Tissue factor-induced fibrinogenesis mediates cancer cell clustering and multiclonal peritoneal metastasis.

Peritoneal metastasis is one of the most frequent causes of death in several types of advanced cancers; however, the underlying molecular mechanisms remain largely unknown. In this study, we exploited multicolor fluorescent lineage tracking to investigate the clonality of peritoneal metastasis in mouse xenograft models. When peritoneal metastasis was induced by intraperitoneal or orthotopic injection of multicolored cancer cells, each peritoneally metastasized tumor displayed multicolor fluorescence regardless of metastasis sites, indicating that it consists of multiclonal cancer cell populations. Multicolored cancer cell clusters form within the peritoneal cavity and collectively attach to the peritoneum. In vitro, peritoneal lavage fluid or cleared ascitic fluid derived from cancer patients induces cancer cell clustering, which is inhibited by anticoagulants. Cancer cell clusters formed in vitro and in vivo are associated with fibrin formation. Furthermore, tissue factor knockout in cancer cells abrogates cell clustering, peritoneal attachment, and peritoneal metastasis. Thus, we propose that cancer cells activate the coagulation cascade via tissue factor to form fibrin-mediated cell clusters and promote peritoneal attachment; these factors lead to the development of multiclonal peritoneal metastasis and may be therapeutic targets.

Aberrant alternative splicing of RHOA is associated with loss of its expression and activity in diffuse-type gastric carcinoma cells.

RhoA is a member of Rho family small GTPases that regulates diverse cellular functions. Recent large-scale sequencing studies have identified recurrent somatic mutations of RHOA in diffuse-type gastric carcinoma (DGC), indicating that RHOA is a driver of DGC. In this study, we investigated the possible abnormalities of RHOA in a panel of gastric carcinoma (GC) cell lines. Pulldown assay and immunoblot analysis showed that the activity and expression of RhoA were detectable in all GC cell lines tested, except for two DGC cell lines, HSC-59 and GSU. RHOA coding region sequencing revealed that aberrant alternative splicing of RHOA occurred in these cell lines. Quantitative real-time PCR analysis showed that the expression of wild-type RHOA was nearly undetectable, whereas splicing variants were almost exclusively expressed in HSC-59 and GSU cell lines. However, the expression levels of RHOA splicing variants were very low and the corresponding proteins were not detected by immunoblotting. Moreover, the splicing isoforms of RhoA protein were neither efficiently expressed nor activated even if ectopically expressed in cells. These results indicate that aberrant alternative splicing of RHOA results in the loss of its activity and expression in DGC cells.

Actinin-1 and actinin-4 play essential but distinct roles in invadopodia formation by carcinoma cells.

Invadopodia are ventral membrane protrusions formed by cancer cells that degrade the extracellular matrix (ECM) during tumor invasion and metastasis. Formation of invadopodia is initiated by the assembly of actin filaments (F-actin) that results from the coordinated activation of several actin regulatory proteins. Actinin-1 and actinin-4 are actin bundling proteins expressed in non-muscle cells and actinin-4 is preferentially associated with malignant phenotypes of carcinoma cells. In this study, we investigated the role of actinin-1 and -4 in invadopodia formation. Expression of both actinin-1 and -4 tended to be higher in invasive and metastatic breast carcinoma cell lines than in non-invasive ones. Immunofluorescence analysis revealed that actinin-1 and -4 colocalized at core actin structures of invadopodia. Time-lapse imaging showed that appearance of both actinins at invadopodia is concomitant with the assembly of F-actin. Knockdown of either actinin-1 or actinin-4 suppressed the formation of invadopodia and degradation of the ECM by carcinoma cells. Interestingly, overexpression of actinin-4, but not actinin-1, significantly promoted the formation of invadopodia and this activity required the actin binding domains and the unique N-terminal motif that exists only in actinin-4. These results demonstrate that both actinin-1 and actinin-4 participate in the assembly of F-actin at invadopodia. Additionally, actinin-4 may have a selective advantage in accelerating invadopodia-mediated invasion of carcinoma cells.