The Mechanosensitive Ion Channel PIEZO1 in Intestinal Epithelial Cells Mediates Inflammation through the NOD-Like Receptor 3 Pathway in Crohn’s Disease

Abstract

Background

Crohn’s disease (CD) is an incurable chronic intestinal inflammatory disease with no recognized cause. It has been reported that the mechanosensitive ion channel PIEZO1 initiates proinflammatory responses. However, little is known about the role of PIEZO1 in CD.

Methods

Ileum biopsies were obtained from 30 patients with CD and 15 healthy volunteers. Clinical data were collected to determine the relationship between CD and PIEZO1. First, HT29 cells were incubated with Yoda1 and GsMTx4 (Grammostola spatulata mechanotoxin 4) to activate and inhibit PIEZO1, respectively. Second, PIEZO1 knockdown was performed using small interfering RNA. Third, calcium imaging, flow cytometry, and immunofluorescence were used to detect intracellular calcium and mitochondrial function. Last, real-time quantitative polymerase chain reaction, immunoblotting, and enzyme-linked immunosorbent assay were used to quantify PIEZO1, proinflammatory cytokines, and NLRP3 (NOD-like receptor 3)–related compounds.

Results

PIEZO1 was highly expressed in the ileum of patients with CD and correlated positively with the Crohn’s Disease Activity Index, platelet count, and hematocrit and fecal calprotectin levels. In HT29 cells, Yoda1 triggered calcium influx, which was inhibited by GsMTx4 treatment and small interfering RNA–mediated PIEZO1 knockdown. Increased calcium concentrations resulted in increased reactive oxygen species accumulation and decreased mitochondrial membrane potential, whereas decreased calcium concentrations caused by GsMTx4 and PIEZO1 knockdown had the opposite effect. Mechanistically, molecules in the NLRP3 pathway were activated in patients with CD and HT29 cells were stimulated by lipopolysaccharide; these effects were reversed by the knockdown of PIEZO1. Finally, PIEZO1 and NLRP3 knockdown decreased proinflammatory cytokine levels in HT29 cells.

Conclusions

PIEZO1 in intestinal epithelial cells caused calcium influx, which resulted in mitochondrial dysfunction and activated the NLRP3 inflammasome, mediating intestinal inflammation.

Introduction

Crohn’s disease (CD) is a chronic intestinal disease that is incurable. The most common symptoms include diarrhea, abdominal pain, and weight loss, with stenosis, intestinal obstruction, abdominal abscess, and extraintestinal manifestations as complications.¹ Therefore, persistent intestinal inflammation is a significant risk factor for disease recurrence and complications. According to a retrospective analysis of patients with CD who had disease recurrence after ileocolic resection, 93% of CD specimens from the anastomosis and neoterminal ileum had CD-like features,² implying that chronic inflammation is the most common source of postoperative recurrence. Meanwhile, lymphocytes, plasma cells, eosinophils, and neutrophils in the intestinal mucosa, submucosa, and lamina propria produce an excessive number of proinflammatory cytokines, such as tumor necrosis factor α (TNF-α), interleukin (IL)-6, IL-17, and transforming growth factor-β1. This causes intestinal strictures and dysfunction due to extracellular matrix accumulation and fibrosis.³ Therefore, it is critical to manage gut inflammation in patients with CD. However, no curative therapy is accessible in clinical practice, and the inflammatory process remains unexplained.

During movement, secretion, and absorption, the digestive tract is permanently subjected to mechanical forces.⁴ Studies have reported that mechanosensors are critical for normal gastrointestinal motility, secretion, and digestion.^4,5 The opening of mechanosensitive ion channels is the first event when mechanical forces stimulate the intestine.⁶ PIEZO proteins are nonselective, converted mechanosensitive cation channels that were identified in 2010,⁷ but their structures were not reported until 2018.⁸ PIEZO1 and PIEZO2 are 2 members of the PIEZO family; PIEZO2 is expressed predominantly in Merkel cells, muscle spindles, and neurons, and it acts as a sensor of light touch and pain.⁹ In contrast, PIEZO1 is ubiquitously expressed in nonsensory tissues and cells, including endothelial cells, erythrocytes, epithelial cells, adipocytes, and pancreatic acinar cells.⁹ Recent studies have revealed that PIEZO1 plays a vital role in various tissues and diseases.^10-14 However, little is known about the function and contribution of PIEZO1 to CD-associated intestinal inflammation.

Intestinal epithelial cells (IECs) act as the first line of defense in responding to the pressure and contents of the gut lumen, including microbiota, metabolites, and other stimuli, and modulating host immune responses.¹⁵ They show various pattern recognition receptors capable of recognizing pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs), including NOD-like receptors, Toll-like receptors, cytosolic DNA sensors, and pyrin.¹⁶ When these receptors are activated, multiprotein complexes, specifically the inflammasome, are formed. Pyroptosis mediated by NOD-like receptor 3 (NLRP3) has been linked to the development of inflammatory bowel disease and experimental colitis.^17,18 Unlike other inflammasomes, NLRP3 inflammasomes can be activated not only by exogenous PAMPs, but also by various endogenous DAMPs, including sodium urate, extracellular ATP, dietary saturated fatty acids, hyaluronic acid, reactive oxygen species (ROS), and Ca²⁺.¹⁹

This study hypothesized that PIEZO1 increases intracellular calcium concentration and mitochondrial dysfunction and, eventually, promotes intestinal inflammation by NLRP3 activation. We discovered that in the ileum of patients with active CD, PIEZO1 and key molecules of the NLRP3 pathway were highly expressed and associated with inflammatory biomarkers. In HT29 cells, the level of proinflammatory cytokines is reduced when PIEZO1 is silenced. PIEZO1 in intestinal epithelial cells caused calcium influx, which resulted in mitochondrial dysfunction and likely activated the NLRP3 inflammasome. NLRP3 silencing reduced proinflammatory cytokines; thus, we hypothesized that PIEZO1 acts as a proinflammatory factor in CD based on our findings.

Methods

Patient Selection

The Institutional Review Board of Anhui Medical University’s First Affiliated Hospital (PJ-2019-15-21) approved this study, and all patients and healthy control subjects provided written informed consent. From June to October 2021, at the Department of Gastroenterology of First Affiliated Hospital of Anhui Medical University, we enrolled patients with CD having active clinical manifestations or lesions under double-balloon enteroscopy (DBE). A thorough examination of clinical symptoms and standard postoperative histological, radiographic, and endoscopic signs led to CD diagnosis. The following exclusion criteria were used: (1) patients with an unconfirmed diagnosis or disease duration inconsistent with CD; (2) patients who received DBE without biopsy; (3) patients with obesity, cardiovascular disease, hematological diseases, tumors, acute infectious diseases, or other autoimmune diseases; and (4) patients with insufficient clinical data. Meanwhile, 15 small bowel deep biopsies were taken from healthy volunteers who underwent endoscopy as part of their physical examination and used as healthy control samples.

Sample Collection

Ileum mucosal tissues were collected from 15 healthy donors and 30 patients with inflammatory lesions. The samples were stored at -80°C or in liquid nitrogen. Additionally, information on sex, age, blood tests (red blood cells, white blood cell count, lymphocyte, neutrophil, hemoglobin, hematocrit, platelets, inflammatory markers [C-reactive protein], fecal calprotectin [FC], and erythrocyte sedimentation rate), Crohn’s Disease Activity Index (CDAI), disease location, disease behavior, and treatments were collected.

Cell Culture and Treatment

The National Collection of Authenticated Cell Cultures, Chinese Academy of Sciences provided HT29 cells. First, cells were cultured in RPMI 1640 Medium (Biological Industries, Kibbutz Beit Haemek, Israel) with 10% fetal bovine serum (WISENT Corporation, Nanjing, China) and 1% penicillin-streptomycin solution (Biological Industries) at 37°C in a humidified atmosphere containing 5% CO₂.

Second, to investigate the role of PIEZO1, we used small interfering RNAs (siRNAs) to inhibit its expression. RiboBio (Guangzhou, China) designed and synthesized siRNA and negative control samples. Following the manufacturer’s protocols, Lipofectamine 3000 reagent (Thermo Fisher Scientific, Waltham, MA, USA) was used to transfect siRNA into HT29 and LPS (Sigma Aldrich, St Louis, MO, USA) was used to treat the cells for 24 hours.

Last, to control PIEZO1 activation, HT29 cells were treated with 5 μM Yoda1 (R&D Systems, Minneapolis, USA) and 5 μM GsMTx4 (Grammostola spatulata mechanotoxin 4) (Abcam, Boston, MA, USA) to activate and inhibit PIEZO1, respectively.

Real-Time Quantitative Polymerase Chain Reaction

The TRIzol reagent was used to extract total RNA from intestinal tissues and cells (Invitrogen, Carlsbad, CA, USA). The RNA was validated using a Nanodrop Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) before being reverse transcribed to complementary DNA with HiScript III RT SuperMix for qPCR (Vazyme, Nanjing, China). Finally, real-time quantitative polymerase chain (qPCR) reaction was executed using the ChamQ SYBR qPCR Master Mix (Vazyme) in an Applied Biosystems Step One Real-time System (Applied Biosystems, San Diego, CA, USA). The expression of genes was normalized to GAPDH and the results were expressed as 2^-ΔΔCt. Table 1 lists the primer oligonucleotide sequences used in this study.

Western Blotting

Tissues and cells were lysed with radioimmunoprecipitation assay buffer (Beyotime, Shanghai, China) and centrifuged for 10 minutes at 12 000 g. The loading buffer was added after the supernatant was collected. Protein samples were separated in sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels and transferred onto polyvinylidene fluoride membranes using transfer buffer after boiling at 100°C for 10 minutes. After blocking with 5% nonfat milk, the membrane was incubated overnight at 4°C with anti-PIEZO1 (1:1000), anti-NLRP3 (1:500), anti-β-Actin (1:50000; Proteintech, Wuhan, China), anti-cleaved N-terminal GSDMD (1:1000; Abcam, Cambridge, United Kingdom), and anti-ASC (1:200; Santa Cruz Biotechnology, Dallas, TX, USA). After 30 minutes of washing, membranes were incubated for 1 hour at room temperature with secondary horseradish peroxidase–conjugated antibodies at 1:5000 dilutions. The photos were quantified using the ImageJ software (1.52v) (National Institutes of Health, Bethesda, MD, USA).

Calcium Imaging

Pretreatment of HT29 cells with si-RNA and 2.5 μM GsMTx4 was followed by loading with 5 μM Fluo-4A (Beyotime) in the dark for 30 minutes at 37°C. After 3 washes with Hank’s Balanced Salt Solution, cells were observed using an Olympus inverted fluorescence microscope. Then, 20 μM Yoda1 was added after the baseline fluorescence value remained stable.

ROS Measurement

Following manufacturer’s protocols for the Reactive Oxygen Species Assay Kit (Beyotime), the intracellular concentrations of ROS were determined. For 30 minutes, HT29 cells were treated with 5 μM Yoda1, 5 μM GsMTx4, and Rosup (assay kit positive control). Next, with DCFH-DA, cells were incubated in the dark for 20 minutes at 37°C, followed by removing the culture medium. The cells were washed thrice in a serum-free culture medium. Finally, CytoFlex (Beckman Coulter, San Diego, CA, USA) was used to determine ROS concentrations.

Mitochondrial Membrane Potential Assay

Mitochondrial membrane potential was determined using a JC-1 mitochondrial membrane potential assay kit (Beyotime). HT29 cells were treated for 30 minutes with 5 μM Yoda1, 5 μM GsMTx4, and carbonyl cyanide-m-chlorophenylhydrazone (positive control in assay kit). Cells were incubated with JC-1 in the dark for 20 minutes at 37°C, followed by 3 washes with phosphate-buffered saline. After 3 washes with JC-1 buffer, the cells were observed using the LSM880 laser confocal microscope (Carl Zeiss, Oberkochen, Germany) and CytoFlex (Beckman Coulter).

Enzyme-Linked Immunosorbent Assay

Human IL-1β, IL-18, TNF-α, and IL-6 were quantified using an enzyme-linked immunosorbent assay (ELISA) kit (Dakewe, Shenzhen, China) according to the manufacturer’s instructions.

Statistical Analysis

SPSS v.19.0 software (IBM Corporation, Armonk, NY, USA)was used for statistical analysis. Statistical graphs were created using GraphPad Prism v.8 (GraphPad Software, San Diego, CA, USA). The mean ± SEM was used for normally distributed continuous variables, whereas median (interquartile range) was used for non-normally distributed data. Two-tailed Student’s t test was used to compare 2 groups of normally distributed independent samples. Pearson’s linear correlation was used for samples with a normal distribution, whereas Spearman’s rank correlation was used for non-normal distributions. The chi-square test was used to compare qualitative data from 2 samples, while a 1-way analysis of variance was used to compare multiple groups. P < .05 was defined as statistically significant.

Results

Demographics and Clinical Characteristics

This study enrolled 15 healthy volunteers and 30 patients with active CD. Each patient underwent DBE, whereas healthy volunteers underwent colonoscopy. Demographic information and disease characteristics are presented in Table 2. The median age of patients with CD was 26.5 years, and 20 of the patients were male. Infliximab, adalimumab, vedolizumab, and ustekinumab were administered to 21 patients, including 16 who received infliximab, 1 who received adalimumab, 1 who received vedolizumab, and 3 who received ustekinumab. Three patients were treated with corticosteroids, whereas 1 was treated with 5-aminosalicylic acid. Three patients underwent surgery, whereas 2 received enteral nutrition therapy. Platelet count was significantly higher (Z = -3.251, P = .001) and hematocrit levels were significantly lower (Z = -2.372, P = .018) in patients with CD. There were no significant differences in sex, age, white blood cell count, neutrophil count, leukocyte count, or hemoglobin levels between patients with active CD and healthy control subjects.

PIEZO1 Expression was Higher in the Ileum of Patients with Active CD Than in That of Healthy Volunteers

To examine PIEZO1 expression in CD, we collected ileum biopsies from patients with active CD (n = 30) and healthy volunteers (n = 15). qPCR analysis of messenger RNA (mRNA) levels revealed increased PIEZO1 expression in active CD intestinal tissues (Figure 1A) and protein levels (Figures 1B, C). Additionally, PIEZO1 was significantly higher in patients with mild and moderate CD according to CDAI than in control subjects (Figure 1D). Furthermore, we investigated the link between PIEZO1 mRNA levels and clinical inflammatory indicators. Interestingly, CDAI, platelets, hematocrit, and FC, which are particular biomarkers of intestinal inflammation, were all positively linked with PIEZO1²⁰ but not with C-reactive protein, erythrocyte sedimentation rate, neutrophil-to-lymphocyte ratio, or platelet-to-lymphocyte ratio (Figures 1E-L). These data indicate a positive correlation between PIEZO1 and intestinal inflammation, implying that PIEZO1 may play a role in promoting inflammation.

PIEZO1 Increased the Secretion of Proinflammatory Factors In Vitro

To determine the role of PIEZO1 in intestinal inflammation, we used siRNA to knock down PIEZO1 expression in HT29 cells. First, we verified the knockdown efficiency using qPCR and immunoblotting (Figure 2A, B), and si-PIEZO1 002 was used in subsequent investigations. Knocking down PIEZO1 reduced the levels of proinflammatory factors, such as IL-6 and TNF-α, after 24 hours of stimulation with 10 µg/mL LPS as measured by qPCR (Figure 2C, D) and ELISA (Figure 2E, F). Then, we activated PIEZO1 using Yoda1, a PIEZO1 agonist. Following LPS stimulation, IL-6 and TNF-α levels were significantly increased (Figure 2C-F). These findings indicate that PIEZO1 promotes inflammation in vitro.

PIEZO1 Increased Intracellular Calcium Concentration

First, we sought to elucidate how PIEZO1 increases intestinal inflammation. PIEZO1 was previously identified as a subunit of the Ca²⁺-permeable nonselective cation channel.²¹ Then, we examined whether PIEZO1 controls intracellular calcium levels in epithelial cells. Calcium imaging experiments were performed in vitro in HT29 cells. After the cells were treated with Yoda1, a considerable increase in calcium influx was noted (Figure 3A). Pretreatment with GsMTx4, a mechanosensitive channel inhibitor, prevented this phenomenon (Figure 3B). As anticipated, silencing PIEZO1 expression suppressed the calcium influx induced by Yoda1 (Figure 3C). Calcium imaging revealed that PIEZO1 could function as a calcium channel, increasing intracellular calcium concentrations (Figure 3D).

Ca²⁺ Promoted Mitochondrial Injury in HT29 Cells In Vitro

We sought to determine how intracellular calcium concentration induces inflammation. Ca²⁺ plays a crucial role in regulating mitochondrial transporters, enzymes, and respiratory complexes.²² Therefore, we hypothesized that intracellular calcium exacerbates inflammation by inducing mitochondrial dysfunction. Therefore, we first determined ROS levels using flow cytometry and discovered that they were significantly increased following Yoda1 stimulation, whereas GsMTx4-treated cells had ROS levels similar to those of the control group (Figure 4A-D). The mitochondrial membrane potential was another sign of mitochondrial injury. We observed an increase in JC-1 monomers and a decrease in JC-1 aggregates after incubation with Yoda1 using a laser scanning confocal microscope (Figure 4E). Flow cytometry revealed the same trend (Figure 4F, G). These findings suggest that activating PIEZO1 releases ROS and decreases mitochondrial membrane potential.

PIEZO1 Activated the NLRP3 Pathway In Vitro

First, we examined how mitochondrial injury contributes to inflammation. A previous study had shown that ROS generation could activate the NLRP3 inflammasome²³ and therefore aid inflammation. Then, we examined whether the NLRP3 pathway was activated in epithelial cells after mitochondrial injury. We stimulated HT29 cells with LPS for 24 hours, and when Yoda1 activated PIEZO1, the mRNA and protein levels of NLRP3, IL-1β, and IL-18 were significantly increased (Figure 5A-E). However, this effect was reversed by inhibiting PIEZO1 expression. Furthermore, qRT-PCR, ELISA, and immunoblotting revealed a decrease in the expression of NLRP3, IL-1β, and IL-18 in siRNA-treated cells compared with the LPS group. Thus, no cleavage of GSDMD or caspase-1 was observed in Western blotting (Figure 5A-F). These findings indicate that PIEZO1 activates the NLRP3 pathway in HT29 cells in vitro.

The NLRP3 Pathway Was Activated in the Ileum Tissue of Patients With CD and Positively Correlated With PIEZO1 Expression

To verify this phenomenon in patients with active CD, we used qPCR to determine the expression of key molecules in the NLRP3 pathway and PIEZO1. We found an increase in the mRNA levels of NLRP3, ASC (apoptosis-associated speck-like protein containing CARD), caspase-1, IL-1β, and IL-18 (Figure 6A-F). Correlation analysis revealed a positive correlation between PIEZO1 and NLRP3, IL-1β, IL-18, and TNF-α (Figure 6G-J). IL-1β and IL-18 were also increased in intestinal tissues as determined by ELISA (Figure 6K, L). As anticipated, NLRP3 and ASC levels were significantly increased in patients with active CD, and Western blotting revealed the cleavage of GSDMD and caspase-1 (Figure 6M, N). According to the clinical data, the NLRP3 pathway was activated in patients with active CD, and PIEZO1 expression was positively correlated with NLRP3 pathway activation.

PIEZO1-Promoted Proinflammatory Cytokines Depend on NLRP3 Expression

We used siRNA to inhibit NLRP3 expression in HT29 cells. qRT-PCR analysis revealed that IL-1β, IL-18, IL-6, and TNF-α levels were decreased 24 hours after LPS stimulation regardless of whether PIEZO1 was activated (Figure 7A-D). Furthermore, ELISA revealed similar trends (Figure 7E-H). As a result, it was concluded that PIEZO1-induced cytokines depend on NLRP3 expression.

Discussion

CD is one of the most common inflammatory bowel diseases, and clinical practice has no cure. Therefore, elucidating the pathogenesis and treatment targets is crucial. Our findings indicated that PIEZO1 was highly expressed in the ileum of patients with active CD and correlated with intestinal inflammation. In vitro, PIEZO1 aided the release of proinflammatory cytokines from epithelial cells. Mechanistically, PIEZO1 regulated calcium influx in HT29 cells, resulting in mitochondrial dysfunction manifested by accumulating mitochondrial ROS and a decrease in membrane potential, which may activate the NLRP3 inflammasome and promote intestinal inflammation.

The study establishes a correlation between intestinal PIEZO1 and CDAI, platelets, hematocrit, and FC, all of which are biomarkers of intestinal inflammation in patients with CD, and intestinal PIEZO1 induces the release of proinflammatory cytokines in vitro. These findings support those of other studies. Aykut B demonstrated that silencing Piezo1 in myeloid cells inhibits inflammation in a sepsis mouse model.²⁴ In atherosclerotic mice, inhibiting Piezo1 with GsMTx-4 reduces the formation of plaques and expression of inflammatory factors.²⁵ Solis et al¹⁴ observed a reduction in lung inflammation in Piezo1^f/fLysmCre mice after infection with bacteria. Following these findings, we used siRNA to knock down PIEZO1 in epithelial cells. We observed a decrease in TNF-α, IL-6, IL-1β, and IL-18 levels following LPS stimulation, which was validated in human intestinal tissues. Thus, our findings suggest that PIEZO1 plays a proinflammatory role in CD.

The findings showed that PIEZO1 regulates calcium influx and increases calcium concentrations in epithelial cells, which is consistent with that of earlier research. Furthermore, Piezo1 mediates calcium activity in endothelial tip cells. Inhibiting Piezo1 with GsMTx4 could significantly suppress calcium activity.²⁶ Additionally, Piezo1 deficiency inhibits intracellular calcium elevation.²⁷ Yoda1 failed to increase the calcium concentration transiently in Piezo1-deficient mice.²⁸ Recently, studies have demonstrated the subcellular localization of Piezo1, implying that Piezo1 mediates extracellular Ca²⁺ influx in the plasma membrane.²⁹ Our results also established that PIEZO1 might function as a calcium influx regulator in intestinal epithelial cells.

Furthermore, our findings indicate that increased intracellular calcium causes mitochondrial dysfunction. Calcium influx is always associated with the accumulation of mitochondrial ROS in various diseases or disease models.^30-32 For example, calcium overload induces oxidative stress and mitochondrial dysfunction in human SW982 cell lines, resulting in cell death.³³ Moreover, calcium influx and increase in intracellular calcium levels stimulate ROS production and depolarization of the mitochondrial membrane potential.³⁴ According to previous studies, we used an agonist to induce calcium influx in HT29 cells. Compared with cells with a low calcium level (GsMTx4), those with a high calcium level (Yoda1) produced more ROS and had a lower mitochondrial membrane potential; this phenomenon could be replicated by PIEZO1 knockdown. Thus, our findings indicate that mitochondrial dysfunction is regulated by intracellular calcium.

Additionally, it has been reported that NLRP3 inflammasomes detect mitochondrial injury,²³ providing the theoretical foundation for our study. Sun et al³⁵ reported that Piezo1 activated the NLRP3 inflammasome in nucleus pulposus cells to increase IL-1β. In this study, we found similar results, indicating that PIEZO1 in intestinal epithelial cells may promote inflammation in the presence of NLRP3.

Moreover, classical NLRP3 inflammasome activation includes 2 signals. In the priming signal, NLRP3 and pro-IL-1β expressions are upregulated,³⁶ whereas ASC pro-caspase-1 is recruited to form an inflammasome complex with NLRP3 to induce active caspase-1 and maturation of IL-1β during the second signal.³⁶ Furthermore, NLRP3 pathway activation has been reported in several chronic inflammatory diseases, including nonalcoholic fatty liver disease, diabetes, obesity, and Parkinson’s disease.^37,38 Also, the mRNA and protein levels of ASC, caspase-1, and GSDMD were higher in patients with inflammatory bowel disease and in experimental colitis models.^17,39 Consistent with these studies, we observed significant increases in NLRP3, IL-1β, and IL-18 in vitro and in patients with CD. Furthermore, immunoblotting revealed that PIEZO1 knockdown suppressed cleaved GSDMD and cleaved caspase-1.

Our findings reveal that PIEZO1 is highly expressed in the intestines of patients with active CD and is associated with intestinal inflammation. Additionally, PIEZO1 regulates calcium, causes mitochondrial injury, and may activate the NLRP3 inflammasome, aggravating intestinal inflammation. These findings suggest a likely target for treatment.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflicts of Interest

The authors declare that they have no competing interests.

References