Zhx2 deletion boosts NK cell antitumor immunity. (A and B) Analysis of 371 biologically independent HCC samples from a TCGA dataset. A shows the results of analysis of ZHX2 expression related to intratumoral NK cell level in HCC patients. B shows the results of Kaplan-Meier analysis of overall survival in patients in the discovery cohort classified by ZHX2 expression. (C–E) Tumor growth and weight in C57BL/6 mice s.c. transplanted with Hepa1-6 cells, which were injected via the tail vein with purified NK cells from Zhx2^+/+ or Zhx2^Δ/Δ mice (representative of two independent experiments). Experimental design, in vitro tumor image (C), tumor growth (D), and weight (E) are shown. Scale bar: 1 cm. (F–H) B16-F10 cells (2 × 10⁵) were tail vein injected into C57BL/6 mice. 3 d later, mice were treated with 1 × 10⁶Zhx2^+/+ or Zhx2^Δ/Δ NK cells with tail vein injection (representative of two independent experiments). The experimental design and in vitro tumor imaging (F), lung nodules (G), and survival curves (H) are shown. Scale bar: 1 cm. (I–K) Human HCC mouse models were established by i.p. injection of human HepG2-luciferase cells (5 × 10⁶) and transfer of 2 × 10⁶ LV-shNC- or LV-shZHX2-NK92 cells. Experimental design (I), tumor burden at day 10 and day 20 (J), and survival curves (K) are shown. ROI, region of interest. (L and M) Human HCC mouse models were established by s.c. injection of human Huh7 cells (5 × 10⁶) and transfer of 2 × 10⁶ LV-shNC- or LV-shZHX2-NK92 cells. Experimental design, tumor image (L), and weight (M) are shown. Scale bar: 1 cm. Dots represent data from individual mice; error bars, SEM per group in one experiment. Data were analyzed using Student’s t test (two-tailed paired t test) in A, E, G, J, and M; using log-rank (Mantel-Cox) test in B, H, and K; and using two-way ANOVA in D. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.