Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Nov 1.
Published in final edited form as: Cancer Discov. 2024 May 1;14(5):752–765. doi: 10.1158/2159-8290.CD-23-0700

TREX1 inactivation unleashes cancer cell STING-interferon signaling and promotes anti-tumor immunity

Tetsuo Tani 1,7, Haritha Mathsyaraja 2,7, Marco Campisi 1, Ze-Hua Li 1, Koji Haratani 1, Caroline G Fahey 1, Keiichi Ota 1, Navin R Mahadevan 1,3, Yingxiao Shi 1, Shin Saito 1, Kei Mizuno 1, Tran C Thai 1, Nobunari Sasaki 4, Mizuki Homme 4, Choudhury Fabliha B Yusuf 1,5, Adam Kashishian 2, Jipsa Panchal 2, Min Wang 2, Benjamin J Wolf 2, Thanh U Barbie 1,6, Cloud P Paweletz 1,5, Prafulla C Gokhale 1,5, David Liu 1, Ravindra Uppaluri 1, Shunsuke Kitajima 1,4,*, Jennifer Cain 2,*, David A Barbie 1,*
PMCID: PMC11062818  NIHMSID: NIHMS1981734  PMID: 38227896

Abstract

A substantial fraction of cancers evade immune detection by silencing STING (Stimulator of Interferon Genes)-interferon (IFN) signaling. Therapeutic reactivation of this program via STING agonists, epigenetic or DNA damaging therapies can restore anti-tumor immunity in multiple pre-clinical models. Here we show that adaptive induction of three prime exonuclease 1 (TREX1) restrains STING-dependent nucleic acid sensing in cancer cells via its catalytic function in degrading cytosolic DNA. Cancer cell TREX1 expression is coordinately induced with STING by autocrine IFN and downstream STAT1, preventing signal amplification. TREX1 inactivation in cancer cells thus unleashes STING-IFN signaling, recruiting T and NK (natural killer) cells, sensitizing to NK cell derived IFNγ, and co-operating with PD-1 blockade in multiple mouse tumor models to enhance immunogenicity. Targeting TREX1 may represent a complementary strategy to induce cytosolic DNA and amplify cancer cell STING-IFN signaling, as a means to sensitize tumors to immune checkpoint blockade (ICB) and/or cell therapies.

Introduction

Immune checkpoint blockade (ICB) has emerged as a critical therapeutic strategy for a variety of malignancies, including lung cancer and melanoma, and has significantly improved clinical outcomes for these patients (1). However, the majority of cancer patients fail to respond durably to ICB, and emergence of resistance poses a major challenge. Several mechanisms of resistance to ICB have been identified, including a low degree of tumor mutation burden (TMB), loss of antigen presentation via MHC class I, and impaired or ineffective immune cell infiltration (2).

STING (Stimulator of Interferon Genes) is a critical mediator of innate immunity that plays a key role in DNA sensing downstream of cGAS, particularly in the context of viral infections. While STING-IFN (interferon) signaling has important functions in the tumor immune microenvironment (TME), the important role of STING signaling in cancer cell immunogenicity has been increasingly appreciated (3-5). Perhaps the best evidence for this is that a variety of cancers epigenetically silence STING as a mechanism of immune evasion (3,4,6-8). By avoiding this type I IFN response, tumor types such as small cell lung cancer, KRAS-LKB1 (KL) mutant non-small cell lung cancer, and certain melanomas are able to resist ICB, despite their elevated TMB (3,4,9). Furthermore, therapeutic epigenetic de-repression of STING signaling in combinations with STING agonists or DNA damaging agents can restore immunogenicity in preclinical models (10,11). These data suggest that developing approaches to unleash cGAS-STING activity in cancer cells with suppressed activity is likely to overcome their escape from immune cell detection.

Conisistent with its critical role in cytosolic DNA-sensing, STING pathway activating mutations in the germline induce autoimmune disorders marked by severe “interferonopathy” (12). Importantly, germline loss of function in TREX1, a 3-5’ DNA exonuclease that can degrade cytosolic DNA, is a cause of the type I interferonopathy Aicardi-Goutieres syndrome (12). Indeed, multiple studies in the auto-immunity field have confirmed that TREX1 is a key negative regulator of cGAS STING signaling and important safeguard against the accumulation of cytosolic DNA (13). Recent work has also begun to suggest a potentially important role of TREX1 in cancer. For example, DNA micronuclei, which can accumulate in cancer cells with genomic instability, are partially safeguarded against cGAS by TREX1 following micronuclear envelope rupture (14) and accumulation of cytoplasmic DNA following p53 mediated TREX1 degradation induces tumor suppression (15). However, the more general role of TREX1 in cancer and its adaptive regulation in cancer cells remains poorly characterized.

We recently reported that KL cells suppress cGAS-STING signaling by epigenetically silencing STING and maintaining low levels of intracellular 2’−3’ cGAMP, which can be overcome by pulse MPS1 inhibition (MPS1i) (11). During the course of experiments to characterize this biology in greater detail, we discovered that TREX1 is adaptively upregulated and restrains STING-IFN signal amplification in cancer cells, prompting us to explore its potential as a target to enhance cancer immunotherapy response.

Results

Adaptive induction of TREX1 in response to tumor cell cGAS-STING activation

To understand the transcriptional program induced by micronuclei generated via pulse MPS1i, we performed RNAseq analysis in KL H1944 cells, which are cGAS intact and STINGLow but hypersensitive to pathway activation (11). We treated cells with DMSO or the MPS1i BAY1217389 for 48 h, followed by 24 h washout and RNA sequencing (Fig. 1A). As expected, BAY1217389 pulse treatment induced a potent transcriptional response indicative of restored cGAS-STING signaling, with gene set enrichment analysis (GSEA) identifying interferon (IFN) response signatures as the top induced gene sets (Fig. 1B and C; Supplementary Fig S1A). Examination of the individual genes that were upregulated by pulse MPS1i in H1944 cells revealed numerous IFN stimulated genes (ISGs), antigen presentation machinery (APM) genes such as TAP1, B2M, and HLA genes, as well as STING-TBK1-IRF3 induced chemokines such as CCL5 (Fig. 1D). We also observed the induction of several genes that could restrain downstream cGAS-STING effects, such as NT5E (encoding CD73), which can interfere with STING mediated immunogenicity by generating immunosuppressive adenosine (16,17) (Fig. 1D). Notably, expression of TREX1, a negative regulator of cGAS-STING signaling that degrades cytoplasmic DNA, was significantly increased upon MPS1i treatment. Since TREX1 has also been reported to antagonize cGAS function at micronuclei, we hypothesized it might play an active role in restraining this activated IFN signaling program in KL cells.

Figure 1. Restoration of cGAS-STING signaling in KL cells induces TREX1.

Figure 1.

Open in a new tab

(A) Schematic of protocol for RNA sequencing. H1944 cells treated with DMSO or BAY1217389 for 48h, followed by 24 h washout and RNA sequencing. Schematic created with BioRender.com.

(B) Gene set enrichment analysis (GSEA) showing significant differentially expressed pathways. GSEA hallmark categories of upregulated hallmark pathways in BAY1217389 treated cells relative to DMSO treated cells.

(C) GSEA of IFN-α signature in H1944 cells treated with DMSO or 100 nM BAY1217389.

(D) Volcano plot of H1944 cells treated with DMSO or 100 nM BAY1217389. p < 0.05 and absolute value fold change > 1 were considered significant. Green indicates IFN stimulated genes, pink antigen presentation machinery genes, and blue negative regulator of immune signaling.

(E) Immunoblot of the indicated proteins in H1944 cells transduced with the indicated vectors and treated with DMSO or 100 nM BAY-1217389.

(F-G) Immunoblot (F) or ELISA of human CXCL10 or IFN-β (G) in conditioned-media derived from H1944 transduced with the indicated vectors and treated with DMSO or 100 nM BAY-1217389.

(H) Structure of TREX1 with dsDNA. The structure of TREX1-DNA complex was obtained from PDB (https://www.rcsb.org) ID 7TQQ.

(I-J) Immunoblot (I) or ELISA of human CXCL10 or IFN-β in conditioned-media (J) derived from H1944 transduced with the indicated vectors.

p-values were calculated by two-way ANOVA followed by Sidak’s post-hoc test (G), or one-way ANOVA followed Tukey’s post-hoc test (J), *p<0.05, **p<0.01.

First, we assessed whether pulse BAY1217389 treatment increases TREX1 protein expression. Indeed, pulse treatment of H1944 cells with MPS1i as described earlier increased TREX1 protein levels, with a concomitant elevation of total and activated phosphorylated STAT1 (pSTAT1) levels (Fig. 1E). This increase in TREX1 expression was prevented by CRISPR-CAS9 mediated knockout of STAT1, suggesting that it represents an adaptive response to this IFN signaling program (Fig. 1E). To examine this further, we next validated multiple different TREX1 CRISPR-CAS9 guides to generate TREX1 knockout in H1944 cells (Supplementary Fig S1B). TREX1 inactivation increased pSTAT1 and total STAT1 levels even in the absence of MPS1i (Supplementary Fig S1B). Consistent with these findings, TREX1 knockout potently enhanced both basal and pulse MPS1i induced pSTAT1, STAT1 levels, IFN-β and CXCL10 secretion in H1944 cells (Fig. 1F and G). Taken together, these data confirm that adaptive induction TREX1 in cancer cells can restrain cGAS-STING-STAT1 pathway activation, limiting the magnitude of downstream type I IFN production.

We next examined whether TREX1 catalytic activity is directly required for suppressing this IFN signaling program. We reconstituted H1944 TREX1 knockout cells with TREX1-WT (wild-type), TREX1-D18N or TREX1-D200N, in which key residues important for TREX1 catalytic function have been mutated (Fig 1H) (12,18). Whereas TREX1-WT re-expression suppressed both pSTAT1 and pTBK1 activation following TREX1 knockout, both mutants failed to rescue this phenotype (Fig. 1I). Similarly, TREX1-WT reconstitution potently inhibited secretion of both IFN-β and CXCL10 in TREX1 depleted H1944 cells, in contrast to TREX1-D18N or TREX1-D200N, which had marginal impact (Fig. 1J). These data highlight the specificity of TREX1 CRISPR-CAS9 knockout and demonstrate that TREX1 DNA exonuclease activity is directly required for its ability to suppress cGAS-STING-IFN signaling.

TREX1 depletion unleashes STING-IFN signal amplification in cancer cells

We next sought to corroborate the dependency of this phenotype on expression of tumor cell cGAS and STING, as well as the impact of downstream autocrine IFN-STAT1 signaling. Since some KL cell lines such as H1944 and H1355 cells retain low levels of STING and can still sense and respond to dsDNA (double stranded DNA), while others robustly silence STING (A549 cells) or lose cGAS expression (H2122 cells), we first determined the consequences of TREX1 knockout across these different cell lines. Whereas TREX1 depletion activated STAT1 in both H1944 and H1355 cells, we did not observe this in A549 or H2122 cells (Fig. 2A). TREX1 knockout also increased IFN-β and CXCL10 secretion in both H1944 and H1355 cells, but not in STINGAbsent A549 cells or cGASAbsent H2122 cells (Fig. 2B). To directly test whether enhancement of IFN signaling downstream of TREX1 depletion requires intact cGAS or STING, we performed dual knockout of cGAS or STING together with TREX1 knockout in H1944 cells. As expected, loss of either cGAS or STING prevented induction of pSTAT1 and total STAT1 in TREX1 depleted H1944 cells (Fig. 2C). In addition, cGAS or STING depletion in H1944 cells completely inhibited secretion of IFN-β and CXCL10 in response to TREX1 loss (Fig. 2D). We also confirmed that cGAS knockout uniquely suppressed the enhanced intracellular 2’−3’ cGAMP production induced by TREX1 knockout, as expected, given the role of cGAS but not STING in generating this secondary messenger (Supplementary Fig. S2A). Collectively, these data confirm that TREX1 mediated suppression of IFN signaling in cancer cells depends upon intact cGAS and STING signaling.

Figure 2. TREX1 Depletion Unleashes an Autocrine Signaling Loop that Amplifies Tumor Cell cGAS-STING Signaling.

Figure 2.

Open in a new tab

(A-B) Immunoblot (A) or ELISA of human CXCL10 or IFN-β in conditioned-media (B) derived from KL cells transduced with the indicated vectors.

(C-D) Immunoblot (C), ELISA of human CXCL10 or IFN-β in conditioned-media (D) derived from H1944 transduced with the indicated vectors.

(E-F) ELISA of human CXCL10 or IFN-β in conditioned-media (E), or Immunoblot in H1944 with the indicated vectors at 24hr, 72hr, or 120hr (F).

(G) Schematic of STING pathway suppression by TREX1

(H) Immunoblot of the indicated proteins in H1944 cells transduced with the indicated vectors treated with IFN-β at the indicated concentration at 24hr.

(I) Immunoblot of the indicated proteins in H1944 cells transduced with the indicated vectors.

p-values were calculated by unpaired two-tailed Student’s t test (B, D, E), *p<0.05, **p<0.01. ns, not significant.

Since IFN-β activates STAT1 signaling in tumor cells in an autocrine fashion to amplify ISG cascades, and adaptive induction of TREX1 was dependent on STAT1 (Fig. 1E), we next examined the role of autocrine IFN-β in this phenotype. Indeed, levels of IFN-β and especially downstream CXCL10 were amplified in conditioned media (CM) of TREX1 deleted cells at 72 or 120 hours as compared to 24 hours (Fig. 2E). Additionally, tumor cell STING levels were substantially increased over time, and preferentially induced following TREX1 knockout (Fig. 2F). We therefore considered the possibility that autocrine IFN-STAT1 signaling might be responsible for co-ordinately increasing tumor cell TREX1 in concert with STING expression, as a means of limiting pathway activation (Fig. 2G). Consistent with this hypothesis, exogenous IFN-β treatment for 24 hours increased STING protein and mRNA expression in multiple STING suppressed KL cell lines (Supplementary Fig. S2B and S2C). IFN-β treatment at multiple concentrations also increased both STING and TREX1 protein levels in H1944 cells in an IFNAR1 and STAT1 dependent manner (Fig. 2H).

These findings suggested that activated STAT1 can compete with the H3K27me3 and/or DNA methylation marks that mediate epigenetic silencing of tumor cell STING. To test this directly, we boosted STAT1 levels in H1944 cells by exogenous over-expression. Indeed, compared with control cells, H1944-STAT1 over-expressing cells exhibited increased responsiveness to IFN-β with respect to both STING and TREX1 induction (Supplementary Fig. S2D). Conversely, pretreatment of H1944 cells with an EZH2 inhibitor, and especially combined pre-treatment of A549 cells with both an EZH2 inhibitor and DNMT inhibitor, potentiated the ability of IFN-β treatment to increase STING and TREX1 expression (Supplementary Fig. S2E). These data confirm that IFN-STAT1 signaling strength can overcome STING silencing, and also suggest that reversing STING promoter repression can potentiate this phenomenon even further.

Finally, to test directly whether autocrine type I IFN signaling regulates this feedback loop, we co-deleted IFNAR1 or a control guide together with TREX1 in H1944 cells, and cultured cells in their CM. Knockout of IFNAR1 prevented the increase in STING levels mediated by TREX1 knockout (Fig. 2I). Together these data demonstrate that co-regulation of TREX1 and STING in cancer cells by STAT1 restrains the degree of type I IFN signaling, and that disrupting TREX1 can unleash autocrine and paracrine STING-IFN signal amplification.

TREX1 knockout in cancer cells recruits T cells and primes NK cell activation

To assess how tumor cell TREX1 knockout might affect the tumor immune microenvironment, we performed multiplexed cytokine/chemokine profiling of H1944 cells, ranking cytokine/chemokines by their degree of induction and cGAS dependency. As expected, direct TBK1-IRF3 targets such as CCL5 and CXCL10 were among the top hits, as well as IFNα2 which was included on the array (Fig. 3A and B). This analysis also identified TREX1/cGAS mediated regulation of additional NF-κB induced cytokines such as IL-6, MIP1a, and G-CSF which can have immune suppressive and tumor promoting functions (Fig 3A and B). We therefore sought to examine the net impact of tumor cell TREX1 knockout on immune effector cell function.

Figure 3. TREX1 knockout in cancer cells recruits T cells and primes NK cell activation.

Figure 3.

Open in a new tab

(A) Heatmap of cytokine/chemokine profiling of H1944 TREX1 sg / TREX1 and cGAS sg conditioned media. Data expressed as relative log fold change (L2FC).

(B) Representative absolute cytokine levels (pg/mL) measured using Luminex. Mean ± s.e.m of duplicate samples shown. # indicates values above assay and max CCL5 value used to calculate L2FC.

(C) Schematic of migration assay in 3D device.

(D) Representative images of NK cells migration.

(E) Immune cells infiltration into peri-tumor region

(F) Schematic depicting paracrine interactions between tumor cells and NK cells.

(G) Representative images at 4 hr post addition of 1.5:1 E:T NK cells to H1944 parental or TREX1 sg cells.

(H) Viability of tumor cells in parental and TREX1 sg co-cultures 22 hr post NK cell addition.

(I) IFN-γ production. Data representative of 3 independent experiments.

(J) Flow cytometric analysis of Annexin V and Helix NP of H1944 cells transduced with the indicated vectors treated with IFN-γ at the indicated concentration for 72 hr (left) and the proportion of dead cells is shown (mean ± SD) (right). Data representative of 3 independent experiments. Ratio of dead cells indicates the propotion of Annexin V- and/or Helix NP-stained cells.

p-values were calculated by unpaired two-tailed Student’s t test (B, E, H), and one-way ANOVA followed by Tukey’s post-hoc test (J), *p<0.05, **p<0.01, ns, not significant. Schematics created using BioRender

Since we previously demonstrated that reactivation of tumor cell STING signaling in KL lung cancer cells and CXCL10-CXCR3 signaling in particular promotes NK (natural killer) and T cell migration, we performed microfluidic 3-dimensional co-culture experiments to assess whether tumor cell TREX1 knockout leads to increased effector cell recruitment (Fig. 3C). We labeled primary NK or CD8+ T cells derived from PBMC and added them to the media containing side channel, measuring the degree of infiltration into the central chamber containing H1944 tumor spheroids embedded in collagen. Whereas negligible NK or T cell infiltration was observed over 2 days towards H1944 spheroids expressing a control guide RNA, NK and CD8+ T cells exhibited significantly increased migratory activity towards H1944 spheroids with TREX1 knockout (Fig. 3D and E; Supplementary Fig. S3A). These findings demonstrate that TREX1 inactivation can promote T/NK cell recruitment even in the presence of potentially immune suppressive cytokines.

We next examined whether TREX1 loss might sensitize cells to immune mediated killing, focusing on the impact of NK cells given their emerging role in STING biology (19-21). Although TREX1 deletion resulted in increased surface levels of MHC Class I, which is inhibitory to NK cells, we also observed upregulation of the NK activating stress ligand MICA (Supplementary Fig. S3B). To assess the functional relevance of these collective changes, we performed tumor killing assays wherein H1944 wild type and TREX1 knockout cells were co-cultured with NK cells (Fig. 3F). The addition of NK cells led to a significant increase in TREX1 knockout tumor cell apoptosis compared to control tumor cells (Fig 3G and 3H; Supplementary Fig. S3C). This was accompanied by enhanced IFN-γ production in the TREX1 knockout co-cultures (Fig. 3I). To understand whether IFN-γ itself could mediate this cytotoxicity, we treated cells in isolation with concentrations of IFN-γ similar to what was generated in the co-culture assay, and measured apoptosis and cell death via dual staining for Annexin V and Helix NP. Remarkably, whereas H1944 KL control cells were largely resistant to IFN-γ, TREX1 depletion resulted in a profound increase in cells undergoing active apoptotic cell death in response to IFN-γ exposure (Fig. 3J). Thus, TREX1 loss not only promotes tumor intrinsic amplification of IFN-β-STING signaling, but also renders cells hypersensitive to parallel STAT1 input from IFN-γ.

TREX1 inactivation broadly primes immunogenicity in tumor cells

Several studies have implicated TREX1 as a negative regulator of type I IFN signaling in the context of inflammatory disease (12) and our data thus far implicate TREX1 in restraining cGAS-STING activation in malignant lung cancer cells. To expand on our findings and assess the broader relevance of TREX1 in cancer, we next interrogated publicly available datasets to examine TREX1 expression in normal tissue (GTex) compared to tumor samples (TCGA). Consistent with the notion that TREX1 acts to dampen tumor immunity, we observed significantly increased TREX1 mRNA expression when normalized to normal tissue expression across a range of solid cancers, including breast, ovarian, uterine and colorectal (Fig. 4A). We also examined the baseline distribution of TREX1 expression across normal tissues, observing already high levels of expression in certain tissues such as lung (Supplementary Fig. S4A). Furthermore, TREX1 expression was maintained in KRAS mutant LUAD and not significantly different on average between KL and KP genotypes within the TCGA cohort (Supplementary Fig. S4B). Additionally, we explored the relationship between TREX1 expression and STING signaling in clinical melanoma tumor cohort data of patients treated with immune checkpoint blockade (ICB) (22). Whereas baseline expression of TREX1 did not predict anti-PD-1 response, this analysis revealed a significant correlation between TREX1 mRNA levels, STING and IFN pathway signatures (Supplementary Fig. S4C and S4D), consistent with our in vitro data demonstrating that TREX1 itself is a STAT1 inducible gene.

Figure 4. TREX1 inactivation broadly primes immunogenicity in tumor cells.

Figure 4.

Open in a new tab

(A) Expression of TREX1 in normal tissue(Gtex). vs. tumor (TCGA) in multiple types of malignant tumors. PRAD, prostate adenocarcinoma; UCEC, uterine corpus endometrial carcinoma; SARC, sarcoma; OV, ovarian cancer; BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; KIRC, kidney renal clear cell carcinoma; CRC, colorectal cancer; STAD, stomach adenocarcinoma; LIHC, liver hepatocellular carcinoma.

(B) Fold induction in IFN-β levels (qPCR) in TREX1 siRNA knockdown and non-targeting (NT) control siRNA in cancer cell lines. Data collected 6 days post siRNA transfection. Mean ± SEM of n = 3 biological replicates are shown.

(C) ELISA of human CXCL10 in conditioned-media from KLE with the indicated vectors at 24hr, 72hr, or 120hr. **p<0.01 using unpaired two-tailed Student’s t test.

(D) Tumor volume of wild-type and TREX1 KO tumors in combination with isotype or anti-PD-1 treatment. n=10 mice per group (WT vs. TREX1 KO isotype, **p=0.006, 2 way Anova at Day 17)

(E-F) CD3+ T cell (E) and NK cell (F) and percentages in all groups indicated above (n=5 for each group). For all comparisons in E and F: **p<0.01, *** p<0.001, **** p<0.0001 using one-way ANOVA.

(G) Gene set enrichment scores for IFN related signaling in the four groups (expression data from RNA-sequencing).

(H) Tumor growth in the CT26 model. Day 0 = start of anti-PD-1 (TV ~100-160mm3). n=15 each group. When TV reached 500-600mm3, n=5 removed from study; n=10 at endpoint. ****p<0.0001 (2-way ANOVA) .

(I) Mean tumor volume of 393P-KL cells after subcutaneous inoculation into 129S2/SvPasCrl mice. Mice were treated with anti-PD-1 antibody on day 7, 10 after implantation. *** p<0.001, ****p<0.0001 (2-way ANOVA).

(J) Mean tumor volume of 393P-KL cells after subcutaneous inoculation into NSG (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) female mice. Mice were treated with anti-PD-1 antibody on days 1 and 4 after randomization. ns, not significant (2-way ANOVA).

To directly confirm the role of elevated TREX1 in diverse cancer types, we suppressed its expression in multiple additional contexts and evaluated type I IFN induction. TREX1 knockdown by siRNA led to increased IFN-β levels in all cancer cell lines tested (Fig. 4B) as well as viability impairment across several of these lines (Supplementary Fig. S4E). To confirm that the effect of siRNA was consistent with the previous data of gene knockout by CRISPR-Cas9, we deleted TREX1 in KLE cells, a cell line derived from a poorly differentiated endometrial carcinoma. Similar to our findings in lung cancer cells, TREX1 knockout increased pSTAT1 (Supplementary Fig. S4F), STING expression (Supplementary Fig. S4G) and CXCL10 over time (Fig. 4C). Thus, TREX1 also suppresses STING-IFN signaling in additional cancer types, limiting feedforward signal amplification.

While these in vitro studies demonstrate that TREX1 knockout activates type I IFN signaling and T/NK effector cell recruitment, deletion of TREX1 in an in vivo context is critical to determine its net impact on tumor immunogenicity. To test the role of TREX1 in mediating immune evasion in vivo, we implanted TREX1 knockout (labeled TREX1 sg) tumor cells using ICB refractory B16F10 melanoma as a model. TREX1 knockout tumors displayed a significant increase in NK cell infiltration (Supplementary Fig. S4H) and an enrichment in immune related transcripts (Supplementary Fig. S4I). Given the strong immune activation observed upon TREX1 loss in this traditionally immune suppressive and refractory model, we asked whether TREX1 inactivation is capable of conferring sensitivity to anti-PD-1 therapy. Anti-PD-1 treatment of mice bearing TREX1 knockout tumors resulted in decreased tumor growth relative to control tumors (Fig. 4D), and resulted in increased infiltration of both CD3+ T cells (Fig. 4E) and NK cells (Fig. 4F). In addition, TREX1 knockout tumors upregulated IFN related pathway gene expression, which was further enhanced upon combination with anti-PD-1 (Fig. 4G). Additionally, similar to our finding in human cell lines, TREX1 inactivation in B16F10 cells cooperated with MPS1 inhibition to further upregulate STAT1 in vitro (Revised Supplementary Figs. S4J). Thus, MPS1 inhibition (pulse BAY1217389 treatment by oral gavage) or direct STING agonism (ADU-S100 intratumoral injection) similarly co-operated with TREX1 knockout to suppress growth of B16F10 tumors in vivo (Supplementary Fig. S4K, L).

To extend these findings, we examined the impact of TREX1 knockout in two additional syngeneic models – CT26 colorectal cancer and 393P-KL lung cancer. Consistent with our findings in B16F10 melanoma, we observed a significant reduction in tumor growth and IFN gene expression in CT26 tumors, which was significantly enhanced by PD-1 blockade (Fig 4H; Supplementary Fig. S4M). Since 393P-KL cells in particular exhibited no anti-proliferative defect following TREX1 inactivation in vitro (Supplementary Fig. S4N), we utilized this model to examine whether tumor growth suppression with or without anti-PD-1 is immune cell dependent. In contrast to treatment studies in syngeneic, immune intact 129S2/SvPasCrl mice, where TREX1 knockout and especially the combination with PD-1 blockade resulted in tumor shrinkage, treatment with an identical regimen in Nod-Scid Gamma (NSG) mice resulted in no impact on tumor growth (Fig. 4I and 4J). These data demonstrate that TREX1 inactivation leads to a functional immune response that directly promotes tumor growth reduction in vivo. Taken together with our in vitro studies, these findings are strongly supportive of a broader role for TREX1 as a negative regulator of anti-tumor immune responses.

DISCUSSION

Tumor cells employ several mechanisms to evade anti-tumor immune responses, including the downregulation or silencing of cytosolic nucleic acid sensing pathway components to dampen type I IFN signaling (23). For example,suppression of STING signaling through epigenetic silencing promotes immune evasion in KRAS/LKB1 mutant non-small cell lung cancer, as well as melanoma (4,10). More generally, activation of tumor cell STING has also been proposed as a critical mechanism in preventing lung cancer cell metastasis (5). Since restoring STING expression and activation via micronuclei induction can restore IFN signal activation and resensitize tumors to ICB (11), we sought therapeutic targets that could restrain this program, uncovering a key role for adaptive TREX1 upregulation. This screen also identified induction of CD73, which can also antagonize cGAS-STING induced immunogenicity via enhanced production of adenosine, along with ENPP1 (16,17). However, in contrast to adenosine, which largely impacts the TME, TREX1 is a critical factor in preventing STING-IFN signal amplification in cancer cells themselves, unleashing tumor cell immunogenicity. Furthermore, irrespective of CD73 induction or the generation of several immune suppressive cytokines following TREX1 knockout, the net impact recruits and sensitizes cancer cells to NK mediated killing and enhances sensitivity to ICB in vivo, highlighting its potential as a cancer therapeutic target.

Malignant cells accumulate aberrant cytoplasmic DNA due to high replication stress and elevated levels of DNA damage (24). Not surprisingly, TREX1 expression is elevated in many types of cancers compared to normal tissues, suggesting that TREX1 induction may act as an alternative to STING silencing as a mechanism of promoting immune escape during tumor development (24). Our finding that TREX1 is adaptively upregulated in response to activation of tumor cell cGAS-STING signaling further suggests that certain cancers might also induce TREX1 expression as a mechanism of escape from therapy induced immune pressure. However, TREX1 inactivation failed to induce pathway activation in tumor cells with robust, DNMT1 mediated silencing of STING, or in the absence of cGAS expression. These data suggest that TREX1 targeting will be most active in contexts where cGAS and STING expression is maintained to overcome anti-PD-1 resistance.

Importantly, STING is almost never mutated at baseline in cancer cells (25), and the preferred pathway of suppressing its expression involves epigenetic silencing. Since TREX1 inactivation could overcome EZH2 mediated silencing of STING in STINGLow H1944 cells, it is possible that feedforward IFN-β-STAT1 signaling can still restore its expression in most contexts. Yet it also possible that TREX1 inhibition could increase selective pressure for cells to eliminate cGAS or STING expression, via DNMT mediated silencing for example. However, such adaptations could be highly reversible through interventions such as DNMT inhibition, which resensitizes cells to IFN mediated STING and TREX1 induction (Supplementary Fig. S2E). Future studies examining how cells adapt to chronic TREX1 suppression in the context of immune selective pressure will help to answer how cancer cells might avoid this immunogenic program.

TREX1 was originally shown to limit activation of IFN signaling in normal tissues in autoimmune diseases such as Aicardi-Goutieres syndrome by suppressing the accumulation of reverse transcribed endogenous retroelements (26). It is thus possible that TREX1 inhibition in cancer cells could also promote accumulation of human endogenous retroviruses (HERVs) and/or LINE elements, especially following DNMT inhibition. Furthermore, while chronic therapeutic targeting of TREX1 may potentiate autoimmunity, cancer cells preferentially de-repress retroelements and survive in the face of DNA damage in contrast to normal tissues, favoring a therapeutic window especially with pulse treatment.

Indeed, our data also reveal that inhibition of tumor cell TREX1 can potently lower the cytotoxicity threshold of cancer cells to IFN-γ exposure. Coupled with the induction of MHC I, this mechanistically supports the enhanced antigenicity and sensitivity to immune checkpoint blockade. We also demonstrate that NK cells, which are known to be recruited by cGAS-STING signaling (19), are able to kill TREX1 depleted cancer cells despite mixed inhibitory and activating signals, and recruited in vivo. Fostering even more robust NK cell recruitment, by coupling TREX1 inhibition with NK cell therapy for example, could result in rapid elimination of tumors, and limit the potential for resistance. Since the effect of endogenous NK cell depletion could not be demonstrated in this study, future studies will be considered to address this issue. Furthermore, we recently demonstrated STING agonist induced CXCL10 can recruit CAR-NK cells targeting mesothelin, which can further overcome NK inhibitory signals by signaling via the CAR (27). Thus, even in cell contexts that may not display IFN-γ hypersensitivity following TREX1 inhibition, unleashing cGAS-STING in this manner could recruit engineered cell therapies. Additionally, while direct STING agonist mediated T cell toxicity limited the ability to couple this approach with CAR-T cells, targeting tumor cell TREX1 would obviate this issue.

More generally, our analysis of TCGA data suggests that TREX1 is widely upregulated in cancer, possibly to counteract micronuclei formation and cytosolic DNA accumulation intrinsic to many cancers. It is also reported that TREX1 is induced upon exposure to extrinsic DNA damage inducers such as radiation and chemotherapy (28,29). Indeed, pulse MPS-1 inhibition or direct STING agonism was also effective in combination with TREX1 knockout in vivo, suggesting additional therapeutic combinations to explore. Thus, TREX1 inhibition has broad potential to re-shape the TME through including cytotoxic immune cell recruitment and activation, both in the context of tumor cell intrinsic and extrinsic sources of DNA damage.

Materials and Methods

RNA-sequencing

RNA extraction was performed using RNeasy Mini Kit (Qiagen, Cat.# 74106). Total RNA concentration was calculated by Quant-IT RiboGreen (Invitrogen, #R11490). To assess the integrity of the total RNA, samples are run on the TapeStation RNA screentape (Agilent, #5067-5576). Only high-quality RNA preparations, with RIN greater than 7.0, were used for RNA library construction. A library was independently prepared with 1ug of total RNA for each sample by Illumina TruSeq Stranded mRNA Sample Prep Kit (Illumina, Inc., San Diego, CA, USA, #RS-122-2101). The first step in the workflow involves purifying the poly-A-containing mRNA molecules using poly-T-attached magnetic beads. Following purification, the mRNA is fragmented into small pieces using divalent cations under elevated temperatures. The cleaved RNA fragments are copied into first strand cDNA using SuperScript II reverse transcriptase (Invitrogen, #18064014) and random primers. This is followed by second-strand cDNA synthesis using DNA Polymerase I, RNase H, and dUTP. These cDNA fragments then go through an end repair process, the addition of a single ‘A’ base, and then ligation of the adapters. The products are then purified and enriched with PCR to create the final cDNA library. The libraries were quantified using KAPA Library Quantification kits for Illumina Sequencing platforms according to the qPCR Quantification Protocol Guide (KAPA BIOSYSTEMS, #KK4854) and qualified using the TapeStation D1000 ScreenTape (Agilent Technologies, # 5067-5582). Indexed libraries were then submitted to an Illumina NovaSeq (Illumina, Inc., San Diego, CA, USA), and the paired-end (2×100 bp) sequencing was performed by Macrogen Incorporated.

Differential gene expression and pathway analysis with RNA-seq data

Differential gene expression between the treatment of BAY1217389 and DMSO was calculated using DESeq2 (RRID:SCR_000154). Differentially expressed genes (DEGs) were selected based on the cut-off criterion (adjusted P value < 0.05 and ∣log2 (fold change)∣ > 0.58). Enrichment analyses for upregulated DEGs were performed using Metascape (http://metascape.org/, RRID:SCR_016620), which processed data combining Gene Ontology (GO)/ Kyoto Encyclopedia of Genes and Genomes (KEGG) terms, Reactome pathways, and Wiki pathways. Additionally, the protein-coding genes were pre-ranked based on log2 (fold change), and Gene Set Enrichment Analysis (GSEA) was conducted using the pre-ranked gene list with 10,000 permutations. The Molecular Signatures Database (MSigDB) hallmark gene set was used for the analysis.

Cell lines

A549 (RRID:CVCL_0023), H1355 (RRID:CVCL_1464), and H2122 cell lines (RRID:CVCL_1531) were obtained from the Broad Institute and authenticated by STR genotyping. H1944 (RRID:CVCL_1508), HEK293T (RRID:CVCL_0063), HCC44 (RRID:CVCL_2060), KLE (RRID:CVCL_1329), DBTRG-05 (RRID:CVCL_1169), and G402 (RRID:CVCL_1221) and the mouse cell lines B16F10 (RRID:CVCL_0159) and CT26 (RRID:CVCL_7256) were purchased from ATCC. COV362 cells (RRID:CVCL_2420) were procured from Sigma-Aldrich. 393P cells were established from KrasLA1/+;p53R172HΔG mice and kindly gifted from Dr. J.M. Kurie (The University of Texas, MD Anderson Cancer Center, Houston, TX). A549 and HEK293T cells were cultured in DMEM (Thermo Fisher Scientific, Cat.# 11965-118) supplemented with 10% fetal bovine serum (FBS) (Gemini Bio-products, Cat.# 100-106), and 1x penicillin-streptomycin (Gemini Bio-products, Cat# 400-109). H1944, H1355, HCC44, H2122, and DBTRG-02 cells were cultured in RPMI 1640 (Thermo Fisher Scientific, Cat.# 11875-119) supplemented with 10% FBS, and 1x penicillin-streptomycin. KLE, COV362 cells were cultured in DMEM/F12 (Thermo Fisher Scientific, Cat.# 11320-033) supplemented with 10% fetal bovine serum (FBS) (Gemini Bio-products, Cat.# 100-106). G402 cells were cultured in McCoy’s 5a Medium Modified (Thermo Fisher Scientific, Cat.# 16600-108) supplemeted with 10% FBS (Gibco Cat #16140-071).CD8+ T cells were isolated from peripheral blood mononuclear cells (PBMCs) (STEMCELL, Cat.# 70025) using the EasySep™ Human CD8+ T Cell Isolation Kit (STEMCELL, Cat.# 17953) following the manufacturer's instructions. The cells were cultured in RPMI 1640 supplemented with 10% human serum (Sigma-Aldrich, Cat.# H5667), 1x penicillin-streptomycin, 2 mM L-glutamine, and 100 IU/mL IL-2, 25 ng/mL IL-7, and 25 ng/mL IL-15. The T cells were immediately activated with 1% T cell TransAct (Miltenyi Biotec, Cat.# 130-128-758) after isolation. NK cells were isolated from PBMCs (STEMCELL, Cat.# 70025) using the EasySep™ Human NK Cell Isolation Kit (STEMCELL, Cat.# 17955) following the manufacturer's instructions. The NK cells were cultured in NK MACS® Medium supplemented with 5% human serum (Sigma-Aldrich, Cat.# H5667), and 500 IU/mL IL-2. Mycoplasma infection was regularly checked by PCR using the condition media from each cell line. The sequences of the primers for checking Mycoplasma infection are listed in Supplementary Table S1. All experiments were performed within 10 passages from the original frozen stocks.

Reagents

The following reagents were used: BAY-1217389 (Selleckchem, Cat.# S8215), Tazemetostat (Selleckchem, Cat.# S7128), Decitabine (Selleckchem, Cat.# S1200), IFN-β (R&D systems, Cat.# 8499-IF), IFN-γ (R&D systems, Cat.# 285-IF-100), and Phorbol 12-myristate 13-acetate (Medchemexpress, Cat.# HY-18739) (Selleckchem, Cat.# S7061).

Cytokine profiling

Luminex cytokine/chemokines profiling was performed using the Bio-Plex ProTM 40-plex human chemokine panel according to the manufacturer’s instructions. Fold changes relative to the control were calculated and plotted as log2-fold change. Lower and upper limits of quantitation were imputed from standard curves for cytokines above or below detection. Human IFN-β (Thermo Fisher Scientific, Cat.# 414101), human CXCL10 (R&D systems, Cat.# DIP100), and 2’3’-cGAMP (Cayman Chemical, Cat.# 501700) ELISAs were performed according to the manufacturer’s instructions. Values represent the average of four replicates from at least two independent experiments (biological replicates).

Generation of lentivirus

3 x 106 HEK293T cells were plated onto a 60-mm dish and transfected using X-tremeGENE HP DNA Transfection Reagent (Roche, Cat.# 06366236001) with 1 μg of lentivirus-based expression vectors together with 1 μg of pCMV-dR8.91 and 1 μg of pCMV-VSV-G. After 48hr incubation, the media containing lentivirus particles were collected, passed through a 0.45 μm filter, and concentrated using Lenti-X Concentrator (Clontech, Cat.# 631231). For selection of virally infected cells, 1 μg/ml of puromycin (pCRISPR-v2 sgRNAs) or 6 μg/ml of blasticidin (plx304-NanoLuc, plx304-TREX1 wild type, plx304-TREX1 D18N, plx304-TREX1 D200N) was used 24 hr post-infection.

Immunoblotting

Cells were lysed in RIPA buffer containing 1x protease inhibitor cocktail tablets (Roche, Cat# 11-836-145-001) and phosphatase Inhibitor cocktail tablets (Sigma, Cat# 4906845001). Immunoblotting was performed as described (30) using following antibodies to: cGAS (#15102, Cell Signaling Technology, RRID: AB_2732795), STING (#13647, Cell Signaling Technology, RRID: AB_732796), phospho-STAT1 (#9167, Cell Signaling Technology, RRID: AB_561284), STAT1 (#9172, Cell Signaling Technology, RRID: AB_2198300), IFNAR1 (A304-290A, Thermo Fisher, RRID: AB_2620486), phospho-TBK1 (#5483, Cell Signaling Technology, RRID: AB_10693472), TBK1 (#3013, Cell Signaling Technology, RRID: AB_2199749), human TREX1 (ab185228, abcam, RRID:AB_2885196), mouse TREX1 (SC113112,Santa Cruz, RRID:AB_2208802) and β-Actin (#3700, Cell Signaling Technology, RRID: AB_2242334). Secondary antibodies were from LICOR Biosciences: IRDye 680LT Goat anti-Mouse IgG (#926-68020), IRDye 800CW Goat anti-Rabbit IgG (#926-32211), or Cell Signaling Technology: Imaging of blots was performed using the LICOR Odyssey system.

CRISPR/Cas9 system and siRNA knockdowns

Target sequences for CRISPR interference were designed using the sgRNA designer (http://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design). A non-targeting sgRNA from the Get-go library v2 was used as a scramble sgRNA. sgRNA target sequences are listed in Supplementary Table S1. sgRNAs were cloned into pCRISPRv2-puro. For tumor killing assays, a transient CRISPR-RNP approach (Alt-R™ CRISPR-Cas9 from IDT technologies sgRNA 29 – Hs.Cas9.TREX1.1.AX) was utilized, followed by expansion of single cell clones. For siRNA knockdown experiments, cells were transfected with non-targeting siRNA (ThermoFisher #4390843) or TREX1 siRNA (ThermoFisher # 4392420 silencer select #535161) using Lipofectamine RNAiMax (ThermoFisher #13778-100) per manufacturer recommendations.

Quantitative RT-PCR

RNA extraction was performed using RNeasy Mini Kit (Qiagen, Cat.# 74106). RNA samples (1 μg) were reverse-transcribed using SuperScript® III First-Strand Synthesis SuperMix (Thermo Fisher Scientific, Cat.# 1683483). Quantitative real-time PCR was performed using Power SYBR Green PCR Master Mix (Thermo Fisher Scientific, Cat.# 4367659). Values represent the average of four technical replicates from at least three independent experiments (biological triplicates). For IFNB mRNA quantification, RNA isolation was performed using the RNeasy Micro kit (Qiagen # 74004). cDNA was made using the high capacity cDNA reverse transcription kit (ThermoFisher #4368813) and the Taqman Fast advanced master mix (ThermoFisher #4444557) was used for qPCR. The sequences of the primers used for qRT-PCR are listed in Supplementary Table S1.

Immune cell migration assay

Immune cell migration assay was performed as previously described (4,31). Briefly, H1944 cancer cell spheroids were generated by seeding 5 x 105 cells in suspension in an ultra-low attachment dish (Corning, Cat.# 3471) for 24 hr. Samples were stained with cell blue dye eFluor 450, (Invitrogen, cat.# 65-0842) pelleted and then resuspended in type I rat tail collagen (Corning) at a concentration of 2.5 mg/mL following the addition of 10× PBS with phenol red with pH adjusted using NaOH. pH 7.0–7.5 was confirmed using PANPEHA Whatman paper (Sigma-Aldrich). Cells and collagen are kept on ice. The spheroids-collagen suspension was then injected into the central gel region of the 3D DAX-1 3-D microfluidic cell culture chip (AIM Biotech, Singapore, Cat.# DAX-1). Microfluidic devices were designed as previously described (32), with a central region containing the cell-collagen mixture in a 3D microenvironment, surrounded by 2 media channels located on either side. After injection, collagen hydrogels containing cells were incubated 40 min at 37°C in humidity chambers, then hydrated with culture media, with 5 x 104 CD8+ T cells or NK cells in one of the side media channels. These immune cells were labeled with Cell Tracker Red (Thermo Fisher Scientific, Cat.# C34552) following manufacturer’s instructions. After 48 hr of incubation, cancer cell spheroids and infiltrated immune cells were captured on a Nikon Eclipse 80i fluorescence microscope equipped with Z-stack (Prior) and CoolSNAP CCD camera (Roper Scientific). Image capture and analysis was performed using NIS-Elements AR software package. Whole device images were achieved by stitching in multiple captures. Quantification of cell migration was performed by measuring the total cell area of cell tracker dye in the entire gel region by region of interest (ROI).

Flow cytometry assay

Cells were examined for surface expression by flow cytometry with BD Biosciences LSRFortessa and the following antibodies: anti-HLA-A,B,C (BioLegend #311436, RRID:AB_2566254), anti-MICA (BD Biosciences #749779, RRID:AB_2874033), or isotype IgG control antibodies (BioLegend #400268, RRID:AB_2734131, BD Biosciences # 562748, RRID:AB_2721018). Cells resuspended in 100 μL PBS containing 3% FBS were stained by each antibody.

Apoptosis assay

Cells (200,000/well) were seeded in 6-well plates and treated with PBS or human IFN-γ (R&D Systems, #285-IF-100) for 72 hours. Annexin V (BioLegend #640912) and Helix NP™ Green (BioLegend #425303) are diluted in Annexin V Binding Buffer (Biolegend, #422201) following manufacture protocol. The cells are stained with this buffer for 15 minutes at room temperature. The BD LSRFortessa™ Cell Analyzer analyzed the cell sample within a half hour. FlowJo v10 (RRID:SCR_008520) was used to perform the analysis of flow cytometry raw data.

Tumor killing assays

H1944 parental and TREX1 sg cells were tagged with Nuclight Red lentivuris (Sartorius, 4476). Primary human NK cells were purchased from Stem Cell Technologies (70036). NK cells were cultured in X-VIVO 15 media (Lonza, 04-418Q) supplemented with heat inactivated FBS (Gibco, 16140-071) and rIL-15 (Biolegend, 570304) . Cell killing was monitored on the Incucyte using the red signal (viability) and green fluorescence Caspase 3/7 dye (Sartorius 4440). Human IFN-γ (MesoScale Discovery K151AEB-2) was assayed as per manufacturer recommendations.

In vivo studies

TREX1 knockout lines were generated using CRISPR-Cas9 by Horizon Discovery Biosciences in CT26 clonal and B-MoGen Biotechnologies (Biotechne) in B16F10. B16 and CT26 studies were run at Crown Biosciences in accordance with IACUC approved protocols. For CT26 studies, 1X105 parental and TREX1 knockout cells in 100ul serum free medium were implanted subcutaneously into BALB/c mice (RRID:IMSR_ORNL:BALB/cRl). When tumors reached 80-120mm3, mice were randomized and dosed with either 10mg/kg isotype control or anti-PD-1. Animals were euthanized when tumor volume reached 3000mm3 or >20% body weight loss occurred. RNA was harvested from tumors at endpoint using the RNeasy Mini Kit (Qiagen #74134) and analyzed using the Immunology panel (Nanostring Technologies). For B16 anti-PD-1 studies, 2 x 105 Puro control and TREX1 knockout cells in 100ul PBS were implanted subcutaneously into C57B6 (RRID:IMSR_JAX:000664, 15 mice per group). When tumors reached 80-120mm3, mice were randomized and dosed with either 10mg/kg isotype control or anti-PD-1 with a Q4D dosing frequency. Five mice per group were removed 24 hrs after final dose to profile tumor infiltrating lymphocytes via flow cytometry. Study was terminated when mean tumor volume of any group reached 2000mm3 or 4 weeks after initiation. Endpoint tumors were harvested and RNA was isolated using RNeasy Mini Kit (Qiagen #74134). RNA-sequencing studies was performed at Q2 solutions. Illumina TruSeq Stranded mRNA sample preparation kits were utilized and 50bp paired-end sequencing was performed. For TIL profiling, tumors were dissociated using a kit from Miltenyi (130-096-730) according to manufacturer’s instructions. The following antibodies were used for profiling: live/dead eFlour780 (eBioscience, 65-0865-18), α-CD3e/17A2-BUV395 (BD Biosciences, 740268, RRID:AB_2687927), α-CD49b/BV911 (BD Biosciences, 740704, RRID:AB_2740388), α-CD45/30-F11-FITC (BioLegend, 103108, RRID:AB_312973).

For B16 BAY-1217389 or ADU-S100 study, 5 x 106 scramble and TREX1 knockout cells in 100ul PBS were implanted subcutaneously into C57B6 and followed by 5mg/kg BAY-1217389 on day 7, 8,13 and 14 or 10ug ADU-S100 on day 7, 10 and 13. n=8 or 9 mice per group.

BAY-1217389 was formulated in 50% PEG 400, 10% ethanol and 40% water and dosed at 5 mg/kg twice daily by oral gavage. ADU-S100 was formulated in HBSS (Gibco, 14025092) and dosed at 10 μg by intratumoral injection.

For 393P-KL cells, all experiments were conducted according to a Dana-Farber Cancer Institute approved protocol. 393P-KL cells were generated by LKB1 knockout (11) and 393P-KL scramble and TREX1 sg were established using CRISPR-Cas9. Five million cells (393P-KL Scramble, 393P-KL TREX1 sg) in PBS with 30% Matrigel (Corning, Cat.#356231, NY) were subcutaneously injected into the flank of 8-week-old female 129-Elite mice (129S2/SvPasCrl, Strain code 476, Charles River Laboratories, RRID:IMSR_CRL:476) or NSG (NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ) female mice (The Jackson Laboratory, ME, RRID:IMSR_JAX:000664). Tumor volume was determined from caliper measurements of tumor length (L) and width (W) according to the formula (L × W2)/2. Animals were randomized (Studylog software, CA) into various treatment groups with an average tumor volume of 159.1 mm3. Both tumor size and body weight were measured twice per week. Mice were treated with InVivoMAb IgG2a isotype control (Bio X cell, Cat.#BE0089, RRID: 1107769) or InVivoMAb anti-mouse PD-1 RMP1-14 (Bio X cell, Cat.#BE0146, RRID: 10949053) on days 1 and 4 after 4 after randomization by intraperitoneal injection.

Supplementary Material

1
2
3
4
5

Significance.

STING-IFN signaling in cancer cells promotes tumor cell immunogenicity. Inactivation of DNA exonuclease TREX1, which is adaptively upregulated to limit pathway activation in cancer cells, recruits immune effector cells and primes NK cell mediated killing. Targeting TREX1 has substantial therapeutic potential to amplify cancer cell immunogenicity and overcome ICB resistance.

Acknowledgments

We thank Minyue Chen, Yurie Yamamoto, Connor Purcell, Pieter J Schol, Madelyn Cueva and Chari Cortez for technical assistance, Ryohei Yoshida for scientific discussion and support, and Chloe Deodato for project management support. Research was supported by Uehara Memorial Foundation Research Fellowship (T.T.), Lilly Oncology Fellowship Program (T.T.)

Footnotes

Declaration of Interest

D.A.B. is a consultant for N of One/Qiagen and Nerviano Medical Sciences, is a founder and shareholder in Xsphera Biosciences, has received honoraria from Merck, H3 Biomedicine/Esai, EMD Serono, Gilead Sciences, Abbvie, and Madalon Consulting, and research grants from BMS, Takeda, Novartis, Gilead, and Lilly. C.P.P is a consultant for DropWorks and XSphera Biosciences, has stock and other ownership interests in XSphera Biosciences. Received honoraria from Bio-Rad and has sponsored research agreements with Daiichi Sankyo, Bicycle Therapeutics, Transcenta, Bicara Therapeutics, AstraZeneca, Intellia Therapeutics, Janssen Pharmaceuticals, Array Biopharma. S.K. has a sponsored research agreement with Boehringer-Ingelheim. H.M., A.K., J.P, M.W., B.J.W. and J.C. are employees of Gilead Sciences.

Data availability

The datasets generated during and/or analyzed in this study are available from the corresponding author upon reasonable request.

Sequence data were deposited at Gene Expression Omnibus (GEO) and the accession number for the RNA-seq data reported in this paper is GEO: GSE252340.

REFERENCES

  • 1.Morad G, Helmink BA, Sharma P, Wargo JA. Hallmarks of response, resistance, and toxicity to immune checkpoint blockade. Cell 2021;184(21):5309–37 doi 10.1016/j.cell.2021.09.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Keenan TE, Burke KP, Van Allen EM. Genomic correlates of response to immune checkpoint blockade. Nat Med 2019;25(3):389–402 doi 10.1038/s41591-019-0382-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Falahat R, Perez-Villarroel P, Mailloux AW, Zhu G, Pilon-Thomas S, Barber GN, et al. STING Signaling in Melanoma Cells Shapes Antigenicity and Can Promote Antitumor T-cell Activity. Cancer Immunol Res 2019;7(11):1837–48 doi 10.1158/2326-6066.Cir-19-0229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Kitajima S, Ivanova E, Guo S, Yoshida R, Campisi M, Sundararaman SK, et al. Suppression of STING Associated with LKB1 Loss in KRAS-Driven Lung Cancer. Cancer Discov 2019;9(1):34–45 doi 10.1158/2159-8290.Cd-18-0689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Hu J, Sánchez-Rivera FJ, Wang Z, Johnson GN, Ho YJ, Ganesh K, et al. STING inhibits the reactivation of dormant metastasis in lung adenocarcinoma. Nature 2023;616(7958):806–13 doi 10.1038/s41586-023-05880-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Xia T, Konno H, Barber GN. Recurrent Loss of STING Signaling in Melanoma Correlates with Susceptibility to Viral Oncolysis. Cancer Res 2016;76(22):6747–59 doi 10.1158/0008-5472.Can-16-1404. [DOI] [PubMed] [Google Scholar]
  • 7.Konno H, Yamauchi S, Berglund A, Putney RM, Mulé JJ, Barber GN. Suppression of STING signaling through epigenetic silencing and missense mutation impedes DNA damage mediated cytokine production. Oncogene 2018;37(15):2037–51 doi 10.1038/s41388-017-0120-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Falahat R, Berglund A, Perez-Villarroel P, Putney RM, Hamaidi I, Kim S, et al. Epigenetic state determines the in vivo efficacy of STING agonist therapy. Nat Commun 2023;14(1):1573 doi 10.1038/s41467-023-37217-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Mahadevan NR, Knelson EH, Wolff JO, Vajdi A, Saigí M, Campisi M, et al. Intrinsic Immunogenicity of Small Cell Lung Carcinoma Revealed by Its Cellular Plasticity. Cancer Discov 2021;11(8):1952–69 doi 10.1158/2159-8290.Cd-20-0913. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Falahat R, Berglund A, Putney RM, Perez-Villarroel P, Aoyama S, Pilon-Thomas S, et al. Epigenetic reprogramming of tumor cell-intrinsic STING function sculpts antigenicity and T cell recognition of melanoma. Proc Natl Acad Sci U S A 2021;118(15) doi 10.1073/pnas.2013598118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Kitajima S, Tani T, Springer BF, Campisi M, Osaki T, Haratani K, et al. MPS1 inhibition primes immunogenicity of KRAS-LKB1 mutant lung cancer. Cancer Cell 2022;40(10):1128–44.e8 doi 10.1016/j.ccell.2022.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Yan N. Immune Diseases Associated with TREX1 and STING Dysfunction. J Interferon Cytokine Res 2017;37(5):198–206 doi 10.1089/jir.2016.0086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Simpson SR, Hemphill WO, Hudson T, Perrino FW. TREX1 - Apex predator of cytosolic DNA metabolism. DNA Repair (Amst) 2020;94:102894 doi 10.1016/j.dnarep.2020.102894. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Mohr L, Toufektchan E, von Morgen P, Chu K, Kapoor A, Maciejowski J. ER-directed TREX1 limits cGAS activation at micronuclei. Mol Cell 2021;81(4):724–38.e9 doi 10.1016/j.molcel.2020.12.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Ghosh M, Saha S, Li J, Montrose DC, Martinez LA. p53 engages the cGAS/STING cytosolic DNA sensing pathway for tumor suppression. Mol Cell 2023;83(2):266–80.e6 doi 10.1016/j.molcel.2022.12.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Li J, Duran MA, Dhanota N, Chatila WK, Bettigole SE, Kwon J, et al. Metastasis and Immune Evasion from Extracellular cGAMP Hydrolysis. Cancer Discov 2021;11(5):1212–27 doi 10.1158/2159-8290.Cd-20-0387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Yoshida R, Saigi M, Tani T, Springer BF, Shibata H, Kitajima S, et al. MET-Induced CD73 Restrains STING-Mediated Immunogenicity of EGFR-Mutant Lung Cancer. Cancer Res 2022;82(21):4079–92 doi 10.1158/0008-5472.Can-22-0770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Zhou W, Richmond-Buccola D, Wang Q, Kranzusch PJ. Structural basis of human TREX1 DNA degradation and autoimmune disease. Nat Commun 2022;13(1):4277 doi 10.1038/s41467-022-32055-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Marcus A, Mao AJ, Lensink-Vasan M, Wang L, Vance RE, Raulet DH. Tumor-Derived cGAMP Triggers a STING-Mediated Interferon Response in Non-tumor Cells to Activate the NK Cell Response. Immunity 2018;49(4):754–63.e4 doi 10.1016/j.immuni.2018.09.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Nicolai CJ, Wolf N, Chang IC, Kirn G, Marcus A, Ndubaku CO, et al. NK cells mediate clearance of CD8(+) T cell-resistant tumors in response to STING agonists. Sci Immunol 2020;5(45) doi 10.1126/sciimmunol.aaz2738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Berger G, Knelson EH, Jimenez-Macias JL, Nowicki MO, Han S, Panagioti E, et al. STING activation promotes robust immune response and NK cell-mediated tumor regression in glioblastoma models. Proc Natl Acad Sci U S A 2022;119(28):e2111003119 doi 10.1073/pnas.2111003119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Liu D, Schilling B, Liu D, Sucker A, Livingstone E, Jerby-Arnon L, et al. Integrative molecular and clinical modeling of clinical outcomes to PD1 blockade in patients with metastatic melanoma. Nat Med 2019;25(12):1916–27 doi 10.1038/s41591-019-0654-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zou SS, Qiao Y, Zhu S, Gao B, Yang N, Liu YJ, et al. Intrinsic strategies for the evasion of cGAS-STING signaling-mediated immune surveillance in human cancer: How therapy can overcome them. Pharmacol Res 2021;166:105514 doi 10.1016/j.phrs.2021.105514. [DOI] [PubMed] [Google Scholar]
  • 24.Kwon J, Bakhoum SF. The Cytosolic DNA-Sensing cGAS-STING Pathway in Cancer. Cancer Discov 2020;10(1):26–39 doi 10.1158/2159-8290.Cd-19-0761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Patel S, Jin L. TMEM173 variants and potential importance to human biology and disease. Genes Immun 2019;20(1):82–9 doi 10.1038/s41435-018-0029-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Stetson DB, Ko JS, Heidmann T, Medzhitov R. Trex1 prevents cell-intrinsic initiation of autoimmunity. Cell 2008;134(4):587–98 doi 10.1016/j.cell.2008.06.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Knelson EH, Ivanova EV, Tarannum M, Campisi M, Lizotte PH, Booker MA, et al. Activation of Tumor-Cell STING Primes NK-Cell Therapy. Cancer Immunol Res 2022;10(8):947–61 doi 10.1158/2326-6066.Cir-22-0017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Wang CJ, Lam W, Bussom S, Chang HM, Cheng YC. TREX1 acts in degrading damaged DNA from drug-treated tumor cells. DNA Repair (Amst) 2009;8(10):1179–89 doi 10.1016/j.dnarep.2009.06.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Vanpouille-Box C, Alard A, Aryankalayil MJ, Sarfraz Y, Diamond JM, Schneider RJ, et al. DNA exonuclease Trex1 regulates radiotherapy-induced tumour immunogenicity. Nat Commun 2017;8:15618 doi 10.1038/ncomms15618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Kitajima S, Asahina H, Chen T, Guo S, Quiceno LG, Cavanaugh JD, et al. Overcoming Resistance to Dual Innate Immune and MEK Inhibition Downstream of KRAS. Cancer Cell 2018;34(3):439–52 e6 doi 10.1016/j.ccell.2018.08.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Ritter JL, Zhu Z, Thai TC, Mahadevan NR, Mertins P, Knelson EH, et al. Phosphorylation of RAB7 by TBK1/IKKepsilon Regulates Innate Immune Signaling in Triple-Negative Breast Cancer. Cancer Res 2020;80(1):44–56 doi 10.1158/0008-5472.CAN-19-1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Aref AR, Campisi M, Ivanova E, Portell A, Larios D, Piel BP, et al. 3D microfluidic ex vivo culture of organotypic tumor spheroids to model immune checkpoint blockade. Lab Chip 2018;18(20):3129–43 doi 10.1039/c8lc00322j. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1981734-supplement-1.xlsx (11KB, xlsx)
2
NIHMS1981734-supplement-2.docx (245.2KB, docx)
3
NIHMS1981734-supplement-3.docx (272.7KB, docx)
4
NIHMS1981734-supplement-4.docx (475.8KB, docx)
5
NIHMS1981734-supplement-5.docx (1MB, docx)

Data Availability Statement

The datasets generated during and/or analyzed in this study are available from the corresponding author upon reasonable request.

Sequence data were deposited at Gene Expression Omnibus (GEO) and the accession number for the RNA-seq data reported in this paper is GEO: GSE252340.

RESOURCES