Butyrate, the four-carbon short chain fatty acid, has special significance for clinicians and scientists interested in large bowel physiology. It is normally present in the colonic lumen at millimolar concentrations as a product of bacterial fermentation of luminal carbohydrates and is readily taken up by the colonic epithelium to be used as a major energy source via β-oxidation. Butyrate affects key functions of the colonic epithelium in vivo or at least in vitro in models of the colonic epithelium. These functions include promotion of sodium and water absorption, improvement of tight junction permeability, and acceleration of epithelial restitution. Thus, butyrate plays an important role in the maintenance of colonic mucosal health.

Butyrate has also been implicated in the pathogenesis of colonic diseases, especially colorectal cancer and ulcerative colitis. Butyrate's role in the pathogenesis of ulcerative colitis has been a fascinating saga. In 1981, Roediger first reported that epithelial cells isolated from the rectum of patients with ulcerative colitis exhibited impaired β-oxidation of butyrate.1 His “energy-deficiency” hypothesis created more attention when diversion colitis, which may histologically resemble ulcerative colitis, was shown to be largely caused by a deficient supply of short chain fatty acids in the lumen.2 Whether a defect in β-oxidation has specificity for ulcerative colitis and indeed whether it is more than an in vitro artefact have been questioned, but attempts to resolve these issues have not really succeeded. The report by Ahmadet al (see page 493), in which deficient β-oxidation of colonic epithelial cells identical to that in ulcerative colitis was shown also to occur in dextran sulphate induced colitis in mice, rekindles debate about this issue. The recent profusion of models of colitis in animals deficient in key immune molecules has tended to draw attention away from the potential primary role of the colonic epithelium in ulcerative colitis. In none of these models, however, do the alterations in crypt architecture at all resemble those that characterise ulcerative colitis. Only models induced by epithelial injury, especially dextran sulphate induced colitis, mimic the shortened and branched crypts of ulcerative colitis, in contrast to the hypertrophic, straight crypts observed in the immune models. We now need information about the metabolic characteristics of the colonic epithelial cells in immune based models of colitis to determine whether the metabolic abnormalities in epithelial cells are indeed secondary to inflammation itself or are specific to the aetiology.

A vexing question has been how this small molecule exerts such a wide array of effects. Butyrate may act indirectly. In cell lines, exposure to butyrate induces a transient intracellular acidification, which itself might trigger many cellular events. Whether this occurs in vivo is uncertain, but would require intermittent rather than continuous exposure to butyrate to occur, conditions more likely to be found in the distal colon. Butyrate may affect cells via the supply of energy from its β-oxidation. This has been shown in vivo in the atrophic colon starved of short chain fatty acids3 and in vitro in a cell line not able to meet its energy needs through other substrates.4 However, apart from stimulation of proliferation under energy deficient conditions, the evidence is scanty for a role of β-oxidation in butyrate's other cellular effects.

Most of butyrate's effects seem to result from a direct action of butyrate itself on intracellular targets. A key target may be histone deacetylase (HDAC). Cells exposed to butyrate exhibit hyperacetylation of core histones, owing to the reversible inhibition of HDAC by butyrate. The importance of butyrate's effect on HDAC has been highlighted by the demonstration that trichostatin A (TSA), which specifically inhibits HDAC, mimics many of the effects of butyrate on specific protein expression, such as interleukin 8 and urokinase receptor, cell proliferation and apoptosis, and epithelial functions, such as paracellular permeability and cell migration. However, Siavoshian and colleagues (see page 507) report that TSA doesnot mimic butyrate's effect on markers of cell differentiation, specifically the activities of brush border hydrolases, in HT29 cells. This observation has also been recently reported in other colonic epithelial cell lines.5 Does this mean that butyrate is acting on hydrolase activities via an intracellular target system that does not involve inhibition of HDAC?

Siavoshian et al examined whether the inhibition of cell proliferation induced by TSA in HT29 cells did truly mimic that of butyrate by comparing their effects on cell cycle events and on key intracellular molecules involved in those events. The effects of butyrate and TSA differed in the changes induced in cyclin dependent kinases and the stage in the cell cycle at which the cells were arrested (G₁ for butyrate, G₁ and G₂ for TSA). Furthermore, the duration that histone H4 was hyperacetylated differed, with TSA having a short action (<15 hours) and butyrate still exerting its effect after 24 hours.

Interpretation of these findings is aided by an understanding of the emerging complexity of HDAC as a transcriptional regulatory system. Histone acetylation precedes transcription and alters nucleosome and chromatin structure. This enhances accessibility of transcription factors to nucleosomal DNA. It is a dynamic process involving two enzyme systems, histone acetyltransferase and HDAC, which catalyse rapid acetylation and deacetylation. HDACs comprise at least two families of proteins that are targeted to specific promoters through sequence specific DNA binding factors.6 It seems likely, though not yet proved, that different HDAC proteins will target different promoters and, therefore, subserve different functions. In turn, different inhibitors may exert different spectra of inhibition of HDAC proteins. Siavoshian et al's findings may reflect such an effect, in addition to the different kinetics of their actions. Inhibition of HDAC leads to increased transcription of HDAC mRNA. The recent report that exposure to butyrate or TSA induces differing patterns of mRNA for HDAC proteins7 further supports the notion of heterogeneity in the patterns of inhibition of specific HDAC proteins.

Thus, whether HDAC is the major intracellular target for butyrate's actions remains unresolved, but it cannot be assumed that, if TSA does not mimic butyrate's action, then HDAC is not involved. Nevertheless, it seems more likely that butyrate has other intracellular targets, particularly as, unlike TSA, butyrate probably does not directly bind to HDAC and requires phosphatase activity to exert its inhibitory effect.8 Resolution of these issues is eagerly awaited as definition of the molecular pathways by which butyrate acts will greatly improve our understanding of multiple cellular processes in general and may be used specifically in the design of new therapeutic agents with—for example, antitumorigenic properties.