Genotoxicity Assessment of Three Nutraceuticals Containing Natural Antioxidants Extracted from Agri-Food Waste Biomasses

Grapes and apples are the most cultivated fruits in the Mediterranean basin and their agricultural processing is responsible for the production of a large amount of bio-waste. The reuse of this food biomass would increase the volume of recyclable/renewable biomaterial and lower the environmental impact due to the increasing demand for these biological products. To this purpose, agri-food waste from grape and apple processing have become an important source of phytochemicals, and many pharmaceutical industries are using it as starting material to produce dietary supplements, functional foods, and food additives for human consumption. In virtue of the chemical diversity and complexity of agri-food biowaste, developers and producers of nutraceuticals are advised to assess the safety of their final nutraceutical products, in compliance with European Food Safety Authority regulation. Here, we use the Ames test to assess the mutagenicity of three nutraceuticals obtained from agri-food waste biomasses: Taurisolo® from grape pomace of Vitis vinifera L. cv ‘Aglianico’, AnnurComplex® from Malus pumila M. cv ‘Annurca’ and Limoncella Apple Extract from Malus domestica B. cv ‘Limoncella’. The results showed that all three nutraceuticals were non-mutagenic.


Introduction
As a consequence of its consumeristic habit, modern society deals with an excessive demand for food, food products, and material of biological origin. The environmental impact caused by the handling of this massive request must be minimized [1]. Taking into account the increase in the world population expected in the near future (9.5 billion people in 2050 and 11.2 billion people in 2100) [2], the World Health Organization considers the achievement of this minimization a mandatory task for modern society. One of the ways environmental impact can be reduced is by enhancing renewable resources, especially those taking part in the production and processing of agricultural biomasses [3].
As described by Xia et al. [4], food waste reaches, globally, a billion tons per year, an amount that can be definitely decreased but that will be never completely erased. During food processing,~75% in weight of the starting biological material becomes waste, contributing 140 billion tons of biomass from the food sectors generated each year around the world. A considerable part of this biomass is made up of agricultural waste (leaves, roots, stalks, bark, bagasse, straw residues, seeds, and woods). Another significant part of the agri-food waste is made by food products (mostly 2nd-best fruits and vegetables), whose weight, shape and/or aesthetic features do not reflect those requested by the modern global market. synergistic or antagonistic effects, the pharmacokinetic parameters, bioavailability, bioaccessibility, bioactivity as well as overall toxicity of the whole product could be different from that of its components. As a consequence, the safety of individual substances cannot be used to draw general conclusions on whole extracts and botanical preparations, whose safety must thus be confirmed.
In 2009, the EFSA published the Guideline entitled "Safety assessment of botanicals and botanical preparations intended for use as ingredients in food supplements" [22]. The suggested approach to assess safety divides the evaluation into core areas and includes the assessment of mutagenicity of the food supplement by means of the Ames test (Organization for Economic Cooperation and Development (OECD) guideline 471) [24][25][26][27]. The same assesment and test is suggested and requested by other regulatory advisory bodies (e.g., European Chemicals Agency; UK Committee on Mutagenicity; the US Environmental Protection Agency).
The Ames test identifies and measures mutations occuring in bacterial genome upon exposure to a test chemical. DNA damage can be considered a surrogate endpoint for carcinogenicity [28,29], since the latter occurs in mammals as a consequence of an accumulation of a series of mutations. The test exposes five bacterial strains (S. typhimurium and E. coli) to a test substance. The bacterial strains used in the Ames test present mutations in genes needed for the synthesis of a specific amino acid (His in Salmonella and Trp in E. coli) [24,28]. Differently from wt strains, these bacteria lose auxotrophy for that amino acid and cannot grow in its absence. Furthermore, the bacterial strains present the rfa mutation, making their cell membrane more permeable to large molecules, and a deletion of the uvrB gene, abolishing the excision repair process, resulting in an increased sensitivity to mutagenic agents [24]. To identify chemicals aquiring genotoxicity upon in vivo metabolization, substances are tested with and without a metabolic activation system derived from rodent liver microsomes (called S9). Genotoxic chemicals revert the mutated gene sequence of the bacterial strain to the wt sequence allowing the growth of revertant colonies in the absence of amino acids [24].
Here, we use the Ames test to assess the mutagenicity of three nutraceuticals obtained from agri-food waste biomasses. The first phytocomplex tested is an extract of a grape pomace from Vitis vinifera cv 'Aglianico'. This extract is used to prepare 'Taurisolo ® ', an antioxidant able to reduce serum levels of the cardiovascular risk factor markers oxidized-low-density lipoprotein (LDL) and Trimethylamine N-oxide (TMAO) in humans and in rodents [30][31][32], and improve microcirculation.
The other two nutraceuticals tested contain extracts of two italian apple cultivar, namely Malus pumila M. cv 'Annurca' and Malus domestica B. cv 'Limoncella'. Malus pumila Miller cv. Annurca is a widespread apple and accounts for 5% of Italian apple production. It is listed as a Protected Geographical Indication (PGI) product from the European Council (Commission Regulation (EC) No. 417/2006)). Annurca Apples are used to prepare 'AnnurComplex ® ', a nutraceutical able to reduce serum cholesterol levels, LDL and lipid uptake; reduce cardiovascular disease (CVD) risk; and promote hair growth in humans [33][34][35][36][37]. Malus domestica cv 'Limoncella' is a juicy and aromatic variety of apple, known since ancient Roman times. Limoncella Apple extract presents high antioxidant activity and it has been shown to reduce colon inflammation and to act as potent inhibitor (in vitro and ex vivo) [38] of the Wingless-related integration site (WNT)-β catenin pathway, a signaling cascade linked to inflammation, oxidative stress, cell proliferation and cancerogenesis.
By means of the Ames test, we here show that the three nutraceuticals are not mutagenic. Our results support their safety and prove that the use of apple and wine waste product to produce polyphenolic mixtures does not enrich the final nutraceuticals of byproducts and secondary metabolites endowed with mutagenic potential.

Ames Test
Ames test was performed following the guidelines of OECD 471 [25,39,40] as described in Appendix A of this manuscript.

Results
Tables A1-A15 (see Appendix B) present the raw data obtained performing the Ames test on the three nutraceuticals. Each table shows the number of revertants per plate, measured in three replicates, their means, the standard deviation (sd) observed in S. typhimurium strains TA98, TA100, TA1535, TA1537 and in the E. coli strain WP2 trp UvrA upon treatment with different doses of Taurisolo  (Tables A1-A5), AnnurComplex (Tables A6-A10) and Limoncella Apple Extract (Tables A11-A15). Treatments were performed in the presence (+S9) and in the absence (−S9) of metabolic activation.
Raw data are summarized in Tables 1-3, where the mutagenic index of Taurisolo (Table 1), AnnurComplex (Table 2) and Limoncella Apple Extract (Table 3) observed in the different bacterial strains are compared to known genotoxic substances. The mutagenic index is the ratio between the average number of revertants per plate measured upon incubation with the test nutraceutical and the average number of revertants per plate measured upon incubation with the negative (solvent) control. A chemical can be considered mutagenic when a two-fold increase in mutagenic index is observed upon treatment with at least one of the tested concentrations.  Mean revertants per plate and mutagenic index measured in bacterial strains TA98, TA100 and TA1535 and 1537 and WP2 treated with Taurisolo at various doses, with (+S9) or without (−S9) metabolic activation. Negative controls consisted of 100 µL water. Positive controls consisted: for S. typhimurium TA100 NaN3 (−S9) and BAP (+S9); for S. typhimurium TA98 2NF (−S9) and BAP (+S9); for S. typhimurium TA1535 NaN3 (−S9) and 2AA (+S9); for S. typhimurium TA1537 9AC (−S9) and 2AA (+S9); for E. coli WP2 trp UvrA NQO (−S9) and 2AA (+S9).

Discussion
The perception that food bio-waste is going to have a huge impact on human health is rapidly growing in modern society. The increasing demand for products of natural origin is exposing air, lands and waters to an unprecedented risk of pollution, with anticipated consequence on the whole ecosystem. Nowadays, agri-food biowaste, owing to their peculiar composition, is being used as starting material for sustainable production of pure chemicals and botanicals to be used in the pharmaceutical industry. Thus, while the processing of grapes and apples, the most cultivated fruits in the Mediterranean basin, are responsible for a large amount of bio-waste [5], the same can be reused for the production of nutraceuticals and food supplements, ultimately contributing to making this biomass recyclable and renewable [6].
By virtue of the peculiar nature of the agri-food biowaste used as starting material for preparation of these nutraceuticals, developers and producers must assess the safety of their products, in compliance with EFSA regulation [22,25]. During the last decades, the non-mutagenicity of several food components (including polyphenols) has been confirmed by means of the Ames test. For a few of them, including wine and apple components quercetin [41], kaempferol [42] caffeic and chlorogenic acid [43], the scientific community is still debating on their mild in vitro genotoxicity. It must be mentioned, however, that these reports tested the genotoxicity of pure polyphenols and not of phytocomplexes. Since the mutagenic activity of individual substances can be altered by the presence of other components of the mixture, their safety (or the unsafety) cannot be thus used to deduce that of a whole botanical preparation. Furthermore, considering that the chemical composition of agri-food biowaste can be influenced by several parameter (agro-geographic factors, type of cultivar, harvesting time, extraction protocol), the genotoxic potential must be assessed for each new nutraceutical formulation.
Genotoxic assessment appears necessary especially for food supplements and nutraceuticals containing Vitis vinifera. The scientific reports assessing the genotoxic potential of phytocomplexes gave indeed controversial results in terms of safety and mutagenicity of grape-derived ingredients. In 1982 and 1984 Stoltz and colleagues [44,45] analyzed the mutagenic potential of a wide variety of food and food products by means of the Ames test. In their surveys, components of polar fractions of raw grapes and grape juice demonstrated potent mutagenic activity, both in the presence and absence of metabolic activation. The work of Stoltz was followed by Patrineli et al. [46], who showed in 1996 a potent mutagenic activity of unfermented white grape juice. The same group suggested that this mutagenic potential of grape juice could be attributed, at least in part, to reactive oxygen species production, which may emanate from the one-electron reduction of quinoid structures present in grape juice [47].
Safety assessment of a grape extracts was assessed in 2002 by Yamakoshi et al. [48], who tested the genotoxic potential of proanthocyanidin-rich extract prepared from grape seeds of Vitis vinifera L. The authors did not measure an increase in the number of revertant colonies in the Ames test, either in the presence or absence of S-9 mix, claiming the non-genotoxicity of the extract. To our knowledge, the first report of the genotoxic assessment of a polyphenols-rich extract obtained from red grape pomace appeared in 2011, in a report published by Lluis et al. [49]. In their work, the authors attribute to their grape pomace extract a weak genotoxic activity. The grape pomace extract was shown to be mutagenic for only one of the tested strain (S. typhimurium TA1537) at the highest dose tested (5 mg/plate).
Here, we show that, following Ames test procedure, Taurisolo does not show signs of mutagenicity. It is hard to compare the results obtained by Lluis and colleagues [49] with those obtained here on Taurisolo. The difference in the outcome might be ascribed to a difference in the cultivar used as starting material, as well as the procedure used to produce the grape pomace extracts, (extraction in water and spray drying for Taurisolo vs. extraction in 50% ethanol followed by chromatographic enrichment of polyphenols for the extract of Lluis et al. [49]).
Regarding apple extracts obtained from agri-food biowaste, the closest food supplements to AnnurComplex and Limoncella Apple Extract has been described by Shoij et al. [19]. In their manuscript, the authors described the genotoxic activity of a polyphenol-rich extract produced from unripe apples. Assayed by means of Ames test, their extract showed a slight increase in the number of revertants at the dose of 2.5 mg/plate on the S. typhimurium TA98 strain without metabolic activation. However, none of the other bacterial strains tested (TA100, TA1535, TA1537, WP2) showed an increase in the number of revertants, with or without S9 mixture, at doses up to 5.0 mg/plate. Here, we show that AnnurComplex and Limoncella Apple Extract did not show sign of mutagenicity. The difference between the results may, again, be attributed to the different cultivar used to produce the supplements as well as the procedures followed (extraction in water and spray drying for AnnurComplex and Limoncella Apple Extract vs. treatment with pectolytic enzyme followed by chromatographic enrichment of polyphenols for the apple extract described by Shoij et al. [19].

Conclusions
The use of agri-food biowaste as starting material to produce nutraceuticals, food supplements and fortified foods can increase the amount of renewable and recyclable biomass, ultimately lowering the environmental impact caused by the high demand for biological products. The safety of the resulting nutraceuticals must, however, always be assessed. The three nutraceuticals Taurisolo, Annutricomplex and Limoncella Apple Extract obtained from grape pomace of Vitis vinifera L. cv 'Aglianico', Malus pumila M. cv 'Annurca' and Malus domestica B. cv 'Limoncella', respectively, were all assayed by means of the Ames tests and resulted to be non-mutagenic. For S. typhimurium TA98 strain, phenotype confirmation was performed by growing an overnight culture in Oxoid No. 2 broth to then challenge 1-2 × 10 8 of bacteria as follows: (a) the strain did not grow on agar minimal plate in the absence of L-His confirming the hisphenotype; (b) the strain manifested zonal growth inhibition on agar minimal plate in the presence of L-His and of a 10 µg Crystal Violet disc. confirming the rfa phenotype; (c) the strain did grow on agar minimal plate containing L-His and a 2 µg Ampicillin disc confirming the presence of the R-factor plasmid; (d) the strain did not grow on agar minimal plate containing L-His, 2 µg Ampicillin and 1 µg Tetracycline disc confirm the absence of the pAQ1 plasmid; (e) the strain did not grow on agar minimal plate containing L-His and 0.2 µg Mitomycin disc confirming the uvrA/B phenotype and thus the absence of an active excision repair. The S. typhimurium TA98 strain yielded spontaneous revertant colony plate counts within the frequency ranges expected from the laboratory's historical control data and within the range reported in the literature [23,28,40] For S. typhimurium TA100 strain, phenotype confirmation was performed by growing an overnight culture in Oxoid No. 2 broth to then challenge 1-2 × 10 8 of bacteria as follows: (a) the strain did not grow on agar minimal plate in the absence of L-His confirming the hisphenotype; (b) the strain manifested zonal growth inhibition on agar minimal plate in the presence of L-His and of a 10 µg Crystal Violet disc confirming the rfa phenotype; (c) the strain did grow on agar minimal plate containing L-His and 2 µg Ampicillin confirming the presence of the R-factor plasmid; (d) the strain did not grow on agar minimal plate containing L-His and of a 2 µg Ampicillin and 1 µg Tetracycline disc confirming the absence of the pAQ1 plasmid; (e) the strain did not grow on agar minimal plate containing L-His and a 0.2 µg Mitomycin disc confirming the uvrA/B phenotype and thus the absence of an active excision repair. The S. typhimurium TA100 strain yielded spontaneous revertant colony plate counts within the frequency ranges expected from the laboratory's historical control data and within the range For S. typhimurium TA1535 strain, phenotype confirmation was performed by growing an overnight culture in Oxoid No. 2 broth to then challenge 1-2 × 10 8 of bacteria as follows: (a) the strain did not grow on agar minimal plate in the absence of L-His confirming the hisphenotype; (b) the strain manifested zonal growth inhibition on agar minimal plate in the presence of L-His and 10 µg Crystal Violet disc confirming the rfa phenotype; (c) the strain did not grow on agar minimal plate containing L-His and a 2 µg Ampicillin disc confirming the absence of the R-factor plasmid; (d) the strain did not grow on agar minimal plate containing L-His and a 2 µg Ampicillin and a 1 µg Tetracycline disc confirming the absence of the pAQ1 plasmid; (e) the strain did grow on agar minimal plate containing L-His and 0.2 µg Mitomycin disc confirming the uvrA/B phenotype and thus the absence of an active excision repair. The S. typhimurium TA1535 strain yielded spontaneous revertant colony plate counts within the frequency ranges expected from the laboratory's historical control data and within the range reported in the literature. Mean revertant per plates: (water, number of colonies 6), Daunomycin (6 µg, number of colonies 5), ICR191 (1 µg, number of colonies 7), Mitomycin C (0.5 µg, number of colonies 0 colonies). NaN 3 (1.5 µg, number of colonies 329 colonies).
For S. typhimurium TA1537 strain, phenotype confirmation was performed by growing an overnight culture in Oxoid No. 2 broth to then challenge 1-2 × 10 8 of bacteria as follows: (a) the strain did not grow on agar minimal plate in the absence of L-His confirming the hisphenotype; (b) the strain manifested zonal growth inhibition on agar minimal plate in the presence of L-His and a 10 µg Crystal Violet disc confirming the rfa phenotype; (c) the strain did not grow on agar minimal plate containing L-His and a 2 µg Ampicillin disc confirming the absence of the R-factor plasmid; (d) the strain did not grow on agar minimal plate containing L-HisAnd a 2 µg Ampicillin and 1 µg Tetracycline disc confirming the absence of the pAQ1 plasmid; (e) the strain did grow on agar minimal plate containing L-His and 0.2 µg Mitomycin disc confirming the uvrA/B phenotype and thus the absence of an active excision repair. The S. typhimurium TA1537 strain yielded spontaneous revertant colony plate counts within the frequency ranges expected from the laboratory's historical control data and within the range Metabolic activation. Metabolic activation of nutraceuticals was achieved by exogenous metabolization using S9 post-mitochondrial fraction. S9 (code 11-402L. LOT NO. 4026) prepared from livers of Sprague Dawley male rats treated with Aroclor 1254 (500 mg/Kg i.p.). Lyophilized S9 was purchased from Trinova Biochem already supplemented with glucose-6-phosphatedehydrogenase (180 mg/mL) Nicotinamide adenine dinucleotide phosphate (25 mg/mL), Potassium chloride (150 mM) mixed in the ratio 2:1:1:1. S9 was reconstituted in deionized water and stored at −80 • C. The protein concentration of S9, assayed with the Lowry Method, was 3.5 mg/mL. To prove S9 able to activate pro-mutagens, we measured the number of revertant colonies of TA98 and TA1535 strains growing in the presence of S9 and of ethidium bromide and cyclophosphamide. TA98 strain yielded 52 colonies in the presence of ethidium bromide and TA1535 yielded 430 colonies in the presence of cyclophosphamide, respectively. Dilution of S9, ranging from 0.6 to 10% were tested for their ability to activate benzo (a) pyrene and 2-aminoanthracene (2-AA) to metabolites mutagenic to TA100. The final concentration of S-9 fraction in the test system was 7% v/v. Cultures treated in the absence of S9 received an equivalent volume of 0.1 M phosphate Buffer pH 7.4 in place of S9 mix.
Nutraceutical test conditions. Taurisolo, AnnurComplex and Limoncella Apple Extract are highly soluble in water that was thus used as vehicle for all the experiments. Mother stocks of Nutraceuticals 50 mg/mL were freshly prepared in water. The recommended maximum test concentration for soluble non-cytotoxic substances is 5 mg/plate. None of the nutraceutical gave precipitation on the surface of the agar plate. However, the grape pomace stained the agar plate in a dark purple color.
Test dilutions were obtained by diluting mother stocks in water. We tested eight dilutions for each of the nutraceuticals (namely, 0.0016, 0.005, 0.016, 0.05, 0.16, 0.5, 1.6. and 5 mg/10 cm plate) (volume 100 µL). Up to 5 mg/plate and on Oxoid Agar Plates, none of the nutraceutical-induced growth inhibition of the bacterial strains here tested, confirming that in the range of dilution here assayed, all the tested nutraceuticals were not cytotoxic for the bacteria strains.
Negative controls consisted of 100 µL water. Positive controls consisted: for S. typhimurium TA100, NaN 3 1.25 µg/10 cm plate in the absence of S9 and BAP 6.0 µg/10 cm plate in the presence of S9; for S. typhimurium TA98, 2NF 2.0 µg/10 cm plate in the absence of S9 and BAP 6.0 µg/10 cm plate in the presence of S9; for S. typhimurium TA1535, NaN 3 1.25 µg/10 cm plate in the absence of S9 and 2AA 2.0 µg/10 cm plate in the presence of S9; for S. typhimurium TA1537, 9AC 50.0 µg/10 cm plate in the absence of S9 and 2AA 2 µg/10 cm plate in the presence of S9; for E. coli WP2 trp UvrA, NQO 1.0 µg/10 cm plate in the absence of S9 and 2AA 20.0 µg/10 cm plate in the presence of S9.
Experimental Procedure. We used the plate incorporation method. Briefly, 0.1 mL of test solutions (the appropriated amount of nutraceutical dissolved in 0.1 mL of water), 0.1 mL of fresh bacterial culture containing 10 8 viable cells and either 0.5 mL of 0.1 M Phosphate buffer (pH 7.4) or 0.5 mL of S9 were mixed with 2.0 mL of Top Agar. For the assay with metabolic activation, 0.5 mL of metabolic activation mixture contained 7% post-mitochondrial fraction. The contents of each tube were mixed and poured over the surface of a minimal agar plate. The overlay agar was allowed to solidify before incubation. Experiments were performed in triplicates for each condition. Plates were incubated at 37 • C for 72 h. After the incubation period, the number of revertant colonies per plate was counted.
Acceptance of the test. Acceptance of the test was based on the following criteria: (a) all experimental conditions requested by OECD 471 were tested; (b) the results obtained for the negative control were consistent with the laboratory's historical negative control database; (c) concurrent positive controls induced responses that were compatible with those generated in the laboratory's historical positive control database and produced a statistically significant increase compared with the concurrent negative control.