E.1.1. Bacterial Mutagenicity Test Protocol

Testing procedures were those reported by Zeiger et al.¹⁶³ Briefly, a commercially obtained sample of trans-resveratrol (RES) (lot number 02-18090-601 from ChromaDex) was sent to the laboratory under code. It was incubated with each of the Salmonella typhimurium tester strains (TA98, TA100, and TA102) either in buffer or S9 mix (metabolic activation enzymes and cofactors from Aroclor 1254-induced male Sprague Dawley rat liver) for 20 minutes at 37°C. Top agar supplemented with L-histidine and d-biotin was added, and the contents of the tubes were mixed and poured onto the surfaces of minimal glucose agar plates. Histidine-independent mutant colonies arising on these plates were counted after incubation for 2 days at 37°C.

Each trial consisted of triplicate plates of concurrent positive and negative controls and at least five doses of RES. RES was tested up to 3,333 μg per plate for each tester strain in the presence or absence of S9 mix.

In this assay, a positive response is defined as a reproducible, dose-related increase in histidine-independent (revertant) colonies in any one strain/activation combination. An equivocal response is defined as an increase in revertants that is not related to dose, is not reproducible, or is not of sufficient magnitude to support a determination of mutagenicity. A negative response is obtained when no increase in revertant colonies is observed after chemical treatment. No minimum percentage or fold increase is required for a chemical to be judged positive or weakly positive, although positive calls are typically reserved for increases in mutant colonies that are at least twofold over background.

E.1.2. Results

RES (33 to 3,333 µg/plate) was not mutagenic in S. typhimurium strains TA98, TA100, or TA102 when tested with or without exogenous metabolic activation provided by Aroclor 1254-induced rat liver S9 and cofactors (Table E-1).

E.2.1. Peripheral Blood Micronucleus Test Protocol

At termination of the 3-month toxicity studies of RES, blood samples were collected from male and female Wistar Han rats and B6C3F1/N mice, placed in ethylenediaminetetraacetic acid (EDTA)-coated tubes, fixed in ultracold methanol, and frozen at −80°C until analysis. Thawed blood samples were analyzed for frequency of micronucleated immature erythrocytes (polychromatic erythrocytes [PCEs], reticulocytes) and mature erythrocytes (normochromatic erythrocytes [NCEs]) using a flow cytometer¹⁶⁴; both the mature and immature reticulocyte population can be analyzed separately by employing special cell surface markers to differentiate the two cell types. Because the very young reticulocyte subpopulation (CD71+ cells) can be targeted using this technique, rat blood samples can be analyzed for damage in the bone marrow that occurred within the past 24 to 48 hours, before the rat spleen appreciably alters the percentage of PCEs in circulation.¹⁶⁵ In mice, both the mature and immature erythrocyte populations can be evaluated for micronucleus frequency because the mouse spleen does not sequester and eliminate damaged erythrocytes. Damaged erythrocytes achieve steady state in the peripheral blood of mice after four weeks of continuous exposure. Approximately 20,000 PCEs and 1 × 10⁶ NCEs were analyzed per animal for frequency of micronucleated cells, and the percentage of immature erythrocytes (% PCE) was calculated as a measure of bone marrow toxicity resulting from chemical exposure.

Prior experience with the large number of cells scored using flow cytometric scoring techniques¹⁶⁶ suggests it is reasonable to assume the proportion of micronucleated reticulocytes is approximately normally distributed. The statistical tests selected for trend and for pairwise comparisons with the vehicle control group depend on whether the variances among the groups are equal. The Levene test at α = 0.05 is used to test for equal variances. In the case of equal variances, linear regression is used to test for a linear trend with dose and the Williams test is used to test for pairwise differences between each dosed group and the vehicle control group. In the case of unequal variances, the Jonckheere test is used to test for linear trend and the Dunn test is used for pairwise comparisons of each dosed group with the vehicle control group. To correct for multiple pairwise comparisons, the p value for each comparison with the vehicle control group is multiplied by the number of comparisons made. If this product is >1.00, it is replaced with 1.00. Trend tests and pairwise comparisons with the vehicle control groups are considered statistically significant at p ≤ 0.025.

In the micronucleus test, it is preferable to base a positive result on the presence of both a positive trend as well as at least one significantly elevated dosed group compared with the corresponding vehicle control group. The presence of either a positive trend or a single significant dosed group generally results in an equivocal call. The absence of both a trend and a significant dosed group results in a negative call. Ultimately, the scientific staff determines the final call after considering the results of statistical analyses, reproducibility of any effects observed (in acute studies), and the magnitudes of those effects.

E.2.2. Evaluation Protocol

These are the basic guidelines for arriving at an overall result for assays performed by the National Toxicology Program. Statistical as well as biological factors are considered. For an individual assay, the statistical procedures for data analysis are described in the preceding protocols. There have been instances, however, in which multiple samples of a chemical were tested in the same assay, and different results were obtained among the samples and/or among laboratories. Results from more than one aliquot or from more than one laboratory are not simply combined into an overall result. Rather, all the data are critically evaluated, particularly those concerning pertinent protocol variations, in determining the weight of evidence for an overall conclusion of chemical activity in an assay. In addition to multiple aliquots, the in vitro assays have another variable that must be considered in arriving at an overall test result. In vitro assays are conducted with and without exogenous metabolic activation. Results obtained in the absence of activation are not combined with results obtained in the presence of activation; each testing condition is evaluated separately. The summary table in the abstract of this toxicity report presents a scientific judgment of the overall evidence for activity of the chemical in an assay.

E.2.3. Results

In rats, the reticulocyte population is the only red blood cell population that can be accurately assessed for micronucleus frequency in peripheral blood due to efficient splenic scavenging of damaged erythrocytes soon after they emerge from the bone marrow. In both sexes of Wistar Han rats in the 3-month study, there were no significant increases in the frequencies of micronucleated PCEs (Table E-2). A positive trend in the percentage of PCEs was observed in female rats; however, the absolute increase (0.38%) in the 1,250 mg RES/kg body weight/day (mg/kg/day) group compared to the vehicle control group was very small and was not considered to be biologically relevant. No increases in the frequency of micronucleated erythrocytes (either immature or mature) were seen in the peripheral blood of female mice in the 3-month study (Table E-3). Significant increases in micronucleated NCEs were observed for every dosed group in male mice; however, the absolute difference in micronucleated NCEs in the dosed groups relative to the vehicle control group ranged from 0.06% to 0.16%. These very small increases were not considered to be biologically relevant.