Frontiers Corrigendum Template for Authors

Marzena Skrzypczak-Zielinska ^1* Marcin Gabryel ² Oliwia Zakerska-Banaszak ¹ Karolina Ladziak ¹ Katarzyna Anna Hubert ¹ Alina Baturo ² Joanna Suszynska-Zajczyk ³ Magdalena Hryhorowicz ³ Agnieszka Dobrowolska ²

• ¹ Institute of Human Genetics, Polish Academy of Sciences, Poznań, Poland
• ² Department of Gastroenterology, Dietetics and Internal Diseases, Poznan University of Medical Sciences, Poznan, Poland
• ³ Department of Biochemistry and Biotechnology, Faculty of Agronomy and Bioengineerring, Poznan University of Life Sciences, Poznan, Poland

1 IntroductionInflammatory bowel diseases (IBDs), including Crohn’s disease (CD) and ulcerative colitis (UC), are a group of chronic inflammatory diseases of the gastrointestinal (GI) tract. These conditions are incurable and characterized by alternating periods of exacerbation and remission. IBD is often referred to as a civilization disease due to its higher prevalence in populations of highly developed countries (1). Unfortunately, the etiology of IBD remains poorly understood, contributing to the lack of causal treatments. Pathogenesis of IBD is complex; many factors, such as genetic, environmental, immunological, and microbial, influence their development (2, 3). However, ongoing research still aims to fully elucidate the molecular pathomechanisms underlying these diseases.Global studies indicate that factors regulating the inflammatory process within the body are critical contributors to the development of various inflammatory diseases, including IBD (4). Interleukin 8 (IL-8, CXCL8) encoded by the CXCL8 gene is one of the first and most intensively studied chemokines and can be released by different cell types, including monocytes, endothelial cells, intestinal epithelial cells, and T lymphocytes (5). Activation of IL-8 by cytokines (such as interleukins 1 and 6, interferon-γ, tumor necrosis factor-alpha, and even IL-8 itself), bacterial molecules, lipopolysaccharide, hypoxia, and other cellular stresses increase its level several folds. The primary function of IL-8 is to activate inflammatory cells and enhance the immune response by acting as a chemotactic factor. This property enables the migration of leukocytes, mainly neutrophils, from peripheral blood to inflamed tissues. IL-8 binds to G-protein-coupled receptors, CXCR1 and CXCR2, present on the surface of neutrophils and triggers cellular signaling leading to neutrophil activation. Consequently, IL-8 plays a pivotal role in the innate immune response (6).The CXCL8 gene (MIM 146930) encoding IL-8 is implicated in the etiology of several chronic inflammatory diseases, including asthma, rheumatoid arthritis, skin inflammation, periodontitis, and cancer (7–9). Sequence variants of the CXCL8 gene are responsible for developing those disorders via two mechanisms. The first mechanism involves the CXCL8 gene expression caused by polymorphisms in the promoter region that lead to alteration in transcription factor binding. Changes in CXCL8 expression levels influence the intensity of the pro-inflammatory response and are often associated with distinct disease phenotypes. The second mechanism involves structural modifications in the receptor binding sites in the IL-8 protein, which can alter receptor affinity (10, 11).More than 700 SNPs (NCBI dbSNP) in the CXCL8 gene have been identified, focusing on those located in the promoter and protein-coding regions. The c.-251A>T (rs4073) variant is situated in the CXCL8 gene at the transcription factor NF-KB binding site and partially at the transcription enhancer-binding position (12). The NF-κB activation pathway has garnered significant interest from scientists due to its role in CXCL8 transcription (13). Both in vitro and in vivo research revealed that allele c.-251T is associated with an increased level of IL-8 (14). Certain studies have demonstrated inappropriate activation of this pathway in gastrointestinal conditions (15). The variant allele (T) frequency in the European population is 57.9% (1000 Genomes Project Phase 3).Meanwhile, variants in the IL-8 protein-coding region can significantly affect the activity and structure, which disturbs the course of the inflammatory reaction cascade. One notable substitution is c.91G>T (rs188378669), which results in the premature termination of the IL-8 protein due to an amino acid change (p.Glu31Ter) (12). The frequency of this rare variant is <1% (1000 Genomes Project Phase 3). However, in our previous investigation on genetic factors determining the response to glucocorticoids, we observed in our study group, i.e., among patients with IBD, an over eightfold higher frequency of this variant compared to the European population (16). This finding prompted us to study a larger IBD patient cohort to determine if either of these two variants in the CXCL8 gene may be associated with the manifestation of IBD and the inflammatory process cascade.2 Materials and methods2.1 Patients and samplesA group of 353 IBD patients (189 with CD and 164 with UC) under the care of the University Clinical Hospital in Poznan, the Department of Gastroenterology, Dietetics, and Internal Medicine, along with 200 subjects (104 women and 96 men) from the Polish population, were randomly enrolled to this study. IBD patients’ middle age was 46.7 ± 17.1 years, and for the population, 43.4 ± 15.2 years. CD and UC were diagnosed based on the clinical, endoscopic, and histological data. All participants provided written informed consent. To conduct the study at the protein level, we used serum material from available three IBD patients with the DNA substitution, 23 IBD patients without the genetic change, and 10 healthy individuals in whom IBD was excluded and who did not have the tested variant. The study was approved by the local Ethics Committee of Poznan University of Medical Sciences (resolution nos. 436/13 and 466/20). All experiments were performed in accordance with the principles of the 1964 Declaration of Helsinki and its subsequent amendments.2.2 DNA isolationGenomic DNA was isolated from peripheral blood samples of all individuals using the guanidine isothiocyanate method. Blood was collected into 5-ml EDTA S-Monovette tubes (Sarstedt, Numbrecht, Germany). The resulting DNA samples were stored at 4°C in TE buffer (10 mM Tris–HCl, 0.1 mM EDTA, pH 8) until use.2.3 CXCL8 locus amplificationThe amplification of CXCL8 gene fragment containing promoter region with variant c.-251A>T was carried out in a total volume of 30 μl using 0.75 U of FIREPol® DNA Polymerase, 2.5 μl 10× buffer, 2.0 μl dNTP mix (2.5 mM each dNTP), 1.5 mM MgCl2 solution, 80 ng DNA, and 0.2 μM of each primer (forward, 5′ATCTTGTTCTAACACCTGCCACTC3′, and biotinylated reverse, 5′AAGCTCCACAATTTGGTGAATTA3 ′). Amplification involved 50 cycles at 95°C for 30 s, 58°C for 30 s, and 72°C for 60 s. All reagents were obtained from Solis BioDyne (Tartu, Estonia). The PCR products with a length of 112 bp were checked in 1.7% agarose gel electrophoresis.2.4 Pyrosequencing and competitive allele-specific PCRGenotyping of rs4073 locus in CXCL8 gene by pyrosequencing was performed with a sequencing primer (5′TAGAAATAAAAAAGCATACA3′) using the PSQ™ 96MA system (Qiagen GmbH, Hilden, Germany) and PyroMark™ Gold Q96 Reagents (Qiagen) as described by the manufacturer.For genotyping of the rs188378669 (c.91G>T; p.Glu31Ter) in the CXCL8 gene, competitive allele-specific PCR (KASP) was used. The reaction mix contained one forward primer, two competitive reverse primers (one complementary to the G allele labeled with FAM and the other complementary to the T allele labeled with HEX), master mix, genomic DNA, and water. The PCR reaction was carried out on CFX Connect Real-Time PCR Detection System (Bio-Rad) according to the manufacturer’s instructions. Fluorescent signals were detected using Bio-Rad CFX Manager. Results were grouped based on genotype.2.5 Sanger sequencingSamples grouped using the KASP method were randomly selected for sequencing to confirm the genotype by Sanger sequencing. Sequencing was performed using BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific, Waltham, MA, USA) on the Applied Biosystems 3500 Series Genetic Analyzer. For this purpose, the following primers were used: forward, 5′ATCACTTTTTCCCCCAACAG3′ and reverse 5′CCTAACACCTGGAACTTTCCTAAA3′ for amplification of 246 bp product.2.6 ELISA assayThe IL-8 concentration in serum was analyzed using an enzyme-linked immunosorbent assay (ELISA), Human IL-8 ELISA Kit (Diaclone, cat. no. 950.050.192), according to the manufacturer’s protocol. The absorbance was measured at 450 nm wavelength using an EL-808 scanner spectrometer (BioTek Instruments Inc., USA). The test sensitivity was 12.3 pg/ml, and the detection range was 31.25–1,000 pg/ml.2.7 In silico prediction of variant consequence and structure-based functionTo evaluate the potential consequences of the rs188378669 (c.91G>T) variant, a comprehensive in silico analysis was conducted using multiple bioinformatic tools and databases. To analyze the structural characteristics of interleukin-8 (IL-8) protein, the AlphaFold AI (17, 18) system developed by Google DeepMind and the PDBe-KB Database (19) were applied.To assess pathogenicity indicators of the variant, the following databases were utilized: NCBI dbSNP (20), ClinVar (VCV003038842.2) (21), ProtVar (22), COSMIC (23), Mutation taster (24), genomeAD browser (25), and VarSome Database (26). To determine the probability of pathogenicity of the tested variant, the subsequent models were used: CADD Score (27), BayesDel Score (28), and EIGEN Score (29).2.8 Statistical analysisThe chi-square test was applied to assess the compliance of the obtained genotype distribution with the Hardy–Weinberg equilibrium. Based on the threshold p-value (p > 0.05), the variants were involved in the linkage disequilibrium analysis using the Haploview v.4.2 software. To investigate the variants’ associations with the disease the chi-square test was also used or Fisher’s exact test for small group sizes (below 5). The quantitative data were analyzed by the Student’s t-test in case data followed a normal distribution (Shapiro–Wilks test). The Mann–Whitney test for comparison of two subjects groups and Kruskal–Wallis test for comparison of three groups was applied if the data did not follow the normal distribution. These analyses were performed using PQStat 1.8.4 (PQStat Software, Poland) and R software 4.3.2, and all tests were considered significant at p value below 0.05. The false discovery rate (FDR) adjusted p-values were calculated according to Bonferroni correction.3 ResultsA total of 353 IBD patients and 200 individuals from the Polish population were genotyped for both loci, c.91G>T and c.-251A>T in the CXCL8 gene. Pyrosequencing was used to identify the c.-251A>T variant in the analyzed samples. All possible genotypes were detected, with a variant allele frequency of 54% in IBD patients and 51.5% in controls. The KASP method was applied for the stop-gain variant (c.91G>T) in CXCL8 gene identification. This technique enabled the grouping of samples according to their genotypes. We identified only the heterozygous form of this rare variant, with a frequency of 0.5% in population of Poland and 4.25% in Polish IBD patients. Sanger sequencing with the reverse primer was used to confirm all heterozygotes.3.1 Frequency and allele distributionThe frequency and distribution of all identified genotypes and alleles in IBD patients and the Polish population group are presented in Table 1.The frequency of promoter variant c.-251A>T of the CXCL8 gene does not differ significantly. It was similar in both patients and control groups and was not much different to the European population in the 1000 Genomes database. For the second variant studied by us, the frequency of the c.91T allele in the CXCL8 gene was higher in the patient group compared to healthy controls (p=0.012) and was also higher than the data from the 1000 Genomes database (which reports a frequency of <1% (30). Therefore, we decided to check whether the differences in frequency would be closely related to CD or UC, comparing both groups of patients (Table 2). Performed analysis of c.91G>T presence in CD and UC patients showed no significant differences in allele and genotype distribution.Table 2www.frontiersin.orgTable 2. Distribution of rs188378669 in CXCL8 gene among CD and UC patients.3.2 Pairwise linkage disequilibrium analysisHaplotype analysis was performed to investigate possible linkage disequilibrium between the examined polymorphic changes in the CXCL8 gene. Results showed that both polymorphisms c.91G>T and c.-251A>T are in strong (D′=1.0) linkage disequilibrium, and three combinations of haplotypes have been observed: TG, AG, and AT, respectively (Figure 1).Figure 1www.frontiersin.orgFigure 1. Linkage disequilibrium between analyzed variants in the CXCL8 gene. Haplotype diagram was prepared with Haploview v.4.2 and show calculated and graphically displayed the absolute value of Lewontin’s measure of linkage disequilibrium, D’ among both studied variants. The intensity of the filled-in box corresponds to the total linkage disequilibrium.3.3 Variant c.91G>T clinical consequencesFor the clinical consequences analysis of the c.91G>T variant, data were collected from the entire group of heterozygous IBD patients (14 subjects) and 23 randomly selected homozygous IBD patients. The characteristics of this group is presented in Table 3.We compared the distribution of GT and GG genotypes in CD and UC locations. Among CD patients with GT genotype, cases with L1, L2, and L3 were found. No homozygotes GG with CD localization in L1 were observed. In turn, among CU carriers of the variant at position c.91, there were only single cases with E2 and E3, and homozygotes GG had E1, E2, and E3. However, no statistically significant differences were found (Figure 2A). Regarding GI surgery due to IBD, more patients with the GT genotype underwent surgical treatment compared to GG homozygotes (Figure 2B), although this difference was borderline significant before statistically correction (p = 0.0851, padj. = 0.2389, Table 3). Genotype GT also predisposed to the need for treatment with corticosteroids. The odds ratio of using corticosteroids therapy in the GT genotype group are 7.8 times higher than in the homozygous GG group (p = 0.0164, padj. = 0.0654, Table 3).IL-8 concentration was measured in serum samples from three IBD patients with the heterozygous c.91G>T variant, 23 IBD patients with homozygous GG, and 10 healthy homozygous GG controls. All those 26 IBD patients were in remission state of disease. We observed increased IL-8 concentration in IBD patients with the GG genotype compared to healthy controls with the same genotype at the studied c.91 position of the CXCL8 gene (p = 0.0018, Figure 2C; padj. = 0.0055, Table 3). Meanwhile, the GT heterozygous IBD patients exhibited lower mean IL-8 concentration than GG homozygous IBD patients but higher than GG healthy controls (61.73 vs. 70.02 vs. 51.50, respectively).3.4 Prediction of variant consequences and structure-based functional analysisTo determine the potential consequences of the rs188378669 c.91G>T variant, an in silico analysis was conducted, focusing on the conservation of the p.Glu31Ter position as a nonsense variant and its predicted pathogenicity. For this purpose, information regarding the structure of IL-8 was retrieved from AlphaFold. The altered amino acid position in rs188378669, Glu31, receives a pathogenicity score of 0.737 (probability scale: 0, benign; 1, highly pathogenic). This suggests that the Glu31 alteration is likely pathogenic. The structures available in the PDBe-KB (Protein Data Bank in Europe—Knowledge Base) database do not describe (except one entry) the first 29 amino acids of IL-8. These initial amino acids are part of a precursor, and the functional protein begins after this segment. In most cases, protein structure simulations start at the 30th amino acid, as the functional part of the protein begins at this point (32). This fact emphasizes the potential significance of a mutation that causes a stop in protein synthesis at the 31st nucleotide site.An in silico analysis of rs188378669 using NCBI dbSNP, ClinVar (accession, VCV003038842.2), ProtVar, and COSMIC databases indicated that premature protein termination is a predicted consequence of the c.91G>T variant. The MutationTester tool classifies this variant as disease causing.ProtVar and the gnomAD browser report a CADD score of 35 for this variant, indicating a high likelihood of deleteriousness (CADD score > 29.9 is highly likely to be deleterious). A BayesDel noAF score of 0.588 also indicated (score ranges from −1.29334 to 0.75731) a strongly pathogenic nature of the variant (23). Another source indicating the variant’s pathogenicity is the EIGEN score, which was found to be 0.9117, interpreted as Pathogenic Moderate.The VarSome database was used to summarize and organize the above data. VarSome assigned the examined variant a high pathogenicity meta-score of 6 (scale: 1, supporting; 2, moderate; 4, strong; 8, very strong). It should be noted, however, that some databases used in the VarSome database, like FATHMM_XF, classify the change as benign, although these represent a minority of the sources (33).4 DiscussionPathogenesis of IBD, one of the fastest-growing and incurable diseases of modern civilization, despite extensive global research, is still not yet fully elucidated. However, it is considered to be a result of an inappropriate inflammatory response to intestinal microbes in genetically predisposed subjects (34). Given the pivotal role of chemokines in the immune system, we investigated in this study whether the CXCL8 gene encoding one of the most important of them, IL-8, could be associated with the disease manifestation and the course of the disease. Because IBD is still an incurable disorder associated with major morbidity in Western countries, searching for new IBD-associated loci and therapeutic targets is an extremely urgent challenge currently the more so that it is a complex, polygenic disease. Up to date, over 200 loci associated with IBD have been identified, mainly through genome-wide association studies (GWAS), from which one of the first was the NOD2 gene (35, 36). Approximately 68% of these loci are common for CD and UC, indicating that these diseases share inflammatory pathways (35). However, the CXCL8 gene has yet to be identified among them.In our IBD patients cohort, heterozygotes GT in locus c.91 of the CXCL8 gene were presented with a frequency of 4.25% compared to 0.5% in the Polish population. We decided to we decided to investigate whether this variant is significantly related either to CD or UC. However, the results showed no association with a specific IBD entity. The presented study has demonstrated for the first time such a correlation of this nonsense substitution c.91G>T in the CXCL8 gene with the presence of IBD. The results also confirmed our previous observations concerning this variant frequency, based on the NGS study of patients with IBD. It is worth noting that c.91 locus of the CXCL8 gene has been already associated with early-onset myocardial infarction, hypertension, chronic kidney disease, or endometrial carcinoma (23, 37).The nucleotide substitution c.91G>T (known as mutation numbered COSV56640974) analyzed in this study has serious consequence—change in the 31st amino acid (Glu) to codon STOP and premature protein chain termination. As a result, a non-functional protein is produced, characterized by reduced stability and receptor-binding affinity, potentially disrupting the proper inflammatory pathway in which IL-8 plays a crucial role (12). Our extensive in silico predictive analysis strongly supports the pathogenic nature of the c.91G>T (p.Glu31Ter) variant.IL-8 activity depends on interaction with the human CXC chemokine receptors CXCR-1 and CXCR-2, atypical chemokine receptor encoded by ACKR1 gene, and glycosaminoglycans (6). In the case of IBD, IL-8 recruits and activates neutrophils into the lamina propria and epithelium by binding to the CXCR-1/2 receptors (38). The axis IL-8-CXCR-1/2 is crucial in the pathogenesis of UC through multiple signaling pathways, including PI3K/Akt, MAPKs, and NF-κB. There are suggestions that targeted inhibition of this essential axis, IL-8-CXCR-1/2, could be a new therapeutic strategy for the disease (39, 40). Research performed on the DSS-induced mice model revealed that treatment with CXCR-1/2 antagonist improved the colonic condition of mice (6). Chapuy and his research team established a link between monocyte-like CD163− MNPs, IL-12, and IL-1β and the detection of colonic memory IL-8-producing CD4+ T cells, which might all contribute to the pathogenesis of UC but not of CD (41). However, IL-8 is only one of the chemokines produced by inflamed epithelial cells in UC patients, besides C-X-C motif chemokines 1 and 9, and others (42, 43). To better understand the exact role of CXCL8 in the genes-based network, we created interaction networks using GeneMANIA (Figure 3) (44). This network visualizes the link of CXCL8 with other genes, not only receptor genes but also genes of many abovementioned interleukins, ligands, matrix metalloprotein, intercellular adhesion molecule, and G protein subunits.Furthermore, recent studies show IL-8 to be a critical factor contributing to tumorigenesis (45). Upregulation of IL-8 was observed at the tumor invasion front in several human cancers. Modulation of IL-8 by CXCR1/2 chemokine receptors promotes various intracellular signaling cascades that lead to tumor-associated inflammation (46). Hence, deciphering the regulatory and signaling function of IL-8 and its further effects may hold prognostic clinical prospects for therapies. Currently, multiple clinical trials are ongoing (6).Previously, we have also analyzed distribution of this variant c.91G>T in CXCL8 gene in a group of cystic fibrosis patients from the Polish population, searching for prognostic markers of clinical form of the disease. We found only one heterozygote c.91GT in a group of 55 individuals (47).Our results here show a borderline association between the GT variant genotype and the need for GI tract surgical treatment, suggesting its relevance to IBD pharmacotherapy effectiveness. Although this study did not explicitly focus on treatment effects, this issue will be addressed in future research. Nevertheless, the observed relationship may indicate the involvement of this variant in the therapy response connected with the disease’s course. Similar association is visible for the GCS treatment; however, further advanced pharmacogenetic investigation are necessary. It should be noted that the mechanism of action of IL-8 is complicated and not obvious. In vitro data demonstrate that chemokine IL-8 concentration gradient determines neutrophil motile behaviors. Namely, neutrophils migrate toward increased IL-8 concentration; however, high concentration of IL-8 can repel neutrophils and induce their reverse movement (48, 49).4.1 Limitations of the studyWe are also aware of another limitation of this study. The IL-8 serum concentrations were measured only in patients in the remission phase to maintain homogeneous conditions with the control group. However, data from patients during exacerbation would also be valuable and should be included in future studies. Second, we had a limited number of GT genotype patients and even fewer protein studies due to the very low frequency of variant c.91G>T in the CXCL8 gene. Namely, in the genetic studies of 353 IBD patients, we showed 14 heterozygotes. For serum testing, only three patients simultaneously meets the criterion of no exacerbation and provided a fresh blood sample. With unbalanced groups (with one group with n = 3), the power of the analysis drops from approximately 87.3% to approximately 80.6%.5 ConclusionThe results of our study, presenting new significant associations with the IBD locus in the CXCL8 gene and defining for the first time IL-8 concentration in IBD p.Glu31Ter carriers, need further confirmation from other populations and functional studies before it may constitute a promising therapeutic value for IBD in the future.