Reciprocal Influences of HIV and Cannabinoids on the Brain and Cognitive Function

Abstract

Globally, cannabis is the most commonly used illicitdrug, with disproportionately high use among persons with HIV. Despite advances in HIV care, nearly half of persons with HIV continue to experience neurocognitive deficits or impairments that may have negative impacts on their daily function. Chronic cannabis use may play a role in the development or exacerbation of these impairments. Here we present a review summarizing existing research detailing the effect of cannabis use associated with the neuropathogenesis of HIV. We examine evidence for possible additive or synergistic effects of HIV infection and cannabis use on neuroHIV in both the preclinical and adult human literatures, including in vitro studies, animal models, clinical neuroimaging research, and studies examining the cognitive effects of cannabis. We discuss the limitations of existing research, including methodological challenges involved with clinical research with human subjects. We identify gaps in the field and propose critical research questions to advance our understanding of how cannabis use affects neuroHIV.

Introduction

Marijuana, or its most commonly abused form cannabis sativa (Aizpurua-Olaizola et al. 2016), is by far the most commonly used illicit drug in the world, although its use is becoming legal in some countries, with estimates showing that nearly 4% of the global population aged 15–64 years use it annually (United Nations 2018). While cannabis is often used recreationally for its psychoactive effects, medicinal use is also commonly reported due to its analgesic, antiemetic, and orexigenic effects. Moreover, many users report both recreational and medicinal use (Pacula et al. 2016). Currently in the United States, 33 states and the District of Columbia (DC) have legal medicinal use, and 11 states and DC have legal recreational use (National Conference of State Legislatures 2019). Given trends toward cannabis legalization in many countries, and in more states within the United States, it is imperative to investigate its effects on human health.

Cannabis use is disproportionately common among persons with HIV. Estimated prevalence rates range from 20–60%, and nearly half of these users report daily use (Crane et al. 2017; D’Souza et al. 2012; Gamarel et al. 2016; Hartzler et al. 2017; Mimiaga et al. 2013; Okafor et al. 2017a; Okafor et al. 2017b). As in the general population, many persons with HIV report both recreational and medicinal use of cannabis (D’Souza et al. 2012; Fogarty et al. 2007; Furler et al. 2004; Woolridge et al. 2005). Persons with HIV report using cannabis for relief of symptoms associated with HIV, including pain, nausea, loss of appetite, and anxious or depressed mood (Abrams et al. 2007; D’Souza et al. 2012; Haney et al. 2007; Haney et al. 2005; Prentiss et al. 2004). While motivations for use are generally similar between seronegative and HIV(+) persons, those with HIV report medicinal use more frequently (Towe et al. 2018). Importantly, research thus far has been focused on perceived benefits of cannabis rather than actual effectiveness.

This review summarizes existing research on how early and chronic cannabis use might render individuals more vulnerable to HIV-associated neuropathology, and how continued cannabis use might affect the ongoing neuropathogenesis of HIV. Despite effective combination antiretroviral therapies (cART), up to half of HIV-infected persons continue to have HIV-associated neurocognitive disorders (HAND), which is based on neuropsychometric testing (Heaton et al. 2010). Compared to the pre-cART era, the more severe form of HAND, HIV-associated dementia, has become less common (from 20% to 6%) in treated HIV persons, but milder forms of HAND, such as asymptomatic neurocognitive impairment and mild neurocognitive disorder persist (Gates and Cysique 2016; Saylor et al. 2016). For individuals with symptomatic neurocognitive impairment, these deficits can cause significant impairments in daily functioning, including activities of daily living, driving, medication management, and employment status (Cattie et al. 2012; Heaton et al. 2004; Laverick et al. 2017; Thames et al. 2013).

Given that HIV confers greater vulnerability to cognitive impairment, chronic marijuana use may exacerbate the development and expression of HAND in persons with HIV. On the other hand, the immunomodulatory properties of cannabinoids may also confer a protective effect. Despite the prevalence of marijuana use among people with HIV, there has been remarkably little research on how it impacts the course of HAND (Skalski et al. 2016; Volkow et al. 2014). Thus, understanding the interplay of cannabis and immunological diseases, such as HIV, is a critical research area. In this review, we examine the empirical evidence from preclinical and adult human studies supporting the unique additive or interactive effects of HIV infection and cannabis use on neuroHIV. We conclude by identifying gaps in the field and providing a framework for future studies on cannabinoids and neuroHIV.

The Endocannabinoid System

Endocannabinoids are an endogenous group of lipid messengers that play an important role in modulating neurotransmitter release and synaptic plasticity (Araque et al. 2017; De Petrocellis and Di Marzo 2009). The two major endocannabinoids best-characterized in mammals are N-arachidonoylethanolamine (anandamide; AEA) and 2-arachidonoylglyercol (2-AG). Rather than being stored, these endogenous agonists are synthesized “on demand” in response to specific stimuli and then immediately released from cells, usually acting on cells in close proximity (De Petrocellis and Di Marzo 2009; Katz et al. 2014). These endocannabinoids are then quickly metabolized or degraded by two major enzymes, fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL); while FAAH degrades AEA to arachidonic acid and ethanolamine, MAGL hydrolyzes and degrades 2-AG to arachidonic acid and glycerol (Figure 1). Hence, inhibitors of these degradation enzymes have also been studied as potential pharmacological targets that may enhance or maintain the extracellular levels of endocannabinoids.

Figure 1. Endocannabinoid synthesis and metabolism in the brain:

The endocannabinoids, diacylglycerol (DAG), 2-Arachidonoylglycerol (2-AG), N-arachidonoylethanolamine (AEA), and N-Arachidonoyl dopamine (NADA) are found in the CNS. Fatty Acid Amide Hydrolase (FAAH) hydrolyzes AEA to arachidonic acid and ethanolamine and is also involved in the synthesis and break down of NADA. The degradation enzymes Monoacylglycerol lipase (MAGL) hydrolyzes 2-AG to arachidonic acid and glycerol while diacylglycerol lipase (DAGL) hydrolyzes diacylglycerol, releasing a free fatty acid. NAAA is located in intraventricular macrophages, but not in microglia and can indirectly influence AEA levels in the brain.

Thus far, two G protein-coupled receptors for endocannabinoids have been identified- cannabinoid type 1 (CB1) and cannabinoid type 2 (CB2) receptors. The CB1 receptors are primarily concentrated in the central nervous system (CNS) and the CB2 receptors are primarily on immune cells; however, both are distributed throughout the body (Cabral and Jamerson 2014). CB1 receptors primarily mediate the psychoactive effects of cannabis and are abundant in the cerebral cortex, hippocampus, amygdala, basal ganglia, cerebellum, and brainstem (Mackie 2005). CB2 receptors, thought to be responsible for the immunomodulatory effects attributed to cannabinoids, are expressed on immune cells such as macrophages, B and T cells as well as microglia in the CNS, with neurons expressing the CB2 receptor, to a much lesser extent than CB1 receptors (Atwood and Mackie 2010; Cabral and Jamerson 2014). Of note, the CB2 receptor has gained attention as a potential therapeutic target because of its expression in activated microglia in the CNS (Atwood and Mackie 2010; Cabral and Griffin-Thomas 2009). Together, these two receptors, their endocannabinoid ligands, and regulatory lipases/hydrolases comprise the endocannabinoid system (De Petrocellis and Di Marzo 2009) (Figure 2).

Figure 2. Locations of Endocannabinoids and Enzymes:

CB1 receptors are found on both neurons and glia, but more on inhibitory than excitatory neurons. CB2 receptors are found on microglia, and to a lesser extent on astrocytes and neurons. They may also be found on some microglial mitochondria. The distribution of DAGL, MAGL and FAAH allow for rapid 2AG and AEA production or degradation in both neurons and glia.

The cannabis sativa plant comprises more than 500 different chemicals, including more than 100 cannabinoids (Aizpurua-Olaizola et al. 2016; Thomas and ElSohy 2016). The most well-known and studied of these cannabinoids, in part because they are the most abundant, are cannabidiol (CBD) and Δ9-tetrahydrocannabinol (THC) (Mechoulam and Parker 2013). While THC is the principal psychoactive component of cannabis, all the cannabinoids are biologically active (Cabral and Jamerson 2014). Like their endogenous counterparts, plant-derived phytocannabinoids signal via the same CB1 and CB2 receptors (Araque et al. 2017; Katz et al. 2014). THC is a partial agonist at CB1 and CB2 receptors, while CBD is an antagonist or inverse agonist (Pertwee 2008). Research on the pharmacology of cannabinoids has supported their immunomodulatory properties (Katz et al. 2014), but rigorous evaluation of the effectiveness of cannabis as a therapeutic agent remains incomplete.

Cannabis became a prohibited drug since the 1920s in the U.S., and the US Drug Enforcement Administration (DEA) currently still classifies marijuana (cannabis) as a Schedule I drug (along with heroin, cocaine), defined as having no currently accepted medical use with high potential for abuse. This strict DEA scheduling of marijuana has impeded evidence-based research on its potential therapeutic effects. In addition, the complex chemical composition of cannabis has made it challenging to draw definitive conclusions regarding the overall beneficial or harmful health effects of marijuana use. However, in conflict with the federal laws, the majority of the states has legalized medicinal cannabis. With the passage of the 2018 Farm Bill, hemp can be legally cultivated under federal law, and the DEA has reclassified CBD with THC content below 0.1% to Schedule 5, provided the product is FDA approved.

Preclinical Research

A small number of studies were designed to determine the underlying causal mechanisms of the complex interactions between the endogenous cannabinoid system, HIV infection, and exogenous cannabinoids used for medicinal or recreational purposes. Some ex vivo and in vitro studies describe the complex interactions between HIV-associated proteins or HIV-mediated neuroinflammation and the endocannabinoid system. Other studies evaluated the interactions between exogenous cannabinoids (e.g., THC), the endocannabinoid system and HIV; both positive and negative effects of THC were reported in cell culture and animal models of HIV (e.g., SIV, FIV, or humanized rodent models), as well as using envelope glycoprotein (gp120) or transactivator protein (Tat) to evaluate the endocannabinoid system, in combinations with agonists or antagonists for CB1 and/or CB2 receptors, or inhibitors of FAAH or AEA.

In vitro studies in Cell Culture Models

Since HIV-associated neurodegeneration may result directly from HIV viral proteins, gp120 and Tat, cell culture studies are often conducted after the administration of these proteins, which are known to induce toxicity in neuronal cell cultures. These in vitro models typically demonstrated a neuroprotective role for endocannabinoids as well as exogenous cannabinoids, such as THC (Table 1). For instance, cannabinoid CB1 and CB2 receptor agonist WIN 55,212–2 protected against HIV-1 Tat-mediated cell toxicity in C6 rat glioma cells, by inhibiting the HIV-1 Tat or HIV-1 Tat + interferon-γ-induced nitric oxide synthase (iNOS) and nitric oxide (NO) release; these effects were also reversed by a CB1, but not CB2, receptor antagonist (Esposito et al. 2002). Similarly, in murine pre-frontal cortical neurons, HIV-1 Tat excessively increased intracellular calcium leading to subsequent neuronal damage, and these effects were ameliorated by prior treatment with anandamide (AEA) (Xu et al. 2017). Paradoxically, in the presence of HIV-1 Tat, CB1 receptor expression increased in prefrontal cortical neurons (Xu et al. 2017), but decreased in HIV-1 Tat treated glioma cells, along with inhibition of reuptake and degradation of anandamide (Esposito et al. 2002), which could further down-regulate CB1 receptor expression. Therefore, the same insult from HIV could lead to divergent effects on the CB1 receptor expression in different models.

Table 1.

Preclinical Studies that Evaluated the Endocannabinoid System and Cannabinoids in HIV Models

	Model(s) Used	Primary Findings	Neuroprotective role of Cannabinoids
In vitro Models
Esposito et al 2002	C6 Glioma Cells + HIV Tat	WIN 55,212-2 (CB1 and CB2 receptor agonist) prevented HIV Tat and pro-inflammatory interferon-γ mediated overproduction of NO and cell damage. HIV tat also inhibited CB1 expression (increase extracelluar endocannabinoids)	Yes
Xu et al, 2017	Murine Prefrontal Cortical Neurons + HIV Tat	Pretreatment with anandamide (AEA) ameliorated HIV Tat-induced increase in intracellular Ca++ and neuronal cell death.	Yes
Hermes DJ et.al. 2018	Murine Prefrontal Cortical Neurons + FAAH inhibitor	FAAH inhibition reduced Tat-mediated excess Ca++ (via CB1R), neuronal cell death and dendritic degeneration (via CB2R).	Yes
Zhang and Thayer 2018	Rat Hippocampal Neurons / HIV gp120 / JZL184	Monoacylglycerol Lipase (MAGL) inhibition with JZL 184 enhanced endocannabinoid signaling, activation of CB2R, decreased gp120-induced inflammatory interleukin-1b (IL-1b) production and hippocampal synapse loss.	Yes
Rodent Models
Maccarrone et.al. 2004	Adult Wistar rats + intraventricular injection of gp120	gp120 activated neocortical FAAH, enhanced AEA degradation but not synthesis, which led to neuronal apoptosis in the brain neocortex.	Yes
Benamar et.al. 2009	Sprague-Dawley rats	The thermoregulatory action of WIN 55,212-2 (CB1 & CB2 receptor agonist) was antagonized by SDF-1a/CXCL12 (a chemokine and CXCR4 ligand) that is often elevated in neuroHIV.	N/A
Gorantla et.al. 2010	hu-PBL/HIVE mice	Gp1a (CB2R agonist) reduced HIV encephalitis (microglia activation), downregulated CCR5 expression but increased CB2R expression.	Yes
Avraham, et al 2015	GFAP/GP120//FAAH−/−	FAAH−/− mice (with decreased endocannabinoid degradation) showed improved neurogenesis in hippocampus and decreased astrogliosis.	Yes
Roth et al 2005	SCID Mice (+/− THC and HIV-infected PBMC)	THC administration prior to HIV infection led to lower CD4 cell counts, while concurrent THC administration led to higher viral load. THC administration after the infection led to upregulated CCR5 expression initially and greater HIV+ cells with longer THC exposure.	No
Simian Models
Benito et al 2005	SIV Encephalitis	SIV-infected macaques showed increased expression of FAAH in cortical white matter with up-regulation of CB2 receptors in microglia/nodules (anti-inflammatory response?).	Yes
Winsauer et.al. 2011	Rhesus Macaques infected with SIV + THC	THC administration prior to SIV infection produced dose-dependent slowing and more errors on cognitive tasks; however, chronic THC produced tolerance to behavioral effects, but not viral load or other markers of disease progression, at 1-year post infection.	Acute (No) Chronic (no effect)
Simon et al 2016	SIV + Δ9-THC	Δ9-THC before and after SIV infection slowed disease progression and decreased inflammation, as well as increased BDNF and decreased proinflammatory cytokines.	Yes

Furthermore, increasing endocannabinoid tone, with the use of inhibitors to the degradation enzymes FAAH and MAGL, demonstrated neuroprotective effects in neuronal cells. In murine prefrontal cortical neurons, inhibition of FAAH (hence maintaining higher levels of endocannabinoids) reduced Tat-mediated increases in intracellular calcium, neuronal cell death, and dendritic degeneration (Hermes et al. 2018). The CB receptors were shown to have differential neuroprotective roles in this model. While activation of CB1 receptors ameliorated toxic intracellular calcium levels, activation of CB2 receptors reduced neuronal damage and degeneration. Similarly, in human and murine neuroprogenitor cells, CB2 agonist AM1241 inhibited gp120-mediated neurotoxicity and apoptosis (Avraham et al. 2014). Furthermore, in rat hippocampal neurons treated with HIV gp120, inhibition of MAGL led to increased endocannabinoid tone and decreased synapse loss as well as decreased inflammatory interleukin-1β (IL-1β) production (Zhang and Thayer 2018). Hence, in neuronal cell cultures, direct or indirect increase of endocannabinoid tone ameliorated neuronal toxicity and ultimately reduced neuronal cell damage caused by HIV Tat or HIV gp120.

The neuroprotective roles of the endocannabinoid system in microglial and neuronal progenitor cells in the presence of HIV Tat or HIV gp120 were also investigated (Fraga et al. 2011; Avraham et al. 2015).. HIV Tat elicits a microglial migratory response, further adding to the inflammatory milieu of HIV-associated brain injury. In a mouse BV-2 microglial-like cell culture model, exogenous partial (THC) and full (CP55940) CB1 and CB2 receptor agonists, as well as endogenous 2-AG, all reduced HIV Tat induced migration of these microglial cells. This reduction was blocked only by a CB2 receptor antagonist (SR144528) but not a CB1 receptor antagonist (SR141716A). In addition, these results accompany an observed reduction in the β-chemokine receptor CCR-3 level (Fraga et al. 2011). Taken together, the findings from in vitro models demonstrated that CB receptor agonists or inhibitors of endocannabinoid degradation enzymes led to neuroprotection, and possibly reduced neuroinflammation. Therefore, these agents may have some potential therapeutic roles in the treatment of HAND.

Ex Vivo and In Vivo Studies with Animal Models

Animal models of HIV/AIDS utilize species-specific viruses (e.g. SIV, FIV, or other murine models) to cause immunodeficiency, since HIV does not infect non-human animal species. In other studies, HIV infection is modeled by using genetically engineered animals that are able to express the HIV viral proteins (e.g., HIV-transgenic rat, HIV-gp120 or HIV Tat mice) or by implanting these animals with HIV-infected macrophages/microglia cells.(e.g., humanized mouse models).. As such, the animals in these studies were immunocompromised before THC or cannabinoid administration. In humans, however, the majority of HIV-infected individuals initiated their cannabis use before the viral exposure although some HIV patients used marijuana for medicinal purposes after the infection. A few studies evaluated the effects of HIV infection in animals with prior chronic cannabinoid exposure, which modeled the marijuana users who become HIV-infected after chronic marijuana use (Winsauer et al. 2011; Simon et al. 2016).

Several in vivo studies assessed how endogenous cannabinoids were impacted in rodent models of HIV/AIDS (Table 1). In the hippocampus of GFAP/gp120 transgenic mice, the CB2 receptor agonist AM1241 enhanced neurogenesis and neuroblast proliferation which are typically impaired due to gp120 (Avraham et al. 2014). Furthermore, with additional genetic deletion of the degradation enzyme FAAH, leading to higher levels of endocannabinoids, the impaired neurogenesis that typically occurs in the GFAP/Gp120 mice was rescued (Avraham et al. 2015). These double GFAP/Gp120//FAAH−/− mice also showed less astrogliosis and gliogenesis (neuroblasts converting into glial cells) in the hippocampus compared to GFAP/Gp120 Tg mice. These transgenic models, however, may not reflect HIV infection of the normally developed brain since maturation occurs in the presence of HIV-viral protein mediated toxicity and/or without FAAH, which may have a role in fetal development and maturation of the central nervous system. This is an important consideration, for instance, gp120 administered intracerebrally to adult rats activated neocortical FAAH, enhanced AEA degradation but not synthesis, which led to neuronal apoptosis only in the brain neocortex but not the hippocampus where ongoing neurogenesis occurs, while inhibition of FAAH reduced gp120-induced apoptosis in rat brain neocortex (Maccarrone et al. 2004). Therefore, the effects of HIV infection on the endocannabinoid system may differ substantially in the developing brain or brain regions that have ongoing neurogenesis compared to the fully developed adult brain.

Only one rodent study evaluated the effects of THC exposure at different time points relative to HIV infection (Roth et al. 2005). In immunodeficient (SCID) mice implanted with human peripheral blood leukocytes infected with HIV, THC administered prior to infection led to lower CD4 counts and fewer IFN-γ-producing cells, while concurrent THC administration had no additional effect on CD4 counts but led to higher viral load. However, THC administration after the infection increased only CCR5 expression but not HIV(+) cells; with more prolonged THC exposure, the percentage of HIV(+) cells but not CCR5 expression was elevated relative to saline-controls. The effects of HIV and THC on IFN-γ-producing cells were additive. This study demonstrated the immunosuppressive effects of THC, which may vary depending on the temporal relationship between THC-administration and HIV infection, as well as the duration of THC exposure, all of which are critically important for modeling the acute or chronic effects of cannabis use in HIV/AIDS.

Studies using an SIV model found both positive and negative effects of cannabinoids (Table 1). Evaluating the endogenous cannabinoid system, a study found that brain tissues from animals that developed SIV encephalitis showed upregulation of CB2 receptors in microglia, and FAAH overexpression in the cortical white matter, while CB1 receptors, which were primarily located in the neurons, were not altered (Benito et al. 2005). These findings suggest that neuroinflammation-mediated or immune cell-activated CB2 expression may reduce the antiviral response, allowing entry of infected monocytes into the CNS; hence, SIV-encephalitis or the consequent neuroinflammation appeared to have a negative influence on the endocannabinoid system. Furthermore, similar to those found in human studies, THC administration prior to SIV infection produced dose-dependent slowing and more errors on cognitive tasks; however, chronic THC use produced tolerance to behavioral effects, but did not adversely affect viral load or other markers of disease progression at 1 year post infection in rhesus macaques (Winsauer et al. 2011). These findings demonstrated that acute, but not chronic, THC administration negatively impacted animals infected with SIV. In addition, a study that assessed THC administration both before and after SIV infection found that THC slowed the disease progression and decreased neuroinflammation (Simon et al. 2016). Furthermore, THC mediated changes in microRNA expression associated with cell signaling, cell cycle, and immune responses were found in the striatum of these SIV-infected macaques, which further suggests that THC may enhance protective neuromodulation. Overall, SIV encephalitis appears to have a negative effect on the endocannabinoid system, while the addition of an exogenous cannabinoid (THC) appears to have negative effects only behaviorally in the short-term but primarily neuroprotective effects with chronic administration, and in an endocannabinoid system already altered by regular THC Exposure.

Despite the recent increased popularity of cannabidiol (CBD) use, no preclinical studies have evaluated the effect of CBD in HIV/AIDS (in vitro or animal models). A few studies evaluated the effects of CBD in immune cells and neuroinflammatory disorders (e.g. multiple sclerosis, Alzheimer’s, stroke). Both THC and CBD showed cell lineage specific effects on chemokine and cytokine levels, suggesting a potential to worsen HIV infection and disease progression (Srivastava et al. 1998). In contrast, CBD attenuated Aβ-induced increases in NO and several inflammatory cytokines from astrocytes, and additionally reduced gliosis and CA1 pyramidal neuronal loss in hippocampal slices of an Alzheimer model (Esposito et al. 2011). CBD also reduced systemic inflammatory cytokines in an in vivo murine model of multiple sclerosis (experimental autoimmune encephalomyelitis) (Elliott et al. 2018). The extent that these immunosuppressive effects from CBD will remain neuroprotective in the long-term in humans remains to be determined.

Table 2 further summarizes the preclinical findings that evaluated the effects of HIV on the endocannabinoid system, as well as how various modulations or activation of the endocannabinoid system might attenuate or mitigate HIV-associated neurotoxcity in different preclinical models.

Table 2.

Reciprocal effects of HIV and the endocannabinoid system

HIV-mediated effects on the endocannbinoid (eCB) system

Effects of eCB on HIV-mediated neurotoxicity

• HIV Tat increased extracellular eCBs in C6 glioma cells • HIV-mediated elevation of cytokines (IL-4, IL-10, TNFα, IL-1, IL-6) reduced eCB functions • HIV Tat inhibited CBR1 expression • Intraventricular gp120 injection in rats led to increased cortical FAAH, degradation of AEA and neuronal apoptosis/ • SIV infection increased FAAH expression and upregulated CBR2 in microglial nodules (anti-inflammatory response)

• Higher levels of eCB (blocking MAGL) blocked gp120-induced synaptic loss and neuroinflammation in hippocampal neurons • eCB protected against HIV Tat-induced overproduction of NO and cell damage in C6 glioma cells • FAAH inhibition (higher AEA) protected against HIV Tat-induced excess calcium-mediated neurotoxicity and neuronal degeneration in cortical neurons • FAAH ablation in gp120 mice led to improved neurogenesis (in hippocampus) and decreased astrogliosis

Clinical research

Neural consequences

Numerous studies evaluated the effects of chronic cannabis use on the adult brain utilizing a variety of magnetic resonance imaging (MRI) techniques. Necessarily, these studies were observational in nature, and most utilized a cross-sectional design. However, very few studies examined the combined effects of HIV and cannabis on brain structure and function (Table 3).

Table 3.

Clinical studies evaluating neural and cognitive effects of marijuana use in PLWH

	Methods/Measures Used	Participants	Primary Findings
Brain Imaging Studies
Thames et. al. 2017	Structural MRI	77 participants across 4 groups:  16 MJ−/HIV−13  MJ+/HIV−  24 MJ−/HIV+  24 MJ+/HIV+	Greater cannabis use (grams/week) was associated with smaller entorhinal and fusiform cortical volumes, while HIV status was associated with thinner cingulate cortices. No interactive or additive effect of cannabis use and HIV infection was observed in brain structure.
Meade et. al. 2018	Functional MRI-Counting Stroop task	93 participants across 4 groups:  25 MJ−/HIV−  19 MJ+/HIV−  29 MJ−/HIV+  20 MJ+/HIV+	Both cannabis use and HIV status were independently associated with abnormalities in brain activation, with greater parietal activation bilaterally in the cannabis users and greater anterior cingulate cortex activation in the HIV(+) participants. Across the 4 groups, HIV(+) cannabis users showed the greatest activation in the insular region, suggesting an additive effect. During the counting Stroop task, insular activation (BOLD signal change) was greater with more years of regular marijuana use, especially in the HIV(+) group.
Chang et. al. 2006a	1H MRS (localized spectroscopy in 6 brain regions)	96 participants across 4 groups:  30 MJ−/HIV−  24 MJ+/HIV−  21 MJ−/HIV+  21MJ+/HIV+	Cannabis use was associated with lower levels of neuronal metabolites (NAA, choline-compounds and glutamate) in the basal ganglia, but higher glial metabolite (total creatine) levels in the thalamus. HIV was associated with trends for lower NAA levels in the parietal white matter and higher choline-compound levels in the basal ganglia. However, HIV+/MJ+ group had normalization of the low glutamate levels in frontal white matter that were observed in HIV+/MJ− and HIV−/MJ+ groups.
Cognitive Studies
Gonzalez et. al. 2011	Procedural learning	86 participants across 4 groups:  21 MJ−/HIV−  23 MJ+/HIV−  25 MJ−/HIV+  17 MJ+/HIV+	Among participants with a history of polysubstance use, HIV status and cannabis dependence (CD) were both independently associated with poorer task performance. While interaction effects were not observed, there was also evidence for additive adverse effects of HIV and CD on task performance.
Skalski et. al. 2018	Learning, memory, executive function, processing speed, motor function, verbal fluency, attention/working memory	69 HIV(+) participants:  12 early onset cannabis users  15 late onset cannabis users  42 non-users	Early onset cannabis users (regular use prior to age 18), compared to non-users and late onset users (regular use at age 18 or later), were more likely to have memory and learning impairment.
Chang et. al. 2006a	Gross motor function, verbal memory, fine motor speed, executive function, verbal intelligence, mood, psychomotor speed	96 participants across 4 groups:  30 MJ−/HIV−  24 MJ+/HIV−  21 MJ−/HIV+  21 MJ+/HIV+	After controlling for age, education, and mood differences, cannabis users did not perform differently from non-users on neuropsychological testing and HIV(+) individuals were slower on some reaction times. No interactive effects between HIV status and cannabis use were observed.
Lorkiewicz et. al. 2018	Memory, attention, self-reported cognitive function	215 HIV(+) participants diagnosed with substance dependence or injection drug use	Current cannabis use was associated with poorer self-reported cognitive function.
Thames et. al. 2016	Premorbid intellectual ability, attention/working memory, processing speed, verbal fluency, learning, memory, executive functioning	89 participants across 6 groups:  12 MJ−/HIV−  14 MJ−/HIV+  12 MJ+ light/HIV−  30 MJ+ light/HIV+  10 MJ+ moderate/HIV−  11 MJ+ moderate/HIV+	Both HIV status and moderate-to-heavy MJ use were associated with poorer test performance. Moderate-to-heavy MJ users performed worse for processing speed, learning/memory, and executive functioning compared to light users and non-users. HIV(+) individuals performed poorer for learning/memory, and executive functioning compared to HIV(−) individuals. There was an interactive effect on learning and memory with HIV(+) moderate-to-heavy users having the poorest learning and memory of all the groups. An interactive effect was observed on verbal fluency: HIV(+) light users performed better that HIV(−) light users.
Thames et. al. 2017	Premorbid intellectual ability, attention/working memory, processing speed, verbal fluency, learning, memory, executive functioning	77 participants across 4 groups:  16 MJ−/HIV−  13 MJ+/HIV−  24 MJ−/HIV+  24 MJ+/HIV+	An interaction between cannabis use and HIV-status was observed. More MJ use was associated with lower global cognitive scores in HIV(−) controls but not in HIV(+) participants. In light MJ users, the HIV(−) group displayed better global cognitive performance than the HIV(+) group. In heavier MJ users, global performance did not differ by HIV status. Processing speed and memory measures drove this effect.
Cristiani et. al. 2004	Intelligence, learning, memory, verbal reasoning, executive functioning, verbal fluency, figure fluency, motor skills, attention/working memory, processing speed	282 participants across 6 groups:  25 MJ−/HIV−  49 MJ+/HIV−  48 MJ−/asymptomatic HIV+  79 MJ+/asymptomatic HIV+  32 MJ−/symptomatic HIV+  55 MJ+/symptomatic HIV+	MJ use was associated with greater cognitive impairment in persons with symptomatic HIV infection, but these effects were not evident in HIV(−) participants or participants with asymptomatic HIV infection. These overall effects appeared to be driven by performance on delayed memory tasks. The effect of marijuana use was greatest in the symptomatic HIV(+) group.

Anatomical or structural MRI studies consistently found that chronic cannabis users had smaller gray matter volumes in the hippocampus, a region rich in CB1 receptors (Brumback et al. 2016; Nader and Sanchez 2018). Morphological changes in other brain regions included the orbitofrontal cortex, striatum, and amygdala, but these results were not replicated in other studies (Brumback et al. 2016; Nader and Sanchez 2018). To date, only one study has examined the combined effects of cannabis use and HIV on brain structure in adult participants (Thames et al. 2017). This study found that higher levels of cannabis use were associated with smaller volumes in the entorhinal and fusiform cortices, but no interactive or additive effect of cannabis use and HIV infection was observed. However, this study had relatively small sample sizes in the HIV(−) groups (13 marijuana users and 16 non-marijuana users), which limited the power to detect significant effects.

Diffusion tensor imaging (DTI) was also used to identify microstructural abnormalities and white matter integrity in specific brain regions of cannabis users (Brumback et al. 2016; Nader and Sanchez 2018). Specifically, lower fractional anisotropy and higher diffusivity in the corpus callosum, superior longitudinal fasciculus, fornix, and internal capsule were reported in cannabis users (Arnone et al. 2008; Becker et al. 2015; Gruber et al. 2011; Zalesky et al. 2012). Although HIV infection is associated with wide-spread disruption in white matter integrity (Ances and Hammoud 2014; Chang and Shukla 2018), no published DTI studies to date have examined the potential additive or interactive effects of cannabis and HIV on brain microstructure.

Functional MRI (fMRI) studies also examined the effects of chronic cannabis use on neural activation during various cognitive tasks. Broadly speaking, cannabis users demonstrated altered or reorganized neural activation patterns in task-relevant regions (Brumback et al. 2016). For example, during a visual attention task, chronic cannabis users showed a reorganized attention network, with lesser activation within many regions of the normal network and the cerebellum, but greater activation (usage of brain resources) in the reserve brain regions, and these effects varied between the active users and abstinent past users, suggesting normalization with abstinence, and greater activation with earlier age of first use (Chang et al. 2006b). Similarly, during a figural memory task, cannabis users had less activation than non-user controls bilaterally in the hippocampus and ventrolateral prefrontal cortices (Dager et al. 2018). In contrast, despite similar behavioral performance, cannabis users showed greater neural activation than non-users in cognitive control and reward-based tasks (Filbey et al. 2014; Manza et al. 2018), as well as a working memory task (Chang and Chronicle 2007). However, these altered activation patterns associated with cannabis use may be more pronounced in persons with HIV infection.

Only one fMRI study examined the independent and combined effects of cannabis use and HIV infection on neural activation during a cognitive interference task (Counting Stroop) (Meade et al. 2018). Individuals with either HIV infection or cannabis use demonstrated independent abnormalities in brain activation, with greater parietal (supramarginal gyrus) activation bilaterally in the cannabis users and greater anterior cingulate cortex activation in the HIV(+) participants. However, HIV(+) individuals who were also chronic cannabis users showed the largest increased activation in the insular region, suggesting a synergistic or additive effect of cannabis use and HIV infection on the compensatory requirement to perform the tasks (Figure 3) (Meade et al. 2018).

Figure 3. Synergistic effect of marijuana use and HIV infection on neural activation during the counting Stroop task:

There was a significant interaction in a cluster in the left insula extending into the central and frontal opercula, inferior frontal gyrus, and orbitofrontal cortex. The Z‐statistic images were thresholded at 2.3 with a cluster P threshold of 0.05. The underlying image is the MNI152 2‐mm standard‐space T1‐weighted structural template, and images are in radiological orientation (left = right, right = left). The bar graph on the bottom shows the mean percent signal change across the four groups (error bars represent the standard error). Adapted from Meade et al. (2018).

A few studies used proton magnetic resonance spectroscopy (¹H MRS) to assess the effects of cannabis use and reported discrete neurochemical alterations. Two studies reported lower levels of the glial marker, myo-inositol in the thalamus (Mashhoon et al. 2013) and the myo-inositol/creatine ratio globally (Silveri et al. 2011), suggesting an anti-neuroinflammatory effect associated with cannabis use (Hellem et al. 2015). Cannabis users were also found to have lower than normal levels of the neuronal marker N-Acetyl-Aspartate (NAA) in the dorsolateral prefrontal and anterior cingulate cortices, which suggested neurotoxic effects of cannabis use in these regions (Sneider et al. 2013). Again, only one study evaluated the independent and combined effects of chronic cannabis use and antiretroviral-treated HIV infection on brain metabolites in six brain regions (Chang et al. 2006a). Specifically, these primarily abstinent cannabis users, regardless of HIV infection, still showed lower levels of neuronal metabolites (NAA, choline-compounds and glutamate) in the basal ganglia, but higher glial metabolite (total creatine) levels in the thalamus relative to non-users. HIV patients, regardless of cannabis use, also showed trends for lower neuronal metabolite (NAA and glutamate) levels in the basal ganglia. However, HIV patients with chronic cannabis use showed normalization of the lower glutamate levels that were seen in HIV(+) individuals or cannabis users in the frontal white matter. These findings are consistent with the lower glutamate levels reported in cannabis users (Blest-Hopley et al. 2019; Colizzi et al. 2016) and in HIV patients (Ernst et al. 2010), suggesting glial/inflammation-mediated neuronal dysfunction. However, the normalized glutamate levels in the frontal region of HIV(+) cannabis users suggest that cannabis use might have attenuated the neuronal dysfunction seen in HIV(+) patients. The possible neuroprotective effect in the frontal brain region of HIV cannabis users is similar to how AEA treatment led to amelioration of neuronal damage that resulted from HIV-Tat mediated calcium release in murine pre-frontal cortical neurons (Xu et al. 2017)

Cognitive consequences

Extensive research has been performed on the cognitive effects of cannabis use, but the findings are mixed largely due to variations in methodologies, including how cannabis use was characterized and how cognitive functioning was assessed. In HIV(−) individuals, acute and residual effects of cannabis use are well-documented and established, showing deficits in memory, but the long-term durability of cognitive effects remain unclear (Broyd et al. 2016; Crane et al. 2013; Crean et al. 2011; Solowij and Battisti 2008). A meta-analysis on the non-acute effects of cannabis found only minor negative effects in the learning and memory domains, and no effects in other domains of function (Grant et al. 2003). However, the authors highlighted that several studies had methodological limitations and sometimes inadequate consideration of potential confounding factors, such as mood disorders or co-use of other substances. A more recent meta-analysis also did not find negative long-term impacts of cannabis use, but they similarly noted the great variability in the methodologies used across studies, and the limitation that most studies did not account for cognitive functioning prior to onset of cannabis use (Schreiner and Dunn 2012). Longitudinal, prospective research in HIV(−) participants has demonstrated that regular cannabis use was associated with broad neuropsychological decline over time, particularly among early initiators (people who began using cannabis before 18 years of age), and that cessation of use did not lead to fully restored cognitive function among early initiators (Meier et al. 2012). However, a longitudinal study that followed >2000 young adult cannabis users for 8 years found that cessation of use was associated with improved or normalized performance on a verbal learning task (Tait et al. 2011).

In HIV(+) individuals, the effects of cannabis use were similarly mixed (Table 3). One study of poly-substance users found independent effects of lifetime cannabis dependence and HIV on procedural learning tasks, and an additive effect with HIV-serostatus and cannabis dependence on worse complex motor skill performance (Gonzalez et al. 2011). Another study found that cannabis use may worsen the effects of HIV-related memory impairment, and that effect was driven by early initiators (Skalski et al. 2018). Amongst a sample of primarily abstinent cannabis users that included antiretroviral-treated HIV(+) individuals, however, cognitive performance in multiple domains were relatively normal (Chang et al. 2006a). Other research has shown that the frequency of cannabis use affected cognitive performance. Among HIV(+) persons, self-reported cognitive function is worse in current marijuana users compared to non-users (Lorkiewicz et al. 2018). In one study, HIV(+) moderate/heavy cannabis users performed worse in learning and memory tasks compared to HIV(+) light users, and HIV(+) non-users, as well as HIV(−) controls (Thames et al. 2016). However, in this same sample, current CD4 counts were lower and current viral load was higher among HIV(+) non-users compared to HIV(+) users, suggesting cannabis use may have been protective in maintaining immune function. Furthermore, amongst cannabis users, HIV(−) participants performed worse with greater marijuana use but not in the HIV(+) participants (Thames et al. 2017); these findings suggest that cannabis use did not additionally impact cognitive function in the HIV(+) patients.

HIV disease progression also appears to be an important factor moderating the effects of cannabis use. One study found that cannabis use was associated with greater memory deficits or impairment in persons with symptomatic HIV infection but not in HIV(−) participants or participants with asymptomatic HIV infection (Cristiani et al. 2004). Therefore, studies that focused on individuals who are asymptomatic or have well-controlled HIV infection may not observe the deficits associated with cannabis use that are typically seen in advanced HIV disease (Kennedy and Zerbo 2014).

Discussion

Challenges to clinical research on impact of cannabis use in the brain

The diverse and sometimes discrepant findings in the extant literature are likely the result of differences in design and research methodology (Byrd et al. 2011; Kennedy and Zerbo 2014), but also challenges inherent to the study of cannabis itself. Heterogeneity in the chemical composition of different strains of cannabis may contribute to confounding effects. Indeed, levels of the different cannabinoids in cannabis are even dependent on the maturity of the plant at the time of harvest (Aizpurua-Olaizola et al. 2016). The percentage of THC in cannabis has risen steadily over the past decades, from ~3.5–7% in the 1990s to 9–17% in recently confiscated samples, and some edible forms or smoking cannabis oil extracts (“dabbing”) may deliver up to 50–80% THC (Center for Behavioral Health Statistics and Quality 2015; Mehmedic et al. 2010). Higher THC levels may increase risk of addiction (Freeman and Winstock 2015). Perhaps one of the most important methodological issues is the varying characterization of cannabis use amongst the studies. While exposure and dosage can be quantified and controlled in the lab setting, accurately assessing these variables outside of controlled experimental studies has significant challenges. For example, operationalization of cannabis use varies from frequency- or quantity-based categorizations to defining groups by diagnostic symptoms associated with use. Some studies took into account the age of onset and lifetime exposure to cannabis, whereas other studies only considered current use patterns. Age of onset and duration of exposure to cannabis are also important factors to evaluate, but they are often not assessed rigorously or accurately. Additionally, many studies rely heavily on self-report without objective biological markers to verify or quantify the cannabinoid levels.

One factor that is typically not considered or reported in most studies is whether HIV infection preceded cannabis use or the reverse. While many people may use cannabis prior to their HIV diagnosis, there are certainly others who initiate use after diagnosis, either recreationally or to manage the adverse symptoms associated with the disease or medication side effects. This distinction is important in terms of understanding the interactive or additive effects. No data is available regarding whether initiating cannabis use after HIV infection increases the likelihood of cannabis abuse or whether ongoing cannabis use at the time of infection worsens HIV-mediated neurotoxicity or hastens overall disease progression. The consequences of HIV in chronic cannabis users with an already-disrupted cannabinoid system may be quantitatively and qualitatively different than the effects seen in persons with HIV who initiated cannabis use after diagnosis.

Poly-substance use further muddles the picture, given the high prevalence of other substance use amongst cannabis users. Additionally, in studies that are focused on cannabis use, persons concurrently using alcohol and tobacco on a daily basis are often not excluded. Furthermore, in many studies that otherwise require negative urine drug results at the time of scanning or cognitive testing, a positive urine drug test for cannabis is typically allowed because cannabis can take weeks to clear from the system, and some of the antiretroviral or pain medications (e.g., efavirenz, ibuprofen, etc.) can lead to a false positive THC toxicology in the urine. Furthermore, a large proportion of HIV(+) individuals are using cannabis medicinally in states that have legalized medical marijuana. Polysubstance use presents a challenge to parsing out the effects of an individual substance, particularly cannabis. In some studies, it may be that cannabis use caused profound effects which are not accounted for due to the confounding effects of other drugs, tobacco and alcohol use.

Cannabis as a therapeutic agent

In the changing landscape of cannabis legalization, the proliferation of CBD products, and the availability of drug formulations such as dronabinol, more scientific data are needed to inform policies. Mechanistic studies on the effects of the individual chemical components of cannabis are needed in order to understand both the short term and long-term effects of these agents on neuroHIV. In particular, it is critical to investigate the independent or combined effects of the two principal components of cannabis, THC and CBD. Observational studies have somewhat limited utility in parsing the precise mechanisms involved in neurocognitive abnormalities because there is no practical way to comprehensively assess the amount and potency of cannabis consumed, and perhaps most importantly, the THC:CBD ratio. Ideal experiments would administer known quantities of THC, CBD, the two combined, or a placebo to determine the individual and combined effects of cannabis on neurocognitive function.

However, observational research on cannabis users in the “real world” remains a critical area of study. Given the abundance of medicinal and recreational use among persons with HIV, cannabis use will likely remain disproportionately common in this population. For example, dronabinol was approved by the US Food and Drug Administration for HIV-associated wasting in the 1990s, with no appreciable decrease in marijuana use rates among HIV(+) persons as a result of its implementation (Levine et al. 2015). Research has since shown that dronabinol, while effective in reducing cannabis withdrawal symptoms, does not appear to help in reducing cannabis use (Copeland and Pokorski 2016). Additionally, cannabis contains many other cannabinoids, albeit in smaller quantities than THC and CBD, and these other cannabinoids are also biologically active. Therefore, it remains important to conduct longitudinal clinical research to identify the long-term impacts of cannabis use on neuroHIV. Similar to research on alcohol use, we may be able to identify threshold levels where use transitions from safe to hazardous, in order to guide recommendations for therapeutic use.

Need for integrative and multi-disciplinary research

Research in other fields has shown that brain alterations can precede the expression of cognitive impairment by years (Buckley et al. 2017; Contreras et al. 2015; Smith et al. 2007). For neuroHIV, this means that the sole use of cognitive performance measures are insufficient in describing or prognosticating the causal mechanisms for HAND in cannabis users. The combination of neuroimaging methodology with neuropsychological assessment and detailed substance use assessment is necessary to advance our understanding of HIV neurologic disease (Table 4). As shown in the studies summarized above, although brain structural changes were documented in cannabis users and HIV(+) individuals, cognitive performance and interactive effects between the combined conditions were not evident. However, other imaging techniques such as functional MRI or ¹H MRS were more sensitive in demonstrating the interactive or additive effects of HIV and cannabis use, even in those who were primarily abstinent. The relatively normal cognitive function in HIV(+) cannabis users is partly due to the brain’s ability to reorganize and use the available brain reserve to compensate for abnormalities in brain networks in order to maintain relatively normal cognition. Lastly, neuropsychological testing alone may be insufficient in determining the combined effects of HIV and chronic cannabis use on everyday activities important for a relatively decent quality of life. As such, further studies with accurate assessments of these measures are needed. Despite the relatively normal cognitive performance on standardized neuropsychological tests, tests that assess real-world functioning (e.g., driving under the influence of cannabis, motivation) are needed to further assess the combined effects of HIV and chronic cannabis use.

Table 4.

Methodological considerations for clinical studies on the combined effects of HIV and cannabis on brain and cognitive function:

Since preclinical studies demonstrated how the activation of endocannabinoids may influence inflammatory cytokines/chemokines that may further contribute to HIV-mediated neurotoxicity, quantification of these inflammatory biomarkers may provide additional insights regarding the combined effects of cannabis (THC or CBD) and neuroHIV.

Conclusion

While the causal relationship between cannabis and HIV neuropathogenesis is unclear, the summarized findings from both preclinical and clinical studies demonstrated that the relationships are complex. Several research priorities for HIV(+) cannabis users remain to be investigated further (Table 5). Such complexity, which is attributed to varied study methodologies, underscores the importance of well-controlled, longitudinal studies that integrate neuroimaging and other biomarkers, as well as “real world” functional assessments to identify the neural mechanisms through which HIV and cannabis interact to disrupt cognitive function or behaviors. Delineation of neural and behavioral deficits associated with concurrent cannabis use and HIV infection are needed to provide future recommendations on safe cannabis consumption in this population, while further elucidation of the potential neuroprotective roles of modulating the endocannabinoid system may guide development of novel therapies for HAND.

Table 5.

Research Priorities

• Determine the long-term neurocognitive consequences of chronic cannabis use

• Identify the role of mediating mechanisms for the additive or interactive effects of HIV and cannabis use (e.g., neuroinflammation)

• Examine the potential moderating role of “time of initiation of use” - before or after HIV diagnosis?

• Investigate the differential or combined effects of CBD versus THC (and other cannabinoids)

• Establish evidence related to the medicinal benefits of cannabis or cannabinoids for HIV disease