S 100 A 4 Protein in Inflammatory Bowel Disease : Results of a Single Centre Prospective Study

A B S T R AC T Introduction: The aim of our study was to assess association of serum S100A4 protein with ulcerative colitis (UC) and Crohn’s disease (CD). Methods: Study included 118 subjects: 93 patients with CD, 16 with UC and 9 controls. In CD group, 20/93 patients had B1 phenotype, 19/93 B2, 20/93 B3 and 34/93 B2 + B3. L1 involvement was present in 15/93, L2 in 14/93 and L3 in 64/93 patients. Serum S100A4 concentration was investigated in peripheral venous blood samples by means of ELISA. Results: Serum S100A4 was significantly higher in UC (158.6 ± 56.2 ng/mL), p = 0.019 and in CD (154.4 ± 52.1 ng/mL), p = 0.007 compared to controls (104.8 ± 40.5 ng/mL). No difference in S100A4 was revealed between UC and CD, p > 0.05. Serum S100A4 in each CD subgroup (according to behaviour) was significantly higher compared to controls, p < 0.05. Serum S100A4 was significantly higher in L2 (144.6 ± 44.2 ng/mL), p = 0.041 and in L3 (163.0 ± 52.8 ng/mL), p = 0.002 compared to controls and in L3 compared to L1 (126.9 ± 47.6 ng/mL), p = 0.017. Conclusion: Association of serum S100A4 protein with UC and CD was confirmed. In CD, disease behaviour did not influence serum concentration of S100A4 protein. In CD, higher levels of serum S100A4 were observed in patients with ileo-colonic and colonic involvement compared to those with isolated small bowel involvement.


INTRODUCTION
Inflammatory bowel disease (IBD), ulcerative colitis (UC) and Crohn's disease (CD) have become a global disease.Molodecky et al. reported in their recent study, that incidence and prevalence of IBD have been increasing worldwide.The highest prevalence of IBD has been documented in Europe and Canada (1).
The family of S100 proteins represents a total of at least 25 small calcium binding proteins.S100 proteins have a broad range of functions -they play an important role in the regulation of cell proliferation, differentiation, apoptosis, energy metabolism, cellular signalling, and calcium homeostasis (16).Involvement of S100 proteins in the pathogenesis of IBD has been clearly documented and the role of individual S100 proteins as biomarkers for CD and UC has been validated in multiple studies (17)(18)(19)(20)(21)(22)(23).
Calprotectin, a heterocomplex of S100A8/9 proteins, plays an important role in the regulation of different inflammatory processes and nowadays, faecal calprotectin is used for assessment of IBD activity on routine basis (17,24).S100A4 (metastatin-1, calvasculin) is localized in the nucleus, cytoplasm, and extracellular space.It is strongly associated with metastatic tumour progression (25).
Boye et al. emphasized that the nuclear expression (not the cytoplasmic one) of S100A4 is a novel prognostic marker for colorectal cancer (26) and further studies showed that S100A4 is not a biomarker only, but mediates the metastatic process itself, too (27).
Recent research has revealed that the role of S100A4 is more complex, including profibrotic effect e.g. in the myocardium, liver and intestine (28)(29)(30).Significant upregulation of S100A4 was observed in certain chronic inflammatory conditions, especially in patients with rheumatoid arthritis.It was shown, that increased secretion of TNF-α, IL-1β and IL-6 was mediated by S100A4 (31).The studies have demonstrated, that Toll-like receptor 4 (TLR-4), which plays a significant role in IBD, was involved in these pro-inflammatory S100A4 mediated processes (32,33).
Fibroblasts represent the key cell type in the pathogenesis of fibrostenosing/stricturing CD.Cunningham et al. investigated S100A4 in ex-vivo culture of resected ileum in patients suffering from fibrostenosing CD.The explant culture of tissue originating from the stricture showed a significant overexpression of S100A4 (30).
We are fully aware, that faecal biomarkers, including calprotectin and S100A12, might have advantage over the serum biomarkers for their anticipated higher sensitivity (being produced by the inflamed mucosa into the intestinal lumen) and higher specificity (as serum biomarkers can be elevated due to non-gastrointestinal disorders) (17).Nevertheless, not a very rare preference of serum sampling to the faecal one by patients with IBD provoked us to investigate S100A4 protein in the serum.
To our best knowledge, there are no studies on S100A4 in different clinical IBD phenotypes in the literature.
The aim of our prospective study was to investigate serum S100A4 protein in patients with IBD and to determine possible association of increased serum S100A4 with complicated forms of CD.

SUBJECTS
A total of 118 subjects were enrolled in the prospective study between 2009 and 2016: 93 patients with CD (44 men, 49 women, aged 22-79, mean 44 ± 14), 16 patients with UC (8 men, 8 women, aged 20-74, mean 39 ± 15) and 9 healthy controls (2 men, 7 women, aged 23-74, mean 52 ± 17).Control group consisted of individuals with normal findings on colonoscopy, who had negative history of IBD and/or colorectal neoplasia.A recent change in bowel habit and symptoms compatible with irritable bowel syndrome were the indications for colonoscopy in that individuals.No patient in control group had any serious comorbidities in relation to the serum S100A4 protein (including rheumatic disorders).
CD group was divided according to the Montreal classification (34) and descriptive statistics is provided in Table 1.
The duration of UC was 3-18 years, mean 10 ± 4, the duration of CD was 1-39 years, mean 15 ± 9. At the time of sampling, three patients with CD were without any treatment, 46 patients were on 5-aminosalicylates (5-ASA) and 44 were treated with immunosuppressive therapy (corticosteroids, azathioprine, anti-TNF, cyclosporine).A total of 18% (17/93) were treated with anti-TNF agents, 4/17 with adalimumab and 13/17 with infliximab.A total of six CD patients were treated with antibiotics (ciprofloxacin and/or metronidazole) including one patient in group with 5-ASA and five CD patients with concomitant immunosuppressive therapy.
Within the UC group, all of the enrolled patients had 5-ASA; 3 patients were also treated with azathioprine.No patient from UC group received anti-TNF therapy.

SERUM CONCENTRATION OF S100A4 PROTEIN: SAMPLE COLLECTION AND MEASUREMENT
Venous blood samples (total amount of 6 mL) were obtained before a standard colonoscopy at the Endoscopy Unit, 2nd Department of Internal Medicine-Gastroenterology. Samples were transferred immediately to the Institute of Clinical Biochemistry and Diagnostics at University Hospital Hradec Králové.Blood centrifugation followed and sera had been stored at −80 °C until the investigation was performed in December 2016.Serum concentration of S100A4 protein was investigated by means of Human Protein S100-A4 ELISA kit, the quantitative sandwich enzyme immunoassay technique (purchased from MyBio Source, San Diego, California, USA).

ETHICAL ISSUES
All subjects included in the study were given the necessary information and provided informed consent via a signed form.The project was approved by the Joint Ethical Committee (Charles University, Faculty of Medicine in Hradec Králové, University Hospital Hradec Králové).For all obtained data, all personal identification information was removed in compliance with the Czech laws for protection of confidentiality.

STATISTICAL ANALYSIS
Obtained data were tested statistically by means of descriptive statistic and non-paired t-test (normal distribution of data was confirmed) using STATISTICA software, version 13, 2013, Tulsa, OK, USA.

RESULTS
Serum S100A4 values were significantly higher in UC compared to controls, p = 0.019.Serum S100A4 were significantly higher in CD compared to controls, p = 0.007.No difference in S100A4 serum levels was revealed between UC and CD group, p = 0.771.See Graph 1.
In CD group, serum S100A4 was significantly higher in patients with all CD phenotypes compared to controls, p < 0.05.No difference in S100A4 was documented between particular subgroups of CD (divided according to behaviour of CD), p > 0.05.See Graph 2.
All CD patients, regardless if treated with anti-TNF agents, had significantly higher serum concentration of S100A4 compared to controls: CD group with anti-TNF: mean 154.7 ± 65.5 ng/mL, p = 0.049; CD group without anti-TNF: mean 154.3 ± 49.2 ng/mL, p = 0.005.There was no significant difference in serum S100A4 between CD patients with and without anti-TNF medication, p > 0.05.

DISCUSSION
In our study, serum S100A4 levels were significantly elevated in patients with ulcerative colitis and in patients with Crohn's disease compared to controls.We are convinced, that S100A4 is not a biomarker of IBD only, but it also plays a crucial role in the development of inflammatory process itself, presumably through the activation of TLR receptors and NF-κβ signalling pathway, too.In patients with rheumatoid arthritis, high levels of S100A4 were associated with a poor clinical response to infliximab and a high rate of anti-infliximab antibodies (35).Based on these data, similar situation could be expected in IBD patients and therefore a study evaluating S100A4 before the patients are started on anti-TNF therapy is being planned in our setting.No difference in serum S100A4 was confirmed between CD patients with and without anti-TNF in our study, however no firm conclusions can be drawn from this, as no one has studied the impact of anti-TNF therapy on serum S100A4 so far.S100A4 protein, known as calvasculin or metastatin-1 (16), was isolated by Ebralidze in 1989 and was considered to be involved in the metastatic tumour cell phenotype (36).A recent metaanalysis carried out by Liu et al. was in agreement: they reported that S100A4 over-expression correlates with tumour progression and poor prognosis of patients with colorectal carcinoma (37).
Recent studies have documented, that S100A4 does not play a role in metastatic cancer only, but it is also involved in inflammatory processes (38).Increased expression of calvasculin was documented in inflamed muscle tissue in patients with idiopathic inflammatory myopathies, where S100A4 may stimulate mononuclear cells to increase synthesis of pro-inflammatory cytokines (39).Significant upregulation of S100A4 was also observed in patients with rheumatoid arthritis: a study performed by Cerezo et al. documented that calvasculin induces inflammatory response (up-regulated production of TNF (tumour necrosis factor)-α, IL (interleukin)-1β and IL-6) of mononuclear cells via the TLR-4 (toll-like receptor) and by the activation of NF-κB signalling pathway (31).Aberrant TLR signalling is known to contribute to intestinal inflammatory processes in IBD and associated carcinoma (32,33).NF-κβ signalling cascade has also been shown to be involved in the development of colitis associated carcinoma (40).
Cunningham et al. found increased expression of S100A4 in fibroblasts and immune cells in the resected ileum in patients with stricturing CD.They also reported that the over-expression of S100A4 was induced by TGF-β1 (transforming growth factor) (30).Therefore we assumed, that patients with stricturing phenotype of CD will have higher serum S100A4 compared to those with a non-stricturing CD behaviour.Nevertheless, our data have shown that patients with all phenotypes of CD had significantly elevated serum S100A4 compared to controls and no significant difference in serum S100A4 was observed between stricturing and non-stricturing CD phenotypes.This might be explained by the fact, that over-expression of S100A4 in fibrotic processes is observed in tissue specific manner reflecting the local situation in the damaged tissue.We hypothesize therefore, that elevation of serum S100A4 protein (found in both, patients with UC and CD), mirrors rather inflammatory properties of calvasculin and its involvement in inflammatory processes in IBD.
Role of S100A4 in fibrotic processes has been investigated recently: Tamaki et al. studied effect of S100A4 on cardiac fibrosis and documented, that S100A4 knockout mice showed reduced interstitial fibrosis, decreased number of myofibroblasts, suppressed expressions of collagen and profibrotic cytokines in the left ventricle.The authors assume, that the S100A4 induces cardiac fibrosis through the modulation of p53 (28).Effect of S100A4 in the ischaemic myocardium seems to be different -protective and regenerative mainly: calvasculin decreased apoptosis of cardiac myocytes via the AKT signalling pathway (41) or via the ERK pathway (42).
We found the presence of highest S100A4 levels in patients with ileo-colonic and colonic forms of CD very interesting.Based on the literature, there is no definitive explanation for this.Nevertheless, (1) reported activation of NF-κβ signalling cascade by S100A4 (31), (2) suggestion that the NF-κβ signalling cascade may be the central mediator of gastrointestinal inflammation in IBD and malignancies (40), (3) known properties of metastatin -being a mediator of metastatic processes ( 27), ( 4) and known association of colorectal carcinoma with IBD (43) are compatible with our observation of higher levels of serum S100A4 in IBD patients with colonic involvement.
We are aware of the possible limits of our current study.On the basis of the available literature, where S100A4 was investigated either in the serum (51 studies; two of them in a relation to a gastroenterology disorderliver fibrosis and cirrhosis (44,45)) or in plasma samples (26 studies; two of them were accomplished in gastric and colorectal cancer (46,47)), we decided to investigate serum S100A4.All individuals enrolled into our study including all healthy controls and all IBD patients were investigated in the same manner.Therefore we assume, that our results are valuable and plausible despite we do not have any comparison between plasma and serum concentrations in our patients.Further, we did not investigate mRNA for S100A4 tissue expression (from the affected areas), therefore it was not possible to correlate it with the serum S100A4.The reasons why we did not investigate the samples were two: a) middle part of the small intestine may not be easily accessible for a routine endoscopy, b) resected specimens might not reflect situation in the whole intestine.
We hypothesize, that serum S100A4 can help to distinguish between IBD and non-IBD population and it possibly might serve as a phenotype marker of IBD (colonic involvement), however further studies are needed to follow in the near future.

CONCLUSIONS
Association of serum S100A4 with inflammatory bowel disease was confirmed.
No difference in serum S100A4 was observed between particular phenotypes of Crohn's disease (CD) including stricturing and non-stricturing forms of CD.
In CD, serum S100A4 was higher in patients with colonic and ileo-colonic involvement compared to patients with isolated small bowel involvement.
Tab. 1: The main characteristics of CD group.