Clinical and Investigative Medicine

 

High frequency of neoplasia in patients with autoantibodies to centromere protein CENP-F

Jerome B. Rattner, PhD
Jennifer Rees, BSc
Clark M. Whitehead, BSc
Carlos A. Casiano, PhD
Eng M. Tan, MD
René L. Humbel, MD
Karsten Conrad, MD
Marvin J. Fritzler, PhD, MD

Clin Invest Med 1997;20(5):308-319

[résumé]

Dr. Rattner, Ms. Rees, Mr. Whitehead and Dr. Fritzler are with the Departments of Anatomy, Medical Biochemistry and Medicine, University of Calgary, Calgary, Alta.; Dr. Casiano and Dr. Tan are with the W.M. Keck Autoimmune Disease Centre, Department of Molecular and Experimental Medicine, Scripps Research Institute, LaJolla, Calif.; Dr. Humbel is with the Laboratoire de Biochimie-Immunopathologie, Centre Hospitalier de Luxembourg, Luxembourg; and Dr. Conrad is with the Institute of Immunology, Dresden Technical University, Dresden, Germany.

(Original manuscript submitted Jan. 15, 1997; received in revised form June 2, 1997; accepted June 20, 1997)

Reprint requests to: Dr. Jerome B. Rattner, Department of Medical Biochemistry, The University of Calgary, 3330 Hospital Dr. NW, Calgary AB T2N 4N1; fax 403 270-0737; rattner@acs.ucalgary.ca

Contents

Abstract

 Objective: To study the clinical features of patients with autoantibodies to centromere protein CENP-F and the frequency of CENP-F autoantibodies in patients with various diseases.

Design: Retrospective clinical and serologic study.

Methods: Thirty-six patients with anti-CENP-F were identified by a characteristic pattern of indirect immunofluorescence (IIF) on HEp-2 cells. Fifty patients with melanoma, 50 with breast cancer, 10 with lung cancer, 354 with systemic sclerosis, 120 with systemic lupus erythematosus and 50 with rheumatoid arthritis were also studied. Recombinant proteins were produced from 5 CENP-F cDNA clones representing amino acids 2192-3317 (p-F1), 5561-7126 (p-F2), 5892-6883 (p-F3), 7538-10116 (p-F4) and 9242-10096 (p-F5). The presence of CENP-F antigen was studied in a breast carcinoma cell line, cryosections of breast carcinoma, normal breast tissue and tonsils.

Results: Twenty-two of 36 patients with CENP-F antibodies had neoplasms; breast (9/22) and lung (5/22) cancer were the most common diagnoses. Thirty-three sera were available for further study; when tested for reactivity to the recombinant peptides, the sera of 21 of 21 patients with neoplasms and 5 of 12 patients with other diseases bound the C-terminal p-F4 peptide. When the terminal third of the p-F4 peptide (p-F5) was studied, a significant difference in pattern of reactivity was not detected. By comparison, the frequency of reactivity with peptides representing other domains of CENP-F was less than that with p-F4 (p-F2 > p-F3 > p-F1). CENP-F autoantibodies were not found in any of the control sera from patients with systemic lupus erythematosus, rheumatoid arthritis or systemic sclerosis or in unselected sera from various malignancies. CENP-F antigens were identified in breast carcinoma tissue but were rarely observed in normal tissues.

Conclusions: A high proportion of individuals with CENP-F antibodies have neoplasia, and there is a bias among their sera for reactivity with determinants in the carboxy terminal domain of CENP-F. CENP-F antigens appear to be highly expressed in malignant tissues.

[ Top of document ]

Résumé

Objectif : Étudier les caractéristiques cliniques de patients qui ont des auto-anticorps de la protéine CENP-F du centromère et la fréquence des auto-anticorps CENP-F chez des patients qui ont diverses maladies.

Conception : Étude clinique et sérologique rétrospective.

Méthodes : On a identifié 36 patients atteints d'anti-CENP-F au moyen d'un tracé caractéristique d'immunofluorescence indirecte (IFI) sur les cellules HEp-2. On a étudié aussi 50 patients atteints d'un mélanome, 50 d'un cancer du sein, 10, d'un cancer du poumon, 354 souffrait de sclérodermie généralisée, 120, de lupus érythémateux disséminé et 50, de polyarthrite rhumatoïde. On a produit des protéines recombinantes à partir de clones de 5 CENP-F de l'ADNc représentant les acides aminés 2192-3317 (p-F1), 5561-7126 (p-F2), 5892-6883 (p-F3), 7538-10116 (p-F4) et 9242-10096 (p-F5). On a étudié la présence de l'antigène CENP-F dans une souche de cellules de cancer du sein, des cryosections de cancer du sein, des tissus normaux du sein et des amygdales.

Résultats : Vingt-deux des 36 patients qui avaient des anticorps CENP-F avaient des néoplasmes; on a diagnostiqué surtout des cancers du sein (9/22) et du poumon (5/22). Trente-trois sérums étaient disponibles pour une étude plus poussée. Lorsqu'on en a analysé la réactivité aux peptides recombinants, les sérums de 21 patients sur 21 qui avaient des néoplasmes et de 5 patients sur 12 qui avaient d'autres maladies ont lié le peptide p-F4 C-terminal. Lorsqu'on a étudié le tiers terminal du peptide p-F4 (p-F5), on n'a pas détecté de différence significative dans la tendance à la réactivité. Comme comparaison, la fréquence de la réactivité avec des peptides représentant d'autres domaines du CNP-F a été inférieure à celle de la réactivité avec le p-F4 (p-F2 > p-F3 > p-F1). On n'a pas trouvé d'auto-anticorps CENP-F dans aucun des sérums témoins provenant de patients atteints de lupus érythémateux disséminé, de polyarthrite rhumatoïde ou de sclérodermie généralisée, ou dans des sérums non sélectionnés provenant de diverses tumeurs malignes. On a identifié des antigènes CENP-F dans des tissus de cancer du sein, mais rarement dans des tissus normaux.

Conclusions : Une proportion élevée d'individus qui ont des anticorps CENP-F ont des néoplasmes et ces sérums présentent un biais en faveur de la réactivité avec des déterminants dans le domaine carboxyle terminal du CENP-F. Les antigènes CENP-F semblent très fréquents dans les tumeurs malignes.

[ Top of document ]

Introduction

 The mammalian centromere is a multifunctional chromosomal domain that plays a major role during cell proliferation, mediating chromosome movement and sister chromatid association. Six unique mammalian centromere proteins (CENPs) have been identified, and all have been shown to be target antigens of human autoantibodies.1­6 While autoantibodies to CENP-A, -B, -C and -E are commonly associated with the limited or CREST variant of scleroderma,1,5,6 autoantibodies to CENP-D and -F appear to be less prevalent. Clinical information has been reported on only a few individuals with these antibodies.2­4

CENP-F is a 367-kDa nuclear matrix protein that accumulates to a maximum in G2 cell nuclei, localizes to the outer layer of the kinetochore at late G2-M, and is subsequently found at the spindle midzone and within the intercellular bridge during the latter part of mitosis.2,3,7 In earlier studies, CENP-F autoantibodies were defined as part of a subgroup of autoantibodies that produced a speckled pattern of staining on metaphase cells that resembled the pattern seen with CENP-A, -B, -C and -D antibodies, but was distinguishable from these by the absence of staining of interphase cells.8 Subsequent studies showed that the CENP-F pattern of reactivity differs in G2- and M-phase cells, thus allowing the identification of 2 distinct populations in cycling cells.2,3 By comparison, CENP-E, a 312-kDa member of the kinesin family of proteins, is first detected in prometaphase cells and is associated with the outer layer of the kinetochore. At anaphase, it appears to move from the centromere to the central region of the mitotic spindle and later to the intercellular bridge, which connects the 2 daughter cells at the completion of cell division.9,10 Finally, of the 6 centromere proteins, 4 (CENP-A, -B, -C and -D) are found in association with the centromere throughout the cell cycle, whereas the other 2 (CENP-E and -F) are found only at the centromere during the latter stages of the cell cycle.

A recent study of 26 individuals with serum autoantibodies to CENP-F revealed that 14 (54%) of these individuals had malignancy of various types.11 This study extends the characterization of the sera of these patients and supplements this group with an additional 10 patients. We show that the sera of these patients contain autoantibodies to multiple epitopes on CENP-F and that malignancy is a frequent clinical feature of CENP-F-reactive individuals. Among CENP-F-positive patients with malignancy, autoantibodies primarily react with C-terminal determinants. Since malignant tissue contains a large number of CENP-F reactive cells, it is a potential antigen source.

[ Top of document ]

Materials and methods

Case identification and clinical information

We performed routine analysis of 2500 sera by indirect immunofluorescence (IIF) using a commercially available HEp-2 cell substrate (Immuno Concepts Inc., Sacramento, Calif.) during an 18-month period, and identified 8 sera that had autoantibodies to CENP-F in the Advanced Diagnostic Laboratory at the University of Calgary. Sera from an additional group of 26 patients from 2 other centres identified in a recent study11 and 2 sera from another centre in Dresden, Germany, were also included in this study. In addition to the typical CENP-F staining pattern on HEp-2 cells,2,3,7 these sera also reacted with a 367-kDa HeLa cell protein on immunoblots, as previously reported.2,7

To determine if CENP-F antibodies were a common feature of malignancy, 50 sera from consecutive patients with breast cancer, 10 from patients with small-cell lung carcinoma (SCLC) and 50 from patients with malignant melanoma were also studied. In addition, to determine the frequency of CENP-F antibodies in rheumatic diseases, 354 sera from patients with systemic sclerosis (SSc), 120 sera from patients with systemic lupus erythematosus (SLE) and 50 sera from patients with rheumatoid arthritis (RA) were studied for the presence of CENP-F antibodies.

Tissue specimens and tissue culture cells

Cryopreserved tissue samples were collected through the Foothills Hospital, Calgary, Alta., from consecutive, unselected patients undergoing either radical mastectomy for treatment of breast cancer or reduction mammoplasty. All breast cancer specimens described in this report were determined by standard pathologic criteria to be ductal carcinoma. As a control for proliferating, non-neoplastic cell populations, tonsil tissue specimens were collected from routine tonsillectomies. Tonsil and breast tissues were quick-frozen in liquid nitrogen, and frozen sections were cut and stored at -20°C. The breast carcinoma cell line HTB-132 (given to us by Dr. D. Fujita, University of Calgary) was grown in DMEM supplemented with 10% fetal calf serum. The cells were seeded onto coverslips 48 hours before use.

Indirect immunofluorescence (IIF)

Frozen sections or coverslips containing HTB-132 cells were fixed for 10 minutes in cold 3% paraformaldehyde in Dulbecco's phosphate buffered saline (D-PBS) solution. Fixed preparations were washed in D-PBS solution and then incubated for 1 hour at 37°C in a 1:100 dilution of a prototype CENP-F serum.2 For some experiments, specimens were incubated in a mixture of a CENP-F serum diluted 1:100 and a human serum containing autoantibodies to centrosomes.12 After 3 washes in D-PBS solution, the samples were incubated for 1 hour at 37°C in secondary antibody: a fluorescein conjugated goat antihuman IgG (H+L) (Dakopatts, Mississauga, Ont.). In some studies a monoclonal antibody to the proliferation antigen Ki-67 (Dakopatts) was used as a reference marker. After incubation, the specimens were washed in D-PBS solution, counterstained with the nuclear dye DAPI (4´,6-diamindino-2-phenyl-indole), mounted in 90% glycerol-containing paraphenylenediamine, and observed with a Nikon Optophot fluorescent microscope. Images were recorded on Ilford HP-5 film.

cDNA cloning and sequence analysis

Cloning of CENP-F has been previously reported (Accession # U19769).7 The fragments used in this study represent regions that are most commonly reactive with human autoimmune sera. These fragments, designated p-F1-5, were obtained by screening a human breast carcinoma using published protocols.13 Briefly, protein expression was induced with isopropylthiogalactoside (IPTG) and the filters screened with 1:1000 dilution of a CENP-F-reactive serum. Reactive clones were detected using the ECL system (Amersham, Little Chalfont, England), and positive plaques were subsequently purified. The phagemid DNA was extracted, the inserted fragment amplified using polymerase chain reaction and the integrity of the fragments verified by the dideoxy chain termination protocol14 using the dsDNA Cycle Sequencing System (Gibco BRL, Life Technologies). DNA sequences were compiled using SeqEd software (ABI, Foster City, Calif.) or the GCG program, as described elsewhere.15 Sequences for the clones p-F1-5 were identical to those registered for CENP-F in GenBank.

Recombinant protein production and generation of CENP-F antibodies

The fragments of CENP-F corresponding to nucleotide positions 2192-3317, 5561-7126, 5892-6883, 7538-10116 and 9242-10096 were subcloned into the EcoR1 site of the vector pGex 5X-3 (Pharmacia, Baie d'Urfé, Que.) and used to produce recombinant proteins containing a 28-kDa glutathione S-transferase leader sequence. The recombinant proteins were prepared from 200-mL cultures of bacteria cells induced with 0.1 mmol/L IPTG. The immunoreactivity of extracted recombinant protein was confirmed by immunoscreening13,15 with the prototype anti-CENP-F serum. Fusion and cleaved recombinants were used; both gave the same result. In some cases, fragments with overlaps were used to confirm results and to identify the reactive determinants more precisely.

Bacterial lysates containing IPTG-induced recombinant protein p-F2 (corresponding to nucleotide positions 5561-7126) were boiled in SDS sample buffer and the proteins separated by SDS PAGE. Recombinant protein was sliced from the gel after staining with in 0.1% Coomassie Brilliant Blue, 40% methanol and 10% acetic acid. Gel slices were macerated and the recombinant protein extracted with a minimal volume of buffer containing 50 mmol/L Tris-hydrochloric acid, pH 8.0, 1 mmol/L EDTA, 1% mercaptoethanol, 1% SDS and 150 mmol/L sodium chloride at 65°C for 2 hours. Extraction was repeated 3 times. Supernatants containing the recombinant protein were pooled, dialyzed against water and lyophilized. Approximately 500 µg of lyophilized protein were resuspended in 0.5 mL PBS, mixed with an equal volume of Freund's complete adjuvant and injected into rabbits. Each rabbit was boosted 4 times at 2- to 3-week intervals before serum was tested by IIF and immunoblotting. A rabbit antiserum, designated 911, displayed a high-titre IIF staining identical to that of human anti-CENP-F autoantibodies and was used for this study. In some cases, the sera were adsorbed with Escherichia coli proteins.

Immunoblotting

Recombinant CENP-F proteins were resuspended in SDS sample buffer and denatured by boiling for 5 minutes. After gel electrophoresis on 8% SDS-PAGE, proteins were transferred to nitrocellulose, blocked and screened with the relevant human sera at 1:1000 dilutions.15 Immunoreactive bands were visualized using the ECL (Amersham, Arlington Heights, Ill.) system. The secondary antibodies were HRP-conjugated goat antihuman immunoglobulins diluted to 1:2000 in D-PBS solution.

In vitro transcription and translation, and immunoprecipitation

The cDNA fragments were subcloned into Bluescript SK (Stratagene, LaJolla, Calif.) and used for coupled in vitro transcription and translation (TnT, Promega, Madison, Wis.) in the presence of T3 RNA polymerase, rabbit reticulocyte lysate and [35S] methionine (trans 35S-label; ICN, Montreal). Translation was carried out at 30°C for 1 hour and was followed by SDS-PAGE of a 2- to 5-µL aliquot to confirm the presence of translation products. Immunoprecipitation of [35S]-labelled in vitro translation products was performed using Protein A-sepharose beads.15

Computer analysis of nucleic acid and protein sequences

Nucleic acid and protein sequences were analysed by the University of Wisconsin Genetics Computer Group (GCG) Sequence Analysis Software Package (ver. 7.2 for UNIX).16 Comparisons with known sequences were performed by BLAST17 on the Internet server.

Statistical analysis

Statistical significance was determined by Fisher's exact test and the 95% confidence interval (CI) calculated using the approximation of Woolf (InStat, ver. 2.05, GraphPad Software, San Diego).

[ Top of document ]

Results

The staining pattern produced by a prototype anti-CENP-F serum is shown in Fig. 1. In a survey of 2500 sequential, unselected sera referred to the Advanced Diagnostics Laboratory at the University of Calgary, 8 (0.3%) had anti-CENP-F reactivity, as determined by IIF. The features that distinguish the CENP-F antigen from the more common CENP-A, -B and -C antigens are intense nuclear staining at G2 phase of the cell cycle (Fig. 1, c and d), staining of centromeres only during prophase and metaphase, and staining within the spindle midzone and intercellular bridge connecting daughters cells at the completion of mitosis. Each of the sera selected for this study produced this characteristic CENP-F IIF pattern.

A survey of the clinical features of patients with CENP-F antibodies, as defined by the distinctive IIF pattern of staining and immunoblotting, indicated that these individuals could be grouped based on the presence or absence of a neoplasm (Table 1). To determine if there was a correlation between patients with malignancy and the fine specificity of their CENP-F autoantibodies, each of 33 available CENP-F sera was reacted with recombinant proteins representing 5 regions of CENP-F (Fig. 2). In Western blots employing the C-terminal fusion protein p-F4 (Fig. 3), 26 of 33 (79%) sera reacted. These results were confirmed when it was observed that all the 26 sera also immunoprecipitated the homologous in vitro translated recombinant peptide p-F4 (Fig. 4). Neither rabbit serum 911 raised to the recombinant peptide p-F2 (see Materials and methods) nor the normal human sera immunoprecipitated the recombinant C-terminal peptide (Fig. 3), verifying the specificity of reaction. When p-F5, a portion of p-F4 representing the most C-terminal of CENP-F, was studied separately, only 1 serum showed a different pattern of reactivity from that seen with the larger p-F4 fragment (Table 1). When Western blot studies were carried out using the N-terminal peptides, 16/30 (53%) reacted with p-F1, 24/33 (73%) reacted with p-F2, and 23/33 (70%) reacted with p-F3 (Fig. 2, Tables 1 and 2). Three sera, all from patients without neoplastic disease, did not react with any of the recombinant peptides tested. This suggests that these sera may bind epitopes that are not represented in the recombinant peptides used in this study. These epitopes are unlikely to reside at the extreme N-terminal region, since none of these sera identified cDNAs spanning this region during screening of 5 × 105 phages. The selectivity of the autoantibodies to the C-terminal portion of CENP-F, as demonstrated by the sera 8 and 32, was confirmed by adsorbing the serum with the p-F1 peptide. After adsorption, these sera failed to produce the CENP-F IIF pattern (data not shown).

Twenty-two of the 30 (73%) sera that reacted with at least 1 of the recombinant CENP-F proteins were from patients who had a neoplasm. Of sera available for testing, all 21 of the individuals with neoplasms but only 5/12 (42%) of the individuals without neoplasms had antibodies to the C-terminal p-F4 peptide (p = 0.0002, odds ratio [OR] 58.636, 95% CI 2.882 to 1192.9). By contrast, 9/19 (47%) of the patients with neoplasms and 7/11 (63%) of those with no neoplasms reacted with p-F1. The difference in reactivity with p-F1 between the patients with and without neoplasms was not statistically significant (p = 0.47). However, the difference in reactivity of the patients with neoplasms with p-F4 compared with reactivity with p-F1 was statistically significant (p < 0.0001, OR 0.02011, CI 0.001 to 0.3793). Similar statistical analyses applied to reactivity with p-F5 (Table 2). None of the patients with neoplasms had antibodies to the N-terminal peptide p-F1 alone. The most common malignancies were breast carcinoma (9/18) and lung carcinoma (5/18), and most of the patients with these types of cancer (15 of 18) were women. By contrast, 5/12 (42%) of the remaining CENP-F-reactive patients, who had a variety of non-neoplastic, inflammatory diseases, had antibodies to the C-terminus, and 4/12 (33%) had antibodies to the N-terminus. The group without neoplasms was distinguished by 2/11 (18%) individuals showing only reactivity to the N-terminus (Tables 1 and 2).

Analysis of control sera from 10 patients with small-cell lung carcinoma, 50 patients with breast cancer and 50 patients with malignant melanoma by IIF on HEp-2 substrates did not identify antibodies to CENP-F. In addition, CENP-F autoantibodies were not identified in the study of 354 sera from patients with SSc, including 5 patients with cancer, 120 with SLE and 50 with RA.

The high frequency of carcinoma in patients with autoantibodies reactive with CENP-F raises questions concerning the basis and the possible source of the autoantigen. Since some individuals in our CENP-F patient group had breast carcinoma, and breast tumour tissues were available, we chose this tumour for a study of CENP-F distribution in malignant tissues. Double-label experiments were carried out using antibodies to Ki-67, a proliferative cell marker,18 and CENP-F on cryopreserved sections of breast cancer tissue counter-stained with DAPI. CENP-F-reactive cells were readily detectable in sections showing abundant Ki-67 reactive cells. Two populations of reactive cells were detected. In one, CENP-F reactivity was found throughout the nucleus, a pattern characteristic of G2 cells (Fig. 5 a, b and c). In the other, the CENP-F pattern appeared punctate (Fig. 5 d, e and f). In cells displaying the latter pattern, the corresponding DAPI image revealed distinct chromosomes, and the corresponding Ki-67 images displayed a halo pattern characteristic of the prometaphase-metaphase staining pattern produced by this antibody (Fig. 5e) owing to its localization to the surface of condensed chromosomes. In some samples, clusters of CENP-F-reactive cells were detected (data not shown), suggesting the presence of clonal populations of cycling cells. CENP-F-reactive cells were not observed in cryosections of tissue obtained from patients who had undergone reduction mammoplasty and were rarely seen in sections of tonsil tissue (data not shown). These results suggest that tissue cells undergoing excessive proliferation, such as those in malignant transformation, are a potential source of immunogenic CENP-F.

[ Top of document ]

Discussion

We have identified 36 sera that display a CENP-F pattern by IIF and authenticated their CENP-F reactivity by showing that most of these sera contain autoantibodies that react with multiple determinants on CENP-F. Since several sera did not show reactivity with the fusion proteins, there may be additional epitopes outside the regions of CENP-F investigated in this study. Nevertheless, it appears that the dominant epitopes are within the determinants in the recombinant peptides used in this study.

The earlier suggestion that malignancy is a frequent finding in individuals with CENP-F antibodies11 is confirmed among the additional 10 patients reported here. In this additional group, 8/10 (80%) had cancer, as did 22/36 (61%) in the entire group. Our study suggests that, in CENP-F-positive patients with cancer, there is a bias toward autoantibodies that bind the C-terminal portion of CENP-F (amino acids 5561-10116).

During the past decade, a growing body of evidence has identified autoantibodies in cancer. The fact that some are markers of a subset of patients with paraneoplastic syndromes has begun to be appreciated.19­23 For example, antibodies to the Purkinje cell cytoplasmic antigen Yo are not commonly found in patients with cancer, yet they serve as a marker for paraneoplastic cerebellar degeneration in patients with breast or ovarian cancer.21,22 Likewise, antibodies to the nuclear antigen Hu are a marker for paraneoplastic encephalomyelitis in patients with small-cell lung cancer and other types of cancer.20,22 These autoantibodies have been shown to antedate the clinical appearance of malignancy and to react with antigens expressed in the tumour tissue.19,22 The early appearance of another autoantibody to the nuclear antigen HCC1, which contains sequence motifs of the SR family of splicing factors, has likewise been shown to antedate the appearance of hepatocellular carcinoma.24 Like CENP-F, topoisomerase II has been shown to be selectively enriched in the centromere of metaphase chromosomes.25 Autoantibodies to topoisomerase II in another patient with hepatocellular carcinoma26 and to the cell cycle-related antigen SG2NA in lung cancer have also been reported.27,28 Autoantibodies to nuclear proteins involved in the control of cell proliferation, such as p53, have been identified in a subset of patients with cancer.29­31

Although only 8 of 2500 sera referred to our clinical reference laboratory produced an IIF pattern of staining consistent with the presence of CENP-F antibodies, and these antibodies were not seen in other patients with systemic rheumatic syndromes or unselected patients with certain types of cancer, the high frequency of cancer seen in this group of patients as well as in our total group of 36 individuals is remarkable. It is interesting that some of the patients in our study came to the physician's attention because of the onset of arthritis. Although it could be argued that joint symptoms are a common medical complaint in patients in this age group, joint disease can herald the onset of cancer.32 Examples include calcific periarthritis in patients with a para-neoplastic syndrome associated with ovarian carcinoma33 and arthritis in patients with leukemia and lymphoma.34 Because the frequency of CENP-F autoantibodies appears to be rare in unselected patients with breast cancer and other types of cancer, multicentre studies will likely be required to determine if this autoantibody is associated with a paraneoplastic syndrome and if its appearance antedates the clinical appearance of cancer. Prospective studies will also be important to determine if the patients who have CENP-F autoantibodies but do not have cancer eventually develop a malignancy, and if the clinical outcome of patients with CENP-F autoantibodies is different from those who do not have these autoantibodies.

One of the problems considered by immunologists is the source of antigens that break tolerance to self-proteins and drive the autoantibody response.35,36 In some paraneoplastic syndromes, there is evidence that autoantibodies are directed against tissue-specific proteins that are highly expressed in cancer cells and normal target tissue. For example, the Hu antigen is highly expressed in tissue and neuronal nuclei in small-cell lung carcinoma.19 In the case of p53, there is evidence that patients with tumours that carry specific missense mutations leading to overexpression of p53 produce autoantibodies to the protein.26,27 Moreover, tumours from these patients also contained complexes between p53 and 70-kDa heat shock protein, whereas none of the tumours from antibody-negative patients contained this complex.27 It was suggested that structural alterations in p53, combined with its overexpression in tumours, might increase the immunogenicity of the protein, thus leading to an autoantibody response against both the wild-type and mutated forms of p53.26,27 These observations suggest that cancer cells may be the source of autoantigen that drives the autoantibody response.19­22 In this study, we have documented that the CENP-F antigen is found in malignant tissue. Since the expression of this protein is closely correlated with cell proliferation, the more active the tumour, the more elevated the expected level of CENP-F antigen. Further evidence of this was recently reported in a study in which 24 different hematopoietic malignancies and 12 breast tumours were analysed by flow cytometry to determine the correlation between the number of CENP-F-positive cells and the S-phase fraction.37 The percentage of CENP-F-positive cells was found to correlate significantly with the fraction of S-phase cells in all human malignancies tested. Thus, CENP-F may be another example of a cancer cell antigen that drives the autoantibody response. These findings, in concert with the recent discovery that the mitosis-specific protein CENP-E is an autoantigen,4 support the hypothesis that mitotic rather than interphase cells may be the source of autoantigen in these patients.

[ Top of document ]

Conclusion

Individuals with autoantibodies to the transiently associated centromere protein CENP-F and antibodies with biased binding to epitopes within the C-terminal portion of the protein have a high frequency of cancer. This finding suggests that the presence of CENP-F antibodies and, specifically, C-terminal CENP-F antibodies in patients who present with joint pain or constitutional symptoms should prompt physicians to consider a diagnosis of cancer. The specificity of this autoantibody for a subset of patients is suggested by its absence in patients with systemic rheumatic diseases, and its presence in patients with small-cell lung carcinoma, malignant melanoma and breast cancer.

[ Top of document ]

Acknowledgements

 This research was supported by the Arthritis Society (Dr. Fritzler), the National Cancer Institute of Canada (Dr. Rattner), with funds from the Canadian Cancer Society, the Alberta Cancer Board (Mr. Whitehead) and National Institutes of Health grant CA56956 (Dr. Tan). Dr. Casiano was supported by a postdoctoral fellowship from the Arthritis Foundation.

[ Top of document ]

References
  1. Earnshaw WC, Rothfield NF. Identification of a family of human centromere proteins using autoimmune sera from patients with scleroderma. Chromosoma 1985;91:313-22.
  2. Rattner JB, Rao A, Fritzler MJ, Valencia DW, Yen TJ. CENP-F is a .ca 400 kDa kinetochore protein that exhibits a cell-cycle dependent localization. Cell Motil Cytoskelton 1993;26:214-26.
  3. Casiano CA, Landberg G, Ochs RL, Tan EM. Autoantibodies to a novel cell cycle-regulated protein which accumulates in the nuclear matrix during S phase and localizes to kinetochores and spindle midzone during mitosis. J Cell Sci 1993;106:1045-56.
  4. Kingwell B, Rattner JB. Mammalian centromere/kinetochore composition. A 50 kDa antigen is present in the mammalian centromere/kinetochore. Chromosoma 1987;95:403-7.
  5. Rattner JB, Rees J, Arnett FC, Reveille JD, Goldstein R, Fritzler MJ. The centromere kinesin-like protein CENP-E. An autoantigen in systemic sclerosis. Arthritis Rheum 1996;39:1355-61.
  6. Fritzler MJ. Autoantibodies in scleroderma. J Dermatol 1993;20:257-68.
  7. Liao H, Winkfein RJ, Mack G, Rattner JB, Yen TJ. CENP-F is a protein of the nuclear matrix that assembles onto kinetochores at late G2 and is rapidly degraded after mitosis. J Cell Biol 1995;130:507-18.
  8. Fritzler MJ, Valencia DW, McCarty GA. Speckled pattern antinuclear antibodies resembling anticentromere antibodies. Arthritis Rheum 1984;27:92-6.
  9. Yen TJ, Compton DA, Wise D, Zinkowski RP, Brinkley BR, Earnshaw WC, et al. CENP-E, a novel human centromere-associated protein required for progression from metaphase to anaphase. EMBO J 1991;10:1245-54.
  10. Yen TJ, Li G, Schaar BT, Szilak I, Cleveland DW. CENP-E is a putative kinetochore motor that accumulates just before mitosis. Nature 1992;359:536-9.
  11. Casiano CA, Humbel RL, Peebles C, Covini G, Tan EM. Autoimmunity to the cell cycle-dependent centromere protein p330d/CENP-F in disorders associated with cell proliferation. J Autoimmunity 1995;8:575-86.
  12. Rattner JB, Martin L, Waisman DM, Johnstone SA, Fritzler MJ. Autoantibodies to the centrosome (centriole) react with determinants present in the glycolytic enzyme enolase. J Immunol 1991;146:2341-4.
  13. Maniatis T, Fritsch EF, Sambrook J. Molecular cloning: a laboratory manual. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1989.
  14. Sanger F, Nicklen S, Coulson AR. DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 1977;74:5463-6.
  15. Fritzler MJ, Hamel JC, Ochs RL, Chan EKL. Molecular characterization of two human autoantigens: unique cDNAs encoding 95- and 160-kD proteins of a putative family in the Golgi complex. J Exp Med 1993;178:49-62
  16. Devereux J, Haeberli P, Smithies O. A comprehensive set of sequence analysis for the VAX. Nucleic Acids Res 1984;12:387-95.
  17. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol 1990;215:403-10.
  18. Gerdes J, Lemke H, Baisch H, Wacker HH, Schwab U, Stein H. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J Immunol 1984;133:1710-15.
  19. Graus F, Elkon KB, Posner JB. Sensory neuropathy and small cell lung cancer. Antineuronal antibody that also react with the tumor. Am J Med 1986;80:45-52.
  20. Dalmau J, Furneaux HM, Gralla RJ, Kris MG, Posner JB. Detection of the anti-Hu antibody in the serum of patients with small cell lung cancer -- a quantitative Western blot analysis. Ann Neurol 1990;27:544-52.
  21. Peterson K, Rosenblum MK, Kotanides H, Posner JB. Paraneoplastic cerebellar degeneration. I. A clinical analysis of 55 anti-Yo antibody positive patients. Neurology 1992;42:1931-37.
  22. Dalmau J, Posner JB. Neurological paraneoplastic syndromes. Springer Semin Immunopathol 1996;18:85-95.
  23. Darnell RB. Onconeural antigens and the paraneoplastic neurologic disorders. At the intersection of cancer, immunity and the brain. Proc Natl Acad Sci U S A 1996;93:4529-36.
  24. Imai H, Nakano Y, Kiyosawa K, Tan EM. Increasing titers and changing specificities of antinuclear antibodies in patients with chronic liver disease who develop hepatocellular carcinoma. Cancer 1993;71:26-35.
  25. Rattner JB, Hendzel MJ, Furbee CS, Muller MT, Bazett-Jones DP. Topoisomerase II a is associated with the mammalian centromere in a cell cycle- and species-specific manner and is required for proper centromere/kinetochore structure. J Cell Biol 1996;134:1097-107.
  26. Imai H, Furuta K, Landberg G, Kiyosawa K, Liu LF, Tan EM. Autoantibody to DNA topoisomerase II in primary liver cancer. Clin Cancer Res 1995;1:417-22.
  27. Landberg G, Tan EM. Characterization of a DNA-binding nuclear autoantigen mainly associated with S phase and G2 cells. Exp Cell Res 1994;212:255-61.
  28. Muro Y, Chan EKL, Landberg G, Tan EM. A cell cycle nuclear autoantigen containing WD-40 motifs expressed mainly in S and G2 phase cells. Biochem Biophys Res Comm 1995;207:1029-37.
  29. Winter SF, Minna JD, Johnson BE, Takahashi T, Gazdar AF, Carbone DP. Development of antibodies against p53 in lung cancer patients appears to be dependent on the type of p53 mutation. Cancer Res 1992;52:4168-74.
  30. Davidoff A, Inglehart JD, Marks J. Immune response to p53 is dependent upon p53/HSP70 complexes. Proc Natl Acad Sci U S A 1992;89:3439-42.
  31. Angelopoulou K, Diamandis EP, Sutherland DJA, Kellen JA, Bunting PS. Prevalence of serum antibodies against the p53 tumor suppressor gene protein in various cancers. Int J Cancer 1994;58:480-7.
  32. Naschitz JE, Rosner I, Rozenbaum M, Elias N, Yeshurun D. Cancer-associated rheumatic disorders: clues to occult neoplasia. Semin Arthritis Rheum 1995;24:231-41.
  33. Latham BA, Fernandes L. Paraneoplastic syndromes in patients with ovarian neoplasia. J R Soc Med 1994;87:368-9.
  34. Caldwell DS, McCallum RM. Rheumatologic manifestations of cancer. Med Clin North Am 1986;70:385-417.
  35. Tan EM. Autoantibodies in pathology and cell biology. Cell 1991;67:841-2.
  36. Theofilopoulos AN. The basis of autoimmunity: Part I. Mechanisms of aberrant self-recognition. Immunol Today 1995;16:90-8.
  37. Landberg G, Erlanson M, Roos G, Tan EM, Casiano CA. Nuclear autoantigen p330d/CENP-F: a marker for cell proliferation in human malignancies. Cytometry 1996;25:90-8.
| CIM: October 1997 / MCE : octobre 1997 |
CMA Webspinners / >