Neutrophil-derived serine proteases induce FOXA2-mediated autophagy dysfunction and exacerbate colitis-associated carcinogenesis via protease activated receptor 2.

Introduction
Colitis-associated cancer (CAC) is a subtype of colorectal cancer (CRC) that develops because of chronic inflammation in individuals with inflammatory bowel disease (IBD). IBD, including ulcerative colitis (UC) and Crohn disease, is characterized by chronic relapsing-remitting inflammation of the intestinal tract. During relapse, the intestinal mucosa is actively inflamed and features neutrophil infiltration, a reliable proxy in IBD disease activity scoring. CAC burden is positively correlated with disease activity, including the severity and extent of inflammation [1]. Although excessive inflammation aggravates tissue damage, the mechanisms whereby inflammation promotes CAC have not been fully elucidated.
Autophagy is an important regulator of cellular homeostasis and is involved in CAC development. It is an evolutionarily conserved lysosome-mediated process that facilitates cell adaptation to various stressors and maintains epithelial cell homeostasis by removing harmful components and damaged organelles [2]. Accumulating evidence indicates that autophagy impairment aggravates inflammation and promotes CAC by regulating intestinal immunity, reactive oxygen species (ROS), barrier function, and cell death [3,4]. Autophagy is a highly controlled process consisting of several distinct stages. Upon formation of the ULK1/Atg1 complex, autophagy is mediated by a series of autophagy-related genes involved in the formation of bilayer autophagosomes containing cellular components. After fusion with lysosomes (autolysosomes), the cargo is degraded by numerous lysosomal hydrolases (such as cathepsins and acid phosphatase). The activity of hydrolytic enzymes highly depends on the acidic environment (pH 4.5-5.5) that is maintained by a proton-pumping vacuolar-type H±translocating ATPase (V-ATPase) in the lysosomal membrane [5]. V-ATPase is a large multi-subunit complex comprising V0 and V1 domains that catalyze ATP hydrolysis and transport protons into the lysosome [6]. Any interruption in the degradation of SQSTM1/p62 (sequestosome 1) impedes autophagic flux. Although the knockdown of different autophagy-related genes can accelerate CAC development in mice [7], the extracellular signaling associated with autophagy and CAC development under inflammatory conditions remains unknown.
Among all cancer types, CRC displays the highest frequency of mutated G-protein-coupled receptors, which suggests its potential role in extracellular signaling associated with colon carcinogenesis [8]. These receptors include F2RL1/PAR2 (F2R like trypsin receptor 1), which is expressed in various cell types, including intestinal epithelial cells (IECs), and plays a critical role in inflammation and carcinogenesis [8]. The extracellular N terminus of F2RL1/PAR2 is cleaved by the activating protease PRSS/trypsin at Arg36/Ser37, exposing the embedded activating sequence, which binds to and triggers the canonical signal. Interestingly, some serine proteases, such as those derived from neutrophils (including ELANE [elastase, neutrophil expressed], PRTN3 [proteinase 3], and CTSG [cathepsin G]), cleave downstream of this site, leading to inactivation of the canonical signal [9]. In addition, autocrine activation of F2RL1/PAR2 promotes CRC progression by enhancing the self-renewal, proliferation, and migration of cancer cells [10–12]. However, systemic knockdown exacerbates CAC in mice [7]. Although reshaping the immunosuppressive microenvironment seems to explain this discrepancy, it is unknown whether there is an epithelial mechanism promoting CAC. In the present study, we used cancer model mice with epithelial-specific deletion of F2rl1 (f2rl1[ΔIEC]) to investigate the role of epithelial F2RL1/PAR2 signaling in autophagy and CAC.

Results

IEC-specific knockout of F2rl1 impairs autophagy and exacerbates AOM-DSS-induced carcinogenesis in mice
To investigate the effect of intestinal epithelial F2rl1 on carcinogenesis, we generated mice with epithelial-specific knockout of F2rl1 (f2rl1[IEC]). There were no differences between f2rl1[IEC] and F2rl1f/f mice in terms of body weight or the architecture and length of colon tissues (Figure S1A-C). To induce CAC, the mice were exposed to AOM (azoxymethane, 12.5 mg/kg) and 2.5% DSS (dextran sulfate sodium) (Figure S1D and E). All mice showed various symptoms of colitis. The disease activity index, based on diarrhea, proctoptosis, and hematochezia, in f2rl1[IEC] mice was higher than that in their littermates (4.8 v.s. 8.4) one week after AOM-DSS administration. Disease activity was reevaluated via colonoscopy five weeks later based on the statuses of edematous mucosa, erythema, intestinal stenosis, and fistula formation. The disease activity score was also higher in f2rl1[IEC] mice than in their littermates, indicating that f2rl1 deletion in IEC further promoted inflammation and damage after AOM-DSS exposure (Figure 1A). At the end of the experiment (nine weeks after AOM-DSS administration), tumors were observed in the proximal and middle colons of almost all model mice; higher tumor numbers and burden were observed in f2rl1[IEC] mice than in their littermates (Figure 1B-D). Real-time PCR analysis of colon tissues revealed higher levels of CAC-related inflammatory mediators, including Nos2/Inos, Il6, Mcpt1, and Tnf/Tnf-α, in f2rl1[IEC] mice than in their littermates (Figure 1E). These results demonstrated that the IEC-specific knockout of F2rl1 exacerbated the AOM-DSS-induced carcinogenesis in mice.

To explore the underlying mechanisms, the mouse colon transcriptome was analyzed. In addition to the inflammation response, differentially expressed genes (DEGs) between f2rl1[IEC] and F2rl1f/f mice were enriched in pathways related to endosomes, autophagy, and macromolecular damage (Figure 1F). Detection of autophagy markers via western blotting and immunohistochemistry revealed the accumulation of MAP1LC3/LC3 (microtubule associated protein 1 light chain 3)-II and SQSTM1/p62 (markers for autophagy function) despite the simultaneous increase in the expression of the autophagy initiation marker MAP1LC3/LC3-I in colon tissues of f2rl1[IEC] mice after AOM-DSS (Figure 1G, H and Figure S1F). The presence of autophagosomes, detected via transmission electron microscopy (TEM) (Figure 1I), strongly suggested that f2rl1 deletion impaired autophagic flux in IECs after AOM-DSS treatment. Consistent with the autophagy defect, fewer MKI67 proliferating cells (Figure 1H and Figure S1G) and smaller tumor size (percentage of d ≤0.2 mm tumor, 68% vs 36%) were observed in f2rl1[IEC] mice than in their littermates. These results indicated that epithelial deletion of F2rl1 exacerbated the CAC associated with autophagic defects in IECs in vivo.

F2RL1/PAR2 deficiency impairs autophagic flux by reducing lysosome acidification in IECs
To investigate whether F2RL1 deficiency impairs autophagy in IECs, we stably knocked down F2RL1 with a specific shRNA (shF2RL1) in different colon cell lines [13] (Figure 2A), most of which showed autocrine/paracrine activation of F2RL1/PAR2 [8] and high levels of basal autophagy [14]. Consistent with our hypothesis, F2RL1 knockdown induced simultaneous accumulation of MAP1LC3/LC3-II and SQSTM1/p62 in various cell lines (Figure 2A), suggesting autophagic flux impairment [15]. TEM revealed increased numbers of double-membrane structures containing organelles and aggregates in shF2RL1 cells (Figure 2B), indicating the accumulation of non-digestive autophagosomes upon autophagy defect in F2RL1-deficient IECs. In addition, transient blockade of F2RL1/PAR2 signaling with inhibitors induced the accumulation of MAP1LC3/LC3-II and SQSTM1/p62 (Figure S2A).

Next, we investigated whether autophagosome-lysosome fusion or lysosome-related degradation was affected by F2RL1 deficiency. We found that shF2RL1 reduced the colocalization of the autophagosome marker MAP1LC3/LC3 with the lysosome marker LAMP1 (lysosomal associated membrane protein 1) (Figure 2C). Moreover, after transfection with the mCherry-EGFP-MAP1LC3/LC3 plasmid, shF2RL1 cells showed strong retention of EGFP fluorescent signals that colocalized with mCherry (Figure 2D and Figure S2B), suggesting the accumulation of nonacidic autophagosomes. Consistently, shF2RL1 treatment reduced the number of acidic lysosomes, as evidenced by LysoSensor Green staining (Figure 2E and Figure S2C). The inhibition of lysosomal acidification was substantiated by the suppressed activities of classical lysosomal enzymes, ACP2 (acid phosphatase 2, lysosomal) (Figure 2F), and CTSB (cathepsin B) (Figure 2G and Figure S2D), both of which depend on acidic environments. In contrast, F2RL1/PAR2 activation with an activating peptide (F2RL1/PAR2-AP) promoted SQSTM1/p62 degradation, autophagosome-lysosome fusion (Figure S2E-G), lysosomal acidification (Figure S2H), and CTSB activity (Figure S2I). These findings indicate that F2RL1/PAR2 signaling is essential for maintaining autophagic flux and that its deficiency interferes with lysosomal acidification in IECs.

F2RL1/PAR2 regulates lysosomal acidification through the lysosomal proton pump subunit ATP6V0E1
As the V-ATPase is responsible for lysosomal acidification, we investigated whether F2RL1 deficiency suppresses its expression. Transcriptome analysis revealed the downregulation of four subunits, including ATP6V0C (ATPase H+ transporting V0 subunit c), ATP6V0E1 (ATPase H+ transporting V0 subunit e1), ATP6V1F (ATPase H+ transporting V1 subunit F), and ATP6V1C2 (ATPase H+ transporting V1 subunit C2), by shF2RL1 in HT29 cells (Figure 3A). We analyzed the mRNA and protein levels of these four subunits and found that shF2RL1 collectively suppressed their expression in different IECs, whereas F2RL1/PAR2-AP increased their expression (Figure 3B,C). Among the four subunits, the change in ATP6V0E1 expression was most consistent with F2RL1/PAR2 activity. Silencing ATP6V0E1 with siRNA impeded lysosomal acidification, autophagosome-lysosome fusion, and impaired autophagic flux (Figure 3D-F). Most importantly, forced expression of ATP6V0E1, but not of the other subunits, effectively reversed the impairment of lysosomal acidification (Figure 3D), autophagosome-lysosome fusion (Figure 3E), and autophagy (Figure 3F and Figure S3A-I) upon F2RL1 silencing. These findings suggested that F2RL1/PAR2 deficiency reduced lysosomal acidity by inhibiting the activity of the lysosomal proton pump subunit ATP6V0E1 in IECs.

F2RL1/PAR2 regulates the transcription of ATP6V0E1 via YAP1-FOXA2
Actinomycin D, an inhibitor of global transcription abrogated the F2RL1/PAR2-AP-stimulated ATP6V0E1 mRNA expression (Figure S4A), we investigated the transcription factors (TFs) that may mediate the regulation of ATP6V0E1 by F2RL1/PAR2. We observed an overlap between the genes downregulated by shF2RL1 and TFs potentially binding to the ATP6V0E1 promoter region, with FOXA2 (forkhead box protein A2) being the most significantly altered among the four TFs (Figure 4A). We verified that F2RL1 depletion significantly reduced the mRNA and protein expression and nuclear localization of FOXA2 in different IECs (Figure 4B-D) whereas F2RL1/PAR2 activation resulted in increased FOXA2 mRNA and protein levels (Figure S4B and C). The siRNA-mediated inhibition of FOXA2 downregulated ATP6V0E1 (Figure 4E and Figure S4D-E) and impaired lysosomal acidification, autophagosome-lysosome fusion, and autophagic flux (Figure 4E,F). FOXA2 overexpression reversed the shF2RL1-induced downregulation of ATP6V0E1 and impairment of lysosomal acidification (Figure 4F and Figure S4F) while recovering autophagic flux (Figure 4E,F). Considering the inability of other TFs (ZNF600 [zinc finger protein 600], RXRB [retinoid X receptor beta], and GATAD2A [GATA zinc finger domain containing 2A]) to reverse the reduction in ATP6V0E1 expression (Figure 4A and Figure S5A-C), FOXA2 emerged as the prominent TF mediating the transcriptional regulation of ATP6V0E1 by F2RL1/PAR2.

Next, we determined whether ATP6V0E1 is a novel target gene of FOXA2 through Chromatin immunoprecipitation sequencing (ChIP-seq) analysis and observed the binding of FOXA2 to the proximal ATP6V0E1 promoter (Figure 4G). The luciferase reporter assay showed that FOXA2 dose-dependently increased the ATP6V0E1 promoter-driving activity, which was completely abolished by deletion of the FOXA2-binding motif (Figure 4H and Figure S5D). Finally, ChIP-qPCR confirmed the direct binding of FOXA2 to the ATP6V0E1 promoter (Figure 4I). These findings suggested that F2RL1 deficiency impeded lysosomal acidification through the transcriptional suppression of ATP6V0E1 via FOXA2.
To answer how F2RL1/PAR2 regulates FOXA2, we found that pretreatment with actinomycin D completely abolished F2RL1/PAR2-AP-induced FOXA2 mRNA, suggesting regulation at the transcriptional level (Figure S5E). Through analysis of the FOXA2 promoter region, we identified a putative YAP1-TEAD binding site (Figure S5F). Our previous study has shown that F2RL1/PAR2 regulates YAP1 protein stability through phosphatase PP1 in DSS-induced colitis [16]. Importantly, forced YAP1 expression reversed FOXA2 and ATP6V0E1 suppression as well as obstruction of autophagy caused by F2RL1 deficiency and/or ECP treatment (Figure 4J). These findings suggested that F2RL1/PAR2 signaling regulates FOXA2 transcriptionally through YAP1-TEAD.

Autophagy impairment caused by F2RL1 deficiency aggravates ROS-induced DNA damage via FOXA2 in IECs
As oxygen stress-induced DNA damage plays a crucial role in CAC pathogenesis [17], we investigated whether F2RL1 deficiency exacerbates DNA damage through impaired autophagy in vitro and in vivo. Consistent with the effects of the autophagy inhibitor BafA1 (bafilomycin A1), shF2RL1 treatment increased the cellular levels of ROS (Figure 5A) and γH2AX (H2A.X variant histone; phosphorylated) (Figure 5B) -a marker of DNA damage. Treatment with NAC (N-Acetylcysteine), a scavenger of ROS, markedly blocked the BafA1-or shF2RL1-induced ROS and γH2AX accumulation but did not affect the autophagy defect (Figure 5A,B), indicating that impaired autophagy led to oxygen stress and, consequently, DNA damage. Forced expression of FOXA2 or ATP6V0E1 mitigated the shF2RL1 -related ROS production (Figure 5C and Figure S6A-B) and γH2AX and 8-oxo-dG accumulation, a marker of oxidative damage (Figure 5D-E and Figure S6C-D). In the model of AOM-DSS-induced CAC, both the mRNA and protein levels of FOXA2 and ATP6V0E1 were lower in the colons of f2rl1[IEC] mice than in those of their littermates (Figure 5F,G). Transcriptome analysis identified the enrichment of the ROS and redox signaling (Figure 5H) and DNA damage response pathways (Figure 1F) in f2rl1[IEC] mice. IHC (Immunohistochemistry) analysis revealed a substantial reduction in FOXA2 expression and γH2AX accumulation in vitro and in vivo (Figure 5I,J). These results indicated that the autophagy impairment caused by F2RL1 deficiency aggravated the ROS-induced DNA damage and contributed to CAC via FOXA2-ATP6V0E1 in IECs.

Neutrophil-derived serine proteases impede autophagy through F2RL1/PAR2
As F2RL1/PAR2 inactivation aggravates carcinogenesis by impairing autophagy, it is important to explore endogenous F2RL1/PAR2-regulating factors in CAC. Activated neutrophils are the most important immune cells related to colitis activity and release a large number and variety of serine proteases into the extracellular space [9], including ELANE, CTSG, and PRTN3, all of which are capable of cleaving and inactivating canonical F2RL1/PAR2 signaling (Figure 6A). Unsurprisingly, treatment with neutrophil-derived serine proteases, individually or in combination, reduced the expression of FOXA2 (Figure 6B,C) and ATP6V0E1 (Figure 6C and Figure S7A) accompanied by impaired autophagy (Figure 6E); all these changes were successfully reversed by FOXA2 overexpression in different IECs, including NCM460, a human immortalized colon cancer cell line (Figure 6B-D and Figure S7A-B). In contrast, canonical activation of F2RL1/PAR2 with PRSS/trypsin or F2RL1/PAR2-AP enhanced the expression of FOXA2 and ATP6V0E1 and facilitated glutamine deprivation-induced autophagy (Figure 6E-G and Figure S7C-D).

Notably, knockout of F2RL1 with CRISPR-Cas9 completely abolished the inhibitory effect of neutrophil-derived serine proteases on FOXA2 and ATP6V0E1 expression, lysosomal acidification, and autophagy obstacle (Figure 6H-K and Figure S7E). Knock-in F2RL1 reversed the downregulation of FOXA2 and ATP6V0E1 expression, lysosomal acidification, and autophagy obstacle caused by F2RL1 knockout, but not those caused by ECP (Figure 6H-K and Figure S7E). These findings suggested that the serine proteases released by neutrophils impeded autophagy via the FOXA2-mediated disruption of lysosomal acidification in a F2RL1/PAR2-dependent manner.

Neutrophil levels correlate with F2RL1/PAR2-FOXA2 expression in colitis and CRC
To determine the roles of neutrophils in disease pathogenesis, we examined the correlation between neutrophils and F2RL1/PAR2-related signals in colitis, CAC, and sporadic CRC specimens using multiple datasets from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases. FOXA2 was downregulated, whereas the neutrophil infiltration index was upregulated in UC specimens compared with those in the healthy controls (Figure 7A and Figure S7F). Compared with sporadic CRC samples, the CAC specimens from patients with a history of IBD showed extremely low FOXA2 expression (Figure 7B). In the CRC TCGA dataset, FOXA2 levels were lower in the high than in the low neutrophil index group (Figure 7C), indicating a correlation between neutrophil activity and FOXA2 suppression in colitis, CAC, and sporadic CRC. FOXA2 expression was strongly correlated with the expression of its target gene, ATP6V0E1, in UC specimens (Figure 7D). The positive correlations among F2RL1, FOXA2, and ATP6V0E1 levels were also corroborated in CRC samples (Figure 7D). Finally, multiple immunofluorescent histochemical staining revealed that the nuclear frequency of FOXA2 in epithelial cells (FOXA2+ PANCK+) was lower in patients with active UC than in those with inactive UC (Figure 7E,F) and was negatively correlated with the density of activated neutrophils (MPO+) (Figure 7G). The density and intensity of FOXA2 accumulation in epithelial cells were significantly reduced within 25 μm from MPO+ neutrophils compared with the total average in UC specimens (Figure 7H), consistent with the observations in CRC specimens with infiltrated neutrophils (Figure 7I,J). Notably, within a certain range, the nuclear frequency of FOXA2 increased as the distance from neutrophils increased in tumor samples (Figure 7K) but not in UC samples (data not shown). Collectively, these results suggest that neutrophil activity is correlated with F2RL1/PAR2-FOXA2 expression in UC and CRC. Inactivation of F2RL1/PAR2 canonical signaling by neutrophil-derived serine proteases may contribute to colonic carcinogenesis through FOXA2-mediated autophagy dysfunction.

Discussion
CAC is a classic inflammation-driven cancer characterized by persistently high levels of ROS and reactive nitrogen species, which irreversibly oxidize DNA and cellular biomolecules, representing the primary carcinogens. Autophagy exerts powerful antioxidant and DNA damage-repair effects by clearing oxidized biomolecules [18]. In the present study, the autophagy disorder induced by F2RL1/PAR2 deficiency in IECs exacerbated DNA damage and ROS production in epithelial and malignant cells, subsequently exacerbating tumor burden in mice. Notably, our findings support the bidirectional role of autophagy in tumorigenesis. On the one hand, autophagy disorder promotes cancer incidence through homeostasis imbalance; on the other hand, it suppresses tumor growth through impaired nutritional adaptability. Our results showed that neutrophil-derived proteases can act as extracellular signals that cause autophagy defects in CAC. Moreover, IBD and CRC data revealed a negative correlation between FOXA2 expression and neutrophils. Neutrophils and their derived proteins and proteases are commonly used biomarkers of disease activity in IBD, which is highly associated with an elevated risk of CAC development [19–22]. This strongly suggests that restoring autophagic flux may effectively prevent CAC, especially in patients with UC and uncontrolled inflammation.
F2RL1/PAR2 participates in various intestinal pathological processes, such as pain, IBD, inflammation, and tumors [23]. Although the F2RL1/PAR2-induced autophagy protects IECs from DSS-induced damage in high-fat diet-fed mice [24], the non-autonomous regulation of F2RL1/PAR2 during inflammation remains unclear. Neutrophils, the most abundant innate immune cells, play a key role in IBD development [25]. In addition to their catabolic and destructive activities, neutrophil-derived serine proteases can also signal pain and inflammation in the gut via the G-protein-coupled receptor F2RL1/PAR2 [23]. Here, we revealed that these enzymes exert novel autophagy-related functions by silencing canonical F2RL1/PAR2 signaling. The canonical activation of F2RL1/PAR2 through proteolytic cleavage of the N terminus by an extracellular serine protease induces Gαq-mediated Ca2+ mobilization. Due to the different cleavage sites [23], neutrophil-derived enzymes either silence F2RL1/PAR2 through disarming the receptor (such as CTSG and PRTN3) or induce biased F2RL1/PAR2 signaling (such as ELANE) mediated by ARRB/β-arrestin, which negatively regulates GNAQ/Gαq-Ca2+ signaling and downstream targets [26]. Here, the opposed effects of neutrophil-derived proteases and trypsin on autophagy strongly suggest the regulation of FOXA2 by GNAQ-Ca2+ signaling. Further research is required to explore the underlying mechanisms. While CD8+ and CD4+ T cells exhibited a significant decrease and the immunosuppressive activity of MDSCs increased in f2rl1 KO tumors compared to WT tumors, RNA-seq analysis of f2rl1[IEC] for immune infiltration indicated no alterations in the infiltration of immune cells, including CD8+, CD4+, MDSCs, and neutrophils (data not shown). These findings do not definitively exclude the potential that F2RL1/PAR2 expressing in epithelial cells has no impact on neutrophil infiltration, which could be attributed to the timing of the sampling, which was conducted during a phase of tumor development when neutrophil infiltration is usually low.
Lysosomal function is regulated globally at the transcriptional level. TFEB, a member of the MiT-TFE family, functions as a master transcription factor in lysosome biogenesis [27]. Conditional deletion of TFEB in IECs increases the susceptibility of these cells to epithelial cell injury and subsequent colitis, indicating the relevance of TFEB in the gut [28]. We found that F2RL1 deficiency increased TFEB expression but did not decrease it. Consistent with its function, the levels of lysosomes and their markers were elevated in shF2RL1 cells compared to those in the control cells (data not shown). However, this lysosomal acidification was disturbed through the transcriptional downregulation of various subunits of V-ATPase, including ATP6V0E1, ATP6V1C2, and ATP6V1F, established via a novel TF, FOXA2. Notably, ChIP-seq analysis revealed more pronounced binding signals for FOXA2 than for TFEB on the promoters of genes encoding the V-ATPase subunits considered in this study. Furthermore, the study findings suggested that different TFs separately regulate the quality and quantity of lysosomes, whereas FOXA2 is more important for lysosomal maturation in IECs.
This study has some limitations. Although we demonstrated that FOXA2 was downregulated in UC and CAC specimens, it is difficult to determine the extent to which CAC can be attributed to long-term autophagy impairment caused by FOXA2 deficiency due to sample size limitations. In addition to neutrophil-derived serine proteases, numerous other proteases, such as mast cell-derived tryptase, coagulation factor FXa, and gut microbiota-derived trypsin-like proteases [23], may regulate F2RL1/PAR2 under physiological and pathological conditions. Since the F2RL1/PAR2-FOXA2 axis enhances lysosomal and autophagy functions in IECs, we hypothesized that these F2RL1/PAR2-activating enzymes may participate in homeostasis maintenance through autophagy, warranting further investigations.
In conclusion, our results suggest that F2RL1/PAR2 deficiency in IEC suppresses FOXA2 expression, leads to V-ATPase-mediated autophagy dysfunction, and exacerbates CAC. Neutrophils may play an essential role in the impairment of autophagy and aggravation of CAC through inactivation of the F2RL1/PAR2 canonical signal via their serine proteases. We suggest that F2RL1/PAR2 signaling participates in the maintenance of intestinal homeostasis through autophagy. These findings provide useful insights into the roles of F2RL1/PAR2 and its-cleaving serine proteases in CAC and would help in developing novel therapeutic strategies for this malignancy.

Materials and methods

Reagents and antibodies

Reagents
AOM (Sigma-Aldrich, A5486), dextran sulfate sodium salt (MP Biomedicals, MFCD00081551), LysoSensor™ Green DND-189 (Thermo Scientific, L7535), LysoTracker Deep Red (Thermo Scientific, L12492), cathepsin B Assay (Abcam, ab270772), GB88 (MedChemExpress, HY-120261), F2RL1/PAR2-IN-1 (MedChemExpress, HY-138558), F2RL1/PAR2-AP (Shanghai Apeptide, N/A), ENMD-1068 (Abcam, ab141699), mitoSox Green (Thermo Scientific, M36005), Lipofectamine 2000 (Thermo Scientific 11668019), DAPI (Sigma-Aldrich, F6057), polyethylenimine (Polysciences 23966–2), TB Green® Premix Ex Taq™ (TAKARA, RR820A), actinomycin D (MedChemExpress, HY-17559), bafilomycin A1 (MedChemExpress, HY-100558), L-Glutamine (Thermo Scientific 25030081), Cell Dissociation Buffer Enzyme-Free PBS-based (Gibco 13151–014).

Antibodies
For western blot, MAP1LC3/LC3 (Cell Signaling Technology 12741), SQSTM1/p62 (Cell Signaling Technology 23214), γH2AX (Abcam, ab81299), F2RL1/PAR2 (Abcam, ab180923), ATP6V0C (Abcam, ab220323), ATP6V1C2 (Abcam, ab176771), ATP6V0E1 (Aviva Systems Biology, OACA00864), ATP6V1F (Abcam, ab190789), FOXA2 (Abcam, ab108422), histone H3 (Cell Signaling Technology, 9715), ZNF600 (Proteintech 20100–1-AP), RXRB (Proteintech 14684–1-AP), GATAD2A (Proteintech 12294–1-AP), TUBB/Beta Tubulin (Proteintech 10068–1-AP), Anti-rabbit IgG, HRP-linked Antibody (Cell Signaling Technology, 7074).
For immunofluorescence, EPCAM (Proteintech, 1050–1-AP), γH2AX (Abcam, ab2893), 8-oxo-dG (R&D Systems, 4354-MC-050), LAMP1 (Cell Signaling Technology, 9091), PANCK (Abcam, ab7753), MPO/myeloperoxidase (Abcam, ab208670), FOXA2 (Abcam, ab256493), Anti-rabbit conjugated to Alexa Fluor 488 (Abcam, ab150077), Anti-rabbit conjugated to Alexa Fluor 594 (Abcam, ab150080).

Plasmids
pCMV-EGFP-mCherry-LC3B (Beyotime, D2816), pcDNA3.1-FOXA2, pCMV-ATP6V0E1, pCMV-ATP6V0C, pCMV-ATP6V1C2, pCMV-ATP6V1F, pCMV-ZNF600, PCMV-GATAD2A, pGL3-basic-ATP6V0E1, and pGL3-basic-ATP6V0E123–33 were synthesized directly by chemical method (Mailgene Biosciences Co., ltd) and verified by sequencing using 5’-CTAGCAAAATAGGCTGTCCC-3’ vector universal primers.

Critical commercial assays
Nuclear and cytoplasmic extracts kit (Thermo Scientific 78,835), Dual-Luciferase Reporter Assay (Promega, E1960), ROS Assay Kit (Beyotime, S0033M), ACP2/acid phosphatase Assay (Solarbio® life sciences, BC2135).

Cell culture and treatment
IEC6, HT29, CaCO2, HEK-293T, DLD1, SW620, and RKO cells were purchased from the American Type Culture Collection/ATCC (CRL1592, HTB-38, HTB-37, CRL-3216, CCL-221, CCL-227, CRL-2577). NCM460 was gifted by Professor Hong Yang (Peking Union Medical College, Chinese Academy of Medical Sciences). IEC6, NCM460, CaCO2, and HEK-293T were grown in DMEM (Hyclone, SH30022.01). DLD1 cells were grown in RPMI 1640 medium (Hyclone, SH30809.01). SW620, RKO, and HT29 cells were grown in DMEM/F-12 (Hyclone, SH30023.01). The medium was supplemented with 10% fetal bovine serum (Cell Technologies 242,332) and 1% penicillin-streptomycin (Hyclone, SV30010).

Assay for LC3+ autophagic vacuoles using mCherry-EGFP- MAP1LC3/LC3 plasmid
Cells were grown on glass chamber slides overnight and transfected with a plasmid encoding mCherry-EGFP-MAP1LC3/LC3 (Beyotime, D2816) for 48 h. The cells were rinsed twice with 1× PBS (Hyclone, SH30022.01) and fixed in 4% paraformaldehyde for 30 min at room temperature. After rinsing twice with 1× PBS, the sections were mounted with DAPI (Sigma-Aldrich, F6057) and analyzed.

Confocal laser scanning microscopy
To detect the pH of lysosomes, mitoSOX, or intracellular CTSB (cathepsin B) activities, cells were grown in 35-mm glass-bottom dishes and cultured to 80% confluence. The cells were then treated with 0.5 μM LysoSensor Green DND-189, 2 μM MitoSox Green, or 20 µL of a reconstituted Magic Red staining solution for 30 min or 1 h, washed twice with PBS, and incubated in fresh medium for another 30 min. Finally, 1 μg/mL Hoechst 33,342, trihydrochloride, trihydrate (Invitrogen, H3570) was added to the medium for nuclear staining before imaging.

Dual-luciferase reporter assay
HEK-293T cells were grown in glass chamber slides overnight and then transfected with the pGL3-basic-ATP6V0E1, pGL3-basic-ATP6V0E123–33, and pcDNA3.1-FOXA2 plasmids for 36 h. Luciferase activity was determined using a commercial kit according to the manufacturer’s instructions (Promega, E1980).

Chromatin immunoprecipitation (ChIP)-qPCR
ChIP assays were performed using a ChIP Assay Kit (EMD Millipore Corporation, 17–295) following the manufacturer’s instructions. Data are presented as the relative enrichment normalized to normal IgG levels and histone H3 modification enrichments normalized to histone H3 levels. The ChIP-qPCR primer sequences used in this study are listed in Table 1.

Animal models and ethical statement
All mice-related studies were approved by the Institutional Review Board of the Chinese Academy of Medical Sciences Cancer Institute (NCC2021A264). The CRISPR-Cas9 technology was used to modify F2rl1; 6- to 8-weeks-old male F2rl1-flox mice (background: C57BL/6JGpt, strain no. T052020) were purchased from GemPharmatech (Nanjing, China). The study design and outcomes followed the ARRIVE guidelines.
For establishing the CAC model, male mice (bred under specific pathogen-free conditions) were intraperitoneally injected with 12.5 mg/kg azoxymethane (AOM) and then administered 2.5% DSS via their drinking water for five days. Colonoscopy was performed to monitor colon tumor formation after five weeks of AOM-DSS administration.

Mouse genotype identification
The brief process is as follows: gRNA was transcribed in vitro. Cas9 and gRNA were microinjected into the fertilized eggs of C57BL/6JGpt mice (strain NO. T052020). Fertilized eggs were transplanted to obtain positive F0 mice which were confirmed by PCR and sequencing. A stable F1 generation mouse model was obtained by mating Positive F0 generation mice with C57BL/6JGpt mice. F2rl1-flox mice were crossbred with Vil1-Cre mice, in order to obtain F2rl1fl/fl
Vil1-icre± (termed f2rl1[IEC]) mice that genetically ablated F2rl1 in intestinal epithelium cells, and the littermate F2rl1fl/-
Vil1-icre± (termed f2rl1f/f). Genotyping was performed using the primers: forward: 5´-GCCTGACTGGATAACCTGGATAG-3´; Reverse: 5´-CGTCTGGAAACACCTGAACATTG-3´.

Human samples
Human colon tissues were obtained from patients in Peking Union Hospital, Chinese Academy of Medical Sciences, Beijing, China. For IHC staining, all tissues were fixed in 10% neutral buffered formalin. This study was approved by the Institutional Review Board of the Chinese Academy of Medical Sciences Cancer Institute (I-24PJ2549). The detailed information on patients is shown in Tables 2–3.

Bioinformatics analysis
For immune cell infiltration analysis, we downloaded the GEO dataset and utilized the “CIBERSORT” package in R (The R Foundation for Statistical Computing, Vienna, Austria) to assess immune cell infiltration using a gene expression matrix; 22 immune cell genes (LM22) were used as the reference set with 1000 permutations [29].
For the analysis of neutrophil infiltration and FOXA2 levels in TCGA database, COAD and READ data were downloaded, and the “CIBERSORT” R package was used to assess immune cell infiltration using the gene expression matrix; 22 immune cell genes (LM22) were used as the reference set with 1000 permutations. The samples were grouped based on the level of neutrophil infiltration, and FOXA2 expression was compared between the high-expression and low-expression 1/4 groups using the Wilcoxon test. Statistical significance was set at p < 0.05. Bar charts were created using the “ggplot2” R package.

Multiple immunofluorescence analysis
Machine learning was employed in sample analysis to recognize the tumor (red), stromal (green), and blank (yellow) areas. The thresholds for cell nuclei and positive cells in each channel were determined, and positive cells were identified within each channel.
Proximity analysis involves examining the in situ data of specific biomarkers, establishing the analysis range, and determining the step size values. To analyze the average fluorescence intensity, the proximity of PANCK+ to MPO+ cells was examined using MPO+ cells as the focal group. The fluorescence intensity of FOXA2 in PANCK+ cells in the raw data was aligned based on the proximity findings, and the average fluorescence intensity of FOXA2 was calculated.

RNA interference
Cells were seeded at 5 × 105 cells per well in 6-well plates prior to the transfection. During the transfection, the cells were transfected with small interfering RNA targeting ATP6V0E1 and FOXA2 purchased from Suzhou GenePharma Co., Ltd. The siRNA sequences used in the study are shown in Table 4.

Reverse transcription-quantitative PCR
Total RNA was isolated with Trizol (Invitrogen 15596-026) according to the manufacturer’s instructions. One microgram of RNA was used to synthesize cDNA with Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Fisher Scientific, K1622) according to the manufacturer’s protocol. The cDNA products were subjected to 40 cycles of PCR. Quantitative PCR (qPCR) was performed by SYBR premix Ex Taq II (TaKaRa Bio Inc., RR820A). The primers used in the study are shown in Table 5. PCR analysis was performed using ABI Step One Plus RealTime PCR detection system, in triplicate for each sample and each gene. The relative levels of mRNA transcripts were normalized to the expression of RNA18S or ACTB/β-actin.

Isolation of cytoplasmic and nuclear fractions
Cytoplasmic and nuclear fractions were isolated following the manufacturer’s instructions (Thermo Fisher Scientific 78833).

Western blot analysis
The total protein of mouse colon tissues or cell lines was extracted using the radioimmunoprecipitation assay lysis buffer with protease inhibitor cocktail and protein phosphatase inhibitor, and total protein concentration was quantified by the BCA (bicinchoninic acid) method (Applygen Technologies Inc., P1511). Equal amounts of protein lysates were resolved by SDS-PAGE and transferred to polyvinylidene difluoride membranes (EMD Millipore Corporation, IPFL00010). The membrane was blocked with 5% skimmed milk in TBST (Sangon Biotech, C520009) for 2 h at room temperature, then incubated with primary antibodies at 4°C overnight and secondary antibodies at room temperature for 1 h. Finally, the subsequent signals were visualized using ECL (enhanced chemiluminescence; Applygen Technologies Inc., P1050) on an Amersham Imager 600 (GE Healthcare).

Electron microscopy
We examined autophagy with electron microscopy in colon tumor tissues and HT29 cells, in which autophagosomes and related autophagic vacuoles were monitored. Mice colon tumor tissues (1 mm3) were fixed with 2.5% glutaraldehyde, and 1% osmic acid was then added for fixation. After dehydration by a graded series of ethanol, the colon tumor tissues and HT29 cells were infiltrated with epoxy resin (Sigma-Aldrich 31190). Then, ultrathin sections were obtained and stained with uranyl acetate and lead citrate to be visualized with a transmission electron microscope. HT29 cells were seeded and treated on glass coverslips. After treatment cells were washed and fixed overnight at 4°C in a solution containing 2% paraformaldehyde (EM grade; Sigma-Aldrich, P6148) and 2.5% glutaraldehyde (EM grade; Sigma-Aldrich, G5882). The cell samples were rinsed with 0.1 M phosphate buffer (pH 7.4) three times. Then fixed with 1% osmium acid (in 0.1 M phosphate buffer) at room temperature for 1 h avoiding light and rinsed with purified water three times. After fixation, samples were successively dehydrated with 30%, 50%, 70%, and 95% alcohol, each for 10 min, and then three times with 100% alcohol with isoamyl acetate for 10 min. The samples were subjected to critical point drying. The dried sample was fixed on the sample stand using double-sided carbon tape, and gold was sprayed under vacuum conditions. The autophagosomes were observed by TEM.

Transcriptome sequencing and data analysis
Total RNA was isolated with Trizol (Invitrogen,15596026) according to the manufacturer’s instructions. A total amount of 1 μg RNA per sample was used as input material for the RNA sample preparations. Sequencing libraries were generated using Hieff NGS Ultima Dual-mode mRNA Library Prep Kit for Illumina (Yeasen Biotechnology,13335ES) following the manufacturer’s recommendations, and index codes were added to attribute sequences to each sample. The raw reads were further processed with a bioinformatic pipelinetool, BMKCloud (www.biocloud.net) online platform.

ACP2/acid phosphatase detection
The lysosomal activity was quantified by determining the activity of acid phosphatase, a lysosomal key enzyme. 1 × 106 cells were lysed in citrate buffer and added to the substrate provided by the Acid Phosphatase Assay Kit (Solarbio® life sciences, BC2135). After incubation for 30 min at 37°C, the absorbance of the p-nitrophenol was determined at 510 nm in a microplate reader (Multiskan, FC).

ChIP-seq analysis
The promoter analysis of the ATP6V0E1 region binding to FOXA2 was conducted using the CISTROME database (http://cistrome.org/db/#/).

Immunohistochemistry (IHC)
Colon sections (5 μm) were dewaxed, rehydrated, and quenched. Antigen retrieval was then performed using citrate buffer. After serum blocking with goat serum (Sigma-Aldrich, G9023) at 37°C for 1 h, slides were incubated with the indicated primary antibodies at 4°C overnight and MaxVision HRP-Polymer anti-Mouse/Rabbit IHC Kit (ZSGB-BIO Corporation, PV-9000). Finally, slides were stained using diaminobenzidine (ZSGB-BIO Corporation, ZLI-9019) and counterstained with Mosaic™ V2.1 (Tucsen Photonics Co., Ltd).

Flow cytometry analysis
The cells were incubated with 0.5 μM LysoSensor Green DND-189 for lysosome pH or 5 μM DCFH-DA for ROS at 37°C for 30 min and then Cell Dissociation Buffer Enzyme-Free PBS-based digestion, washed and resuspended in phosphate-buffered saline to be analyzed by flow cytometry (LSRII, Becton Dickinson).

Disease activity score
For the disease activity score on the 5th week after AOM/DSS exposure, the extent was scored based on the criteria as follows: edematous mucosa and erythema under colonoscopy (1= normal; 2= mild: local or multiple erythema, increased brittleness, no epithelial damage; 3 = moderate: epithelial erosion or superficial small ulcers; 4 = Severe: large ulcer or multiple ulcers, intestinal stenosis, fistula, massive bleeding complications). Each item had a maximum of four points for a total of 12 points (ulcer or multiple ulcers; intestinal stenosis; massive bleeding).
For the disease activity score on the 1st week after AOM/DSS exposure, the extent was scored based on the criteria as follows: body weight loss (1 = 0-5% weight loss; 2 = 5-10% weight loss; 3 = 10-15% weight loss; 4 = more than 15% weight loss); and diarrhea and fecal occult blood (1 = no bleeding; 2 = slight bleeding; 3 = moderate bleeding; 4 = severe bleeding). Each item had a maximum of four points for a total of 12 points (body weight loss; diarrhea; fecal occult blood).

Statistical analysis
GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA) was used to generate graphs and conduct statistical analyses. Statistical significance between groups was assessed using an unpaired Student’s t-test. Comparisons of multiple conditions were performed using a one-way or two-way ANOVA with Tukey’s post-hoc test. Results are presented as mean ± standard deviation. Statistical significance was set at p < 0.05.

Supplementary Material

Supplemental figures R4.docx