A 90 DAY ORAL TOXICITY STUDY OF BLUEBERRY POLYPHENOLS IN OVARIECTOMIZED SPRAGUE-DAWLEY RATS

Abstract

Regular consumption of polyphenol-rich fruits and vegetables is associated with beneficial health outcomes. To increase polyphenol intakes, consumers are increasingly using herbal and botanical dietary supplements containing concentrated polyphenol extracts. However, the safety of this consumption modality has not been vetted. To address this, ovariectomized Sprague-Dawley (OVX-SD) rats were orally gavaged with purified blueberry polyphenols at 0–1000 mg total polyphenols/kg bw/d for 90d. No differences in behavior, body weight, or food consumption were observed. No tumors or macroscopic changes were observed, and histopathological analyses showed no differences among groups. Although several statistically significant differences between treatment and control groups were observed in urine (color and pH) and blood (monocyte count, total cholesterol, and chloride ion concentration) analyses, these parameters were within normal ranges and not considered biologically significant. Intestinal permeability assessed via FITC-dextran showed increased intestinal permeability in the highest dose, though no morphological differences were found throughout the gastrointestinal tract. Given the lack of other systemic changes, this finding is likely of minimal physiological importance. These results indicate a NOAEL for blueberry polyphenols in OVX-SD rats is ≥ 1000 mg total polyphenols/kg bw/d, which translates to a 70 kg human consuming ~10 g polyphenols.

1. INTRODUCTION

Regular fruit and vegetable consumption is part of a healthy diet to promote adequate intakes of essential nutrients and to increase exposure to health-promoting polyphenols. When consumed regularly, polyphenols have been linked to improvements in numerous health endpoints, including cardiovascular and neurocognitive health, while also lowering the potential for developing chronic diseases (Spencer 2010; Boeing et al. 2012; Del Rio et al. 2013; Oyebode et al. 2014).

Blueberries are a rich source of polyphenols, containing large amounts of anthocyanins, chlorogenic acid, and quercetin (Figure 1) (Yousef et al. 2013). Both blueberries and their constituent polyphenols have been linked to the health benefits described above (Ma et al. 2018). In addition, the popularity of blueberries in the US has grown dramatically in the past few decades and growth is expected to continue (USDA 2016).

Figure 1 –

Structure of common polyphenols in blueberries. Anthocyanins account for approximately half of all polyphenols in blueberries and can take different forms, depending on substitutions at the R1 and R2 positions. In planta, anthocyanins and quercetins are generally found in their glycosylated form, with linkages to glucose, galactose, and arabinose moieties (indicated by “gly”) being most commonly found in nature.

In light of these benefits, many consumers have sought to increase their polyphenol consumption, often turning to dietary supplements to meet this need. Recent estimates show that 75% of US adults consume at least one dietary supplement and 30% consume polyphenol-rich herbal and botanical supplements (Bailey et al. 2011; Gahche et al. 2011; Dickinson et al. 2014; Council for Responsible Nutrition 2017). The tacit assumption amongst consumers is that because these supplements are derived from natural sources, they must be safe to consume. However, this hypothesis has not been adequately vetted, and the scientific literature lacks sufficient evidence to validate or nullify this assumption of safety (Anadón et al. 2016; Burton-Freeman et al. 2016). And, given clinical cases of hepatotoxicity resulting from consuming dietary supplements containing high levels of polyphenol-rich green tea extracts (Mazzanti et al. 2009), the presumption of safety for this consumption modality cannot be assumed. In light of this, many regulatory and scientific authorities (e.g., Food and Drug Administration (FDA), National Toxicology Program, and National Cancer Institute), are increasing efforts to test herbal and botanical supplement safety (Zhang et al. 2016).

Most clinical studies regarding the health benefits of polyphenols report that they are well-tolerated at normal to even higher dietary doses, though there are several reports of adverse events and side effects occurring after repeatedly consuming higher doses more in line with supplemental paradigms. These adverse events are dose-dependent and most often associated with some form of gastrointestinal distress (e.g., nausea, abdominal pain, diarrhea) (Brown et al. 2010; Heinonen and Gaus 2015; Burton-Freeman et al. 2016; Coppock and Dziwenka 2016). These effects are often more pronounced in animal models (Heinonen and Gaus 2015) and extend to other organ systems, with indications of altered function in the liver, kidneys, and reproductive systems after oral ingestion of high doses of green tea catechins, quercetin, and soy isoflavones (Coppock and Dziwenka 2016; Esch et al. 2016). However, few of these investigations exist in the literature, despite a call for more studies regarding the safety of polyphenols at high doses (Mennen et al. 2005).

Women ages 51–70y are the most frequent consumers of herbal and botanical dietary supplements (Bailey et al. 2011), but there remains a paucity of available evidence surrounding the safety of this consumption modality in this population. To address this, we tested the safety of increasing doses of blueberry polyphenols in an ovariectomized (OVX) rat model. The OVX rat is frequently used to mimic hormonal changes occurring in postmenopausal women (Thompson et al. 1995), providing direct relevance to the population most likely to consume polyphenol-rich dietary supplements. Additionally, as the OVX model has been used in prior safety studies (Wyde et al. 2000; Wu et al. 2017), we used it to assess the safety of blueberry polyphenols in a sub-chronic toxicity study modeled after OECD 408 guidelines.

2. MATERIALS AND METHODS

2.1. Chemicals/Materials and vendors

Gallic acid, sodium carbonate, fluorescein isothiocyanate-dextran (FITC), and Folin and Ciocalteu’s reagent (2N) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Extraction and LC-MS grade solvents, including methanol, water, acetonitrile, and formic acid, as well as 10% neutral buffered formalin and 0.9% aqueous sodium chloride solution were purchased from Thermo Fisher Scientific (Waltham, MA, USA). VitaBlue Pure American Blueberry Extract was obtained from FutureCeuticals (Momence, IL, USA); freeze-dried blueberries were obtained from the Wild Blueberry Association of North America (WBANA, Old Towne, ME, USA).

2.2. Animal protocols

2.2.1. Study overview

Fifty-four 5-month old, virgin, ovariectomized (OVX), female Sprague Dawley rats were used in this study. Upon arrival, rats were randomized to treatment groups and given one month to stabilize from OVX surgery. Animals then underwent 90d of dosing via oral gavage, modeled after OECD guidelines for sub-chronic toxicity studies (OECD 2018), and were monitored daily. A total of 5 treatment groups plus one non-gavaged group were used in the study. Animals were monitored daily for signs of overt toxicity; body weight and food consumption were monitored weekly throughout the study. Upon completion of the 90d dosing regimen, gut permeability was assessed using FITC, and animals were euthanized via CO₂ asphyxiation and necropsied.

2.2.2. Animal care.

Animal experiments were conducted in adherence to Purdue University Animal Care and Use Committee guidelines, following an approved protocol (1808001790). Rats were purchased from Envigo (Indianapolis, IN, USA) and individually housed in stainless steel, wire-bottom cages in a temperature and humidity-controlled room with a 12h light/dark cycle and ad libitum access to food and water.

2.2.3. Rat chow diet.

All animals were maintained on a polyphenol-free diet throughout the stabilization and study periods. Polyphenol-free diets were based on the AIN-93M diet, using corn oil in place of soybean oil to remove isoflavones and prevent confounding by polyphenols unrelated to blueberry treatment. Diets were prepared by Research Diets (New Brunswick, NJ, USA).

2.2.4. Treatment groups.

A total of 5 treatment groups were used for this study. Because there are no data upon which to reliably base power calculations, OECD 408 guidelines recommend using 10 animals per group, including both male and female animals (OECD 2018). Because women 51–70y are the most frequent consumers of polyphenol-rich dietary supplements (Bailey et al. 2011), we elected to only use female animals in this study. Although this study was modeled after OECD 408 guidelines, it did not complete all tests associated with the guidelines and may not be considered a regulatory study.

Animals were randomized to treatment groups upon arrival and maintained in these groups throughout the study. The first four treatment groups received an oral gavage of purified blueberry polyphenols (see below for details), containing 0, 50, 250, or 1000 mg total polyphenols/kg bw/d (designated as “water”, “low”, “medium”, and “high”, respectively). The low dose corresponds to an adult human consuming approximately 1–2 cups of fresh blueberries per day (i.e., a “dietary dose”, calculated using the FDA’s rat to human conversion factor (FDA 2005)). The medium and high dose groups were 5- and 20-fold higher, respectively, to mimic higher concentrations as may be present in dietary supplements. The fifth group was dosematched to the low-dose group but was dosed with lyophilized, whole blueberries (designated as “BB”) rather than the purified extract to monitor potential differences between whole foods vs. isolated polyphenolics.

2.2.5. Non-gavage control group.

In addition to the 5 treatment groups, a smaller group of animals (n=4) was maintained throughout the study and subjected to all study procedures with the exception of oral gavage. These animals were monitored throughout the experiment to detect potential adverse effects of the daily oral gavage distinct from polyphenol treatment-related effects.

2.2.6. Oral gavage dose preparation and administration.

Concentrated blueberry phenolic extracts, containing 26.7% total phenolics (w/w), were used to prepare oral gavage slurries for the water, low, medium, and high dose groups, respectively, while lyophilized blueberries, containing 3.52% total phenolics (w/w), were used to prepare an oral gavage slurry for the BB group. Administering doses via oral gavage is analogous to how dietary supplements are consumed (i.e., daily ingestion of a concentrated, purified dose at a single time) and maximizes the translatability of our results to dietary supplement consumers. All gavage doses were administered shortly after “lights on” each day (approximately 0800–1000), with the gavage needle dipped in a sugar solution (0.5 g/mL) to minimize stress and induce swallowing (Hoggatt et al. 2010).

2.3. Polyphenol analysis.

Lyophilized whole blueberries, purified blueberry extract, animal diets, and gavage doses were extracted and analyzed in triplicate, as described elsewhere (Furrer et al. 2017). Samples were resolubilized with 2% formic acid in water and purified via solid-phase extraction (SPE) using Oasis HLB 1cc extraction cartridges (Waters, Milford, MA, USA) (Furrer et al. 2017). All samples were analyzed for total and individual phenolics via the Folin method and UPLCMS/MS, respectively, as described previously (Song et al. 2013; Yousef et al. 2013; Yousef et al. 2014). The polyphenolic composition of the starting materials is shown in Table 1.

Table 1 –

Polyphenol content of blueberries and extract used in doses.^a

(Poly)phenol	Raw Materials (mg/100 g)				Gavage Doses (mg/kg bw)
	FD	CE	Water	Low	Medium	High	BB
Total Phenolics b	3517 ± 54.8	26715 ± 384	nd	50.0 ± 3.77	253 ± 6.07	1000 ± 18.4	50.8 ± 7.47
Anthocyanins
Cyanidins
Arabinoside	21.4 ± 0.30	93.8 ± 14.9	nd	0.11 ± 0.01	0.54 ± 0.04	2.27 ± 0.23	0.23 ± 0.07
Galactoside + Glucoside	46.3 ± 0.50	126 ± 12.2	nd	0.44 ± 0.04	2.09 ± 0.17	8.98 ± 0.82	1.00 ± 0.30
Delphinidins
Arabinoside	20.8 ± 0.37	301 ± 50.4	nd	0.35 ± 0.05	1.80 ± 0.17	7.68 ± 0.84	0.15 ± 0.06
Galactoside + Glucoside	100 ± 2.02	960 ± 281	nd	0.81 ± 0.15	3.62 ± 0.71	11.6 ± 2.67	0.54 ± 0.28
Malvidins
Arabinoside	12.6 ± 0.30	209 ± 31.5	nd	0.32 ± 0.03	1.59 ± 0.14	6.67 ± 0.57	0.16 ± 0.04
Galactoside	23.9 ± 1.48	529 ± 66.8	nd	0.78 ± 0.09	3.62 ± 0.49	12.9 ± 2.21	0.15 ± 0.08
Glucoside	34.5 ± 0.69	376 ± 50.0	nd	0.45 ± 0.04	2.17 ± 0.28	7.20 ± 1.27	0.22 ± 0.11
Peonidins
Arabinoside	13.9 ± 0.10	56.6 ± 9.61	nd	0.08 ± 0.01	0.39 ± 0.03	1.66 ± 0.15	0.18 ± 0.05
Galactoside	40.2 ± 0.34	227 ± 22.8	nd	0.35 ± 0.04	1.69 ± 0.14	7.18 ± 0.69	0.58 ± 0.17
Glucoside	75.0 ± 1.40	249 ± 28.5	nd	0.40 ± 0.02	1.85 ± 0.15	8.06 ± 0.55	1.02 ± 0.29
Petunidins
Arabinoside	28.4 ± 0.30	486 ± 71.0	nd	0.72 ± 0.07	3.47 ± 0.27	14.8 ± 1.30	0.35 ± 0.12
Galactoside	59.0 ± 1.42	1121 ± 108	nd	1.86 ± 0.16	8.60 ± 0.57	37.3 ± 3.49	0.68 ± 0.22
Glucoside	74.5 ± 0.86	598 ± 62.8	nd	0.90 ± 0.09	4.28 ± 0.32	18.7 ± 1.79	0.86 ± 0.27
Phenolic Acids
Benzoic Acids
Gallic acid	nd	18.6 ± 2.99	nd	0.029 ± 0.006	0.168 ± 2.00	0.62 ± 0.11	nd
Protocatechuic acid	0.57 ± 0.12	8.82 ± 1.90	nd	0.017 ± 0.003	0.070 ± 0.020	0.35 ± 0.10	0.013 ± 0.004
Cinnamic Acids
Caffeic acid	trace	47.7 ± 11.7	nd	0.064 ± 0.006	36.7 ± 0.035	1.18 ± 0.06	0.007 ± 0.004
Chlorogenic acid	621 ± 21.7	1738 ± 322	nd	2.64 ± 0.17	13.0 ± 1.30	48.9 ± 3.85	4.86 ± 0.68
Ferulic acid	2.01 ± 0.22	43.7 ± 7.17	nd	0.088 ± 0.005	0.41 ± 0.032	16.0 ± 0.18	0.044 ± 0.008
Feruloylquinic acid	8.28 ± 0.45	33.4 ± 5.33	nd	0.074 ± 0.007	0.35 ± 0.040	14.3 ± 0.13	0.143 ± 0.014
Flavan-3-ols
Catechin	10.7 ± 0.17	39.9 ± 12.3	nd	0.092 ± 0.020	0.37 ± 0.15	13.9 ± 0.22	0.089 ± 0.030
Epicatechin	7.78 ± 0.44	11.3 ± 4.04	nd	0.023 ± 0.008	0.10 ± 0.042	0.39 ± 0.05	0.031 ± 0.013
Epigallocatechin	nd	20.9 ± 2.79	nd	0.048 ± 0.004	0.25 ± 0.012	0.78 ± 0.07	nd
Flavonols
Myricetin	1.78 ± 0.15	46.3 ± 12.0	nd	0.062 ± 0.014	0.32 ± 0.046	1.09 ± 0.30	0.031 ± 0.001
Kaempferol	nd	10.8 ± 1.12	nd	0.028 ± 0.001	0.14 ± 0.001	0.69 ± 0.01	nd
Galactoside + Glucoside	96.5 ± 5.26	360 ± 52.8	nd	0.49 ± 0.064	2.37 ± 0.29	9.16 ± 0.99	1.05 ± 0.40
Quercetin	1.80 ± 0.18	242 ± 48.6	nd	0.32 ± 0.050	1.60 ± 0.14	5.02 ± 1.45	0.041 ± 0.009
Galactoside + Glucoside	309 ± 9.78	1165 ± 117	nd	19.3 ± 0.12	9.25 ± 0.099	35.4 ± 2.61	4.45 ± 1.18
Rutin	32.6 ± 0.93	74.6 ± 16.3	nd	0.12 ± 0.011	0.56 ± 0.064	2.51 ± 0.21	0.431 ± 0.068

FD = freeze dried whole blueberries; CE = concentrated blueberry phenolic extract; nd = not detected; trace = compound detected, but below LOQ.

^aData are comprised of three analytical replicates and presented as mean ± SD.

^bMeasured via Folin assay.

To ensure the homogeneity of the test article and the consistency of the polyphenol content of the doses, each dose was analyzed every 10d throughout the study (data not shown). Because a single batch of purified blueberry polyphenols was homogenized prior to the study and used to prepare all doses, minimal differences were observed throughout the study. Blueberry polyphenol powder was kept frozen and dry until immediately prior to gavage administration, at which point, the powder was mixed with water to form a slurry. After administration, remaining gavage slurries were randomly tested to ensure test article stability and consistency throughout the experiment.

2.4. Gut Permeability.

On the last day of oral gavage, gut permeability was assessed using the FITC (Chassaing et al. 2015; Erkens et al. 2018). After receiving their prescribed dose of phenolics, animals were fasted for 4h before receiving FITC via oral gavage (50 mg/100 g bw). After an additional 4h fast, blood was drawn from the jugular vein and coagulated at room temperature for 30 min. Serum was obtained via centrifugation at 2200 g for 90s. The FITC-dextran concentration of serum was determined at 490, 520 nm with BioTek™ Cytation™ 1 Cell Imaging Multi-Mode Reader (Thermo Fisher Scientific Inc., Winooski, VT, USA) to evaluate the intestinal permeability.

2.5. Necropsy and terminal measures.

2.5.1. Urinalysis.

During the last week of the study, urine was collected for 24h using metabolic cages and submitted to the Clinical Pathology lab at Purdue University for analysis. The following parameters were measured: total volume, color, specific gravity, pH, protein, glucose, ketones, bilirubin, blood, urobilirubin, epithelial cells, and triphosphate crystals.

2.5.2. Serum biochemistry and hematology.

Immediately after euthanasia, blood was collected from the abdominal aorta and submitted to the Clinical Pathology lab at Purdue University for analysis. Samples for hematology were collected in tubes containing EDTA as an anticoagulant, centrifuged to separate plasma, and analyzed for the following: total protein (plasma), red blood cell count, hematocrit, hemoglobin concentration, mean corpuscular volume, mean corpuscular hemoglobin concentration, red blood cell distribution width, white blood cell count, segmented neutrophils, lymphocyte count, lymphocyte morphology, monocytes, eosinophils, platelet count (or estimate if too clumped to count), mean platelet volume, anisocytosis, poikilocytosis, polychromasia, target cells, and reticulocyte number. Samples for serum biochemical analyses were collected in tubes without any anticoagulant and allowed to clot for at least 30 minutes prior to separating serum via centrifugation. Serum was then analyzed for the following: glucose, blood urea nitrogen, creatinine, phosphorus, calcium, sodium, potassium, chloride, carbon dioxide, anion gap, total protein (serum), albumin, globulin, albumin:globulin ratio, alanine aminotransferase, alkaline phosphatase, gamma-glutamyl transferase, total bilirubin, cholesterol, amylase, and lipase.

2.5.3. Necropsy.

After drawing blood from the abdominal aorta, animals were flushed with 200 mL saline. Major tissues, including liver, kidney, pancreas, spleen, brain, stomach, small intestine, cecum, large intestine, lungs, heart, femur, tibia, and vertebrae were harvested, patted dry, and weighed. Intestinal contents were removed and the tissues flushed with 3×10 mL saline. To obtain representative samples of the major sections of the intestines, 1 cm segments of tissue were taken from the following locations: 1 cm distal to stomach (duodenum), midpoint of small intestine (jejunum), 1 cm proximal to cecum (ileum), and midpoint of large intestine (colon) (Ruehl-Fehlert et al. 2003). Finally, ovariectomy was checked by visual inspection.

2.5.4. Histopathology.

Harvested tissues were fixed in 10% neutral buffered formalin for ≥1 week. Tissues were embedded in paraffin and sectioned into 4 μm thick slices prior to staining with hematoxylin and eosin. All slides were prepared by the Histology Research Laboratory at Purdue University. Slides of kidney, heart, liver, spleen, lung, and pancreas were analyzed for the water and high dose groups only, while slides of stomach, duodenum, jejunum, ileum, cecum, and colon were evaluated for all dose groups. Histopathology of GI tissues was scored on a histomorphological scale from 0–3 (normal to severe abnormality) for each of the following: goblet cell numbers, mucosal hyperplasia, crypt cell death, erosion, mononuclear infiltrate, polymorphonuclear leukocyte infiltrate, crypt architectural distortion, and involvement of submucosa. Slides were read and interpreted by a board-certified veterinary pathologist.

2.5.5. Bone mineral density.

Bones from euthanized animals, including right femur, right tibia, and L1-L4 vertebrae were removed and manually cleaned before analysis of bone mineral density using a PIXImus 2 mouse densitometer (GE Lunar PIXImus).

2.6. Statistics

Statistics were completed using SAS (SAS Institute, Raleigh, NC). When data were not normal, appropriate transformations were performed before analysis to ensure normality. Differences between dose groups were analyzed via one-way ANOVA. Post hoc analyses were carried out with Dunnett’s test (to compare doses to water control) and Tukey’s HSD test (to compare all dose groups to each other), with significance defined as p<0.05 unless otherwise noted. Guidance in SAS coding was provided by the Statistical Consulting Service at Purdue University.

3. RESULTS

3.1. Clinical findings and survival rate.

No observable change in behavior was noted in any animal throughout the study. Two animals were euthanized prior to study completion due to issues related to oral gavage (e.g., aspiration of dose into lungs). These events were deemed to be related to dose administration and not the phenolic treatment itself. All other animals survived until scheduled necropsy, with no observed signs or symptoms of adverse, treatment-related effects.

3.2. Body weight and food consumption.

Animals were 6 months old and weighed almost 300 g upon starting the dosing protocol. No significant differences in body weight were noted between any of the treatment groups (Figure 2). Due to the large variations in body weight within each group, we also examined the percent change in body weight on a weekly basis as well as total change over 90 days (Figure 3a,b). During the first few weeks of the study, animals in all groups lost weight, though they rebounded from this effect and gradually gained weight throughout the study. We attribute this to the initial stress and added stomach volume associated with daily gavage and consider it unrelated to blueberry polyphenol treatment. This is confirmed by the weight stability of the non-gavaged control group (see section 3.8). While no significant differences were noted week to week, the high dose group gained significantly more weight than the lyophilized blueberry group over 90d. Food consumption among groups was similar throughout the study, though the lyophilized blueberry group tended to eat less than the water control and high dose groups (Figure 3c,d). These differences in body weight change and food consumption were more pronounced over 90 days and were reflected in the food efficiency ratio (FER, defined as weight gain/food intake), with the high dose group having a significantly higher FER than the lyophilized blueberry group (Figure 3e,f).

Figure 2 –

Body weight over 90d. No significant differences in absolute body weight were observed throughout the study. BB = lyophilized blueberry dose group.

Figure 3 –

Week-to-week and total body weight change (a,b), food consumption (c,d), and food efficiency ratio (FER, defined as g body weight gain/g food intake; e,f). In week-to-week comparisons, no significant differences from water control were observed in body weight change (a) or FER (e), though several significant differences in food consumption were observed. Over the course of 90d, the high dose group gained significantly more weight (b), ate more food (d), and had a higher FER (f) than the BB dose group. Data shown as mean ± SEM. Significant differences detected using one-way ANOVA with either Dunnett’s (a,c,e) or Tukey’s HSD (b,d,f) post hoc test (α=0.05) and are indicated with (*) or lower case letters, respectively. BB = lyophilized blueberry dose group; Med = medium dose group.

3.3. Gross necropsy and organ weights.

Upon completion of the study, animals were euthanized and a full necropsy was performed. During necropsy, no tumors or other abnormalities were noted in any of the animals. Harvested organs were weighed and, with the exception of the pancreas, showed no differences in absolute weight, as a percentage of whole-body weight (organ wt/bw), or as a percentage of lean mass (organ wt/brain wt) (Bailey et al. 2004) (Table 2). The differences in pancreatic weights most likely resulted from challenges associated with harvest, as it has a diffuse structure within the mesentery and abdominal fat of rats (Tsuchitani et al. 2016), and is likely not physiologically significant.

Table 2 –

Absolute and relative tissue weights.

			Dose
		Water	Low	Medium	High	BB
Final body wt	(g)	301 ± 17	302 ± 22	298 ± 25	312 ± 37	299 ± 21
Brain	(g)	1.67 ± 0.071	1.64 ± 0.101	1.68 ± 0.135	1.62 ± 0.118	1.60 ± 0.071
	(g/100 g BW)	0.56 ± 0.026	0.54 ± 0.029	0.57 ± 0.039	0.52 ± 0.067	0.54 ± 0.047
Colon	(g)	0.84 ± 0.146	0.78 ± 0.201	0.77 ± 0.099	0.86 ± 0.219	0.78 ± 0.083
	(g/100 g BW)	0.28 ± 0.048	0.26 ± 0.058	0.26 ± 0.021	0.28 ± 0.062	0.26 ± 0.028
	(g/100 g brain)	50.1 ± 8.9	47.3 ± 9.87	45.8 ± 5.62	54.2 ± 17.92	48.5 ± 4.83
Heart	(g)	1.09 ± 0.079	1.17 ± 0.177	1.09 ± 0.086	1.13 ± 0.088	1.10 ± 0.064
	(g/100 g BW)	0.36 ± 0.032	0.39 ± 0.065	0.37 ± 0.026	0.36 ± 0.019	0.37 ± 0.018
	(g/100 g brain)	65.5 ± 5.66	71.7 ± 11.74	65.2 ± 6.82	70.6 ± 8.80	68.6 ± 5.26
Liver	(g)	8.08 ± 0.514	7.76 ± 1.223	7.68 ± 1.199	8.17 ± 1.039	7.68 ± 0.925
	(g/100 g BW)	2.69 ± 0.185	2.56 ± 0.264	2.54 ± 0.237	2.62 ± 0.192	2.56 ± 0.164
	(g/100 g brain)	484 ± 40.6	473 ± 54.3	457 ± 66.5	510 ± 92.1	481 ± 69.4
Lungs	(g)	3.21 ± 0.792	2.70 ± 1.014	2.90 ± 0.971	2.81 ± 0.974	2.45 ± 1.081
	(g/100 g BW)	1.07 ± 0.258	0.91 ± 0.377	0.97 ± 0.318	0.90 ± 0.288	0.81 ± 0.34
	(g/100 g brain)	192 ± 44.5	165 ± 63.1	174 ± 61.9	177 ± 71.4	154 ± 72.7
Kidneys	(g)	1.77 ± 0.151	1.78 ± 0.170	1.75 ± 0.201	1.87 ± 0.230	1.72 ± 0.163
	(g/100 g BW)	0.59 ± 0.028	0.59 ± 0.028	0.59 ± 0.025	0.60 ± 0.026	0.58 ± 0.033
	(g/100 g brain)	106 ± 8.3	109 ± 9.4	104 ± 9.4	117 ± 18.4	108 ± 12.9
R Kidney	(g)	0.89 ± 0.075	0.90 ± 0.082	0.89 ± 0.102	0.95 ± 0.103	0.88 ± 0.087
	(g/100 g BW)	0.30 ± 0.015	0.30 ± 0.012	0.30 ± 0.013	0.31 ± 0.013	0.29 ± 0.016
	(g/100 g brain)	53.4 ± 3.82	54.9 ± 3.97	53.0 ± 4.40	59.2 ± 8.42	54.8 ± 6.84
L Kidney	(g)	0.88 ± 0.081	0.88 ± 0.098	0.86 ± 0.102	0.92 ± 0.129	0.85 ± 0.081
	(g/100 g BW)	0.29 ± 0.016	0.29 ± 0.020	0.29 ± 0.015	0.30 ± 0.016	0.28 ± 0.019
	(g/100 g brain)	52.8 ± 4.8	53.7 ± 5.9	51.1 ± 5.2	57.6 ± 10.1	53.1 ± 6.3
Pancreas	(g)	2.30 ± 0.456	2.09 ± 0.413	2.71 ± 0.488	3.02 ± 0.495*	2.08 ± 0.412
	(g/100 g BW)	0.76 ± 0.143	0.69 ± 0.122	0.91 ± 0.169	1.00 ± 0.196*	0.71 ± 0.133
	(g/100 g brain)	138 ± 28.4	127 ± 23.9	161 ± 28.6	185 ± 35.4*	129 ± 26.3
Small Intestine	(g)	4.67 ± 0.601	4.75 ± 0.481	4.43 ± 0.728	5.27 ± 0.594	4.67 ± 0.598
	(g/100 g BW)	1.55 ± 0.196	1.57 ± 0.107	1.48 ± 0.164	1.70 ± 0.162	1.56 ± 0.100
	(g/100 g brain)	280 ± 36.4	291 ± 23.9	264 ± 41.0	328 ± 44.8	292 ± 43.6
Spleen	(g)	0.64 ± 0.086	0.64 ± 0.064	0.65 ± 0.117	0.68 ± 0.090	0.65 ± 0.073
	(g/100 g BW)	0.21 ± 0.027	0.21 ± 0.020	0.22 ± 0.028	0.22 ± 0.022	0.22 ± 0.021
	(g/100 g brain)	38.3 ± 4.76	39.5 ± 4.24	38.6 ± 7.38	42.2 ± 6.93	40.4 ± 5.17

Values are mean ± SD for 9–10 animals per group.

BB = whole freeze-dried blueberries.

^*Significantly different from water control (p<0.05).

3.4. Histopathological evaluation.

Histopathological analyses were performed by a board-certified veterinary pathologist who was blinded to treatment groups. Analysis compared the water and high dose groups for all tissues (Figure SI-1). No significant histopathological lesions were noted. Several mild lesions that were confined to a focal area were observed and were likely the result of normal biological variability between animals.

3.5. Clinical pathology.

3.5.1. Hematology.

Results from hematological analyses are shown in Table 3. Of all parameters measured, only one (monocyte count for lyophilized blueberries) was significantly different from the water control. All measured values were within the normal range for rats, and the statistically significant difference observed was not considered biologically significant.

Table 3 –

Hematology.^a

		Dose
	Water	Low	Medium	High	BB
n b	9	7	8	10	6
RBC (106/uL)	7.64 ± 0.12	7.58 ± 0.25	7.65 ± 0.13	7.75 ± 0.30	7.53 ± 0.62
HCT (%)	44.1 ± 1.71	45.0 ± 2.28	44.0 ± 0.93	44.6 ± 2.24	42.5 ± 2.74
HGB (g/dL)	14.0 ± 0.48	14.3 ± 0.52	14.1 ± 0.31	14.1 ± 0.62	13.4 ± 1.28
MCV (fL)	61.6 ± 0.98	63.7 ± 1.49	61.8 ± 1.65	61.5 ± 2.24	60.4 ± 3.78
MCHC (g/dL)	31.6 ± 1.09	31.9 ± 0.93	32.0 ± 0.39	31.3 ± 0.76	31.5 ± 1.38
RDW (%)	13.8 ± 1.06	13.8 ± 0.47	13.5 ± 0.70	13.5 ± 0.74	14.8 ± 2.96
WBC (103/uL)	11.3 ± 2.41	10.3 ± 1.18	11.5 ± 1.59	12.1 ± 2.62	11.7 ± 2.20
SEG (103/uL)	0.62 ± 0.22	0.78 ± 0.41	1.01 ± 0.36	1.06 ± 0.46†	0.95 ± 0.38
LYMPH (103/uL)	10.2 ± 2.44	8.9 ± 0.97	10.1 ± 1.64	10.6 ± 2.55	10.1 ± 2.41
MONO (103/uL)	0.32 ± 0.16	0.49 ± 0.17	0.28 ± 0.15	0.38 ± 0.18	0.57 ± 0.21*
EOS (103/uL)	0.23 ± 0.16	0.18 ± 0.14	0.21 ± 0.13	0.17 ± 0.09	0.17 ± 0.10
RETIC # (103/uL)	420 ± 95	475 ± 56	430 ± 86	408 ± 89	423 ± 134

Values are mean ± SD.

^aSeveral additional parameters (polychromasia, target cells, and anisocytosis) were examined and scored qualitatively, with all results considered normal; thus, they are not shown in the chart. Platelets were too clumped to count in most samples, so not shown in the chart.

^bSeveral blood samples were unable to be analyzed due to improper clotting. The total number of samples analyzed for each group is given by n.

BB = whole freeze-dried blueberries; RBC = Red Blood Cell count; HCT = Hematocrit; HGB = Hemoglobin; MCV = mean corpuscular volume; MCHC = mean corpuscular hemoglobin concentration; RDW = red blood cell distribution width; WBC = white blood cell count; SEG = segmented neutrophils; LYMPH = lymphocyte count; MONO = monocyte count; EOS = eosinophils; RETIC # = reticulocyte number.

^*Significantly different from water control (p<0.05).

^†Near significant difference from water control (p=0.072).

3.5.2. Serum biochemistry.

Results from serum biochemistry are shown in Table 4. The only statistically significant differences observed were decreased total cholesterol and chloride ion concentration in the high dose group compared to water control. Despite attaining statistical significance, these values are within normal ranges for rats and not considered physiologically significant.

Table 4 –

Serum biochemistry.

		Dose
	Water	Low	Medium	High	BB
Total protein (g/dL)	6.01 ± 0.21	6.03 ± 0.24	6.03 ± 0.21	5.99 ± 0.17	5.99 ± 0.17
Albumin (g/dL)	3.39 ± 0.14	3.39 ± 0.18	3.42 ± 0.14	3.42 ± 0.16	3.40 ± 0.15
Globulin (g/dL)	2.62 ± 0.10	2.64 ± 0.11	2.61 ± 0.11	2.57 ± 0.07	2.59 ± 0.11
A/G ratio	1.29 ± 0.06	1.29 ± 0.09	1.32 ± 0.04	1.32 ± 0.06	1.32 ± 0.10
BUN (mg/dL)	14.9 ± 2.42	17.6 ± 3.10	16.8 ± 2.49	14.8 ± 1.32	16.3 ± 3.50
Creatinine (mg/dL)	0.67 ± 0.15	0.69 ± 0.09	0.74 ± 0.09	0.70 ± 0.12	0.68 ± 0.10
ALKP (U/L)	123 ± 30.7	112 ± 18.0	111 ± 25.1	117 ± 19.9	114 ± 15.7
ALT (U/L)	50.3 ± 8.3	48.9 ± 9.1	45.7 ± 7.7	53.8 ± 10.7	44.0 ± 5.9
Total bilirubin (mg/dL)	0.12 ± 0.04	0.17 ± 0.05	0.15 ± 0.05	0.14 ± 0.05	0.17 ± 0.07
Amylase (U/L)	1662 ± 318	1517 ± 333	1614 ± 393	1702 ± 304	1673 ± 362
Lipase (U/L)	97.0 ± 23.3	100 ± 35.8	105 ± 27.2	101 ± 15.9	120 ± 35.0
Cholesterol (mg/dL)	135 ± 10.3	128 ± 11.2	123 ± 14.8	118 ± 8.9*	131 ± 20.0
Calcium (mg/dL)	11.6 ± 0.20	11.6 ± 0.30	11.5 ± 0.35	11.6 ± 0.13	11.5 ± 0.26
Chloride (mmol/L)	99.4 ± 1.2	99 ± 1.3	98.6 ± 1.2	97.3 ± 0.9*	99.8 ± 1.6
Phosphorus (mg/dL)	7.53 ± 0.61	7.96 ± 0.83	7.52 ± 0.81	7.40 ± 0.52	7.41 ± 0.35
Potassium (mmol/L)	7.02 ± 0.80	7.27 ± 0.89	7.36 ± 0.47	7.06 ± 0.72	7.10 ± 0.83
Sodium (mmol/L)	141 ± 1.5	141 ± 1.0	142 ± 1.1	141 ± 0.9	141 ± 0.8
Anion gap (mmol/L)	16.1 ± 2.40	16.2 ± 2.41	16.6 ± 1.65	17.4 ± 2.17	16.5 ± 2.80
CO2 (mmol/L)	32.7 ± 1.70	33.2 ± 1.62	33.8 ± 2.28	33.2 ± 1.69	32.0 ± 3.91

Values are mean ± SD for 9–10 animals per group.

BB = whole freeze-dried blueberries; A/G = Albumin:Globulin ratio; BUN = Blood Urea Nitrogen; ALKP = Alkaline Phosphatase; ALT = Alanine Transaminase; CO₂ = Carbon Dioxide.

^*Significantly different from water control (p<0.05).

3.5.3. Urinalysis.

Results from urinalysis are shown in Table 5. Urine from the high dose group was significantly more acidic than the water control. A dose-dependent darkening of urine and feces was observed shortly after the administration of the first oral gavage dose and persisted throughout the study (Figure SI-2). This was most likely due to the high concentration of blueberry polyphenols, as confirmed by the dose-response appearance of blueberry polyphenol metabolites in urine (Table 5). In preliminary experiments, similar results were observed, though within 48h after discontinuing treatment, urine and fecal color returned to normal and urinary polyphenol metabolites were no longer observed (data not shown).

Table 5 –

Urinalysis.

		Dose
	Water	Low	Medium	High	BB
Quantitative measures
Volume (mL)	11.2 ± 4.5	12.4 ± 6.6	10.6 ± 3.2	10.1 ± 3.6	8.2 ± 3.0
Specific Gravity	1.025 ± 0.010	1.056 ± 0.088	1.028 ± 0.016	1.032 ± 0.020	1.031 ± 0.010
Protein (g/L)	0.34 ± 0.37	0.36 ± 0.36	0.41 ± 0.56	0.49 ± 0.36	0.46 ± 0.42
pH	7.6 ± 0.6	7.6 ± 0.7	7.5 ± 0.6	6.6 ± 0.4*	7.4 ± 0.8
24h metab. (umol exc.) a	4.50 ± 0.53 w	12.3 ± 4.87 x	42.7 ± 16.7 y	91.2 ± 34.6 z	11.4 ± 3.40 x
Qualitative measuresb
Color c
Normal	10	10	6	4	9
Darkened	0	0	3	6	0
Glucose
Negative	8	9	7	10	5
Trace	2	1	2	0	4
Triphosphate crystals
None	7	4	2	2	4
Few	1	3	5	5	4
Moderate	0	1	1	2	0
Many	2	2	1	1	1

Values are mean ± SD for 9–10 animals per group.

BB = whole freeze-dried blueberries.

^aQuantitated as sum 24h urinary excretion of phenolic metabolites via LC-MS/MS. Letters indicate significant differences between dose groups using Tukey’s HSD test (p < 0.05).

^bQualitative measures also included ketones, bilirubin, and blood in the urine, none of which were detected in samples, so they are not included in the table.

^cColor was independently graded as pale yellow, yellow, dark yellow, or brown. Samples categorized as “normal” if color was yellow or pale yellow and “darkened” if dark yellow or brown.

^*Significantly different from water control (p<0.05).

3.6. Intestinal permeability.

3.6.1. FITC test.

Intestinal permeability was analyzed using the FITC technique. As shown in Figure 4a, intestinal permeability was significantly higher in the high dose group as compared to other treatment groups (p<0.05).

Figure 4 –

Intestinal barrier function. The integrity of the GI tract was measured using the FITC-dextran method (a). Despite differences in FITC uptake, no histologically significant differences were found (b). Representative histology slides shown at 20x magnification. Data shown as mean ± SEM. Significant differences detected using one-way ANOVA with Tukey’s HSD post hoc test (α=0.05) and indicated with (*). BB = lyophilized blueberry dose group; FITC = fluorescein isothiocyanate-dextran.

3.6.2. Intestinal histopathology.

Based on the results of the FITC test, histopathological analyses of each segment of the GI tract were performed for all animals in all dose groups. Using a histomorphological scale for eight potential markers of toxicity (see Methods), all intestinal segments were scored as “normal” for all dose groups (Figures 4b and SI-3). Although several prominent lymphoid bodies and gastric submucosal edemas were observed in several rats, these were observed in rats from all groups and did not trend towards a specific treatment group, thus, they were considered to be due to biological variation.

3.7. Bone mineral density.

Postmenopausal women are susceptible to rapid bone loss, as modeled by the OVX rat model (Thompson et al. 1995). To assess the effect of increasing doses of blueberry polyphenols on bone, bone mineral density was measured in the femur, tibia, and lumbar spine (Figure SI-4). No significant differences were observed between the water and treatment groups.

3.8. Non-gavaged group.

Non-gavaged animals were no different from water controls for any of the endpoints measured (Tables SI-1–4 and Figures SI-3–6). There was a non-significant trend towards higher body weights during the first few weeks of the study, though this effect resolved later in the study and was most likely due to the initial stress of daily gavage in treatment groups and unrelated to the blueberry polyphenols. Food consumption patterns were significantly different early in the study, though differences disappeared after the first month of the study. This was most likely due to the combined effects of stress associated with and the additional stomach volume from daily gavage. These factors may have initially suppressed appetite in gavage-fed animals, though they did not alter body weight or FER significantly over the course of 90d. The only statistically significant difference was in serum CO₂ concentration, though this difference was not considered biologically significant.

4. DISCUSSION

In the current study, few differences were observed among treatment groups, with almost all measured parameters showing no significant differences between blueberry polyphenol treatments and water control. Although slight differences were observed in 90d body weight change, food consumption, and FER between the high and BB dose groups, these findings did not lead to other significant changes in histopathology. Organ weights were similar among all treatment groups, even after adjustments for body size (organ wt/bw) or brain weight (organ wt/brain wt). Adjusting for brain weight is a surrogate measurement that adjusts for lean body mass, which typically is not affected by xenobiotics like blueberry polyphenols (Bailey et al. 2004). Clinical pathology measures were within normal ranges, though several statistically significant differences were observed between treatment groups and the water control.

Postmenopausal women are the most frequent consumers of herbal and botanical dietary supplements and also the most vulnerable to rapid bone loss, as they lose ~20% of their total bone mass in the first 5 years after menopause, leaving them susceptible to osteoporosis (Hodges et al. 2019). Recent investigations suggest that polyphenols may help slow or even reverse age-related bone loss (Hubert et al. 2014). However, using the OVX rat model to mimic this life stage, we did not see any significant differences in BMD among treatment groups.

The most notable difference was observed in the darkened color of urinary and fecal output for animals treated with blueberry polyphenols. There was a dose-dependent darkening of the urine, and the high dose group had significantly more acidic urine. Additionally, the FITC-dextran test showed evidence of increased intestinal permeability in the high dose group. And, though we did not observe histopathological differences in any portion of the GI tract among groups, this increase in intestinal permeability may indicate mildly damaged gut barrier function, which could lead to increased levels of blueberry polyphenols in systemic circulation. This may be significant because native polyphenols typically demonstrate low levels of bioavailability in vivo (Smeriglio et al. 2016). If gut barrier function, especially in the small intestine, is damaged and increasing levels of polyphenols are found in circulation, this may explain the decrease in urinary pH in the highest dose group, as fruit-derived polyphenols are acidic. Additionally, colonic metabolism of polyphenols produces large quantities of low molecular weight phenolic acids. Colonic metabolites are acidic and may be absorbed at levels 10-fold greater than small intestinal phenolic metabolites (Williamson and Clifford 2017). After entering systemic circulation, phenolic acids are excreted in the urine and may also contribute to decreasing urine pH. If this persists over longer periods of time, it may lead to systemic acidosis in the animals. In the current study, however, we did not observe any other systemic changes in the animals, indicating that if systemic acidosis did eventually develop, it would require longer than 90d.

Several other safety studies of fruit and vegetable extracts are available in the literature. One of the oldest and most frequently cited studies was performed by Pourrat et al (1967) and laid the foundation for toxicity studies of anthocyanin-rich extracts. In this study, extracts of currants, blueberries, and elderberries were tested over varying time periods and found that repeated doses of up to 9 g/kg bw/d did not show evidence of toxicity over three successive generations of rats, mice, or rabbits (Pourrat et al. 1967). In the 1980s, investigations of purple color from grape extracts showed no evidence of toxicity up to 15% (w/w) of the diet in beagle dogs over 90d or Sprague-Dawley rats over two successive generations. The only difference noted in these studies was decreased weight gain in the highest dose group, most likely due to the lack of isocaloricity among diets (Becci et al. 1983a; Becci et al. 1983b). A more recent study of grape skin and grape seed extracts in SD rats over 90d at 0–2.5% (w/w) of the diet showed no differences among treatment groups (Bentivegna and Whitney 2002). Other investigations on the safety of anthocyanins extracts showed that extracts of purple corn (Nabae et al. 2008) and bilberry (Eandi 1987) administered for 3–6 months in various animal models showed no evidence of toxicity, apart from discolored urine and feces. Safety studies of other polyphenol-rich extracts, like apple polyphenols and green tea extract (GTE), are also found in the literature, with varying results in rodents. Apple polyphenols showed no toxicity at up to 2000 mg/kg bw/d over 90d of oral gavage in rats (Shoji et al. 2004), while GTE showed minimal impact at doses up to 1200 mg/kg bw/d over 6-months in one study (Morita et al. 2009) but significant toxicity in a second study over 14-weeks at 1000 mg/kg bw/d (Chan et al. 2010).

Setting safety guidelines for polyphenols and polyphenol-rich extracts is challenging because there is significant heterogeneity in polyphenol content between extracts and standardizing them is challenging. Additionally, the paucity of studies and their varied results add to this challenge. In the present study, we chose to base our doses off total polyphenol content and incorporate them into a gavage-fed preparation to minimize study variability and maximize the translatability of our results, as this is how dietary supplements are typically consumed (i.e., as a concentrated dose at a single time each day). Despite the small, though statistically significant, differences in gut permeability and urine pH in the high dose group, all measured parameters remained in the normal range for this animal model and we, therefore, propose that the NOAEL for blueberry polyphenols is ≥ 1000 mg/kg bw/d for OVX-SD rats. This translates to a 70 kg human consuming ~10 g blueberry polyphenols per day, an amount higher than is currently available in dietary supplements. Thus, we conclude that regular consumption of blueberry polyphenols in foods and supplements is likely safe for consumers.

Highlights.

• Evaluated subchronic oral toxicity of purified blueberry polyphenols in OVX rats.

• No adverse effects induced by purified blueberry polyphenols in rats.

• Proposed NOAEL of ≥ 1000 mg/kg bw/d for blueberry polyphenols in OVX rats.

• Blueberry polyphenol dietary supplements are likely safe to consume.

ABBREVIATIONS

BB

Lyophilized blueberry dose group

BMD

Bone mineral density

FDA

Food and Drug Administration

FER

Food efficiency ratio

FITC

Fluorescein isothiocyanate

GI

Gastrointestinal

GTE

Green tea extract

NOAEL

No observed adverse effect level

OECD

Organisation for Economic Co-operation and Development

OVX-SD

Ovariectomized Sprague Dawley