Reaction Patterns of Herpes Simplex Virus Type 1 and Type 2 Proteins with Sera of Patients with Uterine Cervical Carcinoma and Matched Controls
Steven C. Gilman, John J. Docherty, Aileen Clarke, et al. Cancer Res 1980;40:4640-4647. Published online December 1, 1980.

Updated Version

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/40/12/4640
E-mail alerts Reprints and Subscriptions Permissions
Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at pubs@aacr.org. To request permission to re-use all or part of this article, contact the AACR Publications Department at permissions@aacr.org.
Downloaded from cancerres.aacrjournals.org on June 6, 2011 Copyright 1980 American Association for Cancer Research
[CANCER RESEARCH 40, 4640-4647, 0008-5472/80/0040-OOOOS02.00

December 1980]

Reaction Patterns of Herpes Simplex Virus Type 1 and Type 2 Proteins with Sera of Patients with Uterine Cervical Carcinoma and Matched Controls1
Steven C. Oilman,2 John J. Docherty,3 Aileen Clarke, and William E. Rawls
Department of Microbiology. Cell Biology. Biochemistry, and Biophysics, The Pennsylvania State University, University Park. Pennsylvania 16802 [S. C. G., J. J. D.; epartment of Preventive Medicine and Biostatistics, University of Toronto. Toronto, Ontario M5S 1A8 Canada A.C.]: and Department of Pathology. McMaster D University. Hamilton. Ontario L8N 3Z5 Canada [W. E. R.
ABSTRACT Serum from 105 individuals with diagnosed uterine cervical cancer and 231 matched controls were examined tor their ability to react with a large number of herpes simplex virus type 1 or type 2 (HSV-1, HSV-2) proteins. Radiolabeled HSV-1 or HSV-2 proteins were mixed with test serum and immune com plexes were isolated with staphylococcal protein A. Viral pro teins in the immune complexes were resolved by polyacrylamide gel electrophoresis and visualized by fluorography. When the frequency of precipitation for cancer and control serum was calculated for each HSV-1 and HSV-2 protein, the results demonstrated that four HSV-1 and 11 HSV-2 proteins were precipitated more frequently by cases than by controls (p 0.05). However, since these results could be influenced by the presence or absence of HSV-2-specific antibodies as well as social, economic, and sexual history, the data were grouped and analyzed according to these parameters. This enabled all significant differences between case and control sera in the precipitation of HSV-1 or HSV-2 proteins to be abolished except for two HSV-2 proteins with molecular weights of 38,000 and 118,000. These two proteins appear to be tumor associated and not merely covariables of past infec tion or risk factors alone. INTRODUCTION The association of HSV-2" with CX cancer was suggested
to interpretation due, in part, to the high incidence of HSV-1 and HSV-2 antibodies in the general population and the ab sence of antibodies to these viruses in some patients with this form of cancer (19). It is, therefore, not surprising that a series of reports have appeared (1, 2, 5, 8, 17, 24, 27) describing antigens from HSV-infected or -transformed cells that are tumor associated but vary substantially in characteristics including method of production, method of isolation, molecular weight, and method of assay. Additionally, because of the large number of proteins that HSV has the genetic capacity to code for (10), it is likely that only a fraction of the total have been tested for tumor relatedness. Accordingly, we have solubilized a large number of radiolabeled HSV-1 and HSV-2 proteins and observed their reaction patterns with human sera from CX cancer patients and matched controls using RIP-PAGE and fluorography (13). The results demonstrated initially that a large group of HSV proteins were precipitated more frequently by sera from CX cancer patients than by sera from control women. However, after taking into account the possible influence of differences in HSV-2 typespecific antibodies among cases and controls as well as con trolling for various cancer-related risk factors, only 2 HSV-2 proteins with molecular weights of 118,000 and 38,000 were found to have tumor-associated activity. MATERIALS AND METHODS

by seroepidemiological studies which demonstrated that the frequency and titer of antibody to this virus were higher in the sera of patients than in the sera of controls (16, 20, 23). The significance of these studies has been problematic and open
1This work was supported by Contract N01-CB-74171 from The Division of Cancer Biology and Diagnosis of The National Cancer Institute, NIH. and grants from The National Cancer Institute of Canada and The Ontario Treatment and Research Foundation. 2 Present address: Scripps Clinic and Research Foundation, 10666 North Torrey Pines Road, La Jolla, Calif. 92037. 1 To whom requests for reprints should be addressed, at Department of
Microbiology, Cell Biology. Biochemistry and Biophysics, 101 South Frear Build ing, The Pennsylvania State University, University Park, Pa. 16802. " The abbreviations used are: HSV-2, herpes simplex virus type 2; CX cancer, squamous cell carcinoma of the uterine cervix; HSV-1. herpes simplex virus type 1; HSV, herpes simplex \irus; RIP-PAGE, radioimmune precipitation-polyacrylamide gel electrophores^s; VERO, African green monkey kidney cells; PBS, phosphate-buffered salire (27 mw KCI:14 ITIM KH2PO:137mM NaCI:80 mM Na2HPO<. pH 7.4); RIPA, radioimmune precipitation buffer [ 150 mM NaCI: 10 mM Tris:1 mM MgCI2 (pH 7.4) containing 1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, and 1% aprotinin]; PAGE, polyacrylamide gel electrophoresis; SPA, staphylococcal protein A; BSA, bovine serum albumin; RIA, radioimmunoassay; M-1, tumor antigen from HSV-1 -transformed cells; M-2, tumor antigen from HSV-2-transformed cells; HSV-TAA, herpes simplex virus tumor-associated antigen. Received June 18, 1980; accepted September 8, 1980.
Case-Control Sample. Human sera from women with cervi cal neoplasia and matched controls were collected from To ronto area residents. Cancer sera were obtained prior to ther apy from females between the ages of 20 and 69 years with histologically confirmed invasive cervical cancer diagnosed within a 3-year period ending September 30, 1976 (3). Control sera for this study were acquired by going from door to door starting at the fourth residence to the right of the dwelling of the case and proceeding systematically through the residential block or apartment building until a woman, within 10 years of the age of the patient, was found who would participate in the study. Cases and controls were interviewed using a standard questionnaire designed to obtain information on social, repro ductive, and economic factors. A total of 105 sera from CX cancer patients and 231 sera from controls were coded and then tested in this present study. All sera were stored at 20. Cells and Virus. VERO cells (CCL81 ; American Type Culture Collection, Rockville, Md.) were grown in Medium 199 supple mented with 5% fetal calf serum, 0.075% NaHCO3, kanamycin (100 jug/ml), and penicillin (100 units/ml). HSV-1 (Seibert strain) and HSV-2 (316-D strain) were obCANCER RESEARCH VOL. 40

HSV Proteins in Cervical Carcinoma tained from Dr. F. Rapp, The Pennsylvania State University by centrifugation (12,000 x g, 4, 10 min and washed twice Medical School, Hershey, Pa. Virus stocks were prepared in with RIPA buffer and once with 0.0625 M Tris-HCI buffer, pH VERO cells, titered by the plaque assay (4), and stored at 6.8 (0.5 ml/wash). The cell pellets were resuspended in 150 -70. /il of 0.0625 M Tris-HCI (pH 6.8) containing 4% sodium dodecyl sulfate and 10% 2-mercaptoethanol and heated for 10 min at Radiolabeling of HSV Proteins for Immunoprecipitation. Roller bottle cultures (490 sq cm) of VERO cells at 85 to 95% 75 to release cell-bound immune complexes. The prepara confluence were mock infected with medium or infected with tions were centrifuged (12,000 x g, 4,10 min) to remove the HSV-1 of HSV-2 at a multiplicity of infection of 20. After 1 hr adsorption at 34,the cells were rinsed once with Tris-buffered NaCI solution and overlaid with 5 ml of methionine-free Dulbecco's modified minimal essential medium containing onebacterial cells and particulate debris, and the resultant super natant containing immune complexes were stored at - 20until analyzed by PAGE. Prior to PAGE analysis, all samples were solubilized by heating to 100 for 2 min, and sucrose was added to a final
tenth the normal leucine concentration and 1% fetal calf serum. concentration of 10%. Slab gel electrophoresis was performed The cultures were radiolabeled from 3 to 24 hr after infection with [35S]methionine (20 /iCi/ml; >800 Ci/mmol) and [14C]- on an equal volume (25 /il) of the solubilized immune precipi leucine (5 /iCi/ml; ^300 mCi/mmol). At the end of the labeling tates using the discontinuous buffer system of Laemmli (12) period, the samples were scraped from the culture vessel, except that the stacking gel Tris concentration was 0.25 M and collected by centrifugation (1500 x g, 4,5 min), and washed the gels were cross-linked with diallyltartardiimide. The actual twice with 40 ml of PBS. After the final centrifugation, the cpm for each sample added to the gels varied from less than supernatant was decanted, and the cells were stored as a 100 in controls (i.e., negative HSV serum or any sera reacted pelletat -70. with mock-infected cells) to over 2000 for a strongly positive Solubilization of HSV Proteins. Frozen radiolabeled infected serum. Acrylamide concentrations were 3% in the stacking gel or control cell pellets were thawed by addition of 1 ml of RIPA and 10% in the separating gel. Electrophoresis was carried out buffer. Fifty /gf DNase I (Sigma Chemical Co., St. Louis, Mo.) at 60 V for 2 hr followed by 120 V for 3 hr. The Coomassie o were added, and the cell suspensions were vortexed for 30 Brilliant Blue-stained gels were processed for fluorography and sec, sonicated on ice for 1 min (20 kc, MSE disintegrator), and exposed to preflashed Kodak X-OMAT film for 7 days (13) allowed to stand on ice for 20 min. Insoluble cellular and viral unless otherwise stated. Molecular weight standards included debris were removed by centrifugation at 100,000 x g for 1 hr myosin (M.W. 220,000), /3-galactosidase (M.W. 130,000), at 4,and the resultant supernatant containing RIPA-soluble phosphorylase a (M.W. 94,000), BSA, (M.W. 68,000), ovalbumin (M.W. 43,000), and carbonic anhydrase (M.W. 30,000). material was used immediately as the antigen source for im Controls of SPA-RIP-PAGE Assay. Prior to examining the mune precipitation tests. This procedure routinely solubilized reaction pattern of human sera with radiolabeled HSV-1 and 60 to 80% of recoverable radioactivity representing 44 pro teins, as determined by PAGE analysis and fluorography, with HSV-2 proteins, SPA precipitations were examined for feasi bility and specificity using sera from rabbits immunized with molecular weights of 27,000 to 220,000. The specific activity HSV-1 or HSV-2. These sera were prepared by treating New of the antigen preparations ranged from 3000 to 4000 cpm/ Zealand white rabbits with injections of primary rabbit kidney tig protein. cells infected with HSV-1 or HSV-2 emulsified in Freund's In comparative solubilization studies, the same procedure complete adjuvant. The animals were test bled periodically and described above was followed except RIPA buffer was replaced by Triton X-100 buffer (27 HIM KCI:14 mw KH2PO4:137 mw boosted prior to exsanguination. Rabbit immune sera used to NaCI:80 HIM Na2HPO4 (pH 7.4) containing 1% aprotinin and examine specificity and the acquisition source are as follows: 5% Triton X-100) or PBS. However, these buffers resulted in Anti-adenovirus 12 and anti-simian adenovirus 7, F. Rapp; antithe solubilization of less than 35% of total recoverable counts Streptococcus tecalis, J. Pazur, Biochemistry Program, The Pennsylvania State University; anti-Autographa californica bacand were not suitable for use in studies presented here. SPA Immunoprecipitation. Staphylococcus aureus Cowan ulovirus and anti-gypsy moth baculovirus, W. McCarthy, The I (American Type Culture Collection no. 12598) used as an Pesticide Research Laboratory, The Pennsylvania State Uni versity. Additionally, antiserum BSA was prepared by repeated immunoadsorbent, was grown, fixed, and heat killed as de scribed by Kessler (11 ). The final preparation was stored at 4 s.c. injections of BSA into the haunches of rabbits. When sera as a 10% cell suspension in PBS containing 0.1% sodium from rabbits immunized with HSV-1 or with HSV-2 were reacted with solubilized radiolabeled viral proteins, adsorbed with SPA, azide. Prior to use in radioimmune precipitation experiments, the bacterial cells were washed once with 4 volumes of RIPA and separated in PAGE, a total of 30 HSV-1 antigens and 32 buffer (4)and resuspended to their original 10% concentration HSV-2 antigens were detected (6). The molecular weights of these peptides ranged from 27,000 to 212,000, and crossin the same buffer. In the RIP-PAGE assay, 10 /il of undiluted serum (rabbit or reactions were common. However, when these rabbit sera human) were reacted with 100 /il of RIPA-solubilized extracts were reacted with radiolabeled VERO cell proteins, nonspecific of radiolabeled HSV-1, HSV-2, or mock-infected VERO cells. immunoprecipitation was not detected. Additional control stud The RIPA extracts were diluted to contain in 100 /il 25 to 35 ies included attempts to precipitate in the SPA system radio/tg protein representing 105 cpm of radiolabel. After overnight labeled HSV proteins using rabbit immune serum to the viral, incubation at 4,the primary immune complexes were collected bacterial, and protein agents listed above. However, the results were consistently negative with no bands resolved by PAGE by addition of 200 /I SPA. This amount of SPA was sufficient of and fluorography even when exposure of film to gels was to remove all human IgG as previously determined by quanti tative radial immunodiffusion. Following 30 min incubation at increased to 21 days. Also preimmune rabbit sera did not 4,the SPA with adsorbed immune complexes were collected precipitate radiolabeled HSV-1, HSV-2, or VERO cell proteins.

DECEMBER 1980

S. C. Oilman et al. RIA for HSV-2-specific Antibodies. The RIA was carried out on plastic-coated beads as described previously (18, 25). Briefly, antigens were prepared by infecting VERO cells with HSV-1 or HSV-2 and incubating the cultures at 37 until over 90% of the cells showed cytopathic effect. The cells were then detached and washed with PBS. Mock-infected cells were processed in a similar manner and used as control antigen. The antigens to be used to sensitize the beads were prepared by resuspending the cells to 10% (v/v) in PBS and lysing them by freeze-thawing. The cell pellets resulting from the final wash were used for adsorbing antigen by lysing the cell pellet without prior dilution. The sera were diluted 1:75 in PBS containing 1% BSA, and 5 wells each received 0.15 ml of the diluted serum. Then 0.05 ml of HSV-1-adsorbing antigen was added into each of 2 wells, while 0.05 ml of diluent was added to the remaining 3 wells. The mixtures were incubated for 1 hr at room temper ature after which beads coated with HSV-1, HSV-2, and control antigen were added to the 3 wells receiving diluent. Beads coated with HSV-1 and HSV-2 antigen were added to the wells receiving adsorbing antigen. After overnight incubation, unreacted antibody was removed by washing with H2O, and the beads were then reacted with 125l-labeled antibody to human IgG for 3 hr. Unbound labeled immunoglobulin was then re moved by washing in H2O, and the bound antibody was counted in a gamma counter. Excess radioactivity bound to beads coated with HSV-2 antigens as compared to HSV-1-coated beads by adsorbed serum was taken as evidence of HSV-2specific antibody (18). Statistical Analysis. Differences in the frequencies of pre cipitation of the HSV proteins were examined significance using the x2 test. for statistical
RESULTS Reactivity of CX Cancer and Control Sera with HSV Pro teins. Coded sera from 105 CX cancer patients as well as 231 controls were analyzed for their ability to precipitate radiolabeled HSV-1 or HSV-2 proteins. Each serum was reacted simultaneously with solubilized extracts of radiolabeled HSV1, HSV-2, or mock-infected VERO cell proteins, and SPA immune precipitates were prepared and processed for PAGE and fluorography as described. Standard molecular weight curves were constructed using proteins with known molecular weights, and the molecular weights of all viral proteins precip itated by each human serum were recorded. The data were expressed qualitatively (i.e., presence of a HSV antigen in the immune precipitate) rather than quantitatively. Presented in Figs. 1 to 3 are examples of the reaction patterns of a randomly selected group of cancer and control sera with HSV-1 (Fig. 1), HSV-2 (Fig. 2), and VERO cell (Fig. 3) proteins. It is apparent from Figs. 1 and 2 that differences as well as similarities existed between the reaction patterns of individual sera and that VERO cell proteins were not precipitated by human sera (Fig. 3). Indeed, none of the 336 sera studied here precipitated detect able radiolabeled VERO cell proteins even when the film was exposed 3 times longer (i.e., 21 days) to the gels containing control preparations than to those containing immune precipi tates.

Fig. 1. RIPA-solubilized radiolabeled HSV-1 proteins immunoprecipitated by human sera and SPA. Proteins were resolved by PAGE and visualized by fluorography. Lanes A. C, D, F. and /, sera from cervical cancer patients: Lanes B. E, G, and H, sera from controls. Ordinate, molecular weights x 10~3.

CANCER

RESEARCH

VOL. 40

HSV Proteins in Cervical Carcinoma Individual human sera precipitated up to 23 different proteins from solubilized HSV-1- or HSV-2-infected cells. However, a total of 31 different HSV-1 (Table 1) and 27 HSV-2 (Table 2) antigens were identified when all cancer and control sera were analyzed. The molecular weights of the precipitated HSV pro teins ranged from 30,000 to 220,000 (Tables 1 and 2). The frequency distribution of each protein precipitated by sera from cases and controls was determined, and the differences were examined for statistical significance by the x2 test. Twentyseven of 31 HSV-1 antigens were precipitated with the same frequency (p ^ 0.05) by cancer and control sera. However, HSV-1 antigens with molecular weights of 75,500, 49,000, 46,000, and 31,000 were precipitated by a significantly higher (p < 0.05) proportion of CX cancer sera than by control sera (Table 1). The precipitation frequencies of HSV-2 antigens are pre sented in Table 2. Sixteen of the 27 HSV-2 antigens were precipitated at the same frequency (p 0.05) by both groups of sera. However, 11 HSV-2 antigens were precipitated at a significantly higher frequency (p 0.05) by CX cancer sera than control sera. The molecular weights of the HSV-2 antigens precipitated more frequently by CX cancer sera were 118,000, 101,000, 76,000, 63,000, 51,000, 49,000, 43,000, 38,000, 34,000, 31,000, and 30,000. Three (2.9%) of the cancer sera failed to precipitate any HSV-1 or HSV-2 antigens in RIP-PAGE and were therefore considered negative for antibody to HSV. In contrast, 22 (9.5%) control sera were negative for antibody to HSV-1, and 20 (8.7%) control sera were also negative for HSV-2 antibody.
Table 1 Molecular weights and frequency of precipitation of HSV-1 proteins immune precipitated by cervical cancer and control sera precipitation3M.W. Frequency of Percentage of sera precipitating the protein in question. Cervical cancer sera frequency of precipitation minus control sera frequency of precipitation. Table 2 Molecular weights and frequency of precipitation of HSV-2 proteins immune precipitated by cervical cancer and control sera precipitation8M.W. Frequency of
10~3)2201611451311181071019894837673666359575251494340383735343130Cervical (X cer era43.870.525.797.150.543.895.269.521.957.182.929.519.072.49.559.07.687.687.687.634.34 s sera38.559.326.890.932.936.885.359.322.945.966.226.412.157
Reactivity with HSV Proteins and Past HSV-2 Infections. Thirty-three % of CX cancer sera and 22% of control sera had HSV-2-specific antibodies detected by RIA. Thus, the observed differences in the frequency of precipitation of HSV proteins

can by sera from CX cancer patients and controls could reflect IO"3)21217016013312411210499969388.78275.57366.562575049464340393736353433323130Cervical (x cer era81.013.341.995.255.220.091.495.26.73.89.590.566.737.154.348.665.76.783.846.785.758.141.017.124.819.010.54.821.981.941.0Control s sera72.713.033.889.245.915.284.087.93.53.911.382.347.631.246.844.654.17.472.729.980.151.934.213924.215.28.74.815.266.733.8Difference08.30.38.16.09
Percentage of sera precipitating the protein in question. 0 Cervical cancer sera frequency of precipitation minus control sera frequency of precipitation.
differences in prior infection with HSV-2 rather than being related to the cancerous state. We reasoned that, if cases and controls were selected on the basis of the presence of HSV-2specific antibodies and then analyzed with respect to their ability to precipitate HSV proteins in RIP-PAGE, any significant differences between this subset of cases and controls would more accurately reflect differences related to the cancer itself. When only those sera from cases and controls that contained HSV-2-specific antibodies, as determined by RIA, were exam ined for ability to precipitate the proteins listed in Tables 1 and 2, all significant differences between cases and controls were abolished (Table 3) except for the M.W. 38,000 and M.W. 118,000 HSV-2 proteins. Thus, all the HSV-1 proteins and the HSV-2 proteins with molecular weights of 101,000, 76,000, 63,000, 51,000, 49,000, 43,000, 34,000, 31,000, and 30,000 most probably reflect past infection with HSV-2 and do not appear to be tumor associated. Reactivity with HSV Proteins and Cancer Risk Factors. The controls were matched to the CX cancer patients on the basis of age and residence; however, they differed with respect to certain factors associated with an increased risk of developing CX cancer. For those for whom information was available, first coitus was experienced before 19 years of age by 46% of the cases and 32% of the controls. Marriage before 19 years occurred among 35% of the cases and 25% of controls, while 35% of cases and 24% of the controls had a history of unstable 4643
S. C. Oilman et al. marriages. Forty % of the cases and 29% of the controls had less than 9 years of formal education. These differences were statistically significant for age at first coitus, marital stability, and education at the 95% level. To exclude the possibility that the reactivity of the M.W. 38,000 and M.W. 118,000 proteins was not a covariable of

the risk factors associated with CX cancer, the reactivity to the proteins was examined in relation to the presences of HSV-2specific antibodies among cases and controls with similar risk attributes (Table 4). Among sera with HSV-2-specific antibod ies, both the M.W. 38,000 and M.W. 118,000 proteins were precipitated more frequently by cases than by controls with similar risk attributes. The differences reached statistical sig nificance for marriage at younger and older ages, unstable Table 3 marital history, and 9 or more years of education for the M.W. Molecular weights and frequency of precipitation of possible tumor-associated 38,000 protein. For the M.W. 118,000 protein, all the differ HSV proteins by sera with HSV-2 type-specific antibodies detected by RIA ences reached statistical significance except for first coitus Frequency3 of precipitation after 19 years of age and 1 to 8 years of education. Such by sera differences were not observed among sera without HSV-21CT3)75.549.046.031.0118101766351494338343130Cx cancer8294629182100978297979471569168Control679048885098928396909041389463P>0.05>0.05>0.05>0.05<0.01>0.05>0.05>0.05>0.05>0.05>0.05<0 ProteinsHSV-1HSV-2M.W.(x specific antibodies (Table 4). Substantial differences between cases and controls in the frequency of precipitation of the M.W. 75,500 and M.W. 46,000 proteins of HSV-1 and the M.W. 34,000 protein of HSV-2-specific antibodies were observed, but the differences did not reach statistical significance (Table 3). The reactivity of the sera to these proteins were also analyzed with respect to the risk factors. Differences similar to those shown in Table 4 were not found for the 2 HSV-1 proteins, but a similar pattern was observed for the M.W. 34,000 protein (data not shown). For this protein, the differences reached levels of significance for first coitus younger than 19 years, first marriage in both age groups, and 1 to 8 years of education among women with a Percentage of sera precipitating the protein in question. HSV-2-specific antibodies.
Table 4 of HSV-2 proteins (M. W. 38.000 and 118.000) in relation to HSV-2-specific of cervical carcinoma Frequency of precipitation M.W. 38,000 factorAge Risk coitus<19 1st yrCasesControls5M9yrCasesControlsAge of polypeptide M.W. 118,000

Precipitation

antibodies and risk factors
(34)26/109(24)8/22 1/32 (35)10/11 marriage<19 1st yrCasesControls519 f9/22(41)13/16(81)1 (36)12/33(36)15/46(33)27/134 (91 yrCasesControlsMarital (35)14/22 1/31
(22)30/109(28)9/22(41)15/33(45)14/46(30)34/134(25)13/46(28)
(82)"8/22 (36)16/16(100)15/31
(48)17/22(77)"16/34(47)11/12(92)"8/14(57)8/11 (20)14/46(30)32/134(24)10/24(42)1
statusStableCasesControlsUnstableCasesControlsEducation1-8yrCasesControls59 (64)14/34 (41)10/12(83)"6/14(43)9/11

(28)1 (46)15/40(38)9/24 1/24

(28)9/24 1/40

(38)21/43(49)9/36 (82)10/17(59)12/16(75)c8/27(30)RIA-negative8/23(35)13/45(27)1 (38)15/43(35)12/36(33)24/123(20)RIA-positive12/13(92)"10/17(59)7/10(70)10/20(50)9/11 (73)6/17(35)15/16(94)13/27(48)RIA-negative12/23(52)14/45(31)7/32 yrCasesControlsRIA-positive8/13(62)"9/17(54)7/10(70)7/20 (25)26/123(21) Numbers in parentheses, percentages. " Case-control difference significant at p < 0.05 > 0.01. Case-control difference significant at p < 0.01.
HSV Proteins in Cervical Carcinoma DISCUSSION In the present study, 231 control sera and 105 sera from patients with CX cancer were analyzed for their ability to react with HSV-1 and HSV-2 proteins. We found significantly more sera from CX cancer patients with antibodies to 4 HSV-1 and 11 HSV-2 proteins than sera from control women. The reactivity to 27 HSV-1 and 16 HSV-2 proteins was not significantly different between cases and controls. When the controls were matched to the cases on the basis of age and residence, however, the 2 groups differed with respect to evidence of past infections by the HSV's. When the subset of cases and controls with HSV-2-specific antibodies was analyzed, most of the dif ferences disappeared and only 2 HSV-2 proteins were precip itated significantly more often by sera from cases than by sera from controls. Thus, while antibodies to a number of virus proteins are more often present in sufficient quantities for detection among cases than controls, only the M.W. 38,000 and M.W. 118,000 HSV-2 proteins were tumor associated. This association could not be attributed to differences in cer vical cancer risk factors between cases and controls. It has been repeatedly demonstrated that sera from patients with CX cancer have a higher titer and a greater occurrence of neutralizing antibodies to HSV-2 than did sera from control women (3). Implied in this observation is the existence of an HSV-2 type-specific glycoprotein antigen that is the target site for neutralizing type-specific antibodies and an enhanced im mune response of CX cancer patients to this antigen when compared to controls. While the identity of an HSV-2 typespecific glycoprotein is yet to be resolved, it is of value to note that the HSV A and B glycoproteins have molecular weights (26) similar to the M.W. 118,000 protein observed in our study. Indeed, it has been observed by us that [14C]glucosaminelabeled HSV-2 proteins comigrate with the [35S]methionineand [14C]leucine-labeled M.W. 118,000 protein of this study. Additionally, in SPA immunoprecipitation studies with rabbit anti-HSV-sera, the M.W. 118,000 protein labeled with [14C]glucosamine was efficiently precipitated (6). Further support for the possible tumor association of this protein may be obtained from recent studies which demonstrated that HSVtransformed cells contain the A and B viral glycoproteins.5 While it is tempting to speculate that the M.W. 118,000 protein may be the target site for neutralizing antibodies for HSV-2 in sera of CX cancer patients, further studies are required to substantiate such a claim. Recent in situ hybridization studies using radiolabeled HSV DNA have demonstrated the presence of HSV RNA in a signif icant number of tumor tissue samples from patients with CX cancer (15, 29). Additionally, in vitro transformation of cells with endonuclease-generated fragments of HSV-2 DNA has revealed that the transforming region is contained within 0.582 to 0.628 map unit of the genome (22).6 Contained within this region, or immediately adjacent to it, is the gene for an HSV-2 protein with a molecular weight of 37,800 (14). This is similar in size to the M.W. 38,000 protein which we found more reactive with sera from CX cancer patients than from controls. However, an HSV-2 glycoprotein with a molecular weight of 118,000, which appears similar to the M.W. 118,000 tumor5 J. G. Lewis and R. J. Courtney, personal communication. 6 I. R. Cameron. N. M. Wilkie. and J. C. M. Macnab. personal communication.

associated protein described in this study, is found within genome map limits 0.35 to 0.40 (14). This site is separate from the putative HSV-2 transforming region. If the M.W. 118,000 protein reactive in our study is the same as that located at the 0.35 to 0.40 map location, substantially more than the trans forming region of the viral genome would have to be integrated and expressed for the M.W. 118,000 protein to be eliciting the enhanced antibody responses observed in the cases of this study. Of particular interest are similarities between the M.W. 38,000 and M.W. 118,000 protein observed here and HSV tumor antigens described by others (1, 2, 5, 8, 17, 24, 27). As reported, such tumor antigens are obtained from HSV-1- or HSV-2-infected or -transformed cells and assayed by a variety of methods including immunofluorescence (5), complement fixation (2, 8, 17), RIA (1), and leukocyte inhibition response (24). Viral tumor antigens AG-4, VP143, VP134, and AG-e have molecular weights of 161,000 (2, 27), 143,000 (5), 134,000 (1 ), and 130,000 to 140,000 (24), respectively. In the RIP-PAGE assay used in this study, no HSV-1 or HSV-2 pro teins in the M.W. 130,000 to 160,000 range were precipitated more frequently by cancer sera than by controls (Tables 1 and 2). While the M.W. 118,000 protein observed in the present study was approaching this molecular weight range, it probably does not represent AG-4, VP143, VP134, or AG-e since the molecular weights are quite dissimilar even considering possi ble errors in sodium dodecyl sulfate-PAGE estimates of molec ular weights (28). Tumor antigens of HSV that do have molecular weights similar to the M.W. 38,000 protein observed in this study are HSV-TAA and M-1 and M-2. While the precise molecular weights of HSV-TAA proteins are not clear, they appear to lie between M.W. 40,000 and M.W. 60,000 (9). Tumor antigens M-1 and M-2 which are isolated from HSV-1- and HSV-2transformed cells (1 7) consist of a number of proteins between M.W. 30,000 and M.W. 60,000. Additional studies have dem onstrated that these HSV-transformed cells contain viral pro teins with molecular weights of 30,000, 33,000, and 58,000 as well as the M.W. 59,000 CP-1 protein (7, 21 ). In the present study, the M.W. 38,000 protein is within these molecular weight ranges and demonstrated significant tumor-associated activity. The relationship between the M.W. 38,000 tumor-associated protein and HSV-TAA and M-1 and M-2 is currently under investigation. Hence, 1 of the 2 HSV tumor antigens identified by RIPPAGE in this study is a viable counterpart to some previously described HSV tumor antigens. It is important to note that such comparisons are based primarily on similarities in molecular weight which may vary from laboratory to laboratory and make exact comparisons difficult. Nonetheless, such comparisons can serve as a framework for the elucidation of HSV antigens which have tumor-associated activity. Additionally, it should be noted that, although we were unable to find direct evidence to support HSV tumor antigens AG-4, VP143, VP134, and AG-e, this cannot be interpreted as definitive evidence against the possible role of these proteins as HSV tumor antigens, as these HSV proteins may not be effectively solubilized by RIPA buffer, may not retain antibody binding activity in this buffer, may not be produced by our virus strains in the VERO cells used in this study, or may not be detected by our assay system. Thus, while positive associations between our results and those of 4645

S. C. Oilman et al. others are significant, negative relationships are more difficult to interpret. Additionally, it should be noted that no single unique protein emerged as a specific HSV tumor antigen and reacted consist ently in a significant manner with cancer sera in all risk groups and not all CX cancer patients had HSV-2 type-specific anti body. The lack of a unique HSV tumor protein may in part reflect the commonness of these viruses in the human popu lation. Control sera had precipitating antibodies to all the anti gens examined and were distinguished from patients' sera only by the frequency by which each group precipitated a given viral antigen. The greater reactivity of CX cancer sera to antigens may reflect a more frequent exposure of CX cancer patients to HSV-2 than of controls, a greater number of recur rent infections in patients than of controls, the partial activation of latent HSV-2, or the expression in the tumor cells of certain viral antigens resulting in the observed frequency differences. The observation that not all sera had antibody to HSV-2 has been previously noted (19). Such results suggest that CX cancer may have additional causes besides HSV-2 or HSV-2 is a cofactor or incidental to the cancer process itself. ACKNOWLEDGMENTS
The authors wish to thank Dr. Kendall O. Smith for excellent and continued support. 34: 1122-1125, 1974. 10. Honess, R. W., and Roizman, B. Proteins specified by herpes simplex virus. XI. Identification and relative molar rates of synthesis of structural and nonstructural herpes virus polypeptides in the infected cell. J. Virol., 12: 1347-1365, 1973. 11. Kessler, S. W. Rapid isolation of antigens from cells with a staphylococcal protein A antibody adsorbent: parameters of the interaction of antibodyantigen complexes with protein A. J. Immunol., 115: 1617-1624, 1975. 12. Laemmli, U. K., Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lond.), 227. 680-685, 1970. 13. Laskey, R. A., and Mills, A. D. Quantitative film detection of 3H and '"C in polyacrylamide gels by fluorography. Eur. J. Biochem. 56. 335-341,1975. 14. Marsden, H. S., Stow, N. D., Preston, V. G., Timbury, M. C., and Wilkie, N. M. Physical mapping of herpes simplex virus-induced polypeptides. J. Virol., 20:624-642. 1978. 15. McDougall, J. K., Galloway, D. A., and Fenoglio, C. M. Cervical carcinoma: detection of herpes simplex virus RNA in cells undergoing neoplastic change. Int. J. Cancer, 25. 1-8, 1980. 16. Nahmias, A. J., Josey, W. E., Naib, Z. M., Luce, C. F., and Guest, B. Antibodies to Herpesvirus hominis types 1 and 2 in humans. II. Women with cervical cancer. Am. J. Epidemici. 91: 547-552, 1970. 17. Notier, M. F. D., and Docherty, J. J. Reactions of antigens from herpes simplex virus transformed cells with sera from squamous cell carcinoma patients. Cancer Res., 36: 4394-4401, 1976. 18. Patterson, W. R., Rawls, W. E., and Smith. K. O. Differentiation of serum antibodies to herpesvirus types 1 and 2 by radioimmunoassay. Proc. Soc. Exp. Biol. Med., 757. 273-277, 1978. 19. Rawls, W. E., Clarke. A., Smith, K. O., Docherty, J. J. Gilman, S. C., and Graham. S. Specific antibodies to herpes simplex virus type 2 among women with cervical cancer. In: Viruses in Naturally Occurring Cancers. Cold Spring Harbor, N. Y.: Cold Spring Harbor Laboratory, in press. 1980. 20. Rawls, W. E., Tomkins, W. A. F., and Melnick. J. L. The association of herpesvirus type 2 and carcinoma of the uterine cervix. Am. J. Epidemici. 89:547-555, 1969. 21. Reed, C. L., Cohen, G. H., and Rapp, F. Detection of a virus-specific antigen on the surface of herpes simplex virus-transformed cells. J. Virol., 75: 668670, 1975. 22. Reyes, G. R. La Femina, R., Hayward, S. D. and Hayward, G S. Morpho logical transformation by DNA fragments of human herpesviruses: evidence for two distinct transforming regions in HSV-1 and HSV-2 and lack of correlation with biochemical transfer of the thymidine kinase gene. Cold Spring Harbor Symp. Quant. Biol., in press. 23. Royston, I., and Aurelian, L. The association of genital herpes-virus with cervical atypia and carcinoma in situ. Am. J. Epidemiol., 97: 531-538, 1970. 24. Smith, C. C., and Aurelian, L. Proteins of herpesvirus type 2. V. Isolation and immunologie characterization of two viral proteins in a virus-specific antigenic fraction. Virology, 98: 255-260, 1979. 25. Smith, K. O., and Gehle, W. D. Magnetic transfer devices for use in solidphase radioimmunoassays and enzyme-linked immunosorbent assays. J. Infect. Dis., 736 (Suppl.): S329-S326, 1977. 26. Spear, P. Membrane proteins specified by herpes simplex viruses. I. Identi fication of four glycoprotein precursors and their products in type 1-infected cells. J. Virol., 77: 991-1008, 1976. 27. Strnad, B. C., and Aurelian, L. Proteins of herpesvirus type 2. II. Studies demonstrating a correlation between a tumor-associated antigen (Ag-4) and a virion protein. Virology, 73: 244-258, 1976. 28. Weber, K., and Osborn, M. Reliability of molecular weight determinations by dodecyl sulfate-polyacrylamide gel electrophoresis. J. Biol. Chem., 244: 4406-4412, 1969. 29. Wilkie, N. M., Eglin, R. P., Sanders, P. G., and Clements, J. B. The association of herpes simplex virus with squamous carcinoma of the cervix, and studies of the virus thymidine kinase gene. In: Interactions between Virus and Host Molecules. The Royal Society Discussion Meeting. New York: Academic Press, Inc., in press, 1980.

REFERENCES

1. Anzai, T., Dressman, G. R., Courtney, R. J., Adam, E., Rawls, W. E., and Benyesh-Melnick, M. Antibody to herpes simplex virus type 2 induced nonstructural protein in women with cervical cancer and controls. J. Nati. Cancer Inst. 54: 1051-1059. 1975. 2. Aurelian. L., Strnad, B. C. and Smith, M. F. Immunodiagnostic potential of a virus-coded, tumor-associated antigen (Ag-4) in cervical cancer. Cancer (Phila.), 39: 1834-1849, 1977. 3. Clarke, E. A., and Anderson, T. W. Does screening by pap smears help prevent cervical cancer? A case-control study. Lancet, 2:1-4, 1979. 4. Docherty, J. J., Goldberg, R. J., and Rapp. F. Differential effect of 7,12dimethylbenz(a)anthracene on infectivity of herpes simplex virus type 2. Proc. Soc. Exp. Biol. Med., 736: 328-333. 1971. 5. Flannery, V. C., Courtney, R. J., and Schaffer. P. A. Expression of an early, nonstructural antigen of herpes simplex virus in cells transformed in vitro by herpes simplex virus. J. Virol., 2?. 284-291, 1977. 6. Gilman, S. C. Identification of the immunogenic proteins of herpes simplex virus and their relation to cervical carcinoma. Ph.D. Dissertation, The Penn sylvania State University. University Park. Pa., 1979. 7. Gupta, P., and Rapp, F. Identification of virion polypeptides in hamster cells transformed by herpes simplex virus type 1. Proc. Nati. Acad. Sei. U. S. A., 74:372-374, 1977. 8. Hollinshead. A. C., Lee, D., Chretien. P. B., Tarpley, J. L., Rawls, W. E., and Adam, E. Antibodies to herpesvirus nonvirion antigens in squamous car cinoma. Science (Wash. D. C.), 182: 713-715, 1973. 9. Hollinshead. A. Tarro, G., Foster. W. A. Jr. Segal, L. J., and Jaffurs. W. J. Studies of tumor-specific and herpes-virus nonvirion antigens. Cancer Res.,
Fig. 2. RIPA-solubilized fluorography. Details as for Fig. 3. RIPA-solubilized visualized by fluorography.
radiolabeled HSV-2 proteins immune precipitated by human sera and SPA. Proteins were resolved by PAGE and visualized by Fig. 1. radiolabeled mock-infected VERO cell proteins immune precipitated by human sera and SPA. Proteins were resolved by PAGE and Details as for Fig. 1.