Our findings are the first to show that benfotiamine increases glucose oxidation (CO2 formation) while downregulating NOX4 expression in cultured human myotubes. Benfotiamine also increased overall glucose utilization, which was also shown for thiamine. In contrast to glucose oxidation, we found no significant change in oxidation and overall utilization of oleic acid after incubation with benfotiamine.
Hammes et al. reported that there was no change in TCA cycle activity and that activation of transketolase (TK) was responsible for the effects of benfotiamine on cultured endothelial cells (Hammes et al. 2003
). Our contrasting findings suggest that the increase in glucose oxidation that we observed in myotubes does not occur in endothelial cells. Preliminary data in human umbilical vein endothelial cells (HUVEC) support this, as although we observed a marked increase in cell-associated glucose, we did not see the same marked increase in glucose oxidation as we saw in myotubes (data not shown).
It is unclear whether an increase in mitochondrial glucose oxidation, such as we observed, would have an adverse effect on muscle cell function in vivo. In endothelial cells, it has been proposed that the adverse changes observed in endothelial cells exposed to hyperglycemia result from increased glycolytic flux causing an increase in electron donors (NADH and FADH2
) to the electron transport chain and the subsequent generation of superoxide from the mitochondria (Brownlee 2005
). However, other studies have indicated that an increase in pyruvate entering the mitochondria is beneficial as this increases the NAD/NADH ratio and thereby reduces the intracellular redox potential (Tilton et al. 1992
; Van den Enden et al. 1995
). GSEA analysis did not show any effect of benfotiamine on pathways directly involved in glucose metabolism (e.g., glycolysis). However, gene sets involved in peroxisomal lipid oxidation and organelle membrane function (including mitochondrial OXPHOS genes) were upregulated after benfotiamine treatment.
The marked downregulation of NADPH oxidase 4 (NOX4) we observed after incubation with benfotiamine, if also present in other cell types, may be a key mechanism by which this compound exerts its beneficial effects in animal models of diabetes. NOX4 is upregulated in response to hyperglycemia (Xia et al. 2006
), is a primary source of reactive oxygen species (Basuroy et al. 2009
) and is highly expressed in kidney (Gorin et al. 2005
) and vascular endothelial cells (Chen et al. 2008
). Benfotiamine has previously been shown to prevent genomic damage to cells exposed to prooxidants in vitro (Schmid et al. 2008
) and ex vivo (Schupp et al. 2008
). Inhibition of NOX upregulation has been shown to prevent the development of nephropathy in STZ-induced diabetic rat model (Xu et al. 2009
). Notably, we found that benfotiamine also downregulated NOX4 expression under normoglycemic conditions, suggesting that benfotiamine may have effects on the constitutive expression of NOX4 unrelated to ambient plasma glucose concentrations. The significant downregulation of growth factor midkine (MDK) we observed after benfotiamine treatment may also exert a beneficial effect since it has been shown to play a key role in the development of diabetic nephropathy in a STZ-induced diabetic mouse model (Kosugi et al. 2006
Our finding that the serine proteinase inhibitor SERPINB7 was significantly upregulated by benfotiamine under both normo- and hyperglycemic conditions was unexpected. Recent studies have suggested that upregulation of SERPINB7 expression in response to hyperglycemia increases mesangial matrix accumulation and accelerates diabetic nephropathy in mice (Miyata et al. 2002
). This may be secondary to inhibition of plasmin and matrix metalloproteinase activity (Ohtomo et al. 2008
). This finding does not concord with a previous study, showing that benfotiamine actually protects against diabetic nephropathy in animal models (Babaei-Jadidi et al. 2003
). The reasons for this discrepancy are unclear but may be related to differences in myotubes as compared to mesangial cells where SERPINB7 is predominantly expressed.
In addition to increasing TK activity, it has previously been found that benfotiamine increases the expression of TK in renal glomeruli (Loew 1996
). However, we did not find any change in the gene expression of TK in human myotubes using microarray. In addition, we did not find any significant changes in genes encoding enzymes involved in glycolysis and/or the TCA cycle that might suggest that the increased oxidation we observed may be secondary to changes in enzyme activity rather than gene expression levels.
Using benfotiamine, in vitro differs from in vivo since most of the benfotiamine given orally is partially dephosporylated in the gut to benzoylthiamine. Due to the enhanced membrane permeability of benzoylthiamine, oral benfotiamine leads to higher plasma and intracellular concentrations of the active metabolite, thiamine diphosphate (TPP), than thiamine (Loew 1996
). Although intact benfotiamine is also delivered to cells via the reduced folate carrier-1 (RFC-1) which is then de-benzoylated to thiamine monophosphate (TMP) by cellular and plasma esterases (Thornalley and Babaei-Jadidi 2005
), it is not superior to thiamine in increasing the intracellular concentrations of thiamine in vitro (Volvert et al. 2008
). The exposure of cells in vivo to intact benfotiamine is minimal (peak plasma concentration approx. 10 nM in subjects taking 250 mg benfotiamine/day) as compared to the supra-pharmacological concentrations used in this study (100–200 μM; Ziems et al. 2000
). Although doses as high as 900 mg benfotiamine per day have been used clinically (Alkhalaf et al. 2010
), we must take into account the possibility that the effects we observed were associated with the high concentrations used, which might result in the generation of esterase-derived thiols at concentrations not observed in vivo. It is also possible that benfotiamine, but not thiamine, induces spontaneous rearrangement of S-acylcysteine derivatives to N-acylcysteine derivatives that affects pyrimidine synthesis and related signaling (Edwards et al. 1996
Interestingly, we also found an increase in glucose oxidation in cells exposed to normoglycemia in the presence of benfotiamine. The effects we observed are therefore not simply a result of benfotiamine abolishing the previously documented inhibitory effects of hyperglycemia on glucose uptake in myotubes (Aas et al. 2004
). However, we also found that 200 μM benfotiamine was able to reverse the inhibitory effect of hyperglycemia on glucose oxidation. We have previously shown that chronic hyperglycemia reduces insulin-stimulated glucose uptake and increases triacylglycerol formation from labeled glucose (Aas et al. 2004
). We assessed the effects of benfotiamine on overall glucose uptake under normoglycemic conditions both with and without insulin and only found a small (17%) but significant increase in benfotiamine. Although we cannot say whether this increased uptake preceded or was a result of the increase in glucose oxidation, this increase in uptake was relatively minor compared with the more pronounced increase in glucose oxidation (35–70%). This would suggest that the effects of benfotiamine on glucose oxidation cannot be explained by an increased glucose uptake itself.
The decrease we noted in fractional oleic acid oxidation may reflect the inverse reciprocal relationship between glucose and lipid oxidation. This finding also makes it unlikely that the increase we observed in glucose oxidation is a result of a benfotiamine-induced change in myotube differentiation. Additionally, we found no changes in muscle differentiation markers such as myogenin and myoD. However, our finding that benfotiamine decreased fractional oleic acid oxidation while increasing fractional glucose oxidation suggests that benfotiamine may stimulate a preferential oxidation of glucose at the expense of fatty acids. This may be secondary to increased generation of pyruvate from glycolysis. GSEA analysis also suggested that fatty acids could be channeled to peroxisomes for oxidation after treatment with benfotiamine.
Most previous research into the biochemical effects of benfotiamine has focused on cells, such as endothelial cells, susceptible to hyperglycemic-associated damage (Balakumar et al. 2010
). The clinical implications of the results from the present study are that oral benfotiamine has marked metabolic effects on cells i.e., skeletal muscle, other than those typically associated with diabetic microvascular complications. Whether our results can be extrapolated to the in vivo situation is not clear, as the marked effects we observed were achieved using supra-physiologic concentrations of benfotiamine. If such effects do occur in vivo, our results suggest that benfotiamine may not only be beneficial for microvascular complications associated with type 1 diabetes but could also be of relevance to type 2 diabetes and obesity where reversal of reduced mitochondrial capacity and decreased glucose oxidation is a desirable goal (Aas et al. 2011
In conclusion, our study has shown that benfotiamine, but not thiamine, causes a marked increase in mitochondrial glucose oxidation in human skeletal muscle cells. Our finding that benfotiamine also reduces NOX4 gene expression offers a new insight into the mechanism/s by which this substance exerts its effects on the development of diabetic complications. These findings suggest that benfotiamine has a wider range of effects in different cell types of relevance to both type 1 and type 2 diabetes than has previously been recognized.