Moderate performance of serum S100A12, in distinguishing inflammatory bowel disease from irritable bowel syndrome

Inflammatory bowel disease (IBD) remains a diagnostic challenge for the clinician who faces a repertoire of diverse and fluctuating symptoms and signs. It is mainly due to a subtle or atypical presentation that in some cases the discrimination of Crohn's disease (CD) or ulcerative colitis (UC), from other diseases of the alimentary tract, especially from irritable bowel syndrome (IBS), becomes rather problematic [1]. On the other hand, although several markers are used for diagnostic purposes [2] a simple, economic and reliable test, as an alternative to more complex procedures for IBD diagnosis, is still under investigation. It is in this eager environment that S100A12 has emerged as a promising marker for both diagnosis and disease activity in IBD [3, 4]. The S100A12, also known as calgranulin C, is a member of the S100 protein family which, in humans, consists of twenty five EF-hand (a helix-loop-a helix), calcium-binding proteins, of which the vast majority is in a homodimer, heterodimer ie S100A8/A9, or more complex form [5]. S100A12, like S100A8/A9 (calprotectin), is considered phagocyte-specific, exhibits proinflammatory properties and has already been linked to many different diseases of inflammatory origin, including IBD [3, 4]. Several studies using the determination of S100A12 in feces, revealed a significant association between fecal S100A12 levels and IBD,[6–8] and especially active disease [8], an association that could also aid differential diagnosis by distinguishing IBD from IBS [9]. A limited number of studies was also conducted to examine the diagnostic potential of the S100A12 levels in serum. To our knowledge, three studies have been published so far, two carried out in pediatric [7, 10] and one in adult populations [8]. The present study, however, is the first testing the utility of serum S100A12 levels in discriminating IBD from IBS in adults, while at the same time examining serum S100A12 titers with respect to disease type (UC or CD) or activity, traditional markers of disease activity, patients' characteristics and treatment modalities.

IBD has been long ago recognized as a systemic inflammatory entity and as such, it is anticipated to induce changes exceeding the boundaries of bowel mucosa, being reflected in a broader spectrum of tissues, including blood [16–19]. Examples of such changes are the fluctuations in the levels of CRP, SAA, TNF-α, Interleukins [16], S100 proteins [17], metalloproteinases [18], angiogenins [19] etc. Some of these substances have been introduced as inflammatory markers and used in clinical practice for diagnostic purposes in patients with suspected or confirmed IBD. The sufficiency of these markers, however, has been challenged repeatedly due to a moderate performance, making the study of new markers in IBD mandatory [17, 20, 21]. There are numerous studies examining the significance of several fecal, serum or mucosal markers including members of the S100 protein family such as S100A8/9 and S100A12 [3, 4, 8, 22, 23]. Since the vast majority of available literature is focused on S100A8/9 [22, 23] and to a lesser extent on S100A12 [6–10] and their role in IBD, a subsequent gap in the literature regarding S100A12 studies in IBD patients originated. Moreover, two serum S100A12 studies were carried out in children [7, 10] while only one was conducted in an adult population [8].

Over the past years, several markers emerged to facilitate the differentiation of IBD from IBS. The most promising among these were fecal S100A8/A9 [24–26] and more recently fecal S100A12, which has been shown to discriminate IBD from IBS with a sensitivity of 86% and specificity of 96% [9]. In our study, a significant elevation in serum S100A12 levels correlated well with IBD but not with IBS, thus allowing the distinction between the two entities. The performed analysis also focused on the determination of a cut-off for IBD prediction that would exhibit the highest possible sensitivity and specificity. This optimal cut-off was estimated at 54 ng/mL and was shown to predict both CD and UC with a sensitivity of 66.7% and a specificity of 64.4%, in the population of the study. Although, this performance is superior to the one reported by Sidler et al (21% sensitivity and 81% specificity) [7] still remains moderate and evidently, lower than that reported for fecal S100A12 [9], probably due to the fact that fecal stream is in direct contact with the inflamed mucosa. Perhaps, a better diagnostic performance could be established while assessing simultaneously S100A12 and other IBD-related diagnostic markers, in serum, a model already applied when anti-glycan antibody determinations are carried out [27].

As shown in the results section, a significant correlation between S100A12 serum levels and disease activity or factors (ie treatment) that can alter the activity could not be established. Both UC and CD patients with active disease seemed to have higher values of S100A12 in serum, compared to IBD subjects with inactive disease but this difference was not statistically significant. On the other hand, this "flaw", could also be regarded as a beneficial characteristic, as it implies that the value of serum S100A12 for discrimination of IBD from IBS cannot be challenged by the presence of inactive disease. Although our findings could be the result of the rather small number of patients recruited in the study, they are in accordance with previous reports on the inadequacy of serum S100A12 as a marker for monitoring IBD activity [9], although in an earlier study, S100A12 concentration in serum has been shown to differentiate active from inactive IBD [8]. In that study, however, the levels of serum S100A12 in UC patients with inactive disease were comparable to those in healthy controls [8].

According to our results, serum S100A12 determination could not be used to predict disease extent in UC patients, since the S100A12 serum values were not significantly different in patients with proctitis, left-sided colitis or pancolitis. Similarly, by relying on S100A12 serum levels, disease location could not be predicted, despite the fact that, at first glance, patients with ileocolitis or ileitis seemed to exhibit higher values of serum S100A12, compared to those with CD colitis alone. The data presented by Foell et al, on the other hand, were indicative of a predominant mucosal release of S100A12, from the inflamed colon of CD patients. In the same study, S100A12-stained neutrophils which adhered to the endothelium of blood vessels were detected in the CD-affected terminal ileum [28]. Although no hard evidence favoring this assumption are available, it would be useful to examine the possibility that the S100A12 in serum of CD patients may derive mainly from neutrophils migrating through the blood vessels into the diseased ileum or that ileal inflammation triggers a more "systemic" type of response.

As for the study of S100A12 serum levels with respect to the presence of any extraintestinal manifestations, no significant variations became evident. S100A12 serum concentrations were comparable between IBD patients with or without extraintestinal manifestations.

The effect of current smoking status on serum S100A12 concentration, in IBD individuals was also examined, since smoking has been traditionally regarded as a factor which could improve UC and worsen CD [29]. As shown above, the levels of S100A12 in serum of current smokers, diagnosed with CD appeared to be elevated, compared to CD non-smokers. In the UC group, the levels of serum S100A12 seemed to be higher in non-smokers than in those found in current smokers. These findings, although within reason, lack credibility since, neither statistical significance nor trend could be reached for any of the recorded smoking-related associations.

Similarly, all treatment modalities, leading to a subsequent disease remission, did not seem to have a noticeable effect on serum levels of S100A12 in the IBD group. This result is partly in accordance with an earlier study in which corticosteroids did not alter significantly S100A12 serum levels of IBD adults [8]. The impact in S100A12 serum levels of infliximab-treated IBD subjects recorded in that study [8] was not verified in ours.

What appears to be of interest is that a statistically significant association was found between serum S100A12 levels and the well-known markers of inflammation, CRP and SAA, in the absence of a similar association with IBD activity-determined by conventional inflammatory indices. Before discarding serum S100A12 as a potential marker for the monitoring of IBD activity it would, perhaps be better if possible associations with other indices [30, 31] were also evaluated.

At this point a potential study limitation has to be underlined. The serum S100A12 levels determined in our population, using the prototypic ELISA described above, were lower compared to IBD-oriented studies, using different ELISA assays [7, 8, 10]. According to the manufacturer (CIS bio international, France) this is probably the result of discrepancies, originating from between-studies differences in the pre-analytical phase, as well as the inability of the specific kit to perform equally well when different blood-containing mediums were used. In line with this notion are the results from a study on the impact of diverse sample handling conditions -incubation in tubes with gel/anticoagulants/calcium, dilution with calcium containing buffers, different storing conditions, repeated freeze-thaw cycles etc- revealing great fluctuations in the S100A12 levels, even in blood samples obtained from healthy individuals [32]. According to the authors, however, the most reliable blood sample for S100A12 determinations would be serum stored in gel-containing tubes, as in our study [32]. The use of the most suitable -or least unsuitable- type of blood-derived sample, along with the establishment of identical for all samples "handling" throughout the critical phase that preceded analysis, minimise the odds that the between-subjects differences recorded in our study are due to the use of the specific ELISA.

In conclusion, it is evident that the upregulation of S100A12 is not confined to the boundaries of an IBD-occupied intestine but is also reflected systemically and subsequently detected in serum. This increase in serum S100A12 is associated with IBD and is also well correlated with the "classic" markers of inflammation CRP and SAA. On the other hand, since the diagnostic utility of serum S100A12 was moderate, when used alone, its use in a "palette" of established serological markers, might actually lead to an improved overall diagnostic performance. Thus, it seems that further research in larger populations is mandatory, in order to verify these results, examine the association of serum S100A12 with disease or patients' characteristics and specify whether serum S100A12 could be added in the existing armamentarium, used for the diagnosis of IBD.