Launonen, I.-M. | bioRxiv Science

bioRxiv2024

Multi-Omics Analysis Reveals the Attenuation of the Interferon Pathway as a Driver of Chemo-Refractory Ovarian Cancer

Ovarian high-grade serous carcinoma (HGSC) represents the deadliest gynecological malignancy, with 10-15% of patients exhibiting primary resistance to first-line chemotherapy. These primarily chemo-refractory patients have particularly poor survival outcomes, emphasizing the urgent need for developing predictive biomarkers and novel therapeutic approaches. Here, we show that interferon type I (IFN-I) pathway activity in cancer cells is a crucial determinant of chemotherapy response in HGSC. Through a comprehensive multi-omics analysis within the DECIDER observational trial (ClinicalTrials.gov identifier NCT04846933) cohort, we identified that chemo-refractory HGSC is characterized by diminished IFN-I and enhanced hypoxia pathway activities. Importantly, IFN-I pathway activity was independently prognostic for patient survival, highlighting its potential as a biomarker. Our results elucidate the heterogeneity of treatment response at the molecular level and suggest that augmentation of IFN-I response could enhance chemosensitivity in refractory cases. This study underscores the potential of the IFN-I pathway as a therapeutic target and advocates for the initiation of clinical trials testing external modulators of the IFN-I response, promising a significant stride forward in the treatment of refractory HGSC.

cancer biology↗

bioRxiv2024

Tribus: Semi-automated discovery of cell identities and phenotypes from multiplexed imaging and proteomicdata

MotivationMultiplexed imaging and single-cell analysis are increasingly applied to investigate the tissue spatial ecosystems in cancer and other complex diseases. Accurate single-cell phenotyping based on marker combinations is a critical but challenging task due to (i) low reproducibility across experiments with manual thresholding, and, (ii) labor-intensive ground-truth expert annotation required for learning-based methods. ResultsWe developed Tribus, an interactive knowledge-based classifier for multiplexed images and proteomic datasets that avoids hard-set thresholds and manual labeling. We demonstrated that Tribus recovers fine-grained cell types, matching the gold standard annotations by human experts. Additionally, Tribus can target ambiguous populations and discover phenotypically distinct cell subtypes. Through benchmarking against three similar methods in four public datasets with ground truth labels, we show that Tribus outperforms other methods in accuracy and computational efficiency, reducing runtime by an order of magnitude. Finally, we demonstrate the performance of Tribus in rapid and precise cell phenotyping with two large in-house whole-slide imaging datasets. AvailabilityTribus is available at https://github.com/farkkilab/tribus as an open-source Python package.

bioinformatics↗

bioRxiv2023

Single-cell spatial atlas of tertiary lymphoid structures in ovarian cancer

BackgroundRecent advances in highly-multiplexed tissue technologies and image analysis tools have enabled a more detailed investigation of the tumor microenvironment (TME) and its spatial features, including tertiary lymphoid structures (TLSs), at single-cell resolution. TLSs play a major part in antitumor immune responses, however, their role in antitumor immunity in ovarian cancer remains largely unexplored. MethodsIn this study, we generated a comprehensive single-cell spatial atlas of TLSs in ovarian cancer by extracting spatial topology information from in-situ highly-multiplexed cellular imaging using tissue cyclic immunofluorescence (CyCIF). Our analysis included 44 patients with high-grade serous ovarian cancer (HGSC) from the TOPACIO Phase II clinical trial. We combined spatial and phenotypic features from 302,545 single-cells with histopathology, targeted sequencing-based tumor molecular groups, and Nanostring gene expression data. ResultsWe find that TLSs are associated with a distinct TME composition and gene expression profile, characterized by elevated levels of the chemokines CCL19, CCL21, and CXCL13 correlating with the number of TLSs in the tumors. Using single-cell feature quantification and spatial mapping, we uncover enriched germinal center (GC) B cell infiltration and selective spatial attraction to follicular helper T and follicular regulatory T cells in the TLSs from chemo-exposed and BRCA1 mutated HGSCs. Importantly, spatial statistics reveal three main groups of cell-to-cell interactions; significantly enriched structural compartments of CD31+ cells, myeloid, and stromal cell types, homotypic cancer cell- and cancer cell to IBA1+ myeloid cell crosstalk, and enriched selective Tfh, Tfr, and Tfc communities with predominant Tfh - GC B cell interactions. Finally, we report spatiotemporal gradients of GC-B cell interactions during TLS maturation, with enriched non-GC B cell attraction towards the GC B cells in early TLSs, and avoidance patterns with selective GC B-cell communities in the TLSs with GCs. ConclusionsOur single-cell multi-omics analyses of TLSs showed evidence of active adaptive immunity with spatial and phenotypic variations among distinct clinical and molecular subtypes of HGSC. Overall, our findings provide new insights into the spatial biology of TLSs and have the potential to improve immunotherapeutic targeting of ovarian cancer. What is already known on this topicTLSs play a major part in antitumor immune responses, however, their exact role and mechanisms in antitumor immunity are widely unexplored. What this study addsOur results deepen the understanding of TLS biology including cell-cell interactions and shows how the presence of TLSs is characterized with a distinct TME composition and gene expression profile. How this study might affect research, practice or policyOverall, our findings provide new insights into the spatial biology of TLSs and have the potential to improve therapeutic options for ovarian cancer.

cancer biology↗