Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Numerous studies over recent years have reported the occurrence of simian virus (SV) 40 large T antigen (SV40 Tag) DNA sequences in tissue samples from human tumors (1), leading to the hypothesis that SV40 might represent an etiologic agent or that at least it acts as a cofactor in some human cancers. Several groups, including our own, have focused in particular on malignant mesothelioma, a severe neoplasm most frequently associated with asbestos exposure, and have reported the detection of SV40-like DNA sequences in both frozen human mesothelioma and/ or mesothelioma cell lines (2). The recent findings that SV40 Tag protein isolated from frozen biopsies of human mesothelioma was able to interact with the pRb family proteins and p53 in human tumor specimens (7, 8) has reinforced the interest in the role of SV40 in this human cancer (9). Because p53 plays a significant role in the alteration of cell proliferation induced by various agents, we wanted to determine, in parallel experiments, whether the lack of cell-growth arrest in some human mesothelioma cell lines after treatment with interferon- (10, 27) was dependent on the presence of SV40 Tag. We therefore investigated whether SV40 Tag protein was expressed in human mesothelioma cell lines established in our laboratory and we demonstrated, using Western blotting and immunoprecipitation, that these cell lines do not express detectable amounts of SV40 Tag.

In the course of these experiments, we found that commercial Tag-specific antibody preparations contain a contaminating protein with approximately the same molecular weight as SV40 Tag. This contaminant was also found to react with all horseradish peroxidase (HRP)-conjugated antimouse immunoglobulin (Ig) Gs tested in this study. The results of this study, presented in this paper, suggest that the controversial data obtained in certain studies on SV40 Tag protein expression may be due to the quality of the commercially available anti-Tag monoclonal antibodies (mAbs).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

RPMI 1640 medium, glutamine, penicillin, streptomycin, and fetal bovine serum (FBS) were obtained from GIBCO BRL (Cergy Pontoise, France). Ab-1, a murine monoclonal IgG2_a (clone Pab 419) reactive to the N-terminal region of the SV40 Tag, was obtained from Calbiochem, France-Biochem (Meudon, France). Pab 101, a murine monoclonal IgG_2a specific for a carboxy terminal epitope of the Tag; DO-1, a murine monoclonal IgG_2a specific for human p53; and SMP14, a murine monoclonal IgG₁ specific for human MDM2, were purchased from Santa Cruz Biotechnology (Tebu, Le Perray en Yvelines, France). All chemicals were of analytical grade and were purchased from Sigma Chemicals (L'Isle d'Abeau Chesnes, France) unless otherwise stated.

Human Mesothelioma Cell Lines

The six human mesothelioma cell lines (BL, BR, DC, RV, CR, and FR) used in this study were established in our laboratory from pleural effusions or tumors, and their characteristics have been described elsewhere (10, 11). Cell lines were routinely maintained in RPMI 1640 medium, supplemented with 10% FBS, 5 mM glutamine, 50 µg/ml streptomycin, and 50 U/ml penicillin, and buffered with 10 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes) (ATGC, Noisy-Le-Grand, France). Normal rat pleural mesothelial cells (RPMC) were obtained from rat parietal pleura according to a previously described method (12). They were cultured at 37°C in Ham's F10 medium (ATGC) supplemented with 10% FBS, 10 mM Hepes, 100 IU/ml penicillin, and 50 µg/ml streptomycin (Eurobio, Les Ulis, France). Normal RPMC infected with a recombinant retrovirus encoding SV40 Tag (RPMC/ Tag-SV40⁺) were cultured as uninfected RPMC (RPMC/Tag-SV40) (13).

Immunoprecipitations and Western Blots

Confluent cells were washed twice with ice-cold phosphate-buffered saline (PBS), and scraped in extraction buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% Nonidet P-40, 50 mM sodium fluoride, 1 mM ethylenediaminetetraacetic acid, 1 mM sodium orthovanadate, 10 µg/ml leupeptin, 10 µg/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride). Cells were further lysed by repeated passages through a 21-gauge needle, and lysates were cleared by centrifuging at 16,000 × g for 15 min at 4°C. The supernatants were precleared with protein A-sepharose beads (Pharmacia Biotech, Orsay, France) with rocking for 30 min at 4°C. Protein concentrations were determined using the DC protein assay (Bio-Rad, Ivry-sur-Seine, France) and normalized by addition of extraction buffer. Lysates (500 µg of protein) were incubated for 2 h at 4°C with 3 µg of anti-Tag antibody (PAb-1). Immunocomplexes were collected on protein A-sepharose and washed four times with extraction buffer. Immunoprecipitates were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (12% gel), followed by transfer onto Immobilon-P membranes (Millipore, Paris, France). Membranes were blocked with 5% nonfat dry milk in PBS-0.05% Tween, and probed with PAb-1, followed by HRP-labeled antimouse IgG (Calbiochem). Blots were developed by enhanced chemiluminescence (Amersham, Les Ulis, France). In some experiments, cell extracts were immunoprecipitated with an anti-p53 mAb (HR231, mouse IgG2b; Eurobio) and the resulting immunoprecipitates were immunoblotted with the anti-Tag mAb Ab-1.

Immunocytochemical Analysis

Human mesothelioma cells were cultured in tissue culture chamber slides (LabTeck chambers; Nunc, Naperville, IL). Subconfluent cells were extensively washed with ice-cold PBS, air-dried for 6 h at room temperature, and then fixed with cold acetone (4°C) for 10 min. Subsequently, LabTeck slides were rehydrated with Tris-buffered saline (TBS) (0.05 M Tris-HCl and 0.15 M NaCl, pH 7.6) and incubated in blocking solution (TBS-bovine serum albumin [BSA]: TBS supplemented with 3% BSA) for 30 min. SV40 Tag protein expression was then investigated by a conventional alkaline phosphatase/antialkaline phosphatase procedure using Ab-1. Tris buffer was substituted for the primary antibody in the negative controls.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Human mesothelioma cell lines were screened for the presence of SV40 Tag protein. As shown in the typical example depicted in Figure 1, all cell lines displayed intense cytoplasmic staining whereas no nuclear staining was observed. Detection of SV40 Tag in the cytoplasmic compartment instead of the nucleus led us to investigate expression of this oncoprotein by a second technique. For this purpose, we immunoprecipitated protein extracts from six mesothelioma cell lines with Ab-1, followed by Western blot with the same anti-Tag mAb. As shown in Figure 2A, a protein of about 90 kD was detected. This protein migrated at the same level as the Tag protein immunoprecipitated from RPMC/Tag-SV40⁺ used as positive control (Figure 2B).

View larger version (82K): [in this window] [in a new window]   Figure 1.   Immunocytochemical analysis of Tag expression in a human mesothelioma cell line (DC). (A) Typical cytoplasmic reactivity observed with the murine anti-SV40 Tag mAb, Ab-1. (B) Control experiment in which the primary antibody was omitted.

View larger version (40K): [in this window] [in a new window]   Figure 2.   Western blot analysis of anti-Tag immunoprecipitations. The various samples (cell extracts or cell-free media) were immunoprecipitated with Ab-1 followed by Western blot with the same antibody as described in MATERIALS AND METHODS. (A) The human monocyte cell line U-937 and six human mesothelioma cell lines (BL, BR, DC, RV, CR, and FR) were analyzed. (B) Control experiments were carried out with RPMI-1640 supplemented (RPMI-10% FBS) or not (RPMI-0% FBS) with FBS, and with cell extracts from normal RPMC infected with a recombinant retrovirus encoding the Tag protein (RPMC/Tag-SV40+). HC and LC designate IgG heavy and light chains, respectively.

We then sought to determine whether the 90-kD Taglike protein was physically associated with p53. Cell extracts were immunoprecipitated with a monoclonal anti-p53 (HR231, mouse IgG_2b) and the resulting immunoprecipitates were immunoblotted with the anti-Tag mAb Ab-1. Under these conditions, the anti-p53 mAb coprecipitated the SV40 Tag protein from RPMC/Tag-SV40⁺ lysates used as positive control (Figure 3). In contrast, the anti-p53 did not precipitate the 90-kD Taglike protein initially shown in Figure 2A in the FR mesothelioma cell line (Figure 3) and similar results were obtained with the other five cell lines (data not shown).

View larger version (56K): [in this window] [in a new window]   Figure 3.   Absence of p53 and SV40-Tag coprecipitation from a mesothelioma cell line. Cell extracts were immunoprecipitated with an anti-p53 mAb (HR231) and the resulting immunoprecipitates were immunoblotted with the anti-Tag mAb Ab-1. RPMC/Tag-SV40+: normal RPMC expressing Tag. HL60: human monocyte cell line not expressing Tag. FR: human mesothelioma cell line.

Although the experiments depicted in Figures 1 and 2 suggest that human mesothelioma cell lines contain SV40 Tag protein, these data appeared questionable for the following reasons: (1) this Taglike protein was found in U937 (Figure 2A) and in normal RPMC (not shown), both used as negative controls; (2) the Taglike protein was detected in the CR cell line (Figure 2A), which was previously found negative for SV40-like DNA sequences (4); and (3) the Taglike protein apparently found in the mesothelioma cell lines was not associated with p53 (Figure 3).

To determine whether the detection of the 90-kD protein could be an artifact resulting from uptake by the cells of a protein present in the medium, cell-free media, namely RPMI-1640 and RPMI-1640 supplemented with 10% FBS, were treated according to the same procedure as cell extracts, i.e., immunoprecipitated with anti-Tag. A 90-kD band was observed with both media (Figure 2B), confirming that the protein observed in the cell lines used as negative control was not produced by the cells. Detection of the 90-kD Taglike protein in the cell lines was therefore likely to be an artifact generated by the immunoprecipitation procedure. This protein could be either released from the protein A-sepharose gel or contained in the mAb preparation. To test these possibilities, 30 µl of protein A-sepharose and 30 µl (3 µg) of the Ab-1 anti-Tag antibody (Calbiochem; lot# 349 202 02) were mixed with an equal volume of Laemmli buffer, subjected to Western blotting, and analyzed with the Ab-1 antibody. As shown in Figure 4, a band with an apparent molecular weight of 90 kD was detected in the antibody preparation. This band exhibited an electrophoretic mobility similar to that of the Tag protein detected in RPMC expressing the SV40 Tag protein used as positive control. In contrast, no such protein could be detected in two different batches of protein A-sepharose (Figure 4). Using the same procedure, the 90-kD protein was detected by Ab-1 in another batch of Ab-1 (Oncogene Research Products, France Biochem, Meudon, France; lot# 349 202- 7; data not shown) and in Pab-101 (Santa Cruz Biotechnology; lot# K047), another anti-Tag mAb reactive to a different region of the molecule (Figure 4). The 90-kD Taglike protein was also found in two other batches of Pab-101 (Santa Cruz Biotechnology; lots L037 and I249; data not shown). Interestingly, the protein extract from mesothelioma cell line FR that was included as a control was found to be totally negative (Figure 4) even after a much longer exposure time (not shown). Similar results were obtained with other mesothelioma cell lines (not shown), indicating that if SV40 Tag protein was present in these cells, its concentration was below the limit of detection of the Western blotting procedure.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Commercial preparations of murine anti-SV40 Tag protein mAb contain a 90-kD Taglike protein. Ab-1 and Pab-101 were subjected to Western blotting (3 µg/lane) and probed with Ab-1. Cell extracts from FR, a human mesothelioma cell line (40 µg/lane); normal RPMC (10 µg/lane) expressing (RPMC/Tag-SV40+) or not expressing (RPMC/Tag-SV40) the Tag protein and two different batches of protein A-sepharose (see RESULTS) were analyzed concurrently.

It was then important to determine whether the 90-kD proteins contained in these mAb preparations were immunologically related. For this purpose, both mAbs were subjected to Western blotting and probed with Ab-1 or Pab-101. As shown in Figure 5A, almost identical protein profiles were obtained regardless of the mAb used to probe the blots. More importantly, the 90-kD proteins contained in both mAb preparations were labeled in a control experiment in which the primary antibody was omitted, indicating that detection of this protein was due to the secondary antibody (Figure 5A). Reactivity with the 90-kD proteins was not restricted to the secondary HRP-conjugated antimouse IgG used in this experiment, because similar data were obtained with three other secondary antibodies from different commercial origins and from different sources (Figure 5B). The 90-kD proteins detected by the secondary antibodies in the mAb preparations were unlikely to be immunologically related to SV40 Tag protein because all three secondary antibodies tested failed to interact with the Tag protein expressed by RPMC/Tag-SV40⁺ (Figure 6A). The reactivity of the 90-kD contaminants with the secondary reagents raised the question of whether this protein could be present in other murine monoclonal IgGs directed against different antigens but with the same isotype. As shown in Figure 6B, a protein with a similar apparent relative molecular mass was found in a murine antihuman p53 monoclonal IgG of the same isotype (IgG_2a) as Ab-1 and Pab-101 (Figure 6B). In contrast, no such protein was detected in a murine IgG₁ directed against human MDM2 (Figure 6B).

View larger version (61K): [in this window] [in a new window]   Figure 5.   Detection of the 90-kD contaminant in Ab-1 and Pab-101 preparations by various HRP-conjugated antimouse IgGs. (A) Ab-1 and Pab-101 (3 µg/lane) were subjected to Western blot analysis and probed with either Ab-1 or Pab-101. In the control experiment shown on the right side of the blot, incubation with the primary antibody was omitted and the blot was probed directly with the HRP-conjugated secondary antibody. (B) Blots were probed directly with various HRP-conjugated secondary antimouse IgGs prepared from different species and obtained from different commercial sources.

View larger version (44K): [in this window] [in a new window]   Figure 6.   The 90-kD contaminant is immunologically unrelated to SV40 Tag protein and its presence is not restricted to anti-Tag monoclonal IgG. (A) A cell extract from RPMC/Tag-SV40+ (10 µg/lane) was subjected to Western blotting and probed with different HRP-conjugated antimouse IgGs either directly or after incubation with a murine anti-Tag (Ab-1) mAb (San.-Cr.: Santa Cruz). (B) Several murine monoclonal IgGs were subjected to Western blotting (3 µg/lane) and directly probed with HRP-conjugated goat antimouse IgGs (Santa Cruz). The anti-p53 antibody was an IgG2a, such as Ab-1 and Pab-101, and the anti-MDM2 was an IgG1.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

SV40 and SV40-like DNA sequences have recently been detected in human malignant mesothelioma and several other types of human tumors. These findings raised the possibility that SV40 could contribute to the development of human tumors. The recent finding that SV40 Tag expressed in mesothelioma could bind to known cellular target proteins, namely p53 and pRb, was of special interest because these tumor suppressor genes did not appear to be mutated in mesothelioma (14). As quoted by Wiman and Klein (18), these findings of an interaction between Tag and p53 and pRb proteins increased the suspicion of a possible role of SV40 in tumor etiology, especially by suggesting an alternative mechanism for p53 inactivation in tumor cells expressing an apparent p53 wild-type genotype. Alternatively, other mechanisms could also account for inactivation of p53 and pRb. Alterations and deletions at the INK4a/ARF locus encoding p16^INK4 and p14^ARF cell-cycle regulators appear to be a common feature of mesothelioma (19) and p14^ARF gene transfer in mesothelioma cells results in G1-phase arrest and apoptotic cell death (22). Ablation of p14^ARF may lead to mdm2-mediated inhibition of p53 and absence of p16 may interfere with pRb functioning because of lack of control of pRb phosphorylation.

In this study, we analyzed six human mesothelioma cell lines established in our laboratory and showed that they did not express detectable amounts of SV40 Tag, although some were previously found to contain SV40-like DNA sequences (4). We reached this conclusion despite our initial experiments (see Figures 1 and 2A) suggesting that these cell lines contain a considerable amount of this protein. In fact, these initial findings were artifactual observations caused by the quality of the commercial preparations of anti-Tag mAb. The need for polymerase chain reaction (PCR) to demonstrate the presence of SV40 and/or SV40-like DNA sequences in human tumors raised both technical (7) and conceptual (18) concerns. First, PCR is a very sensitive method that may produce false-positive signals and/or detect other polyomavirus-associated sequences. Second, the presence of SV40 DNA sequences is not sufficient to account for a role in tumor development, because the viral proteins must be expressed and ultimately impair the function of relevant cell proteins such as p53 and pRb. In this study, we show that the reliability of the methods allowing detection of viral proteins must be demonstrated before advocating the role of SV40 in human tumors. We found that both commercial preparations of anti-SV40 Tag mAb, namely, Ab-1 (or Pab 419) and Pab-101, were contaminated by a protein of similar size (90 kD) as SV40 Tag that reacts with the various secondary HRP-conjugated antimouse IgG tested. Consequently, this contaminating protein may be confounded with SV40 Tag and generate false-positive results when these antibodies are used to analyze Tag expression by a standard procedure, i.e., immunoprecipitation followed by Western blotting with the same antibody. The 90-kD contaminating protein was apparently not detected by Carbone and colleagues (2), who used the same procedure with the Ab-1 antibody. One of the reasons for these discrepancies might be the use of different batches of mAbs (although we found the 90-kD contaminant in all batches tested here). Alternatively, these authors used a markedly higher sample-to-monoclonal ratio because they immunoprecipitated 200 mg of frozen tumor with 1 µg of mAbs whereas our experiments were carried out with 500 µg of protein and 3 µg of mAb. Further studies would be needed to clarify these differences. In any event, the data of Carbone and associates are not explained by the contaminating 90-kD protein for the following reasons: (1) PCR and immunoprecipitation were complemented by the immunodetection of Tag in the expected cell compartment (see later), i.e., the nucleus (2); and (2) our immunoprecipitation studies indicated that the contaminating protein did not prevent the detection of Tag in cells that do express this protein (RPMC-SV40 Tag⁺).

The currently available immunohistochemical and immunocytochemical data concerning SV40 Tag expression in mesothelioma are puzzling. Although most studies have been conducted with Ab-1 and/or Pab-101, considerable variations have been observed in both the frequency and the quality of SV40 Tag staining. A low frequency of SV40 Tag expression was reported by Xu and coworkers (23) in a series of cultures of normal mesothelial cells from pleural effusions obtained from noncancerous donors. Two out of 38 cultures exhibited nuclear staining in the vast majority (96 ± 3%) of cells (23). Other investigators reported higher percentages of SV40 Tag in mesothelioma samples. Using the Ab-1 antibody, Carbone and colleagues (2) found that 11 out of 14 mesothelioma showed nuclear staining in 1 to 50% of cells. In another study, Testa and associates (5) observed positive staining in 10 out of 12 mesotheliomas. In this latter study, two types of antibodies were used, Pab 101 and Ab-1, leading to the same results. However, using Pab-101, Dhaene and coworkers (6) did not observe any positivity in a series of 28 Belgian mesothelioma cases. These authors reported cytoplasmic labeling in 10 out of 28 cases with Ab-1. Testa and colleagues (5) also noticed some cytoplasmic staining with Ab-1. In our hands, despite the occurrence of SV40-like DNA sequences in our mesothelioma cases (2), we did not observe nuclear, but did see cytoplasmic, staining with Ab-1. One possible explanation for the observed differences in SV40 Tag expression may be related to the fact that different patient groups from different geographic areas were investigated. Similarly, it should also be noted that Hirvonen and associates (24) did not observe SV40 DNA sequences in 49 mesothelioma cases from the Finnish population. In the present study, we observed cytoplasmic reactivity for SV40 Tag with Ab-1 in six mesothelioma cell lines. Inasmuch as SV40 Tag was undetectable in these cell lines by Western blotting, which is a much more sensitive procedure, our observations suggest that the cytoplasmic reactivity of SV40 Tag was artifactual, providing a possible explanation for some of the controversial data reported in the literature. One possible hypothesis is that the 90-kD protein contaminating anti-Tag mAbs can bind specific cytoplasmic structures in certain cell types. Because this protein is reactive with secondary antimouse IgG, this binding could generate false-positive signals in an unexpected cellular compartment. To test this hypothesis, it would therefore be of interest to determine the identity and origin of the contaminating 90-kD protein. Commercially available murine monoclonal IgGs are usually affinity-purified on immobilized protein A. Although this procedure is believed to achieve good yield and high purity, minor contaminants of approximately 90 kD, including plasmin (25), C1r, and C1s (26), have been reported. Because secondary antimouse IgGs are usually produced by immunizing animals with murine IgG purified on immobilized protein A, their reactivity with the 90-kD contaminants is not surprising. Finally, if this hypothesis is correct it should be possible to eliminate the interference possibly caused by the Taglike contaminant (even without determining its identity) by further purifying the anti-Tag mAbs by epitope affinity chromatography.

In conclusion, this study shows that SV40 Tag was below the limit of detection in six mesothelioma cell lines established in our laboratory. It also demonstrates that commercial preparations of anti-Tag mAb are contaminated by a 90-kD protein, which may generate false-positive responses on Tag expression analysis. Given the central role of Tag in the tumorigenic potential of SV40, these observations should be taken into account in future attempts to detect Tag in human tumors, especially mesothelioma.