Since Burkitt's original observations on the inverse correlation between fibre (non-starch polysaccharides and resistant starch) intake and prevalence of colorectal cancer , a wide range of studies have addressed this relationship and the possible mechanisms by which fibre may protect against bowel cancer. Recent meta-analyses find a strong evidence base to support consumption of fibre-containing foods for prevention of several cancers , and the majority of studies in this area are supportive. There are exceptions, however, and two RCT studies, published in 2000, failed to demonstrate a protective effect [3, 4]. These controversial findings have been the subject of several commentaries [5, 6]. Potential explanations for this conflicting data include: differences between US and EU assays for fibre, different baseline levels of intake and the limitations of adenoma recurrence as a model for primary colorectal cancer prevention.
There are several mechanisms proposed for fibre's proposed cancer-preventive properties. These include dilution of luminal contents; reduction in transit time, which together will reduce exposure of the mucosa to luminal toxin; adsorbtion of bile acids; and production of protective short chain fatty acids (SCFAs: principally acetate, propionate and butyrate) through fermentation of fibre by endosymbiotic bacteria. Studies in rats treated with a colorectal carcinogen, have demonstrated a variable protective effect of different dietary fibre substrates and have linked this with changes in the luminal SCFA profile . Gibson et al for example found that when rats consumed a diet with cellulose, a non-fermentable fibre, as principle fibre source, little protection from DMH-induced carcinogenesis was afforded. Oat-derived fibre, an acutely fermentable fibre which is rapidly turned over to SCFA in the caecum, but yields lower levels of SCFA in the distal colon and rectum, provided improved protection, but maximal protection was conferred by the more weakly fermentable wheat fibre, which yielded higher levels of SCFA in the distal colon and rectum. The study analysed SCFA levels in rats' stools on each regimen and found that the strongest correlation with cancer prevention in this model occurred on diets which gave maximal elevation of faecal butyrate. Not surprisingly this data has led to a resurgence of interest in the actions of butyrate.
Roediger  was first to show that butyrate is the preferred metabolite of colon epithelial cells. In his studies, primary epithelial cells from rat colon were incubated with labelled glucose and labelled butyrate. Butyrate was found to be metabolised in preference to glucose, which is available to colonocyte in vivo through the vasculature. The use of butyrate as an energy souce is inefficient (by comparison with glucose) and it has been suggested that this represents an evolutionary adaptation to recover the maximum energy available from the high-fibre diets consumed by our paleolithic ancestors.
The effect of butyrate on cells grown in vitro is to drive both cell cycle arrest and apoptosis. Both of these alterations in cell fate occur at concentrations of butyrate readily achieved in the colon lumen through fibre fermentation. Cell cycle arrest has variously been reported as G1 arrest, G2 arrest and mitotic bypass [9–11]. Several reports have shown that the apoptosis triggered by butyrate in vitro is associated with dysregulation of Bcl2 family proteins especially upregulation of BAK and downregulation of BclxL [12–14], rather than cellular damage.
These in vitro data contrast with studies on the in vivo or ex vivo effects of butyrate. Takayama's studies investigating the effect of increasing fibre intake after restriction, using a variety of animal models, have shown that switching to a high fibre diet is associated with an increase in colon crypt length, cellularity and proliferation [15–17]. Hass  used ex vivo guinea pig colon mucosa in an Ussing chamber model and monitored rates of cell death. When tissue was maintained in an osmotically balanced chamber, widespread cell death was found on the epithelium and this was associated with up-regulation of Bax. When butyrate was added to the chamber, however, there was reduced cell death and no Bax upregulation. Furthermore, studies of diversion colitis show that widespread cell death occurs after diversion of the faecal stream and loss of luminal content . This condition may be ameliorated by butyrate enema .
Recent studies have shown that elevation of luminal SCFA causes no direct increase in levels of epithelial apoptosis [21, 22], but causes a significant increase in the level of apoptosis after a genotoxic challenge. These in vivo data are suggestive of a model whereby butyrate's antineoplastic action is not in the induction of apoptosis per se, but through sensitization of cells to damage. The observations made in vitro that butyrate elevated levels of pro-apoptotic Bcl2 family proteins, and downregulated their anti-apoptotic counterparts could be predicted to sensitize cells in precisely this way and we have recently proposed this as a model .
How might butyrate alter cell functionality in this way? Although recognised as a metabolite, butyrate is also a potent inhibitor of histone deacetylases (HDACs) - . HDACs are primarily recognised as one part of the regulatory mechanism for governing histone acetylation levels in concert with their agonist enzymes the histone acetyl transferases (HATs). The acetylation state of histones is thought to be a potent governor of gene transcription at both a specific and regional level of the chromatin. A number of publications have shown widespread alteration in gene expression after treatment of cells in vitro with butyrate, indicating as much as 10% of genes may be affected by butyrate either directly or indirectly. However more recently several groups have identified other acetyl proteins in the nucleus and cytosol, and HDAC activities have been found in both cellular compartments [24, 25]. Amongst the acetyl proteins identified are nuclear structural proteins, transcription factors including p53, Sp1, Sp3  and structural proteins including tubulin and cytokeratins [27, 28]. Our own preliminary findings using pan-specific antiacetyl lysine antibodies indicate tens or hundreds of acetyl proteins in cell lines (Leech & Corfe, unpublished). Acetylation has been proposed as being as important as phosphorylation in the regulation of protein function . This is reinforced by the observation that HATs and HDACs are frequently mutated in cancer, which may lead to an alteration in the acetylation landscape of the cell permissive for cancer progression.
Taken together these data allow us to generate an hypothesis that i) colorectal carcinogenesis will be associated with an alteration in global protein acetylation, ii) reduced levels of butyrate will cause alterations in global protein acetylation, which may also be permissive for colorectal cancer progression, iii) that elevation of fibre levels and consequent butyrate levels may reduce or reverse these processes and restore a "normal" profile of protein acetylation.