¹Department of Experimental Medicine and Pathology, University `La Sapienza', Rome, Italy

²Department of Experimental Oncology, European Institute of Oncology, Milan, Italy

³Department of Oncology and Neurosciences, University `G D'Annunzio', Chieti, Italy

^aAuthor for correspondence

Employing NIH3T3 transfectants with individual human ErbB receptor coding sequences as recombinant antigen sources, we detected by immunoblot analysis specific immunoreactivity against all four ErbB receptors among 13 of 41 sera obtained from patients with different types of epithelial malignancies. Overall, serum positivity was most frequently directed against ErbB2 followed by EGFR, ErbB3 and ErbB4. Specificity patterns comprised tumor patients with unique serum reactivity against ErbB2 or ErbB4. Moreover, approximately half of the positive sera exhibited concomitant reactivity with multiple ErbB receptors including EGFR and ErbB2, EGFR and ErbB4, ErbB2 and ErbB3 or EGFR, ErbB2 and ErbB3. Serum reactivity was confirmed for the respective ErbB receptors expressed by human tumor cells and corroborated on receptor-specific immunoprecipitates. Positive sera contained ErbB-specific antibodies of the IgG isotype. Representative immuno-histochemical analysis of tumor tissues suggested overexpression of ErbB receptors for which serum antibodies were detectable in five of six patients. These findings implicate multiple ErbB receptors including ErbB3 and ErbB4 in addition to EGFR and ErbB2 in primary human cancer. Heterogeneity of natural ErbB-specific responses in cancer patients warrants their evaluation in light of immunotherapeutic approaches targeting these receptors.

ErbB receptors; cancer; serum antibody

Introduction

Members of the epidermal growth factor (EGF) receptor family of tyrosine kinases that includes EGFR, ErbB2, ErbB3 and ErbB4 (van der Geer et al., 1994), have frequently been implicated in human neoplasia by overexpression in the presence or absence of gene amplification. Such a role is supported by model systems in which EGFR and ErbB2 have been established in conferring the transformed phenotype (Aaronson, 1991; Di Fiore and Kraus, 1991). More recent mechanistic evidence implies that signal transduction by ErbB family receptors involves an array of ten possible homodimeric and heterodimeric combinations diversifying biologic responses to ligands of the EGF and the neuregulin (NRG) family (Carraway et al., 1994; Burden and Yarden, 1997). Overall, heterodimers were found biologically more active than homodimers. Among multiple heterodimer combinations investigated, EGFR-ErbB2 as well as ErbB2-ErbB3 and ErbB2-ErbB4 heterodimers have been implicated in the neoplastic process (King et al., 1988; Kokai et al., 1989; Alimandi et al., 1995; Pinkas-Kramarsky et al., 1996; Zhang et al., 1996). However, evidence for their involvement in human cancer is currently limited (Alimandi et al., 1995) and apart from frequent co-expression of these receptors in epithelial cancer, has not been established in primary tumors. Furthermore, while overexpression of EGFR, ErbB2 or ErbB3 has been associated with primary human cancer, a similar role for ErbB4 remains elusive.

In spite of these receptors representing autologous proteins, accumulating evidence suggests that tolerance of the EGFR or ErbB2 can be breached in certain cancer patients. Humoral responses directed against the EGFR have been reported (Glushkov et al., 1996). Moreover, it has been reproducibly documented that some breast and ovarian tumor patients mount natural T and B cell responses to ErbB2 (Pupa et al., 1993; Disis et al., 1994a,b, 1997; Fisk et al., 1995; Peoples et al., 1995; Clark et al., 1997). This receptor has triggered particular interest based upon evidence that its overexpression independently predicts a worse disease course in lymph node positive breast cancer (Slamon et al., 1989). In addition, its known oncogenic potential combined with high level overexpression in tumor tissue and cell surface localization render it a suitable target for immunotherapeutic approaches (Disis et al., 1996; Tuttle et al., 1996).

On the basis of mechanistic evidence for ErbB2 activity being regulated by heterodimerization with the other ErbB receptors including EGFR, ErbB3 and ErbB4 (Pinkas-Kramarsky et al., 1996), we explored the presence of antibody responses to all four human ErbB receptors in 41 sera from cancer patients with epithelial tumors including breast, lung, stomach, pancreas, colon, bladder, liver, testis and prostate carcinoma as well as lymphoma. Evidence for immunity to all four ErbB receptors simultaneously affecting multiple members in some sera, provides the basis for identification of immunogenic epitopes on these receptors under disease conditions in vivo and bear relevance for immunotherapy strategies targeting these receptors.

Results

Recombinant antigen sources for human ErbB receptors

NIH3T3 transfectant cell lines provided antigen sources for individual human ErbB receptors including EGFR, ErbB2, ErbB3 and ErbB4. These transfectants have previously been characterized and are known to express individual ErbB receptors in the range of 0.6 - 3´10⁶ molecules/cell (Di Fiore et al., 1987b; Kraus et al., 1993; Fedi et al., 1994; Baulida et al., 1996). A recombinant model system expressing high receptor levels was utilized, to facilitate sensitive and specific detection of single ErbB family members, as frequent co-expression of ErbB receptors in human tumor cell lines may confound definition of a receptor-specific response. Figure 1 illustrates detection of EGFR, ErbB2, ErbB3 or ErbB4 in the 170-185 kDa region of the corresponding transfectant by receptor-specific polyclonal rabbit antisera raised against carboxyl-terminal epitopes of single ErbB receptors (Fedi et al., 1994; Alimandi et al., 1995; Kraus et al., 1993; Baulida et al., 1996). Overexpression of the respective recombinant human ErbB receptor was validated in LTR-EGFR, LTR-ErbB2, LTR-ErbB3, or LTR-ErbB4 when compared with NIH3T3 controls (Figure 1). Detection specificity of the four polyclonal antisera is illustrated by lack of cross-reactivity with other ErbB receptors, respectively (Figure 1). Furthermore, normal rabbit serum lacked reactivity with NIH3T3 controls and transfectants (data not shown). An additional band at about 80 kDa was observed in LTR-ErbB4, but not NIH3T3 controls (Figure 1) suggesting recognition of a proteolytic cleavage product representing ErbB4 cytoplasmic domain in these transfectants (Vecchi et al., 1996). With exception of the anti-EGFR antiserum that upon prolonged exposure faintly visualized endogenous murine EGFR (data not shown), none of the other antisera detected endogenous murine receptors in NIH3T3. Thus, LTR-EGFR, LTR-ErbB2, LTR-ErbB3, and LTR-ErbB4 transfectants proved suitable for screening of serum reactivity to individual human ErbB receptors.

Immunoreactivity of patient sera with human ErbB receptors

Employing these transfectants as unique expression sources for human ErbB receptor proteins, sera from 41 patients with malignancies of different origin including lung, breast, gastrointestinal tract, bladder, liver, prostate, testis and lymphoid system, 21 sera from normal donors and eight sera from patients with non-neoplastic disorders were subjected to qualitative immunoblot analysis. Criteria of positivity comprised appearance of an immunoreactive band in a given transfectant comigrating with the one visualized by polyclonal rabbit antiserum of confirmed receptor specificity. Representative experiments are illustrated in Figure 2. Figure 2a shows reactivity of one of the normal control sera obtained from a healthy donor. This serum detected a 150 kDa band in all transfectants and control NIH3T3 suggesting reactivity with an endogenous mouse protein, whereas other control sera lacked reactivity altogether (data not shown). In contrast, patient sera depicted in Figure 2b and c displayed single reactivity with human ErbB2 or ErbB4, respectively. Specific bands comigrating with similar bands visualized by receptor-specific antiserum, were detected by the patient serum in the respective NIH3T3 transfectant. Intriguingly, several patients exhibited simultaneous reactivity with different ErbB receptors. Figure 2d shows serum of a patient with prostate carcinoma which recognized EGFR and ErbB2, while the patient serum in Figure 2e displayed simultaneous reactivity with EGFR and ErbB4. Finally, concomitant reactivity with EGFR, ErbB2 and ErbB3 is shown in Figure 2f for the serum of a patient with gastric carcinoma.

Patient sera reactivity with immunoprecipitated receptor proteins

To substantiate reactivity of patient sera, ErbB receptor proteins were enriched from transfectant lysates by immunoprecipitation employing receptor-specific antibodies. Immunoprecipitated receptors were then immunoblotted using representative patient and control sera. Immunoprecipitation was conducted in the presence of ionic detergents to minimize protein associations resulting from receptor activation in vivo. Figure 3 illustrates specific reactivity of patient sera with human EGFR, ErbB2, ErbB3 or ErbB4 proteins immunoprecipitated from the respective transfectants. Reactivity with EGFR and ErbB2 are shown for sera of patients AA and BM, respectively, whereas ErbB3 reactivity is verified for serum of patient FB reacting simultaneously with EGFR, ErbB2 and ErbB3. Erb4 reactivity was confirmed for patient serum MM (Figure 3). MabT151, MabE2-1, MabE3-1 and -ErbB4 specifically immunoprecipitated human EGFR, ErbB2, ErbB3 or ErbB4 proteins as demonstrated by immunoblot analysis employing receptor-specific polyclonal antisera. Similar bands were visualized by patient sera, but not negative control serum for EGFR (patient AA), ErbB2 (Patient BM), ErbB3 (patient FB) and ErbB4 (patient MM) immunoprecipitates. These observations validated specific detection of human ErbB receptors by patients' sera, established in immunoblot screening.

Heterogeneous serum reactivity for all ErbB receptors among patients with common epithelial malignancies

Anti-ErbB receptor reactivity was observed in 13 of 41 sera from cancer patients (31.7%) (Table 1). None of the sera from eight patients with nonmalignant disorders showed any specific reactivity with human ErbB receptors under these conditions. Of note, one of 21 sera from apparently healthy donors specifically reacted with LTR-ErbB2. Overall serum positivity was most frequently directed against ErbB2 (11/41) followed by EGFR (5/41), ErbB3 (2/41) and ErbB4 (2/41). Regarding tumor type, anti-ErbB4 reactivity was only observed in two sera originating from breast cancer patients, whereas reactivity to the other ErbB receptors was not restricted to a single tumor type (Table 2).

Among the 13 patients harboring serum reactivity with human ErbB receptors, six patients with tumors of different origin and one breast cancer patient exhibited unique reactivity with ErbB2 or ErbB4, respectively (Table 3). Multiple ErbB receptor reactivity included three patients with antibodies against EGFR and ErbB2 as well as one patient each possessing concomitant serum reactivity with EGFR and ErbB4, ErbB2 and ErbB3 or ternary specificity for EGFR, ErbB2 and ErbB3 (Table 3). None of the sera tested revealed universal recognition of all four ErbB receptors. Furthermore, simultaneous serum specificity for ErbB2 and ErbB4 or ErbB3 and ErbB4 was not detected among the patients analysed. A correlation between specific antibody responses and patients' age, sex, tumor type or stage was not observed.

To determine immunoglobulin isotypes of anti-ErbB receptors antibodies present in serum of cancer patients, separate immunoblots were developed using either goat anti-human IgG or IgM as secondary antibodies. Table 2 summarizes immunoglobulin isotypes for ErbB receptor antibodies in patient sera. All sera testing positive contained ErbB-specific IgG immunoglobulins, implicating an involvement of T helper cell activity in the anti-ErbB response. This assumption is based on the consideration of T helper activity as essential component in heavy chain class switching (Cebra et al., 1984; Zafiropoulus et al., 1997). In addition, eight of those sera revealed the presence of IgM-specific anti-ErbB immunoglobulins suggesting more recent immunogenic events (Table 2).

Reactivity of positive patient sera with ErbB receptors overexpressed by human tumor cells

Overexpression of human ErbB receptors represents a frequent abnormality of human epithelial malignancies. Moreover, evidence is emerging for simultaneous alteration of multiple ErbB receptors affecting the same tumor. Since serum reactivity has been assessed against ErbB receptors recombinantly expressed in a murine cell background where glycosylation could vary, we sought to explore whether patient sera were reactive with ErbB receptors overexpressed by human tumor cells. Human tumor cell lines were selected based upon known relative ErbB receptor expression levels. SK-BR-3 and A431 harbor high level overexpression of ErbB2 and EGFR, respectively associated with gene amplification (Kraus et al., 1987; Ullrich et al., 1984). In addition, SK-BR-3 overexpress ErbB3 at moderate receptor levels (Kraus et al., 1993). Human control serum did not detect any ErbB receptor-specific bands in these cell lines by immunoblot analysis (Figure 4a). Serum from patient BM harboring unique ErbB2 reactivity, visualized a band comigrating with ErbB2 detected by specific peptide antiserum in SK-BR-3, but not A431 cells. Patient CC serum with dual reactivity for ErbB2 and ErbB3 visualized a similar band in SK-BR-3 predicted to consist of ErbB2 and ErbB3. In contrast, serum of patient FB exhibiting ternary reactivity for EGFR, ErbB2 and ErbB3 detected a strong band in A431 and a weaker band of slightly slower migration in SK-BR-3. The latter is likely to be composed of ErbB2 and ErbB3, while the former is consistent with detection of the EGFR (Figure 4a).

To verify receptor-specific reactivity, we analysed two patient sera following immunoprecipitation of EGFR from A431 and ErbB2 or ErbB3 from SK-BR-3 tumor cell lines. Receptor immunoprecipitation using specific monoclonal antibodies was conducted under conditions that disrupt receptor oligomers (Alimandi et al., 1995) and controlled by non-immune IgG. By comparison with EGFR, ErbB2 and ErbB3 visualized by specific antisera, patient BM serum recognized ErbB2 immunoprecipitated from SK-BR-3, whereas anti-EGFR and anti-ErbB3 reactivity was confirmed for patient FB serum on receptor immunoprecipitates from A431 or SK-BR-3 cells, respectively (Figure 4b). These observations established that these patient sera specifically reacted with ErbB receptors including EGFR, ErbB2 and ErbB3 overexpressed by human tumor cell lines. Consistent reactivity with human ErbB receptors expressed by the recombinant model system or human tumor cells, ruled out differential post-translational modification of the receptors in the mouse model as a determinant of the observed immunoreactivity.

ErbB receptor expression in tumor tissues

To assess the state of ErbB receptor expression in tumors of patients showing evidence for anti-ErbB receptor reactivity in their serum, immunohistochemical analysis was conducted for EGFR, ErbB2, ErbB3 and ErbB4. Information could be obtained for six patients including breast carcinoma patients CC and RT, colon carcinoma patient FA, lung carcinoma patient SV as well as gastric carcinoma patients FB and SA. Patients SV, FA, SA or RT displayed unique serum reactivity with ErbB2 or ErbB4, respectively, whereas serum of breast carcinoma patient CC was reactive with both ErbB2 and ErbB3. Patient FB showed concomitant serum reactivity with EGFR, ErbB2 and ErbB3 (Table 4).

Non-immune IgG and normal rabbit serum lacked reactivity with tumor specimen of these patients (data not shown). In contrast, receptor-specific antibodies revealed distinct patterns of reactivity with the tumor tissues. Relative staining intensity when compared to normal epithelium was estimated semiquantitatively: (-) negative, (+/-) weakly positive, (+, ++, +++) increasing degrees of positivity. Figure 5 illustrates receptor staining patterns of a breast carcinoma, a colon carcinoma and a stomach carcinoma. The breast carcinoma (patient RT) positive for EGFR, ErbB2 and ErbB3 expression, showed by comparison very intense ErbB4-specific staining indicative of ErbB4 overexpression (Figure 5a). Strikingly, apparently normal breast epithelium adjacent to the tumor exhibited similarly intense ErbB4 staining. By contrast, in patient CC both normal and neoplastic mammary epithelium were negative for ErbB4 staining (Table 4). The colon carcinoma of patient FA was strongly positive for EGFR and ErbB2, while ErbB3 and ErbB4 reactivities were marginal (Figure 5b). The stomach carcinoma of patient FB exhibited strong positivity for all four ErbB receptors (Figure 5c).

Comparison of ErbB receptor positivity in available tumor tissue and anti-ErbB response in serum of the same patient (Table 4) indicated that tumor tissues generally overexpressed the receptor at the time antibodies were detectable in the patient's serum with the possible exception of patient CC where ErbB2 positivity of the tumor tissue was marginal (Table 4). However, the level of receptor overexpression in the tumor tissue did not necessarily correspond with the antibody response detected in serum suggesting that receptor overexpression in itself may not be the only factor for the observed immune response. Alternatively, the possibility that in some patients presence of serum ErbB receptor antibodies, resulted in steady state down-modulation of the targeted receptor, cannot be excluded.

Discussion

Several studies have previously established that patients with tumors overexpressing the ErbB2 protein, can spontaneously develop natural immunity to this receptor in the apparent absence of structural abnormalities (Pupa et al., 1993; Disis et al., 1994a,b, 1997; Fisk et al., 1995; Peoples et al., 1995; Clark et al., 1997). In the present study, we extend this observation by demonstrating the presence of specific antibody responses against all ErbB receptors including EGFR, ErbB2, ErbB3 and ErbB4 in tumor patients with different epithelial malignancies. Moreover, in approximately half of the tumor patients exhibiting any anti-ErbB serum reactivity, the specificity of such immune responses was concomitantly directed against multiple ErbB receptor family members. Detectability of serum responses by immunoblot analysis suggests reactivity with denatured receptor epitopes. The observation that 13 of 41 patients with breast, lung, stomach, pancreas or prostate carcinoma possessed anti-ErbB reactivity, implies that natural anti-ErbB receptor immune responses can affect all four ErbB receptors, occur rather frequently in cancer patients and are not restricted to a single tumor type.

Discrete peptide epitopes of normal ErbB2 have been identified which can elicit syngeneic cellular and humoral immunity in animal models and in vitro (Disis et al., 1994b, 1996; Nagata et al., 1997). Induction of tolerance is believed to involve elimination of lymphocytes reactive with immunodominant determinants of self-proteins, whereas other epitopes are concealed (Cibotti et al., 1992; Nanda and Sercarz, 1995). Such cryptic epitopes may secondarily be recognized by the immune system under certain conditions including increased amounts and fragmentation of an intact self-protein. Both processes are known to affect ErbB2 in neoplasia (Hynes and Stern, 1994) and might be enhanced by mechanisms associated with the neoplastic process such as sustained proliferation, cell degeneration and tumor necrosis. It is noteworthy that in our study one of 29 human control sera harbored specific anti-ErbB2 reactivity. Sporadic anti-ErbB2 immunoreactivity in apparently normal donors was also observed by others (Disis et al., 1994a, 1997). In the absence of detailed clinical information on normal donors, the reason for such reactivity is currently unknown. However, it cannot be excluded that non-neoplastic conditions associated with proliferation and/or degeneration of tissue contribute to the development of ErbB2 autoimmunity in some normal donors. Moreover, it is known that detectable serum antibody responses to oncogenic proteins can precede the onset of cancer (Lubin et al., 1995).

When compared to 11 of 41 cancer patients possessing anti-ErbB2 immunoreactivity, the prevalence of anti-ErbB2 immunity in tumor patients was statistically significant (P=0.01; see Table 1). Control sera lacked reactivity with EGFR, ErbB3 and ErbB4 or multiple ErbB receptors. Thus, tumor-specific association (13 of 41) of positivity gained significance (P=0.005) when overall reactivity with any of the ErbB receptors was evaluated. Less prevalent immunoreactivities including those against EGFR, ErbB3 or ErbB4 merit analysis of larger study groups for evaluation of statistical significance. Likewise, large-scale screening of patient sera might help to reveal anti-ErbB4 reactivity in tumor patients other than breast cancer.

The presence of ErbB-specific serum antibodies was associated with elevated receptor expression in five of six representative patients as determined by immunohistochemistry in tumor tissues. EGFR, ErbB2 and ErbB3 have been implicated by overexpression and constitutive activation in human neoplasia (Aaronson, 1991), while a similar role for ErbB4 remains elusive. Our detection of anti-ErbB4 serum reactivity associated with intense ErbB4 expression in mammary epithelium of one patient suggests involvement of the most recently discovered family member in primary human cancer. Since the ErbB4-reactive human sera failed to detect a proteolytic ErbB4 fragment representing intracellular domain in the transfectant, it is likely that ErbB4 reactivity was directed against an extracellular domain epitope. Simultaneous reactivity with multiple ErbB receptors prompts the question whether single immunogenic epitopes shared by different receptors or multiple epitopes specific for different receptors, are responsible. While definitive answers will have to await epitope mapping of patients' serum antibody, it should be noted that simultaneous reactivity with all four ErbB receptors was not observed. In addition, concurrent overexpression of different ErbB receptors in representative patients with multivalent ErbB receptor immunity, favors contribution of multiple ErbB receptors to the immunological profile of serum responses. Furthermore, no concurrent reactivity was seen between ErbB3 and ErbB4 which share relatively higher sequence homology of their extracellular domain, whereas unique serum specificity for individual ErbB receptors rules out shared epitopes in those cases.

Accumulating evidence about ErbB receptor signaling indicates that heterodimeric receptor combinations can significantly diversify cellular responses to extracellular ligands. In vitro, ErbB2-containing heterodimers convey a stronger biologic effect than the respective homodimers. In addition, heterodimeric receptor signals have been shown indispensable for physiologic development in vivo, while evidence for a role in human neoplasia currently is limited to tumor cell lines. In this context, our observation of simultaneous syngeneic immune responses to multiple ErbB receptors raises the intriguing possibility that multiple ErbB receptor combinations are involved in primary human neoplasia representing targets of naturally occurring syngeneic immune responses. While simultaneous reactivity to certain receptor combinations including EGFR and ErbB3 or ErbB2 and ErbB4 due to a lower frequency may have escaped detection, the relatively more prevalent immunoreactivity with ErbB2 may reflect its propensity to engage in heterodimeric signaling with each of the other three ErbB receptors.

ErbB2 overexpression detected in a significant fraction of human epithelial malignancies has been correlated with a worse disease prognosis in breast and ovarian cancer (Slamon et al., 1989). Based upon these observations, ErbB2 has been evaluated as target for passive immunotherapy using anti-ErbB2 monoclonal antibodies. Recently completed phase II clinical trials employing a humanized anti-ErbB2 monoclonal antibody resulted in objective responses for five of 43 breast cancer patients with advanced metastatic ErbB2-positive tumors. Lack of a therapeutic response appeared to correlate with high serum levels of shed soluble ErbB2, while pre-existing anti-ErbB2 immune responses were not reported (Baselga et al., 1996). Our findings suggest that tumor patients with ErbB2 overexpression represent an immunologically heterogeneous population involving in some cases pre-existent immunoreactivity with multiple ErbB receptors. Knowledge of anti-ErbB receptor immune status may have relevance for refining immunotherapeutic strategies targeting ErbB receptors in vivo. On one hand, it may provide the basis for predicting individual patient responses to passive immunotherapy. On the other hand, it may help optimizing immunotargeting strategies by identification of immunodominant epitopes, heterodimer-specific antibodies or combined targeting of multiple ErbB receptors.

Materials and methods

Patient sera

Sera samples were collected prior to adjuvant therapy from 41 patients diagnosed of cancer including lung carcinoma (n=10), breast carcinoma (n=13), colon carcinoma (n=7), stomach carcinoma (n=3), pancreatic carcinoma (n=1), hepatocellular carcinoma (n=2), prostate carcinoma (n=2), bladder carcinoma (n=1), seminoma (n=1) and lymphoma (n=1). Control sera were collected from patients with nonmalignant disorders (n=8) and from apparently healthy donors (n=21). All sera were obtained with informed consent of the donor and stored at -20° until use. Statistical association was evaluated by Fisher's Exact Test.

Cell lines

SK-BR-3 breast tumor-derived and A431 epidermoid carcinoma-derived cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) containing 10% Fetal Calf Serum. NIH3T3 cells transfected with expression vectors for human coding sequences of ErbB family receptors have previously been described including LTR-EGFR (Di Fiore et al., 1987a; Fedi et al., 1994), LTR-ErbB2 (Di Fiore et al., 1987b), LTR-ErbB3 (Kraus et al., 1993) and LTR-ErbB4 (Baulida et al., 1996).

Antibodies

Polyclonal antisera and monoclonal antibodies to individual human ErbB receptors have previously been characterized. Anti-EGFR antiserum E7 (Fedi et al., 1994), anti-ErbB2 antiserum M6 (Alimandi et al., 1995) and anti-ErbB3 antiserum MK4 (Kraus et al., 1993) were raised in rabbits against synthetic peptides of the respective human coding sequence [EGFR: aa 1172 - 1186 (Ullrich et al., 1984), ErbB2: aa 1218 - 1232 (Yamamoto et al., 1986), ErbB3: aa 1191 - 1205 (Kraus et al., 1989)]. ErbB4 antiserum was generated against a GST-fusion protein with the carboxyl-terminal ErbB4 domain (Baulida et al., 1996). Mabs T151 (Brice et al., 1993), E2-1 (Alimandi et al., 1995) and E3-1 (Kraus et al., 1993) specifically recognizing the extracellular domains of EGFR, ErbB2 and ErbB3, respectively, have been described elsewhere. Mab CB11 recognizing ErbB2 was purchased from BioGenex (San Ramon, CA, USA). The anti-ErbB4 polyclonal rabbit antiserum C18 was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Purified mouse myeloma protein (MOPC21) (Cappel, Organon Teknika Corp., West Chester, PA, USA), normal rabbit serum or anti-human albumin polyclonal rabbit serum (Cappel, Organon Teknika Corp., West Chester, PA, USA) served as negative controls.

Immunoprecipitation

Cell lysates were prepared in Staph A buffer (10 mM sodium phosphate pH 7.4, 100 mM NaCl, 5 mM EGTA, 1% Triton, 0.1% SDS, 0.5% deoxycholate) containing 100 g/ml aprotinin and 1 mM PMSF. Three g of monoclonal antibodies and 20 l of protein G sepharose were reacted with 300 g lysate of transfectants or tumor cell lines for 2 h at 4°C. Alternatively, 2 l of polyclonal rabbit antiserum were used in immunoprecipitation. The beads were washed three times with StaphA buffer, and the pellets were denatured by boiling for 5 min in 30 l of sample buffer (100 mM Tris pH 6.8, 4% SDS, 0.2% bromophenol blue, 20% glycerol, 50 mM 2-mercaptoethanol). Following electrophoresis, gels were processed for immunoblotting using receptor-specific antisera or patient sera.

Immunoblotting

Electrophoresis of protein lysates or immunoprecipitates was carried out in denaturing 8% Tris-glycine polyacrylamide gels (SDS - PAGE). For direct immunoblotting, 100 g of protein lysates derived from NIH3T3 transfectants or human tumor lines were separated by SDS - PAGE. Following electrophoresis, the proteins were transferred to a nitrocellulose membrane at 40 V for 1 h. The membrane was blocked overnight in wash solution (Tris-buffered saline pH 7.6, 0.1% Tween) containing 5% non-fat dry milk and subsequently incubated overnight at 4°C either with human sera or with specific polyclonal antibodies. Human patient sera were initially titrated by immunoblot analysis at dilutions of 1 : 25, 1 : 50, 1 : 100 and 1 : 200. A dilution of 1 : 100 was chosen for testing as the highest serum concentration that lacked substantial background reactivity. After washing, the membrane was incubated with alkaline-phosphatase-conjugated goat anti-human IgG and IgM or goat anti-rabbit antiserum (Gibco BRL). Bound antibody was visualized using 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and nitro-blue tetrazolium (NBT) (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD, USA).

Immunohistochemical analysis

Expression of EGFR, ErbB2, ErbB3 and ErbB4 was determined by immunoperoxidase staining as previously described (Mariani Costantini et al., 1991). Serial sections (5 m) of formalin-fixed, paraffin-embedded specimen were dewaxed in xylene, hydrated through graded ethanol solutions, and washed in 0.01 M phosphate buffered saline solution (PBS), pH 7.4. Endogenous peroxidase activities were blocked by pre-incubation in methanol containing 0.3% hydrogen peroxidase. Sections were incubated with anti-EGFR Mab T151 (2 g/ml), anti-ErbB2 Mab CB11 (tissue culture supernatant diluted 1 : 20), anti-ErbB3 polyclonal anti-serum MK4 (1 : 250) or anti-ErbB4 polyclonal anti-serum C18 (1 : 100) for 30 min at room temperature. Non-immune IgG (MOPC-21, 2 g/ml) and normal rabbit serum (1 : 250) served as negative controls. Receptor expression was visualized by a streptavidin-biotin system kit for primary murine or rabbit antibodies, respectively (Zymed Laboratories, San Francisco, CA, USA). The peroxidase reaction was initiated by the addition of 0.06% diaminobenzidine (DAB) (Sigma-Aldrich, Milano) in PBS containing 0.01% hydrogen peroxidase. The slides were counterstained with Harris hematoxylin and were permanently mounted under coverslips. Stained sections were independently evaluated by two pathologists.

Acknowledgements

The authors thank Dr Angelo Del Nero, University `La Sapienza', Rome, for statistical analyses. This study was partly supported by grants from the Ministero dell'Universita' e della Ricerca Scientifica (MURST), the Consiglio Nazionale delle Ricerche (project CNR-ACRO), the Associazione Italiana per la ricerca sul Cancro (AIRC) and the European Community (Programme-Biomed II).

Aaronson SA. (1991). Science 254, 1146-1153. MEDLINE

Alimandi M, Romano A, Curia MC, Muraro R, Fedi P, Aaronson S, Di Fiore PP and Kraus MH. (1995). Oncogene 10, 1813-1821. MEDLINE

Baselga J, Tripanthy D, Mendelsohn J, Baughman S, Benz CC, Dantis L, Sklarin NT, Seidman AD, Hudis CA, Moore J, Rosen PP, Twaddel T, Henderson IC and Norton L. (1996). J. Clin. Oncol. 14, 737-744. MEDLINE

Baulida J, Kraus MH, Alimandi M, Di Fiore PP and Carpenter G. (1996). J. Biol. Chem. 271, 5251-5257. MEDLINE

Brice EC, Wu JX, Muraro R, Adamson ED and Willey LM. (1993). Devel. Genetics 14, 174-184.

Burden S and Yarden Y. (1997). Neuron 18, 847-855. MEDLINE

Carraway III KL and Cantley LC. (1994). Cell 78, 5-8. MEDLINE

Cebra JJ, Komisar JL and Schweitzer PA. (1984). Annu. Rev. Immunol. 2, 493-548. MEDLINE

Cibotti R, Kannellopoulus JM, Cabalions JP, Halle-Panenko O, Kosmatopoulus K, Sercarz E and Kourilsky P. (1992). Proc. Natl. Acad. Sci. USA 89, 416-420. MEDLINE

Clark MA, Hawking NJ, Papaioannou A, Fiddes RJ and Ward RL. (1997). Clin. Exp. Immunol. 109, 166-174. MEDLINE

Di Fiore PP, Pierce JH, Fleming TP, Hazan R, Ullrich A, King CR, Schlessinger J and Aaronson SA. (1987a). Cell 51, 1063-1070. MEDLINE

Di Fiore PP, Pierce JH, Kraus MH, Segatto O, King CR and Aaronson SA. (1987b). Science 237, 178-182. MEDLINE

Di Fiore PP and Kraus MH. (1991). Genes, oncogenes and hormones: Advances in cellular and molecular biology of breast cancer. Dickson RB and Lipman ME (eds). Kluwer Academic Publishers: Boston, pp.139-160.

Disis ML, Calenoff E, McLaughlin G, Murphy AE, Chen W, Groner B, Jeschke M, Lydon N, McGlynn E, Livingston RB, Moe R and Cheever MA. (1994a). Cancer Res. 54, 16-20. MEDLINE

Disis ML, Smith JW, Murphy AE, Chen W and Cheever MA. (1994b). Cancer Res. 54, 1071-1076.

Disis M, Gralow JR, Bernhard H, Hand SL, Rubin WD and Cheever MA. (1996). J. Immunology 156, 3151-3158.

Disis ML, Pupa SL, Grawlow JR, Dittadi R, Menard S and Cheever MA. (1997). J. Clin. Oncol. 15, 3363-3367. MEDLINE

Fedi P, Pierce JH, Di Fiore PP and Kraus MH. (1994). Mol. Cell. Biol. 14, 492-500. MEDLINE

Fisk B, Blevins TL, Wharton JT and Ioannides CG. (1995). J. Exp. Med. 181, 2109-2117. MEDLINE

Glushkov AN, Anosova TP, Anosov MP, Garmanchouk LV and Pogrebnoy PV. (1996). Exp. Oncology 18, 81-83.

Hynes NE and Stern DF. (1994). Biochim. Biophys. Acta 1198, 165-184. MEDLINE

King CR, Borrello I, Bellot F, Comoglio P and Schlessinger J. (1988). EMBO J. 7, 1647-1651. MEDLINE

Kokai Y, Myers JN, Brown VI, LeVea CM, Davis JG, Dobashi K and Greene MI. (1989). Cell 58, 287-292. MEDLINE

Kraus MH, Popescu NC, Amsbaugh C and King CR. (1987). EMBO J. 6, 605-610. MEDLINE

Kraus MH, Issing W, Miki T, Popescu NC and Aaronson SA. (1989). Proc. Natl. Acad. Sci. USA 86, 9193-9197. MEDLINE

Kraus MH, Fedi P, Starks V, Muraro R and Aaronson SA. (1993). Proc. Natl. Acad. Sci. USA 90, 2900-2904. MEDLINE

Lubin R, Zalcman G, Bouchet L, Tredanel J, Legros Y, Cazals D, Hirsch A and Soussi T. (1995). Nat. Med. 1, 701-702. MEDLINE

Mariani Costantini R, Muraro R, Ficari F, Valli C, Bei R, Tonelli F, Caramia F and Frati L. (1991). Cancer 67, 2880-2886. MEDLINE

Nagata Y, Furugen R, Hiasa A, Ikeda H, Ohta N, Furukawa K, Nakamura H, Furukawa K, Kanematsu T and Shiku H. (1997). J. Immunol. 159, 1336-1343. MEDLINE

Nanda NK and Sercarz EE. (1995). Cell 82, 13-17. MEDLINE

Peoples GE, Goedegebuur PS, Smith R, Linehan DC, Yoshino I and Eberlein TJ. (1995). Proc. Natl. Acad. Sci. USA 92, 432-436. MEDLINE

Pinkas-Kramarsky R, Soussan L, Waterman H, Levkowitz G, Alroy I, Klapper L, Lavi S, Seger R, Ratzkin BJ, Sela M and Yarden Y. (1996). EMBO J. 15, 2452-2467. MEDLINE

Pupa SM, Menard S, Andreola S and Colnaghi MI. (1993). Cancer Res. 53, 5864-5866. MEDLINE

Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A and Press MF. (1989). Science 244, 707-712. MEDLINE

Tuttle TM. (1996). Cancer Immunol. Immunother. 43, 135-141. MEDLINE

Ullrich A, Coussens L, Hayflick JS, Dull TJ, Gray A, Tam AW, Lee J, Yarden Y, Libermann TA, Schlessinger J, Downward J, Mayes EL, Whittle N, Waterfield MD and Seeburg PH. (1984). Nature 309, 418-425. MEDLINE

van der Geer P, Hunter T and Lindberg RA. (1994). Annu. Rev. Cell Biol. 10, 251-337. MEDLINE

Vecchi M, Baulida J and Carpenter G. (1996). J. Biol. Chem. 271, 18989-18995. MEDLINE

Yamamoto T, Ikawa S, Akiyama T, Semba K, Nomura N, Miyajima N, Saito T and Toyoshima K. (1986). Nature 319, 230-234. MEDLINE

Zafiropoulos A, Andersson E, Krambovitis E and Borrebaeck CA. (1997). J. Immunol. Methods 200, 181-190. MEDLINE

Zhang K, Liu N, Wen D, Chang D, Thomanson A and Yoshinaga SK. (1996). J. Biol. Chem. 271, 3884-3890. MEDLINE

Figure 1 Specific detection of human ErbB receptors overexpressed in NIH3T3 transfectants. Immunoblot analysis of control NIH3T3 and transfectants overexpressing human EGFR (LTR-EGFR), ErbB2 (LTR-ErbB2), ErbB3 (LTR-ErbB3) or ErbB4 (LTR-ErbB4) using receptor-specific polyclonal antisera

Figure 2 Detection of anti-ErbB receptor antibodies in cancer patient sera. Immunoblot analysis of control NIH3T3 and ErbB receptor transfectants employing normal donor serum, cancer patient sera or receptor-specific polyclonal antisera as indicated

Figure 3 Reactivity of patient sera with human ErbB receptor immunoprecipitates. Receptor-specific and control immunoprecipitates from ErbB receptor transfectants were subjected to immunoblot analysis using cancer patient sera, receptor-specific antisera or human control serum as indicated

Figure 4 Reactivity of patient sera with ErbB receptors overexpressed by human tumor cells. (a) Immunoblot analysis of total protein lysates derived from SK-BR-3 or A431 tumor cell lines using patient and control sera. (b) Receptor-specific immunoprecipitates from SK-BR-3 or A431 tumor cell lines were immunoblotted using patient or control sera

Figure 5 ErbB receptors expression in tumor tissues. Immunohistochemical analysis of EGFR, ErbB2, ErbB3 and ErbB4 of a breast carcinoma (patient RT), a colon carcinoma (patient FA) and a stomach carcinoma (patient FB). Immunoreactivity was visualized by immunoperoxidase staining as described in Materials and methods using specific monoclonal and polyclonal antibodies. (a) patient RT; (b) patient FA; (c) patient FB. Original magnification 200´

Table 1 Table 1

Table 2 Table 2

Table 3 Table 3

Table 4 Table 4