INTRODUCTION

Risperidone is a benzisoxazole derivative belonging to the second generation of antipsychotic drugs (Janssen et al, 1988; Leysen et al, 1994). It and its major active metabolite 9-OH-risperidone are effective antipsychotics compared to placebo and to other first- and second-generation antipsychotic drugs (Conley and Mahmoud, 2001; Marder and Meibach, 1994). Furthermore, it has an acceptable side effect profile in humans, particularly at low doses (Marder and Meibach, 1994). One of the distinctive characteristics of this drug is that its motor side effect profile (parkinsonism and akathesia) is low at doses less than 6 mg/day, but emerges at higher doses. Clinically, the antipsychotic dose range of risperidone is 2–16 mg daily, whereas the acute motor side effect dose range generally begins in adults above 6 mg and increases in a linear fashion with higher doses (Marder and Meibach, 1994). Above the daily dose of 8 mg, little motor side effect advantage accrues to risperidone compared to haloperidol, because risperidone shows haloperidol-like extrapyramidal motor effect.

Risperidone doses in the range of 2–6 mg/day induce little parkinsonian motor side effects or akathesia in clinical use. It has remained a clinical question whether risperidone doses at or below 6 mg/day are also minimal risk for tardive dyskinesia (TD). The idea that acute motor side affects predict eventual TD received support when Kane and colleagues (Kane and Smith, 1982; Kane et al, 1986) showed that acute EPS is a risk factor for TD. In addition, several chronic antipsychotic treatment studies have shown a decrease in TD risk with second-generation antipsychotic use, including studies examining risperidone (Correll et al, 2004) and risperidone use in the susceptible elderly (Jeste et al, 1999). Moreover, risperidone in humans has been observed to have an antidyskinetic effect at doses above 6 mg, but not below (Chouinard et al, 1993), consistent with the sparing of hypokinetic (as well as later hyperkinetic) motor effects below 6 mg. Prospective studies showing differential risk between antipsychotics for TD in humans with fixed doses of risperidone would be long and arduous.

Thus, for preliminary screening, we have used a rat model of TD, chronic antipsychotic-induced vacuous chewing movement (VCM), to test the idea that low doses of risperidone will fail to induce VCMs, but that high risperidone doses will produce the TD-like syndrome. Already Marchese et al (2004) have examined subacute risperidone treatment in laboratory rats to show a reduced withdrawal VCM liability at low doses of drug.

Laboratory rats chronically treated with the traditional antipsychotic drug haloperidol or other similar drugs gradually develop VCMs over a time course of 3–6 months (Clow et al, 1980a, 1980b, 1980c; Ellison et al, 1987; Glenthoj and Hemmingsen, 1989; Gao et al, 1998; Andreassen et al, 2003). Several characteristics of these movements, including aspects of their phenomenology, etiology, and pharmacology, resemble TD in humans (Tamminga et al, 1990; Waddington, 1990). This model is consistent with published criteria for evaluating animal models (McKinney and Bunney, 1969; Weiss and Kilts, 1995). Therefore, we have used this model in animals to putatively predict the potential of risperidone to generate TD (Gao et al, 1997; Gao et al, 1998). In this model, neither clozapine, olanzapine, or sertindole are associated with the VCMs, while haloperidol induces these movements across a wide dose range (Tamminga and Woerner, 2002). We have selected doses in animals that produce blood levels within the range of human antipsychotic response. We have dosed laboratory rats using two doses of risperidone over 6 months and compared the development of VCMs to haloperidol and a control condition. The two dose ranges of risperidone were designed to be comparable to human doses of low and high risperidone, matching low and high plasma drug levels across the animal and clinical studies in the human therapeutic dose range.

METHODS

Animals

Male Sprague–Dawley rats were divided into four treatment groups: (1) water ad libitum N=10; (2) haloperidol, 1.5 mg/kg/day, N=10; (3) risperidone 1.5 mg/kg/day, N=12; (4) risperidone 6 mg/kg/day, N=13. Rats were housed in regulated animal quarters, fed ad libitum, and kept on 12 h light/dark cycle. Dosing began when the animals were 200–220 g or approximately 6–8 weeks old.

Drug Administration

The rats were given drinking water containing either no drug, haloperidol, or risperidone (2 concentrations). Haloperidol was dissolved in 10% of glacial acetic acid. The solution was diluted with distilled water and the pH was adjusted to 5.5–6.0, using 10 N sodium hydroxide to produce a stock solution of 0.25 mg/ml. This stock solution was further diluted with distilled water to 0.025 mg/ml and provided as drinking water for 7 days. The test drug, risperidone, was dissolved in a minimum volume of glacial acetic acid. That solution was diluted with distilled water and adjusted to pH 5.5–6.0 using 10 N sodium hydroxide to give a stock solution of 0.33 mg/ml. This stock was used at a 0.1 mg/ml solution for high dose and was diluted to 0.025 mg/ml solution for the low dose every 7 days. These neuroleptic solutions were administered in the drinking water continuously for 6 months. The dose of antipsychotic drug in the drinking water was modified weekly according to the measured daily water intake of the rats to produce the state dose per group.

Trunk blood was collected at the time of killing and the plasma was frozen at −20°C until analysis. Haloperidol and reduced haloperidol were quantified using a modification of the GC/NPD method of (Bianchetti and Morselli, 1978) with chlorohaloperidol as the internal standard. Risperidone and 9-OH-risperidone plasma levels were analyzed using a modification of the liquid chromatographic procedure of Woestenborghs et al (1992). The two major changes were in the extraction procedure from plasma, and the liquid chromatographic column employed. The limit of detection is 2.5 ng/ml for each component. Interassay variability for risperidone and 9-OH-risperidone for the high-, medium-, and low-quality controls is less than 8.6% (n=34 days). These assays were performed in the laboratory at the Nathan Kline Institute.

Assessment of VCMs

Individual rats were observed in a small Plexiglas cage (30 × 18 × 25 cm) that was completely empty of contents, including bedding. Rats were allowed up to 5 min to adapt to the cage before rating. Then, the numbers of VCMs were counted for 5 min by two experienced raters blind to the drug status of rats. The two ratings were averaged for the final score. VCM ratings were carried out at baseline and monthly thereafter for each drug group. Two ratings sessions were done at the end of the last treatment month and averaged for the 6-month end-treatment rating. Where these two final ratings differed by more than 15%, a third rating was done and the three averaged for the final score.

Rat VCMs were analyzed two ways: (1) by a simple average of rat dyskinesia scores across treatment groups; (2) by a dichotomy into dyskinetic and nondyskinetic (or severe and not severe) ‘diagnosis,’ where a score of >8 VCMs/5 min, identified animals as dyskinetic and lower scores, as nondyskinetic.

Statistics

A one-way analysis of variance (ANOVA) was performed to test for effect of drug on number of VCMs. A Dunnett's post hoc test, at a p=0.05 significance level, was performed to compare drug treatments to control.

RESULTS

Rates of VCMs in the last 2 weeks of a 6-month treatment period showed that drug treatment had a significant effect on the number of VCMs (F(3,41)=7.51, p<0.0004). Haloperidol treatment significantly increased the chewing movements almost five-fold after this extended treatment (p<0.05) (Table 1). Moreover, the percent of the rats expressing ‘severe’ chewing movements (>8 VCMs/5 min) was high in the haloperidol group, 70%. These results with haloperidol are consistent with VCM rates and incidence in our previously published studies (Hashimoto et al, 1998; Shirakawa and Tamminga, 1994).

Table 1 VCMs and Drug Plasma Level: Haloperidol and Risperidone
Full size table

Risperidone, when administered at the lower dose (designed to be analogous to a <6 mg/day human dose), did not produce an increase in the rate of VCMs (Figure 1) nor any change from control in the percent of ‘severe’ movements (Table 1). However, when risperidone was administered at the higher dose, VCM rates were significantly increased to the same range as the rates with haloperidol (p<0.05) (Figure 1). Moreover, here there was also an increase in the proportion of animals with ‘severe’ movements, similar to that seen with haloperidol (Table 1). The distribution of VCMs in each treatment group over 6 month period indicated that a lower dose of risperidone has lower incidences of VCMs than higher doses of risperidone (Figure 1).

Figure 1
figure 1

Average number of vacuous chewing movements (VCMs) for each chronic treatement group in a 5 min period measured once a month (Control: (+) (− − −); Haloperidol: (▵) (− − −), 1.5 mg/kg/day; Risperidone: (○) (− − −), 1.5 mg/kg/day; Risperidone: (□) (−), 6.0 mg/kg/day) for a duration of 6 months.

Full size image

Plasma levels of haloperidol assessed in this study at the termination of treatment (6 months) was 3.3±2.0 mg/ml. This level is within the low end of our usual assessed levels (average of 6.8±1.1 mg/ml), but is still on the low side of the clinical therapeutic plasma level range. However, even with this low haloperidol plasma range, the VCM incidence and rate are typical. With risperidone, the metabolism of primary drug to 9-OH-risperidone is even faster than in humans; thus, most of the risperidone is in the form of the 9-OH-risperidone metabolite (Megens et al, 1994). Summing the level of primary drug with the level of metabolite in these rats (as is typically done in humans), the low- and high-dose groups show distinctly different plasma levels (6.1±3.1 and 47.8±1.5 ng/ml, respectively) and those that fall into two ends of the very broad human therapeutic range for risperidone (Table 1).

DISCUSSION

In this experiment we show that, whereas haloperidol produces oral chewing movements with chronic treatment, chronic risperidone at the low dose fails to produce VCMs in rats while a higher risperidone dose produces classic ‘involuntary’ oral movements. This chronic treatment paradigm in rats is used here as a putative predictive animal model for TD liability. Several lines of evidence support the use of this animal preparation as a surrogate for TD (Glenthoj and Hemmingsen, 1989; Gunne et al, 1982; Tamminga et al, 1990; Waddington, 1990; Andreassen et al, 2003), including the observations that drugs which are associated with significant TD in humans all produce VCMs in rats, while drugs which fail to produce TD also fail to produce VCMs outside the placebo range (Ellison et al, 1987; Gunne et al, 1986). Hence, these data suggest that higher doses of risperidone (>6 mg/day) may have liability for generating TD, similar to haloperidol, but low doses may well lack this liability. These results were predicted based on the human studies which consistently show that high but not low risperidone doses produce acute motor side affects (parkinsonism and akathesia) and that the incidence of acute motor side effects appear to determine TD vulnerability.

For these results to be treatment relevant, rat VCMs need to reliably model TD, not parkinsonism. In this experiment, as in our other studies, and studies from other laboratories, these movements have a slow onset, as would be expected for TD, not an immediate onset as in parkinsonism. Characteristically, the movements have a slow rate of reduction after drug discontinuation, as observed in our initial experiments (Tamminga et al, 1990), a jerky and irregular character, and do not show a regular tremor. Since these animals were killed at study end, we cannot specifically document movement reduction after drug discontinuation, but have no reason to believe it would be different than our original observation. Moreover, the movement quality in these animals was jerky and irregular, not a regular tremor. We and others (Gunne et al, 1982; Ikeda et al, 1999; Tamminga et al, 1990; Trevitt et al, 1999) have argued for a connection between these VCMs and TD, and have suggested and tested a systems pathophysiology distinct from parkinsonism to account for the movements (Shirakawa and Tamminga, 1994; Sakai et al, 2001).

Risperidone occupies the serotonin, 5HT2A receptor with high affinity in animal ex vivo binding studies and in human in vitro affinity studies (Leysen et al, 1993). Its dopamine, DA2 affinity is high as well, but 20 times less than the 5HT2 affinity. The rest of risperidone affinities include low but measurable occupancy at the D3, D4, H1, α1, and the α2 receptors. In animals, risperidone blocks serotonin- as well as dopamine (DA)-mediated pharmacologic effects in vitro and in vivo. In the human at a 6 mg/day dose, risperidone occupies 78–88% of 5TH2A receptors and 75–80% of DA2 receptors (Farde et al, 1988). Again, in humans using 2, 4, and 6 mg, there is a linear dose occupancy of the DA2 receptors of 66, 73, and 79%, respectively (Kapur et al, 1995). Whereas, DA2 receptor occupancy was not measured in these animals (Kapur et al, 2003), the dosing was designed to have the 24 h plasma levels match clinical therapeutic levels at low and high doses, with drug delivered continuously. Risperidone shows an ‘atypical’ antipsychotic profile in that it increases fos-like immediate early gene activity in the ventral, but not in the dorsal striatum in rat (Robertson et al, 1994). This potent antiserotonin action of risperidone is often invoked to explain its reduced motor side effect profile in light of its relatively high DA receptor occupancy (Leysen et al, 1994). Whether this may also influence the development of TD is a matter of speculation.

Risperidone is now widely used for the treatment of psychosis in schizophrenia and in other psychotic disorders with considerable success, especially in the elderly (Correll et al, 2004; Jeste et al, 1999). Clinical descriptions suggest that symptom reduction is broader than with first-generation compounds, extending to cognitive and affective areas and is clearly ‘atypical’ with respect to motor side effects particularly in the lower dose range. Naturalistic physician ‘usage’ studies in psychosis show that psychiatrists tend to prescribe risperidone in a biphasic fashion, with one dose peak in the low dose range of 2–4 mg/day and the other in the high 8–12 mg/day range(Conley and Mahmoud, 2001). Risperidone is often combined with other active antipsychotic compounds into a treatment ‘cocktail,’ but this is done without a solid theoretical rationale and no predictions about ultimate TD liability. Use in the elderly is typically at low doses. These data suggest that the low dose use of risperidone may not be associated with elevated TD liability. Further clinical observations are needed to test this conclusion.