Evaluation of the safety and potential genotoxicity of a high-purity mogroside ingredient containing siamenoside I produced from a modified strain of Yarrowia lipolytica Authors: Amy L Mihalchik, Ashley Roberts, Jonathan LaCombe, Mel Lloyd, Candace Doepker, Alex Eapen Published: 2026 ABSTRACT AbstractA fermentation process-derived mogroside mixture (≥91.9% total mogrosides) containing high levels of siamenoside I (≥73%) was assessed in a bacterial reverse mutation assay, in vitro micronucleus assay, 28-day repeat-dose dose range-finding study, and 90-day repeat-dose oral toxicity study. Mogrosides, including siamenoside I, are cucurbitane glycoside derivatives of monk fruit (Siraitia grosvenorii) recognized as natural sweeteners. While monk fruit extracts containing concentrated levels of mogrosides (primarily mogroside V) have been marketed in the United States and evaluated by the European Food Safety Authority (EFSA), a safety assessment of fermentation process-derived mogrosides has not been previously conducted. No biologically-relevant mutagenic or genotoxic effects were reported in a bacterial reverse mutation or in vitro micronucleus assay conducted per Good Laboratory Practices (GLP) and relevant OECD test guidelines (OECD TG 471; OECD TG 487) at concentrations up to 5000 µg/mL in each study. A no-observed-adverse-effect-level (NOAEL) of 50,000 ppm (calculated time-weighted average [TWA] equivalent to 4410 mg/kg-bw/day in males and 4112 mg/kg-bw/day in females) was reported in a non-GLP-compliant 28-day dose-range finding study. Similarly, in the 90-day study conducted per GLP in accordance with OECD TG 408 at concentrations of 0, 12,500, 25,000, or 50,000 ppm in the diet, the NOAEL was 50,000 ppm in male (calculated TWA equivalent to 3309 mg/kg-bw/day) and female rats (calculated TWA 3362 mg/kg-bw/day). In the context of low anticipated consumer exposure in adults and children, these results support the safety of this mogroside mixture in foods.Keywords: Mogrosides, Siamenoside I, Dietary, Rat, Bacterial Mutagenicity Assay, In Vitro Micronucleus Assay Introduction Mogrosides, which include siamenoside I, are cucurbitane glycoside derivatives of monk fruit (Siraitia grosvenorii; also referred to as luo han guo or lo han kuo), a natural sweetener traditionally used in China ([1] Fig. 1). Structure of siamenoside I. Mogrosides are the primary sweeteners of the S. grosvenorii fruit, present at 0.55 – 0.65% w/w of the fresh fruit [2]. Until recently, monk fruit, an herbaceous plant primarily cultivated in specific geographic regions of China (i.e., Guanghxi, Guizhou, Hunan, Guangdong, and Jiangxi provinces), was the primary source of mogrosides. However, more recently, synthesis of mogrosides via microbial fermentation has been explored [3], [4], [5]. While mogroside V is the most abundant mogroside identified in monk fruit, siamenoside I exhibits the highest relative sweetness to sucrose. A 0.01% solution of siamenoside I was found to be 563 times sweeter than a 5% solution of sucrose based on panel of professional tasters and has gained considerable interest for use as a sweetener in food products [6], [1]. Monk fruit extracts and monk fruit juice concentrates are considered generally recognized as safe (GRAS) in the United States2 and are currently undergoing a follow-up review by the European Food Safety Authority (EFSA) which issued an initial opinion in 2019. The opinion concluded that the monk fruit extract database was insufficient to determine the safety of the ingredient as a food additive [2]. In contrast to monk fruit extract and juice concentrate products containing a range of mogrosides (primarily mogroside V) and other fruit components, the current ingredient manufactured by precision fermentation is a purified mogroside ingredient (≥91% purity) containing > 73% siamenoside I. Due to the significant differences in composition and purity of monk fruit extracts, a bacterial reverse mutation assay, in vitro micronucleus assay, 28-day dose range-finding, and 90-day repeat-dose oral toxicity study were conducted on two lots of a mogroside mixture to support the safety profile and establish the no-observed-adverse-effect-level (NOAEL) associated with the mogroside mixture. Thus, herein we discuss the results of genotoxicity assays, which were compliant with Good Laboratory Practice (GLP) standards (OECD ENV/MC/CHEM [98] 17; 21 CFR §58) and followed Organisation for Economic Co-operation and Development (OECD) guidelines (TG 471; TG 487), and were conducted at Labcorp (Harrogate, England, UK). A 28-day non-GLP-compliant dose-range finding study and 90-day study conducted per United States (US) Environmental Protection Agency (EPA) GLP standards (40 CFR §792) following OECD test guidelines (TG 408) were also completed at Labcorp (Greenfield, Indiana, USA). Since mixture safety data should be relevant to unique constituent profiles, this evaluation adds valuable new safety information into the public domain for mogrosides. Materials and methods Genotoxicity assays Dose formulation analyses The test article (Batch No. HY240515–01; 95.6% total mogrosides, 73% siamenoside I) was stored at ambient temperature (15–25°C) upon receipt and transferred to storage at 2–8°C one day following receipt and protected from light. Test article stock solutions were prepared under subdued lighting in purified water with vortex mixing, ultrasonication for 1–3 min for the Ames assay or up to 15 min for the in vitro micronucleus assay and warmed at 37°C to facilitate dissolution. Stock solutions were membrane filter-sterilized. Subsequent dilutions up to 50 mg/mL were made with purified water. Test article solutions were protected from light and used within 4 h of preparation for the Ames assay or 3.5 h for the in vitro micronucleus assay. Two sets of triplicate samples were acquired from the 50 mg/mL test article formulation and vehicle controls for dose formulation analyses from the dose-range finding and pre-incubation studies. Samples were stored at ambient temperature prior to shipping on dry ice under frozen conditions. Using a previously validated analytical method under GLP conditions, all test article samples met specification criteria for concentration. No significant peaks were identified in vehicle control samples. Metabolic activation Mutazyme™-rat liver S9 fraction (Lot 4884) was purchased from Molecular Toxicology, Inc. Mutazyme™ was stored frozen at ≤ 10°C and thawed and reconstituted just prior to use. The S9 mixture was prepared at a concentration of 10% (v/v). Additional constituents in the mixture on a per mL basis included sodium phosphate buffer pH 7.4 (100 µmol), glucose−6-phosphate (disodium) (5 µmol), β-nicotinamide adenine dinucleotide phosphate (disodium) (4 µmol), magnesium chloride (8 µmol), potassium chloride (33 µmol), water (to volume), and 100 µL S9. In the Ames assay, buffer solution contained 100 µmol sodium phosphate buffer pH 7.4 on a per mL basis and water (to volume). Treatments were carried out with and without metabolic activation in both the range-finder and pre-incubation experiments with 0.5 mL 10% S9 mix or buffer solution added to agar. In the in vitro micronucleus assay, treatments were carried out with and without S9 by diluting with either 150 mM potassium chloride or 10% S9 in HEPES-buffered RPMI medium containing 10% (v/v) heat inactivated fetal bovine serum and 0.52% penicillin / streptomycin, so that the final volume following addition of potassium chloride or S9 mix (1% v/v in test system) and the test article in its chosen vehicle was 10 mL. Mitogen phytohemagglutinin (PHA, reagent grade) was included in the culture media at ∼2% to stimulate lymphocytes to divide Bacterial reverse mutation assay (Ames assay) Cellular test systems Bacteria strains were used in accordance with OECD TG 471. Strains included Salmonella typhimurium TA98 (Moltox, US), TA1535 (UK NCTC), TA1537 (UK NCTC), TA100 (Labcorp Early Development Laboratories Inc, US), and Escherichia coli WP2uvrA (Cancer Research Unit, University of York, UK). Prior to all assays, bacteria were incubated for 10 h in nutrient broth (containing ampicillin for TA100 and TA98) at 37°C with shaking to produce bacterial cultures with a range between 108 and 109 cells/mL. Experimental procedures The mutagenic activity of the test article was assessed in two separate experiments at Labcorp (Labcorp Study Number 8523295; Harrogate, England, UK). The study was GLP-compliant and conducted per OECD TG 471. Seven dose concentrations (5, 16, 50, 160, 500, 1600, and 5000 μg/plate) were tested in Experiment 1 using the plate incorporation method. For each concentration level and condition, the mean revertant count and standard deviation (SD) were determined. Purified water was used as the vehicle control for the assay. Positive controls for tests without metabolic activation were 2-nitrofluorene (5 µg/plate; Sigma-Aldrich Inc.) for strain TA98, sodium azide (2.0 µg/plate; Sigma-Aldrich Inc.) for strains TA100 and TA1535, 9-aminoacridine (50.0 µg/plate; Sigma-Aldrich Inc.), and 4-nitroquinoline−1-oxide (2.0 µg/plate; Sigma-Aldrich Inc.) for WP2 uvrA. With metabolic activation, 2-aminoanthracene (Sigma-Aldrich Inc.) was used across TA100, TA1535, and TA1537 (5 µg/plate) and WP2 uvrA (15 µg/plate). Benzo[a]pyrene was used for strain TA98 (10 µg/plate; Sigma-Aldrich Inc.). As the results of Experiment 1 were negative, treatments in the presence of S9 in Experiment 2 included a pre-incubation step. Quantities of test article, vehicle control solution or positive control, bacteria and S9 mix or buffer solution detailed above, were mixed together and placed in an orbital incubator set to 37°C for 20 min, before adding 2 mL molten agar at 45 ± 1°C. Plating of these treatments then proceeded as per the normal plate-incorporation procedure. Data analysis and statistical analysis Under each test condition and concentration level, the mean revertant colony plate counts and SD were determined. Colonies were counted electronically using a Sorcerer Colony Counter (Perceptive Instruments, part of Instem Group) or manually when confounding factors (such as contamination of the test article) affected accuracy of the automated counter. The presence of a concentration response was assessed via non-statistical means. The test article was considered mutagenic 1) if treatments with the test article provided a concentration-related increase in revertant numbers at one or more concentrations in at least one strain with or without metabolic activation system; 2) an increase in mean revertant colony numbers per plate was observed ≥ 2-fold in TA98, TA100, and WP2 uvrA or ≥ 3-fold in TA1535 or TA1537 compared to concurrent vehicle control values; or 3) any increase in revertant numbers was reproducible. In vitro human lymphocyte micronucleus assay Cellular test systems Blood samples used to establish human whole blood cultures were obtained from two healthy, non-smoking males (35 years old or under) in both the range-finding and micronucleus experiments (Labcorp Study Number 8523296). The study protocol was reviewed and approved by the Labcorp-Greenfield Quality Assurance Unit. Per Standard Operating Procedures, the Labcorp Institutional Review Board (IRB) annually reviews and grants approval for human blood sample collection. All volunteer donors were reported to sign a written consent for use of their blood samples and double blinding is used to protect donor identity. Following collection, whole blood cultures were established in HEPES-buffered RPMI medium containing 10% (v/v) heat inactivated fetal bovine serum and 0.52% penicillin/streptomycin. 2% Phytohemagglutinin was included in culture media to stimulate lymphocytes to divide. Blood cultures were incubated at 37°C for approximately 48 h with continuous agitation prior to treatments. Experimental procedures Two GLP-compliant experiments were conducted per OECD TG 487 at Labcorp (Labcorp Study Number 8523296; Harrogate, England, UK). An initial dose-range finding study analyzing for test article cytotoxicity was conducted with and without S9 metabolic activation. Cultures were treated with the test article or vehicle (1.0 mL per culture). Tests were conducted using a range of concentrations from 18.14 to 5000 μg/mL. Positive control treatments were not included. Cytotoxicity was determined based upon the relative replication index (RI). The replication index was derived as follows (Equation 1) following assessment of 200 cells per concentration in the range-finder: Equation 1. Replication index derivation The relative RI (Equation 2) as a percentage was calculated as follows: Equation 2. Relative replication index derivation Based on the range-finder study, the highest concentration selected for micronucleus analysis following all treatment conditions was the highest concentration tested (i.e., 5000 μg/mL) [7]. In the main micronucleus (MNT) assay, human blood cultures were assessed with and without metabolic activation for 3 h with 21 h recovery (3 +21 h) prior to harvest, and without metabolic activation for 24 h plus 24 h recovery (24 +24 h) prior to arvest. Under the 3 + 21 h -S9, 3 + 21 h + S9 experimental conditions, cells were exposed to 0, 250, 500, 1000, 1500, 2000, 3000, or 5000 µg/mL of test article. Under the 24 + 24 h experimental conditions, cells were exposed to 0, 125, 250, 500, 750, 1000, 1250, 1500, 1750, 2000, 3000, or 5000 µg/mL test article. Vehicle control for the assay was purified water. Positive control concentrations were mitomycin-C at 0.3 µg/mL in 3 + 21-h cultures without activation, vinblastine at 0.04 µg/mL in 24 + 24-h cultures without activation, and cyclophosphamide at 5 µg/mL in cultures with activation. Cyto-B (formulated in DMSO) was added to post wash-off culture media for a final concentration of 6 µg/mL per culture. Harvest times were 72 h following culture initiation (i.e., 24 h post-exposure) for 3 + 21 h cultures with and without activation. Harvest time was 96 h following culture initiation (i.e., 48 h post-exposure) for 24 + 24 h cultures without activation. Micronucleus frequencies were analyzed in 1000 binucleate cells per culture. The highest concentration selected for evaluating micronucleated cells based on lack of observed cytotoxicity in the range-finder was 5000 µg/mL. Data analysis and statistical analysis Under each test condition and concentration level in the micronucleus experiment, 1000 binucleate cells from each culture were analyzed for micronuclei and nucleoplasmic bridges between nuclei in binucleate cells were determined. Slide analysis was performed by human analysts trained in applicable Labcorp Early Development Laboratories Ltd. standard operating procedures. The test article was considered clastogenic and/or aneugenic if 1) a statistically significant increase in frequency of micronucleated binucleate (MNBN) at one or more concentrations was observed; 2) an incidence of MNBN at a concentration that exceeded the normal range in both replicates was observed; and 3) a concentration-related increase in the proportion of MNBN cells was observed (positive trend test). Statistical analyses were conducted following scoring completion and slide decoding. The proportion of MNBN cells in each replicate were used to establish heterogeneity between replicates by means of a binomial dispersion test. The proportion of MNBN cells for each treatment condition were compared with the proportion of MNBN cells in vehicle controls using a Fisher’s exact test. The Cochran-Armitage trend test was applied to each treatment condition. In all tests, p ≤ 0.05 was accepted as significant. 28-day dose range study Animals Twenty-two male and 22 female Sprague Dawley (Crl:CD[SD]) rats were received from Charles River Laboratories (Raleigh, North Carolina, US); of these animals, 20 per sex were enrolled in the study. Animals were acclimated to the test facility for seven days prior to initiation. At study initiation, rats were eight weeks old; males and females weighed between 187 and 256 g and 172–252 g, respectively. Animals were randomized to five per sex per group (control, low, intermediate, high) and individually housed in polycarbonate cages with Diamond soft bedding. Water was provided ad libitum. Animals were offered food (Certified Rodent Diet #2014CM), unless fasted for study procedures. Temperature was kept at 20–26°C, and humidity was maintained between 30%–70%, with ≥ 10 air exchanges per hour, and a 12-hour day/night cycle. Animals were provided cage-enrichment devices. Labcorp is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). All procedures in the Protocol were in compliance with applicable animal welfare acts and were approved by the local Institutional Animal Care and Use Committee (IACUC) (Labcorp Study Number 8523298). Experimental procedures Rats were administered 0, 12,500, 25,000, or 50,000 ppm test article (Batch Nos. HY24015–01 [95.6% total mogrosides] or HY240605 [91.9% total mogrosides]) via diet for 28 days in a non-GLP study conducted at Labcorp (Greenfield, Indiana, USA) to identify doses for a subsequent 90-day study (Table 1). Table 1Experimental design of 28-day study.GroupaTest article (ppm)Number of AnimalsMalesFemales1 (Control)0552 (Low)12,500553 (Intermediate)25,000554 (High)50,00055aGroup 1 was administered carrier control item only (Certified Rodent Meal Diet #2014CM) Experimental design of 28-day study. Group 1 was administered carrier control item only (Certified Rodent Meal Diet #2014CM) Rats were observed twice a day for mortality, abnormalities, and signs of pain or distress, and once on days of arrival/sacrifice. Detailed observations and body weight measurements were made once during the predose phase and on Days 1, 8, 15, 22, and 28. Food consumption was recorded for a 1-week period during the predose phase, and Days 1–8, 8–15, 15–22 and 22–28 during dosing. Test item intake in mg/kg bw/day was determined based on food consumption, dietary concentration, and body weight and used to derive time-weighted average NOAEL values for male and female rats. Blood samples for hematology, coagulation, and clinical chemistry parameters were collected from fasted animals on Day 29 of the study. Animals were anesthetized with isoflurane, exsanguinated, and necropsied on Day 29 of the study. Bone marrow smears were prepared but not evaluated based on study findings. Macroscopic histopathologic observations and select organ weights (adrenal [2], brain, cervix, epididymis [2], heart, kidney [2], liver, ovary [2], oviduct [2], pituitary gland, prostate, spleen, testis [2], thymus, thyroid [two lobes] with parathyroid, uterus) were recorded for all animals; microscopic assessments were not conducted based on study findings. Data analyses and statistical analyses Data for each sex were analyzed separately; only data collected on or after the first day of dosing were analyzed statistically. Levene’s test was performed to assess equality of variances between groups prior to using analysis of variance (ANOVA) testing. ANOVA and pairwise comparisons (control compared to low-, intermediate-, and high-dose groups) were used to assess changes in absolute body weight, body weight change, quantitative food consumption, continuous clinical pathology values, terminal body weight, absolute organ weight, organ:body weight percentage, and organ:brain weight percentage. In cases where the ANOVA test was significant (p ≤ 0.05), Dunnett’s test was used for pairwise comparisons to a single control or combined identical controls; t-tests were otherwise used for pairwise comparisons. 90-day dietary toxicity study Dose formulation analyses The test article (Batch No. HY240605–01) was stored at ambient temperature (15–25°C) and had a reported purity of 91.9 % (total mogrosides); 75.8 % was considered to be siamenoside I. All diet formulations met acceptance criteria and were within ±15% of the target concentrations. No test article was detected in the control diet. Animals Forty-four male and 44 female Sprague Dawley (Crl:CD[SD]) rats were received from Charles River Laboratories (Raleigh, North Carolina, US); of these animals, 40 per sex were enrolled in the study. They were received 7 days prior to study initiation to acclimate to the test facility. At study initiation, rats were 8 weeks old; males and females weighed between 193 and 313 g and 153–226 g, respectively. Animals were randomized to ten per sex per group (i.e., control, low, intermediate, high). Animals were pair-housed in polycarbonate cages with Diamond soft bedding and individually housed in stainless steel or polycarbonate cages for study-related procedures. Water was provided ad libitum. Animals were offered food (Certified Rodent Diet #2014CM [Lot Nos. 2014CM−070324MA; 2014CM−080624MA]), unless fasted for study procedures. Temperature was kept at 20–26°C, and humidity was maintained between 30%–70%, with ≥ 10 air exchanges per hour, and a 12-hour day/night cycle. Animals were provided cage-enrichment devices. Labcorp is fully accredited by the AAALAC. All procedures in the Protocol were in compliance with applicable animal welfare acts and were approved by the local IACUC (Labcorp Study Number 8523305). Experimental procedures Rats were administered 0, 12,500, 25,000, or 50,000 ppm test article via diet for 90 days in a study that was GLP-compliant and conducted per OECD TG 408 and US EPA OCSPP 870.3100 at Labcorp (Greenfield, Indiana, USA) (Table 2). Table 2Experimental design of 90-day study.GroupaTest article (ppm)Number of AnimalsMalesFemales1 (Control)010102 (Low)12,50010103 (Intermediate)25,00010104 (High)50,0001010aGroup 1 was administered carrier control item only (Certified Rodent Meal Diet #2014CM) Experimental design of 90-day study. Group 1 was administered carrier control item only (Certified Rodent Meal Diet #2014CM) Rats were observed twice a day for mortality, abnormalities, and signs of pain or distress. Detailed observations and body weight measurements were made once during the predose phase and on Days 1, 8, 11, 18, 25, 32, 39, 46, 53, 60, 67, 74, 81, 88, and 90. Food consumption was recorded for a 1-week period during the predose phase, daily during Week 1 of dosing, weekly from Weeks 2–13 of dosing, and from Days 88–90 during dosing. Test article intake in mg/kg-bw/day was determined based on mean food consumption in g/animal/day, dietary concentration (in ppm), and calculated mid-interval body weight. Based on raw test article intake data (in mg/kg-bw/day) provided in the study report, NOAEL values for male and female rats were derived. For dosing intervals where food consumption data were excluded by the study director (i.e., where food was assumed to be wet, spilled, or contaminated), group test article intake data for the dosing interval were excluded from derivation of the time-weighted average NOAEL. Ophthalmic examinations were conducted for all animals once during the predose phase and once for the control and high-dose groups during Week 12 of dosing. All animals were examined with the functional observational battery (FOB) (i.e., arena observations, elicited behaviors, grip strength) and for locomotor activity during Week 12 of dosing. Estrous stages were determined on Day 91 of the dosing phase via vaginal lavage after thyroid hormone analysis collection. Thyroid hormone analyses (triiodothyronine [T3], thyroxine [T4], and thyroid stimulating hormone [TSH]) were conducted on blood samples on Day 91 of the study. Blood samples for hematology, coagulation, and clinical chemistry parameters were collected from fasted animals on Day 91 of the study; urine samples for urinalysis were also collected during the overnight period before blood collection on Day 91 of the study. Animals were anesthetized with carbon dioxide, exsanguinated, and necropsied on Day 91 of the study. Macroscopic and histopathologic observations and select organ weights (as outlined in OECD 408) were recorded for all animals; microscopic assessments were conducted in control and high-dose animals. A qualitative examination of spermatogenesis was assessed in control and high-dose males. During the scheduled sacrifice, the right vas deferens and epididymis were collected from all males. Sperm analyses (count, morphology) were conducted using these tissues. Data analyses and statistical analyses Several statistical analyses were used to evaluate the data collected in the 90-day rat study; data for each sex were analyzed separately. Levene’s test was performed to assess equality of variances between groups prior to using ANOVA testing. ANOVA and pairwise comparisons (control compared to low-, intermediate-, and high-dose groups) were used to assess changes in absolute body weight, body weight change, quantitative food consumption, continuous clinical pathology values, terminal body weight, absolute organ weight, organ:body weight percentage, organ:brain weight percentage, and continuous data associated with the FOB. In cases where the ANOVA test was significant (p ≤ 0.05), Dunnett’s test was used for pairwise comparisons to a single control; t-tests were otherwise used for pairwise comparisons. Results Test material characterization The test article was produced according to Good Manufacturing Practices (GMPs) relevant to the food industry using fermentation. Briefly, the production process utilized a modified strain of Yarrowia lipolytica to produce mogrosides, primarily siamenoside I. This strain was employed in a bioreactor fermentation process, followed by further processing (filtration, crystallization, and drying) to produce the test article. Lots used in the testing (Batch No. HY240605–01; Batch No. HY240515–01) contained 91.9 and 95.6% total mogrosides, respectively and ≥ 73% siamenoside I based on the available certificates of analysis. The test article was intended to represent a mixture of mogrosides, with siamenoside I as the primary mogroside assessed. Bacterial reverse mutation assay Within Experiment 1, all tester strains were treated in the absence and in the presence of S9, using final concentrations of test article (mogrosides mixture containing siamenoside I) at 5, 16, 50, 160, 500, 1600 and 5000 µg/plate, plus vehicle and positive controls. Following these treatments, no evidence of concentration-related toxicity was observed, as would normally manifest as a thinning of the background bacterial lawn and/or a marked reduction in revertant numbers. A sporadic reduction in revertant colony numbers to 0.5-fold, compared to the concurrent vehicle control value was observed at 500 µg/plate in the presence of S9 in strain TA1537 but there was no evidence of a concentration-related response. Similarly, within the pre-incubation assay (Experiment 2) treatments of all the tester strains were performed in the absence and in the presence of S9. The maximum test concentration of 5000 µg/plate was retained and the same concentrations as Experiment 1 were tested for all strains. Following these treatments, no evidence of concentration-related toxicity was observed in the absence and presence of S9. In addition, the vehicle and positive control treatments included for all strains in both experiments indicated that the mean numbers of revertant colonies were comparable with acceptable ranges for vehicle control treatments and were elevated by positive control treatments. Following test article treatments of all the test strains in the absence and presence of S9 in Experiments 1 and 2, no increases in revertant numbers were observed that were ≥ 2-fold (in strains TA98, TA100 and WP2 uvrA) or ≥ 3-fold (in strains TA1535 and TA1537), compared to the concurrent vehicle controls. The study provided no evidence supporting mutagenic activity elicited by the test article in this assay system. Overall, the results showed that the test article did not induce mutation in four histidine-requiring strains of Salmonella typhimurium (TA98, TA100, TA1535 and TA1537) and one tryptophan-requiring strain of Escherichia coli (WP2 uvrA) when tested under the conditions of this study. These conditions included treatments at concentrations up to 5000 μg/plate (the maximum recommended concentration according to current regulatory test guidelines), in the absence and in the presence of a rat liver metabolic activation system (S9) (see Table 3). Table 3Summary of bacterial reverse mutation assay (Ames assay) results.Concentration (µg/plate)Bacterial Reverse Mutation AssayExperiment 1Test ArticleTA98TA100TA1535TA1537WP2 uvrA-S9+ S9-S9+ S9-S9+ S9-S9+ S9-S9+ S9020.3 ± 2.536.3 ± 4.0153.7 ± 4.7186.3 ± 9.542.7 ± 10.833.0 ± 1.019.7 ± 3.120.7 ± 5.911.3 ± 4.920.3 ± 0.6518.3 ± 7.140.7 ± 7.4194.0 ± 17.8165.7 ± 17.241.3 ± 11.224.0 ± 11.815.7 ± 1.217.7 ± 3.117.0 ± 3.018.7 ± 4.51625.0 ± 3.634.7 ± 7.6153.0 ± 14.7171.3 ± 11.540.3 ± 6.430.0 ± 11.520.0 ± 6.129.3 ± 2.58.3 ± 5.520.0 ± 7.55018.7 ± 2.529.0 ± 5.3170.3 ± 10.0182.7 ± 5.039.7 ± 9.920.7 ± 5.515.3 ± 2.318.3 ± 4.59.3 ± 3.118.0 ± 10.116019.7 ± 4.539.3 ± 5.5170.3 ± 15.3171.0 ± 9.532.3 ± 14.031.7 ± 2.520.0 ± 8.019.3 ± 3.213.3 ± 5.720.0 ± 3.650017.7 ± 2.527.7 ± 6.1144.3 ± 18.5178.3 ± 0.635.7 ± 4.030.7 ± 9.515.0 ± 0.010.3 ± 3.512.0 ± 7.917.7 ± 7.6160020.3 ± 6.126.0 ± 2.0159.3 ± 13.5165.3 ± 27.641.3 ± 4.233.7 ± 6.719.7 ± 3.523.0 ± 6.012.0 ± 2.023.7 ± 5.5500020.3 ± 1.229.0 ± 2.0167.0 ± 14.0153.3 ±¸ 24.431.3 ± 4.535.0 ± 10.119.3 ± 4.217.0 ± 1.711.0 ± 3.018.3 ± 6.5Positive controls*525 ± 48.8270.0 ± 10.5833.7 ± 41.52192.3 ± 186.4823.0 ± 57.4322.7 ± 11.8410.0 ± 46.9445.0 ± 36.4625.3 ± 6.1146.7 ± 23.6Experiment 2Test ArticleTA98TA100TA1535TA1537WP2 uvrA-S9+ S9-S9+ S9-S9+ S9-S9+ S9-S9+ S9020.3 ± 1.243.7 ± 7.8162.3 ± 14.2185.7 ± 12.938.0 ± 5.640.3 ± 1.511.3 ± 2.518.3 ± 3.113.3 ± 5.726.3 ± 4.7516.7 ± 4.244.3 ± 12.7148.3 ± 3.2204.0 ± 10.444.0 ± 10.149.3 ± 3.59.3 ± 6.020.0 ± 5.013.3 ± 2.124.3 ± 6.81615.0 ± 2.652.3 ± 13.3152.7 ± 7.5182.7 ± 4.239.7 ± 6.551.3 ± 13.78.7 ± 2.320.3 ± 2.59.3 ± 2.117.3 ± 9.15026.3 ± 5.041.7 ± 9.0173.3 ± 16.7195.3 ± 6.838.3 ± 4.060.0 ± 11.815.0 ± 1.726.0 ± 7.912.3 ± 4.028.0 ± 11.016025.7 ± 4.243.0 ± 6.6166.7 ± 14.4192.7 ± 23.545.0 ± 7.546.0 ± 7.213.0 ± 2.620.0 ± 2.612.0 ± 4.423.0 ± 4.450021.3 ± 2.538.0 ± 13.0163.3 ± 30.9176.7 ± 11.645.3 ± 10.771.7 ± 41.99.7 ± 1.220.0 ± 1.712.3 ± 4.028.7 ± 5.5160027.0 ± 8.739.0 ± 5.2172.7 ± 10.1196.0 ± 21.055.0 ± 9.544.0 ± 10.610.0 ± 2.013.0 ± 4.414.7 ± 1.223.7 ± 7.8500018.0 ± 2.645.7 ± 1.2175.3 ± 9.8201.0 ± 5.650.7 ± 9.560.3 ± 22.710.0 ± 5.312.7 ± 2.912.7 ± 7.024.0 ± 5.6Positive Controls*913.7 ± 177.5537.0 ± 142.21079.0 ± 37.22070.0 ± 222.41205.0 ± 4.0337.0 ± 16.5257.3 ± 65.1285.7 ± 16.1634.7 ± 34.4136.0 ± 28.7Data are presented as mean ± SD-S9: without metabolic activation; + S9: with metabolic activation*Positive controls without S9 were as follows: TA98 - 2-nitrofluorene (2NF); TA100, TA1535 - sodium azide (NaN3); TA1537 - 9-aminoacridine (ACC); WP2 uvrA - 4-nitroquinoline 1-oxide (NQO); Positive controls with S9 were as follows: TA98 - benzo[a]pyrene (B[a]P); TA100, TA1535, TA1537, WP2 uvrA - 2-aminoanthracene (AAN) Summary of bacterial reverse mutation assay (Ames assay) results. Data are presented as mean ± SD -S9: without metabolic activation; + S9: with metabolic activation Positive controls without S9 were as follows: TA98 - 2-nitrofluorene (2NF); TA100, TA1535 - sodium azide (NaN3); TA1537 - 9-aminoacridine (ACC); WP2 uvrA - 4-nitroquinoline 1-oxide (NQO); Positive controls with S9 were as follows: TA98 - benzo[a]pyrene (B[a]P); TA100, TA1535, TA1537, WP2 uvrA - 2-aminoanthracene (AAN) In vitro human lymphocyte micronucleus assay The results of the initial range-finding assay following administration of the test article at concentrations ranging from 18.14 to 5000 µg/mL under all treatment conditions (3 +21 h with and without metabolic activation and 24 +24 h without metabolic activation) did not elicit clear dose-dependent increases in cytotoxicity as measured by the relative replication index (RI).3 No marked changes in osmolality (≥ 50 mOsm/kg) or pH (≥1 pH unit) were observed in the top three concentrations assessed (1800, 3000, and 5000 µg/mL). In a preliminary assessment of solubility, the test article was shown to be soluble in water at concentrations ≥ 5098 µg/mL based on lack of observed visible precipitates ∼24 h following test article addition at 37°C. Based on results of the range-finding and precipitation assessments, the MNT experiment assessing the 3 + 21-h treatment with and without metabolic activation was conducted using concentrations ranging from 250 to 5000 µg/mL. The top three concentrations (2000, 3000, and 5000 µg/mL) were evaluated for frequency of binucleate cells with micronuclei (MNBN) cells. Statistically significant increases in MNBN cells were not observed. The 24 + 24-h treatment without metabolic activation evaluated concentrations ranging from 125 to 5000 µg/mL. Frequency of MNBN cells at concentrations including 1000, 1750, 3000, and 5000 µg/mL were determined. A statistically significant (p ≤ 0.05) increase in MNBN cells was observed at the 3000 μg/mL dose only but the MNBN cell frequencies of all concentrations analyzed were within the historical vehicle control (normal) ranges for all three treatment conditions. Statistically significant linear trends as assessed using the Cochran-Armitage test were observed for the 3 + 21 h treatment with metabolic activation (p = 0.0483) and 24 + 24 h treatment without metabolic activation (p = 0.0148); however, the mean MNBN cell frequency values were within normal ranges at all concentrations assessed for these treatment conditions. Any statistically significant increases were therefore due to normal variation and these findings were not considered biologically relevant. A summary of results is described in Table 4. Table 4Summary of in vitro micronucleus assay with primary human lymphocytes.TreatmentConcentration (µg/mL)Cytotoxicity (%)bMean MN Frequency (%)Historical Control Range (%)cStatistical Significance3 + 21 h -S9Vehiclea-0.630.10–0.90-200000.45NS300030.70NS500020.30NSMMC, 0.30*252.55p ≤ 0.0013 + 21 h + S9Vehiclea-0.430.10–0.80-200000.40NS300030.80NS500000.65NSCPA, 5.00*533.30p ≤ 0.00124 + 24 h -S9Vehiclea-0.330.10–0.80-100000.30NS1750180.35NS300000.70p ≤ 0.05500010.60NSVIN, 0.04*393.15p ≤ 0.001MN = micronucleated; NS = not significantavehicle was purified waterbbased on replication indexc95th percentile of the observed range*positive control; MMC = Mitomycin C, CPA = Cyclophosphamide, VIN = vinblastine Summary of in vitro micronucleus assay with primary human lymphocytes. MN = micronucleated; NS = not significant vehicle was purified water based on replication index 95th percentile of the observed range positive control; MMC = Mitomycin C, CPA = Cyclophosphamide, VIN = vinblastine 28-Day dietary toxicity study A 28-day non-GLP dose-range finding assay conducted to support identification of doses for the definitive 90-day dietary study, Sprague Dawley (Crl:CD[SD]) rats (n = 5/sex/group) were administered 0, 12,500, 25,000, or 50,000 ppm test article (mogrosides mixture containing siamenoside I in the diet) on a daily basis. No mortality, test article-related clinical observations, changes in body weight or body weight gain, or changes in food consumption were observed in treated animals. No indications of palatability issues were reported. At 50,000 ppm, minimally lower absolute reticulocyte count and minimally lower absolute eosinophil count in males and females, respectively, were reported. No test article-related effects on coagulation parameters or clinical chemistry parameters were observed. At ≥ 12,500 ppm test article, female rats exhibited a test article-related increase in liver weight (absolute, body weight ratio, and brain weight ratio) and reduced thymus weight (brain weight ratio), which were considered adaptive changes. In females administered 12,500 or 50,000 ppm test article, reduced pituitary weight (absolute and brain weight ratio) was noted. Due to the lack of a dose response and low magnitude of change, pituitary weight differences were considered not test article-related and non-adverse. However, based on the mild severity of findings and lack of observed effects on health and well-being of animals administered 50,000 ppm, the no-observed-adverse-effect-level (NOAEL) was considered to be 50,000 ppm (equivalent to a TWA of 4410 mg/kg-bw/day in males and 4112 mg/kg-bw/day in females). 90-Day dietary toxicity study Survival and in-life observations Groups of Sprague Dawley (Crl:CD[SD]) rats (n = 10/sex/group) were administered 0, 12,500, 25,000, or 50,000 ppm test article (mogrosides mixture containing siamenoside I in the diet) on a daily basis. No mortality or test article-related clinical signs were reported throughout the 90-day dosing period. Altered body weights or body weight gain were observed in rats administered up to 50,000 ppm mogrosides. Males administered 12,500 ppm exhibited lower body weights compared to control on Day 88 and Day 90, but findings were considered non-adverse in the absence of a dose-response effect and a lower group mean body weight at group assignment along with a lack of observed lower body weight in females administered 12,500 ppm (Table 5). Reduced body weight gain in all dose groups was anticipated and found non-adverse considering that a percentage of the diet was replaced with a non-caloric ingredient (Fig. 2; Fig. 3). No significant test article-related changes in food consumption were reported (data not shown). Table 5Body weight change over 90-day study in males and females.Dose Level Mogrosides (ppm)Body Weight Change (Day 1–90)MalesFemales0384 ± 52.0112 ± 19.712,500301 ± 58.2*98 ± 9.425,000331 ± 42.4102 ± 8.150,000320 ± 74.3*87 ± 9.5#*p ≤ 0.05 based on ANOVA and Dunnett’s test#p ≤ 0.05 based on ANOVA and Dunnett’s test of rank-transformed data Body weight change over 90-day study in males and females. p ≤ 0.05 based on ANOVA and Dunnett’s test p ≤ 0.05 based on ANOVA and Dunnett’s test of rank-transformed data Weekly mean body weights of male rats administered test article. Weekly mean body weights of female rats administered test article. Clinical laboratory evaluations Hematology and coagulation No test article-related effects on hematology or coagulation test results were reported. Minor differences in test results between control and dosed groups were considered incidental, consistent with normal variation, and were characterized by most or all of the following criteria including small magnitude of change, lack of dose response, inconsistency between sexes, and/or absence of correlative findings (data not shown). Clinical chemistry and thyroid hormone assessments Significant test article-related effects on serum chemistry parameters consisted of dose-dependent increases in total protein in mid- and high-dose females only, as well as increases in globulin and albumin:globulin ratio levels in all dosed females. A significant reduction in triglyceride levels was reported in both low- and high-dose males and all dosed females receiving test article. Additional changes in clinical chemistry parameters (e.g., decreased AST in all dosed females, non-dose dependent changes in ALT in females and SBA in males) were generally considered incidental, consistent with normal variation, and not test article-related. No test article-related effects were observed in thyroid hormone (i.e., TSH, T3, T4) test results (Table S1). Urinalysis No test article-related effects on urinalysis were reported. Minor differences in test results between control and dosed groups were considered incidental, consistent with normal variation, and were characterized by most or all of the following criteria including small magnitude of change, lack of dose response, inconsistency between sexes, and/or absence of correlative findings (data not shown). Ophthalmic findings No remarkable ophthalmic observations were noted following examination once during the pre-dose phase and once for control and high-dose rodents during Week 12 of dosing (data not shown). FOB findings No test-article related effects were noted for approach response, auditory startle response, pinna response, pain response, or forelimb or hindlimb grip strength measurements (data not shown). Locomotor activity No test-article related effects were reported for basic movements, fine movements, ambulation, rearing, or distance traveled (data not shown). Estrous cycle stage No test article-related effects on the estrous cycle were observed at any dose level compared to controls (data not shown). Male reproductive parameters No test article-related effects on sperm motility or sperm density were observed at any dose level compared to controls. Statistically significant decreases in causal epididymal weights in mid- and high-dose males did not result in changes in sperm density (Table S2). Histopathologic findings Organ weights Select organ weights (based on OECD 408 TG) were assessed in control and dosed animals.4 Absolute and relative (body weight) liver weights were increased in a dose-dependent manner in females at all dose levels; brain-to-kidney weight ratios were significantly increased in mid- and high-dose females. Relative liver weights (body weight only) were also increased in mid- and high-dose males. While kidney weights appeared to be unaffected by administration of mogrosides in males, absolute and relative kidney weights were significantly increased in high-dose females. Relative kidney weights (body weight only) were also significantly increased in low- and mid-dose females (Table 6). As noted in the following section, liver and kidney congestion (accumulation of blood cells) was observed in some control and treated animals, which was attributed to incomplete exsanguination, and which may explain the observed organ weight changes. All other organ weight parameter differences were consistent with normal variation and considered incidental (data not shown). Table 6Summary of potential mogrosides-related effects in organ weight parameters.OrganMales (n = 10/group)Females (n = 10/group)0 ppm12,500 ppm25,000 ppm50,000 ppm0 ppm12,500 ppm25,000 ppm50,000 ppmTerminal body weight (g)581 ± 56.0491 ± 54.6*(85)528 ± 42.3(91)535 ± 91.5(92)272 ± 19.9256 ± 17.5(94)271 ± 9.0(110)261 ± 17.8(113)LiverAbsolute weight (g)15.1835 ±1.9750713.4293 ±1.35384(88)15.0286 ±1.63191(99)15.4228 ±2.90684(102)7.4167 ±0.611808.9641 ±0.97380*(121)10.0282 ±1.00743*(135)10.5324 ±0.65131*(142)Body Weight Ratio (%)2.6104 ±0.143292.7418 ±0.15512(105)2.8438 ±0.17440*(109)2.8795 ±0.24612*(110)2.7260 ±0.115403.5131 ±0.36128*(129)3.7034 ±0.36254*(136)4.0405 ±0.25366*(148)Brain Weight Ratio (%)676.7539 ±104.71871606.1904 ±65.48340(90)667.1258 ±82.12723(99)679.0328 ±123.25459(100)387.7803 ±46.36092444.6890 ±46.31306(115)519.0893 ±70.03237*(134)543.1974 ±55.12419*(140)KidneyAbsolute weight (g)3.2428 ±0.387922.9675 ±0.25648(92)3.0569 ±0.40110(94)3.3124 ±0.58543(102)1.6001 ±0.132561.6638 ±0.13960(104)1.7523 ±0.12312(110)1.8149 ±0.17657*(113)Body Weight Ratio (%)0.5588 ±0.044600.6076 ±0.04637 (109)0.5791 ±0.06057(104)0.6200 ±0.05533 (111)0.5884 ±0.034140.6518 ±0.04547*(111)0.6522 ±0.03668*(111)0.7004 ±0.04168*(119)Brain Weight Ratio (%)144.3719 ±19.73315133.8155 ±10.88723(93)136.0790 ±23.26257(94)145.7184 ±23.93208(101)83.5007 ±8.2025082.5247 ±6.52939(99)90.6211 ±9.19016(109)93.5904 ±11.80674*(112)*statistically significant (p ≤ 0.05) difference (absolute or relative), compared with respective control mean value; values for absolute weight and ratio of organ weights (relative to body or brain) for dosed groups expressed as percentage/magnitude of change compared to control mean. Summary of potential mogrosides-related effects in organ weight parameters. statistically significant (p ≤ 0.05) difference (absolute or relative), compared with respective control mean value; values for absolute weight and ratio of organ weights (relative to body or brain) for dosed groups expressed as percentage/magnitude of change compared to control mean. Macroscopic observations No macroscopic findings related to mogrosides were reported in dosed animals. Microscopic observations Incidence of liver and kidney congestion was reported in male and female rats (n = 10/sex/group). In females, liver congestion was observed in all high-dose animals with minimal, slight, or moderate severity, and four control females. Congestion was considered secondary to incomplete exsanguination and also occurred in male rats (zero control males; one mid-dose male; five high-dose males). Similarly, kidney congestion was reported with minimal to moderate severity in female rats (zero control females; seven high-dose females) as well as male rats in a dose-dependent manner (one control male; one low-dose male, three mid-dose males, and three high-dose males). Of animals exhibiting congestion, one mid-dose male, three high-dose males, and seven high-dose females exhibited both liver and kidney congestion (Table 7). Table 7Incidence of liver and kidney congestion.CongestionMales (n = 10/group)Females (n = 10/group)0 ppm12,500 ppm25,000 ppm50,000 ppm0 ppm12,500 ppm25,000 ppm50,000 ppmLiverMinimal00142002Slight00012004Moderate00000004Total001540010KidneyMinimal10330002Slight01000004Moderate00000001Total11330007 Incidence of liver and kidney congestion. Additional microscopic observations were considered spontaneous and/or incidental based on random distribution, incidence, and/or observed severity (data not shown). Discussion Herein, a bacterial reverse mutation assay, an in vitro micronucleus assay, and a 90-day repeat-dose oral toxicity study were conducted to determine the safety profile of a mogrosides mixture containing > 91% total mogrosides, of which ≥ 73% was identified as siamenoside I (CAS No. 126105–12–2). Mogrosides, which are cucurbitane glycoside derivatives of monk fruit, have been used in traditional Chinese medicine and more recently, as natural sweeteners. Mogrosides are present at low levels in monk fruit (0.55 – 0.65%), such that substantial processes to extract and concentrate mogrosides had to be employed to produce monk fruit extracts for consumer use. Considering the resource-intensive nature of processing monk fruit to obtain monk fruit extracts, alternative methods have been developed to produce mogroside-based sweeteners, including bioreactor fermentation. The fermentation derived siamenoside I ingredient was produced using a modified strain of Yarrowia lipolytica to produce the test-article mixture for the toxicological safety assessment described in the present study. Recently, Roberts et al. [8] evaluated the pharmacokinetics and metabolism of a single-dose administration of radio-labeled purified siamenoside I (95.4%) in rats. Results indicated that siamenoside I was well-absorbed (43% in males; 42% in females) and rapidly excreted via bile into the feces (101% in males; 92% in females). Very low levels of radiolabel (approx. 1%) were reported to be available systemically in both males and females with the absorbed test article (identified by study authors to primarily consist of mogrol) found to undergo extensive first-pass metabolism, with oxidized metabolites observed in the bile and subsequently eliminated via the feces. While Tmax, and T1/2 values were similar between sexes, notable differences in plasma Cmax (65 ng/g in males; 352 ng/g in females) and AUC∞ (1500 h*ng/g in males; 5240 h*ng/g in females) were reported. However, the results reported that the overall absorption and excretion were similar between sexes with no evidence of accumulation or retention observed in any tissue over the 168 h study. These findings from the pharmacokinetic study therefore suggest that the mogroside mixture containing siamenoside I administered in the in vivo toxicological studies presented in this report is also likely to result in rapid clearance and metabolism to mogrol and the oxidized metabolites which would be primarily excreted in the feces with a lack of accumulation. The mogrosides mixture assessed did not elicit genotoxic effects at concentrations up to 5000 µg/plate, the highest dose tested, in an in vitro bacterial reverse mutagenicity assay and up to 5000 µg/mL, the highest dose tested, in an in vitro micronucleus assay. While statistically significant linear trends were observed in some treatment groups in the in vitro micronucleus assay, the MNBN cell frequencies of all concentrations analyzed were within the historical vehicle control (normal) ranges for all three treatment conditions and the findings were therefore not considered biologically relevant in the context of historical control data. In the 28-day study, no test article-related clinical observations, body weight changes, or changes in food consumption were observed in rats administered up to 50,000 ppm test article. At the highest dose tested, no test article-related effects on coagulation or clinical chemistry parameters were observed; minimal changes in hematology parameters were consistent with normal variation and considered incidental. At ≥ 12,500 ppm test article, female rats exhibited a test article-related increase in liver weight and reduced thymus weight, which was considered an adaptive response. Based on the mild severity of findings and lack of observed effects on health and well-being of animals administered the high dose, the NOAEL was considered to be 50,000 ppm (equivalent to a TWA of 4410 mg/kg-bw/day in males and 4112 mg/kg-bw/day in females). In the 90-day study, decreased body weight and body weight gain across dose groups were determined to be non-adverse, consistent with replacement of up to 5% of the diet with a non-caloric component. No test article-related effects on hematology or coagulation test results, thyroid hormone levels (i.e., TSH, T3, T4), urinalysis parameters, ophthalmic findings, FOB outcomes, locomotor activity, estrous cycle, sperm motility, or sperm density were reported. Significant test article-related effects on serum chemistry parameters included dose-dependent increases in total protein in mid- and high-dose females only and increases in globulin and albumin:globulin ratio levels in all dosed females. Significantly reduced triglyceride levels were reported in both low- and high-dose males and all dosed females receiving test article. The changes noted in globulin levels are consistent with previously observed adaptive liver changes for other monk fruit extracts [2], [9]. Additionally, as expected with any non-sugar sweetener, there were no adverse changes in blood glucose or urinary biomarkers. Overall, the minor differences in test results reported between the control and dosed groups were considered incidental, consistent with normal variation, and characterized by the changes being small in magnitude with a lack of dose response, inconsistency between sexes, and/or absence of correlative findings. While increases in absolute and/or relative liver or kidney weights were also observed across dose groups, no corresponding macroscopic or histopathological were identified in treated animals. Notably congestion of the liver and kidneys associated with exsanguination and accompanying accumulation of blood cells was observed, which may have contributed at least in part to the observations of elevated organ weights [10]. Although no hepatocellular hypertrophy or other microscopic correlates were identified, the pattern of liver weight change, in the absence of adverse clinical pathology or histopathology findings, is consistent with a non-adverse adaptive response. Overall, in the absence of any corroborating macroscopic or microscopic findings, the observed changes in organ weights were considered non-adverse. This conclusion is also aligned with observed adaptive liver changes observed in 28-day, 13-week, and 90-day studies assessing monk fruit extracts (containing up to 52% mogroside V), although data relevant to other high-purity, fermentation process-derived mogroside ingredients containing high levels siamenoside I were not identified in the literature [2], [9]. Furthermore, based on the mild severity of the findings and overall health and well-being of the animals, 50,000 ppm was considered the NOAEL (equivalent to a TWA of 3309 mg/kg-bw/day in males and 3362 mg/kg-bw/day in females). Therefore, in summary, the 28-day and the 90-day rodent toxicological studies conducted per OECD guidelines supports a NOAEL of 3309 mg/kg-bw/day (i.e., the highest dose tested) for the high purity siamenoside I mogroside ingredient in male and female rats. Taking into account the standard 100-fold margin of safety (MOS) factor, this NOAEL supports an estimated daily intake (EDI) of up to ∼33 mg/kg-bw/day and is likewise supported by the lack of concern regarding mutagenicity/genotoxicity, in line with the current food additive guidelines laid down in recent EFSA guidance [11]. CRediT authorship contribution statement Mel Lloyd: Resources, Investigation, Data curation. Jonathan LaCombe: Resources, Investigation, Data curation. Alex Eapen: Writing – review & editing, Resources, Project administration, Data curation. Candace Doepker: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Ashley Roberts: Writing – review & editing. Amy L. Mihalchik: Writing – review & editing, Writing – original draft, Visualization, Formal analysis. Funding This work was funded by Cargill Incorporated; Author Alex Eapen was an employee of Cargill during the conduct and drafting of this study. Declaration of Competing Interest ToxStrategies, LLC, A BlueRidge LifeSciences company, received consulting fees from Cargill Inc. for conducting and reporting this work. ToxStrategies is a consulting firm that provides services to private and public organizations on toxicology and risk assessment issues. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The authors thank Greg Brorby, Grace Patlewicz, Alexander East, Neepa Choksi, Jesse Lueth, Anna Bottomley, and Jon Urban for contributing subject matter expertise and conducting technical review of the manuscript. Handling Editor:Prof. L.H. Lash Supplementary data associated with this article can be found in the online version at doi:10.1016/j.toxrep.2026.102297. GRAS Notice numbers for monk fruit juice ingredients are as follows: GRN 201, 359, 522, 556, 627, 629, 706, and 784. In the 24 + 24 h treatment range-finding experiment, cytotoxicity was reported as follows for 233.3, 388.8, 648.0, 1080, 1800, 3000, and 5000 µg/mL: 26%, 7%, 31%, 0%, 43%, 92%, and 0%, respectively. Notably, cytotoxicity was not observed at the highest concentration assessed in this experiment conducted in single cultures. In the main experiment (duplicate cultures), additional concentrations between 1000 and 5000 µg/mL were included for the 24 + 24 h treatment condition and no appreciable cytotoxicity was reported (see Table 3), indicating that the cytotoxicity observed at 1800 µg/mL and 3000 µg/mL in the range-finding experiment was non-reproducible. Weighed organs included adrenal gland (2), brain, cervix, epididymis (2), heart, kidney (2), liver, ovary (2), oviduct (2), pituitary gland, prostate (prostate, seminal vesicles, and coagulating gland were weighed together, followed by dissection, and then the prostate gland was weighed separately), spleen, testis (2), thymus, thyroid (two lobes) with parathyroid, and uterus. Supplementary material Supplementary material Data availability Data will be made available on request. References Associated Data Supplementary Materials Supplementary material Data Availability Statement Data will be made available on request.