Binding by Immunoglobulin to the HPV-16 - Cancer Research

(CANCER RESEARCH 51, 4423-4429, August 15. 1991]

Binding by Immunoglobulin to the HPV-16-derived Proteins LI and E4 in CervicalSecretions of Women with HPV-related Cervical Disease1

Kimberly A. Snyder, Shannon R. Barber, Millie Symbula, Peyton T. Taylor,Christopher P. Crum, and James K. Roche2

Departments of Pathology [S. R. B., M. S., C. P. C.J, Microbiology [C. P. C.J, Obstetrics-Gynecology Â¡P.T. T., C. P. C.], and Internal Medicine [K. A. S., J. K. R.],University of Virginia Health Sciences Center, Charlottesrille, Virginia 22908

ABSTRACT

Although DNA of the human papilloma viruses (HPV) can be identifiedin epithelium of a large proportion of patients with genital squamouslesions, relatively little is known about the extent of the local host immuneresponse to this virus. We analyzed cervical secretions from patientsundergoing evaluation because of abnormal Papanicolaou smears (cervical biopsy showed nonspecific atypia, flat condyloma, or intraepithelialneoplasia), as well as controls, for immunoglobulin binding to proteinsproduced in vitro to HPV-16 LI, E4, and E7 open reading frames.

Segments of the HPV-16 genome, including portions of the LI (1111-cleotides 6153-6794), E4 (nucleotides 3399-3648), and E7 (nucleotides686-880) open reading frames, were cloned into pATH vectors andexpressed as tryptophan synthetase E fusion proteins in Escherichia coliand used as a source of study antigens. Fusion proteins containing theHPV LI, E4, and E7 polypeptides were found to be distinct by molecularweight (59,000; 45,000; and 42,000) as well as by immunological determinants recognized by heterologous immune sera. Of 8 cervical intraepithelial neoplasia lesions tested by RNA-RNA in situ hybridization, 7were found to be positive for HPV-16-related nucleic acids, in contrastto none (0 of 4) in the condyloma group (three positive for HPV DNAother than type 16). Immunoglobulin in cervical secretions showed reactivity to HPV type 16 E4 or LI or both, with highest binding in patientswith cervical intraepithelial neoplasia (P < 0.01 for HPV-16 LI and E4compared with controls). Binding was not tryptophan synthetase E dependent and was, in general, coincident for the HPV-16 E4 and LIproteins. We conclude that study of cervical secretions, using a quantitative assay for immunoglobulin binding to HPV-16 proteins producedin vitro, may be useful to document the quality and quantity of theimmune response of the host to this important human pathogen.

INTRODUCTION

HPVs3 are a heterogeneous group of double-stranded DNA

viruses which can infect squamous epithelium of the genitaltract and are associated with genital warts, precancers (intraepithelial neoplasms), and invasive carcinomas (1-6). AlthoughHPV-16 is the principal type in intraepithelial neoplasms ofthe cervix, and HPV nucleic acids have been detected in a largeproportion of genital squamous lesions, relatively little is knownabout the cell biology of host infection by these viruses, specifically which HPV-encoded proteins are produced with infectionand the extent of the host immune response to these viralproteins. These questions have recently been addressed usingtechniques of fusion protein technology, in which DNA se-

Received 12/17/90; accepted 6/7/91.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported in part by USPHS Grants Ca47676. DK35182,

and DK42358, as well as by a grant from the American Cancer Society (MV 395)and by a Physician Scientist Award to C. P. C. (AI00628).

1To whom requests for reprints should be addressed, at Box 1005, MedicalResearch 4 Building, University of Virginia Health Sciences Center, Charlottes-ville, VA 22908.

3The abbreviations used are: HPV, human papillomavirus; CIN, cervicalintraepithelial neoplasia; ELISA, enzyme-linked immunosorbent assay; HPV-16,type 16 of human papillomavirus: LI, E4, and E7. fusion proteins expressingportions of the HPV-16 genome; NSE, normal squamous epithelium; ORF, openreading frames; SDS, sodium dodecyl sulfate.

quence information from previously characterized HPVs hasbeen used to produce plasmici constructs, in which a definedHPV ORF can be expressed as a fusion protein and used togenerate antisera or provide a target for analyzing the hostresponse. Using this approach, investigators have identified ahumoral immune response to HPV-6 and -16 proteins in studypopulations (7, 8), and type-specific epitopes have been definedfor HPV-6b which react with human sera (9). These studiesindicate that it may be possible to identify specific determinantson macromolecules encoded by HPV DNA and which elicit thehost immune response.

Because "genital" HPV DNA types may be recovered in the

conjunctiva, oropharynx, larynx, and subungual regions (10-13), the relevance of a humoral response in serum to genitalinfection remains unclear. One approach to this problem centers on the local immune response in the cervical mucosa.Antibodies reacting to herpesvirus, chlamydial, and gonococcalantigens, principally of secretory IgA isotype, have been isolatedfrom this region (14-17). In a recent report, detection of IgAantibodies reacting to bovine papillomavirus virion proteins incervical mucous specimens of a proportion of women withabnormal Papanicolaou smears was described (18). In thispreliminary study, we analyzed local cervical immune reactivityto fusion proteins produced in vitro to portions of the HPV-16LI, E4, and E7 ORFs. We selected these ORFs because we andothers have been able to demonstrate that HPV-16 LI and E4are expressed in some cervical precancers and E7 expressionhas been documented in cell lines from cervical cancers (19-22). We report the presence of antibody-binding activity to theLI and E4 (but not E7) proteins in a subset of women withcervical precancer lesions. In the process of developing a quantitative assay for binding activity, we have purified each studyprotein (LI, E4, E7) through preparative electrophoresis andelicited specific hyperimmune sera to them. The relationshipsamong HPV-specific binding activity in cervical secretions,lesion type, and associated HPV nucleic acids are discussed.

MATERIALS AND METHODS

Preparation and Characterization of Study Proteins

Expression and Purification. Segments of the HPV-16 genome including portions of the LI, E4, and E7 open reading frames were clonedinto GEM Vectors (Promega, Madison, WI), partially sequenced, andinserted in the appropriate reading frame in selected pATH vectors(23). The LI, E4, and E7 peptides corresponded to nucleotides 6153-6794, 3399-3648, and 686-880 of the HPV-16 genome (Fig. 1) (24),representing 54, 76, and 65% of these ORFs, respectively. Using aprotocol modified from Firzlaff et al. (7), we used constructs to transform into Escherichia coli MM294, and tranformants containing theinsert sequences were incubated overnight with shaking in minimalmedia with tryptophan (7). They were then diluted 1/10 in minimalmedia without tryptophan, shaken for 2 h, and then induced withindoleacrylic acid. Samples were shaken at 32Â°Cfor 2 h, pelleted, and

resuspended in lysis buffer. Whole lysates were resolved on SDS poly-

4423

Research. on December 7, 2018. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

HPV-I6-SPECIFIC ANTIBODY IN SECRETIONS

HPV- 16 GENOME

K H

ORFS

R 1

R2

R3

1.0 20 30 40 50 60 7.0 79

,URR ,

294 1947

E6

PvPS

1_1686 880

1095

E4

BÃ¶PSB PS

6153 6794

Fig. 1. Schematic diagram of the HPV-16 genome. Organization of ORFS RI,R2. and R3 is shown, along with the early (E) and late (L) virus proteins whichthey encode (number of base-pairs transcribed for each protein is indicated underits name). Bottom, subgenomic fragments cloned into pATH vectors for use inthis study. Letters, restriction sites used to procure the fragment; numbers, itsposition along the HPV-16 genome. URR, upstream regulatory (noncoding)region; K.B., kilobase; Pv. Pvu; PS, Pstl; Ba, Bai: B, BamHl.

acrylamide gel electrophoresis, and the size of the fusion proteins wasrecorded.

Fusion proteins were purified by preparative polyacrylamide gelelectrophoresis. A 5.9-12.8 mg of each sonicated preparation waselectrophoresed on a 10% SDS polyacrylamide slab gel, pH 8.8 (200V, 22Â°C,about 4 h) in an LKB model 2001 apparatus as described

previously (25), and electroelution was accomplished by a modificationof the method described by Hunkapiller et al. (26). The purity of eachgel-eluted band was assessed by reelectrophoresis on an analytical 10%SDS polyacrylamide gel, and immuno-reactivity as well as antigenicspecificity studied by immunoblotting, as described below. To determine whether the bacteria-coded portion of the fusion protein containedconstituents also recognized by cervical mucous specimens, 14 mg ofnoninsert containing pATH vector was fractionated and characterizedby the same methods. Protein was measured by the procedure of Lowryet al. (27) using 2-50 Â¿igof twice-recrystallized bovine serum albumin

(fraction V; Eastman Kodak Co., Rochester, NY) as the standard.Analytic Polyacrylamide Gel Electrophoresis. Whole cell lysates and

their electroeluted fractions were examined under SDS-denaturing conditions. A 10% system as described by Laemmli (28) was used. Aliquotsof 1 to 5 Mgprotein were dissolved in 0.5 M Tris buffer, pH 6.8-5%glycerol-0.001% bromphenol blue. Subsequently, the gels were fixed in50% methanol-10% acetic acid (30 min) followed by 10% glutaralde-hyde (30 min). Protein on gels was detected using a silver stain technique as described by Merrill et al. (29). Molecular weight standardsused were: phosphorylase B, 97,400; bovine serum albumin, 66,000;ovalbumin, 44,000, and carbonic anhydrase, 29,000.

Immunological Methods

Elicitation of Specific Antisera. Antibodies to the fusion proteinswere obtained through s.c. injection of 50-150 Â¿igof antigen, dissolvedin saline and emulsified in an equal volume of Freund's complete

adjuvant (Difco Laboratories, Detroit, Michigan), into the hind footpads of New Zealand white rabbits. A panel of sera at peak titer wassubsequently selected for study by a quantitative ELISA describedpreviously (30). Sera were studied at a 1:50,000-1:100,000 dilution,with and without prior absorption with bacterial iysates as indicated.Individual antigenic specificities were identified by immunoblot (seebelow).

Immunoblot Technique. This was carried out by a modification of themethod of Gershoni and Palade (31). Incubation for 2 h at 25Â°Cwith

first antibody (heterologous immune serum or preimmune serum, diluted > 1:50,000) was followed by exposure to a 1:600 dilution ofalkaline phosphatase-conjugated anti-rabbit IgG H and L chain (Kirkegaard and Perry, Gaithersburg, MD) for l h at 25Â°Cand then

substrate (5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetra-zolium, both from Sigma Chemical Co.). To determine the antigenicspecificity of antisera, the fusion protein preparation or control proteins

(sonicated pATH vector after induction or ovalbumin) were added in0- through 50-Mgamounts to antiserum (0.003 M!),incubated for 14 h,and pelleted (5000 rpm, 15 min). Supernatant (with and withoutabsorption) was then tested for residual binding to the study antigensimpregnated onto nitrocellulose, using alkaline phosphatase-conjugated second antibody in the standard immunoblot technique above.

Histochemical Verification of Study Antisera. To verify that the studyantigens elicited antisera which cross-reacted with HPV-related proteins, serial sections from biopsied lesions containing HPV-16 nucleicacids were analyzed by histochemistry, as previously described (21).Sections were incubated with anti-Li, -E4, and -E7 sera at 1:800dilution in 1% goat serum, and reaction products were detected with abiotinylated secondary antibody and avidin-biotin-peroxidase complex(Vector Laboratories, Burlingame, CA), using 3,3'-diaminobenzidine

(0.5 mg/ml; Sigma) as substrate. Controls included sections exposed toantisera but preincubated with their respective antigen.

ELISA. The LI, E4, E7 fusion proteins or control protein, trypEfrom the non-HPV gene insert-containing pATH vector, were adsorbedto the bottom of a 96-well microtiter plate (Falcon 3912; BectonDickinson, Oxnard, CA) by incubating 100 n\ of coupling buffer/wellcontaining 0.2 Mgof antigen. After 3 h at 37Â°C,the plates were washedand blocked with buffer containing 0.5% casein (Sigma) (12 h, 4Â°C).

Study cervical mucous specimens were applied at a 1:25 and a 1:100dilution. After extensive washing, followed by incubation with alkalinephosphatase-conjugated goat anti-human immunoglobulin (Kirkegaardand Perry), substrate (number 104 phosphatase tablets; Sigma) wasadded and the plate read at 2, 5, and 20 h on a Titertek Multiskanapparatus (Flow Laboratories, McLean, VA). A designation of positivewas assigned only when net absorbance at 405 nm of triplicate specimens was >0.140, this finding could be repeated with a second set oftriplicate specimens, and reactivity directed against HPV-16-codedprotein made up at least 35% of the read-out (absorbance) value.

Source and Analysis of Clinical Material

Patient and Sample Selection. Women presenting to the Universityof Virginia Colposcopy Clinic for the evaluation of an abnormal Pa-panicolaou smear were selected as patient cases. A speculum wasinserted, and cervical mucus was aspirated from the cervical os andplaced in normal saline with 1% bovine serum albumin. Samples werethen centrifuged and the supernatants were sonicated and stored with0.01% sodium azide at 4Â°C.As controls, cervical mucous samples from

women presenting for routine gynecological care, without an abnormalPapanicolaou smear or a cervical lesion and age-matched with patientcases (above), were identically processed and tested.

Morphological Analysis of Lesions. Biopsies obtained at the time ofcolposcopical examination were fixed in formalin, sectioned, and classified according to previously defined criteria (32) into the followingcategories: (a) negative or nonspecific atypia, (b) flat condyloma, and(c) CIN. The last two categories correspond approximately to mild-moderate and severe dysplasia/carcinoma in situ, respectively. The lastcategory (CIN) has been closely associated with HPV type 16 nucleicacid sequences in previous reports (3, 32).

Analysis of Specimens for HPV DNA. In order to determine the typeof HPV nucleic acids associated with the lesions under study, RNA-RNA in situ hybridization of biopsy material was performed using 35S-

labeled RNA probes obtained from HPV type 16, as previously described (33). As a control, serial sections were incubated with a mixedprobe consisting of HPV-11 and HPV-18 nucleic acids. The resultinghybridization signals were interpreted as either positive or negative forHPV-16 nucleic acids or HPV nucleic acids of types other than type16 (i.e., HPV-11 or-18).

Statistical Analysis. Since obtained values displayed a nonparametricdistribution, a Wilcoxon rank sum test (34) was used to compareimmunoglobulin binding among the patient groups. The primary datawere analyzed in two ways: (a) as net absorbance at 405 nm, obtainedby subtraction of tryptophan E binding from binding to the fusionprotein and (b) as a ratio, Abs fusion protein/Abs tryptophan E, wherenet absorbance favoring the fusion protein would have a ratio of > 1.00.Only in the case in which both analyses (e.g., net absorption and ratio)

4424



HPV-16-SPECIFIC ANTIBODY IN SECRETIONS

Fig. 2. SDS polyacrylamide gel electropho-resis of bacterial lysates. including those containing fusion proteins LI, E4, E7. used in thisstudy. After electrophoresis of noninduced(Un-I) and induced (I) lysates from pATHvectors containing HPV-16 nucleotide sequences 6153-6794, 3999-3648, or 668-880,the gels were stained as described by Merrill elal. (29). Subsequently, induced componentswere identified (arrows), cut from the gel, andelectroeluted according to the method of Hun-kapiller et al. (26). Tryptophan synthetase Ewas similarly prepared from unmodifiedpATH vector. Molecular weight markers were(in thousands) carbonic anhydrase (29), oval-bumin (45), bovine serum albumin (66), andphosphorylase b (97.4).

GenomicRegionAnolyteio-

97X~

661

45'S2

29LIUn-I

I-â€¢I

I.

t â€¢E4Un-I

Ii

%hE7Un-I

Iâ€¢-1

5pATHUn-I

I!.iI!h

SDS ImmunoblotAnolyte :

io'o

X^ 97-

^ 66-

3 45 HJ)o

LI P

AbsorbentType -Amount (jjg) â€”

LI P LI P LI P

LI25

LI50

LI P

Ovol25

Oval50

Fig. 3. Specificity for an HPV-16-codedprotein of immune sera as demonstrated byimmunoblotting. Twenty >jg of protein Ll-containing pATH lysate (lane LI), in parallelwith a control pATH lysate containing tryp-tophan synthetase E (lane P), was electropho-resed and then transblotted onto nitrocelluloseprior to incubation with a 1:100,000 dilutionof immune serum A, elicited to a fusion proteinencoded by HPV-16 nucleotides 6153-6794.Preexposure of the same serum to progressively increasing amounts of LI protein-containing pATH lysate diminished reactivity tothe LI band (first lane of panels 2 and 3) andless so to the tryptophan synthetase E band(second lane of panels 2 and 3). Exposure tosimilar amounts of ovalbumin did not diminishreactivity to the LI or the tryptophan synthetase E proteins (panels 4 and 5).

of group data demonstrated P < 0.05 was a statistically significantresult designated.

RESULTS

Partial Purification of Fusion Proteins

Bands representing the induced fusion proteins were readilyidentified (Fig. 2), cut from the gel, and electroeluted. Purifiedfrom lysates was 0.4 mg of LI fusion protein, 0.35 mg of E4fusion protein, and 0.32 mg of E7 fusion protein, for yields of6.7, 2.9 and 2.5%, respectively. By their migratory rate throughpolyacrylamide, fusion proteins containing LI, E4, or E7 (Fig.2, components marked by arrows in lanes 2, 4 and 6) could bedistinguished from the inducible tryptophan synthetase E in

noninsert-containing pATH (lane 8). Furthermore, from theirphysicochemical properties, fusion proteins in lysates frompATH with unique gene inserts were quite distinct electrophor-etically. Compared with coincidentally run standards, bandscorresponding to the LI, E4, and E7 fusion proteins demonstrated unique molecular weights (59,000, 45,000, and 42,000,respectively).

Immunological Analysis of Study Antigens

Characterization with Heterologous Immunoglobulin. Reactivity of hyperimmune sera, used below for analysis of determinants on study antigens, was first defined by immunoblot analysis. Insert-containing pATH lysate electrophoretically separated on an SDS-denaturing polyacrylamide gel was used, and

4425




Serum

Fig. 4. QuantitÃ¤ten by an enzyme-linked immunosorbent assay of study serumspecificity for a single fusion protein. Binding by serum A, measured as absorbanceat 405 nm. was >2.5-fold greater for determinants on HPV-16 LI than for thoseon E4 or on the tryptophan synthetase E (Tryp E). Similarly, binding by serumB was directed primarily to determinants on the E4 protein.

study serum A elicited by LI demonstrated a specificity largelyrestricted to a macromolecule at Mr 59,000 present in the lysate(Fig. 3, first panel, lane 1). The antigen-binding characteristicsof serum A suggested, in addition, that the reactive band contained HPV-16-specific determinants, since serum A preincu-bation with LI-containing pATH lysate markedly reduced itscapacity for band recognition (Fig. 3, panels 2 and 3, first lanes).Findings of inhibition were progressive over a range of absorbent (0-50 ng dry weight). Furthermore, the reactivity of serumA was not diminished by prior absorption with a control protein(ovalbumin) (Fig. 3, panels 4 and 5), suggesting that this post-absorptive reactivity for determinants on the M, 59,000 bandwas antigen specific. Binding to the tryptophan synthetase bandin control pATH lysate (P) was only modestly or unaffected bythese absorptions (second lane, all panels). Immunoblot analysisshowed a similar antigenic specificity for the reactivity of anantiserum elicited to E4 (data not shown).

Evidence for Unique Determinants. We next investigatedwhether immunologically unique viral determinants were present on fusion proteins isolated from the LI, E4, and E7 systems.To determine this by a methodology which was quantitative,the fusion proteins isolated by electroelution after electropho-retic separation were examined for binding with specificallyimmune sera by an enzyme-linked immunosorbent assay (Fig.4). While binding of serum A elicited to HPV-16 LI wasstrongest with the homologous macromolecule (e.g., LI) withan end point (A40i)>2.5-fold that of tryptophan synthetase E,it demonstrated minimal binding to E4 constituents (Fig. 4).Evidence for unique determinants was found also with serumB elicited to E4: only a small amount of binding was detectedto LI (Fig. 4). This together with the immunoblot absorptionexperiments (above) indicate the presence of unique, nonbac-terial components on the LI and E4 fusion proteins. Similarly,serum A could not be blocked by preincubation with controlpATH vector lysate (up to 50 Â¿ig)but was markedly inhibitedwith relatively small quantities (5 /ug) of LI-containing pATHlysate (Fig. 5, bottom). Immune sera elicited to E4 reactedsimilarly, being readily inhibited only by E4-containing pATH

lysate (Fig. 5, top).Histochemical Localization of Study Antisera. LI-specific

serum produced nuclear staining in the superficial cells of avariety of genital lesions, paralleling the site of HPV viralreplication (Fig. 6B) and assembly (Fig. 6C). Anti-E4 antiserumproduced cytoplasmic staining in the same cell population as

LI (Fig. 6D), as previously described (21). Anti-E7 antisera,elicited to a construct representing 65% of the E7 ORF, failedto produce specific immunoreactivity in biopsy material (datanot shown), despite previous reports identifying this protein byWestern blot in cancers and their derived cell lines.

HPV-specific Reactivity in Cervical Secretions

Demography/Major Diagnosis of Patients. Patient cases withan abnormal Papanicolaou smear ranged in age from 16-36years (mean, 24 years) and consisted of three with negativebiopsies, four with flat condylomata, and eight with CIN. Twopatients had a history of concurrent or previous genital warts.In all but two cases, the mucosal samples were obtained eitherat the time of biopsy or just prior to cryotherapy. The othertwo were obtained 4 and 6 months following cryotherapy andlaser treatment. Age-matched controls without abnormal Papanicolaou smears were also studied.

Presence of Viral (HPV-16) DNA. Analysis of Cone biopsymaterial performed on patient cases revealed that 0% (0 of 3),0% (0 of 4), and 87% (7 of 8) of NSE, condyloma, and CIN,respectively, were positive for HPV-16 nucleic acids by RNA-RNA in situ hybridization. Hybridization signals were observed,

< 04-0.2-

0.8 -i

0.7

O.I

Absorbent (Â¿ujprotein)

Fig. 5. Specificity of serum for the HPV-16-coded portion, rather than thetryptophan synthetase E portion, of the fusion protein, as demonstrated by antigeninhibition experiments. Bottom, a 1:50,000 dilution of serum A incubated withincreasing quantities of pATH lysate, containing (â€¢)or not containing (O) theLI protein. After centrifugation, the serum was tested by enzyme-linked immunosorbent assay for residual binding to LI. Top, similarly, a 1:50,000 dilution ofserum B, elicited to a fusion protein encoded by HPV-16 nucleotides 3399-3648,was tested for residual binding to E4. subsequent to exposure to pATH lysatecontaining (â€¢)or not containing E4 (O). The end point is absorbance at 405 nm3 h after the addition of substrate.

4426




Fig. 6. Tissue expression of HPV-I6 LI and E4open reading frames. A, hematoxylin- and eosin-stained section of a cervical precancer (C1N) containing numerous vacuolated cells (koilocytotic atypia)(arrowheads), some with enlarged or hyperchromaticnuclei (large arrowheads). Basal (b) and surface (s) celllayers are designated. B-D. higher power magnification of superficial cell layers. B, following RNA-RNAin situ hybridization for HPV-16 RNA with "S-labeled

probe. Strongly positive cells exhibit numerous silvergrains (arrowheads), viewed here by dark-field microscopy. C, following incubation with ami-1 I antibody,with characteristic nuclear staining of capsid protein(arrowheads). D, following incubation with anti-E4antibody, illustrating strong cytoplasmic staining (arrowheads) in essentially the same cell populationwhich contains HPV-16 RNA and capsid (LI) proteins. Incubation with anti-E7 antibody produced nospecific staining (not shown). Bar, 40 /iin.

using the 35S-labeled HPV-16-derived RNA probe, which were

confined principally to the superficial epithelium (similar tothat depicted in Fig. 6Ã„).In the condyloma group, 15% (3 of4) tested positive for HPV DNA other than type 16.

HPV-16-specific Binding by Immunoglobulin in Cervical Secretions. To determine by a quantitative technique whetherimmunoglobulin in cervical secretions was reactive with theHPV-16-coded proteins LI, E4, and E7, an ELISA technique,standardized for this antigen system (described above), wasused to compare binding for each fusion protein, purified byelectroelution. Cervical secretions showed reactivity to HPV-16 E4 or LI or both, with highest binding recorded in patientswith CIN (Fig. 7). Differences were statistically significant forthe HPV-16 LI and E4 antigens, when the control and CINgroups were compared (P < 0.01). A similar significant difference was found for the HPV-16 LI antigen, comparing the CIN

and the NSE groups (P< 0.05). Binding to the bacterial (vector-coded) portion of the fusion protein did not account for thesefindings, since reactivity to tryptophan synthetase E (comprising the bacterial component) was simultaneously determined,and its value subtracted in determining HPV-16 protein-specificbinding. Nonspecific binding was unlikely, given minimal reactivity of cervical mucous specimens with control protein, ÃŸ-lactoglobulin (data not shown) and with HPV-16 E7 (Fig. 7).In general, binding for E4 and LI was coincident in the samecervical specimen.

DISCUSSION

This study was designed to determine whether cervical secretions from women with HPV-associated genital precancerscontained antibodies reactive with proteins associated with

4427



HPV-I6-SPECIFIC ANTIBODY IN SECRETIONS

0.8 -

0.7-

0.6-

S 0-5

04-

0.3-

0.2-

0.1-

Group

FutlonProUIn

C NSECond.CIN

LI

C NSECondCIN

E4

C NSECondCIN

E7

Fig. 7. Binding of immunoglobulin in cervical secretions with HPV-16-codedproteins, as demonstrated by enzyme-linked immunosorbent assay. Secretionswere tested at a 1:25 dilution in triplicate in two consecutive assays. Binding tothe bacterium-coded portion of the fusion protein (tryptophan synthetase E) wasseparately assessed and, when detected, was subtracted to yield a net Am valueshown on the >' axis. A significant difference in LI and E4 reactivity was foundwith cervical secretions from CIN patients, compared with controls (P < 0.01).Bars, SD for all A*Â»values >0.1. C. controls; Cond., condylomata.

HPV type 16. The fusion proteins selected for study, e.g., thosecoded by a portion of the HPV-16 LI, E4, and E7 open readingframes, provided the opportunity to investigate reactivityagainst components of the major capsid protein (LI) of HPV-16, sharing considerable hornology with other HPV types, a"late" protein (E4) with relatively little cross-reactivity to non-type 16 HPV by immunohistochemistry, and a potential onco-protein (E7). Antisera to HPV-16 LI will cross-react withcapsid proteins of a variety of HPV types (22). However, inpopulation studies, serum responses to this region have beenrarely described (8, 9, 35). We have recently localized the HPV-16 E4 protein in tissue in a cytoplasmic distribution but haveobserved no cross-reactivity with other HPV types as of yet(21). Serological responses to this protein in a wide variety ofsubjects have been described, albeit infrequently4 (8). Localiza

tion of E7 by histochemistry has been unsuccessful, althoughthis protein has been detected by Western blot in extracts fromcervical cancers and cell lines (19-20). Serological reactivity tothis protein has been demonstrated as well (35).

The finding of local (genital) antibody reactivity to HPV-16LI and E4 determinants is intriguing and suggests that, inpatients with cervical precancers, immunoglobulins are generated locally in response to these infections and their relatedlesions. This observation would seem confirmed in a new seriesof cervical intraepithelial neoplasia patients just entering alongitudinal study, in which impressive reactivity (/f4o9> 0.50)to HPV-16 LI was found in five additional patients and intermediate reactivity (Atm = 0.2-0.5) to both HPV-16 LI and E4

4 S. A. Jenison, personal communication.

was found in three more cases. One limitation of the currentapproach centers on the volume of cervical secretions availablefor study. In this series, volumes obtainable from individualpatients permitted analysis by a quantitative and sensitive im-munoassay but were insufficient to allow Western (immuno)blot analysis. Nevertheless, standardization of study proteinconcentrations (by Lowry assay), use of purified, characterizedfusion proteins as antigens, and monitoring of reactivity againsttwo control antigens (ÃŸ-lactoglobulin,tryptophan synthetase E)were used to maximize the possibility that reactivity detectedwas due to specific binding of the HPV-16-derived proteins byimmunoglobulin present in the cervical mucous samples. Additionally, the timing of sample procurement with the menstrualcycle could not be addressed. Studies indicate that the amountof mucosal antibody produced (and the amount of cervicalsecretions available for analysis) will vary with the menstrualcycle, being at its maximum just prior to ovulation (36). Hence,the proportion of positive samples may be underestimated ifsecretions are collected at random points in the cycle, as it wasin the current study. It is possible that a bacterial proteincopurified with a study fusion protein accounted for the reactivity which we detected. This possibility is less likely because(a) control patients in our study did not react with any of thefusion proteins (an unlikely event should a common bacterialcomponent have copurified) and (b) cervical secretions frompatients with cervical intraepithelial neoplasia did not reactwith HPV E7, even though the same pATH vector (the onlylikely source of a contaminating bacterial peptide) used toconstruct the HPV-16 LI and E4 fusion proteins was used toconstruct E7.

The finding of specific antibody reactivity, particularly toHPV-16 LI and E4 determinants on fusion proteins, is ofinterest in light of recent studies of serum reactivity to theseantigens in women attending sexually transmitted disease andcolposcopy clinics. Jenison et al. (9) isolated an epitope in thecarboxyl terminal region of HPV-6b which reacted with humansera and may be relatively specific for either HPV-6 or HPV-11. Although rabbit antisera directed to HPV-16 LI proteinsreacted to the HPV-6-derived constructs, a wide variety ofhuman sera from patients with warts or abnormal Papanicolaousmears did not react to the HPV-16 LI proteins. These datawere interpreted as evidence either that HPV-16 LI did notgenerate a Serological immune response during HPV-16 infection in vivo or that the constructs used were unable to identifyan existing immune response to an LI epitope (9).

In this study, several samples of cervical secretions reactedto the fusion protein containing the carboxyl region of theHPV-16 LI protein construct. Although the finding of immunity to this region of HPV-16 is unusual, the study of populations with HPV-related precancerous lesions would likely increase the opportunity to identify immunoreactivity to this openreading frame. Moreover, Dillner et al. (37) recently producedevidence that sera from patients with cervical neoplasms preferentially react to linear epitopes from the HPV-16 LI regionvis-Ã -viscontrol populations. Their results are consistent withthe observations in this study and the hypothesis that immunoreactivity to HPV-16 LI occurs preferentially in individuals

with a genital lesion (37). Furthermore, in a new series ofpatients with cervical intraepithelial neoplasia yet to be reported, we found generous amounts of IgA by a radial immu-nodiffusion technique (Turbo RID; Binding Site Inc., SanDiego, CA), being 410 Mg/ml among 11 case samples withHPV-16 LI reactivity and 420 ng/m\ among 13 samples with

4428



HPV-I6-SPEC1FIC ANTIBODY IN SECRETIONS

HPV-16 E4 reactivity in their cervical mucus. The concentration of immunoglobulins in cervical secretions was markedlydecreased (<150 /ig/ml) at days 12-14 of the menstrual cycle,as previously suggested by Wira et al. (36). Thus, although thedata in this report suggest that this association extends to themucosal region, definitive evidence regarding this issue willrequire comparative analysis of simultaneous serum and cervical samples from the same patient and use of a marker for themucosal origin of immunoglobulin measured (e.g., secretorycomponent). Such a study would confirm the current observations concerning HPV-16-related local immune reactivity anddetermine whether there are fundamental differences in eitherthe source of reactive immunoglobulin or the HPV epitopesrecognized between serum and mucosal sites.

ACKNOWLEDGMENTS

We are indebted to Dr. Barbara J. Burkett for the statistical analysisof this study and to Diane Taylor for her expert typing of this manuscript. We wish to thank Dr. L. Gissman, E-M. deVilliers, and Dr. H.zur Hausen for providing cloned HPV DNA, as well as Dr. Steven Gofffor providing antibody to tryptophan synthetase E and T. J. Koernerfor providing pATH plasmids.

REFERENCES

1. Gissman, L., Wolnick, L., Ikenberg, H., Koldovsky, U., Schnurch, H. G.,and zur Hausen, H. Human papillomavirus type 6 and 11 DNA sequencesin genital and laryngeal papillomas and in some cervical cancers. Proc. Nati.Acad. Sci. USA, 80: 560-563, 1983.

2. Durst, M., Gissman, L., Ikenberg, H., and zur Hausen. H. A papillomavirusDNA from a cervical carcinoma and its prevalence in cancer biopsy samplesfrom different geographic regions. Proc. Nati. Acad. Sci. USA, SO: 3812-3815, 1983.

3. Crum, C. P., Ikenberg, H., Richart, R. M., and Gissman, L. Human papillomavirus type and early cervical neoplasia. N. Engl. J. Med., 310: 880-883,1984.

4. Boshart, M., Gissman, L., Ikenberg, H., Kleinheinz, A., Scheurlen, W., andZur Hausen, H. A new type of papillomavirus DNA and its presence ingenital cancer biopsies and in cell lines derived from cervical cancer. EMBOJ., J: 1151-1157, 1984.

5. Smotkin, D., Gerek, J. S., Fu, Y. S., Hacker, N. F., Major, F. J., andWettstein, F. O. Human papillomavirus deoxyribonucleic acid in adenocar-cinoma and adenosquamous carcinoma of the uterine cervix. Obstet. Gyne-co\.,68: 241-244, 1986.

6. Stoler, M. H., Walker, A. N., and Mills, S. E. Small cell neuroendocrinecarcinoma of the cervix: a human papillomavirus type 18 associated cervicalcancer (Abstract). Lab. Invest., 60: 92A, 1989.

7. Firzlaff, J. M., Hsia, C-N. L., Haibert, C. P. H. L., Jenison, S. A., andGalloway, D. A. Polyclonal antibodies to human papillomavirus type 6b andtype 16 bacterially derived fusion proteins. Cancer Cells, 5: 105-113, 1987.

8. Jenison, S. A., Firzlaff, J. M., Langenberg, A., and Galloway, D. A. Identification of immunoreactive antigens of human papillomavirus type 6b usingEscherichia co/i-expressed fusion proteins. J. Virol., 62: 2115-2123, 1988.

9. Jenison, S. A., Yu, X-P., Valentine, J. M., and Galloway, D. A. Humanantibodies react with an epitope of the human papillomavirus type 6b LIopen reading frame which is distinct from the type-common epitope. J.Virol., 65:809-818, 1989.

10. McDonnell, J. M., Mayr, A. J., and Martin, W. J. DNA of human papillomavirus type 16 in dysplastic and malignant lesions of the conjunctiva andcornea. N. Engl. J. Med., 320: 1442-1446, 1989.

11. De Villiers, E-M., Weidauer, H., Otto, H., and zur Hausen, H. Papillomavirus DNA in human tongue carcinomas. Int. J. Cancer, 36: 575-578, 1985.

12. Dehmezian, R. H., Batsakis, J. G., and Goepfert, H. In situ hybridization ofpapillomavirus DNA in head and neck squamous cell carcinomas. Arch.Otolaryngol. Head Neck Surg., 113: 819-821, 1987.

13. Moy, R. L., Eliezri, Y. D., Nuovo, G. J., Zitelli. J. A., Bennett, R. G., andSilverstein, S. J. Human papillomavirus type 16 DNA in periungual squamous cell carcinomas. JAMA, 261: 2669-2673, 1989.

14. Head, J. R.. and Billingham. R. E. Concerning the immunology of the uterus.Am. J. Reprod. Immunol. Microbiol., 10: 76-81, 1986.

15. Chipperfield, E. J., and Evans, B. A. The influence of local infection onimmunoglobulin formation in the human endocervix. Clin. Exp. Immunol.,//: 219-223, 1972.

16. O'Reilly, R. J., Lee, L., and Welch, B. G. Secretory IgA antibody responses

to Neisseria gonorrhea in the genital secretions of infected females. J. Infect.Dis., 133: 113-125, 1976.

17. Brunham, R. C., Kuo, C-C, CÃes,L., and Holmes. K. K. Correlation of hostimmune response with quantitative recovery ofChlamydia trachomatis fromthe human endocervix. Infect. Immun., 39: 1491-1494, 1983.

18. Dillner, L., Bekassy, Z., Jonsson, N., Moreno-Lopez, J., and Blomberg, J.Detection of IgA antibodies against human papillomavirus in cervical secretions from patients with cervical intraepithelial neoplasia. Int. J. Cancer, 43:36-40, 1989.

19. Smotkin, D., and Wettstein, F. The major human papillomavirus protein incervical cancers is a cytoplasmic phosphoprotein. J. Virol., 61: 1686-1689,1987.

20. Seedorf. K., Oltersdorf, T., Krammer, G., and Rowenkamp, W. Identificationof early proteins of the human papillomaviruses type 16 (HPV 16) and 18(HPV 18) in cervical carcinoma cells. EMBO J.. 6: 139-144, 1987.

21. Crum, C. P., Barber, S., Symbula, M., Snyder, K., Salen, A. M., and Roche,J. K. Co-expression of the human papillomavirus type 16 E4 and LI openreading frames in early cervical neoplasia. Virology, in press, 1990.

22. Firzlaff, J. M., Kiviat, N. B., Beckmann, A. M., Jenison, S. A., and Galloway,D. A. Detection of human papillomavirus capsid antigens in various squamous epithelial lesions using antibodies directed against the LI and L2 openreading frames. Virology, 164:467-477, 1988.

23. Yanofsky. C., Plait, T., Crawford, I. P., Nichols, B. P., Christie, G. E.,Horowitz, H., Van Cleemput, M., and Wu, A. M. The complete nucleotidesequence of the tryptophan operon of Escherichia coli. Nucleic Acids Res.,24:6647-6668, 1981.

24. Seedorf, K., Krammer, G., Durst, M., Suhai, S., and Rowenkamp, W. Humanpapillomavirus type 16 DNA sequence. Virology, 145: 181-185, 1985.

25. Roche, J. K., Fiocchi, C., and Youngman, K. Sensitization to epithelialantigens in chronic mucosal inflammatory disease. Characterization of human intestinal mucosa-derived mononuclear cells reactive with purified epithelial cell-associated components in vitro. J. Clin. Invest., 75: 522-530,1985.

26. Hunkapiller. M. W., Lujan, E.. Ostrander. F., and Houd, L. E. Isolation ofmicrogram quantities of proteins from polyacrylamide gels for amino acidsequence analysis. Methods Enzymol., 91: 227, 1983.

27. Lowry, O. H., Rosebrough, W. J., Fair, A. L., and Randall, R. J. Proteinmeasurement with the Polin phenol reagent. J. Biol. Chem., 193: 265-275,1951.

28. Laemmli, U. K. Cleavage of structural proteins during the assembly of thehead of bacteriophage. Nature (Lond.), 227:680-685, 1970.

29. Merrill, C. R., Switzer, R. C., and Van Keuren, M. L. Trace polypeptides incellular extracts and human body fluids detected by two dimensional electro-phoresis and highly sensitive silver stain. Proc. Nati. Acad. Sci. USA, 76:4335-4339, 1979.

30. Paul, J. W., Kotz, K. W., and Roche, J. K. Sensitization to epithelial antigensin chronic mucosal inflammatory disease: immunochemical similarity ofmacromolecules in human colonie epithelium with murine intestinal epithelial cell-associated components (ECAC). Immunology, 65: 551-557, 1988.

31. Gershoni, J. M., and Palade, G. E. Review: protein blotting; principles andapplications. Anal. Minchcm.. 131: 1-15, 1983.

32. Crum, C. P., Mitao, M., Levine, R. U., and Silverstein, S. J. Cervicalpapillomaviruses segregate within morphologically distinct precancerous lesions. J. Virol., 54: 675-681, 1985.

33. Crum, C. P., Nuovo, G., Friedman. D., and Silverstein, S. J. A comparisonof biotin and isotope-labeled ribonucleic acid probes for in-situ detection ofHPV-16 ribonucleic acids in genital precancers. Lab. Invest., 58: 354-359,1988.

34. Steele, R. G., and Torrie, J. H. Non-parametric statistics. In: Principles andProcedures of Statistics: A Biometrica] Approach, Ed. 2, pp. 544-545. NewYork: McGraw-Hill, 1980.

35. Jenison, S. A., Yu, X-P., Valentine, J. M.. Koutsky, L. A., Christiansen.A. E., Beckmann, A. M., and Galloway, D. A. Evidence of prevalentgenital-type human papillomavirus infections in adults and children. J.Infect. Dis.. 162: 60-69, 1990.

36. Wira, C. R., Sullivan, D. A., and Sandoe, C. P. Estrogen-mediated controlof the secretory immune system in the uterus of the rat. Ann. NY Acad. Sci.,409:534-551, 1983.

37. Dillner, J., Dillner, L., Utter, G., Eklund, C., Rotola, A., Costa, S., andDiluca, D. Mapping of linear epitopes of human papillomavirus type 16: theLI and L2 open reading frames. Int. J. Cancer. 45: 529-535, 1990.

4429



1991;51:4423-4429. Cancer Res Kimberly A. Snyder, Shannon R. Barber, Millie Symbula, et al. Cervical Diseaseand E4 in Cervical Secretions of Women with HPV-related Binding by Immunoglobulin to the HPV-16-derived Proteins L1

Updated version

http://cancerres.aacrjournals.org/content/51/16/4423

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/51/16/4423To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Binding by Immunoglobulin to the HPV-16 - Cancer Research