Systemic sclerosis (scleroderma or SSc) is a heterogeneous disorder characterized by autoantibody subsets, which in turn have their own clinical associations. Much controversy resides in whether these autoantibodies contribute directly to the pathology seen in SSc or whether they are merely epiphenomena of the underlying disease process. Nevertheless, various autoantibodies found in patients with SSc carry significant value in diagnosis and in predicting clinical outcomes (Fig. 1). The autoantibodies classically associated with SSc include anti-centromere antibodies (ACA) and anti-Scl-70 (otherwise known as anti-topoisomerase I or anti-topo I). In addition to these is the less commonly occurring anti-nucleolar antibody (ANoA) system, which comprises a mutually exclusive heterogeneous group of autoantibodies that produce nucleolar staining by indirect immunofluorescence (IIF) on cells from a variety of species 1. The most widely recognized of these include anti-PM-Scl 2, antifibrillarin/anti-U3-ribonucleoprotein (AFA) 3, anti-Th/To 4, and the anti-RNA-polymerase family (anti-RNAP), including anti-RNAP I 5, II 6, and III 7 (although anti-RNAP frequently do not produce nucleolar staining on IIF). In addition to these disease-specific antibodies, anti-Ku, anti-Ro, antiphospholipid antibodies (aPL), anti-Smith (anti-Sm), anti-U1-ribonucleoprotein (anti-U1-RNP), and other autoantibodies are also found in SSc, each with a degree of clinical significance.

The present review details the various autoantibodies associated with SSc, their frequency (including in different ethnic groups), clinical correlates, pathophysiology, and genetic associations.

Since the early 1960s it has been known that ANA are common in the sera of patients with SSc 89, reported in as many as 95% and as few as 75% of patients with SSc with an overall diagnostic sensitivity of 85% and specificity of 54% when tested by IIF as published in a recent meta-analysis 10. The presence of anti-Scl-70 and anti-U1-RNP antibodies in the sera yields a speckled appearance, whereas anti-Th/To, anti-AFA, and anti-PM-Scl give a nucleolar staining pattern. Anti-RNAP I antibodies yield a nucleolar staining, whereas those against RNAP II and III give a speckled appearance or no fluorescence 10. The specificity and sensitivity of ANA vary depending on the antigen substrate used for the assay. The use of HEp2 cells yields a better sensitivity for the detection of nuclear antigens present during cell division (for example centromere antigen) than the use of tissue sections of murine liver or kidney 10. ANA can also be measured by enzyme-linked immunosorbent assay (ELISA), a much less cumbersome technique now employed by many commercial laboratories. Although ANA by ELISA is appealing because the assay is automated, it often produces false positive results 10. In addition, ANA by ELISA can yield false negative results, especially in patients with ANoA, and should not be used in the diagnosis of SSc without corroborative IIF 10.

ACA were initially described in 1980 11 when HEp-2 cells were used as the substrate for the ANA. ACA had not been seen previously with the use of IIF on tissue substrates such as mouse liver, because the tissues in question undergo cell division much less commonly. ACA have been most typically determined by their characteristic staining pattern on immunofluorescence, giving rise to a speckled appearance on HEp-2 cells 11. Subsequently was shown that SSc patients with ACA produce autoantibodies recognized by immunoblotting (IB), which react against six different centromeric proteins 121314151617181920. However, these distinctions have not been shown to have clinical relevance. So far, six centromeric nucleoproteins are known to be bound by sera from patients with SSc, designated CENP-A through CENP-F. Molecular analyses have shown that CENP-A is a 17 kDa centromere-specific histone H3-like protein 13. CENP-B is an 80 kDa haploid DNA-binding protein 141516. CENP-C is a 140 kDa chromosomal component required for kinetochore assembly 1617. CENP-D is a centromere antigen of unknown function, with a molecular mass of 50 kDa 18. CENP-E is a 312 kDa kinesin-like motor protein 19. CENP-F is a nuclear matrix protein that accumulates in the nuclear matrix during S phase, assembling onto kinetochores at late G2 during mitosis 1920.

All sera containing ACA react with CENP-B 21. A solid-phase ELISA has been established by using a cloned fusion protein of CENP-B as antigen 21222324.

The frequency of ACA in patients with SSc has been reported to be 20–30% overall, but it varies depending on the ethnicity of the SSc patient. When determined by IIF, ACA are rather specific for the diagnosis of SSc. They are rarely found in healthy patients 252627. They are likewise seldom found to be positive in patients with other connective tissue diseases (CTD) 25262829 and are rarely found in unaffected relatives 3031 (Table 1). When found in patients evaluated for Raynaud's phenomenon, ACA can predict the future development of SSc 3233343536.

The presence of ACA has long been strongly associated with CREST, a variant of SSc, defined by the presence of calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, and telangectasia 11. Finding ACA can also distinguish CREST serologically from patients with other variants of SSc 34353637, from patients with other CTD 26272834, and from patients with primary Raynaud's phenomenon 3438 (Table 2).

Although IIF remains the 'gold standard' in determining the presence of autoantibodies in SSc, many commercial laboratories have adopted ELISA testing to detect the presence of such autoantibodies. More recently, an ELISA using a cloned fusion protein of CENP-B as an antigen against which ACA are directed has shown no added sensitivity in the diagnosis of CREST compared with other patients with SSc, patients with primary Raynaud's phenomenon and patients with other CTD 2223. One must be very cautious of the specificity of this type of testing, although recent refinements have improved its performance 2324.

The presence of ACA generally carries a better prognosis than many other SSc-associated autoantibodies. In addition, ACA are associated with certain cutaneous and cardiopulmonary manifestations.

ACA are most often seen in the presence of limited cutaneous involvement 3940, and are also correlated with the presence of calcinosis 41 and ischemic digital loss in patients with SSc 42.

Pulmonary disease occurs in more than 70% of patients with SSc, second only to the esophagus in frequency of visceral involvement. The presence of ACA has been associated with a lower frequency of radiographic interstitial pulmonary fibrosis and a lesser severity thereof 37394043.

Lung involvement in SSc is defined by numerous measures, most commonly either by the presence of radiographic interstitial fibrosis, but also by abnormal forced vital capacity (FVC) or diffusion capacity for carbon monoxide (DL_CO) on pulmonary function tests (PFT). Although pulmonary involvement can also be defined by high-resolution computed tomography of the chest (HRCT) or by bronchoscopy with alveolar lavage, no studies have looked at the presence of autoantibodies in SSc-associated lung disease diagnosed by these means.

The lack of uniform criteria employed for the definition of restrictive lung disease (RLD) makes it difficult to compare studies of PFT. ACA have been found in association with a lower frequency of RLD in some studies 3744 but not others 4546. It is noteworthy that ACA-positive patients are more likely to have an abnormal DL_CO but a normal chest radiograph and FVC 47, underscoring that pulmonary hypertension, in the absence of hypoxia from pulmonary fibrosis, is a more common feature of ACA-positive patients with SSc.

Studies on two recent large cohorts of 1321 patients have found there is a lower mortality in ACA-positive patients than in those with positive anti-Scl-70 autoantibodies or AnoA 4148. Within 10 years of diagnosis, patients who are positive for ACA and negative for anti-Scl-70 or negative for AnoA had a significantly better survival 4148.

There seems to be no clinical utility in serially following ACA levels once the SSc patient has been found to be positive for ACA. ACA-positive patients remained positive in nearly all determinations, whether tested by IIF or IB 495051, and no correlation with extent of disease involvement in any organ system has been established with ACA levels as determined by ELISA 51.

The frequency of ACA in patients with SSc varies depending on their ethnicity. It is highest in Caucasians, where they are found in approximately a third of those, compared with a significantly lower frequency in Hispanic, African American and Thai patients with SSc 5253 (Table 3).

HLA-DRB1*01, HLA-DRB1*04, and HLA-DQB1*05 are associated with the presence of ACA 5354 and it seems likely that the generation of ACA is influenced by the presence of both human leukocyte antigen (HLA)-DRB1 and HLA-DQB1 alleles 55. In Caucasian and Japanese patients with SSc, the presence of at least one HLA-DQB1 allele not coding for leucine at position 26 of the first domain was found to be necessary but not sufficient to generate ACA 3155.

In 1979, a basic, heat-labile, chromatin-associated, nonhistone 70 kDa protein against which autoantibodies from patients with SSc are detected was described; it was isolated from rat liver nuclei with a combination of biologic and immunologic methods. This was initially designated Scl-70 56. Subsequent analyses revealed this response to be directed against topoisomerase I 57.

Anti-Scl-70 antibodies have classically been determined by double immunodiffusion techniques against calf or rabbit thymus extract, including Ouchterlony and counterimmunoelectrophoresis (CIE) 49. However, ascertainment of anti-Scl-70 antibodies by immunodiffusion (ID) usually requires 2–3 days and is difficult to automate. To circumvent this problem, IB and ELISA have been introduced more recently 245758. Topoisomerase I, initially purified from calf thymus glands, was used as antigen 59, although more recent studies have used recombinant topoisomerase I fusion proteins as the substrate for the ELISAs 60.

Anti-Scl-70 antibodies are found in 15–20% of patients with SSc by ID 3536. When determined by ID, anti-Scl-70 autoantibodies are virtually never seen in healthy controls 3034353658 or in non-affected relatives of patients with SSc 3031, nor in those patients with other CTD 2829 or primary Raynaud's phenomenon 28293536 (Table 4). As with ACA, the presence of anti-Scl-70 antibodies in a patient initially evaluated for Raynaud's phenomenon can confer an increase in the future development of SSc 33. The identification of anti-Scl-70 antibodies by IB or ELISA carries a similar specificity to that of ID, with overall higher sensitivity in earlier studies 343536525859. However, a recent article raises concern about the specificity of anti-Scl-70 ELISA assays for SSc, reporting positive results in sera of patients with systemic lupus erythematosus (SLE) that were correlated with disease activity, although this was not reproducible by ID 61.

ACA and anti-Scl-70 antibodies are virtually always mutually exclusive, being present in less than 0.5% of all patients with SSc simultaneously 3536414862.

Anti-Scl-70 antibodies are found in about 40% of patients with diffuse cutaneous systemic sclerosis (dcSSc) and less than 10% of patients with limited cutaneous systemic sclerosis (lcSSc) 3536.

The frequency of anti-Scl-70 antibodies in SSc with pulmonary fibrosis is about 45% 35. Anti-Scl-70 antibodies have been associated with both the presence and severity of radiographic interstitial pulmonary fibrosis 3947, whether determined by ID, IB, immunoprecipitation (IP), or ELISA 43. Anti-Scl-70 antibodies have also been found in correlation with RLD 63 and with a higher rate of decline in PFT 64, although this association is not universal 3746.

Anti-Scl-70 antibodies carry an increased SSc-related mortality rate, owing to the higher rate of right heart failure in association with RLD and pulmonary fibrosis 6566. Although no convincing association has been established for anti-Scl-70 and scleroderma renal crisis in other studies, such an association has been shown in one study of Japanese patients with SSc 67.

Repeated testing for anti-Scl-70 antibodies is unlikely to be useful in clinical practice; although several recent studies have examined serial determinations of anti-Scl-70 antibodies in patients with SSc, a clear role for this in patient care has not been established. Patients who are initially positive tend to remain so over time 4568, although in one recent study some patients with milder disease became anti-Scl-70-negative later in their disease course 69. Three studies have shown variations in anti-Scl-70 levels (determined by ELISA) with extent of disease involvement and even seronegative conversion with disease remission 686970, although this was not seen in others 51.

Unlike ACA, anti-Scl-70 antibody frequency has been shown not to vary depending on ethnic distribution. The presence of anti-Scl-70 antibodies is mediated by the presence of the genes for both HLA-DRB1 and DQB1, although primarily by the former in both Caucasian and Japanese patients with SSc 557172 (Table 3). HLA-DRB1*11 is associated with anti-Scl-70 antibodies in all ethnic groups, with HLA-DRB1*1101 and HLA-DRB1*1104 found in anti-Scl-70-positive Caucasians and African Americans and HLA-DRB1*1104 found in anti-Scl-70-positive Hispanics 5355. HLA-DPB1 alleles have also been implicated in the anti-Scl-70 antibody response in patients with SSc, specifically HLA-DPB1*1301 in Caucasians and DPB1*0901 in Japanese 55.

Since at least 1970 it has been recognized that the ANoA staining pattern of ANA was associated with SSc. ANoA actually comprises a group of mutually exclusive and heterogeneous autoantibodies that exhibit a typical nucleolar staining pattern of ANA by IIF on various cells (most often HEp2 cells) 1. They include anti-PM-Scl, anti-Th/To, anti-U3-RNP, AFA, and anti-RNAP I, anti-RNAP II, and anti-RNAP III. Anti-RNAP II and anti-RNAP III do not always yield a nucleolar staining pattern by IIF.

ANoA have been reported in 15–40% of patients with SSc 3973. Unlike with ACA and anti-Scl-70, the number of published studies on frequency of ANoA is relatively small. Nevertheless, specific ANoA are rarely seen in healthy controls 174 nor in healthy non-affected relatives of patients with SSc 75. ANoA are perhaps less specific for SSc than was previously thought, because they can be found in patients with other diseases such as SLE and Sjögren syndrome 7677.

Anti-PM-Scl antibodies were the first of the AnoA to be characterized in 1977. Originally discovered in patients with myositis/scleroderma overlap syndrome (PM/SSc) with the use of Ouchterlony ID techniques 78, anti-PM-Scl are usually identified by IP techniques today 77. Recently, the anti-PM-Scl autoantibodies have been shown to target six human exosome components that make up an RNA-processing complex, namely hRrp4p, hRrp40p, hRrp41p, hRrp42p, hRrp46p and hCs14p. hRrp4p and hRrp42p are most frequently targeted by the anti-PM-Scl antibody 79. The frequency of anti-PM-Scl varies between different ethnic groups, ranging from about 3% of patients with SSc and 8% of patients with myositis in Caucasians 178, to being absent from a large series of 275 Japanese patients with SSc 43.

Anti-PM-Scl antibodies have been associated with the PM/SSc overlap syndrome 8081. As many as 80% of patients with anti-PM-Scl antibodies will have a PM/SSc overlap syndrome 81. Anti-PM-Scl antibodies are found in as many as 50% of patients with PM/SSc overlap in comparison with less than 2% of patients with SSc in general 25. The PM/SSc-associated overlap syndrome is associated with a more benign and chronic course of disease and responds to a low to moderate dose of corticosteroids 80. Anti-PM-Scl antibodies predict limited cutaneous involvement when they are present 437582, although less reliably than ACA. This is likely to be secondary to the relative infrequency of anti-PM-Scl antibodies compared with ACA, reported in less than 15% patients with lcSSc 437582. Anti-PM-Scl antibodies are strongly linked to HLA-DQA1*0501 and HLA-DRB1*0301 54 (Table 3).

After the discovery of anti-PM-Scl antibodies, the refinement of IP assays using [³²P]orthophosphate or [³⁵S]methionine-labeled cell extracts allowed the recognition of another ANoA, anti-Th/To, in 1983 483. The Th/To antigen has recently been identified. Anti-Th/To antibodies are directed against components of the ribonuclease MRP and ribonuclease P complexes, more frequently Rpp25 and hPop1. The Th40 autoantigen is identical to Rpp38 protein 84. Anti-Th/To are present in about 2–5% of patients with SSc, being perhaps more common in the Japanese, and were previously virtually never seen in healthy control patients (less than 1%) 47. This no longer seems to be so, because anti-Th/To antibodies have also been described in patients with SLE, PM and primary Raynaud's phenomenon 7677. Anti-Th/To antibodies are also almost never seen in the presence of ACA 76. Like ACA, their presence most specifically predicts limited skin involvement 47757684, although routine testing is hardly useful as anti-Th/To autoantibodies are found so infrequently (Fig. 2).

Because of the low frequency of anti-Th/To antibodies, few studies have addressed their clinical significance. One report found that no particular clinical features were associated with anti-Th/To 47. In another, anti-Th/To-positive patients with lcSSc carried a worse prognosis 85 with a smaller frequency of joint involvement but a greater frequency of puffy fingers, small bowel involvement, hypothyroidism, and a greater risk for reduced survival at 10 years 85, succumbing primarily to pulmonary arterial hypertension. In still another study, anti-Th/To antibodies were described in those patients with SSc who developed the combination of scleroderma renal crisis and pulmonary hypertension without interstitial lung disease 86. In a study of sera from 172 patients with various CTD 77, anti-Th/To antibodies were increased in those patients with xerophthalmia, esophageal dysmotility and decreased DL_CO. The presence of anti-Th/To antibodies has been associated with HLA-DRB1*11 5587 (Table 3).

Anti-RNAP I, anti-RNAP II, and anti-RNAP III were not discovered until 1987 and 1993 788. Determined by IP techniques, these specific autoantibodies are found in about 20% of patients with SSc 58289 and, like other disease-specific autoantibodies, carry diagnostic and prognostic value. The specificity of anti-RNAP I and anti-RNAP III for SSc is similar and higher than that of anti-RNAP II, which can also be found in patients with SLE/SSc and overlap syndrome 90. Anti-RNAP I and anti-RNAP III almost invariably coexist 58289.

Anti-RNAP antibodies are associated with diffuse cutaneous involvement and have the highest likelihood of being associated with dcSSc than any other disease-specific autoantibodies apart from anti-Scl-70 74382889192. They are found in about 40% of patients with dcSSc. The presence of anti-RNAP II antibodies has been found to independently predict lower lung function, even when ethnicity, age, smoking history, and disease duration were considered simultaneously 64, although this is not uniformly seen 7.

Anti-RNAP antibodies, like anti-Scl-70 antibodies, are correlated with a higher rate of SSc-related mortality, though not independently so. There exists a highly significant association between anti-RNAP antibodies and right heart failure unrelated to pulmonary fibrosis (probably related to pulmonary hypertension), which accounts for this increase 66.

Anti-RNAP I, anti-RNAP II, and anti-RNAP III were found to be associated with HLA-DQB1*0201 in one study, and no HLA association was seen in another 9193 (Table 3).

In 1985, anti-U3-RNP antibodies were isolated by IP techniques 94. More recently it was shown that the mammalian U3 small nuclear RNP (snRNP) is one member of a family of nucleolar snRNPs that are immunoprecipitable by anti-fibrillarin autoantibodies 95. AFA are present in about 4% of patients with SSc and are mutually exclusive with ACA, anti-Scl-70, and anti-RNAP 96. AFA have also been described in patients with SLE, UCTD, and primary Raynaud's phenomenon 77. The frequency of AFA is much higher in patients of African descent with SSc and is reported to be as high as 16–22% compared with only 4% in Caucasian patients with SSc 408895. AFA are highly specific for dcSSc 14043479296 and when found in African American patients with SSc are virtually always associated with dcSSc 408996. Their presence in Caucasian patients with SSc is associated with diffuse skin involvement, but the correlation is not nearly as strong 96. AFA-positivity in those patients with dcSSc also has been associated with myositis, pulmonary hypertension, and renal disease. These autoantibodies also identify a younger subset of SSc patients with frequent internal organ involvement. However, in patients with lcSSc the presence of AFA did not predict pulmonary hypertension. Strangely, for its degree of internal organ involvement, AFA were not associated with a higher mortality rate, although those who died tended to succumb to pulmonary hypertension 96.

Although not seen in all studies 93, the autoantibody response to U3-RNP was associated in one study with HLA-DQB1*0604 40 (Table 3).

Although the autoantibodies discussed in this section are much less specific to SSc than those already described, the following do carry valuable information.

Significant serologic heterogeneity is well known to occur in SSc. Although it remains controversial whether autoantibodies seen in patients with SSc have an actual role in pathogenesis, these serologic markers are useful in the diagnosis and clinical management of scleroderma patients (Table 5). ACA are most often found in Caucasians and in association with limited cutaneous involvement, CREST, and isolated pulmonary hypertension. In contrast, they are infrequently found in patients with pulmonary fibrosis. ACA seem to be a marker for a better prognosis, whereas anti-Scl-70 antibodies, found in patients with dcSSc and pulmonary fibrosis, portend a poor prognosis with increased SSc-related mortality. The following of ACA and anti-Scl-70 levels over time has not been shown to have clinical utility. Of the ANoA, anti-PM-Scl and anti-Th/To antibodies are associated chiefly with lcSSc (with anti-PM-Scl antibodies associated with an overlap syndrome), whereas AFA and anti-RNAP are seen with dcSSc. Anti-Th/To, anti-RNAP and AFA are associated with a less favorable prognosis with a higher frequency of organ involvement, contrary to what is seen in those with anti-PM-Scl antibodies.

Anti-Ku antibodies might have a role in identifying CTD patients with overlap syndrome involving features of scleroderma in the absence of other autoantibodies such as anti-PM-Scl or anti-U1-RNP antibodies. Anti-Ro antibodies are identified in the sera of SSc patients with Sjögren syndrome. Anti-Sm antibodies are rarely seen in patients with SSc unless there are features of SLE overlap. When present, they predict a poor prognosis with frequent renal involvement. Anti-U1-RNP antibodies are usually seen in association with CTD overlaps, specifically with Raynaud's phenomenon, joint involvement, myositis, lcSSc, and a more favorable outcome. Although not seen in association with thrombosis in patients with SSc, inconsistent findings of associations with myocardial ischemia and pulmonary hypertension indicate a need for further study before any clear place of aPL determinations in patients with SSc can be recommended. Similarly, the clinical relevance of more newly recognized autoantibody systems in patients with SSc, particularly ANCA, anti-endothelial cell antibodies and anti-fibrillin-1, needs more study.

Much like the SSc, these disease-associated autoantibodies differ in their frequencies, associated clinical manifestations, pathophysiology, and ethnic and genetic associations. When used correctly they can be a clinically relevant and useful tool in patient management.

None declared.

ACA = anti-centromere antibodies; aCL = anticardiolipin antibodies; AFA = antifibrillarin/anti-U3-RNP; ANA = anti-nuclear antibodies; ANCA = anti-neutrophil cytoplasmic antibodies; ANoA = anti-nucleolar antibodies; anti-RNAP = anti-RNA-polymerase antibodies; anti-Sm = anti-Smith antibodies; aPL = antiphospholipid antibodies; β₂gp I = β₂ glycoprotein I antibodies; CENP = centromeric nucleoprotein; CIE = counterimmunoelectrophoresis; CREST = a variant of SSc defined by the presence of calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, and telangectasia; CTD = connective tissue diseases; dcSSc = diffuse cutaneous systemic sclerosis; DL_CO = diffusion capacity for carbon monoxide; DM = dermatomyositis; ELISA = enzyme-linked immunosorbent assay; FVC = forced vital capacity; HLA = human leukocyte antigen; IB = immunoblotting; ID = immunodiffusion; IIF = indirect immunofluorescence; IP = immunoprecipitation; lcSSc = limited cutaneous systemic sclerosis; MCTD = mixed connective tissue disease; PFT = pulmonary function tests; PM = polymyositis; PM/SSc = myositis/scleroderma overlap; RLD = restrictive lung disease; RNP = ribonucleoprotein; SLE = systemic lupus erythematosus; snRNP = small nuclear RNP; SSc = systemic sclerosis.