Abstract
Chemotherapy-induced peripheral neuropathy (CIPN) is a common and debilitating side effect of cancer treatment, significantly impacting patient quality of life (QoL). Despite its high prevalence, affecting up to 90% of patients, the underlying biological mechanisms remain poorly understood, particularly regarding the heterogeneity of onset, severity, and duration among individuals receiving similar treatments. Genetic predisposition has been proposed as a key contributor; however, previous studies focused primarily on common variants through genome-wide association (GWA) and candidate gene approaches, yielding inconsistent and non-reproducible findings. Emerging evidence has suggested that rare, predicted high-impact coding variants may have a more substantial role in disease susceptibility.This thesis investigated the contribution of rare coding variants to CIPN using whole-exome sequencing (WES) data from 38 female breast cancer patients aged 18+, who underwent anti-neoplastic chemotherapy treatment. Patients were stratified based on patient-reported outcomes (PRO-CTCAE) into those having moderate-severe CIPN; cases (n=32) and those with no reported CIPN symptoms (controls, n=6). For comparison, two high-throughput sequencing platforms (BGI and Ion Torrent AmpliSeq) were utilised, with platform-specific bioinformatic pipelines implemented for variant detection and annotation.
Initially, 21 previously implicated genes were examined in the study cohort, with four rare, predicted high-impact variants (CYP2C8 c.792C>G; SLCO1B1c.263G>T; ABCB1c.1435T>A, and EPHA8 c.893A>G) identified in a subset of individuals. Notably, three of the four genes (CYP2C8, SLCO1B1, ABCB1) have been established to be associated with the paclitaxel absorption, distribution, metabolism, and excretion (ADME) pathway, suggesting a potential role for pharmacokinetic variability in CIPN susceptibility. However, limited recurrence across the cases indicated that no single variant was determinative.
This analysis was then extended to investigate pathway-level convergence of rare variants unique to the Moderate-Severe cases within the cohort. Over-representation analysis (ORA) was performed through PANTHER gene ontology (GO) and identified enrichment in two key biological systems: epithelial structural organisation (hemidesmosome assembly cytoskeletal pathways) and mitochondrial metabolism (β-oxidation). In silico analysis of publicly available gene expression data (GTEx) was used to guide candidate gene prioritisation, followed by variant detection, which identified five rare, predicted high-impact variants (KRT14: rs58378809, KRT5: rs778602599; PLEC:rs782134950, ACADS: rs374726386, HADHA: rs370170143) integrated within these systems, however no recurrence patterns were observed. These findings suggested CIPN susceptibility may arise from coordinated disruption of biological systems rather than through isolated gene effects.
Collectively, these results, characterised by limited variant recurrence alongside convergence at the pathway level, support a multi-axis stress-threshold model in which CIPN emerged from the interaction of three vulnerability domains: systemic drug exposure, epithelial-neuronal interface stability, and metabolic resilience. Notably, epithelial/neuronal interface pathways emerged as a compelling and previously underexplored vulnerability domain for CIPN susceptibility. Within this framework, genetic variations may remain clinically silent until chemotherapeutic stress exceeds individual compensatory capacity, resulting in activation of this pathway and subsequent neuropathy. As such, this model may provide insight into the heterogeneity and limited reproducibility observed in previous studies. However, due to the modest sample size, these findings should be interpreted as preliminary, and future functional studies across larger cohorts will be required to validate the proposed mechanisms.
An improved understanding and examination of CIPN as a latent vulnerability disorder, likely originating from multiple physiological bases may therefore inform the development of integrated risk prediction strategies. Such approaches may enable improved patient stratification, personalised treatment planning, and ultimately enhance the QoL outcomes for cancer survivors.
| Date of Award | 15 Oct 2026 |
|---|---|
| Original language | English |
| Supervisor | Paul Dunn (Supervisor), Joan Roehl (Supervisor), Haupt Larisa (Supervisor) & Kim Alexander (Supervisor) |
Cite this
- Standard