Effects of Haloperidol on the Delta-9-Tetrahydrocannabinol Response in Humans: A Responder Analysis

Abstract

Rationale:

Δ-9-Tetrahydrocannabinol (Δ-9-THC) produces psychotomimetic effects in humans. However, the role of dopamine signaling in producing such effects is unclear. We hypothesized that dopaminergic antagonism would reduce the psychotomimetic effect of Δ-9-THC.

Objective:

The objective of this study was to evaluate whether pre-treatment with haloperidol would alter the psychotomimetic and perceptual-altering effects of Δ−9-THC, measured by the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) and the Clinician Administered Dissociative Symptom Scale (CADSS) in humans.

Methods:

In a 2-test-day double-blind study, 28 healthy individuals were administered active (0.057 mg/kg) or placebo oral haloperidol, followed 90 and 215 min later by intravenous administration of active (0.0286 mg/kg) Δ-9-THC and placebo, respectively. This secondary analysis was conducted because of the observation in other studies and in our data that a significant proportion of individuals may not have an adequate response to THC (floor effect), thus limiting the ability to test an interaction. Therefore, this analysis was performed including only responders to THC (n=10), defined as individuals who had an increase of at least one-point on the PANSS positive scale, consistent with prior human laboratory studies.

Results:

In the 10 responders, Δ-9-THC-induced increases in PANSS positive scores were significantly lower in the haloperidol condition (1.1 ± 0.35) compared to the placebo condition (2.9 ± 0.92).

Conclusion:

This responder analysis showed that haloperidol did reduce the psychotomimetic effect of Δ-9-THC, supporting the hypothesis that dopaminergic signaling may participate in the psychosis-like effects of cannabinoids.

Introduction

The increasing availability of cannabis, for medicinal and recreational purposes, coupled with rapidly changing public perceptions surrounding the use of cannabinoids, highlight the need to probe and understand their substrate, the endocannabinoid (eCB) system. Although the psychotomimetic effects of Δ-9-Tetrahydrocannabinol (Δ-9-THC) have been well established, the neurobiological underpinnings of these effects remain incompletely understood.

Converging evidence demonstrates complex interactions between the dopaminergic (DA) and eCB systems (El Khoury et al. 2012; Fernandez-Ruiz et al. 2010; Gardner 2005; Laviolette and Grace 2006). Cannabinoid (CB) and DA receptors share commonalities in anatomical localization in the brain (Hermann et al. 2002; Julian et al. 2003) as well as in signal transduction mechanisms (Glass and Felder 1997; Kearn et al. 2005; Meschler and Howlett 2001). Accumulating preclinical evidence indicates that Δ-9-THC activates DA mesolimbic neurons (French 1997; French et al. 1997; Gessa et al. 1998) and induces DA release in the striatum (Chen et al. 1990; Tanda et al. 1997). Further, co-administration of haloperidol and Δ-9-THC produce more catalepsy than haloperidol alone in rats, and haloperidol reverses Δ-9-THC-induced reductions in prepulse inhibition and startle response in mice (Marchese et al. 2003; Nagai et al. 2006). Consistent with the notion of functional CB-DA interactions, agonists that bind to these neurotransmitter systems interact to produce varied effects on rotational activity (Sanudo-Pena et al. 1998; Sanudo-Pena et al. 1996), locomotion (Meschler et al. 2000; Meyer and Kunkle 1999) and food intake in rats (Verty et al. 2004).

Possibly owing to the complexity of CB-DA interactions, the data in humans in much less clear, with neurochemical brain imaging studies and human laboratory studies (HLS) initially not supporting the notion Δ-9-THC leads to increased DA signaling (Bloomfield et al. 2016). Recently, however, a reanalysis of neurochemical brain imaging studies did indicate a small release in striatal pre-synaptic DA (Bossong et al. 2015). Reasons why the human data is mixed may include the much lower magnitude of DA release caused by Δ-9-THC compared to other drugs, such as stimulants, combined with technical difficulties in capturing dopamine changes of this magnitude with Positron Emission Tomography (Bloomfield et al. 2016). Similarly to previous HLS, both haloperidol (Liem-Moolenaar et al. 2010) and olanzapine (Kleinloog et al. 2012) were shown to produce a significant reduction of Δ-9-THC-induced psychotomimetic effects measured on the PANSS. Since the effect of olanzapine was demonstrated using a responder analysis (responders were defined as subjects who had an increase of at least one-point on the PANSS positive subscale), we re-analyzed our previously published data on the effects of haloperidol on Δ-9-THC response using this method (D’Souza et al. 2008).

Materials and methods

The methods of this HLS have been described in detail in our previous paper (D’Souza et al. 2008) and are reported here in brief. The study was conducted at the neurobiological studies unit at the West Haven Veterans Affairs Hospital, under IRB approval. The sample consisted of medically and neurologically healthy 18–55 years old men and women, previously exposed to cannabis, evaluated for any DSM-IV axes I or II lifetime psychiatric or substance use disorders (except for cannabis and tobacco), using the structured clinical interview for DSM-IV (SCID DSM-IV) (First and Gibbon 2004). Importantly, no subjects met criteria for cannabis DSM-IV abuse or dependence (DSM-V cannabis use disorder) at the time of inclusion the study. Participants were also excluded if they had a family history of major axis I disorders (i.e., including, but not limited to, psychotic disorders). They completed 2 test days during which they received placebo or active (0.057 mg/kg) haloperidol in a random, counterbalanced order. This was followed by placebo Δ-9-THC (vehicle) 90 minutes later, and active Δ-9-THC (0.0286 mg/kg) 215 minutes later administered intravenously (IV) over 20 minutes, in a fixed order. Psychotomimetic effects of Δ-9-THC were assessed before and after haloperidol (active or placebo) and after placebo and active Δ-9-THC administration, using the PANSS (Kay et al. 1987). Perceptual altering effects were evaluated with the Clinician Administered Dissociative States Scale (CADSS) (Bremner et al. 1998).

Responder analysis

All data from responder subjects was assessed for normality, and transformations/nonparametric tests were performed as previously described (D’Souza et al. 2008). Peak change from baseline (PANSS and CADSS) was analyzed using linear mixed models (α=0.05; two-tailed; SPSS 24, IBM Corp). To evaluate PANSS and CADSS as dependent outcomes, haloperidol (placebo vs. active) and time were included (−90, −30, 15, 65, 140, 190) as a within-subject factor. Cannabis use history (frequent users vs. healthy subjects) was included as a between-subjects factor. Frequent users were defined as (1) lifetime cannabis exposure greater than or equal to 100 times, (2) last exposure to cannabis within the past week, (3) recent cannabis exposure greater than ten times per month as quantified by Time Line Follow Back (Sobell and Sobell 1992), and (4) positive urine toxicological test for cannabis at screening. As defined elsewhere (Liem-Moolenaar et al. 2010), clinically significant positive symptoms were operationalized as higher than one-point increase in the positive-symptom subscale in the PANSS.

Results

Sociodemographic and clinical characteristics

54 subjects were screened and 28 completed this study. Out of 28 completers, only 10 were found to be responders using the criterion of at least a one-point increase in the PANSS positive score in the placebo-haloperidol condition (Kleinloog et al. 2012). The demographic characteristics of the ten subjects are reported in Table 1. The cannabis use patterns including lifetime use and last use are described in Table 2.

Table 1.

Demographics

(N=10)
Age (years ± SD)	26.1 ± 9.26
Gender
Male	8
Female	2
National Adult Reading Test (NART)	158.4 ± 33.51
Years of Education (years ± SD)	15.1 ± 1.19
Race
Caucasian	8
Hispanic	0
African-American	2
Native-American	0
Weight (lbs)	158.3 ± 33.54

Table 2.

Time Since Last Exposure to Cannabis and Lifetime Exposure

Time Since Last Exposure to Cannabis
1 day to 1 week	2
1week to 1 month	5
1 month to 1 year	2
1 year to 5 years	1
Lifetime Exposure to Cannabis
Number of exposures
Less than 5 times	1
5–10 times	0
11–20 times	0
20–50 times	2
51–100 times	3
More than 100 times	4
Number of frequent users	2

Frequent users were defined as (1) lifetime cannabis exposure greater than or equal to 100 times, (2) last exposure to cannabis within the past week, (3) recent cannabis exposure greater than ten times per month as quantified by Time Line Follow Back and (4) positive urine toxicological test for cannabis at screening.

Psychotomimetic effects

Effect of Δ-9-THC and haloperidol:

Administration of Δ-9-THC increased scores on all PANSS subscales, total PANSS scores and clinician-rated and subjective CADSS scores, in both the placebo and the haloperidol condition (Table 3). The increase in the PANSS positive subscale score in the haloperidol condition was significantly lower than in the placebo condition (t=2.77, df=8, p=0.02, d=2.4) (Table 4, Figure 1). The increases in the PANSS general psychopathology and total score, and clinician-rated and subjective CADSS scores in the haloperidol condition were also lower than in the placebo condition. These differences, however, were not significant (Table 4).

Table 3.

Behavioral Outcomes

Mean (SD)
Outcome		Haloperidol						Placebo
		−90	−30	+15	+65	+140	+190	−90	−30	+15	+65	+140	+190
PANSS	Positive	7.78(1.09)	7.89(1.16)	7.78(1.09)	7.78(1.09)	9(1.5)	7.56(1.13)	7.4(0.52)	7.4(0.52)	7.4(0.5 2)	7.5(0.71)	10.3(0.95)	7.8(0.91)
	Negative	6.78(1.3)	6.44(1.01)	6.22(0.44)	6.33(0.71)	8.63(1.85)	6.4(0.72)	6.4(0.7)	6.2(0.42)	6.1(0.31)	6.2(0.42)	7.7(1.16)	6.3(0.67)
	General Psychopath ology	15.67(1)	15.67(0.87)	16(0.71)	15.56(0.73)	18.11(1.69)	15.67(0.71)	15.3(0.48)	15.5(0.71)	15.6(0.7)	15.5(0.85)	18.9(3.03)	15.6(0.7)
	Total	30.22(2.91)	30(2.29)	30(1.58)	29.67(1.32)	35.88(2.52)	29.67(2)	29.1(1.1)	29.1(0.99)	29.1(0.88)	29.2(1.03)	36.9(3.51)	29.7(1.42)
CADSS	Clinician rated	0	0	0.22(0.66)	0	3.33(0.22)	0.125(0.1)	0(0)	0.1(0.33)	0.1(0.32)	0.1(0.32)	4.7(2.6)	0.4(0.7)
	Subjective	0.22(0.67)	0.5(1.67)	0.67(1)	0.2(0.67)	3.7(3.15)	1(2.12)	0.1(0.31)	0.2(0.42)	0(0)	0.2(0.63)	6.7(5.05)	1.9(2.4)

PANSS: Positive and Negative Syndrome Scale. CADSS: Clinician-Administered Dissociative States Scale.

Table 4.

Peak Change from Baseline in Psychotomimetic and Perceptual-altering Effects

	Mean (SD)		t (paired t-test)	df	p-value
	Haloperidol	Placebo
PANSS positive	1.1 (0.35)	2.9 (0.92)	2.77	8	0.024
PANSS negative	2.13 (1.46)	1.5 (1.02)	−0.73	8	0.487
PANSS general psychopathology	2.44 (1.81)	3.4 (3.24)	1.189	8	0.269
PANSS total	5.55 (1.75)	7.8 (2.46)	1.67	8	0.133
CADSS-Clinician Rated	3.33 (1.05)	4.6 (1.45)	1.04	8	0.255
CADSS-Subjective	3.22 (1.01)	6.2 (1.96)	1.24	8	0.331

PANSS: Positive and Negative Syndrome Scale. CADSS: Clinician-Administered Dissociative States Scale.

Fig. 1:

THC-induced Psychotomimetic Effects Change from baseline in Psychotomimetic effects measured by the Positive and Negative Syndrome Scale (PANSS) positive scores.

In this responder analysis (n=10), haloperidol administration reduced THC-induced psychotomimetic effects, compared to placebo *(p=0.02, d=2.4)

Haloperidol Placebo

Haloperidol dose x time interactive effect:

We also found a significant interaction between haloperidol dose and time on the PANSS positive (ATS=6.84; p=0.0016) and the CADSS subjective scores (ATS= 2.87; p=0.0238).

Discussion

The results of this re-analysis are consistent with previous studies where haloperidol (Liem-Moolenaar et al. 2010) and olanzapine (Kleinloog et al. 2012) were shown to reduce Δ-9-THC induced psychotomimetic effects. Specifically, haloperidol reduced the Δ-9-THC-induced increases in the PANSS positive subscale score, but not on other measures, such as the CADSS scores. These findings have methodological implications for future HLS examining interactive effects of THC and other drugs, as well as mechanistic implications for the understanding of CBDA interactions.

Methodological implications:

Due to the high variation in response to cannabinoids (Englund et al. 2013; Morrison and Stone 2011), studies examining interactive effects between THC and other drugs should include individuals who report clinically significant responses to cannabinoids. As there evidence linking a family history of psychosis with higher risk for a psychotomimetic response to THC (Henquet et al. 2008), future HLS should investigate the influence of a family history of psychosis on the cannabinoid-induced psychotomimetic response, and how it may affect the efficacy of antipsychotics. Efforts should be also made to identify predictors of a clinically significant psychotomimetic response to cannabinoids by utilizing polygenic risk scores (Aas et al. 2018; Kuepper et al. 2013) in the context of HLS. Alternatively, post-hoc responder analyses should be conducted aiming to increase the changes of capturing meaningful interactions between THC and other drugs. Finally, a consensus regarding what constitutes clinically significant increase in PANSS should be reached, given that there is variation in the ways different studies have defined responders (D’Souza et al. 2005).

Mechanistic implications:

Along with the data from neurochemical imaging studies indicating small Δ-9-THC-induced releases of striatal DA, this responder analysis and prior HLS provide support for the involvement of the DA signaling in the psychotic effects induced by Δ-9-THC (Bloomfield et al. 2014). Though DA signaling likely participates psychotomimetic response to cannabinoids, there is growing evidence for the notion that such sensitivity to psychotomimetic effects is related to downstream mechanisms, such as functional changes to postsynaptic DA signal transduction, rather than presynaptic DA release (Beaulieu and Gainetdinov 2011; Bloomfield et al. 2014). For instance, cannabis-using individuals with psychosis tend to have higher rates of D2/3 receptor blockade-induced motoric side effects (Potvin et al. 2006). In these individuals, it is conceivable that the blunting effect of chronic cannabis use on DA synthesis/release (Bloomfield et al. 2016) may lead to postsynaptic DA receptor super sensitivity, contributing to the emergence of psychotic symptoms (Howes and Murray 2014). In contrast, in cannabis-naïve individuals and individuals with psychosis without co-occurring substance use, presynaptic mechanisms such as increased DA release/synthesis (Howes and Murray 2014), and other neurobiological underpinnings may have a stronger contribution to psychosis (Howes et al. 2017). Future, adequately powered HLS should examine the influence of cannabis use histories (i.e., having cannabis use disorder, differences in cannabis potency and frequency of cannabis exposure) on the efficacy of antipsychotics for counteracting the psychotomimetic effects of cannabinoids.

The findings from this secondary analysis should be considered as preliminary and interpreted in light of several limitations. First, the small sample size precluded investigation of some important individual factors that may influence the interactive effects of THC and haloperidol, such as cannabis use history, and the adequate application of methods to adjust for multiple hypothesis tests. Second, we did not include individuals with a family history of psychosis. Third, the original HLS design did not allow for examining dose-dependent interactive effects of THC and haloperidol, which could change the affinity of the latter for D1, D2 and α1 receptors. Finally, the behavioral effects of THC were measured repeatedly on each test day, which should be noted in interpreting the effects of haloperidol pretreatment.

Conclusions

Though the precise molecular mechanisms of CB-DA interactions in the psychotomimetic cannabinoid response remain to be fully elucidated, this study lends further support to the cannabinoid model of psychosis and to the testing of antipsychotic agents using this well-validated cannabinoid HLS paradigm. Future studies should include individuals who have a high probability of having a clinically significant psychotomimetic response to cannabinoids, or conduct analogous responder analyses.