Pilot study of small bowel mucosal gene expression in patients with irritable bowel syndrome with diarrhea

We measured mRNA expression by RT-PCR of 91 genes reflecting tight junction proteins, chemokines, innate immunity, ion channels, and transmitters in small intestinal mucosa from 15 IBS-D and 7 controls (biopsies negative for celiac disease). The following genes were significantly upregulated (q < 0.05) in irritable bowel syndrome with diarrhea (IBS-D): INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15. Protein expression of PPP2R5C in nuclear lysates was greater in patients with IBS-D compared with controls. Intestinal mucosal function deserves further study in IBS-D.

Abstract

Prior studies in with irritable bowel syndrome with diarrhea (IBS-D) patients showed immune activation, secretion, and barrier dysfunction in jejunal or colorectal mucosa. We measured mRNA expression by RT-PCR of 91 genes reflecting tight junction proteins, chemokines, innate immunity, ion channels, transmitters, housekeeping genes, and controls for DNA contamination and PCR efficiency in small intestinal mucosa from 15 IBS-D and 7 controls (biopsies negative for celiac disease). Fold change was calculated using 2^(−ΔΔCT) formula. Nominal P values (P < 0.05) were interpreted with false detection rate (FDR) correction (q value). Cluster analysis with Lens for Enrichment and Network Studies (LENS) explored connectivity of mechanisms. Upregulated genes (uncorrected P < 0.05) were related to ion transport (INADL, MAGI1, and SONS1), barrier (TJP1, 2, and 3 and CLDN) or immune functions (TLR3, IL15, and MAPKAPK5), or histamine metabolism (HNMT); downregulated genes were related to immune function (IL-1β, TGF-β1, and CCL20) or antigen detection (TLR1 and 8). The following genes were significantly upregulated (q < 0.05) in IBS-D: INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15. Among the 14 nominally upregulated genes, there was clustering of barrier and PDZ domains (TJP1, TJP2, TJP3, CLDN4, INADL, and MAGI1) and clustering of downregulated genes (CCL20, TLR1, IL1B, and TLR8). Protein expression of PPP2R5C in nuclear lysates was greater in patients with IBS-D and controls. There was increase in INADL protein (median 9.4 ng/ml) in patients with IBS-D relative to controls (median 5.8 ng/ml, P > 0.05). In conclusion, altered transcriptome (and to lesser extent protein) expression of ion transport, barrier, immune, and mast cell mechanisms in small bowel may reflect different alterations in function and deserves further study in IBS-D.

NEW & NOTEWORTHY

We measured mRNA expression by RT-PCR of 91 genes reflecting tight junction proteins, chemokines, innate immunity, ion channels, and transmitters in small intestinal mucosa from 15 IBS-D and 7 controls (biopsies negative for celiac disease). The following genes were significantly upregulated (q < 0.05) in irritable bowel syndrome with diarrhea (IBS-D): INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15. Protein expression of PPP2R5C in nuclear lysates was greater in patients with IBS-D compared with controls. Intestinal mucosal function deserves further study in IBS-D.

the causes of loose bowel movements in patients with irritable bowel syndrome with diarrhea (IBS-D) include organ level alterations of function such as acceleration of colonic transit, documented in ∼45% of patients with IBS-D (13), and intestinal secretory mechanisms (reviewed in Ref. 9). The latter include increased duodenal and rectosigmoid expression of secretory transmitters (e.g., 5-HT) measured by morphological studies, reduced expression of the serotonin reuptake protein, and fecal excretion of secretogranins or chromogranins (18, 19, 38). There was also evidence of reduced expression of proabsorption mechanisms (e.g., mucosal PYY, somatostatin, and neuropeptide Y).

IBS has also been associated with changes in rectosigmoid mucosal mRNA expression of immune factors, barrier function, and mucus secretion (2, 5, 6, 10, 12, 14, 26, 43, 46, 48, 50, 51, 53). The upregulated mechanisms measured initially by RNA sequencing in 9 patients with IBS-D were associated with changes in ion transport and included PDZD3 and GUCA2B (12); these results were largely confirmed in a replication study using RT-PCR in 47 patients with IBS-D (11). PDZ adapter proteins are involved in multiple ion transport functions in the intestine, and GUCA2B is the gene controlling the endogenous uroguanylin that induces chloride secretion. We also demonstrated [with false detection rate correction (FDR)] abnormal expression of mRNA in rectosigmoid mucosal biopsies of receptors or neurotransmitters (P2RY4 and VIP), cytokines and complement components (C4BPA and CCL20), immune function (e.g., TNFSF15), and mucosal repair and cell adhesion (TFF1 and FN1).

Differences in jejunal mucosal expression (at gene and protein levels) and in the distribution of apical junction complex proteins (34, 35), as well as reduced ZO-1 expression in HLA DQ2/8-positive patients with non-celiac IBS-D (46), support the observed alterations in barrier mechanisms seen in colonic mucosa in patients with IBS-D.

In this study, our aim was to compare, in small intestinal mucosa of patients with IBS-D and controls with normal small intestinal mucosa, the expression of a larger complement of genes potentially associated with the pathobiology of IBS-D, including tight junction proteins, chemokines, innate immunity, ion channels, and transmitters.

METHODS

Ethical approval.

The study was approved by Mayo Clinic Institutional Review Board on June 3, 2014 (IRB No. 14-002151), and an amendment was approved on October 28, 2015. Written informed consent was received from participants before inclusion in the study.

Study design.

We appraised bowel functions and descending duodenal mucosal mRNA expression in 15 patients with IBS-D (by Rome III criteria) and 7 controls who had undergone clinically indicated duodenal biopsies (predominantly to exclude celiac disease in patients with iron deficiency), which were morphologically normal.

Patient selection.

Patients were recruited by public advertisement or by invitation to participate from a database of ∼1,200 patients with IBS living within ∼120 miles of the Mayo Clinic in Rochester, MN. Inclusion criteria were based on symptoms using validated diary questionnaire that characterized IBS symptoms and particularly bowel functions (44). Participants also completed the Hospital Anxiety and Depression Inventory (52). These patients had been evaluated at Mayo Clinic, and alternative diagnoses such as inflammatory bowel disease, cancer, and celiac disease were excluded. The main exclusion criterium was intake of medications that could cause bleeding (owing to the need for duodenal biopsies). During a 14-day period (±4 days), the patients with known IBS-D completed a daily bowel pattern diary [that included the 7-point Bristol Stool Form Scale (30), ranging from 1 or hard lumpy stool to 7 or watery diarrhea] to record their bowel habits; they also recorded daily their average abdominal pain severity and worst pain severity using a 100-mm visual analog scale. Clinical details of the seven controls who underwent clinically indicated duodenal biopsies are shown in results (see Table 2).

Table 2.

Clinical features and diagnosis

A. Clinical Features of Patients with IBS-D (at baseline)
Daily Score	Mean (SE)
#Stools/day	2.35 (0.203)
Stool form (Bristol Stool Form Scale, 1–7: 1 = hard lumps; 7 = watery)	5.13 (0.146)
Stool ease of passage (based on scale 1–7: 1 = manual disimpaction; 7 = incontinence)	4.70 (0.095)
Proportion incomplete evacuation (0 = no; 1 = yes)	0.37 (0.070)
Average pain severity (pain scales 0–100 mm, visual analog scale: 1 = none; 100 = worst ever)	18.88 (4.425)
Worst pain severity (pain scales 0–100 mm, visual analog scale: 1 = none; 100 = worst ever)	24.38 (4.561)
Fasting serum FGF-19, pg/ml	115 (17)
Fasting serum 7αC4, ng/ml	33.7 (6.9)

B. Controls Undergoing Duodenal Biopsy for Clinical Diagnosis (All Negative for Celiac Disease)
Control Patients	Age yr	Gender	BMI, kg/m2	Symptoms
1	24	M	21.8	Diarrhea (loose stools)
2	18	F	23.9	Dizziness, postive anti-gliadin IgG and IgA antibodies; no gastrointestinal symptoms
3	67	F	19.3	Diarrhea, weight loss, epigastric pain/discomfort
4	26	F	27.5	Diarrhea (48-h stool weight 800 g), abdominal pain; Past medical illness: rheumatoid arthritis
5	49	F	15.9	Diarrhea, weight loss, iron deficiency anemia
6	54	M	20.8	Vomiting, diarrhea, colon polyps, weight loss, chronic nausea
7	47	F	15.2	Diarrhea, bloating, burping, nausea, weight loss

Normal fasting serum 7α C4 >47.1 ng/ml and serum FGF-19 <131pg/ml, based on 90th percentile of healthy volunteers data. IBS-D, irritable bowl syndrome with diarrhea; BMI, body mass index; M, male; F, female.

Stored biospecimens.

The biopsies from patients with IBS-D were obtained at baseline (before treatment) in a study of the effects of therapy with serum bovine immunoglobulin (results to be published elsewhere).

We used stored samples from seven patients who had consented to the use of biospecimens (for future research) in prior studies (Principal Investigator: Joseph A. Murray) conducted at Mayo Clinic in Rochester, MN. These samples were from patients who underwent biopsies for clinical indications such as iron deficiency, with no evidence of small intestinal abnormality on biopsy. We chose to study “disease controls,” most of whom had diarrhea that was not caused by celiac disease and did not fulfill criteria for IBS-D, to control for the presence of diarrhea. In this pilot study, we considered it was important to identify differences in expression that were not caused by the diarrhea per se.

Biopsies were preserved in a solution of RNAlater and stored at −80°C until the time of assay of gene expression.

Selection of genes of interest.

We developed a custom profile including 89 genes (reflecting tight junction proteins, chemokines, innate immunity, ion channels, and transmitters; Table 1), extending from the previous 19 gene profile (12) as well as housekeeping (HKG) genes for normalization (B2M, ACTB, and GAPDH). Subsequently, we included SCN9A (NHE3) and CFTR expression in view of their known roles in the absorption of Na⁺ and secretion of Cl⁻ ions, respectively. In addition, controls monitored DNA contamination, and first strand synthesis and PCR efficiency were included to check for sample quality and reaction quality.

Table 1.

Genes of interest included in the RT-PCR analysis of duodenal mucosa

Gene Symbol	Refseq No.	Official Full Name
ACTB	NM_001101	Actin, beta (house-keeping gene [HKG])
C4BPA	NM_000715	Complement component 4 binding protein, alpha
CCL20	NM_004591	Chemokine (C-C motif) ligand 20
CLDN1	NM_021101	Claudin 1
FGFR4	NM_002011	Fibroblast growth factor receptor 4
FN1	NM_002026	Fibronectin 1
GAPDH	NM_002046	Glyceraldehyde-3-phosphate dehydrogenase (HKG)
GPBAR1	NM_170699	G protein-coupled bile acid receptor 1
GUCA2B	NM_007102	Guanylate cyclase activator 2B (uroguanylin)
HGDC	SA_00105	Human genomic DNA contamination
IFIT3	NM_001549	Interferon-induced protein with tetratricopeptide repeats 3
NR1H4	NM_005123	Nuclear receptor subfamily 1, group H, member 4
OCLN	NM_002538	Occludin
P2RY4	NM_002565	Pyrimidinergic receptor P2Y, G-protein coupled, 4
PDZD3	NM_024791	PDZ domain containing 3
PPC	SA_00103	Positive PCR control
RBP2	NM_004164	Retinol binding protein 2, cellular
RTC	SA_00104	Reverse transcription control
SLC6A4	NM_001045	5-HT transporter
SLC10A2	NM_000452	Solute carrier family 10 member 2 (sodium/bile acid cotransporter)
TFF1	NM_003225	Trefoil factor 1
TJP1	NM_175610	Tight junction protein 1 (zona occludens 1)
TNFSF15	NM_005118	Tumor necrosis factor (ligand) superfamily, member 15
VIP	NM_003381	Vasoactive intestinal peptide
IFNG	NM_000619	Interferon-gamma
MYLK	NM_053025	Myosin light chain kinase
SLC9A1	NM_003047	Na+-H+-exchange protein
B2M	NM_004048	Beta 2 microglobulin (HKG)
CALR	NM_004343	CALReticulin
CD3E	NM_000733	CD3-epsilon chain
CD74	NM_004355	HLA-DR antigen-associated invariant chain
CLDN2	NM_020384	Claudin 2
CLDN3	NM_001306	Claudin 3
CLDN4	NM_001305	Claudin 4
CLDN7	NM_001307	Claudin 7
CLDN12	NM_012129	Claudin 12
CLDN15	NM_014343	Claudin 15
CLDN16	NM_006580	Claudin 16
CPSF1	NM_013291	Cleavage and polyadenylation specific factor 1
CTNNA1	NM_001903	Catenin (cadherin-associated protein), alpha 1
CTNNB1	NM_001904	Catenin (cadherin-asociated protein), beta1
DLG1	NM_004087	Discs, large homolog 1 (Drosophila)
FOXP3	NM_014009	Forkhead box p3
HAAO	NM_012205	Kydroxyanthranilic acid oxygenase (3-hydroxyanthranilate 3,4-dioxygenase)
HNMT	NM_006895	Histamine N-methyltransferase
HRH1	NM_000861	Histamine receptor 1
HRH2	NM_022304	Histamine receptor 2
IDO1	NM_002164	Indoleamine 2,3-dioxygenase
IDO2	NM_194294	Indoleamine 2,3-dioxygenase 2
IL1B	NM_000576	Interleukin-1beta
IL2RA	NM_000417	(CD25) Interleukin-2 receptor subunit alpha
IL4	NM_000589	Interleukin 4
IL6	NM_000600	Interleukin-6
IL8	NM_000584	Interleukin-8
IL10	NM_000572	Interleukin-10
IL13	NM_002188	Interleukin-13
IL15	NM_000585	Interleukin-15
IL17A	NM_002190	Interleukin-17A
INADL	NM_176877	InaD-like (Drosophila)
AADAT	NM_182662	Kynurenine aminotransferase 2 (alias KAT2)
CCBL2	NM_001008661	Kynurenine aminotransferase 3 (alias KAT3)
GOT2	NM_002080	Kynurenine aminotransferase 4 (alias KAT4)
KITLG	NM_003994	Kit-ligand, stem cell factor
KMO	NM_003679	Kynurenine 3-monooxygenase
KYNU	NM_003937	Kyureninase
MAGI1	NM_004742	Membrane-associated guanylate kinase, WW and PDZ domain containing 1
MPP5	NM_022474	Membrane protein, palmitoylated 5 (MAGUK p55 subfamily member 5)
MPP7	NM_173496	Membrane protein, palmitoylated 7 (MAGUK p55 subfamily member 7)
PPP1CB	NM_002709	Protein phosphatase 1, catalytic subunit, beta isozyme
PPP2R5C	NM_001161725	Protein phosphatase 2, regulatory subunit beta
PRG2	NM_002728	Major basic protein
PVRL3	NM_015480	Poliovirus signaling-related 3
QPRT	NM_014298	Quinolinic acid phosphoribosyltransferase
TDO2	NM_005651	Tryptophan 2,3-dioxygenase
TGFB1	NM_000660	Transforming growth factor beta
TJP2	NM_004817	Zona occludens 2
TJP3	NM_014428	Zona occludens 3
TLR1	NM_003263	Toll-like receptor 1
TLR2	NM_003264	Toll-like receptor 2
TLR3	NM_003265	Toll-like receptor 3
TLR4	NM_138554	Toll-like receptor 4
TLR5	NM_003268	Toll-like receptor 5
TLR6	NM_006068	Toll-like receptor 6
TLR7	NM_016562	Toll-like receptor 7
TLR8	NM_138636	Toll-like receptor 8
TLR9	NM_017442	Toll-like receptor 9
TNF	NM_000594	Tumor necrosis factor-alpha (TNF-a)
TNFSF14	NM_003807	LIGHT/tumor necrosis factor superfamily 14
TPH1	NM_004179	Tryptophan hydroxylase 1
TPSAB1	NM_003294	Tryptase
TPSB2	NM_024164	Tryptase beta 2 (gene/pseudogene)
AHR	NM_001621	Aryl hydrocarbon Rrceptor
SOS1	NM_005633	Son of Sevenless homolog 1 (Drosophila)
MAPKAPK5	NM_003668	Mitogen-activated protein kinase-activated protein kinase 5
MKNK2	NM_017572	MAP kinase interacting serine/threonine kinase 2
SLC9A3	NM_004174	Solute carrier family 9, subfamily A (NHE3, cation proton antiporter 3), member 3
CFTR	NM_000492	Cystic fibrosis transmembrane conductance regulator (ATP-binding cassette subfamily c, member 7)

Thus, the RT²-Profiler-PCR-Array data examined 96 genes on 22 samples (n = 15 IBS-D patients and 7 healthy controls). Out of the 96 genes, two were identical (CCBL2, used for quality control), three were the housekeeping genes (B2M, ACTB, and GAPDH), and another three (HGDC, RTC, and PPC) were assay controls. Therefore, the custom array analysis focused on 89 genes.

Gene expression method by RT² PCR array.

Mature RNA was isolated using the RNAeasy Extraction Kit (Qiagen) according to the manufacturer's instructions. All RNA integrity number values (measured using Agilent QC) were 9.5–10, confirming excellent RNA quality. RNA (500 ng) was reverse transcribed using a cDNA conversion kit, and cDNA in combination with RT² SYBR Green qPCR Mastermix (Qiagen) was used on a Custom RT² Profiler PCR Array. PCR was performed on a ViiA7 thermocycler (Applied Biosystems, Life Technologies, Grand Island, NY) and, finally, relative expression was determined using data from the real-time cycler and the ^ΔΔC_T method.

Individual RT² qPCR Primer Assays (Qiagen) for SCN9A (NHE3) and CFTR expression were used under the same conditions as the custom array plate (i.e., starting amount of RNA, cDNA synthesis, thermocycling conditions, and analysis software). The following assays were applied for the individual assays: SCN9A (PPH11615E), CFTR (PPH01387F), B2M (PPH01094E), ACTB (PPH00073G), and GAPDH (PPH00150F).

Thus a total of 91 genes of interest were evaluated in the mucosal biopsies.

To control for DNA contamination introduced during reaction setup, a no template control (NTC) reaction replacing template with water as well as a no reverse transcription (NRT) reaction were run with the assays.

The mRNA expressions were assayed in triplicate, and the mean value was used for the statistical analysis.

Data and statistical analyses.

Calculation of the threshold cycle (C_T) was determined for each well. Briefly, using the ViiA7 Software on the real-time machine, baseline was defined by choosing the automated baseline option, and threshold was defined manually using the log view of the amplification plots. The threshold value was chosen above the background signal, but within the lower one-third to lower one-half of the linear phase of the amplification plot. In our study, the threshold was chosen at 0.16 for all plates. C_T values for all wells were exported to Excel and analyzed through Qiagen's Data Analysis Center (http://www.qiagen.com/us/shop/genes-and-pathways/data-analysis-center-overview-page).

Delta C_T (ΔC_T) was calculated between gene of interest and an average of housekeeping genes (B2M, ACTB, and GAPDH), followed by delta delta CT (ΔΔC_T) calculations [ΔCT(IBS-D) − ΔCT(health)]. Fold change was calculated using 2^−ΔΔCT formula. The P values (P < 0.05) were reported based on t-test, and the q values were computed to reflect statistical significance with FDR correction (42).

Protein expression in duodenal mucosa.

Nuclear and cytosolic lysates were isolated from human colonic biopsies using a subcellular fractionation protocol supplied by Abcam. Briefly, tissue biopsies were homogenized in RIPA Lysis buffer with added inhibitors (Santa Cruz Biotechnology, Dallas, TX) and centrifuged at 720 G. The pelleted nuclear proteins were then sonicated in additional RIPA lysis buffer to resuspend. The nuclear pellet supernatant was recentrifuged at 10,000 g and the resulting supernatant contained the cytosolic and membrane fraction. Concentrations were determined using the Bradford reagent according to the instructions of the supplier (Sigma, St. Louis, MO).

PPP2R5C protein measurements by Western blots.

Nuclear lysates were separated using 4–15% Mini-PROTEAN TGX gels (Bio-Rad, Hercules, CA) and blotted onto nitrocellulose membranes. The membranes were blocked with 5% milk in phosphate buffered saline (PBS)/0.2% Tween, after which PPP2R5C primary antibody (ab94633_1:1,000; Abcam, Cambridge, MA) was applied overnight at 4°C. α-Actin (A2066_1:1,000; Sigma) was used for normalization of protein loading. Membranes were washed and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (donkey or goat anti-rabbit; Santa Cruz Biotechnology) and visualized with Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare Life Sciences, Pittsburgh, PA) and autoradiography. Band densities were quantified with the ImageJ (40).

Enzyme-linked immunosorbent assay.

All duodenal biopsy extracts were used at a common concentration (1 μg/μl). Fifty microliters were used to assay INADL, MAPKAPK5, and MAGI1, (MyBioSource, San Diego, CA) whereas 100 μl was used for the housekeeping gene phosphoglycerate kinase 1 (PGK1; LifeSpan Biosciences, Seattle, WA), which has been recommended as the preferred gene [over β-actin (ACTB) and β2-microglobulin (β2M)] in the context of inflammatory bowel diseases (29). Enzyme-linked immunosorbent assay (ELISA) kits were used for all these protein assays following the manufacturer's instructions. Samples were run in duplicate and spike recoveries were prepared in kit assay buffer at two concentration levels to verify performance.

Pathway and cluster analyses.

We used a publicly available informatics tool to appraise the potential pathways or clustered mechanisms that are altered in the duodenal mucosa of patients with IBS-D, relative to controls. The Lens for Enrichment and Network Studies (LENS) of proteins provides interactome analysis of the list of human genes examined in this study to explore the network of protein-protein interactions that connects them (23). LENS computes associations of the genes in the interactome to pathways and assesses statistics of network connectivity of these genes compared against connectivity of randomly selected genes. Sources of data for the network generation are direct (biophysical) interactions from HPRD and BioGRID, and pathways are obtained from Reactome (http://severus.dbmi.pitt.edu/LENS/index.php).

Significance analysis of microarray for gene sets analysis of gene expression and pathways.

Gene-set analysis evaluates the expression of biological pathways, or a priori defined gene sets rather than that of single genes, in association with a binary phenotype, and it is of great biologic interest in many DNA microarray studies. Gene Set Enrichment Analysis (GSEA) has been applied widely as a tool for gene set analyses, but it was not applicable for a small gene set in our study. We used an alternative method, more appropriate for targeted gene studies, by extending the single gene analysis method, Significance Analysis of Microarray (SAM), to Gene-Set Analyses (SAM-GS). Briefly, C_T data were converted to expression data as follows: expression = [2^(−Avg.(ΔC_T)], where Δ(C_T) = C_T (Gene Of Interest) − Ave.[C_T(Housekeeping Genes)].

For pathway investigation, we also included the gene set subcatalogs C1 and C2 from MolecularSignatures Database (MSigDB) v5.1, a collection of annotated gene sets for use with GSEA software (http://www.broad.mit.edu/gsea). The C1 catalog includes gene sets corresponding to human chromosomes and cytogenetic bands, while the C2 catalog includes gene sets that are involved in specific metabolic and signaling pathways collected from various sources such as online pathway databases, publications in PubMed, and knowledge of domain experts. SAM-GS divides the C1 and C2 catalogs into 308 pathways, which were included in our analyses, so that, in total, we had 150 pathways/gene sets based on the 89 genes under consideration (16, 17, 32).

RESULTS

Patients.

We studied 15 patients with IBS-D [Rome III positive: 2 male (M), 13 female (F); mean age 40.3 ± 2.3 yr; mean body mass index (BMI) 34.3 ± 3.0 kg/m²] and 7 controls (2M,5F; 40.7 ± 6.9 yr; P = ns) who were negative for celiac disease. Bowel function and abdominal pain scores of the patients with IBS-D are shown in Table 2A. Clinical details of the seven controls who underwent clinically-indicated duodenal biopsies are shown in Table 2B.

mRNA fold change in duodenal mucosa from IBS relative to healthy controls.

The mean (and 95% CI) fold changes (based on 2^{−ΔΔ CT}) in IBS-D and subgroups of IBS-D patients relative to healthy controls are illustrated in Table 3 and Fig. 1. Table 3 shows gene fold expressions (P < 0.05): upregulated [INADL and MAGI1 (possibly related to ion transport)] and barrier (TJP2 and 3) or immune functions (TLR3, IL15, and MAPKAPK5); or downregulated including factors involved in bile acid pathway (FGFR4), immune function (TGFB1 and CCL20), or antigen detection (TLR7).

Table 3.

Duodenal mucosal mRNA fold expression in patients with IBS-D and controls

			mRNA Regulation IBS-D/Controls
Gene Symbol	Official Name	Summary of Functions	Fold	95% CI	P*	q*
INADL	InaD-like (Drosophila)	Gene encoding multiple PDZ domain protein	1.45	1.13, 1.78	0.00291	0.0419
MAGI1	Membrane-associated guanylate kinase	Cell junction, interaction with ASIC and PDZ domain	1.48	1.05, 1.90	0.00607	0.0435
PPP2R5C	Protein phosphatase 2, regulatory subunit beta	Regulation of damage response, cell proliferation, apoptosis	1.24	1.06, 1.41	0.00529	0.0435
TJP1	Tight junction protein ZO-1	Barrier function	1.32	1.00, 1.64	0.0250	0.0673
TJP2	Tight junction protein ZO-2		1.29	1.02, 1.56	0.0325	0.0781
TJP3	Tight junction protein ZO-3		1.38	1.04, 1.72	0.0170	0.0632
CLDN4	Claudin 4		1.46	1.01, 1.92	0.0167	0.0632
MAPK-APK5	Mitogen-activated protein kinase-activated protein kinase 5	Kinase involved in growth factor stimulated cell proliferation and muscle cell differentiation	1.37	1.09, 1.65	0.0028	0.0419
TLR1	Toll-like receptor 1	Modulation of T-reg cells	−1.25	0.63, 0.98	0.0235	0.0673
TLR3	Toll-like receptor 3		1.58	1.13, 2.03	0.0029	0.0419
TLR8	Toll-like receptor 8		−1.90	0.18, 0.87	0.0357	0.0781
IL-1β	Interleukin-1β	Immune function	−1.65	0.18, 1.03	0.0240	0.0673
IL-15	Interleukin-15		1.67	1.18, 2.17	0.0049	0.0435
FGFR4	Fibroblast growth factor receptor 4	Tyrosine-protein kinase activity in cell proliferation, differentiation and migration	−1.33	0.53, 0.97	0.03477	0.0781
CCL20	Chemokine (C-C motif) ligand 20	Immunoregulatory and inflammatory processes	−2.07	0.00001, 1.14	0.01259	0.0632
TGFB1	Transforming growth factor-β1	Cytokine regulation of proliferation, differentiation, adhesion, migration, other functions in many cell types	−1.23	0.63, 1.00	0.0363	0.0781
TNFSF 14	Tumor necrosis factor ligand superfamily member 14	Apoptosis and T-cell proliferation/function	−1.45	0.45, 0.93	0.0156	0.0632
HAAO	3-Hydroxyanthranilate 3,4-dioxygenase	Synthesis of quinolinic acid, which participates in pathogenesis of neurologic and inflammatory disorders	1.50	0.96, 2.04	0.0224	0.0673
QPRT	Quinolinate phosphoribosyltransferase	Biosynthesis of NAD	1.88	0.93, 2.83	0.0176	0.0632
HNMT	Histamine N-methyl transferase	Metabolism of histamine	1.51	1.02, 2.00	0.0383	0.0784
SOS1	Sons of Sevenless (Drosophila) homolog 1	Guanine nucleotide exchange factor	1.33	1.01, 1.66	0.0102	0.0624
SLC9A3	Sodium hydrogen exchange type 3 channel	Sodium absorption	1.54	0.81, 2.28	0.0963	0.1525
CFTR	Cystic fibrosis transmembrane regulator	Chloride secretion	1.46	0.55, 2.37	0.3834	0.2917

ASIC, acid-sensing ion channel; PDZ, common structural domain of 80–90 amino-acids found in signaling proteins that anchor receptor proteins in the membrane to cytoskeletal components; + Numbers, upregulated; − numbers, downregulated.

^*P values are Student's t-test performed on IBS-D patients compared with controls before false detection rate (FDR) correction;

^*q values are FDR adjusted P values.

Fig. 1.

Fold changes in duodenal mucosa of patients with irritable bowel syndrome with diarrhea (IBS-D) compared with controls. mRNA fold expression of genes associated with different functions is shown (P < 0.05 without FDR correction). With false detection rate correction (FDR) correction, the following genes were significantly upregulated (q < 0.05): INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15 in IBS-D relative to controls.

With FDR correction, the following genes were significant (q < 0.05): INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15. All are upregulated in patients with IBS-D relative to controls.

There were other genes that showed numerical (nonsignificant) changes in expression (P > 0.05, < 0.1). Thus there was numerically increased expression in duodenal mucosa from IBS-D patients compared with controls for the following:

1) SLC10A2 (the gene for the bile acid transporter, 2.01-fold, P = 0.093);

2) CLDN1 (claudin, a tight junction protein,1.43-fold, P = 0.066);

3) KITLG (kit ligand, a determinant of development of interstitial cells of Cajal, 1.38-fold, P = 0.074);

4) MPP5 (membrane protein, palmitoylated 5, involved in polarization of differentiating cells, 1.35 fold, P = 0.070);

5) PVRL3 (poliovirus receptor-related 3, 1.22-fold, P = 0.075); and

6) AHR (aryl hydrocarbon receptor, which regulates xenobiotic-metabolizing enzymessuch as cytochrome P450, 1.22-fold, P = 0.053).

In contrast, the following were associated with numerically (not significantly) reduced expression in IBS-D duodenal mucosa compared with controls: CD3E (CD3-epsilon chain involved in T cell surface glycoprotein, −1.38-fold, 0.0.088); and TPSAB1 (tryptase alpha/beta 1, −1.61-fold, P = 0.097).

Clustering analysis of pathways and mechanisms.

Having identified the nominally, univariately significant upregulation or downregulation of genes in the mucosa, we applied LENS analysis, and plotted connectivity among the genes associated with barrier, T lymphocytes, and cytokines (Fig. 2), focusing only on the genes that were nominally (P < 0.05) upregulated or downregulated. Among the 14 nominally upregulated genes, clustering of the barrier-associated mechanisms and related PDZ domains (TJP1, TJP2, TJP3, CLDN4, INADL, and MAGI1) was noted; among the seven downregulated genes, a cluster of immune function (chemokine, Toll-like receptor, and interleukin), specifically CCL20, TLR1, IL1B, and TLR8, was observed. It is relevant to note that FDR-corrected increased expression was identified in INADL and MAGI1.

Fig. 2.

Lens for Enrichment and Network Studies (LENS) interactome of candidate genes evaluated in this study that showed increased or decreased expression. Note the connections between elements involved in ion transport (INADL and MAGI1), barrier function or epithelial integrity (e.g., MAPKAPK5, TJPs), and immune function (e.g., TLR-3 and IL15).

SAM-GS analysis.

The analysis results (Table 4) show the P value and q value of each gene set, based on the permutation test for no association between the gene expression of the gene set and the binary phenotype (IBS-D and control). The results are shown in order of the nominal P value.

Table 4.

SAM-GS results (only top 10 rows are shown)

Gene Set Name/Pathway	GS Size	GS P Value (0 ↔ <1/nb Permutations)	GS q Value	oGS Size	Genes Evaluated in Our Study Belonging to the Stated Pathway
SIG_CD40 PATHWAY MAP	1	0.02	0.386	34	MAPKAPK5
ST_p38_MAPK_Pathway	1	0.02	0.386	34	MAPKAPK5
FRASOR_ER_ DOWN	1	0.045	0.386	68	KYNU
HTERT_DOWN	1	0.045	0.386	64	KYNU
NFKB_ INDUCED	7	0.08	0.386	106	CCL20/FN1/IL15/IL1B/IL6/IL8/TNF
X41bbPathway	2	0.08	0.386	18	IFNG/IL4
MAP00380_ Tryptophan_ metabolism	3	0.085	0.386	42	HAAO/KYNU/TDO2
StemPathway	3	0.09	0.386	15	IL4/IL6/IL8
CR_IMMUNE_ FUNCTION	6	0.105	0.386	50	IFNG/IL10/IL13/IL15/IL1B/IL4
MAP00340_ Histidine_ metabolism	1	0.12	0.386	20	HNMT
89 genes in current study	89	0.185	0.386	89	AADAT/AHR/C4BPA/CALR/CCBL2/CCL20/ CD3E/CD74/CLDN1/CLDN12/CLDN15/ CLDN16/CLDN2/CLDN3/CLDN4/CLDN7/ CPSF1/CTNNA1/CTNNB1/DLG1/FGFR4/ FN1/FOXP3/GOT2/GPBAR1/GUCA2B/ HAAO/HNMT/HRH1/HRH2/IDO1/IDO2/ IFIT3/IFNG/IL10/IL13/IL15/IL17A/IL1B/ IL2RA/IL4/IL6/IL8/INADL/KITLG/KMO/ KYNU/MAGI1/MAPKAPK5/MKNK2/MPP5/ MPP7/MYLK/NR1H4/OCLN/P2RY4/ PDZD3/PPP1CB/PPP2R5C/PRG2/PVRL3/ QPRT/RBP2/SLC10A2/SLC6A4/SLC9A1/ SOS1/TDO2/TFF1/TGFB1/TJP1/TJP2/TJP3/ TLR1/TLR2/TLR3/TLR4/TLR5/TLR6/TLR7/ TLR8/TLR9/TNF/TNFSF14/TNFSF15/ TPH1/TPSAB1/TPSB2/VIP

SAM-GS, Significance Analysis of Microarray to Gene-Set Analyses; GS size, number of genes in the gene set/pathway; GS P value, permutation test for no association between the gene expression of the gene set and the binary phenotype (IBS-D compared with controls); GS q value, to account for multiple comparisons when multiple gene sets are tested, one might consider FDR instead of type I error probability; oGS, original number of genes in the pathway/gene set (for example, there were 106 genes in the NFKB_INDUCED pathway, where 7 genes from our study were found); Genes in Pathway, the genes in our study found in the pathway/gene set (for example, there are 7 genes from our study overlapping with NFKB_INDUCED pathway; the genes are CCL20, FN1, IL15, IL1B, IL6, IL8, and TNF).

The P value (P = 0.185) of the test evaluating the entire 89 gene set shows no association between the gene expression of the gene set and phenotype.

However, there were several genes with univariate significance (see q values in Table 3) that were represented in pathways identified in the SAM-GS analysis, specifically in immune function (IL-15) and MAP kinase-activated protein kinase (MAPAPK5). These genes, found in gene-sets by SAM-GS, showed a trend (P values between 0.02 and 0.105) but not a statistically significant association to the phenotype.

Mucosal protein expression.

Western blot analysis of duodenal mucosa from six healthy controls and seven patients with IBS-D who randomly selected from the same patient group was used to assess protein expression of PPP2R5C. This protein was selected as it showed increased mRNA fold expression (see Table 3). Figure 3 shows the increased protein expression including the appearance of the expression of PPP2R5C relative to the housekeeping protein, actin. The difference in expression between IBS-D group and biopsy-negative controls was statistically significant (P = 0.044).

Fig. 3.

Western blot analysis of protein expression of PPP2R5C in nuclear lysates from duodenal mucosal biopsies of patients with IBS-D and controls. Note the increase in protein (corrected for the housekeeping protein actin) in the patients with IBS-D.

Expression of other mucosal proteins (INADL, MAGI1, and MAPKAPK5) using ELISA measurement was applied to duodenal biopsy extracts from 15 patients with IBS-D and 6 controls (Table 5). There were very similar concentrations of the housekeeping gene PGK1 in the patients and controls. Numerical differences in protein expression of one of those tested, INADL, did not reach statistical significance: median 9.4 ng/ml in patients with IBS-D relative to controls, median 5.8 ng/ml (P > 0.05).

Table 5.

Measurements of mucosal proteins by ELISA

Protein	IBS-D (n = 15)	Control Patients (n = 6)	Estimated Sample Size per Group to Observe 2.5-Fold Difference in Protein Concentration, Based on Pooled SD of Each Protein
INADL	9.39 (2.68, 25.74)	5.83 (2.9, 23.2)	30
MAGI1	0.83 (0.00, 2.75)	0.79 (0.00, 2.26)	13
MAPKAPK5	0.54 (0.04, 1.96)	0.73 (0.00, 1.60)	15
PGK1	4.17 (3.71, 4.53)	4.38 (3.69, 4.83)	Housekeeping gene

Data are median and interquartile range (IQR); all comparisons are not statistically significant; PGK1 is the housekeeping gene. There was insufficient mucosa for protein measurements in 1 control patient. SD, standard deviation.

DISCUSSION

Our study evaluated expression in small intestinal mucosa of genes related to tight junctions, chemokines, innate immunity, ion channels, and transmitters in 15 patients with IBS-D and 7 controls with proven normal biopsies based on histopathology. With FDR correction, the following genes were significant (q < 0.05): INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15, all of which were upregulated in IBS-D relative to controls. The fold expression alterations in mechanisms of ion transport (INADL and MAGI1), epithelial response to damage (PPP2R5C), and immune mechanisms (TLR3 and IL15) in small bowel mucosa suggest that differential expression of small intestinal mucosa in IBS-D may be important in the biological mechanisms in IBS-D and certainly deserve further study. The findings on LENS cluster analysis of barrier and ion transport (both upregulated) and immune function (downregulated) are summarized in Fig. 4. Although the q values in the pathway analysis using SAM-GS do not show significant associations of pathways with phenotype, it is relevant to note that the MAPKAPK5 pathways and CR-immune function pathways show interesting trends (not statistically significant) and include upregulated genes in the comparisons between IBS-D and controls (specifically MAPKAPK5 and IL15).

Fig. 4.

Cartoon summarizing altered mRNA expression in cellular elements of the descending duodenal mucosa in patients with IBS-D, based on nominal significance (P < 0.05) or statistical significance (q < 0.05: INADL, MAGI1, PPP2R5C, MAPKAPK5, TLR3, and IL-15) and including TPSBA1 (P < 0.10). TJ, tight junction; GC-C, guanylate cyclase; CFTR, cystic fibrosis transmembrane regulator. Fold upregulated mRNA is in black; downregulated is in gray.

The upregulated or downregulated gene expressions suggest, in some cases, that the changes may represent reciprocal or compensatory changes within the same biological functions. One example of reciprocal changes in expression is downregulation of TLR1, TLR8, and CD3E in contrast to the upregulation of TLR3; another example where the changes may simply reflect reciprocal changes in expression is downregulated IL-1B and upregulated IL-15. On the other hand, there is consistent upregulation of barrier function gene expression, as shown by the results for ZO-1, 2, and 3, claudin-1 and -4, and MAPKAPK5. ZO and claudin genes are directly related to intercellular junction proteins, whereas MAPKAPK5 controls a kinase involved in growth factor-stimulated cell proliferation and muscle cell differentiation. IL-15 plays a key role in intraepithelial lymphocyte function and in some forms of intestinal inflammation (1, 33).

The biological significance of the upregulation of barrier function is unclear. There are other changes in expression that may conceivably be associated with increased fluid or electrolyte secretion, such as the upregulation of the following:

1) InaD-like (Drosophila) gene, for which there is a human analog INADL gene producing a protein (39), which encodes multiple PDZ domain proteins that may enhance ion transport, and a related tight junction-associated multi-PDZ protein PATJ (PALS1-associated TJ protein) that connects and stabilizes apical and lateral components of tight junctions in human intestinal cells (36);

2) Membrane-associated guanylate kinase MAGI1, which influences cell junction, interaction with acid-sensing ion channel (sensation) and PDZ domain (ion transport);

3) SONS1 [Sons of Sevenless (Drosophila) homolog] 1, which is a guanine nucleotide exchange factor; it is worth noting that guanine nucleotide exchange factor, Epac (Exchange protein directly activated by cAMP), appears to be involved in intestinal secretion (22, 25, 41). However, there is no evidence in the literature that this gene has functional effects in humans.

4) SLC9A3 [Solute Carrier Family 9, Subfamily A, Member 3), which encodes for the NHE3 channel is possibly upregulated (fold change 1.54, q = 0.153). If the increased expression results in increased protein, it would be expected that there would be greater Na⁺ absorption which may conceivably reduce fluid secretion in contrast to some of the other secretory mechanisms that may be upregulated.

Overall, our current studies support upregulation of intestinal ion transport mechanisms. This is consistent with increased mRNA expression of PDZ-related mechanisms previously demonstrated in colonic mucosa (11, 12). These data suggest increased ion transport, possibly leading to water and electrolyte secretion, that is partly counteracted by the increased expression of barrier function mechanisms. These observations emphasize the importance of thorough characterization of the phenotype in patients with IBS-D, given the different peripheral mechanisms that result in the phenotype of IBS-D (8, 9). Future studies will be enhanced by further characterization of the baseline pathophysiological phenotype in patients with the symptom phenotype of IBS-D, such as bile acid diarrhea, rapid small bowel or colonic transit, and upregulated mucosal immune mechanisms.

In addition to the changes in mRNA expression, we also observed significantly increased expression of PPP2R5C as demonstrated by Western blot, and a numerical increase in the protein level of INADL by ELISA, although this did not reach statistical significance. The concentrations of the housekeeping gene in the duodenal biopsy lysates used from the two groups were similar in the ELISA assays, suggesting that the absence of statistically significant differences in the protein concentrations is not the result of technical error in the measurements. On the other hand, the pooled standard deviation of INADL in the two groups was large (19.7 pg/ml), and the estimated sample size to prove statistical difference for 2.5- or 1.5-fold differences in concentration of INADL relative to control (median 5.83 pg/ml) would be respectively 30 and 81 per patient group, respectively. Therefore, our study was underpowered to appraise differences in the proteins measured by ELISA.

The intestinal mucosa contains a pool of histamine that rapidly turns over and that appears to be derived largely from a non-mast cell source. Histamine inhibits chloride absorption, inducing bicarbonate secretion in ileal mucosa and influencing chloride/bicarbonate exchange, largely by actions mediated by histamine H₁ receptors (31). Chromogranin A (ChgA), an acidic protein found in large dense-core secretory vesicles, functions as a monoamine-binding protein that facilitates the regulated endocrine secretion of large amounts of monoamines from enteroendocrine cells (20). The increased gene expression of the histamine metabolizing enzyme, HNMT, may also serve as a compensatory mechanism to reduce the effects of histamine released from enteroendocrine cells or from mast cells whose density is increased in the jejunal mucosa of non-atopic IBS-D patients (47).

The latter investigation by Vicario et al. (47) identified a number of immune-modulating mechanisms, suggesting increased humoral immune function in the jejunal mucosa of patients with IBS-D. This is consistent with several alterations in gene expression for immune regulation and T-cell function in our current study. In addition, the increased expression for the enzyme metabolizing histamine is consistent with increased release or supernantant content of histamine from colorectal mucosal biopsies from patients with IBS (3, 4, 7, 15, 28). This overexpression of HNMT may conceivably provide a natural metabolism of histamine, which may be produced and secreted in larger quantities by the increased numbers of mast cells. This natural reduction in histamine may serve as an alternative to achieve the beneficial effects on symptoms of the mast cell stabilizer and antihistaminic medication ketotifen (28) and the selective H1 receptor antagonist ebastine (45). The numerically increased gene expression for Kit ligand is of significant interest in view of the role of the gene in the function of this tyrosine kinase receptor gene in responding to growth factors such as insulin-like growth factor-1 (IGF-1) in the development of interstitial cells of Cajal, the pacemakers of the gastrointestinal tract.

The numerical downregulation of TPSAB1 (P = 0.097 and q = 0.144, which controls the synthesis of alpha and beta tryptases) and KYNU (P = 0.131 and q = 0.166 kynureninase, which metabolizes tryptophan) is similarly intriguing in the context of IBS. Beta tryptases appear to be the main isoenzymes expressed in mast cells; whereas, in basophils, alpha tryptases predominate. Tryptases are one of the main products of mast cells. Our data suggest the hypothesis that there is downregulation in the synthesis of tryptase, in contrast to several prior publications that have documented increased expression of tryptase in mucosa from patients with IBS, specifically in colonic mucosa (3, 4, 7, 15, 49) and in mRNA (21), as well as protein expression (35) in jejunal mucosal biopsies of IBS patients. The mechanism resulting in this hypothetical reduction in tryptases is the observed numerical downregulation of TPSAB1; however, it is important to confirm these observations given the other data showing increased tryptase and other proteases in supernatants from (particularly) colonic biopsies from patients with IBS-D.

Kynureninase metabolizes tryptophan, and the synthesis of NAD (or NADP) and quinolic acid from tryptophan involves a series of enzymes and the formation of a number of intermediates which are collectively called “kynurenines.” Kynurenic acid is a neuroprotective N-methyl d-aspartate (NMDA) antagonist (e.g., on enteric glutamatergic neurons) and an α₇-nicotinic cholinergic agonist, potentially dysregulating intestinal motility and altering mechanosensitivity (27). In mononuclear cells, including tissue macrophages, quinolinic acid is the main end product of the kynurenine pathway and plays a role in immunoregulatory processes (37). Recent evidence shows kyruneninase plays a role in inflammation (24). Reduced expression of KYNU may conceivably result in greater motor dysregulation and inflammation in IBS-D due to enhanced tryptophan metabolism through the kynurenine pathway. It is also interesting to note that there were other upregulated genes in the quinolinate pathway (HAAO and QPRT).

In summary, our results extend the prior observations recorded in jejunal mucosa (35), confirm the potential upregulation of ion transport processes observed in colonic mucosa (11, 12), and identify potential compensatory changes in immune and barrier functions in small intestinal mucosa. However, our current studies should be regarded as pilot data that require replication in larger numbers, as well as control for BMI, enriching the current study sample with patients with a group of IBS -D patients with normal BMI, since there is evidence that obesity is associated with gastrointestinal symptoms, an inflammatory diathesis in general, as well as possible changes in intestinal permeability or bile acid malabsorption. Neverthelss, this pilot study identifies mechanisms that deserve further study and identifies the fold expression and coefficient of variation that will allow planning with adequate sample sizes in future studies. In fact, our planned hypothesis-testing study will include 60 patients with IBS-D, 30 with IBS-C, and 30 healthy controls. In addition, given the wide variations in protein expression observed, it appears that biological significance of these different mechanisms expressed in intestinal mucosa from patients with IBS-D may be interrogated in future studies using enteroids. These have the potential to provide new insights on the pathobiological mechanisms, allowing different mechanisms and pathways can be interrogated in greater depth. The observations in the current study provide a step to further understand potential etiopathogenetic mediators as well as biological pathways involved in IBS-D.

GRANTS

M. Camilleri is supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK92179; this study (and permission to use data from duodenal biopsies in patients) was supported in part by research funds from EnteraHealth for an ancillary study.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.C. conception and design of research; M.C., J.N., and E.W.K. analyzed data; M.C. interpreted results of experiments; M.C. drafted manuscript; M.C., P.C., N.V., A.A., J.O., R.D., J.N., E.W.K., and J.A.M. edited and revised manuscript; M.C. approved final version of manuscript; P.C., N.V., A.A., J.O., D.J.E., R.D., and J.A.M. performed experiments.