Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Homocysteine, a sulfur-containing amino acid, is involved in many metabolic pathways including trans-sulfuration in cysteine synthesis, re-methylation in methionine synthesis, trans-methylation of DNA, proteins, and lipids, and biosynthesis of small hormonal and neuronal signaling molecules. The normal range for total plasma homocysteine in adults is 5 to 15 µM with an average of 10 µM (Kang et al., 1992). Abnormal elevation of total plasma homocysteine is classified as moderate (16~30 µM), intermediate (31~100 µM), or severe (>100 µM) hyperhomocysteinemia (Kang et al., 1992). Numerous epidemiological studies have reported that as a risk factor, moderate hyperhomocysteinemia is associated with 1) cardiovascular diseases such as atherosclerotic (McCully, 1996; Durand et al., 2001) and ischemic cardiovascular diseases (Brattstrom and Wilcken, 2000; Ueland et al., 2000), stroke (Hankey and Eikelboom, 2001), and venous thrombosis (Makris, 2000); 2) neuropsychiatric disorders such as Alzheimer's disease (Miller, 2000), schizophrenia (Levine et al., 2002), and cognitive dysfunction (Morris et al., 2001); 3) developmental disorders such as neural tube defects (van der Put et al., 2001); and 4) complications of pregnancy (Aubard et al., 2000).

However, the pathogenic mechanisms underlying the moderate hyperhomocysteinemia-associated diseases are not fully understood. Major impediments hindering our understanding of the mechanisms by which hyperhomocysteinemia influences these diseases include the following. 1) Moderate hyperhomocysteinemia, as a risk factor, coexists with other conventional risk factors in these diseases; 2) all the diseases for which hyperhomocysteinemia is a risk factor are polygenetic traits that are also affected by environmental factors; and 3) because homocysteine is also involved in normal metabolic pathways of many biologically functional molecules, abnormalities in homocysteine metabolism may adversely affect many related and unrelated pathways which, in turn, cause diseases and disorders.

Considerable effort has been made to characterize the adverse effects of homocysteine on cells or tissues since McCully's (1969) initial report regarding pathogenic homocysteine. Homocysteine-induced oxidant stress (HIOS) is a major finding from studies that attempt to discover how homocysteine elevations contribute to disease (Loscalzo, 1996). According to the HIOS model, the highly reactive thiol group of homocysteine is readily oxidized to generate reactive oxygen species that in turn cause damage to proteins and lipids (Loscalzo, 1996). The HIOS hypothesis is established mainly on the results from in vitro studies using high concentrations of homocysteine (Kokame et al., 1996; Outinen et al., 1999; Lang et al., 2000; White et al., 2001). Other studies (Weiss et al., 2001) suggest that HIOS might be important in vivo when homocysteine levels are not greatly elevated. It is likely that there are additional, unknown, molecular mechanisms that control moderate hyperhomocysteinemia-associated diseases (Spence et al., 1999; Miller, 2000; Vollset et al., 2001). Given that moderate hyperhomocysteinemia can act as a risk factor for various diseases at low levels in comparison with the high doses used in in vitro studies, we wondered whether a specific interaction between homocysteine and some unidentified functional proteins, such as receptors, might be required to link the moderately elevated level of homocysteine to the pathogenesis of the associated diseases.

Certain nonessential amino acids and derivatives, such as L-glutamate, L-aspartate, and -aminobutyric acid (GABA), have long been known to act as neurotransmitters (Bennett and Balcar, 1999). A group of sulfur-containing amino acids was previously found to exhibit effects similar to those of L-glutamic acid and L-aspartic acid (Thompson and Kilpatrick, 1996). These sulfur-containing amino acid analogs include 1) the L-glutamate analogs, L-homocysteine sulfinic acid (L-HCSA) and L-homocysteic acid (L-HCA), and 2) the L-aspartate analogs, L-cysteine sulfinic acid (L-CSA) and L-cysteic acid (L-CA). Kingston et al. (1998) demonstrated that L-HCSA, L-CSA, and L-CA, but not L-HCA, were capable of stimulating phosphoinositide hydrolysis in the cells transfected with human metabotropic glutamate receptor subtype 1 (mGluR1) or subtype 5 (mGluR5) which belong to group I mGluRs. The potencies of these analogs were similar to that of glutamate. These findings suggest that homocysteine derivatives may interact with group I mGluRs and regulate their signaling function. However, it remains unclear whether these L-glutamate analogs act as full or partial agonists at group I mGluRs, and whether they can interact with other mGluRs (group II and III) that are negatively coupled to adenylate cyclase (De Blasi et al., 2001) or ionotropic glutamate receptors. Recent studies of homocysteine metabolism revealed that homocysteine thiolactone, a homocysteine derivative, is synthesized by methionyl-tRNA synthetase under physiological conditions (Jakubowski et al., 2000). Therefore, other homocysteine derivatives are also candidates for interaction with cell-surface receptors, which may link the effects of homocysteine to its final cellular targets.

The goal of this study is to identify the cell-surface receptors, ion channels, and transporters that act as biological sensors for homocysteine and its derivatives. We screened 60 different membrane-bound receptors, ion channels, and transporters from 15 different families for molecular targets that interact with homocysteine or its derivatives. In second-messenger generation assays, we tested the regulatory activity of the acidic homocysteine derivatives, including L-HCSA, L-HCA, L-CSA, and L-CA, on metabotropic glutamate receptors. We found that metabotropic glutamate receptors are the receptor candidates for several acidic oxidized derivatives of homocysteine.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemical Compounds. DL-homocysteine (DL-HCY), L-cysteine (L-CYS), DL-homocysteic acid (DL-HCA), D-homocysteic acid (D-HCA), L-HCA, L-CA, D-homocysteine sulfinic acid (D-HCSA), L-HCSA, L-CSA, D-homocysteine thiolactone (D-HCYT), L-homocysteine thiolactone (L-HCYT), D-cystine (D-CTN), L-cystine (L-CTN), D-methionine (D-MET), L-methionine (L-MET), 4-aminophosphonobutyrate (AP4), and L-glutamate (L-GLU) were purchased from Sigma-Aldrich (St. Louis, MO). Figure 1 shows the chemical structure of the tested compounds. Radioligands were purchased from PerkinElmer Life Sciences (Boston, MA), except for GR-125,743, which was purchased from Amersham Biosciences Inc. (Piscataway, NJ). myo-[³H]Inositol was purchased from PerkinElmer Life Sciences.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Chemical structure of homocysteine and its derivatives.

Receptors. Cell lines stably transfected with recombinant cDNA encoding receptors or cell lines that express endogenous receptors were used for the drug screen. The recombinant cDNA included 1) human adrenergic receptors, _1A, _1B, _2A, _2B, _2C, and rat adrenergic receptors, ₁ and ₂; 2) rat cannabanoid CB1 receptor; 3) dopaminergic receptors, hD1, hD2, hD3, rD4, and hD5; 4) human histamine receptors, H1, H2, and H4; 5) rat imidazoline receptor; 6) human muscarinic acetylcholine receptors, M1, M2, M3, M4, and M5; 7) human nicotinic acetylcholine receptors, ₂/₂, ₂/₄, ₃/₂, ₃/₄, ₄/₂, and ₄/₄; 8) human opiate receptors, µ, , and ; 9) human peptide receptors, V1, V2, V3, and OT; 10) serotonergic receptors, h5-HT1A, h5-HTB, h5-HTD, r5-HT2A, r5-HT2C, h5-HT3, h5-HT5A, h5-HT6, and h5-HT7; 11) human transporters of serotonin, norepinephrine, and dopamine as previously described (Roth et al., 1998, 2001, 2002; Rothman et al., 2000; Shapiro et al., 2002); and 12) rat metabotropic glutamate receptors, mGluR1a, mGluR2, mGluR4, mGluR5a, mGluR6, and mGluR8 as previously described (Gomeza et al., 1996; Wroblewska et al., 1997; Kozikowski et al., 1998). The endogenous receptors included 1) GABA receptors, GABA_A, GABA_B, and GABA_BZP from rat forebrain; 2) histamine receptor H1 from rat forebrain; 3) rat nicotinic acetylcholine receptor, ₄/₂; 4) ionotropic glutamate receptor NMDA from rat forebrain; and 5) voltage-sensitive Ca²⁺ channel from rat heart.

Radioligand Competitive Binding Assays. Receptor-binding affinity of homocysteine and its derivatives was determined by radioligand binding competitive experiments. The assays were performed with cell lysates extracted from the cells stably transfected with the recombinant receptor expression vectors (Roth et al., 1998, 2001; Rothman et al., 2000; Shapiro et al., 2002), or with forebrain membrane preparation containing the endogenous receptors of interest (Roth et al., 1991). Competitive binding assays were carried out under standard conditions as previously detailed (Roth et al., 2001, 2002; Rothman et al., 2000). The conditions for the different binding assays and K_i values for reference ligands are summarized in Table 1. On-line protocols for binding assays are available at http://pdsp.cwru.edu/nimh/binding.htm. Homocysteine and its derivatives were dissolved in 0.05% L-ascorbic acid and incubated with radioligands as indicated in Table 1. Binding data were determined in duplicate.

Phosphoinositide Hydrolysis Assays Using Recombinant Receptors. mGluR1a and mGluR5a (group I) were stably expressed in Chinese hamster ovary (CHO) cells (Wroblewska et al., 1997). mGluR8 (group III) was stably expressed in a CHO cell line that expressed a chimeric G protein, G_qi9, and experimentally allowed a positive coupling of mGluR8 to phospholipase C (Gomeza et al., 1996). Cultured in 96-well plates, the receptor-expressing cells were incubated overnight in glutamine-free culture medium supplemented with 0.75 µCi myo-[ ³H]inositol to label the cell membrane phosphoinositides. Incubations with homocysteine derivatives were carried out for 45 min at 37°C in Locke's buffer (156 mM NaCl, 5.6 mM KCl, 3.6 mM NaHCO₃, 1 mM MgCl₂, 1.3 mM CaCl₂, 5.6 mM glucose, and 20 mM Hepes, pH 7.4) containing 20 mM LiCl, which blocks the degradation of inositol phosphates (IPs). The reaction was terminated by aspiration and addition of 0.1 M HCl. IPs were extracted with 0.1 M HCl. [³H]IPs were separated by anion exchange chromatography and determined in duplicate by a liquid scintillation counter (LKB, Uppsala, Sweden) as described previously (Raulli et al., 1991).

Cyclic AMP Formation Assay. Receptor-mediated inhibition of the forskolin-induced elevation of cyclic AMP formation was described previously (Wroblewska et al., 1997). In brief, mGluR2 (group II) and mGluR6 (group III) were stably expressed in CHO cells, whereas mGluR4 (group III) was stably expressed in baby hamster kidney cells (Wroblewska et al., 1997). Cultured in 96-well culture plates, the receptor-expressing cells were preincubated for 10 min at 37°C in Locke's medium containing 300 µM isobutylmethylxanthine, which inhibits the activity of phosphodiesterases to prevent the degradation of cAMP. Then, 5 µM forskolin was added without or with the test compounds, and the incubation was continued for 10 min. After incubation, the medium was rapidly aspirated. cAMP was extracted with 0.1 M HCl and measured by radioimmunoassay using a magnetic Amerlex RIA kit (Amersham Biosciences Inc.).

Data Analysis. For binding assays, data analysis and curve generation were carried out with GraphPad Prism 3.02 (GraphPad, San Diego, CA). K_i values were calculated with LIGAND software (Munson and Rodbard, 1981) as previously detailed (Rothman et al., 2000; Roth et al., 2002). For second-messenger assays, EC₅₀ values were determined by fitting the normalized data to the logistic equation by nonlinear regression using SigmaPlot software (SPSS Science, Chicago, IL), whereas curve generation was carried out with GraphPad Prism 3.02.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Receptor Binding Affinity of Homocysteine and Its Derivatives. Radioligand competitive binding assays were performed to test the specific binding affinity of homocysteine and its derivatives against a panel of receptors, ion channels, and transporters via the resources of the National Institute of Mental Health Psychoactive Drug Screening Program (Table 2). The survey was conducted in two steps, a preliminary screen followed by K_i determination. The preliminary screen was performed to determine percentage inhibition of the specific radioligand binding by homocysteine and its derivatives at 10 µM. Determination of K_i values were conducted on the compounds whose inhibition of specific radioligand binding to the corresponding receptors was greater than 50%. As determined in the preliminary screen, significant inhibition (i.e., >50%) by compounds was seen for _1B-adrenergic receptor (by L-HCA, D-HCA, and DL-HCY), imidazoline receptor (by DL-homocysteine), and serotonin receptors of 5-HT1A (by L-CSA, DL-HCA, D- and L-HCYT, and L-CTN), 5-HT1Da (by L-CTN), 5-HT2A (by L-HCA and DL-HCY), 5-HT2C (by L-HCA and DL-HCY), 5-HT3 (by L-CSA and D-HCSA), and 5-HT6 (by DL-HCA) (data not shown). No significant inhibition (i.e., 50%) was found for adrenergic receptors (_1A, _2A, _2B, _2C, ₁, and ₂), dopaminergic receptors (D1~D5), cannabanoid CB1 receptor, GABA receptors (GABA_A, GABA_B, and GABA_BZP), ionotropic NMDA glutamate receptor (except for L-cystine), histamine receptors, muscarinic acetylcholine receptors (M1~M5), nicotinic receptors, opiate receptors (µ, , ), peptide receptors (V1~V3, OT), serotonin receptors (5-HT1Db, 5-HT5A, and 5-HT6), voltage-sensitive calcium channel, and transporters of serotonin, norepinephrine, and dopamine.

Stimulation of Phosphoinositide Hydrolysis through Activation of mGluR1, mGluR5, and mGluR8. Metabotropic glutamate receptors (mGluRs) are well characterized G-protein-coupled receptors composing a large family of eight different mGluR subtypes in three different groups based on their molecular structure and pharmacological behavior (De Blasi et al., 2001). Group I mGluRs include mGluR1 and 5, group II mGluRs include mGluR2 and 3, and group III mGluRs include mGluR4, 6, 7 and 8. Stimulation of mGluRs regulates intracellular signaling through changing the levels of intracellular second messengers. In general, group I mGluRs are involved in activation of phospholipase C (PLC) via G_q or G_o, whereas group II and III mGluRs participate in inhibition of adenylyl cyclase via G_i (De Blasi et al., 2001). In the present study, recombinant cDNAs for mGluR1, mGluR5, and mGluR8 were stably transfected and expressed in cells (see Materials and Methods). It should be noted that the in vitro expressed mGluR8 in the present study is artificially coupled to PLC by coexpression of a recombinant chimeric G_qi9 because the endogenous mGluR8 is not naturally coupled to PLC (Gomeza et al., 1996). L-HCSA, L-HCA, L-CSA, and L-CA were used to investigate whether they were able to stimulate mGluR1, mGluR5, or mGluR8 expressed in the cells. As positive controls, we used L-glutamate, a known endogenous ligand for the mGluRs, and AP4, a known agonist for group III mGluRs (Wroblewska et al., 1997). The stimulatory activity of these homocysteine derivatives at the mGluRs was determined by measurements of their ability to increase intracellular IPs, products from the hydrolysis of membrane phosphoinositides (PIs).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Stimulatory effect of homocysteine derivatives on phosphoinositide hydrolysis in the CHO cells stably expressing metabotropic glutamate receptors, mGluR1 (A), mGluR5 (B), and mGluR8 (C). After preincubation with myo-[ 3H]inositol overnight, cells were treated for 45 min at 37°C with L-HCSA, L-homocysteine sulfinic acid (); L-HCA, L-homocysteic acid (); L-CSA, L-cysteine sulfinic acid (); L-CA, L-cysteic acid (); L-Glu, L-glutamic acid (); and AP4, L-(+)-2-amino-4-phosphonobutyric acid (). Determination of the produced [3H]IPs is described under Materials and Methods. Data were normalized to the maximal response induced by 1 mM glutamate for mGluR1 and mGluR5, or by 0.1 mM AP4 for mGluR8, and are presented as percentage of the maximal response (mean ± S.E.M.). Dose-response curves of the induced IP formation were generated with the IP data from cells expressing (A) mGluR1 in 5~9 independent experiments, (B) mGluR5 in 4~6 independent experiments, and (C) mGluR8 and Gqi9 in 4~15 independent experiments.

Inhibition of Cyclic AMP Formation through Activation of mGluR2, mGluR4, and mGluR6. mGluR2, mGluR4 and mGluR6 are negatively coupled to adenylyl cyclase. To investigate whether the homocysteine derivatives stimulate mGluR2, mGluR4 or mGluR6, we treated the cells that stably expressed each of the three receptors with L-HCSA, L-HCA, L-CSA and L-CA. The stimulatory activity of these compounds at the mGluRs was determined by measurements of their ability to reduce forskolin-stimulated cAMP. Dose response curves of the relative stimulatory activity are shown in Fig. 3 and EC50 values are in Table 3. The relative potencies are ranked as the following, for mGluR2: L-glutamate L-CA  L-HCSA  L-HCA > L-CSA; mGluR4: AP4 L-HCA  L-HCSA > L-CA > L-CSA; mGluR6: AP4 > L-HCSA > L-HCA L-CA  L-CSA. The results demonstrate that the homocysteine derivatives stimulated mGluR2, mGluR4, and mGluR6 with different potencies and efficacies, but they are less potent than L-glutamate and AP4. Of the homocysteine derivatives, L-HCSA and L-HCA were more potent than L-CSA and L-CA at mGluR4 and mGluR6, whereas their potency was similar to that of L-CA at mGluR2 (Table 3 and Fig. 3).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Inhibitory effect of homocysteine derivatives on forskolin-induced cyclic AMP formation in CHO cells stably expressing mGluR2 (A) and mGluR6 (C), and in baby hamster kidney cells stably expressing mGluR4 (B). After blockade of phosphodiesterases for 10 min, cells were treated for 10 min at 37°C with 5 µM forskolin without or with L-HCSA (), L-HCA (), L-CSA (), L-CA (), L-Glu (), and AP4 (). Cyclic AMP concentration was determined by radioimmunoassay. Data were normalized to the maximal response induced by 0.1 mM glutamate for mGluR2, or by 0.1 mM AP4 for mGluR4 and mGluR6, and are presented as percentage of the maximal response (mean ± S.E.M.). Dose-response curves of the inhibited cyclic AMP formation were generated with the cyclic AMP data from cells expressing (A) mGluR2 in 4~12 independent experiments, (B) mGluR4 in 4~12 independent experiments, and (C) mGluR6 in 4~12 independent experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The major finding of the present study is that various acidic homocysteine derivatives, especially L-HCSA and L-HCA, potently and specifically activate cloned rat mGluRs expressed in vitro. These findings imply that groups I, II, and III mGluRs represent specific receptors for acidic homocysteine derivatives. One prior radioligand binding study demonstrated that L-HCSA, L-CSA, and L-CA had high binding affinities for mGluR1 with K_i values of 440, 3510, and 8050 nM, respectively (Kingston et al., 1998). Herein, we further demonstrate that, with respect to the binding site specificity of radioligands, neither homocysteine nor its derivatives have detectable binding affinity for 54 other G-protein-coupled receptors, ion channels, or transporters. Taken together, these results suggest that acidic homocysteine derivatives, if present at high enough concentrations, are likely to preferentially modulate the activity of mGluRs and thereby regulate intracellular signaling events and neural transmission.

G-protein-coupled receptors have a single ligand-binding pocket, although some receptors, channels, and transporters may contain more than one ligand-binding site (Lu et al., 2002; Ma et al., 2002). In the present survey, we mainly used radioligand competitive binding assays to test whether the test ligands have potential capacity to regulate the tested receptors, channels, or transporters. We also tested the phencyclidine (PCP) and MK-801 sites for the NMDA receptor in our competitive binding assay. We found (Table 2) that L-cystine had medium affinity (K_i = 3,240 nM) for the MK-801 site on the NMDA receptor, whereas L-HCA or the other compounds had low binding affinity for the MK-801 site. Moreover, their binding affinity for the PCP site was also quite low. A previous study showed that L-HCA bound to the NMDA receptor through the glutamate-binding site with extremely low affinity (K_i = 9,500 nM) in comparison with that of L-glutamate (K_i = 500 nM) at the same receptor (Olney et al., 1987). In the present paper, we were unable to estimate the binding affinity of homocysteine and its derivatives for the glutamate-binding site on the NMDA receptor because we did not measure significant inhibition (i.e., >50% at 10,000 nM).

Our analysis of the dose-response curves revealed that the agonist properties, i.e., potencies and efficacies, of the acidic homocysteine derivatives (L-HCSA, L-HCA, L-CSA, and L-CA) were similar to each other at mGluR5 but displayed profound differences at the other five mGluRs tested (i.e., mGluR1, 2, 4, 6, and 8). A close examination of the chemical structures reveals 1) L-HCSA and L-HCA are sulfur-containing 4-carbon acidic amino acids, whereas L-CSA and L-CA are sulfur-containing 3-carbon acidic amino acids; and 2) L-HCSA and L-CSA are sulfinic acids, whereas L-HCA and L-CA are sulfonic acids. In comparison to the endogenous neurotransmitter L-glutamate, the chain length of L-HCSA and L-HCA molecules is similar to that of L-glutamate. By contrast, the chain length of L-CSA and L-CA is one carbon shorter than that of L-glutamate. Furthermore, the distal sulfonate groups of L-HCA and L-CA contain one more oxygen atom than the distal carboxylate group of L-glutamate and the distal sulfinate groups of L-HCSA and L-CSA (Fig. 1). The endogenous full agonist glutamate probably fits to a binding pocket on the mGluRs in a "perfect" manner. The closer to glutamate's chemical structure the chemical structure of a glutamate analog is, the better the glutamate analog will fit to the binding pocket and the stronger the agonist property the glutamate analog will have. The structure of L-HCSA is the most similar to L-glutamate among the four tested glutamate analogs; this may explain why L-HCSA is generally the most potent and efficacious metabolite tested. Because all of the present studies were performed with cloned receptors expressed in heterologous expression systems, the issue of receptor reserve and relative agonist potency must be considered. Since there are no adequate commercially available radioligands available for most of these receptors, the absolute receptor number is impossible to quantify. Nonetheless, we have obtained preliminary data with rat cortical neurons in culture that L-HCSA is equipotent with L-glutamate for activation of PI hydrolysis (Q. Shi, S. Hufeisen, J. Nadeau, and B. Roth, manuscript in preparation) as is predicted from our current studies with cloned receptors.

Kingston et al. (1998) reported that L-HCSA, L-CSA, and L-CA stimulated PI hydrolysis through activation of human mGluR1 and mGluR5a expressed in a transformed Syrian hamster cell line. In their study, L-HCSA was less potent than L-glutamate at human mGluR5. In our study, L-HCSA was more potent than L-glutamate and L-HCA was as potent as L-glutamate at rat mGluR5 expressed in Chinese hamster ovary cells. The different results could be due to differences in the host cell lines, the assay conditions, or receptor species used in the two independent studies. Our functional studies demonstrated that the acidic homocysteine derivatives also stimulated group II and III mGluRs. Therefore, all three classes of mGluRs can be regulated by acidic homocysteine derivatives. In the brain, group I mGluRs are located mainly postsynaptically, whereas group II and III mGluRs are located presynaptically (De Blasi et al., 2001); others have suggested that mGluR5 receptors may also be presynaptic (Romano et al., 2002). In terms of spatial distribution, mGluR1 and mGluR5 are located in a variety of brain regions (Shigemoto et al., 1992; Daggett et al., 1995), whereas the distribution of group II and III mGluRs in the brain is more restricted than that of group I mGluRs (Nakajima et al., 1993; Ohishi et al., 1993; Tanabe et al., 1993; Kinzie et al., 1995; Bennett and Balcar, 1999). Knockout mouse studies reveal that mGluRs are involved in a variety of physiological functions including motor coordination and spatial learning (Conquet et al., 1994), associative learning (Aiba et al., 1994), hippocampal long-term depression (Yokoi et al., 1996), complex motor learning (Pekhletski et al., 1996), and visual transmission (Masu et al., 1995). If the acidic homocysteine derivatives are present in high enough local concentrations, they could functionally modulate mGluR activity.

L-HCSA, L-HCA, L-CSA, and L-CA have been long speculated to be endogenous neurotransmitters (Curtis and Watkins, 1960), although there has been no strong support for this hypothesis. High-performance liquid chromatographic studies showed that L-HCSA and L-HCA were detectable in astrocytes (72 and 49 pmol/mg protein, respectively) (Grieve and Griffiths, 1992), and acute exposure to -adrenergic receptor agonists induced a 3.7-fold increase in the concentration of L-HCA (Do et al., 1997). One prior clinical study showed that patients suffering methotrexate-induced neurotoxicity had high concentrations of homocysteic acid (119.1 µM) and cysteine sulfinic acid (28.4 µM) in their cerebrospinal fluid (Quinn et al., 1997), implying that acidic homocysteine derivatives can be produced in the central nervous system under certain circumstances.

The identification of metabotropic glutamate receptors as the molecular targets for acidic homocysteine derivatives could be important for clarifying the molecular mechanisms responsible for the pathogenesis of various neuropsychiatric diseases associated with moderate hyperhomocysteinemia (Spence et al., 1999). Moderate hyperhomocysteinemia is also a risk factor for stroke and cardiovascular disease in general (Beckett and Marsden, 1997; Spence et al., 1999), although it has been unclear how elevated levels of homocysteine increase the risk for stroke and other cardiovascular diseases. Quite recent studies (Collard et al., 2002) in which human brain endothelial cells were demonstrated to have functional mGluRs illuminate a potential mechanism by which elevations of homocysteine and acidic homocysteine derivatives may increase the risk for cardiovascular disease. In these studies, Collard et al. (2002) demonstrated that activation of endothelial mGluRs modulates vascular permeability. Other studies have demonstrated the presence of cardiac mGluR (Gill et al., 1999), the expression of which may be altered by nicotine administration (Hu et al., 2002).

Animal studies have established that the glutamate-mediated excitotoxicity is a cause of neuronal apoptosis in cerebral ischemia (Collingridge and Lester, 1989). Recent studies with cultured neurons showed that activation of group I mGluR can exacerbate NMDA-induced excitotoxic injury (Mukhin et al., 1997), although this exacerbation is not abrogated in mGluR1 knockout mice (Ferraguti et al., 1997). An additional study suggests that noncompetitive mGluR5 antagonists inhibit NMDA-induced apoptosis in primary cortical neuron cultures, indicating the involvement of mGluR5 (Bruno et al., 2000). It is conceivable, therefore, that elevations of acidic homocysteine derivatives may predispose individuals to the neurotoxic consequences of stroke via alterations in mGluR functioning.

Another disease for which moderate hyperhomocysteinemia is a risk factor is dementia, including Alzheimer's disease (Miller, 2000) and schizophrenia (Levine et al., 2002). Recent epidemiological studies reported that serum homocysteinemia concentrations >14.0 µM increase the risk for Alzheimer's disease (Miller, 2000; Seshadri et al., 2002) and schizophrenia (Levine et al., 2002). A previous in vivo study showed that activation of mGluRs induced the intraneuronal production of amyloid beta fragments (A) together with degeneration of pyramidal neurons in the CA1 region of hippocampus (Stephenson and Clemens, 1998). Additional findings showed that the aggregated A-mediated activation of mGluRs and the aggregated A-elicited intracellular inositol phosphate formation are two consecutive events preceding the aggregated A-induced neuronal apoptosis in vitro (Kanfer et al., 1998; Singh et al., 1998; Allen et al., 1999). These prior studies suggest that activation of mGluRs may contribute to the pathogenesis of Alzheimer's disease, although these ideas are still quite speculative. The mechanism by which an elevation of acidic homocysteine derivatives may exacerbate schizophrenia is unknown, although mGluRs are known to affect a number of signaling processes thought to be altered in schizophrena, and mGluRs have been suggested to be a target for antipsychotic drug development (see, for instance, Kalkman, 2002).

In conclusion, we have demonstrated that acidic homocysteine derivatives are selective mGluR family agonists. It is conceivable that if certain acidic homocysteine derivatives are elevated sufficiently under normal and pathological conditions, they could modulate the activity of a variety of mGluRs. Future experiments will be needed to determine whether the local concentrations of acidic homocysteine derivatives are elevated in moderate hyperhomocysteinemia and whether, via mGluRs, they influence other risk factors to contribute to the pathogenesis of the diseases associated with moderate hyperhomocysteinemia.