Introduction

Following the development of the ANA test by indirect immunofluorescence (IIF), the first case of ANoA staining was described in 1961 [1]. Three cases of SSc with ANoA were described [2]. Beck et al. reported results of a large number of patients in 1962 [3]. Eleven sera were found from 358 patients with various connective-tissue diseases, while none of the unselected 490 patients with diseases other than “autoimmune” had ANoA. Six of the 11 patients had scleroderma (SSc, systemic sclerosis), 2 Sjӧgren’s syndrome (SjS), one each of systemic lupus erythematosus (SLE), discoid lupus erythematosus (DLE), and pernicious anemia. ANoA was most common in SSc (19%, 6/32) and found in 5% (2/43) in SjS, 1.5% (1/67) in SLE, 0.67% (1/152) in DLE, 2% (1/52) in pernicious anemia, and none in 90 rheumatoid arthritis (RA). Rodent tissues including rat liver, mouse kidney, and other animal tissues have been used commonly as a substrate of ANA in earlier studies. Animal and human culture cell slides have become commercially available in the late 1970s, and the use of human HEp-2 (laryngeal cancer cell line) has become the standard substrate of immunofluorescence ANA during the late 1980s. Meanwhile, several nucleolar autoantigens were identified and characterized, mainly as cancer-associated antigens.

The use of the ANA slide of human-cultured cells was a major breakthrough in the ANA test. Currently, the HEp-2 ANA slide has been sold by many companies, and virtually, all ANA tests in the laboratory are based on HEp-2 or its variant of cell substrates worldwide. The standardization of methods and interpretation of the ANA test has been an important but difficult issue due to so much variability in the ANA slide, serum dilution, protocol, reagent, in particular, fluorochrome-conjugated secondary antibodies, and equipment used, in addition to the subjectivity of interpreting immunofluorescence reactivity by human eyes.

There have been activities to standardize and harmonize the names and descriptions of the HEp-2 cell ANA patterns. The International Consensus on ANA Patterns (ICAP) was established and a consensus was reached originally on 28 HEp-2 patterns with an alphanumeric code of AC-1 to AC-28 [4]. After the initial classification, the patterns AC-29 [5] and AC-0 (negative) [6] were added (http://www.ANApatterns.org). Historically, ANoA has been recognized and classified into 3–5 fine staining patterns in the literature. In the current ICAP classification, ANoA are defined in 3 AC patterns including AC-8 (homogeneous nucleolar), AC-9 (clumpy nucleolar), and AC-10 (punctate nucleolar) in ICAP classification as shown in Fig. 1.

Fig. 1
figure 1

Immunofluorescence antinuclear antibodies. HEp-2 ANA slides (MBL Inc., Japan) were stained with 1:80 diluted sera from patients followed by incubation with Dylight 488 goat antihuman IgG (γ-chain specific) F(ab)′2. Images with blue color are merged images of Dylight 488 (green) and DAPI staining (blue) of the nuclei to clearly show the staining pattern of mitotic cells. Staining by prototype sera with anti-Th/To, -PM-Scl, -U3RNP, -NOR-90, -RNA polymerases (I > III, anti-RNAP I is strong with weak anti-RNAP III), and -RNAP I/III are shown. Staining by sera with coexistent anti-centromere and -Th/To, PM-Scl, or RNAP I/III is shown

Full size image

There have been many excellent review articles and textbook chapters on specific autoantibodies based on each disease such as autoantibodies in SLE, SSc, inflammatory myopathy, and others; however, review articles based on specific ANA patterns are limited. Clinical relevance of HEp-2 ANA patterns based on ICAP classification was published recently [7]. This review will focus on the clinical relevance of ANoA and specific antibodies that show ANoA.

Detection of Antinucleolar Antibodies by IIF

Currently, the only method to detect ANoA is IIF using human HEp-2 cells. Since the definition of ANoA is based on the physical location of the areas in the cells recognized by human autoantibodies, other methods that have been used to test ANA, including beads assay or enzyme-linked immunosorbent assay (ELISA), are not suitable for the detection of ANoA. In fact, a high prevalence of negative ANA by methods to detect ANA other than IIF has been reported and thought this was because nucleolar autoantigens were not included in the set of recombinant proteins used in these beads assay. The prevalence of ANoA in the ANA-positive population is not high compared with those of nuclear or cytoplasmic staining. Nevertheless, ANoA has been reported in various autoimmune and non-autoimmune conditions and also in healthy individuals like ANA.

Prevalence of Antinucleolar Antibodies by IIF

Selected data on the prevalence of ANA and ANoA using the HEp-2 slide are summarized in Table 1. The prevalence of ANA in the general population in the USA (NHANES, National Health and Nutrition Examination Survey) was ~ 12 to 16% and ANoA was ~ 6 to 7% of the ANA [8, 9]. In a cohort study in Japan, ANoA was 4.7% of ANA positives [10]. Based on these numbers, 0.5–1% of the general population has ANoA. ANoA among ANA positives in regional or hospital laboratories was not so different than those in the general population and 5–9% in various countries [11–14].

Table 1 Prevalence of anti-nucleolar antibodies
Full size table

ANoA has been classically known to be associated with SSc, reflecting its association with several SSc-specific or -associated autoantibodies including anti-RNA polymerase I (RNAP I), U3RNP/fibrillarin, Th/To, and PM-Scl. The prevalence of ANoA in SSc was relatively higher than those in other conditions and ~ 20 to 40%, though the prevalence appears to be different depending on race/ethnicity and disease subset [15–20]. ANoA in CTD was 9.1% [21], and a study reported 20.5% of ANA in RA was ANoA [22]. In patients with interstitial cystitis, ANA was positive in 36% and 7% (20% of ANA positives) were ANoA [23]. ANoA in chronic liver disease, in particular, a high prevalence (63.2% of ANA) in liver transplant patients, were reported.

ICAP Classification of Nucleolar Patterns

The following classification and description are from the ICAP website (www.anapatterns.org) and from the publication [7]. Representative immunofluorescence images for each AC pattern using human autoimmune sera are shown (Fig. 1).

AC-8 Homogeneous Nucleolar

Description: diffuse fluorescence of the entire nucleolus, while the metaphase plate shows no staining.

Antigen association: PM/Scl-75, PM/Scl-100, Th/To, B23/nucleophosmin, nucleolin/C23, and No55/SC65.

AC-9 Clumpy Nucleolar

Description: irregular staining of the nucleoli and Cajal bodies with peri-chromosomal staining at the metaphase plates, e.g., anti-fibrillarin.

Antigen association: U3-snoRNP/fibrillarin.

AC-10 Punctate Nucleolar

Description: densely distributed but distinct grains seen in the nucleoli of interphase cells. In metaphase cells, up to 5 bright pairs of the nucleolar organizer regions (NOR) can be seen within the chromatin body. The cytoplasm of mitotic cells may be slightly positive, e.g., anti-NOR-90, anti-RNA polymerase I.

Antigen association: RNA polymerase I, hUBF/NOR-90.

Homogeneous Nucleolar (AC-8): Anti-Th/To and Anti-PM/Scl Antibodies

  1. 1.

    Anti-Th/To antibodies

Anti-Th/To antibodies were first described as anti-Th antibodies using serum from a patient with SLE [24]. Target antigens of anti-To [24] antibodies using serum from a patient with ANoA-positive SSc and those of anti-Th [25] confirmed that they have the same specificity and recognize RNA–protein complex containing 7-2RNA and 8-2RNA. Since then, this specificity has been called anti-Th/To antibodies.

Th/To Antigens

The Th/To antigen is a multiprotein-RNA complex (human RNase mitochondrial RNA processing (MRP) complex) that contains 7-2RNA and 8-2RNA and at least 9 protein subunits, Rpp14, Rpp20, Rpp21, Rpp29 (hPop4), Rpp25, Rpp30, Rpp38/40, hPop1, and hPop5 [26].

Anti-Th/To Detection Methods

Immunoprecipitation (IP) analysis and the detection of RNA components, 7–2 and 8-2RNA is the original method of identification of anti-Th/To antibodies and has been considered the gold standard (Fig. 2). Originally, HeLa cells metabolically labeled with 32P-orthophosphate were used; however, later studies used IP of non-radiolabeled cell extract followed by urea-PAGE and silver staining to identify RNA components. The IP analysis of protein components using 35S-methionine labeled cells had limited significance, though a ~ 40 kD protein named Th40 was considered a major target and utilized in some studies. One study reported that the ~ 40 kD antigen consists of Rpp20 and/or Rpp25 associated with a nuclease-resistant RNA fragment [27]. Protein IP of the 35S-methionine labeled cell extract has not been much used for the detection of anti-Th/To, and components of Th/To complex have not been clearly shown. Nevertheless, a set of proteins immunoprecipitated by anti-Th/To sera can be recognized (Fig. 3 arrows). Which protein band seen in IP corresponds to which known subunit has not been clarified.

Fig. 2
figure 2

RNA components of anti-U3RNP and Th/To antibodies by immunoprecipitation. Cell extract from K562 cells was immunoprecipitated by human autoimmune sera. RNA components of the immunoprecipitates were analyzed by urea-PAGE and silver staining

Full size image
Fig. 3
figure 3

Immunoprecipitation analysis of protein components of anti-Th/To, -PM-Scl, and -U3RNP antibodies. A 8% SDS-PAGE and B 13% SDS-PAGE. 35S-methionine-labeled K562 cell extract was immunoprecipitated by anti-Th/To, -PM-Scl, or U3RNP human autoimmune sera, and protein components were analyzed by SDS-PAGE and autoradiography. Arrows, components of Th/To complex; blue dots, components of PM-Scl complex; red arrow, fibrillarin; NHS, normal human serum. Anti-U3RNP serum in lane 9 is also positive for anti-Ku (two strong bands of 70–90 kD)

Full size image

Anti-Th/To Antibodies Clinical Significance

It has been believed that anti-Th/To antibodies are relatively specific for the diagnosis of SSc; however, studies reporting disease specificity of anti-Th/To among various systemic rheumatic diseases by immunoprecipitation are very limited, and only a few studies from the USA [28] and Japan [29, 30] are available (Table 2).

Table 2 Disease specificity of antinucleolar antibodies in various rheumatic diseases
Full size table

It appears that the detection of anti-Th/To is relatively specific for SSc (~ 4%) while the number of patients with anti-Th/To and SARDs other than SSc are small compared with those of SSc. Some of these patients have Raynaud’s phenomenon [28, 29, 31–34], maybe considered sine SSc [35–37], early stage of SSc, or undifferentiated connective tissue disease (UCTD) that may evolve into SSc. However, not all anti-Th/To-positive patients have Raynaud’s phenomenon or other signs suggesting they are at the early stage of SSc [29]. Almost all studies reported a few percent prevalence of anti-Th/To in SSc, and the prevalence appears to be somewhat different between different countries (Table 3). When the SSc subsets were compared, virtually, all studies reported a much higher prevalence of anti-Th/To in limited cutaneous SSc (lcSSc) compared with diffuse cutaneous SSc (dcSSc) and the majority of patients had lcSSc regardless of the country of the studies. Also, one study compared the association of anti-Th/To and SSc subsets in African American, Caucasian, and Hispanic to show its association with lcSSc in all races [15].

Table 3 Prevalence of anti-Th/To antibodies identified by IP in various CTDs
Full size table

The prevalence and subset association of anti-Th/To by line immunoassay (LIA) are summarized (Table 4). The prevalence and association with lcSSc appear similar to the results by IP, though the detection in dcSSc may be somewhat more common. The commercial LIA using hPop1 has been increasingly used worldwide due to the limited availability of IP assay; however, the data will need to be carefully interpreted because LIA has not been fully validated compared with a gold standard IP and many anti-Th/To-positive sera do not recognize hPop1 [29, 33]. One study reported that a significant number of anti-Th/To positives (19 in 873 SSc patients) detected by IP were missed by LIA [38]. Thus, although some anti-Th/To antibody immunoassays are commercially available, technical issues relating to the limited sensitivity and the specificity of these immunoassays should be taken into consideration [7, 26, 39].

Table 4 Prevalence of anti-Th/To antibodies in various CTDs detected by LIA
Full size table

Several studies reported the prevalence of anti-Th/To by chemiluminescence immunoassay (CIA) or ELISA using Rpp25 recombinant protein, CIA using Rpp38 peptide, and other assays (Table 5) [39–44]. Immunoassays using a single component would show a lower prevalence compared with IP, which would detect antibodies to all components of multiprotein-RNA complex of Th/To antigen in native form. The relative specificity of SSc, a few percent prevalence, and association with lcSSc of these antibodies were reported. Some studies were reported to show a good correlation with IP results and may become available in the future.

Table 5 Prevalence of anti-Th/To antibodies in various CTDs by CIA and other methods
Full size table

Reported anti-Th/To-positive cases by IP in diseases other than SSc or reports on selected patients are summarized (Table 6). Some studies selected patients based on ANoA to find a few cases of anti-Th/To in patients with SLE, polymyositis (PM), Raynaud’s disease, and others [32, 33]. Fischer et al. reported that 25 of 285 idiopathic pulmonary fibrosis (IPF) patients had ANoA and the majority (13/25) of ANoA were anti-Th/To [37]. Four of them were lcSSc, and 9 could be considered sine scleroderma (ssSSc).

Table 6 Additional anti-Th/To-positive cases (selected patients, non-SSc patients, etc.)
Full size table

The prevalence of anti-Th/To antibodies from the studies with a clear description of the race of the subjects is summarized in Table 7. It appears possible that the prevalence of anti-Th/To in African Americans is lower compared with Caucasian or Latin based on 3 studies directly comparing this point.

Table 7 Prevalence of anti-Th/To antibodies identified by IP and race
Full size table

In IP assay, a whole RNA–protein complex of Th/To is immunoprecipitated regardless of which components are recognized by serum autoantibodies. To examine which components are directly recognized and whether the fine specificity of anti-Th/To is associated with clinical features, several studies reported the reactivity with recombinant proteins (Table 8). Nine components were tested, and Rpp25, Rpp30, and hPop1 were recognized by more than half of the sera [29, 33]. One study correlated the reactivity of the serum with Rpp25 and Rpp38 with AC patterns in immunofluorescence (Table 9) [42]. The majority of patients (43/51) was classified as having AC-8 pattern, consistent with the immunofluorescence staining by anti-Rpp25 antibodies. Overall, only 3 of 51 anti-Rpp25 were without the AC-8 pattern. The prevalence of anti-Rpp38 peptide was lower, but all 8 cases had the AC-8 pattern.

Table 8 Autoantibodies to components of Th/To complex in sera defined by anti-Th/To IP positives
Full size table
Table 9 Prevalence of antibodies to Rpp25 and Rpp38 by ICAP AC pattern [42]
Full size table

Clinical association of anti-Th/To antibodies based on studies where the specificity was defined by IP (Table 10). It is somewhat difficult to define clinical features associated with anti-Th/To antibodies because the methods of comparison in the studies are quite heterogeneous, and how and what should be compared is arguable. Some studies compared the prevalence of features between anti-Th/To positive and anti-Th/To negative in all SSc patients, while others compared between patients with several different SSc autoantibodies. There are studies comparing within lcSSc because anti-Th/To is mainly seen in lcSSc, and other studies compared between anti-Th/To vs ACA because both are mainly associated with lcSSc.

Table 10 Clinical associations of anti-Th/To antibodies detected by IP
Full size table

In anti-Th/To-positive SSc, the majority of patients with anti-Th/To had lcSSc and low skin score [15, 28, 29, 45–49] is a consistent finding. Anti-Th/To was associated with a high prevalence of puffy fingers, small bowel involvement, hypothyroidism [28], and a low prevalence of joint involvement [28, 45, 50]. Anti-Th/To-positive patients had reduced survival due to pulmonary hypertension (PH) [28]. One study reported anti-Th/To was detected in four of 11 patients who had both scleroderma renal crisis (SRC) and PH but none in 23 SRC only patients and 1/15 in PH only patients [51].

By comparing clinical features of patients with two autoantibodies associated with lcSSc, anti-Th/To vs ACA, anti-Th/To-positive patients were younger and more frequently had ILD and restrictive lung disease (reduced % FVC), SRC, and reduced survival [34, 52]. Anti-Th/To patients had a lower prevalence of vascular involvement (pitting scars, digital tip ulcers, digital gangrene), esophageal dysmotility, telangiectasia, and sicca symptoms [34].

Regarding non-SSc patients with anti-Th/To, Okano reported 1.2% (3/244), and all 3 patients had primary Raynaud’s phenomenon [28], and similarly, Kipnis detected anti-Th/To in 2 of 28 patients with Raynaud’s phenomenon [31]. Though anti-Th/To has been mainly reported in SSc or primary Raynaud’s phenomenon, detection in 7 patients with SLE, RA, polymyositis (PM), SjS, and idiopathic thrombocytopenic purpura (ITP) with only one of 7 with Raynaud’s phenomenon, was reported [29]. Other studies also reported a small number of anti-Th/To-positive patients in various diseases including SLE, PM, dermatomyositis (DM), autoimmune hepatitis (AH), and others [33, 34, 53]. One study reported 3 cases of anti-Th/To in 70 patients with localized scleroderma [54].

In a study on patients with IPF who had ANoA, clinical features of 13 anti-Th/To patients were compared with 12 ANoA-positive anti-Th/To negative patients [37]. Anti-Th/To was associated with female sex, Raynaud’s phenomenon, telangiectasia, digital edema, calcinosis, and PH. They stated that 4 of their anti-Th/To-positive patients could be classified as lcSSc, and the remaining 9 were considered ssSSc [37].

In a prospective study on patients with Raynaud’s phenomenon without definite connective tissue disease, anti-Th/To was identified as a predictive factor of progression to definite SSc (HR 5.9, 95% CI 3.2–10.98 by univariate analysis, HR 3.56, 95% CI 1.5–5.3) and an independent factor predictive of progression to microvascular damage by nailfold capillary microscopy (NCM), HR 2.4 (95% CI 1.14–5.06) [55].

In summary, anti-Th/To is one of the ANoA specificity detected in patients with SSc. In SSc, most anti-Th/To-positive patients have lcSSc and less severe internal organ involvement. ILD may be relatively common within lcSSc compared with ACA-positive patients. It can also be detected in patients with ssSSc, idiopathic interstitial lung disease (ILD), Raynaud’s phenomenon, and UCTD but may also occasionally be seen in other diseases or patients without Raynaud’s phenomenon.

  1. 2.

    Anti-PM-Scl Antibodies

Specific autoantibodies in PM/DM were first described as a new precipitin line in double immunodiffusion (DID) using calf thymus nuclear extract (CTNE) as antigen, named anti-PM-1 [56]. Anti-PM-1 was detected in 17 of 28 patients with PM/DM (PM 9/14, DM 1/6, PM-SSc overlap 7/8) but not in 460 patients with other diseases. The following study pointed out the heterogeneity of the anti-PM-1 system and described anti-PM-Scl antibodies, which were detected in 12.5% of PM/DM (21/68) but none in SLE, SSc, RA, SjS, or normal human controls (NHC) [57]. Among the 21 anti-PM-Scl positive cases were 7 SSc-PM and 2 SSc-DM overlap syndrome, suggesting anti-PM-Scl antibodies are associated with SSc-PM/DM overlap syndrome. The target antigen of anti-PM-Scl was characterized by immunoprecipitation in 1986 [58]. RNA component was not observed and 11 polypeptides of molecular weight 110 to 20 kD, named p1–p11 were identified. PM-Scl is a human exosome complex and functions as exoribonucleases during RNA processing. Among the many components of the complex, PM-Scl 100 (~ 110 kD) and PM-Scl75 are considered major targets of autoantibodies and commercial immunoassays utilize these antigens [59].

The prevalence of anti-PM-Scl antibodies in patients with rheumatic diseases by DID or immunoprecipitation is summarized (Table 11). Generally, in US and European countries, anti-PM-Scl antibodies were reported in 2–5% of PM/DM or SSc, and the prevalence is higher in SSc-PM/DM overlap syndrome or scleromyositis than pure SSc or PM/DM. When anti-PM-Scl positive patients had SSc, most of them had lcSSc. In one study on 23 cases of anti-PM-Scl, 8/10 SSc-PM/DM overlap patients had lcSSc, and 5/6 SSc patients had lcSSc [60].

Table 11 Prevalence of anti-PM-Scl antibodies by immunodiffusion or immunoprecipitation
Full size table

In contrast to an association of anti-PM-Scl antibodies with SSc-PM/DM overlap syndrome in the USA and Europe, this specificity was reported absent in Japan until recently, probably related to the low prevalence of HLA-DR3 that is tightly linked with the production of this autoantibodies [30]. Muro et al. reported anti-PM-Scl in 2.3% (3/133) of PM/DM and 0.9% (2/223) of SSc and 25% (4/16) of UCTD [61]. Other cases of anti-PM-Scl positive DM or SSc-DM overlap have been reported in Japan [62, 63]. Thus, anti-PM-Scl also appears to be detected in Japanese patients though at a low prevalence and with a less clear association with overlap syndrome.

In a study on ANoA in SSc, anti-PM-Scl was found in 8 of 37 ANoA positive SSc and associated with SSc-PM/DM overlap syndrome and renal crisis [64].

A strong association of anti-PM-Scl with scleromyositis was reported from Poland. One study reported 83% of the 108 cases with anti-PM-Scl had scleromyositis [65]. Another study reported 19 of 20 anti-PM-Scl positive patients were scleromyositis [66]. In scleromyositis, 90.5% (19 of 21) were anti-PM-Scl positive.

In non-rheumatic diseases, in one study on GVHD, 2 of 19 chronic GVHD (cGVHD) with SSc were positive for anti-PM-Scl and anti-Scl-70 [67].

In summary, anti-PM-Scl antibodies are mainly detected in patients with SSc-PM/DM overlap syndrome [57, 60, 68] or scleromyositis [65, 66]; however, it also can be detected in PM/DM or SSc without overlap syndrome at a 3–5% prevalence. When detected in patients with SSc, most of them have lcSSc.

  1. 3.

    AC-8 Pattern ANoA with Unclear Clinical Significance

There are several other ANoA specificities that show the AC-8 pattern but have unclear clinical significance; anti-nucleophosmin/B23, nucleolin/C23, nucleolar RNA helicase II/Gu, No55, Nop52, and others. They have been reported in various conditions including SARDs, malignancies, GVHD, and others. It is possible that some of these specificities have clinical significance and are useful in diagnosis or predicting clinical features; however, standard immunoassay and systematic screening studies are currently missing, and these autoantibodies are not widely utilized.

Nucleophosmin (NPM)/B23

Anti-NPM Antibodies in Rheumatic and Immune-Mediated Diseases

Nucleophosmin (also called B23, numatrin, No38) is a 37-kD abundant nucleolar phosphoprotein that has roles in preribosomal RNA processing and ribosome assembly.

In a study analyzing autoantibodies in sera from 164 patients (120 provisional diagnoses of SARDs, 44 SSc) with ANoA, 7 sera positive for anti-NPM were detected [69]. Six out of seven anti-NPM positive had anti-cardiolipin (CL) antibodies, and anti-NPM was associated with clinical features of SLE (N = 2) or variants of SLE in 5 out of 7 patients (Table 12).

Table 12 Prevalence of autoantibodies to nucleophosmin (NPM)
Full size table

Ulanet detected antibodies to NPM in 10 of 92 SSc (10.9%) patients and compared clinical features with 82 anti-NPM-negative SSc patients [70]. A total of 70% of anti-NPM positive patients had limited SSc and anti-NPM was associated with low % FVC and a higher prevalence of anticentromere antibodies (ACA) (60% vs 24%), with severe PH. Though anti-NPM has been reported in various malignancies (Table 12), none of the anti-NPM-positive SSc patients had malignancy. Anti-NPM frequently coexisted with anti-U3RNP/fibrillarin (anti-NPM + vs − , 80% vs 18%, P = 0.0002), and 20% had anti-Scl70, while 40% had anti-U1RNP, but none had ACA.

In SLE, anti-NPM antibodies were positive in 28% of sera and significantly associated with anti-cardiolipin (CL) antibodies but not associated with ANA, anti-dsDNA, anti-β2GPI (beta2 glycoprotein I), and anti-β2GPI-dependent CL antibodies [71]. The interaction of CL to NPM in vitro was shown and suggested a possible role in the concomitant production of anti-NPM and anti-CL antibodies. In a mouse study using (NZW × BXSB) F1 mice (WB mice), a model of SLE, anti-NPM antibodies were positive by ELISA in more than 75% of sera from male WB mice [71]. Anti-NPM in female WB mice was less frequent, detected in 25% at 3 months and 40% at 4–6 months. Anti-NPM antibodies in WB mice were associated with anti-CL antibodies.

In a study on allogeneic bone marrow transplant (BMT), 12/19 of patients with extensive GVHD had ANoA and 10/19 were positive for anti-NPM antibodies [72]. Anti-NPM was found in one out of 10 ANoA-positive SSc but not in 8 autologous BMT patients or 48 healthy controls.

Anti-NPM Antibodies and Malignancy Association

In a study on breast cancer patients, levels of anti-NPM antibodies by ELISA were not different between recurrent and nonrecurrent patients at the time of diagnosis or recurrence; however, anti-NPM levels increased significantly between diagnosis and 6 months before a recurrence in recurrent patients [73]. The degree of changes in the levels of anti-NPM antibodies between diagnosis and 6 months before recurrence was associated with the risk of recurrence. Mojtahedi et al. reported anti-NPM antibodies by 2-dimensional western blot (2D-WB) in 44% of HER2-positive or -negative breast cancer patients [74]. Imai et al. reported anti-NPM antibodies in only one out of 184 hepatocellular carcinomas (HCC) patients by WB [74]. In the other two studies, the prevalence of anti-NPM in HCC was 22.4% and 10.5–21.4% [75, 76]. Anti-NPM was also detected in one of 37 lung cancer patients but not in 210 alimentary tract cancer patients [77]. In nonmalignant diseases, anti-NPM was not detected in 187 chronic hepatitis and liver cirrhosis in one study [77], but others reported 3.3% in liver cirrhosis [75] or 5.4% in chronic hepatitis [76] and 1.7% in healthy controls [76].

Immunoreactivity with NPM in sera from 125 ovarian cancer patients was significantly higher than those with 40 patients with the benign ovarian disease and 40 controls by dot blot [78]; however, the prevalence or individual data were not reported. Lu et al. tested levels of antibodies to 14 currently promising ovarian cancer-related biomarkers including B23 by ELISA, in 151 ovarian cancer patients, 23 borderline ovarian tumors, 55 benign tumors, and 75 healthy controls but found no difference in anti-B23 levels [79].

Dai et al. detected anti-NPM antibodies by WB in 28.2% of prostatic carcinoma (PCa) patients but none in benign prostatic hypertrophy (BPH) or NHC [80]. By ELISA using recombinant protein, anti-NPM levels were higher in PCa patients than BPH or NHC. The ROC curve analysis showed a high diagnostic value for PCa to differentiate BPH or NHC. A total of 97.1% of early-stage PCa patients were identified correctly, while 69.2% of BPH patients with elevated prostate-specific antigen (PSA) levels were anti-NPM negative. Anti-NPM levels increased significantly after surgery in patients with early-stage PCa.

Pulford et al. reported autoantibodies to ALK (anaplastic lymphoma kinase)-NPM fusion protein in all 11 ALK-positive ALCL (anaplastic large cell lymphoma) but not in 13 controls (5 healthy, 5 cancer, 3 ALK-negative ALCL) [81]. Damm-Welk et al. detected anti-ALK antibodies in 13 of 21 ALK-positive non-small cell lung cancer (NSCLC) and 13 of 22 ALK translocation positive but NPM-ALK-negative lymphoma patients and one ALK-positive rhabdomyosarcoma patient but not in 20 healthy adults [82]. These sera reacted with ALK-fusion proteins but also with full-length ALK-transfected COS cells, suggesting that antibodies recognize the ALK portion of the fusion protein. Consistent with this study, Knorr et al. showed antibodies from patients with NPM-ALK-positive ALCL patients recognized epitopes on ALK-portion of NPM-ALK by overlapping peptide microarray analysis [83].

In summary, anti-NPM antibodies were reported in rheumatic diseases including SLE and SSc, GVHD, and various types of malignancies including breast cancer, HCC, ovarian cancer, and prostatic cancer. Systematic studies are lacking and clinical association of anti-NPM antibodies does not appear to be established or consistent. Generally, studies by WB tend to show a lower prevalence of anti-NPM antibodies than studies using ELISA. In the majority of studies, whether anti-NPM-positive patients by ELISA or WB showed ANoA by immunofluorescence was not described.

Nucleolin/C23

Nucleolar phosphoprotein originally described as C23 was characterized as a ~ 110 kD protein nucleolin that plays roles in ribosome biogenesis and intranuclear transport of preribosomal particles. Autoantibodies to nucleolin/C23 were reported in SLE and other diseases by WB using serum dilution of 1:10 (IgM) or 1:20 (IgG) [84]. IgM antibodies were positive in 64% (27/42) and IgG in 17% (7/42) adult SLE. The reactivity was mainly IgM and was not specific for SLE. The prevalence of IgM or IgG antibodies were as follows: childhood SLE (15/15), SjS (17/24), SSc (7/16), PM (2/10), RA (1 /20), or polymyalgia rheumatica (PMR) (1 /20), patients with acute hepatitis A (16/20) or infectious mononucleosis (4/20), and none in 10 normal subjects.

In a study on ANA in chronic GVHD after bone marrow transplant (Table 13), 95% of 19 cGVHD with SSc had ANA and 58% had ANoA [67]. Anti-nucleolin antibodies were detected in three of them but not in 18 cGVHD without SSc by WB though ANA and ANoA were positive in 58% and 22%, respectively, in these subsets.

Table 13 Prevalence of anti-nucleolin antibodies in patients with transplant and GVHD
Full size table

Anti-nucleolin antibodies were also reported in kidney or heart transplant patients [85]. The prevalence appears to be high in patients with the irreversible rejection of kidney transplants and heart allografts with transplant-related coronary disease.

In a mouse study, anti-nucleolin antibodies were found by WB in sera from MRL/lpr mice (100%, 10/10) and NZB/W F1 mice (80%, 8/10) but not BALB/c mice (0/10) [86]. Among the 150 kD, 110 kD (nucleolin), 75 kD, and 55 kD proteins often targeted by autoantibodies in these mice, anti-nucleolin antibodies were produced first in all 10 MRL/lpr mice, whereas it was first in 7/10 NZB/W F1 mice.

The clinical association and significance of anti-nucleolin antibodies are not well understood as studies are very limited and no systematic analysis in various diseases is available. Moreover, the relationship between the results by different immunoassays is not known, and thus no gold standard for anti-nucleolin detection is accepted.

Nucleolar RNA Helicase II/Gu

Autoantibodies to nucleolar RNA helicase II (RHA II)/Gu was originally described using a serum “Gu” from a patient with gastric antral vascular ectasia (GAVE, watermelon stomach) that has high titer ANoA [87], and it was shown that Gu antigen is identical to RHA II.

In the following study, anti-RHA II/Gu was screened by WB using the recombinant protein in 108 sera with ANoA from 3408 sera tested [88]. Anti-RHA II/Gu antibodies were found in 11 (10%) of 108 ANoA-positive sera, including 3 of 46 patients with SSc (7%), 3 of 17 patients with SLE (18%), 4 of 9 patients with UCTD (44%), and 1 healthy relative of a SSc patient. It was not found in 11 ANoA-positive SjS, 11 PM/DM, 5 primary antiphospholipid syndromes (APS), and 3 RA. None of the anti-Gu-positive patients had symptoms suggestive of gastric antral vascular ectasia (GAVE). Two other sera positive for ANoA from patients with GAVE were negative for RHA II/Gu [89].

Anti-RHA II/Gu antibodies occur in low frequencies in patients with connective tissue diseases (CTDs) who have ANoA, but they are not specific for SSc or GAVE.

In conclusion, anti-RNA helicase II/Gu antibodies were reported in ANoA-positive SSc, SLE, and UCTD, but their clinical associations will need to be evaluated in the future.

No55

No55 was identified using serum from a female patient with interstitial cystitis from Finland [23]. The antibody was named based on its nucleolar staining in immunofluorescence and ~ 55 to 58 kD doublet proteins recognized in WB. The anti-No55 staining pattern was unique in demonstrating uniform staining throughout the interphase nucleolus, chromosomal staining in mitotic cells, and no apparent staining of coiled bodies.

In a study on patients with prostate cancer, No55 was identified as a protein recognized by 7 out of the 47 sera (14.9%) but not by 20 healthy male controls [90]. Six out of 45 patients with local disease and one of two patients with metastatic disease were positive for anti-No55.

In conclusion, anti-No55 antibodies are reported only in a case of interstitial cystitis and 7 cases of prostate cancer.

Nop52

Nop52 was identified using a human serum from a bone marrow-transplanted male patient as a 52-kD nucleolar autoantigen colocalized with nucleolar proteins involved in the late processing step such as hPop1 and B236. cDNA cloning and sequence revealed that it was identical to NNP1 [91]. A study using a pool of 30 sera including a patient from neuroblastoma identified an autoantigen NNP3 as a novel amino-terminal variant of NNP1 [92]. Antibodies to NNP3 were detected in 1/30 healthy volunteers (titer 1:100), 1/10 neuroblastoma, and 1/10 non-neuroectodermal malignancies (1/2 Hodgkin’s disease) (titer 1:100 and 1:1000) by a plaque assay.

Clumpy Nucleolar (AC-9): Anti-U3RNP/Fibrillarin Antibodies

Autoantibodies to the U3RNA-protein complex were originally described using sera from ANoA-positive SSc patients [24]. A 34-kD protein fibrillarin associated with the U3RNP complex was identified as a target antigen [93]. A monoclonal antibody that recognized a 34-kD nucleolar antigen was established from a lupus-prone New Zealand black × New Zealand white (NZB/NZW) F1 mouse, and the identity of the antigen with fibrillarin was confirmed [94]. Specific induction of ANoA by subcutaneous injection of mercuric chloride in susceptible strains of mice with H-2 s MHC was reported. The target nucleolar antigen was identified as fibrillarin and established as a model of chemical induce autoimmunity [95, 96]. U3 small nucleolar RNP (snoRNP) is a box C/D snoRNPs and contains fibrillarin and several other proteins including Nop58, Nop56, 15.5 K, U3-55 k, Mpp10, Imp3, and Imp4 [97]. The structure of U3 snoRNPs appears quite complex and autoantibodies to components other than fibrillarin have not been fully studied. For clinical purposes, autoantibodies to U3 snoRNP in SSc are called anti-fibrillarin, anti-U3RNP, or anti-fibrillarin/U3RNP.

Anti-U3RNP/fibrillarin antibodies are considered a serological marker of SSc associated with dcSSc. While it is likely to be true, the studies examining the prevalence of anti-U3RNP by IP in various SARDs are limited [98] (Table 2). Anti-U3RNP was detected in 5.8% of SSc but none in SLE, PM/DM, or UCTD, and the only positive non-SSc patients has primary Raynaud’s phenomenon [98]. In a study from Japan, anti-U3RNP was found in 7.1% of SSc, and only one SLE patient was positive in other SARDs [30]. There was a study testing anti-U3RNP by ELISA and western blot (WB); however, the antibodies reported in the article appeared to be very different than conventional anti-U3RNP antibodies [99].

Anti-U3RNP antibodies in SSc patients were tested by standard IP assay in many studies. The majority of studies reported the prevalence of 2–8% in the analysis in different countries. In SSc, association with dsSSc based on the higher prevalence of anti-U3RNP in dcSSc than lcSSc has been shown in most studies (Table 14). There were some other studies that detected anti-U3RNP by WB or other methods or in selected patients [35, 51, 64, 100]. One study using IF and WB reported a case of anti-U3RNP in malignancy [77] (Table 15).

Table 14 Prevalence of anti-U3RNP/fibrillarin antibodies identified by IP in SSc patients
Full size table
Table 15 Additional studies reporting anti-U3RNP/fibrillarin prevalence based on detection methods
Full size table

Discussing the prevalence of anti-U3RNP antibodies in SSc as a whole may be confusing and misleading because the prevalence appears to be quite different depending on race/ethnicity (Table 16). In all studies, the high prevalence of anti-U3RNP antibodies in African Americans is shown (15.8–43.5%), while it is low (0–7.1%) in Caucasians and Asians. In one study in Afro-Caribbean patients, they also showed a high prevalence (22.2%), while Latin appears to be in the middle (6.1–12.5%). Furthermore, the association of anti-U3RNP with lcSSc vs dcSSc subset also appears to be depending on race/ethnicity (Table 17) [15, 98, 101]. In general, the association of anti-U3RNP with dcSSc has been shown in most studies; however, this association is limited to African Americans, and it is not seen in Caucasians or other races in one study [98]. In contrast, another study showed the same clinical association was only in Caucasians and Latin, and it was not seen in African Americans (Table 17).

Table 16 Prevalence of anti-U3RNP antibodies by IP and race
Full size table
Table 17 Prevalence of anti-U3RNP antibodies in SSc patients by race, sex, and disease classification
Full size table

Tormey reported a high prevalence of anti-U3RNP antibodies in Afro-Caribbean (22.2%) compared with Caucasian (3.4%) patients [101]. All 8 Afro-Caribbean patients, 2 Oriental and an Asian patient had dcSSc, whereas only 47% of Caucasians with anti-U3RNP had dcSSc, indicating the association of anti-U3RNP with dcSSc is stronger in Afro-Caribbean and Asian ethnicity. The distinctive clinical association of anti-U3RNP antibodies in different races will need to be carefully evaluated in future studies.

The association of anti-U3RNP antibodies with clinical features is summarized in Table 18. It is not straightforward to describe clinical association within SSc because the method of comparison varies, and also for possible differences between different races, but common features associated with anti-U3RNP include a high prevalence in African Americans, in male patients, with muscle inflammation, calcinosis, esophageal, and small bowel involvement, cardiac involvement, severe ILD, and PH [46, 49, 98, 100, 102]. The reduced prevalence of arthritis or kidney involvement is shown in some studies and also reduced the prevalence of pulmonary fibrosis [46, 102].

Table 18 Clinical association of anti-U3RNP antibodies detected by IP
Full size table

Punctate Nucleolar (AC-10): Anti-NOR-90 and -RNA Polymerase I Antibodies

NOR90/Human Upstream Binding Factor, hUBF

Antibodies to NOR90 were originally described in 6 patients with rheumatic diseases based on immunofluorescence staining of the nucleolar organizing region (NOR) (Fig. 1) and reactivity with doublet proteins of ~ 90 kD in WB [103] and IP (Fig. 4). cDNA clones for NOR-90 were identified as hUBF [31]. From a cohort of rheumatic diseases, the original study reported four of 6 patients with anti-NOR90 were SSc [103]; however, another study showed seven out of 9 patients had SjS including primary and secondary SjS with RA [104]. Anti-NOR90 is not a common specificity, and the number of cases in each report is small; nevertheless, anti-NOR90 has been detected in various rheumatic diseases including SLE, RA, SjS, UCTD, and arthritis (Table 19). Despite the identification of high anti-NOR90 titers, 1:320–1:12,800, no specific association with a particular disease or clinical features has been confirmed yet. The detection of anti-NOR90 in malignancy was also reported, however, at low frequency [105–108]. Anti-NOR-90 tested by LIA has been reported (Table 20); however, technical limitations on sensitivity and specificity of the assay should be noted because this assay has not been fully validated compared with other immunoassays.

Fig. 4
figure 4

Immunoprecipitation using sera with anti-RNA polymerases or -NOR-90 antibodies. 35S-methionine-labeled K562 cell extract was immunoprecipitated by anti-RNA polymerases or -NOR-90 human autoimmune sera and protein components were analyzed by 8% SDS-PAGE and autoradiography. The two largest subunits of RNA polymerase I, II, and III are indicated on the left. Lane I/II/III is a serum with antibodies to all 3 RNA polymerases, whereas lane I/III is a serum with both anti-RNAPI and RNAP III. Lane I > III is a serum with predominant anti-RNAP I compared with anti-RNAP III. NOR-90 proteins are shown with red dots

Full size image
Table 19 Prevalence of anti-NOR90 antibodies
Full size table
Table 20 Prevalence of anti-NOR-90 by line immunoassay
Full size table

ASE-1

ASE-1 was described as a nucleolar autoantigen similar to NOR-90 immunofluorescence pattern and intracellular localization, with similar mobility in SDS-PAGE. ASE-1 was identified as a 55 kD protein based on its sequence; however, it migrates ~ 90 kD close to NOR-90 [109].

In one study, anti-ASE-1 was evaluated by immunofluorescence, WB, and IP of in vitro transcription and translation (TnT) product, compared with anti-NOR90 reactivity [108]. Four of 23 sera were anti-ASE-1 positive, 6 were anti-ASE-1, and NOR-90 positive, 10 were only anti-NOR-90 positive. Thus, in addition to similar localization and mobility, these two specificities also coexist quite often. Clinical features were available in 7 patients with anti-ASE-1 who showed 3 slowly progressive SSc with Raynaud’s phenomenon (one also had lung cancer), 3 patients with malignancy (2 lung cancer, 1 melanoma), and 2 had seronegative arthritis.

Another study reported ASE-1 as an autoantigen in SLE. Anti-ASE-1 was tested in 92 SLE, 35 RA, 50 SSc, and 100 controls by WB using 3 recombinant proteins [110]. The reactivity with each recombinant protein was 42–55% with SLE sera, while it was 11–31% in RA, 7–14% in SSc, and 15–21% in the unaffected community controls. Reactivity to all 3 proteins was seen in 25% of SLE, 8.5% of RA, 6% of SSc, and 1% of control, and in SLE, it was associated with serositis.

Anti-RNA Polymerase I Antibodies

It has been suggested that nucleolar staining pattern by IIF is often seen in SSc; however, the fine specificity was not described until the late 1980s [64, 111]. One of the reasons for the delay in characterizing nucleolar SSc antigen may be because none of the SSc-specific ANoA make precipitin lines in DID that have been used to detect and define autoantibody specificities until the 1980s. The first description of antibodies to RNA polymerases (RNAP) was anti-RNA polymerase I antibodies as one of the specificities for the ANoA in SSc [111]. There were a series of earlier studies reporting anti-RNA polymerase I antibodies by a solid-phase radioimmunoassay (RIA) in the majority of patients with SLE, MCTD, and RA, and lupus-prone MRL/lpr mice [112–115]. However, these data are quite different from the data in all studies after 1987 by radioimmunoprecipitation and ELISA, which show high specificity for SSc [49, 116–118]. Thus, earlier studies by solid-phase RIA appear to be detecting antibodies different than anti-RNAP I antibodies that coexist with anti-RNAP III antibodies and are specific for SSc as we currently detect.

RNA polymerase I localizes to nucleoli with punctate pattern (AC-10), whereas RNA polymerase III and II distribute in nuclei showing a coarse speckled pattern (AC-5) based on the study using monoclonal antibodies (mAbs) [119]. Predominant nucleolar staining by uncommon human sera that show IP of strong RNAP I with weak RNAP III [120] is consistent with the data by mAbs. Commercially available kits for anti-RNA polymerases are only for the detection of anti-RNAP III antibodies [118]. Thus, positive patients are expected to have a nuclear large/coarse speckled pattern (AC-5). Since virtually all anti-RNAP III sera also contain anti-RNAP I, nucleolar staining may be reported; however, reporting punctate nucleolar staining with a background of a coarse speckled pattern may require careful interpretation and experience [121].

Anti-RNA polymerase III immunoassay has become a part of standard screening tests when patients are suspected or diagnosed as SSc [49, 118]. Virtually, all anti-RNAP III sera also have anti-RNAP I, and the screening test is only for RNAP III. Commercial immunoassay for anti-RNAP I is not available. There are no studies separating the clinical significance of anti-RNAP I from that of anti-RNAP III because they coexist in virtually all cases. Thus, a well-established association of anti-RNAP III with SSc, mainly dcSSc often complicated with renal crisis, will be applied for anti-RNAP I.

Many sera immunoprecipitate RNAP III and RNAP I at comparable intensity; however, some sera predominantly immunoprecipitate anti-RNAP I with very weak RNAP III [120, 122]. Some of these patients may show atypical clinical presentation. These sera tend to be negative in anti-RNAP III ELISA, though anti-RNAP III ELISA can be positive up to 80–95% of anti-RNAP III IP positive sera in general [118].

Summary of Clinical Association

The association of each ANoA pattern, specific autoantibodies, and clinical diagnosis are summarized (Fig. 5). The association of anti-Th/To (AC-8), -PM-Scl (AC-8), -fibrillarin(U3RNP) (AC-9), and -RNAP I (AC-10) with SSc or SSc-OL is well established. Anti-Th/To and -PM-Scl are associated with lcSSc, while anti-fibrillarin and RNAP I are associated with dcSSc. Many anti-Th/To-positive patients do not represent typical SSc and may be considered ssSSc or UCTD or show only certain features related to SSc including ILD, PH, or Raynaud’s phenomenon. Limited studies reported anti-RNA helicase II/Gu (AC-8), -NOR90/hUBF, and ASE-1 mainly in patients with SARDs but not necessarily SSc. Several specificities with the AC-8 pattern, including anti-nucleolin/C23, nucleophosmin/B23, No55, and Nop-52/NNP1, have been reported in SARDs, malignancies, GVHD, and other diseases and clinical significance remains to be established.

Fig. 5
figure 5

AC patterns, specific autoantibodies, and clinical diagnosis. Association of each ANoA pattern, specific autoantibodies, and clinical diagnosis is summarized

Full size image

Conclusions

ANoA by immunofluorescence has a long history of over 60 years. Many target antigens of ANoA have been identified and characterized; however, there are still many ANoA-positive sera with unknown specificity. Among the characterized specific ANoA, many articles have been published on anti-Th/To, -PM-Scl, and -U3RNP/fibrillarin, and their clinical association with SSc or SSc overlap syndrome is well established. However, the clinical relevance of other specific ANoA is not well understood despite the fact that many of these specificities were identified decades ago. One important factor and a difference between these two groups are the roles of immunoassays. For anti-Th/To, PM-Scl, and U3RNP, reliable detection by IP (and DID for anti-PM-Scl) contributed significantly to establishing their clinical significance. IP is considered reliable but not available for most clinicians, whereas currently available immunoassays have limitations in sensitivity and specificity. Future important issues will be the development and validation of widely available immunoassay for these. For the latter group of ANoA with unknown clinical relevance, the establishment of widely available assay and systematic screening would help future clinical uses of these specificities.