Multiple effects of butyrate on local (intestinal) and systemic physiological function are presumably due to its actions as a potent regulator of gene expression.4–6
Butyrate metabolism, both biosynthesis by gut microorganisms and utilization in the gut, is tightly regulated and thus, imbalance could play a critical role in the pathology of humans and animals. Typically, increased production of SCFA, especially butyrate, results in lower luminal pH, which, in turn, creates favorable niche for and stimulates proliferation of butyrate-producing bacteria.12
In modern intensive production systems used for ruminants, especially during physiologically critical stages, such as weaning and the transition or periparturient period in dairy cattle, high concentrate rations are fed which can result in abrupt increases in SCFA production. This circumstance often results in production in excess of the utilization capacity, subsequently leading to the development of ruminal acidosis, a prominent digestive disorder with a significant economical impact.27,28
Individual variation in the capacity for uptake of acetate and butyrate among animals fed the same rations can explain the degree of acidosis observed.29
In fact, considerable effort has focused on development of practical nutritional intervention strategies for use during transition periods to enhance rumen epithelial surface area prior to the introduction of high VFA producing rations to facilitate better absorptive capacity and mitigate negative effects of acidosis.30
To date, these strategies have yielded mixed results with respect to increasing papillae surface area and total absorptive capacity (Andersen, personal communication). The development of a better understanding of the regulation of both rumen epithelial growth promotion and VFA transport capacity remains a concern to the industry.29
In this study, we assessed the rumen epithelial transcriptome dynamics when butyrate concentration was elevated due to direct ruminal infusion with buffered butyric acid in order to assess transcriptional effects of butyrate and its potential role in regulating butyrate transport in cattle.
The majority of our knowledge on regulatory impact of butyrate on global gene expression is derived from in vitro studies and observations.6,31,32
However, caution in interpretation of in vitro data and application of knowledge gained in vitro to in vivo models is warranted. SCFAs, such as butyrate, are known to promote rumen development and stimulate the proliferation of rumen epithelial cells in vivo.33
However, butyrate inhibits the proliferation of epithelial cells of the large intestine, rumen and kidney by down- regulating genes controlling cell proliferation in vitro.2,32
Butyrate also induces apoptosis and differentiation of tumor cells.34,35
Moreover, an in vitro study of epithelial cells of different origin (rat small intestine vs. human colon) has demonstrated that the cell type affects butyrate uptake characteristics.36
In fact, opposite effects of butyrate are observed for many cellular processes, such as cell proliferation and division, between in vitro and in vivo models and are clearly reflected in transcriptome characteristics.33,37
A number of genes related to cell proliferation and cell cycle progress were significantly down-regulated by a 24-h 10 mM butyrate incubation of established bovine rumen epithelial cells in long-term culturing (Wu et al 2012, personal communication). In contrast, expression of these same genes in the rumen epithelium in the present data set was not altered despite a 2-fold increase from 19.5 mM to 38.5 mM in intraruminal butyrate concentration (). Interestingly, the abundance of a butyrate transporter, SLC5A8, in the rumen epithelial transcriptome was significantly reduced concomitant with intra-ruminal butyrate concentration increases. However, SLC5A8 expression at the mRNA level was significantly increased ~21 fold by butyrate exposure in vitro (). This apparent opposite effect of butyrate on the expression of its transporter is suggestive of alternative regulatory mechanisms relating to butyrate uptake control and transport by the intact rumen epithelium and cells in culture. Moreover, cellular butyrate metabolism may be different between in vivo
and in vitro models due to changes in the rate of removal of end product as well as changes between cell-cell interactions and micro-environments present in vivo, but disrupted in vitro. To this point, epithelial metabolism of butyrate, especially the pathways leading to ketogenesis, helps to maintain a butyrate concentration gradient in vivo, which in turn facilitates butyrate uptake and affects butyrate intracellular concentrations.30
Rate-limiting enzymes in the ruminal ketogenic process, such as acetyl-CoA acetyl transferase (ACAT) and 3-hydroxy-3-methylglutaryl CoA synthases (HMGCS) 1 (cytoplasmic) and 2 (mitochondrial), play an important role in regulating butyrate metabolism at the substrate level. As depicted in , key enzymes in butyrate metabolic pathways exhibited a different expression pattern between the cell line and the rumen epithelium. As expected, HMGCS2 of mitochondrial origin was significantly up-regulated by butyrate in vitro, in response to increased butyrate concentration. However, expression of HMGCS2 remained unchanged in vivo.
Relative expression of a butyrate transporter, solute carrier family 5 (iodide transporter), member 8 (SLC5A8).
Expression levels of key enzymes involved in butyrate metabolism in the rumen epithelium detected by RNA-seq technology.
The biological interpretation of high-throughput expression data generated using microarrays or RNA-seq technology requires both differential expression and differential network analyses.38
Many transcriptional regulators exert their impact on biological functions via post-transcriptional mechanisms with subtle or no apparent changes at mRNA expression level detectable by tools for assessing differential expression alone. However, differential network analysis relies on powerful computational tools to extract accurate regulatory gene networks reflecting causal interactions underlying biological processes or phenotypes. Of several algorithms available, those based on information theory, including estimating mutual information values, such as ARACNE26,39
perform well in inferring global gene networks, especially for smaller sample sizes.40,41
ARACNE is based on the assumptions that the expression level of a given gene is a random variable and the mutual relationships between them can be derived by statistical dependences.42
Our results provided further support for the utility of this approach in constructing regulatory gene networks that depict phenotypes and regulation of biological processes. An example from this current study is the regulatory network controlled by FBJ murine osteosarcoma viral oncogene homolog, or c-fos (FOS), which was significantly down-regulated by butyrate in both in vitro and in vivo models. As a transcription factor, FOS dimerizes with another oncogene JUN to form the AP-1 complex, which regulates transcription of a diverse range of genes and is implicated in many biological processes including cell proliferation and differentiation as well as tumor transformation and progression. In the current data set, ARACNE inferred a network of four direct interactions (1st neighbors) for FOS and 32 indirect interactions (). All four direct interactions, cingulin (CGN), heparin-binding epidermal growth factor-like growth factor (HBEGF), intermediate filament family orphan 2 (IFFO2), and jun proto-oncogene (JUN), were up-regulated by butyrate (at both P
and FDR < 10−5
). Moreover, of 24 genes in the 2nd neighbors category, all were also regulated by butyrate, including three transcription factors, JUN, upstream transcription factor 2, c-fos interacting (USF2), and REST corepressor 1 (RCOR1). GO analysis identified GO terms significantly enriched in this network, including SMAD binding (GO:0046332 and GO:0070412), SMAD protein signal transduction (GO:0060395), and transforming growth factor (TGF)-β receptor signaling (GO:0007179). FOS binding at the TGF-β1 promoter proximal AP-1 site is required for TGF-β1 production by colon carcinoma cells.43
Indeed, previous studies have shown that SMAD proteins cooperate with FOS/JUN complex to mediate TGF-β-induced transcription.44
Thus, ARACNE correctly inferred a direct interaction between FOS and JUN () as well as interaction between JOS and HBEGF. HBEGF plays a pivotal role in mediating the early cellular response to intestinal injury by serving as a potent cytoprotective factor.45
Other experimental evidence also provides a strong support of a direct interaction between FOS and HBEGF.45,46
A regulatory gene network controlled by FOS.
Tight junctions between epithelial cells regulate the permeability of molecules via para-cellular pathways as well as bacterial translocation across the gut epithelial layer, thus having the potential to strengthen intestinal barrier function and being involved in intestinal pathology.47
In addition to cingulin that was induced by butyrate, at least 16 other tight junction related genes were regulated by butyrate in vitro. A strong up-regulation of the major macromolecular components, such as claudins (CLDN1, CLDN3, CLDN4, CLDN7, CLDN12, and CLDN23), tight junction protein 3 (TJP3), and junctional adhesion molecules 2 and 3 (JAM2 and JAM3), by butyrate supports the contention that these genes may play an important role in maintaining and/or restoring intestinal barrier function. Indeed, evidence indicates that probiotics, such as Lactobacillus plantarum
MB452, enhance intestinal barrier function via increasing expression of genes encoding proteins involved in tight function formation.48
It is known that elevated ruminal butyrate results in a profound change in the ruminal microbial composition, including a potential stimulating effect on butyrate-producing bacteria.12
It is foreseeable that an elevated concentration of butyrate in the lumen of the gut could play a regulatory role in the maintenance of intestinal barrier function via the expression of genes involved in tight junctions. Our future work will include experimental verification of the global regulatory gene networks inferred by computational tools.