Marvin J. Fritzler MD, PhD, Departments of Biochemistry and Molecular Biology, University of Calgary, 3330 Hospital Drive NW, Calgary, AB, Canada T2N-4 N1. E-mail: firstname.lastname@example.org
We studied the autoantigen targets of 75 human sera that had antibodies to the nuclear envelope (NE) as identified by indirect immunofluorescence (IIF) on HEp-2 cells. Several different IIF staining patterns could be identified when antibodies to different components of the nuclear membrane (NM) and nuclear pore complexes (NuPC) were identified: a smooth membrane pattern characteristic of antibodies to nuclear lamins, a punctate pattern typical of antibodies to the nuclear pore complex and more complex patterns that included antibodies to nuclear and cytoplasmic organelles. Western immunoblotting of isolated nuclear and NE proteins and immunoprecipitation of radiolabelled recombinant proteins prepared by using the full-length cDNAs of the Translocated promoter region (Tpr), gp210 and p62 were used to identify specific autoantibody targets. Fifty-two of the 75 (70%) sera bound to Tpr, 25 (33%) bound to lamins A, B or C, 15 (20%) reacted with gp210 and none reacted with p62. Sixteen (21%) did not react with any of the NE components tested in our assays. The clinical features of 37 patients with anti-NE showed that there were 34 females and three males with an age range of 16–88 years (mean 59 years). The most frequent clinical diagnosis (9/37 = 24%) was autoimmune liver disease (ALD; two with primary biliary cirrhosis), followed by seven (19%) with systemic lupus erythematosus (SLE), four (11%) with a motor and/or sensory neuropathy, three (8%) with anti-phospholipid syndrome (APS), two with systemic sclerosis (SSc), two with Sjögren's syndrome (SjS), and others with a variety of diagnoses. This report indicates that Tpr, a component of the NuPC, is a common target of human autoantibodies that react with the NE.
Autoantibodies to intracellular proteins and macromolecular complexes that are components of various cellular structures are a hallmark of systemic autoimmune diseases. Autoantigen targets have been identified in cytoplasmic organelles such as centrosomes [1–3], endosomes [4,5], lysosomes , the Golgi complex [7,8], mitochondria  and glycine (G)- tryptophan (W)- rich cytoplasmic structures (GW bodies) , as well as nuclear structures that include kinetochores, nucleoli and chromosomes (reviewed in [11–13]). The nuclear envelope (NE) is another target of the autoimmune response in patients with autoimmune liver disease, systemic lupus erythematosus and other conditions [14–20].
Structurally, the nuclear envelope is a bilayered membranous structure that can be divided into five distinct components . These include the inner nuclear membrane (INM) with a distinct set of integral membrane proteins; the outer nuclear membrane (ONM); a perinuclear space, which is continuous with the lumen of the endoplasmic reticulum (ER); the pore domains, regions where the INM and ONM come together and fuse; and an underlying nuclear lamina, containing the nuclear lamins. Large channels, termed nuclear pore complexes (NuPCs), reside within the pore domains. Until recently, the ONM was considered to be an extension of the surrounding ER, but the identification of an integral membrane protein distinct to the ONM suggests that this NE component has unique properties . The NE plays key roles in maintaining nuclear structure and integrity, while the NuPCs regulate the bidirectional traffic of molecules between the nucleus and the cytoplasm.
In the past 3 years there has been a tremendous increase in our understanding of the specific components of the NE [23,24]. One of these components is the 265 kDa Translocated promoter region (Tpr) protein identified as a component of the distal ring of the NPC nuclear basket and the intranuclear filaments that, in Xenopus, extend 350 nm into the nucleus . More recent data have also shown that Tpr is a component of the nuclear basket, located near the nuclear ring adjacent to the INM and that it functions in nuclear export of messenger RNA (mRNA) [26–28].
Autoantigens in the NE have been identified in three of the five NE compartments. These include: lamins A, B and C (the nuclear lamina), gp210, p62 complex proteins and Tpr (all associated with the NuPC) and LBR, MAN1, LAP1 and LAP2 (all integral membrane proteins of the INM) [14–19]. In this study we show that the majority of sera in our cohort, which had antibodies to the NE as detected by routine indirect immunofluorescence (IIF) on HEp-2 cell substrates, had antibodies to Tpr.
IIF analyses utilized a commercially prepared HEp-2 cell substrate kit (HEp-2000TM, Immuno Concepts, Sacramento, CA, USA) that included a heavy chain specific, fluorescein-conjugated goat antihuman IgG as described previously . All human sera used in this study were obtained from serum banks at the Advanced Diagnostics Laboratory, University of Calgary. The sera were selected on the basis of an IIF staining pattern of the nuclear membrane and/or nuclear pore complex. Clinical data were obtained by retrospective chart review.
Rabbit polyclonal antibodies directed against the first 150 amino acids of Tpr was a generous gift from Dr Larry Gerace (the Scripps Research Institute, La Jolla, CA, USA). Human autoimmune sera against gp210 were kindly provided by Dr Jean Claude Courvalin (Institut Jacques Monod, Paris, France). Other index autoantibodies to nuclear lamins, GW bodies and mitochondria were from the Advanced Diagnostics Laboratory at the University of Calgary. Secondary antibodies included Cy3-conjugated antirabbit IgG (Jackson Immuno Research Laboratories, Inc., West Grove, PA, USA). Slides were viewed on a Leitz or a Zeiss Universal microscope fitted with a TEC470 CCD video camera system (Optronics Engineering, Goleta, CA, USA) and the appropriate filter sets for FITC, Cy3 and DAPI (4’, 6-diamino-2-phenylindole).
In vitro transcription/translation and immunoprecipitation
Tpr, gp210 and p62 cDNAs (gifts from Dr Brian Burke, University of Florida, Gainesville, FL, USA) were used as templates for in vitro transcription and translation (TnT, Promega, Madison, WI, USA) in the presence of[35S]-methionine as described previously [30,31]. TnT reactions were conducted at 30°C for 1·5–2 h and the presence of translation products was confirmed by subjecting 2–5 µl samples to sodium dodecyl suplphate-polyacrylamide gel electrophoresis (SDS-PAGE) and analysis by autoradiography. The in vitro translated products were then used as substrate in immunoprecipitation (IP) reactions.
IP reactions were prepared by combining 100 µl 10% protein A-Sepharose beads (Sigma), 10 µl human serum, 500 µl NET2 (containing NaCl, EDTA and Tris) buffer (50 mm Tris-HCl, pH 7·4, 150 mm NaCl, 5 mm EDTA, 0·5% nonidet P-40, 0·5% deoxycholic acid, 0·1% SDS, 0·02% sodium azide), and 5–10 µl of labelled recombinant protein obtained from the TnT reaction described above. After 1 h of incubation at 4–8°C, the Sepharose beads were washed five times in NET2 and the proteins eluted in 10 µl of sample buffer. The proteins were analysed by 10% or 12·5% SDS-PAGE as described previously .
Isolation of nuclei and nuclear pore complex components
HeLa cells were grown in T75 tissue culture flasks in Dulbecco's modified Eagle medium (DMEM) (Gibco BRL, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum and antibiotics (Gibco/BRL). When cells were confluent, they were washed once with ice-cold phosphate buffered saline (PBS) containing 8% sucrose and twice with ice-cold water. After this step, all procedures were performed on ice. To each of T75 flask cells, 5 ml of PBS containing 2 mm PMSF and 1% Triton X-100 (buffer A) was added. Cells were then incubated for 5 min with gentle shaking on a shaking platform. The cell lysate was decanted from the flask, an additional 5 ml of buffer A was added to the flasks, and cell suspension incubated for another 5 min with shaking. Cell nuclei were the detached from the flask by vigorous shaking and collected by centrifugation at 1500 r.p.m. (Beckmann bench-top centrifuge) for 10 min at 4°C. The cell nuclei were purified further by sucrose gradient centrifugation as described . Nuclear pore complex components were solubilized and prepared from the nuclear membranes by using Empigen BB (Calbiochem) detergent as described by Cronshaw et al. .
Nuclei and nuclear pore complex components were denatured in SDS sample buffer by boiling for 10 min and separated on a 10% SDS-polyacrylamide gel, and then transferred to nitrocellulose membranes as described previously . Three to four mm strips were cut from the membrane, incubated with the human sera that were diluted 1 : 200 in PBS containing Tween 20 (PBS-T) containing 5% milk and the immune-reactive bands were visualized using an enhanced chemiluminescence kit (Amersham Biosciences, Piscataway, NJ, USA) as described previously . Prototype human sera with antibodies to gp210 and lamins A/B/C, rabbit antibodies to recombinant p62 and Tpr, a monoclonal antibody to lamin B (gift of Dr Michael Pollard, the Scripps Research Institute, La Jolla, CA, USA) were used as controls.
Addressable laser bead assay
The reactivity of the sera with other autoantigens was determined by an addressable laser bead assay (QuantaPlex8 and liver panel kits, Inova Diagnostics, San Diego, CA, USA) using a Luminex100 flow fluorometer (Luminex Corp., Austin, TX, USA). The antigens in the kits provided by Inova included chromatin, Sm, U1-RNP, SS-A/Ro, SS-B/La, ribosomal P protein, Jo-1, soluble liver antigen, liver kidney microsomal antigen and pyruvate dehydrogenase complex (M2). Antibodies to GW-182 were detected by addressable laser bead assay using purified recombinant GW182 coupled to a set of beads as described in detail elsewhere .
The 75 sera selected for this study demonstrated a variety of patterns of reactivity with the HEp-2 NE (Fig. 1). One pattern was represented by a smooth staining of the nuclear membrane with increased intensity of the periphery of the nucleus and accentuation of folds in the NE (Fig. 1a). Another common pattern was punctate or stippled staining of the NE (Fig. 1b). The other patterns included the first two patterns and staining of cytoplasmic and/or nuclear components. In some sera, there was robust staining of the NE and discrete cytoplasmic structures that were shown by co-localization to be GW bodies (Fig. 1c) , while others had staining consistent with antimitochondrial antibodies (Fig. 1d).
When the sera were tested by immunoblotting of purified HeLa nuclei and NuPCs, a variety of reactive protein bands were visualized (Fig. 2). Twenty-five (33%) of the sera bound to 60–70 kDa proteins, and this was consistent with reactivity to lamins A, B or C (Fig. 2b) (Table 1). Other sera bound to >200 kDa proteins that were identified tentatively as gp210 and Tpr (Fig. 2a). Some sera identified other lower molecular weight protein bands that did not co-migrate with prototype sera used in this study.
Table 1. Anti-NE profile of 75 sera with IIF NE staining on HEp-2 cells
When the 75 sera were tested for binding to gp210, p62 and Tpr in a TnT-IP assay that employed the purified cDNAs to produce full-length radiolabelled recombinant proteins, some of the reactivities observed by immunoblotting were elucidated (Table 1). Fifty-two of the sera (70%) bound to Tpr (Fig. 3a), 15 (20%) reacted with gp210 (Fig. 3b) and none reacted with p62 (data not shown). Sixteen (21%) sera did not react with any of the NE components tested in this study and these sera did not demonstrate a unique IIF staining pattern.
When combinations of NE autoantibodies were evaluated by immunoblotting and immunoprecipitation of recombinant proteins, 14/75 (19%) reacted with lamins and Tpr; 10/75 (13%) reacted with gp210 and Tpr; 3/75 (4%) reacted with lamins and gp210; and 3/75 (4%) reacted with Tpr, gp210 and lamins. Twenty-five (33%) reacted with Tpr alone; five (7%) reacted with lamins alone; and two (3%) reacted with gp210 alone. When these reactivities were compared to the IIF staining patterns observed on the HEp-2 cells, no consistent correlations could be identified except for sera that contained antibodies to lamins or Tpr alone. Sera containing antibodies to lamins alone demonstrated the NE IIF pattern illustrated in Fig. 1a and sera that had antibodies to Tpr alone produced the NE IIF pattern illustrated in Fig. 1b.
When the 37 sera that bound recombinant Tpr were tested for other autoantibodies in an addressable laser bead assay, eight reacted with SS-A/Ro, five with SS-B/La, six with chromatin, four with M2 (pyruvate dehydrogenase complex), one with anti-U1RNP and one with ribosomal P protein (Table 2). Three patients had elevated rheumatoid factor, three had a positive anticardiolipin antibody (ACA) test and one had antibodies to both double-stranded DNA and β2 glycoprotein I. Two of the patients with ACA had antibodies to Tpr and none had antibodies to lamins. It was also noted that none of the sera with antibodies to lamins and a peripheral pattern of nuclear staining had antibodies to dsDNA, although three had antibodies to chromatin. There was no correlation of antibodies to chromatin with the IIF rim pattern of staining. Indeed, reference to Table 2 indicates that reactivity with components of the NE or NuPCs were not always correlated with a distinctive staining pattern. For example, sera that had antibodies to lamins and mitochondria often had a punctate NE IIF staining pattern (Fig. 1c).
Table 2. Clinical, demographic and autoantibody profile of 37 patients with antibodies to the nuclear envelope
Retrospective chart review yielded demographic and clinical information on 37 of the patients (Table 2). The age range of the patients was 16–88 years (mean = 59), 34 female and three male. The most frequent clinical diagnosis (9/37 = 24%) was autoimmune liver disease (ALD) [two with primary biliary cirrhosis (PBC)], followed by seven (19%) with systemic lupus erythematosus (SLE), four (11%) with a motor and/or sensory neuropathy, three (8%) with anti-phospholipid syndrome (APS), two with systemic sclerosis (SSc), two with Sjögren's syndrome (SjS) and others with a variety of diagnoses. All APS patients had at least two thrombotic events during the course of their clinical follow-up, two to three had anticardiolipin antibodies, and one (patient 26) was a 36-year-old female with a history of myocardial infarction and a spontaneous abortion (gravida 2, para 1, abortus one, living one). Two had a diagnosis of fibromyalgia but none of the patients had a diagnosis of chronic fatigue syndrome (CFS).
When the clinical diagnosis of the patients was evaluated in the context of specific autoantibodies, it was noted that only one of nine of the ALD patients had antibodies to Tpr and two had antibodies to lamins A/C. Thus, the majority of ALD patients with anti-NE antibodies as detected by IIF had autoantibodies to antigens not measured or detected in this study. By contrast, the majority of patients with anti-Tpr had SLE, SSc or SjS. Of interest, two patients (nos 33 and 37), both of whom had antibodies to Tpr, had antibodies to a recently identified cytoplasmic autoantigen, GW 182 . One of these patients had PBC and arthritis and the other had a neuropathy and SjS.
Autoantibodies directed to the NE target antigens have been detected in at least three of the five domains of the NE. These include antibodies to LAP1, LAP2, MAN1 and LBR of the INM domain; antibodies to Tpr, p62 and gp210 of the nuclear pore domain; and antibodies to the nuclear lamina (lamins A, B, C) [18,19,36–38]. Yet to be identified are autoantibodies directed to the nesprins and NUANCE, which are components of the INM and ONM, respectively [22,39]. Nesprins, NUANCE and Tpr are interesting candidates for an autoantibody target because they have extensive central coiled-coil domains [22,39,40]; and an increasing number of target autoantigens in the cytoplasm are characterized by coiled-coil domains . Nesprins and NUANCE also have a short trans-membrane region at the C-terminus and NUANCE has an actin-binding domain at the N-terminus [22,39]. It is noteworthy that extensive coiled-coil domains and a short trans-membrane region are key motifs identified in gp210  and giantin/macrogolgin, which is also the most common target of sera that have anti-Golgi antibodies . The identity of NUANCE and some nesprins as autoantigens is a challenge due to their high molecular mass, but will probably be elucidated with the availability of complete cDNAs and other specific reagents (i.e. monoclonal antibodies).
From a clinical perspective, the most widely studied NE antigens are the nuclear lamins, which were found primarily in the sera of SLE and PBC . Antibodies to gp210 and p62 have been associated with PBC patients [19,43,44] and other conditions (reviewed in [9,20]). In a study similar to the present one by Gerace and his colleagues, among 55 sera from predominantly rheumatic disease patients with anti-NE as detected by IIF, 31% reacted with lamins A/C, 15% with lamin B, 8% with gp210, and 9% and 29% reacted with the integral membrane proteins LAP1A and LAP2, respectively . In addition, they found that 9/55 sera (16%) reacted with a ‘peripheral’ 175 kDa protein band that was believed to be a proteolytic product of Tpr. They also concluded that antibodies to LAP2 may be among the prominent autoantigens of the NE in rheumatic disease sera. Although our study did not include specific analysis of LAP1 or LAP2, when compared to the Gerace study , we found a remarkably high frequency of antibodies to Tpr (70%versus 16%) and a lower frequency of antibodies to lamins (33%versus 46%) and gp210 (20%versus 8%). However, our study differs in a number of respects. First, Gerace and his colleagues used Western immunoblot and prototype sera to identify reactive bands, whereas we used immunoprecipitation of recombinant proteins derived from the TnT product of full length cDNAs. In general, IP of recombinant proteins is a more sensitive technique than IB (unpublished observations). It is possible that the 175 kDa ‘proteolytic product’ observed by the Gerace group was missing a key reactive epitope. In addition, the patient population studied by Gerace was composed apparently of patients with predominantly rheumatic diseases, whereas our sera were unselected from a serum bank maintained in the Advanced Diagnostics Laboratory at the University of Calgary. The clinical referral base to our laboratory includes a wide spectrum of clinical specialties including neurology, haematology, rheumatology and clinical immunology.
In a study of a cohort of 60 patients with chronic fatigue syndrome (CFS), 52% were found to have antibodies to lamin B1 but antibodies to lamins A/C, gp210, Tpr and LAP2 were not found . In our study, none of the patients had a clinical diagnosis of CFS, although two patients had fibromyalgia, which can be confused with CFS . It is notable that the study of the CFS sera utilized HeLa and HEp-2 cells grown in culture as well as commercially available HEp-2 cells. However, the authors did not indicate which fixatives (if any) were employed, if there were differences in the sensitivity of commercially prepared HEp-2 cells, and if permeabilizing agents (i.e. detergents) were used, both of which can potentially produce results different from our study that used only commercially available HEp-2 cells that are fixed in acetone/methanol and are used widely in routine clinical laboratories. In addition, the titres observed in the sera of CFS sera were generally 1/50 or less, which falls below the dilution of 1/80 used routinely in our clinical laboratory . Thus our study reflects more closely the experience of clinical laboratories that commonly use commercial kits to screen sera for autoantibodies .
Several studies have identified autoantibodies to p62 [18,44,48] where they were thought to be a relatively specific marker for PBC. A recent report used a recombinant fusion protein to identify antibodies to p62 in the sera of a patient with mixed connective tissue disease . Another study of a cohort of 115 PBC patients used the lectin-binding properties of p62 as an approach to the identification of antip62 antibodies in 13% of PBC sera . In our study that used a prototype serum as positive control, antip62 antibodies were not detected when a radiolabelled protein was synthesized from a full-length p62 cDNA. Our cohort included a number of patients with ALD but only 2 had PBC. An explanation for this is that the Advanced Diagnostics Laboratory does not evaluate routinely the sera of patients with PBC but it receives these sera when conventional autoantibody testing [i.e. anti-mitochondrial antibodies (AMA)] is negative or if the patients have other atypical features. This may represent a referral bias, but based on the study of NE/NuPC patterns of staining detected by IIF on commercially prepared HEp-2 substrates, we conclude that antip62 antibodies are rare in cohorts selected on the basis of similar criteria.
In the present study of 75 sera with antibodies to NE components as detected initially by IIF on HEp-2 cells, antibodies directed against Tpr were the most common. Tpr was thought originally to be part of the cytoplasmic face of the NuPC  but was identified subsequently as a component of the inner nuclear face . To date, five proteins have been localized to the cytoplasmic face of the NuPC: Nup358/RanBP2 , CAN/Nup214 [51,52], Nup84/88 ; and possibly NLP1  and ALADIN . Nup358 has been mapped to the cytoplasmic filaments extending out from the NuPC, whereas Nup214 has been mapped to the cytoplasmic coaxial ring . It is probable that Nup84 also resides in the coaxial ring, as it is tightly associated with Nup214 . Two integral membrane proteins have been localized to the pore membrane domain: Pom121  and gp210 . The central channel appears to be made up, at least in part, by the p62 complex of proteins (p62, p58, p54, and p45 ).
In our study, we concentrated on NE antigens of the NuPC but used antibodies to lamins as a reference point for other studies. As noted earlier, a number of other NE proteins, particularly NUANCE and nesprins, are also probable targets of human autoantibodies. With the description of the NE proteome [23,24] and disease associations [59–61], more comprehensive studies of multiple autoantigens will probably be performed in the near future. Technologies such as addressable laser bead arrays  will make studies of large numbers of samples from multiple centres less daunting than earlier technologies of IIF, IP, ELISA and immunoblotting.
We acknowledge the assistance of Meifeng Zhang and Mark Fritzler in laboratory studies, and Whitney Steber who collected the clinical information. This research was supported by the Canadian Institutes of Health Research (grant no. MOP 38034) and a grant from the Natural Sciences and Environment Research Council to J.B.R. M.J.F. holds the Arthritis Society Research Chair at the University of Calgary. P.E. was supported by a MD/PhD studentship from the Alberta Heritage Foundation for Medical Research and Y.O. is supported by a CIHR/Ernst and Young postdoctoral award.