Assessment of Grape , Plum and Orange Synthetic Food Flavourings Using in vivo Acute Toxicity Tests

Globalisation and the development of new technologies have caused signifi cant changes in the human eating habits, who, in recent decades, have frequently consumed foods rich in chemical additives (1). These substances, also called microingredients, include synthetic fl avourings of special importance for the food industry as they provide sensory properties of aroma and fl avour to all kinds of processed foods (2,3). These synthetic additives have a complex formulation containing diverse chemical compounds, such as diluents, preservatives and colourants, among others (4,5) and are classifi ed by the food industry into nature-identical and artifi cial additives (2).


Introduction
Globalisation and the development of new technologies have caused signifi cant changes in the human eating habits, who, in recent decades, have frequently consumed foods rich in chemical additives (1).These substances, also called microingredients, include synthetic fl avourings of special importance for the food industry as they provide sensory properties of aroma and fl avour to all kinds of processed foods (2,3).These synthetic additives have a complex formulation containing diverse chemical compounds, such as diluents, preservatives and colourants, among others (4,5) and are classifi ed by the food industry into nature-identical and artifi cial additives (2).
Worldwide, fl avour and aroma additives are standardised and released for use by food safety agencies Food and Agriculture Organization (FAO) and Flavour and Extract Manufacturer Association (FEMA) (2,3), and nation-ally by the National Sanitary Surveillance Agency (AN-VISA) through RDC Resolution 2 of January 15, 2007 (4).However, these regulatory agencies do not report in detail which compounds and concentrations are found in these microingredients, and to date also have not defi ned the acceptable daily intake (ADI) and the tolerable dose limit of fl avouring for each food type (1,2,4,6,7).Moreover, the enforcement agencies do not require industrialised food manufacturers to display on product labels the description of the substances that make up the fl avouring used in each product (4).Similarly, these agencies do not require the additive manufacturers to give the description on the label of the chemical components found in the aroma and fl avour additives (5)(6)(7).
Because of this lack of information, Konishi et al. (8), Koca et al. (5), Marques et al. (3) and Moura et al. (1) highlight the priority for evaluation of the cytotoxic, genotoxic and mutagenic potential of microingredients of aroma and fl avour.Similarly, the FEMA and ANVISA emphasise the need for studies on acute toxicological eff ect in diff erent bioassays, with the justifi cation for obtaining the data indicating the need for more detailed research on the toxicity of these substances (2,4,9).Cytotoxic, genotoxic and/or mutagenic compounds can alter vital cellular mechanisms, such as gene replication and transcription, and promote mitotic spindle alterations and chromosomal breaks.These changes can signifi cantly impair cell division of the aff ected tissue or organ, and trigger and/or potentiate cancerous processes (10)(11)(12).According to Zaineddin et al. (13), Moura et al. (1) and Santana et al. (14), the development of the most common types of cancer results from the interaction between endogenous and environmental factors, remarkably the diet, particularly when composed of processed foods in excess.
Root meristem cells of Allium cepa L. (onion) are an effi cient test system for initial screening of acute cytotoxicity and genotoxicity of chemicals (14,15).This bioassay provides excellent proliferation kinetic properties and a few chromosomes (2n=16), which facilitates the detection of mitotic spindle changes and chromosomal breaks (16,17).It also allows the verifi cation of changes in the cell division or mitotic index when exposed to chemical compounds with potential cytotoxic action (3,16), and shows mostly a satisfactory similarity to results obtained in other bioassays (1,3).
Among the methods assessing the damage caused by cytotoxic and/or mutagenic substances in animals, the micronucleus test carried out preferably in blood tissues of high proliferative activity stands out, since their cells are frequent targets of clastogenic agents (18).This test allows, by means of micronucleated erythrocyte frequency, to infer whether the tested substances or chemicals cause disorders, such as breaks and chromosome bridges and/ or chromosomal delays during cell division (18,19), and also to evaluate the cytotoxicity of substances through reduction of erythropoiesis in the analysed tissue.In this way, this bioassay is considered the main in vivo evaluation of acute exposure among cytotoxicity tests and therefore a very important parameter in the evaluation of safety in the use of compounds or chemicals in general (20).
Thus, the present study analyses the cytotoxic, mutagenic and genotoxic potential of synthetic, nature-identical, grape, plum and orange fl avourings with the use of root meristem cells of A. cepa and femoral bone marrow of mice.These additives were chosen for analysis because they are used extensively by the food industry in the manufacture of processed sweet foods, such as processed fruit juices, milk drinks, ice cream, jellies, candies, jams, mixed alcoholic drinks and liquors.

Food fl avourings
In this study, no dilution was made to set the doses of fl avourings, that is, the toxicity of these additives was tested directly from the original solutions sold on the market.This is because the aroma additives have complex chemical formulation not described in detail by food safety agencies, and so, the concentration and the action of compounds present in these microingredients could be altered if diluted.It is also important to mention that the formulation of all synthetic food fl avouring standards throughout the word and in Brazil is regulated by ANVI-SA (4).Synthetic, nature-identical, grape, plum and orange liquid fl avourings were obtained from a retailer specialised in national and international marketing of synthetic food additives (Trajano Food Additt ives, Recife, Brazil).The label of the three fl avorings suggested the use of 7.0 mL per kg of mass, so the fi rst dose set for the study was 7.0 mL/kg and the other two were 3.5 and 14.0 mL/kg.Flavouring solutions were stored in 100-mL amber bott les and were used before expiry date.Toxicity tests on A. cepa root meristem were performed between September and November 2015.
As mentioned previously, to date, there is no acceptable daily intake (ADI) set for fl avourings in general and there are no studies in the literature on the toxicological assessment of these substances in animal assays.Thus, for the assessment of the toxicity of grape, plum and orange additives, the doses of 0.5, 1.0, 2.0, 5.0 and 10.0 mL were determined based on the method proposed by Miller and Tainter (21).

Determination of cytotoxicity by Allium cepa assay
Onion bulbs were placed in aerated bott les with distilled water at room temperature (approx.27 °C) until obtaining approx.2.0 cm long roots.Five onion bulbs per each experimental group were used.Before treatments, some roots were fi xed in Carnoy's solution 3:1 (ethanol/ acetic acid) for 24 h and used as control.
Next, the remaining roots were placed in their respective solutions of fl avouring for 24 h and then some roots were taken and fi xed in Carnoy´s solution.Subsequently, the rest of the roots on each bulb were returned to their respective solutions, where they remained for another 24 h.Then, the roots were collected and fi xed again.The exposure times of 24 and 48 h were set to evaluate the eff ects of fl avourings on more than one cell cycle.At each collection, on average, three roots per bulb were taken.
On average, three slides were mounted per bulb, following the protocol proposed by Guerra and de Souza (22), and analysed under an optical microscope (Zeiss, São Paulo, Brazil) with 40× objective lens.Per each bulb, 1000 root cells were analysed, totalling 5000 cells of the control, and root cells treated for 24 and 48 h.Cells in interphase, prophase, metaphase, anaphase and telophase were examined.
The number of interphase and dividing cells was calculated for each control and exposure time and the cell division or mitotic indices for evaluation of the cytotoxic eff ect were determined.Mutagenic action of fl avourings was analysed by determining the number of micronucleated cells, and genotoxicity by mitotic spindle changes.

Cytotoxic eff ect of synthetic fl avourings on mice
In order to evaluate the toxicity of grape, plum and orange fl avourings in an animal bioassay, Swiss mice (Mus musculus L.), males and females, three months of age and 50 g of average body mass, were provided by the Central Animal House, Federal University of Piauí (Teresina, Brazil).During the experiment, mice were kept in plastic cages at a constant temperature of (25±2) °C and 12 h light/dark cycle, and fed standard commercial chow and water ad libitum.
Mice used in this study were treated according to the principles set by the Brazilian College on Animal Experimentation (COBEA) (23) and in accordance with the requirements of Brazilian law (24).The experimental protocol with rodents developed in this study was previously approved by the Ethics Committ ee for Animal Experimentation (CEEA) of the Federal University of Piauí (opinion 008/2015).
A total of seven experimental groups were established for the analysis of fl avourings: two control groups, one consisting of non-treated animals, and the other containing animals treated with the doses of 50 mg/kg of cyclophosphamide (Merck, São Paulo, Brazil), equivalent to 36.45 % lethal dose; and fi ve groups of animals treated with grape, plum or orange fl avourings (in mL/kg): 0.5, 1.0, 2.0, 5.0 and 10.0.
For each experimental group, three mice were randomly selected according to gender.Flavourings were administered via gavage in a single daily application for seven days using a syringe for oral administration of small dosages.Mice of the positive control group received a diff erent treatment consisting of cyclophosphamide at 50 mg/kg intraperitoneally, only 24 h before the sacrifi ce.On the eighth day, aft er sacrifi ce by cervical dislocation, femora were removed by surgery for bone marrow extraction.
Bone marrow extracted from mice was inserted into medium for karyotyping and then centrifuged in a tube twice at 1000×g (model MIKRO 185 centrifuge; Hett ich Lab Technology, Beverly, MA, USA) for 5 min.A drop of the suspension obtained from each animal was spread on a slide.Aft er drying, the material on the slides was fi xed in pure methanol for 10 min, allowed to dry at room temperature with Giemsa (Merck) diluted in phosphate buffer, pH=6.8, at a ratio of 1:10 for 15 min.Aft er staining, the slides were washed in distilled water and air-dried.
The bone marrow material was analysed under oil immersion microscope (Zeiss Microscopy Brasil, São Paulo, Brazil) in which for each smear, 200 polychromatic erythrocytes (PCE) were examined.To determine the cytotoxicity, a total of 400 PCE and normochromatic erythrocytes (NCE) were counted per animal (200 per blade) and the PCE frequency was determined as a ratio of PCE and PCE+NCE.For mutagenic assessment, micronucleated cells were counted in 1000 erythrocytes per animal.

Statistical data analysis
The results obtained in Allium cepa were analysed by χ 2 statistical analysis (p<0.05).Data obtained from the animal bioassay were tested by ANOVA followed by Tukey's post-hoc test using STATISTICA v. 7.0 soft ware (Dell, São Paulo, Brazil), and p<0.05 was adopted as signifi cance level.

Results and Discussion
Based on the results for grape fl avouring in Table 1, it is observed that aft er the exposure of Allium cepa root meristem cells to the dose of 3.5 mL/kg for 24 h, the mitotic index was not signifi cantly diff erent from its control.However, the cell division index aft er the exposure for 48 h to the same dose was statistically lower than the respective control and the above result.At doses of 7.0 and 14.0 mL/kg of this same fl avouring, cell division index aft er the exposure for 24 and 48 h was signifi cantly lower than those of their controls.In addition, comparison of the values of mitotic index between the samples exposed for 24 and 48 h did not show signifi cant diff erences.
Aft er the exposure of A. cepa cells to 3.5 mL/kg of the plum fl avouring for 24 and 48 h, a signifi cant decrease in cell division was observed compared to the mitotic index of its respective control (Table 1).The cell division index of samples exposed for 24 h was signifi cantly diff erent from the value aft er 48-hour exposure, when cell proliferation decreased dramatically with the increase of the exposure time.In turn, the values of cell proliferation aft er the exposure to 7.0 and 14.0 mL/kg of plum fl avouring and 3.5, 7.0 and 14.0 mL/kg of the orange fl avouring, as seen in Table 1, aft er both exposure times were significantly lower than those registered of their respective controls.However, there were signifi cant diff erences in mitotic index values between the samples exposed for 24 and 48 h to the same amounts of plum and orange microingredients.
According to the results in Table 1, all tested amounts of grape, orange and plum fl avouring were cytotoxic to root meristem cells of A. cepa at both exposure times, which is confi rmed by signifi cant antiproliferative eff ect caused to the cells.
Table 2 shows that 3.5 mL/kg of the grape fl avouring did not cause signifi cant damage to the meristem cells after 24 h of exposure.However, this dose aft er 48 h of exposure caused signifi cant alterations to the cells compared to its respective control and the results aft er 24 h of exposure.Nevertheless, the doses of 7.0 and 14.0 mL/kg of grape fl avouring induced the development of a significant number of cellular abnormalities aft er 24 h of exposure.All doses of the plum fl avouring induced a signifi -cant number of mitotic spindle changes and micronuclei in root cells 24 and 48 h aft er the exposure.All three doses of orange fl avouring did not induce a signifi cant number of alterations in the examined cells aft er both exposure times.Therefore, it can be concluded that the amount of 3.5 mL/kg of grape fl avouring aft er 48 h of exposure and 3.5, 7.0 and 14.0 mL/kg of plum fl avouring, aft er both exposure times, proved to be genotoxic and mutagenic to the tested A. cepa cells.
The signifi cant number of micronuclei and chromosomal alterations observed in root meristems treated with grape and plum fl avourings (Table 2) confi rms the antiproliferative eff ect in root meristem cells (Table 1).According to Santana et al. (14), inhibition of cell division is related to cell death caused by disturbances, such as toxic action of chemical substances or compounds to cell division kinetics or essential chromosomes or cells.These events, according to Gomes et al. (25) and Marques et al.
(3), cause signifi cant reduction in cell replacement and alter protein synthesis of the tissue or organ where they occur.
Regarding the evaluation of fl avouring toxicity to mice, animals treated with 2.0, 5.0 and 10.0 mL/kg of grape, plum or orange fl avouring died on days three, four and fi ve of the experiment, respectively.From the third day of treatment, all three animals treated with 10 mL/kg of grape microingredient had severe abdominal swelling.Unfortunately, it was not possible to determine the LD 50 (median lethal dose) of the three additives with the doses evaluated herein.The treatments with 0.5 and 1.0 mL/kg of the three fl avourings (Table 3) altered the maturation cycle of mouse bone marrow cells, reducing the number of polychromatic erythrocytes, and thus, under the study conditions, these doses can be characterised as cytotoxic.These results corroborate those observed for A. cepa (Table 1), where the investigated doses of grape, plum and orange additives drastically reduced the cell division index of root meristems, indicating a strong cytotoxic activity.
Table 3 also shows that the doses of 0.5 and 1.0 mL/kg of grape fl avouring had mutagenic potential because they induced a statistically signifi cant formation of micronucleated erythrocytes.The results observed in A. cepa for this microingredient (Table 2) corroborate the data observed for grape fl avouring in animal cells, where the treatments promoted cellular alterations, such as micronuclei, at a signifi cant frequency.It is known that micronuclei, being acentric fragments that were not incorporated into the nuclei of cells during telophase, can cause cell death due to suppression or reduction of expression of primary or primordial genes.Thus, the signifi cant presence of micronuclei in tissue promotes systemic cytotoxicity, resulting from high mutation rates and, consequently, genetic instability in cells (26).Importantly, although genetic toxicity is not a measure of carcinogenicity, it is oft en related to the onset of cancer, since there is a positive correlation between the increased frequency of micronuclei and the appearance of tumours in mammals (27).
Unfortunately, the chemical composition of grape, plum and orange fl avourings was not found in the literature or on the labels.However, the scientifi c literature demonstrates the toxicity at the cellular level of chemical constituents with diluent and preservative activities, according to RDC Resolution (4) on the basic formulation of fl avourings, and corroborates the data obtained for the three fl avourings evaluated in this study.Among these compounds benzoic alcohol stands out, as it is responsible for maintaining uniformity and facilitating incorporation and dispersion of the fl avour in food products.Analysing the action at the cellular level of this diluent, Demir et al. (28) found that the alcohol promoted signifi cant damage to the mitotic spindle and therefore to cell division in human peripheral blood cells.
Furthermore, diacetyl (butane-2,3-dione) is another diluent found in the formulation of fl avourings.In lymphoma gene mutation assay in rats, Whitt aker et al. (29) reported that this compound caused signifi cant damage to the loci on chromosome 11 of these cells, causing loss of the enzyme thymidine kinase gene expression.Likewise, More et al. (7) observed that the diluent diacetyl had the potential to replace thymine with guanine in euchromatin regions and caused the disruption of hydrogen and disulfi de bonds in the tertiary structure of enzymes involved in the cell division.
In turn, among the chemical constituents responsible for delaying the action of microorganisms, enzymes and physical agents in fl avouring solutions, potassium benzoate, sodium benzoate and potassium nitrate (9) are preservatives that, according to Mpountoukas et al. (30) and Zequin et al. (6), were cytotoxic and genotoxic to normal human peripheral blood cells.The fl avourings also contain the preservatives boric acid, citric acid, potassium citrate and sodium citrate (9), which, in agreement with Türkoğlu (31), resulted in signifi cant reduction of the cell division index of root meristem cells of A. cepa, proving to be cytotoxic.
Currently, the only class of compounds in the formulation of food fl avourings that have restricted use standardised by food safety agencies is the class of extraction solvents, among which the agaric acid, aloin, β-asarone, berberine, coumarin, hydrocyanic acid, hypericin, pulegone, quassin, safrole and isosafrole, santonin and α-and β-thuyone have maximum tolerable limits legally determined (4,9).Nevertheless, according to Moura et al. (1) and Konishi et al. (8), the composition of fl avourings in general includes 11 classes of chemical compounds, where each consists of, on average, 20 chemical compounds, which have not been evaluated for their cytotoxic, mutagenic and genotoxic potential.
In relation to the toxicity of fl avourings in general, ANVISA (4), although not mentioning which studies, concentrations and compounds, or which fl avourings have led to such a conclusion, stated that high doses of such synthetic additives cause annoying and narcotic actions in the organism and may cause toxicity in digestive tract when used chronically and indiscriminately (4,9).Further, Salinas (32) and Polônio and Peres (33) state that the use of fl avourings at low doses does not pose risk to human health; however, when used in doses higher than recommended, these substances can cause annoying and narcotic actions besides chronic cellular toxicity in the long term.Also, these authors did not specify which doses or concentrations of these additives are considered to be high or low, and they do not discriminate which fl avourings or test organisms were used to obtain this information.
In this way, although the use of fl avourings is permitted by EFSA, FEMA (10,25) and ANVISA (4), there is a pressing need for more detailed medium-and long-term studies using diff erent tests, dosages and time of exposure to determine the toxicity of these substances.Furthermore, our fi ndings, although preliminary, together with the results of the evaluation of toxicity at the cellular level of compounds in the formulation of fl avourings already conducted, indicate the need to defi ne the chemical composition of fl avourings in general using high performance liquid chromatography to properly determine the toxicity of these additives and ensure the safety of consumers.

Conclusion
Synthetic grape, plum and orange fl avourings under the study conditions caused acute toxicity to root meristem cells of Allium cepa and to femoral bone marrow cells of mice.The results of toxic activity obtained at the cellular level of grape, plum and orange food fl avourings are of utmost importance because, to date, there are no toxicity studies on these additives and no daily intake level has been set for such microingredients.Also, these fl avouring solutions are present in many processed foods and are freely traded in retail markets and websites specialised in the sale of food additives.

Table 1 .
Number of cells observed in each phase of the cell cycle in root meristem tissue of Allium cepa treated with water and synthetic grape, plum and orange fl avourings V=volume of the fl avouring solution, m=mass (on average 0.2 kg) of onion bulb used, TCII=total number of interphase and undiff erentiated cells, P=prophase, M=metaphase, A=anaphase, T=telophase, TCD=total number of dividing cells, MI=mitotic index.Within the same treatment, MI values followed by diff erent lett ers are signifi cantly diff erent at 5 % by χ 2 test

Table 2 .
Number and types of cellular abnormalities observed in root meristem cells of Allium cepa treated with water and synthetic grape and plum fl avourings V=volume of the fl avouring solution, m=mass (on average 0.2 kg) of onion bulb used, N(TCA)=number of total cellular alterations.Within the same treatment, TCA values followed by diff erent lett ers are signifi cantly diff erent 5 % by χ 2 test

Table 3 .
Cytotoxic and mutagenic potential of grape, plum and orange fl avourings determined in femoral bone marrow cells of mice (Mus musculus) treated orally for seven days