Suppression of non-muscle myosin II boosts T cell cytotoxicity against tumors

Abstract

Increasing evidence highlights the importance of immune mechanoregulation in establishing and sustaining tumor-specific cytotoxicity required for desirable immunotherapeutic outcomes. However, the molecular connections between mechanobiological inputs and outputs and the designated immune activities remain largely unclear. Here, we show that partial inhibition of non-muscle myosin II (NM II) augmented the traction force exerted by T cells and potentiated T cell cytotoxicity against tumors. By using T cells from mice and patients with cancer, we found that NM II is required for the activity of NKX3-2 in maintaining the expression of ADGRB3, which shapes the filamentous actin (F-actin) organization and ultimately attributes to the reduced traction force of T cells in the tumor microenvironment. In animal models, suppressing the NM II-NKX3-2-ADGRB3 pathway in T cells effectively suppressed tumor growth and improved the efficacy of the checkpoint-specific immunotherapy. Overall, this work provides insights into the biomechanical regulation of T cell cytotoxicity that can be exploited to optimize clinical immunotherapies.

Fig. 1.. NM II inhibition enhances T cell cytotoxicity.

(A) Anti-CD3/CD28 bead–activated Pmel-1 CTLs pretreated with vehicle (Veh.), Lat (1 μM), Jas (200 nM), or Ble (2 μM) for 24 hours were cocultured with B16 cells at a ratio of 20:1 for 8 hours. CD45⁻ tumor cell apoptosis was determined by flow cytometry. (B to F) Activated mouse CTLs were treated with Veh., Lat, Jas, or Ble (2 μM) for 24 hours. Flow cytometry analyzed the expression of TNF, IFN-γ (B), granzyme B (GzmB) (C), perforin (D), FasL (E), and CD107a (F). In (A) to (F), n = 3 independent experiments. One-way ANOVA followed by Bonferroni’s test (A). The data represent mean ± SD. **P < 0.01.

Fig. 2.. Blocking NM II enhances the traction force exerted by T cells.

(A) Schematic diagram of TFM used in the following experiments. (B) Traction force of activated mouse CTLs pretreated with Veh., Lat, Ble (2 μM), or Jas for 24 hours (Veh., n = 77; Lat, n = 84; Ble, n = 91; Jas, n = 82). Scale bar, 5 μm. (C) Activated human CD8⁺ T cells were treated with Veh. or Ble (2 μM) for 24 hours. The traction force was analyzed by TFM (Veh, n = 100; Ble, n = 128). Scale bar, 5 μm. (D) Traction force of activated mouse CTLs pretreated with Veh. or Ble (2 μM) for 24 hours, varied with time up to 30 min (n = 25). Scale bar, 5 μm. (E) Schematic diagram of the FluidFM method used to detect the adhesion force between T cells and tumor cells. (F) Adhesion force between activated OT-1 CTLs pretreated with Veh. or Ble (2 μM) and OVA-B16 cells (Veh, n = 19; Ble, n = 22). (G) OVA-B16 cells seeded on the PAA cells were cocultured with OT-1 CTLs in the presence of PI. The traction force of OVA-B16 cell was measured, and the PI staining was recorded during the observation time period. Scale bar, 20 μm. In (B) to (D), (F), and (G), n = 3 independent experiments. In (B), one-way ANOVA followed by Bonferroni’s test. In (C) and (F), Mann-Whitney test. The data represent mean ± SD. *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig. 3.. Inhibiting NM II enhances the efficacy of T cell immunotherapy.

(A) Experimental design for CTL transfer experiments. CD45.1⁺ CTLs from OT-1 transgenic mice were activated as indicated and transferred to OVA-B16–bearing mice. CD45.1⁺ CTLs in tumors were analyzed by flow cytometry and TFM [(B) to (F), n = 5 mice]. (B to E) Expression of CD107a (B), immune inhibiting receptors (C), TNF and IFN-γ (D), and the number of CD45.1⁺ T cells (E) from the intratumors. TIL, tumor-infiltrating lymphocyte. (F) Traction force exerted by CD45.1⁺ T cells. Scale bar, 5 μm. (G to I) OT-1 CTLs were transferred into OVA-B16–bearing mice as indicated (G). The tumor growth was recorded [(H), n = 9], and mouse survival was calculated [(I), n = 10]. (J and K) Pmel-1 CTLs were activated as indicated and transferred to B16F10-bearing mice. Some mice were treated with anti–PD-1 antibodies (J). The tumor growth was recorded [(K), n = 8]. Two-tailed Student’s t test (E), Mann-Whitney test (F), one-way ANOVA followed by Bonferroni’s test [(H) and (K)], or log-rank test (I). The data represent mean ± SD. [(B) to (F)] or mean ± SEM [(H) and (K)]. *P < 0.05, **P < 0.01, ***P < 0.005, ****P < 0.0001.

Fig. 4.. NM II blocking down-regulates the expression of NKX3-2.

(A) Heatmap showing the relative expression of DEGs in T cells treated with Ble (2 μM) for 24 hours. (B) Top 12 enriched GO terms of considerably down-regulated DEGs in T cells treated with Ble. The color and size of the bubbles represent the statistical significance and gene counts, respectively. (C) Activated mouse CD8⁺ T cells pretreated with Veh. and Ble (2 μM) for 48 hours were immunostained for NKX3-2. Scale bar, 10 μm. (D) Activated CD8⁺ T cells from peripheral blood (PB) of the healthy volunteers were treated with Veh. and Ble (2 μM) for 48 hours. T cells were immunostained for NKX3-2. Scale bar, 10 μm. (E) Activated human CD8⁺ T cells were treated with Veh., Ble (2 μM), or Ble/TSA (a specific histone deacetylase inhibitor) for 24 hours. CUT&Tag analysis was performed with NKX3-2 antibody and ADGRB3 promotor–specific primers. Data are presented as relative to the Veh. group (n = 3). In (C) to (E), n = 3 independent experiments. Mann-Whitney test [(C) and (D)] or one-way ANOVA followed by Bonferroni’s test (E). The data represent mean ± SD [(C) to (E)]. *P < 0.05, **P < 0.01, ****P < 0.0001.

Fig. 5.. NKX3-2–ADGRB3 pathway mediates T cell traction force.

(A and B) Activated human CD8⁺ T cells were treated with shRNAs targeting NKX3-2. The expression of ADGRB3 was determined by immunoblot (A) or immunofluorescence (B) analysis. Scale bar, 2 μm. (C to F) Activated human CD8⁺ T cells were treated with shRNAs targeting ADGRB3. GFP⁺ T cells were sorted for real-time PCR analysis (C), immunostaining with F-actin (D), the orientation of F-actin (E), or TFM analysis (F). Scale bar, 5 μm. (G to I) Activated human or mouse CD8⁺ T cells were treated with shRNAs targeting NKX3-2. GFP⁺ T cells were sorted for immunostaining with F-actin (G), the orientation of F-actin (H), or TFM analysis (I). Scale bar, 5 μm. In (A) to (D), (F), (G), and (I), n = 3 independent experiments. One-way ANOVA followed by Bonferroni’s test [(A) to (D), (F), (G), and (I)]. The data represent mean ± SD. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 6.. NKX3-2–ADGRB3 signaling regulates T cell traction force in cancer patients.

(A and E) CD8⁺ T cells from PB of healthy volunteers (n = 6) and patients with colorectal cancer (CRC; n = 10), stomach cancer (STAD; n = 8), and hepatocellular carcinoma (HCC; n = 6) were activated with anti-CD3 and anti-CD28 antibodies for 48 hours. T cells were immunostained for NKX3-2 [(A), n = 35 to 40 cells from each patient] or analyzed by TFM (E). Scale bar, 5 μm. (B to D) Activated CD8⁺ T cells from the PB of CRC (B) or STAD [(C) and (D)] patients were treated with Ble (2 μM) for 24 hours. The expression of NKX3-2 or ADGRB3 was determined by immunofluorescence [(B) and (C)] or immunoblot analysis (D). n = 10. Scale bar, 5 μm. (F and G) The traction force of activated CD8⁺ T cells pretreated with Veh. or Ble (2 μM) from patients with CRC (F) or STAD (G) was detected by TFM. n = 100. (H and I) Activated CD8⁺ T cells from the PB of CRC patients were treated with Veh. or Ble (2 μM) for 24 hours. T cells were immunostained for F-actin [(H), n = 12 from 4 patients], and the orientation of F-actin was shown (I). Scale bar, 5 μm. (J and K) Activated CD8⁺ T cells from CRC patients were treated with shRNAs targeting ADGRB3. T cells were immunostained for F-actin (J), and the orientation of F-actin was shown (K). n = 12 from 4 patients. Scale bar, 5 μm. One-way ANOVA followed by Bonferroni’s test [(A), (E), and (J)], two-tailed Student’s t test [(B) to (D)], or Mann-Whitney test [(F) and (G)]. The data are mean ± SD [(A) to (H) and (J)]. ***P < 0.005, ****P < 0.0001.

Fig. 7.. The expression of NKX3-2 or ADGRB3 is negatively associated with both CD8⁺ T cell infiltration and overall survival in response to immunotherapy in multiple types of human cancers.

(A and B) Negative correlation between NKX3-2 (A) or ADGRB3 (B) expression and overall survival in human cancers in response to anti–PD-1 (A) or all immunotherapies (B) in human cancers. (C and D) Anti-correlation of the expression of NKX3-2 and CD8⁺ T cell infiltration in multiple human cancers, including colon adenocarcinoma [COAD; (C); n = 458], breast invasive carcinoma [BRCA; (D); n = 1100], and pancreatic adenocarcinoma [PAAD; (D); n = 179]. (E) ADGRB3 expression was negatively correlated to CD8⁺ T cell infiltration in patients with glioblastoma multiforme (GBM; n = 153) and skin cutaneous melanoma (SKCM; n = 471). (F) Schematic of the NM II–NKX3-2–ADGRB3 pathway–mediated T cell killing cancer cells. Log-rank test [(A) and (B)] or Pearson’s correlation test [(C) to (E)].