Thèse Caractérisation du Paysage Immunitaire Associé à Ras par Technologies de Cellule Unique dans l'Adénocarcinome du Poumon H/F

Doctorat.Gouv.Fr

Établissement : Université Paris-Saclay GS Life Sciences and Health École doctorale : Cancérologie : Biologie - Médecine - Santé Laboratoire de recherche : Prédicteurs moléculaires et nouvelles cibles en oncologie Direction de la thèse : Sophie DE CARNÉ ORCID 0000000293412586 Début de la thèse : 2026-10-01 Date limite de candidature : 2026-04-28T23:59:59 Lung adenocarcinoma (LUAD) remains the leading cause of cancer-related mortality, and therapeutic success with immune checkpoint blockade or KRAS-targeted inhibitors is limited to a subset of patients. Although oncogenic RAS activity is known to shape the tumour microenvironment through several mechanisms, the immune ecosystems associated with RAS activation remain poorly defined, partly because RAS pathway activation extends well beyond KRAS mutation status and complicates analysis in large patient cohorts. A clearer understanding of the immune consequences of oncogenic RAS signalling is therefore essential to advancing personalised therapy.
Our laboratory developed a transcription-based method (RAS84) to stratify LUAD tumours according to RAS activity. Using this approach and bulk RNA Seq deconvolution, we have shown that distinct levels of RAS activity correspond to distinct immune signatures. This PhD project will extend those findings by characterising the RAS-dependent immune landscape using single-cell and spatial transcriptomics from a Gustave Roussy patient cohort, complemented by CITE-Seq datasets from preclinical models treated with KRAS inhibitors. Applying RAS84 across these multi-omics datasets will enable the identification of immune and stromal states, ligand-receptor communication networks, and spatially organised ecosystems linked to oncogenic RAS. Integration of human tumours with perturbation data from mouse models will allow the validation of immune features directly modulated by RAS signalling.
The PhD candidate will work within a synergistic environment bringing together data scientists and wet-lab experimentalists. They will contribute to advancing our understanding of LUAD, starting with human samples, and collaborate with team members to experimentally validate emerging hypotheses. LUAD and therapeutical problematic
Non-small cell lung cancer (NSCLC) remains the leading cause of cancer-related deaths worldwide (1). Lung adenocarcinoma (LUAD), its most prevalent subtype, poses a major clinical challenge because of its high mortality and limited treatment efficacy. Whilst early diagnosis leads to surgical resection and better outcomes, most patients are diagnosed too late and cannot benefit from surgery. Immune checkpoint blockade (ICB) targeting the PD-1/PD-L1 axis to reactivate the immune system is now the standard of care for these patients. These therapies can offer long-term benefits, but only to a subset of patients (2,3). RAS inhibitors, including recently approved KRASG12C inhibitors, also work, in part, by indirectly reactivating the anti-tumour immune response (4,5). In contrast with immunotherapy, these targeted therapies have shown excellent response rates, but rapid resistance occurs in almost all patients receiving them (6).
KRAS in LUAD
A third of LUAD cases harbour KRAS mutations, leading to constitutive activation of the RAS signalling pathway. Beyond its well-known role in promoting cell proliferation and survival (7-10), oncogenic RAS signalling can directly drive immune evasion by increasing PD-L1 expression via the stabilisation of its CD274 mRNA (11), impairing IFN response, increasing expression of the immunosuppressive enzyme COX2 (12), and promoting immunosuppressive adenosine in the tumour microenvironment (TME) (13). This indicates that oncogenic RAS signalling can profoundly shape the TME composition and its phenotypic activation, thereby promoting tumour progression and influencing the therapeutic response to the abovementioned therapies.
KRAS mutation versus RAS transcriptional activation
Despite the expected impact of KRAS mutations on tumour progression and therapy resistance, there is no consensus in the literature regarding their predictive value for patient outcomes or response to therapy (14-18). Because 74% of lung adenocarcinoma (LUAD) tumours are mutated in one or more genes within the broader RAS pathway, from receptor tyrosine kinases to ERK MAP kinases and PI3 kinase (19), we think this could have complicated the study of oncogenic RAS in large patient cohorts. We therefore developed a stratification method based on RAS-regulated transcriptional activity (RAS84) to account for all alterations that could affect the RAS pathway. Using this method, we could demonstrate that RAS pathway activation was not restricted to KRAS-mutant tumours. We found that more than 80% of LUAD tumours show transcriptional signs of RAS pathway activity, including those with wild-type KRAS. These tumours fall into four groups, each with enrichment in distinct RAS target genes and characterised by concurrent alterations in STK11/LKB1, TP53, or CDKN2A. They also exhibit significant variations in response to therapy (13,20), indicating that refining LUAD patient stratification based on RAS transcriptional activity could improve personalised medicine. Given the central role of RAS signalling in determining clinical outcomes in LUAD, a more precise characterisation of the RAS-dependent tumour microenvironment is required to improve responses to immunotherapy, limit relapse under RAS-targeted treatments, and support more refined approaches to personalised medicine. With this project, we propose to delineate the tumour microenvironment associated with oncogenic RAS activity in LUAD using human tumour samples and a novel stratification method based on oncogenic RAS transcriptional activity. The PhD candidate will have access to the MOSAIC cohort of 120 LUAD patients with Visium ST data, snRNA-Seq, bRNA-Seq, and WES at baseline, as well as CITE-Seq data from a preclinical experiment performed in two murine orthotopic, syngeneic lung cancer models treated or not with a KRAS inhibitor. The PhD candidate will use the human multi-omics spatial and single-cell dataset to correlate RAS activity with immune features identified in the data. The pre-clinical experimental data will be used to test whether modulating the RAS pathway in the tumour directly affects those features. The planned work packages are listed below.
WP1 - Classify Tumours Using the RAS84 Signature (Bulk RNA Seq)
The PhD candidate will work directly with the developer of RAS84 to apply the signature across the MOSAIC cohort. They will generate RAS activity groups, assess classification robustness and link RAS84 labels to genomic and clinical features. They will integrate these findings with single-cell and spatial analyses to build a unified map of RAS activity for the project.
WP2 - Characterise the RAS84 Specific Immune Landscape (snRNA Seq)
The candidate will analyse snRNA Seq data to define how RAS activity shapes immune and stromal composition. They will identify enriched cell populations, quantify pathway activation and relate these features to mutational profiles and known RAS-regulated mechanisms. They will use these results to uncover mechanistic signatures of RAS-dependent immunity.
WP3 - Map RAS Linked Spatial Ecosystems (Visium Spatial Transcriptomics)
The candidate will examine Visium datasets to locate spatially organised immune and stromal structures associated with RAS activity. They will identify spatial expression domains, characterise immune niches and detect stromal remodelling patterns across RAS84 groups. They will integrate spatial patterns with snRNA Seq cell states to produce a spatially resolved view of RAS-driven ecosystems.
WP4 - Reconstruct Cell-Cell Communication Networks
The candidate will integrate single-cell and spatial data to infer ligand-receptor interactions across RAS84 groups. They will identify signalling circuits that promote immune suppression, inflammation or altered therapy response. They will highlight communication pathways that may influence sensitivity to immunotherapy or RAS-targeted treatments.
WP5 - Link RAS Signalling to Microenvironmental Features
The candidate will evaluate how RAS signalling intensity aligns with the cellular and spatial features identified in earlier WPs. They will test associations with processes such as interferon signalling, PD L1 stabilisation, COX2-driven inflammation and adenosine-mediated suppression. Using both human and preclinical datasets, they will identify RAS-regulated pathways that drive specific microenvironmental phenotypes and may offer therapeutic opportunities.