Anti–U3 RNP autoantibodies in systemic sclerosis




To describe the classification, demographic and clinical features, and survival in anti–U3 RNP autoantibody–positive patients with systemic sclerosis (SSc).


Medical records of 108 anti–U3 RNP–positive and 2,471 anti–U3 RNP–negative SSc patients first evaluated during 1985–2003 were reviewed. Anti–U3 RNP antibody was detected by protein and RNA immunoprecipitation. Disease classification, demographic and clinical features, organ system involvement, and survival were compared between the 2 patient groups, by Student's t-test, chi-square analysis, and Mantel-Haenszel test.


The anti–U3 RNP–positive group had a higher proportion of African American patients (27% versus 5%; P < 0.001) and male patients (29% versus 19%; P = 0.021), and was younger at the time of first physician diagnosis (mean age 42.8 years versus 47.4 years; P = 0.001). The 2 groups had similar proportions of patients with diffuse cutaneous involvement (47% and 45% in those with and those without anti–U3 RNP, respectively). However, among patients with diffuse cutaneous involvement, the mean maximum modified Rodnan skin score was significantly lower in the anti–U3 RNP group (22.3 versus 27.9; P < 0.001). Skeletal muscle involvement was more frequent in anti–U3 RNP–positive patients (25% versus 14%; P = 0.002), as was “intrinsic” pulmonary arterial hypertension (PAH) (31% versus 13%; P < 0.001). The frequency of gastrointestinal involvement, cardiac involvement, pulmonary fibrosis, and “renal crisis” did not differ significantly between the 2 groups. Survival was worse in the anti–U3 RNP–positive group (hazard ratio 1.38 [95% confidence interval 1.05–1.82]). PAH was the most common known cause of death in patients with anti–U3 RNP (30%, versus 10% in the anti–U3 RNP–negative group; P < 0.001).


The present findings demonstrate that the frequencies of African American race and male sex are greater among SSc patients with anti–U3 RNP antibody than those without, and the former group is younger at SSc diagnosis. Anti–U3 RNP–positive patients have more frequent skeletal muscle involvement and PAH, the latter being the most common cause of death.

Systemic sclerosis (SSc; scleroderma) is a chronic disorder of connective tissue characterized by inflammation, fibrosis, and degenerative changes in the blood vessels, skin, synovium, skeletal muscle, and multiple internal organs. The clinical features, organ system involvement, natural history, and survival among patients with SSc are highly variable and largely depend on SSc clinical subtype and SSc-associated serum autoantibodies (1). To date, 9 SSc-associated serum autoantibodies have been identified: anticentromere (ACA), anti–topoisomerase I (anti–topo I; anti–Scl-70), anti–RNA polymerase III (anti–RNAP III), anti–U1 RNP anti-Ku, anti–U11/U12 RNP (1–3), and 3 predominantly antinucleolar antibodies (anti–U3 RNP, anti–PM-Scl, and anti-Th/To). Each SSc patient typically is positive for only 1 of these SSc-related autoantibodies (4, 5). The antibodies are recognized to correlate with disease manifestations and prognosis.

Antinucleolar antibodies have been detected in 10–15% of SSc patients in different studies (6–10), and anti–U3 RNP antibody (11) has been previously reported to occur in 5–8% of patients (7, 9, 12–14). Clinical features and prognosis associated with anti–U3 RNP antibodies are not well characterized. In 1992, investigators in our group described the clinical characteristics of 24 anti–U3 RNP–positive SSc patients (13). Two limitations of that study and other prior studies on anti–U3 RNP antibody are small sample size and relatively brief period of followup. In this report we describe the demographic, clinical, and laboratory features, disease classification, natural history, and survival in 108 anti–U3 RNP antibody–positive patients who were first evaluated between 1985 and 2003.



All 2,579 patients from the Pittsburgh Scleroderma Databank with a clinical diagnosis of SSc who were first evaluated between 1985 and 2003 were included in this study. Eighty-one percent met the American College of Rheumatology (formerly, the American Rheumatism Association) classification criteria for SSc (15).

The final study group consisted of all anti–U3 RNP–positive SSc patients, and, as a comparison group, all SSc patients without anti–U3 RNP autoantibody. A standard baseline evaluation was performed on all patients at the first visit, and followup data were collected when available from our scleroderma clinic or from the patients' referring physicians.

Serologic testing.

Antinuclear antibodies (ANAs) were detected using a HEp-2 cell substrate and, if positive, the staining pattern was recorded (speckled, nucleolar, homogeneous, centromere, etc.). ACAs were identified by their characteristic staining pattern at a 1:40 dilution. Anti–topo I and anti–U1 RNP antibodies were tested for by Ouchterlony immunodiffusion (16). Sera negative for these latter 3 autoantibodies were examined by immunoprecipitation for 5 other SSc-associated antibodies (anti–RNAP III, anti–PM-Scl, anti-Th/To, anti–U3 RNP, and anti-Ku), as previously reported (17). In addition, all sera with a prominent nucleolar staining pattern by indirect immunofluorescence were further tested by immunoprecipitation. Sera were considered positive for anti–U3 RNP antibody if they immunoprecipitated a protein with an apparent molecular weight of 34 kd as well as an RNA with the same electrophoretic mobility as U3 small nucleolar RNA. Protein and RNA immunoprecipitation studies were performed as previously described by our group (13, 18).

Definitions for classification of scleroderma.

Diffuse cutaneous SSc (dcSSc) was defined as thickness of the skin proximal to the elbows or knees (upper arms, trunk, or thighs) at any time during the illness (19). Overlap syndrome included patients with SSc who also had evidence of another connective tissue disease (CTD), such as idiopathic inflammatory myopathy (myositis) or systemic lupus erythematosus, according to published diagnostic or classification criteria for these diseases (20, 21). Age at symptom onset was defined as age at development of the first symptom attributable to SSc.

Definitions of organ system involvement.

SSc organ involvement was considered to be present if predefined clinical features in the 8 organ systems listed below were observed during the course of the illness and were not attributable to other diseases.

Peripheral vessels.

Involvement of peripheral vessels was recorded based on the presence of any one of the following: Raynaud's phenomenon, digital pitting scars, digital tip ulcerations, digital gangrene, or abnormal nailfold capillaries.


Skin involvement was recorded if there was any thickening as defined according to the modified Rodnan skin scoring system (22).


Involvement of joints/tendons was recorded if any one of the following was present: joint swelling, carpal tunnel syndrome, one or more palpable tendon friction rubs, or finger joint contractures with finger-to-palm distance in full flexion >1.9 cm.

Skeletal muscle.

Skeletal muscle involvement was recorded if there was proximal muscle weakness found on physical examination, and any one of the following was also present: elevated serum creatine kinase (CK) level, myopathic changes on electromyography, or evidence of myositis on muscle biopsy.

Gastrointestinal (GI) tract.

GI tract involvement was recorded if any one of the following was present: distal esophageal hypomotility (by esophagraphy or manometry), esophageal stricture, radiographic evidence of hypomotility of the small intestine, wide-mouth colonic sacculations, malabsorption syndrome, or death due to a CTD-related GI cause.


Pulmonary involvement was recorded if either interstitial lung disease (ILD) or “intrinsic” pulmonary arterial hypertension (PAH) was present. ILD was defined as the presence of either of the following: restrictive lung disease (forced vital capacity in 1 second [FVC] <70% of predicted plus forced expiratory volume in 1 second/FVC >80% of predicted) or pulmonary fibrosis seen on chest radiography or high-resolution computed tomography scan or listed as the cause of death. PAH was defined as the presence of any one of the following: mean PA pressure of >25 mm Hg detected on cardiac catheterization, PA systolic pressure of >40 mm Hg detected on either right heart catheterization or echocardiogram, or intrinsic PAH listed as the cause of death, in the absence of significant ILD and/or cardiac disease that might lead to secondary PAH.


Cardiac involvement was recorded if any one of the following was present: left-sided congestive heart failure (clinical, or estimated left ventricular ejection fraction <45%), pericarditis (pericardial pain plus pericardial friction rub, pericardial effusion, or electrocardiographic evidence of pericarditis), arrhythmia requiring treatment, complete heart block, or CTD-related cardiac cause of death.


Renal involvement was recorded if there was either clinical evidence of scleroderma renal crisis, defined as the abrupt onset of accelerated arterial hypertension or rapidly progressive oliguric renal failure, or CTD-related renal failure listed as the cause of death.

The percentage of patients with involvement of each individual organ system was determined by the number of patients with documented organ involvement divided by the number of patients who had undergone objective evaluation of the organ. Death and causes of death were determined from information obtained from autopsy reports, medical records, or the patient's family. We used the scleroderma-specific method described by Hesselstrand et al for assigning causes of death (23), with final decisions made by negotiated consensus of 2 of the authors (CVO and TAM).

Statistical analysis.

Student's t-test was used to detect significant differences between distributions of continuous data (means). Chi-square analysis was used to determine significant differences between sets of categorical data, with Fisher's exact test being used when appropriate. Survival was measured from the time of first physician diagnosis of SSc, and survival curves were compared using Kaplan-Meier and Mantel-Haenszel tests. Cox proportional hazard ratios were calculated to compare the survival of the 2 groups after controlling for age at diagnosis and sex.


Study group.

A total of 2,579 SSc patients were studied. The study group (anti–U3 RNP group) comprised 108 patients, and the comparison group (non–anti–U3 RNP group) comprised 2,471 patients. Eighteen anti–U3 RNP–positive patients were also positive for an additional SSc-associated autoantibody, including 6 patients with anti-Th/To, 3 with anti–U1 RNP, 3 with ACA, 2 with anti-Ku, 2 with anti–RNAP III, 1 with anti–topo I, and 1 with both anti–U1 RNP and anti–topo I.

Disease classification.

Disease classification data are shown in Table 1. Almost all anti–U3 RNP–positive patients (92%) had SSc alone. Only 9 (8%) had an overlap syndrome. This distribution was similar to that in the non–U3 RNP group, in which 217 (9%) had overlap syndrome. The anti–U3 RNP group had an equal distribution of patients with dcSSc and patients with limited cutaneous SSc (47% and 44%, respectively). However, among patients with SSc in overlap, those who were positive for anti–U3 RNP had predominantly dcSSc (89%), whereas the anti–U3 RNP–negative patients with overlap had dcSSc only infrequently (28%) (P < 0.001). Eight of the 9 anti–U3 RNP–positive patients with SSc in overlap had myositis, and 1 had lupus.

Table 1. Clinical classification of the SSc patients with and those without anti–U3 RNP*
ClassificationWith anti–U3 RNP (n = 108)Without anti–U3 RNP (n = 2,471)
  • *

    Values are the number (%). CTD = connective tissue disease.

  • P < 0.001 versus patients with diffuse cutaneous systemic sclerosis (SSc) in overlap without anti–U3 RNP.

Diffuse cutaneous51 (47)1,103 (45)
Limited cutaneous48 (44)1,151 (47)
SSc in overlap9 (8)217 (9)
 Diffuse cutaneous8 (89)61 (28)
 Limited cutaneous1 (11)156 (72)
 Overlap with myositis8119
 Overlap with lupus156
 Overlap with other CTD042

Demographic features.

Demographic characteristics of the patients are shown in Table 2. The anti–U3 RNP group had a significantly greater proportion of African Americans compared with the non–anti–U3 RNP group (27% versus 5%; P < 0.001). This difference was found despite the fact that >90% of the population in Pittsburgh and western Pennsylvania is white. The anti–U3 RNP–positive group also had a significantly greater proportion of men (29% versus 19%; P = 0.021). As a group, patients with anti–U3 RNP were significantly younger at the time of first physician diagnosis of SSc (mean 42.8 years versus 47.4 years; P = 0.001) as well as at the time of first presentation at our medical center (45.2 years versus 50.2 years; P < 0.001). Patients with anti–U3 RNP were also younger at disease onset, but this difference did not reach statistical significance (P = 0.056). Anti–U3 RNP–positive patients had a significantly shorter mean disease duration at the time of SSc diagnosis (2.9 years versus 4.7 years; P = 0.015) as well as at first visit to our center (5.3 years versus 7.6 years; P = 0.017). The duration of followup after the first visit did not differ significantly between the 2 groups (mean 5.5 years and 6.0 years in the anti–U3 RNP and non–anti–U3 RNP groups, respectively). Complete followup data (including deaths) were obtained on 78% of the cohort (2,011 of 2,579 patients) through 2004.

Table 2. Demographic and clinical characteristics of the SSc patients with and those without anti–U3 RNP*
CharacteristicWith anti–U3 RNP (n = 108)Without anti–U3 RNP (n = 2,471)P
  • *

    SSc = systemic sclerosis; NS = not significant; UPMC = University of Pittsburgh Medical Center.

Sex, no. (%) male31 (29)475 (19)0.021
Race, no. (%)   
 White76 (70)2,254 (91)<0.001
 African American29 (27)123 (5)<0.001
 Other3 (3)83 (3)NS
Age, mean ± SD years   
 At symptom onset39.8 ± 14.842.7 ± 15.20.056
 At SSc diagnosis42.8 ± 14.947.4 ± 14.60.001
 At first UPMC evaluation45.2 ± 15.650.2 ± 14.3<0.001
Disease duration, years   
 Symptom onset to SSc diagnosis   
  Mean ± SD2.9 ± 5.04.7 ± 7.50.015
 Symptom onset to first UPMC evaluation   
  Mean ± SD5.3 ± 7.67.6 ± 9.40.017

Organ system involvement.

Table 3 shows the findings with regard to organ system involvement. There were no significant differences in peripheral vascular, joint/tendon, GI tract, lung (pulmonary fibrosis/ILD), heart, or kidney involvement between the 2 groups. Skeletal muscle involvement was noted significantly more frequently in patients with anti–U3 RNP (25% versus 14%; P = 0.002). Among patients meeting criteria for muscle involvement, the proportion with elevated serum CK levels was lower among those with anti–U3 RNP (59% versus 77%; P = 0.043). We reviewed muscle biopsy specimens from 9 patients with anti–U3 RNP, and detected no distinctive microscopic findings. The pathologic changes were consistent with the pattern typically seen in myositis associated with CTD, but there were fewer inflammatory cells noted in the samples from anti–U3 RNP–positive SSc patients. This was a nonblinded, uncontrolled observation and does not imply a unique type of muscle injury in patients with anti–U3 RNP.

Table 3. Organ system involvement in the SSc patients with and those without anti–U3 RNP*
Organ systemWith anti–U3 RNP (n = 108)Without anti–U3 RNP (n = 2,471)P
  • *

    Except where indicated otherwise, values are the number (%) of patients in whom objective testing was performed; denominators indicate the number tested. SSc = systemic sclerosis; NS = not significant; GI = gastrointestinal; ILD = interstitial lung disease; PAH = “intrinsic” pulmonary arterial hypertension; dcSSc = diffuse cutaneous SSc.

  • Including patients with SSc in overlap.

Peripheral vessels107/108 (99)2,415/2,471 (98)NS
Joints/tendons89/108 (82)1,924/2,471 (78)NS
Skeletal muscle27/108 (25)351/2,471 (14)0.002
GI tract68/82 (83)1,475/1,883 (78)NS
Lung65/106 (61)1,409/2,446 (58)NS
Fibrosis/ILD35/97 (36)908/2,243 (40)NS
PAH27/86 (31)239/1,823 (13)<0.001
Heart22/94 (23)381/1,936 (20)NS
Kidney7/108 (6)252/2,471 (10)NS
 Skin thickening103/108 (95)2,320/2,471 (94)NS
 Maximum skin score, mean ± SD   
  All patients15.1 ± 12.015.7 ± 14.2NS
  Patients with dcSSc22.3 ± 9.527.9 ± 12.0<0.001

Four categories of pulmonary involvement were detected, as follows: 1) PAH alone (“intrinsic” PAH), with no ILD; 2) “intrinsic” PAH with ILD, i.e., moderate-to-severe PAH, not clinically accounted for by a minor component of ILD; 3) ILD alone, with no secondary PAH; and 4) ILD with secondary PAH. Although pulmonary involvement overall was not significantly different in the 2 groups, the anti–U3 RNP group had a significantly higher frequency of “intrinsic” PAH (31% versus 13%; P < 0.001).

Five anti–U3 RNP–positive patients had no skin thickening (SSc sine scleroderma). The mean maximum modified Rodnan skin score was similar in the 2 groups (15.1 versus 15.7; P not significant.), but among patients with dcSSc, the skin score was significantly lower in the anti–U3 RNP group (22.3 versus 27.9; P < 0.001).

Antibody findings.

ANA patterns differed significantly between the 2 groups. Seventy-six of the 108 anti–U3 RNP–positive patients had nucleolar-only ANA staining (70%), compared with 26 of the 2,196 tested in the non–anti–U3 RNP group (1%) (P < 0.001).

Among the 2,471 anti–U3 RNP–negative patients, the distribution of serum autoantibodies was similar to that which we have found in a consecutive series of new patients evaluated during 1986–1989 (24) and in the entire cohort between 1980 and 1995 (25). Among all 2,579 patients included in the present study, 1,919 (74%) had complete serologic testing performed, with 1 or more of the 9 SSc-associated autoantibodies detected in all 1,919. The frequency distribution of autoantibodies in these patients was as follows: ACA 25%, anti–topo I 25%, anti–RNAP III 25%, anti-Th/To 7%, anti–U3 RNP 6%, anti–U1 RNP 5%, anti–PM-Scl 4%, anti–U11/U12 RNP 2%, anti-Ku 1%, and >1 of the above 3%.


The anti–U3 RNP–positive group had somewhat reduced cumulative survival from the time of first physician diagnosis of SSc (Figure 1). The 5-year and 10-year unadjusted cumulative survival rates among anti–U3 RNP–positive patients were 75% and 56%, respectively, whereas in the non–anti–U3 RNP group, these rates were 81% and 66%, respectively (P = 0.102 at 10 years). After adjustment for age at diagnosis and sex, the anti–U3 RNP group had significantly worse survival than the non–anti–U3 RNP group, with a hazard ratio of 1.38 (95% confidence interval 1.05–1.82) (P = 0.019). “Intrinsic” PAH was the most common cause of death among patients with anti–U3 RNP, accounting for 30% of all deaths (P < 0.001 versus anti–U3 RNP–negative patients) (Table 4). Pulmonary fibrosis was the most common cause of death in the non–anti–U3 RNP group (12%) (P = 0.067).

Figure 1.

Survival in systemic sclerosis (SSc) patients with and those without anti–U3 RNP autoantibody.

Table 4. Causes of death in the SSc patients with and those without anti–U3 RNP*
Cause of deathWith anti–U3 RNP (n = 50)Without anti–U3 RNP (n = 1,072)P
  • *

    Values are the number (%). See Table 3 for definitions.

  • Neared significance (P = 0.067).

SSc-related26 (52)441 (41)0.034
 “Intrinsic” PAH15 (30)105 (10)<0.001
 Pulmonary fibrosis1 (2)121 (12)NS
 Cardiac4 (8)55 (5)NS
 Renal0 (0)69 (6)NS
 GI3 (6)45 (4)NS
 Other3 (6)46 (4)NS
Non–SSc-related10 (20)270 (25)NS
Unknown14 (28)361 (34)NS


The autoantibody profile in patients with SSc is particularly important, because SSc-related autoantibodies are associated with clinical features of the disease, natural history, and survival. These autoantibodies help to identify organ systems that are at risk for involvement.

Human U3 RNP consists of the U3 small nucleolar RNA and at least 6 protein subunits. One of the subunits, fibrillarin (34-kd basic protein), is considered to be the main antigenic determinant (26). Anti–U3 RNP antibody is directed against this subunit (26).

A predominantly nucleolar pattern of ANAs is seen with anti–U3 RNP, anti–PM-Scl, and anti-Th/To antibodies (27, 28). Nucleolar staining has been observed in 8–48% of SSc patient sera (6–10), 12–52% of which have been positive for anti–U3 RNP antibodies (6, 9, 12, 13, 27, 29). Anti–U3 RNP antibodies have been detected in 5–8% of all SSc patients (7, 9, 12–14). Positivity for these antibodies should be suspected in SSc patients with a nucleolar-only staining pattern.

Our group first described the high specificity of anti–U3 RNP antibody for SSc, its frequency in SSc, and its association with SSc in a small study population (24 patients) (13). Anti–U3 RNP was significantly associated with African American race, skeletal muscle disease, and “intrinsic” PAH in SSc. In the present study, there is overlap with the earlier investigation, because 23 of the 108 anti–U3 RNP antibody–positive patients described herein were included in the previous study (13). The same 3 significant associations were found in the 23 pre-1992 and 85 post-1992 patients, and thus, inclusion of the earlier group of patients in the present study cohort is supplementary and does not alter the results or conclusions.

Arnett and coworkers studied 27 anti–U3 RNP–positive scleroderma patients and found that patients with this antibody were more likely to be of African American ethnicity, more likely to be male, and had significantly increased frequencies of GI tract, cardiac, and renal involvement, and pulmonary fibrosis (12). Tormey and colleagues described 42 anti–U3 RNP–positive scleroderma patients (29). In their study, African Americans were disproportionately affected, and diffuse cutaneous involvement was more common than limited skin thickening. Myositis and internal organ involvement (GI tract, pulmonary hypertension, cardiac and renal disease) were also more frequent. Yang et al reported on 31 anti–U3 RNP–positive SSc patients of white German ethnicity (30), noting associations with diffuse skin changes and both esophageal and lung involvement.

Results of our larger study confirm the sex, race, skeletal muscle, and “intrinsic” PAH associations of anti–U3 RNP antibody in SSc, but do not provide evidence of association with involvement of other internal organs. Among patients with dcSSc, the higher mean maximum skin score in the non–U3 RNP group was attributable not to sex or ethnicity, but rather to a high proportion (64%) of patients in the non–U3 RNP group being positive for anti–RNAP III or anti–topo I; the mean maximum skin score in anti–U3 RNP–negative patients who were positive for anti–RNAP III or anti–topo I was 28.8, significantly higher than this score in anti–U3 RNP–positive patients (mean 22.3) (P < 0.001). Although there were no distinct clinical or pathologic differences in skeletal muscle findings in anti–U3 RNP–positive SSc patients as compared with patients with idiopathic inflammatory myopathy, CK elevation was less frequent and inflammation seen on muscle biopsy was minimal in anti–U3 RNP–positive SSc patients with myopathy. Survival analysis revealed reduced survival in the anti–U3 RNP–positive group. Thus anti–U3 RNP antibody in patients with SSc is associated with relatively poor prognosis.

This study has several limitations. First, its design was retrospective, utilizing a patient population from a single referral center. Second, patients positive for >1 SSc-associated autoantibody were included. Third, because sera from some patients with anti–topo I antibody show weak nucleolar staining on indirect immunofluorescence, a few patients with low-titer U3 RNP in combination with anti–topo I may not have been identified. Fourth, followup data (including deaths) were available on only 78% of cases as of 2004. Fifth, our definition of cardiac involvement did not include diastolic dysfunction, which might have led to underrepresentation of cardiac disease in both the anti–U3 RNP and the non–anti–U3 RNP groups. Sixth, the distinction between primary and secondary PAH when a patient has both PAH and ILD is often difficult and is dependent on clinical judgment.

In conclusion, we have shown in this study that the presence of anti–U3 RNP autoantibodies in SSc identifies a well-defined clinical subset of the disease. This antibody should be considered in African American and/or male SSc patients who have positive ANA with pure nucleolar staining. SSc patients with anti–U3 RNP are younger at first physician diagnosis of SSc and have more rapid progression of disease, and therefore they are diagnosed earlier and are referred to a medical center earlier in their disease course. Anti–U3 RNP–positive SSc patients with diffuse cutaneous involvement have less severe skin thickening than anti–U3 RNP–negative dcSSc patients. Patients with anti–U3 RNP have a significantly greater frequency of skeletal muscle involvement and “intrinsic” PAH; the latter is the most common cause of death. Detection of disease-associated autoantibodies in SSc is important for identifying individuals at risk for developing specific types of internal organ involvement, some of which may influence survival, and the present results demonstrate that adjusted cumulative survival after first physician diagnosis of SSc is reduced in anti–U3 RNP–positive patients.


Dr. Medsger 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 design. Medsger.

Acquisition of data. Aggarwal, Lucas, Fertig, Oddis, Medsger.

Analysis and interpretation of data. Aggarwal, Fertig, Oddis, Medsger.

Manuscript preparation. Aggarwal, Medsger.

Statistical analysis. Aggarwal, Lucas, Medsger.