Biguanide-induced mitochondrial dysfunction yields increased lactate production and cytotoxicity of aerobically-poised HepG2 cells and human hepatocytes in vitro

Abstract

As a class, the biguanides induce lactic acidosis, a hallmark of mitochondrial impairment. To assess potential mitochondrial impairment, we evaluated the effects of metformin, buformin and phenformin on: 1) viability of HepG2 cells grown in galactose, 2) respiration by isolated mitochondria, 3) metabolic poise of HepG2 and primary human hepatocytes, 4) activities of immunocaptured respiratory complexes, and 5) mitochondrial membrane potential and redox status in primary human hepatocytes. Phenformin was the most cytotoxic of the three with buformin showing moderate toxicity, and metformin toxicity only at mM concentrations. Importantly, HepG2 cells grown in galactose are markedly more susceptible to biguanide toxicity compared to cells grown in glucose, indicating mitochondrial toxicity as a primary mode of action. The same rank order of potency was observed for isolated mitochondrial respiration where preincubation (40 min) exacerbated respiratory impairment, and was required to reveal inhibition by metformin, suggesting intramitochondrial bio-accumulation. Metabolic profiling of intact cells corroborated respiratory inhibition, but also revealed compensatory increases in lactate production from accelerated glycolysis. High (mM) concentrations of the drugs were needed to inhibit immunocaptured respiratory complexes, supporting the contention that bioaccumulation is involved. The same rank order was found when monitoring mitochondrial membrane potential, ROS production, and glutathione levels in primary human hepatocytes. In toto, these data indicate that biguanide-induced lactic acidosis can be attributed to acceleration of glycolysis in response to mitochondrial impairment. Indeed, the desired clinical outcome, viz., decreased blood glucose, could be due to increased glucose uptake and glycolytic flux in response to drug-induced mitochondrial dysfunction.

Introduction

Extracts of members of the legume family have long been part of the pharmacopeia and used to treat diabetes mellitus. In this century, several guanine analogs in extracts of Galega officinalis were identified as active agents, and several biguanides were developed as therapeutics for diabetes, notably phenformin and buformin. Both drugs improve glycemic control by improving insulin sensitivity, but both were withdrawn from the market because of unacceptably high incidence of potentially-fatal lactic acidosis. Metformin was developed in the 1950s, and is associated with real, but substantially less, risk of lactic acidosis (reviewed by Leverve et al., 2003, Goodarzi and Bryer-Ash, 2005).

The systemic response to the biguanides is pleotropic, and all three analogs decrease intestinal glucose absorption and hepatic glucose efflux, while augmenting peripheral insulin responses. Not surprisingly, the mechanism of action is also pleotropic, and numerous responses have been variously reported for each drug, including inhibition of mitochondrial respiration, diminution of oxidative and nitrative stress, as well as repression of mitochondrial permeability transition that is cytoprotective against several stressors such as t-butyl-hydroperoxide. More recently, all three drugs have been shown to activate AMP-activated protein kinase (AMPK). Such activation increases oxidative phosphorylation (OXPHOS), glycolysis, and mitochondrial fatty acid beta-oxidation, all of which are thought to contribute to mechanism of action for the buguanides (reviewed by Leverve et al., 2003).

It is generally accepted that all three drugs impair respiration via inhibition of Complex I (Wang et al., 2003). Such drug-induced mitochondrial impairment can result in a compensatory acceleration of glycolysis to compensate for reduced ATP production via OXPHOS. Indeed, preservation of ATP levels despite phenformin exposure has been reported (reviewed by Owen et al., 2000). Moreover, the biguanides are taken up into some tissues, notably liver, via the organic cations transporter 1 (OCT-1) (Wang et al., 2003). Moreover, because the guanidinium group would be positively charged at physiological pH, the biguanides are also bio-accumulated 100-fold into mitochondria as a function of the Nernst equation via an uptake mechanism that can be saturated, and that induces respiratory inhibition (Davidoff, 1968, Davidoff, 1971). As a result, depending on the bioenergetic status of the cell, the physiochemical characteristics of the molecule, and plasma membrane transporters, some molecules can be concentrated in the mitochondrial matrix more than 10,000-fold over the plasma (Ross et al., 2008).

In this light, we reasoned that the lactic acidosis characteristic of the biguanides could be a direct reflection of accelerated glycolytic flux, which correspondingly requires increased glucose uptake and which yields lactate. In the face of drug-induced mitochondrial impairment, this lactate effluxes into the circulation rather than being more completely oxidized. Indeed, the potency of inducing lactic acidosis is in direct proportion with potency as Complex I inhibitors (Wang et al., 2003).

To explore this hypothesis, we evaluated all three drugs in a battery of assays specifically developed to reveal drug-induced mitochondrial impairment (Dykens and Will, 2007, Dykens et al., 2007). These include cytotoxicity of cells grown in galactose to render their metabolism more like normal circumstances in the body where metabolism is generally heavily dependent on OXPHOS, function of isolated mitochondria, simultaneous monitoring of intact cell oxygen consumption and media acidification, and inhibition of immunocaptured respiratory complexes.