X

FIGURE 15.15 Structures of glycine (15.58) and an agonist (15.59), two partial agonists (15.60 and 15.61) and two antagonists (15.62 and 15.63) at the glycineB receptor on the NMDA receptor complex.

h2nl h2n

15.7.5 AM PA Receptor Agonists

A large number of selective and potent AMPA receptor agonists have been developed by substituting a heterocyclic bioisosteric group for the distal carboxylate group of Glu. For example, the heterocycles 1,2,4-oxadiazole-3,5-dione, 3-isoxazolol, and uracil, as represented by quisqualic acid (15.64), AMPA (15.65), and (S)-willardiine (15.66), respectively, have been incorporated into numerous AMPA receptor agonists. The natural product quisqualic acid (15.64) was the first agonist in use for pharmacological characterization of AMPA receptors, but due to nonselective action it was later replaced by AMPA (15.65).

Quisqualic acid (15.64) O

N NH

FIGURE 15.16 Structures of AMPA (15.65) and some AMPA receptor agonists.

FIGURE 15.16 Structures of AMPA (15.65) and some AMPA receptor agonists.

The isoxazole-based Glu homolog (S)-Br-HIBO (15.67) also shows AMPA receptor subtype selectivity, preferring GluR1 over GluR3 in receptor-binding and functional assays. Replacing the 3-isoxazolol group of AMPA by a 3-carboxyisoxazole unit gives the Glu homolog ACPA (15.68), which is a selective AMPA receptor agonist that is more potent than AMPA. The potent excitatory AMPA receptor activity of ACPA (15.68) has been shown to reside with the S-enantiomer.

Conformational restriction of the skeleton of Glu has played an important role in the design of selective GluR ligands. However, only few structurally rigid AMPA receptor-selective Glu analogs have been reported. One such example is the cyclized analog of AMPA, 5-HPCA (15.69), which recently has been resolved. Interestingly, the pharmacological effects of 5-HPCA (15.69) reside exclusively with the ^-enantiomer, in striking contrast to the usual stereoselectivity trend among AMPA receptor agonists (Figure 15.16).

15.7.6 Competitive and Noncompetitive AMPA Receptor Antagonists

Early pharmacological studies on AMPA and KA receptors were hampered by the lack of selective and potent antagonists. The discovery of the quinoxaline-2,3-diones CNQX (15.70) and DNQX (15.71) was a breakthrough since these compounds are quite potent antagonists, although nonselective. Subsequently, the more potent analog NBQX (15.72) was shown to be neuroprotective in cerebral ischemia and to have improved AMPA receptor selectivity compared to CNQX (15.70). However, NBQX (15.72) failed in clinical trials because of nephrotoxicity due to a limited aqueous solubility, but nonetheless has become a valuable tool for research. DNQX (15.71) has played a key role in elucidating the binding mode of competitive antagonists, as it was the first antagonist cocrystallized with the GluR2 ligand-binding domain. Attempts to improve the aqueous solubility of such antagonists without losing activity at AMPA receptors, by introducing appropriate polar substituents onto the quinoxaline-2,3-dione ring system have been highly successful, and have resulted in very potent AMPA receptor antagonists, as exemplified by ZK200775 (15.73).

Another series of potent and selective competitive AMPA receptor antagonists based on the isantin oxime skeleton includes NS 1209 (15.74), which shows long-lasting neuroprotection in animal models of ischemia and an increased aqueous solubility compared to NBQX (15.72). At least two classes of amino acid-containing compounds, based on decahydroisoquinoline-3-carboxylic

NC O2N

f3c h

(h3Q2no2s

ZK200775 (15.73)

H2NO2S'

o h2n

GYKI 52466 (15.77)

GYKI 52466 (15.77)

H2NO2S'

ho o

LY293558 (15.75)

LY293558 (15.75)

Talampanel (15.78)

Talampanel (15.78)

FIGURE 15.17 Structures of some competitive (15.70-15.76) and noncompetitive (15.77 and 15.78) AMPA receptor antagonists.

o acid and AMPA, have been found to be competitive AMPA receptor antagonists. LY293558 (15.75), a member of the former class, is systemically active although it shows significant antagonist effects at KA receptors in addition to its potent AMPA receptor blocking effects. The AMPA receptor antagonist (S)-ATPO (15.76), which was designed using AMPA as a lead structure, has like LY293558 a carbon backbone longer than that which normally confers AMPA receptor agonism.

The 2,3-benzodiazepines, such as, GYKI 52466 (15.77) and Talampanel (15.78), represent a class of noncompetitive AMPA receptor antagonists that have enabled the effective pharmacological separation of AMPA and KA receptor-mediated events. These compounds appear to bind to sites distinct from the agonist recognition site, and are thus negative allosteric modulators. Talampanel (15.78), currently under clinical development as a treatment for multiple sclerosis, epilepsy, and Parkinson's disease, may inhibit AMPA receptor function even in the presence of high levels of Glu (Figure 15.17).

15.7.7 Modulatory Agents at AMPA Receptors

The agonist induced desensitization of AMPA receptors can be markedly inhibited by a number of structurally dissimilar AMPA receptor potentiators known as AMPA-kines, including aniracetam (15.79), cyclothiazide (CTZ) (15.80), and in particular CX-516 (15.81), which has been shown to improve memory function in aged rats. These AMPA-kines positively modulate ion flux via stabilization of receptor subunit interface contacts and subsequent reduction in the degree of desensitization. A series of more potent arylpropylsulfonamide-based AMPA-kines have been identified, including LY395153 (15.82) (Figure 15.18).

h3co.

Aniracetam (15.79)

cxto

LY395153 (15.82)

FIGURE 15.18 Structures of some positive allosteric modulators of AMPA receptors.

15.7.8 KA Receptor Agonists and Antagonists

The pharmacology and pathophysiology of KA receptors are far less well understood than for AMPA receptors. However, identification of selective agonists and competitive antagonists has developed the field of KA receptor research during recent years, and has provided insight into the roles of these receptors in the CNS. For a number of years, KA (15.83) and domoic acid (15.84) have been used as standard KA receptor agonists despite their activities at AMPA receptors, characterized by nondesensitizing responses at these receptors. (S)-ATPA (15.85) and (S)-5-I-willardiine (15.86) are more selective KA receptor agonists, and these compounds exhibit some selectivity for the low-affinity KA receptor subtype GluR5 compared to GluR6. (S)-ATPA (15.85) and (S)-5-I-willardiine (15.86) are structurally related to potent AMPA agonists discussed in earlier sections, illustrating that the structural characteristics required for activation of GluR1-4 and GluR5 receptors are quite similar. However, the presence of the relatively bulky and lipophilic ieri-butyl- or iodo-substituents of these compounds is apparently the major determinant of the observed receptor selectivity.

Among the four possible stereoisomers of the 4-methyl substituted analog of Glu, only the 2S,4^-isomer (15.87) shows selectivity for KA receptors. Replacement of the 4-methyl group of (2S,4^)-Me-Glu (15.87) by a range of bulky, unsaturated substituents containing alkyl, aryl, or het-eroaryl groups has yielded a number of interesting GluR5 receptor-selective compounds including LY339434 (15.88). LY339434 shows approximately a 100-fold selectivity for GluR5 over GluR6 and no affinity for GluRl, 2, or 4 receptors.

Whereas a large number of selective competitive AMPA receptor antagonists have been identified, only a few selective KA receptor antagonists have been reported. One of the first reported KA receptor-preferring antagonists was the isantin oxime, NS 102 (15.89), which shows some selectivity toward low affinity [3H]KA sites as well as antagonist effect at homo-meric GluR6. However, low aqueous solubility has limited the use of NS 102 (15.89) as a pharmacological tool. A number of decahydroisoquinoline-based acidic amino acids, including LY382884 (15.90), have been characterized as competitive GluR5-selective antagonists that exhibit antinociceptive effects.

More recently, a series of arylureidobenzoic acids have been reported as the first compounds with noncompetitive antagonist activity at GluR5. The most potent ligands, exemplified by compound 15.91, exhibit more than 50-fold selectivity for GluR5 over GluR6 or the AMPA receptor subtypes (Figure 15.19).

O OH

O OH

Domoic acid (15.84)

N NH

OH O

LY339434 (15.88)

OH O

LY339434 (15.88)

N~OH

N HO

N~OH

o^oh

LY382884 (15.90)

LY382884 (15.90)

Cl Cl

FIGURE 15.19 Structures of KA (15.83) and some KA receptor agonists (15.84-15.88), two competitive (15.89 and 15.90) and one noncompetitive antagonist (15.91).

15.8 METABOTROPIC GLUTAMATE RECEPTOR LIGANDS

The cloning of the mGluRs and the evidence, which has subsequently emerged on their potential utility as drug targets in a variety of neurological disorders, have encouraged medicinal chemists to design ligands targeted at the mGluRs. In analogy to the iGluRs, several x-ray structures of a mGluR ligand-binding construct (see Figure 12.7) including different ligands have been obtained and afforded important structural knowledge of value, e.g., in the design of ligands.

15.8.1 Metabotropic Glutamate Receptor Agonists

The first agonist to show selectivity for mGluRs over iGluRs was (1S,3P)-ACPD (15.92) which has been used extensively as a template for the design of new mGluR ligands. Introduction of a nitrogen atom in the C4 position of 15.92 gave (2P,4P)-APDC (15.93) which displays an increased potency for group II receptors compared to the parent compound while losing affinity for group I and III receptors.

LY354740 (15.94) displays low nanomolar agonist potency at mGluR2 and mGluR3, low micromolar agonist potency at mGluR6 and mGluR8, while showing no activity at the remaining mGluRs.

ABHxD-I (15.95) displays potent agonist activity, comparable to Glu, at all three mGlu groups. This observation has been of key importance in developing early models of the mGluR-binding site. Compound 15.95 is quite a rigid molecule, which adopts a conformation corresponding to an extended conformation of Glu. The observation that the compound is a potent agonist for all three mGluR groups led the suggestion that Glu adopts the same extended conformation at all three receptor groups, and that group selectivity is thus not a consequence of different conformations but rather a consequence of other factors such as, steric hindrance.

HOOC

LY354740 (15.94)

HOOC -V

hooc^^cooh nh2

COOH

COOH

Ibotenic acid (15.46) Quisqualic acid (15.64)

(5)-2-Aminoadipic acid (15.96) (15,3fi)-Homo-ACPD (15.97) (5)-Homo-AMPA (15.98)

FIGURE 15.20 Structures of some mGluR agonists (upper row) and some Glu analogs acting at iGluRs and/ or mGluRs (middle row) and the corresponding homologs acting selectively at mGluRs (lower row).

COOH

COOH

HOOC

HOOC

Apart from Glu itself (1S,3R)-ACPC (15.92), ibotenic acid (15.46) and quisqualic acid (15.64) were among the first potent metabotropic agonists, though fairly nonselective. Synthesis of homologs of these and other Glu analogs afforded compounds with more selective activity at mGluRs. Thus, (S)-aminoadipic acid (15.96) was shown to be a mGluR2 and mGluR6 agonist, (1S,3R)-homo-ACPD (15.97) a Group I agonist, whereas (S)-homo-AMPA (15.98) showed specific activity at mGluR6, and no activity at neither iGluRs nor at other mGluRs. A number of HIBO analogs including (S)-hexyl-HIBO (15.99) show group I antagonistic activity and (S)-homo-quis (15.100) is a mixed group I antagonist/group II agonist. The effect of backbone extension of different Glu analogs is often unpredictable, but chain length is nevertheless a factor of importance (Figure 15.20).

15.8.2 Competitive Metabotropic Glutamate Receptor Antagonists

One of the first potent mGluR antagonists to be reported was (S)-4CPG (15.101), and it has been used extensively as a template for designing further potent and selective antagonists at mGluRl. The a-methylated analog, (S)-M4CPG (15.102), is an antagonist at both mGluRl and mGluR2. It has been shown that the antagonist potency is increased by methylation at the 2-position of the phenyl ring. Thus (+)-4C2MPG (15.103) is approximately fivefold more potent than the nonmethylated parent compounds. It is notable that most 4-carboxyphenylglycines show selectivity for the mGluR1 subtype with no or weak activities at the closely related mGluR5 subtype. One exception to this rule is (S)-hexyl-HIBO (15.99), which is equipotent as an antagonist at mGluRl and mGluR5.

a-Methylation has been widely used to derive antagonists from agonists. Maintaining the selectivity profiles as of their parent compounds, MAP4 (15.104) and MCCGI (15.105) antagonize mGluR2 and mGluR4, respectively, albeit with significantly reduced antagonist potency compared to the parent agonist.

Substituting agonists with bulky, lipophilic side chains has been a much more successful approach to the design of potent antagonists. Two of the early compounds in this class are 4-substituted analogs of Glu such as 15.107 and 15.108, which are potent and specific antagonists for mGluR2 and mGluR3. Interestingly, compounds with small substituents in the same position, such as (2S,4S)-Me-Glu (15.106), are more potent agonists at mGluR2 than Glu, with some activity at mGluRl but without appreciable activity at mGluR4. Thus by increasing the bulk and lipophilicity at the 4-position to give such "flyswatter" substituents, the selectivity for group II is retained, and even increased, but the compounds are converted from agonists to antagonists. One of the most potent

COOH

COOH

COOH

COOH

r"COOH

HOOC

COOH

MCCGI (15.105)

HOOC

COOH

FIGURE 15.21 Structures of some mGluR ligands.

COOH

MCCGI (15.105)

COOH

HOOC

"T

FIGURE 15.21 Structures of some mGluR ligands.

LY341495 (15.109)

compounds of this type is LY341495 (15.109) with a xanthylmethyl substituent. However LY341495 (15.109) also shows affinity for other subtypes, especially mGluR8 (Figure 15.21).

It can be concluded that in their antagonized state receptors from all three mGluR groups can accommodate quite large and lipophilic side chains in a variety of positions. Furthermore, compared with small a substituents such as methyl groups which most often confer antagonists with reduced potency, the large "flyswatter" substituents in most cases confer antagonists with increased potency.

15.8.3 Allosteric Modulators of Metabotropic Glutamate Receptors

CPCCOEt (15.110) is a nonamino acid compound with no structural similarity with Glu and acts as a noncompetitive group I-selective antagonist at the 7TM region rather than the agonist-binding site. A number of other nonamino acid mGluR antagonists have been discovered, e.g., BAY36-7620 (15.111) and EM-TBPC (15.112) which are potent mGluR1 specific antagonists acting at the 7TM domain.

The two compounds SIB-1893 (15.113) and MPEP (15.114) have been reported to be potent and selective, noncompetitive antagonists at mGluR5. Like CPCCOEt (15.110), MPEP (15.114) has been shown to act at the 7TM region rather than the agonist-binding site. MPEP (15.114) also antagonizes NMDA receptors with low micromolar potency, which has led to the design of the analog MTEP (15.115), which is slightly more potent than 15.114 as an antagonist at mGluR5 and with no NMDA antagonist activity. SIB-1893 (15.113) and MPEP (15.114) also act as positive allosteric modulators at mGluR4. The allosteric effect is dependent upon Glu activation, and the compounds are thus unable to activate the mGluR4 receptor directly. Instead, the compounds enhance the response mediated by Glu, causing a leftward shift of concentration-response curves and an increase in the maximum response (Figure 15.22).

CPCCOEt (15.110)

BAY36-7620 (15.111)

CPCCOEt (15.110)

BAY36-7620 (15.111)

SIB-1893 (15.113)

FIGURE 15.22 Structures of some noncompetitive mGluR antagonists and positive allosteric modulators.

SIB-1893 (15.113)

FIGURE 15.22 Structures of some noncompetitive mGluR antagonists and positive allosteric modulators.

15.9 DESIGN OF DIMERIC POSITIVE AMPA RECEPTOR MODULATORS

Many receptors, including the Glu receptors, exist as dimers or higher oligomers and this creates the possibility of having ligand-dimers, which can bind to two binding sites simultaneously. Dimeric ligands have been developed in many different receptor areas and have led to compounds with not only improved potency, but also improved selectivity, solubility, and pharmacokinetic properties can be observed. In spite of numerous examples of dimeric ligands with improved pharmacology compared to their monomeric analogs, no structural evidence have previously been presented for the simultaneous binding of such ligands to two identical binding sites. However, such evidence have been obtained for a dimeric positive allosteric modulator at AMPA receptors.

CTZ (15.80), see Section 15.7.7, is a positive allosteric modulator at AMPA receptors and an x-ray structure of CTZ in complex with the GluR2-binding construct showed a symmetrical binding of two CTZ molecules in two identical binding sites close to each other. Another study showed a number of biarylpropylsulfonamide analogs (15.116) with good activity as positive modulators at the CTZ site, and these structures were used as templates for the design of a symmetrical dimeric ligand. By use of computer modeling, different symmetric dimeric ligands with two pro-pylsulfonamide moieties, a biphenyl linker, and different alkyl substituents were constructed and tested for binding by computer docking. This led to the proposal of dimer 15.118 as a potential ligand to bind, in a symmetrical mode, to two adjacent CTZ-binding sites, with an expected affinity three orders of magnitude better than the monomeric ligand 15.117. Upon synthesis of the two enantiomers of monomer 15.117 and the three stereoisomers of the dimer 15.118 (R,R-, S,S-, and mesoform), these were tested for activity at cloned AMPA receptors expressed in oocytes by electrophysiological experiments. (R,R)-15.118 proved to be the most potent compound with EC50 = 0.79 |M at GluR2 compared to EC50 = 1980 |M for the monomer (R)-15.117. The maximal potentiation of Glu responses for the monomer as well as for the dimer was in the order of 800%-1000%, showing that they are both effective potentiators by blockade of AMPA receptor desensitization. Obviously the dimeric compound was dramatically more potent than the monomer, more than three orders of magnitude, and a similar pattern was observed for the other less active enantiomer (Figure 15.23).

In addition to this was (R,R)-15.118 cocrystallized with the GluR2-binding construct, and from the obtained x-ray structure it was shown that (R,R)-15.118 simultaneously binds to two identical modulatory binding sites at the AMPA receptors. This is shown in Figure 15.24 illustrating the binding of (R,R)-15.118 in comparison with two molecules of CTZ proving the simultaneous binding of a dimeric ligand to two identical binding sites.

15.116

15.117

15.118

FIGURE 15.23 Structure of CTZ (15.80), and other positive allosteric modulators at AMPA receptors including the symmetric dimeric analog 15.118.

FIGURE 15.23 Structure of CTZ (15.80), and other positive allosteric modulators at AMPA receptors including the symmetric dimeric analog 15.118.

FIGURE 15.24 (A) X-ray structure of the upper part of the extracellular amino terminal domain of the GluR2 receptor in complex with CTZ (15.80) (white carbon atoms). (B) Illustration of the binding of dimer 15.118 (cyan carbon atoms) compared to the binding of CTZ (white carbon atoms) in the calculated cavity formed by the binding pocket surface (side-view). (C) Close-up-view of the x-ray structure of the GluR2 receptor in complex with dimer 15.118 (brown carbon atoms) compared to CTZ (white carbon atoms).

15.10 CONCLUDING REMARKS

The cloning of the GABA and Glu receptor subtypes and their pharmacological characterization has been of great importance to the development of the field. For the Glu area a large number of crystal structure determinations has afforded valuable information about these receptors, their mechanism of action and the 3D information about binding sites is used extensively for the design and development of new ligands. Similar detailed information about GABA receptors are still awaited. Many selective ligands have been developed as experimental tools and have been important for the understanding of many brain functions and pathophysiological aspects. However, the success as new therapeutic agents, especially within the Glu area, is still small. The knowledge about what subtypes of receptors is involved in the neurological disorders is still limited, and further elucidation of this and development of new subtype selective ligands may eventually lead to new and better therapeutic agents.

FURTHER READINGS

A.C. Foster and J.A. Kemp, Glutamate- and GABA-based CNS therapeutics, Curr. Opin. Pharmacol. 2006, 6: 7-17.

B.H. Kaae, K. Harps0e, J.S. Kastrup, A.C. Sanz, D.S. Pickering, B. Metzler, R.P. Clausen, M. Gajhede, P. Sauerberg, T. Liljefors, and U. Madsen, Structural proof of a dimeric positive modulator bridging two identical AMPA receptor-binding sites, Chem. Biol. 2007, 14: 1294-1303.

J.N.C. Kew and J.A. Kemp, Ionotropic and metabotropic glutamate receptor structure and pharmacology, Psychopharmacology 2005, 179: 4-29.

C. Madsen, A.J. Jensen, T. Liljefors, U. Kristiansen, B. Nielsen, C.P. Hansen, M. Larsen, B. Ebert, B. Bang-Andersen,

P. Krogsgaard-Larsen, and B. Fr0lund, 5-Substituted imidazole-4-acetic acid analogues: Synthesis, modeling and pharmacological characterization of a series of novel y-aminobutyric acid receptor agonists. J. Med. Chem. 2007, 50: 4147-4161. U. Madsen, H. Brauner-Osborne, J.R. Greenwood, T.N. Johansen, P. Krogsgaard-Larsen, T. Liljefors, M. Nielsen, and B. Fr0lund, GABA and Glutamate receptor ligands and their therapeutic potential in CNS disorders, in S.C. Gad (ed.) Drug Discovery Handbook, Wiley, Hoboken, 2005, pp. 797-907.

Was this article helpful?

0 0

Post a comment