For example, the reduced cytoskeletal reorganization will weaken the cytoskeletal engagement of integrins in tetraspanin-enriched microdomains and subsequently lead to the local attenuation of integrin signaling or even inactivation of integrins. We extrapolate that both imbalance of Rho GTPases and inactivation of integrins result in the aberrancy in cellular morphology and diminishment in cell motility. Consistent with the current understanding of the roles of Rac in cell morphology and movement, the suppressed lamellipodia and cell movement upon KAI1/CD2 expression correlated well with the diminished Rac1 activity in Du145 cells. In some cell types, Rac1 activation negatively regulates RhoA activity by generating reactive oxygen species and subsequently activating p190RhoGAP at the plasma membrane. The delicate balance between the antagonistic activities of Rac1 and RhoA is crucial for proper cell movement and also specifies cell morphology. In addition, KAI1/CD82 was reported to inhibit the activity of Src kinases, which activates p190RhoGAP by tyrosine phosphorylation, and the lower Src activity ultimately leads to the RhoA activation. However, the constitutive RhoA activity appears to be unaltered upon KAI1/CD82 expression. Although HGF markedly enhanced RhoA activity, KAI1/CD82 apparently could block such stimulation, reflected by the defect in retraction and the formation of fewer stress fibers in KAI1/Mepiroxol CD82-expressing cells. It remains unclear that, in Du145KAI1/CD82 cells, RhoA still exhibits a similar level of constitutive activity, although it cannot reach the plasma membrane. As a key effector of both Rac and Rho, cofilin plays an important role in membrane ruffling or lamellipodia formation. Driven by activated Rac or PIP2, cofilin is translocated to or enriched in the cell periphery where it interacts with actin cytoskeleton, generates more barbed ends, and promotes actin cortical meshwork formation and consequently lamellipodia formation. Translocation of cofilin to the plasma membrane is considered to be an indicator of cofilin activation. Interestingly, the subcellular localizations of total and inactivated cofilin in Du145-Mock and -KAI1/CD82 cells displayed distribution patterns similar to the ones in migrating and nonmigrating fibroblast cells, respectively. Upon KAI1/ CD82 expression, cofilin was no longer enriched at the cell periphery, although it could translocate to the peripheral cytoplasm. This observation strongly suggests that KAI1/CD82 expression blocks the translocation and therefore activation of cofilin, underlining a mechanism by which KAI1/CD82 impairs lamellipodia formation. If so, 
one would expect more phosphorylated or inactive cofilin in Du145-KAI1/CD82 cells. However, the level of phosphorylated cofilin proteins remained unchanged upon KAI1/CD82 expression. Possibly, cofilin in Du145-KAI1/ CD82 cells can still undergo de4-(Benzyloxy)phenol phosphorylation during translocation to the peripheral cytoplasm but cannot become enriched at the cell periphery due to aberrant plasma membrane in Du145KAI1/CD82 cells. When KAI1/CD82 is present, HGF cannot upregulate the activities of ROCK and its direct upstream activator RhoA. Besides agreeing with earlier observations that KAI1/CD82 inhibits HGF/c-Met signaling, these findings have further demonstrated that CD82 intercepts the HGF/c-Met signaling leading to the cellular retraction. Because KAI1/CD82 can inhibit the retraction process even without HGF stimulation, we predict that the local, constitutive ROCK activity at the retraction areas in KAI1/CD82-expressing cells.