METHODS

Scanning Electron Microscopy (SEM)

Coded samples were examined. Biopsies were prepared for electron microscopy using standard methods (1) and examined in a LEO 360 IXp Scanning Electron Microscope (LEO Electron Microscopy, Cambridge, UK). Images from each biopsy were recorded at standardized magnifications. Criteria for characterizing the microorganisms as bacteria comprised uniformity in length and diameter, no visible septations, no budding of pseudospores, and a growth pattern with no branches. Measurements of bacterial length and diameter were carried out directly during microscopy, and bacteria were classified as rodshaped if the length/diameter ratio was > 1.5 and as cocciform if the ratio was ≤ 1.5.

Transmission Electron Microscopy (TEM)

Biopsy specimens were fixed by incubation with 2.5% glutaraldehyde in 0.15 M sodium phosphate buffer, pH 7.4, for 5 h. Following three rinses with sodium phosphate buffer, the specimens were post-fixed for 1 h in 1.33% osmium tetroxide in sodium phosphate buffer, rinsed in deionized water, dehydrated using increasing concentrations of acetone and embedded in Epon-Araldite mixture (Fluka, Buchs, Switzerland). Ultrathin sections were stained with 5% uranyl acetate in methanol for 15 min, followed by lead citrate for 5 min. The sections were examined in a Zeiss EM 109 electron microscope at 75 kV.

Bacterial Culture and Identification

The biopsies were weighed, homogenized and diluted ten-fold in Fastidious Anaerobe Broth medium (Lab M, Lancashire, UK) and immediately plated onto selective and non-selective agar plates. The non-selective media used were blood agar (Columbia blood agar-base; Neogen Corp. Acumedia, Lansing, MI) with 5% citrated horse blood for cultivation of aerobic and anaerobic microorganisms and chocolate agar (GC-agar; Difco; Becton Dickinson, Sparks, MD) as a non-selective medium for capnophilic bacteria. Rogosa SL agar (Difco) was employed for detection of lactobacilli, Mitis-Salivarius (MS) agar (Difco) for isolation and differentiation of -streptococci and Veillonella agar (Difco) for detection and identification of Veillonella sp. Blood agar and MS agar plates were incubated at 37 °C aerobically for 24-48 h and chocolate agar was incubated aerobically in 5% CO2 for 48 h. The other set of blood agar plates, Rogosa SL agar and Veillonella agar were incubated in anaerobic atmosphere (10% H2, 5%CO2 in N2) at 37 °C for 48-72 h. After incubation all different colony types were counted and isolated in pure cultures on suitable agar media for further analysis. Identification to genus level was done by Gram-staining and standard biochemical tests. Anaerobic bacteria were classified according to Wadsworth-KTL anaerobic bacteriology manual guidelines (2) and by use of the rapid ID 32A test kit (Biomérieux, Marcy-l’Etoile, France). Selected isolates were subjected to full-length sequencing of the 16S rRNA gene. The lower limit of detection was 102 colony forming units (CFUs)/g.

LEGEND

Supplementary Figure 1

Phylogenetic tree based on 2,247 bacterial 16S rDNA sequences from proximal small intestine biopsies of 63 Swedish children. The tree was generated by Neighbor-Joining-Analysis (MEGA 4.0) showing 6 phyla (A) and 162 Genus Level Operational Taxonomic Units (GELOTUs), (B-D). Alignments and classification were done with BLAST (the Basic Local Alignment Search Tool) and RPD II (Ribosomal Database project II). The GELOTU designation followed by the name and GenBank accession number of the best-matching sequence, percent sequence identity and finally the number of sequences of this type is shown at the end of each branch. Three GELOTUs, one closely related to Treponema sp, are included in the phylum Bacteroidetes (C). Similarly 4 GELOTUs representing other phyla, including TM7, are included in the phylum Proteobacteria (D). The scale bar represents the percentage estimated sequence divergence. The numbers above each node are confidence levels (%) generated from 1000 bootstrap trees. Eighteen biopsies were from clinical controls, 3 of which were from healthy siblings of CD patients. Thirty-three biopsies were from children with CD collected between 2004 and 2007. These fresh biopsies were washed repeatedly with PBS containing DTT to dissolve the mucus layer. Of these: 19 biopsies were from untreated CD patients, 12 from CD patients treated on a gluten-free diet and 2 from challenged CD patients. The remaining 12 biopsies were “historical” biopsies from children with CD born during the CD epidemic and had been stored frozen at -80 °C. They were not washed.

REFERENCES

1. Forsberg G, Fahlgren A, Hörstedt P et al. Presence of bacteria and innate immunity of intestinal epithelium in childhood celiac disease. Am J Gastroenterol 2004;99:894-904.

2. Jousimies-Somer H, Summanen P, Citron DM et al. Wadsworth-KTL anaerobic bacteriology manual 6th ed. Star Publishing: Belmont, CA. 2002.