Seizure suppression via glycolysis inhibition with 2-deoxy-D-glucose (2DG)


Address correspondence to Carl E. Stafstrom, M.D., Ph.D., Department of Neurology, University of Wisconsin, H6/528 CSC, 600 Highland Avenue, Madison, WI 53792, U.S.A. E-mail:


Metabolic regulation of neuronal excitability is increasingly recognized as a factor in seizure pathogenesis and control. Inhibiting or bypassing glycolysis may be one way through which the ketogenic diet provides an anticonvulsant effect. 2-deoxy-D-glucose (2DG), a nonmetabolizable glucose analog that partially inhibits glycolysis, was tested in several acute and chronic seizure models. Acutely, 2DG decreases the frequency of high-K+-, bicuculline- and 4-aminopyridine-induced interictal bursts in the CA3 region of hippocampal slices; 2DG also exerts anticonvulsant effects in vivo against perforant path kindling in rats. Chronically, 2DG has novel antiepileptic effects by retarding the progression of kindled seizures. Finally, 2DG has a favorable preliminary toxicity profile. These factors support the possibility that 2DG or other modifiers of glycolysis can be used as novel treatments for epilepsy.

Conventional anticonvulsant medications reduce neuronal excitability through effects on ionic channels or synaptic function. In recent years, it has become clear that metabolic factors also play a role in the modulation of neuronal excitability. For example, the high-fat, low-carbohydrate ketogenic diet (KD) is effective in controlling seizures in many children whose seizures are refractory to anticonvulsant medications. The mechanism of action of the KD remains unclear (Bough & Rho, 2007). The possibilities include reduction of excitability by an effect of ketone bodies or fatty acids, mitochondrial modulation of excitability, or some other unknown factor. The key observation that ingestion of small amounts of carbohydrate by children on the KD results in loss of seizure control (Huttenlocher, 1976) led to the concept that carbohydrate restriction could exert a protective effect against seizures (Greene et al., 2003). By providing ketone bodies as the proximate energy source, the KD bypasses glycolysis (Fig. 1). Imaging studies have demonstrated that glucose metabolism is increased before and during seizure activity (Pfund et al., 2000).

Figure 1.

Schematic of glucose metabolism and points at which compounds might affect neuronal excitability. Glucose can be diverted to the pentose phosphate shunt (PPP) via fructose-1,6-diphosphate (FDP). 2-deoxy-2-glucose (2DG) inhibits glycolysis by blocking the phosphoglucose isomerase step (see text). The ketogenic diet (KD), via ketone bodies, bypasses glycolysis by providing acetyl-CoA to the TCA (tricyclic acid cycle) after glycolysis. β-OHB, beta-hydroxybutyrate; AcAc, acetoacetate; PDH, pyruvate dehydrogenase.

Using animal models, we evaluated the hypothesis that carbohydrate restriction by glycolysis inhibition exerts a protective effect against seizures. We tested the effects of the glycolytic inhibitor 2-deoxy-D-glucose (2DG) in several models of acute seizures in vivo and in vitro and on kindling epileptogenesis. 2DG is a nonmetabolizable glucose analog, differing only from glucose by removal of a single oxygen atom from the 2-position. 2DG is taken up by cells but cannot participate in glycolysis beyond glucose-6-phosphate since it inhibits phosphoglucose isomerase. 2DG is known to exert neuroprotective actions in some models (Lee et al., 1999; Rejdak et al., 2001).

Anticonvulsant Actions of 2DG

When exposed to elevated extracellular potassium (7.5 mM), hippocampal CA3 neurons in slices from adult rats develop high-frequency interictal (epileptiform) bursts at a frequency of about 30 per minute. Addition of 2DG (10 mM) to the bathing medium reversibly reduces the burst frequency by about one-half (Stafstrom, Ockuly, Murphee, Valley, Roopra, Stables, and Sutula, unpubl. ms.). Similarly, epileptiform bursts induced by bath application of bicuculline or 4-aminopyridine are reduced by 2DG. In slices from juvenile animals (P10–13), 7.5 mM K+ induces prolonged ictal discharges in area CA3 for 10–30 s; 2DG decreases the occurrence of these ictal bursts. Therefore, 2DG has an anticonvulsant effect on both interictal and ictal epileptiform activity in the hippocampal slice.

An anticonvulsant effect was also seen in in vivo seizure models. In the perforant path kindling model, 2DG was administered at a dose of 250 mg/kg intraperitoneally (i.p.) 30 min prior to each kindling stimulation. With successive after-discharges (AD), the mean AD threshold increases, consistent with an elevation in kindled seizure threshold over time.

In addition, an anticonvulsant effect of 2DG is seen in two other acute seizure models in whole animals (Stafstrom, Ockuly, Murphee, Valley, Roopra, Stables, and Sutula, unpubl. ms.). In the 6-Hz stimulation model (Barton et al., 2001), seizures are induced by corneal stimulation using electrical pulses at a frequency of 6 Hz. Pretreatment with 2DG resulted in seizure protection in 75% of rats tested. In audiogenic seizures in Fring's mice, 50% of mice were protected from sound-induced seizures after pretreatment with 2DG. However, 2DG did not protect against pentylenetetrazole (PTZ) or maximal electroshock (MES) seizures in rats. Therefore, anticonvulsant effects of 2DG were demonstrated in several but not all in vivo seizure models, suggesting a novel profile of anticonvulsant effectiveness.

Antiepileptic Actions of 2DG

In both perforant path and olfactory bulb kindling, antiepileptic effects of 2DG on kindling progression were assessed by the number of ADs required to elicit specific stages of seizure severity. Using the Racine scoring system (Racine, 1972), pretreatment with 2DG at a dose of 250 mg/kg imp., 30 min before each kindling stimulation, significantly increased the number of ADs required to achieve classes III, IV, and V seizures by approximately two-fold (Garriga-Canut et al., 2006; Stafstrom, Ockuly, Murphee, Valley, Roopra, Stables, and Sutula, unpubl. ms.), indicating an antiepileptic action.

Possible Mechanisms of 2DG Effects

2DG protects against seizures evoked by a 6-Hz stimulation in mice and audiogenic seizures in Fring's mice, and rapidly suppresses interictal and ictal epileptic discharges in hippocampal circuits in vitro, demonstrating acute anticonvulsant properties. Chronic antiepileptic effects of 2DG are demonstrated by its slowing of kindling progression. These two time courses—acute and chronic—probably involve different cellular and molecular mechanisms.

The chronic antiepileptic effects of 2DG have been associated with repression of brain-derived neurotrophic factor (BDNF) and its receptor tyrosine kinase B (trkB), which are required for kindling progression (He et al., 2004). 2DG suppression of seizure-induced increases in BDNF and trkB is mediated by the transcriptional repressor neuron-restrictive silencing factor (NRSF) and its NADH-sensitive corepressor carboxy-terminal-binding protein (CtBP) acting at the promoter regions of BDNF and trkB genes. In pathological conditions such as seizures, during which glycolysis and glucose production are enhanced, increases in NADH causes dissociation of CtBP from NRSF, thus decreasing transcriptional repression and resulting in increased expression of BDNF and trkB. In the presence of 2DG, which reduces NADH levels as a consequence of glycolytic inhibition, the NRSF–CtBP complex maintains repression of BDNF and trkB, and kindling progression is slowed (Garriga-Canut et al., 2006).

The rapid onset of anticonvulsant effects of 2DG in vitro and in vivo against seizures suggests that 2DG may be exerting direct actions at the synaptic or membrane levels. These mechanisms are presently uncertain, but the different time courses would implicate other mechanisms of network synchronization. The acute anticonvulsant effects of 2DG may be a result of an, as yet, undetermined metabolic or electrophysiological consequence of glycolytic inhibition. For example, there could be an effect on systemic lipid metabolism or mitochondrial metabolism that subsequently influences neuronal excitability.

Another component of glycolysis, fructose-1,6-diphosphate (FDP), has been shown to exert acute anticonvulsant activity in several seizure models in adult rats including kainate, pilocarpine, and PTZ (Lian et al., 2007). In that study, the effectiveness of FDP as an anticonvulsant surpassed 2DG, KD, and valproate. The mechanism of FDP's anticonvulsant effect is unclear. FDP increases glucose flux from glycolysis into the pentose phosphate pathway (PPP). NADPH generated in the PPP reduces glutathione, which has anticonvulsant activity. Therefore, FDP may exert an endogenous anticonvulsant action (Stringer & Xu, 2008). Together, these results identify modification of glycolysis as a possible novel mechanism for treatment of seizures.

2DG as a Potential Clinical Agent

Prior to clinical development of any compound as an anticonvulsant, it is necessary to demonstrate a favorable toxicity profile. Rats treated for 6 months with 2DG at a dose of 500 mg/kg/day, imp., showed no discernible adverse effects. Weight gain was similar in 2DG-treated and control-treated rats. In addition, there was no effect on spatial memory function in the Morris water maze 2 weeks after 2DG treatment (Ockuly, Sutula, and Stafstrom, unpublished data). 2DG has already been approved for phase II clinical trials as an adjuvant chemotherapy agent, attesting to its safety profile (Pelicano et al., 2006).


2DG and the inhibition of glycolysis are novel therapeutic approaches to epilepsy (Table 1). 2DG exerts acute anticonvulsant effects in vitro that are independent of the method of seizure induction. 2DG has acute anticonvulsant effects in vivo against perforant path kindling in rats. 2DG also possesses novel chronic antiepileptic effects against progression of seizures and the adverse consequences of seizure-induced plasticity that are associated with alterations in neuronal gene expression. Finally, 2DG has a favorable preliminary toxicity profile. Together, these observations support 2DG or similar agents as feasible for treating epilepsy in patients. The detailed anticonvulsant and antiepileptic mechanisms of 2DG actions remain to be clarified.

Table 1.  Effects of 2DG in models of seizures and epilepsy
  1. aStafstrom, Ockuly, Murphee, Valley, Roopra, Stables, and Sutula, unpubl. ms.

Anticonvulsant effects
 In vitro
   High Ko+Decreased interictal and ictal bursting (CA3)Stafstrom et al., unpubl. ms.a
   BicucullineDecreased interictal bursting (CA3)Stafstrom et al., unpubl. ms.a
   4-aminopyridineDecreased interictal bursting (CA3)Stafstrom et al., unpubl. ms.a
 In vivo
   6-Hz stimulation in miceSeizure protectionStafstrom et al., unpubl. ms.a
   Audiogenic seizures in Fring's miceSeizure protectionStafstrom et al., unpubl. ms.a
   Kindling of perforant pathIncreased after-discharge thresholdGarriga-Canut et al., 2006
   PilocarpineDecreased seizure severity and duration; increased seizure latencyLian et al., 2007
Antiepileptic effects
 Kindling of perforant pathGreater number of after-discharges to classes III, IV, and V seizuresGarriga-Canut et al., 2006
 Kindling of olfactory bulbGreater number of after-discharges to classes III, IV, and V seizuresStafstrom et al., unpubl. ms.a


This work was supported by The Charlie Foundation (CES), NIH RO1 25020 (TPS), the Epilepsy Research Foundation New Therapy Development Project, and the Wisconsin Alumni Research Foundation.

We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosure: The authors declare no conflicts of interest.