Abstract

Background

   Ulcerative colitis (UC) is a common chronic disease associated with
   inflammation and oxidative stress. This study aimed to construct a long
   noncoding RNA (lncRNA)-microRNA (miRNA)-messenger RNA (mRNA) network
   based on bioinformatics analysis and to explore oxidative
   stress-related genes underlying the pathogenesis of UC.

Methods

   The [36]GSE75214, [37]GSE48959, and [38]GSE114603 datasets were
   downloaded from the Gene Expression Omnibus database. Following
   differentially expressed (DE) analysis, the regulatory relationships
   among these DERNAs were identified through miRDB, miRTarBase, and
   TargetScan; then, the lncRNA-miRNA-mRNA network was established. The
   Molecular Signatures Database (MSigDB) was used to search oxidative
   stress-related genes. Gene Ontology (GO) and Kyoto Encyclopedia of
   Genes and Genomes (KEGG) analyses were performed for functional
   annotation and enrichment analyses. Based on the drug gene interaction
   database DGIdb, drugs that interact with oxidative stress-associated
   genes were explored. A dextran sulfate sodium (DSS)-induced UC mouse
   model was used for experimental validation.

Results

   A total of 30 DE-lncRNAs, 3 DE-miRNAs, and 19 DE-mRNAs were used to
   construct a lncRNA-miRNA-mRNA network. By comparing these 19 DE-mRNAs
   with oxidative stress-related genes in MSigDB, three oxidative
   stress-related genes (CAV1, SLC7A11, and SLC7A5) were found in the 19
   DEM sets, which were all negatively associated with miR-194. GO and
   KEGG analyses showed that CAV1, SLC7A11, and SLC7A5 were associated
   with immune inflammation and steroid hormone synthesis. In animal
   experiments, the results showed that dexamethasone, a well-known
   glucocorticoid drug, could significantly decrease the expression of
   CAV1, SLC7A11, and SLC7A5 as well as improve UC histology, restore
   antioxidant activities, inhibit inflammation, and decrease
   myeloperoxidase activity.

Conclusion

   SLC7A5 was identified as a representative gene associated with
   glucocorticoid therapy resistance and thus may be a new therapeutic
   target for the treatment of UC in the clinic.

   Keywords: Ulcerative colitis, Bioinformatics analysis,
   LncRNA-miRNA-mRNA network, Oxidative stress, Inflammation, SLC7A5

Introduction

   Inflammatory bowel disease, which is a kind of autoimmune disease,
   mainly includes two subtypes: ulcerative colitis (UC) and Crohn’s
   disease (CD) ([39]Ng et al., 2017). In CD, all layers of the bowel wall
   are inflamed; in contrast, UC is generally characterized by mucosal
   layer inflammation and damage to the superficial bowel wall ([40]Hibi &
   Ogata, 2006; [41]Kobayashi et al., 2020). The disease course of UC
   usually involves remission and exacerbation in alternating cycles, and
   if treatment is not performed in a timely, colorectal cancer can
   develop ([42]Ungaro et al., 2017; [43]Lissner & Siegmund, 2013). From
   the year 1955, glucocorticoids have been considered effective for
   patients with UC, and the use of glucocorticoids dramatically decreases
   the mortality of patients with moderate-to-severe UC ([44]Nakase et
   al., 2021). However, long-term use of corticosteroids not only induces
   glucocorticoid-resistance but also leads to numerous adverse effects
   such as depression, cataracts, and osteoporosis ([45]Nakase et al.,
   2021; [46]Magro et al., 2017; [47]Truelove & Witts, 1955). Therefore,
   elucidating the pathogenesis of the initiation and progression of UC is
   imperative to develop novel therapies.

   The development of UC involves multiple mechanisms, and oxidative
   stress caused by the imbalance between antioxidants and oxidants plays
   a vital role ([48]Grisham, 1994). The presence of excessive reactive
   oxygen species (ROS) including peroxynitrite, hydrogen peroxide, and
   superoxide can reduce the productions of endogenous antioxidants,
   eventually resulting in cell death via oxidative damage to DNA,
   membrane lipids, and cellular proteins ([49]Amirshahrokhi, Bohlooli &
   Chinifroush, 2011; [50]Ferrat et al., 2019; [51]Niu et al., 2015). The
   reaction of DNA with ROS leads to the modification of DNA bases and
   contributes to subsequent carcinogenesis. For example, ROS may interact
   with genomic DNA to generate some base modifications with pro-mutagenic
   potentials such as 8-nitro-2′-deoxyguanosine (8-NO[2]-dG) and
   8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) ([52]Kaneko et al., 2008;
   [53]Kondo et al., 1999). One previous study demonstrated excessive
   production of 8-oxodG in patients with UC-associated carcinogenesis
   ([54]Gushima et al., 2009). This series of genetic changes serves as a
   trigger in the pathogenesis of chronic inflammation-associated
   diseases. Other studies have demonstrated local DNA damage in the colon
   and systemic DNA damage involving the hepatocytes, lymphoid organs, and
   blood in mice with dextran sulfate sodium (DSS)-induced UC, and these
   outcomes are considered to be partly mediated by oxidative stress
   ([55]Westbrook et al., 2011, [56]2009; [57]Trivedi & Jena, 2012). In
   addition, the expression of some inflammatory genes, such as TNF-α, is
   reported to be regulated by oxidative stress-associated genes, and the
   application of TNF-α inhibitors has been proven to be effective for the
   treatment of UC ([58]Verhasselt, Goldman & Willems, 1998; [59]Barrie &
   Regueiro, 2007). Therefore, identifying oxidative stress-associated
   genes and establishing a direct linkage between these genes and UC
   pathogenesis may be helpful for the treatment of patients with UC.

   Evidence is now emerging to indicate that non-coding RNAs play vital
   roles in inflammatory bowel disease, especially in the progression of
   UC ([60]Ghafouri-Fard, Eghtedarian & Taheri, 2020; [61]Schaefer, 2016).
   Non-coding RNAs are RNAs that are unable to code proteins and can
   regulate multiple biological processes through modulating the
   expression of coding RNAs ([62]Quinn & Chang, 2016; [63]Schmitz, Grote
   & Herrmann, 2016; [64]Matsui & Corey, 2017). There are mainly four
   types of non-coding RNAs: long non-coding RNAs (lncRNAs), microRNAs
   (miRNAs), circular RNAs (circRNAs), and extracellular RNAs (exRNAs)
   ([65]St Laurent, Wahlestedt & Kapranov, 2015; [66]Ling, Fabbri & Calin,
   2013; [67]Sato-Kuwabara et al., 2015; [68]Ebbesen, Kjems & Hansen,
   2016). [69]Salmena et al. (2011) proposed the notion of a competing
   endogenous RNA (ceRNA) regulatory network that involves interactions
   among these RNAs. lncRNAs are RNA molecules that can affect the
   transcriptional and post-transcriptional expression of genes ([70]Quinn
   & Chang, 2016; [71]Schmitz, Grote & Herrmann, 2016). Increasing
   evidence has demonstrated that lncRNAs have several molecular
   functions, such as being involved in regulatory transcription and
   acting as miRNA sponges and regulatory RNA binding proteins ([72]Jain
   et al., 2017; [73]Zacharopoulou et al., 2017; [74]Khan et al., 2022).
   In general, lncRNAs can sponge miRNA through miRNA response elements,
   which eventually affect the binding between miRNAs and messenger RNAs
   (mRNAs) ([75]Zhang et al., 2019; [76]Ala, 2020). lncRNAs have also been
   reported to be involved in regulating several kinds of human diseases
   including neurological diseases, cancers, cardiovascular diseases, and
   inflammatory bowel disease ([77]Ghafouri-Fard, Eghtedarian & Taheri,
   2020; [78]Canseco-Rodriguez et al., 2022; [79]Chi et al., 2019;
   [80]Poller et al., 2018). Furthermore, a lncRNA-related ceRNA network
   has been identified as a key mechanism involved in immune-related
   diseases including systemic lupus erythematosus ([81]Song et al., 2021)
   and rheumatoid arthritis ([82]Zhang et al., 2020). In addition,
   [83]Dong et al. (2022) constructed a lncRNA-related ceRNA network in UC
   and identified two mRNAs (CTLA1 and STAT1) that are associated with
   immune cell infiltration. However, research on oxidative stress genes
   underlying the lncRNA-related ceRNA network in UC is still in the
   preliminary stages.

   In the current study, the original data from the UC group and control
   group were obtained from the NCBI Gene Expression Omnibus (GEO)
   database. Based on an analysis of differential gene expression,
   differentially expressed (DE)-lncRNAs, DE-miRNAs, and DE-mRNAs were
   identified. This study aimed to probe a complete lncRNA-miRNA-mRNA
   network to determine the roles of oxidative stress genes in the
   pathogenesis and drug resistance mechanism of UC. Subsequently, a
   DSS-induced mouse model was established to validate the results of the
   bioinformatics analysis. These findings clarified the relationships
   among the identified oxidative stress genes and the pathogenesis and
   drug resistance mechanism in UC, providing some references for the