Tetrahydroisoquinolines (THIQs), which belong to a group of cyclized condensation adducts of biogenic amines with aldehydes, are referred to as mammalian alkaloids (Collins et al., 1979). They include salsolinol (1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisoquinoline; SAL) and tetrahydropapaveroline (6,7-dihydroxy-1-(3',4'-dihydroxybenzyl)-1,2,3,4-tetrahydroisoquinoline; THP) that are derived from dopamine through condensation with acetaldehyde and dopaldehyde (3,4-dihydroxyphenylacetaldehyde), respectively (Sandler et al., 1973). THP is a putative dopaminergic neurotoxin that is implicated in the pathology of Parkinson's disease (McNaught et al., 1998; Collins 2004; Soto-Otero et al., 2006). It is known that THP is detected at a high level in the brain (Nagatsu, 1997) and the urine (Sandler et al., 1973; Cashaw, 1993) of parkinsonian patients under 3,4-dihydroxyphenylalanine (L-DOPA) therapy. In addition, significant levels of THP have also been detected in the brain after ethanol uptake (Cashaw, 1993) and/or L-DOPA treatment (Turner et al., 1974; Cashaw et al., 1987; Cashaw, 1993), and this endogenous alkaloid is considered to account for the neurobehavioral abnormalities associated with alcoholism and may act as a neurotransmitter (Sango et al., 2000). In this review, we will focus on the neurochemical/neuropharmacological properties of THP and related THIQs with regards to their possible implications in the pathology of some neuronal disorders including Parkinsonism and alcohol addiction.

Parkinson's disease is the second-commonest neurodegenerative disease and first described by James Parkinson in 1817. The symptoms of the disease involve bradykinesia, resting tremor, and muscular rigidity. The most prominent neuropathology in PD is the progressive loss of dopamine neurons in the substantia nigra pars compacta (SNc). Dopaminergic neurons in the SNc project to the dorsolateral striatum and upon stimuli, they release an important neurotransmitter, dopamine. Therefore, the loss of dopamine neurons of SNc results in reduction of striatal DA level, and further results in symptoms of Parkinson's disease. 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) is a well known exogenous dopaminergic neurotoxin and is converted by monoamine oxidase (MAO) to the 1-methyl-4-phenyl-pyridinium ion (MPP⁺), that selectively kills the nigrostriatal dopamine neurons (Collins et al., 1979; Nagatsu, 1997; McNaught et al., 1998; Storch et al., 2002). THIQs and their N-methylated metabolites may function in a manner similar to MPTP or MPP⁺ and has been demonstrated that they can provoke parkinsonian-like symptoms in monkeys (Nagatsu and Yoshida, 1988). Since THP bears two catechol moieties, the compound may readily undergo redox cycling to produce reactive oxygen species (ROS) as well as toxic quinoids (Surh, 1999).

Although L-DOPA, the natural precursor of dopamine, is the most effective and frequently prescribed drug for controlling symptoms of Parkinson's disease (Marsden, 1994), the dopaminergic neurons continue to die in parkinsonians during the L-DOPA therapy (Ogawa, 1994). L-DOPA treatment may even hasten the underlying neurodegenerative process (Fahn and Cohen, 1992) and also provokes some side effects (Morris, 1978), a phenomenon referred as "L-DOPA paradox". The dopaminergic neurotoxicity and other adverse effects of L-DOPA may be due to the elevated levels of THP (Sandler et al., 1973; Matsubara et al., 1992; Cashaw, 1993; McNaught et al., 1998; Collins, 2004). Accumulation of dopamine after L-DOPA treatment, increases the formation of THP in parkinsonian patients and experimental rats (Turner et al., 1974; Cashaw et al., 1987; Cashaw, 1993). In addition, treatment of PC12 cells with L-DOPA induced apoptosis by generating ROS, which was aggravated by THP cotreatment (Lee et al., 2003).

Therefore, parkinsonian patients treated with L-DOPA for long-term need to be monitored for the relationship between plasma concentrations of THP and any plausible symptoms of neurotoxicity (Kim et al., 2006). Moreover, it would also be worthwhile determining any differences in the levels of THP among untreated parkinsonian subjects, those on L-DOPA medication and normal individuals.

Despite lack of direct evidence supporting the role for THP in neuronal death in Parkinson's diease, accumulating data from cell culture and animal studies suggest that this endogenous THIQ may act as a dopaminergic neurotoxin (Nagatsu, 1997; McNaught et al., 1998; Collins, 2004) and the neurotoxic potential of THP is summarized in Table 1.

THIQs have been also postulated to play a role in the pathogenesis of chronic alcoholism (Collins et al., 1979; Nace, 1986). Collins proposed that oxidative metabolites of endogenous THIQs may be responsible for neuronal damage associated with chronic alcoholism (Collins et al., 1979).

Possible implications of THIQs in alcohol dependence were inferred from the observation that rats which normally rejected alcohol would, following direct delivery of THP, drink ethyl alcohol in increasingly excessive amounts (Myers and Melchior, 1977). When rats were injected with THP, the preference for alcohol lasted up to 10 months (Duncan and Deitrich, 1980; Myers and Privette, 1989). Likewise, microinjection of THP into the ventral tegmental area of anesthetized high-ethanol-preferring (HEP) rats directly altered the function of the mesolimbic pathway and the dopaminergic system. Such a perturbation could account for the induction of alcohol preference (Myers and Robinson, 1999).

Since (S)-THP can serve as an intermediate in de novo synthesis of more complex alkaloids including morphine and codeine in mammals (Haber et al., 1997; Sango et al., 2000), overproduction of THP during alcohol consumption may increase synthesis of morphine and codeine and this may be associated with alcohol addiction (Kirby, 1967; Davis and Walsh, 1970; Nace, 1986). Although the physiologic roles of endogenous morphine and codeine are unknown, they are known to have analgesic effect and to invoke physical dependence (Weitz et al., 1986; Donnerer et al., 1987). The complete morphine and codeine biosynthetic pathways in mammalian species are still unclear, but THP has been suggested as a possible precursor as in the case of plants (Davis and Walsh, 1970; Matsubara et al., 1992). If THP is an intermediate for the morphine and codeine biosynthesis in mammals, the levels of all these endogenous substances in L-DOPA-treated parkinsonian patients should be higher than those in normal control subjects. When, Matsubara and colleagues measured the urinary levels of morphine and codeine as well as THP in parkinsonian patients under L-DOPA therapy, they were found to be significantly higher than those in healthy non-drinker controls (Matsubara et al., 1992). It is also noteworthy that there were very low levels of THP, morphine and codeine in the urine of abstinent alcoholics. Thus, these narcotic substances, endogenously formed from THP as an intermediate, might contribute to alcohol dependence (Davis and Walsh, 1970; Collins, 2004).

THP is considered to be formed spontaneously by non-enzymatic Pictet-Spengler condensation of dopamine with its aldehyde metabolite (3,4-dihydroxyphenylacetaldehyde) produced by monoamine oxidase (MAO) (Fig. 1) (Holtz et al., 1964; Walsh et al., 1970). In parkinsonian patients receiving L-DOPA treatment, L-DOPA is converted into dopamine by DOPA decarboxylase and the excess dopamine may inhibit aldehyde dehydrogenase (Turan et al., 1989), thereby blocking the conversion of 3,4-dihydroxyphenylacetaldehyde into 3,4-dihydroxyphenylacetic acid. The accumulation of a relatively high concentration of 3,4-dihydroxyphenylacetaldehyde, together with the high levels of dopamine derived from L-DOPA administration, may favor the formation of THP (Soto-Otero et al., 2006).

It was postulated that excess acetaldehyde, secondary to ethanol metabolism, would competitively inhibit the breakdown of dopaldehyde by aldehyde dehydrogenase (Davis et al., 1970). The resultant accumulation of dopaldehyde would lead to an additional condensation reaction with dopamine to augment THP formation (Davis et al., 1970; Nace, 1986). THP was detected in specific brain regions of the rat after acute ethanol administration (3 g/kg body weight), but not in the same regions of untreated animals. Most brain regions had detectable levels of THP 100 minutes after the animals received ethanol, and the striatum contained the highest concentration of the alkaloid (Cashaw, 1993). THP can be detected in the brain of rats not only after acute ethanol administration (Cashaw, 1993), but also when animals were subjected to alcohol consumption ad libitum for a longer period (Haber et al., 1997). In the latter study, alcohol administration for 18 months induced formation of the (S) enantiomer of THP only in the striatum of the rat brain.

The stereo-chemically specific THP has also been identified in human brains in a later study (Sango et al., 2000) and interestingly, only the (S) enantiomer of THP was detected, suggesting that this endogenous dicatechol isoquinoline is presumably synthesized in the brain by an enzyme-catalyzed reaction. It is reported that some edible plants adopt the distinct enzymatic pathways for the synthesis of (S)-THP (Rueffer et al., 1983), and exogenous THP consumed by man is anticipated to cross the blood-brain barrier (Cashaw and Geraghty, 1991). Thus, it is possible that the stereo-chemically specific exogenous or peripheral origin of THP is detected in the brain. It is interesting to note that dietary sources of SAL contribute to its detection in biological samples (Smythe and Duncan, 1985), and this may also be the case for THP.

As mentioned previously, THP has been speculated to be implicated in the etiology of several human neurological, behavioral, and psychiatric disorders, such as parkinsonism and alcohol addiction (Collins et al., 1979; Collins, 2004). In this part of review, we will address the possible neurotoxic mechanisms of THP.

As THP has two catechol moieties, it undergoes auto-oxidation or enzymatic oxidation to produce ROS, which may contribute to the THP-induced cytotoxicity. Although overproduction of ROS is cytotoxic, the initial accumulation of moderate amounts of ROS may provoke defense responses of cells. The body's antioxidant defense system serves to protect the cells from excess ROS production and is comprised of both endogenous and exogenous entities (Fisher-Wellman and Bloomer, 2009). The endogenous antioxidant defense include radical scavengers (e.g., GSH, thioredoxin, bilirubin, uric acid, etc.), and antioxidant enzymes, such as SOD, catalase, glutathione peroxidase, glutathione S-transferase, glutamate-cysteine ligase, NAD(P)H:quinone oxidoreductase 1, etc. The exogenous antioxidants (e.g., carotenoids, tocopherols, ascorbate, bioflavonoids, etc.) are provided as diet, especially fruits and vegetables.

Keratinocytes of the human epidermis, a tissue particularly exposed to oxidant stimuli, possess a wide range of antioxidant and detoxifying mechanisms aimed to avoid oxidative damage of the tissue. It is demonstrated that THP and L-DOPA upregulated expression of intracellular antioxidant enzymes to a different extent in normal keratinocytes of human epidermal origin when compared to transformed ones (Foppoli et al., 2005). Normal diploid keratinocytes adequately scavenge toxic substances through the coordinated activation of several concurrent antioxidant pathways. Conversely, in transformed cells, the whole oxidative burden must be neutralized by the limited set of conserved pathways that, accordingly, has to be highly activated (Foppoli et al., 2005).

THP treatment activated the redox-sensitive nuclear factor κB (NF-κB). Preincubation of PC12 cells with NF-κB inhibitors, such as L-1-tosylamido-2-phenylethyl chloromethyl ketone and parthenolide, aggravated THP-induced cytotoxicity (Shin et al., 2004). In addition, THP treatment altered survival or death signals including not only MAPK but also Akt/protein kinase B.

The redox-sensitive transcription factor Nrf2 is known to regulate expression of detoxifying and antioxidant enzymes and other defensive proteins against oxidative stress and other noxious conditions (Motohashi and Yamamoto, 2004). THP treatment elevated nuclear translocation of Nrf2 and its subsequent binding to antioxidant response element. An important enzyme that plays an essential role in cellular survival response to a wide variety of stress is heme oxygenase-1 (HO-1). Treatment of PC12 cells with THP increased expression of HO-1, and THP-induced cytotoxicity was attenuated by the HO-1 inducer (SnCl₂). In addition, pharmacologic inhibition of HO-1 activity exacerbated THP induced cell death (Park et al., 2007). When PC12 cells were transfected with dominant-negative Nrf2 after THP treatment, the cytotoxicity was increased and HO-1 expression was decreased. These results suggest that cells try to protect themselves from THP-mediated cell death by activating Nrf2 and inducing antioxidant/cytoprotective enzymes, particularly HO-1.

Over the past three decades, THP and other structurally related THIQ alkaloids have been studied for neurotoxic potential in parkinsonism and alcoholism. Additional studies are necessary to uncover the molecular targets of this neurotoxin, especially those involved in cell death or cell defense intracellular signaling cascades. In addition, the adverse effects of THP on cellular functions other than neurotoxic activity merit further investigation.

Table 1. Neurotoxic potential of THP

Table 2. Biochemical mechanisms underlying THP-induced neurotoxicity