Effects of Silybin Supplementation on Nutrient Digestibility, Hematological Parameters, Liver Performance, and Liver-specic mi-RNA Concentration in Dogs

The investigation into the effects of the test compounds on the expression of miRNA in blood serum was performed in two experiments: In EXP1, we examined the effect of commercial hepatoprotectant and silybin supplementation on miRNAs expression in healthy dogs (EXP1). In EXP2, we investigated the effect of commercial hepatoprotectant on miRNA expression in dogs with liver disorders. The healthy dogs in EXP1 were used as the control group in EXP2.

in dogs. According to Watson [13], 12% of dogs in United Kingdom were postmortem diagnosed with chronic hepatitis. The etiology may vary from infectious agents such as canine adenovirus 1 (CAV-1) or leptospirosis, to non-infectious factors like neoplasia, poisoning, or inherited malformations [14].
Regardless of the underlying problem, appropriate feeding should be considered as an effective means of handling liver diseases. One relevant question is whether nutraceutical supplementation also affects nutrient digestibility. This is worth considering, as patients with hepatic disorders are reported to show signs of gastrointestinal dysfunctions [15][16][17].
Silybin is a secondary plant metabolite that exhibits health-bene cial properties. Silibinin, composed of silybin A and B isomers, is one of the most active avonolignans present in the extract of milk thistle (Silybum marianum) [10]. Along with other avonolignans (isosilibinin, silidianin, and silicristin), it forms a complex known as silymarin. As far as we know, the literature does not discuss silybin as an antinutrient, and there has been no study of its effects on nutrient digestibility. Furthermore, no study has compared pure silybin with commercial hepatoprotectant containing silybin supplementation. For this reason, in the rst experiment (EXP1) we hypothesise that supplementation of either pure silybin (SIL) or commercial hepatoprotectant containing silybin (HEP) as a bioactive compound in healthy dogs does not affect nutrient digestibility, a control group (CON) of non-supplemented dogs also took part in the study.
We subsequently assumed that supplementation improves liver performance, while not exerting a detrimental effect on general health or blood parameters. In the second experiment (EXP2), we hypothesised that supplementation with commercial hepatoprotectant containing silybin improves liver function in dogs with hepatopathies. The main objectives of this study were thus: 1. to investigate the effects of diet supplemented with either pure silybin or commercial hepatoprotectant containing silybin on nutrient digestibility, general health, immunological parameters (serum cytokines, immunoglobulins, and acute phase protein concentrations), as well as on liver performance in clinically healthy dogs, and 2. to examine the effects of a diet supplemented with commercial hepatoprotectant containing silybin on liver performance in dogs with hepatopathies.

Clinical Observations and Mortality: EXP1
During EXP1, all dogs were in good health. No clinical symptoms or mortality was observed.
Chemical composition and fatty acid pro le of the diet: EXP1 The chemical composition of the diet fed to the dogs in EXP1 is given in Table 1, along with its fatty acid pro le. Body weight and body condition score: EXP1 The treatment affected neither body weight (BW) nor the BCS of the studied dogs (p-value = 0.89). At the end of EXP1, the BW of the CON dogs was about 15.5 kg and was approximately 15.2 kg and 14.7 kg for the HEP and SIL dogs, respectively. The BCS was 5 throughout the whole experiment and did not differ between treatments (Table 2). Apparent Digestibility: EXP1 Nutrient and dry matter (DM) digestibility were not affected by the treatment, but higher ether extract digestibility in the CON group differed from that in the HEP and SIL group (Table 3). Hematology and serum biochemistry: EXP1 SIL and HEP had an effect on hematological parameters (Table 4). WBC was lower in the treatment groups than in the CON group. Monocyte and eosinophil counts were lower in the HEP group than in the CON and SIL groups. Neutrophil count was higher in the CON group than in the HEP and SIL groups. RBC, hemoglobin, and hematocrit were not statistically signi cant, and no signi cant differences were seen in MCV, MCH, MCHC, or PLT between the CON and treatment groups during EXP1. All hematological parameters were within their reference ranges and supplementation did not adversely affect them.
We noted that albumin, alpha-amylase, and AP-alkaline phosphatase were signi cantly lower in the CON group than in the HEP and SIL groups (Table 4). On the other hand, total bilirubin was higher in the HEP group than in the CON group. Finally, the serum activity of GGT was signi cantly higher in the treatment groups than in the CON group.
Considering carbohydrates metabolism, we observed fructosamine to be higher in the HEP group than in the CON group. Further, glucose concentration was higher in CON than in the treatment groups (Table 4).
Additionally, taking into account lipid metabolism, triglyceride concentration was higher in the treatment groups than in the CON group.
Ionograms revealed that Mg 2+ and K + concentrations were higher in the treatment groups than in CON.
The calcium concentration was however lower in the HEP group than in the CON group. Precise results are given in Table 4. Serum fatty acid pro le: EXP1 Supplementation did not affect the serum fatty acid pro le of dogs, with the exception of C20:5 n3, the concentration of which was lower in the treatment groups than in the CON group (Table 5). Serum cytokines, immunoglobulins, and acute phase proteins: EXP1 The supplementation had no signi cant effect on in ammatory proteins, but IL4 concentration was signi cantly higher in the treatment groups than in the CON group (Table 6). IL10 concentration was higher in the SIL group than in the CON group. Urine cortisol to creatinine ratio: EXP1 Supplementation did not affect urine pH, cortisol concentration, or creatinine concentration. Thus, the cortisol:creatinine ratio was not affected either (Table 7). Hematology and serum biochemistry: EXP 2 Supplementation did not affect hematological parameters in the dogs with hepatopathies, and liver markers such as ALT, AST, and GGT signi cantly decreased (Table 8).

Serum microRNA expression: EXP1
We did not observe any differences between the effects of silybin and commercial hepatoprotectant on miRNA expression in healthy dogs (Fig. 1), but we did nd that 28-day supplementation with commercial hepatoprotectant signi cantly decreased the expression level of miR-122 and miR-126. We did not observe any effect of the commercial hepatoprotectant on miR-192 (Fig. 2).

Discussion
Liver disorders are a signi cant group of pathologies in dogs. Patients with hepatopathies are nutritionally demanding and should be provided with high-quality feed supplemented with additives that exhibit liver-bene cial properties and do not interfere with digestion. It is worth mentioning that the presence of antinutritional factors in pet food is important and may lower the digestibility of dogs' diets [18]. Generally, our study (in EXP1) suggests that silybin-the active substance in silymarin-does not interfere with digestion processes. The slightly lower ether extract digestibility in the experimental groups might be due to a slight laxative effect [19] or to the occurrence of subclinical gastroenteritis, which has been reported as a rare adverse effect of silymarin [20]. However, no signs of diarrhoea were observed in the experimental groups. On the other hand, silybin has cholagogic properties, so we might expect lipid digestion to be enhanced [21]. To our knowledge, no experiments have investigated the effects of silybin on nutrient digestibility in dogs. The literature does, however, discuss these effects in other animals: One report suggests that Silybum marianum (L.) as source of silymarin has no effect on nutrient digestibility in buffalos [22], while an experiment with broiler chickens incorporating S. marianum seeds into the diet resulted in increased nutrient digestibility in terms of mycotoxin-contaminated feed [23]. The present study found no clinically signi cant differences in hematological parameters between the groups of dogs. Likewise, Chon and Kim [24] observed no signi cant differences in hematological parameters such as WBC, RBC, MCV, MCH, or MCHC between the control group and the group treated with silybin in the case of giardiasis in dogs. Liver enzymes activity signi cantly decreased in experimental groups (both SIL and HEP), as in other studies with various liver-associated dysfunctions [25][26][27]. Our study shows that, in healthy dogs, liver performance is not negatively affected by silybin supplementation. On the other hand, some reports describe a prophylactic effect of silymarin containing silybin in cats [28] and rats [29]. For example, silymarin protected the liver in healthy cats given acetaminophen. ALT/GPT, AST/GOT, ALP, and LDH did not increase, as happened in cats given acetaminophen alone [30]. It is important to highlight that poisoning with nonsteroidal anti-in ammatory drugs (NSAIDs) happens relatively often in small animal practice, as a result of unauthorised administration by the owner.
Nevertheless, it should be kept in mind that elevated hepatic markers are not always associated with liver injury: they could be a transient effect of the administration of drugs such as phenobarbital in epileptic dogs [31] or glucocorticoids [32]. The literature even describes congenital breed-related causes of elevated liver markers, like benign familial hyperphosphatasemia in Siberian huskies or increased ALP activity in Scottish terriers (where it may be as much as ve times higher than in other breeds) [32]. The supplementation of both HEP and SIL slightly altered the serum fatty acid pro le. These, statistically signi cant, changes were seen in the small amounts of fatty acids physiologically present in serum, and are not clinically relevant. The increase in concentration of C15:0 and C17:0 and the subsequent decrease in C20:5 n-3 may be associated with a moderate interference with lipid metabolism [33], as discussed later. To our knowledge, there are no studies to have investigated changes in the quality of the serum fatty acid pro le in animals supplemented with silybin. In our study, supplementation with HEP and SIL did not affect serum cholesterol concentration, but triglyceride concentration increased. This does not agree with the study of Sun [34], who showed that in a mouse model with nonalcoholic fatty liver syndrome, silybin supplementation signi cantly lowered both serum and hepatic lipid accumulation.
Similar results were obtained by Ramakrishnan in rats [35] with induced hepatocellular carcinoma.
Moreover, the combination of silymarin and n-3 fatty acid supplementation may enhance the antihyperlipidemic effect in rats with metabolic syndrome [36]. These discrepancies with our results are probably due to the fact that different breeds and animals with induced liver disorders were used, rather than the healthy animals in our study (EXP1).
Silymarin, the source of silybin, is believed to protect against renal injury by normalizing the lipid metabolism [37]. Our study suggests that supplementation had no effect on the urine parameters, and thus on the renal function, in healthy dogs (EXP1). Silymarin is eliminated mainly by bile [38], and it does not alter urine pH, which is a signi cant feature in terms of urolith formation and the diagnosis of diseases (such as diabetes mellitus) by urinalysis, as it does not conceal the symptoms. Moreover, since the excretion of some drugs (such as phenobarbital or gentamycin) is related to urine pH [39], the lack of effect of both the commercial preparation and the pure silybin supplementation on urine pH seems to be advantageous.
Silymarin displays anti-in ammatory effects on T-lymphocytes in vitro [40][41][42]. The immunomodulatory properties of oral silymarin (silybin) in vivo in dogs have not previously been described. This study found that neither pure silybin nor commercial hepatoprotectant affected most of the immunological and in ammatory parameters. The supplementation had signi cant effect only on IL4 and IL10 concentration in serum. Immunoregulatory cytokines such as IL4 and IL10 have been described as exerting antiin ammatory properties on various cell types [43][44][45]. The IL4 level was signi cantly higher in the treatment groups than in the CON group, while IL10 concentration was higher in the SIL group than in the CON group. Previous studies involving human subjects have demonstrated a signi cant relationship between greater hepatic in ammation and subsequent brosis progression [46,47]. Thus, control of in ammation may be a useful strategy for ameliorating the sequelae of chronic liver disease. Here we have demonstrated that silybin has the ability to increase anti-in ammatory cytokine concentration in serum and has no effect on the proin ammatory cytokine secretion in vivo, which can be considered a positive effect. Silybin administration in dogs has been well documented as an effective therapeutical tool for different types of liver injuries, such as induced toxaemia, drug poisoning, chemotherapy, and chronic hepatitis [48,[26][27][28]. Several reports also describe its antiviral and antineoplastic properties in laboratory animals or cell cultures [35,49,50]. Similarly, in the current study, silybin supplementation improved liver performance regardless of the underlying hepatic disease. We consequently observed a signi cant decrease in liver enzymatic markers in dogs with liver disorders (EXP2).
In recent years, in addition to traditional markers for liver diseases, such as ALT, AST, ALP and GGT, genetic markers are increasingly used, including changes in the level of microRNA [51][52][53]. Previous studies have shown that miR-122 and miR-126 are highly speci c for the liver in dogs [52,[54][55], while miR-192 is less speci c for the liver in dogs. However, in vitro research on mice has also shown that pathological changes accompanying liver damage may be re ected in miR-192 expression [56]. We therefore decided to examine these three types of miRNA as potential markers for liver metabolism. In EXP1, we investigated whether the administration of SIL and HEP affected the liver metabolism of healthy dogs; since this study showed no negative impact, we decided to investigate the effect of HEP on animals with hepatic disorders. In EXP2, we found that HEP decreased the relative expression of miR-122 after 28 days of supplementation, and also downregulated miR-126; however, in case of miR-126, only a slight trend was observed (H1 vs. H28), and it was not statistically signi cant (P = 0.241).
The usefulness of miRNAs in the diagnosis of liver diseases has been con rmed by data in the literature which show that increased miR-122 expression is noted in almost all liver diseases in dogs, such as acute and chronic hepatitis, hepatocellular carcinoma, lymphoma, and other biliary diseases like extrahepatic bile duct obstruction [51]. These results have also been con rmed by Oosthuyzen et al. [55], who showed that changes in this parameter are not associated with the breed, age, or sex of the dog, and that the number of miR-122 copies increases only during the occurrence of liver disease. They moreover demonstrated a positive correlation between miR-122 and ALT, which is one of the main markers used in the diagnosis of liver diseases [55]. Based on our results in EXP2, and due to the low speci city of changes in miR-122 in relation to various liver diseases (as shown by Dirksen et al. in 2016), we can only conclude that the metabolism of this organ improved [51].
The second type of miRNA investigated in our study was miR-126. Although only minor changes were observed for the HEP group (H1 vs. H28), we decided to study this type of miRNA only because other studies on humans have indicated that miR-126 could also be used as a liver disease marker [56]. This was also indirectly con rmed in dogs by Dirksen et al. (2016) [51], who showed that an increased number of miR-126 copies is typical only of chronic hepatitis, and of other liver diseases, such as hepatocellular adenoma, hepatocellular carcinoma, or acute hepatitis. This may indicate a higher speci city of this marker in the diagnosis of liver disease in dogs. Our results showed only an increased trend in miR-126 (P = 0.053; healthy vs. H1), which along with the biochemical markers may indicate that the dogs were in a transition state from the acute to the chronic phases of hepatitis. It should also be emphasised that, after HEP supplementation, a decrease in the number of miR-126s (H1 vs. H28) was also observed, though this change also lacked statistical signi cance.
Overall, our results in EXP2 showed that liver diseases were accompanied by an increase in miR-122 (H1 vs. H28), while the administration of commercial hepatoprotectant decreases it; this may indicate that treatment with HEP has a positive effect. It should be noted that there is only limited data in the literature on expression changes of miRNAs in the blood during liver diseases in dogs. We thus decided to support our research with further different diagnostic parameters. We also noted a decrease in ALT, AST, and GGT activity after administration of the commercial hepatoprotectant, which also con rms that this supplement improved liver metabolism.
The hepatoprotective properties of silymarin, containing silybin, are mainly associated with its antioxidant, anti brotic, anti-in ammatory and cholagogic effect [14]. Moreover, silymarin accelerates liver regeneration [57]. On a molecular basis, silymarin inhibits lipid peroxidation and synthesis of reactive oxygen species. It has also been found that silymarin interacts with cell and mitochondrial membranes, modifying the ux of substances through them [14]. In a regular small animal practice, it might be challenging to accurately identify the exact liver disorder (following WSAVA guidelines) due to the lack of medical equipment and nancial limitations, so symptomatic treatment with silybin is fully acceptable and reasonable.
An important limitation of our experiments is the fact that we did not study the hormonal pro le of the dogs used for EXP2. Hence leading our study to be considered as a pilot, carried out following the positive result of the tested additives effect obtained during EXP1. Although we have published these results, EXP2 is not a complete study covering all aspects of canine liver disorders. We nonetheless believe that this may point to new directions of research on this issue. We also recognise that the EXP2 results require further research, which we plan to perform.

Conclusion
In conclusion, we have con rmed that, in healthy dogs, supplementation with silybin at 12.75 mg per 10 kg (8.5 mg per 5 kg) BW, or with a commercial hepatoprotectant containing silybin at the same dose, does not interfere with the digestion of nutrients, and subsequently exerts no detrimental effect on liver performance, health, or blood parameters. In dogs with hepatopathies, supplementation with commercial hepatoprotectant containing silybin at a dose of 12.75 mg per 10 kg (8.5 mg per 5 kg) BW resulted in a decrease in the activity of serum liver markers, which was accompanied by a decrease in the concentration of liver-speci c miRNA molecules (mainly miR-122). Liver performance was hence improved. Overall, silybin supplementation has no detrimental impact on healthy dogs and supports liver performance in dogs with hepatopathies.

Animals and experimental design: EXP1 and EXP2
All experimental procedures were performed in accordance with the guidelines of the Local Ethical Committee for Animal Research (Ministry of Science and Higher Education, Poland) as well as in compliance with the ARRIVE guidelines. The study conformed to the 28/2020 statement of Local Ethical Committee in Poznan, Poland. The dog owners gave their informed consent in writing. The research consisted of two consecutive studies. The rst, EXP1, surveyed a group of eighteen healthy laboratory adult beagle dogs (n = 18, nine females and nine males, 2 years old). In EXP1, a 3 × 3 Latin square design with 3 treatments (CON, HEP, SIL) and three periods was used. Each treatment was given to six dogs (three females and three males) in a given period, giving eighteen replicates. A commercial basic diet (Addvena, Poznań, Poland) composed of lamb (including fresh lamb meat 50%), potatoes, peas, beet pulp, animal fat, potato protein, tomato puree, dried alfalfa, axseed, brewer's yeast, salmon oil, sodium phosphate dihydrate, chicory root, glucosamine, and chondroitin sulphate was the control diet (CON). The rst experimental diet (HEP) was commercially available feed (CON) supplemented with commercial hepatoprotectant containing silybin (Hepaxan, Vebiot, Dębica, Poland), while the second diet (SIL) was CON supplemented with pure silybin. The diet for both groups contained silybin, pure or as a preparation, at a dose of 12.75 mg per 10 kg (8.5 mg per 5 kg) body weight. EXP1 was divided into three periods, each lasting 28 days: this consisted of a 23-day adaptation phase (days 1 to 23) and a ve-day total faecal collection phase (days 24 to 28), followed by a 12-day wash-out period. The experiment lasted 108 days (so each dog had three 28-day periods with 12-day wash-out periods between them). Titanium dioxide (TiO 2 ) was included in the diets as a digestion marker at 0.2% of diet. The analysed crude nutrient concentration in the diets and dietary fatty acid pro le are presented in Table 1. Each dog was housed individually in a kennel that enabled social contact among animals, was fed twice a day, and had free access to water. During the adaptation phase, the dogs had access to an outside playground for exercise and socialisation. The maintenance energy requirement (MER) was estimated according to FEDIAF [58] and the diets met the MER of the dogs. Each animal taking part in the experiment was up to date on their vaccination and deworming schedules before beginning.
EXP2 used client-owned dogs (n = 15) referred to the University Centre for Veterinary Medicine at Poznań University of Life Sciences, in which a hepatic disorder was diagnosed. The diagnostic process did not reveal a speci c etiology agent, therefore these cases were considered idiopathic. A pro le of the dogs taking part in EXP2 can be found in Supplementary Table 1. The criteria we established for diagnosing a hepatic disorder were a clinical demonstration of at least one of the symptoms described as most prevalent in dogs with chronic hepatitis [48], including decreased appetite, lethargy/depression, icterus, ascites, PU/PD, vomiting, diarrhoea, or subsequently an increase in at least three out of these four liver markers: alanine aminotransferase (ALT), alkaline phosphatase (ALP), aspartate aminotransferase (AST), and gamma glutamyl transpeptidase (GGT). The exclusion criteria were infectious or parasitic diseases, systemic, neurological or traumatic diseases or general symptoms of food intolerance or allergy in the past. Moreover, individuals with con rmed hepatocarcinoma or other liver-associated cancers were excluded from the study. The dog owners were advised to begin supplementing their pets' diet with commercially available preparation containing silybin (Hepaxan, Vebiot, Dębica, Poland) at the dose recommended by the manufacturer (Supplementary Table 2).

Health status and Body Condition Score (BCS): EXP1 and EXP2
The average weight was 18.1 kg for males and 12.9 kg for females in EXP1, and 28.7 kg for males and 22.1 kg for females in EXP 2. In EXP1, body weight was measured on days 1 and 28 of the experimental period and feed intake was recorded daily. In EXP2, body weight was measured at the beginning (day 1) and at end (day 28) of supplementation with the hepatoprotectant. For all dogs in EXP1, the body condition score (BCS) was assessed throughout the experimental period in line with the recommendations of the World Small Animal Veterinary Association [59]. The dogs in EXP1 and EXP2 underwent weekly check-ups consisting of physical examination, including rectal temperature measurement, mucous membrane inspection, heart and lung auscultation, and stomach palpation (abdominal examination). The dogs were determined to be clinically healthy if the physical examination revealed no pathological ndings.

Blood sample collection: EXP1 and EXP2
Blood samples were collected via cephalic venipuncture as follows: 1) EXP1: on the last day (day 28) of each treatment period at 6.00 AM.
In both EXP1 and EXP2, blood samples were collected in two vacutainer tubes. One of these contained K 3 EDTA anticoagulant and was used for hematological examination; the second tube contained serum separator gel and was used to obtain serum for biochemical and miRNA examination, fatty acid pro les, and serum interleukin, immunoglobulin, and acute phase protein analysis. Blood from the second set of tubes was left at room temperature for blood clot formation and then centrifuged at 3500 rpm for 10 min at 4 °C to obtain serum. The serum samples were transferred to Eppendorf tubes, labeled, sealed, and frozen at -80 ℃ to await analysis.
Hematology and serum biochemistry analysis: EXP1 and EXP2

Urine samples and urinalysis: EXP1
Free catch urine was collected on the last day of each feeding period using a Uripet urine collection device (Rocket Medical, Watford, England). Then 3 ml of urine was stored at -20 °C and analysed for creatinine, cortisol, and pH using VetLab Station (IDEXX Poland) within two weeks of sampling.

Serum miRNA expression: EXP1 and EXP2
The investigation into the effects of the test compounds on the expression of miRNA in blood serum was performed in two experiments: In EXP1, we examined the effect of commercial hepatoprotectant and silybin supplementation on miRNAs expression in healthy dogs (EXP1). In EXP2, we investigated the effect of commercial hepatoprotectant on miRNA expression in dogs with liver disorders. The healthy dogs in EXP1 were used as the control group in EXP2.
Endogenous control was added to the samples during isolation (miRNeasy Serum/Plasma Spike-In Contro; Qiagen, Germany). Following extraction and elution of RNA, the samples were immediately frozen at -80 °C. RNA content and relative purity were determined using the UV-Vis spectrophotometric method with a NanoPhotometer NP80 (Implen, Munich, Germany). Reverse transcription was performed using a miScript II RT kit, following the manufacturer's instructions. The master mix was prepared on cooling blocks and contained 4 µl of 5× HiSpec Buffer, 2 µl of 10× Nucleics Mix, 2 µl of Reverse Transcriptase (RT), and 2 µl of RNase-free water per reaction, giving a total volume of 20 µl. 10 µl of the master mix was added to 10 µl of the total RNA extracted from serum.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and analysed in the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interest.

Funding
This research received no speci c grant from any funding agency in the public, commercial, or not-forpro t sectors.
Author's contributions