We compared gene expression profiles to the c2_all collection of curated gene-sets from the molecular signatures database (version 2·5) [35]. This collection contains gene-sets that are experimentally derived, as well as from expert curated pathway databases. A preranked file was created, containing the average difference between AA and SS for each probeset, sorted from most up-regulated in SS GS-1101 in vivo to most down-regulated. We used the na28 annotation csv file from http://www.affymetrix.com to determine the gene symbol for each probeset and collapsed probesets to unique genes using the default, max_probe option, resulting in 18 600 unique genes. GSEA (version 2·0) [35] was run in preranked mode, using default

parameters (gene-set sizes between 15 and 500 leaving 1387 gene-sets, 1000 permutations, images on the top 50 gene-sets). We used mRNA extracted from distal colons obtained from four

SS and four AA mice for RT–PCR confirmation of our gene expression study. Reverse transcription to produce cDNA was performed using RT2 First Strand Kits (SA Biosciences, Frederick, MD, USA), according to the manufacturer’s instructions. RT–PCR was performed utilizing the LightCycler 480 real-time PCR system (Roche Applied Science, Mannheim, Germany) with RT2 SYBR green PCR master mix according, to the manufacture’s protocol (SA Biosciences). Predesigned primers for genes of interest (slpi, s100A8, lbp, CD68, IL18R1, IL33, ccl8, cxcl10, ccl12, GSK 3 inhibitor pf4, ccl5, ccl7, fpr1 and ccr5) were obtained from SA Biosciences. For reference genes we evaluated three candidates, β-actin, β-glucuronidase and 18S rRNA. Beta-glucuronidase was selected based on similar expression patterns to most of our genes of interest and also because it was expressed invariantly between the groups. Hence, each

sample was normalized on the until basis of its β-glucuronidase content. Thermal cycling was performed as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Each assay was performed in duplicate. The quantification points generated from quantitative RT–PCR (qRT–PCR) were normalized against a reference gene using this formula: normalized value of gene of interest with β-glucuronidase = 2–(QPGOI–QPRG), where QP = quantitative point, GOI = gene of interest and RG = reference gene (i.e. β-glucuronidase). We used the same 14 genes that we used for RT–PCR confirmation of our microarray study. We collected distal colonic samples from 3 days, 14 days and 28 days after the last (second) surgery. For each of the time-points we used four SS and four AA mice. The colons were collected, stored and processed for RT–PCR as described earlier. Group comparisons were analysed using the Mann–Whitney U-test with GraphPad Prism (Graphpad Software, San Diego, CA, USA). The differences were considered to be significant if P < 0·05.