Fig. 7

Interface of TM4SF5 protein and TSIs. a–c Structural modeling of TM4SF5 and TSI binding. a Structural model of TM4SF5 (ribbon form) with simulated phospholipid membrane (stick form) depicted with the docking result at the luminal side of the protein. Refined docking models of ST-5-001 (b) and ST-5-002 (c) show that the isoxazoles fit into the long extracellular loop region positioned at the luminal side of TM4SF5. Both isoxazoles made a hydrogen bond to Arg113, while ST-5-002 made additional favorable interactions toward Asn138 (hydrogen bond) and Trp142 (π–π stacking). d, e Binding of ¹⁴C-ST-5-002 to pulled-down (PD) or immunoprecipitated (IP) Strep-TM4SF5 (left) or HA-TM4SF5 (right) from HEK293FT cells or Huh7_KO cells, respectively, was assayed after a 1 h incubation. Cold ST-5-002 was pretreated to the reaction mixture for 1 h. The indicated cDNAs were transiently (48 h, left) or stably (right) introduced before TM4SF5 PD or IP. Different concentrations of ¹⁴C-ST-5-002 were reacted with TM4SF5 PD or IP at the same protein amount for 1 h at 4 °C before β-particle counting. Data represent three isolated experiments. f A working model of the inhibitory effects of TSIs on TM4SF5-mediated SLAMF7 downregulation and NK cell inactivation. TM4SF5 in cancerous hepatocytes binds SLAMF7, a stimulatory NK cell ligand, and causes SLAMF7 internalization and trafficking toward lysosomes where it can be degraded, leading to less availability of the ligand for the receptor on NK cells (SLAMF7), resulting in NK cell inactivation and HCC progression (left). TSIs can bind to TM4SF5 and thereby interrupt its binding to SLAMF7, thus preventing SLAMF7 degradation and making SLAMF7 available to its receptor on NK cells. Thereafter, NK cells are activated for the secretion of lytic granules including granzyme and perforin to kill TM4SF5-positive HCC (right). It was generated from online tools of Biorenders