Identi cation of Potential Biomarkers for Abdominal Pain in IBS Patients by Bioinformatics Approach

Zhongyuan Lin Sun Yat-sen University First A liated Hospital Yimin Wang Guangdong Second Provincial General Hospital Shiqing Lin Sun Yat-sen University First A liated Hospital Decheng Liu Sun Yat-sen University First A liated Hospital Guohui Mo Sun Yat-sen University First A liated Hospital Hui Zhang Guangdong Second Provincial General Hospital Yunling dou (  sysdyl@163.com ) Sun Yat-sen University First A liated Hospital


Background
Chronic abdominal pain is an important symptom of irritable bowel syndrome (IBS), which affect over 15% of the global population [1]. Recurrent abdominal pain can substantially reduce quality of life in IBS patients and there is no effective and standard treatment currently [1]. In order to identify new therapies, it is important to understand the underlying mechanism of chronic abdominal pain.
Abdominal pain in IBS is thought to be secondary to visceral hypersensitivity (VH), which has been described as exaggerated awareness of gastrointestinal balloon distension [2,3]. The underlying pathogenesis that lead to VH in patients who have IBS is not fully understood. However, the prevailing viewpoint in the pathogenesis involves psychosocial factors, subtle in ammation and alteration in the neuronal excitability of gut sensory pathways, which is responsible for transmitting nociceptive information from the periphery to the central nervous system [4]. Speci c changes of a variety of receptors and ion channels both in the expression and function have been documented in visceral pain conditions relevant to irritable bowel syndrome. Researches have shown genetic factors may play a role in those speci c changes in IBS [5]. Beyder and colleagues have reported that a mutation of the SCN5Aencoded voltage gated sodium channel, type V (alpha subunit) was associated with abdominal pain [6].Other investigators have investigated genetic changes related to enteric infection, epithelial barrier function and immune regulation, serotonin signaling, cannabinoid receptors, and bile acid synthesis, with varying results [7].
Bioinformatics analysis provides tools to produce overviews of genetic networks and potential biological pathways based on large-scale information. This study aimed to explore the molecular mechanisms specially involved in pain transmission underlying IBS and to identify hub genes and potential pathways associated with the pathogenesis of IBS.

Microarray Data Search and Selection of Eligible Data Set
The microarray data of GSE36701 was downloaded from the GEO database (www.ncbi.nlm.nih.gov/geo/) with its microarray platform as GPL570 (Affymetrix Human Genome U133 Plus 2.0 Array). IBS-D patients were studied at baseline, or 6 months after natural evolution (control) or oral cromoglycate treatment. The gene expression les were uploaded by Swan C, Duroudier NP et al. In this microarray,53 samples of rectal mucosa from IBS-D patient with obvious abdominal pain were obtained and 40 samples from healthy volunteer as controls. Bioinformatics analysis was performed to identify the gene expression pattern.

DEGs' Screening
We screened out DEGs using the online tool GEO2R/R package limma. In our study, DEGs between healthy volunteers and IBS-D patients were screened and selected by the cut-off point of FC ≥1.2 and Pvalue <0.05.

Functional Enrichment Analysis
The selected DEGs were submitted to the Database for Annotation, Visualization, and Integrated Discovery (DAVID) for further analysis. The DAVID database, version 6.8 Beta (https://davidd.ncifcrf.gov/), offers biological function annotation for researchers [8]. In this study, Gene Ontology (GO) analysis and The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways of DEGs were obtained from the DAVID database. The biological processes and pathways might contribute to verify the signi cative DEGs and to IBS treatments. P-value <0.05 was chosen as the threshold.

Protein-Protein Interaction (PPI) Analysis
Then we constructed PPI network of DEGs from Search Tool for the Retrieval of Interacting Genes (STRING, http://string-db.org/) [9]. STRING is a public database of known and predicted protein-protein interactions. In this study, we selected con dence score >0.4 as a threshold to construct the PPI network.
Then, the list of PPI pairs was downloaded for further analysis. Module clustering analysis for the network was then performed to identify the potential functional modules in the network, using the Molecular Complex Detection (MCODE ) in Cytoscape 3.7.1 plugin [10,11]. The degree cut-off value to 2 and the node score cut-off to 0.2 were set as criterions in the MCODE process.

Screening and prioritization for DEGs
An available Expression pro ling data set (GSE36701) in NCBI was selected in this study to analyze the association between Irritable Bowel Syndrome with diarrhea (IBS-D) patients and the healthy volunteers (HVs), especially the functional associations related to neurotransmitters/mediators of pain. Each array was normalized by quantile, and then the DEGs analysis was performed (P-value < 0.05, FC ≥ 1.2 or ≤ 0.8). Compared with the HVs, we nd out 9 upregulated genes (CCKBR, CCL13, ACPP, BDKRB2, GRPR, SLC1A2, NPFF, P2RX4, and TRPA1) and 4 downregulated genes (TLX2, MRGPRX3, CXCR1 and PAX2) that may have relevance in the initiation and development of visceral pain. The most signi cant upregulated or downregulated genes specially linked to pain transmission are illustrated in Table 1 and Table 2.  Functional analysis was carried out with GO biological process and KEGG pathways. Genes differentially were classi ed according to their putative Gene Ontology (GO) based on similarity with known genes recorded in public databases (Fig. 1). To further understand the functions of the DEGs, we represented signi cant pathways based on KEGG databases (Fig. 2). The upregulated DEGs were mainly associated with Glutamatergic synapse, Dilated cardiomyopathy, Arginine biosynthesis. The downregulated DEGs were linked to Circadian entrainment, Alcoholism, Aldosterone-regulated sodium reabsorption, Maturity onset diabetes of the young, cGMP-PKG signaling pathway, Dopaminergic synapse, Morphine addiction, Systemic lupus erythematosus, Glutamatergic synapse, Glycine, serine and threonine metabolism, Retrograde endocannabinoid signaling, Renin secretion.

PPI Network
A PPI network of the proteins encoded by identi ed DEGs was constructed by STRING. To identify the key modules of the PPI network, module clustering was then performed through the MCODE plugin of Cytoscape (Fig. 3). The modules showed that several pain or itch related genes like GRPR, NPFF, CCKBR, BDKRB2, CXCR1 and CCL13 genes (Red Rectangle, Fig. 3) were also clustered and most of these genes also occurred in GO terms or KEGG pathways enriched above. Finally, we identi ed that GRPR, NPFF and TRPA1 as potential biomarkers for abdominal pain of IBS patients.

Discussion
Persistent hypersensitivity of sensory pathways innervating the colon is regarded as primary mechanism contributing to the initiation, development, and maintenance of chronic discomfort and abdominal pain in IBS patients (15,16,39). Therefore, determining the mechanisms contributing to these processes is crucial. [12]originally submitted the GSE36701 dataset, performed mRNA expression pro ling study of rectal biopsies from donors with healthy volunteers and IBS-D patients. They collected detailed clinical data (including abdominal pain frequency, anxiety and depression scale), using a combination of bioinformatics and experimental approaches to identify candidate genetic polymorphisms in the IBS [12]. While we further ltered for functional associations related to neurotransmitters/mediators of pain. Most notable among those genes were shown in Table 1(upregulated genes) and Table 2(downregulated  genes). Interestingly, GRPR ,NPFF ,TRPA1, BDKRB2,MRGPRX3, that commonly regarded as factor also regulating itch signaling pathways were found [13][14][15]. Finally, three genes including GRPR, NPFF and TRPA1 were considered to play an essential role in abdominal pain in IBS.
In the colon, afferent sensitization occurs via a variety of processes [16], including histamine-dependent mechanisms and histamine-independent mechanisms [17,18]; Evidence have been accumulated that activation of receptors associated with the above two itch pathways on colon-innervating afferents induces visceral hypersensitivity [17,19]. Gastrin-releasing peptide receptor expressing(GRPR) + neurons have a central role in the spinal transmission of both histaminergic and non-histaminergic itch [20].In our study, GRPR was signi cantly upregulated in IBS-D patients and identi ed as a hub gene by MCODE. GRPR is a G protein-coupled receptor and mediates itch sensation mainly via the PI3Kγ/Akt pathway [21]. At the same time, GRP induces neutrophil chemotaxis through GRPR via p38, ERK1/2 [22] and PI3K activation and p38, ERK and PI3K are important mediator in visceral pain [23][24][25]. Itch and pain signals are conveyed by distinct yet interacting neuronal pathways: pruritogens at higher doses produce pain to suppress itch [26,27]; The frequency of abdominal pain in IBS was higher in patients with chronic pruritus than in healthy controls [28]. In addition, Neuropeptide FF (NPFF), the histamine-independent itch receptors agonist, evoking colonic afferent mechanical hypersensitivity [15,17], was also clustered by Cytoscape module in our study. Therefore, we speculated that altered GRP-GRPR signaling and NPFF in spinal dorsal horn participate in visceral hypersensitivity through sensitization of itch transmission neurons, thereby contributing to abdominal pain or discomfort.
In the present study, another gene related to pain transmission deserve attention. Our results, together with previous studies have indicated signi cantly up-regulated TRPA1(Transient receptor potential ankyrin 1) mRNA expression in biopsies of IBS patients [29]. TRPA1 has been implicated in mechanical hypersensitivity of colonic afferents and both bradykinin and TNF-α induce visceral hypersensitivity through a TRPA1-dependent mechanism [30]. On the other hand, TRPA1 can also induce the in ammatory response via neurogenic in ammation: activation and sensitization of TRPA1 and release of substance P contribute to the initiation and development of colitis in mice, which correspondingly re-sensitises nociceptors [31]. TRPA1 expressed by intestinal enterochroma n cells can serves as the primary detector of intestinal irritants prior to direct sub-mucosal damage [32]. These ndings further highlight TRPA1 as an important integrator of sensory signals in colonic afferents by inducing mechanical visceral hypersensitivity.
As we known, IBS is a multifactorial disease. Abnormal stress response, psychological distress and infectious or in ammatory response in susceptible population may initiate visceral hypersensitivity that results in the development of IBS symptoms. According to the functional enrichment analyses, two pathways were associated with dysregulation of renin-angiotensin system (RAS), which is a potent target of stress-induced intestinal in ammation in a murine model of IBS [33]. Circadian entrainment and disorder of several synapses such as Glutamatergic synapse, Dopaminergic synapse may be risk factors of IBS. Besides, certain pathways directly involved in pain transmission were enriched, including Morphine addiction and Retrograde endocannabinoid signaling. Endocannabinoids are involved in controlling motility, secretion and intestinal in ammation [34]. The endocannabinoid system in DRG neurons mediate stress-induced visceral hypersensitivity in a mouse model of IBS [35].
In conclusion, our results showed that GRPR, NPFF and TRPA1 genes may be potential biomarkers for the diagnosis and new targets for treatment of abdominal pain in IBS. Several synapses modi cation and biological process of psychological distress may be risk factors of IBS. However, further studies are required to con rm the clinical signi cance of these ndings. Authors' contributions ZYL and YMW contributed to this work equally. ZYL: data collection and writing up of the rst draft of the paper; YMW: data analysis. SQL, DCL and GHM did literature research and analysis of the data.
Correspondence and requests for materials should be addressed to YLD and HZ: study design and modifying paper. All authors have read and approved the manuscript and ensure that this is the case.  Biological process (BP) enriched by the differentially expressed genes (DEGs). The x-axis of graph shows the counts, the number of genes clustered in each category, increasing with bar length. The threshold of P-value was set as 0.05.