Scleroderma patients with antibodies against the large subunits of both RNA polymerases-I and -III are protected against cancer

Abstract

Objectives:

While compelling data suggest a cancer-induced autoimmunity model in scleroderma patients with anti-RNA polymerase III large subunit (anti-RPC155) antibodies, ~85% of these patients do not manifest cancer. We hypothesized that anti-RPC155-positive scleroderma patients without detectable cancer target additional autoantigens.

Methods:

168 scleroderma patients with anti-RPC155 antibodies were studied: 80 with a cancer history, and 88 with no cancer diagnosis after >5 years of follow-up. Thirty-five (17 with cancer, 18 without cancer) were randomly selected for autoantibody discovery using immunoprecipitation. An ~194 kDa band was enriched in the subgroup without cancer; this was defined as RNA polymerase I large subunit (RPA194).

Results:

RPA194 generated by in vitro transcription/translation was used for immunoprecipitations performed on the entire cohort to test whether anti-RPA194 was enriched among anti-RPC155-positive patients without cancer. Anti-RPA194 antibodies were significantly more common in the group without cancer (16/88, 18.2%) than the group with cancer (3/80, 3.8%; p=0.003). Patients with anti-RPA194 and anti-RPC155 were significantly less likely to have severe gastrointestinal disease (26.3% vs 51.0%, p=0.043) than patients with only anti-RPC155.

Conclusions:

Anti-RPA194 antibodies are enriched in anti-RPC155-positive scleroderma patients without cancer. Since somatic mutations in the gene encoding RPC155 in scleroderma patient cancers appears to play a role in immune response initiation against RPC155 in those patients, these data raise the possibility that the development of immune responses to both RPC155 and RPA194 may influence clinical cancer emergence. Further study is required to define whether different autoantibody combinations have utility as tools for cancer risk stratification in scleroderma.

Introduction

Emerging data suggest that subsets of systemic sclerosis (scleroderma) patients may have cancer-induced autoimmunity (1). This relationship between cancer and scleroderma emergence has been most striking among scleroderma patients with antibodies against the large subunit of RNA polymerase III (RPC155). Scleroderma patients with these autoantibodies have a significantly higher risk of cancer within a short interval of scleroderma onset compared to scleroderma patients without anti-RPC155 antibodies (2–7). Furthermore, recent data demonstrate that this translates to a 2.8-fold increased risk of cancer within 3 years of scleroderma onset when compared to the expected cancer incidence in the general population (8). Mechanistic studies have demonstrated that genetic alterations (somatic mutations and/or loss of heterozygosity) are present in the gene (POLR3A locus) that encodes for RPC155 in some of these patients’ cancers, with development of both mutation-specific and cross-reactive immune responses (9).

While these data strongly suggest a model of cancer-induced autoimmunity, it is notable that ~85% of scleroderma patients with anti-RPC155 antibodies do not manifest a cancer clinically over extensive follow-up (8). These data raise the tantalizing possibility that cancer may be an underlying trigger for scleroderma in most patients with anti-RPC155 antibodies, with the anti-tumor immune response being variably successful in eliminating the cancer or maintaining it in equilibrium such that it does not emerge (10). In this context, an important relevant property of the immune response is its ability to diversify to additional epitopes within the primary target (intramolecular spreading) and also to additional proteins that bind to the primary target at some stage in its functional cycle (intermolecular spreading) (11). It is noteworthy that many targets of the autoimmune response in scleroderma (e.g. RNA polymerases, the minor spliceosome and the centromere) are multi-component complexes. Furthermore, multiple components of these complexes are recognized by autoantibodies, suggesting antigenic spreading (12).

We hypothesized that the immune response in anti-RPC155 positive scleroderma patients in whom cancer does not emerge might target additional autoantigens. To address this, we initially studied a small group of patients with anti-RPC155 antibodies with and without cancer, and compared the autoantibody specificities in these 2 groups by immunoprecipitation. Interestingly, in anti-RPC155 antibody positive patients without cancer, a 194 kDa protein was enriched. Noting the molecular weight, the prior description of RNA polymerase I as an autoantigen in scleroderma (13), and the observation that an inhibitor inducing destruction of the catalytic subunit of RNA polymerase I (RPA194) is itself an effective anti-cancer agent (14), we pursued whether, and then rapidly confirmed, that the 194 kDa protein was RPA194. When the frequency of RPA194 antibodies was assayed in a large cohort of anti-RPC155-positive scleroderma patients with and without cancer, we confirmed that anti-RPA194 antibodies were enriched among anti-RPC155 patients without cancer.

These data strongly suggest that scleroderma patients targeting the catalytic components of both RNA polymerase I and III complexes (that is, RPA194 and RPC155, respectively) are associated with decreased emergence of cancer, raising the possibility that the combined immune responses may affect cancer survival and fitness. These observations have important implications for understanding the mechanisms underlying the association of cancer and scleroderma, as well as control of cancer by the immune system. Knowing the RPA194 antibody status in anti-RPC155-positive patients may also enable improved precision in cancer prediction in this subgroup.

Methods

Study population.

Patients with systemic sclerosis (scleroderma), as defined by the 2013 ACR/EULAR classification criteria, 1980 ACR criteria, or having at least 3 of 5 CREST (calcinosis, Raynaud’s phenomenon, esophageal dysmotility, sclerodactyly and telangiectasia) criteria, and a banked serum sample were included for study (15, 16). One hundred sixty-eight scleroderma patients with anti-RPC155 antibodies were identified for this study (RPC155 antibody status was determined by clinically obtained assays): 80 with a history of cancer and 88 who had no history of cancer after at least 5 years of follow up. Anti-RPC155 antibody status was subsequently confirmed in all 168 patients using ELISA (INOVA Diagnostics), using 20 units as the cutoff for assigning antibody positivity as recommended by the manufacturer. From this overall study population, 35 sera were randomly selected for initial autoantibody discovery: 17 anti-RPC155 scleroderma patients with cancer and 18 anti-RPC155 scleroderma patients without cancer. The RPC155 antibody status in this subset was validated by a second assay method – immunoprecipitation using ³⁵S-methionine labeled RPC155 generated by in vitro transcription and translation (see below). For the cancer patients, the closest serum sample to cancer diagnosis was studied. For patients without cancer, the most recent available serum sample was studied.

Exposure and outcome assessment.

Demographic data, scleroderma onset dates, clinical characteristics, and cardiopulmonary testing data were abstracted from the cohort database. Age at scleroderma onset was defined as the age at the first scleroderma symptom, either Raynaud’s or non-Raynaud’s. Scleroderma cutaneous subtype was classified as limited if skin thickening was distal to the elbows and/or knees, and diffuse if skin thickening involved the upper arms, thighs, chest or abdomen (17). Measurements of forced vital capacity (FVC) and diffusing capacity (DLCO) were adjusted for age and sex (18, 19). A restrictive ventilatory defect suggestive of interstitial lung disease (ILD) was defined by an FVC ever <70% of predicted. Echocardiographic evidence suggestive of pulmonary hypertension was defined by a right ventricular systolic pressure (RVSP) ever ≥45 mmHg (20). Myopathy was defined by a history of abnormal muscle enzymes, electromyography, muscle magnetic resonance imaging, and/or muscle biopsy. Severe Raynaud’s phenomenon was defined by a maximum Medsger peripheral vascular severity score ≥2 (digital pitting scars, ulcers, or gangrene), and severe gastrointestinal (GI) disease was defined by a modified maximum Medsger GI severity score ≥2 (requirement of high dose anti-reflux medications or antibiotics for small intestinal bacterial overgrowth, malabsorption syndrome, episodes of pseudo-obstruction, or hyperalimentation required) (21). Cancer diagnosis site, histology and date were reviewed in all subjects and confirmed by cancer pathology reports, oncology records and other physicians’ notes. The interval between cancer and scleroderma was calculated as the difference between the cancer diagnosis date and that of the first scleroderma symptom.

Immunoprecipitation using ³⁵S-methionine labeled proteins generated by in vitro transcription and translation (IVTT).

cDNAs encoding full length human RPC155 (purchased from Origene) and RPA194 (human RPA194 cloned into pCMV6-HA-His vector from Origene) were used to generate 35S-methionine-labeled proteins by IVTT, per the manufacturer’s protocol (Promega). Immunoprecipitations using these products were performed (22), and the immunoprecipitates were electrophoresed on 10% SDS-polyacrylamide gels and visualized by fluorography. As we reported previously, IVTT immunoprecipitation and ELISA assays gave identical readouts for RPC155 antibody status (12). For the RPA194 immunoprecipitations, experiments were normalized by including a reference immunoprecipitation performed with the same strong positive anti-RPA194 serum in each set (and electrophoresed on each gel). Densitometry was performed on all autoradiograms, and values were normalized to the reference immunoprecipitation included in each. Antibody positivity was defined as a normalized densitometry value of >1. Immunoprecipitations performed with IVTT RPA194 using sera from 36 healthy controls were all negative.

Immunoprecipitation from radiolabeled HeLa and 624 Melanocyte cells.

HeLa cells (purchased from ATCC) and 624 melanocyte cells were cultured using standard protocols before plating and radiolabeling for 2 hours with 35S-methionine/cysteine. Cells were lysed in RIPA buffer (50 mM Tris pH7.4, 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS), and precleared with immobilized Protein A agarose (Thermo Scientific). Immuno-precipitations were performed by adding 1 μl of patient serum to the lysate (1 hr, 4°C), followed by Protein A agarose (25 mins, 4°C). After washing, the immunoprecipitates were electrophoresed on 10% SDS- polyacrylamide gels and visualized by fluorography. Of note, immunoprecipitations performed using radiolabeled cell lysates detected all antibody specificities in the serum, whereas those performed with IVTT protein read out antibodies only against the input protein.

Statistical analysis.

Demographic, clinical and autoantibody characteristics were compared between (i) anti-RPC155-positive patients with and without cancer and (ii) anti-RPA194-positive patients and anti-RPA194-negative patients using the Student’s t-test for continuous variables and the chi-square or Fisher’s exact test for dichotomous or categorical variables where appropriate. The Mann Whitney U test was used to assess differences in ordinal variables. Statistical analyses were performed using Stata version 13.1 (StataCorp, College Station, TX). Two-sided p-values <0.05 were considered statistically significant.

Results

Clinical characteristics of the anti-RPC155 positive scleroderma cohorts used in this study.

This study was performed using sera from a cohort of 168 well-characterized scleroderma patients with anti-RPC155 antibodies evaluated at the Johns Hopkins Scleroderma Center. Eighty sera were from patients with a history of cancer, and 88 were from randomly selected anti-RPC155-antibody positive patients with no known malignancy after ≥5 years of follow up. Clinical characteristics of these patients were examined by cancer status (Table 1). Scleroderma patients with anti-RPC155 and cancer had a short interval between cancer diagnosis and scleroderma onset (mean 1.1 years (SD 11.3) from the 1^st non-Raynaud’s symptom; 4.6 years (SD 14.9) from the first symptom – either Raynaud’s or non-Raynaud’s). Scleroderma patients with cancer were older at scleroderma onset (mean 51.7 years (SD 14.0) vs. 45.3 years (SD 13.6); p=0.0035) and less likely to have calcinosis (33.8% vs 58.0%, p=0.002) than scleroderma patients without cancer. Both groups had a comparable disease duration at presentation to our center and similar sex, race, subtype, and smoking history. Clinical characteristics including skin severity, baseline pulmonary function, baseline right ventricular systolic pressures, and frequency of renal crisis, myopathy, telangiectasia and tendon friction rubs were not statistically different between groups. The mean baseline left ventricular ejection fraction was statistically, but not clinically, significantly higher in the cancer group. Cancer sites observed in the cancer group are detailed in Supplemental Table 1.

Table 1.

Clinical characteristics of anti-RPC155 positive patients, with and without cancer.

Variable	Cancer (N=80)	No known cancer (N=88)	p-value
Age at scleroderma onset* (years), mean (SD)	51.7 (14.0)	45.3 (13.6)	0.0035
Age at first non-Raynaud’s symptom (years), mean (SD)	55.1 (11.3), N=79	47.4 (12.2), N=87	<0.0001
Disease duration at first visit (years), mean (SD)	6.4 (10.6)	6.2 (7.8)	0.9331
Cancer-scleroderma interval (years), mean (SD)	4.6 (14.9)	NA	NA
Cancer-1st non-Raynaud’s symptom interval (years), mean (SD)	1.1 (11.3)	NA	NA
Male sex, no. (%)	19 (23.8)	12 (13.6)	0.091
Race, no. (%)	N=79		0.519
White	77 (97.5)	82 (93.2)
Black	2 (2.5)	4 (4.6)
Asian	0 (0)	2 (2.3)
Subtype, no. (%)			0.374
Limited	21 (26.3)	18 (20.5)
Diffuse	59 (73.8)	70 (79.6)
Ever smoked, no. (%)	44 (55.0)	39 (44.3)	0.167
mRSS at first visit to Center, mean (SD)	18.7 (12.9), N=79	19.1 (13.8), N=86	0.8359
Maximum ever mRSS, mean (SD)	22.5 (13.5), N=79	22.4 (14.9)	0.9563
Renal crisis, no. (%)	10 (12.5)	11 (12.5)	1.000
Myopathy, no. (%)	13 (16.3)	14 (15.9)	0.952
ILD (FVC ever <70%)**, no. (%)	28 (37.3), N=75	43 (50.6), N=85	0.092
Baseline pulmonary function, mean (SD)
FVC (% predicted)	85.2 (15.2), N=73	83.0 (17.1), N=79	0.4052
DLCO (% predicted)	82.3 (21.2), N=65	83.7 (25.2), N=70	0.7213
Pulmonary hypertension^, no. (%)	18 (23.4), N=77	31 (37.4), N=83	0.055
Baseline RVSP (mmHg), mean (SD)	32.3 (10.0), N=45	33.0 (9.4), N=51	0.7330
Baseline ejection fraction (%), mean (SD)	63.0 (6.8), N=73	59.9 (7.4), N=73	0.0112
Severe Raynaud’s^^, no. (%)	39 (48.8)	54 (61.4)	0.100
Severe GI disease^^, no. (%)	38 (47.5)	43 (48.9)	0.860
Calcinosis, no. (%)	27 (33.8)	51 (58.0)	0.002
Telangiectasia, no. (%)	78 (97.5)	86 (97.7)	1.000
Tendon friction rubs, no. (%)	36 (45.0)	41 (46.6)	0.836

^*defined as the first of either symptom, Raynaud’s or non-Raynaud’s

^**ILD as suggested by an FVC ever <70% of predicted

^{^}Pulmonary hypertension as suggested by RVSP>=45 mmHg ever

^{^^}Severe Raynaud’s and severe GI disease defined by organ specific severity score values ≥ 2 at any time throughout the disease course.

mRSS=modified Rodnan skin score, FVC=forced vital capacity, DLCO=diffusing capacity, RVSP=right ventricular systolic pressure, GI=gastrointestinal

Discovery and identification of anti-RPA194 antibodies in scleroderma patients with RPC155 antibodies, using cancer status as a filter.

To address whether the immune response differed in anti-RPC155-positive scleroderma patients with and without cancer, we initially used an immunoprecipitation approach on a subset of 35 sera randomly selected from the cohorts described above. That is, 17 anti-RPC155 scleroderma patients with cancer and 18 anti-RPC155 scleroderma patients without cancer were used for discovery. These 35 sera were used to immunoprecipitate proteins from radiolabeled HeLa cell lysates, and the profiles were compared (representative examples are shown in Fig 1A, left panel). A subset of the discovery sera was also tested by immunoprecipitation using radiolabeled 624-Melanocyte lysates (Fig 1A, right panel and Fig. 1B), and similar data were obtained.

Figure 1. Discovery and identification of anti-RPA194 antibodies in scleroderma patients with anti-RPC155 antibodies.

(A & B): Immunoprecipitations (IPs) were performed using radiolabeled lysates made from HeLa cells (left) or 624 melanocytes (right). The anti-RPC155 antibody status of all sera in the study was assayed as described in the Methods section. For the 35 sera in the discovery cohort, this was subsequently also confirmed by a third assay (IVTT IP; representative examples are shown in panel C). The sera used for the IPs shown are both from scleroderma patients with antibodies against RPC155. Serum FW-1170 is from a patient without cancer; serum FW-1088 is from a patient with cancer. IPs performed with serum from a healthy donor are also shown. Migration of molecular weight marker standards are shown on the right. (B) is an enlargement of the data shown in the upper section of the 624 melanocyte lysate IPs in A. (C): IPs were performed with the same sera used in A; input material for the IPs was ³⁵S-Methionine-labeled RPA194 (upper) or RPC155 (lower) generated by in vitro transcription and translation (IVTT), as described in the Methods section. The input material (no IP) is shown on the left of each set.

In 7/18 (38.9%) patients without cancer, a band ~194 kDa was detected; this was only seen in 1/16 (6.3%) of the group with cancer (see representative examples in Fig 1A). Additional bands (possibly other protein components of the polymerase complexes) were variably immunoprecipitated by some of the sera (110, 40, 34, 28 kDa). This finding is consistent with elegant work published by Kuwana et al (23), who used an immunoprecipitation approach with radiolabeled cell extracts to show that anti-RNA polymerase antibodies in scleroderma patient sera recognize multiple subunits of the RNA polymerases. Of note, the additional immunoprecipitated bands we detected were observed with similar frequencies in both the groups with and without cancer. Because we were seeking specificities that could distinguish these groups, we focused on the 194 kDa specificity enriched in patients without cancer.

Since (i) RNA polymerases I and III share some common subunits which can enable intermolecular spreading, (ii) RNA polymerase I is a known scleroderma autoantigen (13, 24), and (iii) the large subunit of RNA polymerase I (RPA194) is 194 kDa, we tested whether the 194 kDa band might be RPA194. We generated radiolabeled RPA194 by in vitro transcription/translation from cDNA encoding full length RPA194, and performed immunoprecipitations with the putative anti-RPA194 positive serum (FW-1170) and a negative serum (FW-1088) (Fig. 1C, upper panel). RPA194 pulldown was observed only with the FW-1170 patient serum, confirming that RPA194 antibodies were present in FW-1170, and absent in FW-1088.

Anti-RPA194 antibodies are enriched in anti-RPC155 positive scleroderma patients without cancer

Using this assay, anti-RPA194 antibodies were then determined in the full study cohort of 168 anti-RPC155 positive sera. Anti-RPA194 antibodies were significantly more common in the group without cancer (16/88 patients or 18.2%) than the group with cancer (3/80 patients or 3.8%; p=0.003). Even when secondary analyses were performed defining anti-RPC155 positivity by a more stringent cutoff of ≥ 40 units, the finding that anti-RPA194 antibodies were more common in the group without cancer remained unchanged (18.4% in the group without cancer versus 4.05% in the group with cancer; p=0.008). Of the 3 scleroderma cancer patients with both anti-RPC155 and anti-RPA194, one patient had a basal cell skin cancer 25 years prior to scleroderma onset, one had prostate cancer detected 1.25 years prior to scleroderma onset, and one had a uterine cancer diagnosed 5 years after scleroderma onset. In the 18 sera from anti-RPC155 positive scleroderma patients without cancer that were used for discovery (see above), 7/18 were confirmed to have anti-RPA194 antibodies using the IVTT immunoprecipitation assay, consistent with the initial screening frequency. Thirty-four sera from healthy controls were also tested with IVTT immunoprecipitation assay; none immunoprecipitated IVTT RPA194.

We next assessed whether the clinical phenotype differed between anti-RPC155 patients with and without anti-RPA194 antibodies (Table 2). There were no differences in age at scleroderma onset, race, smoking status, or the frequency of ILD, pulmonary hypertension, myopathy, severe Raynaud’s phenomenon, calcinosis, telangiectasia or tendon friction rubs. While not statistically significantly different, anti-RPA194-positive patients were more likely to be women (94.7% vs. 79.9%, p=0.205), have diffuse scleroderma (89.5% vs 75.2%, p=0.249), and have higher baseline (23.2±14.2 vs. 18.1±13.2, p=0.1358) and maximum (27.5±15.0 vs. 21.8±14.0, p=0.0990) modified Rodnan skin scores than anti-RPA194-negative patients. Anti-RPA194-positive patients were also less likely to have a history of renal crisis (0% vs 14.1%, p=0.134), and were significantly less likely to have severe GI disease (26.3% vs 51.0%, p=0.043). Interestingly, this group was significantly more likely to have a nucleolar pattern on ANA (44.4% vs. 12.5%, p=0.001) than anti-RPA194-negative patients, although ANA titers were not different between groups. This suggests that detection of a nucleolar staining pattern is more likely to capture the anti-RPA194 antibody-positive group. However, the clinical utility of these findings will be greatly enhanced by developing a commercial assay that specifically detects anti-RPA194 antibodies.

Table 2.

Clinical characteristics of anti-RPA194 positive and negative patients, among anti-RPC155 positive patients.

Variable	Anti-RPA194 positive (N=19)	Anti-RPA194 negative (N=149)	p-value
Age at scleroderma onset* (years), mean (SD)	44.5 (15.6)	48.8 (13.9)	0.2112
Age at first non-Raynaud’s symptom (years), mean (SD)	50.3 (11.3), N=18	51.1 (12.5), N=148	0.7916
Disease duration at first visit (years), mean (SD)	6.8 (10.1)	6.2 (9.1)	0.7858
Male sex, no. (%)	1 (5.3)	30 (20.1)	0.205
Race, no. (%)		N=148	1.000
White	19 (100)	140 (94.6)
Black	0 (0)	6 (4.1)
Asian	0 (0)	2 (1.4)
Subtype, no. (%)			0.249
Limited	2 (10.5)	37 (24.8)
Diffuse	17 (89.5)	112 (75.2)
Ever smoked, no. (%)	10 (52.6)	73 (49.0)	0.765
mRSS at first visit to Center, mean (SD)	23.2 (14.2)	18.3 (13.2), N=146	0.1358
Maximum ever mRSS, mean (SD)	27.5 (15.0)	21.8 (14.0), N=148	0.0990
Renal crisis, no. (%)	0 (0)	21 (14.1)	0.134
Myopathy, no. (%)	3 (15.8)	24 (16.1)	1.000
ILD (FVC ever <70%)**, no. (%)	7 (38.9), N=18	64 (45.1), N=142	0.619
Baseline pulmonary function, mean (SD)
FVC (% predicted)	85.2 (16.1), N=16	83.9 (16.2), N=136	0.7740
DLCO (% predicted)	86.3 (22.6), N=14	82.6 (23.4), N=121	0.5767
Pulmonary hypertension^, no. (%)	6 (33.3), N=18	43 (30.3), N=142	0.791
Baseline RVSP (mmHg), mean (SD)	33.6 (9.0), N=10	32.5 (9.8), N=86	0.7422
Baseline ejection fraction (%), mean (SD)	60.3 (9.1), N=17	61.6 (7.0), N=129	0.4836
Severe Raynaud’s^^, no. (%)	12 (63.2)	81 (54.4)	0.468
Severe GI disease^^, no. (%)	5 (26.3)	76 (51.0)	0.043
Calcinosis, no. (%)	11 (57.9)	67 (45.0)	0.287
Telangiectasia, no. (%)	18 (94.7)	146 (98.0)	0.384
Tendon friction rubs, no. (%)	10 (52.6)	67 (45.0)	0.528

^*defined as the first of either symptom, Raynaud’s or non-Raynaud’s

^**ILD as suggested by an FVC ever <70% of predicted

^{^}Pulmonary hypertension as suggested by RVSP>=45 mmHg ever

^{^^}Severe Raynaud’s and severe GI disease defined by organ specific severity score values ≥ 2 at any time throughout the disease course.

mRSS=modified Rodnan skin score, FVC=forced vital capacity, DLCO=diffusing capacity, RVSP=right ventricular systolic pressure, GI=gastrointestinal

Discussion

The existence of antibodies against the RNA polymerases has been recognized for many years in scleroderma patients (23). The recent observation that autoantibodies to RPC155 in scleroderma mark a group of patients with a higher risk of cancer (SIR 2.84, 95% CI 1.89–4.10; OR ranging from 3.85–5.83) occurring within a short interval of scleroderma onset has been confirmed in multiple different cohorts (2–8). Mechanistic studies in scleroderma patients with cancer have demonstrated that somatic mutations at the POLR3A locus are present in the cancers of some anti-RPC155 patients, in whom mutation-specific and cross-reactive T cell immune responses develop (9). These findings strongly suggest that somatic mutations in RPC155 in scleroderma patients’ cancers initiate the anti-RPC155 immune response, which spreads to the wild type protein, exerting both anti-cancer and autoimmune effects. The finding that only 15–20% of patients with anti-RPC155 ever manifest a cancer suggests several possibilities. In a first scenario, different mechanisms may underlie the development of anti-RPC155 immune responses and scleroderma phenotype in those patients with and without cancer. Alternately, it is possible that similar mechanisms (cancer with changes in RPC155 structure due to mutation) drive this immune response and scleroderma phenotype in all anti-RPC155-positive patients, but that the groups with and without cancer represent differences in the efficacy of the anti-cancer effect of the immune response in the two groups, perhaps marked by additional immune responses in patients without cancer, or differential sensitivity of the cancer to the induced immune response. In this study, we sought evidence for the possibility that additional immune specificities are found in anti-RPC155-positive scleroderma patients without a detectable cancer.

The initial screen from radiolabeled cells demonstrated striking enrichment of a 194 kDa protein immunoprecipitated by antibodies from anti-RPC155 patients without cancer. The migration of this protein, coupled with prior descriptions of antibodies to RNA polymerase I in scleroderma (13), and recent evidence that inhibition of RNA polymerase I activity has potent anti-cancer effects, rapidly led us to confirm that RPA194 was the undefined band recognized by the index sera without cancer. Subsequent analysis demonstrated that RPA194 antibodies were found in 18.2% of anti-RPC155-positive patients without cancer, while they were infrequent (3.8%) in anti-RPC155-positive patients with cancer. It is of interest that one of the patients with cancer and antibodies to RPA194 had a basal cell cancer that occurred 25 years prior to scleroderma, suggesting that they might be unrelated. Regardless, anti-RPA194 autoantibodies are enriched in anti-RPC155-positive patients without cancer. That immune responses to RPA194 are associated with the subgroup of anti-RPC155 positive scleroderma patients without cancer is particularly interesting as recent studies have shown that a small molecule (BMH21), which inhibits RNA polymerase I activity, has broad, potent and selective antitumor activities across multiple cancer cell lines in vitro, and represses tumor growth in vivo in mouse models (14). It is of interest that six polymerase subunits are shared amongst RNA polymerases I and III, providing important opportunity for intermolecular spreading. Understanding the targeting of these shared subunits in patients with and without cancer might shed additional mechanistic insights into the origin of the broader immune responses in scleroderma patients without cancer.

Taken together, the observations (i) that many cancers have increased RNA polymerase I activity, (ii) that inhibition of this activity has anti-cancer effects, and (iii) that immune responses to the catalytic subunit of RNA polymerase I (RPA194) are associated with a lower frequency of cancer in the anti-RPC155-positive scleroderma group, raise the possibility that the RPA194 immune response might play a mechanistic role in the immune-mediated control of cancer in scleroderma. Like most autoantigens in the systemic rheumatic diseases, RPA194 is intracellular, raising important questions about how such an effect might be mediated. Several scenarios are possible. Firstly, since RNA polymerase I expression and activity is increased in many cancers, it is possible that an anti-cancer effect is exerted by cell-mediated immunity, which is selectively directed against cancer cells expressing high levels of RPA194. Secondly, it is possible that autoantibodies recognizing intracellular antigens may have direct anti-cancer effects (25). For example, a series of recent studies showed that anti-DNA antibodies induced death of cancer cells in vitro and in vivo, particularly in cancer cell lines with defects in various DNA repair pathways (25). It is similarly possible that autoantibodies recognizing RPA194 might inhibit RPA194 inside cells, thereby exerting anti-cancer effects. Thirdly, a different set of initiating events (other than cancer) might be responsible for initiating the combined immune responses against both RPC155 and RPA194. Confirming these findings in other scleroderma cohorts, and pursuing these mechanistic questions, remain important priorities.

If the association of anti-RPA194 antibodies with decreased cancer incidence in a subgroup of RPC155-positive scleroderma patients is confirmed in other studies, then detecting anti-RPA194 at diagnosis might have clinical utility in identifying the subset of anti-RPC155-positive patients who do not require extensive malignancy workup. The observation that only ~20% of anti-RPC155 positive patients without cancer have anti-RPA194 antibodies suggests that additional mechanisms may be associated with a lower cancer prevalence in these remaining patients. One possibility is that additional immune specificities (other than RPA194) occur in the context of RPC155 immune responses. Finding such specificities would further support the hypothesis that orthogonal immune responses targeting additional cellular machines are associated with decreased cancer frequencies in the high-risk anti-RPC155-positive group.

The subgroup of scleroderma patients with both RPC155 and RPA194 antibodies was not large enough to know with certainty whether they manifested a more severe form of scleroderma than those with anti-RPC155 alone, but the available data does not suggest this. Indeed, none of the anti-RPA194-positive patients manifested scleroderma renal crisis (present in 14% of the anti-RPC155-only group), and severe gastrointestinal disease was present significantly less frequently in the anti-RPC155/RPA194 subgroup than in the anti-RPC155-only subgroup (26% vs 51%). In future studies with larger numbers of RPA194 antibody positive patients, it will be important to define the phenotype and mortality in the anti-RPC155 groups with or without RPA194 antibodies.