Spatiotemporal mapping of hypoxia–immune–ECM interactions driving therapy resistance in MASH-associated hepatocellular carcinoma

Primary supervisor: Gilbert Fruhwirt, King’s College London

Secondary supervisor: Oscar Maiques, Queen Mary University of London

Project

Liver cancers remain a major global health challenge (estimated incidence >1 million/year) with ~6k annual deaths in the UK [Siegel_CACancerJClin_2025]. Hepatocellular carcinomas (HCC) account for ~90% of liver cancer cases. Alcohol misuse and hepatitis B and C virus infection have been the main HCC risk factors, but metabolic dysfunction-associated steatohepatitis (MASH, formerly NASH), linked to metabolic syndrome and diabetes mellitus, is now a major risk factor in the West and drives incidence increases there (projected ~33% increase in the UK over the next 15 years).

Despite new options for advanced/unresectable HCC (immune checkpoint inhibitors (ICI), combination targeted therapy, stereotactic external beam radiotherapy (EBRT)), 10-year survival remains only 8% and overcoming therapy resistance is the main challenge: only 15–20% of patients benefit from ICI and 3-year survival after EBRT is only 23%. Hypoxia is a negative predictive factor for radiotherapy response, lowering the efficacy of ionizing radiation by a factor of 2–3 [Thoday_Nature_1947]. Yet many mechanistic aspects of the interplay between immune evasion, radiation resistance and hypoxia remain elusive, particularly in MASH-driven HCC.

We will define how MASH and hypoxia impact HCC progression and dissemination, and HCC immuno- and radiotherapy response, at the molecular level – harnessing the combined powers of multi-scale imaging, multi-parametric tissue analyses and computational data integration.

We build on existing murine MASH models and our established in vivo cancer tracking technology [Fruhwirth_JNucMed_2014, Maiques_BrJCancer_2021] for 3D quantification of tumour growth and metastasis over time, together with high-dimensional spatial tissue analysis. New in-house hypoxia fate-mapping technology will reveal how hypoxia alters HCC gene expression in vitro and how it shapes tumour establishment and progression in vivo. Ex vivo analyses will show how MASH-associated fibrosis and hypoxia-driven desmoplasia alter gene expression and extracellular matrix (ECM) architecture, building on our understanding of how ECM mechano-sensing drives cytoskeletal and transcriptional memory supporting tumour aggressiveness [Maiques_NatComm_2025].

Tumours will then be challenged with combination therapy (ICI±vascular disrupting agent or ICI±EBRT) and the resulting spatiotemporal data integrated with human patient-derived data – tissue microarray and clinical outcome datasets from collaborating hepatology centres (e.g. Blizard@QMUL, Roger-Williams-Institute-of-Liver-Studies@KCL) providing validation cohorts. Integration will employ AI-driven digital pathology pipelines established by Maiques [Arslan_CommMed_2024], predicting multi-omic biomarkers from routine H&E sections, with spatial transcriptomics correlation and virtual multiplexing to map progression and treatment response.

Deliverables: a MASH-stage-resolved map of HCC progression, a hypoxia-history-resolved spatial atlas of the HCC microenvironment, and an H&E-deployable AI signature of therapy resistance validated in human cohorts.

Figure 1. Project overview and work package interactions (left column) and supporting data (right column). Laboratory expertise: Gilbert Fruhwirth (multi-scale imaging, therapeutic interventions; Data A-B); Oscar Maiques (AI-driven digital pathology, spatial analysis; Data C).

Candidate background

Suitable candidates will have a minimum upper 2nd-class degree in cell biology/immunology/biochemistry or a closely related field with a strong background in cancer biology, an interest in advanced multi-scale imaging and multi-parametric analysis, and a passion for hypothesis-driven research. This is a highly multi-disciplinary project, and the student will employ a range of techniques including cell culture, mouse tumour models, diverse imaging technologies (i.e. SPECT/PET-CT, fluorescence microscopy, immunohistochemistry, spatial biology), fluorescence-activated cell sorting, transcriptomics, and computational methodologies for integrated data analysis. No prior programming experience is required – full training in computational and laboratory methods will be provided.

Potential Research Placements

1. Anita Grigoriadis, Comprehensive Cancer Centre, King’s College London
2. Mirjana Efremova, Barts Cancer Institute, Queen Mary University of London
3. Claude Chelala, Barts Cancer Institute, Queen Mary University of London

References

1. Siegel RL, Kratzer TB, Giaquinto AN, Sung H, Jemal A. Cancer statistics, 2025. CA Cancer J Clin. 2025;75(1):10-45. doi: 10.3322/caac.21871.
2. Fruhwirth GO, Diocou S, Blower PJ, Ng T, Mullen GED. A whole-body dual-modality radionuclide optical strategy for preclinical imaging of metastasis and heterogeneous treatment response in different microenvironments. J Nucl Med. 2014;55(4):686-94. doi: 10.2967/jnumed.113.127480.
3. Maiques O, Fanshawe B, Crosas-Molist E, Rodriguez-Hernandez I, Volpe A, Cantelli G, Boehme L, Orgaz JL, Mardakheh FK, Sanz-Moreno V, Fruhwirth GO. A preclinical pipeline to evaluate migrastatics as therapeutic agents in metastatic melanoma. Br J Cancer. 2021;125(5):699-713. doi: 10.1038/s41416-021-01442-6.
4. Arslan S, Schmidt J, Bass C, Mehrotra D, Geraldes A, Singhal S, Hense J, Li X, Raharja-Liu P, Maiques O, Kather JN, Pandya P. A systematic pan-cancer study on deep learning-based prediction of multi-omic biomarkers from routine pathology images. Commun Med. 2024;4(1):48. doi: 10.1038/s43856-024-00471-5.
5. Maiques O, Sallan MC, Laddach R, Pandya P, Varela A, Crosas-Molist E, Barcelo J, Courbot O, Liu Y, Graziani V, Arafat Y, Sewell J, Rodriguez-Hernandez I, Fanshawe B, Jung-Garcia Y, Imbert PRC, Grasset EM, Albrengues J, Santacana M, Macià A, Tarragona J, Matias-Guiu X, Marti RM, Tsoka S, Gaggioli C, Orgaz JL, Fruhwirth GO, Wallberg F, Betteridge K, Reyes-Aldasoro CC, Haider S, Braun A, Karagiannis SN, Elosegui-Artola A, Sanz-Moreno V. Matrix mechano-sensing at the invasive front induces a cytoskeletal and transcriptional memory supporting metastasis. Nat Commun. 2025;16(1):1394. doi: 10.1038/s41467-025-56299-7.