Direct effect of bevacizumab on glioblastoma cell lines in vitro.

Bevacizumab is a humanized monoclonal antibody directed against the pro-angiogenic factor vascular and endothelial growth factor-A (VEGF-A) used in the treatment of glioblastomas. Although most patients respond initially to this treatment, studies have shown that glioblastomas eventually recur. Several non-mutually exclusive theories based on the anti-angiogenic effect of bevacizumab have been proposed to explain these mechanisms of resistance. In this report, we studied whether bevacizumab can act directly on malignant glioblastoma cells. We observe changes in the expression profiles of components of the VEGF/VEGF-R pathway and in the response to a VEGF-A stimulus following bevacizumab treatment. In addition, we show that bevacizumab itself acts on glioblastoma cells by activating the Akt and Erks survival signaling pathways. Bevacizumab also enhances proliferation and invasiveness of glioblastoma cells in hyaluronic acid hydrogel. We propose that the paradoxical effect of bevacizumab on glioblastoma cells could be due to changes in the VEGF-A-dependent autocrine loop as well as in the intracellular survival pathways, leading to the enhancement of tumor aggressiveness. Investigation of how bevacizumab interacts with glioblastoma cells and the resulting downstream signaling pathways will help targeting populations of resistant glioblastoma cells.

Elastin-derived peptides: matrikines critical for glioblastoma cell aggressiveness in a 3-D system.

In the most common primary brain tumors, malignant glioma cells invade the extracellular matrix (ECM) and proliferate rapidly in the cerebral tissue, which is mainly composed of hyaluronan (HA) along with the elastin present in the basement membrane of blood vessels. To determine the role of ECM components in the invasive capacity of glioma cell lines, we developed a 3-D cell-culture system, based on a hydrogel in which HA can be coreticulated with kappa-elastin (HA-kappaE). Using this system, the invasiveness of cells from four glioma cell lines was dramatically increased by the presence of kappaE and a related, specific peptide (VGVAPG)(3). In addition, MMP-2 secretion increased and MMP-12 synthesis occurred. Extracellular injections of kappaE or (VGVAPG)(3) provoked a pronounced and dose-dependent increase in [Ca(2+)](i). kappaE significantly enhanced the expression of the genes encoding elastin-receptor and tropoelastin. We propose the existence of a positive feedback loop in which degradation of elastin generates fragments that stimulate synthesis of tropoelastin followed by further degradation as well as migration and proliferation of the very cells responsible for degradation. All steps in this ECM-based loop could be blocked by the addition of either of the EBP antagonists, lactose, and V-14 peptide, suggesting that the loop itself should be considered as a new therapeutic target.

Secretion of MMP-2 and MMP-9 induced by VEGF autocrine loop correlates with clinical features in childhood acute lymphoblastic leukemia.

In children with acute lymphoblastic leukemia (ALL), leukemic cells express several members of the VEGF family and the three VEGF receptors which, via an autocrine loop are responsible for secretion of MMP-2/-9. MMP activity and the presence of elements of the autocrine loop are correlated with clinical and prognostic parameters as follows: i) high basal MMP-9 activity with tumoral syndrome, ii) MMP-2 activity with treatment failure, iii) VEGFR-1/-3 co-expression with high hemoglobin level and iv) expression of the VEGF-A 121 isoform and favorable response to treatment. These data implicate autocrine VEGF-induced secretion of MMP-2/-9 in the physiopathology of childhood ALL.

Quantitation of Human herpes virus 6 genome in children with acute lymphoblastic leukemia.

Acute lymphoblastic leukemia is the main type of leukemia in children. An infectious etiology has been suspected and the role of the Human herpesvirus-6 (HHV-6) has been suggested. Several studies have tried to establish a link between HHV-6 infections and hematological malignancies, with discordant results. The potential role of HHV-6 in the pathogenesis of pediatric acute lymphoblastic leukemia was investigated. HHV-6 genome copy number was measured by quantitative real-time PCR (RQ-PCR) in bone marrow or peripheral blood samples obtained from 36 children (median age = 4 years) with B acute lymphoblastic leukemia (n = 31) and T acute lymphoblastic leukemia (n = 5) at diagnosis and during complete remission. Positive samples were further characterized to define viral variant, A or B. A total of 24.7% of samples were positive for HHV-6 genome: 13.9% were leukemia samples and 34.1% were complete remission samples. Viral load was low with values lower at diagnosis (median viral copy number = 22.9) than at complete remission (median copy number = 60.1). Among the 17 patients with positive samples, 15 were typed as B-variant whereas 2 could not be typed. These results argue against a role of HHV6 infection in the development of pediatric acute lymphoblastic leukemia. They also suggest that HHV-6 may infect latently bone marrow progenitors but seems not able to infect leukemic cells, raising again the question of the mechanism of virus fusion and entry. This observation shows that a reactivation may be observed during complete remission supporting the possibility of virus reactivation in immunocompromised hosts.