Paediatric inflammatory bowel disease - bench to bedside and nationwide : a detailed analysis of Scottish children with IBD

Abstract

The inflammatory bowel diseases (IBDs) are a group of chronic conditions affecting the gastrointestinal tract, often presenting with non-specific clinical features such as abdominal pain, weight loss and diarrhoea. Approximately 25% of patients are diagnosed with IBD in childhood. For epidemiological studies, previously collected (1990-1995) and original (2003-2008) Scottish incidence data were used to determine national trends in newly diagnosed paediatric IBD (PIBD). A smaller, geographically defined, prospective 14-year cohort (1997- 2011) in South-East Scotland (SES) was used to assess regional trends in incidence, point prevalence, disease extent, medication use and PIBD surgery rates in 326 children. For the detailed analysis of the role of ICOSLG and CRP in Scottish children with PIBD, haplotype-tagging of both genes in 448 children (and their parents) registered on the Paediatric Inflammatory bowel disease Cohort and Treatment Study (PICTS) database was performed. Further clinical information from this database and previously gathered adult mRNA microarray data were also used to inform the analysis. For the faecal calprotectin (FC) case-control study, all PIBD patients diagnosed in SES between 01.01.05 and 31.12.10 (aged 1- 17yrs) with a FC performed during initial workup were identified; controls were matched non- IBD patients who had similarly undergone endoscopy with a referral FC level available. The systematic review and meta-analysis of FC case-control studies was performed with keywords relating to IBD and calprotectin in electronic resources from 1946 to May 2012. Inclusion criteria were studies that reported FC levels prior to the endoscopic investigation of IBD in children less than 18 years old. Laboratory work used newly derived HEK293 and HCT116 cell lines stably expressing wild-type NOD2 and the CD-associated NOD2 frameshift mutant, as well as utilising previously derived HEK293 and HCT116 cells stably expressing green fluorescent-labelled protein LC3 during the assessment of autophagy. Western blot, immunofluorescent microscopy and flow cytometry were used for analysis. There was a significant rise in PIBD incidence in Scotland since the early 1990s, with 260 new cases between 1990-1995 (4.45/100,000/year) and 436 in the 2003-2008 epoch (7.82/100,000/year) (p<0.001). A five-fold increase in Crohn's disease (CD) in the last 40 years was also demonstrated. SES was shown to have the highest recorded PIBD incidence rate in the UK for the six-year epoch from 2006-2011 (9.50/100,000/year) with a significant rise in ulcerative colitis (UC) to 2.67/100,000/year (p=0.010). Point prevalence rates for PIBD in SES had also risen significantly to 41.2/100,000 between the 2000-2005 and 2006-2011 epochs (p=0.016). With a follow up of 1577 patient years, the severe phenotype in children with PIBD was confirmed; 34% of children with CD presented with pan-enteric disease (44% at follow up), and 76% of children with UC had pancolonic disease at diagnosis (81% at follow up). 26% of patients required methotrexate and 18% were exposed to infliximab/adalimumab, with the time to first exposure of both significantly lower in children diagnosed between 2006-2011 (p=0.001 and p<0.001 respectively). A total of 70% of children were exposed to azathioprine and 20% underwent IBD-related surgery. Using a haplotype-tagging approach and transmission disequilibrium testing (TDT) in 230 PIBD case-parent trios there was significant overtransmission of the rs8126734-A single nucleotide polymorphism (SNP) in ICOSLG following correction (p=0.0467). In the CD TDT analysis the same SNP was overtransmitted (p=0.0084). The strongest susceptibility signal was evident across the two marker haplotype rs762421-A / rs8126734-G (p=0.0072), suggesting that the 3-prime untranslated region in ICOSLG may be targeted for deep sequencing. mRNA microarray data from adult patients showed downregulation of ICOSLG expression in the ascending colon (p=0.023) and upregulation in the descending colon (p=0.0351) in uninflamed biopsies from CD patients and non-IBD controls; no difference in gene expression was shown in UC patients. Using a similar approach, the A allele of two SNPs tagging CRP showed significant over-transmission to affected IBD patients after correction (rs1417938, p=0.006; rs1130864, p=0.015). The six-marker haplotype (ACACAC) showed significant distortion of transmission to affected individuals (p=8x10-4). CD and UC patients demonstrated differences in rs1205 genotype (p=0.0085) and CRP haplotype (p=0.0024), with the influence of the rs1205 SNP on response to anti-tumour necrosis factor-alpha therapy also shown (p=0.021). During the FC case-control study significantly elevated FC levels at diagnosis were demonstrated compared to controls (1265 μg/g vs 65 μg/g; p<0.001). FC also outperformed commonly used blood parameters (e.g. CRP, ESR, platelets), with an area under the curve of 0.93 (95% CI 0.89-0.97) and good sensitivity (0.93 [95% CI 0.86-0.98]) and specificity (0.74 [95% CI 0.64-0.82]) when values above 200μg/g were used. FC levels were not influenced by disease location in CD or UC. The systematic review and meta-analysis highlighted the often poor methodological quality of previous studies and concluded that across all studies FC had a pooled sensitivity of 0.98 (95% CI 0.95-1.00) and pooled specificity of 0.68 (95% CI 0.50-0.86) for PIBD at diagnosis. Characterisation of cells stably-expressing wild-type NOD2 or the CD-associated NOD2 frameshift mutation demonstrated increased cell proliferation compared to empty vector, and an accentuated apoptotic response to serum starvation. The NOD2 frameshift protein had a shorter half-life (at 11 hours) than the wild-type protein, with degradation of the NOD2 protein shown to be mediated through a proteasome-dependent pathway, possibly through lysine residues on the CARD domain. Following the establishment of a robust method of assessing autophagy in a cell culture system, experimental work showed that muramyl dipeptide-induced autophagy is unlikely to signal through the mammalian target of rapamycin, with the intermediate filament vimentin shown to be intimately involved in this pathway; the vimentin gene (Vim) was also shown to be a candidate susceptibility gene for CD. Using a panel of PIBD drugs azathioprine was shown to induce autophagy in a dose-dependent manner through an mTOR-dependent, ERK-independent pathway. It can be seen that with the increasing incidence and prevalence of PIBD in Scotland that a greater understanding of epidemiological trends, the role of genetic susceptibility, the optimal use of biomarkers and translational functional biology are all needed to understand further the aetiopathogenesis of PIBD. This future work will undoubtedly help to inform service design and the clinical care pathways utilised to provide the best care for children in addition to targeting pathways for potential drug development, with these measures helping to prepare for the increasing disease burden generated by PIBD.