Genetic Variations in TrkB.T1 Isoform and Their Association with Somatic and Psychological Symptoms in Individuals with IBS

Abstract

Irritable bowel syndrome (IBS), a disorder of gut-brain interaction, is often comorbid with somatic pain and psychological disorders. Dysregulated signaling of brain-derived neurotrophic factor (BDNF) and its receptor, tropomyosin-related kinase B (TrkB), has been implicated in somatic-psychological symptoms in individuals with IBS. We investigated the association of 10 single nucleotide polymorphisms (SNPs) in the regulatory 3′ untranslated region (UTR) of NTRK2 (TrkB) kinase domain-deficient truncated isoform (TrkB.T1) and BDNF Val66Met SNP with somatic and psychological symptoms and quality of life in a cohort from the United States (U.S.) (IBS n=464; healthy controls n=156). We found that the homozygous recessive genotype (G/G) of rs2013566 in individuals with IBS is associated with worsened somatic symptoms, including headache, back pain, joint pain, muscle pain, and somatization as well as diminished sleep quality, energy level and overall quality of life. Validation using United Kingdom BioBank (UKBB) data confirmed the association of rs2013566 with increased likelihood of headache. Several SNPs (rs1627784, rs1624327, rs1147198) showed significant associations with muscle pain in our U.S. cohort. These 4 SNPs are predominantly located in H3K4Me1-enriched regions, suggesting their enhancer and/or transcription regulation potential. Our findings suggest that genetic variation within the 3′UTR region of the TrkB.T1 isoform may contribute to comorbid conditions in individuals with IBS, resulting in a spectrum of somatic and psychological symptoms impacting their quality of life. These findings advance our understanding of the genetic interaction between BDNF/TrkB pathways and somatic-psychological symptoms in IBS, highlighting the importance of further exploring this interaction for potential clinical applications.

Introduction

Irritable bowel syndrome (IBS) is a disorder characterized by altered bowel habits and chronic abdominal pain, resulting from the complex interaction between the gut and the brain.^1,2 Individuals with IBS often experience extraintestinal symptoms such as backache, headache, joint pain, fatigue, and sleep disturbances.³ They also exhibit a higher prevalence of neurological pain conditions, such as fibromyalgia and migraine, compared to those without IBS.⁴ Psychological symptoms are also common, with a three-fold higher risk of anxiety or depression than in healthy individuals.⁵ Various mechanisms, including serotonin and bile acid metabolism abnormalities, altered intestinal permeability, immune activation, microbiota disturbances, and changes in brain function, contribute to IBS symptoms.⁶ Genetic factors play a significant role in IBS, as shown in epigenetic, genome-wide association, and candidate gene association analyses.^7–9 However, the genetic determinants associated with both gastrointestinal and extraintestinal symptoms that expand IBS symptomatology, such as neuropathic pain or psychological symptoms, as well as their influence on the quality of life of people with IBS, remain elusive. Given the dynamic clinical phenotypes in IBS, we hypothesize that specific genes and variants act as broader genetic regulators, influencing a wide spectrum of neurobiological pathways that encompass not only somatic symptoms but also psychological symptoms.

Brain-derived neurotrophic factor (BDNF) signaling plays a crucial role in diverse somatic pain^10–12 and psychological disorders.^13,14 BDNF belongs to the neurotrophin growth factor family and is vital for neuronal development, survival, and plasticity in the central and peripheral nervous systems.^15–17 BDNF transmits intracellular signaling through the tropomyosin-related kinase B (TrkB) receptor, encoded by neurotrophic receptor tyrosine kinase 2 (NTRK2) gene.¹⁸ The NTRK2 locus encodes multiple TrkB receptor isoforms with distinct functional activities. The primary isoform, full-length TrkB (TrkB.FL), possesses a C-terminal tyrosine kinase domain (exon 20–24) that mediates several canonical signaling pathways, including phosphoinositide 3-kinase-protein kinase B (PI3K/Akt), mitogen-activated protein kinases (MAPK), and phospholipase C-γ (PLCγ) signaling.¹⁹ In contrast, the C-terminal truncated isoform, TrkB.T1, lacks the intracellular kinase domain, resulting in limited canonical or alternative BDNF signaling.²⁰ In both human and animal models, TrkB.T1-mediated non-canonical signaling and differential expression of TrkB.T1 are associated with an increased risk of neuropathic pain^21–23 and mood and psychological disorders.^24–26 Together, these evidences emphasize TrkB.T1’s role in modifying BDNF signaling in somatic and psychological manifestations, which opens new avenues to explore their effects on neuropsychological symptoms in people living with IBS.

Understanding the genetic effects of the TrkB.T1 isoform on somatic and psychological symptoms in IBS could be valuable in uncovering mechanisms and improving diagnostic and therapeutic approaches. The TrkB.T1 isoform is generated via alternative splicing, specifically using exon 16. This exon contains regulatory elements, including H3K4me1 histone mark and polyadenylation sites in its 3′ untranslated region (UTR), essential for gene expression and functional regulation.^24,27–29 We hypothesize that genetic variations in the TrkB.T1 isoform’s 3′UTR may regulate BDNF/TrkB signaling, influencing the manifestation of somatic and psychological symptoms commonly observed in individuals with IBS. This study investigates associations of 14 single nucleotide polymorphisms (SNPs), including 13 in the regulatory 3′UTR region of the TrkB.T1 isoform and well-established functional SNP BDNF Val66Met (rs6265), which impacts BDNF protein release.³⁰ Through genotypic analysis in three domains— (i) abdominal pain and other somatic symptoms, (ii) psychological symptoms, and (iii) quality of life—we aim to comprehensively understand their genetic effects on IBS-related symptomatology. External validation using United Kingdom BioBank (UKBB) data enhances the reliability of our findings.

Materials and Methods

Study Participants

This study examined participant data across five independent research studies conducted at a research center in the Pacific Northwest of the United States (U.S.) (Figure 1). Among these studies, three were observational and focused on understanding the mechanisms of IBS. They recruited women with IBS and women without the condition, who served as healthy controls (HCs) within the age range of 18–50.^31–33 The remaining two studies were randomized controlled trials that investigated a comprehensive self-management intervention. They recruited both men and women with IBS within the age range of 18–70.^34,35 For the current analysis, only the baseline data prior to randomization from the intervention studies was utilized. The study protocols for all five studies were approved by the institutional review board at the University of Washington (97–3895, 29114, 32722, 37126, and 47112) and informed consent was obtained from each subject. Detailed information about the complete study protocols can be found elsewhere.^31–36 Briefly, participants for all studies were recruited through mailings from a gastroenterology practice and community advertisements. The eligibility criteria were similar across the studies. Individuals with IBS had a confirmed diagnosis for at least 6 months by a healthcare provider and met the Rome III research criteria for IBS. The HCs were required to have no moderate to severe diseases, disorders, or symptoms. Additionally, individuals with a history or current experience of IBS-like symptoms were excluded from the HC group. Baseline blood samples were obtained from all participants to conduct targeted genetic testing.

Figure 1. Flow Diagram Illustrating Data Utilization in Five IBS Studies.

*HC individuals were recruited solely for questionnaire and DNA sampling purposes, without any intervention being provided.

BSI-18: Brief Symptom Inventory-18, DSQ: Disease Specific Questionnaire, HC: healthy control

Study Measures

Demographic survey collected participant information on self-identified age, race, ethnicity, gender, marital status, employment, education, and family income using a demographic questionnaire.

The Daily Symptom Diary was used to record the severity of symptoms in a daily diary. Participants were asked to ‘mark the highest severity that you have experienced with each symptom over the past 24 hours’, with symptoms rated as: not present (0), mild (1), moderate (2), severe (3), and very severe (4). Ratings were averaged for each symptom. Daily symptoms included in this analysis are abdominal pain or discomfort, somatic pain (backache, headache, joint pain, and muscle pain), and psychological symptoms (anxiety, depressed/sad or blue). The Brief Symptom Inventory (BSI-18)³⁷ was used to measure psychological distress, including symptoms of anxiety, depression, and somatization. Participants responded on a scale from not at all (0) to extremely (4). Internal consistency of symptom dimensions is reported as .71 to .85, and test-retest reliability from .68 to .91.^38,39

The IBS-related quality of life was evaluated with the Disease Specific Questionnaire (DSQ).⁴⁰ Participants reported the impact of IBS symptoms on their well-being over the past four weeks. Scores range from 0 to 100 for each subscale: emotional well-being, mental health, sleep, energy, physical functioning, food, social interactions, physical role, and sexual relations. Higher scores on these subscales indicate a higher quality of life. The DSQ was administered solely to the IBS group since it is a specific tool for measuring quality of life in individuals with this condition.

Biospecimen Collection and DNA Extraction

Biospecimens were collected per protocol of the above referenced investigations. Genomic DNA (gDNA) was extracted from either frozen buffy coat preparations⁴¹ or fresh whole blood using Qiagen DNeasy Blood & Tissue kits (Qiagen, Valencia, CA) or Puregene DNA Purification kits (Gentra Systems Inc., Minneapolis, MD). Samples were stored at −80°C until shipped overnight on dry ice to the Institute for Genome Sciences at the University of Maryland, Baltimore.

SNP Genotyping and Quality Assurance

Genotyping was performed using TaqMan® SNP Genotyping Assays (Applied Biosystems, Life Technologies Corporation), which include two sequence-specific primers for amplification of sequences containing the SNPs of interest, and two allele-specific TaqMan^™ probes for Allele 1 and Allele 2. The samples were thawed, and gDNA was quantitated. A reaction mix containing TaqMan^® Genotyping Assay (20X), TaqMan^® Genotyping Master Mix, and nuclease-free water was prepared. The reaction mix and samples were aliquoted into 96-well plates, organized by study. The plates were then run on QuantStudioX Real-Time PCR System, with analysis of genotyping experiments performed using QuantStudio^™ Design and Analysis desktop Software. Genotyping was performed using the GRCh38 reference genome as the basis for variant calling and allele identification. Detailed information on the 14 genotyped SNPs, including the closest gene, function, frequency, and minor allele frequency (MAF) in our study and the European population of the 1000 Phase 3 Genome Project,⁴² is provided in Supplemental Table 1. To assess the quality of the genotyping, we controlled the genotypes cluster for each SNP. Two SNPs, rs2013566 and rs3654, located in the 3′UTR of NTRK2 gene, failed genotyping in one out of the five studies, and therefore contained 20% less samples for these SNPs. The remaining SNPs were successfully genotyped in all studies with an average number of missing genotyped samples equal to 5.5 (range 3–11). As an additional quality control measure, considering our sample’s 75% Caucasian ancestry, we compared the MAF of all 14 SNPs with their MAF in the European population of the 1000 Phase 3 Genome Project. We observed a high correlation between the computed allele frequencies of the SNPs in our sample and those of the 1000 Genomes dataset. Two SNPs, rs121434633 and rs74356179, were excluded from the analysis due to their extremely low MAF values in our samples (0.000 and 0.002, respectively; Supplemental Table 1). Next, we measured pairwise squared correlation (r²) linkage disequilibrium (LD) among the SNPs in our samples using Haploview.⁴³ A strong correlation (r² = 0.9) was observed between rs45596934 and rs138535351, leading us to exclude rs45596934 due to its higher number of missing genotypes compared to rs138535351. None of the remaining SNPs in the NTRK2 gene exhibited a significant correlation (r² < 0.6, Figure 2). In conclusion, after conducting a thorough quality check of the 14 genotyped SNPs, we retained 11 SNPs for downstream analyses, with 10 of them located in the NTRK2 gene and 1 in the BDNF gene. The SNP locations within the NTRK2 gene and their association with the most common transcripts are shown in Figure 2.

Figure 2. TrkB.T1 Isoform and SNPs in 3′ untranslated region (UTR).

(Top) Structure of the TrkB isoforms. Protein coding regions are shown as light green boxes and exons are numbered. 5′ and 3′ UTR regions are shown as empty boxes. The location of translational start codon (red) and stop codon (blue) are shown as thin boxes. TrkB.FL contains an intracellular C-terminal tyrosine kinase domain (exon 20–24) and is involved in various signaling pathways, including PI3K/Akt, MAPK, and PLCγ signaling. The truncated isoform, TrkB.T1, lacks the intracellular kinase domain due to the presence of exon 16, which contains 33 nucleotides that encode a unique intracellular 11 amino acid domain, a stop codon, and a unique 3′ UTR sequence with multiple polyadenylation sites. (Bottom) 3′ UTR SNPs in TrkB.T1 Isoform and LD Heatmap. The bottom figure illustrates the pass-filter 11 SNPs located within the 3′ UTR of the TrkB.T1 isoform, where our candidate SNPs are positioned. The heatmap represents R² linkage disequilibrium (LD) for the SNPs in this region. Based on the strong correlation (R²= 0.9) observed between rs45596934 and rs138535351, rs45596934 was omitted from the analysis due to a relatively higher number of missing genotypes than rs138535351. Note that rs1147198 and rs10868235 are situated in an upstream intergenic region and a downstream intronic region, respectively, outside of exon 16. Additionally, rs6265 is located in the BDNF gene on a different chromosome (not shown). The isoform structure in the top figure has been modified from Figure 1 in Timmusk et al. (2010) with permission from John Wiley and Sons.

Data Analyses

The participant data from the five studies were merged, and demographic characteristics were compared between participants with IBS and HCs using t-tests and chi-square tests, with a significance level set at p < 0.05, using R studio.⁴⁴ Logistic regression analysis was initially conducted with 11 SNPs to test for their association with IBS status (IBS n = 464, HC n = 156), controlling for age, sex, and study design. Our primary analysis of the relationship between NTRK2 and BDNF SNPs and symptoms was conducted using an additive linear regression model, considering that symptoms were measured as continuous variables. To investigate potential dominance or recessive effects, we further employed a genotypic modifier, which in addition to testing the additive (ADD) model (number of minor alleles coded as 0, 1, or 2) assessed significant deviation from dominance (DOMDEV) by comparing heterozygote genotypes (coded as 1) to homozygotes (coded as 0). Finally, the genotypic modifier implemented a joint test (GENO_2DF) to evaluate the combined significance of additive and dominance-deviation terms.⁴⁵ The significant effects observed solely in the ADD model indicate a linear dose-dependent relationship between the number of minor alleles and the trait. Additionally, the consistent and significant findings in the DOMDEV and GENO_2DF models, aligning with the ADD model, suggest a potential dominance effect. Moreover, when all three tests are significant and ADD and DOMDEV have effects with opposite directions, it suggests a recessive effect. To maximize the robustness of our findings, we do not consider associations as significant if they are significant only in the DOMDEV and GENO_2DF models but not in the ADD model. To ensure accurate estimation of effect sizes, significant genetic associations are determined based on the criterion that all three genotypes (homozygous dominant, heterozygous, and homozygous recessive) should be represented by a minimum of two individuals. Both logistic and linear regression analyses were performed using PLINK (Version 2).^46,47 To address potential confounding factors arising from independent study recruitment strategies, all regression models were adjusted for age, sex, and study design. To account for the increased risk of false positive results due to multiple testing, we employed the Benjamini-Hochberg procedure to estimate and control the false discovery rate (FDR).

Validation Analyses

To ensure the reliability of our results, we conducted external validation of the associations between 11 SNPs and abdominal/somatic and psychological symptoms. This validation was performed using publicly available databases, namely the ES200kUKB and ImpUKB datasets, which were curated using the Omics Analysis, Search & Information System (OASIS).⁴⁸ The ES200kUKB dataset consisted of participants from the UKBB who underwent whole exome sequencing, enabling comprehensive identification of genetic variants.⁴⁹ The ImpUKB dataset represented imputed genotypes of UKBB participants based on the Haplotype Reference Consortium (HRC) and UK10K haplotype resource, allowing accurate imputation of missing genotypes.⁵⁰ Using International Classification of Diseases, 10^th revision (ICD-10) codes for IBS, pain, and psychiatric diagnoses (Supplemental Table 2), we identified individuals with IBS and symptoms aligning closely with those in our primary investigation. This enabled us to differentiate cases (IBS individuals with one or more comorbid somatic and psychological symptoms based on ICD-10 codes) from controls (IBS individuals without comorbid symptoms based on ICD-10 codes). Of 200,643 individuals in ES200kUKB dataset, 3,060 were diagnosed with IBS according to ICD-10 codes. Similarly, the ImpUKB dataset included 487,409, and 7,568 individuals diagnosed with IBS. Logistic regression using an additive genetic model was used to test the associations between the SNPs and 9 ICD-10 diagnoses, specifically comparing individuals who had comorbid diagnoses of IBS with other somatic symptoms (backache, headache, joint pain, muscle pain, impaired sleep, decreased energy, and/or somatization) and psychological diagnoses (anxiety, depression) to HCs. We applied MAF greater than or equal to 0.005 as criteria to select SNPs and observed 72 associations between the 11 SNPs and 9 symptom traits in ES200kUKB, and 99 associations in ImpUKB.

Results

Demographics and IBS-related symptom characteristics

Genotypes of 11 SNPs were analyzed in 620 participants in the U.S., comprising 464 individuals with IBS and 156 HCs recruited from five independent studies (Figure 1). Table 1 summarizes the baseline characteristics of participants in two groups: IBS Observational Studies (ages 18–50) and IBS Intervention Studies (ages 18–70). In the IBS Observational Studies group, there were 269 participants (IBS n = 163, HC n = 106), with similar mean ages between the groups (p = 0.42). Although the proportion of White individuals was higher among IBS participants compared to HCs, no significant differences were observed in ethnicity, gender, marital status, employment status, education, or family income between the IBS and HC groups. In the IBS Intervention Studies group, there were 351 participants (IBS n = 301, HC n = 50), with similar mean ages (p = 0.62). HCs were recruited for questionnaire and DNA sampling purposes only, no intervention provided. While no significant differences were found in race, ethnicity, gender, education, or employment status, the HC group had a higher proportion of married/partnered individuals (p = 0.002) and lower family income ≤ 50k (p = 0.03) compared to the IBS group. However, the significance of the difference in marital status is more likely to be influenced by the higher amount of missing data in the HC group compared to the IBS group.

Table 1.

Baseline Sample Characteristics

IBS Observational Studies (ages 18–50)
	Total (n=269)	HC (n=106)	IBS (n=163)	p value*
Age, Mean (SD)	28.28 (7.0)	27.87 (6.3)	28.54 (7.5)	0.42
Race, n (%)				0.11
Other	66 (24.5)	32 (30.2)	34 (20.9)
White	203 (75.5)	74 (69.8)	129 (79.1)
Ethnicity, n (%)				0.52
Hispanic or Latino	23 (8.6)	11 (10.4)	12 (7.4)
Not Hispanic or Latino	246 (91.5)	95 (89.6)	151 (92.6)
Gender, n (%)
Female	268 (99.6)
Missing	1 (0.4)
Marital status, n (%)				0.44
Married or partner	82 (30.5)	31 (29.2)	51 (31.3)
Not married	186 (69.1)	74 (69.8)	112 (68.7)
Employment, n (%)				0.98
Full time	84 (31.2)	32 (30.2)	52 (31.9)
Part time	81 (30.1)	33 (31.1)	48 (29.5)
Not working	42 (15.6)	16 (15.1)	26 (15.9)
Missing	62 (23.1)	25 (23.6)	37 (22.7)
Education, n (%)				0.14
≤ 12th grade	21 (7.8)	5 (4.7)	16 (9.8)
Some college	168 (66.2)	70 (66.0)	108 (66.3)
Master or higher	67 (24.9)	31 (29.3)	36 (22.1)
Missing	3 (1.1)		3 (1.8)
Family income, n (%)				0.45
< 50k	103 (38.3)	44 (41.5)	59 (36.2)
50k −100k	73 (27.1)	30 (28.3)	43 (26.4)
>100k	42 (15.6)	12 (11.3)	30 (26.4)
Missing	51 (19.0)	20 (11.3)	31 (19.0)
IBS Intervention Studies (ages 18–70)
	Total (N=351)	HC (n=50)	IBS (n=301)	p value
Age, Mean (SD)	42.91 (14.7)	43.88 (14.8)	42.74 (14.7)	0.62
Race, n (%)				1.00
Other	66 (18.8)	9 (18.0)	57 (18.9)
White	285 (81.2)	41 (82.0)	244 (81.1)
Ethnicity, n (%)				0.13
Hispanic or Latino	12 (3.4)	4 (8.0)	8 (2.7)
Not Hispanic or Latino	339 (96.6)	46 (92.0)	293 (97.3)
Gender, n (%)				0.87
Female	294 (83.8)	41 (82.0)	253 (84.1)
Male	57 (16.2)	9 (18.0)	48 (16.0)
Marital status, n (%)				0.002
Married or partner	149 (42.4)	20 (40.0)	129 (42.9)
Not married	200 (57.0)	28 (56.0)	172 (57.1)
Missing	2 (0.6)	2 (4.0)	0 (0.0)
Employment, n (%)				0.09
Full time	148 (42.2)	14 (28.0)	134 (44.5)
Part time	102 (29.1)	16 (32.0)	86 (28.6)
Unemployed	98 (27.9)	20 (40.0)	78 (25.9)
Missing	3 (0.9)	0 (0.0)	3 (1.0)
Education, n (%)				0.98
≤ 12th grade	29 (8.3)	4 (8.0)	25 (8.3)
Some college	235 (67.0)	34 (68.0)	201 (66.8)
Master or higher	86 (24.5)	12 (24.0)	74 (24.6)
Missing	1 (0.3)	0 (0.0)	1 (0.3)
Family income, n (%)				0.03
≤ 50k	113 (32.2)	24 (48.0)	89 (21.7)
50k −100k	101 (28.8)	12 (24.0)	89 (21.7)
>100k	61 (17.4)	9 (18.0)	52 (17.3)
Missing	76 (21.7)	5 (10.0)	71 (23.6)

^*p values from t-tests or chi-square tests

HC: healthy control, IBS: irritable bowel syndrome, SD: standard deviation, n: number of sample.

Individuals with IBS experience higher levels of somatic pain and psychological distress compared to HCs (Table 2). In the abdominal pain and other somatic symptom domain, IBS participants showed significantly higher mean scores for abdominal pain, back pain, headache, joint pain, muscle pain, and somatization than HCs (p < .0001). Similarly, in the psychological symptom domain, IBS participants exhibited higher mean scores for depression and anxiety than HCs (p < .0001). In the domain of quality of life related to IBS, IBS participants had an overall score of 67.91 (SD = 15.71), indicating a moderate level of quality of life. Together, individuals with IBS experience higher levels of somatic pain and psychological distress compared to HCs.

Table 2.

Symptom comparison between IBS and HCs

Symptom Domains	Daily Symptom Diary			BSI-18			DSQ
	IBS Mean (SD) [Min - Max]	HC Mean (SD) [Min - Max]	p value	IBS Mean (SD) [Min - Max]	HC Mean (SD) [Min - Max]	p value	IBS only Mean (SD) [Min - Max]
Symptom domain i. Abdominal pain and other somatic symptoms
Abdominal pain	1.27 (0.63) [0 – 3.89]	0.13 (0.18) [0– 1.16]	< .0001
Back pain	0.58 (0.64) [0–3.96]	0.18 (0.39) [0– 2.61]	< .0001
Headache	0.52 (0.57) [0–3.96]	0.18 (0.25) [0– 1.57]	< .0001
Joint pain	0.48 (0.64) [0–3.61]	0.05 (0.18) [0– 1.44]	< .0001
Muscle pain	0.56 (0.64) [0–3.19]	0.15 (0.27) [0– 1.82]	< .0001
Somatization				0.63 (0.61) [0– 3.17]	0.08 (0.13) [0–0.50]	< .0001
Symptom domain ii. Psychological symptoms
Depression*	0.48 (0.56) [0–3.96]	0.22 (0.39) [0– 2.58]	< .0001	0.53 (0.59) [0–3.00]	0.29 (0.47) [0–2.83]	< .0001
Anxiety**	0.83 (0.66) [0–3.89]	0.36 (0.47) [0– 2.09]	< .0001	0.66 (0.65) [0–3.67]	0.26 (0.39) [0–2.83]	< .0001
Symptom domain iii: IBS-related quality of life
Overall QoL score†							67.91 (15.71) [8.6–100]
Emotional well-being							52.88 (21.76) [0–100]
Mental health							76.88 (18.09) [0–100]
Sleep							80.85 (18.07) [5–100]
Energy							66.32 (23.49) [8.3–100]
Physical functioning							79.67 (19.03) [0–100]
Food							60.9 (20.06) [0–100]
Social							63.61 (23.25) [0–100]
Physical role							57.68 (25.46) [0–100]
Sexual relations							63.96 (25.91) [0–100]

^*Pearson’s correlation coefficient between depression, as collected through the Daily Symptom Diary and BSI-18, is 0.54 (0.48–0.60; p <.0001).

^**Pearson’s correlation coefficient between anxiety, as collected through the Daily Symptom Diary and BSI-18, is 0.57 (0.51–0.62; p <.0001).

Note: DSQ items only measured in IBS group (n=464); No data available in greyed cells.

BSI-18: Brief Symptom Inventory 18, DSQ: Disease Specific Questionnaire, HC: healthy control, IBS: irritable bowel syndrome, max: maximum, min: minimum, QoL: quality of life, SD: standard deviation

Associations of SNPs in NTRK2 and BDNF genes with IBS phenotype

First, we conducted case-control association analysis using an additive logistic regression model to investigate the potential association between the SNPs and the IBS phenotype. Among the 11 SNPs analyzed, we found that the number of minor alleles at rs41277883 was associated with approximately 3.4 times higher odds of having IBS (95% confidence interval [CI] = 2.30–4.48). However, no significant differences in genotype frequencies were found for the remaining 10 SNPs between individuals with IBS and HCs (Table 3).

Table 3.

Genetics Association of SNPs and IBS phenotypes

SNP ID	Position	REF	ALT	MAF	n	aOR	95% CI	p value
rs1147198	9:84660433	T	G	0.266	613	1.17	0.87, 1.47	0.31
rs1187286	9:84800113	T	G	0.244	611	0.97	0.67, 1.26	0.82
rs1047896	9:84811058	A	G	0.187	613	1.22	0.88, 1.57	0.25
rs2013566	9:84811383	A	G	0.113	494	1.02	0.53, 1.51	0.93
rs41277883	9:84812145	A	G	0.033	611	3.39	2.30, 4.48	0.03
rs138535351	9:84813892	G	A	0.032	615	0.83	0.12, 1.53	0.60
rs1627784	9:84813951	T	C	0.284	612	1.08	0.78, 1.38	0.61
rs1624327	9:84814375	G	A	0.250	607	1.05	0.75, 1.36	0.74
rs3654	9:84815576	T	C	0.069	491	0.82	0.22, 1.41	0.51
rs10868235	9:84878840	C	T	0.466	612	0.96	0.70, 1.22	0.74
rs6265	11:27658369	C	T	0.200	614	0.92	0.61, 1.23	0.60

aOR: adjusted odds ratio relative to the alternative allele, ALT: alternate allele, CI: confidence interval, MAF: minor allele frequency, n: number of samples, REF: reference allele

Uncovering the genetic associations of symptoms in individuals with IBS

Next, we examined the association between 11 SNPs and various quantitative symptom traits, including somatic symptoms, psychological symptoms, and quality of life in individuals with IBS. Using a linear regression model and controlling for age, sex, and study design, we simultaneously tested three genetic models (ADD, DOMDEV, and GENO_2DF; Supplemental Table 3). Notably, the minor allele at rs2013566 demonstrated significant associations with multiple somatic symptom traits, such as back pain, headache, joint pain, muscle pain, and somatization, in the ADD model (p < 0.05; Table 4). Interestingly, these associations were also significant at the DOMDEV and GENO_2DF tests (Supplemental Table 3), indicating the associations between rs2013566 and somatic symptoms follow a recessive pattern (Figure 3a–e). Importantly, contrasting associations were observed within rs2013566 and quality of life domain symptoms. Individuals with a homozygous recessive genotype (G/G) at rs2013566 experienced compromised sleep quality, decreased energy levels, and reduced overall quality of life (p < 0.05, Figure 3f–h). These contrasting associations were supported by DOMDEV and GENO_2DF tests. In the ADD model, the SNP was negatively associated with energy levels and sleep quality, but positively associated in the DOMDEV model (Supplemental Table 3). These findings suggest that individuals with a homozygous recessive genotype at rs2013566 may experience the exacerbation of various somatic symptoms, negatively impacting their quality of life, partly due to poor sleep quality or reduced energy levels. Additionally, three SNPs demonstrated significant associations with specific symptom traits in individuals with IBS, particularly related to muscle pain. For example, the homozygous recessive genotype of rs1147198 showed a positive association with muscle pain (p < 0.05 in three models, Figure 4a), consistent with the findings observed for rs2013566 (Figure 3d). In contrast, the minor alleles of rs1627784 and rs1624327 showed a negative association with muscle pain in the ADD model only (β = −0.12, p = 0.03; β = −0.14, p = 0.03, respectively; Figure 4b, c). Nonetheless, our primary analysis identified no SNPs significantly associated with any psychological symptoms, either depression or anxiety (Supplemental Table 3). Overall, these findings highlight the potential influence of these SNPs on specific symptom traits often associated with IBS, particularly related to somatic pain such as headache and musculoskeletal pain, as well as aspects of quality of life such as sleep quality and energy level.

Table 4.

Significant associations between SNPs and IBS-related symptoms in primary and UKBB data

	Primary Data					ImpUKB				ES200K
Symptoms	SNP ID	R/A	n	β (SE)	p	n	OR	95% CI	p	n	OR	95% CI	p
Symptom domain i. Abdominal pain and other somatic symptoms
Headache	rs2013566	A/G	356	0.37 (0.17)	0.029	7476	1.11	0.93–1.33	0.24	3060	1.34	1.03–1.74	0.028
Back pain	rs2013566	A/G	356	0.49 (0.19)	0.011	7474	0.79	0.65–0.95	0.01	3055	1.02	0.78–1.34	0.894
Joint pain	rs2013566	A/G	344	0.83 (0.17)	<0.001	7445	1.02	0.82–1.28	0.85	2995	1.19	0.85–1.67	0.313
Somatization	rs2013566	A/G	386	0.32 (0.13)	0.012	861	0.83	0.35–1.98	0.67	336	0.69	0.13–3.57	0.654
Muscle pain	rs2013566	A/G	356	0.57 (0.18)	0.002	1176	0.92	0.47–1.81	0.81	492	1.34	0.57–3.15	0.507
	rs1147198*	T/G	427	0.14 (0.06)	0.015	1176	1.41	0.93–2.12	0.1
	rs1627784	T/C	426	−0.12 (0.06)	0.032	1176	1.09	0.73–1.61	0.68	492	1.29	0.73–2.27	0.382
	rs1624327	G/A	424	−0.14 (0.06)	0.034	1176	1.35	0.91–2.01	0.14	492	1.42	0.81–2.52	0.223
Symptom domain iii. IBS-related quality of life
Overall QoL**	rs2013566	A/G	375	−8.85 (3.93)	0.025
Sleep	rs2013566	A/G	375	−16.41 (4.49)	<0.001	1134	0.91		0.78	399	1.17		0.782
Energy	rs2013566	A/G	375	−15.49 (5.88)	0.009	4458	0.88		0.49	1581	1.15		0.598

Note: This table includes only the statistically significant summary statistics in the ADD model from primary data (Supplemental Table 3) and the matched ORs from external validation using UKBB data (Supplemental Table 4). No data are available in the greyed-out cells.

^*SNP not genotyped for muscle pain in the ES 200K cohort

^**Trait does not present in the UKB Imp cohort

CI: confidence interval, ES200k: UKBB participants with whole exome sequencing for comprehensive genetic variant identification, ImpUKB: imputed genotypes of UKBB participants using Haplotype Reference Consortium (HRC) and UK10K haplotype resources, n: number of sample, OR: odds ratio, p: p-value, QoL: quality of life, R/A: reference/alternative alleles, SE: standard error, SNP: single nucleotide polymorphism, UKBB: United Kingdom Biobank, β: beta-coefficient relative to the alternative allele.

Figure 3. Association of rs2013566 with symptom traits in individuals with IBS.

In these bar graphs, the x-axis represents the genotypes at rs2013566, while the y-axis represents the mean scores of each trait (as indicated below). Each bar represents a different genotype group. Error bars extending from each bar indicate the standard error (SE) from the mean. The traits represented by the labels are as follows: (a) back pain; (b) headache; (c) joint pain; (d) muscle pain; (e) somatization; (f) sleep; (g) energy levels; (h) quality of life.

Figure 4. Associations of other SNPs in NTRK2 with muscle pain in participants with IBS.

These bar graphs display the associations between different SNPs in NTRK2 and muscle pain in participants with IBS. The specific SNPs in this figure include: (a) rs1147198; (b) rs1627784; (c) rs1624327.

External validation using UKBB data revealed that rs2013566 is associated with other traits.

To validate our primary findings, we conducted an external validation using independent panels of UKBB data, including ES200k and ImpUKB. Among our 11 candidate SNPs, six (rs2013566, rs1627784, rs1624327, rs1047896, rs1187286, and rs41277883) showed significant associations with various somatic and psychological symptoms in individuals with IBS in the UKBB data (p < 0.05, Supplemental Table 4). Among these six SNPs, three (rs2013566, rs1627784, and rs1624327) were also identified as significant in our primary analysis. Specifically, the minor allele of rs2013566 consistently increased the likelihood of headaches (β = 0.37, p = 0.03), not only in our U.S. study population but also among individuals with IBS in the U.K. in the ES200k dataset (odds ratio [OR] = 1.34, 95% CI = 1.03–1.74; Table 4). However, it exhibited contrasting associations with back pain, being positively associated in our primary data (β = 0.49, p = 0.01), but negatively associated in the ImpUKB (OR = 0.79, 95% CI = 0.65–0.95; Table 4). In our primary analysis, we observed a negative association between the minor allele of rs1627784 and muscle pain (β = −0.12, p = 0.03). Although this association was not significant in both UKBB datasets, a distinct association with increased odds of depression (OR = 1.58, 95% CI = 1.11–2.24; Supplemental Table 4) was identified. Similarly, the association between the minor allele of rs1624327 and reduced muscle pain (β = −0.14, p = 0.03) was not supported by both UKBB datasets. Instead, rs1624327 was associated with decreased somatization (OR = 0.53, 95% CI = 0.3–0.91) and increased depression (OR = 1.57, 95% CI = 1.1–2.24) in the ImpUKB dataset, as well as increased headache (OR = 1.27, 95% CI = 1.07–1.5) in the ES200k dataset. Overall, the external validation not only confirmed the association between rs2013566 and an increased risk of headaches but also revealed a complex relationship between these SNPs and various somatic and psychological symptoms in individuals with IBS, which may vary across specific populations.

Discussion

Although there is growing recognition of the impact of genetic predisposition on physiopsychological symptoms and quality of life in IBS, the role of genetic associations and their mechanistic contribution underlying these associations remain largely unexplored. This study aims to address this gap by investigating BDNF/NTRK2 gene polymorphisms’ role in comorbid symptoms in individuals with IBS. Examining the association between SNPs in these genes and various symptom traits, we found significant associations, including individuals with a homozygous recessive genotype (G/G) at rs2013566 showing positive associations with multiple somatic symptoms and negative associations with quality of life. The association between rs2013566 and headaches in individuals with IBS was validated by UKBB data. Three other SNPs were also identified, associated with muscle pain and reduced quality of life. These findings shed light on complex genetic factors underlying diverse IBS symptoms, providing insights for designing future research and clinical intervention.

The BDNF/TrkB signaling pathway is crucial for neuron development, survival, plasticity, and function in the central nervous system (CNS), with effects influenced by cell type, tissue, age, and context.²⁹ The TrkB gene (NTRK2) is highly expressed in the mammalian brain, primarily undergoing alternative splicing in the brain.^51–54 TrkB.FL is the primary isoform during CNS development, and TrkB.T1 expression rises after birth, reaching its peak in adulthood.⁵⁵ Interestingly, TrkB.T1 is the primary isoform outside the nervous system, found in the heart, kidney, lung, and pancreas.^56–58 Previous studies indicate TrkB.T1 upregulation in astrocytes during nociception and neuropathic pain after spinal cord injury in rodents.^20,21,59 Our latest findings also show that animals with genetic deletion of the variant TrkB.T1 have less pain and better locomotor recovery after spinal cord injury.⁶⁰ Given TrkB.T1’s predominant expression in the adult brain and its upregulation linked to pain, the observed association between rs2013566 and headaches in IBS individuals is biologically plausible. TrkB.T1 signaling contributes to pain through potential mechanisms, including mediating IP3-dependent calcium release from intracellular stores.^56,61,62 Independent mouse studies reveal that calcium-mediated signaling and imbalances between TrkB.FL and TrkB.T1 isoforms impact motor function in amyotrophic lateral sclerosis⁶³ and cardiac contractility in cardiomyopathy.⁵⁶ These findings support the association between rs2013566 and physiological functions, potentially correlating with back pain, joint pain, muscle pain, and somatization in our study. TrkB.T1-dependent alterations in calcium homeostasis and signaling pathways involving IP3 receptors in non-brain tissues may explain this association, contributing to pain pathway sensitization, amplification of nociceptive signals, and musculoskeletal pain development in individuals with IBS.

The TrkB.T1 SNPs are implicated in a spectrum of pain manifestations, likely due to their enhancer element function. The genomic location of most SNPs significantly linked to somatic and psychological symptoms within the 3′UTR suggests their potential as enhancer elements regulating gene expression across diverse tissues. TrkB.T1 shares common exons with TrkB.FL up to exon 15 but has a distinct exon 16 encoding a unique intracellular 11 amino acid domain, a stop codon, and a 3′UTR sequence with multiple polyadenylation sites.²⁹ Importantly, the 3′UTR of TrkB.T1 differs from that of TrkB.FL and other gene isoforms, with combined epigenetic mechanisms impacting its regulation, including histone modifications (H3K4me1 and H3K27ac), DNA methylation, and microRNA binding.^64–66 Histone modification markers and transcription factors, in particular, play a critical role in establishing cell-type-specific gene expression patterns within active chromatin regions.^67,68 According to HaploReg v4.2,⁶⁹ rs2013566, rs1627784, and rs1624327 exhibit enhanced H3K4me1 and H3K27ac markers not only in various brain tissues but also in rectal mucosa-derived cells and fetal leg/truck muscle (Supplemental Table 5a–c). Despite the limited availability of epigenetic data in non-brain tissues, the presence of H3K4me1 and H3K27ac markers near these SNPs suggests that they may be located in close proximity to enhancer elements, potentially influencing the regulation of TrkB.T1 gene expression in gastrointestinal and muscle tissues. Among these, rs2013566 ranks 4 in RegulomeDB, ranging from 1 (most likely functional) to 6 (least likely functional),⁷⁰ further supporting its potential enhancer function and suggesting a probable influence on transcription factor binding (Supplemental Table 5e). Since none of these SNPs are considered expression quantitative trait loci (eQTLs) in any tissues available in a public database (GTExPortal, https://www.gtexportal.org), their potential of functional activity in the TrkB.T1 signaling is anticipated to be indirect rather than direct. The identification of rs2013566 as potentially displaying enhancer regulatory activity in the brain, muscle, and gastrointestinal tissues supports the functional relevance of its associations with a range of symptoms, especially headache and musculoskeletal pain, in individuals with IBS. Conversely, rs1627784 and rs1624327 show limited evidence of regulatory function, underscoring the need for further investigation to clarify their potential significance.

Beside our primary analysis, we uncovered a significant association between the number of minor alleles at rs41277883 and a 3.4-fold higher likelihood of having IBS. Notably, despite this association, this SNP isn’t directly tied to commonly observed somatic or psychological symptoms in IBS. However, rs41277883, located in the 3′UTR, overlaps with evolutionarily Aed elements found in the aforementioned SNPs, suggesting its potential role as an enhancer (Supplemental Table 5d). In a previous study, rs41277883 was associated with heightened pain intensity in individuals aged 65 years or older undergoing hip fracture surgery.⁷¹ This association with the increased likelihood of IBS suggests its potential involvement in mechanisms related to pain sensitivity, but further research is needed to fully understand its implications and uncover specific mechanisms influencing IBS development and manifestation.

The BDNF SNP rs6265, known for substituting valine (Val) with methionine (Met), hampers processes that regulate extracellular BDNF levels.⁷² Extensive studies associated rs6265 with pain and psychological traits in IBS, including migraines,⁷³ fibromyalgic pain,⁷⁴ menstrual pain,⁷⁵ anxiety,^76,77 and depression.^78,79 Regarding IBS, our earlier investigation showed decreased plasma BDNF levels in IBS patients compared to those without IBS.⁸⁰ Others also found an association between lower serum BDNF levels and higher levels of anxiety and depressive symptoms in individuals with IBS.⁸¹ Conversely, preclinical evidence suggests that higher levels of BDNF and TrkB proteins in the thoracolumbar spinal cord of IBS-like rats compared to controls, and administering a selective TrkB antagonist reduced their visceral hypersensitivity.⁸² Despite its association with various conditions, rs6265 is not directly associated with any of the symptom traits observed in our analysis. This suggests that TrkB signaling and TrkB.T1 receptor expression levels, rather than extracellular BDNF concentration, may have a more substantial impact on the somatic or psychological symptoms observed in IBS. Alternatively, the BDNF/TrkB response could differ among tissues, or symptoms may result from multiple interacting factors.

Our findings also suggest that genetic variations in the NTRK2 gene may impact a broad spectrum of somatic and psychological symptoms in individuals with IBS, influencing their quality of life. The variability in scores across the three symptom domains indicates differing levels of impairment among individuals with IBS, necessitating personalized approaches to symptom management and improving their quality of life. Our investigation reveals that individuals with a homozygous recessive genotype (G/G) at rs2013566 are more susceptible to headaches, somatic pains, sleep disturbances, and reduced energy levels. Such genetic association underscores the importance of understanding how the homozygous recessive genotype at rs2013566 is linked to various somatic pain manifestations, as it can contribute to sleep disturbances and hinder energy restoration, ultimately leading to a decline in overall quality of life. Having the genetic information at the outset of the diagnosis could help healthcare providers to be aware of the risk for the above IBS phenotypes and intervene early to potentially prevent them. However, further investigation is required to comprehensively interpret these associations and patterns.

Some aspects of our data require careful interpretation. Despite significantly higher rates of depression and anxiety reported in the IBS group compared to healthy controls (Table 2), we did not find any statistically significant associations with specific SNPs. This suggests that psychological symptoms in IBS are likely influenced by a complex network of multiple genetic factors interacting cumulatively, rather than by single or small groups of SNPs. Consequently, a broader genetic architecture, coupled with environmental factors such as stress, diet, lifestyle, socio-economic status, and comorbid conditions prevalent among IBS patients, might significantly contribute to these psychological symptoms. Moreover, the homozygous recessive genotype of rs2013566 was positively associated with back pain in our primary analysis but negatively associated in the UKBB validation. In addition, significant associations between SNPs and other somatic symptoms or quality of life, initially identified in our primary analysis, were non-significant in the UKBB validation. These discrepancies can be attributed to sample size differences between datasets. We did not observe statistically significant differences in the multiple test analysis due to the small sample size of our primary data. In contrast, the larger sample size of the UKBB provides greater statistical power, potentially leading to more robust associations. Population differences, including genetic heterogeneity and varying environmental exposures, also contribute to these discrepancies by significantly influencing the phenotypic expression of genetic variants. Variations in how somatic pain is defined and measured across studies, as well as differences in genotyping platforms and analytical methods, further contribute to these discrepancies. Therefore, these findings highlight the critical roles of genetic diversity, environmental variability, and statistical power in understanding the genetic basis of IBS symptoms.

This study has limitations. First, our focus was specific to SNPs within the NTRK2 and BDNF loci, potentially overlooking other contributing SNPs in these genes or related ones. The absence of examining susceptibility alleles at multiple loci raises the possibility of unexplored SNPs contributing to IBS symptoms. Additionally, the failure to genotype certain SNPs in one substudy may introduce biases by missing crucial data. Future investigations should prioritize comprehensive genotyping for accurate and thorough interpretation of genetic associations in those with IBS. Second, the study addressed potential bias arising from disparities in marital status and family income between the IBS and HC groups. These demographic factors could confound the study results by impacting the experience and management of IBS-related health conditions, as well as epigenetic signatures. To mitigate the impact of potential bias, interventions, and Hawthorne effects, only baseline data from all five studies were analyzed, and all regression models were adjusted for age, sex, and study design. Third, while environmental and lifestyle factors were not the primary focus, they should be considered in future studies. Finally, participant recruitment from a specific U.S. geographical area, consisting primarily of individuals from a non-Hispanic White background, and validation using U.K. data limit generalizability. This urges caution in applying these findings to diverse populations, and subsequent research should include a broader demographic range.

In summary, our study identified a genetic association between TrkB.T1 SNPs, particularly rs2013566, and symptoms such as headache, musculoskeletal pain, and impaired quality of life in individuals with IBS. These SNPs demonstrated regulatory potential concerning somatic and psychological symptoms, indicating possibilities for personalized treatment and targeted symptom management based on genetic markers. To comprehensively understand these genotypic associations, further research, including large-scale longitudinal observational studies, clinical trials, and exploration of underlying neurogenetic mechanisms, is necessary.

Perspective:

This study aims to understand the genetic effects on IBS-related symptoms across somatic, psychological, and quality of life domains, validated by UKBB data. The rs2013566 homozygous recessive genotype correlates with worsened somatic symptoms and reduced quality of life, emphasizing its clinical significance.

Highlights.

• IBS involves altered gut-brain interactions, potentially regulated by neurobiological pathways.

• NTRK2 (TrkB.T1) variants associated with worse symptoms and lower quality of life in IBS.

• BDNF/TrkB genetic effects on IBS symptoms hold clinical significance, needing further exploration.

Data availability

The original datasets and code generated during this research are available from the authors upon request. Requests for access can be directed to the corresponding author.