Fig. 7: CHMP2A-KO HNSCC xenograft model shows increased sensitivity to NK cells.

a Schematic of in vivo treatment. Non-obese diabetic/severe combined immunodeficiency/γc^−/− (NSG) mice were inoculated subcutaneously with 6 × 10⁶ cells (Cal27-WT or KO) and injected i.v. with 1 × 10⁷ NK cells. NK cells were supported by injections IL-15 daily and IL-2 every other day for 1 week. Tumor volume was monitored every 2–3 days. b Tumor volume progression over 21 days shown as mean ±SEM of n = 5 mice per treatment. Statistical analysis was performed by two-way ANOVA and Bonferroni’s post hoc multiple comparison test (ns, not significant). c Representative flow cytometry dot plots showing increased NK cell infiltration in Cal27 CHMP2A KO tumors. The gating strategy has been defined by combining NK cells and a cell suspension from the untreated tumor. Tumor samples without NK cell treatment were mixed with 2 × 10⁴ human NK cells, stained with anti-human CD45 and anti-human CD56 antibodies or isotype controls, and analyzed. Cells double positive for (human) hCD45 and hCD56 were gated and the gate was used to quantify the number of infiltrating NK cells in Cal27-WT and CHMP2A KO tumors. The graph shows the average NK cell infiltration for all mice in each group (n = 4 mice per group). Error bars represent ±SEM across n = 4 technical replicates and an unpaired one-tailed Student’s t test was performed to determine statistical significance. d Cartoon describing the proposed mechanism of CHMP2A KO in tumor cells. Tumor cells secrete EVs-bearing ligands like MICA/B and TRAIL, which suppress the antitumor effect of NK cells by binding to NKG2D and TRAIL-R to induce apoptosis on NK cells. By CHMP2A KO we reduce EVs secretion from tumor cells decreasing their inhibitory function. Tumor cells CHMP2A KO also secrete CXCL10 and CXCL12, chemokines that increase NK cell migration culminating in enhanced NK cell-mediated cytotoxicity (the cartoon has been created with Biorender.com).