Immunohistochemical Evidence for RNA Virus Related Components in Human Breast Cancer*

Transcription

1 ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 9, No. 3 Copyright 1979, Institute for Clinical Science, Inc. Immunohistochemical Evidence for RNA Virus Related Components in Human Breast Cancer* R. MESA-TEJADA, M.D.,t I. KEYDAR,t M. RAMANARAYANAN,f T. O H N O,t C. FEN O G LIO Í AND S. SPIEG ELM ANf f Institute o f Cancer Research and \Department o f Pathology, College o f Physicians and Surgeons o f Columbia University New York, NY ABSTRACT The localization of antigenic components with cross-reactivity to a 52,000 dalton group specific glycoprotein (gp52) of the mouse mammary tumor virus (MMTV) in paraffin sections of human breast carcinomas is described using an indirect im m unoperoxidase m ethod. This m ethod was first optim ized on paraffin sections of mouse mammary tumors. The specificity of the reaction observed in the hum an tissues was established by absorption of the specific IgG with: a) purified gp52; b) several relevant and irrelevant viral preparations; c) normal hum an plasma, leukocytes, breast tissue, milk, actin, collagen, and hyaluronic acid; d) sheep erythrocytes, bovine mucin and fetal calf serum. Only MMTV and purified gp52 elim i nated the immunohistochemical reaction in hum an breast tumors. Positive reactions were seen in 171 of376 (45.5 percent) randomly selected breast carcinomas of various histopathologic types, while negative reactions w ere obtained in all 137 normal and benign cases tested. In those invasive tumors with an intraductal component, a higher percentage (63.9 percent) of positive cases was seen. A positive reaction of different specificity was observed in foci of apocrine metaplasia. With one exception, 99 carcinomas from 13 organs other than breast and eight cystosarcomas w ere negative. Introduction T he su b seq u en t identification of the Bittner factor as a B-type ribonucleic acid (RNA) virus14 stim ulated a search leading to the ultrastructural finding of particles resem bling m urine B-type and C-type viruses in human breast cancer tissues28 as well as in human milk.27 The eventual isolation and biochem ical characterization of these particles18 revealed further sim ilarities to the oncornaviruses, principally the presence of RN A -dependent d eso x y rib o n u cleic acid (DNA) p o ly m erase (reverse transcriptase) com plexed /79/ $01.50 Institute for Clinical Science, Inc. During the four decades since Bittner4 d e scrib ed a factor in certain high cancer-incidence strains of mice capable of initiating mammary tumors, a considerable experim ental effort has b een directed towards establishing useful correlations betw een the m ouse mammary tum or m odel and hum an breast cancer. * Please send reprint requests to:r. Mesa-Tejada, M.D., Institute of Cancer Research, 701 West 168th Street, New York, NY

2 VIRUS RELATED COMPONENTS IN BREAST CANCER 203 to a 70S RNA m olecule. In addition, m olecular hybridization studies dem onstrated a partial homology betw een RNA molecules found in human breast tumors an d the RNA genom e of th e m ouse mammary tum or virus (MMTV).2, In recent years, much experimentation has been focused on exploring a possible antigenic relationship betw een MMTV p ro tein s and com p o n en ts in hum an breast cancer. Studies w ith sera of breast cancer patients have dem onstrated the presence in such sera of antibodies capable o f neutralizing MMTV7 and of localizing antigens in mouse mammary tumors by im m unofluorescence,19,20 and ultrastru ctu rally, by im m u n o fe rritin 9 and im m unoperoxidase6,10 techniques. The migration inhibition studies of Black et al5 using leukocytes from breast cancer patients suggest a crossreactivity b e tween the major glycoprotein of MMTV and a protein com ponent extracted from human breast cancer tissues. In addition, the immunofluorescent identification of MMTV crossreactive antigens in the MCF-7 human breast carcinoma cell line has b een reported.35 This apparent im m unologic crossreactivity of MMTV proteins and antigens in human breast cancer is of considerable interest in view of the fact that in the mouse mammary tumor model a 52,000 dalton viral glycoprotein (gp52) is an excellent indicator of disease status.24 25'26 Thus, earlier studies in this laboratory have established that plasm a levels of gp52, m easured by radioimmunoassay, could be accurately correlated with the existence,24 size,25 and recurrence after surgical excision26 of m ouse mammary tumors, frequently in the absence of gross physical signs of the disease. Attempts to use radioim m unoassays to d etect crossreacting proteins in the plasma of breast cancer patients,8 36 how ever, have unfortunately not b een very successful, probably ow ing to the thousand-fold blood volum e differences b etw een m ice and humans which would conceivably dilute the signal in human plasma beyond the se n sitiv ity of p re s e n t rad io im m u n o assays. O ur efforts, therefore, have b e e n concentrated on the im m unohistologic localization of crossreactive antigens in the tumors them selves using the very sensitive and convenient immunoperoxidase m ethod of antigen localization.17 After optim izin g our im m u n o h isto ch em ical m ethod in the localization of MMTV antigens in frozen and paraffin sections of mouse mammary tumors,13 we were able to apply successfully the same m ethodology and reagents on routinely processed, paraffin sections o f hum an b reast carcinomas.15,16 The present paper updates our experience and observations w ith respect to the localization in human breast cancer tissues of an antigen which crossreacts w ith gp52, the group-specific glycoprotein of the m ouse m am m ary tum or virus. Materials and M ethods T i s s u e Paraffin blocks of tissues used for diagnostic purposes were selected from the files of the divisions of Surgical, G ynecologic and Anatomic Pathology at the College of Physicians and Surgeons of Colum bia University, as previously d e scrib ed.15 T hese tissues had initially been fixed in Bouin s solution or, in the case of larger specim ens, in 10 percent buffered formalin followed by Bouin s fixation. Additional tumor samples were received fresh from M em orial Sloan- K etterin g C ancer C en ter. O ne sm all block from each of these was fixed in Bouin s and routinely processed. From all of the aforem entioned, 5 /xm serial sections w ere cut for im m u n o h isto chem ical staining; occasional sections were stained with hematoxylin and eosin for comparison.

3 204 MESA-TEJADA, KEYDAR, RAMANARAYANAN, OHNO, FENOGLIO AND SPIEGELMAN T A B L E I Immunoperoxidase Staining of Carcinoma of the Breast Total Cases Classified as to Source Source Cases Positive Percent Positive Columbia-Presbyterian Michigan Cancer Foundation Memorial Sloan-Kettering O ne h u n d re d and four a d d itio n al breast cancer cases were received as unstained paraffin sections from Dr. Marvin Rich, of the M ichigan C ancer Foundation. An t i s e r a a n d IgG P r e p a r a t i o n s The two principal antisera used in the present study were raised in rabbits by using whole (disrupted) MMTV isolated from the milk of Paris RIII mice, and contained antibodies to the MMTV proteins in proportional amounts as reflected by th e ir ra d io im m u n o p re c ip ita tio n p ro files.16 O ther antisera used for com parison included rabbit anti-mmtv (C3H), rabbit anti-gp52 and goat anti-gp52. The IgG fractions from these sera w ere prep a red and c h a ra c te riz e d by rad io im m unoprecipitation as previously d e scribed.16 T A BLE II Immunoperoxidase Staining of Carcinoma of the Breast Histopathologic Classification of Columbia-Presbyterian Cases Percent Type Cases Posi tive Positive Intraductal Intraductal and invasive ,9 Invasive* Medullary Metastaticf Includes all types of invasive carcinoma (e.g., tubular, lobular, small cell, etc.) except those with predominantly medullary features or associated with intraductal lesions, tlncludes metastases to lymph nodes, lungs, liver, adrenal gland and ovaries in proven cases of primary carcinoma of the breast. A b s o r p t i o n o f IgG P r e p a r a t i o n s N um erous substances w ere used to absorb the antibody preparations to en hance or test the specificity of the imm unohistochem ical staining. T he p rep aration of these immunoabsorbants and the conditions of absorption are detailed in a previous report.16 I m m u n o h i s t o c h e m i c a l S t a i n i n g An indirect immunoperoxidase m ethod was used w ith conjugates prepared by the periodate procedure of Nakane and Kawaoi21 w ith some m odifications.16 A detailed description of this procedure has previously been reported.16 Results A positive staining reaction has been observed in 171 (45.5 percent) of 376 cases of hum an breast carcinomas tested, including cases received from other institutions (table I). In table II are sum m arized positive staining reactions seen in 238 C o lu m b ia-p resb y terian b reast can cer cases c la ssified acco rd in g to histopathologic type. The larger percentage of the positive cases in the intraductal and invasive group reflects a trend noted in our original report of 131 cases15 and suggests that the correlation seen in this larger series is indeed significant. It appears, therefore, that an invasive carcinom a associated w ith an intraductal com ponent is m ore likely to contain crossreactive antigen than either a pure intraductal or invasive tumor. The pattern of immunohistochemicai staining in hum an breast tumors tends to be focal, intracellular and cytoplasmic w ith considerable variability even within the same tumor (figures 1 and 2), a situation not altogether unlike that encountered in mouse mammary tumors.13 For example, in the intraductal and invasive carcinom a illustrated in figure 1, only some of the cells in the intraductal lesion

4 VIRUS RELATED COM PONENTS IN BREAST CANCER 205 IB F i g u r e 1. Immunoperoxidase stain o f intraductal and invasive carcinoma with anti-mmtv before (la) and after (lb) absorption with gp52. Some of the cells in the intraductal lesion and most of the invasive cells contain reaction product. (These photomicrographs have been overdeveloped to enhance morpho logic detail; the counterstain may therefore in some areas simulate positive staining.) (M ethylene blue counterstain; x 778).

5 206 M E SA -TEJA D A, K EYDAR, RAM ANARAYANAN, O H N O, F E N O G L IO AND SPIE G EL M A N 2A 2B F i g u r e 2. Immunoperoxidase stain of invasive breast carcinoma with anti-mmtv before (2a) and after (2b) absorption with gp52. Note intense stain of focus o f invasive carcinoma and lack of reaction in neighboring morphologically benign ducts and acini (2a) as w ell as complete elimination of the former reaction after absorption with gp52 (2b). (Methylene blue counterstain; x 609).

6 VIRUS RELATED COMPONENTS IN BREAST CANCER 207 contain reaction product. Likewise, not all of the invasive m alignant cells surro u n d in g the in tra d u cta l lesions are stained. In other sections of tumor from this same case, however, staining was not universal and was even com pletely absent in some intraductal lesions which w ere surrounded by strongly reacting invasive cells. In figure 2, on the other hand, is illustrated an invasive tumor in which practically every m alignant cell gave a strong staining reaction in all the secretions tested. The microscopic field illustrated depicts a focus of intensely stained invasive carcinoma surrounding morphologically benign ducts and nearby uninvolved acini which do not contain reaction product. In figures lb and 2b are shown serial sections adjacent to those illustrated in figures la and 2a, w hich have been treated identically except that the primary antiserum has b e e n previously absorbed w ith purified gp52. Thus, absorption with purified gp52 is sufficient to elim inate the staining reaction in human breast carcinoma, while absorptions with num erous other related and unrelated substances (table III) do not produce any effect. T he sp ecificity of the reaction for breast carcinoma was determ ined by testing 99 carcinomas from other organs and eight cases of cystosarcom a phyllodes (table IV). Only one of these 107 tumors, a m u co ep id erm o id carcinom a o f the parotid gland, gave a positive reaction; two other parotid carcinomas of the same histopathologic type w ere negative. Norm al (resting and lactating) and benign (cystic disease, fibroadenom a, in trad u ctal papillom a, gynecom astia) breast tissues from 137 patients w ere also tested with negative results (table V). The only exception to the absence of staining reaction in nonmalignant breast tissue was the staining of foci of apocrine m etaplasia, one of the m icroscopic features of cystic disease. The different specificity of this reaction follows. Discussion T he su ccessfu l a p p licatio n o f the im m unopero x id ase tech n iq u e to the localization of MMTV antigen in sections of paraffin-em bedded mouse mammary tumor tissues13 made possible the extensive, prim arily retrospective studies d e scribed in this paper. While elim inating the need for fresh tissues in our search for a crossreacting human antigen, we also had at our disposal the same tissues received by the Pathology Laboratory, the paraffin blocks of which could be used after the diagnostic sections had been cut. TABLE I I I A b sorp tion S p e c if i c it y T ests o f Immunoperoxidase S ta in in g o f Human Breast Carcinomas w ith a-mouse Mammary Tumor Virus* C o m p le t e ly E l i m i n a t e d b y N o t E l i m i n a t e d b y MMTV (RIII) from milk MMTV (C3H) from MM5T cell line MMTV (C3H) from CrFeK cell line gp52 (RIII)') purified by concanavalin A gp52 (C3H ) j affinity chromatography gp52 (Rlllf) purified by guanidium gp52 (C3H)_j" chloride chromatography Viruses Rauscher leukemia virus. Simian sarcoma virus, Mason-Pfizer monkey virus. Baboon endogenous virus Human Normal plasma, normal leukocytes, collagen, actin, hyaluronic acid, milk, normal breast tissue Bovine Mucin, fetal calf serum Sheep erythrocytes *The absorptions were performed by using these agents either in the soluble or insoluble form*6.

7 2 0 8 MESA-TEJADA, KEYDAR, RAMANARAYANAN, OHNO, FENOGLIO AND SPIEGELMAN TABLE IV Immunoperoxidase Staining of Malignancies Other Than Breast Carcinomas M a lig n a n cy* C a se s P o s i t i v e Colon 12 Stomach 3 Pancreas 3 Liver 4 Lung 9 Endometrium 22 Ovary 20 Prostate 8 Kidney 5 Urinary bladder 4 Skin 2 Thyroid gland 4 0 Parotid gland Cystosarcoma phyllodes Primary sites of non-breast carcinomas stained with a-mouse mammary tumor virus; cystosarcoma phyllodes is also listed because it is the most common noncarcinomatous malignancy of the breast. T he staining reactions observed in human breast carcinomas w ith antibodies to MMTV differ from the reactions previously observed in the mouse mammary tumors w ith respect to variability and frequency of staining as well as with the results of absorption with gp52. These differences shall be considered in turn. The variability of staining seen among and w ithin the hum an breast carcinomas is greater than that observed in the mouse tum ors,15,16 w hich is not unexpected considering that the etiologic role of MMTV is w ell d o cum ented in the latter. Fur- TABLE V Immunoperoxidase S ta in in g o f Benign and Normal Breast T issues A s s o c i a t e d w i t h B r e a s t P o s i - T y p e C a s e s C a r c in o m a * t i v e Cystic diseaset Fibroadenoma Intraductal papilloma Gynecomastia Resting gland (normal) Lactating gland (normal) *In the same breast. texcluding foci of apocrine metaplasia; the immunologic specificity of this reaction is discussed in the text. 0 1 therm ore, tumors from highly infected mouse strains25 were deliberately chosen in these earlier studies in order to optim ize our m ethod of MMTV antigen localization. Nevertheless, it should be em phasized that cellular variability of antigen expression was not uncommon in the mouse mammary tumors and has also b e e n fre q u e n tly o b se rv e d in hum an tum ors w ith resp e c t to other tum orassociated antigens.23 With respect to frequency, it is evident that there is considerable sampling error in our testing procedure and, therefore, the percentages given in tables I and II represent, at best, minimal values. This conclusion is based on the follow ing facts: (1) diagnostic tissue blocks of a given case are usually representative, but seldom include the entire tum or; (2) these studies were lim ited to an average of less than three (one, in cases from other institutions) representative blocks per case, and as previously noted, a considerable variability in antigen localization was noted am ong and w ithin sections of positive cases; and (3) in contrast to in vitro methods, where a given tissue can be assayed in bulk, our test is lim ited to 5 /im sections which represent only a m inute fraction (less than 1/1000th) of the entire tumor. F in a lly, in co n trast to the m ouse tum ors, w h ere only ab so rp tio n w ith w hole d isru p te d M MTV co m p letely elim inates the staining reaction,13,16 absorption with purified gp52 alone was sufficient to obliterate the reaction in positively staining human tumors (figures lb and 2b). The specificity of this staining reaction was further explored by absorption with the various preparations of MMTV and gp52 listed in table III. All of these blocked the staining reaction, indicating that the species differences that have been previously reported30,31 for gp52 of the CD8 and RIII MMTV do not play a role in this reaction. On the other hand, absorption w ith different virus

8 VIRUS RELATED COMPONENTS IN BREAST CANCER 209 preparations (Rauseher leukem ia virus, simian sarcoma virus, Mason-Pfizer monkey virus, baboon endogenous virus) and w ith several possible crossreacting substances, also listed in table III, fail to elim inate the staining reaction. T he sp ecificity of the reactio n for breast carcinoma is supported by the fact that only one other carcinoma of different origin has b een found to be positive (table IV). In addition, the data presented in table V confirm the specificity of the reaction for m alignant b reast tissues. The troublesom e reactivity observed in apocrine metaplasia, also shared by the m orphologically and histochem ically ind istin g u ish ab le1,32 epithelium of apocrine glands of the axilla and perineum, has been found to differ in specificity from the reaction observed in the carcinomas. For example, absorption with gp52 blocks only a minimal part of the reactio n in the apocrin e glands and m etaplasia w hile it com pletely elim i nates the reaction in the carcinoma. A more complete discussion of experiments w hich further define the nature and differences o f the reactions seen in the hum an tum or and th e apocrine epithelium through separation o f the sugar and protein moieties of the MMTV glycoprotein is in preparation, and will be reported elsew here. O ur co n clu sio n th a t at le a st som e hum an b reast carcinom as contain an antigen immunologically related to gp52 of MMTV is further supported by two recent studies,22,34 in addition to those previously cited. At the present time, it is clearly prem ature to speculate on the etiologic implications of this finding, just as it would be m isleading to draw too close a p a ra lle l b e tw e e n th e m ouse mammary tumor model and the human disease. O ur efforts, therefore, are d i rected rather toward the possibility that these findings can be used to generate clinically useful information. For example, prelim inary evidence is available indicating that breast cancer patients with a family history of the disease have a greater probability of expressing the gp52 crossreacting antigen in their tum or as compared to patients with no such family history. Finally, our laboratory has a well established human breast carcinoma cell line, T47D 12 which also expresses immunohistochem ically detectable antigen(s) related to gp52.n It is hoped that the availability of this source will simplify the task of obtaining the relevant human breast cancer antigen in adequate amounts and stages of purity. This, in turn, may provide the immunologic reagents needed to develop the heterologous radioim m unoassays that can hope to attain the sensitivities required for m easurem ent of a systemic signal in the hum an disease. Acknowledgm ents We are extremely grateful to Dr. Raffaele Lattes for his helpful advice and cooperation. The technical assistance of Peggy Moy, Elizabeth Osze, Susan LaFlamme and Leonard Yu is gratefully acknowledged. This investigation was supported by Grant CA and Contract N01-CP awarded by the National Cancer Institute. References 1. A h m e d, A.: Apocrine metaplasia in cystic hyperplastic mastopathy. Histochemical and ultrastructural observations. J. Pathol. 115: , Ax e l, R., Gu lati, S. C., and Spieg elm an, S.: P articles conta in in g R N A -instructed D N A polymerase and virus-related RNA in human breast cancers. Proc. Nat. Acad. Sci. 69: , Ax e l, R., Sc h l o m, J., and Spieg elm a n, S.: Presence in human breast cancer of RNA homologous to mouse mammary tumor virus RNA. Nature 235:32-36, Bittner, J. J.: Some possible effects of nursing on the mammary gland tumor incidence of mice. Science 284: , Black, M. M., Za ch rau, R. E., D io n, A. S., Sh o re, B., F in e, D. L., Leis, H. P., Jr., and WILLIAMS, C.: Cellular hypersensitivity to gp55 of RHI-murine mammary tumor virus and gp55-like protein of human breast cancer. Cancer Res. 36: , Bo w en, J. M., D m ochow ski, L., Mil l e r, M. F., Priori, E. S., Sem an, G., D o d so n, M. L., and MARUYAMA, K.: Implications o f humoral

9 210 MESA-TEJADA, KEYDAR, RAMANARAYANAN, OHNO, FENOGLIO AND SPIEGELMAN antibody in mice and humans to breast tumor and mouse mammary tumor virus-associated antigens. Cancer Res. 36: , CHARNEY, J. and MOORE, D. H.: Neutralization of murine mammary tumor virus by sera of women with breast cancer. Nature 229: , H e n d r ic k, J.-C., Fr a n ç o is, C., C a l b e r g - Ba cq, C.-M., C o l in, C., F r anchim o n t, P., Go sse l in, L., Kozm a, S., and O sterr ieth, P. M.: Radioimmunoassay for protein p28 of murine mammary tumor virus in organs and serum of mice and search for related antigens in human sera and breast cancer extracts. Cancer Res. 38: , H o l d e r, W. D., Jr., Pe e r, G. W., Bo l o - GNESI, D. P., and We l l s, S. A., JR.: Antibody reacting with mouse mammary tumor virus in serum of breast carcinoma patients. Surg. Forum 27: , HOSHINO, M. and D m o ch ûw ski, L.: Electron microscope study of antigens in cells of mouse mammary tumor cell lin es by peroxidaselabeled antibodies in sera of mammary tumorbearing mice and of patients with breast cancer. Cancer Res. 33: , Key d a r, I., C h e n, L., Karby, S., D elarea, Y., Ram anarayanan, M., Mesa-Tejada, R., Sp ie g e l m a n, S., H a g e r, J. C., and C a l a - BRESI, P.: D etection o f an antigen in a human breast cancer c e ll lin e (T -47D ) im m unologically related to the m ouse mammary tumor virus. P roceedings of the 11th M eeting on Mammary Cancer in Experimental Animals and Man. Detroit, Key dar, I., Ch e n, L., Ka rby, S., We iss, F. R., D e l a r e a, J., Ra d u, M., C h a it c ik, S., and Br e n n e r, H. J.: Establishment and characterization of a cell 1 ine of human breast carcinoma origin. Int. J. Cancer, in press. 13. Ke y d a r, I., M e s a -T e j a d a, R., Ra m a n a r a y a n a n, M., O h n o, T., Fe n o g l io, C., Hu, R., and Spieg elm a n, S.: Detection of viral proteins in m ouse mammary tumors by immunoperoxidase staining of paraffin sections. Proc. Nat. Acad. Sci. 75: , Ly o n s, M. J. and Moore, D. H.: Purification of the m ouse mammary tumor virus. Nature 194: , M e s a -T e j a d a, R., Ke y d a r, I., Ra m a n arayanan, M., Oh n o, T., Fen o g lio, C., and Spieg elm an, S.: Detection in human breast carcinoma of an antigen immunologically related to a group-specific antigen of mouse mammary tumor virus. Proc. Nat. Acad. Sci. 75: , M e s a -T e ja d a, R., Ke y d a r, I., Ra m a n a r a y anan, M., O h n o, T., F enoglio, C., and Spiegelm an, S.: Immunohistochemical detection of a crossreacting virus antigen in mouse mammary tumors and human breast carcinomas. J. Histochem. Cytochem. 26: , M e s a -T e j a d a, R., Pa s c a l, R. R., and F e n o g l io, C. M.: Immunoperoxidase: A sensitive im m unohistochem ical technique as a special stain in the diagnostic pathology laboratory. Human Pathol. 8: , M i c h a l i d e s, R., S p ie g e l m a n, S., and Sc h lo m, J.: Biochemical characterization of putative subviral particulates from human m alignant breast tumors. Cancer Res. 35: , M ü l l e r, M., Ke m m e r, C., Z o t t e r, S., Gr o ssm ann, H., and Mic h a el, B.: Kreuzreaktion zw ischen Bruskrebs und M astopathie des M enschen sow ie murinen Mammakarzinomen. L okalisation des A n tigens im B ereich der A-Partikel d es M ammatumorvirus. Arch Geschwulsforsch 41: , Mü ller, M., Zo tter, S., Gr o ssm a n n, H., and Kem m er, C.: Im m unologische Kreuzreaktion zw isch en Brustkrebs und M astopathie des M en sc h e n so w ie v iru sp ro d u zieren d en M am m ak arzinom en der M aus. Arch G eschwulsforsch 40: , N akane, P. K. and Ka w aoi, A.: Peroxidaselabeled antibody. A new method of conjugation. J. Histochem. Cytochem. 22: , O s t e r r ie t h, P. M., Ko z m a, S., C o l i n, C., H u s t i n, J., F r a n ç o is, C., C a l b e r g - Bacq, C. M., and Go sse l in, L.: Detection by immunofluorescence staining of 4 antigens of the mouse mammary tumor virus in organs of mice, search for related antigens in breast cancer sections. Proceedings of the 11th Meeting on Mammary Cancer in Experimental Animals and Man. Detroit, Pasc a l, R. R., Mesa-Tejada, R., Be n n e t t, S., G a r c e s, A., and F e n o g l io, C.: Carcinoembryonic antigen. Immunohistologic identification in invasive and intraepithelial carcinomas of the lung. Arch. Pathol. Lab. Med. 101: , Rit z i, E., Ba l d i, A., and Spieg elm a n, S.: The purification of a gs antigen of the murine mammary tumor virus and its quantitation by radioimmunoassay. Virology 75: , Rit z i, E., Martin, D. S., St o l f i, R. L., and Spieg elm an, S.: Plasma levels ofa viral protein as a diagnostic signal for the presence of tumor. The murine mammary tumor model. Proc. Nat. Acad. Sci. 73: , Ritzi, E., Ma rtin, D. S., St o l f i, R. L., and Spieg elm a n, S. : Plasma levels of a viral protein as a diagnostic signal for the presence of mammary tumor. The effect of tumor removal. J. Exp. Med. i 45: , Sarkar, N. H. and Moo re, D. H.: On the possibility of a human breast cancer virus. Nature 233: , Sem an, G.,Gallag h er, H.S., Lu k em a n, J. M., and D m ochow ski, L.: Studies on the presence of particles in human breast tumors, pleural effusions, their tissue cultures, and milk. Cancer 28: , Spieg elm an, S., Ax e l, R., and Sc h lo m, J.: Virus-related RNA in human and mouse mammary tumors. J. Nat. Cancer Inst. 48: , 1972.

10 VIRUS RELATED COMPONENTS IN BREAST CANCER T eram oto, Y. A., Ku f e, D., and Sc h o lm, J.: Multiple antigenic determinants on the major surface glycoprotein of murine mammary tumor viruses. Proc. Nat. Acad. Sci. 74: , Teram oto, Y. A., Ku f e, D., and Sc h lo m, J.: Type-specific antigenic determinants on the major external glycoprotein of high and low oncogenic murine mammary tumor viruses. J. Virol. 24: , T r e m b l a y, G.: H istochem ical studies o f oxidative enzymes in apocrine-like cells of the breast and in axillary apocrine glands. J. Invest. Derm. 50: , Vaid y a, A. B., Black, M. M. D io n, A. S., and MOORE, D. H.: Homology between human breast tumor RNA and mouse mammary tumor virus genome. Nature 249: , Verstraeten, A. A., H agem an, Ph. C., and D e Wit, M. Y. L.: Mouse mammary tumor virus and human material: Immunological crossreactions. Proceedings 11th Meeting on Mammary Cancer in Experimental Animals and Man. Detroit, Ya ng, N.-S., So u l e, H. D., and Me Gr a th, C. M.: Expression of murine mammary tumor virus-related antigens in human breast carcinoma (MCF-7) cells. J. Nat. Cancer Inst. 59: , Z a n g e r l e, P. F., Ca l b e r g -B a c q, C.-M., C olin, C., F ranchim ont, P., F rançois, C., G o sselin, L., Kozm a, S., and O sterrieth, P. M.: Radioimmunoassay for glycoprotein gp47 of murine mammary tumor virus in organs and serum of mice and search for related antigens in human sera. Cancer Res. 3 7 : , 1977.