Savioli, Francesca (2024) Furthering our understanding of tumour budding in oestrogen receptor negative breast cancer. MD thesis, University of Glasgow.
Full text available as:
PDF
Download (11MB) |
Abstract
Every year, 55,900 people are diagnosed with breast cancer in the UK alone(1). Treatment modalities for this heterogeneous disease should be individualised and tailored to each case according to an increasing number of molecular, clinical, and personal factors. Molecular subtypes have been classified into Luminal A/B, HER-2 enriched, and basal-like cancers, which overlap significantly with Triple Negative Breast Cancers (TNBC). These cancers are difficult to treat have a poorer prognostic profile(3-5).
Tumour budding, most extensively described in colorectal cancer, represents an isolated group of up to four cancer cells found at the invasive tumour front, separate from the tumour mass(6). Tumour budding in high levels is associated is a marker of poor prognosis in numerous cancers including breast(6-13).
Tumour budding denotes one of the fundamental behaviours of solid tumours in the transition towards metastasis. This finding is felt to be a histopathological depiction of a process belonging under the umbrella term of epithelialmesenchymal transition (EMT) in which tumour cells adopt new characteristics to facilitate the migration from the solid tumour mass and tumour microenvironment to allow seeding to distant sites(14). Tumour budding has been associated with reduced survival in intraductal breast carcinomas when a value of 20 tumour buds/0.385mm2 was used as a threshold to denote “low tumour budding” versus “high tumour budding” phenotype (10). Nonetheless, different thresholds and methods of counting tumour buds have been documented in the literature as having prognostic power, and therefore standardisation has yet to be reached for tumour budding as a prognostic marker in breast cancer(11-13).
A cohort of ER-negative surplus specimens from patients who underwent surgery for primary operable breast cancer was examined and tumour buds scored. A threshold of 28 was identified to have prognostic significance, with “high tumour budding” phenotype associated with reduced cancer-specific survival. Based on this population, 50 patient specimens (25 high tumour budding, 25 low tumour budding) were analysed using TempoSEQ for to evaluate RNA expression and evaluate differential expression. Four highly differential expressed genes were identified as JUNB, ODAM, RFX5 and TBX22.
This thesis aimed to examine JUNB, ODAM, RFX5 and TBX22 histologically, using immunohistochemistry to validate these findings at the protein expression level.
The hypothesis was that differential gene expression at the transcriptional level should be reflected in protein expression in the high vs low tumour budding specimens, and that this would be reflected in a prognostic effect on cancer specific survival.
Protein expression for JUNB, ODAM, RFX5 and TBX22 was assessed using immunohistochemistry in the Glasgow Breast Cancer Cohort (n=850), composed of surplus specimens of breast cancers from patients operated for primary operable breast cancers in Glasgow between 1995-1998. The cohort was validated using full section specimens, where protein expression was compared between tumour mass and tumour buds to confirm that expression was comparable before tissue microarray (TMA) cores were utilised for higher throughput examination of the entire cohort. Only ductal cancers were included in the final analysis (n=736).
Comparisons between quantitative expression of JUNB, ODAM, RFX5 and TBX22 to clinicopathological factors and outcome were made, and molecular subgroups examined in turn to establish whether transcriptional differences in gene expression was comparable to protein expression at the tumour level. Quantitative analysis was manually made using weighted histoscores, and thresholds created by using R to establish prognostic group thresholds for each protein expression in the different cellular compartments: membrane, cytoplasm, and nucleus.
High cytoplasmic JUNB expression was associated with increased cancer-specific survival (CSS) in TNBC and ER-negative cancer. High nuclear expression of JUNB was associated with improved CSS across all patients within this cohort and in the ER+ patients in particular. Membrane expression amongst this cohort was low and could not be analysed. Combined nuclear and cytoplasmic scores suggested that when comparing the high nuclear, high cytoplasmic versus high nuclear, low cytoplasmic JUNB phenotypes, the latter group had poorer CSS.
High cytoplasmic ODAM expression was associated with worse CSS for HER-2 enriched patients in this cohort. Low nuclear ODAM expression was associated with worse CSS across the entire cohort. Combined scoring suggested that low nuclear and high cytoplasmic (combined) expression of ODAM was associated with reduced survival compared to all high expression. Membrane ODAM was not assessed due to low scores throughout the cohort.
Across the whole cohort, low RFX5 cytoplasmic expression was shown to be associated with reduced CSS. Membrane RFX5 WHS did not predict survival across the cohort, but correlated with grade, molecular subtype and Klintrup Makinen score. No effect was observed between nuclear RFX5 expression and survival or clinicopathological factors. Combined membrane, cytoplasmic and nuclear scores allowed the observation that an “all low” phenotype was associated with worse survival, particularly in the Luminal A and HER2-enriched subgroups.
Cytoplasmic TBX22 expression was not associated with prognosis but was found to be particularly high in the HER-2 enriched cohort. Nuclear TBX22 also did not have a prognostic effect on this breast cancer cohort. Membrane scoring was once again low across the whole cohort.
Subsequently, a fresh cohort of TNBCs (n=207) from patients who underwent surgical resection for primary operable invasive cancer at two Glasgow hospitals from 1995-2010 were evaluated. A selection of 50 specimens from this cohort (25 high tumour budding, 25 low tumour budding) from this cohort was examined using GeoMx Digital Spacer Profiling. The GeoMX DSP technology allowed the spatial analysis and exploration of RNA transcription within different portions of the tumour and tumour microenvironment. By examining the differential gene expression in “high tumour budding” versus “low tumour budding” phenotype TNBC specimens, several differentially expressed genes were identified. Amongst the differentially expressed genes identified using GeoMX DSP technology, HIF1alpha, Bcl-2 and CD3 expression was noted to be significantly different between high and low tumour budding TNBC specimens. HIF-1alpha was most differentially expressed in tumour cells in low budding group, together with CD4 expression. When examining the tumour microenvironment (i.e., stroma rich portions of the specimen), Bcl-2 was amongst the most differentially expressed genes. HIF-1alpha, associated with hypoxia, led to the additional examination of CAIX, another well-documented marker of tissue hypoxia, to be examined using immunohistochemistry, and CD4, which had a more stable antibody available for immunohistochemistry, was used as a proxy for CD3, to validate these findings at the protein expression level. Thresholds based on literature search and individual prognostic power were utilised for each protein weighted histoscore. This study suggested that for HIF-1alpha, recurrence-free survival (RFS) was worse in high cytoplasmic HIF-1alpha expressing tumours, and disease-free survival (DFS) was worse in high nuclear HIF-1alpha expressing tumours. High CAIX cytoplasmic expression was associated with poorer RFS, while low Bcl-2 was associated with poorer RFS and DFS. Low stromal and tumoural CD3 expression were each associated with worse RFS, and low combined stromal and tumoral CD3 expression was associated with worse RFS.
In summary, the result from this thesis suggests that both at the transcriptional level and protein expression level, tumour budding may affect prognosis, and that this effect may be more pronounced in certain subgroups, including TNBC. Identification of differential gene and protein expression may therefore allow further targets for therapy to be identified, and subsequently become an additional element in the arsenal of prognostication and treatment of breast cancer.
Item Type: | Thesis (MD) |
---|---|
Qualification Level: | Doctoral |
Subjects: | R Medicine > R Medicine (General) R Medicine > RC Internal medicine > RC0254 Neoplasms. Tumors. Oncology (including Cancer) |
Colleges/Schools: | College of Medical Veterinary and Life Sciences > School of Cancer Sciences |
Supervisor's Name: | Edwards, Professor Joanne and Romics, Dr Laszlo |
Date of Award: | 2024 |
Depositing User: | Theses Team |
Unique ID: | glathesis:2024-84428 |
Copyright: | Copyright of this thesis is held by the author. |
Date Deposited: | 02 Jul 2024 14:15 |
Last Modified: | 05 Aug 2024 10:24 |
Thesis DOI: | 10.5525/gla.thesis.84428 |
URI: | https://theses.gla.ac.uk/id/eprint/84428 |
Related URLs: |
Actions (login required)
View Item |
Downloads
Downloads per month over past year