1

A THESIS FOR THE DEGREE OF

DOCTOR OF MEDIOIHE

UNIVERSITY OF LONDON

Possible Cofactors in the Development of .Cervical Neoplasia;

A Study of Immunologically Active Cells in Cervical Epithelium

Submitted by

Simon Everard Barton

Whittington Hospital London 1 9 8 9 The aetiology of cervical neoplasia is unknown. Although there is evidence that human papillomavirus (HPV) infection may be associated with oncogenesis, it is clear that not all women with genital warts develop either cervical intraepithelial neoplasia (CIN) or invasive cervical cancer. To explain this discrepancy, it has been suggested that one or more factors may be involved in the development of cervical neoplasia. Epidemiological studies suggest that these potential cofactors include cigarette smoking, oral contraception, semen exposure and cervical infections. However, their mechanisms of action are uncertain.

The increased risk of cervical neoplasia in immunosuppressed women has led to the hypothesis that a diminished immune response is important in the aetiology of this cancer. This thesis investigates the immunologically active cells present in the cervical epithelium for any changes associated with the potential cofactors cited above.

Immunocytochemical techniques were used to identify cells involved in the immune response in colposcopic biopsies taken from healthy,

HPV infected and pre-neoplastic cervices. The cell counts per nmr of cervical epithelium were then examined for any association with the putative cofactors being studied.

A significant association was found between cigarette smoking and a decrease in the Langerhans' cell counts in both normal and premalignant cervical epithelium. Such changes in the antigen detecting and presenting arm of the local immune system may facilitate persistent viral infection and permit transformation to neoplasia. A similar effect was noted in cervical epithelium found to be infected by HPV type 16. Other immunological changes associated with cigarette smoking were also noted. However, no significant association was detected between changes in the local immune cell distribution and any other possible cofactor examined.

Supplementary evidence for the role of cigarette smoking in cervical neoplasia is provided by a toxicological study of cotinine in cervical mucus and its relationship to the severity of CIN.

This thesis identifies an immunological mechanism which may explain the observed association between cigarette smoking, HPV type 16 infection and cervical neoplasia. 4

TABLE.OF CONTENTS.

Abstract...... 2

Table of Contents...... 4

List of Tables and Figures...... 6

Acknowledgements...... 9

PART 1: INTRODUCTION

1.1 The Need for Research into the Aetiology of Cervical Neoplasia...... 11

1.2 The Aetiology of Cervical Neoplasia...... 13

1.3 The Role of the Host Immune Response in Cervical Neoplasia... 30

1.4 The Immunological Effects of Possible Risk Factors and Cofactors in the Aetiology of Cervical Neoplasia...... 45

1.5 The Hypothesis to be Examined...... 53

PART 2: GENERAL METHODOLOGY

2.1 Recruitment of Patients...... 54

2.2 Consent and Documentation of History...... 55

2.3 Colposcopy and Specimen Collection...... 55

2.4 Specimen Analysis...... 59

2.5 Statistical Methodology...... 62

2.6 Definitions of Possible Factors and Cofactors to be examined in each of the Study Groups...... 63

PART 3: SPECIFIC EXPERIMENTAL METHODS, RESULTS AND SHORT DISCUSSIONS

3.1 A Study of the effects of Possible Cofactors on the Langerhans' cell counts per mnP in Cervical epithelium...... 64

3.2 A Study of the effects of Possible Cofactors on Lymphocyte cell counts per mm2 in Cervical epithelium...... 80

3.3 A Study of the effects of Possible Cofactors on the Macrophage cell counts per mnF in Cervical epithelium...... 92

3.4 A Study of the effects of Possible Cofactors on Natural Killer (NK) cells in Cervical epithelium...... 100 TABLE OF CONTENTS (Continued)

3.5 Cervical Infections and Alterations in Local Immunological Cell Papulations in the Cervical epithelium...... 107

3.6 A Study of cotinine concentration in cervical mucus...... 112

PART 4: CONCLUSIONS

4.1 Changes in local epithelial immunological Cell Populations in HPV and CIN lesions of the Cervix...... 121

4.2 The effects of the Possible Cofactors studied on the distribution of Immunologoicaliy active cells in Cervical epithelium...... 124

4.3 Future Research...... 137

4.4 A Summary of Claims for Original Contributions to Scientific Knowledge...... 141

APPENDIX:

I Standard History sheet used for all patients studied...... 143

II Standard Examination sheet used for all patients studied.... 146

III A/B Database used in Analysis of Section 3. 1-3. 4...... 147

PART 5: REFERENCES...... 151

SUPPLEMENT:

Papers by the author which are referenced in this thesis and submitted in support of his candidature...... 187 o

LIS.1_.QF TABLES A W L 1IGURES

Tables

Table 1 ,1 Case control studies of Cervical Neoplasia and Cigarette Smoking...... 26

Table 1.2 Summary of findings of Case control study by Trevathan et al 1983...... 27

Table 3.1 Median Langerhans' cell counts per nmr according to the histology of cervical epithelium...... 72

Table 3.2 The effects of Cigarette Smoking on Langerhans' cell counts in Normal cervical epithelium...... 72

Table 3.3 The effects of Cigarette Smoking on Langerhans' cell counts in HPV infection...... 73

Table 3.4 The effects of Cigarette Smoking on Langerhans' cell counts in CIN/HPV...... 73

Table 3.5 Results of Log-linear modelling for a dose dependent effect of cigarette smoking on Langerhans' cell counts per mnP...... 74

Table 3.6 Results of Log-linear modelling for an effect of HPV type 16 DNA detection on Langerhans' ceil counts per unit area in cervical epithelium showing evidence of HPV infection...... 75

Table 3.7 Results of Log-linear modelling for an effect of HPV type 16 DNA detection on Langerhans' cell counts per unit area in cervical epithelium showing histological changes of CIN/HPV ...... 75

Table 3.8 Median UCHT1

Table 3.9 Median UCHL1 positive T cell counts per umr- according to histology of cervical epithelium...... 87

Table 3.10 Median T4 (CD4) positive T cell counts per mm2 according to histology of cervical epithelium as three groups...... 88

Table 3.11 Median T4 (CD4) positive T cell counts per mm2 according to histology of cervical epithelium as five groups...... 88

Table 3.12 Median T8 (CDS) positive T cell counts per mm* according to histology of cervical epithelium...... 89 Table 3.13 Log-linear modelling for dose dependent effect of cigarette smoking on UCHT1 (CD3) positive T lymphocyte cell counts per mm2 in cervical epithelium showing changes of HPV infection...... 89

Table 3.14 The effects of Cigarette smoking on macrophage cell counts per nmfi; according to cervical epithelial histology...... 96

Table 3.15 The effects o f HPV type 16 D M detection on macrophage cell counts per ran2 according to cervical epithelial histology...... 97

Table 3.16 Results of NK cell identification according to histology of cervical epithelium...... 102

Table 3.17 Numbers of women in semen-exposed and non-exposed groups according to histological findings...... 103

Table 3.18 Prevalence of Cervical infections detected in study group...... 107

Table 3.19 The range of histological findings in colposcopic biopsies taken from cervices with detectable HPV type 16 D M ...... 117

Table 3.20 The range of histological findings in colposcopic biopsies taken from cervices without detectable HPV type 16 D M ...... 117

Table 4.1 Carcinogens detected in the smoke of non-filter cigarettes...... 130

Figure 2.1 Colposcopy Clinic at Finsbury Health Centre...... 56

Figure 2.2 MOP-Videoplan system used in measuring ceil counts.56

Figure 3.1 Bright Cryostat used for cutting sections...... 67

Figure 3.2 S100 stained Langerhans' cells in normal cervical epithelium (x250)...... 67

Figure 3.3 T6 (CD1) positive Langerhans' cells in normal cervical epithelium of a non-smoker (xl50)...... 69

Figure 3.4 T6 (CD1) positive Langerhans' cells in cervical epithelium showing changes suggestive of HPV infection, from a twenty per day cigarette smoker (xl50)...... 69 Figure 3.5 T lymphocytes stained using UCKT1 (CD3) in cervical epithelium showing changes suggestive of HPV infection (x250)...... 85

Figure 3.6 Macrophages stained by 3.9 (CDllc) marker in cervical epithelium showing changes of HPV infection

Figure 3.7 Graph showing the means (with 95% confidence intervals) of the 3.9 (CDllc) positive macrophage cell counts per unit area according to histology of cervical epithelium...... 95

Figure 3.8 Graph showing the means (with 95% confidence intervals) of the Ell (CD35) positive macrophage cell counts per unit area according to histology of cervical epithelium...... 95

Figure 3.9 The American Health Foundation Laboratories...... 115

Figure 4.1 Proposed Immunological Model for the role of Possible Cofactors in the aetiology of Cervical Meoplasia.... 136 9

ACKNOWLEDGEMENTS,

The initiation, management and completion of this project would not have been possible without the invaluable support of Albert Singer

FRCOG and David Jenkins MRCPath, both of the Whittington Hospital.

Their collaboration forms an invaluable bridge between the clinician and the pathologist without which studies, such as this thesis, would not be possible. I am indebted to them both.

I gratefully acknowledge the contribution of Peter Maddox FIMLS, whose skill and remarkable patience introduced me to the art of practical immunocytochemistry. The laboratory methods developed by both he and his colleague Margaret Hills MBiol have provided my work with a background of essential technical expertise.

My thanks are due to Dr, Y. Khwaja and her staff at the Finsbury

Health Centre for collaborating with the setting up of the colposcopy clinic and encouraging women to attend. In particular, I am grateful for the invaluable support of Judith Young SEN whose contribution to the smooth running of the clinic was immense.

In the setting up and running of the colposcopy clinic, I fully acknowledge the role of my colleague Tony Hollingworth MRCOG. I thank him for providing me with the results of the HPV type 16 DNA analyses from his PhD thesis, performed under the critical supervision of Dennis McCance PhD at Guys' Hospital. 10 -

In performing the statistical analysis of the results, I was very fortunate to have access to the Imperial Cancer Research Fund computer system and the supervision and advice of Rob Edwards PhD and Jack Cuzick PhD of the department of Mathematics, Statistics and

Epidemiology, Also at this institution, Haney Hogg PhD kindly provided the monoclonal antibodies used for macrophage identification.

The financial support for this study, including my trip to the

American Health Foundation was provided by MedScand, whose president, Prof. M i s Stormby provided constructive criticism and encouragement in equal measure.

I am indebted to Drs. Haney Haley and Daniel Sepkovic of the

American Health Foundation, Hew York, USA, for allowing me into their laboratory and especially to Ellen Louis who taught me the whole of cotinine biochemistry in a week!

I would also like to thank the patients who participated in this study. I humbly hope that their participation has resulted in a study which may contribute a little towards the effort to reduce the incidence of cervical neoplasia.

This thesis is dedicated to my Father, Bless You.

Simon Barton

Whittington Hospital - 11 -

PART l: INTRODUCTION

1,1 The Need for Research into the Aetiology of Cervical Neoplasia

Although the overall mortality rate from cervical cancer in England and Vales is declining (Intercollegiate Working Party on Cervical

Cytology Screening 1987), the age specific analysis of these figures reveals that there is an increased incidence and mortality in women born since 1940 (Cook & Draper 1984). It has been predicted that this will result in a 50% increase in the all-ages cervical cancer rate by the year 2000 (Beral & Booth 1986). As cervical cancer is already the second commonest cancer in women worldwide (Parkin et al

1984), the global implications of such an increase will be of major social and economic significance (Peto 1986).

Although cervical cytology screening programmes have been shown to be responsible for reducing the mortality from cervical cancer in

Nordic countries (Hakama 1984), there is no conclusive evidence that this is the case in England and Vales (Beral & Booth 1986). However, the planned introduction, in 1988, of a national call and recall system for cervical smears is intended to increase the number of women being effectively screened. This may further add to the rising incidence of cases of cervical intraepithelial neoplasia (CIN) being detected (Volfendale et al 1983; Duguid et al 1985). The resources needed to provide adequate colposcopy and therapy for women with abnormal smears will thus require further expansion (Cuzick 1988). - 12

Some authors (Bonham et al 1987) have questioned the cost

effectiveness of cervical screening programmes, as well as their

efficacy in reducing the mortality from cervical cancer.

Thus it may seem inevitable that cervical neoplasia will continue to

cause an increasing health problem for women and that any attempt to

screen effectively for CIU will present an insuperable resource

problem. However, if the aetiology of cervical neoplasia could be

precisely determined, then a different approach might be made. By

aiming to detect and alter those factors which promote the

development of cervical neoplasia, a step towards a truly

preventitive, rather than a 'detect and treat' programme might

become possible. Such an approach could be used to prevent women

from ever developing CII, or alternatively to predict those women

whose CIH lesions would not require treatment because there was no

risk of progression to invasive cancer. There is already extensive

evidence based on both cytological (Fox 1967) and colposcopic

followup (Campion et al 1986), that some cases of CIU may

spontaneously regress. The identification of factors which determine

this process may lead to the possibility of modifying these factors

to produce therapeutic regression of such lesions.

This thesis will investigate the distribution of immunologically active cells in cervical epithelium and search for any evidence of an association between local immunological changes and the possible cofactors which have been proposed in the aetiology of cervical neoplasia. - 1 3 -

1,2 The Aetiology pf .Cervical-Neoplasia

Cervical neoplasia encompasses the premalignant stages of CIU as well as invasive squamous cell cancer. The aetiology of this condition has been extensively investigated but remains elusive.

Despite the accessibility of the cervix, which permits examination and tissue sampling without the need for unduly invasive procedures, our understanding of the natural history of the development of cervical neoplasia is incomplete.

Epidemiological studies have identified many risk factors which may be involved in the aetiology of cervical neoplasia, although their individual importance, interactions and biological effects await further investigation. The evidence that these potential risk factors are involved in the aetiology and natural history of cervical neoplasia will be reviewed. All the relationships reported have been observed with both CIN and invasive cervical cancer unless otherwise stated.

The incidence of CIU begins in late adolescence and increases to a peak at the age of 25 years, subsequently slowly declining (Doll

1986). Invasive cervical cancer, however, increases in incidence from the age of 20 years to a peak at age 50 years, after which it remains fairly constant. This difference has long been held to support the argument that CIN is a true precursor of cervical cancer

(Reagen & Patten 1962). But the large discrepancy between the - 1 4 -

numbers of women with Cl IT and those subsequently developing invasive

disease implies that not all lesions are progressive. The extent to which regression may be promoted by the removal of lesions by punch biopsies or trauma, such as childbirth and instrumentation (Richart

1967), is unknown. But these explanations, even if true, cannot fully explain the finding of regression rates of up to 50% in patients followed by cytology alone without biopsy (Uaisell et al

1976).

Socioeconomic , status

Women from lower socioeconomic classes have a higher incidence of cervical cancer than those in the higher income groups. Several possible explanations including poor hygiene, lower standards of medical care and greater exposure to environmental carcinogens have been suggested as explanations for this difference. However it is

iDrobable . that the association of lower social class with an earlier age at first intercourse may be the most important factor

(Christopherson & Parker 1965), although this association was not found by Harris et al (1980).

Parity

Boyd & Doll (1964) showed that the association between high parity and carcinoma of the cervix disappeared when allowance was made for the correlation between higher parity and an earlier age at marriage, There is no evidence that repeated trauma to the cervix during childbirth predisposes to the development of malignancy

(Anderson 1987). - 15 -

Sexual behaviour.

The venereal origin for cervical neoplasia has been demonstrated by many diverse observations. Originally, Rigoni-Stern in 1842 recorded that the disease was more common in married rather than single women and then, more recently, the observation was made that cervical neoplasia was virtually absent in nuns (Gagnon 1950; Towne 1955).

These findings indicate an association of cervical neoplasia with sexual activity.

The age at first coitus was shown to be important by Rotkin (1967), who found that 42% of cancer patients commenced intercourse before the age of 17 years, compared to less than 20% in a control group.

Such an association is linked to the evidence of Harris et al

(1980), who found that an increased number of sexual partners had the effect of increasing the risk of developing CIU.

If cervical neoplasia is initiated or promoted by a chemical carcinogen transmitted by coitus, then it may be expected that the frequency of sexual intercourse would be associated with the development of disease, however such data is lacking (Doll 1986).

Although an increased risk of cervical neoplasia has been reported in prostitutes (Peryra 1961; Moghissi et al 1968), these studies have been insufficent to ascertain whether this is due to the earlier age of commencing coitus which has been reported in prostitutes (Duenas et al 1972) or their high prevalence of sexually transmitted infections (Barton et al 1987). ~ 1 6 -

Male_£isk Factors...For the Aetiology nf Cervical frpoplasia

There are several studies which support the role of the male in the

aetiology of cervical neoplasia. Firstly, Buckley et al (1981) found

that in a group of women with law numbers of sexual partners, their

risk of cervical cancer increased according to the number of sexual

partners which their male consorts had previously. Secondly, a study

by Kessler (1977) found a 1.7 times increase in the incidence of

cervical neoplasia in the second wives of men whose first wife had

died of cancer of the cervix. Furthermore, Martinez (1969) found an

increased incidence of cervical cancer in the wives of men with

penile cancer.

These studies support the concept of the 'high risk male' proposed

by Singer et al (1976). Further evidence for this model comes from

the importance of the occupation of the patient's husband in a way

which does not only reflect socioeconomic status, but is also linked

to the opportunities for promiscuity of both partners during the

husbands' absence from home. Examples of this include the increased

incidence of cervical neoplasia in the wives of long distance lorry

drivers and merchant seamen (Beral 1974).

Transmissible agents

The role of coitus in the development of cervical cancer has

beenexplained by the transmission of some form of carcinogenic agent

from the male to the female during sexual intercourse. The evidence supporting the claims of the various factors proposed for this role will now be examined. - 1 7 -

Spermatozoa and sem inal fluid

Evidence for the role of spermatozoa in the development of cervical neoplasia was originally put forward by Coppleson & Reid (1967),

Their argument was based on the observation that metaplastic cervical epithelial cells were found to phagocytose spermatozoa. In an experiment to investigate this phenomenon in mice, it was found that the DNA of the host cells had been altered by spermatozoa ingestion. Recently, the same unit (French et al 1987) presented evidence that sperm protamine could induce neoplastic transformation in cultured human cervical cells in vitro.

It is clear that every cervix which comes into contact with spermatozoa and semen does not inevitably undergo malignant transformation. In an attempt to explain this discrepancy, Fish et al (1982) presented evidence, from a study using rats, that the induction of cervical dysplasia by herpes simplex virus infection was significantly facilitated by sperm exposure.

Despite this evidence, further studies of the response to spermatozoa and semen in the female lower genital tract are necessary before any conclusions about the possible role of these agents can be made.

S m e g m

The arguments in favour of this agent came from the evidence of a lower incidence of cervical neoplasia in women from societies where the men are circumcised, such as Muslims (Wynder et al 1954). - 1 8 -

However, no evidence has been found that smegma is carcinogenic,

Indeed, further studies have shown that the role of circumcision may

be secondary to the other sexual characteristics of religious groups

such as orthodox Jewesses (Coppleson 1969).

Treponema Pallidum

Several older studies (Fry 1929; Harding 1942) found an association

between syphilis and cervical cancer. However, when the degree of sexual promiscuity is taken into consideration, the association of

these conditions appears to be due to a common mode of transmission,

rather than a causal link (Rojel 1953).

Neisseria gonorrhoeas

A Danish study by Furgyk & Astedt (1980) followed up a cohort of

women who had been diagnosed as having gonorrhoea twenty four years previously, and compared them to a control group. There was a

fourfold increase in the rate of carcinoma in situ and an eight fold

increase in the rate of invasive cervical cancer in the study group.

Despite these findings, there is no supportive evidence that this constitutes any more than an association through a link with coital exposure to a sexually transmissible agent (Nahmias et al 1985).

Trichomonas Vaginalis

Although trichomoniasis can produce cellular changes which may resemble dyskaryosis (Frost 1961), these changes are reversible following treatment. Besides this phenomenon there is no other evidence linking this protozoal infection with cervical neoplasia, - 1 9 -

despite the assertion that the epithelial injury of carcinoma in

situ favours the growth of T. vaginalis (Koss 8t Volinska 1959).

Chlamydia trachomatis

Naib (1970) first described an association between this infection

and cytological evidence of squamous atypia. Hare et al (1982)

isolated C. trachomatis from 18.2% of women with invasive cancer

compared to 1% of controls. A link was also postulated between the

contemporaneous finding of lymphoid follicles and CII in cervical

biopsies (Hare et al 1981). Although this link was supported by the

findings of Paavonen et al (1982), the evidence for any more than a casual association remains unproven.

Herpes simplex., virus. (HSV.l

Since the initial report by Naib et al (1966) of an increased

incidence of cervical neoplasia in patients with a history of herpetic infections detected in cytological smears, there have been numerous seroepidemiologic studies investigating this association.

Rawls et al (1970) demonstrated antibodies to HSV type 2 in 83% of women with invasive cervical cancer and 33% of women with CIN, compared to a control group with 9 and 22% positivity, for higher and lower social classes respectively. Seth et al (1978) demonstrated that this difference was not common to other malignancies. The prognostic importance of HSV infection was demonstrated by Ory et al (1974, 1977), who showed that women aged

15-24 years with serological evidence of infection had a relative risk of 5.44 times of developing CIN than a control group of women. - 2 0 -

Despite this epidemiological evidence, it remains possible that an

association exists only because both HSV infection and cervical

neoplasia are related to coital exposure to another agent. However,

two further pieces of epidemiological evidence point to a causative

association.

Firstly, the mean age of women with genital HSV infection being 20 years, compared to 25, 31 and 48 years for the development of dysplasia, carcinoma in situ and invasive cervical cancer (Jfaib et al 1969). This suggests that HSV is not a secondary pathogen in women with pre-existing cervical neoplasia.

Secondly, a difference was detected in the quantitative measure of neutralising antibody (using a technique involving the kinetics of a neutralisation assay) between women with CIU and women attending a sexually transmitted disease clinic without CIN (Skinner et al

1971). Although the significance of this antibody assay is uncertain, the higher values in CIIV patients support the argument that HSV infection may constitute one of the aetiological factors necessary for the development of cervical neoplasia. Only in a subpopulation of women infected with HSV, will the presence of other variables determine the development of cervical neoplasia.

Suggestions for the identity of these cofactors have been proposed, such as the basic proteins in spermatozoa (Fish et al 1982), human papillomavirus infection and cigarette smoking (zur Hausen 1982). - 21 -

In addition to epidemiological studies, evidence has been found that

HSV antigens may be detected in exfoliated cells from squamous cervical cancers (Royston & Aurelian 1970). Further studies have detected the genetic sequences of the HSV2 genome in the cells of

40% of CIH lesions and 30% of invasive cervical cancers (Maitland et al 1981; McDougall et al 1982).

Wentz et al (1981) were able to induce cervical neoplasia in mice by using HSV inactivated by either formalin or ultraviolet light. The oncogenic capability was greater with virus inactivated by the latter method. There was no difference between the rates of cancers produced by HSV type 1 or 2. With improvements in the methods of antigen detection, the distribution of different types of HSV antigens in cervical neoplasia is being further elucidated (Fenoglio

& Lefkowitch 1985).

Humah...£ap n iQfflayir.us...(HEV)

The first report of a passible association between HPV infection and cervical cancer was made as recently as 1976 by Meisels & Fortin who described a subclinical form of cervical infection which could be recognised cytologically and histologically. Since that time, there has been a dramatic upsurge in the interest in papillomaviruses which has lead to increased knowledge about the nature of this association.

Firstly, the epidemiological evidence shows that the mean age of women with HPV infection is 24 years, compared with 31 years for dysplasia, 37 years far carcinoma in situ and 55 years for invasive cancer (Meisels et al 1982). The temporal development of

progressively more severe abnormalities with increasing age provides evidence to refute the argument that an association only exists

because tumour cells are in some way more susceptible to secondary viral infection .

Secondly, as techniques of molecular cloning and DNA characterisation improved, evidence has appeared that different viral types are specifically associated with certain types of

lesions in certain sites. In particular, HPV type 16 has been detected in 66% of cases of CIN (Schneider et al 1985) and 90% of cases of invasive cervical cancer (McCance et al 1985), but only rarely in condyloma acuminata (Campion et al 1985). In contrast, HPV types 6 and 11 are more often associated with condylomata or mild grades of CIN only (Gissman et al 1983). HPV types 16 and 18 have also been detected in several cervical cancer cell lines in culture

(Schwarz et al 1985). It is important to note the evidence that HPV

DNA is free and not integrated into the host cell genome in condylomatous and CIN lesions, whereas in most invasive cancers studied, the HPV DNA is integrated into the host chromosomes (Singer

& McCance 1985). This process of integration may be a vital step in the process of malignant cell transformation and tumour development.

Thirdly, the venereal transmission of condyloma acuminata (Oriel

1983) seems to be related to the development of cervical neoplasia.

In a study by Campion et al (1985), 32% of the female sexual - 2 3 -

partners of men with genital warts had evidence of CIS compared to no cases in a matched control group, D M hybridisation studies of these lesions revealed that 78% of the CIN lesions and 66% of the male partners' penile warts contained HPV type 16 DM.

The establishment of an association between HPV type 16 infection and CIN still does not prove a causative link in the aetiology of cervical cancer. The malignant potential of higher grades of CIU has been recognised for many years (Reagan & Patten 1962), with the risk of the development of invasive cancer from CI3T 3 being reported as

30% (Mclndoe et al 1984). The argument that HPV type 16 is a cause of cervical cancer needs proof that cervical infection with this virus confers a particular risk of malignant change. This was examined by Campion et al (1986), who prospectively followed 100 women with cytological and colposcopic evidence of mild cervical atypia consistent with Clfll, for a minimum of 19 months. Progression to CIH3 occurred in 26 women, of whom 22 had evidence of HPV type 16 infection by filter-in-situ D M hybridisation. Although spontaneous regression occurred in 11% of cases, disease recurred in four of these women, including the only woman in this subgroup with HPV type

16 infection. This evidence suggests that HPV type 16 positivity correlates well with worsening CII, but it still does not prove a direct association with invasive disease.

Cantraoeptipa

The results of epidemiological studies of oral contraceptive (OC) use and cervical neoplasia are confusing. The case control studies - 24 - by Ory et al (1977) and. Harris et al (1980) found evidence for an increased relative risk of developing cervical neoplasia with the use of OC for 3 and 10 years respectively, although four similar studies (Worth & Boyes 1972; Thomas 1972; Sandmire et al 1976; Boyce et al 1977) showed no difference. Similarly, the cohort study of

Peritz et al (1977) found a 2.8 times relative risk after 6 years exposure and a similar study by Vessey et al (1983a) found a 2.3 times risk for 8 years use. However, in contrast, another cohort study by Melamed & Flehinger (1973) found no association.

Possible mechanisms for a carcinogenic effect of OC on the cervical epithelium include a direct hormonal influence causing endocervical eversion (Editorial 1977) or an effect on the speed of progression of pre-existing mild cervical dysplasias (Stern et al 1977).

However, other workers have concluded that the association between

OC use and cervical neoplasia is merely due to the increased frequency of cervical smears performed in OC pill users (Harris et al 1980; Irwin et al 1988).

The proposal that an association between OC use and cervical neoplasia may be due to the temporal association between a reduction in the use of barrier contraception and an increase in promiscuity resulting in more venereal diseases was proposed by Singer 8i Jordan

(1977). Evidence that diaphragm users had rates of cervical neoplasia one sixth of those using OC, and one fifth of those with an intrauterine contraceptive device (IUCD) support this argument

(Wright et al 1978). However, it is likely that differences in the - 2 5 - sexual behaviour of women who choose different methods of contraception may account for some of these observed differences.

There is no evidence that the use of an IUCD has any effect on the risk of developing cervical neoplasia (Harris et al 1980). Whereas there is some evidence that women receiving injectable progestogens

(e.g. medroxyprogesterone acetate) may be at a significantly increased risk (Powell & Seymour 1971), though a WHO study (1984) defined only a small relative risk (2.2) to women using the drug for longer than 5 years.

Overall, the evidence for any association of cervical neoplasia with contraceptive use is scanty and all the epidemiological methods used are subject to major criticisms about their failure to properly exclude confounding variables; a problem which some authors claim to be insurmountable (Swan & Petitti 1982).

Cigarette smoking

Summaries of the published epidemiological studies are shown in

Table 1.1 (modified from Vessey 1986). - 26 -

Table—1., 1; Case control studies of Cervical Neoplasia and Cigarette Smoking.

Author Place No. Disease Studied Relative Risks

Naguib et al 1966 USA 31 CIS Current 2,1; Ex 1.3

Thomas 1973 USA 209 CIS Current 1.5

Vigle et al 1980 Canada 508 CIS 1.2 to 2.7 168 INV 2.8 to 3.7

Harris et al 1980 UK 190 CIS 2.2 to 2.4

Stellman et al 1980 USA 332 INV 1. 1 to 1.4

Clarke et al 1982 Canada 181 INV Current 2.3; Ex 1.7

Berggren et al 1983 Sweden 609 CIS Current 3.0

Lyon et al 1983 USA 217 CIS Current 4.1

Marshall et al 1983 USA 513 INV Current 1.6; Ex 0.8

La Vecchia et al 1986 Italy 183 CIN Current 1.8 230 INV Current 1.7

{Where two relative risk figures are shown, this reflects the lower and higher intake groups respectively)

N.B.INV=invasive cancer CIS=carcinoma in situ CIN=all grades of cervical intraepithelial neoplasia.

No.dumber of patients studied. Ex =Ex smoker. - 2 7 -

Table-1,2; Summary of findings of Case Control Study bv Trevathan st al 1963

Figures shown represent relative risks for the development of

cervical lesions identified in left hand column, subdivided by the

effects of increasing cigarette exposure;

Pack Years

Ex <1 4-6 12-

Carcinoma in situ 2. 1 2.3 3.8 12.7

Severe dysplasia 5. 7 2.5 4. 1 10.2

Moderate dysplasia 1.6 0. 7 3.3 4.3

Ex = Ex smoker

JT.B, The histological terms used in this table are those used by the original authors. 2 8 -

All the case control studies indicate an increase in the risk of the disease among smokers, although in some the relative risk is small.

The strongest association was described by Trevathan et al (1983)

who demonstrated a cumulative risk with increased exposure and a

persistent risk for ex-smokers (Table 1.2). This trend was supported

by the findings from the Oxford Family Planning Association study

(Greenberg et al 1985) which found the relative risk of cervical neoplasia to be 1,6 times higher in ex smokers, 1.5 times higher in women smoking 1-14 cigarettes per day, and 2.1 times higher in those smoking more than 15 cigarettes per day. In reviewing this data,

Vessey (1986) concluded that there is 'convincing evidence that smoking increases the risk of cervical cancer' even 'after adjustment for the influence of sexual behaviour* . Although he notes that ' there are essentially no published observations from the laboratory establishing a direct effect of smoking on cervical cells' . The lack of such evidence means that the hypothesis of zur

Hausen (1982) that smoking represents an important cofactor for the development of cervical neoplasia still awaits confirmation.

5unmar.y:

The currently popular model for the aetiology of cervical neoplasia supports the importance of HPV, especially type 16, but emphasises the need for potential cofactors, such as cigarette smoking, cervical infections and systemic immunosuppression to act either alone or together to influence the viral oncogenic potential

(Pfister 1987). - 2 9 -

Recent evidence (Meanwell et al 1987) has suggested that HPV type 16

DNA can be detected in histologically normal cervices, with a frequency which increases with age. This finding has lead to criticism of the importance of HPV type 16 in the aetiology of cervical neoplasia (Editorial 1987). However, this would strengthen the argument for the possible role of cofactors; for if they were unimportant, then all women with HPV 16 cervical infection should develop invasive cancer. Unfortunately, this study did not accurately assess cigarette exposure or other possible cofactors for an effect on the prevalence of HPV type 16 infection, nor were all the women studied using colposcopy. As the colposcopic assessment of women with negative cytology and cervical HPV type 16 infection can reveal histological changes of CIU in up to 17% of cases

(Hollingworth et al 1987), this latter omission may account for an overestimate of 'normal' cervices if cytology alone is used as the indicator.

Until the results of properly controlled prospective studies of women with HPV type 16 cervical infection are known, the debate about the significance of this virus will remain unresolved. In the meantime, the study of the role of one possible cofactor has gained impetus from the increasing evidence of the importance of passive cigarette smoking (Hellberg et al 1986). In pursuing such research into the multifactorial aetiology of cervical neoplasia, it is essential to project a model, which even if it subsequently proves false, serves to allow a hypothesis to be tested and a new model constructed (Popper 1974). - 3 0 -

1,3 The T?n1e nf the Host Immune Response in Cervical Neoplasia

There is clear evidence that patients receiving immunosuppressive therapy following renal transplantation have an increased incidence of malignant tumours (Kinlen et al 1979). The increased incidence of lymphoma, especially reticulum cell sarcoma, is 35 times higher than normal, and skin and lip cancers occur 4 times more often than expected (Hoover & Fraumeni 1973). In women undergoing such immunosuppressive therapy, a fourteenfold increase in the occurrence of carcinoma in situ of the cervix has been reported (Porecco et al

1975). A larger retrospective study (Schneider et al 1983) similarly noted a seven fold increase in the incidence of all grades of CIIT.

Although Sillman et al (1984) found that CIH in immunosuppressed patients is less responsive to conventional therapy and may even have a more rapid progression to invasive cancer, these findings were complicated by the diversity of the causes of immunosuppression in the group of women studied, which included lymphomas, sarcoidosis and lupus syndrome.

There are several passible explanations for these findings. Firstly, the drugs used to produce therapeutic immunosuppression may be directly carcinogenic. Secondly, a response to the graft itself, or the effects of the immunosuppressive drugs, modifies the response to oncogenic virus infections. Thirdly, the impairment of the host immunological surveillance mechanism leads to a failure to recognise and respond to tumour development. - 3 1 -

Although certain immunosuppressive drugs do interfere with cell division there is no evidence of a direct carcinogenic effect, even at high doses. Furthermore, the finding that children with ataxia telangectasia, which features a progressive decrease in cytotoxic T cell production, have an increased incidence of lymphosarcoma and reticulum cell sarcoma (Levis & Datner 1979) demonstrates the

importance of the impairment of cell mediated immunity in leading to the increase in malignant disease. Indeed, the overall frequency of malignancy in patients with primary cell mediated immunodeficiencies is 10,000 times greater than in an age-matched population (Gatti &

Good 1971). In contrast, there is no evidence of an increase in neoplasia in children born with defects of antibody production, such as Bruton's agammaglobulinaemia; instead the major sequelae of this defect is an increased incidence of bacterial infections (WHO 1978).

Experimental evidence has demonstrated that the adoptive transfer of tumour-specific immune T cells between laboratory animals results in the rejection of growing inoculated tumour cells in the recipient, as well as conferring an ability to reject subsequently inoculated tumour cells, so called 'tumour immunity' (Fernandez-Cruz 1980,1982;

Hamaoka & Fujiwara 1987). Although such findings seem to support the model proposed by Burnet (1970) that the major role of the cell mediated immune system is to maintain immunosurveillance for the recognition and deletion of malignant cell clones, it is apparent that not all conditions which involve a depressed cell mediated immune system (e.g. leprosy) result in an increased tendency to malignant disease. Experimental evidence also shows that thymus- - 3 2 -

deprived or congenitally T cell deficient mice do not show an

increased incidence of spontaneous tumours (Ioachim 1976). It is

interesting that after exposure to oncogenic viruses (e.g. polyoma

virus) such mice do show an increased incidence of malignant tumour

development (Mori et al 1964). Although these findings need to be

treated with caution due to the short lifespan of experimental mice

and the dangers of extrapolating evidence across species, this evidence adds weight to the case for examining the role of the

immune system in the development of cervical neoplasia. In

particular, the investigation of the cell mediated arm is indicated because of its role in the immune response to both viruses and tumours.

Systemic Immunity in Cervical Neoplasia

The observations (Mickshe et al 1978; Hancock et al 1979; Alsabti

1980) that a high proportion of women with advanced invasive cervical cancer were anergic to both primary (e.g. dinitrochlorbenzene) or recall (e.g. Candida) antigens has been used as evidence that immunodeficiency plays a role in the progression of this disease. However, these findings are common to other gynaecological malignancies, such as ovarian or endometrial carcinoma (Khoo & Mackay 1974), as well as tumours at other sites

(Robinson 1981). Although there may be an association of low host immunity with poor prognosis, this evidence cannot be used to imply that any abnormality of the immune response was essential for the development of the original tumour. - 3 3 -

In searching for the cell mediated defect which results in the

cutaneous anergy, many studies have reported reduced T cell counts

in peripheral blood in women with advanced cervical cancer (Rand et al 1977; Levy et al 1978; Ishiguro et al 1980; Satam et al 1981),

though, in contrast, Bonneterre et al (1981) found no evidence of a

reduced cell count, even in advanced disease. Likewise in CIH, there are studies which have found the T cell count in peripheral blood is decreased (Sawonabori et al 1977; Ishiguro et al 1980) and others which show no change (Rand et al 1977; Bonneterre et al 1981).

Immunologically, the functional ability of the circulating T

lymphocytes is of greater significance than a simple numerical count of these cells. A reduced response to mitogens, such as phytohaemagglutinin or pokeweed (Levy et al 1978; Stratton & DiSaia

1982) and a depressed lymphocyte rosetting capability (Sawanobori et al 1977) have been shown in patients with cervical neoplasia.

However, Van der Linde et al (1983) did not find any association of these changes with increasing levels of any antibody directed against tumour cells.

Studies of B cell counts in cervical carcinoma are similarly inconclusive. Whereas some workers find no change in B cell counts

(Rand et al 1977; Levy et al 1978; Hancock et al 1979; Satam et al

1981; Bonneterre et al 1981), others have detected a decrease

(Ishiguro et al 1980; Kietlinska 1984). The main function of B cells is to produce antibody and much work has been concentrated on finding an antibody to a tumour specific antigen in cervical - 3 4 - neoplasia. The possibility of detecting such an antibody for use in screening programmes for cervical cancer was first proposed by Levi

(1971), but has since met with scant success. Although carcinoembryonic antigen (CEA) has been shown to be related to tumour volume (Kjorstad & Orjaster 1978) and has a potential use in detecting early recurrences after therapy (te Velde et al 1982), its lack of specificity for cervical neoplasia limits its usefulness in population screening.

In general, the problems of interpreting the significance of systemic changes in immunological parameters mean that their relationship to the development of cervical neoplasia remains uncertain. Indeed, the methodological nature of the study of systemic immunity results in immense difficulties in elucidating the meaning of any detectable change. Whether a decrease in cell count corresponds to genuine immunosuppression or represents a shift of cells to another site is uncertain (Emerson 1978). Similarly, changes in systemic cell activity may be an indicator of immunodeficiency or a reflection of localised immunostimulation. To avoid some of these pitfalls, the study of the host immune response to a tumour at a specific site should begin at that site; in this case the cervical epithelium.

The Local Immune Response in Cervical Neoplasia

The first observations of a chronic inflammatory infiltrate in both carcinoma in situ and invasive cervical cancer (Freidell et al 1960) were interpreted as constituting a form of host resistance or an - 3 5 - autoimmune phenomenon (Glucksman & Cherry 1964). In a review of this subject in 1971, Reid suggested that this was an integral part of the local reaction to epithelial dysplasia and might be implicated

in limiting mutagenic change in the epithelium.

The original attempt to subdivide and even quantify the nature of the observed inflammatory infiltrate was performed by Singer (1973).

He noted that there was a reduced inflammatory infiltrate in dysplastic compared to normal epithelium. The constituents of this infiltrate were identified as predominantly neutrophils and mononuclear cells in the normal epithelium, with an increasing proportion of mononuclear cells seen with increasing grades of cervical dysplasia. It was proposed that the greater degree of inflammatory response seen in lesser grades of abnormality was related to the increased spontaneous regression rate in less severe lesions.

These early studies were limited by their reliance on simple morphology to recognise immunologically active cells. Attempts were made to identify cells by electron microscopy (Younes et al 1968); although such an approach was technically restricted and unsuitable for assessing large areas of epithelium, Langerhans' cells were identified. But it was only in conjunction with the developments in molecular biology and the production of monoclonal antibodies

(Kohler St Milstein 1975), that immunocytochemical techniques for the reliable identification of immunologically reactive cells has become possible. - 3 6 -

Langerhan^1 cells

Langerhans' cells are epithelial dendritic cells, eponymously named

after the medical student who described them in the shin over 100

years ago (Langerhans 1868). Their origin was for a long time a

source of confusion, the main reason for which was the controversy

about the best way of identifying them (Wright & Alison 1984).

However, it is now accepted (Katz et al 1979; Tamaki et al 1980)

that they take their origin from bone marrow and are found in all

squamous epithelia (Choi & Sauder 1986). Functionally, Langerhans'

cells are involved in the process of antigen presentation to produce

specific helper T cell proliferation and the induction of a

cytotoxic T cell response to allogeneic and modified-self

determinants (Shelley & Juhlin 1976; Stingl et al 1978; Streikin et

al 1980; Pehamberger et al 1983; Boog et al 1985). This ability to

present foreign antigens to T lymphocytes depends on the expression

of MHC class II antigens on the surface of Langerhans' cells.

Although the level of expression of class II antigens is known to vary (Berman et al 1985), their exact functional significance remains uncertain (Solheim et al 1986).

Although Langerhans' cells are rarely observed in mitosis on tissue sections, studies using tritiated thymidine show that 1% of them may be proliferating in an epithelium at any one time (MacKenzie 1975).

The factors which control this process are unknown.

The earliest study of Langerhans' cells in the cervix were performed using electron microscopy. Younes et al (1968) studied 10 normal - 3 7 - cervices, and detected Langerhans' ceils in all of them by their distinctive cytoplasmic organelles, called Birbeck bodies (Birbeck et al 1961). Hackerman et al (1968), similarly found these cells in six premenopausal and one postmenopausal woman. Using a zinc iodide osmium technique, Figueroa & Caorsi (1980) described the distribution of Langerhans' cells in the mature squamous epithelium of three normal premenopausal women.

In 1983, Morris et al (1983a), published the results of the first study to make use of new immunocytochemical methods for detecting

Langerhans' cells in the cervix. Using separate markers to HLA-DR, common leucocyte antigen and thymocyte antigen, this study demonstrated the presence of Langerhans' cells in 10 biopsies of cervical epithelium taken from 30 week fetuses, term infants, pubertal adolescents, and both pre and post menopausal women. These same methods were also applied to 12 cervical biopsy specimens showing histological changes of HPV infection and CUT (Morris et al

1983b). In comparison to the normal distribution of Langerhans' cells, tissue showing signs of HPV infection showed a partial depletion or absence, whereas areas of CIM showed a marked increase in these cells. Despite the limited number of samples examined, the authors concluded that there was an immunosuppressive effect of cervical HPV infection which may augment the development of neoplasia, as well as evidence for a host response to such premalignant change. - 3 8 -

In an attempt to quantify the numbers of Langerhans' cells in

cervical epithelium, Vayrynen et al (1984) used T6 (CD1), a

thymocyte marker, to identify Langerhans' cells in the cervices of

263 women with evidence of HPV infection or GIN being followed up

for 16 months in a colposcopy clinic. The fraction of cells

identified in a count of 200 cells in the epithelium was used as a

measure of Langerhans' cell concentration. No significant

difference was found in the cell concentration in HPV infected

tissue or in CIN, Furthermore, there was no difference in the

Langerhans' cell concentrations in those women whose lesions

apparently showed spontaneous regression.

A further study of the distribution of Langerhans' cells in the

whole of the lower genital tract was performed by Edwards & Morris

(1985) . Using both T6 (GDI) and HLA-DR antigen markers to identify

Langerhans' cells, these workers expressed the cell distribution as a count per 100 other cells. The highest concentration of

Langerhans' cells was found in vulval epithelium and the lowest in

vaginal epithelium. In conjunction with their identification of a

variation in the concentration of T lymphocytes in the lower genital

tract, the authors proposed that there was a functional 'cervical

lymphoid tissue' which constituted part of the mucosal associated

lymphoid tissue noted at other anatomical sites (e.g. oesophagus, bronchi and intestine).

Using an antibody to S100 protein as a marker, McArdle & Muller

(1986) examined changes in 79 cervical biopsies with histological - 3 9 - changes ranging from squamous metaplasia through HPV infection to

CIBT 3. They used a cell counting method which expressed the

Langerhans' cells per unit area of cervical epithelium. This showed that the Langerhans' cell concentration was significantly reduced in

HPV infection and increased in areas of CIN. In studying invasive cervical carcinoma and areas of coexistent CIS using an ingenious method of epithelial laminae production, Caorsi & Figueroa (1984,

1986) have shown that the magnitude of the increase in Langerhans' cell concentration is directly related to the severity of the neoplasia.

In direct contrast to these previous findings, Tay et al (1987a) reported a decreased Langerhans' cell count in CIU lesions, as well as in HPV infection. These findings were based on the identification of Langerhans' cells using four separate immunocytochemical markers, all of which demonstrated the same trend in the cell counts. Indeed, the different effects which HPV infection and CII exerted on

Langerhans' cells identified by S100 and T6 (GDI) markers lead to the hypothesis that the S100 staining positive cells represented a functional subpapulation of immune cells.

The findings of Tay et al (1987a) have not yet been conclusively repeated. However, further evidence of a reduction in Langerhans' cell found in CIN lesions has been provided by arecent study by

Hughes et al (1988). This study also examined MHC class II expression in Langerhans' cells in the cervix and demonstrated that, despite the reduction of Langerhans' cells in both HPV infection and - 40 -

CIN, the HLA-DR expression increased. Whether this process is related to the S100 positive subpopulation is not known.

T lymphocytes

The study of T lymphocytes has been facilitated by the development of monoclonal antibodies which allow the identification of functional subclasses (Rung et al 1979; Moretta et al 1982;

Sinkovics & Dreesman 1983). This has lead to a subclassification of

T lymphocytes into two distinct groups. One group, recognised by the presence of the 1'4 (CD4) surface antigen, are mature T lymphocytes with regulatory and antigen recognition functions, restricted by class II HLA antigens, and labelled 'helper' cells because of their capacity to generate a cytotoxic response. A second group express T8

(CD8) surface antigen, have cytotoxic and suppressor functions and are HLA class I restricted; their name derives from their

'suppressor' effect on the immune responses of other lymphocytes to specific or non-specific epitopes (Evans et al 1978; Reinherz et al

1983). Despite the appeal of such a simple classification, the detection of subgroups of functionally heterogenous cells within these classes has recently been proposed (Dalgleish 1986; Blann

1987).

Studies of T lymphocytes in the cervix are scanty. Morris et al

(1983b) using a pan-T cell marker, UCHT1 (CD3), and a cytotoxic/suppressor T cell marker, T8 (CDS), found a depletion of intraepithelial T lymphocytes in HPV infected cervical epithelium, compared to normal controls. Whereas in cases showing histological - 4 1 - evidence of CIU, there was a marked increase in the T cell count in the epithelium, the increase was primarily attributed to T8 (CD8) antigen bearing cells. These findings were interpreted by the authors as indicating that HPV infection exerts an immunosuppressive effect on the cervical epithelium. This study was marred by the lack of detecting T4 (CD4) lymphocytes, the limited number of biopsies studied (12) and the use of a semi-quantitative method of assessing cell counts. Similar findings for the changes in T lymphocyte counts in CIU came from the work of Syrjanen (1983) who studied 150 biopsies of cervical dysplasias, comparing those with evidence of

HPV infection to those without. The major criticisms of this work are the non-specific method used to stain lymphocytes and the lack of a reliable quantitative method to enumerate the cell counts.

Tay et al (1987b) improved on methodology by using monoclonal antibodies to pan-T, T4 (CD4) and T8 (CD8) antigens, as well as introducing an improved quantitative method of cell counting. This study found that in normal cervices, the T4(CD4)/T8(CD8) populations were in similar proportions to the peripheral circulation, but that in HPV and CIN lesions there was a progressive decrease in both the number of T lymphocytes and a reversal of the normal T4(CD4)/T8(CDS) ratio. Such findings may be used to support the argument that a local immune deficiency in HPV infected cervical epithelium is implicated in the aetiology of cervical neoplasia. - 4 2 -

Mscroph&ges

Macrophages are found in most tissues in the body. They are able to perform a variety of functions including non-specific phagocytosis, antigen presentation, direct cytotoxicity to virally infected cells or tumour cells and inhibit lymphocyte proliferation (Dougherty &

McBride 1984). Originating in the bone marrow derived monocyte cell line, one subset of macrophages forms a tissue bound phagocytic network (the reticulendothelial system) whilst another subgroup are found primarily in the skin, lymph nodes, thymus and spleen, functioning as antigen presenting cells (Roitt et al 1985). The role of macrophages in the non-specific immune response to viral infections was recognised first by Mims (1964), and subsequently their further role in the induction, regulation and amplification of specific responses to viral infections has been elucidated (Mogensen

1979). Recent evidence has shown that macrophages do not act in isolation, but are involved in a dynamic interplay with other resistance mechanisms, including UK cells and other specific cell mediated and humoral responses (Morahan et al 1985).

The only published study of macrophages in the cervix was performed by Tay et al (1987c). This used two monoclonal antibodies, 3.9 and

Ell, to detect total macrophages and a subpopulation of activated macrophages respectively. Cervical biopsies from twenty one women were studied with histological features ranging from normality through HPV infection to CIN. Macrophages were found to be scarce in the normal cervix, with few active cells present. However, in HPV infection and CIBT, there was a progressive epithelial infiltration - 4 3 - of macrophages, particularly activated ones. The authors concluded that the presence of this response to uncomplicated HPV infection meant that it was directed against viral antigens rather than a direct anti-tumour response as suggested by Adams & Snyderman

(1979). This hypothesis merits further investigation. yalmr.al-.Killec_gfills

Natural immunity mediated through natural killer (NK) cells is thought to function as a complementary system to the adaptive cell mediated immune system in man (Moore 1985). NK cells constitute a subset of T lymphocytes which do not require specific antigen presentation by accessory cells to become activated; instead they may exert a spontaneous cytotoxic effect against virally infected cells and neoplastic cells (Herberman & Ortaldo 1981; Herberman &

Oldham 1983).

The importance of the role of NK cells in tumour immunosurveillance is emphasised by evidence that patients with congenital NK cell deficiency have an increased risk of developing malignancies (Kersey et al 1980; Roder et al 1980). In particular, Hersey et al (1982) reported the case of a girl suffering from Fanconi's anaemia with selective NK cell defeciency but normal T and B cell function. By the age of 22 years she developed extensive genital warts, vulvar squamous cell carcinoma and carcinoma in situ of the cervix.

Furthermore, both animal (Hanna & Fidler 1983) and human studies

(Forbes et al 1980) have shown that NK cells play an important role in anti-tumour responses in vivo. Systemic studies have shown that - 4 4 -

MK cell activity is diminished in women with invasive cervical cancer (Pulay et al 1982), but not in women with CIN (Neill &

Morvall 1984).

The only published study of local NK cell distribution in the cervix was performed by Tay et al (1987d). Using two markers, Leu-7 and

Leu-11, to detect NK cells by an immunocytochemical method, the examination of 23 cervical biopsies using a semi-quantitative method suggested that an NK cell infiltrate was present in some cases of

HPV infection and CIN. The cells were found primarily in the subepithelial stroma, but their low numbers lead the authors to conclude that they were unlikely to be involved in cytolysis of tumour cells. This conclusion may be premature, in view of the possibility that NK cells can cause immunological amplification by the secretion of interleukin-2 (Kasahara et al 1983), interleukin-1

(Scala et al 1984) and gamma interferon (Djeu et al 1982). Moreover, using the disparity in numbers of cells as a basis for deducing cell function is dubious as this assumes that all cells in a CIN lesion are recognised as tumour cells, which is both unlikely and unproven.

B lymphocytes

The humoral response forms one arm of the adaptive immune system, but is functionally interrelated to the other cell mediated arm. The role of the B lymphocyte response to viral infection is well documented (Mims & White 1984). The generation of an antibody response, the coating of viral antigens with antibody leading to facilitated phagocytosis and lysis by complement, and the long - 4 5 -

lasting plasma cell memory are important, though not essential, methods of protection against viral infection. The measurement of systemic levels of antibodies to viral and tumour specific antigens has been discussed above. The problems of interpretation of temporally uncertain and functionally unknown changes in these studies is difficult.

Two studies have demonstrated the paucity of local B lymphocytes in squamous cervical epithelium (Morris et al 1983b; Tay et al 1987b).

It is possible that improved techniques may lead to the measurement of antiviral antibodies in cervical mucus, but this awaits further investigation.

1.4 The Immunological Effects of Possible Risk Factors and Cofactors in the Aetiology of Cervical Neoplasia

Having established that a defective cell mediated immune response is a risk factor for the development of cervical neoplasia, it seems reasonable to examine each of the risk factors and cofactors suggested by epidemiological evidence for their possible role in producing a local or systemic immune defect which would favour the development of neoplasia.

Susceptibility to viral infection is greatest in the very young and the very old (Mims & White 1984). Furthermore, the occurrence of - 46 -

malignant disease is clustered at these extremes of age, when either

an immature or deteriorating immune system is present. However, the

epidemiology of cervical carcinoma does not fit into either of these

categories. Despite this, it remains possible that age related

changes in the local host response to persistent viral infections

may play a role in the aetiology of cervical neoplasia, although

such changes are more likely to be determined by more specific

factors than the general ageing process,

HPV infection

The current model for a continuum of HPV through to CIH and to

invasive cancer (reviewed by Pfister 1987) has meant that many of the studies previously discussed have examined the effects of HPV on

local cervical immunity. However, there is further evidence from studies of cutaneous HPV lesions that host immunity is important in determining the natural history of HPV induced lesions. Cutaneous warts have been shown to undergo spontaneous regression in between

25-35% of cases in the first six months after infection and in 35-

70% of cases in 2 years (Bunney 1982; von Krogh 1979). In flat warts of the skin caused by HPV 3, the state of regression is always associated with a pronounced mononuclear cell infiltrate (Berman &

Vinkelman 1980; Tagami et al 1974). Furthermore, regression of skin warts has been found to be temporally associated with a rise in the titre of IgG antibodies directed against HPV (Chardonnet et al

1985). 4 7 -

Overall, it is the host cell mediated immunity which has been shown

to be the most important determinant of wart regression (Thivolet et

al 1982), There are strands of evidence which enable the question of

what prevents all warts from regressing to be addressed. Chardonnet

et al (1986) found evidence that cutaneous HPV lesions deplete the

epithelium of Langerhans' cells and Tagami et al (1983) have shown that T8(CD8) positive lymphocytes predominate in cutaneous lesions.

Vhether these changes are reversed during the process of regression

is unknown, and of course, once a biopsy is taken, any changes observed may not reflect spontaneous effects. Ko study has so far reported any differences in the immune response to cutaneous lesions caused by different HPV types, although this possibility has been discussed (Chardonnet et al 1983) as a possible cause for the variation in behaviour of lesions induced by different HPV types.

It is interesting to note that there is experimental evidence that other members of the papillomavirus family may cause malignant disease, such as the Shope virus in rabbits (Rous & Beard 1935). In these experiments, carcinogenesis was found to be enhanced by coal tar or methylcholanthrene applications. More recently, clinical evidence has emerged for the malignant transformation of juvenile laryngeal papillomata containing HPV type 11 after exposure to X- rays (zur Hausen 1982), and epidermodysplasia verruciformis lesions, containing mostly HPV types 5 and 8, after exposure to sunlight

(Orth et al 1980). Such observations suggest that other cofactors may be necessary, via an immunological mechanism, for the oncogenic potential of HPV to be realised in the cervical epithelium. - 4 8 -

Cigarette Smoking

The finding that cigarette smoking can cause changes in the balance

of systemic T lymphocyte subsets indicates the potential effects of

cigarette smoke constituents on organs distant from the main site of

absorption in the lungs (Miller et al 1982). The changes seen in

smokers included a decrease in T4CCD4) positive cells and an

increase in TS(CD8) cells; these changes were reversible by abstinence from smoking for six weeks. Other studies have demonstrated that systemic T cell function may be disturbed

(Silverman et al 1975) and that M cell function may be reduced

(Fearson et al 1979).

The local effects of cigarette smoke in the lung have been

intensively studied. Reversible changes in alveolar macrophages have been documented (Martin 1973; Rasp et al 1978; Hoidal et al 1981), and a reduction in the antibody response to antigenic stimuli has been found in smokers (McSharry & Wilkinson 1986). But how can such local pulmonary effects be relevant to cervical neoplasia?

The finding by Sasson et al (1985) that nicotine concentrations are

56 times higher in the cervical mucus than in the serum of smokers has lead to the suggestion that other hydrophilic basic constituents of tobacco smoke could be selectively concentrated in the cervix.

The possible significance of such a carcinogen in the cervix was demonstrated by Holly et al (1985), who examined the mutagenicity of cervical mucus by using the Ames-salmonella test (Ames et al 1975).

They found that mucus from women who smoked or had concurrent - 4 9 - cervical dysplasia had significantly more mutagenic mucus than control subjects. Singer & Tay (1987) concluded that these findings suggested a direct effect of a cigarette smoke constituent on the cervix, potentially via the local immunological defences. This is one of the hypotheses which this thesis seeks to examine.

Furthermore, there is increasing evidence that non-smokers may have measurable systemic levels of cigarette smoke constituents from environmental contamination (Jarvis et al 1984). The effect of this phenomenon as a potential cause for lung cancer in non-smokers has been clearly demonstrated (Vald et al 1986). It remains to be shown whether passive smoking will be implicated in the aetiology of cervical neoplasia, but an effect on ovarian function, causing a reduced age of menopause, has already been demonstrated (Everson et al 1986).

Hormonal contraception

Although pregnant women have both a tendency to suffer from exacerbations of condylomata acuminata and a demonstrable decrease in cell mediated immunity (Schwartz et al 1987), there is no evidence that either of these phenomena are direct hormonal effects.

However, there have been reports that the addition of sex steroids to human lymphocyte cell cultures impairs cell function, although other workers dispute this finding (Clemens et al 1979). Sulke et al

(1985) demonstrated a periovulatory decrease in UK cell activity in women, which was absent in women using the combined oral contraceptive pill. As they did not find any direct association of - 5 0 - the changes in NK cell function with serum oestriol concentrations, they instead proposed a role for luteinising hormone as the immunosuppressive hormone, although the evidence for this is scanty.

Semen exposure

There is considerable evidence of the in vitro immunosuppressive effects of seminal plasma (James & Hargreave 1984). These include the production of a reduction in both macrophage and NK cell function. Studies of the in vivo effects of semen on the local immunity of the cervix are needed to further investigate this.

Cervical infections

In contrast to the studies of the immune response to HPV infection in the cervix, there has been minimal investigation of the local immunological effects of most other cervical infections.

It is known that bacterial infections such as gonorrhoea initially produce a neutrophil response (Harkness 1948) followed by a lymphocytic and plasma cell infiltrate in the deep cervical glands which form the major site for this organism (Slavin 1976).

Mycotic infections of the cervix have similarly been found to be associated with a predominantly neutrophil exudate (Garcia-Tamayo et al 1982). However, more detailed work by Vitkin et al (1986) has demonstrated that women who suffer from recurrent vaginal candidiasis have impaired candida-directed macrophage responses, - 5 1 -

Chlamydia trachomatis infection has become easier to study since the

development of direct monoclonal antibody kits which have made its

detection more efficient (Thomas et al 1984). Although considered

primarily to only infect columnar epithelium in the endocervix, recent work has suggested that it may also infect squamous epithelium (Barton et al 1985). Studies of the immune response to

Chlamydial infection suggest that monkeys exposed to C. trachomatis eye infection show markedly diminished lymphocyte activation (Young

& Taylor 1986). The ability of this organism to remain latent inside host macrophages raises the possibility of a persistent immunosuppressive action in the cervix (Levitt & Barol 1987).

However despite this experimental evidence, Syrjanen et al (1986) found neither any effect of concurrent C. trachomatis infection on the natural history of HPV cervical lesions nor any changes in the

NK cells, T cell subsets or Langerhans' cell counts in chlamydial infected cervices. The methods for counting immune cells in this study involved calculating a percentage cell count rather than a count per unit area; the former method may not detect subtle changes in the absolute numbers of cells nor be able to detect an effect if the denominator as well as numerator is altered by the agent being studied.

There is no paucity of data on the host response to HSV infection.

Indeed, much of the epidemiological evidence which has implicated

HSV in the aetiology of cervical neoplasia relies on serological studies of antibody to this virus. Specific study of the effect of

HSV infection on CIJF has suggested that there is an increased risk of progression of CUT 1/2 lesions to 01113 in women with antibody to

HSV 2 (Coleman et al 1983). Moreover, Syrjanen et al (1984) found serological or cytological evidence of HSV infection in 10% of women with cervical HPV infection during a 2 year follow up period. They argue that this supports the view of zur Hausen (1982) that KSV and

HPV may act synergistically to initiate cervical neoplasia.

Although such a conclusion may be premature with the currently available knowledge, the evidence does demonstrate the need for more accurate studies of the interaction of viral infections in the same site, particularly as HSV causes a latent infection with the possibility for reactivation at any time. The virus has been shown to reside in the uterosacral ligaments as well as the dorsal ganglia of the nervous system (Kitchener et al 1982). Indeed, asymptomatic cervical recurrences have been shown to occur frequently in some women (Barton et al 1986). The significance of this recurrent cervical ulceration, and resultant metaplasia, on the development of cervical neoplasia is uncertain.

Our knowledge of the host immune response to HSV is primarily based on animal studies (reviewed by Vildy & Cell 1985), with little knowledge of the local immune response in man. The continued investigation into the factors which determine latency and reactivation may produce data which will improve our understanding of viral interactions, the development of cervical neoplasia as well as provide the basis for a vaccine against HSV infection. 1 .JL.Thg Hypothesis to be Examined

Evidence has been presented that impairment of the host immune response, especially the cell mediated arm, may play a significant role in the development of cervical neoplasia. On this basis, an argument has been constructed that certain factors or cofactors which are associated with the development of cervical neoplasia may exert their effect via the immune system. The possibility that these immunological changes may be confined to the cervical epithelium has been proposed.

This thesis will examine the hypothesis that cigarette smoking, the use of the oral contraceptive pill, cervical exposure to semen and the presence of concurrent cervical infections may alter the distribution of cells which are potentially involved in the host immune response in the cervical epithelium. Normal cervical epithelium, as well as that showing histological changes of both HPV infection and CIN with HPV, will be examined for any local immunological effects which might be associated with these potential factors and cofactors. - 5 4 -

PART 2; OHNERAL METHODOLOGY

2,1 Recruitment of Patients

In March 1987, a colposcopy clinic (Fig 2.1) was set up in the

Family Planning Clinic, Finsbury Health Centre, Islington, London. A

Zeiss colposcope was installed in a clinical room and standard supplies and instruments obtained from the Royal Northern Hospital,

London.

Patients were recruited to the clinic from three sources; i) Women referred from the Family Planning Clinic having had an abnormal cervical smear. Those showing any suspicion of invasive cancer were seen immediately at the Royal Northern Hospital colposcopy clinic.

ii) Women referred from the Family Planning Clinic who had normal smears but had an increased concern about cervical disease because of a family history of cervical cancer, a past history of genital warts, or a partner with genital warts.

iii) Women who referred themselves for colposcopic screening. This facility was publicised by posters in the clinics inviting women to attend. No referral by their general practitioner was necessary. - 5 5 -

2.2 Con^PTrt and Documentation of History

An appointment system was organised, allowing 30 minutes for a first visit and 15 minutes for subsequent attendances. Each woman had a standardised history taken and this was recorded on a pre-printed history sheet (Appendix I). A full explanation was given to each patient of the nature of a colposcopic examination, the role of biopsy and the range of infections which were being tested for.

Verbal consent was obtained from all women included in the study.

Ethical approval (Reference EC87/15) was obtained from the Islington

District Ethical Committee prior to commencing the study.

2.3 Colposcopy and Specimen Collection

In the presence of a nurse as both chaperone and assistant, the patient was examined in the dorsal position. A Cuscoe's speculum was passed and a full view of the cervix obtained. The following specimens were taken in order;

An Ayre's spatula was used to obtain a cytological smear by rotation through 360° around the cervical os. After smearing the sample on a clean glass slide and fixing this with ethanol, the remaining material on the smear was transferred into isotonic phosphate buffered saline solution for filter-in-situ hybridisation for HPV type 16 DHA. 56 -

Figure 2,1; Colposcopy Clinic at Finsbury Health Centre

Figure 2,2; MOP-Videoplan System_uaed_liL-aaasmii.ng cell counts - 5 7 -

Two endocervical smears were obtained. One using a nylon cytobrush

(Medscand) was transferred to a welled slide and fixed with acetone for a direct monoclonal antibody test to identify the elementary bodies of Chlamydia trachomatis. A second cotton tipped endocervical swab was placed in Stuart's transport media for microbiological culture.

Using a cotton wool ball and sponge holding forceps, a liberal amount of 5% acetic acid solution was applied to the cervix.

Colposcopic examination of the cervix was then performed. The criteria used for the identification of the lesions of HPV infection and Cl IF were based on those of Walker et al <1983), Reid et al

(1984a) and Reid & Scalzi (1985). In addition to a standardised written description and diagrammatic representation (Appendix II) of the cervical appearance after acetic acid application, all patients had a photographic record produced by a specially designed

'Cervicography' camera (Stafl 1981).

Women with visible cervical lesions, in whom the new squamocolumnar junction was not seen throughout its circumference, were excluded from this study, as was any patiert in whom invasive cancer was suspected. Both these groups of patients had diagnostic cone biopsies arranged forthwith at the Roy1! Northern Hospital.

All patients had two biopsies taken from the cervix using

Eppendorfer punch biopsy forceps. In women with colposcopically visible lesions, the first biopsy was directed to the centre of the - 5 8 - lesion, and the second taken from the adjoining area which contained most abnormal tissue. In women with no visible cervical lesion, the first biopsy was taken from the transformation zone at 12 o'clock, including the original squamocolumnar junction where passible. The second biopsy was taken from the adjacent area clockwise to the first biopsy.

Both biopsies were orientated on a small piece of blotting paper with the epithelial surface parallel and away from the paper. The first biopsy was placed in a modified formalin calcium solution *.

The second was placed in a labelled Nunc tube and snap frozen in liquid nitrogen. Both specimens were carefully labelled and recorded in a laboratory book.

With the exception of the cell suspensions for HPV type 16 identification, all the specimens were transported to their site of analysis by the investigator. Careful records of all specimens and their whereabouts were kept.

* Made by dissolving lg of calcium chloride in 100ml of distiled water, then adding 10ml of 40% formalin, followed by 1ml of ethylene glycol and 7% sucrose. The pH of this solution was maintained at

4. 0. - *sq -

2.4 Specimen Analysis

i) Cytological analysis of cervical smears was performed, by the

Department of Cytology at the Royal Northern Hospital using the

criteria described by Evans et al (1986). All the smears of women

included in this study have been reviewed by Mr. R. Yabsley, Chief

MLSO in this department.

ii) Microbiological culture for Neisseria gonorrhoeae, Anaerobes,

Trichomonas vaginalis, Candida albicans and other yeasts, and group

B streptococci were performed using appropriate media by the staff

of the Microbiology Department of the Whittington Hospital.

iii) Identification of Chlamydia trachomatis was performed by Dr.

B.J. Thomas of St. Mary's Hospital, London, using a direct monoclonal antibody test kit (SYVA, UK) (Thomas et al 1984).

iv) HPV type 16 DNA identification was performed by Dr. A.

Hollingworth at Guy's Hospital Department of Microbiology, under the supervision of Dr. D.J. McCance, using the method described by him and others (McCance et al 1986).

v) The histological analysis of the specimens was performed by the investigator under the supervision of Dr. D. Jenkins, Department of

Histopathology, Whittington Hospital. The preparation of the sections for this analysis is described in the relevant section in

Part 3. The histological criteria used in the reporting of CIN are summarised below (based on Buckley et ai 1982; Intercollegiate

Report on Cervical Cytology Screening (Appendix V) 1987);

-CIN was diagnosed in cervical epithelium containing cells with enlarged, pleomorphic and hyperchromatic nuclei, a failure of normal cellular differentiation and the presence of abnormal mitotic figures.

-The grading of CIN was performed using the fallowing criteria;

-- CIN1: undifferentiated cells were confined to the the basal third of the epithelium. The nuclear abnormalities were most marked in the basal third, but present up to the surface of the epithelium. The mitotic figures were not abundant, mostly of normal form and confined to the basal third of the epithelium.

-- CIN2: undifferentiated cells extended into the middle third of the epithelium, but not into the upper third. The nuclear atypia was more marked than in CIN1 and atypical nuclei extended to the surface. The mitotic figures were fairly numerous, sometimes of abnormal form and were present in the lower two thirds of the epithelium only.

-- CIN3: undifferentiated cells extended into the superficial third of the epithelium. Harked nuclear abnormalities extended throughout the full thickness of the epithelium with abnormal mitotic figures similarly distributed. 6 1 -

-The histological diagnosis of HPV infection (Jenkins et al 1986a)

was based on the appearance of koilocytotic change, with typical

enlarged hyperchromatic nuclei, a granular chromatin pattern and a

wrinkled nuclear membrane. This was often associated with

multinucleation, acanthosis, parakeratosis or hyperkeratosis,

frequent mitotic figures and individual cell keratinisation. These

changes were considered suggestive of HPV infection when seen alone,

in areas adjacent to CIU lesions or superimposed on CUT lesions.

vi) The immunocytochemical techniques used to identify particular

antigens will be described in detail in the relevant experimental

method in Part 3. All were performed by the investigator under the

supervision of Mr. P. H.Maddox, the Chief MLSO.

vii) Quantitative analysis of a specific cell population within a

defined sectional area of cervical epithelium was performed solely

by the investigator. This was done by using a MOP-videoplan image

analyser (Kontron Electronics Group). This is a computer system (Fig

2.2) designed for the measurement and calculation of geometric

characteristics by the tracing of images on a digitising tablet. The

optical linkage of a microscope to the digitising tablet allowed,

with the appropriate magnification factor, the area of a section of epithelium to be measured. The area of epithelium measured was bound superficially by the upper limit of the epithelium, laterally by

imaginary lines perpendicular to the surface and in depth by the basement membrane. By also entering the number of positive cells in that section of epithelium into the computer, the count per unit - 62 - area could be calculated. This system also had a data storage and retrieval facility using floppy discs in an integrated dual disc drive. The cell counts were performed without knowledge of the individual case histories and were all completed before data analysis commenced.

2.5 Statistical Methodology

This was performed by the investigator under the supervision of R.

Edwards PhD and J. Cuzick PhD of the Department of Mathematics,

Statistics and Epidemiology, Imperial Cancer Research Fund, London.

The methods used were:

a) Vilcoxon rank sum test was used to analyse the difference in the median cell counts between different groups of patients.

b) Vilcoxon-type test for trend (Cuzick 1985) was used to test for the presence of a trend in cell counts between several groups of patients.

c) To test for dose dependent effects of particular cofactors on cell counts in the cervical epithelium, a log linear model was constructed using the Generalised Linear Interactive

Modelling (GLIM) computer package (Baker & Nelder 1985). The dependent variable of interest being the cell counts per square millimetre which was assumed to fallow a log linear model with

Poisson error. To allow for over-dispersion, an empirical scale factor was used. All analyses were stratified by age group and the results assessed by the likelihood ratio statistic which was referred to a x": distribution with 1 degree of freedom. - 6 3 -

2,6 Definitions of Possible Factors and Cofactors to be Examined in each of the Study Groups

From the patients' history, the following variables were extracted to be examined as potential factors (acting alone) or cofactors

(acting together) in the aetiology of cervical neoplasia;

Use of hormonal contraception

Use of barrier contraception

Cigarette smoking history

From the specimen analysis obtained from collaborators in this study, the following results were obtained;

Concurrent cervical infections

HPV type 16 D M identification

On the basis of cytology, colposcopy and histology, the patients were divided into the fallowing basic groups;

NORMAL = Normal cytology, colposcopy and histology.

HPV infection = Based on histological appearance.

CIN with HPV = Based on histological appearance.

These groups and the possible factors, or cofactors, cited above were the variables examined in the search for any association with the alterations in the concentrations of immunologically active cells in the cervical epithelium. The full details of which are described in Parts 3.1 to 3.4 of this thesis. - 64 -

PART 3; SPECIFIC EXPERIMENTAL METHODS. RESULTS AND SHORT DISCUSSIONS

3.1 A Study.pf-the effects of Possible Cofactors on the Langerhans1 cell counts per nun- in Cervical epithelium

Methods of Preparation of... sections

The cervical biopsy was placed in a modified formalin calcium solution (section 2.3) for at least 24 hours to ensure complete fixation. After this time, the specimen was rinsed in distilled water to remove any surplus fixative, washed with phosphate-buffered sucrose solution (7% sucrose in 0.1M phosphate buffer at pH 7.4) for

2 hours and then replaced in a labelled Munc tube for storage in liquid nitrogen.

Within two weeks, the specimens were cut using a Bright Cryostat model OTF/AS/D/MR/EC (Figure 3.1). Serial sections were collected on to 3-aminopropyltriethoxysilane (APES) coated slides (Maddox &

Jenkins 1987). Three slides, each with a minimum of two sections, were obtained, one slide for standard haematoxyiin and eosin staining for standard histological assessment (of 5p thickness) and two (of 10p thickness) for immunocytochemical staining. The slides were allowed to dry at room temperature for 30 minutes, then wrapped in tin foil for storage at -20°C. Within two weeks of this, immunocytochemical staining was performed. - 6 5 -

Immunocvtochemical methods used for the identification of Langerhans1 cells

After removal of the slides from the refrigerator, the sections were allowed to thaw to room temperature for 30 minutes. They were then incubated for ten minutes in methanol-hydrogen peroxide to block endogenous peroxidase activity and trypsinised for 3 minutes, using a solution of 0.1% trypsin in 0.1% aqueous calcium chloride at 37*C and pH 7.8.

The avidin-biotin complex (ABC) immunoperoxidase method was performed using the following antisera, incubated for 1 hour each in a humidified incubator at 37°C:

S100: i) S100 protein antibody at 1/100 dilution (z311 DAKQ)

ii) biotinylated anti-rabbit (DAKQ E353) antisera at

1/500 dilution.

T6(CD1): i) T6(CD1) antibody at 1/100 dilution (H721 DAKO)

ii) biotinylated anti-mouse (DAKQ E354) antisera at

1/400 dilution.

To both were added the ABC complex (DAKQ K355): 10pml/ml each of solution A and solution B made up in Tris buffered saline (TBS, pH

7.6) and 1/25 normal human serum (NHS). This was incubated for one hour. The chromogen used to develop the peroxidase was fresh 3 3' diaminobenzidine <0.5 mg/ml in TBS, pH 7.6) with 0.5% hydrogen peroxide, for an incubation period of 5 minutes. The sections were then counterstained using 1% Alcian blue in 3% acetic acid, and mounted in Xamma's xylene.

Details of immunocvtochemical markers used

S100 protein has been shown to be distributed in Langerhans' cells and melanocytes, but not keratinocytes (Cocchia et al 1981; Hakajima et al 1982). The number of melanocytes found in cervical epithelium is negligible and has not been found to affect the identification of

Langerhans' cells (Maddox et al 1987). Any positive staining due to the presence of S100 protein in neural elements (Hyden & Lange 1970) was excluded by using strict morphological criteria to identify

Langerhans' cells. The distribution of cells containing S100 protein has been studied in both normal and neoplastic epithelia in various sites in the body (Hirota & Shimasato 1982).

Chemically, S100 protein belongs to the group of acidic, calcium based proteins such as calmodulin (Isobe & Gkuyama 1978). Although the exact functional significance of S100 protein is unknown, it is known that the calcium signal is an important method of biological communication which may be important in T cell activation

(Cunningham-Rundles 1985). - 6 7 -

" < m f J " V r # ^ r ‘ - 4 * v / ^ > ✓ J r Y - k r ✓ % t Oh'*

V ~

. ^ ' V • ’ ^ ^ * V ♦ s > . * - T ' . ' " ' * m M .-/• - t ; t Jr* „ » - ' * . ♦ . * >' *% f # . . ,. . » c -V > y*JL*'*'**•' * *' ^ *v , ‘ g * « • ; .. M , ' * - ' ^ 0 * - i . „ r % - r * 4- ^ - I t " ••'-... --

F ig ..3., 2.1. SI 00 stainM-Jhangerhans1,, cells in normal .c e r v ic a l epithelium (x25Q) - 6 8 -

Tfi (GDI) antigen has been shown to react with a polypeptide chain of molecular weight 45 kd, termed human thymus antigen 1, which is present on cortical thymocytes and Langerhans' cells (Murphy et al

1981). The ability of this monoclonal antibody to identify

Langerhans' cells has been well demonstrated (Dubertret et al 1982), but no functional, as opposed to ontological, significance has yet been assigned to this marker. An example of T6 (CD1) staining of

Langerhans' cells is shown in Figure 3.3.

The use of two markers for separate and unrelated antigens was designed to improve the capacity of this experiment to detect any changes in Langerhans' cell counts. Although T6 (CD1) identifies more cells (Maddox et al 1988), as discussed above, S100 protein positive cells may constitute a separate functional group of

Langerhans' cells.

Cal cula tion ■QULangechaDs'. cell, .gaunt

Langerhans' cells were identified by their characteristic morphology

(Figures 3.2 and 3.3). The presence of a positively stained cell body with at least one attached dendrite was the criterion used to identify a cell. Using the method previously described (section

2.4), the cell count per mnr: was measured in the cervical epithelium. - 69 -

/ *.

-V

Figure 3,3; T5 (GDI) positive Langerhans1 cells in normal ■cervical epithelium of a non-smoker (x!5Q)

Figure 3.4: T6 (CD1) positive Langerhans1 cells in cervical epithelium, ■S.hQw.i.ng-.cJianges suggestive .of HFV infection, from a twenty per..dav cigarette. s m o k e r . (xl5i)_)_ 70 -

Results.

The 130 women studied were divided into three groups. Thirty six had

normal cervical cytology, normal colposcopy and no evidence of HPV

infection or GDI in the colposcopic biopsy, 39 had colposcopic and

histological evidence of HPV infection without CUT, and 55 had both

colposcopic and histological evidence of Cliff with HPV infection .

The median ages of each these groups were 27, 27 and 28 years respectively. Further demographic data and details of the factors and cofactors recorded for each patient are shown in Appendix IIIA.

The median Langerhans* cell counts by S100 and T6 (GDI) staining

found in normal cervical epithelium, epithelium showing histological evidence of HPV infection, and in CIIT/HPV lesions are shown in Table

3.1. The Langerhans' cells detected in CIIf/HPV lesions were noted to have fewer dendritic processes, and those present were shorter than

in normal epithelium. These changes were also seen in HPV infected epithelium. In normal epithelium, Langerhans' cells were found throughout the lower half of the epithelium, but in areas of CI1T/HPV with depleted cell counts, Langerhans' cells were confined to the region closest to the basement membrane.

The effects of cigarette smoking on the Langerhans' cell counts are shown separately for each histological group in Tables 3.2, 3.3 and

3,4. Current smokers had a highly significant decrease in the numbers of Langerhans' cells, detected by S100 and T6 (GDI), both in - 7 1 - normal cervical epithelium and in all grades of CIU/HPV. Ho such change was found in epithelium showing only histological evidence of

HPV infection.

The results of using a log linear model to test for a dose response relationship between numbers of cigarettes smoked and the

Langerhans' cell count in the epithelium are shown for S100 and T6

(CD1) markers, in Table 3.5. This demonstrates a highly significant association between the daily intake of cigarettes smoked and the reduction in the Langerhans' cell counts in both the normal cervix and CIN/HPV lesions.

Using log-linear modelling, a significant association was found between the detection of HPV type 16 D M (by filter hybridisation) and a reduction in both S100 and T6 (CD1) bearing Langerhans' cells in cervical epithelium showing changes of HPV infection (Table 3.6).

Furthermore, in CIU/HPV lesions, there was also a significant association between the detection of HPV type 16 D M and the reduction in Langerhans' cell counts (Table 3.7). This association was independent of the observed effect of cigarette smoking on the

Langerhans' cell counts per unit area found in CIU/HPV lesions.

Despite extensive statistical analyses, no significant association was found between the use of oral hormonal contraception, exposure to semen, or the presence of concurrent infection and the

Langerhans' cell counts per unit area (all p > 0.5) in either histologically normal or abnormal epithelium. Table 3-.1 Median.,JLangsrhans1 cell counts per.. mm:l_accQrding to the histology of cervical epithelium

S100 T6(CD1)

Formal (n=36) 51.4 115.9

HPV infection (n=39) 41.0 87.7

CIF/HPV (n=55) 12.1 28.3

(p=0,002)* (p<0.001)*

(p=0,0014)+ (p<0.001)+

*Formal vs. CIF +Test for Trend

Tabl3....3,2..The ef£g.c.t.s._Qf.Cigarette Smoking on Langerhans' Cell Count in ijonm i ..GervicaJLEfittbelium (Median cell counts per mm*-)

S100 T6(CD1) n

Fever smoked 72.5 137. 8 18

Ex smokers 63,6 115.7 3

Current smokers 9. 6 45. 4 15

(p=0.001)*

(p=0.001)+

^Current vs. Fever smoked +Test for Trend Table 3,3 The effects o f Cigarette smQki.ng_.on Langerhans' cell counts in HPV infection (Median cell counts per map)

S100 T6

Never smoked 63.6 93.4 13

Ex smokers 59. 8 81.7 8

Current smokers 17.8 75. 1 18

(p=0.12)*

^Current vs. Never smoked +Test for Trend

Table _3.4 The_sifgets of Cigarette smoking .on Langerhans1 cell counts in CIN/HPV (Median cell counts per mir-)

S100 T6

Never smoked 25. 1 44.5 11

Ex smokers 44. 6 37. 0 10

Current smokers 7. 1 21.6 34

(p=0.006)* (p=0.03)*

(p=0.003)+ (p=0.02)+

^Current vs. Never smoked tTest for Trend - 74 -

Table .3_JL Results. o J hogrlinear modelling for a dose dependent effect QI. cigarattg- SfflQki.agL.Qn the Langerhans1 cell counts per umr

31 M .

Category Regression Standard error LR x2 (ldf) LR p-value Coefficient*

Normal -0.0995 0.0252 23.9 <0.001

HPV infection -0.0190 0.0210 0. 9 0.34#

CIN/HPV -0.0717 0.0189 18.2 <0.001

i s i c m

Category Regression Standard error LR X2 (ldf) LR p-value Coefficient*

Normal -0.0734 0.0180 21. 1 <0.001

HPV infection -0.0043 0.0119 0. 12 0.72#

CIN/HPV -0.0311 0.0135 5.6 0. 02

* per daily cigarette.

# Not significant

LR = Likelihood Ratio Table 3,6 Results of Log-linear modelling for an effect of HPV type 16 D M detection on Langerhans1 call counts per unit area in cervical epithelium showing evidence of HPV infection

Antigenic marker Regression Standard error LR X2 (ldf> LR p-value Coefficient

S100 -0.0190 0.0050 15. 8 <0.0001

T6 (CD1) -0.2050 0.0738 7. 7 0. 006

LR = Likelihood Ratio

Table 3.7 Results of Log-linear modelling for an effect of HPV type 16 DffA detection on the LangerhansV cell counts per unit area in cervical epithelium showing histological changes of CIR/HPV

Antigenic marker Regression Standard error LR X2 (ldf) LR p-value Coefficient

S100 -0.2493 0.0967 6. 6 0. 01

T6 (CD1) -0.5609 0.0760 54. 1 <0.0001

LR = likelihood ratio - 76 -

Discussion

This study demonstrates that cigarette smoking results in a highly significant and dose dependent decrease in the concentration of

Langerhans' cells in normal cervical epithelium. This observed decrease in the number of the most prominent antigen presenting cells in normal cervical epithelium (Tay et al 1987a), suggests that a state of local cervical immunosuppression may exist in women who are otherwise clinically healthy. The reduced number of Langerhans' cells available to detect and present viral antigens to T lymphocytes might permit the establishment and persistence of viral infection. This could increase the likelihood of the development of a virally induced neoplastic transformation, as has been proposed for HPV infection (Singer & McCance 1985).

The finding of a significant association between cigarette smoking and a reduction in the Langerhans' cell count per unit area in

CIM/HPV lesions has.not previously been described. This association may partly explain the discrepancies in the previously reported changes in Langerhans' cell distribution in CIII lesions (Morris et al 1983b; McArdle & Muller 1986; Tay et al 1987a), as none of these studies examined differences in smoking habits in either the women studied or their control group. It is essential that an accurate smoking history is obtained and incorporated into future studies of the local immunological changes associated with cervical neoplasia. - 77 -

For cervical epithelium showing histological evidence of HPV

infection, there was no evidence of any significant effect of

cigarette smoking on the Langerhans' cells count per unit area. One

reason for this may be the heterogenous nature of HPV lesions; some

may be progressing, some static and others regressing, some recently

acquired and others longstanding. Unfortunately these are all

factors of which we have only limited knowledge. Other explanations

include the potential error in the diagnosis of HPV infection on the

basis of its histological features alone (Jenkins et al 1986b).

Furthermore, as this study did not detect any C I 3 S T lesions without

associated HPV infection, it was not possible to fully investigate

the nature of the interaction between cigarette smoking and HPV

infection on CIN lesions. Solutions of these problems, along with

the resolution of the debate as to whether GIN can exist without

evidence of HPV infection, await advances in the availability and

accuracy of in-situ diagnostic tests for HPV DM.

The study also found that Langerhans' cell counts per unit area were

significantly reduced in both HPV and CIH/HPV lesions in cervical

epithelium with evidence of HPV type 16 infection. Conclusions based on this data need to be guarded because of the inherent uncertainty

in the accuracy of the test used to detect HPV type 16 D M (McCance

DJ, personal communication). At present, a full assessment of this and other HPV diagnostic techniques is limited by the lack of any method for the viral culture of HPV and uncertainty about the sensitivity and specificity of particular techniques used in different laboratories, A further problem is the limitation of testing far only one HPV type. As the 'non-HPV type 16' infected group will probably include cases of uninfected epithelium as well as that infected by multiple other HPV types, it is possible that these confounding variables may account for the observed findings.

Despite these limitations, the observed association between the detection of HPV type 16 and a reduction in the Langerhans' cell count per unit area of epithelium may begin to explain the particular oncogenic potential of this viral type. The possibilty that HPV type 16 infected epithelium may exert a local immunosuppressive effect on the capacity of the host to recognise the presence of virally infected cells is more likely to lead to persistent infection and an increased likelihood of neoplastic change, which in turn will be less likely to result in an immunological response if HPV type 16 is still present. Further studies of the identification of which, and how many, other HPV types infect cervical epithelium, with both HPV and CIH/HPV lesions will enable a more reliable assessment of whether the observed immunological changes associated with HPV type 16 infection are correct.

This study shows that there is a decrease in Langerhans' cell count in cervical precancer which can be related both to smoking and to

HPV type 16 infection. It is important to determine whether the changes in Langerhans' cells can be reversed if a woman stops smoking. Assessment of this change is hampered by the confounding effect of the trauma of the initial biopsy, which also reduces the - 79 - tumour load and produces an immune response. However, in this study, two smokers who had colposcopic and histological evidence of HPV infection in their initial biopsy stopped smoking and were re- biopsied 3 months later; the counts had increased in persistent HPV lesions. Whereas in six women who had increased their cigarette intake, the counts had fallen and in two cases the disease progressed to CI¥. Although an attempt was made to examine the effect of stopping smoking on local immunological changes in the cervical epithelium, the reticence of women to be recruited to such a study was a problem which proved insurmountable.

The demonstration of Langerhans' cell depletion in the cervical epithelium of cigarette smokers is the first description of a potential mechanism for the role of smoking in the aetiology of cervical neoplasia through local immunosuppression. In conjunction with the observed effect of HPV type 16 infection on Langerhans' cell distribution, this data may be used to postulate that these two potential cofactors may act together, although possibly at different stages, to influence the host immunological response to the oncogenic process.

The main findings of this study have been published as a paper by

Barton SE et al, entitled 'Effect of cigarette smoking on cervical epithelial immunity: A mechanism for neoplastic change ?' Lancet

1988;ii:652-654. 60 -

3.2 A Study of the effects of Possible Cofactors on Lymphocyte cell c o u n t s per mm* in Cervical epithelium

Preparation of sections from frozen cervical biopsy specimens

The biopsy specimen was taken from storage in liquid nitrogen, allowed to thaw and then removed from the Nunc tube. It was then orientated in 1 ml of embedding medium (Cryo-m-bed) on a pre-cooied cryostat chuck. This produced a small capsule, which allowed easier handling of the specimen in a form which facilitated storage until cutting. This was performed within one month of the biopsy being obtained.

Sectioning of the specimen was performed by removing the capsule from the Nunc tube and placing it on t D a pre-cooled cryostat chuck, with a small quantity of distilled water to bond them together. The chuck and specimen were then placed on the pre-cooling plate inside the cryostat (Figure 3.1) at -20°C for 20 minutes to freeze. The temperature of the cryostat was adjusted about 30 minutes before use. It is normally kept at a temperature of -10°C, but for cutting frozen sections was reduced to -30°C in the chamber and -20’C on the plate.

Seven cases were normally cut at one time using a machine sharpened blade and the micrometer set to 7p. The biopsy was trimmed using one end of the blade until a reasonable face of the section was seen.

Then a ribbon of three sections was cut and picked up on an uncoated - 81 -

slide, for staining with haematoxylin and eosin. The middle of the

blade was then used to cut 24 ribbons of 3 sections which were

collected onto microscope slides coated with poly-l-lysine,

The sections were then left to air dry for 30 minutes and then

wrapped in foil and stored in empty slide boxes in the -20 °C

freezer, Immunocytochemical staining was carried out within 2 days

of sectioning the specimens, as prolonged storage may be detrimental to antigen preservation,

Immunocytochemical staining using the Alkaline phosphatase anti- alkaline phosphatase (AFAAP) method,

The sections were removed from storage and allowed to thaw at room temperature for 30 minutes. They were then fixed in acetone for 7 minutes in a Coplin jar at room temperature. After removal from the jar they were placed on blotting paper and allowed to air dry for 2 minutes. They were then placed in a moist chamber and rinsed in Tris buffered saline (TBS, pK=7.6) from a wash bottle.

The TBS was drained from each slide, which was then wiped with a clean tissue to leave moist only the area containing the sections .

This procedure has the effect of concentrating the anti-sera on the section and preventing further dilution of the test solutions.

However it was always ensured that the sections were not allowed to dry out at any time. - 82 -

An aliquot of 50 pi of primary antibody (mouse monoclonal) diluted with TBS was applied using a Jencons digital Finn pipette; the pipette tip was used to smooth out the antibody and ensure an even covering of the sections. The slides were incubated for at least one hour. The sections were then washed by flooding with TBS from a wash bottle, taking care not to dislodge the sections from the slide.

Aliquots of 50 pi of rabbit anti-mouse immunoglobulins (z259 DAKO) at a dilution of 1/20 were applied to each section, in the same manner as before. The sections were incubated for one hour. After which, the sections were washed twice in TBS from a wash bottle.

Aliquots of 50 pi of APAAP complex

Alkaline phosphatase developer was prepared using the following method; 40 mg of Sodium nitrite, 40 mg of Haphthol AS-B1 phosphoric acid and 80mg of levamisole were weighed out. The Sodium nitrite was placed in a test tube and dissolved in 1 ml of distilled water before adding 10 drops of 4% new fuchsin (in hydrochloric acid) and being left to diazatise for 5 minutes. The Naphthol AS-B1 was put into another test tube and dissolved in a few drops of dimethyl formamide. Tris buffer (pH 9.8) was then added to make up 160 ml total volume, in a beaker. The levamisole was then added to this solution and dissolved with the aid of a magnetic stirrer. - 83 -

The diazatised solution was added to the beaker and mixed well. This solution was then placed in a clean staining trough. The slides were then placed into metal slide racks and immersed in the developer in batches of twelve for 5 minutes each.

The slides were then washed in running tap water for 5 minutes before being counterstained using 1% Alcian blue in 3% acetic acid in a staining trough for 2 minutes. Any excess Alcian blue was washed off by immersion of the slide trays in fresh running tap water.

The slides were then placed in a trough of distilled water warmed to

60 °C on a hot plate and cover slips applied using a glycerine jelly, which had been previously heated to 60 °C in the oven. After which, the slides were then allowed to cool in a slide folder prior to labelling.

The slides were then ready for examination and quantitation. The positive sites of antigenic activity were labelled red against the counterstain of the nuclei and cytoplasm a blue-green colour.

Quantitation of cell counts per, unit„area.

This was performed by the candidate using the method described in section 2.4. The counts were made far each labelled cell population at one sitting and without reference to the individual case details. - 84 -

jjnjiQ c lo n a l antibodies, used to l a b e l Lvmphpcvtea_iji^thl5 study,.

IJCHTl belongs to the group of antibodies designated cluster differentiation 3

IJCHL1 is an antibody which reacts with activated T cells in vitro

(Beverley P, personal communication). UCHL1 antigen is the 180 kd polypeptide of CD 45 antigen; this has lead to UCHL1 being classed in the CD 45R group

T4 (CD4) reacts with a protein (mol. wt, 55 kd) present on most human T cells of helper/inducer subtype (Ledbetter et al 1981). A

1/10 solution of this monoclonal was used.

T8 (CDS) reacts with a protein (mol. wt. 33 kd) present on T cells of the cytotoxic/suppressor class (Erber et al 1984). A 1/10 solution of this marker was used in this experiment. - 85 -

MB2 is a murine monoclonal antibody of IgGi subclass, which, has been shown to detect B lymphocytes in both frozen and paraffin fixed tissue sections (Hall et al 1987). It recognises a neuraminidase resistant 28 kd antigen, which is strongly expressed on B cells. It was used in a dilution of 1/5.

Figure 3.5; T lymphocytes stained using UCHT1 (CD3) in cervical epithelium showing changes suggestive pf HPY injection (x.25Q) - 86 -

Easuiis.

The quantitative analysis of T cell populations found in different cervical epithelial lesions showed that the total T cell count per nmr (identified by UCHT1) is reduced in both HPV infection and

CIM/HPV lesions, compared to normal epithelium (Table 3.8). There was no significant change in the UCHL1 cell count in abnormal epithelium (Table 3.9). The reduction in total T cell count appears to be primarily due to a decrease in the T4 (CD4) subset (Table

3.10, 3.11) rather than the T8 (CD8) fraction (Table 3.12).

Using the data on individual patients (see Appendix IIIA/B), a full statistical analysis, as described in section 2.5, revealed only one significant association between changes in the cell counts per mm"' of T cell papulations and one of the possible cofactors being studied. Table 3.13 demonstrates that in tissue showing histological features of HPV infection, cigarette smoking was significantly associated with a decrease in UCHT1 (CD3) positive T lymphocytes. Ho association was found between any change in T cell concentrations and hormonal contraceptive use, cervical semen exposure or concurrent genital infection (p > 0.1).

B lymphocytes were detected in less than 8% of all the sections examined. In all of these, less than 5 cells were identified in the entire section, mostly being seen in areag of metaplastic or columnar epithelium. For this reason, no statistical analysis of local B cell populations was deemed possible. - 87 -

Table 3Ji Median UCHT1 (CD3) positive T cell counts per mm2 according to histology..of cervical epithelium,

Normal (n=16) 75. 10

HPV (n=36> 44. 47

CIN/HPV (n=52) 42. 42

Normal vs. CIN p=0.005 Test for trend p=0,02

Table 3,9 Median UCHL1 positive T cell counts per imr: according to histology of cervical epithelium.

Normal (n== 16) 14. 01

HPV (n=:36) 10. 19

CIN/HPV (n=:52) 11. 77

Normal vs, CIN p=0.70 Test for trend p=0.93 - 88 -

Table 3,10 Median T4 (CD4) positive T cell counts per m 2 according to histology of cervical epithelium asJfcfrree_gr.fliLD£.

normal (n=16) 34.96 n o> HPV 3 00 22.66

Cin/HPV (n=52) 15. 41

Uormal vs. CIU p=0.0005 Test for trend p=0.002

Table 3.11 Median T4 (CD4) positive T cell counts per roup according

normal (n=16) 34.96

HPV (n=36) 22.66

Cini/HPV (n=31) 16.75

Cin2/HPV (n=ll) 15. 19 cm3/HPV

Test for trend p=0.01 Table 3,12 Median T8 (CD8) positive T cell counts per nmr according to.histology. of .c_eryis.al epithelium it Normal Ci 31.55

HPV (n=36) 21.58

CIN/HPV (n=52) 25.45

Normal vs. CIN p=0.91 Test for trend p=0,53

Table 3.13 Log-linear modelling, for dose dependent effect of cigarette smoking on UCHT1 (CD3) positive T lymphocyte cell counts per mar in cervical epithelium showing changes of HPV infection

Regression Coeff S. E. LR statistic P-value

-0.0307 0.0103 9.77 0.002

LR = Likelihood Ratio - 90 -

The progressive reduction in the T lymphocyte cell count per unit area with increasingly severe histological changes of GIF with HPV infection supports the original findings of Tay et al (1987b).

Moreover, this larger series also suggests that this is primarily due to the reduction in T4 (CD4), or helper cells. Whether these numerical changes in the cell papulation are reflected in the local immunological function of the cervical epithelium remains unclear.

But, if it was, then this change would represent a suppression of the lymphocyte component of the cell mediated response to both viral and tumour antigens.

The finding that in epithelium with changes of HPV infection, cigarette smoking was associated with a reduction in total T cell count per mnr may reflect either a direct effect on lymphocytes or on the mechanisms which direct them to a particular site in the cervix. There is no evidence that cigarette smoke constituents exert any direct toxic effect on lymphocytes. However, the results given in section 3.1 describe changes in Langerhans' cells associated with smoking. As class II restricted antigen presentation is important for the recruitment and activation of T4 (CD4) bearing lymphocytes, it is clear that any alteration in Langerhans' cells may also have a marked effect on T lymphocytes.

In the statistical analysis of the data, the biopsies showing histological changes of Cliff with HPV infection were considered both - 91 - as one group and separately for each grade of disease. In the analysis of the results of UCHT1 (CD3) positive cells, the changes seen for GUI overall were not reflected in the changes detected within each grade of CIN. Whereas for T4 (GD4) positive cell counts

(Tables 3.9 and 3.10), the separate analysis of CIN grades 1 to 3 demonstrated a significant reduction in cell counts per mm2 from

CIN1 to CIN3. However, the division of the cases of CIN/HPV into the three separate grades of CIN reduces the numbers of cases in each group unequally. Only eleven and ten cases were present for CUT 2 and 3 respectively. This means that the results of the search for any significant changes between separate grades of CIN need to be treated with caution. In addition, the distinction between individual grades of CIN is a more contentious and potentially inaccurate issue than the overall diagnosis of CIN (Langley 1984).

Thus CIN/HPV constitutes a more appropriate and accurate category for use in the analysis of changes in immune cell counts.

The scarcity of B cells in the squamous epithelium of the cervix is in marked contrast to their occurrence in areas of columnar epithelium of the endocervical crypts. This is reflected by the endocervical glands being the major site of local antibody preduction for the cervical mucus. Further examination of this interesting and unexplored area of cervical immunology was outside the scope of this study. 3,3 A Study of the, effect of Possible Cofactors on the Macrophage cell count per mmP in Cervical epithelium

Immunocvtochemical techniques

The preparation and staining of sections were performed as described in section 3,2. The monoclonal antibodies used to identify the macrophages in the tissue were as follows;

3.9 is an IgG antibody to an antigen called the pl50,95 molecule, which is related to the human leucocyte differentiation antigens.

The expression of this antigen has been shown (Hogg et al 1986) to identify macrophages. Ho cross reactivity with Langerhans' cells or neutrophils has been observed. It has been assigned to the CDilc group of monoclonal antibodies (Hogg & Horton 1987). In this experiment, it was used at a strength of 1/5 dilution. An example of macrophages stained by this marker is shown in Figure 3.6.

Ell is an antibody which has been shown to identify the C3b receptor, CR1, which has been shown to have an increased expression on stimulated myeloid cells (Hogg et al 1984). It has been assigned to the CD35 monoclonal antibody class. It was used at a dilution of

1/2 in this study.

Both antibodies were kindly provided by Haney Hogg PhD of the

Imperial Cancer Research Fund Laboratories, who also provided technical advice. - 93 -

Results.

Both the 3.9 (CDllc) and Ell (CD35) positive macrophage cell counts

per mnr were found to be increased in cervical epithelium showing histological changes of HPV infection. This increase was significantly more marked with increasingly severe changes of CUT

(Figures 3.7 and 3.8).

The macrophage cell counts per unit area were not found to be significantly associated with any of the possible factors or cofactors examined, such as contraceptive use, semen exposure, cervical infections or cigarette smoking (Table 3.14).

However, HPV type 16 detection was significantly associated with an increased Ell (CD35) positive macrophage cell count per nrnr, using the log linear model, in CIH only (Table 3.15). Ho association of viral D M detection with 3.9 positive macrophage counts could be found in any of the lesions studied. 94 -

Figure 3,6; Kacrcpaages stained b.y.JL..£L (CDllcP marker in cervical epitleii.ua snowing .changes..a t .HPX.inla&lion. .

- +------+ ------+ ------+ ------NORMAL <---- * -----) HPV CIN1/HPV (— * -----) CIN2/HPV (------* ------) CIN3/HPV (------* ------) - + ------+ ------+ ------+ 0 16 32 48

Mean Macrophage cell count/mm2

Figure 3.7: Graph showing the means (with 95% confidence Intervals) of the 3.9 (CDllc) positive macrophage cell counts per unit area according to histology of epithelium.

NORMAL (----- *------) HPV (----* — -) CIN1/HPV (---- *--- > CIN2/HPV (-----—*------) CIN3/HPV (------) 0 10 20 30

Mean Macrophage cell count/mm2

Figure 3.8: Graph showing the means (with 95% confidence intervals) of the Ell (CD35) positive macrophage cell counts per unit area according to histology of epithelium*.. - 96 -

Table 3.14 The effects of Cigarette smoking on macrophage cell r.mmts per mnP according to cervical epi-lheliaL-Mstology

3.9 (CDllc) positive median cell counts per mnF

Histology Ion-smokers Smokers P value* lormal 7.59 (n=7) 2.13 (n=8) 0.35 IS

HPV 10.7 (n=15) 9.0 (n=17) 0. 97 IS

CIN/HPV 26.04 (n=13) 28.66 (n=29) 0.64 IS

Ell (CD35) positive median cell counts per mm2

Histology Ion-smokers Smokers P value* lormal 0.0

HPV 0.0 (n=15) 2.86 (n=17) 0. 42 IS

CII/HPV 9.23 (n=13) 4.52 (n=29) 0. 42 IS

* Using Vilcoxon test (IS = not significant) Table 3,15 The effects of HPV type 16 DM detscti.Qii.,._Qn macrophage.. cell counts per mm* according to cervical epithelial histology

3.9 (CDllc) positive median macrophage counts per mnF

Histology HPV 16 POS HPV 16 NEG P value

HPV 7. 91 (n=18) 13.91 (n=18) 0.24 NS

CIN/HPV 28.66 (n=27) 23.61 (n=25) 0.21 NS

Ell (CD35) positive macrophage cell count per nrnP

Histology HPV 16 POS HPV 16 NEG P value

HPV 0. 0 (n=18) 2.13

CIN 9. 27 (n-27) 3.96 (n=25) 0.05 *

* When the effect of HPV type 16 detection was examined for its effect an the Ell (CD35) positive macrophage cell count per mm2;

Regress Coeff S,E. L-R statistic LR p-value

0.8557 0.3705 5.78 0.02

N.B, LR = Likelihood ratio The results of this study support the findings of the smaller study by Tay et al (1987c) that both 3.9 (CDllc) and Ell (CD35) positive macrophage cell counts per mm* are increased in cervical epithelium showing histological evidence of HPV infection. However, this larger study demonstrates that Ell (CD35) positive, activated macrophages, are significantly increased in concentration in lesions showing CIN with HPV infection. The additional finding that HPV type 16 infection is associated with an increase in these same activated macrophage cell population suggests a relationship between the presence of this oncogenic viral type, the local macrophage response and the histological severity of the precancer.

The functional significance of these changes in epithelial macrophage populations can only be tentatively interpreted. The suggestion that macrophages may possess a direct anti-tumour activity (Adams & Snyderman 1979; Harman 1985) may be related to their apparently increased infiltration and activation in higher grade intraepithelial neoplasia. However, whether there is any functional association between the relative increase in macrophages with antigen presenting capability, and the reduction in Langerhans' cells (section 3.1) still remains unknown.

There is also evidence that in certain situations, macrophages may promote tumour growth (Khadim & Rees 1984). The increased stimulated macrophage cell counts per mm* found in CIH/HPV lesions in cervices - 99 -

from which. HPV type 16 D M was detected may be of significance in

this context. But whether the reported higher oncogenic potential

possessed by lesions involving this HPV type is related to

macrophage infiltration must await the development of more detailed

functional and kinetic models for the study of local cervical epithelial immunity. - 1 0 0 -

3.4 A Stndv of the effects of Passible Cofactors.on Natural killer

(M) Cells in Cervical epithelium

Immunocytochernical. methods

The techniques of section preparation and staining were described in detail in section 3.2. The monoclonal antibodies used were;

Leu-7 (HNK-1) has been shown to identify NK cells in peripheral blood (Abo & Balch 1981) and frozen tissue sections (Ritchie et al

1983). It was used at a dilution of 1/2.

Leu-11 (M-15) has been also been shown to demonstrate NK cells

(Phillips & Babcock 1983). It was used at a dilution of 1/2.

The use of two different monoclonal antibodies is useful, because

Leu-7 may also stain some suppressor lymphocytes, whereas Leu-11 does not have such cross reactivity, but has less capability in detecting all of the NK cells present (Lanier et al 1983).

Due to the low numbers of positive staining M cells detected in both normal and abnormal cervical epithelium, cell counts per mnr were not calculated. Instead a grading score was used as follows;

0 = none seen in the epithelium

+ = < 10 cells in the epithelium

++ = 10-20 cells in the epithelium

+++ = > 20 cells in the epithelium -101-

Results.

Whereas HK cells identified by either Leu-7 or Leu-11 were rarely detected in normal epithelium, there was a significant increase in the presence of both cell papulations in CIH/HPV lesions (Table

3.16). In epithelium showing changes of HPV infection alone, a small number of specimens had a few Leu-7 positive cells present, but none of these had any Leu-llB positive cells identified.

The changes in HK cell population were examined for any association with cigarette smoking, hormonal contraception, semen exposure and

HPV type 16 detection. Ho effect of any of these variables could be discerned on the UK cell counts by either monoclonal antibody marker.

The breakdown of the effects of semen exposure on the histological grade of cervical abnormality are shown in Table 3.17. Ho significant difference is observed between the grade of epithelial abnormality in those cervices exposed to recent semen exposure, compared to the non-exposed group. Furthermore, no significant effect was observed of semen exposure on the epithelial HK cell count (p=0.21 for Leu-7 stained HK cells and p=0.59 for Leu-11 stained HK cells). -102'

Table 3,16 KesultS-.Qf. NK cell identification according to histology nf cervical epithelium LfiU-Z.

Normal HPV CIN1/HPV CIN2/HPV CIN3/HPV

0 15 26 15 1 1

+ 1 9 13 5 5

++ 0 1 3 5 2

+ + + 0 0 0 0 2

Amalgamating the ++ ,and +++ groups, the test for trend (McCullagh <

Nelder 1983) for increasing NK cells with increasing disease gives

X*=39. 47, pCO.OOOl.

Leu-U.

Normal HPV CIN1/HPV CIN2/HPV CIN3/HPV

0 15 36 14 0 4

+ 1 0 16 9 4

++ 0 0 1 2 1

+++ 0 0 0 0 1

Amalgamating the ++ and +++ groups, the test for trend (McCullagh &

Nelder 1983) for increasing NK cell counts with increasing disease gives x :2=4 1 . 7 2 , p<0,0001. -103-

Table 3,17 Numbers of women in semen exposed and non-exposed groups according to histological findings.

Normal HPV CIN/HPV

Exposed 13 26 40

Non-exposed 3 10 12

[Exposed=No barrier contraceptive and regular intercourse more than once per week over previous month.

Non-exposed=Barrier contraception or no intercourse for previous month.J -104-

Dis^usslm

This study confirms, using a much larger series of cases, the

preliminary findings of Tay et al (1987d) that UK cell infiltration

occurs in epithelium showing histological changes of premalignancy

with HPV infection. However, in contrast to Tay's findings, there

was no evidence of any Leu-11 positive UK cell infiltrate in

uncomplicated HPV lesions and only minimal evidence of a Leu-7

positive UK cell infiltrate. These findings may be important because

of the significant role played by UK cells in the response to the

development of malignant cells (Moore 1985). Whether or not the

degree of UK cell response is related to the oncogenic potential of

a particular CIU/HPV lesion merits further investigation in more

detailed longitudinal studies. Morever, the need to examine the

possible causes and effects of NK cell infiltrates with cells

bearing different surface antigens is also important.

Also in contrast to Tay et al (19S7d), this study measured UK cells

situated only within the cervical epithelium. Although UK cells were

also seen in the subepithelial stroma, the significance of these

cells in local cervical immunity is even less well established than

for cells within epithelial lesions. The assessment of the changes

in UK cell distribution and their functional significance requires

further studies of the detailed local immunology of epithelial

surfaces. Of particular interest is whether the epithelial and subepithelial responses alter significantly once a CIU lesion has

become microinvasive. This question is currently being addressed by - 1 0 5 - a study in the laboratory where the candidate performed the work for this thesis.

Despite the evidence that seminal plasma may suppress UK cell activity in vitro (James & Szymaniec 1985; Tarter et al 1986), this study found no association between the UK cell numbers and historically reported semen exposure. One explanation for this may lie in the important difference between detectable cell number and cell function, i.e. the presence of immunocytochemically detected UK cells does not mean that UK cell function is normal. The investigation of such basic parameters of immune cell function awaits the development of monoclonal antibodies specific for the antigenic determinants of cell activity and kinetic experimental models of local immunology.

A further reason for this study's failure to show any association between cervical semen exposure and UK cell counts may lie in the method of analysis used. This involved grouping together those women not currently sexually active with those using barrier contraception into the 'non-semen exposed group', and comparing them to all the other sexually active women in the study. This method took no account of the amount and frequency of semen exposure, which may be important in the aetiology of cervical cancer (Turner et al 1988).

Further work into the effects of seminal exposure on pre-existing

CIU/HPV lesions is long overdue since the work of Richardson & Lyon

(1981), who claimed that condom use alone produces regression of the majority of CIU lesions. - 1 0 6 -

In summary, this study provides further evidence for the existence of an NK cell infiltrate in the local immune response to the development of CIU in HPV infected epithelium. However, the functional and prognostic significance of this response must await further investigation. -107-

3.5 Cervical Infections and Alterations in Local Immunological Cell

Populations in the Cervical epithelium

As described in Part 2, all the women studied were screened for the presence of cervical infections. Forty nine infections were detected in 47 women. The results are shown in Table 3.18.

Table 3,1.8 Prevalence at Cervical infections detected in study group

Detection in;

Normals HPV+CIN Total

n=36 n=94 n=130

Anaerobes 2 15 17

C. albicans 3 12 15

C. trachomatis 1 8§ 9

T. vaginalis 1 2 3

£ Haemolytic streptococci 1 1 2

Herpes simplex virus 1* 1* 2

N. gonorrhoeae 0 1 1

* Only 5 normal controls and 20 women with cervical abnormalities were screened for HSV detection.

§ Two patients with C. trachomatis infection also had anaerobes detected. -108-

The overall rate of detecting a woman with a current cervical infection in this study group was 36%. But the incidence of detecting an infection in the group of women with abnormal cervices was 43%, compared to only 25% in those with normal cervices.

Although this sample is small, it demonstrates the clinical value of screening women with one sexually transmitted infection

Even if it is argued that the detection of anaerobes or C. albicans in the lower genital tracts of asymptomatic women may not be clinically important, nor require treatment, this still leaves 11% of the women screened as having a clinically important, transmissible pathogen which does require treatment. Indeed, the often subclinical nature of both gonococcal and chlamydial infection makes such screening of special importance. Furthermore, the value of tracing and treating the sexual partners of patients with transmissible infections has been clearly shown (Bistoletti &

Lidbrink 1988).

In the quantitative immunocytochemical study of cases with concurrent cervical infection, described in sections 3.1-3.4, no statistically signficant effects of particular infections on local cervical immunology could be discerned. However, even though the numbers of cases of particular infections were small, certain observations were made which bear reporting in more detail. -109-

Herpes.Simplex virus (HSV) infection

In both cases in which cervical cultures were positive for HSV, the

patients were currently asymptomatic, However, both had

colposcopically visible areas of cervical ulceration. Although both

women were non-smokers without HPV type 16 D M detected, it was

noticeable that T6 (CD1) positive Langerhans' cells were noted to be

absent from areas of ulceration, but present in areas of normal

epithelium.

This finding is compatible with the evidence that HSV depletes

Langerhans' cells in the skin. The possible role of HSV in the

aetiology and progression of cervical neoplasia has been discussed

in Part 1, but this local immune effect enhances the possibility of

such a role, especially in view of the apparently high prevalence of

subclinical asymptomatic episodes of recurrent cervical HSV

infection (Barton et al 1986; 1987). Whether the metaplasia

initiated by recurrent cervical ulceration or the accompanying local

epithelial immunosuppression might be the important factor in this

association merits further investigation using immunocytochemical

methods. Unfortunately, the number of cervical biopsies available

from women with concurrent culture positive recurrent HSV infection

is limited. This problem might be circumvented by the use of new direct monoclonal antibody tests for HSV specific glycoproteins

(SYVA, UK). -1 1 0 -

Chlamy.dla trachomatis.

All except one of the patients with this pathogen detected were

asymptomatic. No effect of this agent on the Langerhans' ceil count

was apparent, which supports the larger study of Syrjanen et al

(1987). However, an increase in T lymphocytes was noted in both

normal and abnormal epithelium, the latter in deference to the trend

towards a decreased T cell count in epithelium showing changes of

CUT and HPV. In these infiltrates, the majority of cells were found

to be T8 (CD8) positive. Whether this influx of suppressor T cells

might relate to the formation of lymphoid follicles (Hare et al

1981) which have been described in chlamydial cervicitis is

uncertain. The potential role which this change might play in the

local immunosuppression suggested as being associated with

chlamydial infection (Levitt & Barol 1987) merits further

investigation. Furthermore, the potential for suppressor T cells to

enhance the neoplastic process (Hoar 1979) also lends weight to the

need to investigate this observation more closely.

Candida albicans

Hine of the 12 women with CIIT/HPV lesions who had C. albicans

isolated from ectocervical swabs were asymptomatic. In seven of these 9 cases, it was noted that despite a marked 3.9 (CDllc) positive (total) macrophage cell infiltrate, there were no Ell

(CD35) positive (activated) macrophages present in the epithelium.

This is in contrast to the usual finding of a significant fraction of activated cells amongst the total macrophage cell numbers in

CIH/HPV lesions. This apparent reduction in the activation of 1 1 OA-

Table 3, ISA Median CD3. CD4 and CDS T lymphocyte cell counts per mm3 in cervical epithelium according to detection of Chlamydia trachomatis (Ct.) infection,

Ct present (n=9) Ct negative (n=27) p-value

CD3 72,44 54. 80 0. 05

CD4 24.40 23. 65 0.89

CDS 52.80 26.56 0. 04

See Section 3.2 for details of lymphocyte identification techniques.

Table 3.18B Median CD3. CD4 and CDS T lymphocyte cell counts per mm3 in cervical epithelium according to detection of Candida albicans infection,

Candida albicans detected (n=15) llo infection present (n=27) p-value

CD.3 59.60 54.80 0.51

CD4 22,90 23. 65 0.72

CDS 26,48 26, 56 0.93

See Section 3,2 for details of lymphocyte identification techniques. - I l l - macrophages in women with candidal vaginal infection may be a chance

finding in such a small sample. However, Vitkin et al (1986; 1987) has proposed that vaginal candidiasis is directly related to a functional macrophage defect. He also proposes that symptomatic, recurrent candidiasis is due to macrophage mediated hypersensitivity leading to Pg Es production and T lymphocyte responses. It is interesting to note that in the three symptomatic cases there were both 3.9 (CDllc) and Ell (CD35) positive macrophages present, but with no significant difference in their concentrations compared to the asymptomatic group. Ho differences in the T cell counts per mm2 CTtLkirtty were noted. The significance of these findings merits further investigation using a larger number of women, microbiological sampling from both cervix and vagina and more examination of women without CIH/HPV. Further delineation of the relationship between the cervical changes noted here and 'vaginal' candidiasis is indicated.

In summary, although no statistically significant changes in local immunity were identified in this study, local alterations in certain cell types were observed in cases of HSV, C. albicans and C. trachomatis infection. It is important that further studies of these infections are directed towards obtaining a greater understanding of the host immune response to these agents and how this affects the clinical outcome of the infection. Furthermore such changes emphasise the necessity of including screening for sexually transmitted infections in the methodology of all future studies of cervical epithelial immunological cell distribution. -112^-

3.6 A Stndv of cotinine concentration in Cervical mucus

lalTfldurilQIL

Previous studies have demonstrated that significant concentrations of cigarette smoke constituents may be detected in the cervical mucus of women who smoke tobacco. Both nicotine and its metabolite, cotinine, have been shown to be found in higher concentrations in cervical mucus than in serum (Sasson et al 1985). Hellberg et al

(1988) also detected both nicotine and cotinine in the cervical mucus of non-smokers in the absence of measurable serum levels.

These findings provide evidence for a significant local effect of cigarette smoke constituents on the cervix and demonstrate the potential importance of passive smoke exposure. Although there is little evidence for any carcinogenic or immunosuppressive effect of either nicotine or cotinine (Hoffman et al 1986), it has been suggested that their presence in mucus may indicate that other hydrophilic constituents of cigarette smoke, with carcinogenic actions, may be concentrated locally (Sasson et al 1985).

To further investigate the role of the local concentration of cigarette smoke constituents in cervical mucus and the development of Cliff, the following study was performed by the author. The following questions were addressed in the data analysis:

1) whether the cotinine concentration in cervical mucus was significantly associated with the degree of histological abnormality in the cervical epithelium. — 113—

2) whether women who reported that they were ex-smokers or non- smokers had evidence of cotinine in cervical mucus which might be attributed to passive exposure, and whether this was significantly different in women with CIH7HPV.

Patients and Methods

One hundred women attending the same clinic as described in section

2,1 were considered for inclusion into this study. Each subject had a full history taken and gave her informed consent. Colposcopic examination was preceded by taking the following specimens;

-Cervical smear for cytology

-Cervical smear for HPV DUA analysis

-Mucus aspiration using a fine bore pipette and syringe from the cervical os and canal. All available mucus was collected and placed in a Nunc tube and snap frozen in liquid nitrogen.

-Endacervical swabs for the detection of bacterial, yeast and chlamydial infection.

A thorough colposcopic examination was then performed and biopsies taken from colposcopically visible abnormalities or a representative area of the transformation zone as described in section 2.3. - 1 1 4 -

Results of the analysis of the cytology specimens, HPV type 16 D M

detection and microbiological identification were provided by the

co-workers cited in section 2.4. Standard techniques were used to

produce histological sections for staining with haematoxylin and

eosin which were assessed by the author, in conjunction with Dr. D.

Jenkins, using the histopathological criteria described in section

2.4.

The mucus specimens were stored in liquid nitrogen for up to 9

months before analysis. In April 1988, the candidate transported the

samples to the American Health Foundation, Hew York, USA, where the

Biochemistry Department (Figure 3.9) has established a recognised

method for the assay of both nicotine and cotinine. Under the

supervision of Dr. Haney Haley, the candidate was able to perform

the extractions and measurements of cotinine for the following

experiment.

A decision to measure only cotinine was taken because of its greater

significance in representing the overall cigarette smoke exposure,

due to its longer half life than nicotine (22 hours compared to 4

hours) (Sepkovic et al 1987). This feature also makes cotinine a more significant measure of passive smoke exposure.

The mucus was allowed to thaw at room temperature before the volume of each specimen was measured and recorded. The cotinine was extracted from the mucus sample by 20 minutes of sonication in acetone, followed by filtration through a 4p pore filter and -115

concentration under nitrogen. The extracted residues were then

resuspended in acetone. A radioimmunoassay for cotinine was

performed according to the method of Langone et al (1973). Samples

spiked with pure cotinine have been used to demonstrate that this

method retrieves 94% of the available cotinine, whereas blank

samples show no detectable trace (Haley et al 1983). The concentrations of cotinine in each sample were calculated by dividing the amount detected by the volume of mucus in the original specimen, expressing the results in ng/ml.

The Vilcoxon test for trend (Cuzick 1985) was used to test the data for significant associations.

Figure 3.9: The American Health Foundation Laboratories. - 1 1 6 -

R^sulis.

Of the one hundred consecutive women considered for inclusion into this study, 34 were excluded because of the absence of collectable mucus or an identifiable cervical infection; thus sixty six women were studied.

The measured volumes of cervical mucus aspirated ranged from 50pl to

800pl, with a mean of 450pl. The cotinine concentrations measured by radioimmunoassay varied from 3 ng/ml to 1154 ng/ml. In the twenty four women who gave a history of being non-smokers, the mean cotinine concentration in the cervical mucus was 153 ng/ml, with only three women having undetectable levels. Furthermore, in twelve women who claimed to be ex-smokers, the mean cervical mucus cotinine concentration was 311 ng/ml (range 37-755 ng/ml). In stated current smokers, the mean cotinine concentration was 628 ng/ml, but there was no dose related association found between the number of cigarettes smoked and the cotinine concentration in cervical mucus.

Ho significant association was detected between the cigarette smoking history or the cervical mucus cotinine concentration and the grade of cervical histological abnormality (using five groups from normal through HPV to CIH 1/2/3).

To further examine the effects of cigarette smoke on the development of CIU, the cases were subdivided into two groups based on the identification of HPV type 16 D M from the cervical smears; this produced two groups (Tables 3.19 & 3.20). - 1 1 7 -

Table 3. IQ The range of histological findings in colposcopis. biopsies taken from cervices with dgtectablg_HP_l_tvpe 16 ML

Normal 10

HPV 10 c m 4

CIN2 7

CIN3 7

Total 38— consisting of 15 non smokers

8 ex-smokers

15 smokers

Table 3,20 The range of histological findings in_colposcopic biopsies taken from cervices without detectable HPV type lb DM.

Normal 7

HPV 2

CIN1 10

CIN2 5

CIN3 4

Total 28---- consisting of 9 non-smokers

4 ex-smokers

15 smokers “ 117A-

Table 3.20A Median cotinine concentrations (ng/ml) according to HPV type- 16 detection and histology of accompanying cervical epithelial biopsy.

Formal HPV CIF1/HPV CIH2/HPV Cl 113/HPV p-value#

positive 42.6 157. 5 137. 5 253.5 531.3 0. 04

negative 206.7 284.4 143. 3 123.3 147.5 0.57

*Using ViIcoxon test for trend. - 1 1 8 -

It is of n o te that ten women had cytologically, colposcopically and histologically normal cervices in the presence of detctable HPV type

16 DNA. There were no significant differences between the groups in terms of age, contraceptive usage, reported smoking histories or amounts of mucus collected.

In analysing the two groups separately using the test for trend, a significant association was found (p<0.05> between the cotinine concentration in the cervical mucus and the degree of histological abnormality present, which was only significant for those women with detectable HPV type 16 DITA. This association was not found to be due to reported smoking history nor confounded by the volume of the mucus samples obtained.

The findings of this study support the tentative conclusion of

Hellberg et al (1988) that cervical mucus cotinine concentration is not directly related to the number of cigarettes smoked by the subject. This may be due to several explanations. Firstly, that the smoking history is flawed by the subject underestimating the number of cigarettes smoked, especially when telling a doctor who they can reliably expect to disapprove of smoking. Secondly, because the

1 number smoked' does not take into account the actual amount of inhaled products received; this depends on the degree of inhalation and the amount of each cigarette smoked, as well as the ventilation of the area in which the smoker is sited (which influences rebreathing of smoke as additional passive exposure). Thirdly, the - 1 1 9 -

possibility exists that individual women may concentrate cotinine to different extents in cervical mucus, due to unknown variables.

Furthermore because cotinine is a metabolite of nicotine, it is possible that certain cervices may have different enzymatic capacities to produce cotinine locally, via variations in their local oxidation systems. Further studies of smokers under controlled conditions are necessary to further investigate this problem.

The demonstration that in women whose cervices are infected by HPV type 16, there is a dose dependent association between the cervical mucus cotinine concentration and the development (and severity) of

CINVHPV, supports the model for oncogenesis in the cervix (zur

Hausen 1986). It suggests that cervical cotinine concentration may be a useful biological marker for the exposure of cervical epithelium to other locally active agents which promote oncogenesis in the presence of a particular viral infection. However, the lack of any association in women without detectable HPV type 16 DHA produces several unanswered questions, whose answers are confounded by the lack of identification of which HPV types are present in these cervices. Whether the biological significance of cervical cotinine is itself influenced by HPV type 16 infection, or if the progression of CII/HPV in women without HPV type 16 is related to cigarette smoking and whether these lesions have a different prognosis to those women with HPV type 16 infection remains uncertain. The detection of cotinine in cervical mucus, and its possible

relationship to the development of CIS should not be taken to

suggest that this chemical has a direct role in causing malignant

change. In particular, in the search for a constituent or metabolite

of cigarette smoke which might be responsible for the changes found

in cervical Langerhans' cells in smokers (section 3.1), cotinine is an extremely unlikely candidate. However, a future study is planned to compare the epithelial Langerhans' cells per mmr’ to the mucosal and systemic concentrations of cotinine in the same woman.

In conclusion, despite the problems of interpreting the findings, it appears that women with evidence of cervical HPY type 16 infection have an increasing risk of developing more severe CIN/HPV according to their cervical mucus cotinine concentration. The finding of this cigarette derived substance in the cervical mucus of non-smokers and ex-smokers provides supportive evidence for women to be encouraged to both stop smoking and to reduce their environmental exposure to passive cigarette smoke. .21-

EAEJ 4; CONCLUSIONS

4.1 Changes in local epithelial immunological cell populations In

HPV and CIN lesions of the cervix

Before considering the effects of possible cofactors on cervical epithelial immunity, it is necessary to examine the changes which occur in HPV infection and CIN/HPV lesions. Only against this background can the significance of the various changes associated with particular cofactors be placed in perspective.

The results described in sections 3.1 to 3.4 include details of the local populations of immunologically important cells in normal cervical epithelium. It has been demonstrated that in normal cervical epithelium there is a network of interdigitating dendritic

Langerhans' cells. The other main antigen presenting cell group are the macrophages, but there are few of these cells in normal epithelium. Vhereas B lymphocytes are rarely found in normal epithelium, T lymphocytes are often found in the lower third of the epithelium, with both T4 (CD4) and T8 (CD8) bearing cells being found in roughly equal numbers. Natural killer cells were shown to be rarely detectable in histologically normal cervical epithelium.

In epithelium showing evidence of HPV infection, without any degree of ClJT being present, the Langerhans' cell count per unit area was shown to be depleted. Similarly, the total T cell count was - 1 2 2

diminished, primarily due to a loss of T4 (CD4) helper T cells. The

macrophage cell count per unit area was increased, along with the

proportion of activated macrophages. There was no significant

natural killer cell infiltrate noted.

In cases of CIN with HPV infection, there was a more marked

reduction in Langerhans' cells, and those present had a

characteristically altered morphology, with loss of the dendritic

appearance. T cells were significantly reduced in number, with the

T4 (CD4) subset particularly depleted. Natural killer cell

infiltration was found to be associated with increasingly severe grades of CIN.

These findings confirm most of those reported in the smaller studie

of Tay et al (1987a-d). However, certain important differences should be noted. Firstly, the lack of a significant natural killer cell infiltrate in uncomplicated HPV infection, compared to that in

CIN/HPV is important, as this may relate to the role of these cells

in responding to the expression of neoantigens on the surface of epithelial preneoplastic cells. Secondly, the increased activated macrophage cell count per nrnr seen in CIN/HPV is not present in uncomplicated non-HPV type 16 infection. As with natural killer cells, this may reflect changes in the antigenicity of tumour cells

In addition to studying significantly larger numbers of women than previous studies, the methodology used here includes the choice of more relevant controls. The criteria for 'normal control' used in this study specified that each case was cytologicaily,

colposcopically and histologically normal. This is a unique feature

and the inclusion of colposcopy should now be a standard which all

subsequent studies should follow.

Despite the stricter methodological criteria used in this study,

there remain differences in the methods used by different groups of

workers to quantify the numbers of immunological cells in the

epithelium. Two main types of methods have been used for this;

either using the cell count per unit area, as in these studies, or

the cell count per fixed number of cells in the epithelium. Both methods can be criticised. The former on the grounds that any

increase in cell size (e.g. as in koilocytosis) might increase the measured area of the epithelium and hence artefactually result in a decreased cell count per unit area. The latter, on similar grounds due to the cellular proliferation which occurs in both HPV and CIN, which might produce artefactual regional changes in the denominator used in the calculation of cell count, and also because other immunological cell infiltrates may affect the number of cells counted. Indeed, the difficulty in determining what constitutes 'a cell' which should be counted in tissue sections can often be difficult and would be expected to lead to greater error in the latter method.

Overall, neither of these methods should be regarded as perfect, but whichever is chosen for a particular experiment should be performed by one observer with reliability, and with proven repeatability by - 1 2 4 -

others. In the laboratory in which these experiments were performed.,

these criteria have been found to apply to the methods used in the

current studies, and in conjunction with the results discussed in

section 3.6, demonstrate the need to consider as many variables as

possible in searching for potential spurious explanations of

observed associations.

The difficulties in counting cells in quantitative histology are discussed in more detail by Beck & Anderson (1987) who also examine another potential pitfall due to the section orientation on the slide and the effects of freeze-thawing or fixation. Although Reid et al (1987) have examined the effects of freezing on the diffusion of cellular antigens, there is little work on the potential errors induced by these procedures, and though there is no theoretical reason why any errors should be confined to either the study or control groups, it is important that these factors be examined more closely in future studies.

4.2-The.„effects of the Possible Cofactors studied on the distribution of Immunologically active cells in Cervical epithelium

This thesis has sought to examine the local immunological cell pattern of the cervical epithelium for changes associated with the possible factors and cofactors which have been linked, by epidemiological studies, to cervical neoplasia. - 1 2 5 -

The most important finding is the observed association between

current cigarette smoking and a reduction in the Langerhans' cell

counts per unit area in both normal and abnormal cervical

epithelium. The probable functional consequence of this numerical

change in the availability of antigen presenting cells will be to

reduce the capacity for the presentation of viral and tumour

antigens to T lymphocytes. It is possible that the additional

finding of an association between cigarette smoking and a reduced total T cell count per unit area in HPV infected epithelium is related to defective antigen presentation, without which the production of important lymphokines and interleukins responsible for the local recruitment of T lymphocytes may be impaired. However, the

further investigation of the exact functional interaction between

Langerhans' and T cells must await the development of monoclonal antibodies to antigens which represent markers of cellular activation and maturation. In addition, the increasing use of double staining techniques can be used to demonstrate spatial relationships and intercellular contacts between different groups of immunological cells.

A more complex role for Langerhans' cells in the aetiology of epithelial disorders has been proposed (Horhenn 1988). This suggests that Langerhans' cells exert a regulatory effect on keratinocyte and fibroblast proliferation. The mechanism of action is thought to involve chemical mediators; both interleukin-1 (from Langerhan's cells) and gamma interferon (from activated T cells) have been proposed for this role. Any reduction in Langerhans' cells will, 126- both directly and indirectly, reduce the production of both these substances. The net effect would be to reduce the control of keratinocyte replication and maturation; the loss of which are features of both HPV infection and CUT. Whether Langerhans' cell depletion might be related to these histological appearances awaits further investigation.

Having established that there is an association between cigarette smoking and a depletion in the concentration of Langerhans' cells present in both normal cervical epithelium and CIN/HPV lesions, the question arises of which constituent or metabolite of cigarette smoke might cause this effect. From the incandescent tip of a lighted cigarette, burning at a temperature of 800“C, the smoker draws into their mouth a collection of several hundred hot gases and particles with each draw. The major constituents in each of these phases are carbon monoxide, as a gas, and nicotine and tar as particles. After buccal, pulmonary and gastric absorption, several hundred more metabolites are produced in the process of excretion

(Ashton & Stepney 1982). From this plethora of chemicals, several carcinogens have been identified (Table 4.1). Although no evidence exists to directly implicate any of these with a specific toxic effect on Langerhans' cells, other carcinogens, such as

7,12-dimethylbenz(a)anthracene, have been shown to exert such an effect on these cells (Muller et al 1985).

Indeed, many agents have been shown experimentally to result in a reduction in the number of epithelial Langerhans' cells. These - 1 2 7 -

include PUVA therapy (Friedmann et al 1951), topical and systemic

glucocorticoids (Belsito et al 1982), cytotoxic chemotherapy (Pakes

et al 1986), cutaneous HSV and HPV infection (Chardonnet et al

1986). Recently, it has been noted that systemic infection with

human immunodeficiency virus (HIV) is associated with a reduction in

cervical Langerhans' cells (Dreno et al 1988). The data presented above suggest that this association might explain the reported higher incidence of cervical neoplasia in women with HIV infection

(Bradbeer 1987; Stewart et al 1988). Although none of the women studied in this work had clinical evidence of HIV infection or AIDS, it may be necessary to incorporate testing for HIV antibody into future studies of the local cervical immune response, in order to exclude this infection as a variable. This will depend on the background prevalence of HIV infection in the group of women being studied. Indeed, such a precaution may be advisable in studies utilising frozen material to ensure that safe laboratory procedures are used, especially in the light of the detection of HIV in cervical tissue of infected women (Pomerantz et al 1988). It is interesting to note that cigarette smoking has recently been proposed as having a role in determining the risk of progression from asymptomatic HIV infection to AIDS (Berger 1988): whether the effect of smoking on the cervical epithelial immunity is additive to the changes induced by HIV infection is being studied by the candidate at present.

The significant association between the concentration of cotinine in cervical mucus and the severity of GIN in cervical epithelium .28- infected by HPV type 16 has been described in section 3.6. The observation that detectable levels of cotinine were measured in cervical mucus from women who stated that they did not smoke is an important finding, providing evidence that passive smoke exposure may exert an effect in non-smoking women. Although passive smoke exposure has been shown to be important in the aetiology of lung cancer (Kuller et al 1986; Vald et al 1986), it must be remembered that smokers themselves are exposed to much passive smoking as they tend to congregate in places with other smokers, However, the passive exposure of reported non-smokers may be an important variable which has been little explored in epidemiological studies of cervical neoplasia.

If a reduction in Langerhans' cells in cervical epithelium reduces the cell mediated immune response and predisposes to oncogenic change, it is important to consider the potential effects of reversing this change. Maclean (1984a; 1984b) demonstrated that there is a marked appearance of Langerhans' cells in cervical epithelium which is healing after laser therapy. It is passible that this phenomenon constitutes an important therapeutic effect of such local ablative therapy, both in enhancing the immune response to residual HPV infection and preventing new premalignant change.

The only chemical experimentally found to increase the numbers of epidermal Langerhans' cells is the aromatic retinoid, etretinate, administered orally to mice (Shiohara et al 1987). The changes were reversible after cessation of administration. The therapeutic -1 2 9 -

possibilities of this vitamin A derivative await further

exploration. However, it should not be assumed that an excess of

Langerhans' cells will be beneficial; there is increasing evidence

to implicate Langerhans' cells as the major cell line proliferating

in Histiocytosis X (Uezelof et al 1973).

Although this study has concentrated on tobacco smoking, in some

areas of the world tobacco is ingested predominantly by chewing the

leaf. One such area is India, where a study (Chakravarti et al 1987)

examined the incidence of cervical cancer in relation to tobacco

chewing in women. The results showed that tobacco chewing was associated with a threefold rise in the incidence of cervical

neoplasia. This association merits further investigation, especially as the differences in the ingested constituents between smoked and chewed tobacco might be explored to further examine the functional role of carcinogenic or immunosuppressive agents obtained via each method of ingestion. Indeed the work of Hoffmann et al (1986) suggest that nicotine, obtained by either inhalation or ingestion, may play a part in facilitating carcinogenesis by the production of

N' nitrosonornicotine and 4-(methylnitrosamino)-l-(3-pyridyl)-1- butanone. These nitrosamines are then activated to the promutagenic

0s-methylguanine. Whether such substances are produced in cervical mucus of either tobacco smokers or chewers remains to be seen. Table 4,3 Carcinogens detected in the smoke of nonfilter cigarettes

Vapour phase;

Benzene

Hydrazine

Formaldehyde

Vinyl chloride

Acrylonitrite

2-nitropropane

Hitrosodimethylamine

Hitrosodiethylamine

Hi trosodipyrrolidine

Particulate phase;

2-toluidine

4-ami nobiphenyl

1- naphthylamine

2- naphthylamine

Arsenic

Hickel

Cadmium

Polonium 210 Source; Hoffman et al 1987 - 1 3 1 -

A1though the evidence for a true effect of cigarette smoking on the

Langerhans' cell counts per unit area seems probable from the data

presented in section 3, 1, it is important to consider the

possibility of confounding variables which have not been controlled

for in the study design. It is possible that cigarette smoking is

associated with another variable which is really responsible for the

effect; such a variable might be stress. Indeed, Baker et al (1985)

found that stressed individuals had altered systemic T cell subsets.

Furthermore, it is possible that an inherited personality characteristic (which may encourage smoking) might be associated with a particular HLA-type and hence play a role in determining immune function and the response to viral antigens. However, the dose dependent relationship between the number of cigarettes smoked and the Langerhans' cell count per unit area which was detected by log linear modelling would be unlikely to occur if this were so.

It is also possible that the nutritional status of the women studied may be related to their immune response to cervical neoplasia (Orr et al 1985). There is a paucity of data available on this subject, and the methodological problems inherent in such studies are numerous. However, the results of section 3.1 may be used to suggest that vitamin A should be singled out for initial investigation, because of the association of its deficiency with cervical neoplasia

(Romny et al 1981) and the previously noted effect of retinoids on the cutaneous Langerhans' cell count (Shiohara et al 1987). -132-

The lack of any significant changes associated with the other possible factors and cofactors examined in this thesis does not exclude these variables from having a different mechanism of action on the aetiology of cervical neoplasia. As was discussed in Part 1, it is passible that hormonal contraception exerts an effect by increasing the size and/or the susceptibity of the transformation zone to potential mutagens. Moreover, the role of semen may have been obscured by variables in the experimental methodology and analysis which are discussed in section 3.4.

In contrast, cervical infections were observed to have possible effects on certain immunological cell populations in the cervical epithelium. The most important of these was the effects associated with the detection of HPV type 16 DMA using a filter in-situ DMA-DMA hybridisation technique. This analysis was performed by an experienced co-worker in the department in which it was first described (McCance et al 1986). Despite debate over the accuracy of this technique (Lancet 1988), the major weakness of this finding lies in the fact that only one HPV type, of aver 50 described, was tested for. This means that in attributing any particular histological or immunological change to HPV type 16, one is unaware whether the effect is associated with the presence of this viral type alone or in combination with one, or more, other types of HPV.

In particular, this methodology cannot exclude the possibility that the observed epithelial immunological changes are due to dual infections with HPV types 16 and 18, both associated with an increased risk of cervical neoplasia. However, the fact that local immune ceil population changes were found to be significantly associated with the detection of HPV type 16 D M in both HPV and

CIMHPV lesions lends support for its postulated role in the development of cervical neoplasia. Furthermore, the relationship between cotinine concentration in cervical mucus and the grade of

CIU in HPV type 16 infected epithelium suggests that an interaction between these two factors favours oncogenesis. It is interesting to note that there is evidence to implicate papillomavirus infection,

Langerhans' cell depletion and the aetiology of oesophageal squamous cell carcinoma (Morris & Price 1987).

Using the results described in this thesis, it is passible to speculatively construct a model for the mechanism of development of cervical neoplasia (Figure 4.1). This suggests that a reduction in

Langerhans' cells in the normal cervical epithelium, due to active or passive cigarette exposure, facilitates the establishment and persistence of an HPV infection. Local host cellular immunity might be further diminished by the presence of concurrent infection with

HSV or Chlamydia trachomatis.

Once HPV infection is established in the epithelium, its individual subtype may possibly determine whether there is an additional negative effect on Langerhans' cells. Moreover, cigarette smokers will have less T lymphocytes available to respond to viral antigen presented by Langerhans' cells. If the HPV infection contains HPV type 16 DM, then a separate reduction in Langerhans' cells available to detect and present viral antigens will occur. Thus - 1 3 4 -

cigarette smoking and HPV type 16 may act separately, but synergistically, to reduce the available local immune response to

viral antigens.

Assuming that HPV type 16 infection may eventually induce nuclear abnormalities and cellular dedifferentiation, CIIT will result within the existing HPV infection. Ifeoantigens, such as Cal (Fray et al

1984; Lloyd et al 1984) will then become present on these premalignant cells. The recognition of these neoantigens may also be impaired by the lack of available antigen presenting cells in either cigarette smokers or if HPV type 16 infection is present. However, the observed infiltrate of macrophages and NK cells may become available to perform these roles of antigen presentation and cytotoxicity respectively. The evidence that in cases of CIU, this macrophage response may be enhanced by the presence of HPV type 16 infection is of great interest, especially because of its possible significance on the risk of progression of tumour growth (discussed in section 3.3).

Although this model is based on much assumption about the link between cell detection and immune function, it represents an important stage in providing a framework against which to compare further results. In the light of such results, the modification or even rejection of this model would amount to a significant increase in our knowledge of the local immunological interactions in the cervical epithelium and the aetiology of cervical neoplasia. In the meantime, the results described in this thesis provide a mechanism - 1 3 5 - to explain the rale of cigarette smoking in the aetiology of cervical neoplasia. In conjunction with the changes in the local immunological cell populations found to be associated with the identification of HPV type 16 DM, these results support the model for cervical carcinogenesis proposed by zur Hausen (1982,1986). This emphasises the importance of both chemical and viral components of the neoplastic process.

Against this background, it is important to recognise that the increase in the incidence of CIBf and invasive cervical cancer in younger women (described in section 1.1) has occurred contemporaneously with an increase in the numbers of women, especially teenagers, smoking cigarettes (OPCS 1984, 1985; DHSS

1988). These figures, together with the findings of a possible mechanism of action identified in this study, provide an increasingly strong argument that cervical neoplasia should be included in public health education messages as 'a smoking related cancer', and that women, especially those with cervical HPV infection, are actively encouraged to stop smoking. N -- decreased Langerhans ‘ cells in CIGARETTE SMOKERS 0 R M = diminished ability to recognise viral antigens A L >>DECREASED RESISTANCE TO VIRAL INFECTION

H -- also Langerhans' cells reduced in HPV 16 INFECTION P (also by HSV?) V = further reduced antigen presentation to T cells

I N — decreased T cell count in CIGARETTE SMOKERS F E C = reduced T lymphocyte response to HPV T I 0 N >>>PERSISTENT HPV INFECTION (especially type 16)

-- increased macrophage infiltrate in HPV 16 INFECTION C I N = enhanced tumour growth

-- decreased Langerhans* cells in CIGARETTE SMOKERS +

= diminished recognition of neoantigens H P V [-- increased NK cells ?function altered by SEMEN?]

>>>ENHANCED HPV-INDUCED CELL TRANSFORMATION to NEOPLASIA

?????PROGRESSION TO INVASIVE CANCER?????

Figure 4,1 Proposed Immunological Model for the role of Possible Cofactors in the aetiology of Cervical Neoplasia - 137-

4.3 Futurp Research.

Tlie results of this thesis are clearly incomplete and create more questions than answers. Some of these have been addressed in passing, but others merit specific discussion and suggestions of how to solve them.

1) Vhat is the relationship between the detection of immune cells in an epithelium and local immunological function ?

Until cellular antigens associated with particular cell functions are identified and suitable monoclonal antibodies raised to them, this problem will remain unsolved. Although other workers have examined Class II antigen expression as a marker of antigen presentation activity (Hughes et al 1988), the results are indeterminate due to inadequate knowledge of the significance of particular HLA antigenic expression. Further work using multiple staining techniques may clarify the cell to cell interactions (e.g. antigen presentation to T cells). In this department preliminary work combining the use of both APAAP and ABC methods to identify both class II antigens and T6CCD1) antigens on the same tissue section have begun. Furthermore, it will be necessary to investigate the relationship between changes in epithelial immune cell populations and their distribution in local lymphatics and the systemic circulation. Such work is planned using tissue collected by cone biopsy and hysterectomy, as well as post mortem specimens. The investigation of the functional role of immune cell antigens has been improved by the use of the cluster differentiation (CD) - 1 3 8 -

classification for monoclonal antibodies used in

immunocytochemistry. This system has been used throughout this

thesis in conjunction with the original names of the markers to

allow comparisons between different studies, The continuation of

this approach seems essential to the solution of this problem,

2) Are the immunological cell populations found in CIN/HPV lesions predictive of the likelihood of progression ?

The answer to this question is beset by the problem of biopsy induced change and the lack of a suitable endpoint. Because it is possible that a biopsy alters the natural history of a premalignant lesion, but is necessary for histological and immunocytochemical analysis, any project would be biased by its own sampling method.

Equally, the ethical problems of leaving women with proven CIN untreated have been recently highlighted in New Zealand (Paul 1988),

A passible method of circumventing both problems comes from the analysis of cone biopsy material. Increasingly, laser conisation is being used to treat women with all grades of CIN. In a study of laser cone specimens performed by the candidate, the detailed histological analysis of these specimens has been shown to provide a quantitative model of CIN based on the extent as well as the grade of the lesion (Jarmulowicz et al 1988). By examining the local immunological pattern throughout these specimens, any differing responses to different histological grades of disease (with different progressive potentials) may become apparent. Moreover, in those women who have undergone cervical biopsy prior to conisation, the local immunological effects of this may be further studied. - 1 3 9 -

3) Are the immunological changes associated with cigarette smoking related to the concentration of cotinine (or other cigarette smoke markers) in cervical mucus ?

This study did not directly compare cotinine concentrations in cervical mucus and the immunological changes noted in the cervical epithelium due to a failure to obtain and store enough matching samples of both cervical mucus and biopsies from the women being studied. A larger study which combines the measurement of cervical mucus concentrations of cigarette smoke constituents and compares them to the local immunological parameters should be performed. In addition to cotinine and nicotine, it may be possible to measure the local concentrations of carcinogens such as benz-a-pyrene and U- nitrosamines. Another method of examining local carcinogenesis may be to examine DNA-adducts (Perera et al 1987). In addition to further investigating the effect of tobacco constituents on

Langerhans' cells, it may be that a comparison of the cervical changes in smokers and tobacco chewers with CIU/HPV might shed light on the identity of the constituent or metabolites responsible for local immune changes in the cervix.

4) Are the immunological changes associated with cigarette smoking confined to the cervix ?

Study of the immune cell populations, especially Langerhans' cells, in skin biopsies with relation to smoking histories should now be performed. In the gynaecological context, vulval biopsies would be interesting to study in view of the association of this site with

HPV infection and vulval intraepithelial neoplasia. Whether or not -140- smoking contributes to the reduced systemic immunological competence detectable in women with multifocal genital tract intraepithelial neoplasia (Carson et al 1986) should be examined. Moreover, the epidemiological association of cigarette smoking with penile cancer

(Hellberg et al 1987) suggests that males should also be studied in this way; unfortunately, the patient compliance for the acquisition of penile biopsies from normal men is likely to be poor.

5) What are the effects on local immunologically active cell populations in HPV infection with viral types other than 16, and are these changes localised or generalised ?

Without the facility to culture HPV, it is necessary to base evidence of infection on cytology, histology and DMA detection.

Unless there is a major development in either virology or molecular biology, it will not be feasible to screen for every KPV type in each sample. Furthermore, although the use of filter techniques from exfoliated cervical cells is applicable in screening studies, it will be necessary to perform biopsies to localise HPV subtypes within the epithelium and examine the local immune response to particular viral types, both alone and in combination with other infectious agents. The potential interaction between cervical infections merits further examination, both at the immunological and cell transformation levels. It is possible that the study of these interactions in the cervical epithelium of immunocompromised women with HIV infection may shed light on this problem. -141-

4-,4. Summary nf Claims far Original Contributions to Knowledge.

1. Langerhans' cells in the cervical epithelium are diminished in

women who are current cigarette smokers. This effect is present in

both normal epithelium and CIN/HPV lesions.

2. The detection of HPV type 16 DNA from the cervix is significantly associated with a reduction in Langerhans1 cells in cervical epithelium with histological changes of HPV infection or CIN/HPV lesions.

3. Current cigarette smoking was found to be associated with a significant decrease in the total (UCHT1 positive) T lymphocyte count per unit area in cervical epithelium showing changes of HPV infection.

4. In lesions showing histological changes of CIN/HPV, the presence of HPV type 16 DNA was found to be significantly associated with an increase in both the total and activated macrophage counts.

5. A significant Natural Killer cell infiltrate was found to occur only in cervical epithelium showing histological changes of CIN with

HPV infection. An NK cell infiltrate in uncomplicated HPV infection was rarely seen.

6. The necessity to screen women undergoing colposcopy for the presence of sexually acquired infections was further demonstrated by -142- this study and the changes in local immunologically active cells which were observed in association with HSV and chlamydial infection means that such screening is mandatory in all future studies of the local immunopathology of CIK/HPV.

7. Uo significant association was found between the distribution of immunologically active cells in the cervical epithelium and the use of hormonal contraception or history of recent semen exposure,

8. The measurement of cotinine in cervical mucus was found to be associated with the severity of the histological grade of CIK/HPV detected in the group of women found to have HPV type 16 infection of their cervix.

9. The hypothesis that cigarette smoking may affect the concentration of local cervical epithelial immune cells has been demonstrated in normal epithelium and CIK/HPV lesions. A separate effect has been observed on Langerhans* cells in HPV and CIK/HPV lesions in epithelium from which HPV type 16 DKA was detectable. The possible synergistic immunosuppressive actions of these possible cofactors provides supportive evidence for the interaction of viral and chemical agents in the aetiology of cervical neoplasia. A P P E N D I X I

STANDARD HISTORY SHEET FOR CO-FACTORS STUDY

Date Study Number □ □ □ □ □ □ □ □

P lace of b ir th :______

Date of b ir t h : ______

Single / Married / Co-habiting / Separated / Divorced / Widowed

Number of times married: □ □

Race: White / Black / Caribbean or Central American / Oriental/ Middle Eastern / Indian or Pakistani / Other S p ecify______

SEXUAL HISTORY Age at f i r s t in terco u rse: n n T otal number of sexu al p artn ers: □ □

Current partner? YES / NO

What is his occupation?

How many different sexual contacts have you had in the last three months? □ □

Any history of a sexually transmitted disease in the past? YES / NO S p ecif y______

Have you ever had an abortion? YES / NO

When did you last have sexual intercourse?

9.2.87 -144-

CONTRACEPTIVE HISTORY

Have you ever used combined OCP? YES / NO

I f YES: Name of P i l l Age s ta r te d Age stopped

Have you ever used POP or injectable progestogen? YES / NO

I f YES: Name Age s ta r te d Age stopped

Have you ever used IUCD? YES / NO

I f YES: Age in s e r te d ______Age removed______

Have you ever used barrier contraception? YES / NO

Condoms / Cap / Diaphragm / Sponge / Other S p e c ify______

I f YES: Type Age s ta r te d Age stopped

Major type of contraceptive in past 6 months?

Degree of s. exposure: 0 1 2 3 SMOKING HISTORY i^~

Have you ever regularly smoked cigarettes, cigars or a pip^? (Life long non-smoker - 0; current smoker - 1; ex-smoker - 2)

SMOKERS AND EX-SMOKERS ONLY

How old were you when you started to smoke regularly? □ □

Ex-smokers, how old were you when you stopped smoking?

How much do / did you normally smoke (when you last gaveup)? □ □ n

Commercial cigarettes, number per day:

Brand name:______

Hand-rolled cigarettes, number per day: □ □

Small c ig a r s , number per day: □ □

Large c ig a r s , number per week: □ □

Pipe tobacco, oz. per week:

MENSTRUAL HISTORY

LMP______

Menarche: □ □

Cycle length: □ □

Abnormal bleeding: IMB / PCB

Pregnancy details: APPENDIX II

STANDARD EXAMINATION SHEET - 146-

PAT I BIT IUMBER: Date;

VULVA: IORMAL/HICROVARTS/ACBTOVHITB/PUFCTATIOH/MDSAIC/PIGMEHTATIOH

VAGIFA: FORMAL/MICRO VABTS/ ACETOVHITB/ PUFCTATI OB/ KDSAIC/ PIGMEFT AT I OF

CRRVTX: COLPOSCOPY: SAT ISF ACTORY / UFS AT ISF ACTORY

FORMAL/ABFORMAL

MARGII: IIDISTIHCT/REGULAR STRAIGHT/ROLLED PBBLIIG

COLOUR: SIOV VHITB/GREY VHITB/DULL GRBY WHITE

ABFORMAL VBSSBLS: ABSEFT/ UIIFORM FIIB/PUFCTATIOI/MOSAIC

IODIFE TBST: FORMAL/PARTIAL UPTAKE/RBVERSE PUHCTATIOI/MOSAIC

IMPRESS I OF: I 0RMAL/WI/CII1/CII2 /CII3 /MICR0 IIV/IFVASIVB

MAFAGBMEFT/STUDY CODE: 147-

APPENDIX TT T Database used in analysis of Sections 3. 1-3.4

A : Summary of Possible Factors and Cofactors recorded with Langerhans' cell counts used in analysis of Section 3.1.

B : Summary of Possible Factors and Cofactors used in analysis of sections 3.2-3,4.

Guide to Database;

Column(s) Variable recorded

A Patient Code Number

B 0=non-smoker/l=smoker/2=ex-smoker

C number of cigarettes smoked per day

D number of cigarettes smoked per day by partner

E Contraception; 0=none/l=sheath/2=cap/3=0C pill 4=Progesterone only pill/5=Coil/6=Sterilised 7=Withdrawal/8=Depot progestagen/9=sponge

F Age at time of study in years

G/H Current genital infectionActinomyces/06=Gonorrhoea/07=HSV/08=Streptococci/09=Other

I Test for HPV 16 DNA result; l=positive/2=negative

.J Biopsy Histology result; 0=normal/1=HPV infection/2=CIN 1 3=CIN 2/4=CIN 3/5=Microinvasive disease

K S100 stained section; area measured in mm2

L Number of S100 stained cells counted in area K

M T6 (CD1) stained section; area measured in mm2

N Number of T6 (CD1) stained cells counted in area M SECTION A Co 1 umn : A B c D E FG H I JK L M N 1 00 413 05 263 001 1 20 27 00 1 4 05 -148- 002 1 05 3 23 00 00 2 0 322 30 310 50 003 0 3 30 00 00 2 4 277 13 280 20 004 1 10 3 21 02 00 2 3 364 24 315 24 005 1 20 5 30 02 00 1 4 209 01 230 03 006 2 20 20 5 43 00 00 2 2 638 10 339 07 007 2 05 3 26 00 00 1 1 369 34 319 58 009 0 2 22 00 00 2 2 199 05 190 15 010 2 10 20 5 32 01 00 2 2 177 14 160 26 012 1 35 3 23 00 00 1 1 3 3 ' 02 334 15 013 1 20 3 25 00 00 2 0 519 05 556 06 014 2 02 3 26 01 00 1 4 167 02 278 01 015 1 05 3 44 00 00 1 3 234 01 349 01 016 1 15 3 23 03 00 2 1 299 11 368 52 017 1 20 6 40 02 00 1 4 265 17 259 23 018 1 10 1 26 09 00 2 1 651 14 523 14 019 0 0 29 00 00 2 0 512 28 487 60 020 1 20 5 28 03 00 2 0 397 30 348 55 021 1 05 3 24 00 00 2 0 430 31 373 81 022 1 10 3 22 06 00 1 1 371 22 405 50 023 1 20 1 25 00 00 1 4 527 03 416 08 024 1 15 5 26 00 00 2 3 297 02 205 03 026 1 02 0 41 00 00 1 4 214 04 254 12 027 1 20 3 29 00 00 1 1 3 98 05 200 09 028 0 5 30 08 00 2 0 258 43 245 48 029 1 06 3 52 00 00 1 1 316 40 387 41 035 0 1 42 00 00 2 0 395 59 280 80 036 0 8 21 00 00 2 0 307 25 194 33 037 0 2 35 00 00 2 0 116 10 273 33 038 0 0 49 02 00 2 0 320 32 197 32 039 1 15 3 28 02 00 2 2 505 07 274 09 040 0 10 5 27 00 00 1 0 277 13 224 26 043 0 1 41 02 00 1 2 770 15 427 19 044 1 20 0 37 01 00 2 2 268 03 246 05 046 1 20 1 36 00 00 2 4 350 03 348 13 048 1 15 3 22 02 00 1 2 210 07 298 17 049 2 13 3 22 01 00 1 1 324 15 303 11 050 1 40 5 41 00 00 2 2 315 00 223 00 051 1 13 1 18 04 00 2 1 393 00 288 01 052 1 25 1 22 00 00 1 4 225 00 255 04 053 0 3 22 00 00 1 3 689 53 329 28 055 2 10 10 2 22 00 00 2 0 301 40 415 95 056 1 10 5 29 01 04 1 0 482 04 284 10 057 1 10 3 24 02 00 2 3 448 08 440 10 058 0 2 28 01 00 1 4 417 05 348 04 059 2 15 05 5 27 02 00 1 1 263 25 265 30 060 2 10 6 33 00 00 1 1 239 04 510 12 062 1 20 9 26 01 00 2 1 255 57 218 70 066 1 15 0 28 00 00 2 4 345 00 485 04 067 2 10 2 28 00 00 1 1 368 30 364 73 069 0 20 1 31 00 00 1 1 400 00 493 23 078 1 10 3 31 00 00 1 1 133 08 160 10 080 0 3 22 00 00 2 0 252 16 185 28 082 1 10 2 25 00 00 2 2 320 07 212 06 083 1 40 3 30 01 00 1 1 254 02 295 27 087 1 23 0 46 00 00 2 1 391 20 314 44 088 1 15 6 38 00 00 1 1 249 13 251 22 092 1 15 5 22 02 08 2 2 391 10 391 55 A B c D B F G H I J K L M N 093 0 3 27 02 00 2 0 083 13 081 21 094 1 05 3 31 00 00 1 w'T 173 01 191 06 095 2 10 10 3 23 00 00 2 2 470 42 366 60 097 2 20 20 0 32 00 00 2 3 436 00 508 07 100 1 05 1 20 02 00 2 0 528 26 380 48 101 1 01 3 23 00 00 2 4 530 25 408 39 103 0 0 26 00 00 2 1 311 21 503 47 106 1 15 3 24 00 00 1 4 348 04 383 01 113 1 15 3 21 00 00 1 1 458 12 487 22 115 1 20 0 43 07 00 1 3 318 00 318 00 116 1 30 3 31 00 00 1 2 447 00 260 10 117 0 3 19 00 00 1 1 571 40 290 30 118 1 20 5 21 00 00 1 4 265 00 265 12 119 1 20 0 26 00 00 2 0 805 00 806 12 120 2 05 10 3 24 00 00 2 1 521 03 326 05 122 2 20 1 28 00 00 1 2 205 22 347 04 124 2 20 3 24 02 oi 1 3 385 35 254 20 128 0 3 27 00 00 1 4 213 11 433 28 129 0 10 6 34 00 00 1 1 136 15 127 20 130 1 03 3 21 00 00 1 3 207 30 211 31 131 0 20 5 26 00 00 2 3 254 31 185 33 132 1 20 3 30 00 00 2 4 250 01 575 12 133 2 30 4 44 00 00 1 2 319 17 271 14 140 2 10 1 20 02 00 1 3 556 20 336 17 144 2 02 2 23 01 00 2 1 300 22 302 28 151 1 15 2 24 00 00 2 0 987 00 987 22 153 1 20 5 32 02 00 2 2 664 05 350 06 162 0 10 6 50 01 03 1 2 431 10 364 09 169 0 3 27 01 00 1 1 546 06 437 15 171 1 20 3 33 00 00 1 2 591 00 591 00 176 0 20 0 29 00 00 1 1 958 00 958 17 178 0 5 22 00 00 1 2 740 00 744 26 180 0 5 44 00 00 2 4 209 17 556 20 185 2 25 25 5 22 01 00 1 2 685 00 685 16 188 1 25 2 21 00 00 1 1 369 00 369 05 190 0 1 27 00 00 2 0 265 13 270 35 191 1 10 4 25 00 00 2 2 364 00 364 10 197 0 3 24 00 00 2 1 233 17 220 28 202 0 2 24 00 00 1 1 346 22 500 29 250 0 3 18 07 00 1 0 414 03 538 05 257 1 20 0 30 00 00 1 2 460 00 460 05 264 0 3 28 00 00 1 1 401 42 421 50 266 1 20 5 44 01 00 1 5 350 01 350 02 267 1 20 3 22 00 00 2 2 475 05 404 09 273 0 20 2 35 00 00 1 1 390 16 394 39 279 0 0 39 00 00 2 1 375 51 493 70 282 1 05 0 26 00 00 2 1 564 08 366 40 285 1 05 2 22 00 00 2 0 927 03 926 42 286 1 10 3 19 00 00 1 1 716 04 806 17 289 1 20 3 21 02 00 1 2 484 18 500 55 292 1 25 0 55 00 00 2 0 937 00 937 01 293 0 2 29 00 00 1 2 621 11 244 09 297 1 01 0 27 02 00 1 1 300 00 308 42 308 1 20 3 21 00 00 1 2 573 00 573 00 327 0 1 18 00 00 2 0 850 17 505 40 334 0 04 2 29 00 00 1 1 501 03 609 16 335 0 1 28 00 00 1 1 278 15 346 23 336 1 15 3 19 00 00 1 1 315 00 315 08 I 0 LO c~H1

in CO rH co CO o cn C" CO COo in p* z COH CO CO CO COCO CO CO H COCO CO CO V}* CO CO CO in CO o in COin o COCOCD ^ IO CO o o o H cn CO rH o cr\ COrH o CO COCO CO COCO CO H in in CO CO CO CO CO o COCOH CO CO CO 1—1COCOo CO CO _i o o rH CO 1—1 o CO CO 1—11—11—1 COCO CO rH COin CO in in CO o o in in H* o CPi cn CO cr\ CO CO r** CO r- H' CO CO CO COCOCOCO COco CO in CO COCO CO rH o o o o o o o o o o o o o •—Hi—1 COCO CO COCO CO COCO rH CO CO 1—1 CO CO rH CO COCOCO CO CO1—1 1—1 CO1—1 1—1 H rH CO rH rH CO rH CO rH 1—1 CO CO CO CO CO CO CO CO CO CO CO rH CO rH rH CO CO CO o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o X o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o (n o o o CO o o o o o o o rH o o o o o o o o o o o o o CO i—! CO o o o o o o o o o o o o o o o o o o o co o o o o o o V_J o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o 11Ll_ CO co CO H' CO CO CO in o LO in COCOCO r- rH H' o rH CO rH CO CO CO rH CO rH o rH 1—1 1—1 CO o t" CO VO cn o in VO CO in r"- CO cn CO COCOCO CO CO in COCO COCOCO CO CO CO COCOO’ H' COCO CO CO CO COCO CO CO CO CO CO CO CO CO CO CO CO CO CO H 1 CO CO CO H* CO CO CO CO CO CO CO CO CO

CO 1 1 rH CO VO rH rH LU COo o COCO o o CO COCO CO I—iCO o CDH' in o COo CO.CO CO in H CO in in o in CO CO o — CO o CO o o o o CO o CO in CO CO in CO a in o in CO o in o o o rH CO rH CO o CO CO u in o o o in o o o o o o in in o CD o o O' o o o CO in o o in in o o in in o o COCO CO o 1—1 1—11—1 COCO COrH 1—1 CO o CO rH o CO CO CO o H rH CO o r—1 rH rH CO rH CO 1—1 CD CO rH 1—1o o i—! o o iH 1—1 o CO CO o o o o o rH 1—1 H 1—lo 1—1 1—1o 1—l o 03 1—1 rH CO 1—1 o o CO o 1—1 1—1 H o o CO o o o o 1—1 rH rH 1—1 1—1 H H 1—1

r- COCOrH CTi o in CO CO COCO 1—I in H' o in o rH 0s!co o COin f" c o CO in CO CO CO in CTl CO CO in CO f" o cn rH VO H CO o H CO CO in VO CO < CO t> CO CO co p- C" r- CO in co r-' CO o CO t> COCOcn rH COCO in in CO CD CO CO CO cn cr>cn o 1—1 rH VO CO CPk cn o o rH rH 1—1 H* COCOCO o COCOCO CO CO 1—1 rH CO o o o o o o o o rH 1—1rH 1—1 rH 1—1 1—1 1—1 1—1 1—1 rH H 1—1 CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO CO

m z o

u LU co -151-

E E JE E m C E S;

Abo T & Balch CM 1981. A differentiation antigen of human NK and K cells identified by a monoclonal antibody (HNK-1). J Immunol 127:1024-1029.

Adams DO & Snyderman R 1979. Do macrophages destroy nascent tumour? J Natl Cancer Inst 62(6):1341-1345.

Akbar AN Terry L Timms A Beverley PCL Janossy G 1988. Loss of CD45R and gain of UCHL1 reactivity is a feature of primed T cells. J Immunol 140:2171-2178.

Alsabti EA 1980. The immunostatus of untreated cervical carcinoma. Gynecol Oncol 9:6-11.

Ames BN McCann J Yamasaki E 1975. Methods for detecting carcinogens and mutagens with the Salmonella mammalian microsome mutagenicity test. Mutat Res 31:347-363.

Anderson MC 1987. Premalignant and malignant disease of the cervix. In: Haines and Taylors' Obstetrical and Gynaecological Pathology. Ed. Fox H. Churchill Livingstone. Edinburgh.

Ashton H 8s Stepney R 1982. Smoking: Psychology and Pharmacology. Tavistock. London.

Baker GHB Irani MS Byrom NA Nagvekar NM Wood RJ Hobbs JR Brewerton DA 1985. Stress, cortisol concentrations, and lymphocyte subpopulations. Br Med J 290:1393.

Baker RJ 8s Nelder JA 1985. The GLIM system; release 3,77. NAG. Oxford. -152-

Barton SE Thomas BJ Taylor-Robinson D Goldmeier D 1985. Detection of Chlamydia trachomatis in the vaginal vault of women who have had hysterectomies. Br Med J 291:250.

Barton SE Link CM Wright L Munday PE 1986, Screening to detect asymptomatic shedding of HSV in women with recurrent genital HSV infection. Genitourin Med 62:181-185.

Barton SE Davis JM Moss VW Tyms AS Kunday PE 1987. Asymptomatic shedding and subsequent transmission of genital herpes simplex virus. Genitourin Med 63:102-105.

Barton SE Taylor-Robinson D Harris JRW 1987. Female prostitutes and sexually transmitted diseases. Br J Hosp Med 38(1):34-45.

Beck JS & Anderson JM 1987. Quantitative methods as an aid to diagnosis in histopatholagy. In: Recent advances in Histopathology 14. Eds. Anthony PP Sr MacSween MM. Churchill Livingstone. Edinburgh.

Belsito DV Flotte TJ Lim HV Baer RL Thorbecke JG Gigli I 1982. Effect of glucocorticoids on epidermal Langerhans cells. J Exp Med 155:291-302.

Beral V 1974. Cancer of the cervix: a sexually transmitted infection? Lancet i:1037-1040.

Beral V & Booth M 1986. Predictions of cervical cancer incidence and mortality in England and Vales (letter). Lancet i:495.

Berger LR 1988. Cigarette smoking and the acquired immunodeficiency syndrome (letter). JAMA 408:638.

Berggren G & Sjostedt S 1983. Preinvasive carcinoma of the cervix uteri and smoking. Acta Obstet Gynecol Scand 62:593-598. -153-

Berman A & Vinkelman RK 1980. Involuting common warts. Clinical and histopathologic findings. J Am Acad Dermatol 3:356-362.

Berman B Duncan MR Smith B Ziboh VA Palladino M 1985. Interferon enhancement of HLA-DR antigen expression on epidermal Langerhans cells. J Invest Dermatol 84:54-58.

Birbeck MS Breathnach A Everall JD 1961. An electron microscope study of basal melanocytes and high level clear cells (Langerhans cells) in vitiligo. J Invest Dermatol 37:51-64.

Bistoletti P & Lidbrink P 1988. Sexually transmitted diseases including genital papillomavirus infection in male sexual partners of women treated for cervical intraepithelial neoplasia III by conization. Br J Obstet Gynaecol 95:611-613.

Blann AD 1987. T lymphocyte surface molecules: structure and function. Med Lab Sci 44:220-236.

Bonham DG Green GH Liggins GC 1987. Cervical Human Papilloma Virus Infection and Colposcopy. Aust NZ J Obstet Gynaecol 27:131.

Bonneterre J Santaro F Vattre P Beuscart R Adenis L Capron A 1981. Etude de la competance immunitaire chez les malades presentant un cancer du col uterin. Gynecologie 32:257-266.

Boog CJP Kast VM Timmers HTM Boes J de Vaal LP Melief CJM 1985. Abolition of specific immune response defect by immunization with dendritic cells. Mature 318:59-62.

Boyce JG Lu T Nelson JH Fruchter RG 1977. Oral contraceptives and cervical carcinoma. Am J Obstet Gynecol 128:761-764.

Boyd JT & Doll R 1964. A Study of the aetiology of carcinoma of the cervix uteri. Br J Cancer 18:419-434. -154-

Bradbeer CS 1987. Is infection with. HIV a risk factor for cervical intraepithelial neoplasia? Lancet ii:1277-1288.

Buckley JD Harris RVC Doll R Vessey M William PH 1981. Case control study of husbands of women with dysplasia or carcinoma of the cervix. Lancet ii:1010-1015,

Buckley CH Butler EB Fox H 1982. Cervical intraepithelial neoplasia. J Clin Path 35:1-13.

Bunney MH 1982. Viral warts: their biology and treatment. OUP. Oxford.

Burnet FM 1970. The concept of Immunological Surveillance. Prog Exp Turn Res 13:1-5.

Campion MJ Singer A Clarkson PK McCance DJ 1985. Increased risk of cervical neoplasia in consorts of men with penile condylomata. Lancet i:943-946.

Campion MJ McCance DJ Cuzick J Singer A 1986. Progressive potential of mild cervical atypia: Prospective cytological, colposcopic, and virological study. Lancet ii:237-240.

Caorsi I & Figueroa CD 1984. Langerhans' cells in squamous exocervical carcinoma: A quantitative and ultrastructural study. Ultrastruct Pathol 7:25-40.

Caorsi I & Figueroa CD 1986. Langerhans' cell density in the normal exocervical epithelium and in the cervical intraepithelial neoplasia. Br J Obstet Gynaecol 93:993-998. -155-

Carson LF Twiggs LB Fukushima M Ostrow RS Faras AJ Qkagaki T 1986, Human genital papilloma infections: an evaluation of immunologic competence in the genital neoplasia-papilloma syndrome. Am J Obstet Gynecol 155:784-789.

Chakravarti RN Bhattacharya D Ghosh K Sidkar S 1987. Tobacco leaf consumption and the incidence of cervical dysplasia. In: Proceedings of the International Working Party Meeting in India. Eds. Mould RF & Bose A. Adam Hilger. Bristol.

Chardonnet Y Beauve P Viac J Schmitt D 1983. T-cell subsets and Langerhans cells in wart lesions. Immunol Letters 6:191-196.

Chardonnet Y Viac J Staquet MJ 1985. Cell mediated immunity to human papillomavirus. Clin Dermatol 3:156-159.

Chardonnet Y Viac J Thivolet J 1986. Langerhans Cells in human Warts. Br J Dermatol 115:669-675.

Choi KL & Sauder DM 1986. The role of Langerhans cells and keratinocytes in epidermal immunity. J Leukocyte Biology 39:343-358.

Christopherson WM Parker JE 1965. Relation of cervical cancer to early marriage and childbearing. N Engl J Med 273:235-239.

Clarke EA Morgan RW Newman AM 1982. Smoking as a risk factor in cancer of the cervix: Additional evidence from a case control study. Am J Epid 115:59-63.

Clarke EA Hatcher J McKeown-Eyssen GE Lickrish GM 1985. Cervical dysplasia association with sexual behaviour, smoking, and oral contraceptive use. Am J Obstet Gynecol 151:612-616.

Clemens LE_ Sitteri PK Stiles DP 1979. Mechanism of immunosuppression of progesterone on maternal lymphocyte activation during pregnancy. J Immunol 122:1978-1985. - 156-

Cocchia D Michetti F Donato R 1981. Immunochemical and immunocytochemical localization of S100 antigen in normal skin. Mature 294:85-87.

Col eman DV Morse AR Beckwith P Anderson MC Gardner SD Knowles WA Skinner GRB 1983. Prognostic significance of herpes simplex virus antibody status in women with cervical intraepithelial neoplasia (CIM). Br J Obstet Gynaecol 90:421-427.

Cook GA & Draper GJ 1984. Trends in cervical cancer and carcinoma in situ in Great Britain. Br J Cancer 40:367-375.

Coppleson M & Reid BL 1967. Preclinical carcinoma of the cervix uteri, Pergamon Press. Oxford.

Coppleson M 1969. Carcinoma of the cervix: epidemiology and aetiology. Br J Hosp Med 2:348-361.

Cunningham-Rundles S 1985. Approaches to the study of human immunoregulation. Immunology Today 6:251-269.

Cuzick J 1985. A Wilcoxon type test for trend. Statistics in Medicine 4:87-90.

Cuzick J 1988. Cervical Screening. Br J Hosp Med 39(4):265.

Dalgleish AG 1986. The T4 molecule: function and structure. Immunol Today 7(5):142-144.

DHSS 1988. Fourth Report of the Independent Scientific Committee on Smoking & Health. HMSO, London. -157-

Djeu JY Stocks N Zoon K Stanton. GJ Timonen T Herberman RB 1982. Positive self regulation of cytotoxicity in human natural killer cells by production of interferon upon exposure to influenza and herpes viruses. J Exp Med 156:1222-1234.

Boll R 1986. Implications of Epidemiological Evidence for Future Progress. In: Viral Aetiology of Cervical Cancer: Banbury Report 21. Cold Springs Harbor Laboratory.

Dougherty GJ & McBride WH 1984. Macrophage heterogeneity. J Clin Lab Immunol 1:1-11.

Dreno B Milpied B Bignon JD Mousset S Litoux 1988. Prognostic value of Langerhans cells in epidermis of HIV patients. Paper presented to Fourth World AIDS Congress, Stockholm, Sweden.

Dubertret L Picard 0 Bagot M Tulliez M Fosse M Aubert C Touraine R 1982. Specificity of monoclonal antibody anti-T6 for Langerhans' cells in normal human skin. Br J Dermatol 106:287-289,

Duenas A Adam E Melnick JL Rawls WE 1972. Herpes type 2 in a prostitute population. Am J Epid 95:483-489.

Duguid HL Duncan ID Currie J 1985. Screening for Cervical intraepithelial neoplasia in Dundee and Angus 1962-81, and its relation with invasive cervical cancer. Lancet ii: 1053-1056.

Editorial 1977. Cervical Neoplasia and the Pill. Lancet ii:644.

Editorial 1987. Human papillomaviruses and cervical cancer; a fresh look at the evidence. Lancet i:725-726.

Edwards JNT & Morris HB 1985. Langerhans' cells and lymphocyte subsets in the female genital tract. Br J Obstet Gynaecol 92:974- 982. -158-

Emerson EE 1978. Migratory behaviour of lymphocytes with specific reactivity to alloantigens. J Exp Med 147:13-24.

Erber WN Pinching AJ Mason DY 1984. Immunocytochemical detection of T and B cell populations in routine blood smears. Lancet i:1042- 1046.

Evans RL Lazarus H Penta AC Schlossman SF 1978. Two functionally distinct subpopulations of human T cells that collaborate in the generation of cytotoxic cells responsible for cell mediated lympholysis. J Immunol 120:1423-1428.

Evans DM Hudson EA Brown CL Boddington MM Hughes HE MacKenzie EF Marshall T 1986. Terminology in gynaecological cytapathology: report of the working party of the British Society for Clinical Cytology. J Clin Path 39:933-944.

Everson RB Sandler DP Wilcox AJ Schreinemachers D Shore DL Weinberg C 1986. Effect of Passive exposure to smoking on age at natural menopause. Br Med J 293:792.

Fearson M Edwards A Lind A Milton GW Hersey P 1979. Low Natural Killer cell activities and Immunoglobulin levels associated with smoking in human subjects. Int J Cancer 23:603-609.

Fenoglio CM & Lefkowitch J 1985. HSV II and neoplasia of the lower female genital tract. In; Cancer Campaign Vol.8 Cancer of the Uterine cervix. Ed. Grundsmann E. Gustav Fischer Verlag. Stuttgart.

Fernandez-Cruz E Voda BA Feldmann JD 1980. Elimination of syngeneic sarcomas in rats by a subset of T-lymphocytes. J Exp Med 152:823- 841.

Fernandez-Cruz E Gilman SC Feldmann JD 1982. Immunotherapy of a chemically induced sarcoma in rats: characterisation of effector T- cell subset and nature of supppression. J Immunol 128:1112-1117. -159-

Figueroa CD & Caorsi I 1980. Ultrastructural and morphometric study of the Langerhans' cell in the normal human exocervix. J Anat 131:669-682.

Fish EM Tobin SM Cooter NBE Papsin FR 1982. Update on the relation of Herpesvirus Hominis Type II to carcinoma of the Cervix. Obstet Gynecol 59 (2):220-224.

Forbes JT Greco FA Oldham PK 1980. Natural killer cell mediated cytotoxicity in human tumour cell patients. In: Natural cell mediated immunity against tumours. Ed: Herberman RB. Academic Press NY. ppl031-1046.

Fox CH 1967. Biological Behaviour of Cervical Dysplasia and Carcinoma in situ. Am J Obstet Gynecol 99:609-614.

Fray RE Hussain OAN To ACV Vatts KC Lader S Rogers GT Taylor- Papadimitriou J Morris NF 1984. The value of immunohistochemical markers in the diagnosis of cervical neoplasia. Br J Obstet Gynaecol 91:1037-1041.

French PV Coppleson M Reid BL 1987. Effects of sperm protamine on human cervical epithelial cells and BHK 21 cells in vitro. J Roy Soc Med 80:434-437.

Friedell G Hertig A Younge P 1960. Carcinoma in situ of the uterine cervix. Charles C. Thomas. Springfield.

Friedman P Ford G Ross J Diffey B 1981. Reappearance of epidermal Langerhans cells after PUVA therapy. Br J Dermatol 109:30-37.

Frost JK 1961. Cytology of Benign conditions. Clin Obstet Gynaecol 4:1075-1096.

Fry HJB 1929. Syphilis and malignant disease. A serological study. Journal of Hygiene 29:313-322. -160-

Furgyk S & Astedt B 1980. Gonorrhoeal infection followed by an increased frequency of cervical carcinoma. Acta Obstet Gynecol Scand 59: 522'-524.

Gagnon F 1950. Contribution to the study of etiolgy and prevention of cancer of the cervix of the uterus. Am J Obstet Gynecol 60:516- 522.

Garcia-Tamayo J Castillo G Martinez AJ 1982. Human genital candidiasis: histochemistry, scanning and transmission electron microscopy. Acta Cytol (Baltimore) 26:7-14.

Gatti RA & Good RA 1971. Occurrence of malignancy in immunodeficiency diseases. Cancer 28(1):89-98.

Gissman L Wolnik L Ikenberg H Koldovsky U Schnurch HG zur Hausen H 1983. Human papillomavirus types 6 and 11 DNA sequences in genital and laryngeal papillomas and in some cervical cancers. Proc Natl Acad Sci USA 80:560-563.

Glucksmann A & Cherry C 1964. Microinvasive carcinoma of the cervix: histopathological aspect. In: Dysplasia, In situ carcinoma and microinvasive carcinoma of the cervix uteri. Ed. Gray L. Charles C. Thomas. Springfield.

Greenberg ER Vessey MP McPherson K Yeates D 1985. Cigarette smoking and cancer of the uterine cervix. Br J Cancer 51:139-143.

Hackerman M Grubb C Hill KR 1968. The ultrastructure of normal squamous epithelium of the human cervix uteri. J Ultrastructure Res 22:443-457.

Hakama M 1984. Trends in incidence of cervical cancer in the Nordic Countries..In: Trends in Cancer Incidence. Ed. Magnus K. Hemisphere, New York. -161-

Haley Nj Axelrad CM Tilton KA 1983. Validation of self-reported smoking behaviour: biochemical analyses of cotinine and thiocyanate. Am J Public Health 73:1204-1207.

Hall PA Ardenne AJD Butler MG Habeshaw JR Stansfeld AG 1987. New marker of B lymphocytes, MB2: comparison with other lymphocyte subset markers active in conventionally processed tissue sections. J Clin Path 40:151-156.

Hamaoka T & Fujiwara H 1987. Phenotypically and functionally distinct T-cell subsets in anti-tumor responses. Immunology Today 8:267-269.

Hancock BV Bauer L Heath J Sugden P Ward AM 1979. The effects of radiotherapy on immunity in patients with localised carcinoma of the cervix uteri. Cancer 43:728-736.

Hanna N & Fidler I 1983. Relationship between metastatic potential and resistance to natural killer cell mediated cytotoxicity in the murine tumour system. J Natl Cancer Inst 66:1183-1190.

Harding WG 1942. The influence of syphilis in cancer of the cervix uteri. Cancer Res 2:49-61.

Hare MJ Toone E Taylor-Robinson D Evans RT Furr PM Cooper P Oates JK 1981. Follicular Cervicitis - Colposcopic appearances and association with Chlamydia trachomatis. Br J Obstet Gynaecol 88:174- 180.

Hare MJ Taylor-Robinson D Cooper P 1982. Evidence for an association between Chlamydia trachomatis and Cervical Intraepithelial neoplasia. Br J Obstet Gynaecol 89:489-492

Harkness AH 1948. The pathology of gonorrhoea. Br J Vener Dis 24:137-147. -162-

Harris RVC Brinton LA Cowdell RH Skegg DC Smith PG Vessey MP Doll R. 1980. Characteristics of women with dysplasia or carcinoma in situ of the cervix uteri. Br J Cancer 42:359-369.

Hellberg D Valentin J Nilsson S 1983. Smoking as a risk factor in cervical neoplasia (letter). Lancet ii:1497.

Hellberg D Valentin J Nilsson S 1986. Smoking and Cervical Intraepithelial Neoplasia. Acta Obstet Gynecol Scand 65:625-631.

Hellberg D Valentin J Eklund T Nilsson S 1987. Penile cancer: Is there an epidemiological role for smoking and sexual behaviour? Br Med J 295:1306-1308.

Hellberg D Nilsson S Haley NJ Hoffmann D Vynder E 1988. Smoking and cervical intraepithelial neoplasia: Nicotine and cotinine in serum and cervical mucus in smokers and nonsmokers. Am J Obstet Gynecol 158:910-913.

Herberman RB & Ortaldo JR 1981. Natural Killer cells : their role in defence against disease. Science 214:24-30.

Herberman RB & Oldham PK 1983. Cell mediated cytotoxicity against human tumours: lessons learned and future prospects. J Biol Resp Modif 2:111-120.

Hersey P Edwards A McCarthy WH 1980. Tumour related changes in natural killer cell activity in melanoma patients . Influence of stage of disease, tumour thickness and age of patient. Int J Cancer 25:187-192.

Hersey P Edwards A Lewis R Kemp A Mclnnes J 1982. Deficient natural killer cell activity in a patient with Fanconi's anaemia and squamous cell carcinoma, association with defect in interferon release. Clin Exp Immunol 48:205-212. Hirota T & Shimosato Y 1982. An immunoperoxidase study of S100 protein distribution in normal and neoplastic tissues. Am J Surg Pathol 6:715-727.

Hoffmann D Hecht SS Melikian AA Haley NJ Brunnemann KD Adams JD Vynder EL 1986. Tumorigenic agents in Tobacco products and their uptake by chewers, smokers and non-smokers. Biochemical and Molecular Epidemiology of Cancer.

Hoffmann D Vynder EL Hecht SS Brunnemann KD LaVoie EJ Haley NJ 1987. Chemical carcinogens in Tobacco. In: Cancer Risks, Ed. P.Bannasch. Springer Verlag, Berlin.

Hogg N Ross GD Jones DB Slusarenko M Valport MJ Lachmann PJ 1984. Identification of an anti-monocyte monoclonal antibody that is specific for membrane-complement receptor type one

Hogg N Takacs L Palmer DG Selvendran Y Allen C 1986. The pl50,95 molecule is a marker of human mononuclear phagocytes: comparison with expression of class II molecules. Eur J Immunol 16:240-248.

Hogg N & Horton MA 1987. Myeloid antigens: new and previously defined clusters. In; Leucocyte Typing III. Ed. McMichael AJ. OUP. Oxford.

Hoidal JR Fox RB Le Marbe PA Perri R Repine JE 1981. Altered oxidative metabolic responses in vitro of alveolar macrophages from asymptomatic cigarette smokers. Am Rev Respir Dis 123:85-89.

Hollingworth A Barton SE Jenkins D Cuzick J Singer A 1987. Colposcopy of women with cervical HPV type 16 infection but normal cytology (letter). Lancet ii:1148. -1 6 4 -

Holly EA Petrakis NL Friend 1TF Sarles DL Lee RE Flander LB 1985. Mutagenic mucus in the cervix of smokers. J Rati Cancer Inst 76:983- 986.

Hoover R & Fraumeni JF Jr 1973. Risk of cancer in Renal-transplant patients. Lancet ii:55-57

Hughes RG Norval M Howie SEM 1988, Expression of major histocompatibility class II antigens by Langerhans' cells in cervical intraepithelial neoplasia. J Clin Path 41:253-259.

Hyden H & Lange PV 1970. S100 brain protein: correlation with behavior. Proc Natl Acad Sci 67 (4):1959-1966.

Ioachim HL 1976. The stromal reaction to tumours : An expression of immune surveillance, J Natl Cancer Inst 57:465-469.

Intercollegiate Working Party on Cervical Cytology Screening, 1987, RCOG. London.

Irwin KL Rosero-Bixby L Qberle MW Lee NC Whatley AS Fortney JA Bonhomme MG 1988. Oral contraceptives and cervical cancer risk in Costa Rica. JAMA 259:59-64.

Ishiguro T Sugitachi I Katoh K 1980. T and B lymphocytes in patients with squamous cell carcinoma of the uterine cervix. Gynecol Oncol 9:80-85.

Isobe T & Okuyama T 1978. The amino acid sequence of S100 protein and its relations to the calcium binding proteins. Eur J Biochem 89:379-388.

James K & Hargreave TB 1984. Immunosuppression by seminal plasma and its possible clinical significance. Immunol Today 5(12): 357-63. -165-

James K & Szymaniec S 1985. Human seminal plasma is a potent inhibitor of natural killer cell activity in vitro. J Reprod Immunol 8;61-70.

Jarmulowicz M Jenkins D Barton SE Goodall A Hollingworth A Singer A 1988. Cytological status and lesion size; a further dimension in CIF? Br J Obstet Gynaecol (In press).

Jarvis M Tunstall-Peddoe H Feyerabend C Vessey C Salloojee Y 1984. Biochemical markers of smoke absorption and self reported exposure to passive smoking. J Epid Comm Health 38:335-339.

Jenkins D Tay SK McCance DJ Campion MJ Clarkson PK Singer A 1986a. Histological and immunocytochemical study of cervical intraepithelial neoplasia

Jenkins D Tay SK Dyson JL 1986b. Koilocyte frequency and prevalence of cervical HPV infection (letter). Lancet i:557-558.

Kasahara T Hooks JJ Dougherty SF Oppenheim JJ 1983. Interleukin 2 mediated immune interferon (IFN-gamma) production by human T cells and T cell subsets, J Immunol 131:1784-1789.

Katz SI Tamaki K Sachs DH 1979. Epidermal Langerhans' cells are dervied from cells originating in bone marrow. Nature 282:324-326.

Kessler IJ 1977. Venereal factors in human cervical cancer: evidence from marital clusters. Cancer 39:1912-1919.

Khadim SA & Rees RC 1984. Enhancement of tumour growth in mice: evidence for the involvement of host macrophages. Cell Immunol 87:259-261. -166-

Khoo SK & Mackay EV 1974. Relation of cell mediated immunity in women with genital tract cancer to origin, histology, clinical stage and subsequent behaviour of neoplasm. J Obstet Gynaecol Br Commw 81:229-235.

Kietlinska Z 1984. T and B lymphocyte counts and blast transformation in patients with stage 1 cervical cancer. Gynecol Oncol 18:247-256.

Kienzle B Ellis S Verbi V Crumpton M McMichael A Knowles R 1987. Biochemical analysis of the Workshop antibodies with unknown specificities. In; Leucocyte typing III. Ed. McMichael AJ. OUP. Oxford.

Kinlen RJ Shell AGR Peto J Doll R 1979. Collaborative study of cancer in patients treated with immunosuppressive drugs. Br Med J 2:1461-1464.

Kitchener HC Cordiner JV Eglin RP 1982, Latency of herpes simplex virus in uterosacral ligaments. Am J Obstet Gynecol 143:839-840.

Kjorstad KE & Orjaseter H 1978. Carcinoembryonic antigen levels in patients with squamous cell carcinoma of the cervix. Obstet Gynecol 51:536-540.

Kohler G & Milstein C 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-498.

Koss LG & Wolinska VH 1959. Trichomonas Vaginalis cervicitis and its relationship to cervical cancer; a histocytological study. Cancer 12:1171-1193.

Kuller LH Garfinkel L Correa P Haley N Hoffmann D Preston-Martin S Sandler D 1986. Contribution of passive smoking to Respiratory Cancer. Env Health Persp 70:57-69. -1 6 7 -

Kung PC Goldstein G Reinherz EL 1979. Monoclonal antibodies defining distinctive human T cell surface antigens. Science 206:347-349.

Lancet Conference Report 1988. Human Papillomavirus and Cervical Cancer. Lancet 1:756-758.

Langerhans P 1868. Uber die Nerven der menschlichen Haut. Virchows Arch(A) 44:325.

Langley FA 1984. Consistency in the diagnosis of squamous carcinoma of the cervix and its precursors. In: Cancer of the uterine cervix: Biochemical and Clinical aspects. Eds. McBrien DCH & Slater TF. Academic Press. London.

Langone J Gjika HB Vunakis HV 1973. Nicotine and its metabolites. Radioimmunoassays for nicotine and cotinine. Biochemistry 12:5025- 5030.

Lanier LL Lee AM Phillips JH Varner NL Babcock GF 1983, Subpopulations of human natural killer cells defined by expression of the Leu 7 (HNK-1) and Leu 11 (NK-15) antigens. J Immunol 131:1789-1796.

La Vecchia C Franchesci S Decarli A Fasoli M Gentile A Tognoni G 1986. Cigarette smoking and the risk of cervical neoplasia. Am J Epid 123:22-29.

Ledbetter JA Evans RL Lipinski M Cunningham-Rundles C Good RA Herzenberger LA 1981. Evolutionary conservation of surface molecules that distinguish T lymphocytes of helper/inducer and cytotoxic/suppressor subpopulations in mouse and man. J Expl Med 153:310-323.

Levi MM 1971. Antigenicity of ovarian and cervical malignancies with a view towards possible immunodiagnosis. Am J Obstet Gynecol 109:689-670. -1 6 8 -

Levis WR & Dattner AM 1979. Selective defects in T cell function in ataxia telangectasia. Clin Exp Immunol 37:44-47.

Levitt D & Barol J 1987. The Immunobiology of Chlamydia. Immunol Today 8:246-251.

Levy S Kopersztych S Musatti CC Souen JS Sarvatore CA Mendes MF 1978. Cellular immunity in squamous cell carcinoma of the uterine cervix Am J Obstet Gynecol 130:160-164.

Lloyd JM O'Dowd T Driver M Tee DEH 1984. Immunohistochemical detection of Ca antigen in normal, dysplastic and neoplastic squamous epithelia of the human uterine cervix. J Clin Path 37:14- 19.

Lobach DF Singer KH Haynes BF 1987. Characterization of a subset of CD3 antibodies that react with human thymic epithelial cells. In; Leucocyte Typing III. Ed. McMichael AJ. Oxford University Press. Oxford.

Lyon JL Gardner JV Vest DV Standish VM Hebertson RM 1983. Smoking and carcinoma in situ of the uterine cervix. Am J Pub Health 73:558- 560,

McArdle JP & Muller HK 1986, Quantitative assessment of Langerhans’ cells in human cervical intraepithelial neoplasia and wart virus infection, Am J Obstet Gynecol 154:509-515.

McCance DJ Campion MJ Clarkson PK Chesters PM Jenkins D Singer A 1985. The prevalence of HPV type 16 DMA sequence in CIM and invasive carcinoma of the cervix. Br J Obstet Gynaecol 92:1101-1105.

McCance DJ Campion MJ Singer A 1986. Mon-invasive detection of cervical papillomavirus DMA (letter). Lancet i:558-559. -169-

McCullagh P & Nelder JA 1983. Generalized Linear Models. Chapman and Hall, London.

McDougall JK Crum CP Fenoglio CM Goldstein LC Galloway DA 1982. Herpes virus specific RNA and protein in carcinoma of the uterine cervix. Proc Natl Acad Sci USA 79:3853-3857.

Mclndoe V McLean M Janes R Mullins R 1984. The invasive potential of carcinoma in situ of the cervix. Obstet Gynecol 64:451-454.

McSharry C & Wilkinson PC 1986. Cigarette smoking and the antibody response to inhaled antigens (letter). Immunol Today 7 (4):98.

Mackenzie IC 1975. Labelling of murine epidermal Langerhans' cells with 3H thymidine. Am J Anat 144:125-136.

MacLean AB 1984a. Healing of cervical epithelium after laser treatment of cervical intraepithelial neoplasia. Brit J Obstet Gynaecol 91:697-706.

MacLean AB 1984b. Cervical healing and Langerhans cells. Brit J Obstet Gynaecol 91:1145-1148.

Maddox PH & Jenkins D 1987. 3-aminopropyltriethoxysilane (APES): a new advance in section adhesion. J Clin Path 40:1256-1257.

Maddox PH Tay SK Jenkins D 1987. A new fixed cryosection technique for the simultaneous demonstration of T6 and S100 antigens. Histochemical Journal 19:35-38.

Maddox PH Barton SE Jenkins D 1988. Detection of Langerhans' cells in cervical epithelium (letter), Am J Obstet Gynecol (in press) -1 7 0 -

Maitland NJ Kinross JH Busuttil A Ludgate SM Smart GE Jones KV 1981, The detection of DNA tumour virus specific RNA sequences in abnormal human cervical biopsies by in situ hybridisation. J Gen Virol 55:123-137.

Marshall JR Graham S Byers T Swanson M Brasure J 1983. Diet and smoking in the epidemiology of cancer of the cervix. J Natl Cancer Inst 70:847-850.

Martin RR 1973. Altered morphological and increased acid hydrolase content of pulmonary macrophages from cigarette smokers. Am Rev Respir Dis 107:596-601.

Martinez I 1969. Relationship of squamous cell carcinoma of the cervix to squamous cell carcinoma of the penis. Cancer 24:777-786.

Meanwell CA Blackledge G Cox MF Maitland NJ 1987. HPV 16 DNA in normal and malignant cervical epithelium: implications for the aetiology and behaviour of cervical neoplasia. Lancet i:703-707.

Meisels A & Fortin R 1976. Condylomatous lesions of cervix and vagina I. Cytologic patterns. Acta Cytologica 20:505-509.

Meisels A Morin C Casas-Cordero M 1982. Human papilloma virus infection of the uterine cervix. Int J Gynecol Path 1:75-94.

Melamed MR & Flehinger BJ 1973. Early incidence rates of precancerous cervical lesions in women using oral contraceptives. Gynecol Oncol 1:290-294.

Micksche M Luger T Michalica V Tatra G 1978. Investigations on general immune reactivity in untreated cervical cancer patients. Oncology 35:206-210.

Miller LG Goldstein G Murphy M Ginns LC 1982. Reversible alterations in immunoregulatory T cells in smoking. Chest 82:52-56. -1 7 1 -

Mims CA 1964. Aspects of the pathogenesis of virus diseases. Bact Rev 28:30-35.

Mims CA & White DO 1984. Viral Pathogenesis and Immunology. Blackwell. Oxford.

Mogensen SG 1979. Role of macrophages in natural resistance to virus infection. Microbiol Rev 43:1-9.

Moghissi KS Mack HC Porzak JP 1968. Epidemiology of cervical cancer. Am J Obstet Gynecol 100:607-614.

Moore M 1985. Natural Immunity to tumours-theoretical predictions and biological observations. Br J Cancer 52:147-151.

Morahan PS Connor JR Leary KR 1985. Viruses and the versatile macrophage. Brit Med Bull 41(1):15-21.

Moretta L Mingari MC Moretta A Fauci AS 1982. Human lymphocyte surface markers. Semin Haematol 19:273-283.

Morhenn VB 1988. Keratinocyte proliferation in wound healing and skin diseases. Immunol Today 9(4):104-107.

Mori R Nomoto K Takeya K 1964. Tumour formation by polyoma virus in neonatally thymectomised mice. Proc Jpn Acad 40:445-449.

Morris HHB Gatter KC Stein H Mason DY 1983a. Langerhans' cells in human cervical epithelium: an immunohistological study. Br J Obstet Gynaecol 90:400-411.

Morris HHB Gatter KC Sykes G Casemore V Mason DY 1983b. Langerhans' cells in human cervical epithelium: effects of wart virus infection and intraepithelial neoplasia. Br J Obstet Gynaecol 90:412-420. -1 7 2 -

Morris HHB & Price SK 1987. Langerhans' cells, papillomaviruses and oesophageal carcinoma - a hypothesis, SAMJ Mar 21 (Suppl):15-18.

Muller HK Halliday GM Knight BA 1985. Carcinogen-induced depletion of cutaneous Langerhans Cells. Br J Cancer 52:81-85.

Murphy GF Bhan AK Sato S Harris TJ Mihm MC 1981. Characterization of Langerhans cells by the use of monoclonal antibodies. Lab Invest 45:465-468.

Naguib SM Lundin FE Jr Davis HJ 1966. Relation of various epidemiological factors to cervical cancer as determined by a screening program. Obstet Gynecol 28:451-455.

Nahmias A Adelusi B Naib Z Muther J 1985. Changing concepts on the Relation of Genital Herpes and Cervical Cancer. In: Cancer campaign Vol. 8: Cancer of the Uterine Cervix. Gustav Fischer Verlag. Stuttgart.

Naib AM 1970. Cytology of TRIC agent infection of the eye of newborn infants and their mothers genital tracts, Acta Cytologica 14:390- 395.

Naib ZM Nahmias AJ Jo sey WE 1966. Cytology and histopathology of cervical herpes simplex infection. Cancer 19:1026-1031.

Naib ZM Nahmias AJ Josey WE Kramer JH 1969. Genital herpetic infection association with cervical dysplasia and cancer. Cancer 23:940-945.

Naisell K Naisell M Vaclavinkinova V Roger V Hjerpe A 1976. Followup studies of cytologically detected precancerous lesions (dysplasia) of the uterine cervix. In: Skandia International Symposia. Health control in the detection of Cancer. Almqvist and Wiksell International, Stockholm. -1 7 3 -

Nakajima T Vatanabe S Sato S Kameya T Hlrota T Shimosato Y 1982. An immunoperoxidase study of S100 protein distribution in normal and neoplastic tissues. Am J Surg Pathol 6:715-727.

Neill W & Norval M 1984. Natural killer cell activity in patients with abnormalities of the uterine cervix, Gynecologic Obstetric Invest 18:122-128,

Nezelof C Basset F Rousseau MF 1973. Histiocytosis X: histogenic arguments for a Langerhans cell origin. Biomedicine 18:365-371.

Noar D 1979. Suppressor cells: permitters and promoters of malignancy? Adv Cancer Res 29:45-125.

Norman SJ 1985. Macrophage infiltration and tumour progression. Cancer Metastasis Rev 4: 277-291.

Norton AJ Ramsay AD Smith SH Beverley PCL 1986. Monoclonal antibody (UCHL1) that recognises normal and neoplastic T cells in routinely fixed tissues. J Clin Path 39:399-405.

OPCS 1984. General Household Survey 1982. HMSO London.

OPCS 1985. Smoking among secondary school children. HMSO London.

Oriel JD 1983. Condyloma acuminata as a sexually transmitted disease. Dermatol Clin i:93-102.

Orr JV Wilson K Bodiford C Cornwell A Soong SJ Honea KL Hatch KD Shingleton HM 1985. Nutritional status of patients with untreated cervical cancer; Biochemical and immunologic assessment. Am J Obstet Gynecol 151:625-631. - m -

Orth G Favre M Breitburd F Croissant 0 Jablonska S Qbalek S Jarzabek-Chorzelka M Rzesa G 1980. Epidermodysplasia verruciformis: a model for the role of human papillomaviruses in human cancer. In: Viruses naturally occurring in Cancer. Cold Spring Harbor Conferences on Cell Proliferation 7:259-265.

Ory HV Conger B Richart R Barron B 1974. Relation of type 2 herpesvirus antibodies to cervical neoplasia. Am J Obstet Gynecol 43:801-904.

Ory HV Conger B Naib Z Tyler CV Hatcher RA 1977. Preliminary analysis of oral contraceptive use and risk of developing premalignant lesions of the uterine cervix. In: Pharmacology of steroid contraceptive drugs. Eds. Garratini S & Berendes HV. Raven New York.

Paavonen J, Vesterinen E, Meyer B Saksela E 1982. Colposcopic and Histologic Findings in cervical Chlamydial infection. Obstet Gynecol 59:712-714.

Pakes VL Muller HK Schwarz MA Marks R 1986. Langerhans cells - a reduction in numbers and their reappearance following steroid and cytotoxic therapy in humans. Clin Exp Dermatol 11:450-459.

Parkin DM Stjernsward J Muir CS 1984. Estimates of Vorldwide Frequency of twelve major cancers. Bulletin of the Vorld Health Organisation 62:163-182.

Paul C 1988. The New Zealand cervical cancer study: Could it happen again? BMJ 297:533-539.

Pehamberger H Stingl LA Pogantsch S Steiner G Volff K Stingl G 1983. Epidermal cell induced generation of cytotoxic T lymphocyte responses against alloantigens or TNP-modified syngeneic cells: Requirement for la positive Langerhans cells. J Invest Dermatol 81:208-211. -1?5

Perera FP Santella RM Brenner D Poirier MC Munshi AA Fischmann HK Ryzin JV 1987. DNA adducts, protein adducts, and sister chromatid exchange in cigarette smokers and nonsmokers, JNCI 79:449-456

Peritz E Ramcharan S Frank J Brown VL Huang S Ray R 1977. The incidence of cervical cancer and duration of oral contraceptive use. Am J Epid 106:462-465.

Peryra AJ 1961. The relationship of sexual activity to cervical cancer. Cancer of the cervix in a prison population. Obstet Gynecol 17:154-159.

Peto R 1986. Viral Aetiology of Cervical Cancer. Banbury Report 21. Cold Springs Harbor Laboratory. USA.

Pfister H 1987. Relationships of Papillomaviruses to Anogenital Cancer. Obstetrics and Gynecology Clinics of North America 14(2):349-361.

Phillips JH & Babcock CF 1983. NKP-15: a monoclonal antibody reactive against purified human Natural Killer cells and granulocytes. Immunol Lett 6:143.

Pomerantz RJ de la Monte SM Donegan P Rota TR Vogt MV Craven DE Hirsch MS 1988. Human Immunodeficiency virus (HIV) infection of the uterine cervix. Ann Int Med 108:321-327.

Popper K 1974. Normal Science and its Dangers .In: Criticism and the Growth of Knowledge. Eds: Lakatos I Musgrave A. CUP. London.

Porecco R Penn I Droegemueller V Greer B Makowski E 1975. Gynecological malignancies in immunosuppressed organ homograft recipients. Obstet Gynecol 45:359-364.

Powell LC & Seymour RG 1971. Effects of depot medroxyprogesterone acetate as a contraceptive agent. Am J Obstet Gynecol 110:36-40. -1 7 6 -

Pulay TA Benczur M Varga M 1982. Natural Killer lymphocyte function in cervical cancer patients. Neoplasma 29:237-240.

Rand RJ Jenkins DM Bulmer R 1977. T- and B- lymphocyte subpopulations in preinvasive and invasive carcinoma of the cervix. Clin Exp Immunol 30:421-428.

Rasp FL Clawson CC Hoidal R Repine JE 1978. Reversible impairment of the adherence of alveolar macrophages from cigarette smokers . Am Rev Respir Dis 118:979-986.

Rawls WE Gardner HL Kaufman RL 1970. Antibodies to genital Herpes viruses in patients with carcinoma of the cervix. Am J Obstet Gynecol 107:710-716.

Reagan JV & Patten SF Jr 1962. Dysplasia: A basic reaction to injury in the uterine cervix. Ann NY Acad Sci 97:661-682.

Reid BL 1971. Natural History of the origins of cervical cancer. In: Scientific Basis of Obstetrics and Gynaecology. Ed. MacDonald RR. Churchill Livingstone . Edinburgh.

Reid R Stanhope CR Herschman Crum CP Agronow SJ 1984a. A colposcopic index for differentiating subclinical papillomaviral infection from cervical intraepithelial neoplasia. Am J Obstet Gynecol 149:815-823.

Reid R Herschman BR Crum CP Fu YS Braun L Shah KV Agronow SJ Stanhope CR 1984b. The tissue basis of colposcopic change. Am J Obstet Gynecol 149:293-304.

Reid R & Scalzi P 1985, An improved colposcopic index for differentiating benign papillomaviral infections from High grade cervical intraepithelial neoplasia. Am J Obstet Gynecol 153:611-618.

Reid VA Branch T Thompson WD Kay J 1987. The effect of diffusion on the immunolocalization of antigen. Histopathology 11:1277-1284. “ 177-

Re inherz EL Meuer SC Schlossman SF 1983. The human T cell receptor: analysis with cytotoxic T cell clones. Immunol Rev 74:83-112.

Richardson AC & Lyon JB 1981. The effect of condom use on squamous cell cervical intraepithelial neoplasia. Am J Obstet Gynecol 141:909-913.

Richart RM 1967. Natural History of cervical intraepithelial neoplasia. Clinical Obstetrics and Gynecology 10:748-784.

Rigoni-Stern D 1842, Fatti Statistici relativi alle mallattie cancrose che servirono de base alle cose dette dal dott. Gior Servire Progr Pathol Temp 2:507-519.

Ritchie AWS James K Micklem HS 1983. The distribution and possible significance of cells identified in human lymphoid tissue by the monoclonal antibody HNK-1. Clin Exp Immunol 51:439-447.

Robinson KM Haffejee AA Angorn IB 1981. Inhibition of mitogen induced lymphocyte proliferation by autologous serum in oesophageal carcinoma. S Afr Med J 59 (6):187-189.

Roder JC Haliotis T Klein M Korec S Jetts JR Ortaloo J Herberman RB Katz P Fauci AS 1980, A new immunodeficiency disorder in humans involving NK cells. Nature 284:553-555.

Roitt I Brostoff J Male D 1985. Immunology. Gower Medical. London.

Rojel J 1953. Interrelation between uterine cancer and syphilis. Acta Pathologica et Microbiologica Scandinavica 97(Suppl.):66-67.

Rotkin ID 1967. Adolescent coitus and cervical cancer: association of related events with increased risk. Cancer Res 27:603-617.

Romny SL Palan PR Dattagupta C 1981. Retinoids and the prevention of cervical dysplasia, Am J Obstet Gynecol 141:890-893. -1 7 8 -

Rous P & Beard JV 1935, The progression to carcinoma of virus- induced rabbit papillomas (Shape). J Exp Med 62:523-528.

Royston K & Aurelian L 1970.The association of genital herpesvirus with cervical atypia and carcinoma in situ. Am J Epid 91:531-538.

Sandmire HF Austin SD Bechtel RC 1976. Carcinoma of the cervix in oral contraceptive steroid and IUD users and non-users. Am J Obstet Gynecol 125;339-342.

Sasson IM Haley NJ Hoffmann D Vynder E Hellberg D Nilsson S 1985. Cigarette smoking and cervical neoplasia ;smoke constituents in cervical mucus (letter), N Eng J Med 312:315-316.

Satam MN Nadkarni JJ Nadkarni JS Raj pal RM 1981. Immune status in untreated cervical cancer patients. Neoplasma 28:111-116.

Sawanobori S Ashman RB Nahmias AJ Benigno BB LaVia MF 1977. Rosette formation and inhibition in cervical dysplasia and carcinoma in situ. Cancer Res 37:4332-4335.

Scala G Allarena P Djeu JY Kasahar T Ortaldo JR 1984. Human large granular lymphocytes are potent producers of interleukin I. Nature 309:57-59.

Schachter J Hill EC King EB 1975. Chlamydial infection in women with cervical dysplasia . Am J Obstet Gynecol 123:753-757.

Schneider A Kraus H Schuhmann R Gissman L 1985. Papillomavirus infection of the lower genital tract: detection of viral DNA in gynaecological swabs. Int J Cancer 35:443-448.

Schneider V Kay S Lee HM 1983. Immunosuppression as a high risk factor in the development of condyloma acuminatum and squamous neoplasia of the cervix. Acta Cytologica 27 (3):220-224. - 1 7 9 -

Schwarz E Freese UK Gissman L Mayer V Roggenbuck B Stemlau A zur Hausen H 1985. Structure and transcription of human papillomavirus sequence in cervical carcinoma cells. Mature 314:111-114.

Schwartz DB Greenberg MD Daoud Y Reid R 1987. The management of genital condylomas in pregnant women. Obstet Gynecol Clinics of Morth America 14 (2):589-599.

Sepkovic DV Axelrad CM Colosimo SG Haley MJ 1987. Measuring Tobacco smoke exposure: clinical applications and pasive smoking. Presented to the Annual Meeting of the Association dedicated to Air Pollution Control, Mew York June 1987.

Seth P Prakash SS Ghosh D 1978. Antibodies to herpes simplex virus types 1 and 2 in patients with squamous cell carcinoma of uterine cervix in India. Int J Cancer 22:708-714.

Shelley V & Juhlin L 1976. Langerhans' cells form a reticuloendothelial trap for external contact antigens, Mature 261:46-47.

Shiohara T Kobayashi M Marimatsu H Magashima M 1987. Effect of orally administered retinoid on murine Langerhans cells. Arch Dermatol Res 279:198-203.

Si 11man F Stanek A Sedlis A Rosenthal J Lanks KV Buchhagen D Micastri A Boyce J 1984. The relationship between human papillomavirus and lower genital tract intraepithelial neoplasia in immunosuppressed women. Am J Obstet Gynecol 150:300-308.

Silverman NA Potvin C Alexander JC Chretien PB 1975. In vitro lymphocyte activity and T cell levels in chronic cigarette smokers. Clin Exp Immunol 22:285-292.

Singer A 1973. D.Phil Thesis: A Study of the cervix uteri of women in prison. University of Oxford. - 1 8 0 -

Singer A Reid BL Coppleson H 1976, The role of the high risk male in the aetiology of cervical cancer; a correlation of epidemiology and molecular biology. Am J Obstet Gynecol 126:110-116.

Singer A & Jordan JA 1977, Oestrogen like activity and cervical atypias (letter).Lancet ii:359.

Singer A & McCance DJ 1985, The wart virus and genital neoplasia; a casual or causal association? Br J Obstet Gynaecol 92:1083-1085,

Singer A & Tay SK 1986. Smoking and cervical neoplasia; a serious public health problem. J Obstet Gynaecol; 6:989-990.

Sinkovics JG & Dreesman GR 1983. Monoclonal antibodies of hybridomas. Rev Infect Dis 5:9-34.

Skinner GRB Thouless ME Jordan JA 1971. Antibodies to type 1 and type 2 herpes virus in women with abnormal cervical cytology. J Obstet Gynaecol Br Commw 78:1031-1038.

Slavin G 1976. The pathology of cervical inflammatory disease, In:The cervix, Eds, Jordan JA & Singer A. VB Saunders, London.

Solheim BG Moller E Ferrone S (eds.) 1986. HLA class II antigens: A comprehensive review of structure and function. Springer Verlag. Berlin.

Stafl A 1981. Cervicography: a new method for cervical cancer detection. Am J Obstet Gynecol 139:815-818.

Stern E Forsythe AB Youkeles L Ciffelt CF 1977. Steroid contraceptive use and cervical dysplasia: Increased risk of progression. Science 196:1460-1464.

Stellman SD Austin H Vynder EL 1980. Cervix cancer and cigarette smoking: a case control study. Am J Epid 111:383-386. -181-

Stewart GJ Spurrett BR Shelley-Jones D 1988. Cervical intraepithelial neoplasia in HIV infected women. Presented to the Fourth World AIDS congress, Stockholm, Sweden.

Stingl G Katz SI Clement L Green I Shevach EM 1978. Immunologic Function of la bearing epidermal Langerhans cells. J Immunol 121:2005-2008.

Stratton JA & DiSaia PJ 1982. Depressed mononuclear cell function in advanced neoplastic disease. Am J Rep Immunol 2:111-116.

Streikin JW Toews GT Gilliam JM Bergstresser PR 1980. Tolerance or hypersensitivity to 2,4 dinitro-l-flourobenzene: the role of Langerhans' cell density within epithelium. J Invest Dermatatol 74:319-322.

Sulke AM Jones DB Wood PJ 1985. Variation in natural killer cell activity in peripheral blood during the menstrual cycle. Br Med J 290:884-886.

Swan SH & Pettiti DB 1982. A review of the problems of bias and confounding in epidemiological studies of cervical neoplasia and oral contraceptive use. Am J Epid 115:10-18.

Syrjanen K 1983, Immunocompetent cells in uterine cervical lesions of human papillomavirus origin. Gynecol Obstet Invest 16:327-340,

Syrjanen K Mantyjarvi R Varyrynen M 1984. Herpes simplex virus infection of females with human papillomavirus (HPV) lesions in the uterine cervix. Cervix 2:25-32.

Syrjanen K Mantyjarvi R Varyrynen M Parkkinen S Holopainen H Syrjanen SSaarikoski S Castren 0 1986, Coexistent chlamydial infections related to natural history of human papillomavirus lesions in uterine cervix. Genitourin Med 62:345-351. -1 8 2 -

Tagami H Ogino A Tagikawa M Imamura S Ofuji S 1974. Regression of plane warts following spontaneous inflamination: A histopathological study. Br J Dermatol 90:147-154.

Tagami H Oguchi M Ofuji S 1983. Immunological aspects of wart regression with special reference to regression phenomenon of numerous flat warts. An experiment on tumor immunity in man by nature. J Dermatol (Tokyo) 10:1-8.

Tamaki K Stingl G Katz S 1980. The origin of Langerhans' cell. J Invest Dermatol 74:309-311.

Tarter TH Cunningham-Rundles S Koide SS 1986. Suppression of Natural Killer cell activity by human seminal plasma in vitro: identification of 19-OH-PGE as the supppressor factor. J Immunol 136:2862-2867.

Tay SK Jenkins D Maddox P Campion M Singer A 1987a. Subpopulations of Langerhans' cells in cervical neoplasia. Br J Obstet Gynaecol 94:10-15.

Tay SK Jenkins D Maddox P Singer A 1987b. Lymphocyte phenotypes in Cervical Intrapithelial Neoplasia and Human Papillomavirus infection.Br J Obstet Gynaecol 94:16-21.

Tay SK Jenkins D Maddox P Hogg N Singer A 1987c. Tissue macrophage response in human papillomavirus infection and cervical intraepithelial neoplasia. Br J Obstet Gynaecol 94:1094-1097.

Tay SK Jenkins D Singer A 1987d. Natural killer cells in cervical intraepithelial neoplasia and human papillomavirus infection. Br J Obstet Gynaecol 94:901-906.

Thivolet J. Viac J Staquet MJ 1982. Cell mediated immunity in wart infection. Int J Dermatol 21:94-98. ■'183-

Thomas BJ Evans RT Hawkins DA Taylor-Robinson D 1984, Sensitivity of detecting Chlamydia trachomatis elementary bodies in smears by use of a flourescein labelled monoclonal antibody: comparison with conventional chlamydial isolation. J Clin Path 37:812-816.

Thomas DB 1972, Relationship of oral contraceptives to cervical carcinogenesis, Obstet Gynecol 139:52-55.

Thomas DB 1973, An epidemiologic study of carcinoma in situ and squamous dysplasia of the uterine cervix. Am J Epid 98:10-13.

Towne JE 1955.Carcinoma of the cervix in nulliparous and celibate women. Am J Obstet Gynecol 69:606-613.

Trevathan E Layde P Webster LA Adams JB Benigno BB Ory HW 1983. Cigarette smoking and dysplasia and carcinoma of the uterine cervix. JAMA 250:499-502.

Turner MJ White JO Soutter WP 1988. Natural killer cells in cervical intraepithelial neoplasia and human papillomavirus infection (letter). Br J Obstet Gynaecol 95:423,

Van der Linde AW Streefkerk M Schuurmann HJ te Velde ER Kater L 1983. Divergence between the occurrence of antibody and cellular immune reactivity to cervical carcinoma cell lines in preinvasive and microinvasive stages of cervical carcinoma. Br J Cancer 47: 147- 153.

Varyrynen M Syrjanen K Mantyjarvi R Castren 0 Saarikoski S. 1984. Langerhans cells in human papillomavirus (HPV) lesions of the uterine cervix identified by the monoclonal antibody OKT-6. Int J Gynaecol Obstet 22:375-383. te Velde ER Persijn JP Ballieux RE Faber J 1982. Carcinoembryonic antigen serum levels in patients with squamous cell carcinoma of the uterine cervix: clinical significance. Cancer 49:1866-1873. -184-

Vessey MP Lawless M McPherson K Yeates D 1983a. Neoplasia of the cervix uteri and contraception: A possible adverse effect of the pill. Lancet ii:930-932.

Vessey MP Lawless M McPherson K Yeates D 1983b. Oral contraceptives and cervical cancer. Lancet 11:1358-1359.

Vessey MP 1986, Epidemiology of Cervical cancer: Role of hormonal factors, Cigarette smoking and Occupation. In: Viral Aetiology of Cervical cancer. Banbury report 21. Cold Springs Harbor Laboratory.

Von Krogh G 1979. Warts: immunologic factors of prognostic significance. Int J Dermatol 18:195-204.

Wald NJ Nanchahal K Thompson SG Cuckle HS 1986. Does breathing other peoples' tobacco smoke cause lung cancer? Br Med J 293:1217-1222.

Walker PG Singer A Dyson JL Shah KV WiIters J Coleman DV 1983. Colposcopy in the diagnosis of papillomavirus infection of the uterine cervix. Br J Obstet Gynaecol 90:1082-1086.

Wentz WB Louis GC Reagan JW Heggie AD Fu Y-S Antony DD 1981. Induction of uterine cancer with inactivated herpes simplex virus type 1 and 2, Cancer 48:1783-1790.

Wigle DT & Grace M 1980, Re: Smoking and cancer of the uterine cervix: Hypothesis, Am J Epid 111:125-127,

Wildy P & Gell PGH 1985, The Host Response to herpes simplex virus. Brit Med Bull 41(1);86-91.

Witkin SS Hirsch J Ledger WJ 1986. A macrophage defect in women with recurrent Candida vaginitis and its reversal in vitro by prostaglandin inhibitors. Am J Obstet Gynecol 155:790-795. -1 8 5 -

Witkin SS 1987. Transient local immunosuppression in recurrent vaginitis (letter). Immunol Today 8(12):360,

Wolfendale MR King S Usherwood M 1983. Abnormal cervical smears; are we in for an epidemic ? BMJ 287:526-528.

World Health Organization 1978. Immunodeficiency, Technical Reports Series 630. WHO. Geneva.

World Health Organization 1984. Breast cancer cervical cancer and depot medroxy progesterone acetate. Lancet ii:1207-1208.

Worth AJ & Boyes DA 1972. A case control study into the passible effects of birth control pills on pre-clinical carcinoma of the cervix. J Obstet Gynaecol Br Commw 79:673-677.

Wright H Vessey MP Kenward B McPherson K Doll R 1978. Neoplasia and dysplasia of the cervix uteri and contraception: a possible protective effect of the diaphragm. Br J Cancer 38:273-279.

Wright N & Alison M 1984. The Biology of epithelial cell populations Vol 1. Clarendon Press, Oxford.

Wynder EL Cornfield J Schroff PD Dorarswami KR 1954. A study of the environmental factors in carcinoma of the cervix. Am J Obstet Gynecol 68:1016-1052.

Younes MS Robertson EM Bencosme SA 1968. Electron microscope observations of Langerhans cells in the cervix. Am J Obstet Gynecol 102:397-403.

Young E & Taylor HR 1986. Immune mechanisms in chlamydial eye infection: developments of T suppressor cells. Invest Opthalmol Visual Sci.27:615-619. -1 8 6 - zur Hausen H 1982. Human genital cancer: Synergism between two virus infections or synergism between a virus infection and initiating events? Lancet ii:1370-1372. zur Hausen H 1986. Intracellular surveillance of persisting viral infections. Lancet 11:489-491. -1 8 7 -

SUPPLEMEFT: Papers by the author which are referenced In this thesis and submitted In support of his candidature

Barton SE Thomas BJ Taylor-Robinson D Goldmeier D 1985. Detection of

Chlamydia trachomatis in the vaginal vault of women who have had hysterectomies. Br Med J 291:250.

Barton SE Wright LK Link CM Munday PE 1986. Screening to detect asymptomatic shedding of herpes simplex virus (HSV) in women with recurrent genital HSV infection. Genitourin Med 62:181-185.

Barton SE Davis JM Moss VW Tyms AS Munday PE 1987. Asymptomatic shedding and subsequent transmission of genital herpes simplex virus. Genitourin

Med 63:102-105.

Barton SE Taylor-Robinson D Harris JRV 1987. Female prostitutes and sexually transmitted diseases. Br J Hosp Med 38(1) .*34-45.

The clinical work presented in these papers was performed solely by the candidate. In addition, supervised participation in the associated laboratory work was undertaken. This included the cell culture and identification of herpes simplex virus and the use of the direct monoclonal antibody test for Chlamydia trachomatis. The support and cooperation of'the candidate's fellow authors in these papers is fully acknowledged. Reprinted from the BRITISH MEDICAL JOURNAL, 27th July 1985, 291, 250 - 1 8 8 -

Detection of Chlamydia trachomatis that chlamydiae do not infect squamous epithelium. The report of chlamydial vaginitis in a postmenopausal woman with cervical carci­ in the vaginal vault of women who noma and an irradiated atrophic vagina was, therefore, surprising. No have had hysterectomies less surprising was our finding of chlamydiae in 24% of vaginal smears from otherwise healthy premenopausal women who had had hysterec­ tomies. Hitherto such women have not been considered to be at risk of Chlamydia trachomatis is thought to infect the columnar epithelium of harbouring chlamydiae in the vagina. As we believe that our patients the endocervix but not the squamous epithelium of the vagina.1 were an unbiased sample of our clinic population, we would expect After a report of chlamydial vulvovaginitis in a woman who had had a further studies to find a similar prevalence. hysterectomy2 we screened women who had had hysterectomies to Two patients had vaginitis, which resolved after antichlamydial determine the prevalence of C trachomatis in the vagina. treatment. Whether C trachomatis can truly infect the vaginal squamous epithelium and cause vaginitis, however, is uncertain. The organisms may have contaminated the surface of vaginal cells, having been Patients, methods, and results derived from close anatomical sites—namely, the urethra, rectum, and Bartholin’s glands—or from a sexual partner. Indeed, as a urethral In the last three months of 1984, 17 women who had had hysterectomies rather than a vaginal specimen is more likely to be positive forN attended the Praed Street Clinic. In addition to performing microscopy of gonorrhoeae in women who have had hysterectomies5 the prevalence of Gram stained urethral and vaginal smears and dark ground microscopy C trachomatis may have been underestimated in our study because other of a wet smear from the vagina we collected urethral, vaginal, and rectal specimens for culture of Neisseria gonorrhoeae and took a smear from the sites were not sampled. Nevertheless, the identification ofC trachomatis vaginal vault to be examined for C trachomatis using a direct monoclonal in the vaginal vaults of women who had had hysterectomies is impor­ antibody test (Syva).3 tant for two reasons: firstly, the organisms may serve as a source of C trachomatis was detected in smears from four of the 17 women (table). infection for the patients’ sexual partners, and, secondly, they may All had had their ovaries conserved at hysterectomy, and serum follicle cause symptoms in these patients. Sexually active women who have had stimulating hormone and luteinising hormone concentrations of <15 U/l hysterectomies should therefore be screened for vaginal chlamydiae, confirmed that they were premenopausal. particularly as these may be eliminated by treatment. At the next visit the women were examined colposcopically. Of the three who had presented with vaginal discharge, two (cases 3 and 4) had vaginal inflammation most pronounced in the vault, with a white mucous discharge. 1 Schachter J, Grossman M. Chlamydial infections. Ann Rev Med 1981;32:45-61. One patient (case 1) had been treated for gonorrhoea one week previously, 2 Goldmeier D, Ridgway GL, Oriel JD. Chlamydial vulvovaginitis in a postmeno­ pausal woman. Lancet 1981 ;ii :476-7. and examination confirmed that her discharge had resolved. Four ulcerated 3 Thomas BJ, Evans RT, Hawkins DA, Taylor-Robinson D. Sensitivity of detecting areas were seen in the upper third of the vagina in case 2, from which herpes Chlamydia trachomatis elementary bodies in smears by use of a flourescein simplex virus was isolated in cell culture. No cervical remnants were present labelled monoclonal antibody: comparison with conventional chlamydial isola­ tion. J Clin Pathol 1984;37:812-6. in any of the women. 4 Oriel JD, Ridgway GL. Genital infection by Chlamydia trachomatis. London: All patients were treated with doxycycline 100 mg twice daily for seven Edw ard Arnold, 1982. days. The vaginal discharge in cases 3 and 4 resolved; repeat examination 5 Klaus BD, Chandler JE, Dans PE. Gonorrhea detection in posthysterectomised confirmed the absence of vaginitis, and tests for C trachomatis one and four patients. JAMA 1978;240:1360-1. weeks after treatment gave negative results. (.Accepted 8 M ay 1985) The sexual partners of the patients were examined and tested for C trachomatis (table). The husband of one patient (case 2) had urethritis and was the only contact whose urethral smear was positive for C trachomatis. The husband of the patient in case 1 was being treated for gonorrhoea; Praed Street Clinic, St Mary’s Hospital, London W2 IPG, and although C trachomatis was not detected in his urethral smear, he subsequently Division of Sexually Transmitted Diseases, MRC Clinical Re­ needed treatment for postgonococcal urethritis. search Centre, Harrow, M iddlesex HA1 3UJ S E BARTON, bsc, mb, clinical research fellow B J THOMAS, phd , senior research officer Comm ent D TAYLOR-ROBINSON, md , frcpath , consultant microbiologist D GOLDMEIER, md , mrcp , consultant venereologist Vaginal colonisation byC trachomatis in neonates results from mater­ Correspondence to: Dr Barton, Praed Street Clinic. nal contamination during delivery.'* It is generally accepted, however,

Details of women who had had hysterectomies in whom Chlamydia trachomatis was detected in the vaginal vault

Case Age Reason for Concurrent No (years) hysterectomy Presenting complaint infection Sexual partners in previous three months

1 24 Pelvic tuberculosis Vaginal discharge Gonorrhoea Husband had postgonococcal urethritis; 10 contacts untraceable 2 49 Fibroids Genital ulceration Herpes simplex Husband had non-gonococcal urethritis; positive for C trachomatis 3 43 Sterilisation Vaginal discharge Regular boyfriend asymptomatic, negative for C trachomatis 4 44 Menorrhagia Vaginal discharge, dyspareunia Husband asymptomatic, negative for C trachomatis; casual boyfriend untraceable

BMJl1232/85 COPYRIGHT © 1985 ALL RIGHTS OF REPRODUCTION OF THIS REPRINT ARE RESERVED IN ALL COUNTRIES OF THE WORLD Genitourin Med 1986;62:181-5

Screening to detect asymptomatic shedding of herpes simplex virus (HSV) in women with recurrent genital HSV infection

S E BARTON,* L K WRIGHT,t C M LINK,* AND P E MUNDAY*|

From the *Praed Street Clinic, St Mary’s Hospital, London, *f Division and the of Sexually Transmitted Diseases, MRC Clinical Research Centre, Harrow, Middlesex

SUMMARY To investigate the asymptomatic shedding of herpes simplex virus (HSV) from women with recurrent genital herpes infection, and to assess whether inapparent shedding could occur, eight such women were examined thrice weekly for one month. At each visit colposcopy was performed and multiple sites sampled for HSV. During the study four women had no recurrence of HSV infection, but four had at least one positive viral culture. One of these patients was asymptomatically shedding HSV on nine of her 11 clinic visits. Two episodes of urethral shedding were detected. In this group of patients the presence of inguinal lymphadenopathy was appreciably associated with the isolation of HSV from the urogenital tract.

Introduction shedding the virus .5 Though this was claimed to represent true asymptomatic shedding, specimens Genital herpes simplex virus (HSV) infection is the were obtained at visits initiated by patients, which commonest cause of genital ulceration in women in suggested that some patients might have had the United Kingdom. One of the most difficult symptoms that they did not attribute to HSV questions asked by patients with genital HSV infection infection. Furthermore, the effect on viral shedding of is whether they can be infectious in the absence of the treatment under study, photoactivated proflavin symptoms. Most clinicians answer that this treatment, is not known. phenomenon, so called “asymptomatic shedding”, Two other groups of workers have studied women has been observed but is believed to be rare and that attending outpatient clinics. Ferrer et al took cervical there are no data to predict when it might occur. The and vaginal specimens for HSV culture from 123 term “asymptomatic shedding” is open to women in a Puerto Rican sexually transmitted diseases misinterpretation. It may be used to indicate isolation (STD) clinic and described three (2-5%) as of HSV from a clinically obvious lesion of which the asymptomatic vaginal shedders; however, one woman patient is unaware: or from a lesion, such as a cervical presented with symptoms of vaginitis and another lesion, that is accessible to the doctor but not the with secondary syphilis .6 Similarly McCaughtry et al patient; or from an intact, apparently normal, investigated 585 women attending a college health epithelial surface — that is, inapparent shedding. clinic and found five “asymptomatic” cervical There have been several studies of asymptomatic shedders; four had clinical candidiasisand one vaginal shedding in pregnant women ,1"3 but other groups of bleeding.7 Thus serious doubts must be raised about women have been less intensively investigated. whether these workers had detected true Guinan et al obtained cervical swabs once weekly asymptomatic shedding or HSV infection in women from women with proved genital HSV infection and with symptoms that they did not attribute to herpetic recorded asymptomatic shedding in one (1*6%) of 64 infection. Furthermore, the examination findings are patients .4 Adam et al followed up women with HSV not sufficiently detailed to decide whether these infection who were taking part in a treatment trial, episodes were inapparent as well as supposedly and claimed that five (10%) out of 50 women were asymptomatic. In an important study, Rattray et al selected six women with proved genital HSV infection Address for reprints: Dr P E Munday, Pracd Street Ginic, St Mary’s Hospital, Praed Street, London W2 1NY and examined them twice weekly .8 Two women had episodes of asymptomatic viral shedding; one of them Accepted for publication 27 September 1985 had shedding from the cervix on a single occasion, and

181 - 1 9 0 -

182 S E Barton, L K Wnght, C M Link, and P E Munday the other on two occasions had shedding from the naked eye or colposcope. Except for the culture for vulva, which was accompanied once by ulceration Neisseria gonorrhoeae and the serological test for seen on examination. syphilis, all tests mentioned were performed at every To investigate asymptomatic shedding of HSV, we visit. studied prospectively, both virologically and clinically, using the colposcope 9 a group of women with recurrent HERPES SIMPLEX VIRUS ISOLATION genital HSV infection. We also examined the inguinal Swabs were placed in 1 ml of viral transport medium lymph nodes to investigate further the finding of containing nutrient broth, penicillin (1000 IU/ml), Corey et al that about one third of recurrent episodes of streptomycin (1000 IU/ml), and ampicillin (50 HSV infection in women are accompanied by tender mg/ml), and were processed immediately or stored at inguinal lymphadenopathy .10 -70°C. A 0*2 ml volume of the sample was inoculated into two tubes containing monolayers of human embryonic lung fibroblasts (MRC5, Flow Patients and methods Laboratories, Rickmans worth, England), which were incubated at 37°C. The monolayers were observed daily for cytopathic effect. A specimen was considered The same clinician examined thrice weekly for one to be negative if a characteristic change had not month eight women with a history of culture proved occurred after seven days. Monolayers exhibiting a genital HSV type 2 infection. Table I shows the cytopathic effect were harvested by scraping the cells demographic and clinical features of the eight into the medium, and the suspensions were frozen at patients. Women were selected because they were well -70°C until used for virus typing. motivated and able to attend often for the examinations and because they had expressed worries Results about asymptomatic shedding. During the study period each woman kept a diary to record symptoms, Though none of the patients believed that they had sexual activity, and stress events. At each visit this ever infected a sexual partner, all of them had reduced diary was discussed with the patient and she was asked the incidence of sexual intercourse since HSV the specific questions “Do you have any symptoms of infection had been diagnosed. On detailed questioning a herpes attack?” and then “Do you have any this was found to be due to a fear of infecting their symptoms at all?” The patient was then examined with uninfected sexual partner, though the extent to which particular attention being paid to palpation of guilt feelings and other psychological manifestations inguinal lymph nodes, and a scoring system ranging of HSV infection also caused this was uncertain. The from 0 to 3 was used to grade the degree of inguinal motivation of these women to be studied in such an lymphadenopathy. The vulva, vagina, and cervix were intensive way, however, was directly related to their examined with the naked eye and colposcope. Routine worries about asymptomatic shedding. tests to exclude syphilis, gonorrhoea, and candidal During the study period of 234 days and 88 patient and trichomonal infection were performed according visits, four of the women had no clinical or virological to standard clinic procedures. Specimens for HSV evidence of recurrences of HSV infection. The other culture were taken from the urethra, vulva (sweep of four, however, each yielded at least one positive HSV the labia majora), cervix, and any lesions seen with the culture, all of which were confirmed as HSV type 2. table I Characteristics o f patients in study group

Demographic/ clinical features Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8 Age (years) 23 23 25 33 30 23 33 20 Race White White Black White White White White White Marital status Single Single Single Single Single Single Single Single Occupation Psychologist Nurse Secretary Designer Press Officer Nurse Negotiator Nurse Parity 0 + 0 0 + 0 0 + 1 1 + 2 0 + 0 0 + 0 3 + 1 0 + 1 Contraception Cap Sheath None Sheath Oral Cap Sterilised Oral Previous genital infection Candidal NGU PID None NGU Gonorrhoea None None vaginitis contact contact Age at coitarche 18 16 21 16 17 17 17 17 Total lifetime partners 5 20 3 30 15 10 10 1 New partners since diagnosis 0 2 0 0 1 1 2 0 Months since diagnosis 3 8 5 2 5 5 2 5 No of symptomatic recurrences 1 1 1 1 I 2 1 1 NGU = non-gonococcal urethritis. PID = pelvic infammatory disease. - 1 9 1 -

Screening to detect asymptomatic shedding of HSV in women with recurrent infection 183 The details of these patients are presented below. c a s e 3 One patient (case 3) had two reccurrences. On day 8, despite her being asymptomatic, colposcopic examination of her right labium major showed erythema and vesicles from which HSV was isolated. The following day symptoms developed, and ulceration was visible to the naked eye. Viral culture was again positive on day 10. On day 13, though HSV was not isolated from the routine cervical sweep, a colposcopically directed swab of an ulcerated area (figure) gave positive results. The virus was not isolated from the vulva or cervix on day 15, by which time the vulval lesion was a scab and the cervical ulcer had healed. On day 19 of the study the patient was asymptomatic, but a urethral swab was positive for HSV. Thus, this patient had one vulval and cervical recurrence, which was initially asymptomatic but became symptomatic, and one asymptomatic inapparent urethral recurrence.

CASE 5 One patient (case 5) had two recurrences. On day 12 vulval erythema was seen colposcopically; by day 14 the patient had become symptomatic, the lesion being visible to the naked eye and culture positive. The virus was not isolated after day 17 from the typical ulcers, which had healed by day 20. On day 27 a tiny vulval sore, which was culture positive, was seen colposcopically. At this visit the patient had no genital f i g u r e Colpopholograph showing an ulcer and four vesicles symptoms, but felt that she had experienced the on anterior lip of cervix of case 3. prodromal symptoms of an attack for the previous two days. These symptoms, intermittent headache and Thus this patient had an initially asymptomatic backache, had disappeared on the day of the cervical recurrence, followed by symptomatic vulval examination. She reported that these warnings were and buttock recurrences. usually, but not inevitably, followed by vulval sores. By the next visit on day 29, the sores had crusted and CASE 8 the virus was not isolated. Thus this patient had one HSV was isolated from the urogenital tract of this symptomatic vulval recurrence and one apparently patient at nine of 11 examinations (table II). She asymptomatic vulval recurrence during the study experienced symptoms of early morning dysuria on period. table II Clinical and virological findings o f herpes simplex CASE 6 virus (HSV) in case 8 One patient (case 6) had one recurrence at three sites consecutively. On day 13 a virus positive colpos­ Day of study copically apparent cervical lesion was seen, but 1 3 5 8 10 12 16 18 24 26 28 the patient was asymptomatic. Two days later a virus negative vaginal ulcer was seen and a symptomatic Symptoms + Signs on: vulval recurrence, which was virus positive, began. By Cervix + + + + + + + day 18 the cervical, vaginal, and vulval sites had healed Vulva + + + Isolation of HSV from: and the virus was not isolated from them, but on the Cervix + + + + + + + + right buttock a virus positive vesicular eruption Vulva + + + appeared, which caused considerable pain and healed Urethra + after seven days. + = positive findings. -192-

184 S E Barton, L K Wright, C M Link, and P E Munday days 5 and 6 , and on day 6 HSV was isolated from the Sampling multiple anatomical sites and using the urethra. On day 18 she reported a white vaginal colposcope appears to have uncovered the “tip of an discharge, which had increased in amount over the iceberg” of asymptomatic shedding of HSV in women. previous two days. On colposcopic examination virus On at least one day when they did not have symptoms positive cervical and vulval lesions were seen, though of HSV infection, four of our eight patients had HSV they were not visible to the naked eye. isolated from urogenital tract specimens. Thus they At all other times throughout the study period the were a potential source of infection for their sexual patient was asymptomatic and at no time did she think partners, though the degree of risk is uncertain. Mertz that she had a recurrence. She continued to have et at found that 33% of source contacts were sexual intercourse with her regular boyfriend during asymptomatic at the time they infected their sexual the first two weeks of the study, and then their partners," so it seems likely that asymptomatic relationship ended. The lack of sexual contact in the shedding constitutes an important health problem. last two weeks did not affect the asymptomatic There is no evidence to support the assertion that the shedding of HSV. relative risk of infection from an asymptomatic source contact may be less than from a symptomatic source ASSOCIATION OF HSV SHEDDING WITH OTHER FACTORS and may be dependent on the recipients’ immune No appreciable association was found between the status.12 Furthermore, the effect of the site of shedding episodes of symptomatic or asymptomatic shedding of on transmissibility is unknown. HSV and the pattern of sexual intercourse, menstrual The value of the colposcope, both to detect cycle, stress events, or type of contraception used. microscopic genital lesions and to direct swabs to Furthermore, there was no association between virus lesions, is well shown in this study. In addition to the shedding and vaginal pH or the presence of leucocytes typical appearance of HSV infection of the cervix that detected by the examination of Gram stained urethral has been described previously,13 we noted that or cervical smears. There was, however, a strong ulceration was often preceded by increased local association between the presence of inguinal vascularity and vesicle formation. The use of lymphadenopathy and virus shedding. A score of 2 or colpophotography to record these lesions com­ more, which indicated the presence of bilateral or plemented the study records and helped to inform and tender unilateral inguinal lymphadenopathy (table educate the women. III) had a sensitivity of 85-7% and a specificity of The importance of our findings for trials of 92-4% for predicting whether specimens from the antiviral and other drugs intended to reduce the urogenital tract would be virus positive. Furthermore, frequency of HSV recurrences is important, as such the positive predictive value was 78-3% and the drugs might modify the disease and render negative predictive value was 95-4%. This association symptomatic recurrences asymptomatic or was valid for all anatomical sites sampled. All patients inapparent. It seems prudent for all future study had negative serological tests for syphilis, negative designs to include regular frequent (at least twice cultures for N gonorrhoeae, and normal cervical weekly) colposcopic examination, in conjunction with cytology within three months after the study. viral sampling from multiple sites, to exclude the possibility of asymptomatic shedding during and after Discussion the treatment period. The value of the study in providing data for The data on which to base the counselling of women counselling these patients is that certain symptoms with HSV infection about the problem of and signs emerged as being associated with virus asymptomatic shedding are inadequate. We set out to positive episodes in these patients. Thus we could tell obtain information about the patients in our study individual patients that mild dysuria, vaginal group to make individual counselling possible. discharge, and inguinal lymphadenopathy were table in Association between inguinal lymphadenopathy associated and with their recurrences, though the patients isolation o f herpes simplex virus (HSV) from the urogenital had not observed any ulceration. They were therefore tract advised not to have sexual intercourse while these symptoms were present. This was important advice No o f patients whose culture result was: because several of our patients believed that a sore was Lymphadenopathy the only symptom of HSV infection. This advice will score* Positive Negative be necessary until an easily performed reliable test to < 2 3 62 detect the presence of HSV within hours, rather than >2 18 5 days, becomes available. We are currently evaluating • 0 = no lymphadenopathy; 1 = unilateral non-tender the use of a rapid direct monoclonal antibody test to lymphadenopathy, 2 = bilateral non-tender lymphadenopathy or detect asymptomatic shedding of HSV in patients with tender unilateral lymphadenopathy; 3 = bilateral tender lymphadenopathy. recurrent genital HSV infection. - 1 9 3 -

Screening to detect asymptomatic shedding of HSV in women with recurrent infection 185 Whether or not all anxious patients who worry promote carcinogenesis. As sperm may act as such a about asymptomatic shedding can have a programme stimulus,16 it is evident that women who have of frequent regular examinations and virological asymptomatic recurrent cervical HSV infection are at investigation depends on the resources of the most risk because they will be having intercourse while individual clinic and the motivation of the patient. the cervix is infected. Thus barrier contraception, as Though this may be time consuming, the reassurance well as reducing the risk of transmission of HSV, to the patients who completed the study without any might also reduce the woman’s risk of developing recurrence was valuable (even if of uncertain cervical neoplasia. significance), and all patients benefited from knowing how their symptoms related to episodes of viral shedding. Indeed, to counsel patients with HSV References infection properly, larger studies of this type must be performed. Only in this way will a method for predicting which patients become asymptomatic 1. Nahmias AJ, Josey WE, Naib ZM, Freeman MG, Fernandez RJ, Wheeler JH. Perinatal risk associated with maternal genital shedders, deciding what factors influence shedding, herpes simplex virus infection. Am J Obstet Gynecol and assessing the effects of any proposed treatment be 197l;110;825-36. developed. We do not know if one episode of 2- Bolognese RJ, Corson SC, Fuccillo DA, Traub R, Moder R, Sever JL. Herpesvirus hominis type II infection in asymptomatic shedding means that others will asymptomatic pregnant women. Obstet Gynecol 1976;48:507- necessarily occur, or if their incidence relates to the 10. incidence of symptomatic recurrences. 3. Vontver LA, Hickok DE, Brown Z, Reid L,Corey L. Recurrent genital herpes simplex virus infection in pregnancy: infant Non-lesional prodromes have been described by outcome and frequency of asymptomatic recurrences. Am J Sacks in a study that relied totally on patient Obstet Gynecol 1982;143:75-84. 4. Guinan ME, MacCalman J.Kem ER, Overall JC.SpruanceSL. observations of whether a prodrome was followed by The course of untreated recurrent genital herpes simplex genital ulceration .14 One patient (case 5) in our study infection in 27 women. N Engl J Med 1981;304:759-63. had a prodrome that was followed by an episode of 5. Adam E, Kaufman RH, Kirkovic RR, Meinick JL. Persistence of virus shedding in asymptomatic women after recovery from asymptomatic vulval shedding of HSV from a herpes genitalis. Obstet Gynecol 1979;54:171-3. microscopic vulval ulcer that could be seen only with 6. Ferrer RM, Kraiselburd EN, Kouri YH. Inapparent genital the colposcope. It is thus essential for any herpes simplex infection in women attending a venereal disease clinic. Sex Transm Dis 1984;11:91-3. investigators intending to comment on the incidence 7. McCaughtry ML, Fleagle GS, Docherty J J. Inapparent genital of non-lesional prodromes to make regular herpes simplex virus infection in college women. J Med Virol 1982;10:283-90. examinations, and if possible use colposcopy, to 8. Rattray MC, Corey L, Reeves WC, Vontver LA, Holmes KK. ascertain whether lesions really do occur. Recurrent genital herpes among women: symptomatic v. One patient (case 8) was apparently infectious on asymptomatic viral shedding. British Journal of Venereal Diseases 1978;54:262-5. nine of her 11 clinic visits, the virus being cultured 9. Barton SE, Munday PE, Davis JM, Tyms AS. Restriction from a variety of urogenital sites. Our options for enzyme analysis and herpes simplex infections. Lancet management were either barrier contraception, 1985;i:1448-9. 10. Corey L, Adams HG, Brown ZA, Holmes KK. Genital herpes prophylactic acyclovir, or counselling and careful simplex virus infections: clinical manifestations, course and follow up. The patient opted for continued follow up, complications. Ann Intern Med 1983;98:958-972. as she had ended the relationship with her regular 11. Mertz GJ, Schmidt O, Jourden JL, et at. Frequency of acquisition of first-episode genital infection with herpes boyfriend and had no current sexual partner. She simplex virus from symptomatic and asymptomatic source stopped shedding virus two weeks after the study contacts. Sex Transm Dis 1985;12:33-9. 12. Jaratt M. Herpessimplexinfection..4rcAZ)er/7jaf0/1983;119:99- period. Subsequent viral cultures at visits twice weekly 103. for two months have not yielded a positive culture. We 13. Coppleson M, Pixley E, Reid B. Colposcopy: a scientific and intend to continue to see her as often as she agrees and practical approach to the cervix and vagina inheatth and disease. Illinois: Thomas, 1978. to manage her according to her needs. 14. Sacks SL. Frequency and duration of patient-observed The effect of this patient’s frequent asymptomatic recurrent genital herpes simplex virus infection: characterisation of the nonlesional prodrome. J Infect Dis shedding on her risk of developing cervical 1984;150:873-7. intraepithelial neoplasia is not-known. Though HSV 15. Fish EN; Tobin SM, Cooter NBE, Papsin FR. Update on the infection of the cervix has been shown to increase the relation of Herpesvirus hominis type 2 to carcinoma of the cervix. Obstet Gynecol 1982;59:220-4. susceptibility of cervical cells to undergo atypical 16. Coppleson M. The origin and nature of premalignant lesions of transformation,15 additional stimuli are needed to the cervix uteri. Int J Gynaecol Obstet 1970;8:539-41. -194-

Genitourin Med 1987;63:102-5

Asymptomatic shedding and subsequent transmission of genital herpes simplex virus

S E BARTON * J M DAVIS, t V W MOSS * A S TYMS,f AND P E MUNDAY* From the *Praed Street Clinic and the f Division o f Virology, St Mary’s Hospital and Medical School, London

SUMMARY We report the transmission of genital herpes simplex virus (HSV) infection from an asymptomatic woman shedding virus from the cervix to two male sexual partners and further transmission from these two men while their infection was in the prodromal phase. The value of the restriction enzyme analysis of viral deoxyribonucleic acid (D NA ) is presented. Guidelines regarding the management of patients who are found to be asymptomatic shedders of HSV are discussed.

Introduction HSV infection. On examination of her cervix, erythema and a mucopurulent discharge were visible, Epidemiological studies of the transmission of genital and colposcopic examination showed ulceration on the herpes simplex virus (HSV) infection have been made ectocervix. Routine tests for gonorrhoea, candidiasis, more accurate by characterisation of viral trichomoniasis, and chlamydial infection gave deoxyribonucleic acid (DNA) by restriction negative results, but HSV was isolated from the cervix. endonuclease analysis .1*3 We report the use of this Complement fixing,antibody to HSV was detected in method to study the transmission of genital HSV infec­ her serum, with no signficant change in titre three tion by an asymptomatic woman shedding virus from weeks later, which suggested that this was a recurrent the cervix and to investigate viral isolates from subse­ episode. quent recurrences of HSV infection. The next day the first patient’s brother (case B) attended complaining of penile ulceration of three Case report days’ duration, which had been preceded by local irritation, malaise, and headache for four days. HSV A man aged 26 (case A) presented to the Praed Street was isolated, and antibody studies confirmed the Clinic complaining of penile ulceration of 24 hours’ diagnosis of primary genital HSV infection. All three duration. He had had itching and redness of his penis patients gave negative serological test results for and the systemic symptoms of headache, fever, and syphilis and had no concomitant genital infection. As myalgia for five days. He had no history of any the table shows, patients A and B had had sexual inter­ sexually transmitted disease (STD) or of oral HSV course with other partners during the early phase of infection. A clinical diagnosis of genital herpes was their infection while prodromal systemic symptoms made, which was confirmed by isolating the virus. The and local redness and irritation were present, but absence of complement fixing antibody to HSV before the onset of blistering and ulceration. Figure 1 showed that this was a primary episode. At the first shows the subsequent history of HSV infection in their visit, he reported that his elder brother (case B) had sexual partners. similar symptoms and also gave a sexual history (table) that suggested a common source of their t a b l e History of sexual intercourse o f seven patients problems. We traced the female consort (case C), aged 33, Day Event who agreed to attend the clinic on the following day. She was symptomless and had no history of STD or of 0 Case B had sexual intercourse with case C 7 Case A had sexual intercourse with case C 12 Case A had sexual intercourse with case E Address for reprints: Dr S E Barton, Department of Gynaecology, Case B had sexual intercourse with case D Royal Northern Hospital, Holloway Road, London N7 13 Case A had sexual intercourse with cases F and G This paper was presented at the spring meeting of the MSSVD held in 14 Case A attended clinic Uppsala, Sweden, in May 1985. 15 Case C attended clinic Accepted for publication 28 June 1986. 16 Case B attended clinic

102 Asymptomatic shedding and subsequent transmission of genital HSV 103

Case E symptoms during an HSV recurrence in which lesions (GP diagnosed were not visible. In a previous study we had found that primary genital HSV2) asymptomatic shedding from the genital tract was associated with tender inguinal lymphadenopathy, CaseF vaginal discharge, and dysuria .7 Patient C was (STD clinic diagnosed encouraged to return for examination and to have tests primary genital HSV2) for the presence of HSV whenever she wished. In the CaseG six months after diagnosis she returned 12 times: on (STD clinic diagnosed four occasions she was found to have bacterial first episode genital HSV2) vaginosis, and on two candidiasis. HSV was not Case C (Asymptomatic isolated from her urethra, cervix, or vulva at any of cervical shedder) these visits. Advice was also given to the patient regarding the value of barrier contraception, as HSV Case B (Primary genital HSV2) cannot penetrate a sheath .8 Furthermore, the CaseD possibility of antiviral drug treatment was discussed, (GP diagnosed particularly the use of prophylactic continuous oral primary genital HSV2) acyclovir, which has been shown to prevent FIG i Diagram oftransmission of herpes simplex virus typesymptomatic recurrences .910 No data are available on 2 (HSV 2). (GP = general practitioner, STD = sexually the effect of this treatment on genital asymptomatic transmitted diseases.) shedding, but a study using prophylactic acyclovir in immunocompromised patients reported oral We examined the DNA of the HSV isolates from asymptomatic shedding during treatment .11 Patient C patients A, B, and C using restriction endonuclease opted to wait until we had investigated her pattern of analysis as described previously .4 The viruses were infectivity before any treatment was attempted. identified as being type 2, and the electrophoretic Secondly, the transmission of genital HSV infection profiles were identical when cleaved with the enzymes to four sexual contacts during the prodromal phase was Bam HI, Xho 1 (fig 2), and Kpn 1 (results not an important feature in the transmission of the disease. shown). Prevention of such transmission rests on improving The follow up of patients A and B provided interest­ health awareness and encouraging access to STD ing data. In the six months after diagnosis, A and B clinics, as well as on improving diagnostic each had 10 recurrences: four of these began within techniques. five days of each other, a further two within seven days, Thirdly, this case report offers an unusual oppor­ and three more within 10 days. So far, two HSV 2 tunity to investigate the strain epidemiology and isolates from recurrent episodes in patient B have been genome characteristics of a known virus during its examined by restriction enzyme analysis with Bam H1 transmission. Restriction enzyme analysis of viral (fig 2), Xho 1, and Kpn 1 (results not shown). DNA DNA distinguishes unequivocally between HSV types from both isolates was identical, in number and rela­ 1 and 2 and characterises viruses within each serotype tive position of restriction enzyme cleavage sites, to as being epidemiologically related or not by the DNA from the primary isolates of patients B and A presence or absence of restriction sites .12 The site of and the cervical isolate from patient C. We did not find latency in the nervous system has been thought to be DNA fragments differing in their electrophoretic occupiable by only the single virus that caused the mobility, which is a feature of recurrent isolates of primary infection .13 There is now good evidence that HSV.145 mixed populations of HSV showing strain differences (variations in the number of restriction sites) exist Discussion within ganglionic cells in the nervous system .14 Size heterogeneity of restriction fragments due to reitera­ This case report raises several important points tion of sequences in certain regions of the viral genome regarding genital HSV infection. Firstly, the value of without a change in the number or relative position of contact tracing patient C and detecting the presence of restriction sites characterises HSV subpopulations in asymptomatic cervical viral shedding was consider­ cell culture.15 These minor differences can distinguish able. Mertz et al have argued that the identification between HSV type l 16 or type 2 45 isolates from and screening of such source contacts may reduce the successive episodal attacks. Together these findings transmission of the infection .6 Having been traced and provide evidence for virus heterogeneity in the latent identified as the source of the infection, however, this form of HSV. Reactivation of subpopulations of low patient needed counselling on how to prevent further virulence may interrupt the transmission of more transmission of the disease. Our management began pathogenic forms of HSV. Alternatively, virus popula­ with an attempt to improve the awareness of her tions in the latent form, which are homogeneous, might -196-

104 S E Barton, J M Davis, V W Moss, A S Tyms, and P E Munday HG 52 A] Bi Cl B2 B3 HSV1 At Bi Ci HSV1

i------1 i------1 Bam 1 Xho1 FIG 2 Electrophoretic profiles of DNA from type2 reference virus (HG52), isolate from asymptomatic shedder (case C), primary isolates (A and B) from two brothers (cases A andB), and recurrent isolates (B2 and B3 ) from case B after digestion with Bam HI; and profiles of DNA from viruses A, B, and Cj and herpes simplex virus1 (H type SV 1)after digestion with Xho 1. be expected to maintain their inherent characteristics temporal relation of the onset of recrudescence in the during reactivation and transmission. brothers would be unlikely if only exogenous factors The herpes simplex virus strain studied in this case triggered recurrences. Each patient has a different report retained total identity by restriction enzyme occupation and lifestyle, and therefore endogenous analysis after transmission to two people, which characteristics of the virus probably exerted an impor­ resulted in concurrent infections, and during two tant effect on the timing of recurrences. A more episodes of reactivation in one patient. The efficiency frequent rate of recurrence for HSV type 2 is well of transmission of this virus seems to be high. The close documented.17 If the characterisation of HSV by Asymptomatic shedding and subsequent transmission of genital HSV 105 analysis of viral DNA, viral proteins ,13 or other 5. Schmidt OW, Fife KH, Corey L. Reinfection is an uncommon features of virus growth eventually provide markers of occurrence in patients with symptomatic recurrent genital herpes . J Infect Dis 1984;149:645-6. virulence, such information will be useful both for 6. Mertz GJ, Schmidt O, Jourden JL, et at. Frequency of acquisi­ counselling patients about prognosis and for planning tion of first episode genital infection with herpes simplex virus management. from symptomatic and asymptomatic source contacts. Sex In conclusion we have shown that, though early trac­ Transm Dis 1985;12:33-9. 7. Barton SE, Wright JK, Link CM, Munday PE. Screening to ing of source contacts of patients with genital HSV is of detect asymptomatic shedding of herpes simplex virus in women value, it does not prevent the transmission of HSV with recurrent genital herpes. Genitourin Med 1986;62:181-5. infection in the early phase of the disease. To help 8. Con ant MA, Spicer DW, Smith CD. Herpes simplex virus reduce the spread of this disease we need improved transmission: condom studies. Sex Transm Dis 1984;1:94-5. 9. Mindel A, Faherty A, Hindley D, et al. Prophylactic oral health education, faster diagnostic techniques, and acyclovir in recurrent genital herpes. Lancet 1984;ii:57-9. further research into the patterns of asymptomatic 10. Strauss SE,TakiffHE,SeidlinM,e/ al. Suppression of frequently shedding of HSV and the relation of DNA restriction recurring genital herpes. N Engl J Med 1984;310:1545-50. endonuclease profiles to recurrent disease patterns. 11. Shepp DH, Dandliker PS, Flournoy N, Meyers JD. Once-daily intravenous acyclovir for prophylaxis of herpes simplex virus reactivation after marrow transplantation. J Antimicrob We thank the Joint Standing Research Committee at St Chemother 1985;16:389-95. Mary’s Hospital, London W2, for its support. We also thank 12. Roizman B, Toigman M. Restriction endonuclease patterns of Helen Brown for her help in preparing the manuscript. The herpes simplex virus DNA: application to diagnosis and molecular weights for the HG52 DNA fragments, and the molecular epidemiology. In: Bachman PE, ed. New develop­ virus, were kindly supplied by Dr Andrew Davison, Institute ments in diagnositc virology. Berlin: Springer-Verlag, 1983 of Virology, Glasgow. (Cure Top Microbiol Immunol 104):273-86. 13. Lonsdale DM, Brown SM, Subak-Sharpe JH, Warren KG, Kowprowski H. The polypeptide and the DNA restriction enzyme profiles of spontaneous isolates of herpes simplex virus References type 1 from explants of human trigeminal, superi or cervical and vagus ganglia. J Gen Virol 1979;43:151-71. 1. Maitland NJ, Smith IW, Peutherer JF, Robertson DH, Jones 14. Lewis ME, Leung W-C, Jeffrey VM, Warren KG. Detection of KW. Restriction endonuclease analysis of DNA from genital multiple strains of latent herpes simplex virus type 1 within isolates of herpes simplex virus type 2. Infect Immun individual human hosts. J Virol 1984;52:300-5. 1982;38:834-42. 15. Davison A, Wilkie NM. Nucleotide sequences of the joint 2. Roizman B, Tognon M. Restriction enzyme analysis of herpes­ between L and S segment of herpes simplex virus type 1 and 2.J virus DNA; stability of restriction endonuclease patterns. Gen Virol 1981;55:315-31. Lancet 1982; ii:677. 16. Toigman M, Bartoletti A, Constanzo F, Mannis-Palenzoma A, 3. Barton SE, Munday PE, Davis JM, Tyms AS. Restriction Cassal E. Bam H 1 characterization of 15 HSV-DNAs isolated enzyme analysis and herpes simplex infections. Lancet from recrudescent lesions of three individuals. Microbiologica 1985;i: 1448-9. 1984;7:85-90. 4. Davis JM, Jeffries DJ, Tyms AS, Walker D. Molecular biology 17. Reeves WC, Corey L, Adams HG, Vontver LA, Holmes KK. in viral diagnosis: restriction enzyme analysis of viruses from Risk of recurrence after first episode of genital herpes. N E nglJ recurrent genital herpes infections. Analyst 1985;110:605-9. Med 1981;305:315-9. -198-

Female prostitutes and sexually transmitted diseases

S E Barton, Whittington Hospital, London D Taylor-Robinson, MRC Clinical Research Centre, Middlesex J R W Harris, St Mary’s Hospital, London ______• ______

The epidemiology of HTV infection emphasizes the need to be able to identify and communicate with female prostitutes as a group. Health education, screening and the prevention of STD in female prostitutes remains an essential priority for the well-being of the prostitutes, their clients and the health of society. However, these goals will not be achieved by compulsory screening, as this will miss the highest risk group. Instead, increased official recognition and measures to encourage female prostitutes to participate in and communicate with the health care system may result in an improvement in our management of these women. Although female prostitution has certainly existed since ancient the women to first seek and then accept treatment, thus the legal - times (Bess and Janus, 1976), it has functioned more as a service and social attitudes towards prostitution effectively determine industry than the oldest profession. As such, sexually transmit­ the health care delivery as well as the facilities for research. ted infections constitute the prostitutes’ equivalent of industrial Although since 1916 in the UK there has been a free and : diseases. The high incidence of sexually transmitted disease confidential service for the treatment of venereal diseases, no " (STD) and its sequelae in female prostitutes first attracted special provisions have been made for female prostitutes, E organized medical interest in the UK during the early nine­ Indeed, the recent WHO guidelines for the control of STD teenth century. The Victorian era saw the Contagious Diseases (WHO, 1985) specifies female prostitutes among the priority Acts (1864, 1866 and 1869) passed to attempt to reduce the groups to which screening and control activities should be prevalence of syphilis among military personnel by introducing directed. However, in the UK, the only current official statistics the compulsory medical inspection of female prostitutes. The on female prostitution are of convictions for soliciting, which public wrath which this legislation aroused (Tannahill, 1981) are dependent on the vigour of the police to arrest and the resulted in its suspension in 1883. This paved the way for the courts to convict. Although such figures are increasing, they fail development of the system, based on a belief in individual to provide an accurate assessment of the number of women - morality and improved access to medical care, which has existed engaged in prostitution, and no information whatsoever of their j until the present day. The failure of this nineteenth century medical or social problems. The recent interest in the role of the 1 legislation has recently been used (Porter, 1986) to argue against female prostitute as a possible vector for the transmission of 5 the use of legal sanctions to attempt to control the transmission HIV infection (Gillie, 1987) has highlighted this problem, but, :t of human immunodeficiency virus (HIV). However, more as yet, not led to any action. important than the political historical precedent is that this attempt to enforce medical inspection demonstrated a serious Previous studies of STD flaw in its implementation, because the women most easily in female prostitutes identified as prostitutes (i.e. those working in brothels or The results of studies of the prevalence of STD in female ; already voluntarily registered) had the lowest incidence of prostitutes show marked geographical variation, depending on venereal disease. This had previously been shown in Paris by the reliability of the diagnostic tests used, the efficacy of the ? Parent-Duchatelet (1857), in the first study to demonstrate the treatments given, the number of unprotected sexual exposures, * role of the prostitute as a reservoir for STD. This showed that, the prevalence of STD in their clients, and the transmissibility although 26% of prostitutes had been compulsorily treated for of the pathogens to which they are exposed (Darrow, 1984). syphilis, the comparative case rates were 1.9 per 100 public The role of the female prostitute as an important vector for prostitutes, 3.7 per 100 brothel inmates, 1.1 per 100 freelance the transmission of STD was unquestioned until 1958, when prostitutes and 22.3 per 100 unregistered prostitutes. Rosenthal and Vandow concluded that although prostitutes The standard of the medical care of female prostitutes is accounted for 10% of the STD among military personnel in dependent on both the resources available and the inclination of New York, over 80% was acquired from casual sexual contact with “promiscuous amateurs”. Similar conclusions were drawn Dr Barton is Research Fellow in the Department of Gynaecology, Whittington by Willcox (1962), who noted that less than 20% of the STD in ; hospital, London; Dr Taylor-Robinson is Consultant Microbiologist in the Division of Sexually Transmitted Diseases, MRC Clinical Research Centre; the UK and USA was acquired from prostitutes, whereas 30% and Dr Harris is Consultant Physician at The Praed Street Clinic, St Alary’s in France and over 80% in Asia was derived from this source. Hospital, London Furthermore, he noted that the recipients of the STD from : 34 British Journal of Hospital Medians, July 1937 -199- prostitutes in all these countries tended to be travellers, migrant he was conducting. In addition, he collected interesting data on workers, and military personnel, rather than the indigenous the lack of contraceptive use, the high incidence of unwanted population. Turner and Morton (1976) studied female prosti-. pregnancy and the social problems faced by female prostitutes tutes in Sheffield, noting that 60 women accounted for one sixth in Sydney. The limitation of this study was the lack of easily of all cases of locally acquired gonorrhoea in men, this figure transportable diagnostic facilities to the massage parlour, he-ice increasing during times of specific police operations to arrest the absence of tests for C. trachomatis and HSV infections. A prostitutes. They concluded that female prostitutes constituted study by Nayyar et al (1986) in New York City and Rotterdam a special group needing separate contact tracing and health extended the range of sexually transmitted infections that were education services to reduce their role as a reservoir for sought. The results showed that although the incidence of Neisseria gonorrhoeae infections. This view was shared by gonorrhoea infection had not declined since 1958, genital Potterat et al (1979), who showed that 27% of urethral gonorr­ C. trachomatis infections were numerically a more serious hoea in men in Colorado Springs, USA, was acquired from just problem. Furthermore, the high incidence of cervical isolates of 56 female prostitutes. In Central Africa, over 50% of the 86 Mycoplasma hominis and Ureaplasma urealyticum in female female prostitutes studied had cervical gonorrhoea, with 90% of prostitutes corroborated the association of these organisms with these being asymptomatic (Meheus et al, 1974). These findings a high frequency of sexual partners (McCormack et al, 1983). It support the epidemiological model proposed for the transmis­ was also noted that the incidence of trichomoniasis among these sion of gonorrhoea (Morton, 1977) based on a core of highly women was low, and this was attributed to the habit of sexually active individuals acting as both the reservoir and douching with antiseptics between intercourse with clients. primary vectors for infection. The arguments that the identifi­ Although this study widened the spectrum of infections tested cation and treatment of this group would significantly reduce for, it was confined to women in brothels, and follow-up the incidence of gonorrhoea in the overall population have gone examinations were not undertaken. unheeded in the UK. As more accurate diagnostic techniques for STD have been Praed Street Clinic study developed, several studies of the prevalence of conditions other Although many of the aforementioned studies have yielded than gonorrhoea and syphilis have been performed on female interesting data on various groups of female prostitutes, there prostitutes. In a study in Iran, where prostitutes were compul­ has been a conspicuous lack of any detailed data on a group of sorily registered and medically examined, Darougar et al (1983) prostitutes over a period of time. In an attempt to combine found that 94% of the women tested had serological evidence of screening for a wide range of sexually transmitted diseases with infection with Chlamydia trachomatis. In 1972, Duenas et al the longitudinal documentation of the behaviour of female recorded that the prevalence of herpes simplex virus (HSV) prostitutes, a study at the Praed Street Clinic (PSC), London, infection in female prostitutes was twice that in women attend­ was performed (Barton et al, 1986). Fifty women were exam­ ing an STD clinic. Moreover, they noted a higher incidence of ined fortnightly for up to one year. The inherent bias of recurrent genital infection in the former group. recruitment from patients attending this STD clinic was With increasing evidence that sexual transmission of the weighted against the value of long-term follow-up. The sub­ human papillomavirus (HPV) is the most important aetiological groups of prostitutes studied ( Table 1) were representative of the factor in the development of cervical neoplasia (Singer et al, different types of prostitute in London (Paterson-Brown and 1976; Campion et al, 1985), the previous reports of an increased Finnerty, 1986) categorized by their site of work. incidence of cervical carcinoma in female prostitutes (Peryra, 1961; Moghissi et al, 1968) become more explicable. Following the recognition that hepatitis B may be acquired by sexual transmission (Hersh et al, 1971), several studies have Mean age______29.4 years (range 19-48 years) _____ noted an increased prevalence of this disease among female Mean duration of prostitution 4.3 years (range 2 months to 22 years) prostitutes. Frosner et al (1975) found a significant threefold Mean number of dients/week______20 (range 7-200)______increase in the prevalence of hepatitis B virus surface antibody Street 14 in prostitutes compared with that in age-matched controls in Flat 10 West Germany. Moreover, there was a strong positive associ­ . Brothel 4 Club 4 ation between the duration of prostitution, the frequency of Escort 18 other STD and the acquisition of hepatitis B virus infection. A similar study (Papaevengolou et al, 1979) of Greek female prostitutes showed that this group had a 2.5-fold increased risk During the period of the study, 28% of the women had at of acquiring hepatitis B virus infection. However, in neither least one episode of urogenital infection by N. gonorrhoeae. study was any distinction made between prostitutes who did There was no significant difference in the incidence of gonorr­ and did not have a history of parenteral drug abuse. hoea in those prostitues who claimed to insist on condom use by Donovan (1984) surmounted the problem of waiting for self- all their clients. As condoms are impermeable to bacteria recruitment to his project by persuading the manager of a (Felman, 1979), and have been shown to reduce the frequency massage parlour in Sydney, Australia, ^to allow regular on-site of transmission of gonorrhoea (Barlow, 1977), this finding examinations of the women in his establishment. Screening 70 needs explanation. Initial claims of 100% condom use are rarely women for syphilis, gonorrhoea, and trichomoniasis revealed borne out by prospective questioning (Ward et al, 1987). Indeed that new gonococcal infections occurred in 10% of them each half of the women subsequently revealed that they omitted week, but the women affected were not predictable on the basis condom use for “regular clients”. Moreover, few of their of clinical symptoms or signs. Donovan argues that as the regular non-paying consorts used condoms, and the high rate of incidence of gonorrhoea among new prostitutes was 44%, the asymptomatic gonococcal infection (Donovan, 1984) in these figure of 10% among women already in his study demonstrates men may also play an important role in the apparent lack of the beneficial effects of the screening and treatment programme protection from condom use. 36 British Journal of Hospital Medicine, July 1987 -200- C. trachomatis was detected in 46% of the women during the HIV infection does not mean that prostitutes without a history study period. The anatomical site of these isolates was endocer- of intravenous drug abuse are not at high risk of acquiring HIV vical In 20 cases, urethral in four cases, rectal in three cases, and infection in the future. In particular, the epidemic curve used to pharyngeal in one case. All three rectal isolates were from estimate the future number of AIDS cases (McEvoy and Tillett, women who engaged in anal intercourse, two of whom insisted 1985) may not yet have reached the gradient of increase on condom use. However, any possible protection by condom necessary to result in sufficient transmission to new cases to use may have been reduced by the higher incidence of condom allow detection except by mass screening. For this reason, all rupture during anal rather than vaginal intercourse (Barton et female prostitutes have continued to be classified as a high-risk al, 1985). category for the acquisition of HIV infection (Pinching and The incidence of overt genital warts was 6%, but 10% of the Jeffries, 1985; MinkofF and Schwarz, 1986). women had cytological evidence of cervical intraepithelial neoplasia. Moreover, four cases of asymptomatic cervical shed­ Psychosexual characteristics and problems ding of HSV type 2, six cases of pediculosis pubis and one case in female prostitutes of secondary syphilis (Fig. 1) in a street-based prostitute were James and Meyerding (1977) found evidence that street-based diagnosed. Three women had serology consistent with a history female prostitutes enter this occupation because of negative of treated syphilis, and two women had evidence of previous sexual experiences during their childhood. Moreover, they hepatitis B infection. noted a lack of parental guidance leading to early, casual sexual intercourse. There was also a high incidence of incest and rape Fig. 1. during adolescence, and an early formation of the belief that sex Cutaneous could be used to improve social status. The early coitarche of lesions of secondary prostitutes was also noted by Duenas et al (1972) who recorded syphilis in a it as 15.3 years compared with 18.8 years for controls. There is London street strong evidence that childhood prostitution itself is becoming prostitute. increasingly common in most large cities in Europe and America (Sereny, 1984). Although female prostitutes have multiple sexual partners, this does not exclude them from experiencing problems during intercourse with their non-client partners. Some evidence of this was gained in the PSC study, 80% of the women regularly experiencing deep dyspareunia and 50% needing to use vaginal lubricants. However, 50% of the women had experienced orgasm during sexual intercourse with a client, and 12% of these had only experienced orgasm with clients, despite having regular non-paying consorts. Both sexual and psychological problems were more common in the women who worked from the street than in escort agency-based prostitutes. These find­ ings are consistent with the proposal first put forward by Exner et al (1977) that most of the psychopathology associated with prostitution is concentrated in street-based workers. In addition to the problems of intravenous drug abuse (14% of the PSC sample), female prostitutes in the PSC study had a The incidence of gonorrhoea was significantly higher in the high incidence of alcoholism and more than 90% of them street-based women than in escort agency-based prostitutes. smoked more than 10 cigarettes per day. This illustrates that This may reflect the larger number of clients having intercourse these women are in a group at high risk for diseases other than with the former (Table 2). Other STDs showed a trend to being just STDs. more frequent in the street-based prostitutes, but none reached statistical significance. ______Type ot Intercourse______HIV infection Clients/week Vaginal Oral/vag Oral/vag/anal Vag/anal Street 37 3 . 7 2 2 A major feature of the PSC study was an attempt to investigate Fiat 26 ~2 5 3 0 the prevalence of HIV infection in London prostitutes, as Brothel 23 0 2 2 0 studies elsewhere (Table 3) had shown widely varying levels of Ciub 7 3 ' T 0 0 seropositivity. During the study period, one woman was found Escort 8 1 16 1 0 to be infected by HIV; she was asymptomatic but had used intravenous heroin with shared needles and had a history of hepatitis B. Other studies (Tirelli et al, 1985; Brenjty-Fadeux and Fribourg-Blanc, 1985) . have also demonstrated that, in Europe, female prostitutes who are not intravenous drug users Rwanda_____ 33______88______Van de Perre et al, 1985 Seattle _____ 92 5 ______CPC, 1985 ______are not yet at an increased risk of having acquired HIV Florida______25 ______40______CPC, 1985 ______infection. The conclusion that the major source of HIV infec­ Athens______200______6 ______Papaevengolou et al. 1985 tion in an epidemic among heterosexuals will be from infected Paris 56 0 Brenky-Fadeux and ______Fribourg-Blanc. 1985 ____ intravenous drug users has led to a call for preventive measures, Pordenone 14 (IVPA) 71 Tirelli et a!, 1985 . such as free needles and syringes, and HIV antibody testing of ______10 (non-)VDA) P______addicts (Moss, 1987). However, the apparent absence of current IVPA ■ intravenous drug abuser 40 British Journal of Hospital Medicine, July 1987 - 2 0 1 - The clients (Turner and Morton, 1976). Over 90% of the prostitutes in the A rare study of the clients of prostitutes was performed in 1960 PSC study favoured the introduction of compulsory medical by Gibbens and Silberman. They found that only15 % of their examinations for all prostitutes, and 80% felt that it should be sample ever used the same prostitute more than once, and that made an offence for a prostitute to work while she knew she had in making their choice on the street, clients showed a preference an infectious STD. Furthermore, although 68 % of the PSC for girls who had a “motherly” appearance. Most importantly study supported the formation of licensed brothels, only 38% they noted the great difficulty in obtaining accurate histories said they would work in them. It is essential that if the WHO from male clients. This problem accounts for the continued (1985) recommendation of directing health screening and edu­ paucity of data on the characteristics and rates of STD in the cation specifically towards female prostitutes is to be imple­ clients of female prostitutes. Indeed, Donovan (1984) had to mented, a system of licensing or registering will have to be abandon screening the male clients of his study group for introduced, in association with the decriminalization of prosti­ similar practical reasons, although he did detect a reservoir of tution. Until that time, medical contact with prostitutes should asymptomatic urethral gonorrhoea in this group. This finding, be fostered by contact between public health planners and the in association with the omission of condom use for regular representative groups formed by prostitutes themselves. clients, and the concurrent use of several prostitutes by these “regular clients” may make these men an important vector in Specific clinic service for female prostitutes the transmission of STD among prostitutes. The increasing availability of quicker diagnostic tests for STDs is of special benefit to prostitutes. The delay in awaiting a confirmatory culture to exclude an STD or confirm the efficacy Problems in the management of the diseases of treatment is a major problem for prostitutes. Standard advice of female prostitutes from physicians to avoid sexual intercourse until a result is Identification of female prostitutes known means lost earnings and often goes unheeded. The To improve the delivery of health care to female prostitutes it is introduction of direct monoclonal antibody kits for the diag­ essential to be able to identify the women themselves. In the nosis of infections by N. gonorrhoeae, C. trachomatis and HSV UK, this depends on women either attending STD clinics and gives the facility for the “results while you wait” approach to be declaring their occupation or being arrested by the police and made more accurate and extensive. Rapid results and appropri­ being submitted to a compulsory medical examination. How­ ate treatment in this special target group will reduce the ever, many female prostitutes are seen in STD clinics without problems of poor compliance. Furthermore, the increased stating their true occupation, hence appropriate diagnostic tests availability to gynaecologists of quicker diagnostic tests should and health counselling may be omitted. Moreover, although allow more comprehensive screening of female prostitutes for Conrad et al (1981) have demonstrated that screening convicted STD when they attend with other problems, such as an prostitutes can effectively identify untreated cases of STD unwanted pregnancy, abnormal cervical smear or chronic pelvic assiduous pursuance of prostitutes by the police has been shown pain. to result in an increase in the cases of STD in their clients Health education for female prostitutes Although further studies may reveal valuable data on the prevalence of STD in female prostitutes, the existing infor­ mation overwhelmingly demonstrates that the female prostitute is a majgr reservoir and vector for genital infections. Moreover, there is evidence that both hepatitis B and HIV infections do infect female prostitutes by sexual transmission, and thus may be transmitted to their clients. As prostitution cannot be eradicated, appropriate health education must be directed at these women to attempt to reduce their chances of contracting new STD. Such advice should include recommending the use of condoms by all their clients, as both HSV and HIV have been shown to be unable to penetrate sheaths (Conant et al, 1984, 1986). Furthermore, the associated use of spermicides has been found to hasten the inactivation of HIV (Hicks et al, 1985). In addition to reducing the risk of acquiring a new STD, this may also result in a reduction in the number of unwanted pregnan­ cies in these women, although additional hormonal contracep­ tive measures are advised. Indeed, calls have been made (Wellings, 1986) in the UK for the government to provide free condoms, a measure which may encourage prostitutes, among others, to use them. Furthermore, the development of stronger condoms for use by both homosexuals and heterosexuals will lead to a reduction in the risk of condom rupture.

Health education of clients Identification of female prostitute. "Over 90% of the prostitutes Public health measures should be directed at the clients of in the P SC study favoured the introduction of compulsory medical examinations for all prostitutes, and 80% felt that it prostitutes to increase their appreciation of the protection should be made an offence for a prostitute to work while she afforded by condom use and the need for regular medical knew she had an infectious STD.” examinations to detea asymptomatic infections. In the PSC 44 British Journal of Hospital Medicine, July 1987 -202- study, 80% of prostitutes felt that it should be illegal for a client Conant, M A,' Spicer, D W, Smith, C D (1984) Herpes simplex vims transmission: condom studies. Sexually Transmitted Diseases, 11, 94-95 to have sexual intercourse with them if he knew that he had an Conant, M A, Hardy, D, Sematinger, J, Spicer, D, Levy, J A (1986) Condoms infectious STD. Although such a view may seem unrealistic, it prevent transmission of AIDS associated retrovirus. Journal of the American Medical Association, 255, 1706 emphasizes the feeling among prostitutes in this study that Conrad, G L, Kleris, G S, Rush, B, Darrow, W W (1981) Sexually transmitted “something must be done” to protect them from HIV infection. disease among prostitutes and other sexual offenders. Sexually Transmitted Diseases, 8, 241-244 Darougar, S, Aramesh, B, Gibson, J A, Treharne, J D, Jones, B R (1983) Chlamydial genital infection in prostitutes in Iran. British Journal of Venereal Diseases, 59, 53-55 Darrow, W (1984) Prostitution and sexually transmitted diseases. In Sexually Transmitted Diseases (edited by Holmes, K K, Mardh, P A, Sparling, P F, Weisner, P J). McGraw Hill, New York. pp. 109-116 Donovan, B (1984) Gonorrhoea in a Sydney house of prostitution. Medical Journal of Australia, 140, 268-271 Duenas, A, Adam, E, Melnick, J L, Rawls, W E (1972) Herpes virus type 2 in a prostitute population. American Journal of Epidemiology, 95, 483-489 Exner, J E, Wylie, J, Leura, A, Parrill, T (1977) Some psychological characteristics of prostitutes. Journal of Personality Assessment, 41, 474-485 Felman, Y (1979) A plea for the condom especially for teenagers. Journal of the American Medical Association, 241, 2717-2718 Frosner, G G, Bucholz, H M, Gerth, H J (1975) Prevalence of hepatitis B antibody in prostitutes. American Journal of Epidemiology, 102, 241-250 Gillie, O (1987) Sexual behaviour; an issue for the ‘80s. The Independent. 12 Jan 1987 Gibbens, T C N, Silberman, M (1960) The clients and prostitutes. British Journal of Venereal Diseases, 36, 113-115 Hersh, T, Melnick, J L, Ray, D, Goyal, R K (1971) Non-parenteral transmis­ sion of viral hepatitis type B. New England Journal of Medicine, 288, 1363- 1364 Hicks, D R, Martin, L S, Getchell, J P et al (1985) Inactivation of HTLVIII/ LAV infected cultures of normal human lymphocytes by nonoxynol 9 in vitro. Lancet, ii, 1422-1423 James, J, Meyerding, J (1977) Early sexual experience and prostitution. / American Journal of Psychiatry, 134, 1381-1385 McCormack, W M, Moller, B R, Mardh, P A (1983) Mycoplasma hominis: a human pathogen. Sexually Transmitted Diseases, 10, 160-5 McEvoy, M, Tillett, H E (1985) Some problems in the prediction of future numbers of cases of the acquired immunodeficiency syndrome in the UK. Lancet, ii, 541-542 Meheus, A, De Clerq, A, Prat, R (1974) Prevalence of gonorrhoea in prostitutes in a Central African town. British Journal of Venereal Diseases, 50, 50-53 Minkoff, H L, Schwarz, R H (editorial) (1986) AIDS; Time for obstetricians to get involved. Obstetrics and Gynecology, 68, 267-268 Moghissi, K S, Mack, H C, Porzak, J P (1968) Epidemiology of cervical cancer. American Journal of Obstetrics and Gynecology, 100, 607-614 “Public health measures should be directed at the clients of Morton, R S (1977) Gonorrhoea. W B Saunders, London, p.220 prostitutes to increase their appreciation of the protection afford­ Moss A R (1987) AIDS and intravenous drug use; the real heterosexual epidemic. British Medical Journal, 294, 389-390 ed by condom use and the need for regular medical examina­ Nayyar, K C, Cummings, M, Weber, J ct al (1986) Prevalence of genital tions to detect asymptomatic infections.” pathogens among female prostitutes in New York City and in Rotterdam. Sexually Transmitted Diseases, 13, 105-107 Official recognition of the problems of prostitutes, Papaevengolou, G, Trichopolous, D, Kremastinou, T, Papoutsakis, G (1974) Prevalence of hepatitis B antigen and antibody in prostitutes. British Medical as well as of prostitution Journal, ii, 256-258 All measures to improve the medical management of female Papaevengolou, G, Roumeliotou-Karayannis, A, Kallinikos, G, Papoutsakis, G (1985) LAV/HTLVIII infection in female prostitutes (letter). Lancet, ii, 1018 prostitutes have the goal of reducing the prevalence of STD and Parent-Duchatelet, A J B (1857) De la prostitution de la ville de Paris. Baillifere, associated problems in both them and the whole population. New York Paterson-Brown, S, Finnerty, S P (1986) Problems of London prostitutes. Any attempt to achieve this goal will depend on the recognition British Journal of Sexual Medicine, 13, 260-261 that female prostitutes constitute a specific risk group needing Peryra, A J (1961) The relationship of sexual activity to cervical cancer. Cancer of the cervix in a prison population. Obstetrics and Gynecology, 17, 154-159 special resource allocation, health education and medical and Pinching A J, Jeffries, D J (1985) AIDS and HTLVIII/LAV infection: social research. Several studies have demonstrated the willing­ consequences for obstetrics and perinatal medicine. British Journal of Obstetrics and Gynaecology, 92, 1211-1217 ness of female prostitutes to accept regular screening for STD Porter. R (1986) History says no to the policeman’s response to AIDS. British and also their desire for more comprehensive health education. Medical Journal, 293, 1589-90 Potterat, J J, Rothenburg, R, Bross, D C (1979) Gonorrhoea in street Despite this and the current media recognition that HIV prostitutes: epidemiologic and legal complications. Sexually Transmitted infection may be acquired from sexual contact with female Diseases, 6, 58-63 Rosenthal, T, Vandow, J (1958) Prevalence of venereal disease in prostitutes. prostitutes (Wilsher and Hodgkinson, 1986), history shows that British Journal of Venereal Diseases, 34, 94-99 it is unlikely that there will be a major change in the legal status Sereny, G (1984) The Invisible Children. Andre Deutsch, London Singer, A, Reid, B L, Coppleson, M (1976) The role of the high risk male in the of these women in the UK. aetiology of cervical cancer — a correlation of epidemiology and molecular Fig. 1 is reproduced by kind permission of the A VC Department, St Mary’sbiology. American Journal of Obstetrics and Gynecology, 126, 110-116 Hospital Medical School, London W2. Tannahill, R (1981) Sex in History. Abacus, London Tirelli, U, Vaccher, E, Carbone, A, De Paoli, P, Santini, G, Monfardini, S Barlow, D (1977) The condom and gonorrhoea. Lancet, ii, 811-813 (1985) HTLV-III antibody in prostitutes. Lancet, ii, 1424 Barton, S E, Underhill, G S, Gilchrist, C, Jeffries, D J, Harris, J R W (1985) Turner, E B, Morton, R S (1976) Prostitution in Sheffield. British Journal of HTLV-III antibody in prostitutes. Lancet, ii, 1424 • Venereal Diseases, 52, 197-205 Barton, S E, Link, C M, Underhill, G S, Munday, P E, Taylor-Robinson, D, Van de Perre, P, Carael, M, Robert-Guroff, M et al (1985) Female prostitutes: a Harris, J R W (1986) A Longitudinal study of prostitutes in London. risk group for infection with human T-cell lymphotropic virus type III. Presented to Second World Congress on STD. Paris, France Lancet, ii, 524-526 Bess, B E, Janus, S S (1976) Prostitution. In The Sexual Experience (edited by Ward, H, Day, S, Barton, S E, Harris, J R W (1987) Widespread use of Sadock, B J et al). Williams and Wilkins, Baltimore condoms and low prevalence of sexually transmitted diseases in Danish non­ Brenky-Fadeux, D, Fribourg-Blanc, A (1985) HTLV-III antibody in prosti­ drug addicts (letter). British Medical Journal, 294, 444 tutes. Lancet, ii, 1424 Wellings, K (1986) AIDS and the condom. British Medical Journal, 293, 1259- Campion, M J, Singer, A, Clarkson, P K, McCance, D J (1985) Increased risk 1260 of cervical neoplasia in the consorts of men with penile condyloma acumin­ Willcox, R R (1962) Prostitution and venereal disease. British Journal of ata. Lancet, i, 943-945 Venereal Diseases, 38, 37-42 CDC (1985) Heterosexual transmission of human T-cell lymphotropic virus Wilsher, P, Hodgkinson, N (1986) At risk. The Sunday Times. Nov 2 1986 type III / lymphadenopathy-associated virus. Center for Disease Control World Health Organization (1985) Control of Sexually Transmitted Diseases. Morbidity and Mortality Weekly Report, 34, (37), 561-563 WHO, Geneva British Journal of Hospital Medicine, July 1987 45