Overproduction of Gastrointestinal 5-HT Promotes Colitis-Associated Colorectal Cancer Progression via Enhancing NLRP3 Inflammasome Activation

Colorectal cancer is one of the most frequently diagnosed cancers worldwide, with a high death rate only next to lung, liver, and stomach cancer (1). Chronic colitis, including inflammatory bowel disease, ulcerative colitis, and Crohn disease, plays a vital role in colitis-associated colorectal cancer pathogenesis. It can cause multiple dysplastic lesions on the colon epithelium by excessively releasing inflammatory factors, which then form multifocal colorectal lesions, which then progress to multifocal carcinoma (2, 3).

As a key component of innate immune defense, controlled inflammation in the gastrointestinal tract is essential to maintain intestinal physiologic functions via repairing intima damage and balancing the gut microbiota. However, in chronic colitis, inflammatory reactions increase and facilitate an immunosuppressive microenvironment facilitating tumor cell growth (3).

Inflammasomes are molecular complexes that initiate inflammatory reactions by cleaving the pro–caspase-1 into mature caspase-1, which then catalyzes IL1β and IL18 maturation after the cells recognize endogenous factors (i.e., danger-associated molecular patterns) or invading pathogens (i.e., pathogen-associated molecular patterns; refs. 4, 5). Evidence has shown that the NLRP3 inflammasome is the main inflammation mediator during pathologic processes in colorectal cancer (6–8). Although the exact mechanism remains to be elucidated, activation of NLRP3 inflammasome can be divided into two steps: priming and activation. In the priming step, stimulators (e.g., lipopolysaccharide, LPS) activate intracellular NFκB signaling via Toll-like receptor 4 (TLR4) to upregulate NLRP3, IL1B, and IL18 transcription, which provide sufficient substrates for inflammasome assembling. In the activation step, NLRP3 molecules bind with adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD) and pro–caspase-1 to form inflammasome oligomers, which then induce pro–caspase-1 autocleavage and IL1β, IL18 maturation under stimulation of NLRP3 inflammasome activators [e.g., ATP, nigericin (Nig), monosodium urate (MSU); ref. 9]. Serval transcriptional and posttranscriptional regulation mechanisms of NLRP3 inflammasome activation have been proposed. Studies have demonstrated that phosphorylation of NLRP3 (human: Ser198; mouse: Ser194) is indispensable for NLRP3 inflammasome activation by facilitating NLRP3 inflammasome assembling (10, 11).

Increased levels of peripheral nervous neurotransmitters have been found in a variety types of cancers and demonstrated to regulate immune profiles in the tumor microenvironment after perineural invasion. For example, noradrenaline inhibits activation of CD8⁺ T cells by altering glycometabolism via the β₂ adrenergic receptor (12). β₂ adrenergic receptor signaling upregulates expression of programmed cell death protein 1 (PD1) and forkhead box protein P3 (FOXP3) on tumor-infiltrating lymphocytes, as well as programmed cell death ligand 1 (PDL1) on tumor cells (13). Increased adrenergic signaling also induces tumor-associated macrophage recruitment (14).

Because of the lack of cell-to-cell adjunction between most solid tumors and central nervous tissues, the carcinogenesis-related role of central nervous neurotransmitters (e.g., gamma aminobutyric acid, GABA; glutamate; 5-hydroxytryptamine, 5-HT) is not clear, and its contribution to tumor growth and progression remain underappreciated. However, increasing evidence has revealed that various neurotransmitters are produced and secreted by tumor cells in an autocrine/paracrine manner, which function as major regulators of tumor microenvironment remodeling (15). Findings have shown that GABA secreted by pancreatic cancer cells regulate macrophage recruitment by upregulating expression of C-X-C motif chemokine 5 (CXCL5) and C-C motif chemokine 20 (CCL20; ref. 16); glutamate secreted by triple-negative breast cancer cells drive tumorigenesis by inducing expression of hypoxia-inducible factor 1-alpha (HIF1α; ref. 17).

5-HT, also known as serotonin, is one of the most important neurotransmitters in the central nervous system and gastrointestinal tract. Under physiologic conditions, 5-HT is mainly produced by a subtype of intestinal epithelial cells known as enterochromaffin cells, and the physiologic level of 5-HT regulates gastrointestinal motility, secretion, and intestinal microbiota balance (18, 19). In vivo evidence has shown that overproduction of 5-HT may be involved in colorectal inflammatory reactions (20–22), but the exact role of 5-HT in colitis and colorectal cancer remains to be elucidated. In this study, we found that tryptophan hydroxylase 1 (TPH1), the rate-limiting enzyme in 5-HT biosynthesis, was highly expressed in tumor tissues of patients with colorectal cancer, which negatively correlated with overall survival rates. Compared with normal colorectal epithelial cells, production and secretion of 5-HT in colorectal cancer cells were significantly upregulated. Overproduced 5-HT enhanced NLRP3 inflammasome activation in macrophages via the HTR3A receptor, and blocking 5-HT synthesis or HTR3A receptor impaired tumor progression in colorectal cancer mouse models.

Normal and tumor tissues were obtained from surgical resections from 21 patients diagnosed with colorectal cancer from Nanjing First Hospital (Nanjing, P.R. China) between 2017 and 2018 (age > 18). All tissues were sliced into two pieces from the middle axis; one was fixed in 4% paraformaldehyde, and the other was stored at −80°C. The research was conducted according to the principles expressed in the Declaration of Helsinki. Written informed consent was obtained from all the patients. The study was approved by the Ethics Committee of Nanjing First Hospital and the Clinical Research Ethics Committee of China Pharmaceutical University (Nanjing, Jiangsu, P.R. China).

The chemicals and reagents used in this article are listed in Supplementary Table S1.

THP-1 (RRID: CVCL_0006) and CT26 (RRID: CVCL_7254) cell lines were obtained from the Cell Bank of Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences (Shanghai, P.R. China) and cultured in RPMI1640 medium with 10% FBS (Biological Industries). Before experiments, THP-1 cells were differentiated into macrophages by phorbolmyristate acetate (PMA, 500 nmol/L) treatment for 12 hours. The immortalized bone marrow–derived macrophage (iBMDM) cell line was a gift from Feng Shao (ref. 23; National Institute of Biological Science, Beijing, P.R. China) and cultured in DMEM medium with 10% FBS. SW480 (RRID: CVCL_AT68), SW620 (RRID: CVCL_0547), HT29 (RRID: CVCL_0320), and HEK293T (RRID: CVCL_0063) cell lines were obtained from the Cell Bank of Shanghai Institute of Biochemistry and Cell Biology (Shanghai, P.R. China). NCM460 (RRID: CVCL_0460) and HCT116 (RRID: CVCL_0291) cell lines were obtained from ATCC. THP-1, HT-29, HCT116, SW620, SW480, and CT26 cell lines were authenticated by short tandem repeat DNA test within the past year (Shanghai Biowing Applied Biotechnology, Shanghai, P.R. China); iBMDM, NCM460, and HEK293T cell lines were not further authenticated in the past year. The cell lines were used for experiments within 10 passages after thawing. All cell lines were free of Mycoplasma contamination. Primary colon epithelial cells (CEC) were isolated from BALB/c mice as described previously (24). Briefly, colon tissues from BALB/c mice were dissected and washed in cold PBS, then cut into 1.5-mm pieces and digested in RPMI1640 medium containing DNase I (150 μg/mL, Beyotime Biotechnology) and collagenase I (150 U/mL, Sigma-Aldrich) with 10% FBS, and shook for 1.5 hours (37°C, 5% CO₂). Digested tissue suspensions were passed through a cell strainer (70 μmol/L) and separated by a 40%/80% Percoll gradient centrifugation (300 × g, 15 minutes). Separated CECs were cultured in RPMI1640 medium with 10% FBS.

NCM460, SW480, SW620, HCT116, HT29, THP-1, CEC, CT26, and iBMDM cells were seeded in 6-well plates with 1 × 10⁶ cells per well, and THP-1 cells were differentiated into macrophages by PMA (500 nmol/L, 12 hours). After a 24-hour incubation, cells were harvested either for TPH1 mRNA expression detection by real-time PCR (as described below) or for TPH1 protein expression detection by Western blotting (as described below). NCM460, SW480, SW620, HCT116, and HT29 cells were seeded in 6-well plates with 1 × 10⁶ cells per well. After a 24-hour incubation, cells were harvested for aromatic L-amino acid decarboxylase (AADC) and monoamine oxidase A (MAOA) mRNA expression detection by real-time PCR.

The lentiviral pLVX-shRNA2 vector for short hairpin (shRNA) silencing and packaging plasmids PsPAX2 and pMD2.G were purchased from MiaoLing Plasmid Sharing Platform. For human- and mouse-sourced NLRP3, TPH1 and HTR3A shRNA silencing, we designed two shRNA sequences for targeting, and related lentivirus particles were generated in HEK293T cells. Cells were transfected with lentivirus encoded NLRP3, TPH1, or HTR3A shRNA, and stably transfected cells were selected by puromycin (1 μg/mL). Targeting sequences of shRNA for NLRP3, TPH1, or HTR3A are listed in Supplementary Table S2.

A total of 8-week-old wild-type male C57BL/6J (RRID: IMSR_JAX: 000664) and BALB/c (RRID: IMSR_ORNL: IE-BALB/c) mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. NLRP3^−/− C57BL/6J (IMSR, catalog no. JAX: 021302, RRID: IMSR_JAX:021302) mice were purchased from the Jackson laboratory at 5 weeks of age. For the CT26 and iBMDM coimplanted subcutaneous allograft colorectal cancer mouse model, 6 BALB/c mice were randomly assigned to each group: CT26 wild-type + iBMDM wild-type, CT26 shTPH1 + iBMDM wild-type, CT26 wild-type + iBMDM shHTR3A, CT26 wild-type + iBMDM shNLRP3, CT26 wild-type + iBMDM shNLRP3-shHTR3A, CT26 shTPH1 + iBMDM shHTR3A, CT26 wild-type, and CT26 shNLRP3. Cells were mixed with the designated combinations (1 × 10⁶ CT26 cells and 1 × 10⁶ iBMDMs for each mouse), suspended in PBS, and then subcutaneously transplanted into axilla of mice. Tumor sizes were measured every 2 days using a caliper, and tumor volumes were calculated as (width² × length)/2. For the azoxymethane (AOM)/dextran sodium sulfate (DSS)–induced colorectal cancer mouse model, wild-type (Beijing Vital River Laboratory Animal Technology) and NLRP3^−/− (Jackson laboratory) male C57BL/6J mice were co-housed fed for 2 weeks, and then 8 NLRP3^−/− and wild-type mice were randomly assigned to each designated group of AOM/DSS-induced colorectal cancer mouse model (wild-type mice for: control, AOM/DSS, AOM/DSS + pCPA, AOM/DSS + TPS, and AOM/DSS + pCPA + TPS; NLRP3^−/− mice for: AOM/DSS + NLRP3^−/− and AOM/DSS + NLRP3^−/− + TPS). Mice were intraperitoneally injected with AOM (10 mg/kg) as the first day of the experiment, and followed by three cycles of 2.5% (w/v) DSS in the drinking water for 1 week and then normal drinking water for 2 weeks (normal drinking water was used in entire process as negative control), and experiment was ended after three cycles of DSS administration (9 weeks; ref. 25). 4-chloro-DL-phenylalanine (pCPA, 100 mg/kg, i.p.), tropisetron (TPS, 10 mg/kg, i.p.), and combination of pCPA (100 mg/kg, i.p.) and TPS (10 mg/kg, i.p.) were administrated to AOM/DSS-challenged mice every other day from the first day to the end of the experiment. Mice in control group were intraperitoneally injected with normal saline. Colons, serum, heart, liver, spleen, lung, kidney, stomach, duodenum, and small intestine were harvested. Lengths of colons were measured by a caliper, and tumor numbers of colons were counted at the end of experiment. All animal experiments were carried out in accordance with protocols approved by the Animal Care and Use Committee of China Pharmaceutical University (Nanjing, Jiangsu, P.R. China).

Western blot analyses were carried out as described previously (26). Briefly, human and mouse cell and tissue (colorectal cancer tumor and normal colon tissue) lysates were obtained by using RIPA lysis buffer following the manufacturer's instructions (P0013B, Beyotime, Shanghai, P.R. China), and protein concentrations were determined by a BCA protein assay kit (P0012S, Beyotime, Shanghai, P.R. China). Lysates (30 μg of total protein), as well as lysates (10 μL) and cross-linked pellets (10 μL) from the ASC oligomerization assay, and reaction solution (10 μL) from in vitro CaMKIIα kinase assays were then separated by SDS-PAGE gel [separating gel: 1.9 mL ddH₂O, 1.7 mL Acryl/Bis 30% Solution, 1.3 mL 1.5 mol/L Tris-HCl (pH8.8), 0.05 mL 10% SDS (w/v), 0.05 mL 10% APS substitute (w/v), and 0.002 mL TEMED; spacer gel: 1.4 mL ddH₂O, 0.33 mL Acryl/Bis 30% Solution, 0.25 mL 1.0 mol/L Tris-HCl (pH6.8), 0.02 mL 10% SDS (w/v), 0.02 mL 10% APS substitute (w/v), and 0.002 mL TEMED] and transferred to polyvinylidene difluoride membrane for immunoblotting analysis using the following antibodies: anti-NLRP3 (1: 1,000, Cell Signaling Technology, catalog no. 15101, RRID:AB_2722591), anti-cleaved IL1β (1: 1,000, Cell Signaling Technology, catalog no. 2021, RRID:AB_2280224), anti-cleaved caspase-1 (1: 1,000, Proteintech, catalog no. 22915-1-AP, RRID:AB_2876874), anti-pro-IL1β (1: 1,000, Santa Cruz Biotechnology, catalog no. sc-52012, RRID:AB_629741), anti-pro–caspase-1 (1: 1,000, Santa Cruz Biotechnology, catalog no. sc-392736, RRID:AB_2890615), anti-ASC (1: 1,000, Santa Cruz Biotechnology, catalog no. sc-514414, RRID:AB_2737351), anti-HTR3A (1: 1,000, ABclonal, catalog no. A5647, RRID:AB_2766407), anti-TPH1 (1: 1,000, Affinity Biosciences, catalog no. DF6465, RRID:AB_2838427), anti-CaMKIIα (1: 1,000, Santa Cruz Biotechnology, catalog no. sc-13141, RRID:AB_626789), anti-phospho-CaMKIIα (1: 1,000, Affinity Biosciences, catalog no. AF3493, RRID:AB_2834931), anti-human phospho-NLRP3 (1: 1,000, Ser198, Tao Li lab-National Center of Biomedical Analysis, catalog no. LiTao-001, RRID:AB_2890616), anti-mouse phospho-NLRP3 (1: 1,000, Affinity Biosciences, catalog no. AF3555, RRID:AB_2846869), anti-IKKβ (1: 1,000, Affinity Biosciences, catalog no. AF6009, RRID:AB_2834943), anti-phospho-IKKβ (1:1,000, Affinity Biosciences, catalog no. AF3009, RRID:AB_2834448), anti-NFkB p65 (1: 1,000, Cell Signaling Technology, catalog no. 8242, RRID:AB_10859369), anti-phospho-NFkB p65 (1: 1,000, Affinity Biosciences, catalog no. AF2006, RRID:AB_2834435), anti-β-actin (1: 5,000, Bioworld Technology, catalog no. AP0714, RRID:AB_2890614). Goat anti-mouse IgG (H + L, 1:2,000, Proteintech, catalog no. SA00001-1) as well as goat anti-rabbit IgG (H + L, 1:2,000, Proteintech, catalog no. SA00001-2) were used as secondary antibodies. Protein bands were visualized by enhanced chemiluminescence system Tanon 5200-multi (Biotanon) according to the manufacturer's instructions. Densitometry quantification for proteins' expression was carried out using Image J software (RRID:SCR_003070). β-Actin served as an internal control. For Coomassie blue staining of samples from the in vitro CaMKIIα kinase assay, gels were stained with Coomassie blue staining solution (catalog no. P0017B, Beyotime) for 2 hours by gently shaking, and then destained with Coomassie blue staining destaining solution (catalog no. P0017C, Beyotime, destaining solution was exchanged per 30 minutes) for 2 hours by gently shaking.

Total RNA from human and mouse cells and tissues (colorectal cancer tumor and normal colon tissue) were extracted using TRIzol reagent (catalog no. 15596026, Life technologies) with manufacturer's instructions, and RNA was qualified by a micro-spectrophotometer (Nano-100, Hangzhou Allsheng Instruments). Reverse transcription reactions were performed by EasyScript First-Strand cDNA Synthesis Super Mix (catalog no. AE301, Transgen) with a reverse PCR instrument (TP600, TAKARA). mRNA expression was quantified by a real-time PCR instrument (LightCycler 96 System, Roche) using ChamQSYBRqPCR Master Mix (catalog no. Q421-02, Vazyme) following the manufacturer's instructions, and the reaction procedure was: preincubation: 95°C, 30 seconds; amplification: 95°C, 10 seconds and 60°C, 30 seconds for 40 replicates in a 20 μL reaction system (500 ng DNA template). Primers for targeted genes are listed in Supplementary Table S3. The relative expression of targeted genes was normalized by using the internal control GAPDH and calculated using the 2^–ΔΔC_t method.

For supernatants, human and mouse cells (THP-1, NCM460, SW480, SW620, HCT116, HT29, iBMDM, CT26, and CEC) were seeded in 6-well plates with 1 × 10⁶ cells per well, and THP-1 cells were differentiated into macrophages by PMA (500 nmol/L, 12 hours) before further treatments. Supernatants were collected from cell culture medium of THP-1, NCM460, SW480, SW620, HCT116, HT29, iBMDM, CT26, and CEC by centrifugation (1,000 × g, 4°C, 5 minutes) after incubation for 24 hours for basal 5-HT detection.

Treatment 1: Conditional medium from SW620 and HT29 cells (with/without 0.5 μmol/L LP-533401 treatment for 24 hours) was used to incubate THP-1 cells for 24 hours, and then THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour); conditional medium from CT26 cells (with/without 0.5 μmol/L LP-533401 treatment for 24 hours) was used to incubate iBMDM for 24 hours, and then iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour). Treatment 2: THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 5-HT treatment (10 μmol/L) for 5 hours. Treatment 3: Conditional medium from wild-type and stably transfected with vector/shTPH1 of SW620 or HT29 cells was used to incubate THP-1 cells for 24 hours, and then THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour). Treatment 4: THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with 5-HT (10 μmol/L, 5 hours) and sarpogrelate (SGR, a HTR2 antagonist, 10 μmol/L), or TPS (a HTR3 antagonist, 10 μmol/L), or SB-269970 (a HTR7 antagonist, 10 μmol/L) for 5 hours. Treatment 5: THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour), or MSU (500 μg/mL, 1 hour), or Nig (10 μmol/L, 1 hour) with/without 10, 50, 100 nmol/L 2-methyl-5-hydroxytryptamine (2Met5HT) for 5 hours. Treatment 6: THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) or MSU (500 μg/mL, 1 hour) or Nig (10 μmol/L, 1 hour) with/without 5-HT (10 μmol/L) and 10, 50, 100 μmol/L TPS treatments for 5 hours. Treatment 7: THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 2Met5HT (100 nmol/L) or the CaMKIIα specific inhibitor KN-62 (1 μmol/L) for 5 hours. Treatment 8: SW620 and HT29 cells were treated with either recombinant human IL1β (rhIL1β, 10 μg/mL) or rhIL1β (10 μg/mL) and IL1 receptor inhibitor (TLR1, 100 μmol/L) for 24 hours. Treatment 9: THP-1 and iBMDMs were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without IL1β production inhibitor diacerein (Dia, 10 μmol/L) for 5 hours, and then medium was replaced by fresh medium for another 6-hour incubation. Conditional medium from THP-1 cells was then collected and used to incubate SW620 and HT29 with/without recombinant human IL1 receptor antagonist protein (IL1RA, 0.1 μg/mL) or TLR1 (100 μmol/L) or IKKβ inhibitor (BMS-345541, 4 μmol/L) for 24 hours; conditional medium from iBMDMs was collected and used to incubate CT26 with/without TLR1 (100 μmol/L) or BMS-345541 (4 μmol/L) for 24 hours.

After the indicated treatments, supernatants were collected from cell culture medium by centrifugation (1,000 × g, 4°C, 5 minutes). A total of 50 μL of each supernatant was used for ELISAs. For mouse colorectal cancer tumor and normal colon tissue samples, 100 mg tissues were cut into small sections (1 mm) and homogenized by glass homogenizer in 1 mL RIPA buffer in an ice bath. The tissue lysates were collected by centrifugation (3,000 × g, 4°C, 5 minutes), and 100 μL of each tissue lysates were used for ELISA. IL1α, IL1β, IL18, IL33, and 5-HT concentrations were determined by ELISA kits (IL1β: catalog no. MK1198, Boster Biological Technology; IL1α, IL18, and IL33: catalog no. SBJ-H0472, SBJ-H2288, SBJ-H0104, SenBeiJia Technology; 5-HT: catalog no. E-EL-0033c, Elabscience) with the manufacturers' instructions. Concentrations were calculated according to formula obtained from standard curve of absorbance–concentration.

THP-1 cells were collected and lysed by RIPA lysis buffer on ice, and lysates were collected by centrifugation (12,000 × g, 4°C, 15 minutes). Cell lysates (200 μL) were then incubated with anti-ASC (2 μg) at 4°C overnight, and precipitated with protein A+G agarose beads (P2055, Beyotime) for another 4 hours at 4°C. The beads were then washed with PBS five times and collected by centrifugation (1,000 × g, 4°C, 5 minutes). The beads were separated by SDS-PAGE gels and immunoblotting analyses were with anti-NLRP3, anti-ASC, and anti-pro–caspase-1 antibodies as described above.

THP-1 and iBMDM cells were seeded in 6-well plates with 1 × 10⁶ cells per well, and THP-1 cells were differentiated into macrophages by PMA (500 nmol/L, 12 hours) before further treatments. Treatment 1: THP-1 cells (wild-type, stably transfected with vector/shHTR3A) were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with 5-HT (10 μmol/L, 5 hours) treatment. Treatment 2: THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 5-HT (10 μmol/L) and TPS (100 μmol/L) for 5 hours. Treatment 3: THP-1 cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 2Met5HT (100 nmol/L) for 5 hours. Treatment 4: iBMDMs were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) and 5-HT (10 μmol/L, 5 hours) with/without TPS (100 μmol/L, 5 hours). Treatment 5: iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hours) with/without 2Met5HT (100 nmol/L, 5 hours). Treatment 6: THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 2Met5HT (100 nmol/L) and KN-62 (1 μmol/L) for 5 hours. After treatments as indicated, THP-1 and iBMDM cells were washed with cold PBS and lysed in an ice-cold buffer (20 mmol/L HEPES-KOH, pH 7.5; 150 mmol/L KCl, 1% NP-40S, 0.1 mmol/L phenylmethylsulfonylfluoride) with shearing 30 times through a 21-gauge needle. Bulk nuclei were removed by centrifugation (1,800 × g, 4°C, 8 minutes). 50 μL of the lysates were collected for Western blot analysis (as described above) for ASC as an input control. The rest of lysates were then cross-linked with 2 mmol/L fresh disuccinimidyl suberate for 30 minutes at 37°C. The cross-linked pellets were separated by SDS-PAGE gels, and anti-ASC immunoblotting analyses was performed as described above. Qualification of ASC oligomerization was calculated: Σ (optical density of ASC monomer and oligomer in pellets)/optical density of ASC in lysates.

THP-1 or iBMDM cells were seeded in 35-mm cell culture dishes (1 × 10⁵ cells/dish), and meanwhile THP-1 cells were differentiated into macrophages by PMA (500 nmol/L, 12 hours) before further treatments. THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 5-HT (10 μmol/L) plus TPS (100 μmol/L), or 2Met5HT (100 nmol/L) for 5 hours. After treated as indicated, cells were then fixed with 4% paraformaldehyde for 15 minutes, then blocked and permeabilized [PBS, 10% (v/v) FBS, 0.5% Triton X-100] for 60 minutes at room temperature, prior to staining for perinuclear ASC with an ASC primary antibody (1:200, Santa Cruz Biotechnology, catalog no. sc-514414, RRID:AB_2737351) overnight at 4°C. Cells were then stained with secondary goat anti-mouse keyFluor488 antibody (1: 200, KeyGen BioTech, catalog no. KGAB010) for 60 minutes at room temperature, and nucleus was stained with 4′,6-diamidino-2-phenylindole (DAPI, 1 mmol/L) for 15 minutes. Cells were imaged by a laser scanning confocal microscope (Olympus FV3000).

Intracellular Ca²⁺ of THP-1 and iBMDM cells were detected by using Fluo-4 AM fluorescence probe kit (catalog no. S1060, Beyotime) as described previously (27, 28). Cells were seeded in 6-well plates (1 × 10⁶ cells/well) and 96-well plates (5 × 10⁴ cells/well), and meanwhile THP-1 cells were differentiated into macrophages by PMA (500 nmol/L, 12 hours) before further treatments. THP-1 and iBMDM cells were activated by LPS (100 ng/mL, 4 hours) and ATP (5 mmol/L, 1 hour) with/without 5-HT (10 μmol/L) plus TPS (100 μmol/L) or 2Met5HT (100 nmol/L) treatments for 5 hours. After treated as indicated, cells were washed three times with cold PBS and incubated with 5 μmol/L Fluo-4 AM working solution for 30 minutes at 37°C, then Fluo-4 AM was removed, and cells were incubated in PBS for another 30 minutes. Cells in 6-well plates were imaged by an inverted fluorescence microscope (Nikon Ts2R) with an excitation wavelength of 488 nm, and cells in 96-well plates were used for fluorescence intensity detection by a fluorescence microplate reader (Infinite 200 PRO, Tecan) with an excitation wavelength of 488 nm and an emission wavelength of 516 nm. Relative fluorescence intensity was calculated: Fluorescence Intensity_{(treatment group)}/Fluorescence Intensity_{(control group)} × 100%.

E.coli BL21 (catalog no. EC1001, Shanghai Weidi Biotechnology) were transfected with pGEX-4T-3 (RRID: Addgene_79149) bacterial expression plasmids encoding GTS-human NLRP3^ΔLRR (wild type or Ser198Ala mutation) and GST-CaMKIIα (wild-type or Lys42Arg mutation). Recombinant protein expression was induced by 1 mmol/L IPTG for 4 hours at 37°C. When the medium absorbance value of OD600 was 0.6, GST-tagged proteins were purified by using a GST-tagged protein purification kit (catalog no. P2262, Beyotime) with manufacturer's instructions. Briefly, bacteria were collected by centrifugation (4,000 × g, 4°C, 20 minutes) and lysed with ice-cold lysis buffer (catalog no. P2262-2, Beyotime) with ultrasonic fragmentation (procedure: 200 W, 10 seconds, and 10 seconds for interval, six cycles; ultrasonic cell disruptor: TL-250Y, Jiangsu Tianling Instruments). Supernatants were collected by centrifugation (10,000 × g, 4°C, 20 minutes), and then were loaded onto BeyoGold GST-tag resin columns, eluted with glutathione containing (10 mmol/L) elution buffer to obtain the purified GST-tagged proteins.

CaMKIIα kinase activity assay was performed as previously described with modification (29). Briefly, CaMKIIα was added to a final concentration of 100 nmol/L into a reaction system (50 μL) containing 10 mmol/L PIPES pH 7.0, 25 μg/mL CaM, 0.1 mg/mL BSA, 0.25 mmol/L CaCl₂, 50 mmol/L ATP, and 100 μmol/L substrates (wild-type or S198A mutation NLRP3). Reactions were performed for 30 minutes at 37°C. Reactions were stopped by adding SDS-PAGE gel loading buffer (12.5 μL, catalog no. P0015L, Beyotime) and boiling for 5 minutes. Samples were separated by 8% SDS-PAGE gels and then analyzed by immunoblotting as described above.

Tissues from patients with colorectal cancer or mice were fixed with 4% paraformaldehyde overnight and embedded in paraffin, and then cut into 6-μmol/L sections and put onto microscopic slides. For histologic analysis, slides were stained with hematoxylin and eosin (H&E). For IHC analysis, slides were incubated with antibodies against TPH1 (1: 100, Affinity Biosciences, catalog no. DF6465, RRID:AB_2838427), cleaved (C)-casp1 (1: 200, Cell Signaling Technology, catalog no. 2021, RRID:AB_2280224), NLRP3 (1: 200, Cell Signaling Technology, catalog no. 15101, RRID:AB_2722591), phospho-NLRP3 (Ser194, 1: 100, Affinity Biosciences, catalog no. AF3555, RRID:AB_2846869), proliferation marker Ki67 (1: 200, Proteintech, catalog no. 27309-1-AP, RRID: AB_2756525), macrophages marker F4/80 (1: 200, Proteintech, catalog no. 28463-1-AP, RRID:AB_2881149), and iBMDM maker GP70 (1: 100, Santa Cruz Biotechnology, catalog no. sc-65452, RRID: AB_831325; ref. 30) for 24 hours at 4°C, and then stained by using 3, 30-diaminobenzidine followed by hematoxylin counterstain. For immunofluorescence analysis, slides were incubated with antibody against TPH1 (1: 100, Affinity Biosciences, catalog no. DF6465, RRID: AB_2838427) for 24 hours at 4°C, and then incubated with goat anti-mouse keyFluor488 (1: 200, KeyGen BioTech, catalog no. KGAB010) as secondary antibody for another 2 hours at 37°C. DAPI stained for 15 minutes at room temperature. Samples were imaged by an inverted fluorescence microscope (Nikon Ts2R).

RNA sequencing data and patients' clinical information of colorectal cancer (TCGA-COAD) were downloaded from The Cancer Genome Atlas (TCGA) data portal (https://portal.gdc.cancer.gov/) by using the GCD tool. Patients who had TPH1 expression data and integrated overall survival information (status and follow-up time) were selected, and 380 patients were enrolled in this study. For overall survival analysis, patients were divided into TPH1-high expression group (n = 76, top 20%) and TPH1-low expression group (n = 304, bottom 80%), and analyzed by log-rank test using SPSS 18.0 software (RRID: SCR_002865), and Kaplan–Meier curves were built by GraphPad Prism 6.0 software (RRID: SCR_002798). Gene expression profiling of human colorectal cancer cell lines were downloaded from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/, GEO accession ID: GSE59857), which included 153 colorectal cancer cell lines and two normal colorectal epithelial cell lines (FHS and FHC). Relative TPH1 expression was calculated: TPH1 expression/[(TPH1 expression_(FHS) + TPH1 expression_(FHC))/2].

For in silico candidate kinase prediction for NLRP3, the FASTA files of NLRP3 for both human (AAI43360.1) and mouse (NP_665826.1) were downloaded from the National Center for Biotechnology Information (NCBI) protein database (https://www.ncbi.nlm.nih.gov/protein), and then the FASTA files were uploaded to NetPhos 3.1 Server online tool (http://www.cbs.dtu.dk/services/NetPhos-3.1/), and candidate kinases were predicted by following the instructions (31).

Data are presented as mean ± SD. Statistical analysis was carried out by using SPSS 18.0 software (RRID: SCR_002865). Paired Student t test was used to compare the difference in TPH1 expression between normal and tumor tissues from 21 patients with colorectal cancer. Unpaired Student t test were used for comparisons between two groups. One-way ANOVA were used for multiple comparisons. P < 0.05 was regarded as statistically significant.

To understand the role of 5-HT in colorectal cancer progress, we first analyzed the relationship between mRNA expression of TPH1 in tumor tissues and overall survival in 380 colorectal cancer cases from TCGA database. Analysis revealed that patients with colorectal cancer with higher TPH1 expression had lower overall survival rates (Fig. 1A). We next compared mRNA and protein expression of TPH1 in tumor and normal tissues from 21 patients with colorectal cancer and found that both mRNA and protein expression of TPH1 were significantly higher in tumor tissues (Fig. 1B–E). Data from the GEO database (GSE59857) also indicated that TPH1 mRNA expression in colorectal cancer cell lines were higher than normal colorectal epithelial cell lines, although it was not statistically significant (Fig. 1F). We next detected both mRNA and protein expression of TPH1 in five colorectal cancer cell lines (SW480, SW620, HCT116, HT29, and CT26), two normal colorectal epithelial cells (NCM460 and CEC) and two macrophages (THP-1 and iBMDM); and, mRNA expression of AADC (another enzyme in 5-HT biosynthesis) and MAOA (enzyme in 5-HT degradation) in NCM460, SW480, SW620, HCT116, and HT29 cells was also detected. The results showed that expression of TPH1 and AADC were upregulated, whereas MAOA was downregulated in colorectal cancer cells compared with normal colorectal epithelial cells; expression of TPH1 was absent in macrophages (Fig. 1G–I; Supplementary Fig. S1A–S1D). 5-HT concentrations were significantly higher in supernatants from cultured colorectal cancer cells compared with normal colorectal epithelial cells, and not detectable in the supernatants from macrophages (Fig. 1J and K), indicating that 5-HT biosynthesis and secretion are upregulated in colorectal cancer cells.

We next asked whether overproduction of 5-HT by colorectal cancer cells could facilitate inflammation in the colorectal cancer microenvironment. Incubation with conditional medium from SW620 and HT29 cells significantly enhanced cleaved-caspase-1 (C-casp1) and cleaved-IL1β (C-IL1β) production under LPS+ATP stimulation in THP-1 cells (Fig. 2A–C; Supplementary Fig. S1E and S1F), and similar results were observed in mouse-sourced iBMDM cells (Fig. 2D and E; Supplementary Fig. S1G and S1H). To further investigate whether the enhanced inflammation was induced by 5-HT secretion, 5-HT biosynthesis was blocked by a TPH1-selective inhibitor (LP-533401) or an anti-TPH1 shRNA in SW620 and HT29 cells. Both methods significantly blunted enhancement of C-casp1 and C-IL1β production in THP-1 cells (Fig. 2F–L; Supplementary Fig. S1I–S1L; Supplementary Fig. S2A–S2C), and similar results were observed in iBMDM cells (Fig. 2M and N; Supplementary Fig. S2D and S2E). To assess whether 5-HT could enhance IL1β production, THP-1 and iBMDM cells were exposed to exogenous 5-HT, resulting in increased C-casp1 and C-IL1β production (Supplementary Fig. S2F–S2L). In agreement with the in vitro data, co-implantation of iBMDM and CT26 cells increased expression of C-casp1 and C-IL1β in tumor tissues compared with implantation of CT26 cells alone, whereas silencing TPH1 in CT26 cells decreased expression of C-casp1 and C-IL1β in established CT26 and iBMDM coimplanted subcutaneous tumors (Fig. 2O; Supplementary Fig. S2M; Supplementary Fig. S3A and S3B). These data indicated that colorectal cancer cell 5-HT biosynthesis and secretion increases C-IL1β production in macrophages.

To find out how 5-HT enhanced inflammation in macrophages, we analyzed the mRNA expression profile of 5-HT receptors in THP-1 and iBMDM cells. As shown in Fig. 3A, mRNA expression of HTR2/3/7 receptors were detected in both THP-1 and iBMDM cells. Selective antagonists of HTR2, HTR3, and HTR7 were next used to test whether blocking the receptors would impair the inflammation-enhancing effect of 5-HT. The results showed that only blocking HTR3 with TPS, but not HTR2 (by SGR) or HTR7 (by SB-269970), significantly impaired C-casp1 and C-IL1β production (Fig. 3B and C; Supplementary Fig. S3C–S3F), whereas silencing HTR3A had similar effects (Fig. 3D; Supplementary Fig. S3G–S3K). As shown in Fig. 3E–H, Supplementary Fig. S3L, Supplementary Fig. S4, and Supplementary Fig. S5A and S5B, HTR3-specific agonist 2Met5HT dose dependently enhanced production of C-casp1 and C-IL1β, whereas TPS dose dependently impaired enhancement of C-casp1 and C-IL1β production induced by 5-HT in THP-1 cells treated with different NLRP3 inflammasome activators (ATP, MSU, and Nig). Similar results were observed in iBMDM cells (Supplementary Fig. S5C–S5J). TPS also impaired 5-HT–induced IL18 production in THP-1 cells, while having no significant effects on IL1α production. 5-HT had no effects on IL33 production (Supplementary Fig. S5K–S5M). In agreement with in vitro data, silencing HTR3A or NLRP3 in iBMDM cells decreased expression of C-casp1 and C-IL1β in tumor tissues, whereas silencing both HTR3A and NLRP3 in iBMDM cells did not further decrease C-casp1 and C-IL1β in the coimplanted allograft colorectal cancer model (Fig. 3I; Supplementary Fig. S2M; Supplementary Fig. S3A and S3B). These results indicated that 5-HT induces NLRP3-mediated inflammation via HTR3A in the colorectal cancer microenvironment.

We next investigated how 5-HT/HTR3A regulated NLRP3 inflammasome activation. We found that neither TPS nor 2Met5HT had any significant influence on mRNA expression of NLRP3, IL1B, and IL18 or protein expression of NLRP3, ASC, pro–caspase-1, and pro-IL1β in THP-1 cells (Fig. 4A–C; Supplementary Fig. S5N–S5R); colorectal cancer cell conditional medium also had no significant impact on mRNA expression of NLRP3, IL1B, and IL18 in THP-1, or Nlrp3, Il1b, and Il18 in iBMDM cells (Supplementary Fig. S6A and SS6B), indicating that 5-HT/HTR3A is not involved in the priming step of NLRP3 inflammasome activation. We then examined whether 5-HT/HTR3A would affect NLRP3 inflammasome assembling using coimmuno-precipitation, ASC oligomerization, and ASC speck staining assays. Silencing HTR3A with anti-HTR3A shRNA or blocking HTR3A with TPS significantly impaired NLRP3 inflammasome assembling, whereas activating HTR3A with 2Met5HT enhanced assembling in THP-1 cells (Fig. 4D–J; Supplementary Fig. S6C–S6E). Similar results were observed in iBMDM cells (Fig. 4J; Supplementary Fig. S6F–S6I). These results indicated that 5-HT/HTR3A enhances NLRP3 inflammasome assembling, without affecting the priming step.

Because NLRP3 phosphorylation (human: Ser198, mouse: Ser194) is indispensable for NLRP3 inflammation activation and because HTR3A, a ligand-gated ion channel, has been reported to lead to Ca²⁺ influx and Ca²⁺/calmodulin-dependent protein kinase IIα (CaMKIIα) activation (32), we hypothesized that the Ca²⁺/CaMKIIα cascade possibly involved in 5-HT/HTR3A–mediated NLRP3 inflammasome activation by phosphorylating NLRP3. We used an on-line bioinformatics tool to predict the candidate kinases for NLRP3 phosphorylation at Ser198 (mouse: Ser194), and CaMKIIα was predicted as one of the possible kinases. The effect of 5-HT/HTR3A on Ca²⁺ influx was next assessed in macrophages, and the results indicated that blocking HTR3A by TPS impaired 5-HT–induced Ca²⁺ influx, whereas activating HTR3A by 2Met5HT increased Ca²⁺ influx (Fig. 5A–D). We then investigated the effects of 5-HT/HTR3A on phosphorylation activity of CaMKIIα in macrophages and found that activating HTR3A by 2Met5HT time dependently induced CaMKIIα phosphorylation (Thr286), which was blocked by TPS treatment (Fig. 5E; Supplementary Fig. S6J and S6K), whereas neither 2Met5HT nor TPS had any effects on mRNA expression of CAMK2A (Supplementary Fig. S6L). Effects of 5-HT/HTR3A signaling, as well as CaMKIIα on NLRP3 phosphorylation, were next investigated in macrophages, and results showed that 2Met5HT increased NLRP3 phosphorylation (human: Ser198, mouse: Ser194), which was impaired by KN-62 (a specific inhibitor for CaMKIIα; Fig. 5F and G). Kinase assays indicated that wild-type CaMKIIα induced NLRP3 phosphorylation, whereas phosphorylated NLRP3 was not detected in the context of kinase-dead CaMKIIα or phosphorylation site-mutated NLRP3 (Mut) proteins (Fig. 5I). The results also indicated that inhibition of CaMKIIα (by KN-62) impaired C-casp1 and C-IL1β production, as well as ASC oligomerization induced by 2Met5HT, but had no influence on protein expression of NLRP3 inflammasome components (NLRP3, ASC, pro–caspase-1) and pro-IL1β (Fig. 5H and J; Supplementary Fig. S6M–S6U; Supplementary Fig. S7A–S7F). In line with in vitro data, coimplantation of iBMDM and CT26 cells increased phosphorylated NLRP3 (Ser194) in tumor tissues compared with implantation of CT26 cells alone, whereas silencing TPH1 in CT26 cells or silencing HTR3A in iBMDM cells decreased phosphorylated NLRP3 compared with coimplantation of wild-type CT26 and iBMDM cells in the coimplanted allograft colorectal cancer model (Supplementary Fig. S2M; Supplementary Fig. S7G and S7H). These results demonstrated that the Ca²⁺/CaMKIIα cascade is involved in 5-HT/HTR3A–induced NLRP3 inflammasome activation by inducing NLRP3 phosphorylation.

Because TPH1 expression was upregulated in colorectal cancer cells, as well as in tumor tissues from patients with colorectal cancer and because IL1β has been reported to induce 5-HT secretion in gut mucosal cells (33), we next investigated whether IL1β could induce 5-HT biosynthesis in colorectal cancer cells. Conditional medium from activated macrophages was used to incubate colorectal cancer cells as shown in Fig. 6A. Incubation with conditional medium from activated macrophages significantly increased mRNA and protein expression of TPH1, as well as 5-HT concentrations in the supernatant of colorectal cancer cells, which was impaired by Dia (an inhibitor for IL1β production) treatment, whereas incubation with conditional medium from inactivated macrophages had no significant effects on TPH1 expression (Fig. 6B–D; Supplementary Fig. S7I–S7L). To further demonstrate whether IL1β could induce TPH1 expression, rhIL1β was used to treat SW620 and HT29 cells, and rhIL1β significantly increased TPH1 expression and 5-HT concentrations in supernatants of SW620 and HT29 cells, which was reversed by TLR1 (Supplementary Fig. S7M and S7N; Supplementary Fig. S8A–S8C). Given IL1β has been reported to mainly activate the intracellular NFκB signaling cascade to regulate downstream events by binding to IL1 receptor (34), we next examined phosphorylation NFκB and IKKβ in colorectal cancer cells after incubation with conditional medium from activated macrophages and found that phosphorylation NFκB and IKKβwere upregulated in colorectal cancer cells, whereas conditional medium from Dia-treated macrophages had no significant influence on them (Fig. 6D; Supplementary Fig. S8D–S8F). IL1RA, TLR1, and IKKβ inhibitor (BMS-345541) were then used to block IL1β/IL1R/NFκB signaling, and results showed that conditional medium from activated macrophages no longer increased expression of TPH1, as well as 5-HT production in colorectal cancer cells (Fig. 6E–I; Supplementary Fig. S8G–S8J). In line with our in vitro findings, co-implantation of iBMDM and CT26 cells increased TPH1 expression and 5-HT concentrations in tumor tissues compared with implantation of CT26 cells alone, whereas silencing HTR3A or NLRP3 in iBMDM cells decreased mRNA and protein expression of TPH1, as well as 5-HT concentrations, compared with implantation of wild-type CT26 and iBMDM cells. Silencing both HTR3A and NLRP3 in iBMDM cells did not further decrease TPH1 expression and 5-HT concentrations compared with silencing HTR3A or NLRP3 alone (Fig. 6J; Supplementary Fig. S2M; Supplementary Fig. S8K–S8M). These results showed that IL1β production induces 5-HT biosynthesis by upregulating TPH1 expression in colorectal cancer cells. On the basis of above results, overproduced 5-HT in colorectal cancer cells enhances NLRP3 inflammasome activation in macrophages, and in return, IL1β produced by the NLRP3 inflammasome activation induces 5-HT biosynthesis in colorectal cancer cells, demonstrating a positive feedback loop between 5-HT and NLRP3 in colorectal cancer microenvironment.

We next asked whether the 5-HT–NLRP3 positive feedback loop played a role in colorectal cancer progression. The CT26 and iBMDM coimplanted subcutaneous allograft colorectal cancer mouse model was used, and macrophages were identified by positive staining of an iBMDM-specific marker (envelope glycoprotein GP70; ref. 30), as well as F4/80 in tumor tissues (Fig. 7A; Supplementary Fig. S8N and S8O; Supplementary Fig. S9A). We found that coimplantation with iBMDM cells significantly promoted CT26 tumor growth (increased tumor volume and weight), as well as increased Ki67 expression in tumor tissues compared with implantation of CT26 cells alone, whereas silencing TPH1 in CT26 cells or silencing HTR3A or NLRP3 in iBMDM cells impaired tumor growth. Silencing both TPH1 in CT26 cells and HTR3A in iBMDM cells further impaired tumor growth. Tumor volume and weight, as well as Ki67 expression, in mice with silenced NLRP3 in iBMDM cells was comparable with silencing both HTR3A and NLRP3, indicating that the tumor growth–promoting effects of 5-HT/HTR3A were NLRP3 dependent (Fig. 7B–D; Supplementary Fig. S9A). No significant differences were observed between wild-type and NLRP3 silencing on tumor volume and weight, expression of Ki67, TPH1, C-casp1, C-IL1β, and 5-HT concentrations in tumor tissues when CT26 cells were implanted alone; protein expression of NLRP3 was much lower in colorectal cancer cells than in macrophages (Fig. 7B–D; Supplementary Fig. S9A–S9D; Supplementary Fig. S2M; Supplementary Fig. S3A and S3B; Supplementary Fig. S8K–S8M), which implies that NLRP3 expressed by macrophages, but not by colorectal cancer cells, promotes colorectal cancer tumor growth.

To further evaluate the role of 5-HT/HTR3A in colorectal cancer progression, we established an AOM/DSS-induced colorectal cancer mouse model using wild-type and NLRP3^−/− mice (Fig. 7E). We found that NLRP3 knockout, pCPA (TPH1 inhibitor), or TPS (HTR3A antagonist) treatment significantly ameliorated AOM/DSS-induced colorectal cancer progression, with reduced tumor numbers, alleviated pathologic progression, and increased colon lengths, whereas the combination of pCPA and TPS had a more substantial effect (Fig. 7F–H). NLRP3 knockout, as well as pCPA or TPS treatment, significantly deceased protein expression of C-casp1, C-IL1β, TPH1, as well as 5-HT concentrations, in tumor tissues compared with AOM/DSS group, and a further decrease was observed in combination group. In agreement with results from the coimplanted allograft colorectal cancer model, TPS treatment in NLRP3^−/− mice had no significant decreases in tumor progression, as well as expression of C-casp1, C-IL1β, TPH1, and 5-HT concentrations in tumor tissues. Phosphorylation of NLRP3 in tumor tissues was downregulated by pCPA or TPS treatment (Fig. 7I; Supplementary Fig. S9E–S9J). These findings demonstrated that the 5-HT–NLRP3 positive feedback loop in the colorectal cancer microenvironment enhances tumor growth and progression, whereas blocking the feedback loop could slow down progression.

Peripheral 5-HT has been reported to play a promoting role in gastrointestinal tract inflammation and carcinogenesis (35–37). In our study, we found that 5-HT biosynthesis and secretion was upregulated in colorectal cancer cells, and overproduced 5-HT–enhanced NLRP3 inflammasome activation in macrophages; on the other hand, NLRP3 inflammasome–mediated IL1β release induced 5-HT biosynthesis in colorectal cancer cells, forming a positive feedback loop between 5-HT and NLRP3 signaling in the colorectal cancer microenvironment, thus facilitating colorectal cancer tumor growth and progression. Mechanistically, 5-HT bound to the HTR3A receptor on macrophages, which led to Ca²⁺ influx, subsequently induced CaMKIIα phosphorylation and activation, which then induced NLRP3 phosphorylation and activation. In return, the NLRP3 inflammasome activation led to IL1β production and secretion, which induced TPH1 transcription via IL1R/NFκB signaling, resulting in increased 5-HT biosynthesis (Supplementary Fig. S10).

Consistent with our findings, TPH1 expression and 5-HT biosynthesis have been demonstrated to be upregulated in several types of cancer, such as breast cancer, hepatocellular cancer, pancreatic ductal adenocarcinoma, and cholangiocarcinoma (35–38). However, the role of overproduced 5-HT in innate immune regulation in the tumor microenvironment is not well investigated, with some data indicating that TPH1 might be indirectly involved in immune regulation by competing with indoleamine 2, 3-dioxygenase 1 (IDO-1) in tryptophan metabolism (39, 40). Data from animal experiments have shown that 5-HT can enhance colitis and inflammatory bowel disease (41–43), indicating that there might be a more initiative role of 5-HT signaling in innate immune regulation. On the other hand, uncontrolled and distensible inflammatory reactions mediated by inflammasomes has been well documented to play a key role in colorectal cancer carcinogenesis. Therefore, we evaluated the effects of 5-HT, overproduced by colorectal cancer cells, on inflammation in the colorectal cancer microenvironment. Our data indicated that 5-HT can enhance NLRP3-mediated inflammation via activating its HTR3A receptor on macrophages. In line with our findings, data from animal experiments have demonstrated that HTR3A antagonists can attenuate colitis, pancreatitis, sepsis, and mucositis (22, 44–46). However, the detailed mechanism of how HTR3A accelerates inflammation still remains unclear. Our study showed that intracellular Ca²⁺/CaMKIIα signaling was responsible for the proinflammatory effects of 5-HT/HTR3A via upregulating NLRP3 phosphorylation, which is an essential step during NLRP3 inflammasome activation. In line with our finding, as a ligand-gated ion channel, HTR3A-induced cellular Ca²⁺ influx and CaMKIIα activation has been detected in vivo (32). Ca²⁺/CaMKIIα signaling is reported here as a novel regulator for NLRP3 phosphorylation, in addition to IL1 receptor–associated kinase 1/4 (IRAK1/4)-c-Jun N-terminal kinase (JNK1) signaling reported by Song and his colleagues (11). We also investigated how 5-HT biosynthesis was upregulated in colorectal cancer cell lines, as well as in tumor tissues from patients with colorectal cancer. We demonstrated that IL1β produced by macrophages could induce TPH1 expression via IL1R/NKκB signaling, which resulted in increased 5-HT biosynthesis in colorectal cancer cells. In agreement with our finding, it has been reported that IL1β or LPS treatment can upregulate 5-HT production (33, 47).

We further evaluated the therapeutic potential of blocking 5-HT biosynthesis or HTR3A receptor on colorectal cancer progression. The in vivo study indicated that both TPH1 inhibitor and HTR3A antagonist treatment had significant amelioration effects on colorectal cancer tumor growth and progression, whereas combination treatment had a more substantial effect. We speculate that this was because of blocking the two pivotal nodes of the 5-HT–NLRP3 positive feedback loop, which could result in stronger inhibition. Nevertheless, other possibilities cannot be excluded, as research has demonstrated that 5-HT autosecreted by pancreatic ductal adenocarcinoma cells effectively enhance cell proliferation via Warburg effects, and 5-HT autosecreted by triple-negative breast cancer cells promote breast cancer invasion and proliferation via upregulation of matrix metalloproteinase-9 and VEGF (35, 38). Further attempts to investigate whether autosecreted 5-HT has direct effects on colorectal cancer cell proliferation and migration will be interesting.

In conclusion, we found that there is a 5-HT–NLRP3 positive feedback loop which helps to maintain persistent inflammation to promote tumor growth in the colorectal cancer microenvironment. Targeting the 5-HT–NLRP3 feedback loop offers a potential strategy for colorectal cancer therapy.

No disclosures were reported.

T. Li: Data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft. B. Fu: Validation, investigation, visualization. X. Zhang: Investigation. Y. Zhou: Validation. M. Yang: Investigation. M. Cao: Investigation. Y. Chen: Investigation. Y. Tan: Investigation. R. Hu: Conceptualization, resources, supervision, funding acquisition, validation, writing–review and editing.

This work was supported by the National Natural Science Foundation of China (numbers 81672816, 81872337, and 82073185). The authors thank Feng Shao from National Institute of Biological Science (Beijing, P.R. China) for providing the iBMDM cell line and Tao Li from National Center of Biomedical Analysis (Beijing, P.R. China) for providing the anti-human phospho-NLRP3 (Ser198) antibody for this work.

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.