Early systemic sclerosis: marker autoantibodies and videocapillaroscopy patterns are each associated with distinct clinical, functional and cellular activation markers

Early systemic sclerosis (SSc) is a condition characterized by Raynaud's phenomenon (RP) associated with SSc marker autoantibodies (anti-Scl-70, anticentromere antibodies, anti-RNA polymerase III, anti-fibrillarin, anti-PmScl, and anti-Th/To) and/or nailfold videocapillaroscopy (NVC) findings typical of SSc (namely, megacapillaries and avascular areas - commonly referred to as an 'NVC scleroderma pattern'). By definition, patients with early SSc do not show any distinctive clinical manifestation of the disease (namely, sclerodactyly, digital ulcers/scars, two or more teleangectasias, clinically visible nailfold capillaries, cutaneous calcinosis, X-ray bibasilar lung fibrosis, X-ray esophageal dysmotility, electrocardiographic signs of myocardial fibrosis, or a serum creatinine increase suggestive of scleroderma renal crisis) except puffy fingers and/or arthritis [1–4].

Various attempts have been made to find a correct classification for these patients who do not meet the preliminary 1990 American College of Rheumatology (ACR) SSc criteria [5] in order to foster their inclusion in clinical trials. In fact, in 2001 LeRoy and Medsger proposed that RP patients with SSc marker autoantibodies or an NVC scleroderma pattern should be classified as being affected by limited SSc [1]. Seven years later, Koenig and colleagues validated the LeRoy and Medsger criteria in a large 20-year prospective study [2]. They found that, at the last follow-up, patients presenting with SSc marker autoantibodies and a scleroderma pattern at capillaroscopy (referred to as subset 1 in the present paper), with no other manifestation distinctive of the disease other than puffy fingers and/or arthritis, had developed definite SSc 60 times more frequently than patients presenting with only RP. Patients with either only SSc marker autoantibodies (subset 2) or a capillaroscopic scleroderma pattern (subset 3) had developed definite SSc five and eight times more frequently, respectively, than patients presenting with only RP. Koenig and colleagues therefore suggested that such cases be labeled collectively as early SSc.

Two years ago, we reported that 42% of thus defined early SSc patients admitted to our tertiary Rheumatology Unit had preclinical functional alterations of the esophagus, lung or heart [3]. Subsequently, we showed that a high percentage of early SSc patients develop definite SSc within 5 years of presentation and found that some circulating fibroblast, endothelial and T-cell activation markers could be used as predictors of disease evolution besides marker autoantibodies and NVC findings [4]. In their study, Koenig and colleagues did not systematically investigate visceral involvement and neither did they consider circulating activation markers [2]. In our previous papers we were unable to try any subsetting among our early SSc patients because of the low number of early SSc patients with only an NVC scleroderma pattern and negative for SSc marker autoantibodies admitted to our unit [3, 4].

In the present study, we evaluated whether clinical, preclinical functional or cellular activation markers are related to distinct early SSc subsets at presentation by merging our early SSc cohort with that recruited at a secondary Angiology Unit, where a significant number of patients with early SSc were SSc marker antibody-negative and capillaroscopy-positive.

By definition, no patient presented any clinical feature of SSc other than puffy fingers or arthritis. The prevalence of puffy fingers was significantly higher in both subset 1 and subset 3 Angiology Unit patients than in the respective Rheumatology Unit patients (6/6 vs. 4/26, P = 0.004 for subset 1; 19/20 vs. 1/3, P = 0.04 for subset 3). The prevalence of arthritis did not differ statistically between the two centers (1/26 (3.85%) for Rheumatology Unit patients vs. 2/6 (33.3%) for Angiology Unit patients, P >0.05 for subset 1; 0/3 (0%) for Rheumatology Unit patients vs. 2/20 (10%) for Angiology Unit patients, P >0.05 for subset 3). Because of the low number of subset 3 patients recruited at the Rheumatology Unit and of subset 1 patients admitted to the Angiology Unit, in order to increase the power of the analysis we merged the two series.

Figure 1 shows the prevalence of puffy fingers (Figure 1A) and arthritis (Figure 1B) in the three subsets. The prevalence of puffy fingers was significantly higher in subset 3 patients (20/23, 87%) from the pooled series than in subset 1 (10/32, 31.3%) and subset 2 (2/16, 12.5%) patients (P = 0.0001) (Figure 1A). Thirty out of the 32 (93.75%) patients across the three subsets with puffy fingers had a positive NVC, and 25/39 (64.1%) patients without puffy fingers had a positive NVC (P = 0.0037) (Figure 1C).

The prevalence of functional abnormalities did not differ between each subset of patients from the two centers. A forced vital capacity <80% of the predicted value and an inverted E/A ratio were found only in 2/32 (6.25%) and 1/32 (3.13%) subset 1 patients, respectively, and in no subset 2 and subset 3 patients. However, an impaired DLCO was more prevalent in subset 1 (11/32, 34.4%) and subset 2 (6/16, 37.5%) patients than in subset 3 (2/23, 8.7%) patients (P = 0.057) (Figure 2A). In addition, out of the 48 marker autoantibody-positive patients (that is, subset 1 + subset 2), DLCO <80% of the predicted value was detected in 17 (35.4%) versus 2/23 (8.7%) marker autoantibody-negative subset 3 patients (P = 0.0217) (Figure 2B). Lastly, 6/14 (42.9%) patients with SSc marker autoantibodies other than anticentromere antibodies (three anti-Scl-70; three anti-RNA polymerase III) had a reduced DLCO as compared with 11/34 (34.4%) anticentromere antibody-positive patients. The difference did not reach statistical significance (Figure 2C).

A ground glass opacity involving <5% of inferior lobes and an interlobular septal thickening were found in two patients and two different patients, respectively, by lung high-resolution computed tomography.

Sera collected at admission were available for 22/45 (48.9%) patients from the Rheumatology Unit and for 22/26 (84.6%) patients from the Angiology Unit. The clinical, serological and NVC features of the 22 Rheumatology Unit patients reflected the characteristics of the whole series of the center (data not shown). Figure 3 shows serum levels of the investigated activation markers in the three early SSc subsets and in controls. Two main findings emerged from this analysis. Firstly, serum concentrations of ICTP were significantly higher in patients with SSc marker autoantibodies (median = 4.025 µg/l in subset 1 + subset 2, range 1.883 to 10.53 µg/l) than in subset 3 patients (median = 2.51 µg/l, range 1.55 to 6.972 µg/l; P = 0.0108) (Figure 3A). In detail, 4/15 (26.7%) subset 1 patients and 3/9 (33.3%) subset 2 patients had ICTP levels that exceeded the 95th percentile (5.05 µg/l) of the values measured in controls with respect to 2/17 (11.8%) subset 3 patients. Secondly, levels of sE-selectin were significantly higher in patients with an NVC scleroderma pattern (median = 1.031 pg/ml in subset 1 + subset 3, range 0.425 to 7.51 pg/ml) than in subset 2 patients with only SSc marker autoantibodies (median = 0.67 pg/ml, range 0.486 to 1.203 pg/ml; P = 0.033) (Figure 3B). In detail, sE-selectin levels exceeded the 95th percentile of the values recorded in controls (1.709 pg/ml), in 11/17 (64.7%) subset 3 patients, in 2/15 (13.3%) subset 1 patients and in 0/9 subset 2 patients (P = 0.0003). Noteworthy, we found that among all three subsets sE-selectin levels were significantly higher in patients with puffy fingers than in those without puffy fingers (2.013 pg/ml, range 0.55 to 7.51 vs. 0.725 pg/ml, range 0.425 to 1.203 pg/ml; P < 0.0001) (Figure 3C). Finally, sIL-2Rα levels were lower in SSc marker autoantibody-negative patients (subset 3: median = 24.41 pg/ml, range 0 to 411.2) than in SSc marker autoantibody-positive patients (subset 1 + subset 2 patients: median = 239.8 pg/ml, range 66.74 to 1,003 pg/ml; P = 0.0015) (Figure 3D). sIL-2Rα levels were lower than the 5th percentile of control values (25.7 pg/ml) in 11/17 (64.7%) subset 3 patients but not in any subset 1 or subset 2 patient (P = 0.0001).

The mean capillary number was >9/mm in 41 patients, 9 to 7/mm in 24 patients and 6 to 4/mm in six patients. No statistical difference was found among these three subgroups in the prevalence of each clinical, preclinical functional and laboratory abnormality investigated (data not shown). Finally, none of the investigated clinical (puffy fingers, arthritis), preclinical (DLCO/forced vital capacity reduction, E/A ratio inversion), or serological (SSc marker autoantibodies, ICTP, sE-selectin, sIL-2Rα) parameters was associated with the disease duration, as evaluated by RP onset.

To our knowledge, this is the first study to systematically explore clinical and preclinical organ involvement and circulating activation markers in early SSc patients subdivided into three subsets, depending on the presence of both marker autoantibodies and NVC scleroderma pattern or either of these features.

Firstly, we found that our early SSc patients, which were enrolled by two distinct clinical units, were similar in sex and age distribution and disease duration, but differed in terms of clinical features and serum levels of fibroblast, endothelial, and T-cell activation markers at presentation.

Given the different target patients of the two units, it is not surprising that a higher prevalence of SSc marker autoantibody-positive patients was enrolled by the Rheumatology Unit whereas a higher prevalence of RP patients presenting with puffy fingers was enrolled by the Angiology Unit. These differences are in line with results reported in patients with definite SSc by Walker and colleagues, who analyzed data entered into the European League Against Rheumatism (EULAR) Scleroderma Trials and Research database and found that differences among patients from various centers depend mainly on the clinical attitude of the respective medical team [16]. Notwithstanding this possible center-related recruitment bias, in our study early SSc patients with marker autoantibodies had a higher prevalence of impaired DLCO and higher circulating levels of ICTP irrespective of an NVC scleroderma pattern (megacapillaries and/or avascular areas). Similarly, early SSc patients with an NVC scleroderma pattern had a higher prevalence of puffy fingers and higher circulating levels of E-selectin irrespective of circulating marker autoantibodies, the two features being closely related each other. These associations have never previously been searched for. Moreover, four out of the 14 Rheumatology Unit patients with DLCO <80% of the predicted value already presented mild alterations of lung anatomy as detected by high-resolution computed tomography.

In this context, it is noteworthy that an impaired DLCO is considered the first alteration in SSc interstitial lung disease [14, 17]. Increased ICTP levels have been reported in SSc patients [18] and suspected SSc patients (that is, patients not meeting the preliminary 1990 ACR criteria for SSc) [4]. Moreover, elevated ICTP levels have been found to correlate with modified Rodnan skin score and to be associated with a diffuse disease and a reduced pulmonary function [19]. Given these observations, it is conceivable that early SSc patients with SSc marker autoantibodies are at higher risk of developing fibrotic organ complications.

Puffy fingers and RP are considered the main clinical features of the undifferentiated connective tissue syndrome [20]. sE-selectin, which reflects endothelial activation, has been reported to be increased in patients with definite SSc [21]. Our study demonstrates that puffy fingers are detectable in early SSc patients, mainly in those with an NVC scleroderma pattern and increased sE-selectin levels. Serum IL-2Rα levels, which reflect the activation of the adaptive immune response through T-cell recruitment, have been found to be strongly associated with mortality and inversely correlated to disease duration in SSc patients with diffuse disease [22]. Intriguingly, we found that sIL-2Rα levels were strikingly lower in early SSc patients who were marker autoantibody-negative. This result could indicate a lack of T-cell recruitment in the peripheral blood of autoantibody-negative early SSc patients.

Taken together, the results of our study seem to suggest that the major cellular pathways involved in the pathogenesis of SSc, as evaluated by serum concentration of relevant activation markers, differ among early patients at presentation, in relation to autoantibody and microvascular status. Nevertheless, the asynchronous appearance of autoantibody positivity and capillaroscopic alterations has been reported in two recent papers. First, Englert and colleagues described the case of one patient in whom anti-Scl-70 positivity, associated with fatigue, weight loss and puffy fingers, predated the appearance of RP and the development of capillaroscopic abnormalities [23]. Second, Moinzadeh and colleagues reported the late appearance of serum autoantibodies in three patients with an NVC scleroderma pattern at admission [24]. Whether SSc marker autoantibody positivity or the presence of an NVC scleroderma pattern really do identify the dominating activation of a given cellular pathway at the disease onset in early SSc patients can only be determined by investigating circulating and tissue markers of cell activation contemporaneously.

Our study has some limitations. First, it might be affected by a differential verification bias because the patients were examined by clinicians with different medical expertise. To reduce this possible bias, however, capillaroscopic images and sera were made anonymous and centralized to the Rheumatology Unit. Capillaroscopic images were blindly evaluated by one investigator experienced in NVC (MI); sera were blindly analyzed in duplicate by one investigator experienced in immunoassays (SV). Moreover, the clinical (puffy fingers) and preclinical (DLCO and E/A ratio) parameters used to assess both series of patients are well defined and routinely applied in clinical practice. Second, given the small number of patients, our conclusions may be susceptible to be changed by future studies on large number of patients, such as the ongoing Very Early Diagnosis of Systemic Sclerosis (VEDOSS) project [25, 26]. In addition, one must underline that using the new ACR/EULAR classification criteria presented at the ACR 2012 meeting [27, 28], a number of our early SSc patients could already meet the criteria for SSc at presentation. Indeed, in a post-hoc analysis 10/71 (14.1%) patients met the new EULAR/ACR classification criteria for SSc and belonged to subset 1 (that is, patients with both SSc marker autoantibodies and capillaroscopic abnormalities). These patients met the new criteria because of the presence of puffy fingers along with SSc marker autoantibodies and the NVC scleroderma pattern. However, our major findings - that is, early SSc patients with SSc marker autoantibody have higher ICTP levels, regardless of the NVC pattern; and early SSc patients with an NVC scleroderma pattern have a higher prevalence of puffy fingers and higher circulating levels of E-selectin, regardless the autoantibody profile, the two features being closely related each other - did not differ significantly when we excluded these 10 patients from the analysis (see Additional files 1, 2, 3 and 4).

In conclusion, we found that early SSc patients show distinct clinical and preclinical features, and distinct circulating markers of cellular activation at presentation, according to their autoantibody and capillaroscopic status. Longitudinal large studies are required to verify whether the pathophysiological (cellular activation markers), functional (DLCO) and clinical (puffy fingers) differences in early SSc subsets that we detected in a small yet well characterized cohort really do translate into the development of different disease patterns over time.