Piracetam and TRH analogues antagonise inhibition by barbiturates, diazepam, melatonin and galanin of human erythrocyte D-glucose transport

Abstract

1 Nootropic drugs increase glucose uptake into anaesthetised brain and into Alzheimer's diseased brain. Thyrotropin-releasing hormone, TRH, which has a chemical structure similar to nootropics increases cerebellar uptake of glucose in murine rolling ataxia. This paper shows that nootropic drugs like piracetam (2-oxo 1 pyrrolidine acetamide) and levetiracetam and neuropeptides like TRH antagonise the inhibition of glucose transport by barbiturates, diazepam, melatonin and endogenous neuropeptide galanin in human erythrocytes in vitro. 2 The potencies of nootropic drugs in opposing scopolamine-induced memory loss correlate with their potencies in antagonising pentobarbital inhibition of erythrocyte glucose transport in vitro (P<0.01). Less potent nootropics, D-levetiracetam and D-pyroglutamate, have higher antagonist Ki's against pentobarbital inhibition of glucose transport than more potent L-stereoisomers (P<0.001). 3 Piracetam and TRH have no direct effects on net glucose transport, but competitively antagonise hypnotic drug inhibition of glucose transport. Other nootropics, like aniracetam and levetiracetam, while antagonising pentobarbital action, also inhibit glucose transport. Analeptics like bemigride and methamphetamine are more potent inhibitors of glucose transport than antagonists of hypnotic action on glucose transport. 4 There are similarities between amino-acid sequences in human glucose transport protein isoform 1 (GLUT1) and the benzodiazepine-binding domains of GABAA (gamma amino butyric acid) receptor subunits. Mapped on a 3D template of GLUT1, these homologies suggest that the site of diazepam and piracetam interaction is a pocket outside the central hydrophilic pore region. 5 Nootropic pyrrolidone antagonism of hypnotic drug inhibition of glucose transport in vitro may be an analogue of TRH antagonism of galanin-induced narcosis.

Figure 1

Effect of pentobarbital (2 mM) and piracetam (4 mM) on D-glucose exit rate from human red cells at 25°C into solutions containing varied (glucose) concentrations. Pentobarbital reduced the maximal rate of zero-trans net glucose exit to 0.031±0.001 s⁻¹ from a control rate of 0.061±0.002 s⁻¹ (P<0.001). The infinite cis K_m was 6.55±1.0 mM with pentobarbital (2 mM) and 2.78±0.25 mM in control. Piracetam (4 mM) in the copresence of pentobarbital (2 mM) significantly increased the maximal rate of zero-trans exit (P<0.001) and reduced the infinite cis K_m of glucose to 3.25±0.45 mM (P<0.025). On its own, in comparison with control rates, piracetam (4 mM) had no effect. Error bars are the s.e.m.'s of four to five separate determinations of rates. This experiment was repeated five times with similar results. The lines drawn are the least-squares best-fit nonlinear regression lines to the equation for competitive inhibition (see Methods).

Figure 2

Effect of varying concentrations of piracetam on the pentobarbital-dependent inhibition of zero-trans net glucose exit from human erythrocytes. Increasing concentrations of piracetam caused a linear increase in the apparent K_i of pentobarbital inhibition of glucose exit rates. Error bars are s.e.m.'s of four to five separate estimates of the rates of glucose exit. The lines drawn between the points are the least-squares best-fit nonlinear regression lines to the equation for competitive inhibition (see Methods).

Figure 3

Dixon plots of piracetam, levetiracetam and its D (+) stereoisomer ucb L060 on the apparent K_i of pentobarbital-dependent inhibition of zero-trans glucose exit. This is a replot of data shown in Figure 2. Increasing concentrations of piracetam, levetiracetam (ucb L059) and ucb L060 significantly compete with pentobarbital action on glucose transport. The lines are the least-squares linear regression lines. The secondary K_i's are estimated as outlined in Methods.

Figure 4

A Effects of bemigride and methamphetamine on zero-trans glucose exit rates from erythrocytes. Bemigride and methamphetamine caused partial inhibitions in glucose exit. Bemigride inhibits glucose exit by 30% and methamphetamine causes a maximal inhibition of glucose exit of 15%. Bemigride inhibits glucose with a relatively high potency K_i=10.5±2.5 μM but low efficacy. Methamphetamine inhibits glucose with lower potency, K_i 118±44 μM and very low efficacy. (b) The antagonist effects of bemigride and methamphetamine on the apparent K_i of pentobarbital-dependent inhibition of glucose exit from erythrocytes. The Dixon plots showing hyperbolic relationships of apparent K_i with increasing inhibitor concentrations. At low concentrations of bemigride and methamphetamine, there is a rapid increase in apparent K_i of pentobarbital-dependent inhibition. This increase in K_i plateaus at raised concentrations of bemigride or methamphetamine. The lines fitted to the data are hyperbolas of the form ax/(b+x)+z, where x is the concentration of inhibitor (μM), a is the maximal apparent K_i of pentobarbital inhibition of glucose exit, b is the concentration of x giving half-maximal increase in apparent K_i and z is the K_i of pentobarbital with zero antagonist present. The K_i of bemigride or methamphetamine at low concentrations=zb/a. The best-fit lines and error estimates of a, b and z were obtained using Kaleidagraph 3.6, as in Methods. The antagonist K_i's of bemigride=4.5±0.3 μM and for methamphetamine=130 30 μM.

Figure 5

Correlation between the in vivo potency of nootropic effects of drugs and their antagonist K_i against pentobarbital-dependent inhibition of glucose transport in red cells. The linear regression line is weighted by the s.e.m.'s of K_i (Table 1), and shows the relationship between in vivo nootropic potency and K_i of in vitro antagonism of nootropics against pentobarbital inhibition of glucose transport in erythrocytes (d.f.=(N-2)=7; correlation coefficient r=0.804; P<0.01). The circled numbers refer to the following drugs. DM235 (Ghelardini et al., 2002); CG370 (Ogasawara et al., 1995); Cyclo (pro-gly) (Seredenin et al., 2002); TRH (Brooks et al., 1987); (Abou-Khalil et al., 2003; Brooks et al., 1987; De Reuck & Van Vleymen, 1999; Devinsky & Elger, 2003); Piracetam (Koskiniemi et al., 1998; De Reuck & Van Vleymen, 1999; Genton et al., 1999); Aniracetam (Spignoli & Pepeu, 1987); L-Pyroglutamate (Spignoli et al., 1987); L060 (Noyer et al., 1995). The in vivo potencies of the drugs were obtained from literature reports of the i.p. dose required to observe a significant nootropic response against scopolamine-induced memory loss. The curvilinearity of the regression line is introduced by using the logarithmic axes required to show the full range of the regression. The regression line is weighted to the errors of the antagonist K_i's of the drugs using the software package in Kaleidagraph 3.6.

Figure 6

Histograms showing the ratios of K_i's for indirect antagonism of pentobarbital-dependent inhibition of glucose exit and the K_i of inhibition of glucose exit by the drug. Also shown are the K_i's of drug antagonism against pentobarbital-dependent inhibition. The inhibitions of glucose exit by TRH, piracetam, L-pyroglutamate and ucb L060 are very small, so the K_i's of inhibition of glucose are likely to be underestimates. Thus, the K_i ratios of direct inhibition of glucose: indirect antagonism of pentobarbital by these drugs may be underestimates.

Figure 7

Map of regions of similarity between GABA_A receptor and GLUT1 on the basic skeleton of GLUT1 showing the relative positions of the transmembrane helices and the exofacial and endofacial linkers. Regions of similarity between GABA_A receptor and GLUT1. The labels α, β and γ refer to the GABA receptor subunits. A, B and C are positions of benzodiazepine-binding loops in subunit α and E and D in subunit γ forming a single benzodiazepine-binding pocket within GABA_A receptor (Sigel and Buhr 1997; Ernst et al., 2003). The 3-D structure of GLUT1 is shown using the known topology of the 12 transmembrane helices of the major facilitator superfamily (Hirai et al., 2003). The positions of the similarities A–E with equivalent regions on GABA_A receptor are shown. The exofacial linkers between the transmembrane helices are shown either as a dotted line where similarities with GABA_A receptor are present or as a continuous line. The endofacial linkers are not shown. Glucose is thought to bind within the central hydrophilic pore (Mueckler & Makepeace, 2004). The rhomboid shape represents the position where benzodiazepine may bind. It is surrounded by six regions of similarity within the benzodiazepine-binding domain of GABA_A receptor (see Table 3). This region could form a noncatalytic site of inhibition of glucose transport by preventing glucose-induced conformational changes.