This work comprises part of a project dedicated to the study of age-related cognitive decline in humans and dogs (the Neurophenols Consortium). We aim to complete a novel study into the efficacy of mixed extracts of grape and blueberry in counteracting age-related deterioration of function. In aged dogs, polyphenol ingestion (including grape pomace), and vitamin use, are both thought to ameliorate the effects of aging on learning ability [1]; similarly beneficial effects have been reported for humans [2]. A popular hypothesis is that protection against oxidative stress explains these effects. In aged mice, the consumption of a mixed grape and blueberry extract, has also been shown to improve spatial navigation; one of the skills that declines with age. In this scenario, increased expression for hippocampal nerve growth factor mRNA [3] may play a causal role [4].

Studies reporting grape extract consumption in dogs (using grape seed and skin extracts, or grape seed proanthocyanidins), or grape pomace in aged dogs have, to date, not reported any side effects [1, 5, 6]. However, acute renal failure has been reported in dogs after their consumption of grapes, with kidney histopathology revealing tubular degeneration leading to necrosis, particularly in the proximal tubule [7–9]. In a retrospective study involving a cohort of 43 dogs, all of whom had eaten grapes, raisins, or both, animals presented with clinical signs consistent with kidney deterioration during a window extending from 24 h, until 5 days, after consumption. Vomiting, diarrhea, lethargy, and either olig- or, anuria, were the common clinical signs. A diagnosis of renal damage was supported by biochemical abnormalities showing higher plasma creatinine (Creat), blood urea nitrogen (BUN), an altered ionogram, glycosuria, and proteinuria. Again, histopathology revealed severe diffuse renal tubular degeneration, especially in proximal cells, with glomerular deterioration. Half of the dogs died [9]. In these cases, the precise amount of fruit eaten varied greatly (from 3 g/kg BW of raisin, to 150 g/kg BW of grape), as did the type of fruit (grape, raisin, seedless grape), and the affected breed [7–9]. Hepatic toxicity has also been associated with the consumption of plants such as greater celandine, green tea, valerian, or ayurvedic products. In these cases, higher concentrations of alanine transaminase (ALT), alkaline phosphatase (ALP), aspartate aminotransferase (AST), and bilirubin, were all demonstrated (reviewed in [10, 11]). Abnormal values of ALT and ALP provoked by grape consumption also point to the liver being a target of grape toxicity [9], although the factors responsible for hepatic damage, as well as the acute renal failure, have yet to be identified.

The Neurophenols Consortium is a Europe-North America research collaboration dedicated to the research, and development of natural ingredients and products to prevent age-related cognitive decline in humans and pets. The Consortium brings together scientists in the fields of phytochemistry, neuroscience, psychology and nutrition with companies specialized in the development of active ingredients and food supplements. The specific aims of the program are to characterize and formulate fruit extracts from blueberry and grape, to evaluate their safety and efficacy in pre-clinical and clinical trials.

The aim of this study was to assess the safety of a polyphenol-rich extract from grape and blueberry (PEGB; from the Neurophenols Consortium). We studied the safety of this extract following chronic use in dogs, by monitoring renal and hepatic health, using early biomarkers of renal damage as well as a biochemical approach.

Our aim was to assess the safety of a polyphenol-rich extract from grape and blueberry (PEGB; from the Neurophenols Consortium) for dogs, by monitoring early biomarkers of renal damage over a 24-week period. This work considerably extends the previous study periods reported, where platelet effects, and gene expression profiles, were interrogated after 7 days, or 3 months of supplement use [5, 6].

After PEGB consumption, biomarker values exceeded the reported maximal limits in no dog, with no differences observed at the end of the 24-week period, compared to beginning, for plasma CysC, and urinary CysC/Creat, Clu/Creat, or NGAL/Creat ratios. When considering these data, we conclude that the dogs neither presented with renal, nor hepatic injury, at the end of the study.

While bioavailability of the Neurophenols Consortium PEGB had never been evaluated in dogs, our evaluation of the safety of this supplement necessitated measurement of PEGB derivatives in plasma. The main polyphenols in the extract were flavan-3-ols, resveratrol, anthocyanins (malvidin, petunidin, peonidin, petunidin, cyanidin), and flavonol (quercetin). Some polyphenols and polyphenol metabolites were found in plasma. Malvidin, which is present in blueberry but not in grape, has been the only anthocyan detected, but it is known that anthocyanins are less absorbed than other flavonoids. The finding of resveratrol derivatives (which are grape specific) is in accordance with a study that also showed appearance of resveratrol conjugates (sulfate & glucuronide) in the plasma of dogs after resveratrol administration [13]. The valerolactones detected resulted from the metabolization of flavan-3-ols by gut microflora. Quercetin and isorhamnetin sulfate, which are present in both fruits, were also found. Other compounds may have been absorbed, but either they have not been identified, or their concentration was under the detection threshold, or they were rapidly metabolized and excreted. Very few data on polyphenols pharmacokinetics in dogs are available. Regarding resveratrol, Cmax could not be compared since in previous report [13] it was given to dogs at much higher doses than the intended dose in the present study (200–1200 mg/kg/d, compared to 4 mg/kg/d). When anthocyanins were given to pigs at 1 to 4 % of the diet (w/w), several metabolites were measured in liver, eye and brain while there were not detected in plasma [14], and again the doses were far higher than in the present study. Catechin and epicatechin glucuronides from a grape extract given to mice were measured in plasma [15], which was not the case in our study, but the dose used was still much higher (grape-derived polyphenols: 80 mg/kg/d). When green tea catechins (13 mg/kg/d, [16], 170 mg/kg/d [17]) and epigallocatechin gallate (EGCG; 250 mg/kg/d [18]) were given to dogs, respective metabolites were found in plasma, which was not the case after PEGB consumption where only valerolactones were detected. The difference could be explained either by the catechin sources or higher doses or both. Another possible explanation is that dogs were given the PEGB at the same time of their daily meal, and the plasma measurements were done after a relatively short period of exposure. Indeed in dogs given EGCG at 300 mg/kg/d, plasma area under the curve (AUC) for EGCG was higher in unfed than fed dogs [19]. When EGCG was given at 500 mg/kg/d, authors reported, although the difference did not reach the significance level, that the AUC for EGCG was 1.6 time higher after 28 days of dosing than after 14 days [19]. The data of the present study demonstrated that the polyphenols of the PEGB extract were, at least in part, bioavailable, and this is the first report on the appearance of valerolactones as well as quercetin, isorhamnetin sulfate and malvidin in the plasma of dogs after consumption of a mixture of polyphenols.

The origin of the grape toxicity described in the literature for dogs is still obscure, but numerous hypotheses have emerged. Among them, it was reported that exogenous compounds on grapes, such as mycotoxin, pesticides, or herbicide residues, could be responsible for the kidney toxicity, with histopathology indicating that the proximal cells are the primary target [8]. These findings provoked further hypotheses, such as the toxic accumulation of a foreign chemical (a xenobiotic), with a particular affinity for tubular specific transporters. Additionally, the expression of a perinuclear golden brown pigment [8], could imply its cytotoxic accumulation, with failed cellular clearance. Hypercalcemia and renal mineralization induced by the high sugar content of grapes are also current hypotheses.

The resveratrol concentration in grapes could also be responsible for renal damage. A previous study described that the no-observed-adverse-effect level of resveratrol consumption was 600 mg/kg BW/d in dogs. Consumption of twice this dose (1200 mg/kg BW/d) induced a loss of appetite, and weight [20]. Given that grapes contain 1.5 to 7.8 μg of total resveratrol per gram of fresh weight [21], it is highly unlikely that resveratrol is responsible for the acute kidney injury observed in clinical cases in dogs.

Plasma creatinine and urea are the most frequently measured parameters used to evaluate renal damage. High creatinine concentrations are seen when at least 75 % of renal function has already been lost [22]. In previous studies describing acute renal failure after grape consumption, symptoms appeared rapidly [9]. Therefore, we reasoned that to monitor kidney health, earlier biomarkers of renal damage would be required. In 2010, the Nephrotoxicity Working Group established a consortium between the European Medicines Agency, and the Food and Drug Administration. They listed seven biomarkers needed to detect the early development of renal injury [23]. Among these, we chose to assess CysC and Clu, because of their ease of use in dogs. In addition, NGAL was measured, as a promising early biomarker of drug-induced kidney injury. Collectively, these early biomarkers of renal damage are ideal for monitoring renal health, before irreversible damage, as they survey different renal functions, and compartments of the kidney.

Ordinarily, cystatin C, which is a low molecular weight protein produced at a constant rate by all cells, is completely reabsorbed and catabolized in proximal tubular epithelial cells [24]. Following renal injury, CysC concentration increases in the plasma, as the glomerular filtration rate declines [25]; an increased concentration in urine reflects tubular impairment [26]. Plasma CysC has previously been measured in healthy dogs (urea and creatinine concentrations within reference intervals), with the highest reported values of 2 μg/ml [27]. For all dogs that had received PEGB, at any dose, plasma CysC concentrations were beneath this upper limit. To the best of our knowledge, the referenced study [27] is the only one in which plasma CysC concentrations have been measured in healthy dogs by canine ELISA. We therefore conducted the same tests, in our study. In other studies, CysC was measured in serum and/or with a different ELISA kit or technique (i.e. Particle-Enhanced Turbidimetric Immunoassay), which may explain the slightly different reference ranges reported [28–30]. In our study, the maximum urinary CysC/Creat ratio that we measured in dogs following PEGB consumption (regardless of dose) was 79 μg/g, whereas reported urinary CysC/Creat ratios have been as higher as 0.11 ± 0.02 mg/g [31]. Therefore, we conclude that our CysC results revealed no glomerular or tubular impairments.

Clusterin is a high molecular weight glycoprotein expressed in epithelial cells (reviewed in [32]); in cases of acute renal failure, clusterin is found at high concentrations in the urine, indicating glomerular damage [33]. The highest urinary Clu/Creat ratio previously reported in healthy dogs was 4.87 μg/g [33]. In this study, the urinary Clu/Creat ratio measured in dogs after PEGB consumption (4 to 40 mg/kg/d) was far lower, ranging from 10 to 437 ng/g. Therefore, clusterin analyses also revealed no evidence of glomerular damage after PEGB consumption.

NGAL is a protein that has raised some interest since its mRNA and protein were detected in urine after induction of acute kidney injury in rodents [34]. NGAL mRNA has been found in the ascending limb of Henle, and in collecting duct cells after ischemia-reperfusion [35]. NGAL is ordinarily reabsorbed by the proximal tubule [35, 36]. However, in case of renal injury, reabsorption may decrease, which results in higher urinary concentrations. Tubular damage and reduced filtration may also cause the accumulation of plasma NGAL [37]. The reported ranges of urinary NGAL/Creat ratio have varied greatly in healthy dogs from 10 to 460 ng/g, or from 40 to 3660 ng/g [38, 39]. These variations could reflect reporting from client-owned dogs of various breeds, age, and gender, fed with various diets. In our study, the urinary NGAL/Creat ratios after PEGB consumption (at any dose), ranged from 0.9 to 10 ng/g, leading us to conclude that there was no evidence of tubular damage. Recently, it was found that plasma NGAL was not an absolute criterion with which to discriminate between a healthy dog, versus a dog with either chronic, or acute kidney disease [38] contrary to urinary NGAL [39] and this shows how we must be cautious when interpreting these values. Moreover, increasing plasma NGAL would reflect tubular and filtration dysfunction, data already provided by other early biomarkers of renal damage (Plasma CysC, and urinary CysC/Creat, NGAL/Creat, and Clu/Creat ratios). Therefore, we suggest that plasma NGAL measurements represent redundant data and can be omitted.

Intermediate measurements were also taken during the 24-week study period for all biomarkers; these did not reveal any significant differences.

The PEGB doses ranged from 4 to 40 mg/kg/d, the intentional dose for dogs facing cognitive decline being 4 mg/kg/d [as recommended by the Neurophenols Consortium]. In studies where dogs were fed supplements with grape seed/skin extract at 20 mg/kg/d [5], or grape seed proanthocyanidins at 5 mg/kg/d [6], symptoms related to acute renal failure were not reported. In the group given the PEGB at 4 mg/kg/d, the dose of grape extract was beneath these previously reported doses. In addition, dogs consuming five or even ten times the intentional PEGB dose, showed no alteration of kidney or hepatic damage at 24 weeks. These data corroborated the 2013 European Pet Food Industry Federation (FEDIAF) advice that dogs could safely consume grape extract.

We have considered why our extract, consumed long-term, as described in this study, appears to be entirely safe for consumption by dogs, in stark contrast to reports of acute renal failure in pets following their consumption of whole grapes or raisins. We can envisage some possibilities. The extract developed by our consortium is actually a complex mix of different extracts. How these extracts are derived (i.e. extracted from the grape), may have reduced, denatured, or eliminated, potentially toxic compounds. These factors may underlie the lack of any discernable toxicity when dogs consume the Neurophenols Consortium extract, even at high doses.

Following consumption of the PEGB at all doses, conventional biomarkers of renal and liver damage were within the reference range throughout the study, with values of early biomarkers of renal damage CysC, Clu, NGAL unremarkable. To our knowledge, this is the first study demonstrating that chronic consumption of the PEGB extract can be achieved with neither renal, nor hepatic damage, at least based on plasma and urine analyses. Of note, renal health was monitored using a panel of parameters encompassing both early biomarkers of renal damage, as well as conventional biochemistry; this complementary approach is recommended in future studies. To conclude, dogs can safely consume a polyphenol-rich extract from grape and blueberry (PEGB; from the Neurophenols Consortium).

ALP, Alkaline phosphatase; ALT, Alanine transaminase; AST, Aspartate aminotransferase; AUC, Area under the curve; BCS, Body condition score; BUN, Blood urea nitrogen; BW, Body weight; Clu, Clusterin; Cmax, maximum concentrations; Creat, Creatinine; CysC, Cystatin C; EGCG, Epigallocatechin gallate; NGAL, Neutrophil gelatinase-associated lipocalin; PEGB, Polyphenol-rich extract from grape and blueberry; UHPLC-MS/MS, Ultra high-performance liquid chromatography coupled to tandem mass spectrometry