1. Introduction
Resveratrol is a stilbene-based, polyphenolic compound occurring naturally in nuts, berries, and the skin of grapes, and is present in many food items, particularly red wine [
1]. In recent years, the clinical effects of resveratrol have been studied intensively owing to a plethora of promising effects in cells and animals; among these are improvement in terms of the various consequences of obesity, e.g., impaired insulin sensitivity and low-grade inflammation [
2,
3,
4]. A landmark study demonstrated a shortened life span in diet-induced obese mice and a resveratrol-mediated improvement in longevity despite obesity [
5]. Owing to its purported salutary effects, resveratrol is already being sold as an over-the-counter dietary supplement. Unfortunately, translating the effects of resveratrol to a human setting has proven difficult. Some clinical studies have confirmed the pre-clinical results of resveratrol-mediated improved insulin sensitivity, diminished low-grade inflammation, reduced blood pressure, and reduced amounts of liver fat accumulation [
6,
7]. In contrast, other studies fail to detect any positive physiological effects [
8,
9]. These conflicting results may in part be attributable to differences in study designs, populations, and resveratrol formulations. Some studies have been performed in healthy individuals, and the ability of resveratrol to improve glucose handling in situations with normal glucose homoeostasis has been questioned [
8]. In contrast, it is feasible that the effect of a food ingredient is too weak to demonstrate any measurable consequences in full-blown type 2 diabetes. Thus, subjects suffering from modest degrees of insulin resistance are most likely optimal for the purpose of investigating the possible effects of resveratrol on insulin sensitivity, and probably most comparable to the original mice studies using diet-induced obese mice. Study length may also be of importance. Most studies include intervention periods of only a few weeks [
6,
9]. Obviously, longer studies provide a better determination of the possible effects of resveratrol on chronic conditions such as insulin resistance and low-grade inflammation. Finally, the methods currently used in most studies on resveratrol are targeted towards whole-body effects. Insulin sensitivity, for instance, has been assessed via measurements of specific targets such as plasma insulin, glucose, and HbA1c [
6,
8,
9]. These techniques have their strengths, such as being very specific and applicable in daily clinics; however, the drawback is that findings will be limited to the pre-specified objectives. Likewise, when confining assessment to plasma samples, intracellular effects may be missed.
Recently, the so-called metabolomics approach has been developed. Metabolomic analysis profiles small-molecule metabolites using either a targeted or non-targeted approach [
10]. Through systematic identification and quantification of metabolites in biological systems, e.g. tissues, organs, or fluids at a certain time point, metabolomics directly sample metabolic modulations. Targeted metabolomics allows for very precise measurement of a specific target or an isolated group of related metabolites, whereas non-targeted metabolomics assesses as wide a range of different metabolites as possible. This technique provides a simultaneous screening of multiple metabolic pathways and captures an instantaneous integrated snapshot of the entire physiology of an organism. Metabolomics has the advantage over other “omics” that it integrates changes in gene expression, protein levels, enzymatic activity, and post-translational changes. Previously, metabolomics made it possible to determine changes in the human plasma metabolome caused by various conditions and diseases like sleep restriction [
11,
12], diabetes [
13], and bariatric surgery [
14].
Studies have demonstrated a weak correlation between plasma and muscle metabolite levels, indicating that plasma measurements are poor indicators of skeletal muscle metabolism [
15]. Consequently, metabolomic analysis requires tissue biopsies to allow a detailed examination of changes in the intracellular pathways within this specific tissue. Such metabolic profiling offers the possibility to identify biochemical signatures of cellular metabolism involved in different pathways like glucose, lipid, and protein handling.
Massive research interest has gone into resveratrol as a potential protective compound in the management of obesity-associated complications, e.g., metabolic syndrome and low-grade inflammation. However, the particular pathways involved in potential salutary effects are by and large undetermined. Non-targeted metabolomic profiling may provide new insight into these matters. We have recently conducted a four-month randomized, placebo-controlled clinical trial describing the effects of resveratrol treatment in middle-aged males with metabolic syndrome and reported the findings on bone [
16], circulating steroids [
17], and glucose metabolism and inflammation [
18]. The aim of the present study is to provide a comprehensive metabolomic analysis of the changes caused by four months of high-dose resveratrol treatment in these middle-aged men with metabolic syndrome. As the effects of resveratrol in humans are incompletely characterized, we used a non-targeted metabolomics approach based on four different matrices: plasma, urine, adipose tissue, and skeletal muscle tissue from each study subject. In our study on circulating steroids and prostate size we established a dose-dependent decrease in plasma androgen precursors by resveratrol [
17]. In order to maximize the chances of detecting even subtle resveratrol-induced pathway changes, we therefore decided to employ the metabolomics analysis in the high-dose resveratrol (hRSV) (500 mg twice daily) group compared to the placebo group.
3. Discussion
The effects of resveratrol in human trials have proven to be very inconsistent. Most data represent standard biochemical analyses of selected parameters. The focus has been on inflammation [
8,
19], glucose metabolism [
9,
20], lipids [
21], liver function [
22], and steroids [
17], assessed by measurement in blood samples, thus representing a picture of whole body metabolism. The present paper describes an exhaustive metabolomic profile of blood, adipose tissue, skeletal muscle tissue, and urine from a comprehensive clinical trial of high-dose resveratrol-treated, middle-aged males with metabolic syndrome. This approach allows for a more elaborate investigation of subtle resveratrol effects on distinct intracellular pathways than what is possible by measurements confined to blood. In addition, this approach of unbiased measurements in various body compartments may help us to determine the cause of some of the effects we have previously detected. Recently, we reported that the androgen precursors in blood were significantly reduced after resveratrol treatment for four months [
17]. The presented metabolomic analysis was performed in the same study cohort, and by using this advanced technique we are able to verify the decrease in sulfated androgen precursors in the blood [
17]. In addition, we find significantly lower intracellular amounts of sulfated androgen precursors in skeletal muscle tissue as well as adipose tissue. Sulfation is a common modification to control levels of active steroid hormones at target sites by increasing solubility and renal excretion. The metabolomic profiling of urine revealed increased urinary excretion of the majority of the measured sulfated steroid hormones. In the original paper on reduced androgen precursors, we discussed whether the reduction was caused by decreased formation or increased excretion. The present results propose that at least part of the decrease in sulfated androgen precursors in blood and tissues is caused by an increase in urinary excretion of the sulfated steroids.
The metabolomic analysis of lipids points to a striking accumulation of long-chain saturated, monounsaturated, and polyunsaturated (n3 and n6) free fatty acids in adipose tissue. However, in skeletal muscle two long-chain lipids were significantly changed; one was elevated (10-heptadecenoate (17:1n7)) whereas the other was significantly reduced (arachidate (20:0)). Deducing how resveratrol induces the increase in many of the long-chain polyunsaturated fatty acids (n3 and n6) in adipose tissue is complex. Usually these long-chain polyunsaturated fatty acids derive from increased intake of seafood or special plant oils. However, the two essential fatty acids alpha-linolenic acid (ALA; 18:3n3) and linoleic acid (LA; 18:2n6) can be metabolized into longer-chain polyunsaturated fatty acids through a series of desaturation and elongation processes [
23]. The literature describes a gender difference in the conversion of ALA to eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), with women having a much higher conversion rate than men [
24,
25]. This difference depends upon sex steroids, and testosterone supplementation to female transsexuals reduced DHA concentration by 22% [
26]. Clearly, this indicates that the enzymes responsible for the desaturation and elongation conversion of ALA to DHA are negatively regulated by androgens. Based on this knowledge, we speculate that resveratrol might stimulate the conversion process of ALA and LA, resulting in an increase in long-chain polyunsaturated fatty acids through the robust reduction in androgen precursors.
The health impacts of changes in levels of long-chain polyunsaturated fatty acids are still not fully determined. Long-chain polyunsaturated fatty acids have been associated with a positive effect on insulin signaling via the insulin receptor [
27], and even a minor increase of ALA content in adipose tissue of 1% has been associated with a 5 mmHg decrease in blood pressure [
28]. However, intervention studies providing ALA have yielded conflicting results on blood pressure, cholesterol levels, and triglyceride levels, as reviewed by Baker et al. [
23]. Also, despite the increase in these fatty acids in adipose tissue after resveratrol treatment, no clinically relevant improvements in blood pressure or insulin sensitivity were observed in our study population [
18].
In the adipose tissue we also detected a significant increase in intracellular glycerol and free fatty acids in the resveratrol-treated subjects. In support of this notion, previously published cell culture studies revealed a resveratrol-mediated enhancement of adrenergic stimulated lipolysis in fat cells [
29,
30].
Moreover, resveratrol treatment caused consistent reductions in several dicarboxylic fatty acids in plasma. These lipids can be produced from peroxisomal-mediated degradation of very long chain fatty acids (>22 carbons) and subsequently be metabolized via beta-oxidation. Collectively, a decrease in the dicarboxylic fatty acids in plasma could be a result of increased beta-oxidation in adipose tissue.
Regarding glucose metabolism, we found trends toward elevation in glycolytic intermediates and the pentose phosphate pathway (PPP) intermediates (
p < 0.10) in adipose tissue in response to resveratrol. Stimulation of the PPP pathway is associated with increased insulin resistance, increased NADPH production, and thus, increased production of free fatty acids [
31]. Hence, our metabolomics findings may indicate increased glucose availability and utilization through glycolysis and the pentose phosphate pathway, owing to decreased insulin sensitivity in the males receiving resveratrol. Clinical parameters in the same study subjects support this assumption, as resveratrol increased fructosamine levels after four months of treatment, indicating poorer insulin sensitivity [
18].
In the resveratrol-treated group we demonstrated elevation in metabolites related to metabolism of the amino acid histidine. In particular, histidine was significantly increased in the blood and a trending elevation in the related biochemical marker of the muscle protein turnover 3-methylhistidine along with the acetylated derivative
N-acetyl-3-methylhistidine was also observed. Increase in 3-methylhistidine, which is a post-translationally modified amino acid derived from the muscle contractile proteins actin and myosin, and
N-acetyl-3-methylhistidine indicates increased muscle turnover in men treated with resveratrol. This finding ties in well with the study by Olesen et al. [
32], in which resveratrol blunted the positive effects of exercise training. In contrast, Scribbans et al. found that resveratrol in addition to exercise had positive effects on muscle metabolism with an increase in mitochondrial capacity [
33], while others found no effect in humans [
34]. Most rodent studies have found positive effects of resveratrol on muscle function [
5,
35,
36]. Clearly, more clinical studies are needed to determine the exact effects of resveratrol on skeletal muscle function.
Finally, we found urinary changes in several tyrosine-derived, tryptophan-derived, phenylalanine-derived, and histidine-derived metabolites. Tryptophan, histidine, and phenylalanine are classified as essential amino acids, which cannot be synthesized de novo by the body. However, gut bacteria may contribute to their production and degradation [
37]. Fermentation of tyrosine and tryptophan by colonic bacteria is found to produce phenols and indoles, which are excreted in the urine [
38]. Also, imidazole propionate is known to be excreted in the urine and direct degradation of urocanate to imidazole propionate by the intestinal flora was demonstrated in an animal study [
39]. Changes in microbiota-related urinary metabolites are proposed to be of relevance to human health. For instance hippurate, phenylacetylglutamine, 4-cresylsulfate, and 4-hydroxyphenylpropionate have been related to body weight, blood pressure, and metabolic syndrome [
40], while urinary imiadazole propionate, which we found to be elevated by resveratrol, has been related to intestinal dysfunction [
39]. Therefore, the changes in urinary amino acids and derivatives revealed by the metabolomic analysis may imply a resveratrol-induced modulation of the gut flora in men with metabolic syndrome. This is in accordance with studies in rodents where resveratrol has been shown to modulate the gut microbiota directly [
41,
42,
43].
The strengths of the present clinical trial are a carefully selected, strong study design and the advanced metabolomics approach. A high dosage of resveratrol in a relatively long-term study was chosen. Participants were selected because they had metabolic syndrome, a condition considered a prime target for resveratrol. Furthermore, the advanced metabolomics technique was applied in an unbiased approach to four different matrices. However, a limitation of our study is the lack of liver tissue samples. The liver is a highly metabolic organ and a number of experimental studies have provided evidence of the hepatic benefits of resveratrol treatment [
44]. Metabolomic profiling of liver tissue from the same study subjects would help complete the metabolic puzzle of resveratrol-induced biochemical effects. Also, metabolomic analysis on feces would have provided insight into possible beneficial effects on the gut microbiota. Finally, it must be kept in mind that the present study used a very high dose of resveratrol (1000 mg), which is impossible to reach from natural food items. Resveratrol has a very low bioavailability owing to a rapid metabolism to glucuronide and sulfate conjugates in the liver. However, it is possible that ingestion of food items rich in resveratrol and other polyphenols might inhibit the metabolism or increase the absorption of resveratrol, resulting in a better bioavailability from food items than from purified resveratrol products [
45,
46].