Le carcinome hépatocellulaire (CHC) est un cancer de mauvais pronostic. Les maladies chroniques du foie provoquent le développement d'une cirrhose qui évoluera éventuellement en CHC. Le récepteur de l'insuline (IR) est exprimé sous deux isoformes, IR-A et IR-B, qui résultent de l'épissage alternatif de l'ARNm. IR-B est l'isoforme majoritairement exprimé dans l'hépatocyte adulte et IR-A dans l'hépatocyte fœtal. Plusieurs travaux montrent une surexpression de IR-A dans différents cancers, associée à une signalisation oncogénique en réponse à l'IGF-II. Le rôle de la signalisation IGF-II/IR-A a peu été investigué dans le CHC. Notre étude porte sur le rôle biologique de la surexpression de IR-A dans les cellules de CHC au regard de la surexpression d'IGF-II. Une augmentation du ratio d'expression IR-A/IR-B est observée dans 70% des tumeurs de CHC par rapport au foie adjacent non tumoral, qui est significativement associée à l'expression de marqueurs clinicopathologiques de mauvais pronostic dont des marqueurs de cellules souches/progénitrices (CSP) et à une faible survie post-hépatectomie. Seules 9,4% des tumeurs surexpriment IGF-II. In vivo, la surexpression stable de IR-A, et non de IR-B, augmente la tumorigénicité de deux lignées de CHC disposant (Huh7) ou non (PLC/PRF5) d'une boucle autocrine d'IGF-II, et est associée à l'induction de différentes signatures pro-inflammatoires. Sans effet sur la prolifération, la surexpression de IR-A stimule la migration et l'invasion in vitro et augmente l'expression de marqueurs CSP. En conclusion, ces résultats identifient IR-A comme un nouvel acteur de la progression du CHC de façon indépendante d'une boucle autocrine d'IGF-II. / Hepatocellular carcinoma (HCC) is a poor prognosis cancer. Chronic liver diseases induce cirrhosis which will eventually evolve into HCC. The insulin receptor (IR) exists in two isoforms, IR-A and IR-B, resulting from mRNA alternative splicing. Adult hepatocytes predominantly express IR-B while fœtal hepatocytes mainly express IR-A. The overexpression of IR-A associated with IGF2 oncogenic signaling has been abundantly described in cancer cells. Little is known about IGF2/IR-A signaling in HCC. Our study aims to evaluate the biological functions associated to IR-A overexpression in HCC in relation to IGF2 overexpression. We observed that 70% of 85 HCC tumors showed upregulation of IR-A/IR-B ratio compared to adjacent nontumor tissue, which was significantly associated with clinicopathological markers of aggressive tumours such as stem/progenitor cell (SPC) markers and correlated with reduced patient survival after surgery. IGF2 upregulation was observed in only 9.4% of HCC. Stable overexpression of IR-A, but not IR-B, increased in vivo tumorigenicity in two HCC cell lines presenting (Huh7) or not (PLC/PRF5) an autocrine IGF2 secretion loop and was associated with the induction of different pro-inflammatory gene signatures. While IR-A overexpression did not promote cell proliferation in vitro, it stimulated migration and invasion and increased some SPC markers expression. Altogether these results highlight IR-A as a novel player in HCC progression irrespective of an IGF2 autocrine loop.

Umbilical cord blood (UCB) is used as a source of haematopoietic stem cells (HSCs) for transplantation but shows defective homing to the bone marrow niche and delayed haematological reconstitution. Following transplantation, HSCs will home to the bone marrow in response to the CXCL12 chemokine, adhere to the bone marrow sinusoidal endothelial cells and then migrate into and lodge in bone marrow niches. In addition to CXCR4, a variety of molecules have been described as being important in these processes. In this laboratory, junctional adhesion molecule-A (JAM-A) was shown to be expressed on human UCB CD133⁺/CD34⁺ cells and regulated by hypoxia. In this thesis, further phenotypic studies show that this molecule is most highly expressed on human CD41a⁺ megakaryocytes and CD14⁺ monocytes/macrophages in UCB. JAM-A was also found to be expressed on all human UCB CD133⁺ cells, which have been shown by others to encompass the HSCs and early myeloid-lymphoid precursors and on the majority of CD34⁺ haematopoietic progenitor cells (HPCs). While it is also present on bone marrow sinusoidal endothelium (BMEC), JAM-A is not detected on cultured bone marrow mesenchymal stromal cells (MSCs). JAM-A blockade, silencing and overexpression experiments showed that JAM-A contributes to, but is not solely responsible for, the adhesion of CD34⁺ haematopoietic progenitor cells to IL-1β activated BMEC-60 cells and fibronectin. Lack of significance in cell migration suggested that JAM-A is more likely to act as an adhesion molecule or a regulator of adhesion rather than as a migratory molecule in such cells. Further functional studies using the proximity ligation assay highlight a potential association of JAM-A with CXCR4 and the adhesion molecules, tetraspanin CD82 and integrin β1. Mechanistic studies were commenced to establish if JAMA could modulate CXCR4 signalling following CXCL12 stimulation, but time constraints prevented these from being completed. These preliminary experiments which were carried out first in the Jurkat cell line lacking JAM-A or transduced to express JAM-A, however, suggest that JAM-A may modulate CXCL12-induced Rap1 phosphorylation and ERK1/2 phosphorylation. The former pathway is important for integrin function and the latter pathway is important in cell adhesion. The results described here, although requiring finalisation, support the hypothesis that JAM-A acts as an adhesion molecule and also may fine tune CXCR4 and integrin mediated functions on human CD34⁺ cells, thereby potentially regulating engraftment of these cells to the bone marrow niche.