Trichuris muris-induced colitis represents a tractable murine model for understanding the patho-biological mechanisms of chronic intestinal inflammation
[9, 19]. The use of Quantitative Trait Loci (QTL) mapping based on continuous phenotypic variation has proved a useful technique in many murine polygenic traits including intestinal inflammation
[12, 20, 21]. Yet, of more than 2000 QTL documented within the mouse genome database
 fewer than 1% of studies have actually been characterized at a gene or molecular level, due to the small effect size of the susceptibility locus in question (<10% penetrance), or the large interval size defined
. New multi-factorial approaches have been discussed in the literature
 and demonstrate that understanding complex genetic traits requires an integrative analysis.
Specific steps were taken in our experimental design to consider recent reports concerning the QTL/microarray approach in the identification of QTL candidate genes
. First, QTL were defined with regards to experimental phenotype (pQTL), and correlated with transcriptional expression activity in parental strains. Second, the use of high density Affymetrix exon array, which targets approximately 40 exonic probes per gene, overcame any problem of potential allelic-biased probe binding. Third, a hypothesis-free pathway analysis, backed up by additional text-mining, was employed in the secondary filtering of potential candidate genes to reduce bias. Fourth, any genes lacking polymorphisms (coding and non-coding) between parental strains were excluded from analyses, and lastly, positional overlap with a previously replicated major colitis susceptibility quantitative trait locus (Cdcs1) prioritised Tm3 for targeted analysis.
With regards to this shared locus, Cdcs1 on chromosome 3 was first noted in a QTL study of spontaneous colitis using IL10 deficient mice
. This locus has been shown to contain at least 3 distinct regions (Cdcs1.1, 1.2 & 1.3) that contribute to a severe colitic phenotype
[14, 15]. Interestingly, all three regions contribute to caecal and proximal colonic inflammation strongly suggesting that this locus is a colitis ‘hotspot’ for susceptibility and/or regulation. Here we show complete overlap of Tm3 with Cdcs1.1 (Figure
2). Although NF-kB1 has been suggested as a candidate gene for the Cdcs1 locus, it is clear that it is not responsible entirely for the severe pathology observed
. To date, FcgR1 remains the key candidate gene described in the Cdcs1.1 region
[14, 15] and is corroborated by our findings. Additional association with colitis susceptibility in Gnai2-/- mice
 suggests that this locus may govern key inflammatory pathways in disease development, irrespective of trigger. QTL mapping specifically highlighted the Cdcs1.3 region in the spontaneous colitis and colorectal cancer development of TRUC mice
. However, more distal colonic disease (distal third of the colon) or the potential for malignant transformation may not be represented at this sub-locus.
We have shown that at least 6 biologically significant and polymorphic candidate genes lie within the Cdcs1.1 autosomal region. Importantly, 4 of these candidate genes are key in pathways relevant in the context of human Crohn’s disease (FcgR1, Vav3, Vcam1 and Ctss), a disease with highly similar pathology to both the IL10 deficient and T. muris models of colonic inflammation
[9, 24]. The remaining 2 genes are highly polymorphic and known to be important in inflammation (RORc, Ptpn22). As individual candidate genes, each demonstrates interesting biological functionality that could play a role in mucosal inflammation. For instance, FcgR1 codes for a high affinity IgG receptor, key to IgG2a-induced phagocytosis and antigen specific immune responses. In the mouse, FcgR1 has been associated with autoimmune disorders such as rheumatoid arthritis and bacterial infection
. In humans the closely related FcgR2a and FcgR3 have been associated with IBD
. The protein tyrosine phosphatase gene (Ptpn22) is of particular interest, as in humans a mis-sense SNP (C1858T) has already demonstrated strong correlation with rheumatoid arthritis
, type-1 diabetes mellitus
, and other autoimmune disease
. Interestingly, the C1858T gene variant is not associated with the establishment of human Crohn’s disease
 and may even represent protection
. In our study, Ptpn22 demonstrated progressively increased expression within the colonic tissue of susceptible mice following the establishment of colitis.
The unbiased approach we have used to select candidate genes has also highlighted a gene whose currently assigned pathway (circadian rhythm) does not overtly relate to mucosal inflammation. The Retinoic acid-related orphan receptor-C (RORc/RORγ) gene encodes for RORγt (RORγ2), a lineage-specific transcription factor of CD4+ TH17 cell differentiation
. Excessive TH17 cell activity has been implicated in both autoimmune
 and inflammatory bowel diseases
Finally, Vav3 was the primary candidate revealed by integrative pathway and SNP analysis and is of particular interest, as in six week old Vav1/2/3 triple knockout mice altered gut enterocyte differentiation and morphology has been shown, along with spontaneous colitis and ulceration in the caecum and ascending colon
. Vav3 is also involved in at least 7 known biological pathways, all of which could play a role in mucosal homeostasis and regulation. Some of these pathways involve other candidate genes in this region, for instance FcgR1 (Fc-gamma receptor mediated phagocytosis), Vcam1 (leukocyte transendothelial migration, focal adhesion) and Ptpn22 (negative regulation of T-cell receptor signalling)
[36, 37]. We hypothesise therefore that Cdcs1 is in fact a ‘colitis hotspot’ containing several genes which if dysregulated through genetic variation, could adversely affect gut inflammation. It is possible that the specific candidate genes for each colitis model are not the same. However, the biological interaction between genes at this locus, demonstrates the importance of Cdcs1 and why this region appears in unrelated models of gut inflammation.
Interestingly, Tm3 (Cdcs1) does not correlate with any known nematode infection susceptibility QTL, but instead appears exclusive to colonic inflammatory disease.
For instance, expulsion and resistance to the small intestinal nematode Heligmosomoides bakeri in mice has been characterized at murine chromosome 1 and 17
 corresponding to Tm1 and Tm17. Similarly, a study of Trichinella spiralis infection in rats, which causes acute and transient small bowel inflammation, identified a single significant QTL region homologous to the murine chromosome 1 locus (Tm1)
. Lastly, resistance to small bowel and abomasum/gastric nematode infections of sheep, have highlighted a number of suggestive QTL
[40–42], homologous to Tm17, and downstream of Tm10. All studies demonstrated that resistance/susceptibility to GI nematode infection is under multi-genetic control, with MHC and non-MHC loci important in outcome
. However, these studies also highlight the importance of the Cdcs1 locus with the establishment of a large bowel inflammatory phenotype, separate to precise anti-parasitic mechanisms.
In conclusion, we have corroborated three previously published studies which associate the locus Cdcs1 with colonic mucosal inflammation in the mouse. Furthermore, we have shown that in the AKR and BALB/c, genetic variation in this region has the potential to affect mucosal homeostasis through several different pathways. Most importantly, we have demonstrated that an unbiased integrative analysis can be beneficial in candidate gene identification and prioritization, particularly cis-regulated genes, even in large regions. This approach is particularly useful for hypothesis generation, and has positionally implicated Vav3 as a biologically relevant gene candidate in colitis.