Association between anti-SSSCA1 antibodies and cancer in systemic sclerosis

Abstract

Objective

To define the clinical phenotype of SSc patients with antibodies against Sjogren’s syndrome (SS)/scleroderma autoantigen 1 (SSSCA1), and to examine the association between these antibodies and cancer in SSc patients.

Methods

We conducted a case–control study using data from 209 patients with SSc and cancer, and 205 SSc patients without cancer. All were randomly selected from the Johns Hopkins Scleroderma Center Research Registry. Antibodies against SSSCA1 were assayed by immunoprecipitation of ³⁵S-methionine-labelled protein generated by in vitro transcription and translation. We performed logistic regression analysis to examine the relationship between anti-SSSCA1 antibodies and cancer.

Results

Among the 414 study patients, 31 (7%) were anti-SSSCA1 antibody positive. Antibody-positive patients were more likely to have severe RP, a lower minimum ejection fraction, a trend towards more severe heart involvement and a lower baseline diffusing capacity of the lungs for carbon monoxide percent predicted than anti-SSSCA1-negative patients. Patients with cancer were significantly more likely to be anti-SSSCA1 positive compared with those without cancer [22/209 (11%) vs 9/205 (4%), respectively; P = 0.018]. Among patients with cancer, there was a trend towards longer cancer–SSc interval in anti-SSSCA1-positive patients compared with anti-SSSCA1-negative patients. Patients with anti-SSSCA1 antibodies had an increased adjusted risk of cancer (odds ratio 2.46, 95% CI 1.06, 5.70) compared with anti-SSSCA1-negative patients.

Conclusions

These data suggest anti-SSSCA1 antibody status may be of utility as a cancer biomarker in SSc. Anti-SSSCA1-positive patients with SSc may be more likely to have severe Raynaud’s and cardiac involvement.

Rheumatology key messages.

• Approximately 7% of patients with SSc were anti-SS/scleroderma autoantigen 1 (SSSCA1) antibody positive.

• Anti-SSSCA1 positivity was associated with increased cancer risk and later cancer emergence in SSc.

• Anti-SSSCA1 antibody positive patients with SSc may have more severe Raynaud’s and cardiac involvement.

Introduction

Emerging data suggest that distinct subsets of SSc patients, identified by the presence of specific autoantibodies, may have an increased or decreased risk of cancer around SSc onset or during the disease course. For example, patients with anti-RNA polymerase III antibodies have a heightened cancer risk at SSc onset, while those with anti-centromere or anti-Th/To antibodies may have a lower risk [1–3]. Identifying cancer-associated immune responses may aid in cancer risk stratification and enable targeted cancer screening in patients with SSc. Of great interest are immune responses targeting autoantigens that are highly expressed in cancers.

Anti-Sjogren’s syndrome/scleroderma autoantigen 1 (anti-SSSCA1; also called anti-p27) antibody was discovered in the serum of an anti-centromere positive patient with SS in the late 1990s [4]. Four additional patients with anti-SSSCA1 antibodies were identified by screening a cohort of 83 anti-centromere-positive (three patients) and 215 anti-centromere-negative patients (one patient) with various CTDs. All five patients had SS and/or SSc. Further research demonstrated increased expression of SSSCA1 in cultured dermal fibroblasts from patients with SSc compared with controls [5].

SSSCA1 is overexpressed in several cancers, including breast, colorectal, oral squamous cell, cervical, lung and prostate cancer, and is thought to be involved in the adenoma-to-carcinoma progression in colorectal cancer [6, 7]. The association between SSSCA1 and both SSc and cancer raises the possibility that anti-SSSCA1 antibodies could serve as an additional cancer biomarker in patients with SSc. In this study, we report the clinical phenotype associated with anti-SSSCA1 positivity in patients with SSc, and examine the association between anti-SSSCA1 positivity and cancer risk.

Patients and methods

Study population

We performed a case–control study using data and banked serum samples from 209 patients with SSc and cancer, and 205 SSc patients without known cancer after 5 years of disease. Controls were randomly selected, using a random number generator, for anti-SSSCA1 antibody testing from the Johns Hopkins Scleroderma Center Research Registry. Patients gave written informed consent to participate in the Registry and provide banked serum samples. The Registry prospectively collects demographic information and longitudinal disease metrics in the context of routine clinical care. Cancer diagnoses are prospectively collected and confirmed by medical record review. We excluded patients with non-melanoma skin cancer.

Patients were classified as without cancer if they were cancer free at least 5 years after SSc onset. SSc onset was defined as the development of the first Raynaud’s or non-Raynaud’s symptom, whichever occurred first. For patients with cancer, the time interval between cancer diagnosis and SSc onset was calculated. The following clinical characteristics (captured in the Center’s Registry) were also examined: cutaneous subtype, modified Rodnan skin score, Medsger modified organ-specific disease severity scores, renal crisis, interstitial lung disease, pulmonary hypertension, tendon friction rubs, synovitis, calcinosis and sicca symptoms (Supplementary Data S1, available at Rheumatology online) [8–11]. Sera from 36 healthy individuals were also studied. The Johns Hopkins Institutional Review Board approved the study protocol, which complies with the Declaration of Helsinki.

SSSCA1 antibody assays

Antibodies against SSSCA1 were assayed by immunoprecipitation (IP) as follows. cDNA encoding full-length human SSSCA1 was used to generate ³⁵S-methionine-labelled protein by in vitro transcription and translation (IVTT), according to the manufacturer’s protocol (Promega, Madison, WI, USA). IPs were performed by diluting 1 ml of IVTT product in 1 ml of ice-cold buffer A (20 mM Tris pH 7.4/150 mM NaCl/1 mM EDTA pH 7.4/1% Nonidet P40 and protease inhibitors). One millilitre of serum was added to each tube, and mixtures were rocked (1 h, 4°C) before adding 35 ml of protein A agarose beads (Pierce, Rockford, IL, USA; 20 min, 4°C). Washed IPs were electrophoresed on 10% SDS–polyacrylamide gels and visualized by autoradiography. An IP performed using an anti-FLAG mAb (Sigma, St. Louis, MO, USA, clone M2) was included as a positive control in each dataset. All positive sera were tested twice to confirm antibody status. The IP assay readout using IVTT input was validated by IP/blot using a subset of the sera as described [12]. Briefly, cultured human salivary gland cells were lysed in Buffer A and precleared with protein A agarose beads. IPs were performed by rocking 1.5 ml of serum with 100 mg of lysate (90 min, 4°C). After adding protein A agarose beads the IPs were electrophoresed, transferred to nitrocellulose membrane and immunoblotted with an anti-SSSCA1 mAb (Santa Cruz, Dallas, TX, USA, #515430; 1:500), followed by horseradish peroxidase-labelled anti-mouse secondary antibody (Jackson Immunoresearch, West Grove, PA, USA; 1:10 000). Detection was performed with chemiluminescence (Pierce, Rockford, IL, USA) and images were acquired using a Protein Simple Fluorochem-M digital imager. For details of Euroimmun, U1RNP and RNPC3 autoantibody assays, see Supplementary Data S2, available at Rheumatology online.

Statistical analyses

Analyses were performed using Stata software, version 17 [13]. Our primary analysis examined whether anti-SSSCA1 antibody positivity was associated with cancer or a short cancer–SSc interval. Multivariable logistic regression was performed to explore the relationship between anti-SSSCA1 antibody positivity and cancer, after adjusting for key demographic and clinical factors that may associate with cancer risk, including age at SSc onset, sex, race, cutaneous subtype, follow-up time, history of smoking and anti-RNA polymerase III antibody status [2, 3, 14, 15]. Secondary analyses explored phenotypic differences between anti-SSSCA1-positive and -negative patients. For comparison between groups, χ², Fisher’s exact test, Student’s t-tests and Wilcoxon rank sum tests were used as appropriate. Due to the exploratory nature of these analyses, adjustment for multiple comparisons was not performed.

Results

Four hundred fourteen patients were studied; 396 (96%) met 2013 ACR SSc classification criteria and 155 (37%) had diffuse disease. The mean age at SSc onset was 43.0 (s.d. 14.9) years. Most patients were female (86%), with median (interquartile range) baseline disease duration of 5.6 (1.9, 14.8) years and follow-up after first symptom of 14.4 (8.7, 23.7) years. Clinical characteristics by cancer status are presented in Supplementary Table S1, available at Rheumatology online.

Anti-SSSCA1 antibodies were assayed using a gold standard IP assay. Antibodies against SSSCA1 were not detected in any healthy control sera (0/36). Of the 414 SSc patients, 31 (7%) were anti-SSSCA1 antibody positive. Anti-SSSCA1 antibody readout with this assay was validated using an alternate assay (IP/blot, see methods section and Supplementary Fig. S1, available at Rheumatology online).

The primary analysis demonstrated that patients with cancer were significantly more likely to be anti-SSSCA1 positive compared with those without cancer [22/209 (11%) vs 9/205 (4%), respectively; P = 0.018]. After adjusting for age at SSc onset, sex, race, cutaneous subtype, follow-up time, history of smoking and anti-RNA polymerase III antibody status, the relative odds of cancer was 2.46 (95% CI 1.06, 5.70) among anti-SSSCA1-positive vs -negative patients. The most common cancers among the anti-SSSCA1-positive patients were breast (8/22 patients, 36.4%) and lung (3/22 patients, 13.6%) (see Supplementary Data S3, available at Rheumatology online for a complete list of cancers). Among patients with cancer, there was a trend towards a longer cancer–SSc interval in anti-SSSCA1-positive patients compared with anti-SSSCA1-negative patients [15.6 (S.D. 13.8) vs 11.4 (S.D. 12.0) years, respectively; P = 0.133, Fig. 1]. The mean time between first non-Raynaud’s SSc symptom and cancer differed significantly by anti-SSSCA1 antibody status: 12.3 (S.D. 10.4) years for anti-SSSCA1-positive compared with 6.9 (S.D. 9.3) years for anti-SSSCA1-negative patients (P = 0.012).

Figure 1.

Among patients with cancer, comparison of cancer–SSc interval by anti-SSSCA1 antibody status. SSc onset is defined as date of RP or first non-RP symptom, whichever was first. On the horizontal axis, positive values denote cancer diagnosis after SSc onset, and negative values indicate cancer preceding SSc onset. SSSCA1: SS/scleroderma autoantigen 1

Patients with and without anti-SSSCA1antibodies had similar demographics and frequency of diffuse cutaneous disease, RP, echocardiographic evidence of pulmonary hypertension, interstitial lung disease, severe gastrointestinal involvement, renal crisis, severe muscle disease, synovitis, tendon friction rubs and calcinosis (Table 1). There were no differences in immunosuppressive medication exposure between patients with and without anti-SSSCA1 antibodies (Supplementary Table S2, available at Rheumatology online).

Table 1.

Demographic and phenotypic characteristics of patients with SSc stratified by anti-SSSCA1 antibody status

	Anti-SSSCA1 positive	Anti-SSSCA1 negative	P-value
Overall	31 (7%)	383 (93%)
Demographics
Female sex	28 (90%)	330 (86%)	0.784
Race
AI/AN	1 (3%)	4 (1%)	0.324
Black	3 (10%)	46 (12%)	1.000
East Asian	0 (0%)	2 (1%)	1.000
Middle Eastern/Arabian	0 (0%)	2 (1%)	1.000
Indian subcontinent	0 (0%)	6 (2%)	1.000
White	27 (87%)	310 (81%)	0.397
Latin American/Hispanic	1/22 (5%)	12/187 (6%)	1.000
History of smoking	19 (61%)	188/380 (49%)	0.206
SSc disease characteristics
Met 2013 ACR criteria	31 (100%)	365 (95%)	0.383
Age of SSc onset (years)	42.8 (14.9)	43.0 (14.9)	0.941
Disease duration at cohort entry (years), median (IQR)	7.5 (1.8, 20.3)	5.5 (1.9, 14.6)	0.402
Follow-up duration (years), median (IQR)	18.6 (9.8, 29.5)	14.1 (8.7, 22.7)	0.165
Diffuse cutaneous disease	12 (39%)	143 (37%)	0.879
Baseline mRSS, median (IQR)	4.0 (2.0, 15.0)	4.0 (2.0, 14.0)	0.831
Maximum mRSS, median (IQR)	6.0 (4.0, 17.0)	7.0 (4.0, 17.0)	0.987
RP	31 (100%)	379 (99%)	1.000
Renal crisis	1 (3%)	22 (6%)	1.000
Pulmonary hypertensiona	11/30 (37%)	128/339 (38%)	0.906
Interstitial lung disease	15/30 (50%)	188/368 (51%)	0.909
Pulmonary functionb
Baseline FVC (% pred)	76.6 (18.2)	80.1 (18.6)	0.333
Baseline DLCO (% pred)	65.8 (20.3)	75.9 (23.5)	0.042
Minimum FVC (% pred)	69.0 (17.4)	69.5 (20.5)	0.909
Minimum DLCO (% pred)	55.5 (20.1)	58.8 (25.4)	0.492
Baseline ejection fraction (%)	60.6 (5.2)	61.4 (6.8)	0.532
Minimum ejection fraction (%)	53.0 (11.5)	56.5 (8.6)	0.045
Weight loss/anemiac	11 (35%)	121 (32%)	0.655
Raynaud’s syndromec	24 (77%)	226 (59%)	0.044
Lung diseasec	27 (87%)	267/366 (73%)	0.084
Heart diseasec	14/30 (47%)	111/366 (30%)	0.064
GI involvementc	18 (58%)	216 (56%)	0.857
Kidney diseasec	1 (3%)	33/367 (9%)	0.499
Muscle diseasec	2 (6%)	23/375 (6%)	1.000
Tendon friction rubs	3 (10%)	73 (19%)	0.194
Synovitis	6 (19%)	86 (22%)	0.690
Calcinosis	15 (48%)	180/382 (47%)	0.892
Sicca symptoms	21 (68%)	291 (76%)	0.306
SSc antibodies
Anti-centromere	8 (26%)	112 (29%)	0.685
Anti-topoisomerase I	9 (29%)	80 (21%)	0.288
Anti-RNA polymerase III	4 (13%)	84 (22%)	0.237
Anti-Ro52	13 (42%)	99 (26%)	0.052
Anti-Ku	3 (10%)	12 (3%)	0.061
Anti-PM/Scl	0 (0%)	8 (2%)	1.000
Anti-Th/To	2 (6%)	30 (8%)	1.000
Anti-Nor90	2 (6%)	17 (4%)	0.645
Anti-Fib	0 (0%)	27 (7%)	0.247
Anti-RNPC3	0 (0%)	22 (6%)	0.394
Anti-U1RNP	2 (6%)	29 (8%)	1.000
Association with cancer
Cancer	22 (71%)	187 (49%)	0.018
Interval: cancer–SSc (years)d	15.6 (13.8)	11.4 (12.0)	0.133
Interval: cancer–first non-RP sx (years)	12.3 (10.4)	6.9 (9.3)	0.012
Survival
Deceased	18 (58%)	140/380 (37%)	0.020
Interval: SSc–death (years)d, median (IQR)	19.2 (11.3, 33.7)	14.1 (9.6, 23.3)	0.267
Interval: 1st non-RP symptom–death (years) median (IQR)	18.2 (11.3, 33.7)	13.2 (9.2, 23.3)	0.245

Categorical variables are presented as n (%); continuous variables are presented as mean (s.d.) unless otherwise noted. A P-value in bold indicates a statistically significant difference between anti-SSSCA1+ and anti-SSSCA1–. Unless otherwise stated, values represent whether the manifestation has ever emerged or maximal disease severity over the course of follow-up.

^aPulmonary hypertension defined by estimated right ventricular systolic pressure (RVSP) ≥45 mmHg on echocardiogram.

^bPFT (pulmonary function test) data were available for 30/31 anti-SSSCA1-positive patients and 368/383 anti-SSSCA1-negative patients. Echocardiogram data was available for 28/31 anti-SSSCA1-positive patients and 342/383 anti-SSSCA1-negative patients.

^cModerate–severe disease: weight loss/anaemia = weight loss ≥10.0 kg or haematocrit ≤32.9; Raynaud’s syndrome = digital pits, ulceration or gangrene; lung = FVC and/or DLCO <70% predicted, mild-severe PH or oxygen dependence; heart = left ventricular ejection fraction <45%, clinical signs of left or right heart failure or sustained clinically important arrhythmia; GI = high dose acid reflux medications, antibiotics for bacterial overgrowth, malabsorption syndrome, episodes of pseudo-obstruction or total parental nutrition requirement; kidney = serum creatinine ≥1.7 mg/dl, or at least 2+ urine protein; muscle = strength of ≤3/5 in upper or lower extremities, or requirement of ambulatory aids.

^dSSc onset is defined as date of RP or non-RP symptom, whichever was first. For the cancer–SSc interval, positive values denote cancer diagnosis after SSc onset, and negative values indicate cancer preceding SSc onset. AI/AN: American Indian/Alaska Native; DLCO: diffusing capacity of the lungs for carbon monoxide; FVC: forced vital capacity; GI: gastrointestinal; IQR: interquartile range; mRSS: modified Rodnan skin score; pred: predicted; SSSCA1: SS/scleroderma autoantigen 1; sx: symptom explanations.

However, anti-SSSCA1-positive patients were more likely to have severe RP (frequency of digital pits, ulcers or gangrene 77% vs 59%; P = 0.044). They also had a lower minimum ejection fraction during follow-up [53.0% (S.D. 11.5) vs 56.5% (S.D. 8.6); P = 0.045], a trend towards more severe heart involvement (frequency of clinically significant arrhythmia, left ventricular ejection fraction <45% or clinical signs of heart failure: 47% vs 30%; P = 0.064), and a lower baseline diffusing capacity of the lungs for carbon monoxide percent predicted [65.8 (s.d. 20.3) vs 75.9 (s.d. 23.5); P = 0.042]. While anti-SSSCA1 antibodies have been described in SS, the frequency of sicca symptoms did not differ significantly between anti-SSSCA1-positive and -negative patients (68% vs 76%, respectively; P = 0.306). The frequency of anti-centromere, anti-topoisomerase I and anti-RNA polymerase III antibodies was similar between anti-SSSCA1-positive and -negative patients. Anti-SSSCA1-positive patients had higher mortality during follow-up than anti-SSSCA1-negative patients (58% vs 37%; P = 0.020), but there was no significant difference in the time from SSc onset to death [19.2 (11.3, 33.7) vs 14.1 (9.6, 23.3) years; P = 0.267].

Discussion

To our knowledge, this is the first study to report an association between anti-SSSCA1 antibodies and cancer in SSc, and the first to describe the phenotypic associations of this specificity in a large (N = 414), well-characterized cohort of patients with SSc.

SSSCA1 is an ∼22 kDa protein that is widely expressed in most normal tissues. Given data showing overexpression of SSSCA1 in several cancers and a role for SSSCA1 in adenoma-to-carcinoma progression [6, 7], we examined the association between anti-SSSCA1 antibodies and cancer in SSc. A case–control study design was used, in which the cancer status and the time interval between SSc diagnosis and cancer emergence were well-defined in 414 patients. We found a 7% prevalence of these antibodies in our sample. This frequency is similar to the 4% prevalence found by Muro and colleagues in a Japanese cohort [4]. In our study, anti-SSSCA1 antibodies were significantly associated with an increased risk of cancer. Of the 209 patients with SSc and cancer in the study cohort, cancer emergence was significantly delayed in those with anti-SSSCA1 antibodies.

While there are multiple possible reasons for delayed cancer onset, including exposure to cytotoxic therapies and malignant transformation in inflamed tissues, another intriguing hypothesis is that scleroderma is a byproduct of anti-tumour immunity which successfully eliminates or delays clinical cancer onset [3]. In support of this hypothesis, patients with autoimmune paraneoplastic syndromes tend to have smaller tumours than those without autoimmune paraneoplastic syndromes [16]. Additionally, anti-SSSCA1 antibodies are not the first to be associated with delayed-onset cancer. In patients with TIF1-γ-positive DM, the presence of autoantibodies recognizing cell division cycle and apoptosis regulator protein 1 (CCAR1) is associated with lower risk of cancer around the time of DM onset. When cancers emerged in these patients, they were delayed and at an earlier stage than in anti-CCAR1-negative patients [17]. Careful study of the late-emerging cancers in anti-SSSCA1-positive patients with SSc will likely provide important new insights into the mechanisms targeting this autoantigen. Such studies may also further our understanding of a possible SSc/cancer continuum in the context of immunoediting, as has recently been proposed for patients with DM and late-emerging cancers [17].

In addition to the association with cancer, our data suggest that anti-SSSCA1 antibodies provide prognostic information about risk of myocardial involvement and severe Raynaud’s syndrome in patients with SSc. Interestingly, while anti-SSSCA1 antibodies have been described in patients with SS, SSc and SSc/SS, anti-SSSCA1 antibodies were not associated with more frequent sicca symptoms in our patients with SSc.

A major strength of our study is the use of extremely well characterized patients with SSc, all with well-defined cancer status, including timing. A limitation of the study is the possible misclassification of cancer-free patients, since patients designated as cancer free may not have been followed for long enough to be diagnosed with long-interval cancer. A further limitation is that anti-SSSCA1 antibody status was determined using baseline serum for patients without cancer, while using the closest serum to cancer diagnosis for patients with cancer. Additionally, the phenotypic analyses were exploratory and require validation in other SSc cohorts.

We demonstrate, for the first time, an association between anti-SSSCA1 antibodies and malignancy in patients with SSc and delayed cancer emergence. In the context of prior data showing overexpression of SSSCA1 in several cancers, the association with a long cancer–SSc interval raises the intriguing possibility that anti-SSSCA1 antibodies could be part of a protective anti-tumor immune response that delays cancer emergence in patients with SSc. Future studies are warranted to validate these findings in other SSc cohorts with well-defined cancer status, to examine the potential anti-tumor effects of anti-SSSCA1 immune responses, and to determine whether SSSCA1 is highly expressed in cancerous tissues from these patients.

Supplementary data

Supplementary data are available at Rheumatology online.

Data availability statement

The data used in this article cannot be shared publicly in order to protect the privacy of the participants in this single-center study. The data will be shared on reasonable request to the corresponding author.

Funding

This work was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases at the National Institutes of Health [grant numbers R01 AR073208, K24 AR080217, P30 AR070254, and T32 AR048522-18]; The Jerome Greene Foundation; The Donald B. and Dorothy L. Stabler Foundation; The Chresanthe Staurulakis Memorial Fund; and The Johns Hopkins inHealth initiative.

Disclosure statement: Dr Shah has received clinical trial grants from Eicos Sciences, Arena Pharmaceuticals, Kadmon Corporation and Medpace LLC. Drs Shah and Casciola-Rosen are co-inventors on a patent, “Autoimmune Antigens and Cancer,” and co-inventors on a pending patent application, “Materials and Methods for Assessing Cancer Risk and Treating Cancer.”