CAR-T cell therapy has transformed the treatment of hematologic malignancies but remains largely ineffective in solid tumors, mainly because of limited infiltration, immunosuppressive environments, and strong intratumoral heterogeneity. This heterogeneity promotes antigen escape, a major cause of relapse. One way the immune system can compensate is through bystander killing, where activated T cells eliminate neighboring antigen-negative (Ag⁻) tumor cells. Although this phenomenon has been described, most existing models fail to capture the cellular architecture and complexity of human tumors. Patient-derived organoids (PDOs) retain these key features, making them a powerful system to explore bystander killing in a human context.

We engineered colorectal cancer PDOs to express a defined CAR target antigen and combined them at controlled Ag⁺:Ag⁻ ratios to generate mosaic tumors. CD8 CAR-T cells (CAR8) were co-cultured with these PDOs from three patients. We tracked tumor cell death by real-time caspase-3 imaging and endpoint viability assays, quantified cytokine release, and performed time-course bulk RNA-seq on sorted Ag⁻ cells. To functionally dissect the mechanism, we disrupted IFNGR1, TNFR1, and FAS individually or in combination using CRISPR-Cas9.

CAR8 cells showed strong antigen-specific killing, eliminating up to 90% of Ag⁺ tumor cells in a dose-dependent manner. In mosaic PDOs, Ag⁻ cell death followed after a ~10-hour delay, coinciding with high IFNγ and TNFα levels. RNA-seq revealed early induction of IFNγ- and TNFα-responsive genes in Ag⁻ cells, detectable within 4 hours. Knocking out IFNGR1, TNFR1, or FAS each partially reduced Ag⁻ killing, and triple disruption provided greater but incomplete protection, suggesting multiple converging cytokine pathways.

Our data show that CAR8-mediated bystander killing is driven by cytokine signaling through IFNγ, TNFα, and Fas, but also likely involves additional non-redundant factors. The PDO platform captures the spatial and molecular complexity of human tumors, enabling mechanistic insights that are difficult to obtain in conventional models.

Using patient-derived colorectal cancer organoids to model antigen heterogeneity, we show that CD8 CAR-T cells can eliminate Ag⁻ tumor cells through multifactorial cytokine-driven bystander killing. We now aim to identify other contributors to this process, including the role of epithelial architecture, to guide next-generation CAR-T therapies against heterogeneous solid tumors.

Resistance to anthracycline- and taxane-based chemotherapy remains a major hurdle in breast cancer treatment. Building on previous work linking FAN1 overexpression to residual disease in ER−/HER2− tumors treated with anthracycline/taxane combinations, we reanalyzed eight public datasets (1,148 patients) receiving anthracycline-based neoadjuvant chemotherapy. High FAN1 expression was associated with poor response only when taxanes were included, particularly in HER2+ and triple-negative breast cancers (TNBC), suggesting a role for FAN1 in dual replication and mitotic stress adaptation.

FAN1 function was investigated using genetic and biochemical approaches. FAN1 knockout and mutant cell lines lacking the ubiquitin-binding (UBZ) or nuclease (NUC) domains were generated. Cells were exposed to doxorubicin or paclitaxel to induce replication or mitotic stress. Proximity-dependent biotinylation (BioID) and mass spectrometry were performed to identify FAN1-interacting partners, followed by biochemical validation.

FAN1 knockout cells exhibited increased chromosomal instability and hypersensitivity to doxorubicin and paclitaxel. Complementation assays revealed that both the UBZ and NUC domains are essential for replication fork protection and mitotic stress tolerance. BioID and mass spectrometry identified ANLN, a cytokinesis-associated actin-binding protein, as a novel FAN1 interactor upon paclitaxel treatment. This FAN1–ANLN pathway appears to safeguard cells against paclitaxel-induced mitotic catastrophe.

Our findings identify FAN1 as a crucial regulator of genome stability and chemotherapeutic response. By preserving chromosomal integrity under replication and mitotic stress, FAN1 contributes to resistance against combined anthracycline and taxane treatment. The dual role of FAN1—mediating replication fork protection and mitotic resilience—highlights its importance as a molecular determinant of treatment outcome in breast cancer.

FAN1 expression represents a potential predictive biomarker of chemoresistance in breast cancer. Targeting FAN1 or its functional domains may offer a novel therapeutic approach to sensitize tumors to anthracycline/taxane-based chemotherapy and improve clinical response.

Tumor growth and metastasis depend on tissue remodeling, a process driven by extracellular matrix (ECM) degradation. Glycosylation plays a key role in regulating ECM remodeling, primarily through the activation of the GALA pathway which could relocate GalNAc Transferases (GALNTs) from the Golgi to the endoplasmic reticulum (ER) and then drive protein hyperglycosylation. One of these proteins is the ER resident protein Calnexin (Cnx) essential for proper protein folding and also vital to ECM degradation in cancer cells. The importance of Cnx in tumor progression is highlighted by findings that anti-Cnx antibodies inhibit liver tumor growth and lung metastasis in breast and liver cancers. Interestingly, antibodies targeting the intracellular C-terminal region of Cnx also block ECM degradation, suggesting that this portion of the protein is exposed on the cell surface. Thus, we hypothesize that O-glycosylation induces a topology shift in Calnexin, transitioning from an "I-type" topology to a "U-type" topology.

To explore the topological change of cell surface glycosylated Calnexin, and the mechanisms underlying this phenomenon, we attempted to detect the extracellularly exposed C-terminus of Calnexin through IF and WB experiments. HaloTag (HT) was inserted at the C-terminus of Calnexin as a marker for the Calnexin C-terminal (Cnx-HT), while HaloTag can be recognized by its specific HaloTag ligands (HL). The pancreatic ductal adenocarcinoma cell line, KPC, was used as a cellular model. GALA levels were controlled by expressing an ER-localized GALNT1 under the control of a doxycycline (DOX) inducible promoter (ER-G1) to simulate high and low GALA conditions.

We investigated the presence of ‘U-type’ topology of Calnexin by using C-terminal HaloTag and the cell impermeant ligand HL-Alexa Fluor 660. Through the results, we can clearly say that induction of GALA (ER-G1) increased signal from the cell-impermeant ligand, which was not observed in WT, non-induced cells or by expression of HaloTag alone, suggesting an increase in ‘U-type’ calnexin at the cell surface and a link between this topological change and O-glycosylation.

1. How does CHX lead to a significant decrease in CS Cnx levels? 2. Glycosylation might regulate the ribosome translocon complex (RTC)

Calnexin undergoes a significant change in its topology after O-glycosylation. We provide preliminary evidence that the topology change of glycosylated Calnexin occurs, with both the N-terminal and C-terminal domains localized outside the cell membrane. This surprising and unique topological change of glycosylated Cnx indicates a structural alteration that may be key to its escape to the cell surface.

Aberrant glycosylation drives tumour progression and metastasis. The GalNAc transferase activation (GALA) pathway, where O-GalNAc glycosyltransferases (GALNTs) relocate from Golgi to endoplasmic reticulum (ER), uniquely glycosylates ER-resident proteins. GALA is active in breast and liver cancers, contributing to ECM degradation, tumour growth, and invasion. This study investigates GALA in pancreatic ductal adenocarcinoma (PDAC), focusing on tumour growth and defining a GALA-driven glyco-signature for disease detection.

Pancreatic tumour tissues and cell lines were analysed for GALA activity. Human tumour microarrays were stained with Vicia villosa and Helix pomatia lectins to detect Tn antigen, a GALA hallmark. Genetic models included ER-localised GALNT1 (ER-G1), GALNT1-overexpressing (WT-G1), wild-type (WT), and GALA-inhibited (ER-2Lec) cells. Orthotopic injections of KPC47 cells evaluated tumour growth and metastasis. Glycoproteomic profiling using jacalin-agarose lectin weak affinity chromatography followed by data-independent acquisition mass spectrometry identified glycopeptides with T and Tn glycans. Eighteen patient-derived xenografts (PDXs) from the PaCaOmics cohort validated findings.

Pancreatic tumours showed elevated Tn levels with ER-like distribution versus normal tissue, indicating GALA activation. Tn localised primarily in CK19-positive epithelial cells, though some patients showed high stromal Tn expression, suggesting microenvironment heterogeneity. GALNT2 colocalised with ER marker GRP78, and calnexin was O-glycosylated, confirming GALA activity. ER-localised GALNT1 increased Tn levels more than overexpression alone, indicating ER localisation, not expression level, drives hyperglycosylation. GALA promoted ECM degradation in vitro, and ER-G1 cells enhanced tumour growth and metastasis in orthotopic models, while GALA-inhibited tumours failed to establish. Glycoproteome profiling identified a distinct signature driven by ER-localised GALNT1. Hyperglycosylation targeted ER-resident proteins, ECM components, and membrane proteins. GALA increased Tn glycopeptides and poly-T/poly-Tn structures, indicating clustered ER O-glycosylation. PDX tumours validated the ER-like glyco-signature.

This study highlights GALA's role in driving a unique pancreatic cancer glyco-signature. Hyperglycosylation of ER-resident, extracellular, and membrane proteins contributes to ECM degradation and tumour progression. Cleaved hyperglycosylated proteins could serve as serum biomarkers for PDAC detection.

Future work will assess this glyco-signature's cancer specificity and diagnostic utility in patient serum.

Growing evidence highlights the primary role of endothelial cells (ECs), fibroblasts, and stromal cells, in regulating the tumour microenvironment (TME) to control tumour growth and modulate immune responses. Furthermore, ECs are considered non-professional antigen-presenting cells due to the expression of adhesion molecules and immune checkpoint proteins, positioning them as critical intermediaries in cancer immunosurveillance.

In this study, we exposed human lung microvascular endothelial cells (HLMVECs) and human umbilical vein endothelial cells (HUVECs) to conditioned media (CM) derived from A549 (CM-A549) and Calu-3 (CM-Calu-3), two lung adenocarcinoma cell lines, to mimic the humoral cancer-endothelium crosstalk. Endothelial cell functional properties were evaluated through capillary-like tube formation and wound-healing migration assays. The phenotypic profile was characterized by flow cytometric analysis of key immunoregulatory molecules.

Our results demonstrated that both CM-A549 and CM-Calu-3 significantly enhanced the migratory capacity and promoted capillary-like structure formation on Matrigel® of both endothelial cell types, indicating pro-angiogenic effects. Notably, neither EC line expressed HLA-DR constitutively or following CM treatment, suggesting no antigen presentation capacity under these experimental conditions. HLMVECs exhibited increased PD-L1 and ICAM-1 expression following exposure to both CM, with CM-A549 inducing more pronounced PD-L1 upregulation. HUVECs exposed to CM-Calu-3 demonstrated significant upregulation of both PD-L1 and ICAM-1 proteins, while a negligible rise was observed following CM-A549 exposure, unveiling distinct responsiveness patterns among endothelial cell populations derived from anatomically distinct circulatory districts. CD275/ICOSL expression levels remained consistently low or undetectable across all conditions, indicating that this co-stimulatory pathway is not modulated by the tested tumour secretomes. Furthermore, the addition of 5µM ruxolitinib, a JAK1/2 inhibitor, effectively reduced PD-L1 and normalized ICAM-1 expression in both EC types across all experimental conditions. Similarly, 1µM niclosamide, a STAT3 inhibitor, showed potent inhibitory effects, particularly in suppressing CM-induced PD-L1 upregulation and completely preventing ICAM-1 modulation.

Our findings demonstrate that lung cancer secretomes significantly alter endothelial cell functional and immunoregulatory properties in a cell type-specific manner, promoting pro-angiogenic behaviours while inducing immunosuppressive phenotypes. The differential responsiveness between HUVEC and HLMVEC populations highlights the heterogeneity in cancer-endothelium interactions. The efficacy of both ruxolitinib and niclosamide documents the critical role of JAK/STAT signalling in reversing these tumour-induced alterations and suggests that their combination with targeted therapy drugs could represent a promising strategy to restore endothelial function and enhance the effectiveness of cancer immunotherapy.

These culture-based approaches provide valuable tools for exploring novel therapeutic compounds to advance current immunotherapeutic outcomes.

Breast cancer screening reveals many pre-neoplastic lesions, yet their risk of progression remains unpredictable due to limited understanding of early tumorigenesis. Cellular plasticity is increasingly considered as a key driver of tumor initiation, but its molecular mechanisms—studied mainly in transgenic mouse models—remain poorly defined. Lineage-tracing studies suggest that early lineage infidelity increases susceptibility to oncogenic transformation and contributes to intertumoral heterogeneity. However, the human mammary gland is more complex and displays greater diversity of pre-neoplastic lesions than murine tissue, underscoring the need for human-relevant systems. Currently, no experimental model faithfully captures how oncogenes reprogram the cell of origin during the earliest stages of human breast tumorigenesis, representing a major barrier and an opportunity for innovation

In this project, we address this gap by implementing the first lineage tracing strategy in human mammary epithelium. This system allows direct tracking of luminal cells and luminal progenitors (LCs)—the presumed cells of origin for multiple breast cancer subtypes—after activation of oncogenes such as PIK3CAH1047R, MYC, ERBB2, and H-RASG12V. By combining organoid culture with mouse mammary fat pad transplantation, we model oncogene-induced lineage plasticity, preneoplastic lesion formation, and tumor initiation both in vitro and in vivo.

We identified Mucin 4 (MUC4) as a faithful promoter to target LCs in the human mammary epithelium. Using CRISPR-knock in strategy, we developed a new lineage tracing system: MUC4:CreERT2. In the control condition, upon tamoxifen administration, reporter GFP is specifically expressed in LCs in 2D, 3D organoids and in vivo mouse mammary fat pad transplantation. In contrast, oncogene expression such as H-RASG12V induced lineage plasticity and tumor formation.

This new lineage-tracing model provides a powerful framework to dissect the molecular drivers of lineage plasticity and early transformation using single-cell multi-omics. The identified candidates will offer a valuable resource of targets to intercept tumor initiation. As a next step, CRISPR-based functional screens in organoids carrying this system will directly test their roles. Overall, this work is expected to reveal biomarkers for risk stratification of pre-malignant lesions and establish a unique platform for developing innovative strategies for breast cancer interception and prevention

This new lineage tracing system is an indispensable tool It will provide fundamental insights into the cellular and molecular mechanisms underlying malignant transformation of mammary epithelial cells, while identifying key regulators of oncogene-induced plasticity.

To sustain their growth, cancer cells often increase their demand for iron, which also makes them more vulnerable to ferroptosis—a form of iron-catalyzed cell death driven by lipid peroxidation. Drugs that target ferroptosis, such as sulfasalazine and sorafenib, have shown promise in enhancing the effectiveness of chemotherapy and immunotherapy, offering a potential strategy to eliminate therapy resistant cancer cells . However, lung cancers frequently activate pathways that enable them to evade ferroptosis. In particular, protective mechanisms like GPX4 and SLC7A11 play a critical role in shielding cancer cells from this form of cell death, supporting their survival. Developing new strategies to sensitize lung tumors to ferroptosis is therefore crucial for improving the efficacy of patient therapies. The Ubiquitin Proteasome System (UPS) is the primary pathway responsible for protein degradation, with E3 ubiquitin ligases recognizing specific target substrates. The hypothesis of my project starts off recent evidence suggesting that Cullin ligases are involved in regulating iron metabolism and oxidative stress, thereby influencing ferroptosis sensitivity . During my training, I identified the Cullin ligase complex CUL2-FEM1B as a novel regulator of ferroptosis in lung cancer by degrading BACH1, a transcription factor that maintains redox and iron homeostasis. In particular, BACH1 represses the expression of ferroptosis-protective genes such as SLC7A11, HMOX1, and FERRITIN, thereby sensitizing cells to ferroptosis. My goal is to investigate the CUL2-FEM1B – BACH1 axis as a therapeutic target to enhance the efficacy of ferroptosis inducers, offering new strategies to overcome lung cancer resistance.

Animal Experiments Transcriptomics Proteomics CRISPR, ShRNA, SiRNA Protein Purification and Pull Down Assay Immunohistochemistry Immunofluorescence and Stereomicroscopy Lentivirus- and Retrovirus- Mediated Gene Transfer Cell Viability Assay Lipid Peroxidation Measurement Quantitative PCR Cell Fractionation GSH Quantification

My findings are -CUL2 regulates ferroptosis in lung cancer heme-dependent degradation -FEM1B binds and degrades BACH1 via a heme-binding motif in BACH1’s C-Terminal region -The FEM1B-BACH1 pathway controls the expression of ferroptosis regulators -FEM1B knockdown sensitizes lung tumors to ferroptosis inducers in mouse

Heme binding enables a degron for direct E3 ligase recruitment and proteolysis. The CRL2FEM1B–BACH1 axis as a therapeutic target to elicit ferroptosis in lung cancer.

Our study identifies the CRL2FEM1B–BACH1 axis as a potential therapeutic target to enhance ferroptosis induction in vivo. These findings provide new opportunities to increase the effectiveness of ferroptosis-inducing treatments that, in combination with standard therapies, may help improve patient outcomes.

Melanoma is an aggressive skin cancer that metastasizes if not detected early, starting by the invasion of the lymph nodes. This lymphatic invasion is a critical step in melanoma progression, as it allows cancer cells to enter the bloodstream and spread to other organs. Understanding the mechanisms underlying lymph node invasion could lead to earlier and more effective interventions. During the pre-metastatic phase, lymph nodes are reprogrammed by factors secreted by melanoma cells in the skin, creating a niche favorable to tumor invasion and proliferation. During this phase, lymph node fibroblasts, known as Fibroblastic Reticular Cells (FRCs), are reprogrammed. In healthy lymph nodes, FRCs play a key role in organizing the structure of lymph nodes and in regulating T cell recruitment, survival, and activation. In many tissues, fibroblasts in the tumor microenvironment, also known as cancer-associated fibroblasts, are known to promote cancer progression, but little is known about the role of FRCs in the lymph node.

To mimic pre-metastatic reprogramming in the lymph node, healthy human FRCs are incubated with melanoma-secreted factors. The reprogrammed FRCs are then cocultured with T cells or tumor cells to assess dysregulated interactions using flow cytometry and real-time microscopy. In parallel, a murine model is used in which melanoma-secreted factors are injected into healthy mice. Draining lymph nodes are then collected and analyzed by flow cytometry to characterize stromal and immune remodelling.

We identified by RNAseq analysis that FRCs are transcriptionally reprogrammed by factors secreted by melanoma cells. Indeed, our previous work demonstrated that IL-1 secreted by dedifferentiated melanoma cells inhibit the contractility of reprogrammed FRCs, facilitating melanoma cell invasion. My research also reveals that reprogrammed FRCs enhance the proliferation of tumor cells, their motility and their resistance to targeted therapies used in the clinic. Additionally, these reprogrammed FRCs disrupt the anti-tumor immune response by altering T cell motility and upregulating immune checkpoint molecules (PD1, LAG3, and CTLA4) on T cells.

These results show the role of reprogrammed FRCs in the initiation of the pre-metastatic niche. Future studies will focus on deciphering the molecular mechanisms by which FRCs acquire their tumor-promoting phenotype and on identifying potential targets that could be used to prevent lymph node reprogramming.

These findings highlight the critical role of FRCs in creating a tumor-permissive microenvironment in early melanoma progression, and suggest new therapeutic approaches based on targeting the interactions between reprogrammed FRCs, tumor cells, and immune cells.

Breast cancer is the leading cause of cancer death in women worldwide. It is an heterogeneous disease with several molecular subtypes. The basal-like subtype has the poorest prognosis. It is characterized by increased expression of basal differentiation markers and genetic instability, frequently due to alterations in homologous recombination. Mutations in the BRCA1 gene are the best-known risk factor for developing a basal breast tumor. The basal subtype is also enriched in cancer stem cells. These cells appear to be involved in the early stages of carcinogenesis but also in resistance to cytotoxic treatment and relapse. Several signaling pathways influence their biology, notably the bone morphogenetic protein (BMP) pathway. Dysregulation of this pathway has been demonstrated in certain luminal tumors, but its involvement in the emergence of basal-like tumors remains to be explored.

Using primary samples and public databases, we searched for BMP pathway abnormalities in basal-like tumors and BRCA1-mutated predisposed tissues. We also used the MCF10A human mammary epithelial stem cell line to model in vitro stem cell tumor initiation processes in the mammary gland.

We show that BMPR1A receptor and BMP4 ligand expression are deregulated in basal-like tumors and predisposed tissues. In addition, we demonstrate an inverse correlation between the level of BMPR1A protein and BRCA1 mRNA in BRCA1 WT basal breast tumors from patients. Using a human mammary stem cell model, we show that BMP4 signals through BMPR1A to repress the BRCA1 gene transcription. The functional consequences of this BRCA1 repression include preferential differentiation toward the basal phenotype, increased stemness and alterations in the homologous recombination pathway associated with an increased sensitivity to PARP inhibitors, genomic instability and transformation.

The regulation by BMP4/BMPR1A signaling of BRCA1 expression and consequently of homologous recombination could be a new mechanism of BRCAness, i.e. any situation mimicking a BRCA1 mutation. Further studies are required to understand the function of the BMP4/BMPR1A/BRCA1 axis in the physiology of the normal mammary gland as well as the precise molecular mechanisms that could favor basal breast cancer formation when BMP4/BMPR1A signaling is altered. Identifying BRCA1 WT patients with basal breast cancer showing a BRCAness phenotype due to altered BMP4/BMPR1A could make them eligible to PARPi therapies.

We suggest a role for the BMP4-BMPR1A axis in the early stages of basal-like breast tumor carcinogenesis, through a phenocopy of BRCA1 mutations inducing basal differentiation of stem cells and promoting the genetic instability necessary for their transformation.

Colorectal cancer (CRC) represents the 2nd deadliest cancer in the world. There is therefore a real necessity to identify new biomarkers to promote the development of more efficient therapies. The Wnt/β-catenin signaling pathway has a major role in CRC due to frequent dysregulations. PTK7 (Protein Tyrosine Kinase 7), a transmembrane receptor required during embryo development and involved in Wnt/β-catenin and Planar Cell Polarity (PCP) pathways, has been identified both as a poor prognosis marker and a therapeutic target in CRC . As a cell surface protein, PTK7 oncogenic properties have mainly been associated with its membrane localization. Indeed, cell surface PTK7 has already been characterized to act as a positive activator of Wnt pathways by interacting with Wnt ligands and membrane components, such as LRPs and Frizzled. Recently, other studies described PTK7 membrane activity through its interaction with key surface receptors, including EGFR and FGFR1, and interfering with their signaling. Despite their surface location, some tyrosine kinase receptors have been described to translocate into the nucleus, and to interact with transcription factors for regulating gene expression and promoting tumor progression in various cancers. Here, using cell fractionation followed by western blot and mass spectrometry analyses, we have evidence for PTK7 basal localization in both soluble nuclear and chromatin bound fractions in colon cancer cell lines. Additionally, mass spectrometry data suggest HMGA1, a transcription factor positively regulating Wnt/β-catenin pathway by initiating LRP5 transcription, as a potential PTK7 nuclear interactor. Finally, thanks to bulk/single-cell RNAseq analyses, we propose a model in which nuclear PTK7 interacts with HMGA1 at the LRP5 promotor region in order to potentially enhance the Wnt/β-catenin signaling in CRC.

We used cell fractionation followed by western blot to detect PTK7 in CRC cell lines nuclear compartment. Mass spectrometry/BioID was used to identify PTK7 nuclear interactors including transport proteins such as XPO1, KPNB1, KPNA2, transcription factor such as HMGA1. To confirm the implication of those proteins in PTK7 nuclear transport, we used inhibitors selectively directed against KPNB1 and XPO1 activities. To evaluate the correlation of co-expression between LRP5, HMGA1 and PTK7, we analyzed bulk/single cell RNA seq data extracted from publically available sequenced CRC tumors. Then, we additionally confirmed the implication of both PTK7 and HMGA1 in the positive regulation of LRP5 expression in CRC cell lines by using si-RNA strategies followed by Western-Blot and qRT-PCR analysis.

The Wnt/β-catenin and planar cell polarity (PCP) signaling pathways are involved in various physiological processes and are often dysregulated in cancer. Within the Wnt pathway, the receptor tyrosine kinase (RTK) PTK7 (Protein Tyrosine Kinase 7) is associated with metastasic progression and reduced survival in patients with colorectal cancer (CRC). The receptor is now considered both a poor diagnostic marker and as a promising therapeutic target in CRC. Although PTK7 lacks catalytic activity, it ensures key biological functions through protein-protein interactions. At the cell surface, its pseudokinase domain acts as a scaffold for cytoplasmic partners such as β-catenin or Rack1, while its extracellular domain (ECD) serves as a co-receptor for proteins including Ror2 or Vangl1/2. The structure of the pseudokinase domain has recently been solved by crystallography. In contrast, the structure of the PTK7 ECD is currently only predicted by AlphaFold. The elucidation of this structure will be crucial to understand how PTK7 engages its various co-receptors. The PTK7 ECD consists in seven immunoglobulin-like domains that undergo alternative splicing events, giving rise to different isoforms whose specific functions remain unknown. Several structures of PTK7 co-receptors have recently been elucidated by crystallography or cryo-electron microscopy (cryo-EM). This includes EphA2, an active RTK involved in cancers and identified by the J.-P. Borg team as a partner of the PTK7 ECD. PTK7 negatively regulates EphA2 activity and modulates its oligomerization state.

The strategy consists in producing a soluble form of the PTK7 ECD, fused to a Twin-Strep-Tag, in human HEK293T cells, followed by purification on Strep-Tactin resin. The successful purification of the recombinant protein was confirmed by SDS-PAGE and mass spectrometry, allowing us to perform initial cryo-EM assays. The resulting medium-resolution structure of the PTK7 ECD suggests a trimeric organization, consistent with Blue Native PAGE data and with an AlphaFold3 model. In parallel, co-immunoprecipitation experiments allowed us to map the PTK7:EphA2 complex, identifying the EphA2 ligand-binding domain (LBD) and Cystein-rich domain as key regions involved in the interaction. Now, the objective is to characterize, using the same workflow, the structures of the different PTK7 ECD isoforms and subsequently the structure of the PTK7:EphA2 complex. Additionally, in order to validate the formation of this complex and complete the mapping of the interaction interface, we will conduce biophysical assays such as Bio-Layer Interferometry.

Matrix metalloproteinases (MMP) play a crucial role in the remodelling of the extracellular matrix (ECM) during cancer progression. The membrane-type 1 MMP (MT1-MMP) in particular, is thereby strongly associated with poor prognosis. MT1-MMP localizes at invadopodia, specialized actin-rich structures developed by cancer cells to degrade the ECM, and at the surface of secreted small extracellular vesicles (sEV). The role of sEV-associated MT1-MMP in ECM degradation and the mechanisms supporting MT1-MMP loading into sEV are unknown. We previously established that the PDZ protein syntenin and the syndecan heparan sulfate proteoglycans (SDC) syntenin associates with, are major players controlling sEV biogenesis.

The interaction between syntenin and MT1-MMP was assessed by surface plasmon resonance using the BIAcore technology.Their localization was analyzed by immunofluorescence in MDA-MB-231 triple negative breast cancer (TNBC) cells. MT1-MMP surface levels and endocytosis were examined by biotinylation of cell surface proteins following syntenin or SDC depletion using siRNA. In addition, sEV collected by sequential ultracentrifugation of conditioned media from MDA-MB-231 cells depleted of syntenin or SDC were used to evaluate the impact of the syntenin-SDC pathway on MT1-MMP sorting into sEV. Finally, the effect of syntenin depletion on the matrix degrading activity of MDA-MB-231 cells and their immunocaptured sEV was determined using a gelatin degradation assay.

In present study, we demonstrate that ECM degradation via invadopodia depends on syntenin. We also show that syntenin colocalizes with MT1-MMP in late endosomes, the cellular compartment from which sEV in part originate, and that syntenin directly interacts with MT1-MMP. Interestingly, this interaction does not involve the PDZ-binding motif, but a PRR sequence in MT1-MMP. Moreover, we demonstrate that syntenin and SDC are essential determinants of the sorting of MT1-MMP to sEV. Additionally, sEV contribute to ECM degradation through MT1-MMP activity, a process dependent on syntenin.

These findings provide evidence that syntenin-SDC-MT1-MMP complexes orchestrate ECM degradation in a TNBC model and pave the way for innovative rational approaches to control cancer cell invasion.

PTK7 is a pseudo-tyrosine kinase involved in cell adhesion, polarity, and migration. Its overexpression is associated with tumor progression, metastatic dissemination and poor prognosis in multiple cancers, including hepatocellular carcinoma, melanoma, and triple-negative breast cancer. A PTK7-targeting antibody–drug conjugate has demonstrated therapeutic efficacy by inducing durable tumor regression (Damelin et al., Sci. Transl. Med., 2017). This study also highlighted PTK7 expression on plasmacytoid dendritic cells (pDCs) in both blood and non-small cell lung carcinoma samples, leading our team to investigate the role of PTK7 in dendritic cells (DCs). We have shown that PTK7 is physiologically expressed mainly by Langerhans cells (LCs), while other DC subsets, particularly cDC2 can express it within B16F10 melanoma tumors, suggesting regulation by the tumor microenvironment. Moreover, our in vitro data indicate that PTK7⁺ DCs display impaired T-cell priming ability, characterized by reduced proliferation of CD4⁺ and CD8⁺ T cells and increased differentiation of regulatory T cells (Jaeger, PhD thesis 2024; Jaeger et al.). These findings suggest that PTK7 expression may contribute to the establishment of an immunosuppressive microenvironment and to the induction of immune tolerance. Given the central role of DCs in activating antitumor T-cell responses, it is plausible that cancers exploit escape mechanisms by altering DC functionality. Previous studies have described tumor-infiltrating DCs capable of supporting tumor progression (Hanks et al., 2013; Scarlett et al., 2012), although the underlying mechanisms remain poorly defined. In light of this and based on our preliminary observations, we hypothesize that PTK7 expression in DCs is modulated by the tumor microenvironment, thereby promoting an immunosuppressive state conducive to cancer progression. To address this hypothesis, our project is structured around three main objectives: 1. Evaluate in vivo the impact of PTK7 deficiency on DC function in physiopathological contexts (cancer, aging). 2. Identify inflammatory signals in the microenvironment regulating PTK7 expression in DCs. 3. Define the PTK7 interactome and associated signaling pathways in DCs under inflammatory conditions.

Hepatocellular carcinoma (HCC) is the most common primary liver tumour and represents a major public health concern. Its incidence continues to rise, notably with the increasing prevalence of metabolic-associated steatohepatitis. HCC is characterized by pronounced heterogeneity, both inter- and intra-tumoural, which poses significant challenges for the development of effective therapeutic strategies. Protein Tyrosine Kinase 7 (PTK7) is a pseudo-tyrosine kinase whose functions have been extensively studied in several biological contexts, notably by our team. In cancer, PTK7 has been associated with therapeutic resistance, metastatic progression, and overall poor prognosis. Despite its well-established roles in other cancers, limited data are available regarding PTK7 involvement in HCC. The few existing studies suggest that PTK7 overexpression may correlate with worse clinical outcomes in HCC and the activation of pro-metastatic pathways. In this project, we aim to elucidate the role of PTK7 in HCC tumorigenesis, disease progression, and therapeutic resistance.

This study is structured around four major axes. First, we analyzed public datasets to characterize the status of PTK7 across liver disease and cancer. Second, in relation to pathways relevant to PTK7 functions, we employed hydrodynamic tail-vein injection (HTVi)-based mouse models to reconstitute these signaling circuits in vivo and evaluate their capacity to trigger and drive liver cancer. Third, we will apply state-of-the-art -omics approaches in order to better elucidate the mechanism by which PTK7 induces tumoral growth and treatment resistance. Finally, we will assess the therapeutic efficacy of anti-PTK7 CAR-T cells we have developed as therapeutic agents for HCC treatments.

While PTK7 is poorly expressed in HCC compared to other malignancies, its expression is greatly increased in damaged liver, and correlates with the level of fibrosis in patients. This increased expression spikes in cirrhotic and liver tumours, where elevated PTK7 levels are found in distinct patient clusters, distinguished by specific gene expression patterns, pathway enrichments, and prognosis. In mice, we found that PTK7 triggers HCC formation and progression by cooperating with other oncogenic signals, such as MYC and WNT/β-Catenin. Intriguingly, these PTK7-driven tumours are composed of distinct cellular subtypes, whose characterization is currently ongoing.

This study aims to decipher the mechanism by which PTK7 induces HCC development, and support treatment resistance. Using our expertise, study the ability of anti-PTK7 CAR Tcells to induce tumoral regression, that could help establish a robust preclinical framework and potentially support the stratification of patients based on PTK7-associated co-alterations, ultimately paving the way towards more personalized therapeutic strategies.

Triple-negative breast cancer (TNBC) is an aggressive and heterogeneous disease in which immune checkpoint inhibitors (ICI) provide inconsistent benefit. Understanding how tumor-immune interactions shape therapeutic responses is crucial for developing treatment strategies. We hypothesized that the MMTV-R26Met spontaneous TNBC model could help uncover how immune microenvironmental features contribute to differential sensitivity to chemo/immunotherapy.

We performed histological analyses, multi-omics profiling (transcriptomics, genomics, proteomics), and immune characterization of spontaneous MMTV-R26Met tumors, and established syngeneic grafts derived from primary tumors. Results were compared with patient datasets and tissue microarrays from human TNBC cohorts. We evaluated the efficacy of epirubicin and anti-PD-1 in MMTV-R26Met syngeneic models. We explored mechanisms driving the recruitment of specific immune cell populations by analyzing cytokine and chemokine expression profiles in cell lines established from MMTV-R26Met tumors, while immune infiltrates were characterized in tumors arising from orthotopic grafts of these cell lines or tumoroids.

Multi-parametric analysis identified four molecular-immune clusters, each characterized by specific immune features: lymphocyte-, macrophage-, neutrophil-, or dendritic cell-enriched. These immune infiltrates were conserved across serial syngeneic transplantations, indicating that tumor-intrinsic properties dictate the recruitment of specific immune subsets. This classification aligned with immune signatures found in TNBC patients. In vitro cell and tumoroid models established from MMTV-R26Met primary tumors maintained stable inflammatory identities, reflected by cytokine and chemokine expression patterns characteristic of their dominant immune subtype. Their orthotopic engraftments recapitulated the same immune infiltrates as the original tumors, demonstrating that they intrinsically encode signals capable of driving selective immune recruitment in vivo. Focusing on neutrophil- and macrophage-enriched tumors, immunosuppressive subtypes associated with poor prognosis, we found marked differences in sensitivity to ICI, alone or combined with epirubicin. Specifically, epirubicin induced a myeloid subtype-specific remodeling of the immune microenvironment, with variable ability to convert immunologically “cold” tumors into “hot” ones.

Our findings show that the immune microenvironment is a key determinant of therapeutic outcome in TNBC. The dominant myeloid context - neutrophil- or macrophage-enriched - shapes chemotherapy efficacy and its ability to potentiate ICI responses. This myeloid-driven heterogeneity provides a mechanistic basis for interpatient variability and supports immune-informed therapeutic stratification. Ongoing work aims to map the temporal dynamics of myeloid infiltration to understand how shifts in immune states influence treatment sensitivity.

The bone marrow (BM) niche is a complex and dynamic microenvironment composed of diverse cell types, including stromal, endothelial, immune, and hematopoietic cells. Dysregulation of the interactions occurring between cells influence malignant transformation. Existing in vitro and in vivo models incompletely recapitulate human BM physiology, due to species-specific differences, limited immune representation, and poor preservation of cellular heterogeneity. To overcome these limitations, we developed a standardized three-dimensional human bone marrow model (3D-BOM) capable of mimicking key structural and functional features of the BM niche and we assessed its capacity to reproduce physiological and pathological interactions, including those occurring in acute myeloid leukemia (AML).

The 3D-BOM system was generated culturing HMEC-1 endothelial cells and mesenchymal stromal cells (HS27A or primary MSCs) with biphasic calcium phosphate beads in osteogenic medium for 3 weeks allowing self-assembly of the 3D structure. THP-1 monocytic cell line, primary CD14+ monocytes, hematopoietic progenitors cell line or primary HSPCs from normal or AML samples were subsequently introduced. Flow cytometry, ELISA, tubules formation assays, long-term serial 3D cultures, and single-cell RNA sequencing were performed after enzymatic dissociation of the disk.

Single-cell transcriptomics revealed stromal heterogeneity comparable to native human BM, identifying MSCs, osteochondral progenitors, osteoblasts, and chondrocyte progenitors, along with sinusoidal and type S endothelial cells. Long-term culture preserved endothelial functionality, while endothelial cells grown in 3D with AML progenitors exhibited enhanced tubule formation, mirroring the increased angiogenesis observed in AML-BM. In addition, HSPCs maintained stemness features across serial 3D culture. THP-1 and iPSC-derived macrophages spontaneously adopted pro-inflammatory features within the 3D-BOM. Importantly, THP-1 and primary CD14+ cultured with AML progenitors acquired anti-inflammatory phenotype such as increased IL-10 secretion and CD209 expression.

3D-BOM faithfully recreates key features of the human BM niche, including stromal and endothelial heterogeneity, long-term maintenance of hematopoietic progenitors and endothelial cells functions, and supporting functional immune activation. Its ability to capture niche remodeling in AML context highlights its relevance as a human physiopathological tool for dissecting mechanisms of leukemic transformation and for preclinical testing of therapies targeting both AML cells and their microenvironment, while reducing reliance on animal models.

The 3D-BOM functions as a self-organizing human BM niche, supporting stromal, endothelial, immune, and hematopoietic dynamics, and reproducing AML-driven niche alterations observed in patients BM.

Pancreatic ductal adenocarcinoma (PDAC) remains among the most lethal malignancies, with a 5-year overall survival of ~12%, largely due to late diagnosis, limited resectability, and resistance to standard therapies. Over 90% of PDACs harbor oncogenic KRAS mutations, which arise early and are key drivers of tumour initiation and maintenance. Recent studies reveal that quantitative differences in KRAS mutant dosage through allelic gain, loss of wild-type allele or copy-number imbalance may amplify oncogenic signaling and reshape tumour cell behaviour, potentially favouring evolution toward more aggressive and therapy-resistant states.

We hypothesised that comparing KRAS-balanced versus KRAS-imbalanced PDAC models would reveal evolutionary and adaptive mechanisms underlying tumour aggressiveness and treatment resistance, beyond subtype association. Specifically, we aimed to determine whether allelic imbalance generates distinct metabolic and transcriptional programs. KRAS allelic imbalance was quantified using droplet digital PCR (ddPCR) across 333 preclinical PDAC models derived from 219 patients, encompassing patient-derived xenografts (PDXs, n=125), primary cell cultures (PDCs, n=51), and organoids (PDOs, n=157). Transcriptomic subtyping was performed using bulk RNA-seq. To specifically assess the biological effect of allelic dosage, differential expression analyses were conducted within tumours models. Gene Set Enrichment Analysis (GSEA), along with pathway enrichment, was employed to explore metabolic distinctions.

KRAS allelic imbalance was conserved across model types from the same patient, indicating clonal stability. Imbalance was significantly enriched in basal-like PDX (p=0.00037), but intriguingly, we also identified a subset of classical PDAC models exhibiting KRAS imbalance. Within these classical models, GSEA revealed significant up-regulation lipid transport (p-adj.=0.009), and mitotic spindle (NES=1.8 , p-adj.=0.003) while pro-inflammatory pathways (N=-1.78, p-adj.=0.007) are down-regulated. Complementary pathway analyses reinforced enrichment in lipoprotein particle organisation, acylglycerol metabolism, and lipid remodeling, independently of basal-like markers.

These results suggest that KRAS allelic imbalance may act as an evolutionary driver shaping metabolic rewiring beyond canonical subtype boundaries. The enrichment of lipid uptake and fatty acid metabolic programs in KRAS-imbalanced classical tumours suggests subtype-transcending metabolic reprogramming. This highlights the need for therapeutic strategies tailored not only to transcriptomic subtype but also to specific metabolic vulnerabilities.

KRAS dosage should therefore be considered as a complementary stratification layer to transcriptomic subtype, and metabolic rewiring particularly lipid-dependent programs may represent exploitable vulnerabilities in KRAS-imbalanced classical PDAC.

Several studies on breast cancer have shown a link between activation of the WNT/Planar Cell Polarity (WNT/PCP) pathway, tumor development, and resistance to treatment. CELSR2 is a receptor in the WNT/PCP pathway that belongs to the family of G protein-coupled receptors involved in cell adhesion. This receptor has a very large extracellular N-terminal region that includes a proteolytic site where the receptor is cleaved in an autoproteolytic manner, leading to its activation. We have demonstrated, through two transcriptomic analyses, that CELSR2 overexpression is associated with reduced metastasis-free survival and decreased response to treatment in triple negative breast cancer (TNBC), a subtype of breast cancer. TNBC, which affects about 15% of patients, is the most aggressive subtype due to its strong propensity to develop metastases. Moreover, lacking hormone receptors and HER2, TNBC does not benefit from hormone therapy or Trastuzumab. It is therefore essential to identify new therapeutic targets to optimize its treatment and prevent metastatic development.

To investigate how CELSR2 may contribute to TNBC progression, we examined its role in tumor growth (in vitro, in vivo), in cell migration and epithelial–mesenchymal transition (EMT). We also sought to elucidate the molecular mechanisms through which CELSR2 promotes tumor cell aggressiveness.

Here, we experimentally demonstrate that CELSR2 promotes cancer cell proliferation by contributing to tumor growth in orthotopic xenografts of TNBC cell lines implanted in immunodeficient mice, through activation of the cAMP/PKA/p-CREB signaling pathway. We further show that, via this pathway, the transcription factor p-CREB induces the expression of genes involved in oxidative phosphorylation, which may underlie the enhanced proliferation of primary tumors overexpressing CELSR2. In contrast, CELSR2 does not appear to play a role in cell migration. However, CELSR2 is highly expressed in epithelial cells and is downregulated during EMT, becoming less expressed in mesenchymal cells.

Taken together, these results suggest that CELSR2 overexpression in TNBC patients is associated with tumor growth through enhanced oxidative phosphorylation, and its underexpression may facilitate tumor cell detachment from the primary tumor. The receptor is no longer expressed in circulating tumor cells but is re-expressed in metastatic lesions, to promote tumor growth. It is therefore important to further investigate the potential of CELSR2 as a predictive biomarker of patient survival and/or treatment response in TNBC, by using of blood-based assays to monitor soluble CELSR2 in pre-neoplastic lesions or to track residual disease.

Protein phosphorylation is one of the most common and important post-translational modification. This reversible mechanism drives various cellular processes such as cell-cycle regulation, intracellular signaling, cell proliferation and metabolic regulation. Thus aberrant phosphorylation-mediated signaling networks can contribute to cancer progression and aggressiveness. This is the reason why phoshoproteome studies became an important field to make the research going further.

To overcome the challenges of low stoichiometry of phosphoproteins compared to non-phosphoproteins and peptide loss during enrichment, we use an automated sample preparation with the AssayMap Bravo from Agilent. Our study aimes at determining the capacity of Fe(III)-NTA cartridges to enrich phosphopeptides with a dilution series of breast cancer line (SKBR3) lysates quantity from microgram to milligram.

The Fe(III)-NTA cartridges effectively captured phosphorylated peptides at input quantities ranging from 50µg to 500µg, with a reduction in efficiency observed beyond 500µg due to cartridges saturation. The results showed robust performance with high sensitivity and reproductibility, enabling the identification of thousands of phosphopeptides even with low starting material.

Future optimization could futher enhance efficiency, particulary for low-abondance samples.

These advancements hold promise for oncology research.

Protein Tyrosine Kinase 7 (PTK7) is a Wnt pathway co-receptor whose overexpression is associated with poor prognosis in several cancers, making it an attractive therapeutic target. Although PTK7 expression has recently been reported in dendritic cells (DCs), its functional role in these cells remains unknown.

Here by using multiparametric flow cytometry, bulk RNA-seq analysis and functional experiments in vitro and in vivo, we provide the first comprehensive study of PTK7 expression and function in dendritic cells.

At steady state, PTK7 expression was restricted to a subset of DCs: the Langerhans cells, in the skin and cutaneous lymph nodes (CLNs). PTK7⁺DCs displayed increased CCR7 expression and were enriched among recently migrated DCs in CLNs in a skin sensitization model. Functionally, PTK7⁺ DCs showed a reduced capacity to activate CD4⁺ and CD8⁺ T cells and to induce their proliferation, while promoting regulatory T cell differentiation. In inflammatory conditions, particularly in melanoma and hepatocellular carcinoma models or during aging, PTK7 expression was acquired by additional DC subsets, potentially including recently described mreg DCs, suggesting modulation by the tumor microenvironment. Using a DC-specific PTK7-deficient mouse model, we further observed that PTK7 may contributes to immune regulation, as its deletion leads to increased inflammation in aged mice. RNA-seq analyses revealed altered expression of antigen presentation-related genes, providing mechanistic insight.

Altogether, our data identify PTK7 as a novel marker of regulatory dendritic cells involved in immune homeostasis and cancer-associated inflammation.

O-glycosylation is a post-translational modification catalyzed by N-acetylgalactosaminyltransferases (GALNTs) in the Golgi apparatus, where a glycan is added to proteins to regulate their maturation and function. In some cancers, GALNTs are aberrantly translocated from the Golgi to the endoplasmic reticulum (ER), leading to enhanced glycosylation of newly synthesized proteins. This shift alters the global glycosylation landscape and is known as the GalNAc transferase activation (GALA) pathway. Reported in several solid tumors, including liver cancer, GALA-driven hyperglycosylation can profoundly impact protein secretion, cellular signaling, and interactions within the tumor microenvironment.

This project investigates how increased O-glycosylation modulates protein secretion in liver cancer cells. To this end, secretome analysis was performed using two HUH7-derived cell lines differing in GALA activity. The first, HUH7 ERG1-GFP, expresses an inducible ER-targeted GALNT enzyme that elevates O-glycosylation upon induction, modeling GALA activation. The second, HUH7-GFP, serves as a control with low GALA activity. Mass spectrometry-based proteomic analysis was carried out on the secreted proteins from both cell lines to determine how O-glycosylation influences secretion profiles.

Analysis of five biological replicates revealed that GALA alters the secretion of nearly 400 proteins, with both increases and decreases observed. Notably, angiotensin-converting enzyme (ACE) was highly and exclusively secreted by high-GALA cells, suggesting that hyperglycosylation can selectively enhance the secretion of specific proteins. Such differential secretion may result from increased shedding of membrane-associated or surface receptors, potentially enhancing cell–cell communication and contributing to tumor progression. To understand these phenomena further, we propose that hyperglycosylation modifies the conformation or trafficking behavior of proteins, thereby influencing their secretion or degradation. Excessive glycosylation could target some proteins for degradation while stabilizing or enhancing the release of others, reflecting a fine balance between secretion and turnover. Interestingly, several key regulators of secretion and intracellular trafficking also displayed hyperglycosylation, indicating that GALA might broadly disrupt the cellular secretory machinery.

Future investigations will aim to manipulate glycosylation patterns to delineate their effects on protein stability, secretion, and degradation. Understanding how GALA-mediated hyperglycosylation shapes the secretome will provide new insights into tumor biology and may uncover novel targets for modulating cell–cell signaling and microenvironment interactions in liver cancer.

Pancreatic ductal carcinoma (PDAC) is one of the most aggressive and lethal malignancies, largely due to its late diagnosis and resistance to chemotherapy. Nuclear protein 1 (NUPR1), a stress-inducible transcription factor, is overexpressed in many cancers and associated with poor prognosis, making it a promising therapeutic target. We have developed a NUPR1 inhibitor, ZZW-115, which demonstrates strong anticancer effects both in vitro and in vivo. Recent studies, including our own, have shown that ZZW-115 induces ferroptosis in a mitochondria-dependent manner, ferroptosis is an iron- and lipid-dependent cell death that is particularly effective against cancer cells resistant to conventional therapies. Given that dysregulated iron metabolism has been described as marker of poor prognosis PDAC, we investigated the role of iron in tumor progression and in ZZW-115-mediated cell death.

Experiments were conducted using the pancreatic cancer cell line MiaPaCa-2. Functional studies for iron homeostasis and lysosomal activity were performed using microscopy, flow cytometry, and Western blotting measuring various parameters. Mitochondrial respiration was assessed using Seahorse XF technology.

We found that, unexpectedly, iron supplementation tends to attenuate ZZW-115-induced cell death. Mechanistically, NUPR1 inhibition alters lysosomal function and induces the accumulation of ferritin, the iron storage protein, within lysosomes, independently of cellular iron level. This leads to iron imbalance within the cells, and ultimately affects mitochondrial functions which mainly rely on iron level, ultimately leading to cell death. Importantly, iron supplementation might fuel mitochondria and rescue mitochondrial damages caused by ZZW-115.

Targeting NUPR1 emerge as a novel strategy to disrupt both iron homeostasis and autophagy, and offering a promising strategy to modulate metabolism and treat PDAC.

In summary, our study highlights a critical connection between NUPR1, lysosomal activity, and the regulation of iron homeostasis in PDAC. We demonstrate that intracellular iron level plays a protective role against damages resulting from NUPR1 inhibition. Our results uncover an exploitable iron-dependent metabolic vulnerability in PDAC that could be targeted to improve therapeutic outcomes

Glutamylation is a post-translational modification which adds glutamate side chains onto the γ-carboxyl groups of glutamic acid residues in the primary sequence of target proteins. The glutamate required for this mechanism can be produced by glutaminolysis, the conversion of glutamine to glutamate by glutaminase enzyme (GLS), feeding biosynthesis and energy production which has emerged as a key metabolic route in mechano-dependent diseases such as breast cancer and pulmonary hypertension. Although discovered in the 1990s, glutamylation has mainly been reported as a post-translational modification of tubulin. Yet, glutamylation is not restricted to tubulin, and whether mechano-induced glutamine catabolism drives protein glutamylation to promote cancer cell aggressiveness remains unknown.

Using mass spectrometry, we reported that matrix stiffening may induce the glutamylation of several metabolic enzymes, such as GLS. We first determined the precise glutamylated sites on GLS and then identify enzymes mediating the GLS glutamylation by performing a siRNA screening of tubulin tyrosine ligase-like (TTLL) enzymes, which adds glutamate on the protein, and cytosolic carboxypeptidase (CCP) enzymes which remove them.

We demonstrated that knockdown of TTLL5 or TTLL9 significantly reduce GLS glutamylation, whereas knockdown of CCP4, CCP5, or CCP6 increased it. In addition, we showed that mutations preventing GLS glutamylation impair its function, reduce glutamine catabolism and blunt cancer cell’s aggressiveness.

The high metabolic adaptability of pancreatic ductal adenocarcinoma (PDAC) is a key factor that drives tumor evolution in disseminating cells throughout the metastatic cascade and contributes to successful colonization of the liver. We hypothesize that metastatic tumor cells modify their metabolism to establish a bi-directional pro-tumoral crosstalk with resident liver cells, especially with hepatocytes, the most abundant and metabolically active cell population in the liver. We theorize that this intercellular metabolite exchange between hepatocytes and tumor cells promotes tumor cell survival in the new environment and facilitates metastatic expansion.

To explore the global metabolic alterations in liver metastases compared to primary PDAC tissue we performed transcriptomic analysis on bulk RNA extracted from liver metastases and from the primary tumor of a murine PDAC model. This approach was coupled with integration of RNAseq and semi-targeted metabolomics in tumor cells extracted from liver metastases or from primary PDAC and cultivated in vitro to identify alterations specific to tumor cells. To simulate the hepatocyte – tumor cell metabolic crosstalk, we used an in vitro co-culture model of medium exchange, accompanied by proteomic analysis and steady state metabolic tracing. Key elements of the identified metabolic pathway were finally genetically and pharmacologically inhibited in functional assays in vitro to better discern the metabolite exchange between hepatocytes and tumor cells.

Using our high-throughput integration analysis alongside the tissue transcriptomics we identified that tyrosine metabolism is a major metabolic dependency of metastatic cells in the liver. More specifically, we found that hepatocytes convert phenylalanine to tyrosine, which is then secreted and absorbed by tumor cells. Tumor cells then break down tyrosine into simple compounds like fumarate which feeds the TCA cycle promoting metastatic cell growth and ultimately metastatic expansion. Pharmacological inhibition or genetic ablation of the tyrosine degradation enzyme HPD in metastatic tumor cells hinders their tumorigenic and metastatic potential.

These results demonstrate the cooperating action between hepatocyte and tumor cell metabolism in the liver: Tumor cells re-wire hepatocytes to produce more tyrosine which tumor cells use to promote cell growth. Interestingly, high expression of enzymes involved in tyrosine metabolism correlates with worse survival of PDAC patients. Pharmacological targeting of tyrosine metabolism may be an attractive approach to improve patient outcomes.

The interface between normal and transformed epithelial cells at the tumor edge remains poorly understood. In solid tissues, cancer cells outcompete normal cells for space, but the mechanisms behind this process remain unknown. Cell competition, observed in multicellular organisms, is thought to eliminate defective cells and may act as an "epithelial defense against cancer". However, cancer cells function as "super-competitors," allowing them to form tumor nodules. The molecular processes underlying cell competition implicate various known pathways and oncogenes, such as the Hippo and/or Wnt pathways, the Myc, Src and EGF-R oncogenes and others. Cell surface proteins are in their vast majority glycoproteins, carrying various N- and O-glycans. Cancer is associated with massive changes in glycosylation. Cell surface protein glycosylation occurs in the secretory pathway, a complex and compartmentalized system composed of two main organelles: the ER and the Golgi apparatus. GALNTs, that normally localised in the Golgi, are relocated to the ER upon activation of signaling molecules such as the Src and EGF-R kinases. This relocation leads to increased O-glycosylation and higher cellular Tn levels. We nicknamed GALA pathway, this highly regulated activation of O-glycosylation.

We discovered that when the GALA pathway is activated, tumor growth accelerates as compared to tumors with lower GALA levels. In contrast, decreasing both GALNT1 and 2 greatly reduces the risk of developing liver cancers. In contrast, activating the GALA pathway across the liver protected it from tumor growth. As a result, GALA greatly boosts growth when activated only in tumor cells but inhibits tumor development when expressed in surrounding, non-transformed cells. These findings suggest that GALA promotes a type of cell competition required for tumor growth: high GALA cells outcompete low GALA cells. To examine the nature of this competition, we attempted to duplicate it in an in vitro environment. We began coculture of high and low GALA cells to see if low GALA cells exhibit an enhanced rate of apoptosis when cocultured with high GALA cells, as opposed to situations when they are cocultured with cells with equivalent levels of GALA. In timelapse microscopy, we observed distinct interactions between high GALA and low GALA cells, which resulted in apoptosis in the latter.

Our findings revealed that when low GALA cells are exposed to high GALA cells, they undergo more apoptosis than when cocultured with cells with the same level of glycosylation. Additional research is needed to determine the molecular actors in these interactions.

In most cancers, death is caused by metastases rather than the primary tumour. This is the case with colorectal cancer, where metastases account for over 90% of mortality. This cancer metastasizes to the liver in over 50% of cases, making it an interesting model for the study of metastatic spread and our choice for this project. Metastases occur when certain cells break away from the tumour, enter the bloodstream, where they are known as Circulating Tumour Cells (CTCs), before eventually leaving and nestling in new organs. CTCs form a heterogeneous population, many of which will not survive exposure to the stressful conditions of the bloodstream (flow, frictional forces, lack of contact with other epithelial and matrix cells, …). Nor are all surviving cells have the capacity to leave the bloodstream (extravasation), migrate into a new tissue, and proliferate. Only some CTCs are therefore at the origin of metastases, but there is currently no marker that can distinguish CTCs according to their metastatic potential. The impact of CTC analysis could be enhanced by ability to distinguish cells with metastatic potential from harmless cells.

I will present an in vitro model I implement to determine the metastatic potential of CTCs and study its characteristics.

This in vitro test is based on 1- the cells' ability to survive in a flow similar to the bloodstream mimicked by a fluidic system, 2- their ability to cross a physical barrier and migrate to a host site mimicked by a porous membrane, 3- to invade and proliferate in a host tissue represented by a matrix where invasion and proliferation are assessed. I will present the early work on establishing and testing of this system.

We have laid the foundations for a simple fluidic system that can mimic the escape of CTCs from circulation to reach a host compartment. The relevance of the system to reflect the metastatic potential of the cells needs to be confirmed by further experiments with other cell lines. The simple system presented here will also be refined to better reflect the physiological situation.