Phytochemicals induce an array of biological responses in animal cells, including enzyme inhibition, protection against oxidation and regulation of cellular signaling pathways. Among phytochemicals with xenobiotic activities, the stilbenes are of great interest for their potential medical impact. Stilbene compounds, particularly pterostilbene and resveratrol, have been associated with antidiabetic, anticarcinogenic and antilipogenic activities in vitro and in vivo 1234. Specific enzymatic targets are known for some of these xenobiotic activities. Pterostilbene and related compounds inhibit the cytochrome P450 enzymes, CYP1A1 and CYP1B1, which induce carcinogenicity of environmental teratogens 5. Pterostilbene also inhibits cyclooxygenases COX-1 and COX-2, interfering with endocrine functions and supporting analgesic activity 6. Significant increase in activity of hexokinase, while significant decrease in activity of glucose-6-phosphatase and fructose-1,6-bisphosphatase, were observed after oral administration of pterostilbene in diabetic rats 7. The related compound, resveratrol, stimulates sirtuin enzymes to promote longevity in response to dietary restriction 89. Resveratrol has also been demonstrated to protect against obesity- and diet-related disease in rodents 10. Additional interactions with new targets may be found to underlie other activities of these compounds.

While dietary stilbenes and flavonoids are of particular interest for their potential health benefits, in vivo activity is limited by low bioavailability due to rapid metabolism and excretion. Studies have found that methoxylation can protect flavonoids from derivatization, thereby improving biostability 1112. However, there is also evidence that methoxylation can alter bioactivity. For example, methoxylated flavones induced different cell-cycle arrest points in cultured human cells than hydroxylated flavones 13. Thus, methoxylation may alter interactions with target proteins, as well as increase bioavailability. These reports on altered bioactivity in methoxylated flavonoids triggered our interest to investigate the effects of methoxylation on stilbene bioactivity, particularly in a whole organism, to assess effects on postmitotic cells in adult tissues as well as on actively dividing cells.

Use of invertebrates provides one approach to address these questions rapidly and economically in a whole organism. In particular, the roundworm, Caenorhabditis elegans, offers many advantages for such analyses. C. elegans has a short generation time of 1 week and are easily maintained in large numbers with ordinary laboratory equipment. The adult body is approximately 1 mm in length, and contains most of the tissues present in higher vertebrates, such as a complete nervous system, striated muscles and an intestine with digestive, detoxification and innate immune functions. It is of note that the adult somatic tissues are post-mitotic and therefore lack cell-replacement capacity. C. elegans are hermaphrodites and able to self-fertilize, which has greatly facilitated the genetic analysis of developmental and cell biological processes in this organism. Previous uses of C. elegans for pharmacological studies have taken advantage of mutant strains for mechanistic studies, in addition to basic studies of compound-induced gene expression or phenotypic changes 141516.

In our study, we used C. elegans to compare the bioactivities of methoxylated and hydroxylated stilbenes, based on the parent compound resveratrol. Seven stilbenes differing in hydroxylation and methoxylation patterns were examined for effects on survival of fem-1(hc17) sterile adults and growth of germline tumors in gld-1(q485) mutants. Methoxylated stilbenes exhibited greater bioactivity in both assays, as compared with hydroxylated stilbenes. Steady-state levels for three of these compounds (one mono-, one di-, and one tri-methoxylated) were measured in treated worms and were not found to correlate with relative bioactivity. This suggests that the greater bioactivity of methoxylated stilbenes does not reflect differential uptake of the compounds. These findings demonstrate that methoxylation substitution enhances stilbene bioactivity in this whole-organism model. Methyl group-specific target site interactions may be one factor accounting for the differential bioactivities of these compounds. Alternatively, methoxylation may protect stilbenes from metabolic modification and excretion, leading to higher potency in vivo in C. elegans than hydroxylated stilbenes.

Seven methoxylated or hydroxylated stilbenes were compared for effects on C. elegans survival and tumor growth. This collection included two hydroxylated structures (resveratrol and piceatannol), two monomethoxylated structures (pinostilbene and desoxyrhapontigenin), two dimethoxylated structures (3-hydroxy-5,4'-dimethoxystilbene and pterostilbene) and one trimethoxylated structure (resveratrol-trimethylether) (Figure 1). Bioactivity in C. elegans has been reported only for resveratrol, which extended adult lifespan 9. It was of interest, therefore, to compare adult survival in animals treated with resveratrol and the related compounds.

For survival assays, young adult (day 0) animals were transferred onto fresh medium supplemented with the test compound at a range of concentrations. The compounds were tested at indicated concentrations in the growth medium (below). C. elegans nematodes are surrounded by a thick collagenous cuticle and are therefore relatively impermeable to environmental compounds. Consistent with this, our measurements demonstrated that internal concentrations of these compounds were significantly lower than their concentrations in the medium. Survival was scored every 2–3 days as the ability to move in response to a gentle mechanical stimulus. During the first week of adulthood, C. elegans hermaphrodites are reproductively competent and normally produce 200–300 progeny. Since progeny production would complicate survival analysis, these experiments were performed using fem-1(hc17) animals that are sterile due to a spermatogenesis defect 1718. Our prior analyses have confirmed that fem-1(hc17) and wildtype animals have similar lifespan characteristics 15.

Under our experimental conditions, both hydroxylated and one monomethoxylated stilbene, pinostilbene, had negligible or modestly beneficial effects on adult survival (Figure 2). Consistent with previous reports, we observed that resveratrol had a modestly beneficial effect on adult lifespan (0–15% at 100 μM dose) (Table 1). However, the effect was variable and did not attain statistical significance in our experiments. In addition, piceatannol and pinostilbene had beneficial effects on lifetime survival. The effects of these compounds were also modest (5–15%) and were not statistically significant.

The methoxylated stilbenes, desoxyrhapontigenin, 3-hydroxy-5,4'-dimethoxystilbene, pterostilbene and resveratrol-trimethylether, each exhibited significant detrimental effects on adult survival (Figure 3). At a 100 μM dose initiated on the first day of adulthood, survival was reduced 11–26% by desoxyrhapontigenin, 48–57% by 3-hydroxy-5,4'-dimethoxystilbene, 21–28% by pterostilbene and 18–51% by resveratrol-trimethylether (Table 1). The toxicity of resveratrol-trimethyether increased steeply within a relatively limited dose range, between 5–20 μM. In contrast, the toxicity of desoxyrhapontigenin, 3-hydroxy-5,4'-dimethoxystilbene and pterostilbene gradually increased between the 20–100 μM doses. The observed difference in activity between the monomethoxylated stilbenes, desoxyrhapontigenin being toxic while pinostilbene not, may be due to differences in steric hinderance introduced by the bulky methoxy groups at different positions in the stilbene ring. Without knowing the target site(s) in C. elegans for these stilbenes, it can be speculated that the position of the methoxy group at ring A in pinostilbene causes unfavourable fit for target site inhibition. Differences in activity of these monomethoxylated stilbenes have been demonstrated in other studies; e.g., pinostilbene moderately inhibited the catalytic activity of cytochrome P450 2E1, while desoxyrhapontigenin was not inhibitory 19. We noted that animals treated with either methoxylated or hydroxylated stilbenes did not display any other noticeable phenotypes, such as altered feeding or movement rates.

The somatic tissues of C. elegans adults are comprised solely of post-mitotic cells and lack the ability to renew by stem cell replacement. To more closely examine the effects of methoxylated stilbenes on mitotically-active cells in this organism, we examined their effects on growth of germline tumors produced in gld-1(q485) animals. The C. elegans hermaphrodite gonad is a U-shaped structure that is bisymmetric and contains the germ cell pool (Figure 4A, gld-1(+)). Within each gonad arm, germ cells in the distal region undergo mitotic proliferation in response to Notch-like signals produced by cells at the distal tips of the gonad arms. As germ cells migrate along the gonad arm, they exit from mitosis, undergo meiosis and subsequently differentiate into mature oocytes. In gld-1(q485) mutants, the germ cells fail to exit from mitosis and continue to proliferate throughout the gonad forming a germline tumor that is lethal to the animal (Figure 4A, gld-1(q485)) 2021.

Several stilbenes have demonstrated antiproliferative effects on mammalian cancer cells in vitro 22232425. To assess the effects of methoxylation on antitumor activity in vivo, tumor area was measured in gld-1(q485) adults treated for four days with either resveratrol trimethylether, desoxyrhapontigenin or pterostilbene, which are methoxylated and toxic to adults, or resveratrol, which is hydroxylated and nontoxic. As controls, germline tumor area was also measured in animals treated with carrier only (ethanol) or with 5-fluoro-2'-deoxyuradine (FUDR), a thymidylate synthetase inhibitor that blocks cell proliferation 2627. FUDR was chosen as a positive control for its known antiproliferative effects in C. elegans, which have been demonstrated by inhibition of embryonic and larval development 28. After four days of treatment, animals were collected and germline tumors were visualized by DAPI staining to visualize germcell nuclei. Because the level of DAPI staining was not quantitative between animals, we DAPI signal intensity was not used as a measure of tumor size. Rather, tumor size was directly measured as the area of the DAPI-stained gonad.

For the initial tests of tumor growth suppression, each compound was tested at a 100 μM dose in the growth medium (Figure 4B). At this concentration, FUDR exhibited a modest inhibitory effect on glp-1(q485) germline tumor growth in two of three experiments (-4%, -7%, -19%; p ≤ 0.05 for trials 2, 3, t-test vs untreated). At 100 μM dose, the methoxylated stilbenes resveratrol trimethylether and desoxyrhapontigenin were also associated with consistent reductions in tumor size. Pterostilbene also modestly reduced tumor size, and this effect was statistically significant in 2 of 3 trials (p ≤ 0.05, t-test vs untreated). In contrast, resveratrol had a negligible effect on gld-1(q485) germline tumor size which was not statistically significant (-4%, p = 0.35). As expected from its known antiproliferative effect, FUDR treatment also appeared to decrease tumor size, although the effect was statistically significant in only 2 of 3 trials.

To better quantify the effects of stilbenes and FUDR on gld-1(q485) tumor growth, two stilbenes, resveratrol and pterostilbene, and FUDR were examined at a series of additional doses between 50–400 μM. Over this range of concentrations, both FUDR and pterostilbene were associated with 10–20% reduction in tumor area (Figure 5A, B). The effects of pterostilbene were somewhat variable between trials, and were statistically significant for 3 of 4 experiments at both the 200 and 400 μM dose. The 200 and 400 μM doses of FUDR were consistently associated with smaller tumor area and these effects were significant in all 4 trials. Resveratrol at the 200 and 400 μM dose was not associated with statistically significant reductions in tumor area for any of the three trials conducted. These data demonstrate that the enhanced bioactivity of methoxylated stilbenes in vivo in C. elegans can be detected as a reduction in glp-1(q485) tumor area as well as by adult toxicity.

The differential bioactivities of methoxylated and hydroxylated stilbenes in C. elegans adults might reflect differences in uptake or bioavailability. In general, intact C. elegans adults take up compounds relatively poorly due to their thick, impermeable cuticle. This impermeability has impeded pharmacodynamic analysis in this organism. However, steady-state levels of compounds can be measured in treated animals to provide one measure of pharmacodynamics. To assess whether differential uptake correlated with bioactivity in C. elegans, steady-state levels for pterostilbene, resveratrol-trimethylether and desoxyrhapontigenin were determined in animals treated for 2 days with 100 μM of each compound in the culture medium. The levels of these stilbenes in C. elegans nematodes ranged from 20–80 μM, indicating relatively efficient uptake (Figure 6). These concentrations were within the range of serum stilbene levels detected after dietary stilbene supplementation with human volunteers (approx. 2 μM) and rabbits (42.8 μM) 29. It is also within the dose range where resveratrol was observed to induce dilation of isolated retinal arterioles (50 μM) 30. These three compounds, while all methoxylated, exhibited different levels of toxicity in adult C. elegans. In treated worms, levels of pterostilbene were slightly elevated compared with resveratrol-trimethylether and desoxyrhapontigenin (Figure 6). However, the most toxic compound in this group, resveratrol-trimethylether, was present at less than one-third the level of pterostilbene. These data suggest that relative bioactivity did not solely reflect steady-state levels in intact animals, although, this experiment does not directly compare the pharmacokinetic profile of each compound in C. elegans.

This study compared the effects of methoxylated and hydroxylated stilbenes on adult lifespan and tumor growth in C. elegans nematodes. In this organism, stilbene methoxylation was associated with toxicity under chronic treatment conditions and reduced tumor size in vivo. Overall, this study corroborated other findings of enhanced bioactivities in methoxylated versus hydroxylated stilbenes in other experimental systems 31. However, ours is the first study to systematically evaluate the effects of stilbene methoxylation on bioactivity in a whole organism, using C. elegans.

The mechanisms for enhanced bioactivity of methoxylated stilbenes may reflect increased biostability in vivo due to reduced efficiency of metabolic modification leading to excretion. In the case of another group of compounds, the flavonoids, substituting the hydroxyl groups with methoxy groups was shown to improve flavonoid stability in cultured cells and liver microsomes by blocking metabolic modification and excretion 1211. In mammals, hydroxylated flavonoids were shown to undergo glucuronidation by hepatic enzymes to promote excretion 11. Thus, methoxylation may block this modification and may also promote biostability. Indeed, one naturally methoxylated stilbene, pterostilbene, is cytotoxic in several cellular systems 32. In contrast, the hydroxylated stilbene, resveratrol, has also been shown to be cytotoxic, but is poorly retained in plasma compared with pterostilbene 33. Thus, methoxylation differences between pterostilbene and resveratrol may improve bioactivity by increasing in vivo stability.

The basis for the toxicity and antiproliferative activities of methyoxylated stilbenes in C. elegans adults is not known. The somatic cells of C. elegans adults are entirely postmitotic, ruling out antiproliferative effects as the basis for methoxylated stilbene toxicity. However, the toxicity and antitumor activities of methoxylated stilbenes in C. elegans adults may reflect interactions with common biological targets. Studies in other experimental systems show that stilbenes are potent antioxidants 3435. However, resveratrol and pterostilbene, which have strikingly different activity in vivo in adult C. elegans, demonstrated similar antioxidant activity in vitro, suggesting that antioxidant activity is not the underlying reason for the difference in stilbene bioactivities in C. elegans. Stilbenes are also known to inhibit specific cellular enzymes, such as cytochrome P450 and COX enzymes. One study has compared the relative inhibition of mammalian CYP1 cytochrome P450 enzymes by pterostilbene, pinostilbene and desoxyrhapontigenin 5. In this study, these three methoxylated stilbenes demonstrated increased CYP1 inhibition relative to resveratrol, in partial agreement with our results. However, pinostilbene demonstrated slightly stronger CYP1 inhibition than pterostilbene and desoxyrhapontigenin, a finding not consistent with our C. elegans toxicity data. Stilbenes also inhibit COX-1 and -2 monooxygenases. C. elegans nematodes do not have any obvious COX-1 or -2 homologs, although the C. elegans genome sequence does contain numerous monooxygenases that might have similar mechanisms of action to mammalian COX-1 and -2 enzymes. Further characterization of the in vivo targets for methyoxylated stilbenes in C. elegans may provide insight into other bioactivities of these compounds in human cells.

Stilbene methoxylation was associated with increased in vivo bioactivity as tested by toxicity and tumor growth in C. elegans. In general, increasing degree of methoxylation was correlated with increasing bioactivity, with the exception of pinostilbene. However, methoxylation and toxicity were not correlated with steady-state levels of these compounds in treated animals. These findings suggest that methoxylation enhances bioactivity possibly through increased interactions with in vivo targets. A working hypothesis to account for these results is that stilbene methoxylation protects the compound from conjugation and subsequent excretion, thereby increasing biostability and bioavailability in C. elegans.

MAW and CAW conceived of and designed the C. elegans experiments, which were performed entirely by MAW. AMR prepared the stilbene derivatives and performed the uptake determinations. CAW, MAW and AMR prepared the manuscript. All authors have read and approved the manuscript.