Sex hormones in gastric mucosal physiology and the development of oesophageal, gastric and colorectal adenocarcinoma

Abstract

The aim of this thesis was to continue to test the hypothesis that sex hormones play some part in the aetiology and seem to protect women from the development of oesophageal, gastric, and colorectal adenocarcinoma. All three of these cancers display an as yet unexplained male predominance in their respective worldwide incidence rates. The overall global male:female ratio in oesophageal adenocarcinoma (OAC) is just over 4:1, whereas the male:female incidence ratio for gastric adenocarcinoma (GAC) is somewhere around 2:1. Also, men have about a 1.5-fold higher risk of developing colorectal adenocarcinoma (CRA) compared to women.

In the first study, we evaluated the potential impact of sex hormones, using age as a proxy, on the incidence of oesophageal, gastric, and colorectal adenocarcinoma. Data from the Swedish Cancer Register yielded cases which were stratified into five-year age groups and employed to calculate incidence rates and sex ratios. The sex ratio of OAC declined steadily with age from around 10:1 to around 4:1. The sex ratio for non-cardia GAC increased with age to around 2:1, one to two decades following menopause, before levelling off and declining thereafter. The sex ratio for CRA was stable with age. The consistent decline of the sex ratio with age in OAC suggested a mechanism separate from menopause. The observed pattern in the sex ratio of non-cardia GAC indicated a protective effect in females prior to the onset of menopause. The lack of an age-dependent sex ratio in CRA suggested mechanisms other than menopause.

In the second study, we investigated potential sex differences in rat gastric mucosal defence, including sex differences and effects of oestrogen stimulation on gastric mucosal blood flow. These investigations were carried out in an in vivo and an ex vivo rat experimental model. Absolute blood flow in the organs of the gastrointestinal tract was generally higher in male rats, but only statistically significant in the corpus part of the stomach, where blood flow in females was about half that of males (males: 1.12 ± 0.12 mL/min/g; females: 0.51 ± 0.03 mL/min/g) (p=0.002). Compared to baseline, oestrogen administration (0.1 increasing to 1 μg/kg/min) was followed by a 31 % (68 ± 13 PFU) statistically significant decrease in mean gastric mucosal blood flow in male rats (p=0.02), whereas mean gastric mucosal blood flow remained virtually unchanged with a 4 % (5 ± 33 PFU) decrease in female rats. Gastric mucosal blood flow was higher in male than in female rats, contrary to our hypothesis, and was reduced in male rats by oestrogen stimulation. The first finding was unexpected but could, speculatively, imply that females have the ability to respond to a mucosal challenge with a greater increase in gastric mucosal blood flow from baseline than males. The second finding is in opposition to the generally accepted physiological effect of oestrogen as a vasodilator but could potentially be the result of a shift in the vasodilator-vasoconstrictor balance towards vasoconstriction caused by supra-physiological levels of oestrogen, leading to a reduction in blood flow towards the level seen in females. The lack of oestrogen effects in females might be due to higher baseline oestrogen levels, a different expression of oestrogen receptors as well as other factors, individually or in combination, leading to females being more resistant to exogenous oestrogen effects. However, these findings provide no explanation for the male predominance and no support for the oestrogen hypothesis in human gastric carcinogenesis.

In the third study, we assessed the risk of OAC, cardia GAC and non-cardia GAC as well as for oesophageal squamous-cell carcinoma (OSCC) (included only as a comparison outcome) in men treated for prostate cancer using androgen deprivation therapy (ADT). A nationwide population-based matched cohort study was conducted based on a national Swedish database of prostate cancer patients. Prostate cancer cases receiving ADT were the exposed group, whereas cancer-free men from the general population formed the unexposed control group. The risk among ADT-exposed compared to unexposed was calculated by multivariable Cox proportional hazard regression and expressed as hazard ratios (HRs) with 95 % confidence intervals (CIs). We found a significantly reduced risk of non-cardia GAC among ADT-users compared to non-users (HR 0.49 [95 % CI 0.24 – 0.98]). No such association was found for any of the other types of cancer explored. Our study thus indicated that ADT decreased the risk of non-cardia GAC, while no such risk reduction was found for OAC, cardia GAC or OSSC. Whereas the role of androgens in oesophageal carcinogenesis remain largely unknown, androgens seem to play an aggravating role in male gastric carcinogenesis.

In the fourth study, we assessed the risk of CRA in men treated for prostate cancer using androgen deprivation therapy (ADT). As for study III, we employed a nationwide population-based matched cohort study based on the same national Swedish database of prostate cancer patients, albeit a later version. Prostate cancer patients receiving ADT were the exposed group, whereas cancer-free men from the general population formed the unexposed control group. The risk of CRA among ADT-exposed compared to unexposed was calculated by a flexible parametric survival model and expressed as hazard ratios (HRs) with 95 % confidence intervals (CIs). We found an increased risk of CRA among ADT-users compared to non-users (HR 1.27 [95 % CI 1.15 – 1.41]), in particular an increased risk of adenocarcinoma of the colon (HR 1.33 [95 % CI 1.17 – 1.51]) and more specifically, we uncovered an increased risk of adenocarcinoma of the distal colon (HR 1.53 [95 % CI 1.26 – 1.85]). However, evaluation of latency effects showed significantly decreased HRs with time for CRA overall (p=0.049 for trend) and for rectal adenocarcinoma (p=0.034 for trend). We concluded that exposure to ADT was followed by an increased risk of CRA, specifically in the distal part of the colon. Although, these findings seemed to be in line with a protective effect of androgens in male colorectal carcinogenesis, the lack of a positive dose-response trend led us to question whether this was indeed a true causal effect.

In conclusion, this thesis supports the notion that sex hormones have some part to play in the aetiology and/or development of non-cardia GAC. It does not offer the same level of support regarding the association of sex hormones and CRA. Apart from establishing that the sex ratio in OAC incidence rates seem to be age-dependent, we found no further evidence to support the association between sex hormones and OAC or cardia GAC. The potential mechanisms of these associations remain unclear and further research is warranted.