[CANCER RESEARCH 52, 4009-4016, July 15. 1992]

HLA Class II Antigen Expression in Human Papillomavirus-associated CervicalCancer1

Susan S. G lew,2 Margaret Duggan-Keen, Teresa Cabrera, and Peter L. Stern3

Cancer Research Campaign Department of Immunology, Paterson Institute for Cancer Research, Christie Hospital, Manchester, M20 9BX, United Kingdom[S. S. G., M. D-K., P. L. S.], and Servicio De Analisis Clínicos,Hospital Virgen De Las Nieves, Avenida Coronel Muño:2, 18014 Granada, Spain [T. C.¡

ABSTRACT

The observation that tumor cells of some neoplasms display majorhistocompatibility complex (MHC) class II molecules may be of functional significance, influencing the progression of malignancy by allowing the cancer cells to present antigen to the immune system. In thenormal cervix, class II molecules are expressed by columnar but notsquamous epithelium. The pattern of MHC class II expression in cervical carcinomas has been documented using immunohistochemicalmethods. Of 53 cervical squamous carcinomas examined for MHC classII expression, only 17%maintained a negative phenotype characteristicof the epithelium from which they were derived, while the remainingtumors exhibited either uniform (45%) or heterogeneous (38%) expression. Tumor areas which were class II positive also express class IIassociated invariant chain and the adhesion molecules lymphocyte function antigen 3 and intercellular adhesion molecule 1. The DR, DP, andDQ class II MHC subitici are differentially expressed, suggesting independent regulation. There is a trend for tumors with the uniform classII phenotype to predominantly express DR antigen, whereas tumors ofthe heterogeneous class II phenotype express with equal frequency either DR or DP antigens dominant)}'. There is no apparent influence of

class II status on lymphocyte infiltrâtinn of the tumors. The presence ofhuman papillomavirus 16 DNA in the cervical carcinoma specimens wasanalyzed by Southern blotting of restriction enzyme digested DNA andno correlation between the presence of human papilloma virus andMHC class II expression was found.

INTRODUCTION

Cervical cancer is estimated to be the second most commonfemale cancer with an estimated 500,000 new cases per annumworld wide (1). Sexually transmitted infections are recognizedas one of the major risk factors and the active agents are believed to be specific types of HPV4 (2-4). HPV 16 is the type

most frequently detected in cervical carcinomas. However, it isclear that HPV infection alone is insufficient for cancer development and that the pathogenesis of cervical carcinoma is mul-tifactorial (3). In this context, the immunogenicity of HPVinfected and/or transformed cells and the role of the immuneresponse may be relevant to the development or progression ofcervical malignancy (5).

The major histocompatibility complex class I and class IIproducts are central to the functioning immune response byacting to restrict T-cell recognition of foreign peptide antigens.The HLA-A, -B, and -C loci encode the MHC class I polymorphic MT 45,000 heavy chains which associate with an invariantA/r 12,000 02- microglobulin chain. They are expressed by most

Received 1/27/92; accepted 5/7/92.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 in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by the Cancer Research Campaign and a British-Spanish Joint Research Programme Grant, ref/2398.

2 Supported by a Joseph Starkey Clinical Research Fellowship.3 To whom requests for reprints should be addressed.4 The abbreviations used are: HPV. human papillomaviruses; LFA, lymphocyte

function antigen; ICAM, intercellular adhesion molecule: MHC. major histocom-palibility antigen; SCC, squamous cell carcinomas; li, invariant chain; itiAb. monoclonal antibody.

somatic cells and present endogenous antigens as processedpeptides to CDS cytotoxic T-cells (6). The HLA- DR, -DP, and-DQ loci encode the MHC class II heterodimeric a (A/r 32,000)

and ß(A/r 28,000) chain complexes and are expressed primarilyby antigen presenting cells but also by some epithelia. They canpresent processed peptides from both exogenous (7) and endogenous antigens (8) to T-helper cells. This process is initiated inthe endosóme where class II molecules are intracellularly targeted by an associated li. Here, in the low pH, li is released andthis allows the peptides to interact with the a and ßchaincomplex of the class II molecules; these complexes are subsequently transported to the cell surface (9). It is likely thatchanges in either MHC class I or II expression may significantly influence tumor cell evasion of immune surveillance. Inview of the association between carcinoma of the cervix andHPV infection, changes in expression of MHC class II molecules in the tumors might have important implications for presentation of viral target antigens to the immune system.

A previous study has established that a significant proportionof cervical carcinomas show down-regulation in MHC class Iexpression (10). Altered MHC class II expression has been welldocumented in malignant melanomas (11) and hepatocellular(12), gastric (13), colorectal (14), and laryngeal carcinomas(15). The relationship to prognosis varies with the tumor typeand can be favorable (15), disadvantageous (11), or of no apparent significance (14). Since a HLA class II positive tumorcan apparently be derived from tissue normally negative forclass II, we have examined the class II MHC status of malignant lesions of the cervix. The cellular adhesion moleculesICAM-1 and LFA-3 are known to be involved in the mechanisms of antigen presentation by both HLA class I and HLAclass II molecules (16); therefore the tumor expression of thesemolecules was also examined. The possibility that HPV infections associated with cervical carcinomas might influence MHCexpression was investigated by Southern blotting for HPV 16DNA in the tumors. Our results indicate that the majority ofthe squamous tumors are class II positive in contrast to thenormal epithelium of the cervix. There is no apparent correlation between MHC class II phenotype and the presence of HPVtype 16 DNA.

MATERIALS AND METHODS

Clinical Material. Normal cervix specimens were obtained from patients undergoing total abdominal hysterectomy for benign disease ofthe corpus uteri with no evidence of prior cervical cytological abnormality. Biopsies of primary cervical carcinomas were obtained frompatients undergoing disease staging under anesthesia prior to radiotherapy at the Christie Hospital, Manchester, United Kingdom. Followingexcision, all specimens were immediately divided, with one portionbeing fixed in formalin for routine processing and histológica!analysisand the other snap frozen in liquid nitrogen and stored at —¿�70°C.

Pathology. 12 control cervix specimens (patient age 27-42 years;mean, 34 years) were confirmed normal by histopathological

4009

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

examination. Tumor type and grade were assessed according to established criteria in hematoxylin-eosin stained sections (17). In addition,periodic acid-Schiff (with or without diastase treatment)-Alcian bluestaining was used to assign tumors with mucin production and squa-mous morphology to the adenosquamous category ( 17). Seventy tumorswere categorized as follows: 53 (76%) squamous cell carcinomas(patient age 29-79 years; mean 55 years); 11 (16%) adenosquamouscarcinomas (patient age 33-73 years; mean 55 years); 5 (7%) adeno-carcinomas (patient age 30-50 years; mean 43 years); 1 tumor waspoorly differentiated and unclassifiable.

Immunohistochemistry. Cryostat sections (7 /¿m)of cervical biopsies were thawed, fixed in acetone, and then labeled in a two stepimmunoperoxidase procedure with a primary mouse antibody andDako (Copenhagen, Denmark) rabbit anti-mouse immunoglobulin-peroxidase. Negative (no first layer antibody) and positive (LP34 cy-tokeratin specific mAb or W6/32 recognizing MHC class I molecules)controls were included for each specimen. In addition, MHC class IIexpression of the leukocytes infiltrating the stroma served as an internalcontrol in each section. The location of the epithelial tumor cells wasassessed by labeling with LP34 recognizing cytokeratins 6 and 18 whichare expressed by both normal and neoplastic epithelial cells (18). Thefollowing murine mAbs were used to investigate the expression ofMHC class II molecules by normal and malignant cervical epithelialcells: CR3/43 (Dako) recognizing the ßchain of HLA-DR, DP, and DQmolecules (19). The sublocus specific antibodies DA6.164 (20) (giftfrom Dr. K. Guy, Edinburgh, United Kingdom), B7/2.1 (21), Tu22(22), and LN2 (23) (Janssen, Beerse, Belgium) recognizing HLA-DR,-DP, and -DQ and the class II associated li molecules, respectively.ICAM-1 expression was detected with mAb RR1/1 (24) and LFA-3 bymAb TS2/9 (25), both gifts from Dr. T. Springer (Boston, MA).

Analysis of Antigen Expression. Location of the epithelial tumorcells was confirmed by staining with mAb LP34. Specimens labeledwith anti-class II, LFA-3, or ICAM-1 antibodies were classified asuniformly positive (+), heterogeneous (+/—),and uniformly negative(—).The number of +'s is proportional to intensity. The relative dom

inance of expression of the class II sublocus antigenic products wasassessed by area and intensity of the labeling in sequential sections.

Detection of HPV DNA. The presence of HPV 16 DNA in 60 of 70of the cervical carcinoma specimens was analyzed by Southern blottingof restriction enzyme digested DNA as described previously (10). TenMgof BamHl digested DNA from each specimen were Southern blottedto Hybond N membrane (Amersham, Little Chalfont, United Kingdom) and hybridized with the labeled HPV 16 probe prepared fromHPV DNA cloned into pAT153. Sensitivity, assessed using Caski cellDNA, was usually <1 viral genome copy/cell when 10 Mgof DNA wereanalyzed. Specimens 2, 18, 24, 46, 62, 64, and 75, where no HPV 16DNA was detected, had respectively 4, 4, 5, 7, 3, 5, and 5 Mgof DNAloaded and therefore a corresponding reduction in sensitivity.

Analysis of Tumor Infiltrating Lymphocytes. Six specimens of eachdifferent class II phenotype were analyzed with reagents recognizingseveral markers of lymphocyte subsets by counting ten fields at xSOO.All reagents were from Dako: anti-CD45, a pan leukocyte marker;anti-CD3, a pan T-cell marker; anti-CD4 and anti-CD8, helper andcytotoxic T-cell markers, respectively; and anti-CD19, a pan B-cell

marker.

RESULTS

Cervical Carcinomas Express MHC Class II Antigens. Thesquamous epithelium of the 12 normal cervix specimens didnot express MHC class II molecules but infiltrating leukocytesand Langerhans cells are strongly labeled with CR3/43 mAb(Fig. 1). Columnar epithelium was present in 9 of these biopsiesand always stained positively for HLA-DR, -DP, and -DQ(not shown). Of the tumor biopsies examined 76, 16, and7% were squamous, adenosquamous, and adenocarcinomas, respectively. Overall 83% of these tumors showed evidence ofMHC class II expression. The pattern of labeling of these spec-

Fig. 1. Immunohistochemistry of normal cervical squamous epithelium illustrating the absence of MHC class II expression, a, no labeling seen in negativecontrol; b, MHC class II expression detected by mAb CR3/43 only on infiltratingleukocytes and Langerhans cells; c, epithelial cells labeled by mAb LP34. Bar, 100urn.

imens could be subdivided uniformly (47%) or heterogeneously(36%) positive and negative (17%). These patterns were represented in each type of tumor pathology. Squamous carcinomasof homogeneous, heterogeneous, and negative class II pheno-types were in the proportion 46, 37, and 17%, respectively.Examples of the three patterns of class II expression by squamous cervical carcinomas are illustrated in Fig. 2 and 3.

4010

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MIIC CLASS [I ANTIGENS

«J.

\

**

. *.

Fig. 2. Patterns of M HC class II expression in squamous carcinomas of the cervix. Examples of a tumor helerogeneously positive (a-c) and negative (</-/) for MHCclass II antigens are shown; a, b, c, and d, e. and/are labeled, respectively, with negative control, for MHC class II with CR3/43, and for epithelial cytokeratins withLP34. There is no significant difference between the control (</)and class II labeling (e) except for some class II positive infiltrating cells and the latter are in markedcontrast to the positive control (/). The specimen exhibiting heterogeneous class II labeling (ft) is clearly different from the control (a). Bar. 100 turn.

The data in Table 1 summarize the clinical details of thetumors examined for HLA expression. The class I expression ofthe tumors is also documented. There were two tumors exhibiting complete and three exhibiting heterogeneous MHC class Iloss of expression in this series but there is no systematic correlation with any class II phenotype. Table 2 shows that there isno difference in the proportions of the different class II pheno-types of the squamous carcinomas compared with all the tumors and no significant association between the MHC class IIphenotype and the stage or differentiation of the tumor.

401

HPV 16 DNA Detection and Class II Expression. The relationship between class II phenotype and the presence or absenceof HPV 16 is shown in Table 1 and summarized in Table 2. Theproportion of specimens positive for HPV 16 was 66.7%. Allclass II phenotypes are represented in both the HPV 16 negative and positive tumors. There appears to be an increasedfrequency of uniform class II positive tumors in the HPV negative category and a corresponding decrease in the frequency ofheterogeneous class II labeling tumors. However, this is notstatistically significant for either all or the SCC alone.

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

Fig. 3. A homogeneously MHC class II and ICAM-1 positive squamous carcinoma of the cervix. Immunohistochemistry with (a) negative control; (b) for MHC classII with CR3/43; (c) for ICAM-1 with RR1/1; and (</)for epithelial cytokeratins with LP34. Bar, 100 /im.

Expression of MHC Class II Subloci HLA-DR, -DP, and-DQ by Cervical Carcinomas. Some of the tumors with eitheruniform or heterogeneous MHC class II expression as measured by CR3/43 reactivity were analyzed for expression of theindividual HLA sublocus products. The relative areas of staining by the sublocus specific reagents in a particular specimenwere ascertained and this measure was used to rank any dominance of expression by HLA-DR, -DP, and -DQ. The relativeintensity of staining was a second factor used to allocate dominance where there were equal areas labeled. In some specimensthere is evidence of preferential sublocus expression. In thehomogeneously labeled group of tumors HLA-DR is most frequently dominant and in the heterogeneous group there is anequal preference for HLA-DP to be dominant. DQ expressionis never dominant. Fifty-seven % of the uniform but only 5% ofthe heterogeneously labeled group of specimens show no dominance of any sublocus expression. These results indicate thatthe individual sublocus products can be independently regulatedin these carcinomas. Table 3 summarizes the results and showsthat there is a significant difference in HLA sublocus dominance between uniformly and heterogeneously labeled tumors.Codominance of HLA-DR and -DP is also significantly lessthan individual dominance for heterogeneously labeled tumors(P = 0.005).

Expression of Invariant Chain. Five specimens from each ofthe three phenotypes of class II expression in cervical squamouscarcinomas were analyzed for invariant chain expression. In allcases, apart from one the pattern of li staining was essentiallycongruent with the pan class II labeling of an adjacent tissue

section (data not shown). This suggests that the class II expression is potentially functional since invariant chain expressionwould be expected to deliver class II molecules to the intracel-lular compartment where processed peptides can bind to theclass II molecules.

Expression of ICAM-1 and LFA-3. ICAM-1 expression wasexamined on 4 specimens of normal cervical epithelium and 28squamous carcinomas representative of different class II phenotypes. Normal cervical squamous epithelium did not label forICAM-1, although endothelial cells in the lymphatics of thecervical stroma were positive. Six of ten class II negative carcinomas showed no ICAM-1 expression, three exhibited heterogeneous labeling (in one case the periphery of tumor onlylabeled), and the other was homogeneously ICAM-1 positive.Often specimens selected as examples of HLA class II homogeneously positive carcinomas, nine were labeled homogeneously for ICAM-1 (e.g., Fig. 3) although in a single case therewas considerable variation in intensity throughout the sections.The other specimen showed no expression of ICAM-1 despitestrong uniform staining for HLA class II antigens on an adjacent tissue section. Six of eight tumors with class II heterogeneous phenotype showed a pattern of ICAM-1 expressionwhich was essentially congruent with the HLA class II stainingin adjacent tissue sections. One tumor expressed ICAM-1 in ahomogeneous pattern and another specimen exhibited anICAM-1 negative phenotype. In contrast with ICAM-1, LFA-3was expressed on normal squamous epithelium and all 14 carcinomas tested, including representatives of all the differentclass II phenotypes.

4012

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

Table l HL4 expression in cervical cancer in relation to clinical parameters ana IIPl I fi detectionThe sublocus specific reagent DA6.164 is known to not detect the HLA-DR7 alíele(20). Differential reactivity to individual allelic MHC class II products by

the locus specific reagents may contribute to the differences in labeling patterns of some individual tumors with the pan class II reagent CR.1/4.1 and the former.

No.Adenocarcinonias13.16.175153Adenosquamous

carcinomas12162529.14434856596174Squamous

carcinomas12346791011141517IS192021222.124262728.1132333538.1940414245464749505254555758606263646566677071727375Unclassifiablecarcinoma30Age

(yr)504334443033505937347.171444842666251495147383231776664596765477764494159595236455476673669553664485553344853295757714766637.17074406070554077StageIBIBIBIIBIBIBIBIBIIBIBIIBIIIA1MB1MBIBIBIBIIBIB1IIBIBIIBIBIBIBHA1MBIIIAIVBIBIIBIVHA1MBHAHAHIIBIIBIIBI1AHAIBIBIIB1MBHIIBIBIBIBIIBIBIVIIBIBIVIB1MBIIB1MBIBHIIBIBIBIBHIIBIIBTyp«Pan IIDRM

++++W+++NT"P

NTW+/—+/-P+++P

--P+++++P

++/-*M++++P+/-++/—MW

++++/-M+/-+/-P+++/-P+/—+/-M+4-++M

++++/—M++/-+/-MM

++/-+/-M++/-+/-P

+++++M++/-+/-M

++++P+++++M+++++pM

+++++SC++++P+/—+/-M

NTM+++/-+/-P

NTW+/-+/-W

NTM+++++W+++++M+++++P+++/-P++NTP

NTP+++NTP

+/-M+/-+/-M+/-+/-M

+/-P+++SC++++W

++/-+/-P++++W+++/-M+/-+/-M

+/-P+/-+/-M+/-+/-P

++M+MNTP+/-MNTP

+++++cp4--Õ-4-o\~W++/-++/-W++/-+/-W

+++/-M+/-+/-P+++NTW+++++MP

++ ++DP

DQ Pan IHPV+

++++NTNT ++++NTNT+++/-++ +++/-+/- +++NT+++

+++++ ++++++

NT+++++/-

+++++NT++

+/-++++/—+— ++++/—+++NT

++++-I-/-+/- ++++/-

++NT+/-++++

++/—+/— +++++/—+/— ++++++/-+/- ++NT+/—

++++++/- ++++++ ++++

++-1-NT+NT++

++ +++++/—++ +++—¿�+/—+/- +++NTNT +++++/—+/-+NTNT +++++

+NTNT+++/-+/- +++++

+ +++++/-—¿� +++++/-

++NTNT +++NTNT +++NTNT +++++/-

+/-++/—+/—++/—+/— +/—-t-+/—

++/—++++++

++ ++++/—+/- +++—¿�++

++++/—H—¿� +++/—-H/— ++—¿�+

+++++/—+/— ++-f+•*-/—+/— ++++-t-

+ +++-t-+++NT

NT ++++/——¿� ++—¿�NT

NT +++++++++—¿�++

++ +++NT+/-+++NT+/-+++NT++

+/- +++NT+++NTNT

NT +++NT+++ +++—¿�++++

+/- +++°NT, not tested.bOnly 5% tumor cells positive.

4013

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

Table 2 MHC class II phenolype of cervical cancers in relation lo HPV 16, disease stage and differentiation

TumorphenotypcExpression

of MHC Class II antigens in cervical cancer(%)AllClass II (N =70)SCC

Class II (A153)Human

papillomavirus type 16 (66.7% tested positive of alltumors)ALLHPV positives (A/ =38)SCCHPV positives (N =31)ALLHPV negatives (N =19)SCCHPV negatives (N =12)Disease

stage alltumorsStageI (A/ =33)StageII (A' =19)StageIII(A'=14)StageIV (N =4)Differentiation

alltumorsWell(N=12)Moderate

(A1=29)Poor(A7= 29)Homogeneous

(+)47.245.336.8»38.763.2°50.048.542.150.050.050.034.558.6Heterogeneous(+/-)35.737.744.7°41.921.1«33.336.426.342.950.041.744.824.1Negative(-)17.217.018.519.415.716.715.131.67.1—8.320.717.2PStatisticNot

significantNot

significantNot

significantStage

1 vs.2:Stage2 vs.3:Not

significantNot

significant

There are no differences between the proportions of all tumors of the different MHC class II phenotypes and the SCC. There is a trend for HPV 16 negative tumorsto show an increased frequency of homogeneously (uniform) class II expressing tumors with a corresponding decrease in the proportion of heterogeneous class II labelingtumors. The HPV positive tumors appear to show a small increase in the heterogeneously labeled phenotype frequency and slight reduction in the uniform labeledproportion. These differences are not significant.

There are no significant associations with any class II phenotype with either stage or degree of differentiation of the tumor. Statistical analysis was by either x2 or Fisher's

exact test.•¿�Fisher's exact test, P = 0.07.

Table 3 Differences in HLA class II sublocus expression dominance betweenhomogeneous and heterogeneous labeling tumors

There are no significant differences in the proportions of carcinomas exhibiting cither uniform or heterogeneous labeling for the class II sublocus phenotypes

Class II phenotype No. HLA DR HLA DP HLA DQ

HLA class II sublocusexpression incervicalsquamous

carcinomaUniformHeterogeneous212020(95%)°16

(80%)"21(100%)»17(85%)»13(62%)»9

(45%)»

Class II dominance No. HLA-DR HLA DP HLA DQ

HLA class II sublocusdominance in cervicalsquamous carcinoma

Uniform (21)Heterogeneous (20)

12(57%)» 7(35%)» 2(10%)*1 (5%)» 9 (45%)» 10 (50%)»

0 (0%)0 (0%)

»x2 test, P = 0.97. However, there are significant differences in the proportionsof tumors either class II positive phenotype and the pattern of sublocus dominance.

* Fisher's exact test. P = 0.003.

Tumor Infiltrating Lymphocytes. There was no apparent difference in the leukocyte infiltration in 6 representatives of eachof the different tumor class II phenotypes as judged by countingthe number of CD45 positive cells in ten fields of tumor associated stroma at xSOO. However, the range between specimenswas very variable. The estimated proportion of these leukocyteswhich were CD3 positive T-cells was in the range of 22-75%.The ratio of CD4/CD8 T subsets was not different betweenclass II phenotypes and averaged 1.7 ±0.19 (SD) in the tumorspecimens which is slightly less than that of normal peripheralblood. Where leukocytes were detected within the tumor per se,they were predominantly CDS rather than CD4 T-cells. Therewere very few B-cells in the specimens and no differences evident among the various phenotypes.

DISCUSSION

HLA (MHC) class II molecules are immunoregulatory molecules found predominantly in those cells of the immune systemconcerned with antigen presentation, although they are alsoexpressed on a limited range of epithelia (26), and expression in

a wider range of cell types is inducible by cytokines (27). Ourfinding of the lack of HLA class II expression by squamousepithelium compared to the positive columnar epithelium ofnormal cervix is in keeping with previously published results(28, 29). In contrast to normal squamous cervical epithelium,the data presented here demonstrate that the vast majority(83%) of squamous cervical carcinomas express HLA class IIantigens in all or a proportion of the tumor cells. Althoughtumors in the adeno- and adenosquamous carcinoma groups

may be glandular in origin, approximately the same distributionof class II phenotypes is found as for the SCC. Most SCC arethought to develop from the transformation zone of the cervix,an everted area of ectocervical columnar epithelium which subsequently undergoes the physiological process of squamousmetaplasia (30) during which class II expression is presumablyregulated. It is possible that the distribution of tumor class IIphenotypes is the result of selection and fixation at the stage atwhich transformation takes place and/or a subsequent regulation in class II expression. Adenocarcinoma //; situ arises in thecolumnar cells of the endocervix and is thought to be the precursor lesion of at least a proportion of adenocarcinomas (31).The fact that the patterns of class II expression are similar forsquamous and adeno/adenosquamous tumors could be the result of common features in the origins and/or natural history ofthese different tumor pathological types.

As the DNA of HPV type 16 is found in a large proportion ofcervical carcinomas and there is strong in vitro evidence fortheir involvement in the neoplastic process (4), an associationbetween HPV 16 DNA and HLA class II expression wassought. However, our study could find no correlation betweentumor HLA class II expression and the presence of HPV 16DNA as detected by Southern blot analysis. This does not exclude a relationship between HPV gene function and class IIexpression. For example in situ hybridization could be used todetect any relationship between specific viral transcription andMHC class II expression in cervical tumors. The pathogenesisof cervical cancer is regarded as a multifactorial, multistageprocess and other factors, both infectious and otherwise, are

4014

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

likely to be involved and may also influence MHC expressionand immune recognition of such neoplasms.

It is interesting that tumor areas expressing class II determinants usually label positively for the cellular adhesion moleculeICAM-1 and class II associated invariant chain. The congru-

ency of expression usually seen in these three classes of molecule could suggest some common regulatory factor. ICAM-1, aA/r 90,000 surface glycoprotein member of the immunoglobulinsuperfamily, is not present in normal cervical squamous epithelium. It promotes adhesion in immunological interactions byacting as the ligand for the integrin molecule LFA-1, expressedon T-lymphocytes (32). LFA-1 is required for a wide range ofleukocyte functions including T-cell-mediated killing, T-helper,and B-lymphocyte responses (16). In comparison to ICAM-1,LFA3 has a wider distribution and is found in both normal andneoplastic cervical epithelium. It acts as the ligand for the T-cellspecific LFA-2 (CD2) antigen (25). Both the LFA-1/targetICAM-1 pathway and the LFA-2/target LFA-3 pathways are

required for optimal effector/target conjugation, in this caseeffective T-helper or cytotoxic T-cell/tumor cell conjugation(17). The role of invariant chain is believed to control the in-tracellular transport of class II MHC molecules from the en-doplasmic reticulum, through the Golgi apparatus into an en-dosomal compartment where it is suggested class II moleculesmeet endocytosed exogenous antigens (9). In cervical carcinomas, li expression usually accompanies HLA class II expression, which is in contrast to colorectal carcinomas where thereis frequently noncoordinate expression of HLA-DR antigensand li (33).

It seems likely that cervical carcinomas could present antigensince they express the necessary immune function related molecules. If this is the case then which, if any, peptides might bepresented? As discussed above, either exogenously or endoge-nously derived peptides might theoretically be involved. Thegenes coding for the HPV oncoproteins E6 and E7 are the only-

regions of the viral genome that are consistently retained andexpressed in cervical carcinomas (34). It is possible thereforethat the class II expression by cervical carcinomas might lead topresentation of HPV derived peptides. There is evidence ofspecific anti-E6 and E7 serological responses in cervical cancer(35)5 which might reflect specific helper T-cell function elicited

by carcinoma presentation. We are currently investigatingwhether there is any correlation between antibodies to HPVand the class II phenotype of the tumor in cervical cancer patients.

Our study found that for those squamous carcinomas whichexhibited uniform class II expression, the order of sublocusexpression is generally DR, DP, and DQ. This is in keepingwith other solid tumors (12, 13, 36) and is suggestive of independent regulation of the sublocus genes. Interestingly, cervicaltumors with a heterogeneous pattern of HLA class II expression are more likely to rank DP, DR, DQ than those with auniform pattern but the significance of this is uncertain. Littleis known about the differing immunoregulatory functions of theMHC class II subloci products. The main restriction determinants for antigen presentation appear to be associated withHLA-DR and not with -DP and -DQ (37). HLA-DQ is expressed at low levels on antigen presenting cells (38) and whilenever dominant!) expressed in preference to -DR and -DP, itwas present in 54% of cervical cancers. It has been shown that

HLA-DQ molecules may stimulate immunosuppressionthrough restriction of CDS suppressor T-cells (39). In both ourstudy and others (28) the majority of infiltrating T-cells of thetumor itself are CDS positive; however, a specific relationshipwith HLA-DQ expression was not investigated. Since there is

no obvious relationship between the degree of leukocyte infiltration in the tumor associated stroma and class II expression,the latter is unlikely to be only the result of cytokine induction.However, is possible that cytokines are released from responding T-cells which can induce or enhance tumor cell class II

expression (11).We have shown6 that cervical squamous epithelia with his-

tological evidence of HPV infection or cervical intraepithelialneoplasia also express class II antigens. Thus MHC phenotypemay influence the likelihood of regression or progression ofthese preinvasive lesions. Whether class II expression renderscervical cancer cells more immunogenic and to what extentclass Il-positive neoplastic cells can present viral or tumor associated antigen are unknown. Animal experiments have shownthat transfection of HLA class II genes can completely abrogatemalignancy in otherwise lethal tumor cells (40). In human neoplasia, the significance of class II expression appears peculiar toeach tumor type. For example it seems to bestow a favorableprognosis in breast (41) and laryngeal (15) carcinomas, has noprognostic significance in colorectal carcinoma (14), and is associated with tumor recurrence and metastasis in malignantmelanomas (11). It may be highly relevant that antigen presentation by human keratinocytes, derived from the cervix, caninduce tolerance in HLA-DR restricted T-cells (42). A longterm follow-up study of our cervical carcinoma patients is underway to assess whether tumor MHC class II expression has anyprognostic value in this disease.

Finally, it must be remembered that HLA class II antigens oncervical cancer cells may have no antigen presentation functionand their expression may reflect the internal state of differentiation and metabolic activity of the cells. HLA class II molecules are unlikely to be differentiation antigens/wse, inasmuchas they are absent in normal proliferating cervical squamousepithelium and there is no pattern of expression consistent withthis in the different grades of differentiation of the tumors.MHC class II molecules could participate in other immuneinteractions, such as directing the traffic of lymphoid cells (43).

ACKNOWLEDGMENTS

We are extremely grateful to Dr. R. D. Hunter for allowing us toexamine patients under his care. Dr. R. Harris and Dr. H. Buckley forexpert pathology. Dr. M. Bromley and G. Ashton for aspects of theimmunohistochemistry. and Dr. M. Connor and Dr. J. R. Arrand forhelpful advice and materials.

REFERENCES

1. Parkin, D. M.. Laara. E.. and Muir, C. S. Estimates of the world frequencyof sixteen major cancers in 1980. Int. J. Cancer, 41: 184-197, 1980.

2. McCance, D. J. Human papillomavirus (HPV) infections in the aetiology ofcervical cancer. Cancer Surv., 7: 499-506, 1988.

3. Kitchner. H. C. Does HPV cause cervical cancer? Br. J. Obstet. Gynecol., 75:1089-1091. 1968.

4. Vousden. K. H. Human papillomaviruses and cervical carcinoma. CancerCells. /: 43-50. 1989.

5. Moore. M. Natural immunity to tumors—theoretical prediction and biological observations. Br. J. Cancer, 52: 147-151, 1985.

6. Bjorkman. P. J.. Saper, M. A., Samraoui. B., Bennet, W. S., Strominger, J.

5 A. K. Ghosh, N. Smith, S. Stacey, M. Connor. J. R. Arrand, and P. I. Stern, 6 S. S. Glew, M. Connor, H. Buckley, C. Stanbridge. J. M. M. Walboomers, F.manuscript in preparation. Cromme, C. J. L. M. Meijer, manuscript in preparation.

4015

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CERVICAL CARCINOMAS EXPRESS MHC CLASS II ANTIGENS

L., and Wiley, D. C".The foreign antigen binding site and T cell recognition 25.

regions of class I histocompatibility antigens. Nature (Lond.), 329: 506-512,

1987.7. Fleisher. B.. Bechi. H., and Roti. R. Recognition of viral antigens by human

influenza A virus-specific T lymphocyte clones. J. Immunol.. l3Si 2800- 26.

2804. 1985.8. Nüchtern.J. G.. Biddison, W. E.. and Klausner. R. D. Class II MHC' mol

ecules can use the endogenous pathway of antigen presentation. Nature 27.(Lond.). 343: 74-76. 1990.

9. Hammcrling, G. J., and Moreno, J. Function of invariant chain in antigenpresentation by MHC class II molecules. Immunol. Today. //: .137-340, 28.

1990.10. Connor. M. E., and Stern, P. L. Loss of MHC class I expression in cervical

carcinomas. Int. J. Cancer. 46: 1029-1034. 1990.11. Ruiter. D. J.. Brocker. E. B., and Ferrane. S. Expression and susceptibility to 29.

modulation by interféronsof HLA class I and II antigens on melanoma cells:immunohistochemical analysis and clinical relevance. J. Immunogenet.. 13:229-234, 1986. 30.

12. Patcrson. A. C, Sciot, R.. Kew, M. C., Callea, F. Duskeiko, G. M., andDesmet, V. G. HLA expression in human hepatocellular carcinoma. Br. J. 31.Cancer, 57: 369-373, 1988.

13. Sakai. K.. Takiguchi. M.. Mori. S.. Kobori. O.. Morioka. V.. Inoko, H..Seriguchi. M.. and Kano. K. Expression and function of class II antigens on 32.gastric carcinoma cells and gastric epithelia: differential expression of DR.DQ and DP antigens. J. Nati. Cancer Inst.. 79: 923-932. 1987.

14. Van den Ingh, H. F.. Ruiter. D. J.. Griffioen. G.. Van Muijen. G. N. P., andFerrone, S. HLA antigens in colorcctal tumours—low expression of HLA 33.class I antigens in mucinous coloréela!carcinomas. Br. J. Cancer. 55: 125-

130. 1987. 34.15. Esteban. F., Ruiz-Cabello. F.. Conchas. A.. Perez-Ayala, M.,Sanchez-Rozas,

J. A., and Garrido. F. HLA-DR expression is associated with excellent prognosis in squamous cell carcinoma of the larynx. Clin. Exp. Metastasis. 8:319-328. 1990. 35.

16. Springer, T. A. Adhesion receptors of the immune system. Nature (Lond.),346: 425-434, 1990.

17. Buckley. C. H.. and Fox. H. Carcinoma of the cervix. In: P. P. Anthony andR. N. M. Macswecn (eds.). Recent Advances in Histopathology. Vol. 14. pp.63-78. Edinburgh: Churchill Livingstone. 1989. 36.

18. Moll, R.. Franke, \V. W., and Schiller. D. L. The catalog of human cytok-eratins: patterns of expression in normal epithelia. tumors and cultured cells.Cell. 31: 11-24. 1982.

19. Vinci. G., Tabilo. A.. Dcschamps. J. F.. van Haeke, D., Henri, A.. Guichard, 37.J.. Tetleroo. P., Londsdorp. P. M.. and Hercend. T. Immunological studyin-\'itro of the maturation of human megakaryocytes. Br. J. Haematol., 56:589-601. 1984.

20. Van Heyningen. V., Guy. K.. Newman. R.. and Steel. C. M. Human MHC 38.class II molecules as differentiation markers. Immunogenctics. 16:459-469,

1982.21. Watson, A. J.. Demars. R.. Trowhridge. I. S.. and Bach. F. H. Detection of 39.

a novel HLA antigen. Nature (Lond.). 304: 358-361. 1983.22. Zielger. A.. Uchanska-/iegler, B.. Rosenfelder, G.. Braun, D. G., and \Ver-

net, P. Heterogeneity of established human hematopoietic cell lines: surface 40.antigens detected by monoclonal antibodies and glycosphingolipid patterns.In: W. Knopp(ed.), Leukaemia Markers, pp. 317-320. New York: Academic-

Press, 1981. 41.23. Epstein. A. L.. Marder, R. J.. Winter. J. N.. and Fox. R. I. Two new mon

oclonal antibodies (LN-1. LN-2) reacting in formalin fixed, paraffin embedded tissues with follicular center and mantle zone human B lymphocytes and 42.derived tumors. J. Immunol.. 133: 1028-1036. 1984.

24. Rothlein, R.. Dustin. M. L.. Marlin, S. D., and Springer, T. A. A humanintercellular adhesion molecule (ICAM-1) distinct from LFA-1. J. Immunol., 43.137: 1270-1274, 1986.

Sanchez-Madrid, F., Krensky, A. M., Ware, C. F., Robbins, E., Strominger,J. L., Burakoff, S. J., and Springer, T. A. Three distinct antigens associatedwith human T-lymphocytc-mediated cytolysis: LFA-1, LFA-2, and LFA-3.Proc. Nati. Acad. Sci. USA, 79: 7489-7493. 1982.I Lui A. S., Fuggle, S. V., Fahre, J. W., Ting, A., and Morris, P. J. Thedetailed distribution of MHC class II antigens in normal human organs.Transplantation (Baltimore). 38: 229-234. 1984.Sullivan. K. E.. Caiman. A. F.. Nakanishi. M.. Tsang. S. Y., Wang. Y., andPeterlin. B. M. A model for the transcriptional regulation of MHC class IIgenes. Immunol. Today, *: 289-293, 1987.

Ferguson, A.. Moore. M.. and Fox, H. Expression of MHC products andleucocyte differentiation antigens in gynaecological neoplasms: an inumimihistological analysis of tumour cells and infiltrating lymphocytes. Br. J. Cancer. 52: 551-563, 1985.Hughes, R. G., Norval, M., and Howie, S. E. M. Expression of major histocompatibility class II antigens by Langerhans cells in cervical intracpithe-lial neoplasia. J. Clin. Palhol., 4l: 253-259, 1988.C'oppleson, M., and Reid, B. L. Prcclinical Carcinoma of the Cervix Uteri,

321 pp. London: Pergamon Press, 1967.Boon. M. E., Baak. J. P. A.. Kurner. P. J. H., Overdiep, S. H. and Verdonk,G. W. Adenocarcinoma in situ of the cervix: an underdiagnosed lesion. Cancer (Phila.). 48: 768-773. 1981.

Dustin. M. L.. Rothlein, R., Bhon. A. K.. Dinarello. C. A., and Springer, T.A. Induction by ILI and interferon (•>):tissue distribution, biochemistry, andfunction of a natural adherence molecule (ICAM-1). J. Immunol., 137: 245-

254. 1986.Momburg, F.. and Moller. P. Non-co-ordinate expression of HLA-DR antigens and invariant chain. Immunology, 57: 551-553, 1988.Shwarz, E., Freese, U. K., Gissmann. L.. Mayer. W.. Roggenbuck, B.. Strem-lau. A., and zur Hausen. H. Structure and transcription of human papillo-mavirus sequences in cervical carcinoma cells. Nature (Lond.). 314:111-113,1985.Jochmus-Kudielka, I.. Schneider, A.. Braun, R.. Kimmin, R.. Koldovsky, I ..

Schneweis, K. E., Seedorf. K., and Gissmann. L. Antibodies against thehuman papillomavirus type 16 early protein in human sera: correlation ofanti-E7 reactivity with cervical cancer. J. Nat!. Cancer Inst.. 81: 1698-1703,1989.Moore. M.. and Ghosh, A. K. MHC status of primary human coloréela!carcinoma. In: F. Gimino. J. Birkmayer, E. Kliavins. R. Pimental, and F.Salvatore (eds.). Human Tumor Markers, pp. 421-433. Berlin: Walter deGruyer and Co., 1987.Ottenhof, T. H. M.. Elferink, D. G., Hermans, J.. and de Vries. R. R. P. HLAClass II restriction repertoire of antigen-specific T cells. I. The main restriction determinants for antigen presentation are associated with HLA-D/DRand not with DP and DQ. Hum. Immunol.. 13: 105-106. 1985.Altmann, D. M.. Sansoni. D.. and Marsh. S. G. E. What is the basis forHLA-DQ associations with autoimmune disease? Immunol. Today, 12:267-270, 1991.Sálgame, P., Convit, J., and Bloom. B. R. Immunological suppression byhuman CD8* T cells is receptor dependent and HLA-DQ restricted. Proc.Nati. Acad. Sci. USA. 88: 2598-2602, 1991.Ostrand-Rosenberg. S.. Clements, V. K.. Thakur. A., and Cole. G. A. Trans-fection of major histocompatibility complex class I and class II genes causedtumour rejection. J.Immunogenet., 16: 413-423, 1989.Concha. A.. Cabrera. T.. Ruiz-Cabello. and Garrido. F. Can the HLA phe-notypc be used as a prognostic factor in breast carcinomas. Int. J. Cancer,Suppl.. 6: 145-154. 1991.Bal, V.. Mclndoe, A., Dentón, G., Hudson, D.. Lombardi, G.. Lamb, J., andLechler, R. Antigen presentation by keratinocytes induces tolerance in human T cells. Eur. J. Immunol.. 20: 1893-1897. 1990.Pankratova. E. V. Role of histocompatibility antigens in nonimmune intercellular interactions. Mol. Biol., 24: 231-235, 1990.

4016

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1992;52:4009-4016. Cancer Res   Susan S. Glew, Margaret Duggan-Keen, Teresa Cabrera, et al.   Papillomavirus-associated Cervical CancerHLA Class II Antigen Expression in Human

  Updated version

  http://cancerres.aacrjournals.org/content/52/14/4009

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

  .pubs@aacr.orgDepartment 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/52/14/4009To request permission to re-use all or part of this article, use this link

on May 26, 2020. © 1992 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from