Studies to characterise ovarian tumours in the mare

AL Ibrahim, Abdulqader Awadh (2019) Studies to characterise ovarian tumours in the mare. PhD thesis, University of Glasgow.

Abstract

Reason for performing this study: Granulosa cell tumours (GCT) are the most common type of ovarian tumours associated with mare infertility requiring highly invasive treatment, yet knowledge of GCT prevalence, pathogenesis and early diagnostic options is lacking. Text mining of UK veterinary practice data was used to enable the estimation of equine GCT (and ovarian pathology) prevalence in the UK. Gene sequencing studies were undertaken to investigate an important point mutation found in the coding region of the forkhead transcription factor gene FOXL2 (C402G) in adult human GCT biopsies. Antral follicle wave development in mares is highly regulated and leads to differing outcomes in follicles (differentiation versus atresia) with consequences for cellular morphology and function. As rapidly growing GCT must transform from normal ovarian follicles, specifically the granulosa cell, but also the theca compartment, a detailed histological and functional comparison of normal (control) follicle walls to diseased, solid or cystic GCT tissues was also carried out. While the influence of the time of year on follicular function has previously been studied in mares, the cellular effects of other factors such as disease or age of the mare are less clear, yet important for understanding and controlling mare reproduction. In addition, the mare is also considered an excellent model for ovarian function in women, thus our understanding of both regulated and dysregulated follicle growth may benefit from our studies using mare ovaries.
Objective: The general aims were to determine the prevalence of ovarian pathology and particularly GCT in UK equine practice; determine the effects of GCT formation in mare ovaries, which may dysregulate follicular granulosa (GC) or theca cell (THC) proliferation in very healthy (VH), healthy (H), early atretic (EA) and late atretic (LA) antral follicles using morphometric and functional histological approaches; determine whether the FOXL2 gene mutation identified in human GCT also exists in the equine GCT; and finally determine the FOXL2 protein localization in GCT samples compared with medium and large healthy and atretic antral follicles.
Materials and Methods: Data mining of clinical records (515,832 records) from seven veterinary practices from around the UK was used in this study.
Sixty-three follicles H&E sections were dissected from the normal ovaries of 19 mares classified as healthy (n= 35), or suffering from chronic clinical disease (n= 4) or severe systemic illness (n= 24) unrelated to the reproductive system. Follicle walls were processed for histological categorisation and (granulosa and theca layer, and a subset for granulosa cell nuclear circumference) measurements using H&E sections. Forty follicle walls recovered from 16 mares were functionally examined for hormone (inhibin), enzyme (aromatase, CYP17A), receptor (AMHR2) or transcription factor (FOXL2) expression parameters (stain intensity, distribution and % area covered) using immunohistochemistry. Similarly, ten GCT tissue samples from five mares (each mare provided two tissue section samples, one from cystic and one from solid areas of the GCT) were histologically categorised using H&E sections, followed by measurement of the nuclear circumference of granulosa cells within the GCT and the same immunohistochemical evaluations. Data were analysed using non-parametric statistical analyses, including correlation analyses between protein expression (% area covered) and functional follicle parameters (diameter and follicular fluid oestradiol concentration).
In addition, the FOXL2 gene was partially sequenced from a subset of GCT (from seven mares) and control follicle wall samples (from five mares) using in house molecular approaches followed by next generation sequencing (Illumina NextSeq 500 at Glasgow Polyomics), to analyse the region spanning the FOXL2 CG mutation discovered in human adult GCT.
Results: Data mining of clinical records from 26,019 mares revealed 115 potential GCT cases, but with only 11 cases being true GCT positive on further reading, resulting in a prevalence estimate of 0.04% (1: 2,365 mares); 143 potential ovarian pathology cases were found of which 65 cases were truly positive resulting in a prevalence estimate of 0.25% (1 in 400 mares).
Follicles measured on average 21 mm in diameter, which was not associated with mare disease status (p>0.05), and were allocated to very healthy (VH), healthy (H), early atretic (EA) and late atretic (LA) categories based on the histological appearance of the GC and THC layers. Mare disease status was not associated with the follicle health category (p>0.05), and was not associated with percentages of basal, intermediate or antral GC, or GC layer thickness (p>0.05). However, the THC layer thickness was greater (p=0.01) in follicles from mares with severe disease compared with follicles from healthy mares. More healthy than diseased mares were in seasonal transition at the time of ovary recovery (p= 0.001), and transitional follicles showed reduced THC layer thickness and percentages of basal GC and large THC, compared with follicles recovered from those in deep anoestrus (all p-values <0.05). Follicle categorization into VH, H or EA was correlated with follicular fluid oestradiol concentrations, but not associated with follicle diameter or any GC layer characteristics (all p-values >0.05); however, H follicles had a thicker THC layer than EA follicles (p=0.001). In LA follicles the GC layer was not present, and the percentage of large THC was reduced, while the percentage of small THC was increased compared to VH and H follicles (p<0.001), reducing the THC layer thickness in LA versus H follicles (p=0.001).
Inhibin expression was detected only in GC of normal follicular walls (FW) and GCT categories, not in THC. Similarly, aromatase expression was detected only in GC of normal FW; no aromatase stain was detected in any of the GCT samples and categories. Conversely, CYP17 expression was only detected in THC of normal FW and in interstitial TH like cells in diseased GCT samples. Both compartments of GC and THC in normal FW demonstrated the expression of AMHR2, which was also highly expressed in GC of cysts and solid areas in GCT samples, where it showed only little expression in the interstitial stromal tissue.
Within normal ovaries the nuclear circumference of GC nuclei decreased from basal (62.9±3.1) to intermediate (54.5±1.6) and antral cells (44.7±1.9) (p-0.001), with follicle atresia being associated with reduced basal GC nuclear size (p=0.04). In GCT1 and GCT2 cysts, basal GC nuclei (73.3±3.9) were also larger than antral (53.7±3.1) (p=0.001) and intermediate cell nuclei (63.9±4.7) (p=0.03), and antral GC nuclei of GCT1 and 2 (52.8±2.8) were larger than those in normal ovaries (44.7±1.9) (p=0.03). The nuclear circumference of solid GC nests (74.0±4.1) varied more between tumours and was larger than antral GC nuclei within cysts (53.7±3.1) (p=0.001).
The FOXL2 CG (C402G) human mutation was detected very rarely, and considered to be a result of sequencing error. However, a second, proximal CT mutation was identified in 25% of all reads (mean ± SEM: GCT 27.2± 4.3, controls 25.5± 3.0, P=0.4), even though this was not predicted to change the amino acid within the protein. Nuclear expression of FOXL2 was seen in GC and THC of all control FW, and expression declined with follicle atresia; FOXL2 was very highly expressed in GC of GCT cysts and solid areas, yet showed only a low level of expression in interstitial stromal tissue of GCT.
Conclusion: A very low prevalence of GCT in horses attending UK veterinary practices was identified. Differences in histomorphological and functional GC and THC measurements in equine large antral follicles appear more associated with season and follicle atresia than with disease. Location within normal antral follicle and tumour cyst walls, tumour transformation, and GCT compartment are associated with GC nuclear size, likely reflecting differing nuclear activity. While this study could not link the specific human FOXL2 SNP to equine GCT in the samples that were collected, a further frequent mutation in the small region amplified was identified. Therefore, further investigation of the complete equine FOXL2 gene may identify other mutations with functional significance in GCT formation in the horse and thus potentially also in the woman.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: mare ovarian tumours, granulosa cell tumour, GCT, ovarian pathology, mare follicle wall histology, follicle wall classification, follicle wall health status categorisation, follicular fluid oestradiol concentration, E2 ELISA, follicle wall immunohistochemistery, Inhibin, AMH, AMHR2, CYP17, Aromatase, FOXL2 protein, FOXL2 gene, mare FOXL2 gene, Text mining, content mining, Smistat/Wordstat, data mining.
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Biodiversity, One Health & Veterinary Medicine
• Supervisor's Name: Parkin, Professor Timothy and Mihm Carmichael, Dr Monika
• Date of Award: 2019
• Depositing User: Dr Abdulqader AL Ibrahim
• Unique ID: glathesis:2019-70981
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 13 May 2019 08:55
• Last Modified: 05 Mar 2020 17:34
• Thesis DOI: 10.5525/gla.thesis.70981
• URI: https://theses.gla.ac.uk/id/eprint/70981