Objectif : Les traitements répétés des lésions précancéreuses du col utérin (HSIL), nécessaires en cas d’échecs de traitement, sont associés à des issues obstétriques négatives, telle qu’une augmentation de la mortalité néonatale. Nous avons investigué l’association entre un grand nombre de facteurs de risque potentiels pour l’échec de traitement des HSIL dans le but d’identifier des prédicteurs potentiellement modifiables de l’échec de traitement. Méthodes : La population source était constituée de 1 548 femmes canadiennes qui ont subi un premier traitement pour HSIL. L’échec de traitement a été défini comme étant un diagnostic histologique de HSIL ou cancer au cours des deux années suivant le traitement. Nous avons mené une étude cas-témoins nichée incluant les 101 cas d’échec de traitement ainsi que les témoins appariés 1 :1 par centre de traitement et par date d’échec. Nous avons calculé des rapports de cotes (OR) et intervalles de confiance (CI) à 95% à l’aide de régressions logistiques conditionnelles, pour les associations entre l’échec de traitement et l’âge, le nombre d’accouchements, le statut tabagique, le nombre de partenaires sexuels, l’utilisation du condom, la méthode de contraception, les marges, le nombre de passages, le diagnostic sur le spécimen de traitement, le génotype du VPH, et le nombre de types. Nous avons aussi estimé l’association entre la charge virale et les variants du VPH16 et du VPH18 et l’échec de traitement. Résultats : Les marges positives vs négatives (OR ajusté=4.05, 95% CI 1.57-10.48), la positivité pour le VPH16/18 vs autres types (OR ajusté=2.69, 95% CI 1.32-5.49), et avoir un variant similaire au prototype du VPH16 vs le prototype (OR ajusté=2.49, 95% CI 1.07-5.83) étaient des prédicteurs de l’échec de traitement des HSIL. Être plus âgé, avoir des lésions plus sévères, avoir une infection monotype, et avoir une variation à la position 7521 chez celles avec le VPH16 pourraient augmenter le risque d’échec de traitement, mais les associations n’étaient pas statistiquement significatives. Les estimations pour les autres facteurs étaient proches de la valeur nulle. Nous n’avons pas observé de modification d’effet du génotype sur le risque de l’échec de traitement par le tabagisme, ni par les marges. Conclusion : Seules les marges positives, la positivité pour le VPH16/18 et avoir un variant similaire au prototype étaient des prédicteurs d’un échec de traitement au cours des deux années suivant le traitement. Malgré l’aspect non-modifiable des prédicteurs identifiés, ils sont informatifs et pourront éclairer la prise en charge et le suivi clinique. / Objective: Repeated treatments for high-grade squamous intraepithelial lesions (HSIL), which are necessary in the case of treatment failure, are associated with negative obstetric outcomes, such as an increased risk of neonatal death. We investigated the association between a large number of potential risk factors and HSIL treatment failure in an effort to identify potentially modifiable predictors of treatment failure. Methods: The source population included 1,548 Canadian women who received a first treatment for HSIL. Treatment failure was defined as the histological diagnosis of HSIL or cancer within the two years following treatment. We conducted a nested case-control study that included all 101 cases of treatment failure and controls that were matched 1:1 on treatment center and date of failure. We used conditional logistic regression to calculate the odds ratios (OR) and 95% confidence intervals (CI) between treatment failure and age, parity, smoking status, number of sexual partners, condom use, method of contraception, margins, number of passes, diagnosis on the treatment specimen, HPV genotype and number of types. We also estimated the association between HPV16 and HPV18 viral loads and variants and HSIL treatment failure. Results: Having positive vs. negative margins (adjusted OR=4.05, 95% CI 1.57-10.48), being positive for HPV16 and/or HPV18 vs. any other type (adjusted OR=2.69, 95% CI 1.32-5.49), and having a prototype-like variant of HPV16 vs. the prototype (adjusted OR=2.49, 95% CI 1.07-5.83) were predictors of HSIL treatment failure. Older age, more severe lesions, single-type infections and a variation at the 7521 position of the HPV16 genetic sequence may lead to a higher risk of treatment failure but were not statistically significant. Estimates for all other factors were near the null value. The effect of genotype on the risk of treatment failure was not modified by smoking status, nor by margin status. Conclusion: Only positive margins, HPV16/18 positivity, and having a prototype-like variant of HPV16 were predictors for HSIL treatment failure within two years of treatment. Despite being non-modifiable, the identified predictors are clinically significant in regards to management and follow-up of patients.

At present, prostate cancer is a significant public health issue throughout the world and is the second leading cause of cancer deaths in older men. The prostate specific antigen or PSA (which is encoded by the kallikrein 3/KLK3 gene) test is the current most valuable tool for the diagnosis and management of prostate cancer. However, it is insufficiently sensitive and specific for early diagnosis, for staging of prostate cancer or for discriminating between benign prostatic hyperplasia (BPH) and prostate cancer. Recent research has revealed another potential tumour marker, glandular kallikrein 2 (KLK2 gene/hK2 protein), which may be used alone or in conjunction with PSA to overcome some of the limitations of the PSA test. Twelve new kallikrein gene family members have been recently identified and, like hK2 and PSA, many of these genes have been suggested to be involved in carcinogenesis. In this study, the cellular localisation and level of expression of several of these newer kallikreins (KLK4, KLK5, KLK7, KLK8 and KLK11) was examined in prostate tissue, to provide an understanding of the association of their expression with prostatic diseases and their potential as additional biomarkers. Like PSA and hK2, the present observation indicated that each of these proteins, hK4, hK5, hK7, hK8 and hK11, was detected within the cytoplasm of the secretory cells of the prostate glands. For the first time, all of these newly-identified proteins were shown to be expressed in prostatic intraepithelial neoplasia (PIN) lesions, in comparison to normal glands and cancer lesions. In addition to cytoplasmic secretory cell expression, the localisation of hK4 to the basal cells and nuclei in prostatic lesions was intriguing. The intensity of hK4 staining in prostate tissue was strongest in comparison to the other newly-identified kallikrein proteins (hK5, hK7, hK8 and hK11). Therefore, KLK4/hK4 expression was characterised further to define this cellular localisation and examined in non-prostatic tissue and also in a larger number of prostate tissues in an attempt to determine its potential value as a biomarker for prostate disease. Three hK4 antipeptide polyclonal antibodies, derived against N-terminal, mid-region and C-terminal hK4 amino acid sequences, were used. The hK4 N-terminal antipeptide antibody was used to demonstrate the cellular localisation of hK4 in kidney, salivary glands, liver, testis, colon carcinoma, heart, endometrium and ovarian cancer, for the first time. The presence of hK4 in these non-prostate tissues was consistent with the previous reports using RT-PCR. The dual cytoplasmic and nuclear localisation of hK4 observed in the prostate above was also seen in these tissues. Although hK4 was found widely expressed in many human tissue types, indicating that it is not prostate specific in its expression, the highest expression level of hK4 was seen in the prostate. Therefore, detailed expression patterns and levels of KLK4 mRNA and hK4 protein in the normal prostate and prostatic diseases and histopathological lesions were investigated and reported for the first time in this study. Twelve benign prostatic hyperplasia (BPH), 19 adenocarcinoma (Gleason grade 2-5) and 34 bone metastases from prostate cancer were analysed. Using in situ hybridisation, the expression of KLK4 mRNA was detected in the cytoplasm of the secretory cells of both normal and diseased prostate tissue. KLK4 mRNA was also noted in both secretory and basal cells of PIN lesions, but the basal cells of normal glands were negative. Using the hK4 N-terminal and mid-region antipeptide antibodies, hK4 was predominantly localised in the cytoplasm of the secretory cells. The intensity of hK4 staining appeared lowest in normal and BPH, and increased in PIN lesions, high Gleason grade prostate cancer and bone metastases indicating the potential of hK4 as a histopathological marker for prostatic neoplasias. Further studies are required with a larger cohort to determine its utility as a clinical biomarker. Small foci of atypical cells, which were found within normal glands, were also intensely stained. Surprisingly, hK4 protein was found in the nucleus of the secretory cells (but not the basal cells) of high grade PIN and Gleason grade 3 prostate cancer. The detection of KLK4 mRNA and hK4 protein in PIN lesions and small foci of atypical cells suggests that up-regulation of KLK4 expression occurs early in the pathology of prostate carcinogenesis. The finding of basal cell expression is not typical for the kallikreins and it is not clear what role hK4 would play in this cell type. With the use of the hK4 C-terminal antipeptide antibody, the staining was mainly localised in the nuclei of the secretory cells of the prostate glands. Although the nuclear localisation was readily noted in more than 90% of epithelial cells of the prostate gland with the C-terminal antibody, no difference in staining intensity was observed among the histopathological lesions of the prostate. The prominent nuclear localisation with the C-terminal antipeptide antibody was also shown to be distributed throughout the nucleus by using confocal microscopy. Further, by using gold-labelled particles for electron microscopy, the intracellular localisation of these hK4 antipeptide antibodies was reported here for the first time. Similar to the immunohistochemical results, the cytoplasm was the major site of localisation with the N-terminal and mid-region antipeptide antibodies. To further characterise the involvement of KLK4/hK4 in human prostate cancer progression, the transgenic adenocarcinoma mouse prostate (TRAMP) model was used in this study. In this study, mouse KLK4 (also known as enamel matrix serine protease -1, EMSP-1) was shown to be expressed in the TRAMP prostate for the first time. Previous studies had only shown the developing tooth as a site of expression for EMSP-1. The level of EMSP-1 mRNA expression was increased in PIN and prostate cancer lesions of the TRAMP model, while negative or low levels of EMSP-1 mRNA were seen in normal glands or in control mouse prostate tissue. The normal mouse prostate did not stain with any the three hK4 antipeptide antibodies. hK4 N-terminal and mid-region antipeptide antibodies showed positive staining in the cytoplasm of the epithelial cells of PIN and cancer lesions of the mouse prostate. The C-terminal antipeptide antibody showed distinctively nuclear staining and was predominantly localised in the nuclei of the glandular cells of PIN and cancer lesions of the mouse prostate. The expression patterns of both the mRNA and protein level for mouse KLK4 strongly supported the observations of KLK4/hK4 expression in the human prostate and further support the utility of the TRAMP model. Overall, the findings in this thesis indicate a clear association of KLK4/hK4 expression with prostate cancer progression. In addition, several intriguing findings were made in terms of cellular localisation (basal as well as secretory cells; nuclear and cytoplasmic) and high expression in atypical glandular cells and PIN, perhaps indicating an early involvement in prostate disease progression and, additionally, utility as basal cell and PIN histological markers. These findings provide the basis for future studies to confirm the utility of hK4 as a biomarker for prostate cancer progression and identify functional roles in the different cellular compartments.