Macrophage-Stimulating 1 Polymorphism rs3197999 in Pediatric Patients with Inflammatory Bowel Disease

Background and Objectives: Inflammatory bowel disease (IBD), which includes Crohn’s disease (CD) and ulcerative colitis (UC), often necessitates long-term treatment and hospitalizations and also may require surgery. The macrophage-stimulating 1 (MST1) rs3197999 polymorphism is strongly associated with the risk of IBD but its exact clinical correlates remain under investigation. We aimed to characterize the relationships between the MST1 rs3197999 genotype and the clinical characteristics in children and adolescents with IBD within a multi-center cross-sectional study. Materials and Methods: Clinical data included serum C-reactive protein (CRP), albumin, activity indices (PUCAI, PCDAI), anthropometric data, pharmacotherapy details, surgery, and disease severity. Genotyping for rs3197999 was carried out using TaqMan hydrolysis probes. Results: The study included 367 pediatric patients, 197 with Crohn’s disease (CD) (40.6% female; a median age of 15.2 years [interquartile range 13.2–17.0]) and 170 with ulcerative colitis (UC) (45.8% female; a median age of 15.1 years [11.6–16.8]). No significant relationships were found between MST1 genotypes and age upon first biologic use, time from diagnosis to biological therapy introduction, PUCAI, PCDAI, or hospitalizations for IBD flares. However, in IBD, the height Z-score at the worst flare was negatively associated with the CC genotype (p = 0.016; CC: −0.4 [−1.2–0.4], CT: −0.1 [−0.7–0.8], TT: 0.0 [−1.2–0.7)]). The TT genotype was associated with higher C-reactive protein upon diagnosis (p = 0.023; CC: 4.3 mg/dL [0.7–21.8], CT 5.3 mg/dL [1.3–17.9], TT 12.2 mg/dL [3.0–32.9]). Conclusions: This study identified links between MST1 rs3197999 and the clinical characteristics of pediatric IBD: height Z-score and CRP. Further studies of the associations between genetics and the course of IBD are still warranted, with a focus on more extensive phenotyping.


Introduction
Inflammatory bowel disease (IBD), which includes Crohn's disease (CD) and ulcerative colitis (UC), affects approximately 0.3% of the population of Poland [1] and up to 1% in the most highly developed countries [2].The course of IBD varies individually [3], but a large share of patients require hospitalizations, with a trend for increased in-hospital management in many countries globally [4].The disease affects both pediatric and adult populations, burdening the caregivers and the healthcare systems [5].Its long-term treatment focuses on controlling the symptoms and does not always prevent the need for surgery.Moreover, pharmacotherapies have limited efficacy, exhibit side effects, and may lose their beneficial effects with time as in the case of anti-tumor necrosis factor alpha (TNF-alpha) biologics [6].Furthermore, dietary treatment, whilst efficacious in pediatrics, remains relatively challenging in practice in adult patients with CD [7,8] and still requires further research in UC [9].The disease not only results from lifestyle but can also impact it by limiting physical [10] and social activities.The most recent push to predict response to individual IBD treatments is not always bearing fruit [11], but efforts continue to enable precision medicine in IBD, with some promising results [12].These findings build on IBD omics, including IBD genomics, where the rapid growth in knowledge [13] has not always been accompanied by independent investigations of relationships between individual single-nucleotide polymorphisms (SNPs) and characteristics of IBD.
The SNP (single-nucleotide polymorphism) rs3197999 (C/T; missense) is located in the last exon of MST1 (encoding macrophage-stimulating 1; 487 aminoacids), and is less than 1 kbp following acylaminoacyl-peptide hydrolase (APEH), encoded on the opposite strand [14].It was found to be the top SNP in this genomic region following an IBD genetics metaanalysis [13], with a similar strength of relationships for both CD and UC.The overall statistical significance for the genetic relationship between rs3197999 and IBD was the 8th strongest of 241 loci, giving MST1 the place among such key IBD genes as IL23R, NOD2, CARD9, and IL10.Of note, rs3197999 is also a risk factor for primary sclerosing cholangitis, which may accompany UC [15,16] and is associated with a gain of function of MST1 [17].Yet, MST1 also attracts attention across a broad spectrum of disciplines not related to the strongest findings from genome-wide association studies, including cardiology, diabetology, neurology, and oncology [18][19][20].Especially in this last context, much stress is placed on the role of MST1 in the Hippo tumor suppressor pathway, which regulates cell proliferation [21].
There are, however, also reports of the role of MST1 in immunity, with reduced peripheral T cells described in Mst1-null mice, potentially due to early apoptosis, especially in the case of more easily proliferating naïve T cells, potentially providing a link to the Hippo pathway [22].Mst1−/− mice may be susceptible to T cell-related autoimmunity because of insufficient antigen recognition during thymic selection [23], although data on this topic are conflicting [22].Importantly, the development of FoxP3+ CD4+ Tregs was found to be impaired in Mst1-null mice through the insufficient post-translational acetylation of Foxp3 [24].The influence of MST1 on immunity may involve known IBD actors such as LFA-1 and ICAM-1 and processes such as cytoskeleton dynamics [25].
Therefore, MST1 is one of the most important genetic contributors to the development of IBD and its role is not entirely clear [26].We aimed to provide new data on the relationship between MST1 and the clinical picture of pediatric IBD.

Materials and Methods
The study is based on samples gathered in the years 2016-2019 at seven pediatric university hospitals in Poland, in the cities of Pozna ń (n = 1), Warsaw (n = 2), Wrocław (n = 1), Zabrze (n = 1), Katowice (n = 1), and Bydgoszcz (n = 1).The inclusion criteria were (i) IBD diagnosed according to current guidelines, (ii) age between 3 and 18 years, and (iii) the absence of life-threatening symptoms [27].Informed consent was obtained from parents or legal guardians of all the children and adolescents and from adolescents aged 16 years or older.The study was carried out within the scope of formal approval by the Bioethical Committee at Poznan University of Medical Sciences (960/15, with amendments).The study respected the rules set out in the Declaration of Helsinki [28].
The basic collected data were age upon diagnosis and inclusion, gender, type of IBD (UC or CD, without IBD-unspecified), and a range of parameters upon diagnosis and at the worst flare: serum C-reactive protein (CRP); albumin; Pediatric Ulcerative Colitis Index (PUCAI) [29] or Pediatric Crohn's Disease Activity Index (PCDAI) [30]; and Z-scores for body mass, height, and body mass index (BMI), based on local reference values.Severity was thought to be reflected by the employed treatment: systemic steroids, azathioprine, methotrexate, ciclosporin, biologics (infliximab, adalimumab), and also more specifically, the time from diagnosis to starting biological therapy, age at the introduction of therapy with a biologic, the need for any IBD-related surgery, the number of hospitalizations for IBD flares, and in the case of patients with a disease duration > 1 year, also the number of hospitalizations (and days of hospitalization) for IBD flares per year.
DNA was obtained from whole blood sampled in tubes with EDTA using a microcolumnbased kit (Blood Mini, A&A Biotechnology, Gdynia, Poland).Genotyping was based on the use of a TaqMan probe with the TaqMan Genotyping Master Mix (both from Thermo Fisher Scientific Inc., Waltham, MA, USA).The CFX96 thermocycler system (Bio-Rad, Hercules, CA, USA) was used to carry out quantitative PCR reactions and discriminate alleles based on plots.The four runs included negative control.The number of cycles was 41.Samples with fluorescence close to zero were excluded prior to analyses (7 samples).The plot of the sample fluorescence and genotyping is attached as Supplementary Figure S1.
Data were analyzed to explore if the homozygosity of the major or minor allele (dominant/recessive model) was associated with the phenotype.The analyses were carried out using the Mann-Whitney U test because of the encountered non-normal distributions.Additionally, the Kruskal-Wallis test was applied to compare variable values between genotypes.Two-way tables were analyzed using Fisher's test.Two-sided p values are reported.The significance (alpha) threshold was set at 0.05.Statistical analyses were carried out using PQStat (PQStat, Poznan, Poland).The data were analyzed separately for CD, UC, and also in the entire IBD group to achieve the maximum power of analysis with regard to common variables.
The list of abbreviations can be found in the Supplementary Materials (Supplementary Table S1).

Results
Three hundred sixty-seven children and adolescents were included in the analysis (197 with CD; 170 with UC).The CD and UC groups were similar in age upon diagnosis (or inclusion), sex, and frequency of rs3197999 alleles.The median duration of the disease was above 2 years and the data concerning hospitalizations were available for more than half of the patients in both CD and UC.Over half of pediatric patients in both groups received azathioprine.However, the frequency of the use of anti-TNF biologics was lower in UC, which is discussed in limitations.
No relationships were found between MST1 genotypes and age upon the first biological therapy or time from diagnosis to biological therapy, PUCAI, PCDAI, or hospitalizations for IBD flares (Table 1).In UC, a marginally lower height Z-score at the worst flare was found in the homozygous rs3197999 CC genotype relative to other genotypes (Table 2).In CD, the CC genotype was both positively related to the use of systemic steroids and more common in female patients (Table 3).In IBD overall, the height Z-score at the worst flare was negatively associated with the CC genotype but the TT genotype was correlated with higher C-reactive protein upon diagnosis (Table 4).

Discussion
This cross-sectional study investigated the relationship between MST1 SNP rs3197999 and the clinical characteristics of IBD in pediatric patients across Poland.The need for study resulted from the high burden of disease and strong links between the aforementioned SNP and the risk of IBD, with compelling molecular evidence for the involvement of MST1 in immunity.The identification of relationships between specific clinical characteristics and the polymorphism could help in the understanding of its role and the involved mechanisms (with potential implications for treatment) and might provide a new tool for risk assessment in the pediatric population.
The rs3197999-CC genotype was negatively associated with height Z-score at the worst flare in IBD, possibly reflecting a greater impact of disease on development, potentially confounded by the more frequent use of steroids in CD (which negatively impacts growth [31]).However, this result implicating greater severity in rs3197999-CC appeared discordant with higher CRP upon diagnosis in patients with IBD with the rs3197999-TT genotype.These observed differences, albeit relatively small, provide new insights into the role of MST1 in IBD.
Previous studies provide strong evidence for the link between MST1 and IBD.In a study of almost 35 000 patients with IBD, MST1 rs3197999 was associated with the risk of UC and younger age upon its diagnosis, as well as with the ileal location of CD (vs.colonic, OR 1.10 [1.02-1.19] in [32] and vs. control OR 1.5 [1.1-2.1] in [33]).In yet another study, rs3197999 was associated with early-onset UC, distal colitis, and a reduced risk of extraintestinal manifestations [34].In patients with CD, MST1 rs3197999 conferred a higher risk of perianal disease (OR 1.19 [1.01-1.41)[35].Importantly, MST1 IBD risk loci were replicated in populations globally, including China [36], South Asia [37] and Eastern Europe [38].
Genomic analyses of three large cohorts suggested that higher MST1 levels may be protective against IBD [26].The minor allele of rs3197999 correlated negatively with the serum concentration of MST1 [39], and the relationship between SNPs in the MST1 locus and their relationship with IBD risk overlapped.The lower serum concentration may be explained by the lower stability of the mutant protein [40].The SNP may have gain-of-function effects with regard to macrophage inflammatory activity [17].
The investigated SNP was also found to be a risk factor for extrahepatic cholangiocarcinoma, and this was independent of the diagnosis of primary sclerosing cholangitis [41].This suggests a role for MST1 in the bile ducts, which is possibly related to inflammation that may underlie different types of disease.Additional insight into rs3197999 is provided by results implicating it as a modifying factor (relationship quantitative trait locus) of genetically predicted insulin concentration and sensitivity to insulin [42].Analyses of MST1 beyond rs3197999 link it also to other diseases, which is illustrated by MST1 rs9858542-a risk factor for both CD (more frequent homozygosity of the minor allele) and ankylosing spondylitis (more frequent heterozygous genotype, compared with controls) [43].
GPX1 (glutathione peroxidase 1) neighbors MST1 (within 300 kbp), and some studies pointed towards its involvement in IBD, possibly through its antioxidant functions.Evidence comes from Mendelian randomization [44] and genetic analyses [45].Notwithstanding the limitations related to potential linkage disequilibrium and difficulties in meeting assumptions of Mendelian randomization, in a specific Gpx1-null and Gpx2-null murine model (B6), the development of ileocolitis was facilitated [46].Also, in patients with IBD, the activity of GPX1 decreased, but this may be an epiphenomenon [47] similar to the reduced MST1 expression in the rectum in UC [48].Although there is some evidence for the involvement of GPX1 (or potentially also BSN and APEH), research supports the role of MST1 most strongly.
In the earlier part of the study period, the treatment with anti-TNF agents was not easily accessible for pediatric UC in Poland, and, therefore, their use could discriminate the most severe cases of IBD.One of the limitations is the lack of a healthy control group to compare the prevalence of the SNP, and the lack of study of MST1 serum levels or MST1 methylation.This study utilized data like height Z-scores, CRP levels, and the use of specific treatments, which might not capture the full clinical complexity of IBD.Moreover, it does not include fecal calprotectin level assessment, which was not systematically collected.The median disease duration was above 2 years, which might not be sufficient to observe long-term outcomes and the full impact of MST1 genotypes on disease progression and treatment response.However, this study also provides interesting insights into such aspects as the age of IBD diagnosis in Poland and the most commonly used medications, including the high frequency of azathioprine use.

Conclusions
This study has found few links between MST1 rs3197999 and clinical characteristics of IBD in pediatric patients.The MST1 rs3197999-CC genotype was associated with a reduced height Z-score upon the worst flare and in CD, it could be related to the more frequent use of steroids.A link was also identified between this polymorphism and CRP upon diagnosis.However, no relationships were found between rs3197999 and PUCAI, PCDAI, hospitalizations, and the time from diagnosis to biological treatment.Further studies of the links between genetics and the course of IBD are still warranted, with a focus on more extensive phenotyping.They are likely to yield insights useful for understanding disease complexity and personalizing clinical care.
BMI-body mass index; CC-cytosine/cytosine, CRP-C-reactive protein; CT-cytosine/thymine, IBDinflammatory bowel disease; PCDAI-Pediatric Crohn's Disease Activity Index; PUCAI-Pediatric Ulcerative Colitis Activity Index; TT-thymine/thymine, Z-score-number of standard deviations that the result differs by from the mean of the population at given age.

Table 2 .
Characteristics of children and adolescents with ulcerative colitis dependent on MST1 rs3197999 genotype.The Mann-Whitney U test, the Kruskal-Wallis test (K-W: comparison of three groups), and Fisher's test were used for comparisons.Q1-Q3: quartiles 1-3.Statistically significant results are marked with asterisk (*).

Table 3 .
Characteristics of children and adolescents with Crohn's disease dependent on MST1 rs3197999 genotype.The Mann-Whitney U test, the Kruskal-Wallis test (K-W: comparison of three groups), and Fisher's test were used for comparisons.Q1-Q3: quartiles 1-3.Statistically significant results are marked with an asterisk (*).

Table 4 .
Characteristics of pediatric patients with inflammatory bowel disease dependent on MST1 rs3197999 genotype.The Mann-Whitney U test, the Kruskal-Wallis test (K-W: comparison of three groups), and Fisher's test were used for comparisons.Q1-Q3: quartiles 1-3.Statistically significant results are marked with an asterisk (*).