Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

There are several sites on the NMDA receptor complex at which antagonists can interact (for a review, see Harris, 1995). Interaction at different binding sites, and by different mechanisms of action, might account for the diverse clinical properties of NMDA receptor antagonists and present potential targets for the development of new pharmaceutical agents. Uncompetitive NMDA receptor antagonists acting at the phencyclidine site produce a block only after the channel enters its open state after activation (Huettner and Bean, 1988), although this use-dependent blockade does not necessarily imply occlusion of the open pore or preferential binding to the open state (Orser et al., 1997). Use-dependent antagonism could be advantageous in treating disease states where excessive activation of NMDA receptors occurs, such as cerebral ischemia or epilepsy, because it allows for preferential block of excessive activation over normal synaptic activity. Consistent with this idea, both MK-801 and phencyclidine, which bind with high affinity to a site within the channel pore, exhibit neuroprotective effectiveness in various models of focal and global ischemia (Kemp et al., 1987; Olney et al., 1989). However, their therapeutic potential is negated by their serious neurobehavioral side effects. Low-affinity, use-dependent NMDA receptor antagonists may have reduced toxicities because they reach a steady-state block more rapidly due to their rapid on-off kinetics (Rogawski, 1993), thus preventing significant calcium entry before equilibrium is reached without producing a supramaximal blockade. It has been suggested that the low-affinity, use-dependent NMDA antagonists, memantine (Muller et al., 1995; Parsons et al., 1995), amantadine (Parsons et al., 1995), and ADCI (Rogawski et al., 1991, 1995), lack serious side effects due to their relatively rapid kinetics (Chen et al., 1992; Parsons et al., 1993, 1995). However, a comparison of a series of uncompetitive NMDA receptor antagonists for efficacy as antiepileptic drugs, using electroshock in mice, demonstrated that some less potent antagonists actually had worse therapeutic indices (Parsons et al., 1995). Furthermore, although ADCI and ketamine are both low-affinity, use-dependent NMDA receptor antagonists, ketamine induces psychotomimetic effects (Ginski and Witkin, 1994; Krystal et al., 1994) despite substantially faster on-off kinetics (Parsons et al., 1995, 1996).

Recently, another low-affinity, use-dependent NMDA receptor antagonist, AR-R15896AR, was developed. It blocks NMDA-induced toxicity in primary cultures of cortical neurons and, at neuroprotective concentrations, rapidly decreases the NMDA-induced calcium influx and subsequent rise in intracellular calcium (Black et al., 1995). Electrophysiological studies have shown it to be a use- and voltage-dependent blocker with rapid kinetics (Mealing et al., 1997). AR-R15896AR also exhibits neuroprotection in rodent models of global and focal ischemia and has a favorable pharmacokinetic profile (Cregan et al., 1997; Palmer et al., 1997). It is presently in phase IIa clinical trials with stoke patients.

The use-dependent antagonists MK-801, ADCI, memantine, amantadine, and AR-R15896AR all exhibit some degree of trapping (Jones and Rogawski, 1992; Blanpied et al., 1997; Chen and Lipton, 1997; Mealing et al., 1997). Trapping channel blockers permit agonist dissociation and channel closure while the antagonist is bound to its site in the channel, whereas sequential blockers prevent the channel from closing while blocked. Blanpied et al. (1997) suggested that some drugs may show a combination of these effects. With MK-801 and ADCI, trapping appears to be complete, but in the case of memantine, about one sixth of the blocked channels release, rather than trap, the blocker (Blanpied et al., 1997). This partial trapping could not be attributed to different mechanisms of action of memantine on a heterogeneous population of NMDA receptor subtypes. In pathophysiological situations where there is repetitive NMDA receptor stimulation, trapping could potentially result in an undesirable accumulation of antagonist, producing a supramaximal blockade. The extent of this accumulation of block would be interdependent on the degree of trapping, on-off kinetics, and the frequency of repetitive stimulation. Here, we investigated the trapping characteristics of three low-affinity, use-dependent NMDA receptor antagonists: AR-R15896AR, ketamine, and memantine. These antagonists have similar intermediate blocking kinetics and strong voltage dependence, yet ketamine and memantine are examples of compounds that exhibit a small and large therapeutic index, respectively. The purpose of this study was to determine whether differences in the degree of trapping exist among use-dependent NMDA antagonists with similar kinetics and, if so, whether they correlate with differences in therapeutic potential.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals and Reagents. PBS, HEPES, MEM, poly-L-lysine, tetrodotoxin, 9-aminoacridine, and strychnine hydrochloride were purchased from Sigma Chemical (St. Louis, MO). Heat-inactivated fetal bovine serum was purchased from GIBCO Laboratories (Grand Island, NY), and heat-inactivated horse serum was from Hyclone Laboratories (Logan, UT). NMDA, AP-5, and ketamine were purchased from Research Biochemicals International (Natick, MA). Memantine was purchased from Tocris Cookson (St. Louis, MO). EGTA was purchased from Fluka Biochemika (Ronkonkoma, NY). AR-R15896AR was provided by Astra Arcus USA (Rochester, NY).

Cell Culture. Rat cortical neurons isolated from E18 fetuses were grown in primary culture as described previously (Black et al., 1995). Briefly, timed-pregnant Sprague-Dawley rats were purchased from Charles River Canada (St. Constant, Quebec, Canada). After the mother was killed through cervical dislocation while under halothane anesthesia, the fetuses were removed from the uterus on day E18; their brains were removed and placed in ice-cold PBS; and the cortices were isolated. The cortical neurons were dispersed by trituration with a 10-ml pipette, and the cells were centrifuged at 250g for 5 min at 4°C. The cells were gently resuspended in plating medium, and viable cells, as determined by trypan blue exclusion, were counted. The cells then were plated at 10⁵ cells/cm² on poly-L-lysine-coated 35-mm culture dishes (Nunc; Roskilde, Denmark) in 2 ml of plating medium at 37°C in an atmosphere of 5% CO₂/95% air. Plating medium consisted of 80% MEM, with 20 mM glucose, 10% heat-inactivated fetal bovine serum, and 10% heat-inactivated horse serum containing 2 mM L-glutamine. Because the cultures contained both neurons and glial cells, they were treated with 15 mg/ml 5-fluoro-2'-deoxyuridine and 35 mg/ml on day 5 to minimize glial cell proliferation. On day 6, half of the medium was removed and replaced with growth medium consisting of 90% MEM and 10% horse serum. Experiments were performed on cultures after 14 to 19 days in vitro.

Whole-Cell Recording. Cortical neurons grown on 35-mm culture dishes were mounted on the stage of an inverted Zeiss microscope equipped with Hoffman Modulation optics in a perfusion system flowing continuously at 1 ml/min at 22°C. The bathing solution contained (in mM): 140 NaCl, 5 KCl, 1 CaCl₂, 10 HEPES, and 3 glucose, pH 7.4. Bathing solutions also contained 1 µM tetrodotoxin to block Na⁺ currents, 10 µM glycine to saturate the glycine site on NMDA receptors, and 1 µM strychnine to block glycine-activated Cl currents.

Data Analysis. The block of NMDA-evoked currents was calculated according to the formula:

(1)

where I was determined by curve fitting the decay of the NMDA-evoked current during the NMDA application and extrapolating the fit to the end of the NMDA/antagonist coapplication, and I_B was the current measured at the end of NMDA/blocker coapplication. Current decays were fit to first-order exponential curves using a Chebyshev fit method and pClamp software.

(2)

(3)

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Onset and Relief of Antagonism. The kinetic properties of AR-R15896AR, ketamine, and memantine have been previously reported and, as can be seen from Table 1, are quite similar. However, there is considerable variability among some of this data, possibly due to differences in cell preparations and experimental conditions. For example, the k₁ reported for ketamine varies from 0.075 to 0.21 s¹. Therefore, before designing an agonist/antagonist exposure protocol to study trapping, we examined the development and relief of block by these NMDA antagonists under comparable experimental conditions.

View this table: [in this window] [in a new window]   TABLE 1 Kinetic properties of NMDA receptor antagonists k1 and k1 were determined from macroscopic current measurements.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Onset and relief of block of NMDA-induced whole-cell currents. NMDA (10 µM) was applied for 2 to 3 s followed by a 30-s coapplication of NMDA/antagonist and a second extended NMDA application. Neurons were voltage-clamped at 60 mV. Top trace, 50 µM AR-R15986AR. Middle trace, 10 µM memantine. Bottom trace, 10 µM ketamine. Dashed lines indicate the curve fit of the decay of the initial NMDA-induced current and the extrapolation of this fit to the end of the trace. Dotted lines indicate the resting current level. Solid lines superimposed on the data indicate the first-order exponential curve fits for onset and relief from block. Light shaded bar indicates NMDA application. Dark shaded bar indicates antagonist application. Scale bar, 100 pA, 10 s.

Degree of Trapping. To study trapping, we used a low agonist concentration (10 µM NMDA) and sufficiently high concentrations of each antagonist to produce a near-maximal block, thus optimizing the opportunity for trapping by minimizing the proportion of blocked receptors that were liganded at steady state (Blanpied et al., 1997). A 3-s NMDA application immediately followed by an NMDA/antagonist coapplication 2 to 60 s in duration was followed 120 s later by a second 20-s NMDA application. During the washout period, 80 µM AP-5 was included in the perfusate to prevent possible "occult" channel opening due to NMDA contamination but was removed 1 s before NMDA application to permit its dissociation from the ligand-binding site. The difference in degree of trapping by AR-R15896AR 120 s after 2-, 10-, and 60-s coapplications with NMDA is shown in Fig. 2. Transient artifacts similar to those in Fig. 1 were observed on coapplication of NMDA/antagonist that were temporally distant from the regions in the recordings important to determination of I_{B, I1st}, or I_2nd. We rejected all data that contained such perfusion artifacts during application of NMDA, where I_1st, or I_2nd were measured. Trapping by 50 µM AR-R15896AR, 10 µM ketamine, and 10 µM memantine 120 s after a 30-s coapplication are shown in Fig. 3. The residual block that remained trapped 120 s after washout was calculated according to eq. 2 by comparing the maximal current during the first 200 ± 25 ms of the first NMDA exposure with that of the second NMDA exposure. The fractional block trapped was then determined according to eq. 3. The increase in degree of trapping as a result of extending the duration of agonist/antagonist coapplication is plotted in Fig. 4A. The development of trapping with all three NMDA receptor antagonists could be fit to single exponential curves with s for AR-R15896AR, ketamine, and memantine of 3.0, 6.0, and 3.5 s, respectively. In all cases, there was no significant (p > .05) increase in trapping by extending the agonist/antagonist coapplication duration beyond 30 s. The blocks caused by a 30-s coapplication of 50 µM AR-R15896AR, 10 µM ketamine, or 10 µM memantine were not significantly different, being 82 ± 1% (n = 5), 80 ± 2% (n = 5), and 81 ± 2% (n = 7), respectively. However, after 120 s of washout, there were significant differences in the degree of trapping between each of the antagonists, as 54 ± 3% of the AR-R15896AR block, 86 ± 1% of the ketamine block, and 71 ± 4% of the memantine block remained trapped (Fig. 4B). Trapping was also examined using higher antagonist concentrations. Significantly (p < .05) higher initial blocks were produced by 100 µM AR-R15896AR, 20 µM ketamine, and 20 µM memantine, being 88 ± 2% (n = 3), 96 ± 1% (n = 3), and 92 ± 1% (n = 5), respectively. However, with all three compounds, the percentage of block trapped was not significantly different from that measured at the lower antagonist concentrations, being 57 ± 5%, 87 ± 5%, and 76 ± 3%, respectively. The 76 ± 3% trapping caused by a 30-s coapplication of 20 µM memantine agrees well with the 19 ± 2% release from block reported by Blanpied et al. (1997) using a 60-s coapplication and a holding potential of 67 mV. Trapping was not investigated at lower antagonist concentrations because the magnitude of change in current amplitude was too small to resolve.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Trapping by 50 µM AR-R15896AR after 2-, 10-, and 60-s coapplications with 10 µM NMDA. A 3-s NMDA application immediately followed by an NMDA/AR-R15896AR coapplication 2 s (top left trace), 10 s (middle left trace), or 60 s in duration (bottom left trace) was followed 120 s later by a second NMDA application 20 s in duration (middle traces). The left and middle traces are superimposed to show the difference in initial current amplitude between the first and second NMDA exposures (indicated by arrows) after 2-s (top right trace), 10-s (middle right trace), or 60-s (bottom right trace) exposure to AR-R15896AR. During the washout period, 80 µM AP-5 was included in the perfusate but was removed 1 s before NMDA application. Light shaded bar indicates NMDA application. Dark shaded bar indicates antagonist application. Dashed lines indicate curve fit of current decay during the initial NMDA application extrapolated to the end of the NMDA/antagonist coapplication. Scale bar, 100 pA, 10 s.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Degree of trapping exhibited by several NMDA antagonists 120 s after an extended agonist/antagonist coapplication. The first NMDA/antagonist coapplication is superimposed over a second NMDA application 120 s later to show the difference in initial current amplitude (indicated by arrows). During the washout period, 80 µM AP-5 was included in the perfusate but was removed 1 s before NMDA application. Top left trace, 20-s application of 50 µM 9-AA. Top right trace, 30-s application of 50 µM AR-R15896AR. Bottom left trace, 30-s application of 10 µM memantine. Bottom right trace, 30-s application of 10 µM ketamine. Dashed lines indicate curve fit of current decay during the initial NMDA application extrapolated to the end of the NMDA/antagonist coapplication. Scale bar, 50 pA, 5 s.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   A, The increase in degree of trapping as a result of extending agonist/antagonist coapplication duration. With all three NMDA antagonists, the development of trapping was fit to single exponential curves. The s for AR-R15896AR (), ketamine (), and memantine () were 3.0, 6.0, and 3.5 s, respectively. In all cases, there was no significant (p > .05) increase in trapping by extending the agonist/antagonist coapplication duration beyond 30 s. Data points are the mean ± S.E.M. of observations from three to nine cells. B, The block caused by a 30-s coapplication of 50 µM AR-R15896AR, 10 µM ketamine, or 10 µM memantine was not significantly different, being 82 ± 1% (n = 5), 80 ± 2% (n = 5), and 81 ± 2% (n = 7), respectively. However, after 120 s of washout, there were significant differences in the degree of trapping between each of the antagonists, as only 54 ± 3% of the AR-R15896AR block, 86 ± 1% of the ketamine block, and 71 ± 4% of the memantine block remained trapped. Significantly (p < .05) higher initial blocks were produced by 100 µM AR-R15896AR, 20 µM ketamine, and 20 µM memantine, being 88 ± 2% (n = 3), 96 ± 1% (n = 3), and 92 ± 1% (n = 5), respectively. However, the percentage of block trapped was not significantly different from that measured at the lower antagonist concentrations, being 57 ± 5%, 87 ± 5%, and 76 ± 3%, respectively.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Trapping channel blockers permit agonist dissociation and channel closure when the antagonist is bound to its site in the channel, whereas sequential blockers prevent the channel from closing while blocked. Sequential and trapping channel blockers differ in the dependence of their block on agonist concentration. Although the mechanisms underlying trapping are not well understood, conformational changes in gating could sterically prevent movement of a trapping channel blocker out of the channel or, on the other hand, the binding of a sequential blocker might prevent movement of a channel's gate.

In this comparative study, we examined NMDA receptor antagonist trapping. Three low-affinity, use- and voltage-dependent antagonists were chosen whose kinetics were sufficiently similar that, at the concentrations used, they produced an equivalent, near-maximal, steady-state block within 30 s, which could be completely relieved within 50 s of exposure to NMDA. We report that despite the similarities of the initial block produced by these compounds, ketamine exhibited significantly more trapping than memantine or AR-R15896AR and that memantine exhibited significantly more trapping than AR-R15896AR.

The 76 ± 3% trapping observed using 20 µM memantine agrees with the 19 ± 2% release from block reported by Blanpied et al. (1997), although we found that a 30-s, rather than a 60-s, agonist/antagonist coapplication duration was sufficient to produce a steady-state block. We also demonstrated an absence of trapping using the sequential NMDA-receptor antagonist 9-AA. The large inward tail currents observed immediately after cessation of NMDA/9-AA coapplication are also consistent with a sequential mechanism of block (Benveniste and Mayer, 1995). Together, the confirmation of comparable trapping values to those published for memantine and the absence of trapping by 9-AA verify our ability to measure trapping over a broad spectrum. We did not measure currents (I_1st and I_2nd) until 200 ± 25 ms after agonist application due to the technical limitations mentioned previously. During the delay before the measurement of I_2nd, some degree of untrapping probably occurred, which may have affected the accuracy of our trapping measurements. However, the off rates for all three antagonists are slow by comparison with the brief delay in current measurement. There was very little relief from block within 200 ms for any of the antagonists, as is shown in Fig. 1, so the error in determining trapping of block due to unblock during this short time delay in current measurement was negligible.

All three NMDA receptor antagonists show use- and voltage-dependent antagonism, suggesting activity at a site within the channel pore. Trapping was not complete with any of these antagonists. This is not consistent with linear models of trapping block (Lingle, 1983; Blanpied et al., 1997) but can occur in cyclized models of trapping block (Gurney and Rang, 1984; Benveniste and Mayer, 1995). Consider the following model, adapted from Gurney and Rang (1984):

View larger version (17K): [in this window] [in a new window]

where Ag is agonist, B is blocker, C is closed channel, O is open channel, OB is open-blocked channel, CB is closed-blocked channel, l± and k± are the rate constants between transition states, and _B is rate of agonist unbinding from an open-blocked channel.

For a sequential block to occur, k > 0, _B = 0, and l = 0. For a trapping block, k > 0, _B > 0, and l = 0. For a nonsequential block, k > 0, _B > 0, and l > 0. In this model, it was assumed that the rate constant l for the escape from the closed channel is very small. However, if this rate constant is much larger than zero, then closed-channel egress occurs. It has also been suggested that some blockers may exhibit a combination of trapping and sequential block (Blanpied et al., 1997). This could account for the differences in trapping observed in the present study. However, the appearance of tail currents with 9-AA, but not with AR-R15896AR, memantine, or ketamine, does not favor a sequential mechanism of block by these latter compounds. In this regard, the cyclic models imply that channel closure and occlusion of blocker are separable events. It then is possible to envision that a blocker may retard or escape the conformational relaxation that could result in trapping while still allowing channel closure to occur. As pointed out by Benveniste and Mayer (1995), it may be that trapping and sequential channel block are opposite ends of a spectrum rather than two fundamentally different mechanisms. The lipophilic nature of many open-channel blockers can result in the blocker leaking from the antagonist site when the channel is closed, resulting in variable degrees of trapping (Chen and Lipton, 1997). However, the rank order of degree of trapping, ketamine > memantine > AR-R15896AR, does not correlate with the compound lipophilicity [c-logP values for ketamine, memantine, and AR-R15896AR of 2.15, 3.18, and 1.63, respectively (R. Murray, personal communication)], suggesting that variation in diffusion is an inadequate explanation for differences in trapping.

Activity at other binding sites on the NMDA receptor also varies with antagonist. For example, memantine antagonism is not affected by glycine, suggesting that there is no interaction with the strychnine-insensitive glycine modulatory site (Parsons et al., 1993), whereas AR-R15896AR does show a potentially competitive interaction at the glycine modulatory site (Mealing et al., 1997). Ketamine inhibits the NMDA receptor by an open-channel block and by closed-channel block from the membrane phase (Orser et al., 1997). Therefore, the differences in trapping that we observed may not be entirely due to differences in antagonist action within the channel pore; there may also be a minor contribution from block at other sites on the NMDA receptor complex. In the case of AR-R15896AR, which shows the least trapping, at 60 mV, the contribution to the block due to antagonism at the glycine site is very small (Mealing et al., 1997).

Among the three low-affinity, use-dependent NMDA-receptor antagonists tested in this study, a correlation can be drawn between degree of trapping and therapeutic safety margin. Ketamine, which has been reported to induce psychotomimetic effects (Ginski and Witkin, 1994; Krystal et al., 1994), showed 86 ± 1% trapping, whereas memantine and AR-R15896AR, which have improved therapeutic safety profiles (Parsons et al., 1995; Palmer et al., 1996, 1997), trapped 71 ± 4% and 54 ± 3% of their initial block, respectively. This significant reduction in trapping, from ketamine to memantine to AR-R15896AR, is the same order as that for reduction of severity of side effects as measured with the Inverted Screen Test or Delayed Matching to Sample Test, which provide an indication of neuromuscular or memory impairment, respectively. For example, therapeutic ratios (ED₅₀ for neuromuscular impairment divided by the ED₅₀ for protection against maximal electroshock) for ketamine, memantine, and AR-R15896AR of 2.28, 3.28, and 8.55, respectively, have been observed (C. Crammer, personal communication). Therefore, use-dependent NMDA receptor antagonists that exhibit less trapping may provide the safest compounds for therapeutic use in disease states where repetitive NMDA receptor activation could potentially lead to an undesirable supramaximal accumulation of block. It remains to be demonstrated, however, that the differences in trapping observed in this study result in differences in the accumulation of block during a repetitive NMDA receptor stimulation protocol that mimics a pathophysiological situation.