IMMUNOHISTOCHEMICAL LOCALIZATION OF A CHORIOGONADOTROPIN-LIKE PROTEIN IN ISOLATED FROM PATIENTS HERNANF. ACEVEDO,PHD, MALCOLMSLIFKIN, PHD GAILR. POUCHET,BA, AND MATIASPARDO, BA

By the use of specific antibody to human chorionic gonadotropin (CG) as well as to its P-subunit, and the application of the indirect fluorescein-labeled and peroxidase-labeled antibody techniques, we have demonstrated the presence of a membrane (wall)-associated CG-similar immunoreactive protein in 15 strains of bacteria isolated from tissues of patients bearing malignant neoplasms. These microorganisms were classified as S. epzdermidis, (12), E. coli (2), and a single strain ofP. maltophilia (ATCC 13637). The absence of the CG-like antigen in other ‘‘cancer associated bacteria”, Streptococcus faecalis (ATCC 12818) and Pseudomonas aeruginosa (from patient with cancer of colon), demonstrated that not every ‘‘cancer associated bacteria” has the capability to synthesize the trophob- lastic-like protein. The negative results obtained with a number of “noncancer control” bacteria of known origin, obtained from ATCC and from clinical samples, strongly supported the idea that the existence of these CG-like protein producing microorganisms is not a ubiquitous finding. The demonstration of a de nouo bacterial biosynthesis of a protein having similar antigenic and biophysi- cal properties to those of the human trophoblastic hormone, has great biologi- cal implications, especially if its biosynthesis is proven only in bacterial strains growing in the presence of cancer cells in which we have already demonstrated the presence of a similar antigen. The explanation of the phenomenon is unknown. Because of their origin, the potential of “genetic exchange” with subsequent expression of the mammalian gene by the bacterial cells becomes a possibility. It is also possible that the gene coding for the CG-like protein is normally present but inactive or repressed in all bacteria. Cancer 4 1: 12 17-1 229, 1978.

ECINNING IN 1948, VIRGINIA(WUERTHELE- plasms. In 1970, she and Eleanor Alexander- B Caspe) Livingston and her associates have jackso on published a full description of the orga- published a series of papers describing a pleo- nism(s) and proposed the name Progenitor crypto- morphic, acid-fast bacterium which she ob- cides as a new taxon within the Actinomycetales. served in, and isolated from, cancer tissues and They reported that the P. cryptocides cross-reacts body fluids of patients with malignant neo- antigenically with M. , with which it shares the property of acid fastness. 29 Based on the proven presence of a human From the Division of Experimental Pathology and Section of Microbiology and Immunology, Department of Labora- chorionic gonadotropin (choriogonadotropin, tory Medicine, William H. Singer Memorial Research Insti- CG)-like protein in the serum of some patients tute, Allegheny General Hospital, Pittsburgh, Pennsylvania. with different types of cancer, 13,14,21-23,34,37,40,48 The authors thank to Drs. Virginia Livingston, The Liv- and on her views concerning the nature of can- ingston Clinic, Herman Cohen and Alice Strampp, Prince- cer, primarily in respect to her consideration of ton Laboratories, Inc., and Lewis F. Affronti, George Wash- ington University Medical Center, for making available to us P. cryptocides as a possible etiologic agent of the their strains of “cancer associated bacteria,” and Dr. Ver- disease, Livingston and her associates tested non C. Stevens, The Ohio State University College of Medi- their bacterial isolates for the production of the cine, and Mr. David R. Crockford, Serono Laboratories, trophoblastic-like protein. Their investigations Inc., for supplying us with some of the antiserum. demonstrated that these bacteria, when cultured Address for reprints: H. F. Acevedo, Ph.D., Department in vitro produced a substance immunologically of Laboratory Medicine, 320 East North Avenue, Pitts- burgh, Pennsylvania 15212. similar to the human trophoblastic hormone. 30 Accepted for publication December 2, 1977. In 1976, Cohen and Strampp17reported that 0008-543X-78-0400-1217-0120 @ American Cancer Society 1217 1218 CANCERApril 1978 Vol. 41

TABLE1. “Control bacteria,” from American Type chemical techniques we are employing, the pro- Culture Collection duction and localization of the antigen, and by 1. Aeromonas hydrophila iubsp. formicans ATCC 13137 electron microscopy, their ultrastructural char- 2. (htridium haemolilicum ATCC 9650 acteristics. 3. (.%oslridiumnougi Type A ATCC 19402 While the testing for the presence of the anti- 4. Eschrrichiu coli ATCC 25922 gen in the cancer cells as well as in all the 5. l~urohacterii~mriucleatum ATCC 25586 aforementioned bacteria was being carried out, 6. Huemophilus aqyptus ATCC 11116 7. I,aclobardlur casei ATCC 7469 Dr. Robert J. Ryan at the Mayo Medical 8. .Iljcoplamia huminis Type 1 ATCC 23114 School, in collaboration with Dr. Nancy D. 9. .Ilr/raxrlle lacunata ATCC 17067 Richert, isolated from the ovarian fluid of a sow 10. I’reudomonas aeruginosa ATCC 27853 two Gram negative, motile bacteria, classified as 11. PrriLrionionaJ mallophilta ATCC 13270 ATCC 13636 Pseudomonas maltophilia and Enterobacter cloacae on ATCC 17448 the basis of physiological, nutritional and bio- ATCC 17666 chemical characteristics.39 Subsequent out- 12. l’seudomonar lrrlosteroni ATCC 11996 growth of the isolated strains in liquid media 13. Serratia rriarcrsrens ATCC 13880 14. Stafihylocriccus epidermtdi.s ATCC 14970 demonstrated that this porcine P. maltophilia was 1.5, Strr~~tobecillirsmonill/ormis ATCC 14647 capable of binding ‘251-labeledhCG, whereas E. 16. ,Struptococcur mirtans ATCC 2.5175 cloacae showed no gonadotropin binding. Among several strains of different bacteria utilized as controls, the investigators used a P. maltophilia they had obtained two cultures of P. cryptocides obtained from American Type Culture Collec- from Dr. Livingston and confirmed the produc- tion (ATCC 13637). Their preliminary data tion of a CG-like glycoprotein not only by radio- showed that the P. maltophilia culture media con- immunoassay (KIA) using specific antibody to tained a protein molecule crossreactive with the P-subunit of the trophoblastic but antisera to total CG as well as to its P-subunit. also by two assays for specific biologic activity, But the most important aspect concerning the radioreceptor analysis41 and the testosterone ATCC 13637 strain of P. maltophilia was that stimulating assay. l9 The glycoprotein nature of Richert and Ryan did not disclose in their pub- the material was demonstrated by its chromato- lication the fact that this bacterium was originally graphic behavior on Concanavalin A-Sepharose isolated from the oropharyngeal region of a patient with columns. ‘O mouth cancer. 25 Cohen and Strampp stated in their pub- Because of all these findings, a systematic in- lication that the bacteria received from Dr. Liv- vestigation for the presence of the CG-like anti- ingston were Gram positive cocci. These investi- gen in bacteria from proven sources (known ori- gators also reported that they isolated bacteria gin) was initiated in order to demonstrate if the from the urine of a patient with terminal carci- CG-like antigen is ubiquitous in nature, or is noma of the colon which also produced CG-like present only in bacteria isolated from humans material. In contrast to the material obtained and animals with cancer, spontaneously devel- from Dr. Livingston, these bacteria were Gram oped and/or experimentally induced. The pre- negative rods, identified as Escherichia coli by fer- liminary results of our investigations are hereby mentation criteria. reported. Working independently, Affronti and his as- sociates have also described the isolation of two MATERIALSAND METHODS CG-like protein producing microorganisms. One was classified as a epidermidis “Control Bacteria” and “Cancer Associated strain obtained from patients with advanced Bacteria” carcinoma of the breast, and the other was an E. The “control bacteria” utilized in this work colz strain isolated from patients with adeno- has been obtained from different sources. First, carcinoma of the COIO~.~-~ from American Type Culture Collection Because of our findings of the de noun biosyn- (ATCC), using strains which origin has been thesis of a CG-like glycoprotein by every cell established with complete assurance (Table l), classified as a cancer cell thus far tested, in and second, from clinical specimens from pa- tissue culture, in the experimental animal and in tients in which the presence of cancer was ruled the human being, 1.2,42 we decided to include out with almost absolute certainty (Table 2). these bacteria in our ongoing investigations in Young males, aged 8 to 18 years were the best order to study, by the specific immunohisto- source, since malignant lymphomas, brain neo- No 4 CG-LIKEPROTEIN IN BACTERIA Acevedo et al. 1219

plasms and testicular neoplasms, the most com- TABLE2. “Control bacteria.” laboratory isolates mon malignancies associated with this age group, can be easily ruled out. 1. Acinefobacfer anztralus 2. Aeromonas hydrophila Nine samples of “cancer associated bacteria” 3. Uordetella bronchisepfica were obtained from Dr. Virginia (W-C) Liv- 4 .Veisseria gonorrhoeae (10 strains) ingston, San Diego, . One of the cul- 5. I’seudomonads: tures was obtained from a neoplastic tissue, and a) P. aeruginosa h) P. cepacia the remaining were isolated from urine samples. c) P.,Juorescens All the patients were very advanced cases repre- d) P. maltophilia senting carcinomas, sarcomas and lymphomas. e) P. putrefacten5 Six additional cultures were obtained from Dr. f) P. Jtutzeri Lewis F. Affronti, George Washington Univer- 6. Staphylococcus: a) S.aureus (2 strains) sity Medical Center, Washington, D.C. All b) S. epidermzdis (8 strains) these samples were isolated from cancer tissue, three of them from patients with metastatic car- cinoma of the breast and the other three from patients with carcinoma of the colon. The proce- quality control and reproducibility, since sam- dures for the primary isolation of the bacteria ples of the same type of cells can be tested with were those previously described by Livingston antisera from different sources. and Alexander-Jacksonm and by Diller and Do- Specific rabbit antiserum to p-CG was sup- nelly. l’ plied by Dr. Vernon C. Stevens, The Ohio State We also obtained from ATCC, strain 13637 of University College of Medicine, and by Serono P. maltophilia, described as having been isolated Laboratories, Inc., Boston, Massachusetts. Spe- from a patient with mouth cancer, and an addi- cific rabbit antiserum to total choriogonadotro- tional bacterium, Streptococcus faecalis (ATCC pin was supplied by the same sources, and was 12818) which according to the ATCC Catalog also obtained from Cappel Laboratories, Down- was originally isolated from a patient with carci- ingtown, Pennsylvania, and from Miles Labora- noma of the gingival area.’ tories, Elkhardt, Indiana. All these antisera are All bacteria were classified by standard bacte- utilized as first antibodies. Rabbit antihorse riological procedures, that is by morphological, antiserum utilized for control, replacing the first physiological, nutritional and biochemical char- antibody was obtained from Cappel Laborato- acteristics, 49 including gas chromatographic ries. procedures as described by Holdeman and Fluorescein-labeled goat antirabbit antiserum Moore. 24 Gram staining and the Kynyoun acid- was obtained from Cappel Laboratories and fast procedure was performed in all isolates. from Behring Diagnostics, Rochester, New Biotyping of S. epidermidis was also performed York. Peroxidase-labeled goat antirabbit anti- according to the method of Baird-Parker. “Anti- serum and the peroxidase-antiperoxidase com- biograms were done by the Bauer-Kirby plex were obtained from Cappel Laboratories. method. l2 All cultures received from the differ- All these antisera are utilized as second anti- ent sources were stored at 4 C on Trypticase soy bodies. All antibodies were purified according to agar (BBL) and maintained in Trypticase soy actual standard procedures at their point of ori- broth containing 15% v/v glycerol at -70 C gin. (43). The indirect immunoperoxidase reaction requires 3,3’-diaminobenzidine-tetrachloride Cell Preparation (DAB) as peroxidase stain, 3.5 mg in 10 ml of The two immunohistochemical reactions, the phosphate buffered saline (PBS) with 0.015 ml indirect fluorescein-labeled and the indirect per- of 30% hydrogen peroxide. Light Green (Yel- oxidase-labeled reactions, were carried out in lowish SP) in 0.2% absolute methanol is used as unfixed air dried cells. The use of unfixed sam- counterstain, and Permount as mounting medium. ples is of importance, since the cells maintain Ninety-five percent glycerine in PBS, pH 7.0 f their morphology as well as their antigenicity. 0.2, 0.01 M is used as mounting media for the indirect immuno-fluorescein reaction. Reagents Reagent Controls: Utilization of CG-absorbed Highly purified and specific antibodies are first antibody (rabbit anti-serum to 0-CG and to obtained from different sources. This has the total CG) and elimination and/or replacement advantage of providing a way for monitoring of the first antibody by rabbit serum, rabbit 1220 CANCERApril 1978 Vol. 41 antihorse antiserum and/or PBS were used as were subsequently fixed in denatured alcohol for reagent controls. 15 minutes, washed in PBS for S-10 minutes, In addition to the aforementioned standard and stained with freshly prepared DAB for 10 controls for immunohisto-chemistry we also em- minutes. PBS washings were repeated again fol- ployed, as an additional test for specificity, what lowed by a one minute rinse in distilled water. we have called “cell controls,” which were Dehydration was then performed by washing proven CG-producing cells utilized as “positive twice for one minute in 95% alcohol. The sam- controls.” These were the nonclonal human ples were then counterstained with Light Green choriocarcinoma BeWo cell line given to us by for 15-20 seconds (3 slow dips), destained in Dr. Roland A. Patillo, from the Department of absolute methanol for 10-15 seconds (2 slow Obstetrics and Gynecology, Medical College of dips) and cleared with xylene. With the xylene Wisconsin, Milwaukee, Wisconsin, 38 and a sub- still on the slide, permount and coverslip were line obtained from ATCC, strain CCL 98. then applied. The presence of the trophoblastic- At the same time, proven “non-CG-produc- like protein is indicated by a brown to almost ing” cells were utilized as “negative controls. ” black granular or continuous staining depend- An example of such “negative cell control” is the ing on the concentration of the antigen. MA-169, a tissue cultured clone of human ade- The trophoblastic-like protein(s) is considered nomatous prostatic hyperplasia, obtained from to be present only when the indirect fluorescein- Microbiological Associates. In every instance, labeled and the indirect peroxidase-labeled im- reagents giving a positive reaction with the “pos- munohistochemical reactions are positive by itive cell control” have given a negative reaction light microscope. This is essential because with “negative cell controls”. equivocal reactions in the immunoperoxidase re- agent controls can sometimes be observed in Immunohistochemical Methods some cells, possibly due to intrinsic peroxidase The specimens were labeled following pre- content. viously described techniques. 1,10*28,32,33938,42 For Nonspecific fluorescence was not a problem the indirect immunofluorescein reaction, one or since it is easily recognized by an experienced more drops of the first antibody were added in observer, and can be eliminated by the use of order to cover all the preparation on the slide. appropriate filters and/or counterstaining. The samples were then incubated at room tem- perature in moist atmosphere for 30 minutes, Instruments using a moist ambient temperature incubator. The optical system for examining the prepa- The slides were then gently agitated on a Clini- rations subjected to the indirect immuno- cal Rotator (Eherbach Corporation, Ann Arbor, fluorescein reaction consists of a Leitz-Ortholux Michigan) for a period of 10 minutes. After microscope fitted with a darkfield condenser, 10 completion of this time, the slides were removed Xocular and achromatic oil immersion objective and drained briefly. Without allowing the prep- (95 X; NA l.O), and a lamp housing with an aration to dry, each sample was then treated Osram HBO 200 mercury vapor bulb. A BG-12 with one or more drops of fluorescein-labeled excitor filter is used in combination with a bar- second antibody. The 30 minute incubation was rier filter No. 47. Photographs are made with a repeated, followed by rinses and washes as 35 mm Leitz camera using Eastman Kodak above. The slides were then drained and air Photomicrography color film with exposure dried. A drop of BACTO mounting fluid (Difco times from 40 to 60 seconds. Laboratories, Detroit, Michigan) and coverslip Transmission electron microscopy studies were applied. Samples were then ready for ex- were performed using Philips 300 electron mi- amination with the ultraviolet microscope. croscopes, according to the method described by When present, the CG-like antigen is recognized Slifkin et a1.“ in order to study the ultrastruc- by an apple green fluorescence at the foci of the tural characteristics of the CG-producing bac- protein. From the moment of reconstitution of teria in comparison to those of the “non- the fluorescein-labeled second antibody, all producer^". work must be done in a darkened laboratory. The procedure followed for the indirect im- RESULTS munoperoxidase reaction was identical to the method used for the indirect immunofluorescein Figures 1 and 2 illustrate the results of the reaction up to the rinses and washes done fol- immunohistochemical reaction as applied to the lowing the incubation with the second antibody, “positive cell control” and to the “negative cell this time peroxidase-labeled. The preparations control” respectively. These “cell controls” No. 4 CG-LIKEPROTEIN IN BACTERIA Acevedo et al. 1221

of the great production of the antigen, CG, non- specific attachment of the fluorescein-tagged second antibody molecules to the cell did not occur. It is necessary to stress here that all the afore- mentioned controls with both markers as well as with “control cells” are performed every time cell tests are done, and every time new reagents are received. With respect to the “cancer associated bac- teria,” all the cultures received from Dr. Liv- ingston as well as 2 of the 3 isolated from breast carcinoma patients by Dr. Affronti were classi- fied as Staphyloccus epidermidzs, biotype 111. The other culture corresponded to a S. epidermidis biotype I. The results of the indirect immunofluorescein reaction and of the indirect immunoperoxidase- antiperoxidase reaction done on the S. epidermidis isolated from the cancer patients are illustrated by Figs. 4 and 5 respectively. The doughnut- shaped reaction of the CG-like antigen on the membranes and/or capsule is clearly demon- FIG. 1. “Positive cell control.” BeWo cell, nonclonal hu- strated in these microphotographs. Figures 6 man malignant trophoblast. Indirect immunofluorescein re- and 7 show the corresponding negative “reagent action with antiserum to the &subunit of CG as first anti- controls,” body. Note the intensity of the immunohistochemical ~i~~~~~ 8 and 9 illustrate the of the two reaction at the level of the plasma membrane (X 1520).

demonstrate the specificity of the entire tech- nique, in these examples, the indirect immuno- fluorescein reaction. The high concentration of the antigen, choriogonadotropin, at the level of the plasma membrane in the human malignant trophoblast is notable (Fig. 1). In contrast, the nonmalignant human pros- tatic cells showed an absolute lack of reactivity (Fig. 2). Since the indirect immunofluorescein reaction is capable of detecting a minimum of one to two thousand molecules of antigen in the pla~malemma,~~the lack of reactivity of this “negative cell control” demonstrate that the nonmalignant prostatic cells have only a minute expression of the CG-like antigen and/or do not express the information for its biosynthesis at all. Figure 3 illustrates the results of one of the standard controls for the specificity of the sec- ond antibody, in this experiment, the fluores- cein-tagged antibody. A complete negative reac- tion was obtained because the first antibody (specific antirabbit antiserum to the trophoblas- tic hormone) was replaced by PBS, and the second antibody is a fluorescein-labeled goat antirabbit antiserum. Since the cell used in this F,c, 2, cell control,^^ MA-169, human non- experiment Was the CG-producer malignant tro- malignant prostatic cells. Indirect immunofluorescein reac- phoblast, the results demonstrate that, in spite tion with antiserum to total CG as first antibody (X 1520). 1222 CANCERApril 1978 Vol. 41 Tables 1 and 2 were also completely negative for both immunohistochemical reactions, thus dem- onstrating the lack of and/or extremely low ex- pression of the information for the synthesis of the CG-like antigen. Possible morphologic differences between the CG-like protein producing “cancer associated bacteria” and the “normal” non-producing forms were analyzed by a serial study of their ultrastructural characteristics by transmission electron microscopy (Figs. 14-21). Only an in- creasing thickness of the cell wall of the CG-like protein producers was apparent. Gas chromatographic analysis of short and long chain fatty acids, used as “finger~rinting”’~ did not reveal any difference between the “pro- ducers” and the “nonproducers. ” Acid fastness was not demonstrated by the CG-like protein producing bacteria.

DISCUSSION By the use of a specific antibody to 0-CG as well as to the complete hormone, CG, and the application of the highly sensitive indirect fluo- rescein-labeled and peroxidase-labeled antibody FIG. 3. “Reagent control” for specificity of the second antibody, BeU’o cell. Indirect immunofluorescein reaction. techniques (sandwich technique, Or anti- First antibody was replaced by PHS. Negative results dem- onstrate that in spite of great production of the antisen, nonspecific attachment of the Huorescein-tagged second antibody to the cell does not occur (X 1520). immunohistochemical reactions performed with ATCC 13637 strain of P. maltophilia and figs. 10 and 11 their corresponding reagent controls. The presence of the CG-like antigen is clearly evident, thus demonstrating the synthesis of the trophoblastic-like protein (Figs. 8 and 9). Two of the three microorganisms isolated by Dr. Affronti and his associates from patients with carcinoma of the colon were classified as Escherichia coli. These two strains also revealed an active synthesis of the CG-like protein (Figs. 12- 13). Their corresponding reagent controls were totally negative. It is necessary to note at this point, that all the CG-like protein producing microorganisms clas- sified as S. epidermidis showed Gram positive characteristics. In contrast, the P. maltophilia and E. coli were demonstrated to be Gram negative bacteria. The third bacteria isolated from a patient with carcinoma of the colon were classified as heudomonas aeruginasa. This microorganism, as well as ATCC 12818 of Streptococcusfaecalis FIG. 4. S. epzdermzdzs, biotype 111, isolated from urine of patient with metastatic carcinoma. Indirect immuno- did not synthesize the trophoblastic-likeprotein, fluorescein reaction with antiserum to the P-subunit of CG since both immunohistochemical reactions were as first antibody, N~~~the “doughnut”shape as well as the totally negative. The “control bacteria” listed in intensity of the reaction (x 1520). No. 4 CG-LIKEPROTEIN IN BACTERIA Aceuedo et al. 1223 body technique), we have demonstrated the presence of a membrane (capsule)-associated CG-similar or -identical immunoreactive pro- tein in twelve samples of Staphylococcus epidermidis, two samples of Escherzchia coli, and ATCC 13637 strain of P. maltophzlia. In addition to the synthesis of the CG-like common antigen, the only other common fea- ture of such a diverse group of bacteria uas the fact that every one of these microorganisms was isolated ,from patients bearing malignant neoplasms. On the other hand, the absence of the CG-like antigen in the Pseudumunas aeruginosa isolated from a patient with cancer of the colon as well as in ATCC 12818 strain of S. faecalis, demon- strated that nut eriery “cancer associated bacteria” has the capability to synthesize the trophoblastic- like protein at least in detectable amounts. The negative results also obtained with a great num- ber of “noncancer control” bacteria, strongly supported the idea that the existence of these CG-like protein producing microorganisms is not a ubiquitous finding. The demonstration of the de nuuo bacterial biosynthesis of a glycoprotein so similar to the human trophoblastic hormone that it shares not FIG i. s epdermtdu. biotype I. isolated from tumor of only antigenic sites but also membrane receptor patient with metastatic adenocarcinoma of breast. Indirect t,inding sites, as in the of ATCC 13637 p, immunoperoxidase-antiperoxidase reaction with antiserum lo hCX as first antibody. Note the intensity of the immuno- maltophilia, 39 has tremendous biological implica- histochemical reaction at the level of plasma membranes tions. It may become one of the most important (cell wail anti/or capsule) (X 1600).

FIG. 6-7. Reagent controls of Figs. 5 and 6 for specificity of the second antibody. First antibody was replaced by PBS. The negative results indicate the absence of nonspecific attachment of the second antibody to the cells in the absence of the specific first antibody (X 1520 and X 1600). 1224 CANCERApril 1978 Vol. 41

discoveries in the biomedical field if the biosyn- thesis of significant amounts of the glyco-protein is proven to occur only in bacterial strains that are growing in or have been exposed to the presence of cells and/or tissues associated with ectopic production of the CG-like protein, in other words, with cancer cells. The explanation of the phenomenon is quite difficult, Since until now the biosynthesis of the protein(s) has been proved only in strains grow- ing in the presence of CG-producing cells, the potential of “genetic exchange” with subsequent expression of the mammalian gene by the bacte- rial cells becomes an interesting possibility. Nevertheless, such natural genetic recombina- tion is difficult to explain for several reasons. One of them, for instance, is that CG is a sia- loglycoprotein with a molecular weight of about 40,000, constituted by two asymmetrical sub- 4 = Fie: 8. I’. niultuphzlia, ATCC 13637 strain, isolated from the oropharyngeal region of a patient with mouth cancer. Indirect immunofluorescein reaction with antiserum to the &subunit of hCG as first antibody. The intensity of the immunohistochemical reaction is notable at the peripheral level (X 1520).

FIG. 9. P. maltuphdia, ATCC 13637 strain. Indirect peroxidase- anti-peroxidase reaction. Again note the intensity of the reaction at the level of the bacterial wall. (X 1600). No. 4 CG-LIKEPROTEIN IN BACTERIA Aceuedo et a/. 1225

FIGS. 10-1 1. Reagent controls of Figs. 8 and 9 for specificity of the second antibody 1;irst antibody was replaced by PBS (X 1520 and X 1600).

FIG. 12. E. colz, isolated from tumor of patient with carci- noma of colon. Indirect immunofluorescein reaction with antiserum to total CG. Note the intensity of the reaction in Frc;. 13. I<. rolz, as above. Indirect peroxidase-anti- the cell wall (X 3040). peroxidase reaction (X 1600). 1226 CANCERApril 1978 Vol. 41

FIGS.14-17. Electron microscopy ultrastructural characteristics of a mitotic. cycle of “CGlike protein producers” S. ppihnidz,, biotype 111, (Figs. 14, 16, 18, 20) as compared to that of the “non-producers” (Figs. 15. 17, 19, 21). The preparations were ohtained from a 15-hour culture. It can be noted that theonly apparent difference is the thickness of the cell wall (original magnification. X t00,OOO). Fiq. 14-upper left; Fig. 1.5-upper right; Fig. 16-lower left; Fig. 17-lower right. units not covalently linked. Moreover, the avail- gests that the two subunits are made by separate able information concerning the biosynthesis of genes. A natural process is apparently needed, the hormone by the human trophoblast as well which would involve plural recombinant events. as by cancer cells in tissue culture’6~26~45~48sug- Adelberg, in his Report to the Recombinant No. 4 CG-LIKEPROTEIN IN BACTERIA Aceuedo et al. 1227

F~c;s.18-21. See legend for Figs. 14-17. Fig. 18-upper left; Fig. 19-upper right; Fig. 20-lower left; Fig. 21-lowrr right. 1228 CANCERApril 1978 Vol. 41

DNA Molecule Program Advisory Committee of brane antigen. The cancer cells and sperma- the NIH4 also suggested that the phenomenon tozoa are the only other cells in which such may also be explained if the gene coding for the membrane-associated protein has been local- CG-like protein is normally present in all bac- ized. This protein has been demonstrated to be teria as a result of conservation during evolution immunologically and biologically similar to or as as example of convergent evolution. choriogonadotropin, CG, the glycoprotein hor- Such an explanation may be supported by the mone synthesized by mammalian tropho- demonstration that bacterial and mammalian blasts, 1-3,9,35,42 cell membranes possess a number of glycopro- The biological and physiological implications teins which are immunochemically and biologi- of such findings in respect to cancer etiology, cally cross-reactive. 31 For instance, ABH (0) diagnosis, prevention and treatment are obvi- substances, known to be present in human epi- ous. The impact of these findings in the fields of thelial secretions, are also present in over half of oncology, bacteriology, epidemiology, genetics some of 300 species of Gram negative organisms and molecular biology is so great that a detailed in the same or closely related chemical forms. description will be beyond the scope of this com- The explanation is also supported by our own munication. Since none of the current theories work demonstrating the synthesis of the CG-like relating to these fields can explain the phenome- protein by cancer cells in rats and non observed with the bacterial systems, it is These rodents do not synthesize CG during their apparent that this phenomenon exposes the gestation. The synthesis of the trophoblastic-like need for a new approach to the analysis as well protein that takes place after the malignant as to our current concepts of cancer. transformation occurred left no doubt that the coding for the glycoprotein synthesis was avail- Finally, the feasibility of a practical appli- able and that it was very possible that malignant cation may be one of the most significant aspects transformation was the event that derepressed of our investigations. The presence of a common or activated such information. membrane antigen expressed or detectable only Such explanation in the case of the bacteria in cancer cells, spermatozoa, trophoblasts and will require the existence of a repressing com- these “cancer associated” bacterial strains, ex- ponent or absence of a positive effector molecule poses the potential for a specific immunological in an organism that, as far as we know, only approach to the prevention and treatment of the contains one double strand of DNA. One will disease as well as for fertility control. A bacterial have to consider therefore, mRNA directives as vaccine elaborated with these CG-like protein an intrinsic part of such a mechanism, if it does producing microbes may theoretically produce exist. antibodies sharing antigenic determinants The results that we have obtained leaves little against the common membrane antigen. The doubt of the existence of a biological association significance of this possibility in the fields of among the bacteria, cancer cells and sperma- cancer control and reproduction control is obvi- tozoa through the presence of a common mem- ous.

REFERENCES

1. Acevedo, H. F., Slifkin, M., Pouchet, G. R., and Rak- 5. Affronti, L. F., Grow, L., and Begell, F.: Character- hshan, M.:Human chorionic gonadotropin (hCG) in cancer ization of bacterial tumor isolates. Fed. Proc. 34(3): 1043, cells. I. Identification in zn uztro and in vivo cancer cell sys- 1975. (Abst. 4676.) tems. Proc. 3rd. Int. Symp. DPCa., New York, April 26-29, 6. Affronti, L. F., Grow, L., Brumbaugh, R., and Orton, 1976. Abst. 356. K.: Bacterial tumor isolates. Abst Ann. Meeting Am. SOC. 2. Acevedo, H. F., Slifkin, M.,Pouchet, G. R., and Rak- hlicrobiol., Atlantic City, NJ, May 3-8, 1976. Abst. E-56. hshan, M. : Choriogonadotropin (CG) in cancer cells. 111. 7. Affronti, I,. F., Grow, L., Brumbaugh, R., and Orton, Further studies in zn mtro and in vzuo cell systems. Prog. K. : Production of human gonadotropin-like substance by 58th Annual Meet. The Endocrine Society, San Francisco, bacterial tumor isolates. Abst. Ann. Meeting Am. SOC.Mi- CA, June 23-25, 1976. Abst. 420. crobiol., New Orleans, hlay 8-13, 1977. Abst. E-23. 3. Acevedo, H. F., Slifkin, M., Pouchet, G. R., and Rak- 8. American Type Culture Collection Catalog of Strains hshan, M.: Identification of the @-subunit of choriogona- I, 12th Edition. H. D. Katt, Ed. Rockville, hID.. 1976. dotropin (CG) in human spermatozoa. In The Testis in Nor- 9. Asch, R. H., Fernandez, E. O., and Pauerstein, C. J.: mal and Infertile Men. P. Troen and H. R. Nankin, Eds. Immunodetection of hCG-like substance in human sperm. New York, The Raven Press, 1977, pp. 185-192. Frrlzl. Slerzl. 28:1258-1262, 1977. 4. Adelberg, E. A. : The production of a human hormone 10. Bain, J., and Ezrin, C.: Immunofluorescent local- (chorionic gonadotropin) by bacteria isolated from cancer ization of the LH of the human adenophypophysis. 3. Clin. patients: A possible instance of gene transfer in vivo between Endocrul. Melab. 30:181-184, 1970. human and bacterial cells. In Report to members of the 11. Baird-Parker, A. C. : The basis for the present classifi- Recombinant DNA Molecule Program Advisory Committee cation of staphylococci and micrococci. Ann N.T. Acad. Scz. of the NIH. Bethesda, January 6, 1977. 236:7-13, 1974. No. 4 CG-LIKEPROTEIN IN BACTERIA Aceuedo et al. 1229

12. Bauer, A. W., Kirby, M. N., Sherris, J. C., and Turk, bedded in polyethylene glycol with a peroxidase-labeled M. : Antibiotic susceptibility testing by a standardized disc antibody method. -7. Htstochem. Cytochem. 20:969-974, 1972. method. Am. 3. Clin. Path. 45:493-496, 1966. 33. Moriarty, G. C., Moriarty, C. M., and Sternberger, 13. Beck, J. S., Porteous, I. B., and Ullyot, J. L.: Gona- A.: Ultrastructural immunocytochemistry with unlabeled dotropin-secreting bronchial carcinoma: Aberrant endo- antibodies and the peroxidase-antiperoxidase complex. A crine activity or trophoblastic differentiation? 3. Path. technique more sensitive than radioimmunoassay. 3. Hzsto- 101:59-62, 1970. chem. Cytochem. 21 325-833, 1973. 14. Braunstein, G. D., Vaitukaitis, J. L., Carbone, P. P., 34. Muggia, F. M., Rosen, S. W., Weintraub, B. D., and and Ross, G. T. : Ectopic production of human chorionic Hansen, H. H.: Ectopic placental proteins in non- gonadotropin by neoplasms. Ann. Int. Med. 78:39-45, 1973. trophoblastic tumors. Serial measurements following 15. Castro, A. E., AIon60, A., and Mancini, R. E. : Local- chemotherapy. Cancer 36: 1327-1337, 1975. ization of follicle-stimulating hormone and luteinizing hor- 35. McManus, L. M., Naughton, M. A,, and Martinez- mones in the rat testis using immunohistological tests. 3. Hernandez, A. : Human chorionic gonadotropin in human Endocrtnol 52:129-136, 1972. neoplastic cells. Cancer Res. 36:3476-3481, 1976. 16. Chou, J. Y., Robinson, J. C., and Wang, S-S.: Effects 36. Patillo, R. A.: Trophoblastic : Chorionic of sodium butyrate on synthesis of human chorionic gona- gonadotropin hormone production, antigenic expression, dotropin in trophoblastic and nontrophoblastic tumors. Na- and trophoblastic redifferentiation in multiple forms of ture 268:543-544, 1977. malignancy. In Pathobiology Annual. Meredith Corpora- 17. Cohen, H., and Strampp, A.: Bacterial synthesis of tion, New York, 1973, pp. 241-268. substance similar to human chorionic gonadotropin. PTOC. 37. Perlin, E., Engeler, J. E., Edson, M., Karp, D., et al: Soc. Exp. Bid. Med. 152:408-410, 1976. The value of serial measurement of both human chorionic 18. Diller, I. C., and Donnelly, A. J.: Experiments with gonadotropin and alph-fetoprotein for monitoring germinal mammalian tumor isolates. Ann. N. Y. Acad. Sci. 174:655- cell tumors. Chncer 37:215-219, 1976. 674, 1970. 38. Petrusz, P. : Light microscopic localization of binding 19. Dufau, M. L., Catt, K. J., and Tsuruhara, T.: Sensi- sites for human chorionic gonadotropin in luteinized rat tive gonadotropin responsive system and radioimmunoassay ovaries by a peroxidase-labeled antibody method. 3. Hzsto- of testosterone production by the rat testis in uttro. Endocri- chem. Cytochem. 21 :279-282, 1972. nology 90: 1032- 1040, 1972. 39. Richert, N. I)., and Ryan, R. J.: Specific gonadotro- 20. Dufau, M. L., Tsuruhara, T., and Catt, K. J.: Inter- pin binding to Pseudomonas maltophilia. Proc. Natl. Acad. Scz. action of glycoprotein hormones with Agarose-Con- 1I.S.A. 74:878-882, 1977. canavalin-A. Biochim. Btophys. Acta 278:281-292, 1972. 40. Samaan, N. A,, Smith, J. P., Rutledge, F. N., and 21. Goldstein, D. P., Kosasa, T. S., and Skarin, A. T.: Schultz, P. N.: The significance of measurement of human Clinical application of a specific radioimmunoassay for hu- placental lactogen, human chorionic gonadotropin and car- man chorionic gonadotropin in trophoblastic and non- cinomembryonic antigen in patients with ovarian carci- trophoblastic tumors. Surg. Gynecol. Obstet. 138:747-751, noma. Am. J. Obstet. Gynecol. 126:186-189, 1976. 1974 41. Saxena, B. B., Hasan, S. H., Haour, F., andSchmidt- 22. Goldstein, D. P., and Kosasa, T. S.: The subunit Collwitzer, M. : Radioreceptor assay of human chorionic radioimmunoassay for hCG-Clinical application. In Prog- gonadotropin: Detection of early pregnancy. Science ress in Gynecology. M. L. Taymor and T. H. Green, ,Jr., 184:793-795, 1074. Eds. Grunne & Stratton, New York NY, 1975, 6:145-184. 42. Slifin, M., Acevedo, H. F., Pardo, M., Pouchet, G. R., 23. Goldstein, D. P. : Chorionic Gonadotropin. Cancer and Rakshan, M. : Human chorionic gonadotropin in cancer 38:453-459, 1976. cells. 11. Ultrastructural localization. Proc. 3rd. Int. Symp. 24. Holdeman, L. V., and Moore, W. E. C. Eds.: Anaero- DPCa., New York, April 26-29, 1976. Abst. 357. bic Laboratory Manual. The Virginia Polytechnic Institute 43. Slifkin, M., and Pouchet, G. R.: Rapid carbohydrate and State University Anaerobic Laboratory, Blacksburgh, fermentation test for confirmation of the pathogenic Neisseria VA. 1972. using a Ba(OH)zindicator.J. Cltn. hficrobiol. 5:15-19, 1977. 25. Hugh, R., and Ryschenkow, E.: Pseudomonas malto- 44. Slifkin, M., Merkow, L. P., Kreuzberger, S. A., Eng- phzlia, an alcaligenes-like species. 3. Gen. Micmbiol. 26: 123- wall, C., and Pardo, M.: Characterization of COZdependent 132, 1961. microcolony variants of Staphylococcus aureus. Am. 3. Clin. Path. 26. Hussa, R. 0.: Immunological and physical character- 56:584-592, 1971. istics of human chorionic gonadotropin and its subunits in 45. Tashjian, A. H., Jr., Weintraub, B. C., Barowsky, cultures of human malignant trophoblast. 3. Clin. Endocrol. N. J,, Rabson, A. S., and Rosen, S. W.: Subunits of human Metab 44:1154-1162, 1977. chorionic gonadotropin. Unbalanced synthesis and secretion 27. Kawamura, A,, Jr. : Fluorescent antibody techniques by clonal cell strains derived from a bronchogenic carci- and their applications. University of Tokyo Press, Tokyo, noma. Prac. ,Wall. Acad. Sri. U.S.A. 70:1419-1422, 1973. University Park Press, Baltimore, London, Tokyo. Second 46. Tormey, D. C., Waalkes, T. P., Ahmann, D., Gehrke, Edition, 1977. C. W., Zumwatt, R. W., Snyder, J., and Hansen, H.: Bio- 28. Leleux, P., and Robin, C.; Immunohistochemistry of logical markers in breast carcinoma. I. Incidence of abnor- individual adenohypophyseal cells. Acta Endacr. (Xbk) Supp. malities of CEA, hCG, three polyamines and three minor 153:168-189, 1971. nucleosides. Cancer 35:1095-1100, 1975. 29. Livingston, V. W-C., and Alexander-Jackson, E.: A 47. Vaitukaitis, J. L., Braunstein, G. D., and Ross, G. T.: specific type of organism cultivated from malignancy: bacte- A radioimmunoassay which specifically measures human riology and proposed classification. Ann. .N.Y. Acad. Sci. chorionic gonadotropin in the presence of human luteinizing 174:636-654, 1970. hormone. Am. 3. Obstet. Gynecol. 113:751-758, 1972. 30. Livingston, V. W-C., and Livingston, A. M.: Some 48. Vaitukaitis, ,J. L. : Changing placental concentrations cultural, immunological and biochemical properties of P70- of human chorionic gonadotropin and its subunits during genitor Cryptorides. Trans. .NY Acad. Sci. 36:569-582, 1974. gestation. J. Clin. Endocrind. Metab. 38:755-760, 1974. 31. Markowitz, A. S.: Protoplasmic and plasma mem- 49. Washington, J. A,, II., Martin, W. J., and Karlson, brane relationships. Trends Btochem. Sci. 1 :161-163, 1976. A. G.: Identification of bacteria. In Laboratory Procedures 32. Mazurkiewicz, J. E., and Nakane, P. K.: Light and in Clinical Microbiology. T. A. Washington, 11, Ed. Little, electron microscopic localization of antigens in tissue em- Brown and Co., Boston, 1974, pp 53-124.