Activation of STING by SAMHD1 Deficiency Promotes PANoptosis and Enhances Efficacy of PD-L1 Blockade in Diffuse Large B-cell Lymphoma

Abstract

Genomic instability is a significant driver of cancer. As the sensor of cytosolic DNA, the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway plays a critical role in regulating anti-tumor immunity and cell death. However, the role and regulatory mechanisms of STING in diffuse large B-cell lymphoma (DLBCL) are still undefined. In this study, we reported that sterile alpha motif and HD domain-containing protein 1 (SAMHD1) deficiency induced STING expression and inhibited tumor growth in DLBCL. High level of SAMHD1 was associated with poor prognosis in DLBCL patients. Down-regulation of SAMHD1 inhibited DLBCL cell proliferation both in vitro and in vivo. Moreover, we found that SAMHD1 deficiency induced DNA damage and promoted the expression of DNA damage adaptor STING. STING overexpression promoted the formation of Caspase 8/RIPK3/ASC, further leading to MLKL phosphorylation, Caspase 3 cleavage, and GSDME cleavage. Up-regulation of necroptotic, apoptotic, and pyroptotic effectors indicated STING-mediated PANoptosis. Finally, we demonstrated that the STING agonist, DMXAA, enhanced the efficacy of a PD-L1 inhibitor in DLBCL. Our findings highlight the important role of STING-mediated PANoptosis in restricting DLBCL progression and provide a potential strategy for enhancing the efficacy of immune checkpoint inhibitor agents in DLBCL.

Keywords: DLBCL; DMXAA.; PANoptosis; SAMHD1; STING.

Figure 1

SAMHD1 expression is up-regulated in DLBCL and related to tumor growth. A, B. The mRNA levels of SAMHD1 in DLBCL samples from the Oncomine (A) and TCGA database (B). C. IHC staining revealed the expression of SAMHD1 in GCB-like DLBCL, ABC-like DLBCL, and RHL tissues (upper), followed by relative quantitative analysis (lower, p=0.04). P value came from the Chi-square test. D. Kaplan Meier plots for DLBCL patients enrolled in IHC staining (n=100, p=0.03). E. Protein levels of SAMHD1 in normal CD19⁺ B-cells (N1, N2, N3) and DLBCL cell lines (LY1, LY3, LY8) were detected by western blotting. F, G. LY1 cells were transfected with LV-Con or LV-SAMHD1 sequences. After stable transfection, transfection efficiency was examined by immunoblot (F), while cell viability was detected by CCK-8 assay (G). H, I. LY1 and LY3 cells were transfected with Ctrl or SAMHD1-KD sequences (sh1#, and sh2#). After stable transfection, knockdown efficiency was examined by immunoblot (H), while cell viability was detected by CCK-8 assay (I). J-L. Schematic of in vivo tumor growth investigation (J). SCID beige mice were injected with Ctrl or SAMHD1-KD LY1 cells (n=6/group). Tumor bodies were taken on day 28 (K). Tumor growth curves were shown from day 18 to day 28 (L). Immunoblot images in E, F, and H were the representation of 3 independent experiments. Vertical bars indicated mean ± SD. P values came from Kruskal-Wallis test followed by Dunn's test (A), Log-rank test (D), unpaired two-tailed t-test (B, F, G, H, I), and Two-way ANOVA with Sidak correction (L). ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, and ^****p < 0.0001.

Figure 2

SAMHD1 deficiency induces DNA damage and dsDNA accumulation. A-C. RNA-seq was performed in Ctrl and SAMHD1-KD LY1 cells (three biological replicates for each group). Volcano plot revealed DEGs in SAMHD1-KD cells compared with Ctrl cells (A). Bubble plots revealed the significantly up-regulated GO terms (B) and Reactome pathways (C) in SAMHD1-KD LY1 cells. D. Fluorescence plots (upper) and fluorescence density (lower) of H2AX in Ctrl and SAMHD1-KD DLBCL cells. Scale bar=10μm. E. Representative images (upper) and qualification (lower) of neutral comet assay revealed the abundance of DNA fragments in Ctrl and SAMHD1-KD DLBCL cells. Scale bar=20μm. Individual dots represented single cells. F. Fluorescence plots (upper) and fluorescence density (lower) of dsDNA in Ctrl and SAMHD1-KD DLBCL cells. Scale bar=10μm. Vertical bars indicated mean ± SD. P values from unpaired two-tailed t-test (D, E, F). ^*p<0.05, ^****p<0.0001.

Figure 3

Down-regulation of SAMHD1 promotes the expression of the STING-related DNA-sensing pathway. A. GSEA plot of the cytosolic DNA-sensing pathway (NES=1.36, p<0.05, FDR=0.07). B. Immunoblot showed the protein levels of cGAS and STING proteins in Ctrl and SAMHD1-KD DLBCL cells. Immunoblot images were the representation of 3 independent experiments. C. Heatmap revealed the DEGs in the cytosolic DNA-sensing pathway between Ctrl and SAMHD1-KD LY1 cells.

Figure 4

STING activation induces multiple forms of cell death to suppress DLBCL cell growth. A. STING overexpression model was established in LY1 and LY3 cells utilizing LV-STING sequences. Transfection efficiency was determined by immunoblot. B. STING-KO model was constructed in LY1 and LY3 cells by transfecting three independent CRISPR/Cas9-mediated STING-KO sequences (STING-KO #1, #2, and #3). Transfection efficiency was determined by immunoblot. C. Annexin V-PE/7AAD double staining flow cytometry revealed the scatter plots (left) and quantitative apoptosis rates (right) in LV-Con and LV-STING DLBCL cells. D. Annexin V-FITC/PI double staining flow cytometry revealed the scatter plots (left) and quantitative apoptosis rates (right) in WT and STING-KO DLBCL cells. E. Microscopic images (left, scale bar=50μm) and supernatant LDH levels (right) of LV-Con and LV-STING DLBCL cells. White arrows indicated cell membrane swelling. F. Microscopic images (left, scale bar=50μm) and supernatant LDH levels (right) of WT and STING-KO DLBCL cells. G. CCK-8 assay revealed the cell viability of LV-STING (left) and STING-KO (right) DLBCL cells, which were compared with empty vectors. H, I. Ctrl and SAMHD1-KD sequences were transfected in STING-KO DLBCL cells. Transfection efficiency was determined by immunoblot (H). Cell proliferation of Ctrl, SAMHD1-KD, Ctrl+STING-KO, and SAMHD1-KD+STING-KO DLBCL cells was compared by CCK-8 assay (I). Immunoblot images in A, B, and H were the representation of 3 independent experiments. Vertical bars indicated mean ± SD. P values from unpaired two-tailed t-test (A, C, D, E, F, and G) and Two-way ANOVA with Sidak correction (I). ^*p<0.05, ^**p<0.01, ^***p<0.001, ^****p<0.0001, and ns=no significance.

Figure 5

STING activates MLKL, CASP3, and GSDME to induce PANoptosis.A-C. Immunoblot revealed the expression of different cell death effectors in LV-Con and LV-STING DLBCL cells, including the effectors of necroptosis (RIPK3, p-RIPK3, MLKL, and p-MLKL) (A), apoptosis (CASP8, cleaved-CASP8, CASP3, and cleaved-CASP3) (B), and pyroptosis (GSDME-FL and GSDME-N) (C). D. Interactions between CASP8, RIPK3, and ASC in LV-STING DLBCL cells were detected by CO-IP assay. E, F. Immunoblot revealed the expression of cell death effectors in WT and STING-KO DLBCL cells, including the effectors of necroptosis (RIPK3, p-RIPK3, MLKL, and p-MLKL) (E), apoptosis and pyroptosis (CASP3, cleaved-CASP3, GSDME-FL, and GSDME-N) (F). G. Immunoblot revealed the protein levels of p-RIPK3, p-MLKL, cleaved-CASP3, and GSDME-N in LY1 cells transfected with Ctrl, SAMHD1-KD, Ctrl+STING-KO, and SAMHD1-KD+STING-KO. Immunoblot images were the representation of 3 independent experiments.

Figure 6

DMXAA inhibits DLBCL cell growth by inducing cell death. A. LY1 and LY3 cells were treated with the concentration gradients of DMXAA for 24 hours. CCK-8 assay revealed the cell viability and IC50 value in DMXAA-treated cells (LY1 IC50=177μM, LY3 IC50=165μM). B-F. LY1 and LY3 cells were treated with DMSO or 177μM DMXAA for 24 hours. STING expression was measured by immunoblot analysis. Wild-type cells without treatment were the blank control (Blank) (B). Scatter plots (left) and quantitative apoptosis rates (right) were revealed by Annexin V-FITC/PI double staining flow cytometry (C). Supernatant LDH levels were detected by LDH release assay (D). Immunoblot showed the protein levels of cell death effectors in LY1 (E) and LY3 (F) cells. G. WT and STING-KO DLBCL cells were treated with DMSO or 177μM DMXAA for 24 hours. Cell viability was determined by CCK-8 assay. H. LV-Con and LV-SAMHD1 LY1 cells were treated with DMSO or 177μM DMXAA for 24 hours. Cell viability was determined by CCK-8 assay. Vertical bars indicated mean ± SD. P values from unpaired two-tailed t-test (A, C, D, F, G, H). ^*p<0.05, ^**p < 0.01, ^***p < 0.001, and ^****p < 0.0001.

Figure 7

DMXAA enhances the efficacy of BMS1166 in DLBCL. A. LY1 and LY3 cells were treated with DMSO or BMS1166 (3, 6, 9, 12μM) for 24 or 48 hours. Cell viability was detected by CCK-8 assay. B. Immunoblot revealed the expression of PD-L1 proteins in LV-Con, LV-STING, WT, and STING-KO DLBCL cells. Wild-type cells without treatment were the blank control (Blank). C, D. DLBCL cells transfected with LV-Con, LV-STING, WT, and STING-KO were treated with DMSO or 9μM BMS1166 for 24 hours. Cell viabilities of LV-STING (C) and STING-KO (D) groups were determined by CCK-8 assay. E. LY1 cells were treated with DMXAA and BMS1166 at a concentration ratio of 88.5:6. Cell viability (upper) and CI values (lower) of different regiments were presented after 24 hours of incubation. F-H. Balb/c nude mice were injected with LY1 cells and randomized into 4 groups. Groups 1-3 were set for monotherapy, where mice were injected with drug-free control or 250μg/ml BMS1166 or 20mg/kg DMXAA every two days. Group 4 was applied for drug combination, where mice were alternately injected with 250μg/ml BMS1166 and 20mg/kg DMXAA every other day (F). Tumor volumes were measured every other day from day 10 (G). Images of tumor bodies were taken on day 18 (H, n=5/group, scale bar=1cm). Immunoblot images were the representation of 3 independent experiments. Vertical bars indicated mean ± SD. P values from unpaired two-tailed t-test (C, D), Welch's one-way ANOVA test with Dunnett's T3 test (E), and Two-way ANOVA analysis with Sidak correction (A, G). *p<0.05, ^**p<0.01, ^***p<0.001, and ^****p<0.0001.

Figure 8

A proposed model of STING-mediated PANoptosis in DLBCL. SAMHD1 deficiency induced DNA damage to promote STING activation. Activation of STING led to the formation of CASP8/RIPK3/ASC complex, further activating MLKL, CASP3, and GSDME to induce PANoptosis. Specifically, MLKL phosphorylation induced necroptosis. CASP3 cleavage not only induced apoptosis but also cleaved GSDME to induce pyroptosis.