The present study shows that green tea powder (GT) alone, or in combination with L. plantarum DSM 15313 (Lp), exerts beneficial metabolic effects in C57BL/6J mice fed an HFD. These results are in agreement with and an extension of previous studies of metabolic effects of green tea in rodents. However, in contrast to most previous studies, in which the green tea extract or the phenolic constituent epigallocatechin gallate (EGCG) has been evaluated, we have here studied the effect of whole tea leaves milled to a powder, which is suggested to contain higher quantities of polyphenols compared to a water extract of green tea . Indeed, our analysis of the green tea powder showed a higher amount of catechins compared to a water extract from the same green tea powder. The present study additionally shows that when a tannase active strain of L. plantarum was supplemented to the GT-diet (Lp + GT), both the load of lactobacilli and the bacterial diversity increased significantly in the small intestine. It seems likely that the increase in diversity was due to effects exerted by the given Lactobacillus strain in combination with GT rather than just the addition of one more species to the microbiota.
The observed reduction in adiposity after green tea intake is in agreement with previous studies, including studies in HFD-fed C57BL/6J mice. Increased faecal lipid excretion is likely to be a major factor contributing to the reduced adiposity. It is known that green tea extracts possess the ability to inhibit intestinal lipases . In addition to lipases, green tea catechins have been shown to inhibit glucose transporters and enzymes involved in carbohydrate digestion, raising the possibility that also carbohydrate absorption could be reduced [47, 48]. EGCG has previously been shown to slightly increase energy content of faeces as a consequence of reduced digestibility . There were indications that components of the microbiota might interfere with lipid metabolism. The amount of Akkermansia in the small intestine was shown to correlate negatively with the total body fat content, the periovarian fat depots as well as the circulating levels of leptin, suggesting a role for Akkermansia in reducing fat accumulation. Interestingly, Akkermansia muciniphila was recently shown to be increased in prebiotic-treated ob/ob mice which had lower fat mass compared to control ob/o b mice . In a study of pregnant women, the number of Akkermansia and Bifidobacterium was higher in women with normal weight gain compared to those with excessive weight gain . Akkermansia has also been shown to be increased in normal weight and post-gastric-bypass individuals compared to obese . In the present study, a negative correlation was also detected between the total number of bacteria and periovarian fat mass. However, it remains to be proven that the correlations point at the cause of the alterations in lipid metabolism.
In accordance with the reduced adiposity after GT intake, GT reduced hepatic lipid content and circulating levels of ALT suggesting that GT either directly, or indirectly, perhaps via an altered microbiota, ameliorates the liver damage imposed by the HFD. The involvement of an altered microbiota is supported by the negative correlation between the amount of Akkermansia and the hepatic TAG content. Ma et al. demonstrated that the probiotic mixture of different Lactobacillus and Bifidobacterium strains (labelled VSL#3) reduced a HFD induced hepatic steatosis in C57BL6 mice . On the other hand Velayudham et al. demonstrated that the same probiotic mixture did not prevent liver steatosis but modulated the progression to liver fibrosis . Here, the supplement with one strain of L. plantarum did not have the capacity to reduce the hepatic TAG accumulation. SREBP-1c is a key transcription factor in hepatic lipogenesis and is partly regulated by PPARγ and LXR . The down-regulation of the mRNA expression of SREBP-1c, PPARγ and LXR suggests that decreased hepatic lipogenesis contributes to the reduced hepatic lipid content observed in the mice fed GT. Also, the hepatic fatty acid transporter CD36, another lipogenic target of PPARγ and LXR  was shown to be down-regulated in the GT groups compared to control at 11 weeks, however, the decrease compared to control was abolished after 22 weeks. In addition, the hepatic mRNA expression of the lipogenic enzymes ACC and FAS was reduced and trended towards a diminuition, respectively, in mice fed GT alone compared to control mice, indicating a decreased de novo lipogenesis in these mice. Surprisingly, the decreased expression of the lipogenic enzymes could not be detected in mice from the Lp + GT group. The expression of the transcription factor PPARα was down-regulated in mice in the GT group compared to control mice, indicating a decreased hepatic fatty acid oxidation as well. The mechanisms behind this finding need further elucidation. It should be emphasized that this is a large screening of the alterations induced by the dietary supplements and more detailed analyses, including expression at the protein level, need to be undertaken.
The choice of the present strain of probiotics is partly based upon its ability to increase the barrier effect of the gut-mucosa  but mainly on its ability to degrade polyphenols as tannins, and produce compounds as substituted phenyl propionic acids, phenyl valeric acids and benzoic acid derivates . These compounds are more easily absorbed than longer molecular chains of phenolics, and are often also more bioactive [55, 56]. They can have anti-inflammatory effects as well as antimicrobial effects. One hypothesis is that Lp in the large intestine is able to convert polyphenols in GT to more easily absorbed metabolites with antioxidative effects in organs such as the liver. Generally, polyphenols possess powerful antimicrobial activities  which can have growth suppressing effects on many bacteria but which are better endured by others.
Feeding mice with Lp alone did not affect inflammatory markers. A reduced spleen weight was observed already after 11 weeks in the Lp + GT group and Lp + GT significantly reduced the inflammatory tone at 22 weeks, both systemically, as indicated by reduced circulating PAI-1 and decreased spleen weight , as well as locally in the liver, as shown by decreased mRNA expression of TNF-α and MCP-1. This, together with the higher number of Lactobacillus both in caecum and small intestine in the Lp + GT group support the hypothesis of Lactobacillus converting polyphenols to more active anti-inflammatory components. The higher number of lactobacilli in the Lp group and higher diversity in the GT group in the ceacum seemed not to affect the inflammatory markers at 11 weeks. The effect seen after 22 weeks might partly be explained by an increased number of bacterial taxa over time in all groups since the number of T-RFs increased at 22 weeks while no difference in diversity between the groups was seen in ceacum.
It has previously been shown that a probiotic-supplemented diet decreases levels or expression of liver TNF-α in animal models. Ma et al.  found a decreased expression of TNF-α in the liver of HFD-fed C57BL/6 mice when the diet was supplemented with the probiotic VSL#3 mixture. Furthermore, in an acute liver injury model in rats, TNF-α levels decreased in the liver when rats had been pre-treated with L. plantarum DSM 15313 before inducing liver injury . Also, green tea extract alone has been shown to reduce hepatic mRNA levels of both TNF-α and MCP-1 as well as NFκB binding activity . Additionally, a negative correlation between Akkermanisia and plasma PAI-1 was seen, indicating that Akkermansia may have a beneficial influence on the inflammatory state. It has been shown in germ-free mice inoculated with Akkermansia municiphila MucT that the colonization altered the mucosal gene expression towards a profile involved in immune responses and cell fate, which led to the conclusion that the tested strain of Akkermansia modulated pathways involved in establishing homeostasis for basal metabolism and immune tolerance towards commensal bacteria .
Dietary administration of GT reduced plasma cholesterol as well as hepatic cholesterol content. The supplementation with Lp alone had a significant cholesterol-lowering effect in the liver after 22 weeks and additionally the total number of bacteria in the small intestine was negatively correlated with both liver and plasma cholesterol. An increased faecal excretion of cholesterol was observed in the GT groups, indicating that this is at least one of the mechanisms underlying the cholesterol-lowering effect of GT administration. EGCG as well as other polyphenols have previously been shown to inhibit cholesterol absorption in rodents [61, 62]. The underlying mechanisms are not fully elucidated but green tea catechins have been suggested to reduce the absorption of cholesterol from the intestine by reducing the solubility of cholesterol in mixed micelles . SREBP2, the major regulator in cholesterol biosynthesis, and its downstream target HMG-CoA reductase, were up-regulated when GT was combined with Lp. An explanation for the up-regulated cholesterol biosynthesis might be a response to the very efficient GT-induced cholesterol excretion in an attempt of the system to restore cholesterol homeostasis. This rescue mechanism of cholesterol is further supported by the upregulation of the hepatic HDL receptor SR-B1 mRNA in the GT groups.
The lower fasting plasma glucose and insulin, resulting in a lower HOMA index, as well as the lower levels of fructosamine, mirroring the blood glucose concentration over several weeks, indicate a more insulin sensitive state in the mice fed GT. In contrast, the oral glucose tolerance test showed deteriorated glucose elimination despite increased insulin secretion in the GT groups, implying decreased oral glucose tolerance compared to the control group and the Lp group. The cause for this discrepancy is not known.
A component of the microbiota that seems to be of relevance for several of the metabolic effects studied here is Akkermansia. Akkermansia muciniphila is a newly described species which has been shown to be an efficient mucin degrader found in the intestines of humans and animals  and it has been associated with healthy gut mucosa [64, 65]. However, the amount of Akkermansia in the small intestine did not differ between the groups indicating that it is not the diet per se but instead the response of the specific microbiota in an individual that might be of importance. It is clear from the present results that the dietary supplements GT and Lp exercise effects on the composition of the microbiota in both the small intestine and in caecum as well as on metabolism. However, a key-question that remains to be answered is whether the changes in gut-microbiota affect the metabolic markers or whether the changes in the gut-microbiota result from metabolic alterations. With the assumption that certain components of the microbiota exert metabolic effects, it is clear from the present results that the microbiota of the individual mice varied widely in spite of the fact that they are from an inbred strain. Especially the individuals of the control group differ while the dietary supplements appear to standardize the microbiota to some degree. Surprisingly, the standardization differed between the small intestine and caecum, i.e. green tea made the microbiota of the individual mice more uniform in the former while Lp did the same in the latter. The multivariate analysis revealed that certain, relatively few components of the microbiota (T-RFs) in both the small intestine and in caecum, had a considerably higher effect on the correlation model built for comparing T-RFLP-data with metabolic test-parameters, some showing positive correlations and some negative. It was also clear that GT and Lp in some cases affected the abundance of these bacterial components differently. One of these critical components (T-RF) could be identified as Akkermansia. However, the fact that several correlations have been found between Akkermansia and metabolic parameters do not necessarily implicate a causal role for this taxum but at least reflects that markers for inflammation and lipid metabolism are linked to the microbiota, and especially the microbiota of the small intestine. Mice in the Lp + GT group had higher bacterial diversity in the small intestine compared to both control and the GT group. The higher diversity may as such be a positive health factor or at least a health marker. It was shown in humans that the bacterial diversity of the gut microbiota was higher in lean individuals than in obese ones  and that neonates with low diversity at one week of age had increased risk for developing atopic eczema at 18 months of age . As expected the number of T-RFs were smaller in the small intestine compared to caecum as the sampling were done close to the pylorus. In humans, the diversity in the jejunum has been shown to be lower than in the colon . The T-RFLP method was chosen here even if it has a lower resolution compared to high through-put sequencing, but it gives comparable results to pyrotag sequencing regarding diversity measures .