GI highlights from the literature

Dietary influences on the gut microbiome

▶ Wu GD, Chen J, Hoffman C, et al. Linking long-term dietary patterns with gut microbial enterotypes. Science 2011;334:105–8.

Mounting evidence suggests that the gut microbiota may influence the development of human diseases such as autoimmunity, obesity, and inflammatory bowel disease. Diet may serve as a modifiable factor that can influence the composition of the host microbiome. Wu et al recently reported findings of a cross sectional study of 98 healthy volunteers investigating dietary influences on the gut microbiota. Short-term and long-term dietary intake was assessed by the use of food frequency questionnaires. Massively parallel sequencing of 16S ribosomal bacterial DNA sequences was performed on DNA isolated from stool samples. Gut bacterial populations were clustered into ‘enterotypes’, largely as a result of the relationship between the two dominant bacterial genera. Enterotype was associated with long-term, but not short-term, nutrient intake. There was a significant association between the Bacteroides enterotype and consumption of animal protein and saturated fat. The Prevotella enterotype was, in contrast, associated with vegetarian, carbohydrate-based diets. In a controlled-feeding experiment, 10 subjects were randomised to receive either a high-fat/low-fibre diet, or a low-fat/high-fibre diet. Stool samples were then collected over 10 days. Interestingly, changes in the gut microbiome occurred rapidly but were small in comparison to the magnitude of inter-individual variability. No stable switching between enterotypes occurred. Whether gut enterotype is associated with specific disease predisposition in humans remains to be seen. If such associations exist, stable modification of enterotype may represent a prophylactic or therapeutic target in clinical practice.

Is multiple sclerosis a disease of the gastrointestinal microbiota?

▶ K Berer, M Mues, M Koutrolos, et al. Commensal microbiota and myelin autoantigen cooperate to trigger autoimmune demyelination. Nature. Published Online First: 26 October 2011. doi:10.1038/nature10554.

Berer et al investigate multiple sclerosis (MS) in a mouse model of experimental autoimmune encephalomyelitis (EAE). These mice express an antigen receptor for MOGpeptide 92–106 in their CD4₁ T cells with 35%–90% spontaneously developing EAE similar to MS. The initiating step involves infiltration of the CNS by T cells, which then recruit anti-MOG B cells. Although mice raised in specific germ-free conditions develop EAE within 3–8 months, germ-free mice never do. This was not due to an immunodeficiency because germ-free mice develop EAE when re-colonised. Immune competence was confirmed by immunisation of germ-free mice with MOG resulting in EAE. Critically, EAE was due to auto-immunisation rather than microbial mimicry. MOG knock-out mice did not develop auto-antibodies despite normal colonisation but did when immunised against MOG. Finally, B-cell recruitment appears to occur as a result of exposure to MOG in the cervical lymph nodes, evidenced by the accumulation of labelled MOG-reactive B cells within this site on their introduction. Notably, the crucial step for EAE development in mice is the presence of the gastrointestinal microbiota. Studies are required in humans to explore the role of the commensal gut microbiota in MS and other autoimmune conditions.

Fusobacterium, the next Helicobacter pylori?

▶ AD Kostic, D Gevers, CS Pedamallu, et al. Genomic analysis identifies association of Fusobacterium with colorectal carcinoma. Genome Res. Published Online First: 18 October 2011. doi:10.1101/gr.126573.111.

▶ M Castellarin, RL Warren, JD Freeman, et al. Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma. Genome Res. Published Online First: 18 October 2011. doi:10.1101/gr.126516.111.

In addition to emerging association between the gut microbiota and extra-intestinal disease, significant interest has focused on the potential relationship between a dysbiotic mircobiota and pathogenesis of colorectal cancer (CRC). Two papers on this topic, by independent groups, were published simultaneously in a recent edition of Genome Research. Kostic et al analysed DNA from nine CRC and normal tissue pairs by whole genome sequencing. Using a metagenomic biomarker discovery approach, they observed enrichment for Fusobacterium and Streptococcaceae sequences in tumour samples. In a larger sample of 95 paired specimens, 16S sequencing confirmed the relative abundance of Fusobacterium, but not Streptococcaceae DNA in tumour tissue. Quantitative real-time PCR analysis of 98 paired samples confirmed an absolute increase in the amount of Fusobacterium DNA, as opposed to a relative increase due to depletion of other bacterial species. Quantitative PCR detected Fusobacterium DNA in two of 11 resected CRC metastases but not in any of 59 CRC cell lines examined. These findings were complimented with fluorescent in situ hybridisation data.

Castellarin et al used RNA sequences for their initial screen in order to detect live, transcribing, microbes. Following host sequence subtraction, Fusobacterium sequences were again over-represented in tumour compared to normal mucosa. This observation was confirmed by quantitative PCR in a validation set of 99 paired samples. These studies present exciting novel findings on the potential microbial contribution to the risk of developing CRC. Further studies are required to investigate whether Fusobacterium has a causal role in the initiation and progression of CRC.

Barrett's oesophagus—continuing the debate on surveillance endoscopy

▶ Hvid-Jensen F, Pedersen L, Drewes AM, et al. Incidence of adenocarcinoma among patients with Barrett's esophagus. N Eng J Med 2011;365:1375–83.

Endoscopic surveillance for Barrett's oesophagus is a contentious topic. Although the incidence of oesophageal adenocarcinoma has risen, the vast majority do not occur in individuals with a prior diagnosis of Barrett's oesophagus. A large population-based Danish cohort study involving over 11 000 patients with Barrett's oesophagus, diagnosed between 1992 and 2009 and with a median follow-up of 5.2 years, identified 197 cancers overall. One hundred and thirty-one of these were diagnosed in the first year after index endoscopy and only 66 new cancers were found during subsequent years of follow-up. The overall incidence rate of adenocarcinoma following an initial diagnosis of Barrett's oesophagus was 2.9 per 1000 person-years (95% CI 2.6 to 3.4). After excluding all cancers diagnosed in the first year after initial diagnosis of Barrett's, the incidence rate of adenocarcinoma fell to 1.2 cases per 1000 person-years equating to an annual risk of 0.12%, much lower than the previously estimated 1.5%. Low-grade dysplasia was associated with a slightly higher rate of annual progression—0.5% per year, but in individuals without dysplasia the rate of development of adenocarcinoma was only 1 per 1000 person-years. The overall RR of adenocarcinoma was 11.3 (95% CI 8.8 to 14.4) compared to the general population.

This data may have potential implications for clinical practice and the following points are worthy of consideration. With the low rates of cancer development in the absence of dysplasia, is endoscopic surveillance necessary for this group? Second, the rate of cancer appears to be significantly higher in the first year after index endoscopy, therefore, another strategy may be to consider increasing the number of endoscopic procedures carried out in year one, with subsequent surveillance limited to those with evidence of dysplasia. The debate continues.

Interferon for hepatitis B; how much, how long and how to stop

▶ Liaw Y-F, Jia J-D, Chan HLY, et al. Shorter Durations and Lower Doses of Peginterferon alfa-2a Are Associated with Inferior Hepatitis B e Antigen Seroconversion Rates in Hepatitis B Virus Genotypes B or C. Hepatology 2011;54:1591–9.

Interferon (IFN) has an important role in the treatment of e-antigen positive hepatitis B due to efficacy at inducing seroconversion and prolonged reductions in viral load after a relatively short course of treatment. However, the optimal dose and duration of treatment is unclear leading to variations between recommendations from expert bodies (including NICE and hepatology organisations) and the approved licensed dose. In addition (in contrast to hepatitis C) we lack reliable early predictors of treatment outcome to allow prescribers to stop ineffective treatment early. Laiw et al present a large (n=792) multicentre 2×2 factorial RCT comparing two doses (90 μg vs 180 μg) and durations (24 vs 48 weeks) of IFNα-2a in e-antigen positive hepatitis B. The primary outcome measure was e-antigen seroconversion 6 months after treatment cessation. Seroconversion rates were better in both the higher dose (29.3% vs 19.7%) and longer treatment (30.9% vs 18.4%) arms. These effects were independent with the best results with 180 μg for 48 weeks (36.2%) and the worst with 90 μg for 24 weeks (14.2%). Similar advantages were reported for secondary outcome measures including ALT normalisation and viral load suppression. No patients in any arm achieved seroconversion if their quantitative s-antigen level exceeded 20 000 IU at week 12 (negative predictive value 100%). 18.4% of patients would have stopped treatment for futility at 12 weeks if this rule were applied. This phase 4 study is one of the most important investigations of IFN in e-antigen positive hepatitis B since the initial licensing reports. It confirms optimal dose and duration of therapy and suggests a highly reliable stopping rule for guiding therapy in clinical practice.

Can thalidomide reduce refractory bleeding from gastrointestinal angiodysplasia?

▶ Ge ZZ, Chen H-M, Gao Y-J, et al. Efficacy of thalidomide for refractory gastrointestinal bleeding from vascular malformation. Gastroenterology 2011;141:1629–37.

Chronic blood loss from gastrointestinal vascular malformations creates a healthcare burden due to recurrent hospital admissions and transfusion dependency. The widespread use of single or dual anti-platelet therapy and an ageing population may exacerbate this. Efforts at endoscopic treatment are often ineffectual. Ge and colleagues performed a prospective, open-label study assessing long-term efficacy and safety of thalidomide for refractory bleeding from intestinal vascular malformations. Fifty-five patients were randomly allocated to either 100 mg thalidomide or 400 mg iron daily for 4 months, with at least one year of follow-up thereafter. The primary end-point was defined as the proportion of patients in whom bleeding episodes—measured by immunoassay faecal occult blood tests—decreased by ≥50%. Secondary end-points included reduction in blood transfusion and hospital admissions along with stabilisation of haemoglobin levels. Serum levels of vascular endothelial growth factor (VEGF) were measured in the 28 individuals randomised to thalidomide. A significant reduction in bleeding episodes was reported in the thalidomide group (71.3% vs 3.7%; p<0.001). Thalidomide was also significantly more effective in all secondary outcome measures. Adverse events occurred in 20 (71.4%) of the thalidomide group and 9 (33.3%) of the iron group, with many in the thalidomide group reporting more than one adverse effect. On the whole these were mild and only two patients discontinued thalidomide treatment. Serum VEGF levels were significantly reduced with thalidomide treatment (p<0.001). This suggests thalidomide is a potential treatment in refractory bleeding from GI vascular malformations and its effects may extend beyond cessation of treatment, although in practice due to teratogenicity and toxicity, needs to be confined to a selected patient cohort.

Reviewers

Dr Paul Lochhead, Dr Richard Hansen, Dr Jonathan MacDonald, Dr Martin Prince.

Journals reviewed

Science, Nature, Genome Research, New England Journal of Medicine, Hepatology, Gastroenterology.