I. Introduction

Top Next References

Axelrod et al. (1958) first described the enzyme-catalyzed O-methylation of catecholamines and other catechols in the late 1950s. The enzyme responsible for the O-methylation, catechol-O-methyltransferase (COMT; EC 2.1.1.6),² was partly purified and characterized by the same group (Axelrod and Tomchick, 1958). The subsequent basic research on COMT and the first COMT inhibitors introduced between 1958 and 1975 have been extensively reviewed by Guldberg and Marsden (1975).

The interest in COMT was rekindled in the late 1980s when the potent and selective second-generation COMT inhibitors were developed (Männistö and Kaakkola, 1989, 1990), and soon the structures of the two isoforms of COMT and the gene were characterized and COMT polypeptide cDNAs were cloned (Salminen et al., 1990; Bertocci et al., 1991; Lundström et al., 1991). Several review articles have recently dealt with this development (Männistö and Kaakkola, 1989, 1990; Männistö et al., 1992b, 1994; Roth, 1992; Kaakkola et al., 1994a; Dingemanse, 1997), culminating in the marketing of two new COMT inhibitors. This review concentrates on the recent information on the biochemistry and molecular biology of COMT and on the pharmacology and clinical efficacy of the new selective and relative nontoxic COMT inhibitors.

	II. COMT Gene and Proteins

Top Previous Next References

The structural organization of the COMT gene has been reviewed in detail by Lundström et al. (1995). There is one single gene for COMT, which codes for both soluble COMT (S-COMT) and membrane-bound COMT (MB-COMT; Salminen et al., 1990; Lundström et al., 1991). In the application of the in situ hybridization and cell hybrid techniques, the COMT gene in humans was localized to chromosome 22, band q11.2 (Grossman et al., 1992a; Winqvist et al., 1992). The human COMT gene contains six exons with the two first exons being noncoding. The expression of the COMT gene is controlled by two distinct promoters located in exon 3 (Salminen et al., 1990; Lundström et al., 1991). The distal 5' promoter (P2) regulates synthesis of 1.9-kb (rat) and 1.5-kb (human) mRNA species. This mRNA can code for both MB-COMT and S-COMT proteins by using the leaky scanning mechanism of translation initiation (Tenhunen and Ulmanen, 1993; Tenhunen et al., 1993, 1994). The expression of the shorter transcript (1.6 kb in rat and 1.3 kb in human) is regulated by the P1 promoter, which is located between S-COMT and MB-COMT ATG start codons and partly overlaps the MB-COMT coding sequence. MB-COMT AUG translation initiation codon is not included in these transcripts, which therefore can code only for S-COMT polypeptide (Tenhunen et al., 1993, 1994; Tenhunen and Ulmanen, 1993).

In most human tissues, there are both transcripts, but in the human brain, only the longer transcript was found in 16 regions studied (Hong et al., 1998). When S-COMT and MB-COMT polypeptides of various rat and human tissues were quantified through Western blot analysis, S-COMT was usually dominant by a factor of 3 or higher. The only exception was the human brain, where 70% of the total COMT polypeptides was MB-COMT and 30% was S-COMT, demonstrating the bifunctionality of the 1.5-kb transcript (Tenhunen et al., 1993, 1994). The varying expression levels of the human COMT promoters suggest regulation by some tissue-specific transcription factors. In fact, human COMT promoters contain several putative binding sites for such factors that may cause the variability of COMT gene expression in different tissues (Tenhunen et al., 1994).

Both rat and human S-COMTs contain 221 amino acids, and the molecular masses are 24.8 and 24.4 kDa, respectively. The human S-COMT is 81% identical with the respective rat enzyme (Lotta et al., 1995; Lundström et al., 1995). Rat MB-COMT contains 43 additional amino acids, and human MB-COMT contains 50 additional amino acids. The corresponding molecular masses of MB-COMT are 29.6 and 30.0 kDa. Of these extra amino acids, 17 (rat) and 20 (human) function as hydrophobic membrane anchors (Salminen et al., 1990; Tilgmann and Kalkkinen, 1990, 1991; Bertocci et al., 1991; Lundström et al., 1991). The remainder of the MB-COMT molecule is suspended in the cytoplasmic side of the intracellular membranes (Bertocci et al., 1991; Lundström et al., 1991; Ulmanen and Lundström, 1991).

Rat S-COMT has been recently crystallized at 1.7- to 2.0-Å resolution (Vidgren et al., 1991, 1994), and the critical atomic structures have been described in detail (Vidgren and Ovaska, 1997). Therefore, only some of the most important aspects are repeated here.

COMT has as a single domain /-folded structure in which eight -helices are arranged around the central mixed -sheet. The active site of COMT consists of the S-adenosyl-L-methionine- (AdoMet)-binding domain and the actual catalytic site. The binding motif of the AdoMet site is similar to the Rossman fold, which is a common feature of many nucleotide binding proteins. The crystal structures of several of the methyl transferases that have been characterized are strikingly similar in the AdoMet-binding regions. The catalytic site is formed by a few amino acids that are important for the binding of the substrate, water, and Mg²⁺ and for the catalysis of O-methylation (Fig. 1). The Mg²⁺, which is bound to COMT after AdoMet binding, converts the hydroxyl groups of the catechol substrate to be more easily ionizable. Near one of the hydroxyl groups of the substrate, there is a lysine residue (Lys144) in COMT that accepts the proton from that hydroxyl, and subsequently the methyl group from the AdoMet is transferred to the hydroxyl group. Lysine acts as a general catalytic base in this base-catalyzed nucleophilic reaction. Mg²⁺ has an octahedral coordination to two aspartic acid residues (Asp141 and Asp169), to one asparagine (Asn170), to both catechol hydroxyls, and to a water molecule. Hence, Mg²⁺ ions control the orientation of the catechol moiety. In addition, the "gatekeeper" residues Trp38, Trp143, and Pro174 that form the hydrophobic "walls" and that define the selectivity of COMT toward different side chains of the substrate participate directly in the methylation reaction by keeping the planar catechol ring in the correct position (Fig. 1). They contribute significantly to the binding of the substrates (and inhibitors) of COMT (Vidgren et al., 1991, 1994, 1999; Vidgren and Ovaska, 1997). Although Mg²⁺ ions are crucial for COMT, most other small molecule methyltransferases are not Mg²⁺ dependent (Fujioka, 1992).

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Schematic view of the catalytic site of COMT demonstrating the necessary amino acids (rectangular boxes), Mg2+, and water molecule, with AdoMet and the catechol substrate. Among the amino acids, there are "gatekeepers" and catalytic amino acids, notably Lys144, which acts as a general catalytic base. The activated hydroxyl, to be methylated, is shown in bold. Details are from Vidgren and Ovaska (1997) and Vidgren et al. (1999).

COMT catalyzes the transfer of the methyl group of AdoMet to one of the hydroxyl groups of the catechol substrate in the presence of Mg²⁺ (Guldberg and Marsden, 1975). Methylation of the 3'-hydroxyl is much more common than that of the 4'-hydroxyl, for reasons that are discussed later. The mechanism and kinetics of the O-methylation reaction have been studied using partially purified enzyme preparations from various sources, most commonly from rat liver and rat or human brain, and recently also with the use of recombinant enzymes. The stereochemical course of the reaction has shown that the methyl transfer proceeds through a direct nucleophilic attack by one of the hydroxyl groups of the catechol substrate on the methyl carbon of AdoMet in a tight S_N2-like transition state (Woodard et al., 1980).

A sequential ordered mechanism has been proposed based on product inhibition studies (Jeffery and Roth, 1987), but there is some doubt as to their accuracy. AdoMet does seem to be the first substrate to bind, and S-adenosyl homocysteine is the last product to dissociate from the enzyme (Rivett and Roth, 1982; Tunnicliff and Ngo, 1983). However, now that the crystal structure of S-COMT has been resolved, it is possible to extend the kinetic studies to include recombinant S-COMT and MB-COMT with many different substrates. Thus, Lotta et al. (1995) proposed a reformulation of the kinetic behavior and the ordered mechanism of O-methylation. The active site of COMT, which is located in the outer surface of the enzyme, is a shallow groove on the surface of COMT. This is the same in both S-COMT and MB-COMT (Vidgren et al., 1994; Vidgren and Ovaska, 1997). AdoMet is able to bind even without Mg²⁺. The binding pocket of the methionine portion of AdoMet is deeper within the protein than the Mg²⁺ site, and it would be impossible for AdoMet to bind after Mg²⁺. Moreover, the catechol substrate cannot bind before AdoMet because then it would also be impossible for AdoMet to reach its binding site in the narrow groove located deep in the COMT molecule (Vidgren et al., 1994; Vidgren and Ovaska, 1997). Therefore, the order in which the compounds bind is as follows: AdoMet binds first, followed by Mg²⁺ and, finally, the catechol substrate. This reaction cycle differs from the previous suggestion that Mg²⁺ binds to COMT in a rapid equilibrium before the addition of AdoMet (Jeffery and Roth, 1987).

Although S-COMT and MB-COMT have identical kinetic mechanisms (Ca²⁺ inhibition, Mg²⁺ requirement, pH optimum, a similar K_m value for AdoMet, recognition by S-COMT antiserum), S-COMT and MB-COMT are certainly different enzymes, and MB-COMT is not a precursor of S-COMT (see above).

Based on early studies with crudely purified enzymes, S-COMT has a high K_m value for dopamine but a very high capacity (V_max from 50 pmol/min × mg protein in skeletal muscle to values as high as 14,690 pmol/min × mg protein in the liver). MB-COMT has a much lower K_m value but a low capacity (2-40 pmol/min × mg protein; Guldberg and Marsden, 1975; Roth, 1992). It is noteworthy that the V_max values are strongly dependent on the enzyme activities in various tissues rather than on the basic kinetic constants of these enzymes. Physiological substrate concentrations and possible differences in substrate selectivity have to be considered when the relative importance of either enzyme subtype is assessed. Dopamine levels in striatum and hypothalamus of brain homogenates are about 65 and 3 µM, respectively. The striatal and hypothalamic noradrenaline concentrations are 0.8 and 12 µM, respectively. It seems that at the concentrations of catecholamines naturally present, MB-COMT may be more important in their metabolism (Roth, 1992). According to Roth and associates (Rivett et al., 1982; Rivett and Roth, 1982; Roth, 1992), MB-COMT is the predominant enzyme at dopamine concentrations of <10 µM and at noradrenaline concentrations of <300 µM.

In addition, the recombinant human MB-COMT (K_m = 10 µM) has a higher affinity for catechol substrates than S-COMT (K_m = 108 µM; Malherbe et al., 1992). Lotta et al. (1995) also reported that the catalytic number (V_max) of rat recombinant S-COMT was slightly higher than that of MB-COMT with all four substrates studied. It is notable that in the study of Lotta et al. (1995), V_max values were calculated taking into account the actual enzyme concentration. Thus, V_max can be given as U/min, and it represents the catalytic number (k_cat). At saturating substrate concentrations, S-COMT functioned two times more efficiently, but the catalytic number for both COMT isoforms was similar for all substrates, nor was there any substrate selectivity. The K_m and k_enzyme values varied greatly between the two isoforms and were strongly dependent on the substrate. Malherbe et al. (1992) reported that S-COMT has a 15-fold higher K_m value and a 5-fold higher k_enzyme value for catecholamines than MB-COMT. However, this is not the case for dihydroxybenzoic acid (DBA) and L-3,4-dihydroxyphenylalanine (L-dopa; Table 1).

The catalytic sites of S-COMT and MB-COMT have identical amino acid sequences, but the membrane-bound portion of MB-COMT or the charged membrane itself causes more favorable, although poorly characterized, binding interactions even though there is no conformational change in the basic enzyme structure of MB-COMT.

New data generated using recombinant enzyme isoforms have confirmed earlier results but also added further detail to the picture, particularly regarding the binding differences of the various COMT substrates (Lotta et al., 1995; Tables 1 and 2). The atomic structure and the sequence comparison reveal that all residues that are important for the binding of the substrates and for the catalytic activity are similarly conserved in human and rat COMT. Only two amino acids are different in the active site. When the three substrates (L-dopa, dopamine, and DBA), which have different affinities for the active site of COMT, are compared, it is apparent that the kinetic differences are due to interactions of the substrate side chains with COMT residues. Generally, the binding of the catechol ring is similar to binding of enzyme to nitrocatechol-type inhibitors (e.g., OR-486 or OR-1840) when this is viewed with crystallized rat S-COMT (Vidgren et al., 1994). DBA has a charged carboxyl moiety, but the molecule is planar and fits well between the gatekeepers Trp38 and Pro174 (Fig. 1), and thus it has high affinity. Dopamine has a positively charged amino group that, despite its rotational freedom, still makes a repulsive contact with one of the gatekeepers. L-Dopa has the largest, double-charged side chain, and therefore the propulsions are strong and its affinity to COMT is lower.

COMT is able to methylate only one of the two catechol hydroxyls. Both COMT isoforms favor 3-O-methylation, and MB-COMT is even more regioselective than S-COMT (Table 2). The meta/para ratio is higher, 22 to 88 (depending on the substrate) in MB-COMT than in S-COMT, 4 to 15 (depending on the substrate; Lotta et al., 1995). The reason for favoring 3-O-methylation over 4-O-methylation may be that as the p-hydroxyl group (i.e., 4-O-hydroxyl) approaches the AdoMet; this forces the side chain to become orientated in an unfavorable position with the hydrophobic protein residues of the catalytic site. As stated, L-dopa has the strongest repulsive interactions, which are reflected not only in its high K_m value but also as in the high 3-O-methylation/4-O-methylation ratio (Table 2).

Recently, molecular dynamic simulation studies have been used to explain the preference of meta-O-methylation over para-O-methylation. The catechol ring has a tilt of about 30 degrees compared with that of the X-ray structure of the active site. This directs any substituent at the para-position of the catechol ring into a hydrophobic pocket formed by Trp38 and Tyr200. Hydrophobic substituents are accommodated in this pocket so that para-O-methylation is favored, whereas polar substituents are repelled, making meta-O-methylation favorable (Lan and Bruice, 1998). In addition, the distances of the hydroxyls to the active methyl group of AdoMet are very different (2.6 versus 4.8 Å), so no competition for the methylation site is possible (Vidgren et al., 1999). Different meta/para ratios of substituted catechols are solely a consequence of their relative ability to bind in two dissimilar orientations to the active site of COMT (Vidgren et al., 1999).

Those catechols that contain electronegative substituents (like NO₂, CN, and F) are potent inhibitors, but poor substrates, of COMT (Bäckström et al., 1989; Borgulya et al., 1989). This point has been recently clarified in a semiempirical study using dinitrocatechol (OR-486) as a model inhibitor (Ovaska and Yliniemelä, 1998). As mentioned, Lys144 acts as a general base in the methylation and can activate one of the catechol hydroxyls before the nucleophilic attack of the methyl group of AdoMet. The electronegative nitro groups reduce the nucleophilicity of the ionized catechol hydroxyls (Fig. 1). Therefore, dinitrocatechol itself is not methylated at all but instead is a potent COMT inhibitor. Among the mononitro catechols, the decrease in nucleophilicity is evidently less than that in dinitrocatechol, and the ultimate effect depends on the side chain. For instance, entacapone is to all intents not O-methylated at all, whereas tolcapone is O-methylated by about 3% (Dingemanse, 1997).

COMT is not easily induced or suppressed. The capacity of COMT in the peripheral tissues is probably so high that only a minor fraction of the protein is ever needed. Therefore, general peripheral COMT inhibition may be difficult to achieve. This may be crucial for the safety of COMT inhibitors. Locally, such as in the gastrointestinal mucosa (Schultz and Nissinen, 1989), the new powerful COMT inhibitors can suppress COMT activity sufficiently to cause significant changes in the metabolism of L-dopa.

Some special treatments or situations may increase COMT activity, but most of them cause at best only a doubling of activity. Very early, it was found that long-term treatment with pyrogallol could elevate the COMT activity in the liver (Guldberg and Marsden, 1975). Dipyridamole (Li et al., 1991), exogenous AdoMet (Baldessarini, 1987), and butylated hydroxyanisole (Lam, 1988) can slightly increase COMT activity.

Pregnancy and progesterone treatment seem to induce COMT activity in the uterus. In the rat, estrogen exposure results in decreased hepatic COMT activity (Cohn and Axelrod, 1971; Parvez et al., 1975). On the other hand, subchronic estrogen treatment increases COMT immunostaining in hamster kidney (Weisz et al., 1998b). There also is a gender difference. The COMT activity in the liver of male subjects is about 30% higher than that in females (Boudikova et al., 1990). During aging, COMT activity in the liver increases by about 10-fold from birth to adulthood (Guldberg and Marsden, 1975). A similar trend occurs in the kidney, where K_m values increase about 5-fold during aging (Vieira-Coelho and Soares-da-Silva, 1996). This could be explained by assuming that the predominant form of COMT is not the same in newborn rats and adult animals. Some diseases seem to be associated with altered COMT activities. COMT activity in the spinal cord appears to be decreased in Huntington's disease (McGeer et al., 1993), but it is slightly increased in amyotrophic lateral sclerosis where monoamine oxidase (MAO)-B activity was substantially elevated (Ekblom et al., 1993). Chromosomal microdeletion at 22q11-13 (where the human COMT gene is also located) has been associated with a number of defects or syndromes (Lachman et al., 1996a). This phenomenon and genetic COMT polymorphism in various diseases are discussed later (see Genetic Polymorphism: Variants with Different Thermostability).

We also made an anecdotal finding that high concentrations of ethanol can reduce MB-COMT activity but increase S-COMT activity in recombinant enzyme forms. However, the concentrations needed are from 50 to 1000 mM, which could scarcely be attained even in cases of ethanol intoxication (Reenilä et al., 1995).

The level of COMT enzyme activity is genetically polymorphic in human tissues with a trimodal distribution of low (COMT^LL), intermediate (COMT^LH), and high (COMT^HH) activities (Weinshilboum and Raymond, 1977; Boudikova et al., 1990; Jeanjean et al., 1997). This polymorphism, which according to segregation analysis of family studies is caused by autosomal codominant alleles, leads to 3- to 4-fold differences in COMT activity in human erythrocytes and liver (Weinshilboum and Raymond, 1977; Boudikova et al., 1990; Jeanjean et al., 1997). Low COMT activity is associated with enzyme thermolability, even at 37°C (Scanlon et al., 1979; Spielman and Weinshilboum, 1981; Boudikova et al., 1990). Recently, the molecular basis of the thermolability was revealed with the baculovirus expression system; namely, substitution of Val108 by Met108 in the S-COMT (or the corresponding amino acids 158 in the MB-COMT) is caused by transition of guanine to adenine at codon 158 of the COMT gene (Grossman et al., 1992b; Lotta et al., 1995). Although there are some other mutations in the COMT gene (Lachman et al., 1996b), it seems quite established that the low thermal stability and the low COMT activity go together. Because there is only one COMT gene without any known tissue-specific splice variants (Tenhunen et al., 1994; Lundström et al., 1995), it is likely that the codon 108/158 polymorphism, causing the change in thermostability, also leads to functional alterations of COMT in all tissues (Lachman et al., 1996b). The catalytic activity per se is the same with the two variants, and the labile enzyme variant is fully stabilized by AdoMet binding. In the following discussion, we do not necessarily go to the methodological details when describing the association of COMT activity with a number of diseases. In publications printed after 1996, the molecular difference may be demonstrated. Earlier, either the thermostability or only the COMT activity was reported.

Polymorphism of COMT activity could have clinical implications, but the true relationships in about 30 genetic mapping studies with a number of diseases have not been very impressive. There are, however, three or four disorders in which some relationship has been observed. First, there is an obsessive-compulsive disorder in which the affected persons follow various anxiety-reducing rituals, which seems to be correlated to low COMT activity allele (Karayiorgou et al., 1997). Second, low COMT activity allele appears to have some association with aggressive and highly antisocial impulsive schizophrenia (Strous et al., 1997a,b; Lachman et al., 1998). A special velo-cardio-facial syndrome in which the chromosome 22q11 is deleted (including the COMT gene) also carries with it similarly bizarre behavior, but several types of other symptoms, including schizophrenia, may appear (Lachman et al., 1996a,b; Papolos et al., 1996). However, generally there is only a weak (Ohmori et al., 1998) or no association between COMT activity alleles and schizophrenia (Chen et al., 1996; Daniels et al., 1996; Riley et al., 1996; Wei et al., 1996; Karayiorgou et al., 1998). Paradoxically, in one study, the high activity allele was preferentially transmitted from healthy parents to their schizophrenic children (Li et al., 1996).

Recently, Tiihonen et al. (1999) reported a clear indication of the association of the late-onset alcoholism (type 1) and the low-affinity allele of COMT in a Finnish population. When 123 alcoholics were compared with 246 race- and gender-matched controls, the odds ratio for alcoholism of subjects with COMT^LL was 2.51 over those with COMT^HH. The frequency of low allele was also significantly higher in the alcoholics than in 3140 Finnish blood donors representing a general population. The estimate for population etiological percentage of the COMT^LL in type 1 alcoholism was 13.3%.

With respect to depression, the results are variable. Several studies do not show any relationship between depression and COMT (BIOMED European Bipolar Collaborative Group, 1997; Gutíerrez et al., 1997; Kunugi et al., 1997b), although some show a low COMT activity allele or a low COMT activity in the erythrocytes in patients with major depression but not in those with the bipolar form (Karege et al., 1987; Ohara et al., 1998b). In some patients, the link seems to be with bipolar manic-depressive illness (Li et al., 1997). Recently it has been shown that the low COMT activity allele is dominant in rapid-cycling (cycle 1-2 days) bipolar manic-depressive disorder (Kirov et al., 1998; Papolos et al., 1998). These findings confirm the results of an earlier report that patients with velo-cardio-facial syndrome, which includes a rapid-cycling bipolar disorder, have low COMT activity (Lachman et al., 1996a; Papolos et al., 1996). Interestingly, polysubstance abusers have been reported to more commonly have the high COMT activity allele than the controls (Vandenbergh et al., 1997). No association has been found between anxiety disorders and COMT polymorphism (Ohara et al., 1998a).

Some ethnic differences have been recognized. For instance, the frequency of the low COMT activity allele is lower in Kenyan than in Caucasian or South-West Asian individuals (McLeod et al., 1998). However, black Americans have higher COMT activity than white Americans (McLeod et al., 1994). Low COMT activity is found in a Saami population (Klemetsdal et al., 1994).

The association of the COMT alleles with Parkinson's disease (PD) has been extensively studied, but usually no association has been found (Hoda et al., 1996; Syvänen et al., 1997; Xie et al., 1997). However, some Japanese individuals with the low COMT activity allele may have an increased risk for PD (Kunugi et al., 1997a; Yoritaka et al., 1997). As cited above, different populations have different frequencies of COMT^LL and COMT^HH alleles and therefore may have variation in individual responses to therapy with a drug preparation containing L-dopa (levodopa; Reilly et al., 1980; Rivera-Calimlim and Reilly, 1984; Klemetsdal et al., 1994). It would be interesting to know whether the confirmed COMT^LL PD patients really can benefit more from the levodopa therapy than COMT^HH patients, as has been suggested based on COMT activity analysis from erythrocytes (Reilly et al., 1980). It is also known that patients with high COMT activity in their erythrocytes have encountered more adverse effects during levodopa therapy and have had more frequent on-off effects than patients with low COMT activity (Reilly et al., 1983; Rivera-Calimlim and Reilly, 1984).

Finally, low COMT activity allele is found more frequently in patients with breast cancer than in healthy controls (Lavigne et al., 1997), particularly in women with menopausal symptoms (Thompson et al., 1998). This phenomenon may be related to the decreased metabolism of catecholestrogens because COMT also metabolizes these compounds (see Estrogen Metabolism and Role of COMT and COMT Inhibitors). However, a recent extensive case-control study on patients with invasive breast cancer in North Carolina (654 cancer patients and 652 controls) did not show any relation of the COMT genotype to the breast carcinoma (Millikan et al., 1998).

In addition to humans, other mammals can have different genetically determined COMT activity. There are two rat strains with different activity levels (Weinshilboum et al., 1979; Roth et al., 1990); however, the rat polymorphism has some other structural basis from the human polymorphism because both rat and pig have Leu108 (not Met108 or Val108; Salminen et al., 1990) in the critical site of the COMT polypeptide, which determines the level of activity and thermal stability (Salminen et al., 1990; Malherbe et al., 1992).

In mammals, COMT is widely distributed throughout the organs of the body. COMT is an intracellular enzyme. As discussed (see One COMT Gene and Two Proteins), the COMT protein in vertebrates appears mostly in a soluble form (as S-COMT), and only a minor fraction is in the particular form (as MB-COMT; Guldberg and Marsden, 1975; Roth, 1992).

1. Brain COMT. The cellular localization of COMT has been studied in several ways. A number of lesion studies (Rivett et al., 1983; Kaakkola et al., 1987) and immunohistochemical studies (Karhunen et al., 1995a; Lundström et al., 1995) have demonstrated that there is no significant COMT activity in presynaptic dopaminergic neurons, but some activity is present in postsynaptic neurons and substantial activity is located in glial cells.

2. COMT in Other Tissues. COMT has been found in practically all mammalian tissues investigated. The highest COMT activity in both rat and humans is in the liver, followed by the kidneys and gastrointestinal tract (both stomach and intestine; Nissinen et al., 1988b; Schultz and Nissinen, 1989; Männistö et al., 1992b). These findings have been confirmed and expanded to spleen and submaxillary glands through the use of immunohistochemistry (Karhunen et al., 1994; Lundström et al., 1995). In pancreas, COMT immunoreactivity was found in  and  cells but not in  cells (Karhunen et al., 1994). The importance of pancreatic COMT activity has also been demonstrated through other means. After the pretreatment of conscious rats with OR-486 (dinitrocatechol), a potent nitrocatechol-type COMT inhibitor, the uptake of radioactivity from [¹³C]L-dopa into the pancreas was increased by 4-fold. Most of the radioactivity was derived from [¹³C]dihydroxyphenylalanine ([¹³C]DOPAC; Bergström et al., 1997). Positron emission tomography (PET) studies have demonstrated high COMT activity in kidney, liver, intestine, stomach, spleen, lungs, and heart of mice (Ding et al., 1996); they also confirmed high COMT activity in baboon liver and kidney.

G. General Importance of COMT

1. Substrates of COMT. The physiological substrates of COMT include L-dopa, catecholamines (dopamine, norepinephrine, epinephrine), their hydroxylated metabolites, catecholestrogens (Ball and Knuppen, 1980), ascorbic acid, and dihydroxyindolic intermediates of melanin. Several dietary and medicinal products are also COMT substrates, such as triphenols and substituted catechols, dobutamine, isoprenaline, rimiterol, -methyldopa, benserazide, carbidopa, dihydroxyphenyl serine (Maruyama et al., 1996), flavonoids, and dihydroxy derivatives of tetrahydroxyisoquinolones. The general function of COMT is the elimination of biologically active or toxic catechols and some other hydroxylated metabolites. During the first trimester of pregnancy, COMT protects the placenta and the developing embryo from activated hydroxylated compounds formed from aryl hydrocarbons by hydroxylases (Barnea and Avigdor, 1990). COMT also acts as an enzymatic detoxicating barrier between the blood and other tissues shielding against the detrimental effects of xenobiotics (e.g., in the intestinal mucosa and the brain). COMT may also serve some unique or indirect functions in the kidney and intestine tract by modulating the dopaminergic tone; the same may be true in the brain: COMT activity may regulate the amounts of active dopamine and norepinephrine in various parts of the brain and therefore be associated with mood and other mental processes.

2. Quantitative Role of COMT in Metabolism of Catecholamines. The relative importance of enzymes in metabolizing catecholamines and the uptake processes of catecholamines have been clarified. Without exogenous levodopa loading, the high-affinity neuronal reuptake (uptake₁) is an efficient elimination system for the released catecholamines, being responsible for most of their elimination both in peripheral tissues and the brain (Kopin, 1985; Männistö et al., 1992b; Cass et al., 1993). The role of the extraneuronal transport is less clear (Friedgen et al., 1996b). The contribution of metabolism, including COMT, is unimportant, and therefore COMT inhibition does not affect dopamine levels to a detectable degree. It is equally clear that inhibition of MAO (by pargyline) and COMT (by tolcapone), each separately, has little, if any, effect on the removal of norepinephrine, epinephrine, and dopamine on passage through the systemic and pulmonary circulation. The pulmonary, but not the systemic, clearance of catecholamine can be reduced by a combined blockade of MAO and COMT (Friedgen et al., 1996a). Thus, inhibition of COMT does not greatly alter the elimination of infused catecholamines or exercise-induced elevation of catecholamines, which is an important safety-increasing factor in the use of potent COMT inhibitors (see later). However, the elimination pathways are adaptable. When MAO is blocked, both COMT and phenolsulfotransferase activities are increased, but these two pathways do not compete with each other (Buu, 1985).

3. COMT Knockout Mice. The ultimate importance of COMT will probably be clarified in a new strain of animals lacking the COMT gene. Such mice were recently produced by Gogos et al. (1998). The mice appeared to be normal, they had only minor changes in their behavior, and the brain neurochemistry of catecholamines was virtually unaltered. The mice were also able to breed normally. Surprisingly, all of the changes were sex and region specific. Mutant male mice, totally lacking COMT activity, had an almost 3-fold increase of dopamine levels in the frontal cortex but not in the striatum or hypothalamus, but females showed no changes. It is noteworthy that despite the complete lack of COMT gene and protein, residual homovanillic acid (HVA) levels were detectable in several brain areas. This points to the possibility that there exists a still-unidentified methylation pathway in the brain. Female knockout mice had impaired emotional reactivity in the black-white box. On the other hand, heterozygous males (but not homozygous males or any females) were more aggressive than their wild-type counterparts. These findings suggest that 1) the importance of COMT in the behavior has probably been underestimated and 2) a complete lack of COMT can be effectively compensated, at least in mice.

	III. COMT Inhibitors

Top Previous Next References

In 1975, Guldberg and Marsden reviewed early COMT inhibitors. Several of these compounds (e.g., gallates, tropolone, U-0521, 3',4'-dihydroxy-2-methyl-propiophenone) have been used as in vitro tools; however, their efficacy in vivo is low, and they are short acting. Moreover, they lack selectivity and are rather toxic. For instance, U-0521 depressed the function of the rat vas deferens at 1 µM and above, which apparently invalidates its use as a COMT inhibitor in smooth muscle preparations (Rice et al., 1997). Also, the limited clinical experiences with gallates, tropolone, ascorbic acid, and U-0521 were disappointing (Ericsson, 1971; Reilly et al., 1983; Reches and Fahn, 1984). It is worth noting that millimolar concentrations of ascorbic acid may well reduce COMT activity significantly, as was recently shown in primate spinal meninges (Kern and Bernards, 1997).

Three laboratories independently developed very potent, highly selective, and orally active COMT inhibitors. Nitrocatechol is the key structure in most of these molecules (Fig. 2). Only CGP 28014 is a chemically distinct compound (Bäckström et al., 1989; Borgulya et al., 1989; Waldmeier et al., 1990a).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Chemical structures of some second-generation COMT inhibitors.

In principle, the new COMT inhibitors can be divided into three groups: 1) mainly peripherally acting compounds, 2) broad-spectrum compounds working in both the periphery and the brain, and 3) atypical compounds, probably acting preferably in the brain. This classification was derived from our comparative studies (Männistö et al., 1992a; Törnwall and Männistö, 1993), where the decrease in the plasma and brain 3-OMD was used as a marker for peripheral COMT inhibition. The decline of the brain HVA [and 3-methoxytyramine (3-MT) after pargyline treatment] was used as a signal of COMT inhibition in the brain.

Because the amount of COMT is high in the liver, kidney, and intestinal mucosa (Sharpless et al., 1973; Nissinen et al., 1988b), peripherally acting COMT inhibitors would be primarily interesting. The clearest clinical application of these COMT inhibitors would be as adjuncts to levodopa in PD (Männistö and Kaakkola, 1989, 1990). L-Dopa is a catechol, and it is predominantly O-methylated to 3-OMD, especially when the peripheral decarboxylation of L-dopa is inhibited by benserazide or carbidopa (Nutt and Fellman, 1984). Although the elevation in 3-OMD levels was not harmful as such, peripheral COMT inhibition should enhance the brain penetration of L-dopa. Those COMT inhibitors that gain access to the brain would provide additional information about the role of metabolism to potentiate brain dopaminergic and adrenergic functions.

C. Properties of New Compounds

1. COMT Inhibition. Nitrocatechols are so-called tight-binding inhibitors of COMT, although their binding to COMT is fully reversible (Schultz and Nissinen, 1989; Lotta et al., 1995; Borges et al., 1997). Prolongation of the preincubation time with an inhibitor markedly reduces the IC₅₀ values of nitrocatechol-type inhibitors (Schultz and Nissinen, 1989). This particular property has been the source of many problems when the kinetic parameters have been characterized. It is necessary to perform inhibition kinetic studies in a special way to avoid invalid conclusions. In the presence of varying inhibitor concentrations, the reaction velocity of COMT increases progressively with increasing enzyme concentrations and parallels the velocity curve without an inhibitor if sufficiently high amounts of enzyme are used. Also, the effect of the nitrocatechol-type inhibitors can be gradually abolished by dialysis (about 50% in 5 h, fully reversed within 24 h; Schultz and Nissinen, 1989).

2. Effects on L-Dopa and Catecholamine Metabolism. In rats, oral administration of entacapone and nitecapone (3-30 mg/kg) in combination with levodopa and carbidopa effectively reduce 3-OMD formation and elevate serum and brain L-dopa, dopamine, DOPAC, and HVA levels (Männistö et al., 1988, 1992a; Nissinen et al., 1988a, 1992). Similar findings have been reported in Cynomolgus monkeys in whom the i.v. efficacy of entacapone and nitecapone was equipotent (Cedarbaum et al., 1991). Part of the levodopa saved from COMT is metabolized through alternative pathways because the increase of serum L-dopa is smaller than the decrease in 3-OMD would lead one to expect.

1/2

1/2

3. Microdialysis Studies. In vivo microdialysis technique is a widely used method for the estimation of extracellular levels of various substances, such as neurotransmitters and their metabolites. The effects of entacapone, tolcapone, and CGP 28014 alone or in combination with levodopa and DDC inhibitors on brain L-dopa and dopamine metabolism have been investigated in rats. The aims have been, on one hand, to clarify the mechanisms of brain dopamine metabolism and, on the other hand, to confirm the beneficial effect of combination of a COMT inhibitor with levodopa on brain dopamine formation. Only a few studies have dealt with norepinephrine levels and metabolism (Li et al., 1998).