Proc. Nall. Acad. Sci. USAVol. 81, pp. 5227-5231, August 1984Medical Sciences

Monoclonal antibodies of predefined specificity detect activated rasgene expression in human mammary and colon carcinomas

(onc genes/synthetic peptide/breast cancer/tumor-associated antigens)

P. HORAN HAND, A. THOR, D. WUNDERLICH, R. MURARO, A. CARUSO, AND J. SCHLOM*Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, MD 20205

Communicated by Hilary Koprowski, May 8, 1984

ABSTRACT Monoclonal antibodies (MAbs) of predefinedspecificity have been generated by utilizing a synthetic pep-tide reflecting amino acid positions 10-17 of the Hu-ras'"4 geneproduct as immunogen. These MAbs, designated RAP-1through RAP-5 (RA, ras; P, peptide), have been shown to re-act with the ras gene product p2l. Since the Hu-ras reactivedeterminants (positions 10-17) have been predicted to be with-in the tertiary structure of the p21 molecule, it was not unex-pected that denaturation of cell extracts or tissue sections withFormalin or glutaraldehyde enhanced binding of the RAPMAbs. When paraffin-embedded Formalin-fixed tissue sec-tions and the avidin-biotin complex immunoperoxidase meth-od were used, the RAP MAbs clearly defined enhanced ras p21expression in the majority of human colon and mammary car-cinomas. The majority of all abnormal ducts and lobules fromfibroadenoma and fibrocystic disease patients were negative,as were all normal mammary and colonic epithelia examined.The findings reported here form the basis for quantitative ra-dioimmunoassays for a ras translational product and provide ameans to evaluate ras p21 expression within individual cells ofnormal tissues and benign, "premalignant," and malignant le-sions.

ras oncogenes were originally recognized as the transform-ing genes of two rat-derived viruses, the Harvey (Ha) andKirsten (Ki) strains of murine sarcoma virus (1). Molecularcloning studies have identified Ha-ras, Ki-ras, and subse-quently N-ras, as members of a gene family present in a widerange of species, including humans (2). At least two mecha-nisms have been identified by which ras activation can medi-ate transformation of NIH 3T3 mouse cells: (i) a point muta-tion in the ras gene resulting in an alteration of a single aminoacid in the 21,000-dalton ras gene product, p21 (3), or (ii)increased expression of p21 (1, 4, 5).

ras activation has been detected in certain human tumorcell lines and selected biopsy materials from a variety of can-cers by the demonstration of the transforming potential ofDNA from those specimens for NIH 3T3 cells; only a smallpercentage (usually less than 10-30%) of specimens testedfrom any given human tumor system, however, have provenpositive for activated ras in this system (ref. 6; S. Aaronson,personal communication). One possible explanation for thisis that the mechanism of ras activation in these tumors mayinvolve the increased expression of a normal cellular rasgene and gene product [e.g., hormone activation of ras via apromoter sequence (7)].An alternative approach to determining whether correla-

tions exist between enhanced ras gene expression (activa-tion) and a particular neoplastic state would be the use ofimmunologic assays directed against p21. Radioimmunoas-says (RIAs) would provide a quantitative determination ofras gene expression, while immunohistochemical studies

would define ras gene expression in circumstances in whichother methods are not suitable-i.e., the determination ofras gene expression in normal epithelium, hyperplastic le-sions, and adjacent malignant tissues from the same patient.In light of the provocative findings on ras gene activation inselected colon carcinoma cell lines (6) and a human breastcarcinosarcoma cell line (S. Aaronson, personal communica-tion) and the recent finding of ras gene activation in carcino-gen-induced rat mammary tumor models (8, 9), we under-took the development of monoclonal antibodies (MAbs) to ahuman ras (Hu-ras) gene product. Since the DNA sequenceof the entire human Ha-ras gene has been defined (3) and thestudies of Lerner (10) have demonstrated an extremely effi-cient utilization of polyclonal antibodies prepared againstsynthetic peptides, the approach of preparing MAbs to a pre-defined set of ras amino acid sequences was undertaken.

MATERIALS AND METHODSPeptides. The following peptides (positions 10-17 of the

Hu-ras gene product) were obtained from Peninsula Labora-tories (San Carlos, CA): Hu-rasT24, Gly-Ala-Val-Gly-Val-Gly-Lys-Ser, and Hu-rasHa, Gly-Ala-Gly-Gly-Val-Gly-Lys-Ser. Tyrosine was added to the amino terminus of each pep-tide for ease of iodination and lysine was added to thecarboxyl terminus to facilitate solubilization of the peptides.Peptides were coupled to thyroglobulin as described (11).Hybridoma Methodology. BALB/c mice were immunized

by intraperitoneal inoculation with 10 ,ug of the Hu-rasT24peptide coupled to 120 ,ug of thyroglobulin on day 1 (in com-plete Freund's adjuvant) and day 7 (in incomplete Freund'sadjuvant). At day 14, 1 ,ug of peptide coupled to 12 ,ug ofthyroglobulin was administered intravenously. Spleens wereremoved at day 17 for cell fusion (12, 13). All hybridoma celllines were cloned twice. MAbs B1.1 [directed against carcin-oembryonic antigen (14)], UPC-10 [a purified mouse myelo-ma protein of the IgG2a isotype (15)], and YA6 259 [a ratanti-rat ras, supplied by M. Furth (16)] have been described.Purified immunoglobulins were prepared as described (17).Immunoperoxidase Studies. The avidin-biotin complex

immunoperoxidase assay was performed on 5-,m sectionsfrom Formalin-fixed paraffin-embedded tissues as described(18, 19). In all cases an isotype-identical primary antibody(UPC-10) was used as a negative control.

RESULTSGeneration of MAbs. Hybridoma cultures were assayed by

solid-phase RIA for synthesis of Ig demonstrating differen-tial reactivity with the Hu-rasT24 peptide vs. the Hu-rasHapeptide, and no reactivity with thyroglobulin. Five doublycloned hybridoma cultures, representing the widest spec-trum of differential reactivities to the Hu-rasT24 peptide vs.the Hu-rasHa peptide, were chosen and designated RAP-1

Abbreviations: MAb, monoclonal antibody; IDC, infiltrating ductalcarcinoma.*To whom reprint requests should be addressed.

5227

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 4,

202

1

5228 Medical Sciences: Horan Hand et al. Proc. Natl. Acad. Sci. USA 81 (1984)

through RAP-5 (RA, ras; P, peptide). Each was determinedto be the IgG2a isotype. Fig. 1 shows the differential bindingof Purified Ig of MAb RAP-5 to the Hu-rasT24 vs. the Hu-ras a peptide. As little as 250 pg of purified RAP-5 Ig candetect 10 ng of the Hu-rasT24 peptide, while no reactivity hasbeen observed to as much as 120 ng of thyroglobulin with 32times more RAP-5 MAb. Similarly, MAb UPC-10 showed nobinding to either peptide (Fig. 1). In solid-phase RIAs em-ploying a constant MAb amount (1:100 dilution of cell cul-ture supernatant) vs. dilutions of peptides, MAb RAP-5could detect as little as 400 pg of Hu-rasT24 peptide vs. 2000pg of Hu-rasHa peptide. Since these RIAs utilize a solid-phase substrate, one must view with caution the differentialreactivity observed to equal quantities of Hu-rasT24 vs. Hu-rasHa (see Discussion).

Initial attempts to immunoprecipitate native p21 from ex-tracts of transformed cells proved unsuccessful. This wasnot unexpected, because the determinants detected by theRAP MAbs have been predicted to be internally located inthe native p21 molecule (22, 23). Immunoblotting experi-ments in which extracts were treated with NaDodSO4/mer-

4 ,_'Gdo (l I

FIG. 1. Reactivity of MAb RAP-5 with Hu-rasT24 and Hu-rasHa

peptides. By using a solid-phase RIA (13), increasing amounts of

purified IgG of MAb RAP-5 were assayed for binding to10 ng of Hu-

rasT24 (e), 10 ng of Hu-rasHa(a), and 120 ng of thyroglobulin (o).Purified IgG of MAb UPC-10 was also assayed for binding to ras

peptides (o). (Inset) Detection of p21 by MAb by immunoblotting of

tumor cell extracts. Lane a, molecular weight (x 10-') markers.

Lane b, p21 detected in a dimethylbenzanthracene-induced rat

mammary tumor (8) by MAb RAP-5. One hundred micrograms of

tumor extract was subjected to NaDodSO4/PAGE and transferred

to nitrocellulose as described (20). The nitrocellulose filter was incu-

bated for 2 hr with 200tug of purified RAP-5 and washed extensive-

ly, and bound MAb was detected by using 251I-labeled goat anti-

mouse IgG followed by autoradiography. Lane c, binding of MAb

RAP-2 to the same p21 molecule immunoprecipitated by MAb YA6

259 from T24 cells. ras p21 was immunoprecipitated by MAb YA6

259 from a nonradioactive extract of T24 cells as described (21). The

immunoprecipitate was subjected to NaDodSO4/PAGE and trans-

ferred to nitrocellulose. The nitrocellulose filter was incubated for 2

hr with undiluted MAb RAP-2 tissue culture supernatant and

washed extensively, and bound RAP-2 MAb was detected by using

125"-labeled goat anti-mouse IgG followed by autoradiography.

captoethanol were thus used to demonstrate binding to p21(Fig. 1 Inset).RIAs for ras p21. Several RAP MAbs showed weak bind-

ing in solid-phase RIAs (13) to undenatured cell extracts (13)of T24 human bladder carcinoma cells and T24 DNA trans-fected NIH 3T3 cells (3); no binding to NIH 3T3 cells wasdetected. As discussed above, one would not predict effi-cient binding of these MAbs to native ras p21; therefore, ex-tracts were treated with 10% buffered Formalin in an at-tempt to alter the configuration of p21 and thus expose thedeterminants detected by the RAP MAbs. As shown in Fig.2, Formalin treatment of extracts of either the T24 line or theMCF-7 human mammary carcinoma cell line appreciably en-hanced detection of ras p21 by MAb RAP-1. It is interestingto note that Formalin had no effect on the binding of eitherMAb UPC-10 to T24 cell extracts (Fig. 2A Inset) or MAbB1.1 to extracts of MCF-7 cells (Fig. 2B Inset). Formalin dideliminate, however, the binding of MAb YA6 259 to T24cells. Preliminary studies using extracts of biopsy material orcell lines indicate that fixation of extracts with 2% glutar-aldehyde is even more efficient than Formalin in detection ofras p21 by the RAP MAbs.

Analysis of Human Colon Carcinomas. Formalin-fixed 5-,m tissue sections of human colon carcinomas and normalcolon were assayed for Hu-ras gene expression by using theavidin-biotin complex immunoperoxidase method. Equalconcentrations of MAb RAP-1 detected cytoplasmic ras p21expression in 97% of 32 colon carcinomas (Fig. 3A) vs. noneof5 normal colon samples (Fig. 3B). It is interesting to notethat the rapidly dividing normal colonic epithelium (deep incrypts) scored routinely negative for ras p21. Sixteen of the32 colon carcinomas demonstrated ras expression in greaterthan 20% of carcinoma cells, with several showing rasexpression in over 90% of carcinoma cells. An 8- to 32-folddilution of MAb RAP-1 supernatant fluid still resulted in apositive reaction for most colon carcinomas. Several his-tologic types of hyperplastic and inflammatory colon lesionswere also examined. Some of these lesions contained a smallpercent of cells expressing ras. It will thus be important toconduct detailed studies concerning ras gene expression innumerous histologic types of benign, "premalignant," andmalignant colon lesions.

Analysis of Human Mammary Carcinomas. Thirty Formal-in-fixed infiltrating ductal carcinomas (IDCs) from 30 pa-tients were examined for ras gene activation with each ofseveral RAP MAbs (RAP-1, -2, and -5), all with similar re-sults. Twenty-seven of 30 (90%) of theIDCs scored positive(see Fig. 3 C and D). RAP MAb dilution experiments demon-strate that most "positive" mammary carcinomas containcells that score positive for cytoplasmic ras expression atMAb endpoint dilutions 32-fold to 320-fold higher than cellsin the vast majority of benign lesions (Table 1). As seen inFig. 4, however, a natural separation of this group of carci-nomas is seen, with those tumors scoringsS% of carcinomacells positive (11/30) vs. those with .20% of tumor carcino-ma cells scoring positive (19/30). By this criterion-i.e.,.20% of tumor cells scoring positive-only 2 of the 21 be-nign breast lesions (0/11 fibrocystic disease and 2/10 fibro-adenomas) scored positive (Fig. 4). It is interesting to notethat the two fibroadenomas positive for ras gene expression(Fig. 4B, denoted a and b) were from two patients with multi-ple fibroadenomas (see Discussion). The 1 of 11 fibrocysticdisease lesions examined (Fig. 4B, patient c) with 10% ofcells scoring positive for ras expression was from a patientwith a clinical history of severe chronic mastitis and a histo-logical diagnosis of fibrocystic disease and focal chronic in-flammation (see Discussion). MAb dilution experimentsdemonstrated that the cells positive for ras expression in thelesions of patients a, b, and c (Fig. 4B) reacted as strongly asthose found in most carcinomas. However, the vast majority

Dow

nloa

ded

by g

uest

on

Nov

embe

r 4,

202

1

Proc. Natl. Acad. Sci. USA 81 (1984) 5229

0.

U

0

0

a

-6

0.0c0

E

0-4

x

SU

0-6c-0

0S

1/antibody dilution

FIG. 2. Reactivity of MAb RAP-i to cell extracts. In a solid-phase RIA, MAb RAP-i was used to detect ras p21. (A) Undenatured T24 cellextract (o), and Formalin-fixed (10% buffered Formalin for 1 hr at 25°C) T24 cell extract (e). (B) Undenatured MCF-7 cell extract (o), Formalin-fixed MCF-7 cell extract (n). (A Inset) MAb UPC-10 reactivity to unfixed (A) and Formalin-fixed (A) T24 cell extract. (B Inset) MAb B1.1reactivity to unfixed (v) and Formalin-fixed (v) MCF-7 cell extract.

of all abnormal ducts and lobules from fibroadenoma pa-tients (Fig. 3E) and fibrocystic disease patients (Fig. 3F)were negative. Normal lobules and ducts from breasts of fi-broadenoma patients and fibrocystic disease were also rou-tinely negative for ras gene expression.

In most primary mammary carcinomas, a heterogeneity inthe number of tumor cells expressing ras was observed. Sev-eral mammary tumor metastases from five patients with ras-positive primary mammary tumors were allowed to reactwith MAb RAP-i and also showed some degree of heteroge-neity of ras gene expression within individual metastaticmasses. The possible reasons for this will be discussed be-low. However, all five regional lymph node metastases as-sayed, as well as all four distal metastases (skin, chest wall,rib, adrenal), demonstrated ras gene expression.

DISCUSSIONThe percent of human colon carcinomas in which activatedras was detected with the RAP MAbs was much higher thanthat reported with DNA transfection experiments. The find-ing of activated ras gene expression in a majority of human

Table 1. Titration of MAb RAP-5 with malignant and benignbreast lesions

ras-positive tumor cells, %

MAb RAP-5, Malignant BenignAg/200 dul IDC-1 IDC-2 IDC-3 FA1 FA2 FD1

16 85 95 70 <1 <1 <58 100 45 90 0 0 04 70 65 90 0 0 02 90 1 70 0 0 01 30 1 70 0 0 00.5 30 1 60 0 0 00.1 0 0 5 0 0 00.05 0 0 <1 0 0 0

Sequential tissue sections were assayed by the immunoperoxidasemethod. FA, fibroadenoma; FD, fibrocystic disease (percent posi-tive cells based on the percent of epithelial cells scoring positive).

mammary carcinomas was even more unexpected in light ofthe fact that DNA transfection experiments looking for acti-vated ras expression in human breast cancers have been thusfar unsuccessful. Two possible explanations for these obser-vations could be that (i) the NIH 3T3 foci indicator system isnot efficient for this particular ras gene, or (ii) the mecha-nism of ras activation in many colon and mammary carcino-mas involves the enhanced expression of a normal c-rasgene, perhaps at times, via a hormonally controlled promot-er sequence; DNA transfection experiments would not nec-essarily score positive for this type of mechanism. It is inter-esting to note that the only two benign breast lesions thatscored with percentages of tumor cells positive similar tothose of the mammary carcinomas were from patients (Fig.4B, patients a and b) with multiple fibroadenomas (a total of7 tumors removed from the two patients prior to 24 years ofage). Hormonal factors have been implicated in the develop-ment of these lesions and for severe chronic mastitis (Fig.4B, patient c) (24).The studies reported here demonstrate that MAbs pre-

pared against a set of predefined amino acid sequences, be-lieved to be internally located in a molecule and thus notefficiently exposed for antibody binding in the native confor-mation, can be rendered available by simply altering the ter-tiary structure of the molecule with a substance such as For-malin (Fig. 2). Fortunately, Formalin-fixed tissue sectionshave several advantages over frozen sections: finer histologicdetail, ease of processing, and availability. Conversely, wehave been unable to detect ras p21 with the YA6 259 anti-body (prepared against native p21) in Formalin-fixed tissues.The RAP MAbs do appear to react differentially to the

Hu-rasT24 vs. the normal Hu-rasHa peptide. The solid-phaseRIAs and solid-phase immunohistochemical assays reportedhere, however, are well suited for quantitative assessments.Qualitative evaluations will require liquid competition RIAs,using the RAP MAbs vs. the Hu-ras synthetic peptides,which can record differences in slopes with different com-petitors; these assays may thus detect the one amino acidchange out of the six to eight needed for effective antigen-antibody union.

Medical Sciences: Horan Hand et aL

Dow

nloa

ded

by g

uest

on

Nov

embe

r 4,

202

1

5230 Medical Sciences: Horan Hand etalP

.~~~~~~~

I~~~~~~~~~~~~f w

-~~~~~~~~~~~~k so ,

F I

/. V

1' w~W,PtC^;lj rf;ej

7/i: It.,'I C'.' S

*~~~~~~~~~~~~~~;I. t

4. 4v.s.44l S

t ~ ~4 at

~~~~~~4 r t

- * 4.~

-V0

4t,:IS

...,A(-.-Amw9,FX.1-. X-.,i_.

4

4%

.4D E

*4! :: J

1Pi:..

% p Q*\1.. 1.

c-At, -'i: *,$ -**

$I

*1= *

2~~~~~~~~~~~~.i £ S S * ,~~~~~~

4e a-rtc. I-t~~~~~~~~~~~~~~~~~~~~~~~~t

,§ _

i* . O.-

FIG. 3. Immunoperoxidase staining of paraffin-embedded Formalin-fixed tissue sections with MAb RAP-1. (A) Adenocarcinoma of thecolon, invasive component. The reddish-brown reflects the reaction of the diaminobenzidine substrate and thus MAb binding to cytoplasmicras p21 in carcinoma cells. The blue is the hematoxylin counterstain-i.e., no MAb reactivity. (x11O.) (B) Normal colonic mucosa. Normalcolonic epithelium (arrow), stroma, and inflammatory cells are all negative. (x 17.) (C) Human mammary infiltrating ductal carcinoma (IDC),positive for cytoplasmic reactivity with MAb RAP-1. (x 110.) (D) Human mammary IDC with tubular (glandular) features (arrow) positive forreactivity with MAb RAP-1 from a different patient than C. (x 110.) (E) Fibroadenoma, ductal epithelium (arrow) negative for reactivity withMAb RAP-1. (x70.) (F) Fibrocystic disease with epithelial hyperplasia (arrow) negative for reactivity with MAb RAP-1. (x70.)

The detection of ras p21 in solid-phase RIAs by the RAPMAbs provides a quantitative RIA for an onc gene product;we know of no other such assay. The purified Ig dilutionexperiments shown in Table 1 also provide a semiquantita-tive evaluation of the relative amounts of p21 in individual

cells in tissue sections, since we have determined that bind-ing to purified peptides is linear within the range of Ig con-centrations employed (data not shown).The RAP MAbs generated have not been shown to dis-

criminate among the Hu-ras gene products (i.e., Ha-, Ki-,

Proc. NatL Acad Sd USA 81 (1984)

m I ,

Dow

nloa

ded

by g

uest

on

Nov

embe

r 4,

202

1

Proc. NatL. Acad. Sci. USA 81 (1984) 5231

and N-ras). Thus, the immunoassays employed here shouldbe termed "group specific" in that they recognize transla-tional products from members of a group of related geneswithin a given species. Clearly, "type-specific" immunoas-says, employing antibodies directed against peptides reflect-ing variable regions of the ras gene, will eventually be uti-lized to determine which of the Hu-ras genes is being ex-pressed in a given cell.The studies reported here demonstrate heterogeneity of

ras gene expression among cells in a given colon or mamma-ry carcinoma. Thus, if ras activation has anything to do withtransformation of human colon or mammary epithelium,continuous expression in all tumor cells is not necessary formaintenance of the transformed phenotype. Several possibleexplanations for this phenomenon are as follows: (0) a cellcycle dependency for ras gene expression, as is seen for oth-er tumor antigens (25, 26); (ii) the milieu of the individual cell(i.e., its exposure to endocrine, paracrine, or autocrine fac-tors); or (iii) ras activation was required at the time of trans-formation, but subsequently, during the evolution of the tu-mor, other mechanisms of immortality have come into play.This may particularly be true in the case of colon or mamma-ry carcinomas, in which the interval between the transforma-tion event and tumor detection is believed to be from 5 togreater than 10 years. The quantitative ras p21 RIAs devel-oped, as well as the use of Formalin-fixed tissue specimensfor identification of particular cell types expressing ras p21,now make possible the direct evaluation (both retrospectiveand prospective) of clinical material in an attempt to definethe precise role, if any, of ras gene translational products in

Q

0.4-

._

1o

Malignant Benign

FIG. 4. MAb RAP-i (at a 1:8 dilution of supernatant fluid) wasallowed to react with mammary carcinomas (e) and benign lesions:*, fibroadenoma; A, fibrocystic disease. Each symbol represents aparaffin-embedded Formalin-fixed tumor or lesion from a differentpatient. The percent cells positive denotes, for carcinomas and fi-broadenomas (A), the number of tumor cells scoring ras-positive di-vided by the total number of tumor cells x 100; for fibrocystic dis-ease lesions (B), the number of epithelial cells x 100. Tumors a andb are from patients with multiple fibroadenomas; lesion c is from apatient with severe chronic mastitis.

the processes of carcinoma initiation, promotion, and pro-gression.

We thank P. Freeman, E. Smith, D. Poole, M. Taylor, A. Sloan,and K. McNaughton for excellent technical assistance and Drs. M.Weeks, K. Robbins, and S. Aaronson for helpful discussions. A.C.was supported, in part, by the Instituto di Microbiologia, Univer-sita' di Brescia, Consiglio Nazionale delle Ricerche, Progetto, Fina-lizzato Controllo Malattie de Infezione.

1. Ellis, R. W., DeFeo, D., Shih, T. Y., Gonda, M. A., Young,H. A., Tsuchida, N., Lowy, D. R. & Scolnick, E. M (1981)Nature (London) 292, 506-511.

2. Shimizu, K., Goldfarb, M., Suard, Y., Perucho, M., Li, Y.,Kamala, T., Feramisco, J., Stavnezer, E., Fogh, J. & Wigler,M. H. (1983) Proc. Natl. Acad. Sci. USA 80, 2112-2116.

3. Reddy, E. P., Reynolds, R. K., Santos, E. & Barbacid, M.(1982) Nature (London) 300, 149-152.

4. Tabin, C. J., Bradley, S. M., Bargmann, C. I., Weinberg,R. A., Papageorge, A. G., Scolnick, E. M., Dhar, R., Lowy,D. R. & Chang, E. H. (1982) Nature (London) 300, 143-149.

5. Schwab, M., Alitalo, K., Varmus, H. E. & Bishop, J. M.(1983) Nature (London) 303, 497-501.

6. Pulciani, S., Santos, E., Lauver, A. V., Long, L. K., Aaron-son, S. A. & Barbacid, M. (1982) Nature (London) 300, 539-542.

7. Huang, A. L., Ostrowski, M. C., Berard, D. & Hager, G. L.(1981) Cell 27, 245-255.

8. Sukumar, S., Notario, V., Martin-Zanca, D. & Barbacid, M.(1983) Nature (London) 306, 658-661.

9. Cho-Chung, Y. S. & Huang, F. L. (1984) in Proceedings ofthe16th Miami Winter Symposium, eds. Ahmad, F., Black, S.,Shultz, J., Scott, W. A. & Whelan, W. J. (ISCU Press, Miami,FL), pp. 142-143.

10. Lerner, R. A. (1982) Nature (London) 299, 592-596.11. Kagan, A. & Glick, M. (1979) in Methods ofHormone Radio-

immunoassay, eds. Jaffe, B. M. & Behrman, H. R. (Academ-ic, New York), pp. 328-329.

12. Herzenberg, L. A., Herzenberg, L. A. & Milstein, C. (1979) inHandbook of Experimental Immunology, ed. Weir, D. M.(Blackwell Scientific, London), pp. 25.1-25.7.

13. Colcher, D., Horan Hand, P., Teramoto, Y. A., Wunderlich,D. & Schlom, J. (1981) Cancer Res. 41, 1451-1459.

14. Colcher, D., Horan Hand, P., Nuti, M. & Schlom, J. (1983)Cancer Invest. 1, 127-138.

15. Potter, M. (1972) Physiol. Rev. 52, 631-719.16. Furth, M. E., Davis, L. J., Fleurdelys, B. & Scolnick, E. M.

(1982) J. Virol. 43, 294-304.17. Colcher, D., Zalutsky, M., Kaplan, W., Kufe, D., Austin, F. &

Schlom, J. (1983) Cancer Res. 43, 736-742.18. Nuti, M., Teramoto, Y. A., Mariani-Constantini, R., Horan

Hand, P., Colcher, D. & Schlom, J. (1982) Int. J. Cancer 29,539-545.

19. Hsu, S. M., Raine, L. & Fanger, H. (1981) J. Histochem. Cy-tochem. 29, 577-580.

20. Towbin, H., Staehlin, T. & Gordon, J. (1979) Proc. Natl.Acad. Sci. USA 76, 4350-4354.

21. Shih, T. Y., Weeks, M. O., Gruss, P., Dhar, R., Oroszlan, S.& Scolnick, E. M. (1982) J. Virol. 42, 253-261.

22. Wierenga, R. K. & Hol, W. G. J. (1983) Nature (London) 302,842-844.

23. Pincus, M. R., van Renswoude, J., Harford, J. B., Chang,E. H., Carty, R. P. & Klausner, R. D. (1983) Proc. Natl.Acad. Sci. USA 80, 5253-5257.

24. Robbins, S. L. & Cotran, S. R. (1979) Pathologic Basis ofDis-ease (Saunders, Philadelphia), 2nd Ed., pp. 1317-1318.

25. Kufe, D. W., Nadler, L., Sargent, L., Shapiro, H., HoranHand, P., Austin, F., Colcher, D. & Schlom, J. (1983) CancerRes. 43, 851-857.

26. Campisi, J., Gray, H. E., Pardee, A. B., Dean, M. & Sone-shein, G. E. (1984) Cell 36, 241-247.

Medical Sciences: Horan Hand et aL

Dow

nloa

ded

by g

uest

on

Nov

embe

r 4,

202

1