mRNA expression of NLRP3 and IL1B in inflammatory bowel diseases: a systematic review and meta-analysis

Abstract

Background

Inflammatory bowel diseases (IBD) are chronic inflammatory conditions of the gastrointestinal tract. Clinical studies of IBD robustly show elevated interleukin 1β (IL-1β) levels in intestinal tissue, implicating the NLR family pyrin domain containing 3 (NLRP3) inflammasome, which controls IL-1β secretion in myeloid cells. In this study, we aimed to ground potential NLRP3 involvement in IBD in molecular evidence from human studies.

Methods

This systematic review and meta-analysis investigated NLRP3 and IL1B mRNA expression in biopsies of intestinal tissue and cells isolated from unstimulated blood of IBD patients compared to healthy controls (HCs), using standardized mean differences (SMD) [95% confidence interval (CI)], p-values and heterogeneity (I²), and a meta-correlation of the meta-analyses.

Results

A systematic search in Medline and Embase identified 2218 records, with 30 studies describing 33 separate datasets meeting inclusion criteria. Meta-analysis revealed consistently higher NLRP3 and IL1B mRNA expression in intestinal tissue of IBD patients (24 datasets, n = 3912; NLRP3: SMD = 0.53 [0.36, 0.70], p < 0.001, I²=66%; IL1B: SMD = 0.95 [0.66, 1.24], p < 0.001, I²=89%). In blood, IL1B expression was significantly elevated (9 datasets, n = 1921; SMD = 0.20 [0.06, 0.33], p = 0.003, I²=15%), while NLRP3 expression showed no significant increase (SMD = 0.12 [-0.16, 0.40], p = 0.40, I²=78%). SMDs of NLRP3 and IL1B in tissue strongly and positively correlated with each other (rho = 0.854 [0.687, 0.935], p < 0.001).

Conclusion

We found upregulation of both NLRP3 and IL1B mRNA in intestinal tissue of IBD patients, which highly correlated across 24 studies, reinforcing the role of the NLRP3 inflammasome in IBD and its potential as a therapeutic target.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-025-04566-8.

Introduction

Inflammatory bowel diseases (IBD), including ulcerative colitis (UC) and Crohn’s disease (CD), are chronic, inflammatory conditions of the gastrointestinal (GI) tract. Patients typically experience abdominal pain, (bloody) diarrhea, weight loss, and fatigue, necessitating lifelong treatment. IBD is thought to arise from an interplay of genetic predisposition, microbiome dysbiosis, and environmental factors [1]. The modern lifestyle, an environmental factor, includes ultra processed food diets high in animal protein, sugar, salt, fat, chemical additives like emulsifiers, and low in dietary fiber. These dietary components are associated with increased risk of IBD [2, 3]. IBD is still increasing in prevalence and one of the most costly disorders, both in health-care related costs and disease-related work productivity loss [4]. Understanding of the pathogenic mechanisms underlying IBD will inevitably aid the management of the debilitating disease.

Clinical studies have long reported high levels of the proinflammatory cytokine interleukin 1 beta (IL-1β) in the inflamed mucosa of IBD patients [5, 6]. IL-1β translation and secretion is tightly controlled by the NLR family pyrin domain containing 3 (NLRP3) inflammasome in several steps, indicating its potent inflammatory activity [7]. The NLRP3 inflammasome, a large cytosolic multiprotein complex broadly expressed in monocytes, macrophages, dendritic cells, neutrophils, and colonic lamina propria [8], requires a 2-step activation. First, a priming signal through the detection of danger-associated molecular patterns (DAMPs) and/or pathogen-associated molecular patterns (PAMPs) by toll like receptors (TLRs) and Nucleotide-Oligomerization Domain 2 (NOD2) induces NF-κB–mediated transcription of NLRP3 and IL1B mRNA [9]. Second, activation through stimuli such as ATP or microbial components results in inflammasome assembly and caspase 1 activation [9, 10]. Caspase 1 cleaves the inactive precursor protein of IL-1β into a biologically active, mature form. It similarly cleaves IL-18 and gasdermin D, whose N-terminal fragment forms membrane pores to allow IL-1β and IL-18 secretion and to induce pyroptotic cell death, amplifying inflammation [9, 10]. NLRP3 is a crucial complex in defense against microbial components, tissue damage, and a wide range of metabolic or stress-related danger signals [10]. It operates as a critical early responder to harmful stimuli and, under physiological conditions, is only active during acute stress [11]. However, in chronic inflammatory environments such as in IBD, continuous exposure may provoke sustained NLRP3 “priming” and partial activation. As the NLRP3 inflammasome is both primed and activated by environmental components [10], it follows that NLRP3 can be an important driver of inflammation in IBD.

Case-in-point, genome studies found NLRP3 gene polymorphisms associated with CD risk [12, 13]. A CARD8 loss-of-function mutation (V44I) causes increased IL-1β production and, correspondingly, severe CD that is resistant to anti-TNF therapy but responsive to IL-1 blockade [14]. Specific commensal bacteria, including Enterobacteriaceae and Proteus mirabilis, activate NLRP3-dependent IL-1β release from monocytes upon epithelial injury [15]. Furthermore, microbiome-derived short-chain fatty acids butyrate and propionate act as DAMPs that promote NLRP3 priming in macrophages [16].

Given the elevated levels of IL-1β in inflamed mucosa and the importance of NLRP3 as a sensor of danger, we conducted a meta-analysis to evaluate whether NLRP3 and IL1B mRNA expression levels differ in blood and intestinal tissues of IBD patients versus healthy controls (HCs). This analysis seeks to ground potential therapeutic strategies in the molecular evidence from human studies.

Methods

A systematic review and meta-analysis of publicly available mRNA expression data has been conducted in accordance with the PRISMA 2020 guidelines [17].

Eligibility criteria

Studies that met the following inclusion criteria were included: (1) randomized controlled trials and nonrandomized controlled trials that included mRNA expression measurements at baseline; (2) mRNA measured in cells isolated from unstimulated blood or in biopsies of intestinal tissue; (3) mRNA expression data is publicly accessible; (4) study participants are humans; and (5) IBD patients defined as case group and HCs defined as control group. Studies with small sample sizes (n < 30) or non-standard microarrays were excluded.

Information sources and search strategy

Medline [Ovid; 1946] and Embase [Ovid; 1947] were searched for eligible studies using predefined terms. The reference lists of eligible articles were also checked to identify additional studies. All searches were conducted from inception to October 24, 2022, and results were restricted to the English language.

MEDLINE(R) search strategy

1. exp Inflammatory Bowel Diseases/ or ((Crohn* adj disease) or (inflammatory-bowel adj2 disease*) or (regional adj enteritis) or ileitis or colitis or proctosigmoiditis or rectocolitis or rectosigmoiditis or (ulcerative adj proctocolitis) or (h?emorrhagic adj proctocolitis) or proctitis).ti, ab, kf.

2. exp Sequence Analysis, RNA/ or exp Gene Expression Profiling/ or exp Microarray Analysis/ or exp NLR Family, Pyrin Domain-Containing 3 Protein/ or exp Interleukin-1beta/ or (transcriptome or (whole-exon adj2 sequenc*) or (micro adj array) or microarray or exposome or (rna adj2 sequenc*) or nlrp3 or il1b).ti, ab, kf.

3. exp Leukocytes, Mononuclear/ or exp Blood/ or exp Serology/ or (blood or serolog* or intestinal or mucosal or PBMC* or serum or biopt*).ti, ab, kf.

4. 1 and 2 and 3.

5. (exp animals/ not humans/) or (mice or mouse or rat or rats or pig or pigs or dog or dogs or cat or cats).ti.

6. 4 not 5.

EMBASE search strategy

1. exp inflammatory bowel disease/ or ((Crohn* adj disease) or (inflammatory-bowel adj2 disease*) or (regional adj enteritis) or ileitis or colitis or proctosigmoiditis or rectocolitis or rectosigmoiditis or (ulcerative adj proctocolitis) or (h?emorrhagic adj proctocolitis) or proctitis).ti, ab, kf.

2. exp *RNA sequencing/ or exp *transcriptome/ or exp *’gene expression profiling’/ or exp *’rna microarray’/ or cryopyrin/ or exp *’interleukin 1 beta’/ or (transcriptome or (whole-exon adj2 sequenc*) or (micro adj array) or microarray or exposome or (rna adj2 sequenc*) or nlrp3 or il1b).ti, ab, kf.

3. exp peripheral blood mononuclear cell/ or exp blood analysis/ or exp serology/ or (blood or serolog* or intestinal or mucosal or PBMC* or serum or biopt*).ti, ab, kf.

4. 1 and 2 and 3.

5. ((rat or rats or mouse or mice or swine or porcine or murine or sheep or lambs or pigs or piglets or rabbit or rabbits or cat or cats or dog or dogs or cattle or bovine or monkey or monkeys or trout or marmoset*).ti, ot. and exp ‘animal experiment’/) or (exp ‘animal experiment’/ not (exp ‘human experiment’/ or exp ‘human’/)).

6. 4 not 5.

7. limit 6 to embase.

Selection strategy

After removal of duplicates, 2 authors independently screened records by title and abstract. In the second phase, the full text of all articles identified in the first phase were screened to confirm suitability according to the in- and exclusion criteria, again by both authors. Consensus was reached by discussion in case of disagreement. All screening steps were performed in Rayyan [18].

Data extraction and data items

Independently, 2 authors performed data extraction. Data items included 1. Gene Expression Omnibus (GEO) [19] series- and platform accession number; 2. first author, year of publication and PubMed identifier (PMID); 3. sample characteristics (sample size, age, disease status, medication status); 4. tissue information; and 5. raw NLRP3 and IL1B mRNA expression values. Published summary effect measures remained unused. GEO series accession (GSE) numbers were entered in GEO-2R to extract raw gene expression values.

Quality of evidence

The Newcastle-Ottawa Scale for case-control studies [20] or assessing the quality of nonrandomized studies in meta-analyses was used to judge for risk-of-bias independently by 2 authors. With the scale, studies are judged on 3 domains by awarding a maximum of 9 stars: selection of cases and controls (4 stars), comparability of cases and controls (a star for each controlling factor with a maximum of 2 stars), and outcome (3 stars). Studies are considered of good quality when awarded ≥ 3 stars in selection, and ≥ 1 stars in comparability, and ≥ 2 in outcome. Fair quality studies receive 2 stars in selection, and ≥ 1 in comparability, and ≥ 2 in outcome. Studies awarded less stars are considered of poor quality.

Effect measures

Check for normality of the data was performed in IBM SPSS Statistics version 28.0. Non-normally distributed data was analyzed according to the central limit theorem, which states that the distribution of a sample means approximates a normal distribution as the sample size increases [21]. The sample size for this to hold is n = 30, hence the corresponding exclusion criterion. The mean difference (MD) in expression between IBD and HC and the corresponding standard deviation (SD) was calculated with the independent samples t-test. MD, SD and sample size of each study were entered into Review Manager version 5.4 [22] and the standardized mean difference (SMD) as meta-effect measure was calculated to standardize the results of the studies to a uniform scale. SMDs were estimated in Review Manager by applying Cohen’s d (i.e. dividing MD by SD) and Hedges’ g, providing a bias correction [23]. The SMD with corresponding 95% confidence intervals (CIs) were presented as effect measure.

Statistical analyses

All meta-analyses were conducted using Review Manager. Separate analyses were performed for NLRP3 and IL1B mRNA expression, both in blood and in intestinal biopsies. A random-effects model was adopted based on anticipated clinical- and methodological heterogeneity and hence to correct for between-study variation [23]. Statistical heterogeneity was quantified with I² statistics. I² is the proportion of the observed variance that reflects real differences in effect size and is computed as An I² > 75% was considered indicative of considerable heterogeneity [23]. A bivariate random-effects meta-analysis model with an unstructured covariance matrix of SMDs and corresponding 95% CIs was conducting using the metafor package in R (version 4.3.2). The resulting meta-correlation described by rho was visualized in a scatterplot using R, with study-level SMDs and horizontal and vertical error bars representing 95% CIs.

Results

Study selection

The initial search resulted in 2218 records (Fig. 1). After removal of duplicates, 1610 records were screened of which 241 were read full text. We excluded 221 studies because they contained no publicly available data, employed a method considered exclusionary for this meta-analysis (RNA retrieved from cells, other tissues, animals or RNA read with a non-standard microarray), included no HCs, included fewer than 30 participants, concerned a review or reported on a non-unique dataset, in which case we only included 1 article per dataset to prevent duplicate participants in the final analysis. Finally, 30 studies were included in the analysis, describing 33 separate datasets (Table 1) with a total of 5833 participants (4711 IBD patients and 1122 HCs). IBD patients were adults or children, with CD or UC, with active or inactive disease, and on medication (including corticosteroids, 5-aminosalicylic acids, immunomodulators, biologic agents, and small molecules) or treatment naïve. mRNA expression was measured with microarrays or RNA sequencing, from biopsies taken from inflamed or uninflamed sections of the intestines (ileum, colon, rectum) or from cells isolated from untreated whole blood or freshly isolated peripheral blood mononuclear cells (PBMCs).

Fig. 1.

PRISMA flow chart of study selection

Table 1.

Characteristics of included databases

Database	Database described in		Description of subjects included							RNA derived from
GEO accession number	Author Year	PMID	HC	CD	UC	Total	Age	Disease status	Medication status	Tissue	Biopsy site and status / Type of blood	Platform
GSE193677	Argmann 2022	36109152	241	492	404	1137	Adults	Active and inactive	On medication	Biopsy	Inflamed and uninflamed ileum, colon, or rectum	RNA sequencing
GSE16879	Arijs 2009	19956723	12	37	24	73	Adults	Active	On medication	Biopsy	Inflamed ileum or colon	Microarray
GSE73661	Arijs 2018	27802155	12		67	79	Adults	Active	On medication	Biopsy	Inflamed colon	Microarray
GSE153974	Ashton 2021	33232439	17	57	22	96	Children	Active and inactive	Treatment naïve and on medication	Biopsy	Inflamed and uninflamed ileum	RNA sequencing
GSE9686	Carey 2008	18069684	8	20	5	33	Children	Active and inactive	Treatment naïve and on medication	Biopsy	Inflamed colon	Microarray
GSE128682	Fenton 2021	32322884	16		28	44	Adults	Active and inactive	Treatment naïve and on medication	Biopsy	Not specified: “mucosal biopsy”	RNA sequencing
GSE57945	Haberman 2014	25003194	42	218	62	322	Children	Active and inactive	Treatment naïve	Biopsy	Inflamed and uninflamed ileum	RNA sequencing
GSE109142	Haberman 2019	30604764	20		206	226	Children	Active	Treatment naïve	Biopsy	Inflamed rectum	RNA sequencing
GSE117993	Haberman 2019	30604764	55	92	43	190	Children	Active	Treatment naïve	Biopsy	Inflamed rectum	RNA sequencing
GSE10616	Kugathasan 2008	18758464	11	32	10	53	Children	Active	Treatment naïve and on medication	Biopsy	Inflamed colon	Microarray
GSE52746	Leal 2015	24700437	17	22		39	Adults	Active and inactive	On medication	Biopsy	Inflamed colon	Microarray
GSE87466	Li 2018	29401083	21		87	108	Adults	Active	On medication	Biopsy	Inflamed colon	Microarray
GSE36807	Montero-Meléndez 2013	24155895	7	13	15	35	Adults	Active and inactive	On medication	Biopsy	Uninflamed colon	Microarray
GSE126124	Palmer 2019	31618209	19	37	18	74	Children	Active	Treatment naïve	Biopsy	Inflamed and uninflamed colon or cecum	Microarray
GSE206285	Pavlidis 2022	36170836	18		550	568	Adults	Active	On medication	Biopsy	Not specified, sigmoid	Microarray
GSE83687	Peters 2017	28892060	60	42	32	134	Adults	Active	On medication	Biopsy	Inflamed ileum, colon, or rectum	RNA sequencing
GSE38713	Planell 2013	23135761	13		23	36	Adults	Active and inactive	On medication	Biopsy	Inflamed and uninflamed sigmoid or rectum	Microarray
GSE48634	Smith 2014	25171508	17	9	20	46	Adults	Inactive	On medication	Biopsy	Uninflamed descending colon	Microarray
GSE85499	Toyonaga 2020	32561494	11	21		32	Adults	Unknown	On medication	Biopsy	Uninflamed ascending colon	RNA sequencing
GSE112366	Van Dussen 2018	29782846	26	141		167	Adults	Active	On medication	Biopsy	Uninflamed ileum	Microarray
GSE75214	Vancamelbeke 2017	28885228	22	75	97	194	Adults	Active and inactive	On medication	Biopsy	Inflamed and uninflamed ileum or colon	Microarray
GSE59071	Vanhove 2015	26313692	11	8	97	116	Adults	Active and inactive	On medication	Biopsy	Inflamed sigmoid or rectum	Microarray
GSE102133	Verstockt 2019	30657881	12	65		77	Adults	Active	Treatment naïve and on medication	Biopsy	Inflamed ileum	Microarray
GSE83448	Zabana 2020	32508154	14	19		33	Adults	Active	On medication	Biopsy	Inflamed ileum	Microarray
GSE186507	Argmann 2022	36109152	209	432	389	1030	Adults	Active and inactive	On medication	Blood	Whole blood	RNA sequencing
GSE3365	Burczynski 2006	16436634	42	59	26	127	Adults	Active	On medication	Blood	PBMCs	Microarray
GSE169568	Juzenas 2022	35022690	30	58	52	140	Adults	Active and inactive	Treatment naïve	Blood	Whole blood	Microarray
GSE112057	Mo 2018	29950172	12	60	15	87	Children	Unknown	Not specified: “on diagnosis”	Blood	Whole blood	RNA sequencing
GSE119600	Ostrowski 2019	31076612	47	95	93	235	Both	Active and inactive	On medication	Blood	Whole blood	Microarray
GSE126124	Palmer 2019	31618209	32	39	18	89	Children	Active	Treatment naïve	Blood	Whole blood	Microarray
GSE94648	Planell 2017	28981629	22	50	25	97	Adults	Active and inactive	On medication	Blood	Whole blood	Microarray
GSE33943	Van Lierop 2013	24260248	13	24	21	58	Children	Inactive	On medication	Blood	Whole blood	Microarray
GSE86434	Ventham 2016	27886173	13	23	22	58	Adults	Active	Treatment naïve and on medication	Blood	Whole blood	Microarray

GEO Gene Expression Omnibus, PMID PubMed Identifier, HC healthy controls, CD Crohn’s disease, UC ulcerative colitis, PBMCs peripheral blood mononuclear cells

NLRP3 and IL1B mRNA expression in intestinal tissues

Both NLRP3 and IL1B mRNA expression was significantly higher in intestinal tissue of IBD patients compared to HCs (24 datasets, n = 3912). For NLRP3 mRNA expression, we calculated an SMD of 0.53 (95% CI [0.36, 0.70], p < 0.001; Fig. 2) and for IL1B mRNA expression an SMD of 0.95 (95% CI [0.66, 1.24], p < 0.001; Fig. 3). Heterogeneity was moderate for the NLRP3 analysis (I² = 66%) and considerable for the IL1B analysis (I² = 89%).

Fig. 2.

Meta-analysis of NLRP3 mRNA expression in intestinal tissue of inflammatory bowel disease (IBD) patients versus healthy controls (HC). SD, standard deviation; IV, inverse variance; Random, random-effects model; CI, confidence interval; Std. Mean Difference, standardized mean difference (SMD); I², heterogeneity statistic

Fig. 3.

Meta-analysis of IL1B mRNA expression in intestinal tissue of inflammatory bowel disease (IBD) patients versus healthy controls (HC). SD, standard deviation; IV, inverse variance; Random, random-effects model; CI, confidence interval; Std. Mean Difference, standardized mean difference (SMD); I², heterogeneity statistic

NLRP3 and IL1B mRNA expression in blood

IL1B mRNA expression was significantly higher in blood of IBD patients compared to HCs (9 datasets, n = 1921), presenting an SMD of 0.20 (95% CI [0.06, 0.33], p = 0.003; Fig. 4). Heterogeneity was low (I² = 15%). In contrast, NLRP3 mRNA expression in blood of IBD patients was not significantly different from HCs (9 records, n = 1921), presenting an SMD of 0.12 (95% CI [-0.16, 0.40], p = 0.40; Fig. 5) with considerable heterogeneity (I² = 78%).

Fig. 4.

Meta-analysis of IL1B mRNA expression in blood of inflammatory bowel disease (IBD) patients versus healthy controls (HC). SD, standard deviation; IV, inverse variance; Random, random-effects model; CI, confidence interval; Std. Mean Difference, standardized mean difference (SMD); I², heterogeneity statistic

Fig. 5.

Meta-analysis of NLRP3 mRNA expression in blood of inflammatory bowel disease (IBD) patients versus healthy controls (HC). SD, standard deviation; IV, inverse variance; Random, random-effects model; CI, confidence interval; Std. Mean Difference, standardized mean difference (SMD); I², heterogeneity statistic

Assessment of quality of included studies

The methodological quality of the included studies was good (18/33, 55%) or fair (11/33, 33%), with 4 studies (12%) rated as poor (Table 2). All studies received 2 stars for case selection, as IBD diagnoses were based on established international criteria. Likewise, every study was awarded all 3 stars in the outcome domain because cases and controls underwent identical procedures for tissue sampling and gene expression measurement. In contrast, the selection and definition of control groups were insufficiently described. Controls were typically hospital-based and labeled as “normal,” “healthy,” or “non-IBD,” but only the latter explicitly indicated an absence of disease history. Comparability between cases and controls was limited, only 1 study received 2 stars. Most studies described age matching, except for the 4 studies rated as poor.

Table 2.

Newcastle-Ottawa scale assessment of quality

Database described in Author Year	Case definition	Representativeness cases	Selection of controls	Definition of controls	Comparability	Outcome ascertainment	Same method for outcome	Non-response rate	Stars (total)	Quality
Argmann 2022	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Arijs 2009	★	★	-	-	★	★	★	★	★★★★★★ (6)	Fair
Arijs 2018	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Ashton 2021	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Carey 2008	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Fenton 2021	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Haberman 2014	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Haberman 2019	★	★	-	★	★★	★	★	★	★★★★★★★(8)	Good
Haberman 2019	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Kugathasan 2008	★	★	★	★	★	★	★	★	★★★★★★★★(8)	Good
Leal 2015	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Li 2018	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Montero-Meléndez 2013	★	★	-	-	-	★	★	★	★★★★★(5)	Poor
Palmer 2019	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Pavlidis 2022	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Peters 2017	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Planell 2013	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Smith 2014	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Toyonaga 2020	★	★	-	★	-	★	★	★	★★★★★★ (6)	Poor
Van Dussen 2018	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Vancamelbeke 2017	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Vanhove 2015	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Verstockt 2019	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Zabana 2020	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Argmann 2022	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Burczynski 2006	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Juzenas 2022	★	★	-	-	★	★	★	★	★★★★★★ (6)	Fair
Mo 2018	★	★	-	-	-	★	★	★	★★★★★(5)	Poor
Ostrowski 2019	★	★	-	-	-	★	★	★	★★★★★(5)	Poor
Palmer 2019	★	★	-	-	★	★	★	★	★★★★★★(6)	Fair
Planell 2017	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Van Lierop 2013	★	★	-	★	★	★	★	★	★★★★★★★(7)	Good
Ventham 2016	★	★	★	★	★	★	★	★	★★★★★★★★(8)	Good

Exploration of heterogeneity

To investigate potential sources of heterogeneity, we conducted a series of subgroup and sensitivity analyses. Separate meta-analyses were performed for CD and UC in both intestinal tissue and blood (Appendix Figs. 1, 2, 3, 4, 5, 6, 7 and 8), and the results were comparable across IBD subtypes. Likewise, separate analyses for adults and children yielded similar effect sizes (Appendix Figs. 9, 10, 11, 12, 13, 14, 15 and 16). When at least 5 studies were available within a subgroup (Table 1), we further examined NLRP3 and IL1B mRNA expression in studies including only patients with active disease (Appendix Figs. 17 and 18), patients receiving medication (Appendix Figs. 19, 20, 21 and 22), studies using microarrays (Appendix Figs. 23, 24, 25 and 26), and studies using RNA sequencing (Appendix Figs. 27 and 28). Across all subgroup analyses, the direction and significance of effects remained consistent and heterogeneity decreased only modestly, with the exception of a significant SMD for NLRP3 mRNA expression in blood among medicated patients (Appendix Fig. 21).

Sensitivity analyses were also conducted by excluding specific studies and by comparing fixed-effect and random-effects models. Removing the largest dataset (Argmann 2022) or the 4 studies rated as poor quality from the main meta-analyses did not affect the direction and significance of effects or meaningfully reduce heterogeneity (data not shown). Using a fixed-effect model produced similar findings to the random-effects model (Appendix Figs. 29, 30, 31 and 32), with the exception that NLRP3 mRNA expression in blood reached statistical significance under the fixed-effect model (Appendix Fig. 31).

Meta-correlation of NLRP3 and IL1B mRNA expression across studies

A bivariate random-effects meta-analysis across 24 studies in intestinal tissue, estimated a strong positive correlation between SMDs of NLRP3 and IL1B (rho = 0.854, 95% CI [0.687, 0.935], p < < 0.001; Fig. 6). This analysis could not be performed for gene expression in blood, due to insufficient data (9 studies) for a stable bivariate model.

Fig. 6.

Study-level standardized mean differences (SMD) for NLRP3 and IL1B (blue dots) with horizontal and vertical error bars representing 95% confidence intervals (CI). The red line depicts the regression trend across 24 studies in intestinal tissue (n = 3912), summarized by rho, with its 95% CI in grey shading

Discussion

This study systematically analyzed NLRP3 and IL1B mRNA expression in intestinal tissue and blood of patients with IBD compared to HCs, synthesizing results from 33 publicly available datasets including 5833 samples. We found that both NLRP3 and IL1B mRNA expression was significantly upregulated in intestinal tissue of IBD patients. Whereas in blood, IL1B, but not NLRP3 mRNA, was significantly higher.

Our meta-analysis of 24 studies in intestinal tissue (n = 3912) showed robust IL1B mRNA upregulation in intestinal tissue, supporting prior evidence of increased IL-1β production in IBD mucosa [24–26]. Our analysis also found a significant NLRP3 mRNA upregulation in intestinal tissue. Moreover, a meta-correlation of these meta-analyses indicated a strong positive relationship between NLRP3 and IL1B mRNA expression, consistent with the role of NLRP3 protein in regulating IL-1β maturation and release. It is important to keep in mind that increased transcription does not necessarily mean functional inflammasome activation. The NLRP3 inflammasome is strongly regulated at the post-translational level through formation of the NLRP3-ASC-Caspase 1 complex [7]. Nonetheless, the presence of elevated NLRP3 and IL1B transcript levels in inflamed mucosa highlights a pro-inflammatory environment in which inflammasome activation is more likely.

Our pooled analysis of 9 blood transcriptomic studies (n = 1921) indicated higher IL1B mRNA expression in IBD. Previous individual studies have reported inconsistent results [26, 27]. Unstimulated PBMCs do not show significant IL1B mRNA upregulation, whereas studies demonstrate exaggerated IL1B mRNA induction and IL-1β secretion upon LPS stimulation in IBD-derived PBMCs [28, 29]. In contrast, NLRP3 mRNA expression was not significantly increased in blood. Gole et al. (2023) [28] found significantly more IL1B mRNA expression in macrophages once differentiated from monocytes, this could hold true for NLRP3 transcripts as well. Since macrophages accumulate at sites of intestinal inflammation [8], the local upregulation of NLRP3 mRNA in intestinal tissue but not blood can be explained. Together, these findings suggest that systemic immune cells in IBD may be “primed” toward a hyperinflammatory response, even if baseline transcript levels are not uniformly elevated.

NLRP3 activity is tightly controlled under physiological conditions, as excessive activation can drive pyroptosis and amplify proinflammatory innate (e.g. TNFα and IL-6) as well as adaptive (IFN-γ by IL-18) cytokine cascades [7, 11, 30]. Our analysis suggests a shift toward constitutive inflammasome priming in IBD tissue. Environmental, dietary, and host-derived factors, including ATP, saturated fatty acids, and microbiota-derived signals, are known activators of NLRP3 [10]. While direct evidence in IBD is limited, such pathways may contribute to a self-sustaining cycle of inflammasome activation and chronic intestinal inflammation. This hypothesis needs further mechanistic investigation.

Therapeutically, our findings add to a growing body of evidence implicating IL-1β and, by extension, the NLRP3 inflammasome in IBD pathogenesis, particularly in patients unresponsive to anti-TNF therapy. Biologics blocking IL-1 signaling may offer alternatives for anti-TNF–refractory IBD, as suggested by non-responder studies wherein both infliximab and adalimumab response can be predicted by IL-1β levels in patients [28, 31, 32]. Evidence from clinical studies of anakinra and canakinumab in IBD remains limited, with neither agent showing benefit compared to standard care in randomized controlled trials [33]. However, anakinra and canakinumab did benefit specific individuals as reported by case reports [34, 35]. Given the differential expression patterns between blood and intestinal tissue, our analyses suggest that local targeting of the NLRP3 inflammasome, subsequentially limiting IL-1β, TNFα, IL-6, IL-18, and INF-γ production, can offer a more precise therapeutic approach with fewer systemic side effects. However, no NLRP3 inhibitors have yet been tested in IBD patients. MCC950, a potent small-molecule NLRP3 inhibitor, demonstrated efficacy in preclinical IBD models but was discontinued in early-phase trials due to hepatotoxicity [36, 37]. MCC950 derivatives are under investigation now [38]. Oral NLRP3 inhibitor DFV890 is currently in phase II studies in other inflammatory conditions but not in IBD (NCT04886258) [39, 40]. The caspase 1 inhibitor VX765 also showed promise in preclinical IBD models [41]. Thus, while pharmacological targeting of the priming or activation steps of the NLRP3 inflammasome might be promising, clinical translation in IBD remains speculative. Prospective, randomized controlled trails in IBD patients are needed to validate causality and therapeutic potential.

The datasets analyzed herein included heterogeneous patient populations differing in age, sex, disease type, disease activity, treatment status, biopsy location, and transcriptomic platform, contributing to both clinical and statistical heterogeneity. Medication exposure is a particular consideration, as aminosalicylates for example are known to suppress cytokine production, including IL-1β, potentially affecting gene expression levels [42].

When sufficient data were available, we performed subgroup analyses to explore sources of heterogeneity, including disease subtype, age group, treatment status, disease activity, and transcriptomic platform. Across all subgroups, the direction, magnitude, and significance of effects remained consistent, suggesting that heterogeneity reflects the cumulative diversity of patient populations, study designs, and platforms rather than any single factor. For disease subtype and age group, this observation aligns with Ostrowski et al. (2019) [43], who reported largely overlapping gene expression profiles between UC and CD, as well as between pediatric and adult cohorts. Sensitivity analyses, including exclusion of select studies and comparison of fixed-effect versus random-effects models, also did not alter the results, reinforcing the robustness of the observed effect. Overall, despite variability in population characteristics, we consistently observed robust and statistically significant upregulation of intestinal NLRP3 and IL1B mRNA expression in IBD patients, underscoring the strength of the signal.

A limitation of this study is the use of bulk intestinal tissue and blood samples, which prevents identification of specific cell types expressing NLRP3 and IL1B mRNA, necessitating further studies to interrogate which cells may be responsible for the observed differential expression patterns in IBD. In addition, the definition of “healthy” controls varied across datasets, ranging from non-inflammatory individuals to patients undergoing endoscopy for other indications or adults screened for colorectal cancer. Quality assessment indicated that unclear control selection could introduce bias. While obtaining true healthy intestinal biopsies is ethically challenging, this variability should be considered when interpreting the differences between IBD patients and controls.

In conclusion, our meta-analysis provides robust evidence for the correlated upregulation of NLRP3 and IL1B mRNA in intestinal tissue of IBD patients compared to HCs. Our findings reinforce the role of inflammasome-related pathways in intestinal inflammation and suggest that therapeutic targeting of IL-1β and NLRP3 merits further investigation in IBD.

Authors’ contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by PIM and RCHS. The first draft of the manuscript was written by PIM and RCHS. All authors commented on previous versions of the manuscript and all authors read and approved the final manuscript.

Funding

No funding has been received for the preparation of this manuscript.

Data availability

All public datasets analyzed during the current study are available in GEO: GSE193677, GSE16879, GSE73661, GSE153974, GSE9686, GSE128682, GSE57945, GSE109142, GSE117993, GSE10616, GSE52746, GSE87466, GSE36807, GSE126124, GSE206285, GSE83687, GSE38713, GSE48634, GSE85499, GSE112366, GSE75214, GSE59071, GSE102133, GSE83448, GSE186507, GSE3365, GSE169568, GSE112057, GSE119600, GSE126124, GSE94648, GSE33943, GSE86434.

Declarations

Competing interests

The authors declare no competing interests.