vulval carcinoma derived cell line A431 (6). Direct sequence analysis of the EGF receptor has revealed extensive sequence. Received 4/29/85; revised 9/10/85; ...
[CANCER
RESEARCH
46, 285-292, January 1986]
Expression of Epidermal Growth Factor Receptors on Human Cervical, Ovarian, and Vulval Carcinomas William J. Gullick,1 Judith J. Marsden, Nigel Whittle, Bruce Ward, Linda Bobrow, and Michael D. Waterfield Protein Chemistry Laboratory [W. J. G., J. J. M., N. W., M. D. W.], Interferon Laboratory [B. W.¡,and Histology Unit [L B.], Imperial Cancer Research Fund Laboratories, Lincoln's Inn Fields, London WC2A 3PX, United Kingdom
homology with the transforming protein v-enb B present in several isolates of avian erythroblastosis virus (6). Comparison of the complete sequence of the human EGF receptor obtained by molecular cloning (7) with that of v-e/t> B showed that the latter
ABSTRACT We describe the properties of two monoclonal antibodies produced to a synthetic peptide consisting of residues 985 to 996 from the cytoplasmic domain of the epidermal growth factor (EGF) receptor. We have examined a group of ten human tumors including cervical, ovarian, and vulval carcinomas for expression of EGF receptors by immunohistological staining using one of these antibodies and another monoclonal antibody to the extra cellular domain of the molecule. The tumors were examined using a sensitive amplified enzyme system and a less sensitive indirect staining method. There was generally a good correlation in staining intensity with the two monoclonal antibody reagents. Both antibodies showed strong staining of squamous cell carci nomas and usually weak or heterogeneous patterns with the adenocarcinomas. Samples of each tumor were solubilized in detergent and analyzed for the presence of functional EGF receptors by ¡mmunoprecipitation and autophosphorylation. Three of the squamous cell tumors gave labeled bands, M, 170,000, on sodium dodecyl sulfate:polyacrylamide gels. DMA was extracted from seven of the tumors and digested with two restriction endonucleases, and the fragments were analyzed on Southern blots using probes representing the extracellular and cytoplasmic domains of the molecule. The tumor DMA showed no apparent rearrangements or amplifications when compared to the EGF receptor gene in human placental DNA. These results suggest that there is a high level of EGF receptors on some squamous cell tumors.
was homologous only with the transmembrane and cytoplasmic domains of the EGF receptor and lacked the majority of the extracellular NH2-terminal sequences (6, 7). It has been hypoth esized that expression of the v-enb B protein, which cannot bind EGF, in a constitutively active state contributes to uncontrolled growth and division of infected cells (6, 7). While this model remains to be proved it has stimulated examination of human malignant cell lines and primary tumors for their level of EGF receptor expression. A variety of different approaches have been taken including direct binding of iodinated EGF to breast cancer cells in culture (8), to human lung cancer cells (9), and to breast tumor biopsies (10) and by immunoperoxidase staining of cryosections of human malignant breast using a monoclonal anti body to EGF receptor (11). This antibody, called EGFR1, binds to the native folded external domain of the human and rat EGF receptors (12). It does not react with the v-ert> B protein presum ably since it recognizes sequences not present in the truncated molecule. Using the same mAb Ozanne and co-workers surveyed the expression of EGF receptors on samples of several normal human tissues (13) and squamous lung cancers (14). The level of EGF receptor protein has also been determined by ¡mmuno precipitation from detergent solubilized brain tumors (15,16) and malignant cell lines (17) using monoclonal and polyclonal antiEGF receptor antibodies. The organization of the gene(s) encod ing EGF receptor has also been analyzed in detail using cDNA EGF receptor clones as probes on Southern blots (16, 17). The vulval carcinoma cell line A431, which has been studied exten sively (18), possesses very high levels of EGF receptors per cell (2x106) and has manifold amplification of the gene(s) encoding
INTRODUCTION The EGF2 receptor is a M, 170,000 transmembrane glycoprotein found on many cell types (1-3). Its function is to bind the mitogen EGF and to transduce this signal across the cell mem brane to the cytoplasm. EGF binding induces a variety of receptor mediated effects including ion transport, morphological changes, and ultimately mitogenesis in appropriate cell types (2, 3). The only intrinsic property of the EGF receptor described to date is tyrosine protein kinase activity (4, 5). Whether this activity alone is responsible for the pleiotropic effects that EGF binding can induce has not been determined. Recently the human EGF receptor has been purified to ho mogeneity from normal human placenta and from the human vulval carcinoma derived cell line A431 (6). Direct sequence analysis of the EGF receptor has revealed extensive sequence
EGF receptors (7,19, 20). Libermann ef a/. (16) have shown that in some primary tumors of glial origin the EGF receptor gene is amplified. The A431 cell line also synthesizes a truncated form of the EGF receptor comprising the external glycosylated EGF binding domain which is excreted into the growth medium (21, 22). It has, however, been difficult to examine cell lines or primary tumors for truncated EGF receptors representing the cyto plasmic domain since they would bind neither EGF nor the mAbs used to date. To address this problem we have made two mAbs to a synthetic peptide from the cytoplasmic domain of the EGF receptor. These mAbs are used together with EGFR1 mAb to examine the immunohistological localization of EGF receptors on ten tumor samples of cervical, ovarian, and vulval origin. We also examined the tyrosine protein kinase activity of the EGF receptor by immunoprecipitation of solubilized tumor followed by
Received 4/29/85; revised 9/10/85; accepted 9/30/85. 1To whom requests for reprints should be addressed. 2The abbreviations used are: EGF, epidermal growth factor; ELISA, enzyme linked immunosorbent assay; mAb, monoclonal antibody; PBS, phosphate buffered saline (10 mM Na2HPO4:2.5 mw KH2P0.4:170 mm NaCI:3 mm KCI); cDNA, comple mentary DNA; IAP, immunoalkaline phosphatase; APAAP, alkaline phosphataserinti alkaline phosphatase.
CANCER RESEARCH
VOL. 46 JANUARY 1986 285
EOF RECEPTORS
ON HUMAN TUMORS with 9 volumes of phosphorylation
autophosphorylation and gel electrophoresis. Finally we isolated DNA from the tumors and investigated the structure and number of copies of the EGF receptor gene by Southern blot analysis.
buffer [50 mM 4-(2-hydroxyethyl)-1 -
piperazineethanesulfonic acid buffer, pH 7.4, containing 150 mw NaCI, 0.2% (v/v) Triton X-100, 2 mw MnCI2, and 100 //M sodium o-vanadate] and spun in an Eppendorf centrifuge for 2 min. The supernatant was removed, purified murine EGF (4 Mg/ml) was added, and the mixture was incubated for 30 min at room temperature. The solution was cooled to 0°C, [7-32P]ATP (Amersham, United Kingdom) was added to a final
MATERIALS AND METHODS
concentration of 10 ^M reaction was allowed to mixed with 50 n\ of 10x nQ of purified mAb and
Monoclonal Antibody Production. The peptide representing residues 985 to 996 of the human EGF receptor (7) with the sequence DVVDADEYLIPQ was synthesized as described previously (23). It was coupled to keyhole limpet hemocyanin as follows. Peptide (5 mg) in 350 pi of phosphate buffered saline was mixed with keyhole limpet hemocyanin (5 mg) in 145 n\ of PBS and 5 ^l of 25% (v/v) glutaraldehyde (BDH, Poole, United Kingdom) was added, and the mixture was stirred at room temperature for 15 min. A further 2.5 ¡Aof glutaraldehyde were then added and the reaction was continued for 15 min. A 1 M solution of glycine-HCI (100 ¡t\;pH 6.6) was then added and stirred for an additional 10 min to stop the reaction. The conjugate (600 n\) was mixed well with 3.9 ml PBS and 4.5 ml adjuvant (Cappel, Cochranville, PA). BALB/c mice, 6 weeks old, received on day 0 200 n\ of conjugate, mixed with complete Freund's adjuvant, s.c. at several sites. The animals were
suspension (20 ^l) of Protein A:Sepharose (Pharmacia, Uppsala, Sweden) in PBS was added and the incubation was continued at 4°Cfor 30 min. The pellets were washed by centrifugaron as described (26) and run on 7% polyacrylamide gels containing sodium dodecyl sulfate. The gels were stained, destained, dried, and then autoradiographed for approxi mately 18 h at -70°C using Kodak XAR5 film. Immunoprecipitation of v-erb B Protein. The chicken erythroblast cell line AEV 6C2 infected with AEV ES4 virus was labeled with [MS]methionine as follows. Logarithmically growing cells (5 x 107) were incubated for 1 h at 37°C in 1.0 ml of Dulbecco's modified Eagle's medium containing 10% of the normal methionine concentration, 10 mw 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer (pH 7.2), and 0.5 mCi [MS]methionine (3000 Ci/mmol; Amersham). The cells were then washed with 10 ml Dulbecco's modified Eagle's medium at 0°Cby gentle
boosted twice at 2-week Intervals in the same way but using incomplete Freund's adjuvant. Serum samples were taken from the tail on days 28 and 35 and were tested in ELISA assay against immunizing peptide and by immunoprecipitation of autophosphorylated EGF receptor as de scribed below. The mouse used in the fusion described here received an i.p. injection of 200 pi of conjugate in PBS on day 39 and was sacrificed on day 42. The spleen cells were fused with NS-1 myeloma cells using
centrifugation. The cell pellet was lysed with 2 ml of 50 mM Tris-HCI buffer, pH 7.4, containing 1% v/v Triton X-100, 150 mw NaCI, 25 mM benzamidine, 5 mM ethylene glycol bis(/3-aminoethyl ether)-/v,/v,/v',A/'tetraacetic acid, 0.1% bovine serum albumin, and phenylmethylsulfonyl fluoride, 50 ng/ml. Protein A:Sepharose (Pharmacia), 20 p\ of a 1:1 suspension in PBS, was mixed with 2 ng of mAb F4 and incubated at room temperature for 10 min. The pellets were washed with 1.0 ml PBS by centrifugation. Erythroblast lysate (100 n\) was added, in either the presence or the absence of 5 mM peptide 2E, and the tubes were agitated at 4°Cfor 1 h. The pellets were washed as described (26) and
polyethylene glycol 4000 (Merck, Darmstadt, West Germany) as de scribed (24). The fused cell pellet was resuspended in 50 ml of medium and distributed into approximately 500 wells in 96-well tissue culture plates. After 14 days the supernatant from each well was tested for the production of immunoglobulin specific for the immunizing peptide by ELISA. ELISA. A concentrated solution of peptide (10 mg/ml in PBS) was diluted to 10 ¿jg/mlin 200 mw NaHCO3 buffer, pH 9.2, and 50 ¿ilwere added to wells of 96-well tissue culture plates. Plates from several
sample buffer (50 M!) containing /3-mercaptoethanol was added. The samples were heated to 100°Cfor 2 min, then electrophoresed on 10% polyacrylamide gels containing sodium dodecyl sulfate, and treated with AMPLIFY (Amersham) before drying. The dried gels were exposed to Kodak XAR5 film for 16 to 18 h at -70°C using intensifying screens.
manufacturers were tested; those from Nunc Inter Med (Komstrup, Denmark) were found to give the best signalinoise ratio at a given serum dilution. Coating with the peptide alone had the advantage of excluding the contribution of anti-carrier and anti-glutaraldehyde antibodies to the
Tissue Preparation. Fresh samples of five ovarian, two cervical, and three vulval carcinomas were obtained from surgical excision biopsies. Several portions of each sample were immediately snap frozen and stored in liquid nitrogen, and one portion was routinely processed for conventional histological diagnosis. Similar preparations of normal skin and cervix were used as controls for the immunohistological studies. Immunohistology. Immunostaining was performed on frozen tissue sections using the indirect IAP and unlabeled antibody-enzyme (APAAP)
color reaction. The peptide solution was left in the plates overnight at 4°C.Another set of plates was incubated with buffer alone (blank plates). The peptide or buffer was then discarded and the plates were washed three times by inundation with PBS containing 0.05% Tween 20 (Sigma). Nonspecific binding was blocked with PBS, 200 ¿¿l/well, containing 0.5% bovine serum albumin for 1 h at room temperature and then the plates were washed again as above. Supernatant (25 ¿/I)from each fusion well was transferred to an equivalent position on plates coated with peptide and blank plates and incubated at room temperature for 2 h. The plates were then washed and 50 /¿I of a 1:500 dilution in PBS of rabbit immunoglobulin:horseradish peroxidase conjugate (Dako Immu-
techniques (27). Air dried cryostat sections (5 ^m) were prefixed in cold (4°C) chlorofornracetone (1:1) for 5 min and incubated with ascites preparation of mAb EGFR1 and concentrated supernatant of mAb F4 at dilutions of 1:100 and 1:10, respectively. Subsequent layers consisted of sheep anti-mouse IgG (Fab2):alkaline phosphatase conjugate (1:50; Sigma) in the IAP test and rabbit anti-mouse immunoglobulin (Dakopatts,
noglobulins, Denmark) were added and incubated for a further 2 h at room temperature. The plates were then washed and 100 fil of substrate solution [o-phenylenediamine, 2 mg/ml (Sigma), in 50 HIM sodium phos phate, to pH 6.0 with 1.0 M citric acid] were added. The plates were left to develop color in the dark for 20 min at room temperature, and their absorbance at 450 nw was measured in an automated plate reader (Titertek Multiskan, Titertek, Finland). Only wells with supernatants that were positive on peptide coated plates and negative on blank plates were examined further. Cells from these wells were transferred to 2-ml
Denmark) at a final dilution of 1:25 in normal mouse serum followed by monoclonal (mouse) APAAP (1:10; Dakopatts) reagent in the unlabeled antibody-enzyme test. All antibody dilutions and intervening washes were made in 0.1 M PBS, pH 7.2. The red enzyme reaction was developed with Naphthol AS Bl phosphate and Fast Red TR in Veronal acetate buffer, pH 9.4, containing 1 mM levamisole to block endogenous alkaline phosphatase. Sections were counterstained in Mayer's hemalum and examined by light microscopy. Immunoprecipitation from and Phosphorylation of Human Tumor Extracts. This was done essentially as described (15) with the modifi cations that purified murine EGF, 4 ng/m\, was included in the phos phorylation mixture and the mAb used was EGFR1 at 5 ^g/immunopre-
wells and retested by ELISA. Positive cultures were expanded and cloned twice without feeder cells by limiting dilution. Autophosphorylation and Immunoprecipitation of EGF Receptor. Membranes were prepared from A431 cells (25); an aliquot was mixed
CANCER
RESEARCH
and a specific activity of 5 ^Ci/Mmol, and the continue for 10 min; 100 n\ of this were then concentrated hybridoma supernatant or 1 to 5 incubated with agitation at 4°Cfor 1 h. A 1:1
VOL. 46 JANUARY 286
1986
EOF RECEPTORS ON HUMAN TUMORS cipitate. Isolation of DNA and Southern Blot Analysis. Samples of frozen tumors (approximately 100 mg) were homogenized in liquid nitrogen and the resulting powder was added directly to 10 ml of 100 mw Tris-HCI buffer, pH 7.6, containing 100 HIM EDTA, 150 HIM 0-mercaptoethanol, and 1% (v/v) Sarkosyl. The solution was mixed gently for 2 h at 50°C and then overnight at 37°C with the addition of proteinase K (100 ^9/ ml; BRL, Cambridge, United Kingdom). The solution was then made 100 Mg/ml with ethidium bromide and 90% (w/v) cesium chloride was added. Samples were centrifuged at 110,000 x g for 48 h in a Beckman 70 Ti rotor and the DNA band was isolated using a 15-gauge needle. The DNA
Table 1 Properties of the monoclonal antibodies to the EGF receptor used for immunohistological staining and immunoprecipitation
mAbEGFR1
cytoplasmic domain of the EGF receptorExternal
chainK
(12)
6x 10-'
F4 Cytoplasmic lgG1igeiLight KKAffinity0Kd(M)NO6 NDC D10External/ CytoplasmicSubclasslgG2b 8 Determined by radioimmunoassay using solubilized A431 membranes (26). 0 ND, not determined. c The binding affinity of D10 to native EGF receptor was too low to measure
sample was extracted with butanol and dialyzed extensively against 10 rriM Tris-HCI buffer, pH 8.0, containing 1 mw EDTA, and the volume was reduced by dehydration using Sephadex G-50 superfine (Pharmacia).
reliably.
DNA (15 M9) was digested overnight in 400 ¿ilbuffer using 200 units of the indicated enzyme (Fig. 4) (BRL). After ethanol precipitation the samples were electrophoresed on 0.8% agarose gels at 2 V/cm and then transferred to Biodyne A transfer membranes (Pall Ultrafine Filtration Corporation, NY). The filter bound DNA was prehybridized and hybridized for 16 h with nick-translated 32P-labeled cDNA inserts (28). Washing
MWxIO
included 1 h in 30 mw sodium chloride, 3 mw sodium citrate, and 0.1% (w/v) sodium dodecyl sulfate at 50°C. Filters were subject to autoradiographic analysis using Kodak XAR5 with intensifying screens at -70°C
B
-3
-3
MWx
10
200 r
200r
for the indicated times.
RESULTS 116
116 Preparation of Anti-Synthetic Peptide mAbs. In order to
93
develop reagents that could detect the cytoplasmic domain of the EGF receptor that were suitable for immunohistological ex amination of thin sections of tumors, we prepared mAbs to a synthetic peptide from the EGF receptor sequence. This peptide comprises residues 985 to 996 from the complete EGF receptor sequence of 1206 amino acids (7) and is thus about 200 residues from the COOH terminus but just outside the region of sequence homology shared by the src oncogene family (5). Polyclonal serum to this peptide raised in rabbits can recognize the native and denatured EGF receptor and v-erb B protein and reacts with the cytoplasmic domain of the molecule (23). We have produced two mAbs, designated F4 and D10, to this sequence which recognize only the appropriate peptide in ELISA (Chart 1). Both antibodies are of the lgG1 subclass (Table 1) and can immuno-
93 68
..
68 L
Fig. 1. Immunoprecipitation of EGF receptor and v-erb B by mAbs F4 and D10. In A. EGF receptor from A431 cell membranes was labeled with {y-^P^TP and immunoprecipitated by normal mouse serum, 5 ^l (track 1); 2 »gmAb F4, preincubated with 5 mM peptide 2E (track 2): 2 ¿igF4 mAb (track 3); 2 ¿¿g mAb D10 preincubated with 5 HIM peptide 2E (track 4);2t¡g mAb D10 (track 5). In B, cells infected with AEV ES-4 were labeled with ["SJmethionine and immunoprecipitated with 2 fig mAb F4, preincubated with 5 mM peptide 2E (track 7) or 2 //g mAb F4 (track 2). Arrows, position of the v-erb B protein.
0.5rprecipitate native autophosphorylated human EGF receptor (Fig. ÃŒA)and [35S]methionine labeled v-erb B (Fig. 1B) but have 0.4
relatively low binding affinities (Table 1). Immunohistology of Human Tumors. We obtained ten human tumors and examined them for their reaction with EGFR1 mAb to the extracellular domain and the F4 mAb to the cytoplasmic domain of the EGF receptor. The histological diagnosis and immunostaining results are shown in Table 2. Using a combina tion of the high sensitivity APAAP staining and the less sensitive IAP method, a clear distinction could be made between strong and weak reactivity patterns. The product of the alkaline phosphatase enzyme reaction is deposited as a red color. In general both antibodies produced stronger staining of the squamous cell carcinomas regardless of the anatomical state. Tumors 2, 3, 5, 9, and 10 (squamous cell carcinomas) were relatively strongly stained, although tumor 8 (adenocarcinoma), also gave strong staining with the F4 mAb. Tumors 1, 4, 6, and 7 (adenocarcino-
A450nm 0.3
0.2
0.1
640
2560
10240 40960
163840
DILUTION Chart 1. ELISA F4 p. A), D10 (*, »),and a peptide with the sequence
of mAbs F4 and D10 against immunizing and irrelevant peptides: •)against plates coated with the immunizing peptide, 2E (D, including the P3 phosphorylation site of the EGF receptor (30) DDTFLPVPEYINQS (A, •).
CANCER
RESEARCH
mas) gave weaker staining or were negative. Sections from VOL. 46 JANUARY
287
1986
EGF RECEPTORS
ON HUMAN TUMORS
Mr (x10-3)
Table2 Immunohistological
analysis of tumor samples using monoclonal antibodies to the EGF receptor Immunoenzyme reaction with mAbs
F4
EGFR1
Tumor12345678910tionOvaryCervixVulvaOvaryVulvaOvaryOvaryOvaryVulvaCervixHistological diagnosisSerous IAP± IAP APAAP
±+ papillary ad enocarcinomaSquamous —+ cell car cinomaSquamous cell car ±± cinomaInfiltrating -+ adeno carcinomaSquamous ++ cell car cinomaPoorly -differentiatedadenocarcinomaPoorly
—
+
±
+
+
+
-± differentiatedadenocarcinomaPoorly
NT*
+
+
+
+
±
+±
116
±
93 +
+
±
±
-
±
differentiatedadenocarcinomaSquamous cell car ++ cinomaUndifferentiatedsquamous
200
68
1234
5
6789
10
±
cellcarcinomaAPAAP 0 NT, not tested.
Fig. 3. Immunoprecipitation and phosphorylation of EGF receptor from extracts of human tumors. Tracks 7 to JO correspond to tumors 1 to 10 as indicated in Table 2. The labeled immunoprecipitation was run on 7.5% sodium dodecyl sulfate:polyacrylamide gels.
tumors 3 and 5 (squamous cell carcinomas) and tumor 7 (ade nocarcinoma), stained with both antibodies using the APAAP method, are shown in Fig. 2. Diffuse uniform staining was demonstrated on the tumor cells in both groups apart from sample 6 which contained foci of strong mAb EGFR1 positive cells. Immunoprecipitation and Phosphorylation from Tumor Ex tracts. We next examined the ten tumors for the presence of functional EGF receptors using the system developed in Ref. 15. Pieces of each tumor were dissolved in detergent and the EGF receptor was precipitated by mAb EGFR1. The immunoprecipitates were incubated with EGF and carrier free [7-32P]ATP so that any enzymically active EGF receptors would become autophosphorylated. Monoclonal antibody F4 was not used here because it is less sensitive than EGFR1 partly due to its lower affinity for the EGF receptor and partly because in binding close to the kinase domain it has an inhibitory effect on autophos phorylation on immunoprecipitates. This emphasizes that the radioactivity in the gel bands is a product of the amount of EGF receptor present and its autophosphorylating activity. Three tumors (tumors 3, 5, and 10; Fig. 3) gave labeled bands, M, 170,000, characteristic of the human EGF receptor. Tumor 10 (Fig. 3, track 10) also has a band at M, 150,000 which probably represents the breakdown product of the EGF receptor gener ated by a ubiquitous calcium activated neutral protease (29). In the tissue section studies these tumors were also strongly stained with the mAbs (Fig. 2; Table 2). The remaining tumors gave no apparent EGF receptor band by immunoprecipitation and showed variable immunohistological staining patterns (Table 2). This behavior may therefore reflect differences in the sensitiv ity of the assays or loss of enzyme activity during storage. Southern Blotting. DNA was prepared from tumors 1 to 7 (Table 2) and digested with the restriction endonucleases EcoRI and H/rtdlll. For comparison equal amounts of DNA from normal human placenta and from A431 cells [which are known to have CANCER
RESEARCH
amplified sequences coding for the EGF receptor (7, 19)] were included on the Southern blots. Two cDNA clones representing the 5' and 3' ends of the coding region of the human EGF receptor gene (Fig. 4A) were used to probe the blots (Fig. 4, 8 to £).The tumor DNA and normal placental DMA showed essen tially the same pattern of fragments with each enzyme and each probe and gave similar intensities of hybridization. The DNA from A431 cells, however, gave a much greater extent of hybridization (Fig. 4) as reported previously (7, 17, 19, 20). There were, however, several small differences in fragment patterns that could be discerned on close examination of the autoradiographs. In Fig. 4E, the material digested with Hind\\\ and probed with the cytoplasmic domain cDNA probe showed some differences in mobility between the 6-kilobase fragments. This behavior has, however, been observed when comparing DNA isolated from groups of normal individuals and thus probably represents poly morphism.3 An additional band of about 9 kilobases was present in Fig. 40, track C, which was not present in normal DNA. This tumor (a squamous cell vulval carcinoma, tumor 3) displayed high levels of EGF receptor both by ¡mmunostaining (Fig. 2; Table 2) and by immunoprecipitation and autophosphorylation (Fig. 3, track 3). We are currently investigating whether this behavior represents significant rearrangement or polymorphism of EGF receptor coding sequences. These results, however, indicate that there has been no gross amplification of the EGF receptor genes in the tumors examined. DISCUSSION Little is known in detail about the pattern and mechanism of EGF receptor expression on human tumors and its relevance to the transformed state. In this work we have examined ten tumors 3 N. Whittle, unpublished results.
VOL. 288
46 JANUARY
1986
EGF RECEPTORS ON HUMAN TUMORS lem. It would also be desirable to develop probes that extend 5'
for their localization and level of expression of EGF receptor protein and its intrinsic kinase activity and gene organization. As a first level of analysis we have localized EGF receptors on thin cryosections of selected normal and malignant tissue using mAbs to the extracellular and cytoplasmic domains of the molecule. Variable reactivity patterns were obtained using the IAP staining method and the more sensitive (APAAP) technique. Both mAbs exhibited strong staining of squamous cell carcinomas (13) and usually weak or heterogeneous patterns in the adenocarcino-
to the sequence coding for the EGF receptor proteins signal sequence so that the structure of the normal and tumor transcriptional promotors could be analyzed. Additional methods of increasing gene transcription include chromosomal translocations and polysomy, both found in A431 cells (31) and believed to be responsible for the activation of cmyc in Burkitt's lymphoma (32). Other mechanisms than those increasing transcription rates may be envisaged. Either increased mRNA stability or decreased EGF receptor degradation rates could lead to higher levels of EGF receptor protein. Indeed the EGF receptor in A431 cells appears to be degraded more slowly than in human fibroblasts (33) but since the EGF receptor gene is amplified in A431 cells it is difficult to say what effect this has on absolute receptor protein levels. Finally as suggested by the v-erb B model, alterations in receptor signaling activity may be important. Subtle structural changes such as point mutations are difficult to detect by the methods outlined above but could have significant effects on EGF receptor function. Such changes would not necessarily involve increased EGF receptor expres sion. It should be emphasized that there is no direct evidence that overexpression or aberrant expression of EGF receptors is in volved in the transformation of human cells. Transfection of cells with constructs of the EGF receptor gene by which the level of receptor protein can be varied, as has been done with c-src and v-src (34-36), would help to analyze this question. Examination
mas. A wider and more detailed study is at present under way, using the sensitive procedure described, to provide more infor mation on EGF receptor expression on normal human tissues. There was generally good correlation in staining intensity with the two mAbs which suggests that both reagents recognize the same molecule and no individual tumor within this group ex presses exclusively truncated EGF receptors. It is not possible to detect by this method cells which express a low level of truncated EGF receptors (should such a situation exist in vivo as it clearly does in the A431 cell in culture). It would be informative to separate tumor derived EGF receptors by size and examine them by Western blotting with polyclonal serum to many epitopes to detect all the species present and to describe these in detail using site specific antipeptide antibodies. To date our attempts to achieve this have been unsuccessful, partly due to the small amount of protein that can be loaded on a single gel lane and partly due to the inefficiency of transfer of the high molecular weight form of the EGF receptor. Immunoprecipitation and autophosphorylation of tumor ex tracts confirmed that several tumors positive immunohistologi-
of human tumors with high levels of EGF receptor expression for the levels of transcription of other oncogenes such as c-myc and c-fos would also be valuable. Meanwhile, however, exami
cally contained functional EGF receptors. The detection of such a function is contingent on the receptor reacting with the mAb, possessing an active autophosphorylating activity and the sites at which phosphate may react covalently. The three major sites of EGF receptor autophosphorylation have been mapped (30) to the extreme carboxy terminus of the molecule. Nonetheless EGF receptor which lacks its COOH-terminal 20,000 daltons of se quence, cleaved off by the calcium activated protease, will be come phosphorylated in an immunoprecipitate (Fig. 3, track 10, for instance) probably at a fourth site that has not yet been localized. Some tumors which were positive by immunostaining (Table 2) did not give visible immunoprecipitates. Whether this results from a difference in sensitivity of the assay, lack of kinase activity, or autophosphorylation sites is not clear. DNA from the first seven tumors was analyzed by Southern blotting and showed no gross amplifications or rearrangements of the 5' or 3' EGF receptor coding sequences. Libermann er al.
nation of a wide spectrum of human tumors may aid in tumor classification (notably in the brain). A recent report (37) has indicated a correlation between invasiveness of human bladder tumors and a high level of EGF receptor expression suggesting that such analysis may be useful in predicting tumor behavior. Finally the relative levels of EGF receptors on some tumors versus normal tissues and the battery of anti-EGF receptor mAbs available (38) suggest that it might be worthwhile testing cytotoxic antibody conjugates in model systems. ACKNOWLEDGMENTS We are grateful to Dr. R. Sainsbury for providing his unpublished results and to Audrey Becket for typing the manuscript.
REFERENCES
(15, 16) have clearly demonstrated that such changes have occurred in some glial cell derived tumors. However, not all tumors that had high levels of EGF receptor protein had amplified EGF receptor genes. In addition several cell lines with relatively high levels of EGF receptor protein showed little or no alteration in their EGF receptor gene structure (17). As stated by these authors gene amplification "is not the sole mechanism whereby cell types generate enhanced levels of EGF receptor protein."
1. Adamson, E. D., and Rees, A. R. Epidermal growth factor receptors. Mol. Cell. Biochem., 34: 129-152, 1981. 2. Carpenter, G. The biochemistry and physiology of the receptor-kinase for epidermal growth factor. Mol. Cell. Endocrinol.. 3Õ:1-19, 1983. 3. Schlessinger, J., Schreiber, A. B., Levi, A., Lax, l., Libermann, T., and Yarden, Y. Regulation of cell proliferation by epidermal growth factor. CRC Crit. Rev. Biochem., 74:93-111,1983. 4. Ushiro, H., and Cohen, S. Identification of phosphotyrosine as a product of epidermal growth factor activated protein kinase in A431 cell membranes. J. Biol. Chem., 255: 8363-8365, 1980. 5. Hunter, T., and Cooper, J. A. Protein tyrosine kinases. Annu. Rev. Biochem., 54:897-930,1985. 6. Downward, J., Yarden, Y.. Scrace, G., Totty, N., Stockwell, P., Ullrich, A., Schlessinger, J., and Waterfield, M. D. Close similarity of epidermal growth factor receptor and v-erb B oncogene protein sequences. Nature (Lond.), 307: 521-527,1984. 7. Ullrich, A., Coussens, L, Hayflick, J. S., Dull, T. J., Gray, A., Tarn, A. W., Lee, J., Yarden, Y., übermann,T. A., Schlessinger, J., Downward, J., Mayes, E. L.
The presence of EGF receptors in these cell lines did, however, correlate with elevated levels of EGF receptor mRNA (17). De spite the good morphological preservation of the tumors studied in this work we could not obtain intact mRNA from them. In situ hybridization of EGF receptor specific nucleic acid probes to cryosections might be an alternative way to address this prob CANCER
RESEARCH
VOL. 289
46 JANUARY
1986
EGF RECEPTORS
8. 9. 10.
11. 12.
13. 14. 15. 16.
17. 18. 19.
20.
21. 22.
ON HUMAN TUMORS
V., Whittle, N., Waterfield, M. D., and Seeburg, P. H. Humanepidemial growth factor receptor cDNA sequence and aberrant expression of the amplifiedgene in A431 epidermoid carcinoma cells. Nature (Lond.), 309: 418-425,1984. Fitzpatrick, S. L., LaChance, M. P., and Schultz, G. S. Characterization of epidermalgrowth factor receptor and action on human breast cancer cells in culture. Cancer Res., 44: 3442-3447,1984. Sherwin, S. A., Minna, J. D., Gazdar, A. F., and Todaro, G. J. Expression of epidermal and nerve growth factor receptors and soft agar growth factor production by human lung cancer cells. Cancer Res., 41: 3538-3542, 1981. Fitzpatrick, S. L., Brightwell, J., Wittliff, J. L, Barrows, G. H., and Schultz, G. S. Epidermalgrowth factor binding by breast tumor biopsies and relationship to estrogen receptor and progestin receptor levels. Cancer Res., 44: 34483453, 1984. Sainsbury,J. R. C., Farndon,J. R., Sherbet, G. V., and Harris, A. L. Epidermal growth factor receptors and estrogen receptors in human breast cancer. Lancet, 7:364-366, 1985. Waterfield, M. D., Mayes, E. L. V., Stroobant, P., Bennett, P. L. P., Young, S., Goodfellow, P. N., Banting, G. S., and Ozanne, B. A monoclonal antibody to the human epidermal growth factor receptor. J. Cell. Biochem., 20: 149-161, 1982. Gusterson, B., Cowley, G., Smith, J. A., and Ozanne, B. Cellular localization of human epidermal growth factor receptor. Cell Biol. Int. Rep., 8: 649-658, 1984. Hendler, F. J., and Ozanne, B. Human squamous cell lung cancers express increased epidermal growth factor receptors. J. Clin. Invest., 74: 647-651, 1984. Libermann, T. A., Razón,N., Bartel, A. D., Yarden, Y., Schlessinger, J., and Soreq, H. Expression of epidermal growth factor receptors in human brain tumors. Cancer Res., 44: 753-760,1984. Libermann, T. A., Nusbaum, H. R., Razón,N., Kris, R., Lax, I., Soreq, H., Whittle, N., Waterfield, M. D., Ullrich, A., and Schlessinger, J. Amplification, enhanced expression and possible rearrangement of EGF receptor gene in primary human brain tumors of glial origin. Nature (Lond.), 373: 144-147, 1984. Xu, Y-H., Richert, N., Ito, S., Merlino, G. T., and Pastan, I. Characterizationof epidermal growth factor receptor gene expression in malignant and normal human cell lines. Proc. Nati. Acad. Sci. USA, 87: 7308-7312,1984. Stoscheck, C. M., and Carpenter,G. Biologyof the A431 cell: a usefulorganism for hormone research. J. Cell. Biochem.,23: 191-202, 1983. Lin, C. R., Chen, W. S., Kruiger. W., Stolarsky, L. S., Weber, W., Evans, R. M., Verma, I. M., Gill, G. N.. and Rosenfeld,M. G. Expressioncloning of human EGF receptor complementary cDNA: gene amplification and three related messenger RNA products in A431 cells. Science (Wash. DC), 224: 843-848, 1984. Merlino, G. T., Xu, Y-H., Ishii, S., Clark, A. J. L, Semba, K., Toyoshima, K., Yamamoto, T., and Pastan, I. Amplification and enhanced expression of the epidermalgrowth factor receptor gene in A431 humancarcinomacells. Science (Wash. DC), 224: 417-419,1984. Mayes, E. L. V., and Waterfield, M. D. Biosynthesis of the epidermal growth factor receptor in A431 cells. EMBO J., 3: 531-537, 1984. Weber, W., Gill, G. N., and Spiess,J. Production of an epidermalgrowth factor
CANCER
RESEARCH
receptor-related protein. Science (Wash. DC), 224: 294-297, 1984. 23. Kris, R. M., Lax, I., Gullick, W., Waterfield, M. D., Ullrich,A., Fridkin, M., and Schlessinger, J. Antibodies against a synthetic peptide as a probe for the kinase activity of the avian EGF receptor and v-erb B protein. Cell, 40: 619625,1985. 24. Galfre, G., Howe, S. C., Milstein, C., Butcher, G. W., and Howard, J. C. Antibodies to major histocompatibility antigens produced by hybrid cell lines. Nature (Lond.), 266: 550-552,1977. 25. Thorn, D., Powell, A. J., Lloyd, C. W., and Rees, D. A. Rapid isolation of plasma membranes in high yield from cultured fibroblasts. Biochem. J., 768: 187-194,1977. 26. Gullick, W. J., Downward, D. J. H., Marsden, J. J., and Waterfield, M. D. A radioimmunoassay for human epidermal growth factor receptor. Anal. Biochem., 747. 253-261, 1984. 27. Mason, D. Y., Abdulaziz, Z., Falini,B., and Stein, H. In: J. M. Polak and S. van Noorden (eds.), PracticalApplications in Pathologyand Biology, pp. 113-128. Bristol, London, Boston: Wright PSG, 1983. 28. Maniatis, T., Fritsch, E. F., and Sambrook, J. MolecularCloning:A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1982. 29. Yeaton, R. W., Upari, M. T., and Fox, C. F. Calcium mediated degradation of epidermal growth factor receptor in dislodged A431 cells and membrane preparations.J. Biol. Chem., 258: 9254-9261,1983. 30. Downward, D. J. H., Parker, P., and Waterfield, M. D. Autophosphorylation sites on the epidermal growth factor receptor. Naure (Lond.), 377: 483-485, 1984. 31. Shimizu, N., Kondo, I., Gamou, S., Behzadian,M. A., and Shimizu, Y. Genetic analysis of hyperproduction of epidermal growth factor receptors in human epidermoid carcinoma A431 cells. Somatic Cell Mol. Genet., 70: 45-53,1983. 32. Klein,G. Specificchromosomaltranslocationsand the genesisof B cell derived tumors in mice and men. Cell, 32: 311-315,1983. 33. Krupp, M. N., Connolly, D. T., and Lane, M. D. Synthesis, turnover and downregulation of epidermal growth factor receptors in human A431 epider moid carcinoma cells and skin fibroblasts. J. Biol. Chem., 257: 11489-11496, 1982. 34. Jakobovitz, E. B., Majors, J. E., and Varmus, H. E. Hormonalregulation of the Rous sarcoma virus src gene via a heterologous promoter defines a threshold dose for cellular transformation. Cell, 38: 757-765,1984. 35. Shalloway, P., Coussens, P. M., and Yaciuk, P. Overexpression of the c-src protein does not induce transformation of NIH 3T3 cells. Proc. Nati. Acad. Sci. USA, 87:7071-7075, 1984. 36. Iba, H., Takeya,T., Cross, F. R., Hanafusa,T., and Hanafusa,H. Rous sarcoma virus variants that carry the cellular src gene instead of the viral src gene cannot transform chicken embryo fibroblasts. Proc. Nati. Acad. Sci. USA, 87: 4424-4428,1984. 37. Neal, D. E., Marsh, C., Bennett, M. K., Abel, P. D., Hall, R. R., Sainsbury, J. R. C., and Harris, A. L. Epidermalgrowth factor receptors in human bladder cancer: comparison of invasive and superficial tumors. Lancet, 7: 366-368, 1985. 38. Gullick, W. J., Marsden, J. J., and Waterfield, M. D. The production and use of antibodies to the epidermal growth factor receptor using natural and synthetic immunogens. Hormones Cell Régulât., 9: 37-46, 1985.
VOL. 46 JANUARY 290
1986
EGF RECEPTORS ON HUMAN TUMORS
H/E
F4
EGF R1
CO
i-
o 3
• .• '
'V •;-;
ID
i_ O 3
O 3
Fig. 2. Immunohistologicalstaining of frozen sections of human tumors. A, D, G, H & E stained sections; B. E, H, stained with EGFR1 mAb; C, F. I. stained with F4 mAb. The tumor numbers correspond to those in Table 2. Tumors 3 and 5, vulva, squamous cell carcinomas; tumor 7, ovary, poorly differentiated adenocarcinoma.
CANCER
RESEARCH
VOL. 46 JANUARY 291
1986
A.
extracellular
intracellular protein
6~
5
EcoRI
EcoRI
EcoRI
EcoRI
kbp
EcoRI cDNA
B.
a
b c
, D.
V
C.
a
d e
f
g h
i
j
a
b c d
e
f
g
g
h
i
-23.1
-23.1
-9.4
-9.4
6.6
-6.6
--
bed
e f
-4.4
-4.4
-2.2
-2.2
E.
h
abcdefgh
t -23.1
i -9.4 -6.6 -4.4
l
2.2
-2.2
Fig. 4. Analysis of EGF receptor gene sequences by Southern blotting. The EcoRI subclones of the EGF receptor cDNA used as probes are shown diagramatically in A. Probe I corresponds to most of the external and transmembrane domains, while probe II represents the COOH terminus and part of the kinase domain (7). Genomic DMA (15 tig) was digested with EcoRI (B and C) or H/ndlll (0 and £).The blots were hybridized with probe I (B and D) or probe II (C and E). Slight differences in intensity of hybridization signals can be accounted for by variations in the amount of DMA loaded. Tracks a to g correspond to tumors 1 to 7 (Table 2); 'rack h, A431 cell DMA; track i, human placental DNA; track j, A431 cell DMA. Filters were autoradiographed at 70°Cfor 16 h except for tracks j. B and D. which were exposed for 2 h at -70°C.
CANCER
RESEARCH VOL. 46 JANUARY 1986 292
