Hypoglycemic and hypolipidemic effect of S-allyl-cysteine sulfoxide (alliin) in DIO mice

Alliin (S-allyl cysteine sulfoxide) is a bioactive sulfoxide compound derived from garlic. To evaluate the preventive effect of alliin against metabolic risk factors in diet induced obese (DIO) mice, we treated the C57BL/6J DIO mice with drinking water with or without alliin (0.1 mg/ml) for 8 weeks. Results showed that alliin had no significant effect on the body weight, adiposity or energy balance. However, alliin treatment enhanced glucose homeostasis, increased insulin sensitivity and improved the lipid profile in the DIO mice. This was, at least partly, attributable to alliin induced modulation of the intestinal microbiota composition, typically decreased Lachnospiraceae and increased Ruminococcaceae. From above, we conclude that alliin has nutraceutical or even medicinal potential in prevention of diabetes and lipid metabolic disorders.

. Alliin improves glucose tolerance and insulin sensitivity in DIO mice. (a and b) Alliin significantly reduced glucose level challenged with insulin in DIO mice. (c) Alliin significantly improved insulin sensitivity to glucose injection in DIO mice. Data are means ± SEM (n = 10). *P < 0.05, **P < 0.01. GTT, glucose tolerance test; ITT, insulin tolerance test. Alliin and energy balance of DIO mice. The alliin treated mice consumed 4.35 ± 0.44 ml water (equivalent to approximately 0.44 mg alliin) the day before the whole-body energy metabolism of the DIO mice was examined. Results showed that the physical activity, respiratory exchange rate, energy intake or total energy expenditure was not changed by alliin treatment in DIO mice (Fig. 4a-c). Similarly, alliin did not change the thermography, the body core temperature or the associated activity of brown adipose tissue in DIO mice ( Fig. 4d-g).
Alliin modifies intestinal microbiota of DIO mice. Alliin modified the intestinal microbiota, which was significantly different between the alliin treated mice and the control (ANOSIM = 0.0008) (Fig. 5a). Specifically, alliin increased the dominant bacteria of Actinobacteria and Firmicutes but decreased Bacteroidetes and Proteobacteria at phylum, although the deference failed to reach the significant level (Fig. 5b). We selected two groups of bacteria with significant changes, of which the averaged reads were over 100 and their percentage was over 1% in their annotated phylum (Fig. 5c). The other changes with all annotated bacteria were plotted in Fig. 5d.

Discussion
The beneficial effect of garlic or derived products has been widely reported, especially for those garlic commercial products publicized with garlicin. As a result, many suggest chop or crush garlic before consumption which can maximize the conversion of alliin into allicin. Indeed there are sufficient evidence showing the positive effect of allicin on metabolic disorders 7,9,[19][20][21][22] . In fact the main bioactive ingredient of both fresh or dried garlic is alliin, the precursor of allicin 3 . Many medicinal effect of garlic products are probably, or at least partly, attributable to alliin, whereas the extent that alliin contributes to the beneficial function of garlic is not quite clear. Systematic studying of the nutraceutical or medicinal effect of alliin is lacking. Hereby our research systematically investigated the effect of the garlic derived compound alliin on the metabolic factors including body weight, glucose metabolism, adiposity, energy balance and associated metabolic risk factors.
We failed to observe any significant effect of alliin on the body weight or energy balance in the DIO mice. However, alliin was still a strong bioactive compound as it effectively reduced fasting glucose level, increased insulin sensitivity and improved the lipid profile of the DIO mice. As these symptoms were the key features of metabolic diseases, our result suggested that alliin had optimistic potential in nutraceutical or medicinal use. This was consistent with the previous report that alliin was comparatively effective as the anti-diabetic drugs such as glibenclamide and glyclazide 10,11 .
As reported we found alliin had a remarkable lipid lowing effect since it significantly decreased serum triglycerides and free fatty acids 13,23,24 . The different thing was the animal model we used were obese mice generated with high fat diet that imitated the food so readily available in society. Additionally, we found alliin increased serum HDL which is often referred as a type of "good" cholesterol as they can remove fat molecules. Previous reports showed that administration of garlic derived alliin at the dose of 200 mg/kg or 500 mg/kg body weight significantly decreased serum lipids and serum enzymes like liver glucose-6-phosphatase, intestinal HMG CoA reductase and liver hexokinase in addition to its hypoglycemic effect 23,25 . We obtained the similar effect in mice with free access to water containing a low concentration of alliin, under which condition the gavage associated stress was avoided as weight gain was thought to be facilitated by increased glucocorticoids 26 .
To ensure that the beneficial function of alliin was not caused by any toxic effect, we examined the organs as well as several serum biochemical factors. No apparent pathological change was observed in any organ examined in the DIO mice. Interestingly, we noticed that alliin might have the ability to protect against the liver damage evidenced by that the serum level of several markers like AST, ALP, TP and ALB were decreased within normal range by alliin treatment, which phenomenon was also found in alliin treated diabetic rats 11,25 .
Alliin contributed its hypoglycemic effect with no clear mechanism. Its ability of inhibiting glycolysis 23 , or stimulating insulin secretion 10 may contribute to its hypoglycemic effect, while the nitric oxide was also involved in its regulation of glucose and lipid 27 . Although diallyl disulfide is one of the metabolites of allicin, it is not able of managing glucose metabolism alone 28 .
Another factor that is closely linked with metabolic diseases is the intestinal microbiota structure [29][30][31] . To examine this possible mechanism, we further analyzed the microbiota of the alliin treated DIO mice. We found several changes in intestinal bacterial composition after alliin treatment. Typically, alliin decreased Lachnospiraceae and increased Ruminococcaceae. Both Lachnospiraceae and Ruminococcaceae are the main gut bacteria family of mammal microbiota that have a high number of genes equipped to degrade a wide variety of polysaccharides including cellulose and hemicellulose components 32 . Lachnospiraceae is closely associated with insulin signaling in response to nutrient availability via the mammalian target of rapamycin (mTOR) and contributes to the development of diabetes 33 . This indicated that alliin regulated the glucose metabolism possibly via reducing the composition of Lachnospiraceae in the gut. On the other hand, the evidence associating Ruminococcaceae with metabolic disease is lacking. Ruminococcaceae is maybe involved in lipid metabolism as the high fat diet can decrease this family of bacteria 34 . Alliin that tended to increase Ruminococcaceae seemed to inhibit the negative effect caused by high fat foods. More evidence is still required to better understand the role of Ruminococcaceae in metabolic risks.
By systematically studying the metabolic factors in DIO mice model, we confirmed the hypoglycemic and hypolipidemic effect of alliin, the main bioactive compound in garlic. The beneficial function was at least partly attributed by modulation of the intestinal microbiota structure. As a relatively stable bioactive compound, alliin has high potential in use as nutraceutical and medicinal agents for hyperglycemia and hyperlipidemia. Glucose tolerance test and insulin tolerance test. Glucose tolerance test (GTT) and insulin tolerance test (ITT) were conducted at the 6th week. GTT procedure was reported previously 35 . ITT was carried out as follows: Mice were fasted for 4 h (9:00 A.M. to 1:00 P.M.) and insulin (1 U/kg Humulin R; Novo Nordisk) was administered intraperitoneally. The blood glucose levels were measured immediately before and after 15, 30, 45, and 60 min of insulin injection with an Accu-Chek glucose monitor (Roche Diagnostics, Indianapolis, IN, USA). Histological examination and hematoxylin-eosin staining. Tissues fixed with 4% paraformaldehyde were sectioned after embedded in paraffin. Multiple sections were prepared for hematoxylin-eosin staining, photographed (×20) and analyzed by a light microscope (DS-RI1; Nikon).
Metabolic rate and physical activity. Respiratory exchange rate (RER) and physical activity were determined as previously described 35 . Briefly, the mice (n = 6) were acclimated to the TSE labmaster for approximately 24 h with dietary intake and drinks recorded, and then Vo 2 and Vco 2 were measured during the next 24 h. Voluntary activity of each mouse was measured with an optical beam technique (Opto-M3, Columbus Instruments, Columbus, OH, USA) over 24 h. Digested energy was analyzed as described previously 36 .
Body composition measurement. The total fat mass, lean mass of free water of mice after 7 weeks treatment with high fat diet co-administered with either vehicle or alliin were assessed with the Small Animal Body Composition Analysis and Imaging System (MesoQMR23-060H-I; Nuimag Corp., Shanghai, China), according to the manufacturer's instructions.
Cold-induced thermogenesis. A cold tolerance test was performed in 13-week-old mice. The mice were placed in a cold chamber (4 °C) for up to 4 h with free access to food and water. Whole body temperature was monitored using infrared thermography, while core body temperature was measured with a rectal probe connected to a digital thermometer (Yellow Spring Instruments, Yellow Springs, OH, USA). Intestinal microflora analysis. Feces from mice were collected after alliin treatment and stored at −80 °C until use. Seven samples from each group were used for the intestinal microbiota analysis. Microbial genomic DNA was extracted from each fecal sample (0.1 g) using the method as previously described 37 . The V3 + V4 region of the 16S rRNA was amplified by PCR and sequenced using a HiSeq platform (Illumina, San Diego, CA, USA) at Novogene Bioinformatics Institute (Beijing, China). Sequence analyses were performed using Uparse software (Uparse v7.0.1001) 38 . Sequences with ≥97% similarity were assigned to the same operational taxonomic unit (OTU). Taxonomic annotation was conducted using a RDP classifier (Version 2.2) 39 . Nonmetric multidimensional scaling (NMDS) plots and ANOSIM were applied in analyzing the variation between each group utilizing the PAST version 2.17 software program. Community structure variance analysis was conducted with LEfSe software using the default parameters (http://huttenhower.sph.harvard.edu/galaxy/) 40 .
Statistical analysis. Student t test was used to evaluate the differences between groups using IBM SPSS Statistics 20.0 software (IBM SPSS Inc., USA). Significant differences were defined as p < 0.05. Data availability. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.