Serum Amyloid A Is a Marker for Pulmonary Involvement in Systemic Sclerosis

Inflammation in systemic sclerosis (SSc) is a prominent, but incompletely characterized feature in early stages of the disease. The goal of these studies was to determine the circulating levels, clinical correlates and biological effects of the acute phase protein serum amyloid A (SAA), a marker of inflammation, in patients with SSc. Circulating levels of SAA were determined by multiplex assays in serum from 129 SSc patients and 98 healthy controls. Correlations between SAA levels and clinical and laboratory features of disease were analyzed. The effects of SAA on human pulmonary fibroblasts were studied ex vivo. Elevated levels of SAA were found in 25% of SSc patients, with the highest levels in those with early-stage disease and diffuse cutaneous involvement. Significant negative correlations of SAA were found with forced vital capacity and diffusion capacity for carbon monoxide. Patients with elevated SAA had greater dyspnea and more frequent interstitial lung disease, and had worse scores on patient-reported outcome measures. Incubation with recombinant SAA induced dose-dependent stimulation of IL-6 and IL-8 in normal lung fibroblasts in culture. Serum levels of the inflammatory marker SAA are elevated in patients with early diffuse cutaneous SSc, and correlate with pulmonary involvement. In lung fibroblasts, SAA acts as a direct stimulus for increased cytokine production. These findings suggest that systemic inflammation in SSc may be linked to lung involvement and SAA could serve as a potential biomarker for this complication.


Introduction
Systemic sclerosis (SSc) is a chronic multisystem disease associated with immune dysregulation, vascular injury and fibrosis [1]. Progressive fibrosis in the skin and lungs are prominent, and ultimately leads to organ failure accounting for the substantial mortality of SSc [2]. Inflammatory infiltrates are observed in a variety of affected organs in early-stage disease [3][4][5] and are accompanied by elevated circulating levels of inflammatory cytokines and chemokines [6,7]. Serum amyloid A (SAA) is an evolutionarily conserved ̴ 12 kDa acute phase protein [8]. Circulating levels of SAA increase >1000-fold during inflammatory responses in a manner analogous to that of CRP [9]. There are four human isotypes of SAA. Systemic SAA1 and SAA2 are induced in the liver upon stimulation by interleukin-6 (IL-6), interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) [10]. Moreover, SAA can also be produced by macrophages and other extrahepatic cells [11][12][13] as well as in the lung [14].
SAA was shown to regulate expression of TGF-β, the master regulator of connective tissue remodeling and fibrogenesis [30]. In mice, SAA adenoviral transfer leads to increased plasma TGF-β, and increased biglycan expression. Interestingly, it has been reported that the SAA receptor FPRL-1/FPR2 was involved in TGF-β, as well as in biglycan expression [30]. These studies implicate SAA in extracellular matrix remodeling.
The role of inflammation in SSc, and biomarkers for identifying inflammation, have received scant attention to date. Previous studies showed that erythrocyte sedimentation rate (ESR) is elevated in SSc and predicts mortality [31,32]. ESR is one of the parameters comprising the modified Medsger SSc Disease Severity Scale [33,34]. Levels of CRP are also elevated in SSc, correlate with disease activity and pulmonary function [7,35], and predict pulmonary decline and survival [35]. In contrast to CRP and ESR, little is known to date about SAA in SSc or its role in disease pathogenesis. A small pilot study over three decades ago showed elevated SAA levels in 24% of SSc patients; marked elevations predicted poor survival [36]. In the present study, we sought to determine circulating levels of SAA in SSc, and to correlate these levels with clinical features of the disease. Our findings indicate that SAA levels are elevated in a subset of SSc patients, and correlate with pulmonary involvement and patient-reported outcomes, in particular symptoms related to respiratory dysfunction. In vitro, recombinant SAA induced enhanced IL-6 and IL-8 production in fibroblasts explanted from normal human lungs. These findings provide evidence for the occurrence of a systemic inflammatory process in SSc, and suggest a potential for SAA as a biomarker in evaluating patients with SSc.

Patients
One hundred twenty nine consecutive adult patients with SSc, evaluated at the Northwestern Scleroderma Program between February 2009 and April 2010 were included in the study. The study was approved by Northwestern University Institutional Review Board. All patients met the ACR criteria [37]. Serum samples were obtained at scheduled visits after patients signed informed consent approved by Northwestern University Institutional Review Board. Serum was also collected from 98 healthy Caucasian volunteers (65% male, 35% female; median age 43.3 yrs), and processed in a manner identical to that for SSc serum. Samples were stored at −80°C until assayed. Clinical information obtained on the SSc patients at the time of serum collection included demographics, body mass index (BMI), disease duration (defined as interval between first non-Raynaud event and blood sampling as early (up to 36 months) and late (above 36 months)), and modified Rodnan skin score (mRSS, range 0-51). Two-dimensional echocardiography with tissue Doppler, pulmonary function tests (PFT) and high-resolution computed tomography of the chest (HRCT) were performed as clinically indicated. Pulmonary arterial hypertension was diagnosed if the estimated pulmonary artery systolic pressure was 40 mm Hg on echocardiography or with mean pulmonary artery pressure 25mm Hg and pulmonary capillary wedge pressure 15mmHg on right heart catheterization [38]. Antinuclear antibodies in the serum were detected by indirect immunofluorescence, and antibodies against topoisomerase-1, centromere and RNA polymerase III by latex immunoassay, immunofluorescence and ELISA, respectively (S1 Dataset).

Determination of serum SAA levels
Serum SAA levels were determined using Milliplex Cardiovascular Disease Panel 2 multiplex assay kits (Millipore, Billerica, MA), according to manufacturer's instructions. Briefly, samples (1:500 or 1:2000 dilution) and standards, along with sonicated beads were added to the wells. After incubation and washing, antibodies followed by streptavidin-phycoerythrin were added. Wells were then washed and fluorescence measured on a Luminex 100 platform (Luminex, Austin, TX). Two control samples with known concentrations of SAA were included in each analysis. Results were calculated from six standard samples, ranging from 0.08-250 ng/ml in concentration, with four parameter curve fit. The average coefficient of variation from replicates in all analyses was 9%.

Effects of SAA on lung fibroblast in vitro
Fibroblasts explanted from healthy adult lungs (Lonza, Walkersville, MD, USA) were used. Fibroblasts were seeded in 6-well plates and maintained in fibroblast basal medium with growth supplements (Lonza, Walkersville, MD, USA) and 20% FBS in a humidified atmosphere of 5% CO2 at 37°C. Sub-confluent low passage fibroblasts were incubated with human recombinant SAA (Peprotech EC Ltd, London, UK) at indicated concentrations for 24h. Levels of IL-6 were determined in culture supernatants by ELISA (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. RNA was isolated from confluent fibroblasts using RNeasy Plus Micro Kits (Qiagen, Hilden, DE) and reverse transcribed using Reverse transcription System (Promega, Madison, WI, USA). StellARray platforms were used to measure gene expression (Bar Harbor BioTechnology, Trenton, ME, USA).

Statistical analysis
The normality of distribution of SAA levels was determined by Kolmogorov-Smirnov test. Due to non-normal distribution of the data, summary statistics are expressed as medians and interquartile ranges and nonparametric tests were performed. Mann-Whitney U tests or Kruskal-Wallis tests were used to compare SAA levels. Cut-off was defined as 95 percentile of healthy controls. Spearman's rank correlations were calculated to measure the correlation between SAA levels and various clinical/laboratory parameters, which accommodated skewedness in measures of SAA. Because SAA was found not to correlate with age, gender or ethnicity, partial correlation was not used for adjustment. Odds ratios of increased SAA were calculated with 95% confidence interval (CI). Data were analyzed using SPSS Statistics 17 (Chicago, IL). P < 0.05 was considered statistically significant.

SAA levels are elevated in SSc
Levels of SAA were significantly higher in patients with SSc compared to healthy controls (U = 3419, p<0.000) (Fig. 1A). Gender, ethnicity, age or clinical SSc subtype (dcSSc or lcSSc) did not significantly influence levels of SAA. Using a cut-off value of 19.5 μg/ml determined in 98 healthy controls, 37% patients with early dcSSc, but none of patients with early lcSSc, were found to have elevated SAA levels (U = 80, p = 0.007). Early stage dcSSc (defined as disease duration < 36 month from the first non-Raynaud symptom of SSc) was associated with higher SAA levels compared with late-stage disease (U = 222, p = 0.08), whereas an opposite trend was seen in patients with lcSSc (U = 269, p = 0.02) (Fig. 1B). No significant differences in SAA levels were detected when patients were classified based on their scleroderma specific autoantibody profiles.

SAA levels correlate with other markers of inflammation
The ESR and serum levels of CRP were elevated in 37.3% and 28.8% of patients, respectively. Both inflammation markers correlated with SAA (SAA/CRP r = 0.433, p = 0.001; SAA/ESR r = 0.282, p = 0.030) ( Table 3 and Fig. 3), as well as with FVC and DLCO (Table 3).

SAA increases IL-6 expression in pulmonary fibroblasts
To explore the biological activities of SAA on fibroblasts, the primary effector cells of fibrosis linked to the pathogenesis of SSc, low-passage fibroblasts explanted from healthy lungs were incubated with recombinant SAA, followed by determination of secreted IL-6 and changes in fibroblast gene expression in culture. The results indicated that SAA caused a dose-dependent increase in IL-6 secretion (Fig. 4). Moreover, SAA enhanced the expression of IL-6 and IL-8 mRNA (Table 4). In addition, SAA also stimulated the production of matrix metalloproteinases MMP-1 and MMP-12. In contrast, no consistent effect of SAA on collagen gene expression was observed (Table 4).

Discussion
We show here that circulating levels of the inflammatory marker SAA are elevated in patients with SSc. Elevated SAA levels are associated with signs and symptoms of pulmonary involvement, as well as health-related quality of life measures. In particular, levels of SAA were found to correlate with measures of pulmonary function and radiologic evidence of SSc-associated interstitial lung disease. Furthermore, SAA levels were significantly correlated with PA pressure, in a manner analogous to recent findings in patients with idiopathic pulmonary arterial hypertension [42]. Exposure of healthy lung fibroblasts in culture to SAA resulted in stimulation of the expression of IL-6 and IL-8, two cytokines previously implicated in the pathogenesis of SSc. The levels of SAA were only modestly correlated with those of the inflammatory markers CRP and ESR. While levels of CRP and ESR were elevated in 29 and 37% of SSc patients, respectively, the correlation with SAA was less than 0.5, revealing unexpected differences in these three inflammatory parameters in SSc. Our observations are broadly consistent with previous studies examining CRP and ESR in SSc [7,35]. Chronic inflammation and fibrosis are often linked, particularly in interstitial lung disease. For instance, in patients with sarcoid lung disease, SAA correlated with collagen deposition and lung fibrosis [12] and negative correlation of lung functions and SAA was found [43].  Recombinant SAA potently stimulated the production of IL-6 and IL-8 in lung fibroblasts in culture. Importantly, these stimulatory effects of SAA on cytokine gene expression occurred at physiologic concentrations of SAA.
We previously reported that SAA stimulated IL-6 in human endothelial cells in culture [44,45] and IL-8, MMP-3 proteins and NF-ƘB DNA binding activity were up-regulated by SAA in fibroblast-like synoviocytes [46]. In this study we report stimulation of IL-6 in lung fibroblasts at the mRNA level and secreted cytokine production. IL-6 is emerging as a potentially important mediator of fibrosis in SSc. In fibroblasts, SAA has been recently shown to trigger a TLR2-dependent innate immune pathway, contributing to induction of IL-6, and potentially linking SAA to innate immunity and fibrosis in SSc [47]. IL-6 is implicated in the regulation of  collagen gene expression and extracellular matrix production [48,49]. Furthermore, levels of IL-6 are elevated in serum and lesional tissue of patients with SSc [50,51]. Treatment of SSc patients with anti-IL-6 intervention was shown to have beneficial effects in a small clinical trial [52]. IL-8 is a multifunctional chemokine produced primarily by macrophages, and exerting potent effects on chemotaxis and angiogenesis. Scleroderma fibroblasts spontaneously secrete IL-8 [53]. We and others have shown that levels of IL-8 are elevated in the serum, as well as in bronchoalveolar lavage fluid, from patients with SSc [54][55][56].
The present results demonstrate elevated circulating SAA levels in a subset of SSc patients that are correlated with symptoms and signs of SSc-associated pulmonary involvement. The biological implications of these findings remain to be elucidated. It is noteworthy, however, that in lung fibroblasts, SAA acts as a direct stimulus for the synthesis of IL-6 and IL-8, mediators implicated in the pathogenesis of SSc and its pulmonary complications. Longitudinal studies to determine if baseline SAA levels in SSc predict disease activity or progression, and whether changes in SAA levels over time correlate with changes in measures of disease activity, seem warranted.
Supporting Information S1 Dataset. (XLSX) S1 Table. Patient-reported outcomes in patients with normal and elevated SAA. Short form-36 (SF-36), patient-reported outcomes measurement information system (PROMIS-29) and health assessment questionnaire-disability index (sHAQ-DI) were collected within 6 months of serum collection. (DOCX) S2 Table. SAA levels and chest radiologic patterns. SAA median levels associated with different radiologic ILD patterns on high resolution computerized tomography (HRCT) of the chest. IQR, interquartile range. Kruskal Wallis test to compare SAA levels among different HRCT patterns was significant (p = 0.03), so Mann Whitney pairwise comparisons were performed and adjusted for overall p-value using Bonferroni correction. (DOCX) S3 Table. SAA levels are associated with radiologic patterns and pulmonary function tests. HRCT, high resolution computerized tomography; FVC, forced vital capacity; DLCO, carbon monoxide diffusing capacity. (DOCX)