As a cytoplasmic sensor of double-stranded DNA (dsDNA), the cyclic GMP-AMP synthase–stimulator of IFN genes (STING) pathway plays an important role in antitumor immunity. In this study, we investigated the effect of Src homology-2 domain-containing protein tyrosine phosphatase-2 (SHP2) on tumor cell–intrinsic STING pathway activity and DNA repair in colon cancer. SHP2 interacted with and dephosphorylated PARP1 after DNA damage. PARP1 inhibition by SHP2 resulted in reduced DNA repair and accumulation of dsDNA in cells, thus promoting hyperactivation of the STING pathway. The SHP2 agonist lovastatin was able to enhance SHP2 activity and promote STING pathway activation. Moreover, lovastatin significantly enhanced the efficacy of chemotherapy in colon cancer models, in part via STING pathway-mediated antitumor immunity. These findings suggest that SHP2 exacerbates STING pathway activation by restricting PARP1-mediated DNA repair in tumor cells, providing a basis for the combined use of lovastatin and chemotherapy in the treatment of colon cancer.
Dephosphorylation of PARP1 by SHP2 simultaneously suppresses DNA repair and enhances STING pathway-mediated antitumor immunity, highlighting SHP2 activation as a potential therapeutic approach in colon cancer.
Colon cancer is the third most common malignancy and one of the most common causes of cancer-related deaths worldwide (1). Systemic therapy for advanced colon cancer typically includes a chemotherapy regimen paired with a biological medicine (2). First-line chemotherapy consisting of a two- or three-drug regimen (i.e., fluoropyrimidines, oxaliplatin, and irinotecan) significantly improves the survival and quality of life of patients with advanced-stage disease (2, 3). Chemotherapeutic drugs are primarily used because of their cytotoxic properties that are partially mediated by DNA damage; increasing evidence suggests that they could also promote an antitumor immune response (4). Therefore, targeting tumor immunity provides a novel strategy for improving the efficacy of chemotherapy for colon cancer.
Src homology 2 (SH2)-containing protein tyrosine phosphatase-2 (SHP2), encoded by the protein tyrosine phosphatase, nonreceptor type 11 (PTPN11) gene, is a nonreceptor protein tyrosine phosphatase (5). SHP2 protein contains two SHP2 domains (i.e., N-SH2 and C-SH2), which play a critical role in its subcellular localization, and a PTP domain, which plays a pivotal part in its enzymatic activity (6). SHP2 is involved in various functions related to tumor development, such as cell proliferation, differentiation, metastasis, and drug resistance (7). Increasing evidence indicates that SHP2 is also involved in the immune response by regulating the T lymphocytes and macrophages (8–11). However, SHP2 plays opposite immunomodulatory functions in different immune cells, and the effect of tumor-intrinsic SHP2 on immunity remains to be explored.
The cyclic GMP-AMP synthase (cGAS)–stimulator of IFN genes (STING) signaling pathway has emerged as a key regulator of innate and adaptive immunity. It is activated by binding of cGAS to both the exogenous and endogenous DNA in the cytoplasm (12). STING then phosphorylates TANK binding kinase 1 (TBK1) and IFN regulatory factor 3 (IRF3), whose activation initiates the expression of type I IFNs and inflammatory cytokines (12). Increasing evidence supports the vital function of this axis in antitumor immunity to prevent the occurrence and metastasis of various cancer types (13). The evasive effect of some tumors on STING-mediated immune response has led to the application of STING agonists in oncotherapy. Interestingly, the STING pathway and its mediated antitumor immunity can also be activated by DNA-damaging drugs such as etoposide, teniposide, cisplatin, and adriamycin (14–17). Nonetheless, chemotherapy-mediated STING pathway activation is insufficient for inducing adequate immunogenic cell death (18, 19). Therefore, amplifying the STING pathway may potentiate the efficacy of chemotherapeutic drugs.
DNA double-strand break (DSB), one of the most severe forms of DNA damage caused by chemotherapy, give rise to mutations, genomic instability, and even cell death if left unrepaired (20). Deficiency in the DNA repair of DSBs leads to the accumulation of DNA damage, accompanied by leakage of damaged DNA into the cytoplasm (21). It has been reported that the SHP2 aggravates genomic instability, which may be related to its interference with cell mitosis (22). However, SHP2's involvement in DNA repair in tumors remains unknown.
In this study, we aimed to investigate the effects of SHP2 on the STING pathway and DNA repair as well as their potential regulatory mechanisms, to elucidate their implications in colon cancer.
Materials and Methods
Chemicals, reagents, and antibodies
Irinotecan (CPT-11, I1406), 5-fluorouracil (F6627), oxaliplatin (O9512), lovastatin (1370600), simvastatin (S6196), pravastatin (P4498), and rosuvastatin (SML1264) were purchased from Sigma-Aldrich. Etoposide (S1225), cGAMP (S7904), olaparib (AZD2281), rucaparib (S1098), and atorvastatin (S5715) were purchased from Selleck Chemicals. The C-176 (314054-00-7), SHP099 (HY-100388) and PMA (HY-18739) were bought from MedChemExpress. NSC-87877 (565851) and PHPS1 (540213) were purchased from Calbiochem. ELISA kit for murine IFNβ (70-EK2236) was purchased from MultiSciences. Anti-GFP-tag (2956, RRID: AB_1196615, 1 : 1,000 dilution), anti-HA-tag (3724, RRID: AB_1549585, 1:1,000 dilution), anti-cGAS (15102, RRID: AB_2732795, 1:1,000 dilution), anti-STING (13647, RRID: AB_2732796, 1 : 1,000 dilution), anti-p-TBK1 (5483, RRID: AB_10693472, 1 : 1,000 dilution), anti-p-IRF3 (29047, RRID: AB_2773013, 1 : 1,000 dilution), anti-SHP2 (3397, RRID: AB_2174959, 1 : 1,000 dilution), anti-p-SHP2 (5431, RRID: AB_10693803), and anti-PARP1 (9532, RRID: AB_659884, 1 : 1,000 dilution) were purchased from Cell Signaling Technology. Anti-dsDNA (double-stranded DNA; sc-80772, RRID: AB_2191337, 1 : 50 dilution) was purchased from Santa Cruz Biotechnology. Anti-γH2AX (ab26350, RRID: AB_470861, 1 : 1,000 or 1 : 100 dilution) was purchased from Abcam. Anti-TBK1 (28397-1-AP, RRID: AB_2881132, 1:1,000), anti-IRF3 (11312-1-AP, RRID: AB_2127004, 1:1,000), anti-H2AX (10856-1-AP, RRID: AB_2114985, 1 : 500 dilution), anti-CD8 (66868-1-Ig, RRID: AB_2882205, 1 : 100 dilution), anti-PCNA (60097-1-Ig, RRID: AB_2236728, 1 : 100 dilution), and anti-HMGCR (13533-1-AP, RRID: AB_2877957, 1:1,000) were purchased from Proteintech. Anti-IFNβ (PA5-102429, RRID: AB_2851834, 1:100 dilution) was purchased from Thermo Fisher Scientific. Anti-β-Actin (M20010, 1 : 2,000 dilution) and anti-GAPDH (M20005, 1 : 2,000 dilution) were purchased from Abmart. Anti-human CD69-FITC (11711-50, 1:100 dilution) for flow cytometric analysis was purchased from Biogems. Alexa Fluor 488 goat anti-rabbit IgG (A11008, RRID: AB_143165, 1:100 dilution), Alexa Fluor 488 Donkey Anti-Goat IgG (A11055, RRID: AB_2534102, 1:100 dilution), and Alexa Fluor 594 Goat Anti-Mouse IgG (A11032, RRID: AB_2534091, 1:100 dilution) were purchased from Thermo Fisher Scientific. All other chemicals were obtained from Sigma-Aldrich.
Plasmids and lentivirus
The pCMV-SHP2 (Plasmid 8381, RRID: Addgene_8381), pDRGFP (Plasmid 26475, RRID: Addgene_26475), pimEJ5GFP (Plasmid 44026, RRID: Addgene_44026), and pCBASceI (Plasmid 26477, RRID: Addgene_26477) were purchased from Addgene. The pCMV-PARP1-GFP (Plasmid HG11040) and pCMV-SHP2-GFP (Plasmid HG12318) were purchased from Sino Biological. The pCMV-SHP2-HA, pCMV-SHP2-D61A-HA, pCMV-SHP2-C459S-HA, pCMV-SHP2-ΔPTP-HA, pCMV-SHP2-ΔSH2-HA, pCMV-PARP1-Y907D-GFP, and pCMV-PARP1-Y907A-GFP were obtained by PCR-based mutation and amplification of wild-type (WT) expression vector. The lentivirus for short hairpin (shRNA) targeting human SHP2 and shRNA-scramble were purchased from Obio Technology. The sequences were 5′-TTCTCCGAACGTGTCACGT-3′ (shRNA-scramble), 5′-ACACTGGTGATTACTATGA-3′ (shRNA-SHP2-#1), and 5′-GGTCCAGTATTACATGGAA-3′ (shRNA-SHP2-#2), respectively. The lentivirus for shRNA targeting mouse SHP2 and shRNA-scramble were purchased from GeneChem. The sequences were 5′-TTCTCCGAACGTGTCACGT-3′ (shRNA-scramble), 5′-CGTGTTAGGAACGTCAAAGAA-3′ (shRNA-SHP2), respectively. The PARP1 siRNA, HMGCR siRNA, and siRNA-scramble were purchased from RiboBio. The sequences were 5′-TTCTCCGAACGTGTCACGT-3′ (siRNA-scramble), 5′- GCAGCAGAGTATGCCAAGT-3′ (siRNA-PARP1-#1), 5′-GGAACCAACTCCTACTACA-3′ (siRNA-PARP1-#2), and 5′-UCUUCAUGUUAAAGGUGCUUCUGAA-3′ (siRNA-HMGCR), respectively.
Cell culture and transfections
Human colon cancer cell lines (HCT116, HT29, and SW620), mouse colon cancer cell lines (CT26 and MC38), human gastric cancer cell line (SGC-7901), human esophageal cancer cell lines (KYSE-30 and TE-1), human lung cancer cell line (A549), human cervical cancer cell line (Hela), human neuroblastoma cell line (SH-SY5Y), human breast cancer cell lines (MDA-MB-231 and ZR-75-1), human T lymphocytic leukemia cell line (Jurkat), human acute monocytic leukemia cell line (THP-1), human diffuse large B lymphoma cell line (OCI-LY3), and human embryonic kidney cell line (HEK-293) were purchased from Shanghai Cell Bank of Chinese Academy of Sciences (Shanghai, P.R. China). These cells were cultivated in McCoy's 5A (KeyGen Biotech), DMEM high-glucose (Biological Industries) or RPMI1640 medium (Gibco) supplemented with 10% FBS and cultured in a humidified incubator with 5% CO2 at 37°C. Cells were transfected with plasmid DNA, shRNA, or siRNA by Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Stable SHP2 knockdown cell clones after lentivirus interference were selected using puromycin and confirmed by Western blot analysis.
Total RNA was extracted with TRIzol reagent (TaKaRa) and reverse transcribed with FastQuant RT Kit (Tiangen) according to manufacturer's instructions. Then, qPCR analysis was performed using the SYBR Green Kit (Invitrogen) and quantified by the Real-Time PCR Detection system (Roche). Each sample was detected in triplicate and relative mRNA levels normalized to the expression of β-actin were calculated using the 2–ΔΔCt method. The primer sequences were indicated in Supplementary Table S1.
RNA sequencing analysis
Total RNA extraction, library preparation, and RNA sequencing (RNA-seq) were carried out by Novogene. Each group consists of three replicates and the RNA-seq data have been uploaded to the Gene Expression Omnibus database (GSE164730). Differentially expressed genes (DEG) were analyzed with a DEGseq algorithm by Novogene. The list of genes induced after stimulation was obtained by filtering the DEG list, and a heatmap of DEGs was generated using the heatmap package in R.
Western blot analysis
Whole cells were harvested in RIPA lysis buffer containing protease inhibitor cocktail (Beyotime, P.R. China) on ice. Equivalent loading proteins were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Millipore). Next, the membranes were blocking with 5% nonfat milk or 3% BSA before incubation with antibodies at 4°C overnight. Finally, immunoreactive bands after incubation with secondary antibodies were detected using an ECL kit (Beyotime) according to the manufacturer's instructions.
Coculture and function analysis
Coculture assay was performed using trans-well inserts (BD Biosciences). HCT116 colon cancer cells (2 × 105) were inoculated in upper chambers. After pretreated with CPT-11 for 24 hours, the HCT116 cells were washed with PBS to get rid of the CPT-11 and then cocultured with Jurkat cells, which were inoculated in lower chambers. After additional 24 hours, the cells were incubation CD69-FITC antibody and detected by FACS. Data were analyzed using FlowJo software (RRID: SCR_008520, Tree Star Inc).
Extracellular dsDNA detection
The supernatant was collected after centrifugation at 5,000 rpm for 5 minutes at 4°C for dsDNA quantification. The dsDNA concentration of the supernatant was measured by PicoGreen (Invitrogen) according to the manufacturer's protocol.
Neutral cell comet assay
The neutral comet assay was performed using the Comet Assay Kit (Keygenbio) following the manufacturer's instructions. In detail, cells embedded on slides in 0.6% low melting point agarose gels were lysed overnight then incubated in electrophoresis buffer for 30 minutes before being electrophoresed for 20 minutes (25 V). They were then neutralized and propidium iodide stained prior to imaging with a fluorescence microscopy (Olympus). Finally, the olive tail moment was analysis using comet assay software project.
Cells or tumor sections were fixed in 4% formaldehyde, permeabilized in 0.5% Triton-100 in PBS or saponin (Beyotime), and incubated in blocking solution of 5% BSA. Samples were stained with primary antibodies at 4°C overnight followed by incubation with fluorescently labeled secondary antibodies. Coverslips were mounted in antifade solution containing DAPI and imaged with a confocal laser scanning microscope or fluorescence microscopy (Olympus).
Homologous recombination and nonhomologous end joining measurement
Homologous recombination (HR) or nonhomologous end joining (NHEJ) activity was measured using pDRGFP or pimEJ5GFP reporter system as described previously (23, 24). Briefly, cells were transfected with pDRGFP or pimEJ5GFP and pCBASceI plasmid, which induces double strand break using Lipofectamine 2000, followed by FACS for GFP recovery at 48 hours latter. Data were analyzed using FlowJo software.
Cells with different treatments were lysed by RIPA buffer and the supernatants of the whole-cell lysates (1 mg) were incubated with 1 μg of appropriate primary antibodies on a rotator at 4°C overnight, and then the immune complexes were added with protein A/G beads (Santa Cruz Biotechnology) for 4 hours at 4°C. After washed five times with lysis buffer, the immune complexes with beads were subjected to SDS-PAGE, followed by Western blot analysis or mass spectrometry analysis.
Mass spectrometry analysis
After co-precipitated protein digestion, the peptides were subjected to analyses with a Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific), and the resulting raw files were preliminary searched by using the Mascot search engine (Version 2.5.1, Matrix Science) and quantified with Maxquant software (RRID: SCR_014485).
Catalytic activity analysis of SHP2
The catalytic activity of SHP2 was detected using the surrogate substrate DiFMUP in a prompt fluorescence assay format as described previously (25). The fluorescence intensity was examined by a fluorescence microplate (Thermo Fisher Scientific).
Microscale thermophoresis assay
The interaction between small molecule and protein or between proteins was investigated by microscale thermophoresis (MST) using a NanoTemper Monolith NT.115 instrument and the MO. Recombinant human SHP2 protein was labeled by Monolith Protein Labeling Kit RED-NHS 2nd Generation. Affinity Analysis software (NanoTemper Technologies) was used for combined analysis.
Full-length human PARP1 was purified as reported previously (26). Briefly, full-length human PARP1 was cloned into PET-28a with 6*His on N-terminal and transfected into Rosetta (DE3). A single colony was inoculated into Luria-Bertani (LB) medium containing kanamycin; cultures were incubated in 37°C at 200 rpm. When OD600 was between 0.4 and 0.6, 100 mmol/L ZnCl2 to a final concentration of 0.1 mmol/L was added. Once cell density reached to OD = 0.6–0.8 at 600 nm, remove cultures from incubator and chill on ice for 1 hour. Add 1 mol/L IPTG to a final concentration of 0.2 mmol/L to induce protein production. Inoculate with shaking at 16°C, 210 rpm for 24 hours. Cells were lysed with high-pressure crusher. The target protein was first purified by affinity chromatography using an Ni-NTA column (QIAGEN) and further purified by gel filtration using a Superdex 200 10/60 GL column (GE Healthcare).
Cellular thermal shift assay
Cellular thermal shift assay (CETSA) was performed as described previously (27). PBS diluted cells, which were treated with DMSO or lovastatin for 4 hours, were divided into smaller aliquots (50 μL) and heated individually at different temperatures for 3 minutes (Veriti thermal cycler, Applied Biosystems/Life Technologies) followed by cooling for 3 minutes at room temperature. Then lysates were freeze-thawed three times with liquid nitrogen and separated from the cell debris by centrifugation at 20, 000 g for 20 minutes at 4°C. A total of 40 μL of the supernatant was mixed with 10 μL of 5 × loading buffer and then separated on a 10% SDS-PAGE for Western blot analysis of SHP2.
Cell viability analysis
The in vitro cytotoxicity was measured by Cell Counting Kit-8 (CCK-8; Vicmed). Cells were plated in 96-well plates and treated with indicated drugs for 24 hours. Then, CCK-8 solution was added into each well and incubated at 37°C for 2 hours. The absorbance (450 nm) was measured in microplate reader (BioTek).
Cell-cycle and apoptotic analyses
The Cell Cycle Detection Kit (KGA512, KeyGEN) and the Annexin V-PE Apoptosis Detection Kit (KGA1013, KeyGEN) was used to detect cell cycle and apoptotic, respectively. Cells were collected and mixed with Annexin V-FITC buffer. FACS data were analyzed using FlowJo software.
Six- to eight-week-old female BALB/c or C57BL/6 mice were purchased from Model Animal Genetics Research Center of Nanjing University (Nanjing, P.R. China). All animal experimental procedures were approved by the Animal Ethical and Welfare Committee of Nanjing University (Nanjing, P.R. China; IACUC-2010010). CT26 or MC38 cells (1 × 106 cells in 100 μL PBS per mouse) were injected into mice by subcutaneous injection. Tumor dimensions were measured using calipers, and tumor volume was calculated using the following formula: (shortest diameter)2 × (longest diameter) × 0.5. After continuous observation, the tumor tissues were separated from sacrificed mice. Tumors were collected for protein and RNA isolation and paraffin histologic analysis.
For the lovastatin experiment, BALB/c mice with CT26 or shSHP2 CT26 tumors received one of the following treatments: (i) control; (ii) lovastatin, 20 mg/kg, orally administration, daily; (iii) CPT-11, 40 mg/kg, intraperitoneal administration, twice a week; or (iv) combination of lovastatin and CPT-11. For the STING inhibition experiment, BALB/c mice with mouse CT26 tumors received one of the following treatments: (i) control; (ii) lovastatin; (iii) C-176 13.4 mg/kg (28), intraperitoneal administration, daily; (iv) CPT-11; (v) C-176 and CPT-11; (vi) lovastatin and CPT-11; or (vii) combination of lovastatin, C-176 and CPT-11. For the CD8 knockout (KO) experiment, C57BL/6 WT or CD8 KO mice with mouse MC38 tumors received one of the following treatments: (i) control; or (ii) lovastatin and CPT-11.
IHC and TUNEL assays
Paraffin-embedded tumor sections were firstly deparaffinized with 100% xylene, followed by rehydration using gradient ethanol. After inactivation of endogenous peroxidase and retrieval antigen, IHC staining was performed using Immunohistochemical Detection kit (Proteintech). TUNEL staining was performed using the Cell Apoptosis Detection Kit (Trevigen) according to the manual. All stained specimens were photographed using 200- or 400-fold magnification by a light microscope (Olympus).
IFNβ protein level in tumor tissue was quantified using ELISA Kit (MultiSciences) according to manufacturer's instructions. Absorbance was measured at 450 nm and corrected against absorbance at 570 nm. Relative IFNβ level was calculated by normalizing to protein concentration of each sample.
Patient-derived tumor organoid culture
Intestine tumor tissue was obtained from colon cancer subject providing written informed consent for tissue in accordance with the Declaration of Helsinki and studies were approved by an Institutional Review Board at The First Affiliated Hospital of Nanjing Medical University (Nanjing, P.R. China; 2020-SRFA-080).
Tumor tissue was washed with cold Hank's Balanced Salt Solution. After removing the muscle tissue, the epithelial tumor tissues were cut into small pieces and vigorously suspended. Then the tissue suspension was centrifuged (300 g, 3 minutes) and tumor fraction was enriched at the bottom. The tumor fraction embedded in 50 μL Matrigel (BD Biosciences) were seeded into 24-well plates and cultured with conditioned medium after the Matrigel solidified, which contains 50% WNT3A conditioned medium, 10% R-spondin conditioned medium, 38% advanced DMEM/F12 (Gibco), 1 × B27, 1.25 mmol/L N-acetyl cysteine, 10 mmol/L nicotinamide, 50 ng/mL EGF (Sigma), 500 nmol/L A83-01, 100 ng/mL Noggin, and 10 μmol/L Y-27632 dihydrochloride (R&D). Colon organoids were digested into single cells with dispase II (Roche) and TrypLE (Gibco). A total of 4 × 103 cells were embedded in Matrigel and cultured in 24-well plates for 5 days, then organoids were treated with lovastatin or CPT-11 for 5 days. All organoids were photographed using a light microscope. The viability of organoids was measured by CellTiter-Glo luminescent cell viability assay (Promega).
Statistical analyses were performed using the GraphPad Prism software (version 8). All data were presented as the mean ± SEM. One-way or two-way ANOVA was used to analyze statistically significant differences between multiple-group comparisons. A two-tailed Student t test was used to analyze statistically significant differences between two groups. P < 0.05 was considered statistically significant.
DNA damage chemotherapeutic drugs activated the STING pathway in colon cancer cells
STING expression was detected in different cancer cells and was highly expressed in gastrointestinal cancer, including colon, gastric, and esophageal cancer, compared with other tumor types (Supplementary Fig. S1A). CPT-11 contributed markedly to the mRNA expression of IRF3-responsive genes (IFNβ, CCL5, CXCL10, and IFIT1) in the STING-expressing cancers whereas its induction was impeded in the STING-deficient cancers, such as cervical cancer, neuroblastoma, and lymphoma (Supplementary Fig. S1B–S1O).
We then stimulated the HCT116 cells with cGAMP (a STING agonist) and found the STING pathway to be activated in colon cancer (Supplementary Fig. S1P). RNA-seq assay showed upregulation of type I IFNs and IFN-stimulated genes in the HCT116 cells treated with CPT-11 (Fig. 1A). Similarly, CPT-11 significantly induced a dose- and time-dependent increase in the mRNA expression of IRF3-responsive genes (Supplementary Fig. S1B and S1Q), accompanied by increased phosphorylation of TBK1 and IRF3 levels in the HCT116 cells (Fig. 1B). Meanwhile, 5-fluorouracil, oxaliplatin, and etoposide also demonstrated a dose-dependent stimulation of IRF3-responsive genes (Supplementary Fig. S1R).
Furthermore, HCT116 cells pretreated with C-176 (STING inhibitor) eliminated the CPT-11–induced expression of IRF3-responsive genes (Supplementary Fig. S1S). In addition, the activated Jurkat cells after coculture with CPT-11–treated HCT116 cells, expressed as upregulated CD69 expression, were reversed by C-176 pretreatment (Supplementary Fig. S1T). These data suggest that DNA damage drugs could activate the STING pathway in colon cancer cells.
SHP2 inhibition decreased activation of the STING pathway in colon cancer cells
To explore whether SHP2 could regulate the STING pathway activation triggered by DNA damage drugs, we used gene-specific shRNA to knockdown SHP2 in colon cancer cells. Western blot analysis and qPCR showed that SHP2 silencing inhibited the phosphorylation of TBK1 and IRF3 and downregulated the expression of the IRF3-responsive genes stimulated by CPT-11 (Fig. 1C and D; Supplementary Fig. S2A and S2B). Consistently, SHP2 knockdown significantly weakened the IRF3-responsive genes expression in CT26 and KYSE30 cells (Supplementary Fig. S2C and S2D). In contrast, SHP2 overexpression enhanced the IRF3-responsive gene expression in HT29 cells (Supplementary Fig. S2E). Importantly, CD69 expression was attenuated in the Jurkat cells after coculture with CPT-11–treated shSHP2 HCT116 cells compared with the WT cells (Fig. 1E and F). In addition, pharmacologic inhibition of SHP2 with SHP099, NSC87877, or PHPS1 resulted in a similar decrease in the IRF3-responsive gene production (Supplementary Fig. S2F–S2H). Collectively, these data indicate that SHP2 could positively regulate the STING pathway activation in colon cancer.
SHP2 inhibited DNA damage repair in colon cancer cells
DNA-damaging drugs lead to an accumulation of cytosolic DNA (29), which is the initial activator of the STING pathway. After selective plasma membrane permeabilization and anti-dsDNA antibody staining, we found that SHP2 knockdown decreased the cytosolic dsDNA levels in HCT116 cells after CPT-11 treatment (Fig. 2A and B). Similarly, dsDNA release into the supernatant was also diminished in the SHP2-silenced HCT116 cells (Fig. 2C). These results suggest that SHP2 might exacerbate the degree of DNA damage. DNA damage of DSB can be aggravated after a deficient DNA repair (20). Thus, we assessed whether SHP2 affected the DSB repair using the comet assay. Notably, DSB repair could be achieved within 9 hours in SHP2-silenced cells, whereas the nonsilenced cells accumulated more DSBs (Fig. 2D and E; Supplementary Fig. S2I). Western blot analysis confirmed a decrease in γH2AX levels in cells with SHP2 knockdown (Fig. 2F and G; Supplementary Fig. S2J). Subsequently, the changes in the formation of γH2AX foci were measured. Three hours after CPT-11 withdrawal, the control cells displayed prominent γH2AX foci, significantly more than that in the SHP2-depleted cells (Fig. 2H and I).
HR and NHEJ are the two mechanistically distinct repair pathways for DNA DSBs (20). The effect of SHP2 on the two different pathways was examined using the two well-characterized GFP-based reporting systems. SHP2 depletion improved the repair efficiency of the NHEJ pathway but did not affect the HR pathway (Fig. 2J; Supplementary Fig. S2K). Collectively, these data indicate that SHP2 plays a role in the NHEJ pathway for DNA damage repair.
SHP2 interacted with PARP1
To identify the molecular basis of the DNA damage inhibition of SHP2 in colon cancer, we purified an SHP2-containing complex from the HCT116 cells following the CPT-11 treatment. Mass spectrometry profiling of binding protein partners of SHP2 identified PARP1 as a high-confidence hit, which plays a pivotal role in DNA repair (Supplementary Table S2). Western blot analysis showed that PARP1 could be co-precipitated by the anti-SHP2 antibody in the HCT116 cells exposed to CPT-11 (Fig. 3A). Reciprocal co-immunoprecipitation (co-IP) with anti-PARP1 antibody also pulled down SHP2 in CPT-11–treated cells (Fig. 3B). Confocal immunofluorescence analysis showed that SHP2 colocalized with PARP1 in the nucleus upon CPT-11 treatment (Fig. 3C). By using the purified SHP2 and PARP1, we confirmed that SHP2 could directly bind to PARP1 (Fig. 3D; Supplementary Fig. S3A). These data provide evidence that SHP2 and PARP1 interact with each other during DNA damage.
To further investigate which domain of SHP2 interacts with PARP1, we performed protein–protein docking for interaction prediction. The results showed that the PTP and N-SH2 domains but not the C-SH2 domain, interact with the PARP1 catalytic domain. Furthermore, the PTP domain exhibited a larger interaction surface than the N-SH2 domain (Fig. 3E). In addition, co-IP in HCT116 cells showed that either the PTP or SH2 domain is dispensable for the interaction between SHP2 and PARP1. Moreover, the PTP domain had a superior binding capacity to the N-SH2 domain (Fig. 3F). These findings demonstrate that PARP1, an endogenous SHP2-interacting protein, is mainly bound to its PTP domain.
SHP2 promoted STING pathway activation in a PARP1-dependent manner
The interaction of SHP2 and PARP1 prompted us to investigate the role of PARP1 in the STING pathway activated by SHP2. Several studies have shown that PARP1 inhibitors enhance the function of CTLs through the STING pathway (30, 31). Thus, we explored the relationship between PARP1 and the chemotherapy-triggered STING pathway in colon cancer. Compared with the nonsilenced control group, the phosphorylation of IRF3 and mRNA expression of IRF3-responsive genes was significantly enhanced in PARP1-silenced HCT116 cells exposed to CPT-11 (Fig. 4A and B; Supplementary Fig. S3B and S3C). Similarly, the PARP1 inhibitor olaparib or rucaparib also magnified the expression of CPT-11–induced IRF3-responsive genes (Supplementary Fig. S3D–S3F). These results indicate that PARP1 inhibition activates the STING pathway. Furthermore, we compared the effects of SHP2 overexpression alone, PARP1 overexpression alone, and SHP2-PARP1 cooverexpression on the STING pathway. Remarkably, the overactivation of the STING pathway resulting from SHP2 overexpression and subsequent activation of Jurkat cells after coculture was partially reversed by the PARP1 and SHP2 cooverexpression (Fig. 4C and D; Supplementary Fig. S3G and S3H). These findings indicate that PARP1 is required for SHP2-mediated regulation of the STING pathway.
SHP2 dephosphorylation of PARP1 at Tyr 907 was essential for STING activation
Given that SHP2 is a nonreceptor tyrosine phosphatase (5), we sought to determine whether its phosphatase activity is involved in the STING pathway regulation. As shown in Fig. 5A and B, overexpression of SHP2-WT or SHP2-D61A (Asp61 mutated to Ala in the SH2 domain, gain-of-function mutant) but not SHP2-C459S (Cys459 mutated to Ser in PTP domain, loss-of-function mutant) facilitated the CPT-11–induced phosphorylation of TBK1 and IRF3 as well as the expression of IRF3-responsive genes. These findings suggest that the phosphatase activity of SHP2 is required for STING pathway activation.
Phosphorylation of PARP1 Tyr907 has been reported to enhance its catalytic activity (32). The protein–protein docking experiment showed that the Tyr907 site was located on the interaction surface where PARP1 binds to the PTP domain of SHP2 (Fig. 3E). To further determine whether Tyr907 (Y907) of PARP1 is dephosphorylated by SHP2, we generated a plasmid-expressing PARP1-Y907A (Tyr907 mutated to Ala, loss-of-function mutant). The results showed that PARP1 Y907A loses the ability to reverse the STING pathway activation (Fig. 5C and D). Meanwhile, PARP1-Y907D (Tyr907 mutated to Ala Asp, gain-of-function mutant) attenuated the activation of the STING pathway (Fig. 5C and D). These data suggest that SHP2-mediated dephosphorylation of PARP1 at Tyr 907 might be critical for STING pathway activation.
Lovastatin activated the STING pathway in an SHP2-dependent manner
Our previous study on SHP2 agonist screening showed that the antilipemic agent lovastatin exerts a unique function in enhancing SHP2 activity (33). Here, we confirmed that lovastatin was able to elevate the SHP2 activity in colon cancer cells (Fig. 6A). The CETSA demonstrated that lovastatin affects the thermal stability of SHP2 but not β-actin in colon cancer cells (Fig. 6B). Furthermore, MST assay demonstrated that lovastatin binds to SHP2 at a Kd value of 20.9 μmol/L (Fig. 6C).
Next, we explored whether lovastatin affects DNA repair and the STING pathway activation. The results showed that lovastatin exacerbates the degree of DNA damage and prolongs the DSB repair time (Supplementary Fig. S4A and S4B). In addition, lovastatin strengthened the mRNA expression of IRF3-responsive genes induced by CPT-11 (Fig. 6D and E; Supplementary Fig. S4C and S4D). Furthermore, SHP2 knockdown and its inhibition impaired, while HMGCR (the target of lovastatin in cholesterol-lowering) knockdown did not alter, the expression of IRF3-responsive genes amplified by lovastatin (Fig. 6E; Supplementary Fig. S4E and S4F).
In addition, we explored the effects of other statins on SHP2 and found that most statins had a similar effect on the SHP2 activity (Supplementary Fig. S4G). CETSA confirmed lovastatin to have a better binding ability to SHP2 than other statins (Supplementary Fig. S4H). Similarly, lovastatin was superior to other statins in amplifying the CPT-11–induced IRF3-response gene expression (Supplementary Fig. S4I). The above results reveal that lovastatin induced the STING pathway activation by promoting the SHP2 phosphatase activity.
Lovastatin enhanced the chemosensitivity of colon cancer via SHP2-mediated DNA damage aggravation
Our observations indicate that SHP2 could inhibit DNA repair, which prompted us to investigate the impact of SHP2 on the antitumor effect of cytotoxic drugs. In vitro, SHP2 silencing reduced the sensitivity of colon cancer cells to CPT-11, and lovastatin partially reversed the chemoresistance process by apoptosis induction and cell-cycle arrest (Supplementary Fig. S5A–S5D). The in vivo data showed that lovastatin alone did not exert a significant antitumor effect. However, when administered with CPT-11, it significantly inhibited the tumor growth compared with CPT-11 alone. Moreover, the CPT-11 sensitization effect of lovastatin was reduced when SHP2 was knocked down (Fig. 6F and G). Consistent with the tumor volume data, PCNA and TUNEL assays showed that knockdown of SHP2 attenuated the effects of lovastatin and CPT-11 on inhibiting the proliferation and apoptosis induction (Supplementary Fig. S5E and S5F). In addition, SHP2 knockdown resulted in fewer γH2AX foci in the xenografts treated with CPT-11, indicating less DNA damage. In contrast, lovastatin aggravated the CPT-11–mediated DNA damage as indicated by more γH2AX foci (Fig. 6H). Collectively, these findings demonstrate that lovastatin sensitized chemotherapy for colon cancer via SHP2-induced DNA damage aggregation.
STING pathway inhibition reversed the lovastatin-mediated chemosensitization of colon cancer
Although our results suggested that lovastatin had a chemosensitization effect (Fig. 6), it was unclear whether this effect was also associated with the enhanced antitumor immunity in vivo. Thus, we examined the STING pathway and T-lymphocyte infiltration in these tumors. The results showed that SHP2 deficiency partially inhibited the expression of IFNβ and CXCL10, and infiltration of CD8+ T lymphocytes induced by CPT-11 treatment. On the other hand, lovastatin enhanced these effects (Supplementary Fig. S5F–S5I).
To further investigate the effects of lovastatin on antitumor immunity mediated by STING pathway, we used C-176 to inhibit the STING function in a xenograft model. The in vivo assay confirmed the chemosensitization effect of lovastatin. Although C-716 did not affect tumor growth, it partially reversed the efficacy of CPT-11 sensitization induced by lovastatin cotreatment (Fig. 7A and B; Supplementary Fig. S5J). Furthermore, we found that IFNβ expression and infiltration of CD8+ T lymphocytes were abrogated by C-176 in the lovastatin plus CPT-11 group (Fig. 7C and D).
To further confirm whether CD8+ T lymphocytes cells are involved in the sensitization effect of lovastatin on chemotherapy, the effect of lovastatin plus CPT-11 was observed in WT and CD8 KO C57BL/6 mice with transplanted tumors. The synergistic effect of the two drugs was abrogated in CD8 KO mice compared with the WT mice (Fig. 7E and F). Hematoxylin and eosin staining showed that tissue necrosis was more obvious in the WT group than the CD8 KO group after lovastatin and CPT-11 treatment (Fig. 7G). These data indicate that lovastatin partially enhanced the chemosensitivity of colon cancer via the STING pathway–mediated antitumor immunity.
Lovastatin promoted the sensitization of CPT-11 in colon cancer–derived organoids
Finally, we examined the clinical potential of lovastatin combined with chemotherapy using the patient-derived organoids (PDO) ex vivo. PDOs were generated from sequential biopsies of a patient with primary colon cancer. As shown in Supplementary S5K and S5L, CPT-11 inhibited the survival of the PDOs compared with the control group. Furthermore, cotreatment with lovastatin and CPT-11 resulted in better inhibition of the tumor growth. Compared with CPT-11 alone, DNA damage as indicated by more γH2AX foci (Supplementary Fig. S5M), and the expression of IFNβ, CCL5, and CXCL10 were enhanced after lovastatin and CPT-11 treatment (Supplementary Fig. S5N). These findings confirm that lovastatin could improve the sensitivity of colon cancer to CPT-11 chemotherapy, while simultaneously activating the tumor cell–intrinsic STING pathway.
DNA damage chemotherapeutic drugs exert immunosuppressive functions (34). With the advent of immunotherapy for cancer, the relationship has been re-recognized, that is, chemotherapy stimulates antitumor immunity (4). The results of the current study show that DNA damage drugs activate the cell-intrinsic STING pathway of colon cancer, which mediates the adaptive antitumor immune. SHP2 was identified as a positive regulator of the STING pathway. As shown in Fig. 7H, chemotherapy triggers SHP2 to interact with and dephosphorylate the PARP1. This molecular event prevents DNA repair and thereby activates the STING pathway. Moreover, lovastatin-induced SHP2 activation leads to excessive DNA damage and enhanced STING pathway, resulting in sensitization to chemotherapy in colon cancer.
The function of the STING pathway is to monitor the presence of dsDNA in the cytoplasm and trigger the expression of inflammatory genes (12). Several positive regulators that regulate different nodes of the STING pathway have been demonstrated to amplify the immune response, such as SENP7, IFI16, or PCBP1 potentiating cGAS and USP20 deubiquitinating STING (35–38). In this study, we show that SHP2 could positively regulated the STING pathway in colon cancer. Dong and colleagues (39) reported that STING phosphorylates SHP2 to inhibit the STAT1-mediated expressions of IFN-stimulated genes in B cells. Several studies have shown that SHP2 exerts completely opposite effects in different conditions or cells via different substrates (11). Here, we discovered that SHP2 dephosphorylated PARP1, which explains the positive regulation of SHP2 on type I IFN expression in cancer cells, as opposed to B cells. Besides, no binding of SHP2 to STING was observe on mass spectrometry of colon cancer (Supplementary Table S2). Therefore, it remains to be further verified whether STING can indirectly phosphorylate SHP2 to positively activate, as feedback, the STING pathway in colon cancer.
SHP2 has been reported to aggravate genomic instability by a mechanism associated with interference with cell mitosis (22). Here, we found that SHP2-depleted accelerated the DNA repair and then reduced dsDNA accumulation in the cytoplasm of colon cancer. Moreover, SHP2 primarily suppressed NHEJ repair rather than HR repair. Therefore, we concluded that SHP2 plays a crucial negative role in DNA damage repair, which in turn leads to genomic instability. The molecular functions of SHP2 are largely dependent on its phosphatase activity to dephosphorylate its substrates. For example, SHP2 dephosphorylates Gab1 (40) or Sprouty (41) in the cytoplasm to participate in inflammation and innate immunity, ANT1 in the mitochondria to destroy mitochondrial homeostasis (42), and STAT1 in the nucleus to inhibit its transcriptional activity (43). Using the gain-of-function or loss-of-function mutations in the PTP domain and inhibition of enzymatic activity, we showed that the chemotherapy-mediated activation of the STING pathway depends on the SHP2 enzyme activity. However, the substrate of SHP2 in DNA damage repair is unclear. Our data suggest that SHP2 is rapidly translocated into the nucleus and interacts with PARP1. PARP1 is an abundant nuclear protein that has been implicated in the classical NHEJ pathway by interacting with DNA-PKcs and recruiting XRCC4/CHD2 to sites of DSBs (44). Several studies have shown that PARP1 inhibition activates the STING pathway in BRCA-deficient triple-negative breast cancer (30) and ERCC1-deficient non–small cell lung cancer (31). This is in line with our finding that PARP1 inhibition enhanced the chemotherapy-induced STING pathway. The key regulatory role of SHP2-PARP1 interplay in the STING pathway is supported by the robust rescue effect of PARP1 overexpression on SHP2 overexpressed cells. Furthermore, we identified that the PTP domain with enzyme activity, which is better than the N-SH2 domain, is required for SHP2 interaction with the dephosphorylated substrate PARP1. Du and colleagues (32) showed that c-MET translocate into the nucleus where it directly binds and phosphorylates the Y907 of PARP1 to stimulate its enzymatic activity, thus enabling tumor cells to survive under the lethal influence of DNA damage. Interestingly, Y907 is located on the interaction region of the PARP1 and PTP domain. On the basis of this, we found that the Y907D mutant (gain of activity), not the Y907A mutant (loss of activity), mimicked the effect of PARP1 on the STING pathway deactivation. Future studies are required to elucidate how SHP2 regulates dephosphorylation of PARP1.
SHP2, which is usually inactive in its basal state, can undergo conformational changes under different stimuli, thus activating its phosphatase activity (45). Using the SHP2 enzyme activity screening system, we found that lovastatin, a serum cholesterol-lowering drug (46), is a potent small-molecule agonist of SHP2, directly binding to SHP2. To analyze the structure–activity relationship between SHP2 and lovastatin, a series of statins including simvastatin, atorvastatin, pravastatin, and rosuvastatin were employed. These statins could interact with SHP2 and activate its PTPase activity to different extents. Nevertheless, lovastatin showed the highest affinity for SHP2 and the greatest amplification effect on the STING pathway. Importantly, lovastatin-mediated STING pathway activation is SHP2 dependent but HMGCR independent.
Increasing evidence has shown that statins exhibit different beneficial effects both in cancer prevention at low concentrations and antitumor effects at high concentrations (47, 48). Several studies have shown that statins can aggregate the DNA damage response and sensitize chemotherapy (49–52). Consistent with the findings above, we found that lovastatin could delay the chemotherapy-induced DNA damage repair in colon cancer. The amplification effect of lovastatin on STING pathway activation was also confirmed. These results not only deepen the mechanistic understanding of lovastatin in DNA repair inhibition but also provide a new perspective for its regulation of antitumor immunity (53). Meanwhile, in vitro and in vivo studies showed that lovastatin potentiates the efficacy of CPT-11 in colon cancer. SHP2 knockdown, STING inhibition, and CD8 KO attenuated the chemosensitization effect of lovastatin. We speculate that by activating SHP2, lovastatin aggravates the DNA damage and activates the tumor cell–intrinsic STING pathway, thus stimulating CD8+ T lymphocytes mediated antitumor immunity. Collectively, these findings support that the combination of lovastatin and chemotherapy is a promising therapeutic strategy for colon cancer.
In conclusion, this study revealed a previously unknown negative regulatory effect of SHP2 on DNA repair and the STING pathway in colon cancer. SHP2 activation via lovastatin treatment suppressed the PARP1-mediated DNA repair and further induced the STING pathway-mediated antitumor immunity leading to sensitization of colon cancer to chemotherapy. Thus, promoting SHP2-PARP1–mediated intrinsic immunity of the tumor cell may provide a potential strategy for chemotherapy and immunotherapy sensitization, particularly in colon cancer.
No disclosures were reported.
B. Wei: Data curation, formal analysis, methodology, writing–original draft, writing–review and editing. L. Xu: Data curation, formal analysis, methodology. W. Guo: Formal analysis, methodology, writing–original draft, writing–review and editing. Y. Wang: Data curation, visualization. J. Wu: Data curation, visualization. X. Li: Data curation. X. Cai: Data curation. J. Hu: Data curation. M. Wang: Data curation. Q. Xu: Writing–review and editing. W. Liu: Supervision, funding acquisition, project administration. Y. Gu: Conceptualization, supervision, funding acquisition, project administration.
We thank Lilin Ye (Third Military Medical University) for sharing CD8 KO mice with C57BL/6 background.
This study was supported by National Natural Science Foundation of China (82072675 to Y. Gu; 81922067 to W. Guo; 81871944 to Y. Gu; 21937005 to Q. Xu), Jiangsu province Key Medical Talents (ZDRCA2016026 to Y. Gu), Natural Science Foundation of Jiangsu Province (BK20190306 to W Liu), Science and Technology Development Foundation of Nanjing Medical University (NMUB2019346 to J. Wu), and Fundamental Research Funds for the Central Universities (14380114 to W. Guo).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.