Team members

Signalling networks for stemness and tumorigenesis

Our team studies resilience versus vulnerability of cells to signalling fluctuations to understand how cancer starts, evolves, and responds to treatments.

Our team studies resilience versus vulnerability of cells to signalling fluctuations to understand how cancer starts, evolves, and responds to treatments. A key tool in our studies is a unique mouse genetic setting in which slightly increased wild-type receptor tyrosine kinase levels trigger the tumorigenic program in vulnerable tissues: liver and mammary gland. This “inside-out” genetic setting is a predisposing cancer model open to a variety of spontaneous alterations, thoroughly recapitulating disease features occurring in human patients. By combining complementary tools within interdisciplinary approaches, we identify new mechanisms at the root of tissue homeostasis destabilisation, tumour initiation, progression, as well as vulnerabilities to exploit for targeted therapies. Furthermore, we study how qualitative and quantitative signalling changes lead to diversified cell behaviours using stem cell cultures, organoids, and tumoroids, to document how perception of environmental signals and modulation of epithelial integrity impact cell behaviour, cell-cell interaction, and the acquisition of diversified cell identities.

Drosophila suzukii assesses the quality of a ripe cherry before deciding where to lay an egg


Our last publications


of the team

Team members

They drive our research


They contributed to our research
Yannan Fan
Postdoctoral Researcher INSERM, France
Francesca Zaccagnino
Postdoctoral Researcher at Aix-Marseille Univ, France
Serena Corti
Stem Cell Scientist at University of Tübingen, Germany
Sehrish Khan Bazai
postdoc at the Institute of Genetics and Molecular and Cellular Biology (IGBMC), France
Diane Rattier
Scientific Researcher at Poietis-Pan3D, France
Ahmed Abdouni
Senior Scientist, Veracyte, Inc., France
Fahmida Ahmad
Lecturer at SBK Women’s at the University of Quetta, Pakistan
Rémi Bonjean
Ingénieur d’études at Aix-Marseille Univ, France

Funding bodies

They support our research


Signalling networks for stemness and tumorigenesis

Our achievements illustrate how a combinatorial use of complementary tools and approaches led to uncover new mechanisms regulating tissue homeostasis, triggering its perturbation towards tumorigenesis, and conferring vulnerabilities of cancer cells for new treatment options.

Subtle increases in wild-type MET-RTK levels are sufficient for spontaneous occurrence of tumours in restricted sensitive tissues, conceptually illustrating how the shift from physiological to pathological conditions occurs overtime as a result of slight perturbations in signalling dosage.

A tissue vulnerable to perturbations in RTK levels is the liver, with spontaneous tumour formation evolving into hepatocellular carcinoma (HCC; Alb-R26Met mice). This model recapitulates inter-/intra-tumour heterogeneity and primary resistance to drugs used in the clinic. We undertook an unbiased approach to analyse hundreds of signalling proteins simultaneously combining phosphokinome with bioinformatics and identified the signalling network reproduced by the Alb-R26Met tumorigenesis, which led to the identification of new, deleterious synthetic lethal interactions for a HCC subset and their mechanism of action (Hepatology 2017). We showed how enhanced MET levels in mouse hepatocytes attenuates insulin-mediated signalling (Cells 2022).

Tracking dynamics of spontaneous tumours in mice using Photon Counting Computed Tomography.

In collaboration with researchers at the CPPM (Marseille), we applied a new Photon Counting micro-Computed Tomography (PC-CT) approach for in vivo longitudinal monitoring of tumour formation and evolution overtime in Alb-R26Met mice. We achieved great capability to detect neoplastic nodules starting from about 2mm of diameter. Results illustrate how combining physics, medical imaging, genetics, immunology, and molecular therapy for cancer treatment allows uncovering tumour dynamics at initiation/evolution phases, and regression following treatment (iScience 2019). We are exploiting these approaches to identify alterations diversifying quiescent versus evolving preneoplastic lesions, and to document specificity in immune cell type remodelling according to distinct anticancer treatments.

Hypermethylation of gene body CpG islands predicts high dosage of functional oncogenes in liver cancer.

We reasoned that the Alb-R26Met predisposing cancer model open to a variety of spontaneous alterations could offer the possibility to examine the contribution of the epigenetic reprogramming associated with tumorigenicity. By exploring the impact of DNA methylation on transcriptional switches associated with tumorigenesis, we identified a striking enrichment in genes simultaneously overexpressed and hypermethylated in CpG islands (CGIs) located in gene body regions, similar to the HCC proliferative-progenitor subgroup. We demonstrated how these genes act together as an “oncogene module”. Thus, hypermethylation of gene body CGIs is predictive of elevated oncogene levels in cancer, offering a novel stratification strategy and perspectives to normalise cancer gene dosages (Nat Commun 2018).

Evaluating the landscape of gene cooperativity with RTKs in liver tumorigenesis.

The vulnerability of tissues to slightly enhanced RTK dosage offers a unique possibility to address another challenging topic related to the specificity in genetic cooperativity associated with tumorigenesis. We therefore applied the Alb-R26Met mouse model to explore functional cooperation in combination with a screening paradigm, in order to identify alterations that accelerate tumour initiation, combining the Sleeping Beauty (SB) transposon (T2/onc2) mutagenesis system with Alb-R26Met genetic setting. We reasoned that such an unbiased mutagenesis screen would offer a unique way to identify non-predictable genetic interactions underlying gene cooperativity with RTKs during tumour initiation. We identified 275 genes, most of which are altered in HCC patients. Surprisingly, these genes are not restricted to a small set of pathway/cellular processes, but cover a large spectrum of cellular functions. Focussing on 15 tumour suppressor candidates, we demonstrated their action in tumour initiation in an enhanced RTK context, as shRNA-mediated targeting confers tumorigenicity to RTK-sensitized cells, but not to cells with basal RTK levels. Our study identified unanticipated genetic interactions underlying gene cooperativity with RTKs in HCC. Moreover, results document how subtly increased levels of wild-type RTKs provide a permissive context allowing a large spectrum of deregulated mechanisms to initiate liver cancer (J Hepatol 2019).

From screen outcomes to new key regulators of tumorigenesis and biomarkers of HCC: the case of ADAMTSL5.

Outcomes from screen studies we performed offer a valuable set of new genes to consider as potential regulators of tumorigenesis and/or biomarkers for HCC. We focussed our attention on one of these genes still uncharacterised, ADAMTSL5, not previously linked to cancer, and obtained solid evidence for its implication in HCC tumorigenesis. ADAMTSL5 is upregulated in a large percentage of HCC patients (44-70% according to analysed cohorts), characterised by a shorter overall survival and of disease-free interval. Augmented ADAMTSL5 expression correlates with hypermethylation of its gene body CGI. Notably, ADAMTSL5 expression classifies 26% of HCC patients otherwise not identified by AFP, thus permitting to cover up to 80% of HCC patients (manuscript in prep). ADAMTSL5 silencing leads to a striking switch from an epithelial-like to a fibroblast-like cell shape, supported by the acquisition of fibroblast markers. Additionally, ADAMTSL5 targeting interferes with tumorigenic properties of HCC cells, depleting cells of several oncogenes, including RTKs. Consequently, HCC cells become sensitive to drugs used in the clinics. Collectively, these results indicate that ADAMTSL5, produced and secreted by HCC cells, may be a key factor determining their tumorigenicity (J Hepatol 2021; Patent). We have preliminary data indicating ADAMTSL5 action in other cancer types. We are developing ADAMTSL5 targeting agents.

Subtle increases in wild-type MET-RTK levels are sufficient for spontaneous occurrence of tumours in the mammary gland, recapitulating resistance to treatment, and uncovering new combinatorial therapies.

The mammary gland is another tissue vulnerable to subtly increased MET levels, as illustrated by aggressive, infiltrating breast carcinomas, exclusively defined as Triple-Negative Breast Cancer (TNBC; MMTV-R26Met mice). Proteomic profiling, machine learning and drug screen approaches showed that the MMTV-R26Met model recapitulates intra-/inter-tumoral heterogeneity and primary resistance to chemotherapeutic treatments used in clinics. Further signaling network analysis highlighted potential druggable targets, of which co-targeting of WEE1 and BCL-XL synergistically killed TNBC cells and efficiently induced tumor regression. Mechanistically, BCL-XL inhibition exacerbates the dependency of TNBC cells on WEE1 function, leading to Histone H3 and phosphoS33RPA32 upregulation, RRM2 downregulation, cell cycle perturbation, mitotic catastrophe and apoptosis (Adv Sci 2021). We further exploit these findings by exploring the vulnerability of TNBC cells to targeting cell cycle regulators, as they are frequently altered TNBC. Results illustrated that specificity matters in targeting cell cycle regulators for combinatorial anticancer therapies (Theranostics 2021). Through collaborative studies, we contributed to identify two additional new very potent drug combinations, one specific to mesenchymal TNBC and the other effective for all TNBC subtypes (coll. with Pr. Lev, Weizmann Institute, Israel; Sci Adv 2020; Life Sci Alliance 2021). Collectively, our study introduces a unique, powerful mouse model for studying TNBC formation and evolution, its heterogeneity, and for identifying efficient therapeutic targets. These TNBC project are co-leaded in the team with Dr. F. Lamballe.

Integration of signalling by pleiotropic factors in developmental and disease processes: the Glypican4 case.

We established that it is possible to manipulate mouse embryonic stem cell (mESC) and human induced pluripotent stem cell (hiPSC) fate by acting on mechanisms regulating perception of instructing environmental cues. Down-regulation of the morphogen modulator Glypican-4 (GPC4) confers to cells a unique biological state characterized by: a) maintenance of self-renewal/pluripotency in stemness conditions; b) oriented/accelerated cell lineage entry in differentiation conditions; c) loss of tumorigenicity after in vivo transplants. We named this unique biological state as a “safe-PSC state”. We demonstrated that hiPSCs with downregulated GPC4 levels undergo a more efficient commitment towards ventral midbrain dopaminergic neuron (DA) fate at the expense of self-renewal and serotonergic identity and they generate a larger amount of DA neurons. Through transplantation experiment in rat brains showed that grafts of GPC4mut cells are enriched in markers of ventral midbrain precursors. Our results highlight GPC4 downregulation as a powerful approach to enhance generation of VMDA neurons while limiting side effects like teratoma formation when cells are implanted in vivo (Stem Cells Transl Med 2021). This project is leaded in the team by Dr. R. Dono.

Disruption of epithelial integrity drives mesendoderm differentiation in human induced pluripotent stem cells by enabling TGFβ Protein Sensing.

The processes of primitive streak formation and fate specification in the mammalian epiblast rely on complex interactions between morphogens and tissue organization. We interrogated the interplay between tissue organization and morphogens by using GPC4mut hiPSCs, in which defects in tight junctions result in areas of disrupted epithelial integrity. This phenotype does not affect hiPSC stemness, but impacts on cell fate acquisition processes. Strikingly, cells within disrupted areas become competent to perceive BMP4 and ACTIVIN A/NODAL gastrulation signals and thus, differentiate into mesendoderm. Yet, disruption of epithelial integrity prolongs temporal activation of BMP4 and ACTIVIN A/NODAL downstream effectors and correlates with enhanced hiPSC endoderm/mesoderm differentiation potential. Altogether, our results disclose epithelial cell integrity as a key determinant of TGFβ activity and highlight a new mechanism guiding morphogen sensing and spatial cell fate change within an epithelium (bioRxiv 2021). Furthermore, through a CNRS pre-maturation project, we have developed Llama Nanobodies with properties to recognise GPC4; one is also able to block GPC4 function (patent). This project is leaded in the team by Dr. R. Dono.