Abstract
The 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-lesioned primate is the gold-standard animal model of Parkinson disease (PD) and has been used to assess the effectiveness of experimental drugs on dyskinesia, parkinsonism, and psychosis. Three species have been used in most studies—the macaque, marmoset, and squirrel monkey—the last much less so than the first two species; however, the predictive value of each species at forecasting clinical efficacy, or lack thereof, is poorly documented. Here, we have reviewed all the published literature detailing pharmacologic studies that assessed the effects of experimental drugs on dyskinesia, parkinsonism, and psychosis in each of these species and have calculated their predictive value of success and failure at the clinical level. We found that, for dyskinesia, the macaque has a positive predictive value of 87.5% and a false-positive rate of 38.1%, whereas the marmoset has a positive predictive value of 76.9% and a false-positive rate of 15.6%. For parkinsonism, the macaque has a positive predictive value of 68.2% and a false-positive rate of 44.4%, whereas the marmoset has a positive predictive value of 86.9% and a false-positive rate of 41.7%. No drug that alleviates psychosis in the clinic has shown efficacy at doing so in the macaque, whereas the marmoset has 100% positive predictive value. The small number of studies conducted in the squirrel monkey precluded us from calculating its predictive efficacy. We hope our results will help in the design of pharmacologic experiments and will facilitate the drug discovery and development process in PD.
Introduction
Since its accidental discovery (Davis et al., 1979; Langston et al., 1983), 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) has been used extensively to model Parkinson disease (PD) in a breadth of nonhuman primates, enabling the investigation of anatomic (Jan et al., 2000; Zeng et al., 2008), behavioral (Pessiglione et al., 2004a,b), neurochemical (Soghomonian et al., 1994; Huot et al., 2008), and other aspects of the disease. The MPTP-lesioned primate has also been invaluable in the search for effective antidyskinetic and antiparkinsonian agents. A review article previously examined the translational predictive value of the MPTP-lesioned primate for the development of antidyskinetic drugs, but it was published more than a decade ago (Fox et al., 2006a), and many drugs have since been tested both in the primate and in clinical settings. This previous review did not examine the translational predictive value of the MPTP-lesioned primate when it relates to the effect of drugs on parkinsonian disability or dopaminergic psychosis, nor did it look at differences between different primate species when it comes to predicting the clinical effectiveness of an experimental molecule.
We have therefore conducted a thorough and unbiased review of the literature to compare the predictive effectiveness of different MPTP-lesioned primate species on the clinical effectiveness of potential antidyskinetic, antiparkinsonian and antipsychotic agents.
We believe this review comes at a critical time given that it has recently been suggested that some primate species might be more suited than others to conduct behavioral pharmacologic research (Porras et al., 2012); also, there have been several failures of high-profile drugs in clinical trials (Cook et al., 2014; Bespalov et al., 2016), emphasizing the need to use the best animal model possible in preclinical settings to maximize chances of success when translating to the clinic.
Materials and Methods
Three nonhuman primate species have been used in most experiments to determine the antidyskinetic, antiparkinsonian, and antipsychotic effect of experimental drugs: the macaque [both cynomolgus (Di Paolo et al., 1986) and rhesus (Burns et al., 1983), Macaca fascicularis and Macaca mulatta, respectively], the common marmoset (Jenner et al., 1984) (Callithrix jacchus), and the common squirrel monkey (Langston et al., 1984) (Saimiri sciureus). For each of these three primate species, we have reviewed all the studies that reported their effect, or lack thereof, on dyskinesia, parkinsonism, and psychosis. We then sought which of these drugs tested in the nonhuman primate underwent clinical testing and compared their preclinical and clinical effects and calculated different predictive values (see Endpoints).
Inclusion Criteria.
The literature search spans from 1983, when the first MPTP-lesioned nonhuman primate was engineered (Burns et al., 1983), until December 10, 2017. Studies published after this date, either online or in print format, are not included.
Only drugs whose preclinical and clinical efficacy, or lack thereof, were reported in peer-reviewed articles are included in this article. Results disseminated solely through abstracts or conference proceedings are therefore not reported here. We are aware that not all studies conducted on parkinsonian primates have been published and that the results of some clinical trials have not been disseminated through peer-reviewed journals; by choosing to include only studies and reports that were published in peer-reviewed journals, we may be introducing a selection bias in our analysis. For instance, although it has been studied in humans (NCT00034814, NCT00108667), the clinical effectiveness of talampanel will not be discussed here, as the results of the clinical trials where it was assessed were not published in peer-reviewed scientific journals. In addition, we have excluded l-3,4-dihydroxyphenylalanine (l-DOPA) and inhibitors of l-aromatic amino acid decarboxylase [e.g., benserazide (Rinne et al., 1975) and carbidopa (Marsden et al., 1973)] from our analysis, as these were used in almost all studies cited here to reverse parkinsonism and induce dyskinesia.
Research Methods.
Literature review was conducted primarily through the United States National Library of Medicine database, accessed via PubMed. The search engine Google was used to complete the literature review and to access articles not indexed in the National Library of Medicine database. The following terms were used to perform the literature search: abnormal involuntary movements, chorea, cynomolgus, dyskinesia, dyskinetic, dystonia, hallucinations, hyperactivity, hyperkinesia, l-DOPA, levodopa, macaque, marmoset, monkey, MPTP, nonhuman primate, PD, parkinsonian, parkinsonism, primate, psychosis, rhesus, squirrel monkey, visual hallucinations.
Definition of Efficacy.
Throughout our literature search, a reduction in dyskinesia/psychosis was considered to have occurred only if statistical significance was reached; trends were not considered. A drug was considered to have antiparkinsonian benefit if it alleviated parkinsonism as monotherapy or as an adjunct to l-DOPA. Impulse-control disorders (Weintraub et al., 2010) were not considered a psychotic manifestation.
Several studies, especially in the macaque, were performed that administered drugs (e.g., cabergoline) to reduce the development of dyskinesia and then conducted postmortem experiments; we have included these studies in our analysis, even if their primary endpoint was not determination of pharmacologic efficacy. Studies reporting the effects of experimental drugs on rotational behavior in the hemi-MPTP-lesioned primate (e.g., Vermeulen et al., 1995) were not included.
In some instances, conflicting results were obtained with some drugs (e.g., studies encountered a therapeutic benefit), whereas others did not reach similar conclusions; whenever this happened, we calculated the predictive values using “best result” (i.e., the one that showed a therapeutic effect), weighing in the methods used in the studies (Gad, 2009), if applicable. For instance, in the case of preladenant, phase 2 studies have found antiparkinsonian efficacy (Hauser et al., 2011; Factor et al., 2013), whereas phase 3 studies did not (Hauser et al., 2015; Stocchi et al., 2017); but in one instance, the active comparator, rasagiline (Hauser et al., 2015), was also ineffective, despite proven antiparkinsonian benefit in randomized-controlled trials (Parkinson Study Group, 2005; Rascol et al., 2005), somewhat casting doubt on study conclusion. In this case, we considered that preladenant exerted antiparkinsonian benefit.
Endpoints.
By comparing the outcomes of the primate studies on dyskinesia, parkinsonism, and psychosis, we have calculated, for each species, the positive predictive value, the negative predictive value, and the false-positive rate.
Here, we define the positive predictive value of a species as the percentage of cases for which a primate species has correctly predicted that an antidyskinetic or antiparkinsonian benefit would be achieved in the clinic. For the positive predictive value (eq. 1), the denominator was the number of drugs for which a therapeutic effect was achieved in the clinic, whereas the numerator was the number of these drugs that showed efficacy in the primate:
(1)Here, we define the negative predictive value of a species as the percentage of cases for which a primate species has correctly predicted that a lack of antidyskinetic or antiparkinsonian benefit would be achieved in the clinic. For the negative predictive value (eq. 2), the denominator was the number of drugs for which a therapeutic effect was not achieved in the clinic, and the numerator was the number of these drugs that did not show efficacy in the primate:
(2)Here, we define the false-positive rate of a species as the percentage of cases for which a primate species has incorrectly predicted that a lack of antidyskinetic or antiparkinsonian benefit would be achieved in the clinic. For the false-positive rate (eq. 3), the denominator was the number of drugs for which a therapeutic effect was not achieved in the clinic; the numerator was the number of these clinically ineffective drugs that were deemed to be effective in the primate:
(3)Results
Pharmacologic Targets
Several pharmacologic targets have been modulated in studies performed in the MPTP-lesioned macaque and the MPTP-lesioned marmoset. In contrast, only dopaminergic, opioidergic and cholinergic targets have been assessed in studies conducted in the MPTP-lesioned squirrel monkey (see Supplemental Table 1 for the pharmacologic profile of all molecules that have been tested in the MPTP-lesioned primate).
A total of 98 different molecules were assessed in the MPTP-lesioned macaque, 97 in the MPTP-lesioned marmoset, and nine in the MPTP-lesioned squirrel monkey (Fig. 1). Of all the molecules tested in the MPTP-lesioned primate, 64 have been assessed at the clinical level or are clinically approved. Of these, 44 were tested in the macaque, 38 in the marmoset, and one in the squirrel monkey (Fig. 2). Because very few drugs have been tested in the MPTP-lesioned squirrel monkey compared with the macaque and the marmoset, we do not discuss it further in the text, but we have nevertheless included it in the tables.
We have summarized our research results in tables:
Table 1: drugs tested in the MPTP-lesioned macaque
View this table:Table 2: drugs tested in the MPTP-lesioned marmoset
View this table:Table 3: drugs tested in the MPTP-lesioned squirrel monkey
View this table:Table 4: drugs tested in the MPTP-lesioned primate that were tested in the clinic or are clinically-available
View this table:
In each table, the drugs are listed in numerical or alphabetical order.
We have also summarized our results in figures:
Figure 1: number of drugs tested in the MPTP-lesioned nonhuman primate
Figure 2: number of drugs assessed in the clinic or clinically approved that were tested in the MPTP-lesioned nonhuman primate
Figure 3: antidyskinetic positive predictive value of the MPTP-lesioned nonhuman primate
Figure 4: antidyskinetic negative predictive value of the MPTP-lesioned nonhuman primate
Figure 5: antidyskinetic false positive rate of the MPTP-lesioned nonhuman primate
Figure 6: antiparkinsonian positive predictive value of the MPTP-lesioned nonhuman primate
Figure 7: antiparkinsonian negative predictive value of the MPTP-lesioned nonhuman primate
Figure 8: antiparkinsonian false positive rate of the MPTP-lesioned nonhuman primate
Figure 9: worsening of parkinsonism positive predictive value of the MPTP-lesioned nonhuman primate.
Prediction of Antidyskinetic Effect
Of the 64 drugs that were tested in the clinic and in the MPTP-lesioned primate, 22 showed antidyskinetic effect in clinical trials, no antidyskinetic effect was found or reported for 36 drugs, and four drugs showed a deleterious effect on dyskinesia severity (see Supplemental Table 2 for details).
MPTP-Lesioned Macaque.
Of the 22 drugs that demonstrated an antidyskinetic effect in clinical settings, 16 were tested in the macaque, and an antidyskinetic effect was obtained with 14 (87.5% positive predictive value, Fig. 3). Of the 36 drugs for which no antidyskinetic effect was found or reported in the clinic, 21 were tested in the macaque. No antidyskinetic effect was encountered or reported with 13 (61.9% negative predictive efficacy, Fig. 4); an antidyskinetic action was found with eight (38.1% false-positive rate, Fig. 5). Of the four drugs that had a deleterious effect on dyskinesia severity in the clinic, two were tested in the macaque, and an exacerbation of dyskinesia could not be demonstrated in either case.
MPTP-Lesioned Marmoset.
Of the 22 drugs that demonstrated an antidyskinetic effect in clinical settings, 13 were tested in the marmoset, and an antidyskinetic effect was obtained with 10 (76.9% positive predictive value, Fig. 3). Of the 36 drugs for which no antidyskinetic effect was found or reported, 22 were tested in the marmoset, and the absence of antidyskinetic effect was identified in 19 (86.4% negative predictive value, Fig. 4), whereas an antidyskinetic action was found with three (15.6% false-positive rates Fig. 5). Of the four drugs that had a deleterious effect on dyskinesia severity, three were tested in the marmoset, but an exacerbation of dyskinesia could not be demonstrated in any case.
Prediction of Antiparkinsonian Action
Of the 64 drugs that were tested in the clinic and in the MPTP-lesioned primate, 34 showed an antiparkinsonian effect in clinical trials; no antiparkinsonian effect was found or reported for 24 drugs, and five drugs were found to have a deleterious effect on parkinsonian disability (see Supplemental Table 3 for details).
MPTP-Lesioned Macaque.
Of the 34 drugs that showed antiparkinsonian effect in clinical trials, 22 were tested in the macaque, and an antiparkinsonian benefit was obtained with 15 (68.2% positive predictive value, Fig. 6). Of the 24 drugs for which no antiparkinsonian effect was found or reported, 18 were tested in the macaque, and no antiparkinsonian effect was encountered or reported with six (33.3% negative predictive value, Fig. 7), whereas antiparkinsonian action was found with eight (44.4% false-positive rate, Fig. 8). Of the five drugs that had a deleterious effect on parkinsonian disability, four were tested in the macaque, and this deleterious effect on parkinsonism was correctly identified in three (75% predictive value, Fig. 9).
MPTP-Lesioned Marmoset.
Of the 34 drugs that showed antiparkinsonian effect in clinical trials, 23 were tested in the marmoset; an antiparkinsonian effect was obtained with 20 (86.9% positive predictive value, Fig. 6). Of these, 24 drugs did not find or did not report an antiparkinsonian effect, 12 were tested in the marmoset, and no antiparkinsonian effect was encountered with six (50.0% negative predictive value, Fig. 7), whereas antiparkinsonian benefit was found with five (41.7% false-positive rate, Fig. 8). Of the five drugs found to have a deleterious effect on parkinsonian disability, three were tested in the marmoset, all of which hindered parkinsonism (100% predictive value, Fig. 9).
Prediction of Antipsychotic Action
Of the 64 drugs that were tested in the clinic and in the MPTP-lesioned primate, five showed an antipsychotic effect in clinical trials/reports (clozapine, mianserin, mirtazapine, pimavanserin, quetiapine). Compared with dyskinesia and parkinsonism, the effect of experimental drugs on dopaminergic psychosis has been far less studied in the MPTP-lesioned primate. No drug that underwent clinical testing or that is clinically available has demonstrated antipsychotic effect in the MPTP-lesioned macaque or the MPTP-lesioned squirrel monkey. Pimavanserin was not tested in the MPTP-lesioned marmoset, but an antipsychotic benefit was achieved with clozapine, mianserin, mirtazapine, and quetiapine (100% positive predictive value).
Discussion
Here, we have reviewed all the literature published in peer-reviewed scientific journals that reported the results of pharmacologic studies conducted in the MPTP-lesioned macaque, marmoset, and squirrel monkey in which the effects of experimental drugs on dyskinesia, parkinsonism, and psychosis was assessed. By comparing the results obtained at the preclinical level with those obtained in clinical settings, we have calculated the predictive value of each primate species for these disease manifestations/treatment-related complications.
There are limitations to our analysis that must be mentioned. First, as mentioned in the Introduction, the results of several studies, both preclinical and clinical, have not been published; and, although we aimed for exhaustiveness, our review is necessarily incomplete, which may have affected the various rates presented. Second, the methods of the clinical trials cited is highly variable, ranging from observational reports to randomized controlled trials; they were weighed equally here. Third, the method used in preclinical studies is, at times, different from the one used in clinical settings; one example is when a low dose of l-DOPA is administered to primates in combination with an agent with potential antiparkinsonian effect as adjunct therapy. Lowering the l-DOPA dose administered may be poorly tolerated by patients, which is why this approach is seldom used in clinical trials. Fourth, some clinical trials were performed in early stage PD patients, whereas the degree of parkinsonism after MPTP administration is severe and would correspond to advanced-stage PD. Finally, as several types of trials are part of our review, in some, PD patients were taking antiparkinsonian medication, in addition to l-DOPA, which is generally not the case in primate studies, and the extent to which these molecules affected the results is undetermined.
Keeping these limitations in mind, the following general conclusions can be drawn:
Relative to the antidyskinetic effect of drugs, the macaque has higher positive predictive value than the marmoset, but the marmoset has fewer false-positive results than the macaque. Both the macaque and the marmoset appear limited when it comes to predicting a detrimental effect of experimental drugs on dyskinesia.
Relative to the antiparkinsonian action of drugs, the marmoset has a greater positive predictive value and fewer false-positive results than the macaque; both species have high predictive values when it comes to forecasting a potentially deleterious effect of drugs on parkinsonism.
Relatively to the antipsychotic effect of drugs, comments can be made only for the marmoset, which has high positive predictive value.
Compared with the macaque and the marmoset, the squirrel monkey has been used in a small number of studies, and few pharmacologic targets have been assessed in this primate species, which makes it impossible to calculate its predictive value.
At a time when the discovery and development process for drugs acting at the central nervous system level are facing challenges and have been marred by failures of high-profile candidates, it is our hope that this review will help in the planning and design of preclinical experiments aimed at testing the effects of drugs on l-DOPA–induced dyskinesia, parkinsonian disability, and dopaminergic psychosis by helping experimenters and sponsors plan their experiments in the animal model of PD with the highest translational potential for their specific endpoint.
Authorship Contributions
Participated in research design: Veyres, Huot.
Performed data analysis: Veyres, Huot.
Wrote or contributed to the writing of the manuscript: Veyres, Hamadjida, Huot.
Footnotes
- Received December 13, 2017.
- Accepted March 6, 2018.
P.H. holds research support from Parkinson Canada, Fonds de Recherche Québec – Santé, Natural Sciences and Engineering Research Council of Canada, and the Weston Brain Institute.
↵This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- l-DOPA
- l-3,4-dihydroxyphenylalanine
- MPTP
- 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
- PD
- Parkinson disease
- Copyright © 2018 by The American Society for Pharmacology and Experimental Therapeutics