capacity (n = 66; r = −0.229, P = 0.06), and 6MWD (n = 22; r = 0.189, P = 0.40) did not correlate with sCD90 levels, whereas there was a correlation between DLCO and sCD90 (n = 65; r = −0.348, P = 0.005).

Figure 2.

Soluble CD90 (sCD90) serum concentrations of patients with A, systemic sclerosis and B, pulmonary involvement. ** = P < 0.005; * = P < 0.05; PH = pulmonary arterial hypertension; PF = pulmonary fibrosis.

Table 2. Multivariate logistic regression analysis for pulmonary arterial hypertension*
 OR (95% CI)P
  • *

    OR = odds ratio; 95% CI = 95% confidence interval.

Age1.111 (1.036–1.191)0.003
Sex0.331 (0.075–1.460)0.144
Soluble CD901.004 (1.001–1.007)0.017
Pulmonary fibrosis0.569 (0.137–2.363)0.438

With regard to the cellularity of the BAL fluid, there was no correlation between sCD90 levels and the percentage of lymphocytes (n = 18; r = 0.082, P = 0.747), macrophages (r = 0.095, P = 0.708), neutrophils (r = 0.042, P = 0.867), and eosinophils (r = 0.045, P = 0.858). Active digital ulcers (n = 8) were also not associated with higher median sCD90 serum levels (491 ng/ml, IQR 306–651 versus 398 ng/ml, IQR 266–662; P = 0.99). We found a correlation between sCD90 and skin fibrosis assessed using the MRSS (n = 66; r = 0.249, P = 0.04).

sCD90 and pulmonary involvement.

To address the diagnostic value of sCD90 for the occurrence of PAH and pulmonary fibrosis in SSc, ROC curves were calculated. For pulmonary fibrosis in SSc, sCD90 concentrations ≥439 ng/ml showed a sensitivity of 68% and a specificity of 60% (area under the curve 0.686, 95% confidence interval [95% CI] 0.565–0.806; P = 0.006). A cutoff value of ≥626 ng/ml revealed a sensitivity of 68% and a specificity of 83% for PAH (area under the curve 0.773, 95% CI 0.648–0.898; P < 0.001) (Figure 3). Interestingly, all patients with sCD90 serum concentrations ≥932 ng/ml revealed PAH (n = 5). In comparison to sCD90, an NT-proBNP value ≥407 pg/ml (n = 27) showed a sensitivity of 80% and a specificity of 91% (area under the curve 0.873, 95% CI 0.713–1.000; P = 0.01), and a DLCO ≤54% revealed a sensitivity of 53% and a specificity of 26% for PAH (area under the curve 0.291, 95% CI 0.151–0.431; P = 0.02).

Figure 3.

Receiver operating characteristic curves. A, A cutoff value for soluble CD90 (sCD90) serum concentrations ≥439 ng/ml had a sensitivity of 68% and a specificity of 60% for pulmonary fibrosis in systemic sclerosis (area under the curve 0.686, 95% confidence interval [95% CI] 0.565–0.806; P = 0.006). B, For pulmonary arterial hypertension, sCD90 serum concentrations ≥626 ng/ml revealed a sensitivity of 68% and a specificity of 83% (area under the curve 0.773, 95% CI 0.648–0.898; P < 0.001).

sVCAM-1 and sE-selectin serum concentrations in SSc patients.

There were no differences in median sVCAM-1 (723 ng/ml, IQR 570–987 versus 828 ng/ml, IQR 573–1,086; P = 0.742) and sE-selectin serum concentrations (10.0 ng/ml, IQR 7.6–16.0 versus 11.3 ng/ml, IQR 5.6–17.9; P = 0.985) in SSc patients with pulmonary fibrosis (n = 12 versus n = 16) compared to patients without pulmonary fibrosis (n = 35 versus n = 39) (Figure 4). Moreover, in SSc patients with PAH (n = 8 versus n = 10), there were no differences in median serum concentrations of sVCAM-1 (707 ng/ml, IQR 482–1,007 versus 828 ng/ml, IQR 607–1,086; P = 0.506) and sE-selectin (11.9 ng/ml, IQR 9.5–24.8 versus 10.0 ng/ml, IQR 6.1–15.0; P = 0.214) in comparison to SSc patients without PAH (n = 39 versus n = 45). There was a correlation between sVCAM-1 and sCD90 (n = 47; r = 0.301, P = 0.040), whereas sE-selectin neither correlated with sCD90 (n = 55; r = 0.146, P = 0.289) nor correlated with sVCAM-1 (n = 47; r = −0.057, P = 0.705).

Figure 4.

Serum concentrations of soluble vascular cell adhesion molecule 1 (sVCAM-1) and sE-selectin of patients with systemic sclerosis and pulmonary involvement. PH = pulmonary arterial hypertension; PF = pulmonary fibrosis.

DISCUSSION

The present study shows that sCD90 serum concentrations are increased in patients with SSc compared to healthy control subjects and patients with IRF. Moreover, sCD90 levels are elevated in SSc patients with pulmonary involvement. Patients with pulmonary fibrosis and in particular PAH revealed highly increased sCD90 serum concentrations. Parenchymal and vascular pulmonary involvement accounts for the majority of deaths in SSc (1). The pulmonary vasculature in patients with SSc is affected by remodeling processes, which are preceded by EC activation and dysfunction (3). Anti-EC antibodies may contribute to this process (25). Anti-EC antibodies are elevated in the sera of patients with vascular manifestations of SSc such as digital ulcers or PAH, but also in idiopathic PAH (25, 26). In the development of PAH, activated ECs proliferate and lead to vessel obstruction (3, 27). Activated ECs express several adhesion molecules (28). CD90 is a member of the immunoglobulin superfamily (7) and as such, like other adhesion molecules, is involved in cell–cell adhesion and the recruitment of leukocytes via Mac-1 (CD11b/CD18) (29). It has been shown that CD90 is expressed by activated ECs (9, 15, 30). In mice, CD90 is expressed by T lymphocytes, which can shed the molecule spontaneously from their cell surface (31). Also, other adhesion molecules can be released from the cell surface by enzymatic cleavage (32). The soluble adhesion molecules sVCAM-1, sE-selectin, and sICAM-1 are elevated in the sera of patients with SSc (12, 13). There are some small studies regarding the association of these soluble adhesion molecules with disease severity and PAH in SSc: Denton and colleagues showed that changes of sVCAM-1 and sE-selectin concentrations paralleled decline of lung function and improvement of skin score and renal function in 6 of 12 patients with SSc (12). A study of 19 patients with SSc revealed that sVCAM-1 correlates with abnormal left ventricular filling patterns in echocardiography and the degree of dyspnea. However, only 2 patients in this cohort had signs of PAH in echocardiography, and serum concentrations of sVCAM-1 and sICAM-1 in these 2 patients were not discussed (13). Moreover, in a study of 31 SSc patients, it has been shown that sVCAM-1 and sE-selectin are related to internal organ involvement of SSc (16 patients with internal organ involvement: 2 patients with PAH, 13 patients with pulmonary fibrosis, and 1 patient with cardiac involvement) (33). Iannone et al showed that sICAM-1, sP-selectin, soluble platelet EC adhesion molecule 1, and sVCAM-1 were elevated in 10 patients with SSc and PAH in comparison to healthy controls (34). Our study shows for the first time that sCD90 serum concentrations are also elevated in patients with SSc and are related to pulmonary involvement. In particular, patients with PAH revealed high serum levels of this adhesion molecule. Endothelin-1 expression is increased in patients with PAH and is an important therapeutic target in this disease (35). Endothelin-1 is linked to EC activation in patients with PAH and SSc and is associated with serum levels of soluble adhesion molecules. Therefore, it has been shown that sVCAM-1 and sICAM-1 serum levels were elevated in SSc patients with PAH and decreased after initiation of treatment with endothelin receptor blockers (34). In this study, we could not find an association between serum levels of sVCAM-1 or sE-selectin with the presence of PAH or pulmonary fibrosis in SSc. In contrast, sCD90 serum levels were associated with PAH and also correlated with systolic PAP and DLCO, both known markers of PAH (24). Although NT-proBNP was superior to sCD90 in predicting the presence of PAH, sCD90 might differentiate PAH from pulmonary hypertension due to disease of the left side of the heart. However, the present study is limited by the fact that it is based on echocardiography only. Further studies focused on right-sided heart catheterization, and hemodynamic parameters are needed. Since VCAM-1 and E-selectin expressions are also elevated in EC activation (28), and only a weak correlation of sVCAM-1 and sCD90 was observed, another potential source of sCD90 in SSc might be fibroblasts. In fibroblast foci of patients with idiopathic pulmonary fibrosis, only CD90 fibroblasts/myofibroblasts could be detected, whereas in healthy controls, CD90+ fibroblasts predominate (14, 36). In vitro stimulation of CD90+ fibroblasts with profibrotic cytokines such as tumor necrosis factor α and interleukin-1 leads to a shedding of CD90 from the cell surface and a differentiation toward a highly active myofibroblast phenotype (14, 37). Presumably, CD90 is shed from the cell surface in the course of such a differentiation process of fibroblasts, therefore causing increased serum concentrations of sCD90 in SSc patients.

In consideration of the high sCD90 levels in patients with PAH and SSc, the known role of sCD90 in endothelial activation, and the spatial proximity between ECs and the blood stream, we primarily suspect an endothelial source of sCD90. Nevertheless, sCD90 might also be, at least partially, released by fibroblasts in SSc.

In conclusion, sCD90 serum concentrations were elevated in patients with SSc. Pulmonary fibrosis and specifically PAH were associated with elevated sCD90 levels. The exact source of sCD90 has to be elucidated in further investigations. Moreover, the applicability of this soluble adhesion molecule for diagnostic and monitoring purposes of PAH in SSc versus idiopathic PAH and idiopathic pulmonary fibrosis and PAH and pulmonary fibrosis associated with other rheumatic diseases must be determined in further validation cohorts and prospective clinical trials.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kollert had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Kollert, Budweiser, Binder, Zissel, Prasse.

Acquisition of data. Kollert, Christoph, Probst, Budweiser, Bannert, Binder, Walker, Prasse, Saalbach.

Analysis and interpretation of data. Kollert, Christoph, Budweiser, Binder, Sehnert, Voll, Warnatz, Walker, Prasse, Saalbach.

Acknowledgements

The authors thank Birgit Nettlenbusch, Kathleen Fischer, and Danny Gutknecht for the excellent technical assistance.

Ancillary