NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Authors: Nadia Hitchen, Philippe L Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M Kwan, Ben Tran Published: 2026 ABSTRACT AbstractBackgroundPlatinum-based chemotherapy cures most patients with advanced testicular germ cell tumours (TGCTs), yet a subset develops resistance or relapse. The molecular basis of platinum sensitivity and resistance remains unclear, particularly in archival formalin-fixed paraffin-embedded (FFPE) samples that limit broad genomic analyses.ObjectivesTo evaluate whether differential expression of genes involved in apoptosis, DNA repair, and cell-cycle regulation distinguish platinum-resistant TGCTs.DesignRetrospective exploratory molecular profiling study using archival TGCT tissue specimens.MethodsWe profiled the expression of 30 genes related to apoptosis, DNA repair, and cell cycle regulation in 50 archival TGCT samples from 41 patients using the NanoString nCounter platform. Samples included platinum-sensitive, platinum-resistant, and post-chemotherapy teratomas. Platinum-resistant disease was defined as radiological progression and/or rising tumour markers after first line platinum-based chemotherapy. Platinum-sensitive cases included patients achieving complete response as well as those with residual teratoma requiring post-chemotherapy resection. Differential expression analysis and principal component analysis (PCA) were performed to compare transcriptional patterns across groups.ResultsPost-chemotherapy teratomas demonstrated a distinct expression profile, characterised by lower POU5F1 (encodes OCT4 protein) expression and higher levels of apoptosis and cell-cycle related genes. In contrast, platinum-resistant and platinum-sensitive primary TGCTs showed substantial overlap in global expression patterns, and no individual gene met thresholds for differential expression. These findings suggest that the transcriptional features of teratoma differentiation are readily detectable with targeted profiling, while platinum resistance is not captured by this gene set.ConclusionsNanoString-based targeted expression profiling identifies biologically distinct expression patterns in post-chemotherapy teratomas but did not reveal a unifying transcriptional signature associated with platinum resistance in TGCTs. These negative results underscore the multifactorial and heterogenous nature of platinum resistance and highlight the need for broader genomic, immune and epigenomic profiling to better elucidate mechanisms of platinum resistance.Keywords: testicular germ cell, testicular cancer, platinum-resistant, nanostring Introduction Advanced testicular germ cell tumours (TGCTs) are highly curable, with platinum-based chemotherapy achieving 10-year survival rates of approximately 95%.1–3 Despite this, approximately 10% of patients will demonstrate primary platinum resistance, or experience disease recurrence after completion of platinum therapy, necessitating intensive salvage therapies.4,5 Salvage therapies, such as high-dose chemotherapy and surgical intervention can be curative, although success rates vary by histological subtype. Teratomas, for instance, exhibit primary chemotherapy resistance, rendering surgical resection the only curative option for these patients. 6 Due to the substantial morbidity associated with salvage therapies and the reduced prognosis for patients progressing after platinum-based treatments, elucidating the mechanisms of platinum resistance in TGCTs is crucial for improving outcomes. Despite the high overall cure rates, this minority of non-responders remains a critical clinical challenge, highlighting the need to identify reliable molecular markers that could guide new therapeutic strategies. Several potential mechanisms of platinum resistance in TGCTs have been identified, including alterations in DNA damage repair genes, reduced sensitivity to apoptosis, and mutations in TP53.7–9 However, current high-throughput methods, ranging from targeted panel sequencing 8 to whole-exome sequencing8,10 and microRNA profiling 11 are often constrained by the quality and quantity of available samples. Patients with recurrence of TGCT frequently present with disease at challenging biopsy sites or exhibit biochemical-only recurrences, restricting sampling to previously collected formalin-fixed-paraffin-embedded (FFPE) specimens. These FFPE samples can suffer degradation and other processing artefacts, which may complicate comprehensive molecular analysis. We hypothesised that specific gene expression signatures distinguish platinum-resistant TGCTs (including teratomas, which are intrinsically chemotherapy-resistant) from platinum-sensitive tumours. To test this, we used the NanoString nCounter gene expression platform. 12 Unlike traditional RNA-based techniques that often require larger sample volumes and suffer from amplification bias, the NanoString nCounter system enables direct digital quantification of target mRNA from minimal input, making it ideally suited for probing critical pathways implicated in TGCT platinum resistance. We aimed to identify potential biomarkers and molecular pathways that may elucidate the mechanisms of chemotherapy resistance in TGCTs, ultimately informing new strategies to improve outcomes for this minority of patients whose TGCTs fail to respond to first-line platinum therapy. Methods Study population and sample collection Our study used data collected as part of the previously described institutional multidisciplinary TGCT database (eTestis) at the Princess Margaret Cancer Centre. 13 We identified all patients with platinum-resistant TGCT between January 1, 2000, and December 31, 2010, from the eTestis database, and included those with available archival specimens. Platinum-resistant disease was defined as radiological progression and/or rising tumour markers after first line platinum-based chemotherapy. Platinum-sensitive disease was defined as either complete response or residual teratoma treated with post-chemotherapy surgical resection. Eligible patients with platinum-resistant disease were then matched 1:1 (balancing for IGCCCG risk and seminoma or non-seminoma based on orchidectomy histology) with a comparator group of patients with platinum-sensitive disease to form the analytical cohort (Figure 1(a)). Study design and sample characterization of testicular germ cell tumours for gene expression profiling. (a) Flow diagram illustrating sample stratification. (b) Custom NanoString nCounter gene panel comprising 36 genes across seven functional pathways. (c) Distribution of primary histological subtypes across tissue classification groups. Annotations indicate specimen types within groups. The inset pie chart shows anatomical distribution of post-chemotherapy metastasis specimens. (d) Principal component analysis (PCA) of normalized gene expression data. Each point represents one sample; dashed ellipses indicate 80% confidence intervals for each group. For gene-expression analysis, specimens were grouped into four categories reflecting clinical response and specimen context: i) chemotherapy-sensitive TGCT, ii) chemotherapy-resistant TGCT, iii) post-chemotherapy teratoma, and iv) post-chemotherapy metastasis. Although post-chemotherapy teratoma arises in platinum-sensitive patients, it was analysed as a separate subgroup throughout given its distinct biology. To address inter-patient variability, we analysed a small cohort of patients with matched pre- and post-chemotherapy specimens, encompassing both teratoma and non-teratoma histologies. Given the rarity of platinum-resistant TGCT and the retrospective nature of archival tissue collection, formal prospective sample size calculations were not feasible. This exploratory study was therefore designed to identify potential transcriptional differences to inform potential future larger studies. All living patients had previously consented to tissue retrieval for exploratory molecular analysis, and a waiver of consent was granted to retrieve samples for deceased patients. The study was approved by the local institutional research ethics committee. The reporting of this study conforms to the REMARK statement 14 (Supplementary File 1). Sample processing and NanoString nCounter expression data generation Formalin-fixed paraffin-embedding (FFPE) tissue blocks were selected based on confirmed presence of TGCT. All pathology reports were reviewed by expert genitourinary pathologists, and samples were chosen to ensure that representative tumour histology for the study cohort was present. Tissue sections (5µm depth) were prepared for tumour profiling. Matched hematoxylin and eosin slides were subject to macrodissection guided by histologic review to ensure tumour cellularity of ≥80%. Total RNA was extracted from FFPE tissue sections according to institutional protocols. RNA quality and quantity were assessed prior to expression analysis, with samples meeting quality thresholds proceeding to NanoString profiling. Gene expression was quantified using the NanoString nCounter Analysis System with a custom CodeSet consisting of 36 genes: 30 target genes selected for their potential relevance to platinum resistance, and six housekeeping genes used for normalization (Figure 1(b)). The custom probe set included genes from pathways relevant to DNA damage repair (nucleotide excision repair, base excision repair, homologous recombination repair, mismatch repair), apoptosis/survival, stress response, and drug detoxification. Gene expression data processing and normalization Raw nCounter expression data were processed using the NanoTube package (version 1.12.0) in R. 15 Quality control metrics were assessed for all samples prior to normalization. Field of view (FOV) imaging efficiency was evaluated by comparing the number of FOVs successfully imaged (FOV Counted) to the total attempted, with samples requiring >75% successful imaging. Binding density, which measures the number of optical features per square micron and indicates reporter distribution on the cartridge surface, was assessed with acceptable values ranging from 0.05 to 2.25. Positive control linearity was verified by examining the correlation between observed and expected counts for the four highest concentration positive control probes (128 fM to 2 fM); probes at 0.5 fM and 0.125 fM were excluded from linearity assessment as counts approached background noise levels. All samples demonstrated R-squared values >0.98 for the evaluated positive controls. Background correction was performed using a t-test based approach against eight negative control probes included in each CodeSet. For each endogenous gene, a one-sided t-test compared the gene’s expression across samples against all negative control probe counts. Genes with p-values >0.01 were considered indistinguishable from background noise and were flagged for cautious interpretation. Expression values were normalized using the nSolver algorithm, which implements a two-step normalization procedure. First, technical variation was corrected by calculating a lane-specific positive control normalization factor based on the geometric mean of the six positive control probes, adjusting all counts to account for differences in hybridization efficiency, purification, and binding between lanes. Second, biological normalization was performed using the geometric mean of the six housekeeping genes to correct for differences in sample input between assays. Principal component analysis was used to assess normalization performance and identify potential batch effects or outliers. Differential expression analysis was performed using the limma package with empirical Bayes moderation. 16 For the primary comparisons between response groups, analyses were performed across the full cohort. Multiple testing correction was applied using the Benjamini-Hochberg method across the 30-gene panel within each comparison, with significance defined as adjusted p-value < 0.05 and |log2 fold change| > 1. Histology-stratified analyses (seminoma and non-seminoma) were performed separately within each subgroup using. The same limma framework and Benjamini-Hochberg correction. Given the limited sample size of these subgroups, these analyses were considered exploratory in nature and were underpowered to detect differential expression after multiple testing correction. Results Patient cohort and sample composition A total of 50 TGCT tissue samples from 41 patients were included in this study (Figure 1(a)). Of 41 patients identified, one platinum-sensitive patient was excluded owing to a lack of evaluable expression data, leaving 40 patients across four groups for analysis. The cohort comprised primary TGCT specimens and post-chemotherapy samples, categorized into four groups based on clinical outcomes: chemotherapy-sensitive TGCT (n=21), chemotherapy-resistant TGCT (n=16), post-chemotherapy teratoma (n=5), and post-chemotherapy metastasis (n=8) (Figure 1(b)). Primary histologic subtypes were similar between groups, with non-seminoma germ cell tumours predominating across all categories (Figure 1(c)). Baseline demographics and clinical characteristics were similar between platinum-resistant and platinum-sensitive groups (Table 1), including median age (34 years in the resistant group vs. 28 years in the sensitive group), IGCCCG risk classification (poor risk: 32% vs 27%), and histologic subtype (non-seminoma: 84% vs 86%). Ten patients contributed paired pre- and post-chemotherapy specimens: five platinum-resistant cases with matched primary tumours and post-chemotherapy metastatic samples, and five platinum-sensitive cases with matched primary tumours and post-chemotherapy teratoma. Table 1.Patient characteristics.​Platinum-resistant (n=19)Platinum-sensitive (n=22) Median age, years 3428 Histology, n (%) Non-seminoma/mixed16 (84)19 (86) Seminoma3 (16)3 (14) IGCCCG risk group, n (%) Good5 (26)8 (36) Intermediate8 (42)8 (36) Poor6 (32)6 (27) Stage at initial TGCT diagnosis, n (%) Stage 13 (16)5 (23) Advanced16 (84)17 (77) First-line chemotherapy, n (%) BEP16 (84)19 (86) EP2 (11)3 (14) VeiP1 (5)0 (0) Specimens analysed, n Pre-chemotherapy primary tumour1621 Post-chemotherapy teratoma05 Post-chemotherapy metastasis80One platinum-sensitive patient was excluded owing to a lack of evaluable expression data. Specimens from remaining 40 patients (n=50) were analysed. Post-chemotherapy teratoma specimens were obtained from the platinum-sensitive group. Patient characteristics. One platinum-sensitive patient was excluded owing to a lack of evaluable expression data. Specimens from remaining 40 patients (n=50) were analysed. Post-chemotherapy teratoma specimens were obtained from the platinum-sensitive group. Gene expression landscape To characterize the overall gene expression patterns across our cohort, we performed principal component analysis (PCA) on the normalized expression data from our 36-gene panel (Figure 1(d)). Notably, we observed distinct separation of post-chemotherapy teratomas from all other sample groups. In contrast, chemotherapy-sensitive and chemotherapy-resistant primary tumours showed substantial overlap. These data suggest that our targeted gene panel captures transcriptional differences associated with teratoma differentiation but does not distinguish primary TGCT based on subsequent chemotherapy response. Grouping genes by functional pathway, teratomas displayed a visually distinct expression signature characterized by markedly reduced expression of the pluripotency marker POU5F1 and altered expression across apoptosis-related genes (Figure 2). The heatmap further reveals heterogeneity within the primary tumour groups, consistent with the overlapping PCA clusters, while highlighting the more homogenous expression profile of teratomas. Expression patterns did not differ significantly between seminoma and non-seminoma histologies within clinical categories (Supplementary File 2), indicating that our findings are not confounded by histological subtype. Gene expression heatmap across TGCT tissue classifications. Heatmap displaying expression of 30 target genes across 50 samples. Expression values are Z-score normalized per gene (row-scaled) and capped at ±3 for visualization. Rows represent genes grouped by functional pathway (right annotation); columns represent samples grouped by tissue classification (top annotation). Within each tissue classification, samples are hierarchically clustered. Differential gene expression analysis Differential expression analysis revealed strikingly different findings between the two primary comparisons. Comparing post-chemotherapy teratoma to chemotherapy-sensitive TGCT, twelve genes met our significance threshold (Figure 3(a)). In contrast, comparing chemotherapy-resistant to chemotherapy-sensitive primary tumours yielded no significant differences (Figure 3(b)). Differential gene expression analysis. (a) Volcano plot comparing gene expression in post-chemotherapy teratoma (n = 5) versus chemotherapy-sensitive TGCT (n = 21). The x-axis shows log2 fold change; the y-axis shows -log10 q-value (Benjamini-Hochberg adjusted across the 30-gene panel). Differential expression was assessed using limma with empirical Bayes moderation. Dashed lines indicate significance thresholds (|log2 fold change| > 1, q < 0.05). Genes meeting both thresholds are labelled; upregulated genes (orange) and downregulated genes (blue) are highlighted. (b) Volcano plot comparing chemotherapy-resistant TGCT (n = 16) versus chemotherapy-sensitive TGCT (n = 21). (c) Violin plots showing normalized expression of four key dysregulated genes: POU5F1 (pluripotency marker, downregulated in teratoma), RAD51 (DNA repair, downregulated), CDKN1A (cell cycle arrest, upregulated), and FAS (apoptosis receptor, upregulated). Violins show distribution density; embedded boxplots show median and interquartile range; individual points represent samples. Statistical comparisons between chemotherapy-sensitive and post-chemotherapy teratoma groups were performed using Wilcoxon rank-sum test. Among the teratoma findings, POU5F1 showed the largest magnitude of change (Log2FC = -3.6, q = 0.0052), with near-complete loss of expression in teratomas compared to chemotherapy-sensitive tumours (Figure 3(c)). This dramatic downregulation of the canonical pluripotency marker confirms the differentiated phenotype of mature teratoma. DNA repair pathways showed divergent changes between the post-chemotherapy teratoma and chemotherapy-sensitive TGCTs. Homologous recombination genes RAD51 (Log2FC = -1.9, q = 3.7x10-4) and BRCA1 (Log2FC = -1.2, q = 0.0015) were downregulated, as was the mismatch repair gene MSH2 (Log2FC = -1.4, q = 5.6x10-4). Yet the nucleotide excision repair gene ERCC1 was upregulated (Log2FC = 1.2, q = 0.0027), suggesting pathway-specific remodelling of DNA damage response capacity rather than global suppression. Apoptosis and cell cycle genes showed coordinated upregulation. FAS (Log2FC = 1.9, q = 7.9 x10-6), BCL2 (Log2FC = 2.2, q = 5.6 x10-4), and BCL2L1 (Log2FC = 1.1, q = 0.0022) were all increased, as were the cell cycle checkpoint regulators CDKN1A (Log2FC = 2.7, q = 2.0 x10-5) and 14-3-3 sigma (Log2FC = 2.4, q = 0.0015). The concurrent upregulation of both pro-apoptotic FAS and anti-apoptotic BCL2 family members suggests a rebalanced apoptotic threshold, while elevated CDKN1A and 14-3-3 sigma indicate cell cycle arrest consistent with the indolent clinical behaviour of teratoma (Figure 3(c)). Stress response genes HSP27 (Log2FC = 1.2, q = 0.015) and AKT1 (Log2FC = 1.1, q = 0.0084) were also upregulated. Notably, TP53 expression did not differ significantly between any of our comparison groups. The absence of significant expression differences between chemotherapy-resistant and chemotherapy-sensitive primary tumours suggests that resistance mechanisms in TGCT are not captured by transcriptional changes in the DNA repair and apoptosis pathways examined here. Within-patient paired analysis of the five chemosensitive patients with matched primary and teratoma specimens confirmed consistent directional changes in POU5F1, FAS, and CDKN1A (Supplementary File 3). Discussion Despite high cure rates, patients with TGCTs who develop platinum-resistant disease account for nearly all TGCT-related mortality. This small but clinically critical subgroup has driven sustained interest in identifying molecular determinants of treatment failure. Although recent genomic profiling studies confirm that TGCTs are highly aneuploid, characterised by near-universal 12p gain, low mutational burden, and relatively few recurrent driver mutations, these large-scale analyses have not revealed a consistent genetic basis for cisplatin resistance. 17 This genomic simplicity highlights the need to explore additional layers of biology, including transcriptional, epigenetic, and pathway-level dysregulation, using approaches such as gene-expression profiling to better understand mechanisms of platinum sensitivity and resistance. In our study, the most consistent and biologically coherent signal was teratoma-specific, characterised by marked downregulation of POU5F1 (OCT4) together with coordinated upregulation of apoptosis-related and cell-cycle arrest genes (FAS, CDKN1A and BCL2). This transcriptional pattern reflects the differentiated, low-proliferative phenotype of mature teratoma rather than features of chemotherapy resistance. OCT4 has been shown to maintain high levels of the pro-apoptotic protein NOXA, which keeps TGCT cells primed for chemotherapy-induced death. 18 Mechanistically, OCT4-dependent NOXA priming has been established as a key determinant of the profound cisplatin sensitivity of pluripotent TGCT components; thus, loss of the OCT4/NOXA axis plausibly raises the apoptotic threshold and contributes to the intrinsically chemotherapy-inert nature of mature teratoma. 19 Our finding that BCL2 is upregulated in teratomas provides direct support for this model: increased BCL2 would shift this balance away from apoptosis, contributing to the chemotherapy-inert phenotype of differentiated teratomas. These findings are consistent with previous reports of gene expression changes in teratomas, 20 though comparison with more recent transcriptomic studies is limited, 21 as most genes identified in those larger panels were not included in our 36-gene targeted panel. In contrast to the distinct transcriptional patterns in mature teratoma, we did not detect a robust molecular signature differentiating platinum-sensitive from platinum-resistant TGCTs. Genes frequently implicated in platinum resistance in preclinical models, such as ERCC1, XRCC2, or those along the PI3K/AKT pathway7,22 did not exhibit differential changes in our analyses to distinguish between the platinum-resistant or sensitive mechanisms. This result mirrors findings from large genomic profiling studies, which similarly report that resistant tumours do not harbor a consistent set of recurrent mutations, copy-number events, or structural alterations that reliably predict treatment failure.23,24 The absence of a single predictive biomarker aligns with mounting evidence that platinum resistance may be driven by multifactorial processes, encompassing p53-independent pathways, epigenetic reprogramming, and potentially an interplay with the tumour microenvironment (e.g., IL6-related immune pathways).25–29 Emerging epigenomic studies provide an additional explanation, demonstrating that cisplatin resistance in TGCT may be associated with global DNA hypermethylation and methylation remodelling.30,31 The substantial overlap between platinum-sensitive and platinum-resistant tumours on PCA further supports the absence of a dominant transcriptional pathway within this dataset. This overlap may reflect extensive biological heterogeneity between tumours, relatively subtle resistance-associated transcriptional effects or resistance, mechanisms mediated through pathways or non-transcriptional processes not captured by the current assay. Importantly these findings should not be interpreted as evidence for a true biological absence of differences between platinum-sensitive and platinum-resistant TGCTs but rather suggest that resistance-associated biology may be insufficiently resolved by targeted mRNA profiling. Collectively, these findings support the emerging view that platinum resistance in TGCT is likely driven by complex and interconnected processes extending beyond DNA repair and apoptosis represented in our targeted panel. The central role of the p53 pathway in TGCT biology underscores why transcriptional profiling alone may be insufficient to capture the mechanisms underpinning platinum resistance. In our NanoString analysis, no differential TP53 mRNA expression was observed between platinum-sensitive and platinum-resistant primary tumours, and no unified resistance signature emerged within our targeted DNA-repair/apoptosis panel. This aligns with previous studies 32 and genomic evidence that primary TGCTs are uniformly TP53 wild-type, 19 with the exception of mediastinal GCTs which are enriched for TP53 mutations. 33 Taken together, these results suggest that resistance does not hinge on steady-state TP53 transcript levels but rather on post-transcriptional regulation of p53 activity and apoptotic threshold, including mitochondrial priming and MDM2-mediated p53 suppression, which was not captured by our gene set. Primary TGCTs show a unique “p53 hypersensitivity” in which p53 activation triggers apoptosis rather than the cell cycle arrest typical of other tumour types.18,34,35 Notably, TP53 mRNA expression does not directly reflect p53 pathway activity, as missense mutations typically cause protein stabilisation rather than altered transcript levels. 36 It is also noteworthy that we did not observe consistent mRNA-level differences in DNA repair or cell-cycle genes that segregate platinum-resistant from platinum-sensitive tumours. In particular, XRCC2 and CCND1, previously linked to cisplatin resistance in TGCT models and retrospective series,37,38 did not distinguish resistance in our cohort at the transcript level. This aligns reports suggesting that while such genes can be present in chemotherapy-naïve tumours, they may be selected or modulated under treatment pressure rather than serving as baseline discriminators of resistance across patients. 39 Additionally, the role of OCT4 (encoded by POU5F1), previously associated with platinum sensitivity and apoptosis, remains ambiguous. Although low OCT4 expression is linked to NOXA-mediated mitochondrial priming in pluripotent TGCT components,18,34,40,41 we did not observe significant differences between platinum-sensitive and resistant non-teratoma samples. This suggests OCT4’s impact is context-dependent and resistance in clinical samples may reflect heterogeneous, patient-specific rewiring of apoptotic control that is not captured by steady-state OCT4 mRNA levels. The absence of a unifying molecular signature for platinum resistance in our focused gene-expression panel is informative. Large-scale genomic profiling studies likewise report no single recurrent genomic hallmark of resistance across gonadal TGCTs, with the notable exception of TP53/MDM2 lesions enriched in specific contexts (e.g., mediastinal non-seminoma). Together with our data, these observations argue that clinically actionable biomarkers will likely not emerge from narrow candidate lists of DNA repair/apoptosis genes alone. Rather, discovery efforts should prioritise multi-omic integration (genome, transcriptome, methylome, small RNAs, proteome) and functional assays to capture apoptotic priming and post-transcriptional regulation that escape detection by mRNA-only approaches. Our study has several limitations. First, we analysed a targeted 30-gene panel designed to interrogate pre-defined pathways related to DNA repair, apoptosis and cell-cycle regulation; while this focused approach was well suited to potentially degraded FFPE-derived RNA, it may not fully capture the complex, multifactorial, and interconnected biology underlying platinum resistance. Several regulators now implicated in resistance, including epigenetic modifiers, immune-related pathways, non-coding RNAs, and post-transcriptional regulators such as MDM2, were not represented in our assay. Second, although the overall cohort size was acceptable for a rare disease, subgroup analyses were limited by small numbers, particularly in the post chemotherapy teratoma cohort and paired-sample analyses, reducing statistical power. Additionally, the platinum-sensitive cohort included both complete responders and patients with residual teratoma following chemotherapy, which may represent biologically distinct clinical entities and contribute to additional heterogeneity. The inherent molecular heterogeneity within NSGCTs further complicates the identification of shared resistance-associated pathways. Third, reliance on archival FFPE tissue, despite the robustness of NanoString, can limit dynamic range and degrade some analytes. Fourth, our analysis was restricted to mRNA expression profiling without orthogonal protein-level or functional validation, consequently, transcriptional differences may not directly reflect pathway activity or biological function. Finally, the limited availability of longitudinal clinical data and absence of external validation cohorts constrained our ability to correlate expression patterns with survival outcomes or refine prognostic models. Larger-scale studies integrating circulating tumour DNA, microRNA profiling, and comprehensive immune analyses will likely be necessary to capture the full landscape of platinum resistance. Integration with independent validation cohorts and complementary multi-omic approaches, including longitudinal clinical outcomes, proteomic analyses and epigenetic profiling, may further improve the translational relevance of these findings and help identify clinically meaningful resistance mechanisms not detectable through targeted mRNA profiling alone. International collaboration, as well as multicentre consortia, will be instrumental in obtaining sufficiently large cohorts of resistant TGCTs to detect subtle, clinically relevant signals. These efforts could pave the way for developing personalised therapies that target the molecular underpinnings of platinum resistance and improve outcomes for the subset of patients who currently receive limited benefit from standard treatments. Conclusion In summary, our NanoString-based analysis reveals that post-chemotherapy teratomas in TGCTs exhibit a biologically distinct expression profile, characterised by lower POU5F1 (OCT4) expression and higher levels of apoptosis and cell-cycle related genes, compared to platinum-sensitive and resistant tumours. Despite this distinction, we did not identify a unifying signature of cisplatin resistance to distinguish from platinum-sensitive TGCTs within the constraints of our targeted gene panel and cohort size. These findings should not be interpreted as evidence for a true biological absence of differences between resistant and sensitive disease states, but rather highlight the likely multifactorial, heterogenous and potentially non-transcriptional nature of platinum resistance biology. These negative findings underscore the need for more comprehensive investigations incorporating immune and epigenetic profiling to ultimately improve outcomes for the subset of TGCT patients with platinum-resistant disease. Supplemental material Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Acknowledgments The authors would like to acknowledge the contributions of the Princess Maragret Cancer Centre multidisciplinary testicular cancer program and the patients whose samples contribute to this research. Author contributions: All authors contributed to the study conception, design, writing and editing of the manuscript. AG and EK also performed data curation. Funding: The authors received no financial support for the research, authorship, and/or publication of this article. The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: B.T has received research Funding from Amgen, Astellas, AstraZeneca, Bayer, BMS, Genentech, Ipsen, Janssen, Pfizer, Movember, MSD; Honorarium from Amgen, Astellas, AstraZeneca, Bayer, BMS, Ipsen, Janssen, Merck, MSD, Pfizer, Sanofi, Tolmar; and has provided consulting Advisory services for Amgen, Astellas, AstraZeneca, Bayer, BMS, Ipsen, IQVIA, Janssen, Merck, MSD, Novartis, Pfizer, Roche, Sanofi, Tolmar, VIR. E.M.K. has consulted or served in an advisory role for Astellas Pharma, Janssen and Ipsen, received travel funding from Astellas Pharma, Pfizer, Ipsen and Roche, received honoraria from Janssen, Ipsen, Astellas Pharma and Research Review, and received research funding from Astellas Pharma (institutional) and AstraZeneca (institutional). P.B has received honoraria from Boehringer Ingelheim, consulted or served in an advisory or consulting role for Bayer and Daiichi Sankyo and received research funding from Bristol-Myers Squibb, Sanofi, AstraZeneca, Genentech/Roche, GlaxoSmithKline, Novartis, Merck, Seattle Genetics, Lilly, Amgen, Bicara, Zymeworks, Medicenna, Bayer, Takeda, Gilead Sciences, LegoChem Biosciences, Boehringer Ingelheim, ProteinQure, Daiichi Sankyo. All other authors have no conflicts to declare in relation to this study. Supplemental material: Supplemental material for this article is available online. ORCID iDs Nadia Hitchen https://orcid.org/0000-0001-9327-4171 Edmond M. Kwan https://orcid.org/0000-0002-7053-680X Ben Tran https://orcid.org/0000-0001-9124-354X Ethical considerations This study was conducted in accordance with the Declaration of Helsinki and approved by the University Health Network Research Ethics Board in Toronto, Canada. All living patients had previously provided consent for tissue retrieval and exploratory molecular analysis. A waiver of consent was granted for deceased patients in accordance with institutional research ethics approval. Consent for publication Consent for publication was not required as all data were de-identified and analysed in accordance with institutional research ethics approval. Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding authors on request.* References Associated Data Supplementary Materials Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Supplemental material - NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes Supplemental material for NanoString profiling of DNA repair and apoptosis genes in testicular germ cell tumours with divergent chemotherapy outcomes by Nadia Hitchen, Philippe L. Bedard, Neil Winegarden, Joan Sweet, Amanda Grubb, Lynn Anson-Cartwright, Peter Chung, Padraig Warde, Malcolm Moore, Edmond M. Kwan and Ben Tran in Therapeutic Advances in Medical Oncology. Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding authors on request.*