- Email: [email protected]
Sera from patients with malignant mesothelioma can contain autoantibodies
Lung Cancer 20 (1998) 175 – 184
Sera from patients with malignant mesothelioma can contain autoantibodies Cleo Robinson, Bruce W.S. Robinson, Richard A. Lake * Uni6ersity Department of Medicine, Queen Elizabeth II Medical Centre, 4th Floor, G Block, Nedlands, Perth, Western Australia 6009, Australia Received 29 September 1997; received in revised form 9 January 1998; accepted 26 January 1998
Abstract Malignant mesothelioma (MM) is resistant to all conventional forms of therapy though there is considerable evidence from clinical trials and animal models of the disease that an immune response can be elicited to the tumour. In order to define those target antigens expressed by MM cells which might provide a focus for an effective immune response we tested patients’ sera for the presence of MM autoantibodies by Western blot analysis. Eight of 29 (28%) patients with MM had serum antibodies of the IgG class in high titre and each antiserum recognised different protein antigens. In those individuals where sequential samples were available, the antibody titre increased with the progression of the disease though the number of target antigens remained constant. Sera from the eight patients were studied further: six of the antigen complexes were expressed at least partially in the nucleus; two showed some specificity for the tumour in that they discriminated antigens that were highly expressed in all human MM cell lines, but were not expressed in a human SV40 transformed mesothelial line; four of the antisera recognised a homologue in mouse tissue and each of these had a different pattern of expression. Collectively, these antisera define a subset of nuclear autoantigens that are over-expressed in dividing cells. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Malignant mesothelioma; Tumor associated antigens; Autoantibodies; Tissue specificity; Western blot
1. Introduction Malignant mesothelioma (MM) is a tumour involving serosal surfaces; generally those of the pleura and less commonly of the peritoneum [1]. * Corresponding author. Tel.: + 618 9 3463127; fax: + 618 9 3462816; e-mail: [email protected]
The occurrence of MM is singularly associated with exposure to asbestos. The high levels of asbestos fibres in particular workplace environments and around the mining areas has resulted in population pockets with a correspondingly high incidence of the disease [2]. The disease is unresponsive to all conventional therapies, but there is some hope that in the future MM may be effec-
0169-5002/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0169-5002(98)00014-2
176
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
tively treated by immunological approaches [3]. This expectation arises from two types of evidence, firstly, clinical trials of various immunotherapeutic regimens in patients with MM have shown some capacity to ameliorate the disease [4–6]. Secondly, transplantable syngeneic murine MM cell lines induce a disease that is pathologically identical to the human disease [7], and immunological processes have been shown to moderate tumour growth in this model [8 – 11]. Target antigens in the immune response to MM have not been characterised although much work has been directed towards the production of mesothelioma specific monoclonal antibodies in the expectation that these will lead to the cloning of useful antigens. An example of this is the K1 antibody and its recently cloned target, the mesothelin molecule [12]. This protein has a restricted tissue distribution but is not unique to mesothelial tissue. Other monoclonal antibodies have been reported to distinguish MM from carcinoma but none so far distinguish the tumour from reactive mesothelial tissue. A polyclonal antibody that reacts strongly with MM and only weakly with reactive mesothelium has been described and this suggests that differentially expressed antigens do exist [13]. It seems clear that useful antitumour immunity will depend upon T cell recognition of tumour associated antigens (TAA) and it seems likely that the effector cells in such immunity will include cytotoxic T lymphocytes (CTL). The recent cloning of TAA from other immunotherapy sensitive cancers has resulted in an increased interest in the potential of active immunisation as a strategy to control cancer [14 – 16]. In particular, TAA have been cloned and characterised from melanoma [17–21] and antigen-specific CTL can be expanded in vitro using synthetic epitope peptides derived from such cloned sequences [22]. Recent work has shown that appropriate presentation of whole antigen as well as peptide fragments can generate a tumouricidal CTL response [16]. Thus, there is now ample evidence that an antitumour immune response can lead to tumour regression and the challenge for MM research is to define a useful set of target antigens.
Although the immune response to MM is usually ineffective, in many tumour biopsies, lymphocytes, macrophages and dendritic cells can be found infiltrating the tumour mass. Despite the presence of these cells, the tumour continues to grow. A number of possibilities have been suggested to account for the failure of the immune system to clear the tumour: these include; anergy induction; apoptosis induction; suppression or a deficiency in antigen presentation function. It is also possible that in some patients an immune response does occur, but it is ineffective. We believe that these patients might fail to generate a CTL-dominated immune response because a process of immune deviation allowed the generation of a humoral response rather than a cellular response with cytotoxic effectors. The presence of antitumour autoantibodies in the serum of these patients may therefore point towards potentially useful immunological targets. We therefore screened for such activity by Western blot.
2. Patients and methods
2.1. Patients Blood was obtained by venipuncture from patients attending the respiratory clinic at the QEII Medical Centre over the calender year 1995. The type and stage of MM at diagnosis was identified by histological examination of tumour samples and by radiology. The patients (25 male and 4 female) were aged between 40 and 81 years at diagnosis (mean 60.6910.1 years).
2.2. Sera Blood was allowed to clot at room temperature and the clots retract overnight at 4°C. Sera were clarified by centrifugation and stored in aliquots at − 20°C.
2.3. Cells Various transformed cell lines were maintained in liquid culture by growing them at 37°C in a water saturated atmosphere of 5% CO2 in air. Cell
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
lines derived from pleural effusions of patients with MM were adapted for growth in vitro [23]. All other cell lines were sourced from the American Type Culture Collection (ATCC), Rockville, MD. Cells were grown in RPMI 1640 containing 10% fetal calf serum supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine (all tissue culture media supplied by Life Technologies, MD). Adherent cells were washed gently with PBS and then removed by scraping with a sterile policeman.
2.4. Preparation of cell extracts Nuclear protein extracts were made by swelling cells on ice for 30 min in IB (10 mM Tris – HCl pH 7.5, 10 mM NaCl, 5 mM MgCl2), NP40 was added to 0.5% and the cells were then Dounce homogenised with 20 strokes of a ‘B’ type pestle. The lysate was overlaid on 3 ml 250 mM sucrose, 5 mM MgCl2, 25 mM Tris – HCl pH 7.5 and centrifuged at 1500 g for 10 min. The nuclear pellet was resuspended in a small volume of TE buffer (50 mM Tris– HCl pH 7.6, 0.5 mM EDTA) and an equal volume of 2× SDS sample buffer was added, the samples were boiled for 5 min before SDS-PAGE. Whole cell extracts were made by washing cells in PBS, resuspending in a small volume of TE buffer, then adding an equal volume of 2 ×SDS loading buffer. Before loading on an SDS-PAGE gel, samples were sonicated for 5 min in a water bath sonicator, boiled for 5 min and debris was pelleted at 14000 rpm in an Eppendorf centrifuge for 2 min.
2.5. Western blot analysis Protein samples separated by SDS-PAGE were transferred onto nitrocellulose filters (Transblot, Bio Rad Laboratories, CA). The filters were blocked in 5% low-fat milk powder in PBS for 2 h at 4°C. Antisera were diluted 1/100 in the blocking solution and incubated at 4°C overnight with constant agitation. Blots were then washed four times for 30 min in PBST (PBS containing 0.1% Tween 20), then incubated with goat antihuman IgG conjugated to horseradish peroxidase (Sigma, NJ) at a concentration of 1/2000 in PBST for 1 h
177
at room temperature. After washing four times with PBST for 15 min, bound antibody was detected by the ECL chemiluminescence reaction (Amersham) according to the manufacturer’s protocol.
3. Results
3.1. Screening of patients’ sera Sera from 29 patients diagnosed with MM were collected over the calender year 1995. These were screened for the presence of autoantibodies by Western blot analysis at a dilution of 1/100 on MM cell lysates (Fig. 1). Eight of the patients’ sera were positive (28%) recognising only one or a restricted number of antigens. Six normal sera (healthy laboratory volunteers, not shown) showed no reaction in this assay (likelihood ratio 0.063, x 2-test). The age and sex distribution of patients with antibodies (599 6.2 years, one female) was not significantly different from those without antibodies (61.2911.4 years, three female), nor was the presence of antibodies a prognostic indicator of survival (data not shown). There were no clear differences in the type of MM between the two groups nor was there any indication of differences in exposure to asbestos as judged by the clinical history. The molecular weight distribution of the antigens strongly suggested that no two sera recognised the same antigenic complex. This was confirmed by analysing the cellular distribution of the target antigens, by separating nuclear and cytoplasmic fractions from MM cells and running parallel samples in Western blots (Fig. 2). Six of the autoantigenic targets were at least partially localised to the nucleus. The various patterns of subcellular distribution were often more complex than might be guessed from Fig. 1. For example, serum c 27 detected three major molecular weight antigenic variants in a whole cell extract, one of these was exclusively located in the nucleus, the other two were exclusively located in the cytoplasm. Paradoxically, the cellular fractionation procedure exposed two other antigenic forms of the serum c 27 antigen also located exclusively in the cytoplasmic com-
178
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 1. Western blot analysis of 38 sera from 29 patients with MM (positive reactions are scored []; serum c3 and c 27 are from the same patient and similarly c 7 and c 36). Ju77, MM cells were lysed and loaded equally across the top of a gel; the proteins were separated by PAGE, then transferred to nitrocellulose membranes. The membranes were blocked, cut into strips and exposed to individual sera at a dilution of 1/100. Antibody binding to particular proteins was visualised by exposure of the strips to an HRP conjugated antihuman IgG with development and autoradiography according to standard ECL protocols (Amersham). Precise realignment of the strips was not possible so the estimated molecular weights are a guide only.
partment. These findings are most easily explained as resulting from cleavage products after the manipulation of the cells. This, and other blots (Fig. 3, for example) expose microheterogeneity in these antigens which is likely the result of posttranslational modifications such as phosphorylation.
3.2. Further characterisation of autoantigenic targets defined by patients’ sera The tissue specificity of the eight antigens was partially determined by testing the antisera in Western blot against human cell lines of diverse embryological origin (Fig. 3). These included a panel of human MM (Ju77, DeH, STY, ONE and Lo68), an SV40 transformed mesothelial cell line (Met5a), an epithelial cell line (HeLa), a T cell (ALL) line (CH7, a subclone of Jurkat) and a lung carcinoma cell line (A549). Several of the antigens were universally expressed (those detected by sera c5, c 15 and c27) although some of these antigens were expressed at different levels in the cell lines. Other antigens showed a more restricted pattern of expression; the higher molecular weight antigen detected by serum c 18 and the antigen detected by serum c36 for example are detected in all the MM cell lines but not the mesothelial cell line nor the two lines representing non-mesothelial tissues.
3.3. The titre of antibody can rise with disease progression Three patients with antibodies were bled on separate occasions and the sera titrated in Western blot analysis. The blots were scanned and the data plotted to reveal that in one patient (the donor of sera c 3 and c 27), a ten-fold increase in titre had occurred over a period of some 30 weeks (Fig. 4). One of the other patients experienced a two-fold increase in titre over 19 months. The third, who was bled on four separate occasions over a 7-month period reproducibly maintained the same titre (data not shown).
3.4. Patterns of tissue specific expression The eight antisera were tested in Western blot against lysates prepared from tissues recovered from a mouse at necropsy (Fig. 5). These blots are characterised by a high background and non-specific staining particularly in the kidney and liver. Nevertheless, four antisera detected a homologous murine antigen in some tissues and these findings are summarised in Table 1. It is notable that two of the four antigens are expressed in the brain but the dominant pattern of expression was in lymphoid tissues. These tissues are exceptional because of the high frequency of mitotic cells that
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
179
Fig. 2. Subcellular localisation of the antigens detected by eight of the sera identified in Fig. 1. Nuclei from NP40 cell lysates were separated from cytoplasmic constituents by centrifugation and washing. These two fractions and a third total cellular extract were separated by SDS PAGE and developed as for Fig. 1.
they contain. The four antisera that did not react in this assay may not cross react with a murine homologue or it may be that there is no murine homologue. Alternatively, these antigens may be expressed at low levels or in tissues that were not efficiently recovered (for example, mesothelium). The finding of generally high levels of expression in cells growing in vivo and the dominant pattern of expression in the thymus suggests that at least a subset of these autoantigens are overexpressed in dividing cells.
4. Discussion The search for tumour associated antigens (TAA) and the idea of using an antibody response to probe for such antigens has a long history [24]. Antibodies have been described that react with normal cellular proteins that are over-expressed in tumours, though those that have excited most interest identify antigens of oncofetal origin or with restricted tissue distribution. Historically,
there has been some expectation that mutated proteins involved in tumourigenesis would provide a unique target to activate the immune system. At least for some oncogene products, this expectation was ill founded, though there are isolated reports of novel T cell epitopes associated with activating mutations of some genes [25]. In general, it seems that the B-cell response to TAA is focused on regions located outside the mutational hot spots [26], indicating that a sustained response may depend upon over expression of a protein rather than de novo expression of a neo epitope. It is important to note, however, that the experiments described here were not designed to find unique antibody specificities because the screening process was always against heterologous tissue. To investigate the possibility of unique specificities, we plan to screen patients against their own primary tumour. A key observation from our studies is that no two individuals react to the same antigenic complex and most patients recognise only a very limited number of distinct antigens. In general,
180
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 3. Tissue specific expression of the antigens recognised by patients’ sera. Lysates were prepared from cells of diverse tissue origin, loaded onto gels and separated by electrophoresis as before. Western blotting was performed as described in Fig. 1.
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
181
Fig. 4. Antibody titre measured by limiting dilution in a Western blot assay. Sera were titrated with dilutions as indicated (a) and the assay performed as described in Fig. 1. The information was quantitated using a densitometer and plotted for serum, c3 and serum c 27 (b).
the major reactivity of each serum was to antigens expressed by a range of tumour cells of various cell lineage and primary tissues that contained proliferating cells. This, when taken together with the lack of any reactivity to these antigens in normal serum is strong evidence that the antibodies were elicited by the tumour. The stringent screening conditions used in these experiments also increases the probability that these antigens are related to the presence of tumour. Importantly, the fact that the detection system limits the
screen to antibodies of the IgG class predicates T cell recognition of these proteins. Nevertheless, it is clear that few if any of the antigens detected in this study are uniquely expressed in MM. This finding accords with those of other investigators [27] and supports the notion that the immunogenicity of particular tumours is a consequence of over-expression of a normal self protein rather than the novel expression of TAA. Recently, two strategies using T cells as probes have yielded a rich harvest of tumour antigens.
182
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 5. Tissue specific expression of the antigens recognised by patients’ sera. Lysates were prepared from mouse tissues and separated by SDS PAGE as before. Western blotting was performed as described in Fig. 1. Arrows indicate the position of the antigen in Ju77 lysate.
The first has made use of antigen loss variants [28]; whereas the second is dependent upon acid elution of peptides from the MHC [29,30]. Most of these studies have revealed antigens in
melanoma and relatively little is known about the frequency of antigen recognition in other tumours with the exception of one recent study which found a high frequency of autoantibodies to au-
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
183
Table 1 Patterns of tissue specific expression of murine homologues of autoantigens detected by four antisera from patients with MM Brain c12 c15 c27 c36
ã ã
Heart
Kidney
ã
Liver
ã
Lung
ã ã
Thymus
Spleen
ã ã ã
ã ã
LN
ã
Mouse tissue lysates were separated by electrophoresis and blotted onto nitrocellulose membranes. The blots were probed with patients’ sera (Fig. 5). Tissues showing a positive reaction at the same molecular weight as the human autoantigen were scored positive (tick) in this table.
tologous tumours of diverse origin in patients’ sera [27]. Our approach has yielded results consistent with these findings; we found eight potential tumour antigens for MM, though clearly molecular cloning and further characterisation of these antigens is necessary before their suitability as targets for immunotherapeutic intervention can be realistically assessed. CTLs are likely to constitute the most potent antitumour effector mechanism because of their ability to kill their target directly [31]. These cells are normally restricted by class I MHC molecules and express CD8 as a coreceptor. Unfortunately, the isolation of MM reactive T cell clones or lines has been difficult to achieve and this maybe because MM are capable of producing prodigious amounts of immunosuppressive molecules. These tumours and the cell lines derived from them characteristically secrete TGFb which has powerful antimitotic activity [32]. Antigen processing and presentation to class I restricted T cells might be expected to occur through the endogenous pathway, though it now seems clear that tumour cells are actually very poor at presenting antigen to prime CTL. Priming in vivo may require specialised antigen presenting cells; it has now been clearly shown that class I-restricted processing and presentation of exogenous antigen ‘can’ occur following immunisation with cell-associated antigen [33]. In one analysis, tumour antigens were exclusively presented by the host bone marrow-derived cells showing that tumour antigens are efficiently taken up by such APC in vivo [34]. Collectively, these observations indicate that immunotherapy for MM is likely to depend upon immunisation with multiple antigens
in a manner designed to elicit a maximal CTL response. We are now using the sera identified in this study to molecularly clone antigens which we expect to be useful in this process. The evidence that these antigens are potential T cell targets can be deduced from the knowledge that the isotype of the antibodies (IgG) requires cognate B and T cell interaction.
Acknowledgements We acknowledge financial support from James Hardie Industries, Australia.
References [1] Ruffie P. Pleural mesothelioma. Curr Opin Oncol 1992;4:334 – 41. [2] Anderson HA, Hanrahan LP, Schirmer J, Higgins D, Sarow P. Mesothelioma among employees with likely contact with in-place asbestos-containing building materials. Ann NY Acad Sci 1991;550 – 72. [3] Robinson BWS. Development of new therapeutic strategies for mesothelioma. Lung Cancer 1993;9:413 – 8. [4] Christmas T, Manning L, Garlepp M, Musk A, Robinson B. Effect of Interferon alpha-2a on malignant mesothelioma. J Interferon Res 1992;9 – 12. [5] Robinson B, Bowman R, Manning L, Musk A, van Hazel G. Interleukin-2 and lymphokine-activated killer cells in malignant mesothelioma. Eur Respir Rev 1993;3:220 – 2. [6] Upham J, Musk A, van Hazel G, Byrne M, Robinson B. Interferon alpha and doxorubicin in malignant mesothelioma. Aust New Zealand J Med 1993;23:683 – 7. [7] Davis MR, Manning LS, Whitaker D, Garlepp MJ, Robinson BW. Establishment of a murine model of malignant mesothelioma. Int J Cancer 1992;52:881 – 6.
184
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
[8] Bielefeldt OH, Fitzpatrick DR, Marzo AL, Jarnicki AG, Himbeck RP, Davis MR, Manning LS, Robinson BW. Patho- and immunobiology of malignant mesothelioma: characterisation of tumour infiltrating leucocytes and cytokine production in a murine model. Cancer Immunol Immunother 1994;39:347–59. [9] Leong CC, Robinson BW, Garlepp MJ. Generation of an antitumour immune response to a murine mesothelioma cell line by the transfection of allogeneic MHC genes. Int J Cancer 1994;59:212–6. [10] Bielefeldt OH, Marzo AL, Himbeck RP, Jarnicki AG, Robinson BW, Fitzpatrick DR. Interleukin-6 involvement in mesothelioma pathobiology: inhibition by interferon alpha immunotherapy. Cancer Immunol Immunother 1995;40:241 – 50. [11] Garlepp MJ, Leong CC. Biological and immunological aspects of malignant mesothelioma. Eur Respir J 1995;8:643 – 50. [12] Chang K, Pastan I. Molecular cloning of mesothelin, a differentiation antigen present on mesothelium, mesotheliomas, and ovarian cancers. Proc Natl Acad Sci USA 1996;93:136 – 40. [13] Donna A, Betta PG, Cosimi MF, Robutti F, Bellingeri D, Marchesini A. Putative mesothelial cell growth-promoting activity of a cytoplasmic protein expressed by the mesothelial cell. A preliminary report. Exp Cell Biol 1989;57:193 – 7. [14] Boon T, Cerottini JC, Vandeneynde B, Vanderbruggen P, Vanpel A. Tumour antigens recognized by T lymphocytes. Ann Rev Immunol 1994;12:337–65. [15] Pardoll DM. Tumour antigens—a new look for the 1990s. Nature 1994;369:357–8. [16] Paglia P, Chiodoni C, Rodolfo M, Colombo MP. Murine dendritic cells loaded in vitro with soluble protein prime cytotoxic T lymphocytes against tumour antigen in vivo. J Exp Med 1996;183:317–22. [17] Brichard V, Van PA, Wolfel T, Wolfel C, De PE, Lethe B, Coulie P, Boon T. The tyrosinase gene codes for an antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas. J Exp Med 1993;178:489–95. [18] Coulie PG, Brichard V, Van PA, Wolfel T, Schneider J, Traversari C, Mattei S, De PE, Lurquin C, Szikora JP, et al. A new gene coding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas [see comments]. J Exp Med 1994;180:35–42. [19] Haraguchi M, Yamashiro S, Yamamoto A, Furukawa K, Takamiya K, Lloyd KO, Shiku H, Furukawa K. Isolation of GD3 synthase gene by expression cloning of GM3 alpha-2,8-sialyltransferase cDNA using anti-GD2 monoclonal antibody. Proc Natl Acad Sci USA 1994;91:10455– 9. [20] Kawakami Y, Eliyahu S, Delgado CH, Robbins PF, Rivoltini L, Topalian SL, Miki T, Rosenberg SA. Cloning of the gene coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumour. Proc Natl Acad Sci USA 1994;91:3515–9. [21] Kawakami Y, Eliyahu S, Sakaguchi K, Robbins PF, Rivoltini L, Yannelli JR, Appella E, Rosenberg SA.
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31] [32]
[33]
[34]
Identification of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the majority of HLA-A2-restricted tumour infiltrating lymphocytes. J Exp Med 1994;180:347 – 52. Salgaller ML, Weber JS, Koenig S, Yannelli JR, Rosenberg SA. Generation of specific anti-melanoma reactivity by stimulation of human tumour-infiltrating lymphocytes with MAGE-1 synthetic peptide. Cancer Immunol Immunother 1994;39:105 – 16. Manning LS, Whitaker D, Murch AR, Garlepp MJ, Davis MR, Musk AW, Robinson BW. Establishment and characterization of five human malignant mesothelioma cell lines derived from pleural effusions. Int J Cancer 1991;47:285 – 90. Old LJ. Cancer immunology: the search for specificity — G.H.A. Clowes Memorial Lecture [Review]. Cancer Res 1981;41:361 – 75. Noguchi Y, Chen YT, Old LJ. A mouse mutant p53 product recognized by CD4 + and CD8 + T cells. Proc Natl Acad Sci USA 1994;91:3171 – 5. Schlichtholz B, Legros Y, Gillet D, Gaillard C, Marty M, Lane D, Calvo F, Soussi T. The immune response to p53 in breast cancer patients is directed against immunodominant epitopes unrelated to the mutational hot spot. Cancer Res 1992;52:6380 – 4. Sahin U, Tureci O, Schmitt H, Cochlovius B, Johannes T, Schmits R, Stenner F, Luo G, Schobert I, Pfreundschuh M. Human neoplasms elicit multiple specific immune responses in the autologous host. Proc Natl Acad Sci USA 1995;92:11810– 3. De Plaen E, Lurquin C, Van Pel A, Mariame B, Szikora JP, Wolfel T, Sibille C, Chomez P, Boon T. Immunogenic (tum-) variants of mouse tumour P815: cloning of the gene of tum-antigen P91A and identification of the tummutation. Proc Natl Acad Sci USA 1988;85:2274 – 8. Van Bleek G, Nathenson SG. Isolation of an endogenously processed immunodominant viral peptide from the class I H-2Kb molecule [see comments]. Nature 1990;348:213 – 6. Falk K, Rotzschke O, Stevanovic S, Jung G, Rammensee HG. Allele-specific motifs revealed by sequencing of selfpeptides eluted from MHC molecules. Nature 1991;351:290 – 6. Roth C, Rochlitz C, Kourilsky P. Immune response against tumours. Adv Immunol 1994;57:281 – 351. Fitzpatrick DR, Bielefeldt Ohmann H, Himbeck RP, Jarnicki AG, Marzo AL, Robinson BW. Transforming growth factor-beta: antisense RNA-mediated inhibition affects anchorage-independent growth, tumourigenicity and tumour-infiltrating T-cells in malignant mesothelioma. Growth Factors 1994;11:29 – 44. Carbone FR, Bevan MJ. Class I-restricted processing and presentation of exogenous cell-associated antigen in vivo. J Exp Med 1990;171:377 – 87. Huang AY, Golumbek P, Ahmadzadeh M, Jaffee E, Pardoll D, Levitsky H. Role of bone marrow-derived cells in presenting MHC class I-restricted tumour antigens. 1994;264:961 – 65.
Sera from patients with malignant mesothelioma can contain autoantibodies Cleo Robinson, Bruce W.S. Robinson, Richard A. Lake * Uni6ersity Department of Medicine, Queen Elizabeth II Medical Centre, 4th Floor, G Block, Nedlands, Perth, Western Australia 6009, Australia Received 29 September 1997; received in revised form 9 January 1998; accepted 26 January 1998
Abstract Malignant mesothelioma (MM) is resistant to all conventional forms of therapy though there is considerable evidence from clinical trials and animal models of the disease that an immune response can be elicited to the tumour. In order to define those target antigens expressed by MM cells which might provide a focus for an effective immune response we tested patients’ sera for the presence of MM autoantibodies by Western blot analysis. Eight of 29 (28%) patients with MM had serum antibodies of the IgG class in high titre and each antiserum recognised different protein antigens. In those individuals where sequential samples were available, the antibody titre increased with the progression of the disease though the number of target antigens remained constant. Sera from the eight patients were studied further: six of the antigen complexes were expressed at least partially in the nucleus; two showed some specificity for the tumour in that they discriminated antigens that were highly expressed in all human MM cell lines, but were not expressed in a human SV40 transformed mesothelial line; four of the antisera recognised a homologue in mouse tissue and each of these had a different pattern of expression. Collectively, these antisera define a subset of nuclear autoantigens that are over-expressed in dividing cells. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Malignant mesothelioma; Tumor associated antigens; Autoantibodies; Tissue specificity; Western blot
1. Introduction Malignant mesothelioma (MM) is a tumour involving serosal surfaces; generally those of the pleura and less commonly of the peritoneum [1]. * Corresponding author. Tel.: + 618 9 3463127; fax: + 618 9 3462816; e-mail: [email protected]
The occurrence of MM is singularly associated with exposure to asbestos. The high levels of asbestos fibres in particular workplace environments and around the mining areas has resulted in population pockets with a correspondingly high incidence of the disease [2]. The disease is unresponsive to all conventional therapies, but there is some hope that in the future MM may be effec-
0169-5002/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0169-5002(98)00014-2
176
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
tively treated by immunological approaches [3]. This expectation arises from two types of evidence, firstly, clinical trials of various immunotherapeutic regimens in patients with MM have shown some capacity to ameliorate the disease [4–6]. Secondly, transplantable syngeneic murine MM cell lines induce a disease that is pathologically identical to the human disease [7], and immunological processes have been shown to moderate tumour growth in this model [8 – 11]. Target antigens in the immune response to MM have not been characterised although much work has been directed towards the production of mesothelioma specific monoclonal antibodies in the expectation that these will lead to the cloning of useful antigens. An example of this is the K1 antibody and its recently cloned target, the mesothelin molecule [12]. This protein has a restricted tissue distribution but is not unique to mesothelial tissue. Other monoclonal antibodies have been reported to distinguish MM from carcinoma but none so far distinguish the tumour from reactive mesothelial tissue. A polyclonal antibody that reacts strongly with MM and only weakly with reactive mesothelium has been described and this suggests that differentially expressed antigens do exist [13]. It seems clear that useful antitumour immunity will depend upon T cell recognition of tumour associated antigens (TAA) and it seems likely that the effector cells in such immunity will include cytotoxic T lymphocytes (CTL). The recent cloning of TAA from other immunotherapy sensitive cancers has resulted in an increased interest in the potential of active immunisation as a strategy to control cancer [14 – 16]. In particular, TAA have been cloned and characterised from melanoma [17–21] and antigen-specific CTL can be expanded in vitro using synthetic epitope peptides derived from such cloned sequences [22]. Recent work has shown that appropriate presentation of whole antigen as well as peptide fragments can generate a tumouricidal CTL response [16]. Thus, there is now ample evidence that an antitumour immune response can lead to tumour regression and the challenge for MM research is to define a useful set of target antigens.
Although the immune response to MM is usually ineffective, in many tumour biopsies, lymphocytes, macrophages and dendritic cells can be found infiltrating the tumour mass. Despite the presence of these cells, the tumour continues to grow. A number of possibilities have been suggested to account for the failure of the immune system to clear the tumour: these include; anergy induction; apoptosis induction; suppression or a deficiency in antigen presentation function. It is also possible that in some patients an immune response does occur, but it is ineffective. We believe that these patients might fail to generate a CTL-dominated immune response because a process of immune deviation allowed the generation of a humoral response rather than a cellular response with cytotoxic effectors. The presence of antitumour autoantibodies in the serum of these patients may therefore point towards potentially useful immunological targets. We therefore screened for such activity by Western blot.
2. Patients and methods
2.1. Patients Blood was obtained by venipuncture from patients attending the respiratory clinic at the QEII Medical Centre over the calender year 1995. The type and stage of MM at diagnosis was identified by histological examination of tumour samples and by radiology. The patients (25 male and 4 female) were aged between 40 and 81 years at diagnosis (mean 60.6910.1 years).
2.2. Sera Blood was allowed to clot at room temperature and the clots retract overnight at 4°C. Sera were clarified by centrifugation and stored in aliquots at − 20°C.
2.3. Cells Various transformed cell lines were maintained in liquid culture by growing them at 37°C in a water saturated atmosphere of 5% CO2 in air. Cell
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
lines derived from pleural effusions of patients with MM were adapted for growth in vitro [23]. All other cell lines were sourced from the American Type Culture Collection (ATCC), Rockville, MD. Cells were grown in RPMI 1640 containing 10% fetal calf serum supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin, 2 mM L-glutamine (all tissue culture media supplied by Life Technologies, MD). Adherent cells were washed gently with PBS and then removed by scraping with a sterile policeman.
2.4. Preparation of cell extracts Nuclear protein extracts were made by swelling cells on ice for 30 min in IB (10 mM Tris – HCl pH 7.5, 10 mM NaCl, 5 mM MgCl2), NP40 was added to 0.5% and the cells were then Dounce homogenised with 20 strokes of a ‘B’ type pestle. The lysate was overlaid on 3 ml 250 mM sucrose, 5 mM MgCl2, 25 mM Tris – HCl pH 7.5 and centrifuged at 1500 g for 10 min. The nuclear pellet was resuspended in a small volume of TE buffer (50 mM Tris– HCl pH 7.6, 0.5 mM EDTA) and an equal volume of 2× SDS sample buffer was added, the samples were boiled for 5 min before SDS-PAGE. Whole cell extracts were made by washing cells in PBS, resuspending in a small volume of TE buffer, then adding an equal volume of 2 ×SDS loading buffer. Before loading on an SDS-PAGE gel, samples were sonicated for 5 min in a water bath sonicator, boiled for 5 min and debris was pelleted at 14000 rpm in an Eppendorf centrifuge for 2 min.
2.5. Western blot analysis Protein samples separated by SDS-PAGE were transferred onto nitrocellulose filters (Transblot, Bio Rad Laboratories, CA). The filters were blocked in 5% low-fat milk powder in PBS for 2 h at 4°C. Antisera were diluted 1/100 in the blocking solution and incubated at 4°C overnight with constant agitation. Blots were then washed four times for 30 min in PBST (PBS containing 0.1% Tween 20), then incubated with goat antihuman IgG conjugated to horseradish peroxidase (Sigma, NJ) at a concentration of 1/2000 in PBST for 1 h
177
at room temperature. After washing four times with PBST for 15 min, bound antibody was detected by the ECL chemiluminescence reaction (Amersham) according to the manufacturer’s protocol.
3. Results
3.1. Screening of patients’ sera Sera from 29 patients diagnosed with MM were collected over the calender year 1995. These were screened for the presence of autoantibodies by Western blot analysis at a dilution of 1/100 on MM cell lysates (Fig. 1). Eight of the patients’ sera were positive (28%) recognising only one or a restricted number of antigens. Six normal sera (healthy laboratory volunteers, not shown) showed no reaction in this assay (likelihood ratio 0.063, x 2-test). The age and sex distribution of patients with antibodies (599 6.2 years, one female) was not significantly different from those without antibodies (61.2911.4 years, three female), nor was the presence of antibodies a prognostic indicator of survival (data not shown). There were no clear differences in the type of MM between the two groups nor was there any indication of differences in exposure to asbestos as judged by the clinical history. The molecular weight distribution of the antigens strongly suggested that no two sera recognised the same antigenic complex. This was confirmed by analysing the cellular distribution of the target antigens, by separating nuclear and cytoplasmic fractions from MM cells and running parallel samples in Western blots (Fig. 2). Six of the autoantigenic targets were at least partially localised to the nucleus. The various patterns of subcellular distribution were often more complex than might be guessed from Fig. 1. For example, serum c 27 detected three major molecular weight antigenic variants in a whole cell extract, one of these was exclusively located in the nucleus, the other two were exclusively located in the cytoplasm. Paradoxically, the cellular fractionation procedure exposed two other antigenic forms of the serum c 27 antigen also located exclusively in the cytoplasmic com-
178
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 1. Western blot analysis of 38 sera from 29 patients with MM (positive reactions are scored []; serum c3 and c 27 are from the same patient and similarly c 7 and c 36). Ju77, MM cells were lysed and loaded equally across the top of a gel; the proteins were separated by PAGE, then transferred to nitrocellulose membranes. The membranes were blocked, cut into strips and exposed to individual sera at a dilution of 1/100. Antibody binding to particular proteins was visualised by exposure of the strips to an HRP conjugated antihuman IgG with development and autoradiography according to standard ECL protocols (Amersham). Precise realignment of the strips was not possible so the estimated molecular weights are a guide only.
partment. These findings are most easily explained as resulting from cleavage products after the manipulation of the cells. This, and other blots (Fig. 3, for example) expose microheterogeneity in these antigens which is likely the result of posttranslational modifications such as phosphorylation.
3.2. Further characterisation of autoantigenic targets defined by patients’ sera The tissue specificity of the eight antigens was partially determined by testing the antisera in Western blot against human cell lines of diverse embryological origin (Fig. 3). These included a panel of human MM (Ju77, DeH, STY, ONE and Lo68), an SV40 transformed mesothelial cell line (Met5a), an epithelial cell line (HeLa), a T cell (ALL) line (CH7, a subclone of Jurkat) and a lung carcinoma cell line (A549). Several of the antigens were universally expressed (those detected by sera c5, c 15 and c27) although some of these antigens were expressed at different levels in the cell lines. Other antigens showed a more restricted pattern of expression; the higher molecular weight antigen detected by serum c 18 and the antigen detected by serum c36 for example are detected in all the MM cell lines but not the mesothelial cell line nor the two lines representing non-mesothelial tissues.
3.3. The titre of antibody can rise with disease progression Three patients with antibodies were bled on separate occasions and the sera titrated in Western blot analysis. The blots were scanned and the data plotted to reveal that in one patient (the donor of sera c 3 and c 27), a ten-fold increase in titre had occurred over a period of some 30 weeks (Fig. 4). One of the other patients experienced a two-fold increase in titre over 19 months. The third, who was bled on four separate occasions over a 7-month period reproducibly maintained the same titre (data not shown).
3.4. Patterns of tissue specific expression The eight antisera were tested in Western blot against lysates prepared from tissues recovered from a mouse at necropsy (Fig. 5). These blots are characterised by a high background and non-specific staining particularly in the kidney and liver. Nevertheless, four antisera detected a homologous murine antigen in some tissues and these findings are summarised in Table 1. It is notable that two of the four antigens are expressed in the brain but the dominant pattern of expression was in lymphoid tissues. These tissues are exceptional because of the high frequency of mitotic cells that
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
179
Fig. 2. Subcellular localisation of the antigens detected by eight of the sera identified in Fig. 1. Nuclei from NP40 cell lysates were separated from cytoplasmic constituents by centrifugation and washing. These two fractions and a third total cellular extract were separated by SDS PAGE and developed as for Fig. 1.
they contain. The four antisera that did not react in this assay may not cross react with a murine homologue or it may be that there is no murine homologue. Alternatively, these antigens may be expressed at low levels or in tissues that were not efficiently recovered (for example, mesothelium). The finding of generally high levels of expression in cells growing in vivo and the dominant pattern of expression in the thymus suggests that at least a subset of these autoantigens are overexpressed in dividing cells.
4. Discussion The search for tumour associated antigens (TAA) and the idea of using an antibody response to probe for such antigens has a long history [24]. Antibodies have been described that react with normal cellular proteins that are over-expressed in tumours, though those that have excited most interest identify antigens of oncofetal origin or with restricted tissue distribution. Historically,
there has been some expectation that mutated proteins involved in tumourigenesis would provide a unique target to activate the immune system. At least for some oncogene products, this expectation was ill founded, though there are isolated reports of novel T cell epitopes associated with activating mutations of some genes [25]. In general, it seems that the B-cell response to TAA is focused on regions located outside the mutational hot spots [26], indicating that a sustained response may depend upon over expression of a protein rather than de novo expression of a neo epitope. It is important to note, however, that the experiments described here were not designed to find unique antibody specificities because the screening process was always against heterologous tissue. To investigate the possibility of unique specificities, we plan to screen patients against their own primary tumour. A key observation from our studies is that no two individuals react to the same antigenic complex and most patients recognise only a very limited number of distinct antigens. In general,
180
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 3. Tissue specific expression of the antigens recognised by patients’ sera. Lysates were prepared from cells of diverse tissue origin, loaded onto gels and separated by electrophoresis as before. Western blotting was performed as described in Fig. 1.
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
181
Fig. 4. Antibody titre measured by limiting dilution in a Western blot assay. Sera were titrated with dilutions as indicated (a) and the assay performed as described in Fig. 1. The information was quantitated using a densitometer and plotted for serum, c3 and serum c 27 (b).
the major reactivity of each serum was to antigens expressed by a range of tumour cells of various cell lineage and primary tissues that contained proliferating cells. This, when taken together with the lack of any reactivity to these antigens in normal serum is strong evidence that the antibodies were elicited by the tumour. The stringent screening conditions used in these experiments also increases the probability that these antigens are related to the presence of tumour. Importantly, the fact that the detection system limits the
screen to antibodies of the IgG class predicates T cell recognition of these proteins. Nevertheless, it is clear that few if any of the antigens detected in this study are uniquely expressed in MM. This finding accords with those of other investigators [27] and supports the notion that the immunogenicity of particular tumours is a consequence of over-expression of a normal self protein rather than the novel expression of TAA. Recently, two strategies using T cells as probes have yielded a rich harvest of tumour antigens.
182
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
Fig. 5. Tissue specific expression of the antigens recognised by patients’ sera. Lysates were prepared from mouse tissues and separated by SDS PAGE as before. Western blotting was performed as described in Fig. 1. Arrows indicate the position of the antigen in Ju77 lysate.
The first has made use of antigen loss variants [28]; whereas the second is dependent upon acid elution of peptides from the MHC [29,30]. Most of these studies have revealed antigens in
melanoma and relatively little is known about the frequency of antigen recognition in other tumours with the exception of one recent study which found a high frequency of autoantibodies to au-
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
183
Table 1 Patterns of tissue specific expression of murine homologues of autoantigens detected by four antisera from patients with MM Brain c12 c15 c27 c36
ã ã
Heart
Kidney
ã
Liver
ã
Lung
ã ã
Thymus
Spleen
ã ã ã
ã ã
LN
ã
Mouse tissue lysates were separated by electrophoresis and blotted onto nitrocellulose membranes. The blots were probed with patients’ sera (Fig. 5). Tissues showing a positive reaction at the same molecular weight as the human autoantigen were scored positive (tick) in this table.
tologous tumours of diverse origin in patients’ sera [27]. Our approach has yielded results consistent with these findings; we found eight potential tumour antigens for MM, though clearly molecular cloning and further characterisation of these antigens is necessary before their suitability as targets for immunotherapeutic intervention can be realistically assessed. CTLs are likely to constitute the most potent antitumour effector mechanism because of their ability to kill their target directly [31]. These cells are normally restricted by class I MHC molecules and express CD8 as a coreceptor. Unfortunately, the isolation of MM reactive T cell clones or lines has been difficult to achieve and this maybe because MM are capable of producing prodigious amounts of immunosuppressive molecules. These tumours and the cell lines derived from them characteristically secrete TGFb which has powerful antimitotic activity [32]. Antigen processing and presentation to class I restricted T cells might be expected to occur through the endogenous pathway, though it now seems clear that tumour cells are actually very poor at presenting antigen to prime CTL. Priming in vivo may require specialised antigen presenting cells; it has now been clearly shown that class I-restricted processing and presentation of exogenous antigen ‘can’ occur following immunisation with cell-associated antigen [33]. In one analysis, tumour antigens were exclusively presented by the host bone marrow-derived cells showing that tumour antigens are efficiently taken up by such APC in vivo [34]. Collectively, these observations indicate that immunotherapy for MM is likely to depend upon immunisation with multiple antigens
in a manner designed to elicit a maximal CTL response. We are now using the sera identified in this study to molecularly clone antigens which we expect to be useful in this process. The evidence that these antigens are potential T cell targets can be deduced from the knowledge that the isotype of the antibodies (IgG) requires cognate B and T cell interaction.
Acknowledgements We acknowledge financial support from James Hardie Industries, Australia.
References [1] Ruffie P. Pleural mesothelioma. Curr Opin Oncol 1992;4:334 – 41. [2] Anderson HA, Hanrahan LP, Schirmer J, Higgins D, Sarow P. Mesothelioma among employees with likely contact with in-place asbestos-containing building materials. Ann NY Acad Sci 1991;550 – 72. [3] Robinson BWS. Development of new therapeutic strategies for mesothelioma. Lung Cancer 1993;9:413 – 8. [4] Christmas T, Manning L, Garlepp M, Musk A, Robinson B. Effect of Interferon alpha-2a on malignant mesothelioma. J Interferon Res 1992;9 – 12. [5] Robinson B, Bowman R, Manning L, Musk A, van Hazel G. Interleukin-2 and lymphokine-activated killer cells in malignant mesothelioma. Eur Respir Rev 1993;3:220 – 2. [6] Upham J, Musk A, van Hazel G, Byrne M, Robinson B. Interferon alpha and doxorubicin in malignant mesothelioma. Aust New Zealand J Med 1993;23:683 – 7. [7] Davis MR, Manning LS, Whitaker D, Garlepp MJ, Robinson BW. Establishment of a murine model of malignant mesothelioma. Int J Cancer 1992;52:881 – 6.
184
C. Robinson et al. / Lung Cancer 20 (1998) 175–184
[8] Bielefeldt OH, Fitzpatrick DR, Marzo AL, Jarnicki AG, Himbeck RP, Davis MR, Manning LS, Robinson BW. Patho- and immunobiology of malignant mesothelioma: characterisation of tumour infiltrating leucocytes and cytokine production in a murine model. Cancer Immunol Immunother 1994;39:347–59. [9] Leong CC, Robinson BW, Garlepp MJ. Generation of an antitumour immune response to a murine mesothelioma cell line by the transfection of allogeneic MHC genes. Int J Cancer 1994;59:212–6. [10] Bielefeldt OH, Marzo AL, Himbeck RP, Jarnicki AG, Robinson BW, Fitzpatrick DR. Interleukin-6 involvement in mesothelioma pathobiology: inhibition by interferon alpha immunotherapy. Cancer Immunol Immunother 1995;40:241 – 50. [11] Garlepp MJ, Leong CC. Biological and immunological aspects of malignant mesothelioma. Eur Respir J 1995;8:643 – 50. [12] Chang K, Pastan I. Molecular cloning of mesothelin, a differentiation antigen present on mesothelium, mesotheliomas, and ovarian cancers. Proc Natl Acad Sci USA 1996;93:136 – 40. [13] Donna A, Betta PG, Cosimi MF, Robutti F, Bellingeri D, Marchesini A. Putative mesothelial cell growth-promoting activity of a cytoplasmic protein expressed by the mesothelial cell. A preliminary report. Exp Cell Biol 1989;57:193 – 7. [14] Boon T, Cerottini JC, Vandeneynde B, Vanderbruggen P, Vanpel A. Tumour antigens recognized by T lymphocytes. Ann Rev Immunol 1994;12:337–65. [15] Pardoll DM. Tumour antigens—a new look for the 1990s. Nature 1994;369:357–8. [16] Paglia P, Chiodoni C, Rodolfo M, Colombo MP. Murine dendritic cells loaded in vitro with soluble protein prime cytotoxic T lymphocytes against tumour antigen in vivo. J Exp Med 1996;183:317–22. [17] Brichard V, Van PA, Wolfel T, Wolfel C, De PE, Lethe B, Coulie P, Boon T. The tyrosinase gene codes for an antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas. J Exp Med 1993;178:489–95. [18] Coulie PG, Brichard V, Van PA, Wolfel T, Schneider J, Traversari C, Mattei S, De PE, Lurquin C, Szikora JP, et al. A new gene coding for a differentiation antigen recognized by autologous cytolytic T lymphocytes on HLA-A2 melanomas [see comments]. J Exp Med 1994;180:35–42. [19] Haraguchi M, Yamashiro S, Yamamoto A, Furukawa K, Takamiya K, Lloyd KO, Shiku H, Furukawa K. Isolation of GD3 synthase gene by expression cloning of GM3 alpha-2,8-sialyltransferase cDNA using anti-GD2 monoclonal antibody. Proc Natl Acad Sci USA 1994;91:10455– 9. [20] Kawakami Y, Eliyahu S, Delgado CH, Robbins PF, Rivoltini L, Topalian SL, Miki T, Rosenberg SA. Cloning of the gene coding for a shared human melanoma antigen recognized by autologous T cells infiltrating into tumour. Proc Natl Acad Sci USA 1994;91:3515–9. [21] Kawakami Y, Eliyahu S, Sakaguchi K, Robbins PF, Rivoltini L, Yannelli JR, Appella E, Rosenberg SA.
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31] [32]
[33]
[34]
Identification of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the majority of HLA-A2-restricted tumour infiltrating lymphocytes. J Exp Med 1994;180:347 – 52. Salgaller ML, Weber JS, Koenig S, Yannelli JR, Rosenberg SA. Generation of specific anti-melanoma reactivity by stimulation of human tumour-infiltrating lymphocytes with MAGE-1 synthetic peptide. Cancer Immunol Immunother 1994;39:105 – 16. Manning LS, Whitaker D, Murch AR, Garlepp MJ, Davis MR, Musk AW, Robinson BW. Establishment and characterization of five human malignant mesothelioma cell lines derived from pleural effusions. Int J Cancer 1991;47:285 – 90. Old LJ. Cancer immunology: the search for specificity — G.H.A. Clowes Memorial Lecture [Review]. Cancer Res 1981;41:361 – 75. Noguchi Y, Chen YT, Old LJ. A mouse mutant p53 product recognized by CD4 + and CD8 + T cells. Proc Natl Acad Sci USA 1994;91:3171 – 5. Schlichtholz B, Legros Y, Gillet D, Gaillard C, Marty M, Lane D, Calvo F, Soussi T. The immune response to p53 in breast cancer patients is directed against immunodominant epitopes unrelated to the mutational hot spot. Cancer Res 1992;52:6380 – 4. Sahin U, Tureci O, Schmitt H, Cochlovius B, Johannes T, Schmits R, Stenner F, Luo G, Schobert I, Pfreundschuh M. Human neoplasms elicit multiple specific immune responses in the autologous host. Proc Natl Acad Sci USA 1995;92:11810– 3. De Plaen E, Lurquin C, Van Pel A, Mariame B, Szikora JP, Wolfel T, Sibille C, Chomez P, Boon T. Immunogenic (tum-) variants of mouse tumour P815: cloning of the gene of tum-antigen P91A and identification of the tummutation. Proc Natl Acad Sci USA 1988;85:2274 – 8. Van Bleek G, Nathenson SG. Isolation of an endogenously processed immunodominant viral peptide from the class I H-2Kb molecule [see comments]. Nature 1990;348:213 – 6. Falk K, Rotzschke O, Stevanovic S, Jung G, Rammensee HG. Allele-specific motifs revealed by sequencing of selfpeptides eluted from MHC molecules. Nature 1991;351:290 – 6. Roth C, Rochlitz C, Kourilsky P. Immune response against tumours. Adv Immunol 1994;57:281 – 351. Fitzpatrick DR, Bielefeldt Ohmann H, Himbeck RP, Jarnicki AG, Marzo AL, Robinson BW. Transforming growth factor-beta: antisense RNA-mediated inhibition affects anchorage-independent growth, tumourigenicity and tumour-infiltrating T-cells in malignant mesothelioma. Growth Factors 1994;11:29 – 44. Carbone FR, Bevan MJ. Class I-restricted processing and presentation of exogenous cell-associated antigen in vivo. J Exp Med 1990;171:377 – 87. Huang AY, Golumbek P, Ahmadzadeh M, Jaffee E, Pardoll D, Levitsky H. Role of bone marrow-derived cells in presenting MHC class I-restricted tumour antigens. 1994;264:961 – 65.