You are on page 1of 9

Current Pharmaceutical Design, 2002, 8, 835-843

835

Amino-Alkyl-Cyclohexanes as a Novel Class of Uncompetitive NMDA Receptor Antagonists


W. Danysz*1, C.G. Parsons1, A. Jirgensons2 , V. Kauss2 and J. Tillner1
1 Merz+Co., Eckenheimer Landstrasse 100-104, 60318 Frankfurt am Main, Germany; 2Institute of Organic Synthesis, 21

Aizkraukles str., 1006 Riga, Latvia Abstract: Because of its widespread involvement in the physiology and pathology of the CNS, the glutamatergic system has gained considerable attention as a potential target for development of new agents for a number of therapeutic indications. In this respect, the glutamate receptor subtype of the NMDA type has been most intensively studied. The present review describes the rational for developing amino-alkyl-cyclohexanes, as new uncompetitive NMDA receptor antagonists based on our positive experience with memantine which has been used clinically for many years for the treatment of neurodegenerative dementia. Many amino-alkyl-cyclohexane derivatives have been evaluated in vitro and in animal models, and in turn, one structure, namely neramexane HCl (MRZ 2/579) was selected for further development. This agent shows some similarity to memantine e.g. channel blocking kinetics, voltage dependency, and affinity. Preclinical tests indicated particularly good activity in animal models of alcoholism (self-administration, withdrawal-induced audiogenic seizures etc.) and pain (chronic pain, inhibition of tolerance to the analgesic effects of morphine). It turn, this agent has recently entered phase II of clinical trials in alcoholism after a favourable profile seen in phase I studies.

INTRODUCTION In general, the development of CNS active treatments, based on rational screening has not resulted in overwhelming quantities of new drugs. The probable reason is that our knowledge about brain function is still quite fragmentary and we actually often still learn more from effective treatments per se developed by chance than by use of our theoretical knowledge to develop new innovative treatments. One example of this paradox is the field of antagonists acting at glutamate receptors of the N-methyl-D-aspartate (NMDA) type which are cationic channels permeable to Ca2+ , Na+ and K+ . The potential therapeutic application of agents interacting with the glutamatergic system in the CNS is enormous and ranges from neuroprotection through pain to psychiatric diseases such as schizophrenia and depression [1]. Despite this fact, no new treatment targeting this receptor has been introduced since it was recognized as an attractive target over a decade ago. However, several agents currently in clinical use have subsequently been found to block NMDA receptors e.g. memantine, dextromethorphan, ketamine, or amantadine. Memantine is the 3,5-dimethyl derivative of amantadine and has been used successfully in Germany for the treatment of dementia and spasticity for over a decade. The therapeutic potential of memantine extends far beyond these two applications as shown by several clinical studies and suggested by a wealth of preclinical evidence [2]. This popularity of memantine as a scientific tool is probably related to the fact that it proves the concept that a clinically well tolerated NMDA receptor

antagonist could be used successfully in therapy. In turn, such studies created perspectives for many therapeutic uses of NMDA antagonists apart from those indications already selected. This perspective was a major trigger for the development of new NMDA receptor antagonists possessing some features similar to memantine but based on a different structural moiety. An outcome of these efforts was the synthesis of over a hundred agents based on the cyclohexane structure one of which, neramexane HCl (MRZ 2/579, 1amino-1,3,3,5,5-pentamethyl-cyclohexane HCl) is in phase II of clinical trials. The present review describes the hallmarks of its development and highlights further perspectives.

SYNTHESIS The synthesis of neramexane hydrochloride (7) is shown in Fig. (1). The commercially available isophorone (1) served as starting material to prepare 3,3,5,5tetramethylcyclohexanone (2) by CuCl catalyzed conjugate addition of methylmagnesium iodide [3]. 1,3,3,5,5pentamethylcyclohexanol (3) was obtained from ketone (2) by the Grignard reaction with methylmagnesium iodide [4]. In the first published approach toward neramexane (7), cyclohexanol (3) was converted to 1-azido-1,3,3,5,5pentamethylcyclohexane (4) and the former was reduced to amine ( 7) [5]. This method was limited to the preparation of only small amounts of amine (7) due to the explosive hydrazoic acid used. Alternatively, the Ritter reaction of cyclohexanol (3) with hydrogen cyanide and subsequent hydrolysis of the resulting N-formyl-1,3,3,5,5pentamethylcyclohexan-1-amine (5) could be successfully used to obtain amine (7) (Jirgensons and Kauss, unpublished). However, very toxic hydrogen cyanide
2002 Bentham Science Publishers Ltd.

*Address correspondence to this author at Merz+Co., Eckenheimer Landstrasse 100, 60318 Frankfurt am Main, Germany; Tel: +49-69-1503564; fax: +49-69-596-2150; e-mail: wojciech.danysz@merz.de

1381-6128/02 $35.00+.00

836

Current Pharmaceutical Design, 2002, Vol. 8, No. 10 O MeMg, cat, CuCl CH3 1 H3 C 3 NH 3, TiCl4 67% H3 C H3 C 4 O H3 C 3 TMSCN, H2 SO4 H3 C H3 C 5 O H3 C 3 ClCH 2CN, H2SO4 86% H3 C H3 C 6 CH3 CH3 HN Cl 1. thiourea, AcOH 2. HCl 89% CH3 CH3 HN H 1. 20% aq. H2SO4 2. HCl 88% in 2 steps H3 C H3 C H3 C 7 NH 2* HCl CH3 CH3 CH3 CH3 N 78% H3 C H3 C 2 1. LiAlH4 2. HCl 82% O H3 C CH3 CH3 MeMgI 85% H3 C H3 C 3 OH CH3 CH3

Danysz et al.

H3 C H3 C

Fig. (1). .Synthesis of neramexane hydrochloride. Summary four step procedure for large scale synthesis.

involved in the process was recognized as a drawback for the scale-up. This prompted us to develop a method for the preparation of neramexane (7) avoiding dangerous and expensive reagents [6]. Using this approach cyclohexanol (3) was converted to N-chloroacetyl-1,3,3,5,5pentamethylcyclohexan -1-amine (6) by the Ritter reaction with chloroacetonitrile. Subsequent cleavage of chloroacetyl group in amide (6) provided amine (7) in good overall yield. In summary, a four step procedure for the preparation of Neramexane 7 well suited for large scale synthesis was developed.
NH 3+Cl NH 3+Cl -

IN VITRO PHARMACOLOGY STRUCTURE ACTIVITY RELATIONSHIP (SELECTIVITY) All of the amino-alkyl-cyclohexanes tested were uncompetitive NMDA receptor antagonists as evidenced by: their displacement of equilibrium [3H](+)-MK-801 ((+)-5methyl-10,11-dihydro-5H-dibenzocyclohepten-5,10-imine maleate) binding to cortical membranes, their use- and voltage-dependent blockade of NMDA-induced currents in patch clamp experiments on cultured hippocampal neurones, and antagonism of NMDA receptor-mediated currents in
NH 3+Cl NH 3+Cl NH 3+Cl -

MRZ 2/579 NH 3 Cl
+ -

MRZ 2/600 NH 3+Cl -

MRZ 2/601 NH 3+Cl -

MRZ 2/615 NH 3+Cl -

MRZ 2/623

MRZ 2/632

MRZ 2/633

MRZ 2/641 NH + Cl
-

MRZ 2/621 NH + Cl -

NH 2+ Cl -

NH 2+ Cl -

MRZ 2/640

MRZ 2/639

MRZ 2/642

MRZ 2/705

Fig. (2). Chemical structures of selected amino-alkyl-cyclohexanes having antagonistic properties at NMDA receptors.

Uncompetitive NMDA Receptor Antagonists

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

837

Xenopus oocytes [7, 8]. The structures of selected aminoalkyl-cyclohexanes are shown in Fig. (2). Inward current responses of cultured hippocampal neurones to NMDA (200 M with glycine 1 M at -70mV) were antagonized by the amino-alkyl-cyclohexanes tested with a wide range of potencies correlating very well with their potency assessed in binding experiments (r=0.763, p<0.00001). The use-dependency of this open channel blockade was reflected in a change in kinetics of the first response to NMDA following pre-incubation with these antagonists. A similar use-dependence was apparent for recovery from the effects of these antagonists, with responses after removal of more potent compounds taking longer to reach control levels. Further support for the uncompetitive nature of this antagonism was provided by the clear voltage-dependency of the blockade. Kinetic experiments were performed by applying various concentrations of amino-alkyl-cyclohexanes for 5-10 seconds in the continuous presence of NMDA (200 M and glycine 1 M) for 20-30 seconds. Neramexane HCl showed concentration-dependent open channel blocking kinetics at -70mV (K on 10.67 0.09 *10 4 M -1 sec-1 ) whereas the offset rate was essentially concentration-independent (Koff 0.199 0.002 sec -1 ). Calculation of the ratio Koff / Kon revealed an apparent Kd value of 1.87 M which was somewhat greater than the IC50 calculated at equilibrium (1.29 0.20 M). In agreement with our previous data with well characterized uncompetitive NMDA receptor antagonists [9] the offset kinetics of NMDA receptor blockade by the amino-alkyl-cyclohexane derivatives was strongly correlated with their affinity at equilibrium (r = 0.86, p < 0.0001). Moreover, affinity calculated on the basis of the kinetics experiments (Kd = Koff/Kon) showed the strongest correlation with equilibrium potencies assessed in patch clamp experiments (r = 0.981, p < 0.001). In contrast, there was little or no correlation between equilibrium potencies and Kon or voltage-dependency. Moderate affinity uncompetitive NMDA receptor antagonists might be predicted to block other Ca2+ permeable channels. We therefore tested memantine and neramexane on L- and N-type VACCs in freshly dissociated hippocampal neurones and P-type VACCs in freshly dissociated cerebellar Purkinje neurones. Both were very weak antagonists at all three channel types (IC50s>170 M) [10]. Memantine and neramexane were also very weak antagonists of voltage-activated Na + channels in freshly dissociated DRG neurones (TTX-sensitive and TTXresistant, IC50s>100 M for both). In contrast, neramexane HCl and memantine had similar potency against NMDAinduced currents in these freshly dissociated and cultured hippocampal neurones [11, 10]. In conclusion, neramexane HCl seems to be a selective uncompetitive NMDA receptor antagonist despite moderate potency. The potency of these agents on native receptors in cultured hippocampal neurones agrees well with that accessed on NR1a/2A receptors expressed in Xenopus oocytes [8]. Several amino-alkyl-cyclohexanes showed the following rank order of potency 2D >= 2C > 2B >2A

although there were notable exceptions as e.g. with neramexane HCl which showed no subtype selectivity (IC50s of 0.49 0.11, 0.56 0.01 M, 0.42 0.04 and 0.49 0.06 M on NMDA1a/2A, 2B, 2C and 2D respectively)[8]. The biggest difference in potency between NR2B and NR2A receptors was seen with agents with additional alkylation of the amino-moiety such as MRZ 2/639, 2/640 and 2/705. However, this very modest differences are unlikely to account for the promising therapeutic profile of these agents [7]. Low M concentrations of most of the tested aminoalkyl-cyclohexane derivatives were effective neuroprotectants against glutamate toxicity in cultured cortical neurones [7]. Neramexane HCl had an IC50 of 2.16 0.03 M in this assay. Neramexane HCl concentrationdependently facilitated recovery of population fEPSPs in the CA1 of hippocampal slices following 7 min. of hypoxia / hypoglycaemia with an EC 50 of 7.01 M [7]. IN VIVO EFFECTS INDICATIVE OF RECEPTOR ANTAGONISM IN THE CNS NMDA

One of the crucial aspects of development of CNS active drugs is their accessibility to the site of action i.e. good penetration through the blood-brain barrier. There are a number of methods allowing to screen for these feature, but they either dont take into account all of the multiple mechanisms responsible for brain accessibility (transport in and out of the brain, saturation of these processes, metabolism, elimination) which holds true for in vitro methods or are very time and cost consuming such as in vivo methods e.g. brain microdialysis. Both of these methods require the involvement of an analytical department for quantitative analysis of the drug of interest. Fortunately, a very common feature of NMDA receptor antagonist is their anticonvulsive activity. Although this effect will probably never find therapeutic use (enhanced side effects in epileptic patients see [12] for discussion), it is ideally suited to assess CNS accessibility in a very fast and efficient way whilst being very cost effective. Hence, we assessed anticonvulsive effects of all cyclohexanes active as NMDA antagonists in vitro tests using the maximal electroshock (MES) convulsion model in mice [7]. We used a standard post injection time interval of 30 min and 3-5 doses (4-7 mice per dose) for each agent. Using this design we are able to obtain a reliable quantal dose response allowing calculation of ED 50 according to Litchfield and Wilcoxon [13]. Additionally, to have an initial insight into the relationship of anticonvulsive potency to possible side-effects, we assessed myorelaxation (wire test) and ataxia/myorelaxation (rota-rod test) in parallel in the same mice. The therapeutic index (TI) obtained this way by dividing ED50 for side effect over ED50 for MES seems to have value as an indicator of very preliminary tolerability e.g. (+)MK-801 has a considerably lower TI than memantine [7]. Neramexane HCl was also tested using in vivo electrophysiological techniques and selectively antagonized responses of single spinal neurones to electrophoretic NMDA with an ID50 of 3.97 0.34 mg/kg i.v. (McClean & Headley, unpublished). In the majority of cases the potency of amino-alkyl-cyclohexanes corresponded

838

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

Danysz et al.

Table 1.

In vitro Characterization of Selected Amino-alkyl-Cyclohexanes as NMDA Receptor Antagonists in Comparison to Memantine


MK-801 Ki (M) 0.79 1.27 2.28 8.09 2.42 32.20 3.16 2.88 5.18 4.17 4.87 143.33 12.72 7.09 PC IC50 (M) 1.39 1.29 3.49 10.00 2.90 92.40 3.74 6.40 13.90 7.38 14.04 >100 42.50 20.80 Glut Tox MES ED50 IC50 (M) (mg/kg) 1.40 2.16 2.10 3.50 2.29 >100 4.50 11.20 6.10 6.33 8.35 >100 29.30 NT 6.9 3.6 22.6 15.6 6.1 36.9 13.1 11.0 8.8 5.9 8.2 >50 8.0 9.5 NR1a/2A IC50 (M) 0.89 0.49 3.69 9.05 1.80 36.85 3.36 1.69 2.50 6.82 8.13 416.18 42.22 26.94 NR1a/2B IC50 (M) 0.40 0.56 1.96 4.99 1.10 30.77 1.94 1.52 2.65 2.97 2.78 147.39 26.18 7.67 NR1a/2C IC50 (M) 0.32 0.42 1.16 1.22 0.49 22.00 0.49 1.45 1.52 1.75 2.10 206.81 16.29 5.69 NR1a/2D IC50 (M) 0.28 0.49 0.48 0.86 0.53 16.87 1.00 0.66 1.65 2.90 4.65 179.76 20.87 6.39 Kon 104 M -1 s-1 7.52 10.67 6.43 3.71 NT 3.21 1.72 3.61 NT 7.99 4.01 NT NT NT Koff s-1 0.21 0.20 0.29 0.28 NT 2.97 0.22 0.25 NT 0.51 0.47 NT NT NT delta

MRZ 2/

Memantine 579 600 601 615 621 623 632 633 639 640 641 642 705

0.80 0.82 0.84 0.79 NT 0.60 0.87 0.70 NT 0.68 0.70 NT NT NT

MK-801 is the potency expressed as Ki in M in displacing the binding of [3 H]MK-801 to cortical membranes; PC is the potency expressed as IC50 in M in antagonizing NMDA induced currents in patch clamp recordings from cultured hippocampal neurones at -70mV; Glut Tox is the potency expressed as IC50 in M in antagonizing glutamate induced toxicity in cultured cortical neurones and MES is the potency expressed as ED50 in mg/kg i.p. in reversing maximal electroshock convulsions in mice [7]. NR1a/2A, NR1a/2B, NR1a/2C and NR1a/2D is the potency expressed as IC50 in M at -70mV in antagonizing glutamate-induced currents in two electrode voltage-clamp recordings from Xenopus oocytes expressing these NMDA receptor subtypes [8]. K on is the onset kinetics of blockade (104 M-1 s -1) Koff is the offset kinetics of blockade (s-1) and delta is the depth of the binding site within the channel i.e. reflection of the voltage-dependency (theoretical maximum 1) [7].

to their potency as NMDA antagonists in binding experiments and patch clamp studies (Table 1), however the correlation (r=0.515 and 0.597, both p < 0.005) is far from perfect. The reason could be additional actions of some amino-alkyl-cyclohexanes, which increase or decrease their anticonvulsive potency (e.g. sodium and potassium channels respectively) or differences in metabolism, elimination and penetration to the brain. Out of all agents tested, (neramexane HCl) seemed to have the most favourable profile. However, in initial characterization in animal models several other agents with similar properties have been tested as well (see later).

summarized in Table 1 and 2 and include voltagedependency and on/off kinetics of the blockade. Neramexane was tested in a battery of tests/models indicative of its efficacy in a wide range of indications (see below). In some cases, several other selected amino-alkyl-cyclohexanes were tested in parallel but none turned out to be better. During a recent meeting on amino acids in Bonn, Germany, a list of features was presented by Michael Rogawski [14], which likely determine the good tolerability of moderate affinity NMDA receptor antagonists (Table 3). Taking into account factors important for good therapeutic tolerability, the following should be considered: 1. Moderate affinity is only one of the determinants of good clinical tolerability. Other factors such as kinetics, voltage-dependence, partial trapping, subtype selectivity could play a role as well, although these are probably related to the former. Decreasing affinity leads unequivocally to a decrease in selectivity, namely actions on other ion channels. This could be beneficial supporting therapeutic efficacy (good dirty drug) or could be detrimental if the wrong targets are hit. Hence, in our opinion, there is a certain range of affinity for channel blockers at NMDA receptor which can be considered optimal -

SELECTION OF NERAMEXANE

THE

LEAD

COMPOUND

Based on experience with memantine we established certain criteria, which should be fulfilled in order to achieve reasonable clinical tolerability, moderate affinity being a prerequisite. Moderate means 0.5-5 M and agents more potent than this range may show almost irreversible blockade due to slow channel blocking kinetics while agents less potent than this range will likely affect other ionotropic and voltage-activated channels, possibly leading to unwanted side effects. Other factors that were determinants in the selection of neramexane for further development are

2.

Uncompetitive NMDA Receptor Antagonists

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

839

Table 2. Selectivity of Neramexane for NMDA Receptor Channels


MRZ 2/579 (M) Channel type NMDA (Hippocampus) NMDA (Hypothalamus) AMPA (Hippocampus) Ca2+, L-type Ca2+, N-type Ca2+, P-type Na+, TTX-Resistant Na+, TTX-Sensitive IC50 0.66 0.57 10M = 99.4% 619.3 178.0 399.9 153.7 185.9 SD 0.03 0.05 0.9% 20.1 17.8 11.5 17.5 20.9

assessment of concentrations in brain homogenates might be misleading resulting in overestimation (15-45 fold) of free extracellular brain levels due to accumulation in intracellular compartments [15]. Bearing this in mind we performed brain microdialysis experiments with in vivo recovery to assess free ECF levels of neramexane HCl in comparison to memantine. Memantine at a dose of 5 mg/kg produces free extracellular concentrations in the rat brain around the therapeutic concentration of 1M while neramexane HCl at 5 mg/kg leads to 0.8 M ECF levels [15, 16]. Infusion of 40 mg/kg/day leads to 0.73 M levels in plasma and 0.36 M in the extracellular fluid [16]. Selection of Animal Models Neramexane HCl has been tested in a wide battery of animal models in which NMDA receptor antagonists are expected to have promising activity. Our selection included animal models of acute and chronic pain with special attention to the interaction with morphine (morphine tolerance), acute and chronic models of brain insult, alcohol and opioid abuse, Parkinsons disease (including L-DOPAinduced dyskinesia) and depression. For all of these indications there is either a strong mechanistic rational, promising preclinical data or preliminary clinical studies with other agents. Neuroprotective activity Neramexane HCl shows strong neuroprotective activity. It was very potent against NMDA-induced lesion of the NBM and was protective at 1 mg/kg [17]. Much higher doses of neramexane HCl are needed to provide neuroprotection in a transient model of MCA occlusion. In our own studies a bolus of 6 mg/kg at the time of occlusion followed by infusion of 6 mg/kg/hr leading to 7 M levels in the ECF and 10 M in plasma, provided significant protection (over 50 %) in the cortex but not striatum (Sopala et al., 1998 internal report). Even better effects were obtained by Ginsberg and colleagues [18], since not only were neuroprotective effects seen in the striatum

Data obtained using patch clamp experiments from freshly dissociated neurones [10] as described in [11].

approx. from 0.5 M to 5 M. The design of drugs with low affinity for multiple "correct" targets seems, in most cases, to be destined to failure.

ACTIVITY IN ANIMAL MODELS Pharmacokinetics dose Selection An important element of any in vivo experiments with CNS active substances is predetermination of some pharmacokinetic aspects such as the relation between dose, serum levels and free brain concentration. Such comparison allows the establishment of serum levels achieved at doses effective in animal models and free brain levels. The former then makes selection of dosing in phase I clinical studies much easier. Furthermore, the free brain levels give information whether the presumed mechanism of action found through in vitro studies - at the doses used can be verified. A very important and often ignored issue in the field of NMDA receptor channel blockers is the fact that the
Table 3.

Features Which Determine Good Tolerability of NMDA Channel Blockers (According to Rogawski [14]).
Property Comment General feature of many well tolerated uncompetitive antagonists Relevant for only selected antagonists High affinity agents (MK-801) produce irreversible-like blockade Differences in extent of partial trapping are small; requires further validation NR2B site antagonists seem to have low toxicity (except hERG), differences small Importance unknown Common feature of many agents, more relevant as affinity decreases - +ve and -ve effects possible Examples All low affinity antagonists IEM-1754 Most moderate affinity antagonists WIN 63480, Memantine, AR-R15896AR Also ADCI, Felbamate - but via channel blockade Remacemide, AR-R15896AR, ADCI Felbamate, ADCI, remacemide, ARL12495AA

Rapid association rate (related to low affinity) Rapid association rate (intrinsic) Rapid dissociation rate Reduced (partial) trapping NMDA receptor subtype selectivity Voltage independent (allosteric) block Multiple receptor targets

840

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

Danysz et al.

(which is rather rare in this model) but the neuroprotective effect was obtained also when the treatment started 2 hr after occlusion induction i.e. at the time of reperfusion. Moreover, in this study also a clear-cut improvement in a functional parameter i.e. neurological score was found. Memantine has previously been found to be active in the MCA model after a single injection of 20 mg/kg [19]. Interestingly, neramexane HCl also prevented MCAo-induced transneuronal secondary impairment in long term potentiation (LTP) in the hippocampus, a region remote from the area of primary insult (striatum and motor cortex) [20]. In spite of these encouraging data, we decided not to continue further development as a therapy against stroke, in line with the decision of several pharmaceutical companies to discontinue the development of several NMDA antagonists already at the clinical stage of assessment in this indication e.g. selfotel, cerestat, GV 150526, eliprodil, ACEA 1021 [1]. In these cases either recently completed clinical trials were very disappointing or there was no sufficient risk-benefit ratio. Opioid Abuse Although opioid abuse (morphine in particular) in general does not seem to be regarded as a major target for drug development, it is important to investigate this indication considering the fact that neramexane HCl might be profiled for inhibition of morphine tolerance. Thus, assessing the action of neramexane HCl in animal models focussing on morphine abuse seemed of primary importance. Neramexane HCl attenuated morphine place preference [21, 22] and also inhibited morphine self-administration in mice at very low doses (starting at 1-3 mg/kg) [23]. In contrast, (+)MK-801 was ineffective at 0.1 mg/kg providing an example of a qualitative difference between low and high affinity NMDA channel blockers. These studies not only indicate that neramexane HCl should not increase the risk of morphine abuse in patients but might even have an opposite effect. Alcohol Abuse Alcohol has inhibitory effects on NMDA receptor function at concentrations known to be achieved in alcoholics [24]. This finds support in behavioural studies showing that NMDA receptor antagonists (including memantine and neramexane HCl) generalize to alcohol in drug discrimination studies at low - therapeutically relevant doses [25, 26, 27]. In operant conditioning studies responding to obtain alcohol was inhibited by neramexane HCl at a dose of 5 mg/kg [28] while no selective effect for memantine could be found in the same model [29]. Repetitive treatment with neramexane HCl had similar inhibitory effects on alcohol self-administration [30]. Rats with long lasting (several month) exposure to alcohol show relapse behaviour upon withdrawal. This relapse drinking was selectively inhibited by continuous sc. infusion of neramexane HCl [26, 31].

Neramexane HCl was also very potent in inhibiting audiogenic seizures following alcohol withdrawal, supporting its potential utility in alcohol detoxification [30]. Parkinsons Disease Neramexane HCl showed antiparkinsonian like activity in rodent models such as haloperidol-induced catalepsy, reserpine-induced sedation and rotation in rats with unilateral lesion to the nigro-striatal system [32, 33] However, effective doses (except for the first model) were above the predicted maximal therapeutically relevant range (10 mg/kg) and in none of these models was efficacy better than following (+)MK-801 treatment. Synergism with LDOPA was obtained at somewhat lower doses (ibid.) but the necessity of combination in order to obtain sufficient efficacy is not encouraging for development in this indication. Depression Some experimental evidence and recent clinical studies suggests that NMDA receptor antagonism may be a valid approach in the treatment of depression [34, 35]. The clinical observation that ketamine produces short lasting psychotomimetic effects (hours) but long-lasting antidepressive action (days) following i.v. infusion was recently reported by Krystals group [35]. Neramexane HCl showed clear, dose-dependent antidepressive-like activity in the forced swim test but only the dose of 5 mg/kg was selective i.e. devoid of locomotor stimulatory properties in the open field test, which is a necessary control experiment (Maj et al. in preparation). Even more interesting is the observation that neramexane HCl at a very low dose (2.5 mg/kg) enhanced in a synergistic manner (over additive effect) the antidepressive action of imipramine and fluoxetine in the forced swim test (Rogoz, submitted). The enhancement of venlafaxine action was additive only (ibid.). This interaction was not accompanied by overall increase in locomotor activity as assessed in the open field supporting the specificity of this effect. Pain Tolerance to morphine, and likely to other drugs can be viewed as a kind of synaptic plasticity. NMDA receptors have been implicated in various plasticity processes, hence the fact that NMDA receptor antagonists inhibit tolerance is not unexpected. A similar type of plasticity could play a role in the induction of hyperalgesia and / or allodynia in chronic pain states. Neramexane HCl at low doses (2-5-5 mg/kg) attenuated thermal hyperalgesia in rats produced by carrageenan injection into the paw [37] while memantine was active at higher doses of 10-15 mg/kg [38]. Many studies also demonstrated that NMDA receptor antagonists inhibit tolerance to the analgesic action of morphine [39]. Similar effects were observed when morphine was combined with neramexane HCl [40, 37]. Interestingly, neramexane HCl also reversed existing morphine tolerance both when morphine treatment was

Uncompetitive NMDA Receptor Antagonists

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

841

terminated before neramexane HCl treatment started (enhanced extinction) and also when morphine treatment continued (co-treatment, reversal) [40]. The above-cited studies were based on repetitive treatment with both morphine and NMDA antagonists. A different approach was used by Houghton and colleagues [37] who found that continuous s.c. co-infusion of morphine and neramexane HCl leads to substantial inhibition of tolerance to the analgesic effects as compared to rats treated with morphine and saline. Generally, a similar profile has been described previously for (+)MK-801 [41]. Acute interaction experiments revealed that the acute analgesic effect produced by the combination of neramexane and morphine depends on the test used and experimental setup, but in general is less reliable than inhibition of tolerance [42, 43]. Side Effects The major concern for the clinical use of NMDA receptor antagonists is their side-effect profile, in particular psychotomimetic activity. Psychotomimetic potential can be monitored in animals using prepulse inhibition of the acoustic startle reflex (PPI) a model of sensory gating. (+)MK-801 and PCP produce clear impairment of PPI while neramexane HCl was not effective at the maximal tested dose of 20 mg/kg (Wedzony et al. , unpublished). Moreover, in the open field test PCP and (+)MK-801 enhance locomotor activity from the beginning of the test while neramexane action is rather seen as a decrease in habituation and not detected in the first observation period (5 min) (Danysz et al., unpublished). Phase I So far, four phase I clinical trials have been conducted including a total of 61 healthy male volunteers treated with neramexane HCl. The initial study was a first-in-man trial with four increasing dosages (5, 10, 20 and 40 mg) in 12 healthy male volunteers. The aim of the trial was to obtain the first information about safety and tolerability and to evaluate pharmacokinetic parameters after single dose oral administration of neramexane HCl. A second study was conducted in order to evaluate the absolute bioavailability following oral and i.v. application of 10 mg neramexane HCl in 12 healthy male volunteers. Safety and tolerability of neramexane HCl as well as pharmacokinetics upon repeated oral administration of 20, 30 and 40 mg o.i.d. and 2 x 20 mg b.i.d. over a time period of 12 consecutive days in 45 healthy male volunteers were assessed in two further clinical trials. Pharmacokinetics Absolute bioavailability was 90% as calculated from AUCinf values obtained after oral and intravenous

application (infusion) of 10 mg neramexane HCl in man. Neramexane tmax was c.a. 3.5 hours post oral application whereby C max reached 37 ng/ml for 5 mg up to 51.2 ng/ml for 40 mg. Following multiple oral applications of 20, 30, and 40 mg o.i.d., Cmax at steady state was 55, 75 and 101 ng/ml, respectively. The average concentrations at steady were 41, 59 and 77 ng/ml (20, 30 and 40 mg neramexane HCl) and time to steady state was approximately 120 hours. Like C max the area under the curve (AUC) was always dose dependent and was about 2101 ngxh/ml following 40 mg single application and 1837 ngxh/ml at steady state following multiple application of 40 mg neramexane HCl. Taking together the results for Cmax and AUC, there is no evidence for dose non-linearity. Following single applications of neramexane HCl (5, 10, 20, 40 mg) V sys/f could be calculated with 850 L, suggesting extensive distribution into peripheral tissue (high volume of distribution). The volume of distribution at equilibrium Vss following i.v. application was 701 L. Approximately 30-40% of the dose given was excreted unchanged via kidneys. However, no data on alternative elimination routes (bile, liver) are available and so no final conclusion about the extent of metabolism could be drawn. In human urine, chromatography and mass spectrometry have, so far, identified 4 metabolites. Cumulative renal clearance CLR was calculated preliminary for single applications (5-40 mg) at 90 ml/min and with multiple applications (30, 40 mg) at 140 m/ml. Since the rate of plasma protein binding has yet not been evaluated, no conclusion on tubular secretion or reabsorption could be drawn. The total plasma clearance CLtot/f was consistent during each trial and could be calculated as approximately 300 ml/min. CL tot after infusion was 262 ml/min. A cross correlation of creatinine clearance, measured in healthy male volunteers, with t1/2, MRT and CLtot did not reveal any influence of decreased filtration rate on the pharmacokinetic properties of neramexane HCl. However, one should bear in mind that CL CR was not pathological. An impairment of elimination in subjects/patients suffering from renal insufficiency might be assumed and will be clarified in an additional approach. The terminal half-life t1/2, calculated from the terminal rate constant by log-linear regression, was in a range between 30 up to 38 hours. Safety The total number of adverse events reported was 241. 204 events were observed under verum treatment, including events which have been categorised as unlikely or not related to the study drug by the investigator. The number of adverse events under verum treatment which have been categorised as possibly or probably related was 167. Fatigue, dizziness and daze (somnolence) were observed most often. However, no SAE occurred and no subject was withdrawn due to an

842

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

Danysz et al.

adverse drug reaction. No consistent effects on laboratory parameters, vitals signs, ECG and EEG could be seen. The intensity for the vast majority of adverse events observed so far was mild. Approximately one half (101) of all events reported under verum treatment occurred during the 40 mg multiple application period. Although peak plasma levels for both treatments, 40 mg repeated single application and 20 mg b.i.d over 12 days, were in the same range (100 ng/ml), the splitting of the daily dose significantly reduced the total number of adverse events. The conclusion is that b.i.d. dosing considerably improves tolerability. There are no clinically relevant findings for vital signs. Trends including increases or decreases of values could not be observed. ECGs were retrospectively evaluated with respect for a QT interval prolongation by experienced cardiologists. No QT interval prolongation was found. Further Clinical Development Based on preclinical research, Merz +Co. has developed a comprehensive phase Ib/II programme for further clinical investigations of neramexane HCl. The first trial (phase Ib) has been set up recently. The overall aim of the trial is to evaluate if neramexane HCl is able to reduce secondary hyperalgesia, allodynia and wind up following capsaicin injection in healthy volunteers. A positive effect is expected since NMDA receptor antagonists have shown their potential benefits in neuropathic pain in several clinical trials as well as in clinical routine (e.g. ketamine). As written earlier in this paper, treatment of alcohol craving is a special focus of interest for NMDA receptor antagonists. Encouraging results from preclinical tests and a medical need for new anticraving substances were the primary reasons to consider this indication. A multicentre, randomized, placebo-controlled phase II clinical trial is ongoing.

4.

Based on animal models, neramexane has been profiled for alcohol addiction and the phase II is currently ongoing.

REFERENCES
[1] [2] [3] [4] [5] Parsons, C.G., Danysz, W., Quack, G. Drug News Perspect., 1998, 11, 523. Parsons, C.G. , Danysz, W., Neuropharmacology, 1999, 38, 735. Quack, G.

Kharasch, M.S.T.P.O. J. Am. Chem. Soc., 1941, 63, 2308. Chiurdoglu, G.M.A. Bull. Soc. Chim. Belg., 1954, 63, 357. Jirgensons, A., Kauss, V., Kalvinsh, I., Gold, M.R., Danysz, W., Parsons, C.G., Quack, G. Eur. J. Med. Chem., 2000, 35, 555. Jirgensons, A.K.V.K.I.G.M.R. Synthesis , 2000, 1709. Parsons, C.G. , Danysz, W., Bartmann, A., Spielmanns, P., Frankiewicz, T., Hesselink, M., Eilbacher, B., Quack, G. Neuropharmacology, 1999, 38, 85. Spielmanns, P., Parsons, C.G., Ruppersberg, J.P., Plinkert, P. Soc. Neurosci. Abst., 2000, 26, 42.6. Parsons, C.G. , Quack, G., Bresink, I., Baran, L., Przegalinski, E., Kostowski, W., Krzascik, P., Hartmann, S., Danysz, W. Neuropharmacology, 1995, 34, 1239. Tsinsadze, T., Lozovaya, N., Krishtal, O.A., Parsons, C.G. Soc. Neurosci. Abst., 1999, 25, 1716. Parsons, C.G., Panchenko, V.A., Pinchenko, V.O., Tsyndrenko, A.Y., Krishtal, O.A. Eur. J. Neurosci., 1996, 8, 446. Loscher, W. Prog. Neurobiol., 1998, 54, 721. Litchfield, J.T., Wilcoxon, F. J. Pharmacol. Exp. Ther., 1949, 96, 99. Rogawski, M. Amino Acids, 2000, in press. Hesselink, M.B., DeBoer, B.G., Breimer, D.D., Danysz, W. Pharmaceut. Res., 1999, 16, 637. Hesselink, M.B., Parsons, C.G., Wollenburg, C., Danysz, W. Naunyn-Schmied. Arch. Pharmacol., 1999, 360, 144. Wenk, G.L., Baker, L.M., Stoehr, J.D., HaussWegrzyniak, B. , Danysz, W. Eur. J. Pharmacol., 1998, 347, 183. Liu, Y., Belayev, L., Zhao, W., Busto, R., Ginsberg, M.D. Brain Res., 2000, 862, 111. Chen, H.S.V., Wang, Y.F., Rayudu, P.V., Edgecomb, P., Neill, J.C., Segal, M.M., Lipton, S.A., Jensen, F.E. Neuroscience, 1998, 86, 1121. Sopala, M., Frankiewicz, T., Parsons, C., Danysz, W. Neurosci. Lett., 2000, 281, 143. Popik, P., Danysz, W. J. Pharmacol. Exp. Ther., 1997, 280, 854.

[6] [7]

[8] [9]

[10] [11]

[12] [13] [14] [15]

CONCLUSIONS
[16]

1.

The rational search for a moderate affinity NMDA receptor antagonist resulted in discovering aminoalkyl-cyclohexanes. They were all uncompetitive and voltage-dependent antagonists (channel blockers) but one, neramexane seemed to have desired features as revealed by electrophysiological studies. This prediction was supported by in vivo studies in animal models where a favourable profile was obtained. Neramexane seems too be well tolerated in humans as shown by phase I studies.

[17] [18] [19]

2.

3.

[20] [21]

Uncompetitive NMDA Receptor Antagonists

Current Pharmaceutical Design, 2002, Vol. 8, No. 10

843

[22]

Popik, P., Mamczarz, J., Fraczek, M., Widla, M., Hesselink, M., Danysz, W. Neuropharmacology, 1998, 37, 1033. Semenova, S. , Danysz, W., Bespalov, A. Eur. J. Pharmacol., 1999, 378, 1. Faingold, C.L., NGouemo, P., Riaz, A. Prog. Neurobiol., 1998, 55, 509.

[33] [34]

Karcz-Kubicha, M., Lorenz, B., Neuropharmacology, 1999, 38, 109.

Danysz,

W.

Skolnick, P. Eur. J. Pharmacol., 1999, 375, 31. Berman, R.M. , Cappiello, A., Anand, A., Oren, D.A., Heninger, G.R., Charney, D.S., Krystal, J.H. Biol. Psychiatry, 2000 , 47, 351. Moryl, E., Danysz, W., Quack, G. Pharmacol. Toxicol., 1993, 72, 394. Houghton, A.K., Parsons, C.G., Headley, P.M. Pain, 2000, in press. Eisenberg, E., Lacross, S., Strassman, A.M. Neurosci. Lett., 1995, 187, 17. Trujillo, K.A. Psychopharmacology, 2000, 151, 121. Popik, P., Kozela, E., Danysz, W. Naunyn-Schmied. Arch. Pharmacol., 2000, 361, 425. Herman, B.H., Vocci, F., Bridge, Neuropsychopharmacology, 1995, 13, 269. P.

[23] [24] [25] [26] [27] [28] [29] [30] [31] [32]

[35]

[36] Hundt, W., Danysz, W., Holter, S.M., Spanagel, R. Psychopharmacology, 1998, 135, 44. [37] Holter, S.M. , Danysz, W., Spanagel, R. J. Pharmacol. Exp. Ther., 2000, 292, 545. [38] Koros, E., Kostowski, W., Danysz, W., Bienkowski, P. Naunyn-Schmied. Arch. Pharmacol., 1999, 359, 117. [39] Bienkowski, P., Koros, E., Kostowski, W., Danysz, W. Alcohol, 1999, 18, 131. Piasecki, J. , Koros, E., Dyr, W., Kostowski, W., Danysz, W., Bienkowski, P. Eur. J. Pharmacol., 1998, 354, 135. Bienkowski, P., Krzascik, P., Koros, E., Kostowski, W., Scinska, A., Danysz, W. Eur. J. Pharmacol., 2001, 413, 81. Holter, S.M. , Danysz, W., Spanagel, R. Eur. J. Pharmacol., 1996, 314, R1-R2. [43] Danysz, W., Essmann, U., Bresink, I., Wilke, R. Pharmacol. Biochem. Behav., 1994, 48, 111. [40] [41] [42]

Belozertseva, I.V., Dravolina, O.A., Neznanova, O.N., Danysz, W., Bespalov, A.Y. Eur. J. Pharmacol., 2000, 396, 77. Kozela, E., Danysz, W., Popik, P. Eur. J. Pharmacol., 2001, In press.

RELATED ARTICLES RECENTLY PUBLISHED IN CURRENT PHARMCEUTICAL DESIGN Komarova, S.V.; Dixon, S.J. and Sims, S.M. Osteoclast Ion Channels: Potential Targets for Antiresorptive Drugs. Curr. Pharm. Design, 2001, 7(8), 637-654. Sahuquillo, J.; Poca, M.A. and Amoros, S. Current Aspects of Pathophysiology and Cell Dysfunction after Severe Head Injury. Curr. Pharm. Design, 2001, 7(15), 1475-1503. Skerry, T.M. and Taylor, A.F. Glutamate Signalling in Bone. Curr. Pharm. Design, 2001, 7(8), 737-750.