Clinical and Genetic Aspects of Prolactin Hypersecretion

Abstract

Hyperprolactinaemia is the commonest pituitary endocrinopathy. The degree of prolactin elevation is integral to patient assessment, necessitating vigilance in serum prolactin measurement. Treatment with dopamine agonists is usually highly effective; however, some patients experience intolerable side effects or fail to achieve normoprolactinaemia and/or adequate prolactinoma shrinkage. The side effect profile has traditionally focused on gastrointestinal and cardiovascular symptoms, but mounting evidence suggests that hyperprolactinaemic patients are at risk of the same dopamine agonist-induced impulse control disorders frequently observed in the Parkinson's disease and restless legs syndrome populations. Extrapolating from limited data, the overall prevalence of prolactinoma patients with dopamine agonist treatment failure is at least 50 per million population. Additionally, prolactinomas are one of the commonest subtypes of aggressive pituitary tumours and pituitary carcinomas. Prolactinomas also feature prominently in the familial pituitary tumour syndromes, including the recently recognised '3P' association of pituitary adenomas, phaeochromocytomas and paragangliomas due to germline mutations in the succinate dehydrogenase genes. In contrast, the somatic mutational events underlying prolactinomas are unknown. By unclear mechanisms, prolactin hypersecretion may also coexist with Cushing's disease and carotid aneurysms, which may cloud patient assessment. This thesis evaluates pitfalls in the assessment and management of patients with prolactin excess due to prolactinomas or related disorders and the pathogenesis of prolactin excess. The key finding of the clinical section is that treating prolactinomas with dopamine agonists poses a high, previously underestimated risk of impulse control disorders. We documented this in a case series of eight men with prolactinomas and dopamine agonist-induced hypersexuality, which we referred to as 'dopa-testotoxicosis' to highlight the apparent additive effects of dopamine receptor stimulation and testosterone normalisation. We subsequently undertook the largest reported cross-sectional analysis of the risk of impulse control disorders in 113 hyperprolactinaemic patients vs. 99 controls. Our findings highlight the need for screening for dopamine agonist side effects, particularly those that may drive behavioural disturbances, as they may not be appreciated in routine practice. The other clinical studies of this thesis were also practice-changing. In a study of 58 cases, we showed that serum prolactin is overestimated by the Roche platform, with the potential for patient mismanagement as physiological and statistical prolactin variations may be misclassified as tumoural hyperprolactinaemia. A retrospective cohort study of 13 patients who underwent inferior petrosal sinus sampling in the evaluation of Cushing's syndrome revealed consistent co-lateralisation of prolactin and adrenocorticotrophic hormone and demonstrated how prolactin-corrected adrenocorticotrophic hormone concentrations may threaten test accuracy, arguing against routine prolactin measurement in petrosal sinus samples. The major molecular study of this thesis involved next generation sequencing {NGS} of paired germline and tumour DNA from 12 patients with sporadic prolactinomas. This was the first pangenomic study of a pure prolactinoma cohort investigating both point mutations and copy number variants. We found a high burden of copy number variation and a paucity of point mutations. In another NGS study of two families with familial paragangliomas and other tumours including a prolactinoma, we demonstrated a novel SDHC deep intronic mutation which is the first reported deep intronic mutation amongst the succinate dehydrogenase genes. We also employed NGS in the first molecular study of cyclical Cushing's disease, finding a putative novel role for the aryl hydrocarbon receptor {AHR} gene as a link between pituitary tumorigenesis and clock genes. Targeted RNA sequencing was employed in a study of two patients with the rare association of marked hyperprolactinaemia and carotid aneurysms. We introduced the term 'vasculogenic

hyperprolactinaemia' to describe this association and performed the first molecular study of the disorder; however, our investigation of candidate prolactin secretagogues was negative. Taken together, these studies have produced new knowledge in the important clinical field of prolactin hypersecretion; each of the studies either impacts upon diagnosis and/or therapy in this field, or points towards new strategies for further scientific study.