The study was performed during the spring of 2021 at the Obstetric, Gynecological, and Andrological Clinic of the Veterinary Hospital of the Veterinary Medicine Department of the “Aldo Moro” University of Bari (Italy). Dogs' experimental management is reported in Figure 1. The dogs were clinically examined to ensure their health status 30 days before the commencement of the trial. Each dog has submitted for a clinical examination and blood analyses (hematocrit and total protein) (25, 26). An ultrasound exam of the reproductive tract was also performed, and it was verified that all dogs reacted positively to sperm collection by digital manipulation. Only normospermic male pet dogs kept in the house were included. In the end, forty mixed-breed dogs were included in the present study (Supplementary Table S1 reported all animals characteristics). Dogs were healthy with a mean age of 3.8 years, a mean body weight of about 27 kg, and a mean body condition score (BCS) of 3. All the dogs started to be fed with the same commercial feed [10 g/kg of body weight (BW) daily, composition reported in Table 1] two times a day to ensure a 30-days period of adaptation to the new commercial feed before starting the trial.

The dogs were randomly assigned to one of two equal-sized groups (n = 20) using www.randomizer.org (27) (Figure 1).

One group was the experimental one (PTHL) and the other was the control group (CON). The PTHL group received the supplement containing PTHL (Oxilem®, I-Green, Padua, Italy). Table 2 reports the chemical composition and antioxidant activity of PTHL (24). The PTHL group received 500 mg/kg of PTHL each day during the 90 days of the trial. The dose was orally administered in powder according to the manufacturer's instructions, based on empirical trials (IGreen, Padua, Italy). The supplement was orally administered. During the trial, semen and blood samples collection were performed before starting the trials (day 0) and 30, 60, and 90 days.

Blood was aseptically collected via cephalic vein puncture using disposable needles (22G), with a negative pressure 4 ml tube system for serum (without anticoagulant) and plasma (with 15 USP U/ml of heparin) (Becton, Dickinson Canada Inc, Vacutainer 1, Oakville, Canada). Tubes for plasma were immediately centrifuged (1,500 × g for 10 min) while tubes for serum were allowed to clot at a refrigerated temperature for 10 min prior to being centrifuged (1,500 × g for 10 min). All plasma and serum aliquots were stored at −80°C until analyses. Clinical biochemistry parameters were obtained from the serum samples using an automated biochemistry analyzer (CS-300B; Dirui, Changchun, China) as described by De Palo et al. (30). The following parameters were analyzed: alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatine phosphokinase (CPK), lactate dehydrogenase (LDH), alkaline phosphatase (ALP), glucose (Glu), blood urea nitrogen (BUN), creatinine (Crea), total serum protein (TP), albumin (Alb), cholesterol (Chol), triglycerides (Trig), non-esterified fatty acids (NEFA), calcium (Ca), phosphate (P), magnesium (Mg), chloride (Cl), (Gesan Production Kit, Campobello di Mazara, Trapani, Italy). Besides, globulins (Glob) and albumin/globulin ratio (Alb/Glob) were calculated starting from total protein and albumin parameters. Standard assay kits were used to calibrate the multi-parameter analyzer (Seracal, Gesan Production Kit, Campobello di Mazara, Trapani, Italy) before each analytical session. After setting the calibration curve, two multi-parameter control sera, a normal and a pathological one (Seracontrol N and Seracontrol P, Gesan Production Kit, Campobello di Mazara, Trapani, Italy), were used to verify internal accuracy, considered satisfactory when the measured value deviated by no more than 3.00% from the manufacturer's declared values. Each sample was analyzed in triplicate, and the value used in the raw dataset was the arithmetic mean of the three recordings for each item. Plasma samples were used for oxidation parameters and antioxidant enzyme activities assays. Thiobarbituric acid reactive substances (TBARS) were determined spectrophotometrically as described by Maggiolino et al. (31), by adding 100 ml of plasma to a 3.7 μl/ml thiobarbituric acid solution. Plasma reactive carbonyl derivative (RCD) levels were determined according to Faure and Lafond (32) using the carbonyl reagent DNPH. Plasma (200 ml) was mixed with 1 ml of water and 2 ml of 200 μl/ml trichloroacetic acid and centrifuged at 1,000 × g for 10 min. The pellet was resuspended in 1 ml of 10 mmol/L DNPH and incubated for 60 min at 37.8°C. For control, 1 ml of 1 mol/L hydrochloric acid was used instead of DNPH. Subsequently, 1 ml of 200 μl/ml trichloroacetic acid was added, and the sample was centrifuged at 1,000 × g for 10 min. The pellet was washed with 1:1 ethanol-ethyl acetate solution and centrifuged at 1,000 × g for 10 min. The pellet was mixed with 1 ml of 6 mol/L guanidine (diluted in 20 mmol/L dihydrogenphosphate at pH 2.3). Finally, the sample was incubated for 40 min at 37.8°C. The absorbance was measured at 380 nm.

Hydroperoxides (Hy) were determined spectrophotometrically by an iodometric method as described by Maggiolino et al. (33). Aliquots (90 ml) of plasma were put into eight microcentrifuge vials (1.5 ml). Then, 10 μl of 10 mM TPP in methanol were added to four of the vials to reduce ROOHs, thereby generating a quadruplicate of blanks. Methanol (10 ml) was added to the remaining four vials to produce a quadruplicate of test samples. All the vials were then vortexed and incubated at room temperature for 30 min prior to the addition of 900 ml of FOX2 reagent. After mixing, the samples were incubated at room temperature for 30 min. The vials were centrifuged at 2,400 × g for 10 min with a swing-out rotor (Hettich Rotenta/RP centrifuge, Hettich-Zentrifugen, Tuttlingen, Germany). The absorbance of the supernatant was measured at 560 nm using an Ultraspec 2000 spectrophotometer (Pharmacia Biotech, Uppsala, Sweden). The ROOH concentration in the plasma samples was calculated using the mean absorbance difference between quadruplicates of test samples and blank samples.

Protein carbonyls (PC) levels were determined spectrophotometrically as reported by Salzano et al. (21). The superoxide dismutase (SOD) (EC 1.15.1.1) activity was examined according to Misra (34), and the enzymatic activity was based on the 50% inhibition rate of epinephrine auto-oxidation at 480 nm (35). The SOD activity was assessed as the 50% inhibition rate of epinephrine auto-oxidation at 480 nm. The epinephrine autoxidation stimulation by traces of heavy metals present as contaminants in the reagents was prevented by adding 10⁻⁴ M EDTA in the buffer to chelate those ions. The SOD activity was expresses as U/ml.

The catalase (CAT) (EC 1.11.1.6) activity was assayed by the method of Clairborne (36) as described by Tateo et al. (37). The amount of enzyme required to degrade 1 μmol of H₂O₂ in 60 s was defined as 1 unit of enzyme activity by following the decrease in absorbance of H₂O₂ at 240 nm (e = 40 M⁻¹ cm⁻¹). Its activity was expressed as U/mg of protein.

The glutathione peroxidase (GPx) (EC1.11.1.9.) activity was measured according to Gunzler (38) as described by Dinardo et al. (39). The analysis was based on the measure of the rate of reduced glutathione oxidation by tertbutyl hydroperoxide, catalyzed by GPx. The constant concentration of reduced glutathione was ensured by the addition of exogenous glutathione reductase and NADPH, which converted the oxidized glutathione to reduced glutathione. The rate of oxidized glutathione formation was then measured by the change in the absorbance of NADPH at 340 nm. Its activity was expressed as nmol of NADPH oxidized/min per ml.

Dog's semen was collected into an artificial vagina by manual stimulation, while the dogs sniffed swabs of bitches in estrous (using natural estral pheromones). The ejaculate collection was performed using 3 different Falcon tubes, one for each part of the semen: urethral, spermatic, and prostatic (40, 41). The ejaculation analysis was performed as described by Alonge et al. (42). The second seminal part was analyzed by the Computer Assisted Sperm Analyzer (43) (CASA, IVOS-Sperm CASA system, Version 12.3, Hamilton Thorne, MA, USA). The CASA software (IVOS 12.3 version) was set up for canine semen-speficic parameters as reported in Table 3. According to the manufacturer's instructions, for each analysis, a 3 μl drop from each sperm sample was diluted 5 times in Tris-Fructose extender and put on a Leja slide 4 chambers of 20 μm (Leja Products B.V. Nieuw Vennep, The Netherlands). The Leja slide was positioned in the dedicated chamber of the microscope, allowing it to settle for a few seconds before analysis. The computerized analyzer scanned five random non-consecutive microscopic fields. The parameters evaluated were: ejaculate volume, concentration, total motility, and percentage of motile spermatozoa (progressive motility), velocity average pathway (VAP), straight-line velocity (VLS), curvilinear velocity (VCL), amplitude of lateral head displacement (ALH), beat-cross frequency (BCF), straightness (STR), linearity (LIN), and total number of counted cells (TSC). VAP was elaborated by the software as the average velocity of smoothed cell path, expressed in μm/s. Then, the overall sperm population was divided into 4 groups based on the velocity, according to low VAP cut-off (LVV) and medium VAP cut-off (MVV). Thus, sperms were classified as follows: rapid spermatozoa, with VAP > MVV; medium spermatozoa, with LVV < VAP < MVV; slow spermatozoa, with VAP < LVV; and static spermatozoa, represented by the fraction of those cells not moving during the analysis (44).

The dataset was tested for normal distribution and variance homogeneity (Shapiro–Wilk test). Afterward, all data were subjected to analysis of variance (ANOVA) using the general linear model (GLM) procedure as reported the following model:

where Y_ijkl represents the dependent variables, μ is the overall mean; α_i is the ith dog random effect (i = 1,…40), O_j is the effect of the jth oral administration treatment (j = 1, 2), T_k is the effect of the kth time (1,…4), (O × T)_jk is the binary interaction effect of jkth (1,…8) treatment × time and ε_ijk is the error term. A Tukey test for repeated measures with respect to time was applied to evaluate the differences among means when the effect of time or the binary interaction of treatment × time was significant. The significance level was set at p < 0.05, and the results were expressed as means and mean standard error (SE). Data analysis was performed using SAS software (46).

The overall data obtained from the serum biochemical and electrolyte profiles in dogs of both experimental and control groups showed ranges almost comparable with those considered physiological for this species (47, 48). The diet inclusion of PTHL in dogs did not affect the dogs health, considering that all blood constituents alteration makes it possible to hypothesize and potentially diagnose organs functionality alterations (49). For example, ALT activity is strongly related to liver function, AST to muscle and liver cell function (50), and ALP has a widespread tissue distribution, with a lot of isoenzymes isolated from other different tissues, such as liver, bone, intestinal mucosa, kidney, and leucocytes, but all of them showed values within ranges reported in the literature (51), demonstrating that treatment did not cause any tissue damages. Animals' antioxidant defense can be raised in vivo (enzymatic) or can derive from the diet (non-enzymatic) (52). For this, TBARs, Hy, and protein carbonyls are considered biomarkers of cell damage: their increase can be generated by oxidative stress and a boost of oxygen reactive substances (ROS). Additionally, FRAP values allow an overall evaluation of plasma antioxidant capacity (53, 54). It is well known that the dietary assumption of high amounts of substances with antioxidant activity results in the transfer of these molecules to different animal tissues, usually followed by a significant increase in their total antioxidant capacity (55), and that several polyphenolic molecules in plants possess antioxidant activities (23). Research is raising interest in these compounds due to their beneficial effects on health as anticarcinogenic (56), anti-inflammatory (57), immune and microbial modulating (58, 59), especially if taken through the diet. Several studies confirm a positive correlation between total antioxidant activity and the total phenol content assumed with diet, for example, spinach and broccoli (60, 61). However, the pharmacokinetics and antioxidant activity of polyphenols are still unclear in dogs (48, 62). Generalized stress induces a systemic increase in oxidation products.

The patterns that do not foresee the activation of the endogenous antioxidant systems, and specifically the catabolites of lipid and protein oxidation (TBARS, hydroperoxides, and protein carbonyls) measured in the plasma, did not show any difference between the two groups of animals during the entire experimental test. However, this does not exclude the hypothesis that the inclusion of PTHL in the diet of dogs with different stressful conditions (work, running, specific pathologies, etc.) may induce a variation in the production of these metabolites. On the other hand, the maximum antioxidant capacity measured by laboratory techniques is different, i.e., the animals that have taken PTHL show greater enzymatic activity and higher levels of FRAP. These results may indicate a potential positive effect of PTHL dietary supplementation against cellular oxidative stress and we can hypothesize that animal under different conditions, i.e., living stress challenges, could show greater differences on these parameters after PTHL assumption. In fact, the enzymatic activity evaluated through the used techniques highlights a capacity to react to an in vitro insult and therefore potential greater antioxidant response. These results agree with what observed by other authors studying the inclusion of polyphenols in the diet (62). On the other hand, increased enzymes activity represented a positive result for dog health considering that these enzymes serve as an endogenous defense against oxidative stress phenomena by determining free radical neutralization (63).

Infertility is a common problem in the reproduction of dogs. For this reason, several authors suggested different food supplements to improve the quality of the ejaculate (64). According to recent studies (5, 42), the dietary approach by balanced food integration is an effective method for the improvement of fertility in dogs. Present results showed enhancing effects of a 2-month natural antioxidant diet supplementation on canine spermatozoa, as already observed in dogs' sperm quality (42), as well as in other species (65) by other authors. Specifically, the integration of PTHL can significantly increase the semen volume and spermatozoa motility in healthy normospermic dogs. In fact, after only 30 days, the PTHL group generally showed considerable improvements from the feed integration. Furthermore, in the PTHL group, the progressive motility and the percentage of rapid-movement sperms significantly increased, while the percentage of slow and static sperms significantly decreased. Indeed, thanks to PTHL feed supplementation, the sperm quality was significantly improved in all dogs in the treated group. Although the first evidence is evaluable after 30 days of supplementation, the best results have been reached at 60 days of administration, which approximately corresponds to the physiological length of the total duration of spermatogenesis in dogs (61.9 ± 0.14 days; (66)). The improvements in sperm motility depend probably on the antioxidant power of polyphenols contained in PTHL, which is able to improve sperm quality at the beginning of the sperm differentiation and development. It is well known that oxidative stress is one of the major issues associated with sperm function (67, 68) and sperm motility is one of the factors limiting male fertility (69). Although all dogs enrolled were healthy without problems of fertility, the PTHL group showed higher total sperm count, concentration, and progressive motility, compared with the CON group. Moreover, the integration of antioxidants, compared with the normal diet, is associated with a higher number and better motility of spermatozoa (70). In fact, spermatozoa are sensitive to oxidative stress because they are well endowed with polyunsaturated fatty acids, which are vulnerable to free radical attack at the alpha methylene carbons adjacent to the carbon–carbon double bonds (6). In addition, the spermatozoa's capacity for antioxidative defense is relatively low/vulnerable in comparison to other cells/tissues that are susceptible (71). Furthermore, spermatozoa are attacked by the oxidative effect of leukocytes, such as neutrophils, which are present in the male genitals for infection and other causes (67). The decrease in concentration, motility, and function of canine spermatozoa in dogs can be attributed to poor food intake, reduced absorption, increased losses, or augmented demand for microelements. Recent literature in farm animals show that natural polyphenols are used to reduce oxidative stress because of their antioxidant properties (15, 21, 72, 73). The results of our study are biologically plausible. The production of reactive oxygen species (ROS) causes loss of motility and a decreased capacity for sperm–oocyte fusion, inducing harmful chemical and structural changes to sperm DNA, proteins, and lipids of plasma and mitochondrial membranes (9). ROS induces changes in the sperm membrane altering its fluidity, resulting in loss of motility, and impaired events, such as acrosome reaction (74). Several studies have shown a significant reduction in oxidative stress or DNA damage after treatment with antioxidants (8). Indeed, the literature shows that antioxidants may play a primary role in protecting male germ cells against oxidative action (75). Our results encourage considering the alimentary approach of balanced food supplementation for the solution of dog subfertility. Balanced feed supplementation and integration could mitigate the negative impact of infertility in canine breeding.