In the present study, we showed that activation of the nuclear receptor PPARα in peripheral blood mononuclear cells results in a considerable change in gene expression profiles, as 10.5% of the genes expressed exhibited altered gene expression levels after incubation with the specific PPARα agonist WY14,643. The main function of PPARα in PBMCs appeared to be the regulation of fatty acid β-oxidation and other lipid metabolism related functions, which is in line with results from mice studies in liver  and intestine , and human cell line studies [18, 19]. Moreover, the observed down-regulation of amino acid metabolism has been shown before in liver in studies comparing wild type mice to the PPARα knock out mouse model .
Besides the possible roles of PPARα in PBMCs, this study also demonstrates strong individual variability between the subjects in gene expression responses to activation with WY14,643. It appears that each donor has its own specific gene expression profile response to PPARα activation, which results in distinct differences in the expression of certain genes after WY14,643 incubation. Beck et al. also reported differences in responsiveness in gene expression between individuals, after incubation of endothelial cells with LPS. However, endothelial cell cultures were already divided beforehand into type I or type II responders based on their LPS mediated IL8 production . In another study, incubation of cultured macrophages with oxidized low-density lipoprotein resulted in a person-specific inflammatory gene expression response that could be correlated to changes in gene expression of scavenger receptors . However, we did not find a correlation between basal PPARα expression or changes in PPARα expression and the observed variation in gene expression changes. In addition, the differences observed are probably not caused by the nutritional status of the subjects at baseline, as we did not observe differences in expression changes of selected PPARα target genes between the postprandial and the fasted state of PBMCs incubated with WY14,643. However, it should be noted here that only four subjects were studied. A reason for the difference in response of the donors in the first study could be genetic variation, such as single nucleotide polymorphisms (SNPs) in the PPARα gene, its target genes or PPARα co-factors involved in activation of gene transcription. Furthermore, epi-genetic variation such as methylation status of the PPARα promoter or its target genes may have caused between-subject differences in gene expression levels. Additional studies are required to elucidate whether gene expression profiles can be clustered in different response profiles, simplifying the identification of factors responsible for these individual responses. With respect to personalized nutrition these individual responses are of great interest as it can be expected that nutrients such as fatty acids can induce similar variations in response as WY14,643, which in the end might lead to personalized dietary advice.
The PPRE analyzes of the genes changed showed that approximately 8% of the genes changed after incubation with the PPARα ligand WY14,643 contained a predicted or reported PPRE, using the list as described by Lemay et al . However, Lemay et al. report that they tolerate a low false-positive, and a high (60%) false-negative rate, suggesting that their list of PPREs is far from complete. Our additional transcription factor binding site analysis increase the number of genes that contain a PPRE to a total of 17% of the genes changed. A network search showed that, besides PPAR, five other transcription factors were involved in direct regulation of at least 10 out of the 1,373 changed genes. Interestingly, all these transcription factors are known to be affected by PPARα activation [22–26]. Transcription factor binding site analysis revealed that, out of the changed genes that did not contain a PPRE, 27% contained a binding site for at least one of the other five selected transcription factors These genes appear not to be regulated by PPARα directly, but indirectly, via these other transcription factors, a mechanism which has been suggested before [27, 28]. The role of PPARα in this respect seems to be extensively larger than expected based on the outcome of PPRE analyzes alone.
An interesting observation is the decrease in expression level of genes containing a PPRE. Activation of PPARα by a ligand may result in a negative regulation of genes by means of transrepression as has been reported in several studies and reviewed by Ricote and Glass (2007) . This transrepression, however, does not require the presence of PPREs in the promoter regions of the target genes. Apparently, negative regulation of these genes, regardless of its mechanism, is stronger than the transcriptional activation of PPARα. Previously, Degenhardt et al. (2006) also showed down regulation of an insulin-like growth factor-binding protein gene (IGFBP-6) that contained a predicted PPRE, in response to the presence of a PPARα ligand .
The overlap in gene expression profiles between fasting and incubation with WY14,643 shows that PPARα in PBMCs carries out a substantial part of its function during fasting, when concentrations of its natural ligands, free fatty acids, are elevated in the blood. The main role of PPARα in PBMCs during fasting is fatty acid β-oxidation, most likely to cope with the reduced availability of glucose for utilization in energy production and the increase of fatty acids.
Direct comparison between the two array analysis should be examined with care, since the two studies are distinctly different in set-up. The fasting intervention study was conducted in vivo, while the WY14,643 incubation experiments were performed ex vivo. Moreover, fasting involves many more changes in physiology, apart from the before-mentioned increase in plasma free fatty acids, including changes in plasma insulin, glucose and leptin concentrations. The PPARα ligand incubations were set-up to elucidate the specific effects of activation of one nuclear factor, controlling for all other parameters.
Summarizing, this study gives us valuable information on the extent of the effect of PPARα activation, during fasting and in general, on human PBMC gene expression. It also shows that persons respond differently to PPARα activation with respect to their gene expression changes, indicating a possible person-specific nutrient response. It seems justified to conclude that human PBMCs are a suitable model to study changes in PPARα activation. This opens up the possibilities for more specific PPARα signaling studies in healthy humans using these relatively easily obtainable blood cells.