The Affymetrix GeneChip Mouse Genome 430A array comprises 22,690 probesets, representing 12,453 unique genes. Annotation information from Affymetrix was queried to compile a list of transporter and phase I/II metabolism genes present on the array (for details, see Methods). This list consisted of 665 probesets, encoding for 436 unique genes, and was used in the remainder of our analyses (Table 1). Under basal, chow-fed conditions, our filtering protocol identified 5,993 significantly expressed genes in the small intestine (i.e. having an absolute expression signal >20), of which 253 were transporters and phase I/II metabolism genes. After 24 hours of fasting, 713 genes, including 33 transporter and phase I/II metabolism genes, were differentially expressed (fold change >1.3, p-value < 0.01), corresponding to 13% of the expressed transporter and phase I/II metabolism genes. For selected genes additional qRT-PCR analyses were performed, which confirmed the array results (Table 2).

To get insight into the time-dependent effects of fasting, we performed an additional experiment in which mice were deprived of food for various time points up to 24 hours. Intestinal weight, blood glucose and plasma free fatty acid levels were measured (Figure 1). As expected, blood glucose levels decreased until 12 hours of fasting, after which it stabilized 2. Free fatty acid levels rapidly increased at the onset of fasting, and remained constant after 18 hours. Intestinal weight decreased already after 3 hours of fasting, significant changes were found as from 12 hours of fasting. For selected transporters and phase I/II metabolism enzymes gene expression was measured using qRT-PCR (Figure 2). Genes were selected based on function and array data. Several genes were regulated gradually in time and seemed to follow the same pattern as the free fatty acid concentration (Fatp2, Znt2, G6pt1, Zip4, Cyp27a1, Cyp2j6, and Sult1d1). Other genes were acutely regulated (Nadc1, Znt2, Zip4, Cyp27a1, Sult1d1, Gstm3, and Abca1.) In this group Nadc1 and Gstm3 were most drastic regulated after 3 hours fasting. Finally, a group of genes responded only after prolonged fasting (G6pt1, Fatp2, Mct4, Cyp4a10, Cyp3a11, Cyp2j6, and Abcg8). These data indicate that different mechanisms underlay the response to fasting.

It has been shown that PPARα is an important mediator of the hepatic adaptive response to fasting 2. Therefore we evaluated the role of this transcription factor in regulating transporter and phase I/II metabolism genes in small intestine during fasting. To this end, the effect of 24 hours fasting was compared in wild-type and PPARα knockout mice (Table 6). Eight of the genes identified in wild-type mice were PPARα-dependently regulated, which corresponded to 24% of all genes regulated. No genes were found to be suppressed in a PPARα dependent manner. qRT-PCR was used to confirm the differential expression of Cyp4a10, Abca1, and Slc25a36 (Figure 3).

Expression of the serotonin transporter Sert was increased after 24 hours of fasting. Serotonin, a neurotransmitter secreted by enterochromaffin cells, is considered to play a key role in normal functioning of the gut, initiating peristaltic reflex pathways and facilitating propulsive activity. Sert-mediated uptake of serotonin in enterocytes is responsible for the termination of the action of serotonin in the intestine 28. After a meal, the small intestine exhibits a pattern of phasic contractions of various amplitudes 44. During fasting, these phasic contractions are replaced by a cyclic pattern with less frequent contractions, enough to propel undigested food residues and sloughed enterocytes. As it is known that serotonin increases the frequency of intestinal contractions 45, removal of serotonin by Sert preserves these lower motility reflexes during fasting. Sult1d1 is involved in the sulfation of serotonin for the serotonin removal in enterocytes 46. For this conjugation the activated form of sulphate, PAPS, is needed. Sult1d1, the apical sulphate transporter Dtd, and the PAPS transporter Papst2 were all upregulated during fasting. We believe that the coordinated induction of these enzymes is required to maintain the lower peristaltic reflexes during fasting (Figure 4B).

Fasting increased the expression of Cact and Fatp2, and G6pt1, three intracellular SLC transporters involved in transport of substrates of mitochondrial β-oxidation and glycogenolysis, respectively 30313347. Changes in gene expression of these transporters coincided with differential gene expression of enzymes involved in both processes (data not shown). In addition, the increased expression of Cyp4a10 and Cyp4b1 points to enhanced peroxisomal oxidation of fatty acids 11, and an elevated level of Nadc1 implies increased uptake from luminal dicarboxylates (Krebs cycle intermediates) as well as citrate secreted by pancreatic and gastric juice 48. Thus, these changes reflect the molecular events caused by the switch of fuel utilization from carbohydrates to fatty acid oxidation (Figure 4A).

Cholesterol can be secreted from enterocytes by chylomicrons and by the efflux transporters Abca1 and Abcg5/g8. Expression of the cholesterol efflux transporters Abca1 and Abcg8 was increased upon fasting. Since both carriers are LXR target genes 4950 we believe this is due to the profoundly increased expression levels of Cyp27a1, which results in enhanced levels of the potent LXR agonist 27-hydroxycholesterol 51. Since during fasting no cholesterol is required for chylomicron formation, we speculate this may be a compensatory mechanism for the enterocyte to balance its intracellular cholesterol concentrations (Figure 4C).

Various CypP450s with well-known drug-metabolizing capacity were differentially expressed after fasting (Cyp2j6, Cyp3a11, and Cyp4b1). Increased expression of Cyp3a11 after fasting has also been observed in rat liver 52. Although Cyp3a11 does not have a direct human ortholog, it has similar substrate specificity as human CYP3A4. CYP3A4 is considered to be the major metabolizing enzyme of approximately half of the drugs in use today 53. Our results indicate that upon a fasting period drugs may be more efficiently detoxified (Figure 4D), which is of relevance e.g. during surgical interventions in patients.

Except for Gsts, which were expressed at reduced levels, fasting had no denoting effect on expression of phase II metabolism genes. GSH plays an important role in the defense against oxidative stress 54, and is required as cofactor for glutathione peroxidase (GPx) and glutathione S-transferase (Gst) activity. GPx detoxifies peroxides using GSH as an electron donor, producing GSSG as end product which in turn is converted back into GSH by glutathione reductase (Gsr). Gsts catalyze the conjugation of GSH to a wide variety of endogenous and exogenous electrophilic compounds. It is known that upon fasting levels of GSH are reduced in small intestine and liver 555657, likely as a result of increased oxidation and decreased presence of its dietary precursory amino acids 58. We believe these reduced GSH levels may be responsible for the observed reduced mRNA levels of the glutathione S-transferase (Gsts), since it is known that expression levels of Gsts are directly dependent on the presence of GSH 59. In addition, it is known that dietary electrophiles are able to activate the transcription factor Nrf2, which can activate among others Gsts 60. During fasting these dietary electrophiles are not available, which may be an additional explanation for the reduced expression levels of Gsts. The time dependent fasting experiment showed that the reduction of Gstm3 is a very acute process. Thus, regardless the underlying mechanism, the reduced GSH and Gsts levels will render the fasting gut more sensitive towards electrophilic stressors and other Gst substrates, which may have implications for humans on drug therapy during fasting and the period directly thereafter (Figure 4E).

We showed that eight of the transporters and phase I/II metabolism genes were PPARα dependently regulated during fasting (Figure 4F). Genes that were most prominently regulated by PPARα were involved in lipid metabolism (Cyp4a10, Abca1, and Smct1). In liver Cyp4a10 and Abca1 are known to be regulated via PPARα during fasting 6162. Furthermore, Cyp4a10 is a known PPARα target gene 63. It has also been reported that Abca1 is regulated by PPARα 64, although no PPAR responsive elements have been identified in its promoter region 65. Sert and Dtd were PPARα dependently regulated, which indicates that PPARα may link nutritional status to peristaltic movement. Sert, Dtd, Slc25a36, Smct1, Chst4, and Mgst1 have all not been identified yet as PPARα target genes. We conclude that PPARα is required for the adaptive response of a subset of genes.

Pure bred wild-type (129S1/SvImJ) and PPARα-null (129S4/SvJae) mice 66 were purchased from Jackson Laboratories (Bar Harbor, ME) and bred at the animal facility of Wageningen University. Mice were housed in a light- and temperature-controlled facility and had free access to water and standard laboratory chow (RMH-B, Hope Farms, Woerden, the Netherlands). All animal studies were approved by the Local Committee for Care and Use of Laboratory Animals.

Three to four month old male PPARα-null and wild-type mice were fasted for several time points up to 24 hours. Fasting experiments started at the onset of the light cycle. Mice were anaesthetized with a mixture of isofluorane (1.5%), nitrous oxide (70%) and oxygen (30%). Blood was collected via orbital puncture, plasma was obtained by centrifuging at 200 g for 10 minutes and stored at -80°C until use. The small intestines were excised, flushed with ice-cold PBS, and weighted. Remaining fat and pancreatic tissue was carefully removed and the small intestine was snap-frozen in liquid nitrogen and stored at -80°C until RNA isolation.

Total RNA was isolated from small intestinal samples using TRIzol reagent (Invitrogen, Breda, the Netherlands) according to the manufacturer's instructions. RNA was treated with DNAse and purified using the SV total RNA isolation system (Promega, Leiden, the Netherlands). Concentrations and purity of RNA samples were determined on a NanoDrop ND-1000 spectrophotometer (Isogen, Maarssen, the Netherlands). RNA integrity was checked on an Agilent 2100 bioanalyzer (Agilent Technologies, Amsterdam, the Netherlands) with 6000 Nano Chips according to the manufacturer's instructions. RNA was judged as suitable for array hybridization only if samples exhibited intact bands corresponding to the 18S and 28S ribosomal RNA subunits, and displayed no chromosomal peaks or RNA degradation products.

For microarray analyses, we used RNA isolated from the full-length small intestine. RNA was hybridized on an Affymetrix GeneChip Mouse Genome 430A array. This array detects 22,690 transcripts that represent 12,453 known genes. For each experimental group, three biological replicated were hybridized, thus in total 12 arrays were used. Detailed methods for the labeling and subsequent hybridizations are available on request. Arrays were scanned on a GeneChip Scanner 3000 (Affymetrix). Array data have been submitted to the Gene Expression Omnibus, accession number GSE6864.

Scans of the Affymetrix arrays were processed using packages from the Bioconductor project 67. Expression levels of probe sets were estimated using the library GCRMA 68, where after differentially expressed probe sets were identified using linear models 69. The library LIMMA implements an empirical Bayes method to assign differential gene expression, an approach repeatedly shown to be the most appropriate 70717273. To compile a list of transporter and phase I/II metabolism genes present on the array, annotation information from Affymetrix (release of July 2006) was queried for SLC transporters, ABC transporters, CypP450s, the phase II metabolism enzymes glutathione S-transferases, sulfotransferases, epoxide hydrolases, aldo-keto reductases, N-acetyltransferases, and glucuronosyl transferases. Also glutathione reductase, glutathione synthetase, and glutathione peroxidases were included in this set. The final set consisted of 665 probesets, encoding for 436 unique genes (Table 1). To study significantly expressed genes, only probesets with an expression value higher than 20 in the WT control group were selected for further analysis. This filtering was done after normalization. Probe sets that satisfied the criterion of >1.3 fold change with a p-value < 0.01 were considered to be significantly regulated. Of these, probe sets that were not changed in fasted PPARα-knockout mice, were designated PPARα regulated. Interpretations of functional outcomes of fasting focused on groups of genes that are known to be functionally related (i.e. participating in the same pathway or having a similar function). Although at first sight the fold change threshold may seem to be low, we could confirm all changes in gene expression identified on the microarray (Table 2). Moreover, we would like to stress that it is generally accepted that effects of nutritional interventions on gene expression are subtle, in contrast to pharmacological-type of interventions 747576. A clear example of this is found in Patsouris et al. 10, in which the effect of pharmacological, physiological, and nutritional intervention on expression of the same set of genes were compared.

Single-stranded complementary DNA (cDNA) was synthesized from 1 μg of total RNA using the Reverse transcription system (Promega, Leiden, The Netherlands) following the supplier's protocol. cDNA was PCR amplified with Platinum Taq DNA polymerase (all reagents were from Invitrogen). Primer sequences used in the PCR reaction were chosen based on the sequences available in GenBank. The sequence of primers used is available in Table 7. PCR was carried out using SYBR green on a MyIQ thermal cycler (Bio-Rad laboratories BV, Veenendaal, The Netherlands) with the following thermal cycling conditions: 8 min at 94°C, followed by 45 cycles of 94°C for 15 s and 60°C for 1 min. All samples were performed in duplicate and normalized to cyclophilin expression.

Free fatty acids were measured with the Free fatty acids half-micro test (Roche Diagnostics, Almere, The Netherlands) according to the manufacturer's instructions. Blood glucose levels were determined by the Accu-Chek Compact Glucose (Roche Diagnostics, Almere, The Netherlands) with 1 drop of blood obtained by orbital puncture.