Effects of acute THC challenge on behavior and neuroinflammation in HIV-1 Tg26 mice vary based on HIV status, chronic THC history, and sex

Abstract

Δ⁹-tetrahydrocannabinol (THC) has been studied for its neuroprotective benefits in disease and its ability to improve HIV-1-related symptoms in clinical and preclinical models. Chronic THC administration may cause reduced sensitivity to antinociceptive, hypothermic, and anxiolytic effects following acute THC administration, and HIV status may further influence these effects. Thus, the present study investigated the effects of an acute THC challenge after chronic THC exposure on behavioral and neuroinflammatory measures using the HIV-1 Tg26 neuroHIV mouse model. HIV-1 Tg26 transgenic [Tg26(+/−), n=32(16f)] mice and their control littermates [Tg26(−/−), n=31(16f)] received subcutaneous injections of vehicle solution or THC (3 mg/kg), once a day, 5 days per week, for 90 days. After a 7-day drug-free period, all mice were given a high dose of THC (10 mg/kg; intraperitoneally), and their body temperature, antinociception, locomotor activity, and elevated plus maze data were collected. To assess inflammation, cytokine/chemokine levels were assessed via Bio-Plex, and microglial quantification and microglial CCL3/MIP-1α were assessed via immunohistochemistry in various brain regions. THC metabolite levels in the plasma were also collected. A chronic THC history resulted in minor behavioral/physiological changes (e.g., increase in body temperature but no effects on antinociception or locomotor activity) but overall decreases in proinflammatory and anti-inflammatory cytokines/chemokines in various brain regions of female mice. Importantly, across behavioral measures, a chronic THC history attenuated the efficacy of the acute high THC challenge dose, resulting in reduced THC-induced hypothermia, antinociception, and hypolocomotion, especially in females, and occasionally in a genotype-dependent manner. In the elevated plus maze, the acute THC challenge increased anxiety-like behavior in female mice with a chronic THC history compared to chronic vehicle history females, whereas no effects were noted in males. Further, regardless of microglial quantity, Tg26(+/−) mice showed high microglial-CCL3/MIP-1α co-occurrence in a sex- and brain-region dependent manner (e.g., BLNa in females & DS in males). The data suggest that female mice may develop reduced sensitivity to THC’s hypothermic, antinociceptive, and anxiolytic effects, and that this insensitivity development may depend on HIV genotype. The sex and genotype effects seen in the behavioral assays may be elucidated by differential effects in the inflammatory measures.

1. Introduction

Human immunodeficiency virus (HIV) is considered a global health issue that has impacted millions of individuals since the epidemic began in the 1980s, and as of 2022, roughly 38 million people were living with HIV (Hemelaar, 2012). With the development of combined antiretroviral therapy (cART) in the 1990s, the mortality rate, lifespan, and health outcomes for people living with HIV (PLWH) have drastically improved (Nakagawa et al., 2013). Despite cART, the central nervous system (CNS) serves as a major HIV viral reservoir, contributing to the disruption of the blood-brain barrier (BBB) and allowing the circulation of HIV-infected CD4+ T-cells and monocytes (Ash et al., 2021; Osborne et al., 2020). This induces the infection and activation of macrophages such as microglia that release neurotoxic viral proteins and proinflammatory chemokines and cytokines, including macrophage inflammatory proteins (MIP), interleukins (IL), and interferons (IFN; Saylor et al., 2016; Wang et al., 2017). Thus, the prolonged disease state leads to issues such as HIV-associated neurocognitive disorder (HAND), which encompasses neuroinflammation and neuronal dysfunction, and psychiatric disorders such as anxiety and depression (Ash et al., 2021; Cañizares et al., 2014; Clifford & Ances, 2013; Costiniuk & Jenabian, 2019; Kovalevich & Langford, 2012). Given the wide range of pathologies associated with HIV, gene therapies, pharmacological medications, and adjunct therapies continue to be studied to treat HIV infection or alleviate its symptoms (Littlewood & Vanable, 2008; Rossi et al., 2007; Stolbach et al., 2015). One such adjunct therapy involves the use of cannabis; multiple studies have explored the effects of medical cannabis on HIV and its symptoms (Lutge et al., 2013; Mills et al., 2021; Woolridge et al., 2005). Notably, in North America, cannabis use is more prevalent among PLWH than it is among the general population, and as of 2022, roughly 34–37% of PLWH currently used cannabis (Basova et al., 2022; Wardell et al., 2023). Preceding literature suggests that the anti-inflammatory effects of cannabis can improve chronic neuroinflammation, pain, and anxiety (Mills et al., 2021; Saito et al., 2012; Woolridge et al., 2005). In PLWH, chronic cannabis use has been shown to improve systemic inflammation by reducing levels of proinflammatory cerebrospinal fluid markers and inflammatory monocytes (Watson et al., 2021). Furthermore, cannabis use has been associated with lower levels of proinflammatory chemokines and cytokines, including CCL3/MIP-1α and IFN-γ, which regulate the immune functions of microglia (Giorgi et al., 2021; Rock et al., 2005; Watson et al., 2021). Interestingly, there are few studies focusing on chronic cannabis (i.e., cannabidiol [CBD], Δ⁹-tetrahydrocannbinol [THC]) administration in pre-clinical neuroHIV models. However, acute THC administration has been shown to have neuroprotective effects and enhance recovery in HIV pre-clinical models. A recently published study from our laboratory reported differential effects of acute THC on antinociception and memory in the HIV-1 Tat transgenic mouse model, where Tat expression increased antinociceptive effects of THC and lacked the limiting effects on recognition memory (Yadav-Samudrala et al., 2024). Additionally, depending on levels of cannabinoid type receptors 1 (CB₁R) and 2 (CB₂R), THC has been shown to improve immune response to HIV viral proteins via leukocyte modulation in an in vitro cell culture model (Chen et al., 2015; Klein et al., 1998).

While there are apparent effects of acute THC administration, chronic THC administration is a more clinically relevant model for medical cannabis use since PLWH use cannabis chronically. Hence, it is important to understand the effects of chronic cannabis exposure in the context of HIV therapies, and cannabis-based therapies must consider the effects of cannabinoids or cannabis administration after a history of chronic, long-term cannabis use. As such, literature has shown benefits of chronic THC dosing in non-HIV pre-clinical models. One study showed that a chronic low dose of THC can help rescue the effects of non-apoptotic neuronal injuries in aging mice and stabilize dendritic spines (Komorowska-Müller et al., 2023). In another study that utilized an Alzheimer’s mouse model, chronic THC treatment improved abnormal protein processing (Wang et al., 2022). While these studies involve non-HIV models, HIV in the CNS contributes to faulty synaptodendritic pruning and abnormal protein processing, demonstrating the potential efficacy of chronic THC administration in HIV models (Madden et al., 2020; Saylor et al., 2016).

Consequently, the goal of the current study was to explore the effects of chronic THC exposure as well as THC-based therapies in the context of a THC history, on behavioral and inflammatory measures in pre-clinical HIV models. This study utilized the HIV Tg26 mouse model, which is a well-established, non-infectious viral protein model for neuroHIV, since the model’s pathologies mirror those observed in PLWH undergoing cART with HAND. Behavioral assays included assessments of anxiety-like behavior, locomotor activity, antinociception, and body temperature. Additionally, we assessed cytokine and chemokine levels via a Bioplex assay, quantified inflammation via immunohistochemistry staining, and measured plasma THC metabolite levels using a custom liquid chromatography mass spectrometry (LC-MS) method.

2. Materials and methods

2.1. Animal subjects

This study included 3–7 month old (M = 127 days, SD = 33 days, or M = ~4 months, SD = ~1 month) inbred HIV-1 Tg26 transgenic mice, Tg26(+/−), n = 32(16f), and their sex and age-matched control littermates, Tg26(−/−), n = 31(16f). The Tg26 transgenic mouse model is a well-established neuroHIV mouse model that expresses multiple viral proteins, including the gag(p17), rev, vif, tat1 & 2, gp120(env), vpr, vpu, and nef proteins. This model was originally developed on an FVB/N background with the insertion of a transgene containing the NL4–3 HIV-1 genome with a deletion of a 3kb region encompassing the gag/pol genes, rendering the model non-infectious (Dickie et al., 1991). Due to the early mortality and multiple pathologies present in the original mice, they were backcrossed onto the C57BL/6J background for at least eight generations; the resulting mice demonstrate increased life expectancies and do not develop organ failure or kidney disease (Putatunda et al., 2019; Putatunda et al., 2018). As such, this modified HIV Tg26 mouse model developed on the C57BL/6J background was utilized for the current study.

Animals were kept in standard cages, with 2 to 4 mice per cage, on a reversed 12-hour light:dark cycle (lights on at 6:00 am). Behavioral testing took place during the dark cycle (9:00 am to 3:00 pm) and food and water were available to the mice ad libitum. All research procedures were conducted in strict accordance with the guidelines outlined in the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85–23), and the University of North Carolina at Chapel Hill Institutional Animal Care and Use Committee approved all procedures utilized.

2.2. Drug administration and behavioral testing

Mice were chronically given subcutaneous (s.c.) injections of Δ⁹-Tetrahydrocannabinol (THC, 3 mg/kg, Cayman Chemical, Cat# 12068) or vehicle (1:1:18; ethanol, Kolliphor^® EL, saline) once a day, five days a week, for ninety days (Figure 1). A dose of 3 mg/kg was utilized in this study as it has been shown to have antinociceptive and anxiolytic effects in adult mice without inducing locomotive deficits (Feliszek et al., 2016; Yadav-Samudrala et al., 2024c). A 3 mg/kg dose of THC is roughly equivalent to the average THC content in a joint, while a 10 mg/kg dose of THC is roughly equivalent to the average THC content in four joints (Zamberletti et al., 2012). The chronic 3 mg/kg dosing regimen is reflective of chronic (3 months) and heavy recreational use, and the 10 mg/kg acute challenge presents a high enough dose to challenge chronic, heavy users (Alvarez-Roldan et al., 2023). Drug treatments were randomized, and the doses were administered at a volume of 10 μL per gram of body mass.

Figure 1. Schematic of the timeline for drug administration and behavioral testing.

LA, locomotor activity; BT, body temperature; HP, hot plate; EPM, elevated plus maze; s.c., subcutaneous; i.p., intraperitoneal; D, day.

Injections were discontinued on day 90, and on day 95, post-chronic vehicle/THC administration body temperature and hot plate data were collected. Post-chronic vehicle/THC administration locomotor activity data were collected on day 96. On the following day, animals received an acute THC “challenge” consisting of a 10 mg/kg intraperitoneal (i.p.) injection; 30 minutes after the acute dose, the challenge locomotor activity data were taken. While both subcutaneous (s.c.) and i.p. routes of THC administration produce comparable potency in mice, s.c. injections are associated with reduced discomfort and stress, and were therefore used for the chronic dosing paradigm (Levin-Arama et al., 2016; Marusich & Wiley, 2023). In contrast, i.p. administration has produced more immediate and robust behavioral effects in our laboratory, thus making it the preferred route for the acute challenge (Levin-Arama et al., 2016). Additionally, this acute challenge i.p. dose was given after a 7-day washout period so that the animals did not experience the immediate effects of the drug while behavioral measures were taken for the first time. This was done to ensure that the measures showed the effects of a chronic period of usage instead of the immediate, acute dose in both drug-naïve (chronic vehicle history) or heavy drug use history (chronic THC history) individuals. Literature shows the importance of a long drug washout period (e.g., 5 days, 14 days, 21 days) following chronic THC treatment in order to minimize possible direct effects that result from THC treatment (Bilkei-Gorzo et al., 2017; Hasegawa et al., 2023; Lopez-Cardona et al., 2018; Nidadavolu et al., 2021). Thus, this 7-day washout period before the acute THC challenge dose was meant to facilitate the exploration of the dosing regimen efficacy in both drug-naïve and chronic drug history individuals.

After a 72-hour period, to allow clearance of the first acute THC challenge dose, on day 100, the animals were given a 10 mg/kg acute THC challenge dose once more, and after 30 minutes, their challenge body temperature and hot plate data were collected. After another 72-hour period, on day 103, the animals were given a final 10 mg/kg acute THC challenge dose, and after 30 minutes, challenge elevated plus maze (EPM) data were collected. On day 105, after more than 48 hours had passed, the animals were sacrificed for tissue collection.

2.3. Plasma and tissue collection

Plasma collection occurred on day 90, 30 minutes after s.c. vehicle or THC injections were administered. Submandibular bleed was performed as previously described, with no more than 100 μL of blood being taken in order to collect plasma (Yadav-Samudrala et al., 2024b). After submandibular bleed, the animals were returned to their cages. Plasma was subsequently used to quantify the concentrations of Δ⁹-THC and two of its metabolites, 11-Nor-9-carboxy-Δ⁹-tetrahydrocannabinol (THC-COOH) and 11-hydroxy-Δ⁹-tetrahydrocannabinol (THC-OH).

Tissue collection was performed on day 105, 48 hours after the acute THC injection on day 103. Animals were deeply anesthetized with isoflurane and sacrificed by decapitation. The brains were removed and sagittally bisected. The right half of the brain was dissected into seven brain regions (prefrontal cortex, striatum, amygdala, hippocampus, brainstem, cerebellum, spinal cord) according to the protocol described in Aboghazleh et al. (2024), and snap frozen in liquid nitrogen. The prefrontal cortex, hippocampus, and spinal cord are regions that are not directly related to the behavioral outcomes in this study, but as they are involved in neuroHIV pathology, they are included in Supplementary Table S4. These samples were stored at −80°C until they were sent to University of North Carolina at Chapel Hill Respiratory TRACTS Core for cytokine/chemokine analyses.

The left half was postfixed in 4% paraformaldehyde (ThermoScientific, ParaFormaldehyde, 4% in PBS, Cat# J19943-K2) for 6 hours on a shaker (ThermoScientific, Compact Digital Rocker, Cat# 88880019) at 4°C. Brains were then washed three times (1 hour per wash) in phosphate buffered saline (PBS) on a shaker at 4°C. Brains were placed in a 30% sucrose solution and agitated for 24 hours on a shaker at 4°C. Finally, brains were embedded in Tissue-Tek OCT compound and frozen on dry ice. Samples were stored at −80°C until preparation for immunohistochemistry.

2.4. Behavioral tests

2.4.1. Body temperature

Body temperature was taken with a non-contact infrared thermometer (Fisher Scientific, Cat# 12-894-005). Mice were held by the tail adjacent to an open cage such that their forepaws rested on the edge of the open cage. When mice were in this position and relatively less mobile, the thermometer was held near the exposed abdominal surface, and body temperature measurements were recorded. Additionally, the percent change in body temperature from post-chronic vehicle/THC administration measures to acute THC challenge measures was calculated and is shown in Supplementary Figure S1.

2.4.2. Hot plate assay

The hot plate test was utilized to assess spontaneous heat-evoked nociception, and testing was based on previously outlined procedures (Yadav-Samudrala et al., 2024b). Mice were placed on the surface of a hot plate (55 ± 0.1°C; IITC, Inc., MOD 39, Woodland Hills, CA), which was equipped with Plexiglas walls that were 15 cm high and 10 cm in diameter; this prevented possible escape. After the mouse was placed on the surface of the plate, it had to exhibit jumping behavior or lick/withdraw a paw before being removed. If an animal did not respond within 15 seconds, it was removed from the plate to avoid damage to tissue. The hot plate latency was recorded as time spent on the hot plate. Additionally, the percent change in hot plate latency from post-chronic vehicle/THC administration measures to acute THC challenge measures was calculated and is shown in Supplementary Figure S2.

2.4.3. Locomotor activity

Behavioral testing was based on previously outlined procedures (Yadav-Samudrala et al., 2024a). The apparatus used for testing was an open activity chamber (SD Instruments, Photobeam Activity System–Open Field, San Diego, CA, USA) which was 41 cm wide and 41 cm long and featured Plexiglas walls on each side. A video camera was mounted above the apparatus to record behavior, and testing was conducted in the dark cycle under red lights. For each session, a mouse was placed in the center of the chamber and recorded while it was allowed to explore the maze for 10 minutes. Each mouse had one session in the open activity chamber. Each of these recorded sessions was analyzed by the ANY-maze software (Version 4.3; Stoelting Co., Wood Dale, Il). During the video analysis, ANY-maze recorded a variety of measures, including distance travelled and mean speed. Additionally, the percent change in distance from post-chronic vehicle/THC administration measures to acute THC challenge measures and the percent change in mean speed from post-chronic vehicle/THC administration measures to acute THC challenge measures were calculated and are shown in Supplementary Figure S3.

2.4.4. Elevated plus maze

Behavioral testing, utilized to assess anxiety-related behavior, was based on previously outlined procedures (Arabo et al., 2014; Yadav-Samudrala et al., 2022). The apparatus used for testing was an elevated plus maze (San Diego Instruments, Cat# 7001–0316, San Diego, CA), a plus-shaped apparatus raised 38 cm from the ground. The apparatus features four arms that are 30 cm long and 5 cm wide. Two of the arms are closed arms with 15 cm tall beige walls, while the other two arms are unsheltered open arms. A video camera was mounted above the apparatus to record behavior, and testing was conducted in the dark cycle under red lights. The apparatus itself was placed under indirect lighting so that all four walls of the apparatus were lit in the same manner in each test (Yadav-Samudrala et al., 2024a). For each session, a mouse was placed in the center of the maze facing towards an open arm and recorded while it was allowed to explore the maze for 10 minutes. Each mouse had one session in the elevated plus maze. Each of these recorded sessions was analyzed by the ANY-maze software (Arabo et al., 2014). During the video analysis, ANY-maze recorded a variety of measures including total distance traveled and distance traveled in open/closed arms. Additionally, the percent distance traveled in the open arms of the maze was calculated by the following formula:

$$\mathit{\text{percent distance traveled in open arms}}=\frac{{\mathit{\text{distance traveled}}}_{\mathit{\text{open arms}}}}{\mathit{\text{distance traveled}}}\times 100$$

The total time spent in open/closed arms was also measured and is shown in Supplementary Figure S4.

2.5. Mass spectrometry

2.5.1. Standard preparation

THC, THC-OH, THC-COOH, and their respective deuterated standards were obtained from Cayman Chemical Company (Ann Arbor, MI) and prepared in Optima-grade methanol (Fisher Scientific, Hampton, NH). The calibration curve consisted of 10 standards, with the highest concentration at 250 ng/mL and subsequent levels prepared by serial 1:1 dilutions in methanol, yielding a lowest calibrator concentration of 0.488 ng/mL. All calibration standards contained internal standards at a final concentration of 100 ng/mL. In addition, two quality control (QC) standards were prepared at concentrations of 5 and 75 ng/mL.

2.5.2. Sample preparation

For sample preparation, 100 μL of plasma were transferred to a 2.0 mL Eppendorf tube for extraction, followed by 600 μL of 80:20 methanol:water. Samples were shaken for 15 minutes and then centrifuged at 20,000 rcf to form the protein pellet. The supernatant was transferred to another tube and dried down. Samples were reconstituted with 100 μL of 80:20 methanol:water, shaken for 10 minutes, and centrifuged again. The supernatant was again transferred to a separate tube and dried down. Samples were reconstituted with 100 μL of methanol with internal standard at 100 ng/mL.

2.5.3. Mass spectrometry procedure

Mass spectrometry analysis was conducted at the University of North Carolina at Chapel Hill Department of Chemistry Mass Spectrometry Core Laboratory to quantify Δ⁹-THC and two of its primary metabolites, THC-COOH, and THC-OH. Samples were analyzed using a TSQ Vantage triple quadrupole mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) coupled to a Waters Acquity H-Class UPLC system (Waters Corporation, Milford, MA, USA). Samples were introduced via a heated electrospray source (HESI) at a flow rate of 0.25 mL/min. HESI source parameters were set as follows: spray voltage, 3.5 kV; sheath gas (nitrogen), 35 arbitrary units; auxiliary gas (nitrogen), 30 arbitrary units; sweep gas (nitrogen), 0 arbitrary units; nebulizer temperature, 300°C; and capillary temperature, 350°C. Analytes were detected using multiple reaction monitoring (MRM) mode.

Chromatographic separation was performed on a Waters Acquity UPLC BEH C18 column (2.1 × 150 mm, 1.7 μm particle size). The mobile phases consisted of (A) water with 0.1% formic acid and (B) acetonitrile with 0.1% formic acid. The column temperature was maintained at 45°C, and the autosampler was held at 10°C. The gradient elution program was as follows: initial conditions of 65% A; decreased to 45% A at 2.0 minutes and held for 1.0 minute; further decreased to 20% A at 7.0 minutes; to 5% A at 8.0 minutes; then to 0% A at 10.0 minutes and held for 3.0 minutes. The mobile phase composition was returned to the initial conditions at 13.5 minutes and held for 2.5 minutes, resulting in a total run time of 16.0 minutes. The injection volume for all samples was 10 μL.

Data acquisition and analysis were performed using Xcalibur software (Thermo Fisher Scientific, Breman, Germany). All detected species were singly charged, as confirmed by the unit mass-to-charge (m/z) separation between monoisotopic and ¹³C isotopologue peaks for each compound.

2.6. Cytokine and chemokine analyses

2.6.1. Homogenization

At University of North Carolina at Chapel Hill Respiratory TRACTS Core, tissue homogenization was performed based on the protocol utilized in Hermes et al. (2020). Tissue homogenization was performed by adding one tablet of Roche cOmplete, EDTA-free Protease Inhibitor (Sigma-Aldrich, Cat# 11873580001) to 10 mL of Pierce IP Lysis Buffer (ThermoFisher, Cat# 87788), which was then vortexed to mix the tablet with the lysis buffer. Then, 300 μL of the IP buffer was added into a labeled soft tissue homogenizing bead tube CK-14 (Bertin, Cat# P000912-LYSK1-A.0). Tissue sample was added to the bead tube containing the lysis buffer. The tubes were shaken for 1 minute at 3700 rpm using a bead beater (Biospec, Mini Bead Beater 24, Cat# 112011). Then, the tubes were placed on ice for 2 minutes, after which they were shaken and iced twice more. The bead tubes were rotated for 30 minutes at 4°C using a platform rocker. The homogenate was then transferred into a new, labeled microcentrifuge tube, leaving the beads in the lysing tube. The homogenate was centrifuged at 14,000 rpm for 15 minutes at 4°C. The supernatant was then transferred to a new, labeled microcentrifuge tube, leaving the pellet behind. The supernatant was stored in a −80°C freezer until use.

2.6.2. Luminex assay

Bio-Plex Pro Mouse Cytokine, Chemokine, and Growth Factor assays (Cat# M60009RDPD) were performed at University of North Carolina at Chapel Hill Respiratory TRACTS Core to quantify cytokine/chemokine levels in the plasma. First, the assay buffer, wash buffer, and sample diluent were brought to room temperature, and the homogenized samples were thawed. Samples were prepared with a 1:4 dilution (sample:Bio-Plex sample diluent). A 1x wash buffer was prepared by mixing 10x stock by inversion, and then adding 60 mL (1 part) of 10x wash buffer to 540 mL (9 parts) distilled water. The Bio-Plex system was calibrated at this point. A single vial of standards was reconstituted in 500 μL of standard diluent; this was vortexed for 5 seconds and incubated on ice for 30 minutes. A fourfold standard dilution series and blank were prepared, ensuring that each vial was vortexed for 5 seconds before being transferred to the next vial in the dilution series. The 10x coupled beads were vortexed for 30 seconds, diluted to 1x in Bio-Plex Assay Buffer, and protected from light exposure.

The diluted 1x beads were vortexed for 10 to 20 seconds. To each well of the assay plate, 50 μL of vortexed diluted 1x beads were added. The assay plate was washed twice with 100 μL Bio-Plex Wash Buffer. Then, the samples, standards, and blank were vortexed, and 50 μL were added to each well. The plate was covered with sealing tape and incubated for 30 minutes on a shaker at room temperature at 850 ± 50 rpm. Separately, during the last 10 minutes of sample incubation, the 10x detection antibody was vortexed for 5 seconds and quick-spun to collect liquid. It was then diluted to 1x by adding 300 μL of 10x detection antibody to 2,700 μL of detection antibody diluent. After incubation was complete, the assay plate was washed three times with 100 μL wash buffer. The diluted (1x) detection antibody was then vortexed, and 25 μL were added to each well. The plate was once more covered with sealing tape and incubated on the shaker for 30 minutes at room temperature at 850 ± 50 rpm. With 10 minutes remaining in the detection antibody incubation, 100x streptavidin-phycoerythrin (SA-PE) was vortexed for 5 seconds and quick-spun to collect liquid. It was diluted to 1x by adding 60 μL 100x SA-PE to 5,950 μL assay buffer and protected from light exposure. After incubation, the assay plate was washed three times with 100 μL wash buffer. The diluted (1x) SA-PE was vortexed and 50 μL were added to each well. The plate was once more covered with sealing tape and incubated on the shaker for 10 minutes at room temperature at 850 ± 50 rpm. The plate was washed three times with 100 μL wash buffer, and then the beads were resuspended in 125 μL assay buffer. The plate was covered with sealing tape and shaken for 30 seconds at 850 ± 50 rpm. Finally, the sealing tape was removed, and the plate was read using low photomultiplier tube RP1 setting on the Bio-Plex 200 system for optimal sensitivity.

2.7. Immunohistochemistry

Using a Leica CM300 cryostat (Leica, Deerfield, IL), the O.C.T-embedded, frozen, post-fixed left hemisphere samples were cut in the lateral direction into 30 μm thick sagittal sections. Sections were placed in PBS in 12-well plates, sealed with ParaFilm (Amcor), and stored in a 4°C fridge for immunohistochemistry (IHC).

All the following steps were performed at room temperature, unless otherwise specified. Six tissue sections spaced around 350 μm apart were taken from each animal subject and transferred to a labeled, corresponding well in a 12-well plate. For each step, tissue was agitated in 2.5 mL of solution in 74 mm mesh inserts (Corning Life Sciences, Netwell Inserts, Cat# 29442–132) on a shaker (ThermoScientific, Compact Digital Rocker, Cat# 88880019) at 20 rpm. The staining protocol was adapted from Osman et al. (2020). First, the tissue was washed in PBS for 5 minutes, three times. Then, the sections were incubated in 10 mM sodium citrate solution pH 6.0 (Sigma Aldrich, Citrate Buffer, pH 6.0, 10x, Antigen Retriever, Cat# C9999) at 80°C for 30 minutes to enable antigen retrieval. This was followed by a PBS wash, the incubation of the tissue in 0.5% H₂O₂ solution for 30 minutes, and another PBS wash (3 times, 5 minutes each). The tissue was incubated in goat blocking solution (5% normal goat serum, 0.1% Triton X100) for 60 minutes before incubation for 48 hours at 4°C in primary antibodies (Mouse CCL3/MIP-1a R&D Systems, Biotechne, #AF-450-NA, 1:500; Rabbit Iba1, Wako #019–19741, anti Iba1, 1:500) mixed into goat blocking solution. CCL3/MIP-1α antibodies were used to detect CCL3 chemokines; Iba1 antibodies, which bind to ionized calcium binding adaptor molecule 1, were used to label microglia. It should be noted that Iba1 is not a specific marker for microglia and that it may also label macrophages, including infiltrating macrophages (Uff et al., 2022). However, since microglia are the primary resident macrophages of the CNS, in closer proximity to neurons, and more numerous than non-parenchymal brain macrophages, Iba1 staining is utilized for the purposes of this study (Lee et al., 2021; Jurga et al., 2020; Prinz et al., 2021; Sun & Jiang, 2024).

After incubation with primary antibodies, the tissue was washed in PBS and incubated in the dark for 60 minutes in secondary antibodies (Anti-mouse Alexa 488, Invitrogen #A21121, green, 1:500; Anti-rabbit Alexa 594, Invitrogen #A11012, red, 1:500) mixed into goat blocking solution. The tissue was kept covered and in the dark for every step hereafter. Following incubation, the tissues were washed in PBS, incubated in Hoechst 33342 (Molecular Probes, H3570; 1:2000) for 3 minutes, and then washed in deionized water for 5 minutes, three times. The sections were mounted on glass slides (Fisherbrand Superfrost Plus Microsope Slides, Cat# 12-550-15) and coverslipped with antifade reagent (Invitrogen ProLong Gold #P36930). For each animal, a total of 3–6 tissue sections were mounted. Once dry, the edges of the coverslip were sealed with nail polish and left to dry in the dark. To complete IHC for all 63 animals, this process was performed in three separate batches, with animals from different groups randomly spread across the three batches.

2.8. Confocal microscopy

Tissue was labeled with CCL3/MIP-1α and Iba1 primary antibodies in order to quantify microglia and the colocalization of these markers, aiming to identify one of the markers of a dystrophic microglial state among Iba1-labeled microglia (Johnson et al., 2011). A Zeiss LSM800 T-PMT laser scanning confocal microscope and ZEN 2018 Blue Edition software (Carl Zeiss, Inc., Thornwood, NY) were used to image the immunolabeled tissue at 20x and 63x. Images were taken in the dorsal striatum (DS), which is involved with movement and locomotor activity, the caudal periaqueductal grey (cPAG), which is involved with nociception such as pain propagation and modulation, and the basolateral nucleus of the amygdala (BLNa), which is involved with anxiety-like behavior (Asim et al., 2024; Knight & Goadsby, 2001; Kravitz & Matikainen-Ankney, 2022). These regions were determined in part by areas delineated in the Allen Mouse Brain Atlas; further information regarding the demarcation of these regions can be found in Supplementary Table S7.

For microglia quantification, one image was taken from each of the ROIs (DS, cPAG, BLNa) in 3–4 tissue slices per animal. For the colocalization analyses, one image was taken from each of the ROIs (DS, cPAG, BLNa) in each of the 3–6 tissue slices per animal. Imaging and channel parameters were held constant for each IHC batch, with slight variation in parameters (i.e., pinhole and wavelength intensity measures) between batches. These minor adjustments to the parameters, which were kept consistent across all images in each batch, were meant to better visualize staining, and these changes were made due to slight differences in staining intensity across the three separate IHC batches. Z-stack images were collected at 10 sections per stack, with a stack step thickness of 7.83 μm, and orthogonal projections were generated for each Z-stack, with one orthogonal projection for each region of interest per tissue section per animal. For microglial quantification, the resulting orthogonal projections were exported as 8-bit TIFF files of merged channel images and counted by a researcher blinded to the conditions. For colocalization analyses, the resulting orthogonal projections were exported as 8-bit TIFF files of both merged channel images and individual channel images for analysis in CellProfiler.

2.9. Image analysis

Images were analyzed using CellProfiler (Version 4.2.8). The IdentifyPrimaryObjects module was used to identify CCL3/MIP-1α staining and the IdentifySecondaryObjects module was used to identify Iba1 staining. The MeasureColocalization module was used to obtain metrics for colocalization between CCL3/MIP-1α and Iba1. Manders’ overlap coefficient (MOC) was used to quantify cooccurrence, which was used for colocalization analyses (Aaron et al., 2018). For more detailed information, see Supplementary Table S6.

2.10. Statistical analyses

Statistical analyses were performed using the Statistical Package for Social Sciences (Version 29.0.2.0 (20); IBM SPSS Statistics), and data graphing was performed with Prism (Version 10.4.0 (527); GraphPad Software). The alpha level for all analyses was set at 0.05. All data are expressed as mean ± SEM. To assess a history of drug use, three-way analyses of variance (ANOVAs) with sex (2 levels: female, male), drug status (2 levels: chronic vehicle history, chronic THC history), and genotype (2 levels: Tg26(−/−) mice, Tg26(+/−) mice) as between-subjects factors were performed for the data collected following the chronic vehicle/THC administration period (hereby referred to as “post-chronic vehicle/THC administration”), including body temperature, hot plate, locomotor activity, and elevated plus maze. To assess the acute THC challenge, mixed four-way ANOVAs were performed with timepoint (2 levels: post-chronic vehicle/THC administration, post-acute THC challenge) as a within-subjects factor, and sex, drug status, and genotype as between-subjects factors for body temperature, hot plate, and locomotor activity. Three-way ANOVAs were also performed for the cytokine assay and immunohistochemistry data (i.e., microglial quantification and microglial CCL3/MIP-1α co-occurrence) with sex, drug status, and genotype as between-subjects factors. For the plasma metabolite measures, only samples form animals that were chronically treated with THC were analyzed; thus, two-way ANOVAs were performed with sex and genotype as between-subjects factors. To further explore the specific treatment effects for each sex in these analyses, the dataset was split by sex and analyzed with two-way ANOVAs with drug status and genotype as between-subjects factors. Tukey’s post hoc tests were utilized following ANOVA analyses when applicable.

3. Results

3.1. Body temperature

The three-way ANOVA that was performed to assess the effects of a chronic vehicle or THC history demonstrated that after the chronic vehicle/THC administration period, mice with a chronic THC history showed significantly higher body temperature, F(1, 55) = 15.27, p < .001, η² = 0.217, compared to mice with a chronic vehicle history (Fig. 2A). This higher body temperature was noted for both male and female mice [females: F(1, 28) = 7.44, p = .011, η² = 0.210; males: F(1, 27) = 8.82, p = .006, η² = 0.246] (Fig. 2A, 2B). There were no other significant effects or interactions. A detailed summary of the statistical results can be found in Supplementary Table S1.

Figure 2. Effects of chronic THC drug administration (3 mg/kg, s.c.) on body temperature following acute THC challenge (10 mg/kg, i.p.).

Body temperature of female mice (A) and male mice (B) as observed after chronic vehicle/THC administration. Both female mice and male mice with a chronic THC history show significantly higher body temperature compared to their chronic vehicle history counterparts. Change in body temperature of female mice (A’) and male mice (B’) from post-chronic vehicle/THC administration measures to post-acute THC challenge measures. The acute THC challenge significantly reduced body temperature of both female and male mice compared to post-chronic vehicle/THC administration, but acute THC challenge-induced hypothermia was attenuated only in female mice with a chronic THC history compared to female mice with a chronic vehicle history. ^αmain effect of drug status in two-way or mixed three-way ANOVA (p < 0.05); ^#main effect of timepoint in mixed three-way ANOVA (p < 0.05); ^ωinteraction effect of timepoint and drug status in mixed three-way ANOVA (p < 0.05). n = 7–8 per group.

The mixed four-way ANOVA that was performed to assess the effects of an acute THC challenge demonstrated a significant timepoint effect, F(1, 55) = 97.47, p < .001, η² = 0.639, such that body temperature was higher post-chronic vehicle/THC administration (prior to the acute THC challenge) compared to post-acute THC challenge. When separating by sex, this higher body temperature post-chronic vehicle/THC administration was observed in both females, F(1, 28) = 48.27, p < .001, η² = 0.633, and males, F(1, 27) = 49.11, p < .001, η² = 0.645. Further, a timepoint x drug status interaction effect was noted, F(1, 55) = 11.86, p = .001, η² = 0.177, indicating that chronic vehicle/THC history influenced the efficacy of the acute THC challenge. In other words, THC-induced hypothermia due to the acute THC challenge was attenuated in mice with a chronic THC history compared to mice with a chronic vehicle history. Interestingly, this interaction effect was noted for female mice, F(1, 28) = 13.17, p = .001, η² = 0.320 (Fig. 2A’), but not for male mice (Fig. 2B’). Additionally, a main effect of drug status was noted, F(1, 55) = 36.48, p = .001, η² = 0.399, such that mice with a chronic THC history showed significantly higher body temperature compared to mice with a chronic vehicle history. This higher body temperature in mice with a chronic THC history was noted for both female, F(1, 28) = 25.96, p < .001, η² = 0.481, and male mice, F(1, 27) = 11.46, p = .002, η² = 0.298. A detailed summary of the statistical results can be found in Supplementary Tables S2 and S3.

3.2. Antinociception

The three-way ANOVA that was performed to assess the effects of a chronic vehicle or THC history demonstrated that after the chronic vehicle/THC administration period, mice with a chronic THC history showed significantly higher antinociception, F(1, 55) = 3.93, p = .053, η² = 0.067, compared to mice with a chronic vehicle history (Fig. 3A). When separating by sex, there were no significant differences. Additionally, Tg26(+/−) mice showed significantly higher antinociception, F(1, 55) = 15.42, p < .001, η² = 0.219, compared to Tg26(−/−) mice (Fig. 3A). This increased antinociception in Tg26(+/−) mice was observed in both male and female mice [females: F(1, 28) = 9.83, p = .004, η² = 0.260; males: F(1, 27) = 5.83, p = .023, η² = 0.178] (Fig. 3A, 3B). There were no other significant effects or interactions. A detailed summary of the statistical results can be found in Supplementary Table S1.

Figure 3. Effects of chronic THC drug administration (3 mg/kg, s.c.) on antinociception following acute THC administration (10 mg/kg, i.p.).

Hot plate latency of female mice (A) and male mice (B) as observed after chronic vehicle/THC administration. Both female and male Tg26(+/–) mice show significantly higher antinociception compared to their Tg26(–/–) counterparts. Change in hot plate latency in female mice (A’) and male mice (B’) from post-chronic vehicle/THC administration measures to post-acute THC challenge measures. The acute THC challenge significantly increased antinociception in both female and male mice compared to post-chronic vehicle/THC administration. Additionally, acute THC challenge-induced antinociception was attenuated in female and male mice with a chronic THC history compared to their respective counterparts with a chronic vehicle history. Interestingly, female Tg26(+/−) mice also demonstrated an attenuation of THC-induced antinociception compared to female Tg26(−/−) mice, which was not observed in male mice. ^βmain effect of genotype in two-way ANOVA and mixed three-way ANOVA (p < 0.05); ^#main effect of timepoint in mixed three-way ANOVA (p < 0.05); ^ωinteraction effect of timepoint and drug status in mixed three-way ANOVA (p < 0.05); ^ψinteraction effect of timepoint and genotype status in mixed three-way ANOVA (p < 0.05);^δinteraction effect of drug status and genotype status in mixed three-way ANOVA (p < 0.05). n = 7–8 per group.

The mixed four-way ANOVA that was performed to assess the effects of an acute THC challenge demonstrated a significant timepoint effect, F(1, 55) = 69.78, p < .001, η² = 0.559, such that antinociception was lower post-chronic vehicle/THC administration (prior to the acute THC challenge) compared to post-acute THC challenge. When separating by sex, this lower antinociception post-chronic vehicle/THC administration was observed in both females, F(1, 28) = 42.34, p < .001, η² = 0.602, and males, F(1, 27) = 30.40, p < .001, η² = 0.530. Further, a timepoint x drug status interaction effect was noted, F(1, 55) = 15.75, p < .001, η² = 0.223, indicating that chronic vehicle/THC history influenced the efficacy of the acute THC challenge. In other words, THC-induced antinociception due to the acute THC challenge was attenuated in mice with a chronic THC history compared to mice with a chronic vehicle history. This interaction effect was noted for both female, F(1, 28) = 9.49, p = .005, η² = 0.253 (Fig. 3A’), and male mice, F(1, 27) = 6.91, p = .014, η² = 0.204 (Fig. 3B’). Further, a timepoint x genotype interaction effect was noted, F(1, 55) = 39.40, p = .005, η² = 0.137, demonstrating that THC-induced antinociception due to the acute THC challenge was attenuated in Tg26(+/−) mice compared to their Tg26(−/−) counterparts. Interestingly, this interaction effect was noted for female mice, F(1, 28) = 10.49, p = .003, η² = 0.272 (Fig. 2A’), but not male mice (Fig. 2B’). Additionally, a main effect of genotype was noted, F(1, 55) = 10.78, p = .002, η² = 0.164, such that Tg26(+/−) mice showed significantly higher antinociception compared to Tg26(−/−) mice. This higher antinociception in Tg26(+/−) mice was noted for both female, F(1, 28) = 4.94, p = .035, η² = 0.150, and male mice, F(1, 27) = 5.94, p = .022, η² = 0.180. Also, a significant genotype x drug status interaction effect was noted, F(1, 55) = 4.53, p = .038, η² = 0.076, demonstrating that in Tg26(−/−) mice a chronic THC history increases antinociception (M = 11.13, SEM = 0.37) compared to a chronic vehicle history (M = 10.52, SEM = 0.30), whereas in in Tg26(+/−) mice a chronic THC history decreases antinociception (M = 11.55, SEM = 0.51) compared to a chronic vehicle history (M = 12.68, SEM = 0.38). This genotype x drug status interaction effect was noted for female, F(1, 28) = 39.36, p = .012, η² = 0.206, but not male mice. A detailed summary of the statistical results can be found in Supplementary Tables S2 and S3.

3.3. Locomotor activity

The three-way ANOVA that was performed to assess the effects of a chronic vehicle or THC history demonstrated that after the chronic vehicle/THC administration period, there were no significant differences for distance traveled during locomotor activity (Fig. 4A, 4B). A detailed summary of the statistical results can be found in Supplementary Table S1.

Figure 4. Effects of chronic THC drug administration (3 mg/kg, s.c.) on locomotor activity following acute THC administration (10 mg/kg, i.p.).

Distance traveled by female mice (A) and male mice (B) during locomotor activity as observed after chronic vehicle/THC administration. There were no significant differences. Change in distance traveled by female mice (A’) and male mice (B’) during locomotor activity from post-chronic vehicle/THC administration measures to post-acute THC challenge measures. The acute THC challenge significantly decreased distance traveled in both female and male mice compared to post-chronic vehicle/THC administration. Additionally, acute THC challenge-induced hypolocomotion was attenuated in female and male mice with chronic THC history compared to female and male mice with chronic vehicle history. Also, greater distance was traveled by Tg26(+/−) male mice compared to Tg26(−/−) males, which was not observed in female mice. ^βmain effect of genotype in mixed three-way ANOVA (p < 0.05); ^#main effect of timepoint in mixed three-way ANOVA (p < 0.05); ^ωinteraction effect of timepoint and drug status in mixed three-way ANOVA (p < 0.05). n = 7–8 per group.

The mixed four-way ANOVA that was performed to assess the effects of an acute THC challenge demonstrated a significant timepoint effect, F(1, 55) = 23.06, p < .001, η² = 0.295, such that distance traveled was greater post-chronic vehicle/THC administration (prior to the acute THC challenge) compared to post-acute THC challenge. When separating by sex, this greater distance traveled post-chronic vehicle/THC administration was observed in both females, F(1, 28) = 13.47, p = .001, η² = 0.325, and males, F(1, 27) = 9.82, p = .004, η² = 0.267. Further, a timepoint x drug status interaction effect was noted, F(1, 55) = 11.34, p = .001, η² = 0.171, indicating that chronic vehicle/THC history influenced the efficacy of the acute THC challenge. In other words, THC-induced hypolocomotion due to the acute THC challenge was attenuated in mice with a chronic THC history compared to mice with a chronic vehicle history. This interaction effect was noted for both female, F(1, 28) = 6.02, p = .021, η² = 0.177 (Fig. 4A’), and male mice, F(1, 27) = 5.34, p = .029, η² = 0.165 (Fig. 4B’). Additionally, a main effect of genotype was noted, F(1, 55) = 7.67, p = .008, η² = 0.122, such that Tg26(+/−) mice showed significantly greater distance traveled compared to Tg26(−/−) mice. This greater distance traveled by Tg26(+/−) mice was noted for male, F(1, 27) = 7.95, p = .009, η² = 0.227, but not female mice. Additionally, a main effect of drug was noted, F(1, 55) = 4.62, p = .036, η² = 0.078, such that mice with a chronic THC history showed significantly greater distance traveled compared to mice with a chronic vehicle history. When separating by sex, there were no significant effects. Finally, there was a significant genotype x drug x sex interaction effect, F(1, 55) = 3.91, p = .053, η² = 0.066, thereby indicating that the manner in which a chronic vehicle/THC history affects the efficacy of an acute THC challenge is sex- and genotype- specific. A detailed summary of the statistical results can be found in Supplementary Tables S2 and S3.

3.4. Elevated plus maze (EPM)

For EPM measures, testing was conducted only once to avoid exposure effects that occur with repeated testing. Hence, data were not collected following the chronic drug administration period, and mice were only tested after the acute THC challenge. A detailed summary of the statistical results can be found in Supplementary Table S1. After the challenge, Tg26(−/−) mice showed significantly reduced distance traveled during EPM measures, F(1, 55) = 4.48, p = .039, η² = 0.079, compared to their Tg26(+/−) counterparts. When separating by sex, no genotype or drug status effects were noted (Fig. 5A).

Figure 5. Effects of chronic THC drug administration (3 mg/kg, s.c.) on anxiety-like behavior in the elevated plus maze following acute THC administration (10 mg/kg, i.p.).

(A) Distance traveled in elevated plus maze by female and male mice as observed after acute THC challenge administration. There were no significant differences. (B) Percent of total distance traveled in open arms by female and male mice. Female mice with a chronic vehicle history traveled a higher percent distance in the open arms compared to females with chronic THC history, indicating less anxiety-like behavior. (C) Percent of total distance traveled in closed arms by female and male mice. Female mice with a chronic THC history traveled a higher percent distance in the closed arms compared to females with chronic vehicle history, indicating more anxiety-like behavior. ^αmain effect of drug status in two-way ANOVA (p < 0.05). n = 7–8 per group.

For percent distance traveled in the open arms, mice with a chronic vehicle history traveled a significantly higher percent distance in the open arms, F(1, 55) = 4.11, p = .047, η² = 0.070, than mice with a chronic THC history. When separating by sex, female mice with a chronic vehicle history traveled a significantly higher percent distance in the open arms, F(1, 28) = 5.27, p = .029, η² = 0.158, than females with a chronic THC history (Fig. 5B). There were no significant effects for male mice.

For percent distance traveled in the closed arms, mice with a chronic THC history spent significantly more time in the closed arms, F(1, 55) = 4.62, p = .036, η² = 0.078, than mice with a chronic vehicle history. When separating by sex, female mice with a chronic THC history spent significantly more time in the closed arms, F(1, 28) = 5.56, p = .026, η² = 0.166, than females with a chronic vehicle history (Fig. 5C). There were no significant effects for male mice.

3.5. Cytokines and chemokines

Anti-inflammatory cytokine IL-10, proinflammatory cytokines IL-12(p70) and IFN-γ, and proinflammatory chemokine CCL3/MIP-1α are represented in Figure 6. Other cytokines and chemokines, including IL-2, IL-17, TNF-α, IL-1α, IL-1β, IL-3, IL-5, IL-6, IL-9, IL-12(p40), MCP-1, KC, MIP-1β, RANTES, Eotaxin, GM-CSF, G-CSF, IL-4, and IL-13 are represented in Supplementary Table S4.

Figure 6. Expression levels of anti-inflammatory cytokine IL-10, proinflammatory cytokines IL-12(p70) and IFN-γ, and proinflammatory chemokine MIP-1α in brain regions of interest.

Cytokine and chemokine expression levels as denoted by log(MFI*) for IL-10, IL-12(p70), IFN-γ, and MIP-1α in female and male mice in the amygdala, brainstem, and striatum. IL-10, an anti-inflammatory cytokine, is denoted by the boxed panel. Overall, a chronic THC history appears to reduce cytokine and chemokine expression in the amygdala, brainstem, and striatum, especially in female mice, but not particularly in male mice. ^αmain effect of drug status in two-way ANOVA (p < 0.05); ^βmain effect of genotype in two-way ANOVA (p < 0.05); ^δinteraction effect of drug status and genotype status in two-way ANOVA (p < 0.05). Followed by Tukey’s post hoc tests (p’s < 0.05): ^aVehicle Tg26(+/−) vs. THC Tg26(−/−); ^bVehicle Tg26(+/−) vs. THC Tg26(+/−); ^cVehicle Tg26(+/−) vs. Vehicle Tg26(−/−). n = 7–8 per group.

For IL-10 (Fig. 6A), there was a significant drug effect such that mice with a chronic vehicle history showed significantly higher corrected mean fluorescence intensity (MFI*) values in the amygdala, F(1, 55) = 7.80, p = .007, η² = 0.124, and in the brainstem, F(1, 55) = 7.94, p = .007, η² = 0.126, compared to mice with a chronic THC history. When separating by sex, these higher MFI* values were noted in females in the amygdala, F(1, 28) = 16.49, p < .001, η² = 0.371, the brainstem, F(1, 28) = 16.42, p < .001, η² = 0.370, and the striatum, F(1, 28) = 6.58, p = .016, η² = 0.190. There was a significant sex x drug interaction effect in the amygdala, F(1, 55) = 6.65, p = .013, η² = 0.108. Tukey’s post hoc tests demonstrated that females with a chronic vehicle history showed higher MFI* values compared to males with a chronic vehicle history (p = .008).

For IL-12(p70) (Fig. 6B), there was a significant drug effect such that mice with a chronic vehicle history showed significantly higher MFI* values in the brainstem, F(1, 55) = 21.28, p < .001, η² = 0.279, and in the striatum, F(1, 55) = 7.52, p = .008, η² = 0.120, compared to mice with a chronic THC history. When separating by sex, these higher MFI* values were noted in females in the amygdala, F(1, 28) = 8.35, p = .007, η² = 0.230, the brainstem, F(1, 28) = 39.96, p < .001, η² = 0.588, and the striatum, F(1, 28) = 14.24, p < .001, η² = 0.337. There was a significant sex effect in the brainstem, F(1, 55) = 8.38, p = .005, η² = 0.132, such that females showed higher MFI* values compared to males. There was a significant sex x drug interaction effect in the brainstem, F(1, 55) = 10.45, p = .002, η² = 0.160. Tukey’s post hoc tests demonstrated that females with a chronic vehicle history showed significantly higher MFI* values compared to females with a chronic THC history (p < .001), males with a chronic vehicle history (p < .001), and males with a chronic THC history (p < .001). There was a significant drug x genotype interaction effect in the brainstem, F(1, 55) = 5.86, p = .019, η² = 0.096. Tukey’s post hoc tests demonstrated that Tg26(+/–) mice with a vehicle history showed significantly higher MFI* values compared to Tg26(–/–) with a THC history (p = .029), and Tg26(+/–) mice with a THC history (p < .001). Also, Tg26(–/–) mice with a vehicle history showed significantly higher MFI* values compared to Tg26(+/–) mice with a THC history (p = .023). When separating by sex, for males, there was a significant drug x genotype interaction effect, F(1, 27) = 4.20, p = .050, η² = 0.135. Tukey’s post hoc tests revealed no significant differences between groups.

For IFN-γ (Fig. 6C), there was a significant drug effect such that mice with a chronic vehicle history showed significantly higher MFI* values in the amygdala, F(1, 55) = 7.24, p = .009, η² = 0.116, compared to mice with a chronic THC history. When separating by sex, these higher MFI* values were noted in females in the amygdala, F(1, 28) = 12.66, p = .001, η² = 0.311, the brainstem, F(1, 28) = 5.31, p = .029, η² = 0.159, and the striatum, F(1, 28) = 4.28, p = .048, η² = 0.133. There was a significant sex x drug interaction effect in the striatum, F(1, 55) = 4.31, p = .043, η² = 0.073. Tukey’s post hoc tests revealed no significant differences between groups. When separating by sex, for females, there was a significant drug x genotype interaction effect in the amygdala, F(1, 28) = 5.29, p = .029, η² = 0.159. Tukey’s post hoc tests demonstrated that Tg26(+/–) females with a chronic vehicle history had significantly higher MFI* values than Tg26(–/–) females with a chronic vehicle history (p = .025), Tg26(–/–) females with a chronic THC history (p = .003), and Tg26(+/–) females with a chronic THC history (p = .002).

For CCL3/MIP-1α (Fig. 6D), there was a significant drug effect such that mice with a chronic vehicle history showed significantly higher MFI* values in the amygdala, F(1, 55) = 18.21, p < .001, η² = 0.249, compared to mice with a chronic THC history. When separating by sex, these higher MFI* values were noted in females in the amygdala, F(1, 28) = 15.01, p < .001, η² = 0.349, and in the striatum, F(1, 28) = 5.99, p = .021, η² = 0.176; higher MFI* values were also noted in males in the amygdala, F(1, 28) = 5.27, p = .030, η² = 0.163. There was a significant sex effect in the striatum, F(1, 55) = 4.82, p = .032, η² = 0.081, such that females showed higher MFI* values compared to males. There was a significant sex x drug interaction effect in the striatum, F(1, 55) = 5.90, p = .018, η² = 0.097. Tukey’s post hoc tests demonstrated that female mice with a chronic vehicle history showed significantly higher MFI* values compared to male mice with a chronic vehicle history (p = .008). When separating by sex, for females, there was a significant genotype effect in the amygdala, F(1, 28) = 4.80, p = .037, η² = 0.146, such that Tg26(+/–) females had significantly higher MFI* values than Tg26(–/–) females. Additionally, for females, there was a significant drug x genotype interaction effect in the amygdala, F(1, 28) = 5.77, p = .023, η² = 0.171. Tukey’s post hoc tests demonstrated that Tg26(+/–) females with a chronic vehicle history had significantly higher MFI* values than Tg26(–/–) females with a chronic vehicle history (p = .015), Tg26(–/–) females with a chronic THC history (p = .001), and Tg26(+/–) females with a chronic THC history (p < .001).

3.6. Plasma

Plasma was used to quantify the concentrations of Δ⁹-THC and two of its metabolites, 11-Nor-9-carboxy-Δ⁹-tetrahydrocannabinol (THC-COOH) and 11-hydroxy-Δ⁹-tetrahydrocannabinol (THC-OH). For Δ⁹-THC and THC-OH, no genotype effect was noted, but females showed significantly lower concentrations of Δ⁹-THC, F(1, 27) = 7.18, p = .012, η² = 0.210, and THC-OH, F(1, 27) = 19.13, p < .001, η² = 0.415, compared to males (Fig. 7). Separate analyses for each sex did not reveal significant differences. For THC-COOH, no significant differences were noted overall and when split by sex. A detailed summary of the statistical results can be found in Supplementary Table S5.

Figure 7. Concentrations of Δ⁹-THC, THC-COOH, and THC-OH in plasma for female and male mice.

Concentrations of Δ⁹-THC and its metabolites, THC-COOH, and THC-OH, in female and male mice. Female mice showed significantly lower concentrations of Δ⁹-THC and THC-OH compared to male mice. ^γmain effect of sex in two-way ANOVA (p < 0.05). n = 7–8 per group.

3.7. Immunohistochemistry

Microglial quantification analyses are shown in Figure 8. In the BLNa (Fig. 8A), there were no significant effects. When separating by sex, however, there was a significant effect of drug for females, F(1, 27) = 6.21, p = .019, η² = 0.187, such that female mice with a chronic THC history had significantly higher microglial counts compared to their chronic vehicle history counterparts. In the cPAG, there were no significant results. In the DS, (Fig. 8C), there was a significant drug effect such that animals with a chronic THC history had significantly higher microglial counts compared to their chronic vehicle history counterparts. When separating by sex, these higher microglial counts were noted in females, F(1, 27) = 5.37, p = .028, η² = 0.166, but not in males. No other significant effects were noted.

Figure 8. Microglial quantification of Iba1⁺ cells in brain regions of interest.

Counts of Iba1⁺ cells in the (A) basolateral nucleus of the amygdala (BLNa), (B) caudal periaqueductal grey (cPAG), and (C) dorsal striatum (DS) for female and male mice as imaged with confocal microscopy following immunohistochemistry experiments. Each data point is the average of counts from 3–4 sections per animal. Female mice with a chronic THC history had significantly higher microglial counts in the BLNa (A) and the DS (C) compared to their chronic vehicle history counterparts. (D) Representative sagittal image denoting the imaging brain regions of interest (i.e., BLNa, DS, cPAG). Image was taken from a male Tg26(+/–) chronic THC history mouse. (E-H) Representative images of Iba⁺ cells (red) in the BLNa for female mice in each respective group. (I-L) Representative image of Iba⁺ cells (red) in the BLNa for male mice in each respective group. ^αmain effect of drug status in two-way ANOVA (p < 0.05). n = 7–8 per group. Scale bar = 20 μm.

Colocalization analyses are shown in Figure 9, as represented by Manders’ overlap coefficient (MOC) values that quantify co-occurrence of the microglia marker Iba1 and the proinflammatory chemokine CCL3/MIP-1α. In the BLNa (Fig. 9A), there was a significant genotype effect, F(1, 53) = 8.06, p = .006, η² = 0.132, such that Tg26(+/–) mice showed significantly higher MOC values compared to Tg26(–/–) mice. When separating by sex, these higher MOC values were noted in females, F(1, 27) = 14.90, p < .001, η² = 0.356, but not males. There was also a significant sex x genotype interaction effect, F(1, 53) = 4.08, p = .049, η² = 0.071, with Tukey’s post hoc tests demonstrating that female Tg26(+/–) mice showed significantly higher MOC values compared to female Tg26(–/–) mice (p = .005) and male Tg26(–/–) mice (p = .047). In the cPAG (Fig. 9B), there were no significant effects. In the DS (Fig. 9C), there was a significant genotype effect, F(1, 53) = 4.64, p = .036, η² = 0.081, such that Tg26(+/–) mice showed significantly higher MOC values compared to Tg26(–/–) mice. When separating by sex, these higher MOC values were noted in males, F(1, 26) = 6.94, p = .014, η² = 0.211, but not females. No other significant effects were noted.

Figure 9. Presence of CCL3/MIP-1α co-occurrence in Iba1-labeled microglia in brain regions of interest.

Co-occurrence of Iba1⁺ cells with CCL3/MIP-1α in the (A) basolateral nucleus of the amygdala (BLNa), (B) caudal periaqueductal grey (cPAG), and (C) dorsal striatum (DS) for female and male mice as imaged with confocal microscopy following immunohistochemistry experiments. Each data point is the average of Manders’ Overlap Coefficients (MOC) from 3–6 sections per animal. Female Tg26(+/–) mice showed significantly higher MOC values compared to their Tg26(–/–) counterparts in the BLNa (A), while male Tg26(+/–) mice showed significantly higher MOC values compared to their Tg26(–/–) counterparts in the DS (C). (D) Representative image of CCL3/MIP-1α (green) co-occurrence in Iba⁺ cells (red) in the BLNa in a female Tg26(–/–) chronic THC history mouse. (E) Magnified inset of CCL3/MIP-1α (green) co-occurrence in an Iba⁺ cell. Individual red (F), green (G), and blue (H) channels from the magnified inset from the representative image. (I) Co-occurrence overlay for the representative image as produced by CellProfiler pipeline. Iba⁺ staining is outlined in magenta while CCL3/MIP-1α staining is outline in cyan. (E) Magnified inset of co-occurrence overlay for the representative image. ^βmain effect of genotype in two-way ANOVA (p < 0.05). n = 7–8 per group. Scale bar = 20 μm.

4. Discussion

The current study investigated the effects of chronic THC exposure as well as THC-based therapies in the context of THC history, on behavioral and inflammatory measures in an HIV Tg26 mouse model. The data revealed that chronic THC history had significant effects across most behavioral assays, especially for females, and often not for males. Similarly, a Tg26(+/–) genotype significantly affected a variety of behaviors, and results suggested that females may develop a reduced sensitivity to certain chronic THC effects depending on their HIV genotype. The sex and genotype effects seen in the behavioral assays may be elucidated by differences present in the inflammatory measures in the central nervous system.

In line with previous literature, the current study showed that chronic THC history resulted in significant reduction of body temperature, reduced antinociception, and increased anxiety-like behavior (Long et al., 2010; Metna-Laurent et al., 2017). These results were expected since both CB₁R and CB₂R are involved in these behaviors, and CB₁R in particular is thought to mediate cannabinoid-induced hypothermia, antinociception, anxiogenesis/anxiolysis, and catalepsy (Nealon et al., 2019; Rawls et al., 2002; Rodgers et al., 2003; Tambaro et al., 2013). Interestingly, however, after the acute THC challenge, female mice with a chronic THC history did not show significant differences in these behavioral measures, while their chronic vehicle history and male counterparts did. These results suggest that, after chronic THC treatment, female mice may develop reduced sensitivity to the hypothermic, antinociceptive, and anxiolytic effects of THC. While THC binds with and activates CB₁Rs to produce some of the effects associated with cannabis use, chronic THC administration has been shown to increase G-protein coupled receptor kinase (GRK) and β-arrestin levels in a brain region dependent manner (Haney et al., 2023; Howlett et al., 2011). These GRKs and β-arrestins are involved in the regulation and desensitization of CB₁Rs and may contribute to the development of tolerance to hypothermic, anxiolytic, and antinociceptive effects (Leo & Abood, 2021; Simone et al., 2015).

Notably, previous literature has shown that disruption of the β-arrestin mediated desensitization of CB₁Rs delays tolerance to certain CB₁R agonists, including THC, in male mice, but not in female mice (Henderson-Redmond et al., 2021). As such, it is possible that the speculated CB₁R desensitization, and any subsequent tolerance or insensitivity development, occurs via different mechanisms in females. This may explain the differential effects in THC-induced hypothermia, antinociception, and anxiogenesis/anxiolysis between males and females in our study, and quantitative measures of receptor levels and signaling should be assessed in order to build upon this research. Furthermore, in a previous study published by the lab which studied acute administration of THC in Tat transgenic mice, female Tat(+) mice showed a region-specific upregulation of CB₁Rs, indicating that there may be a role of changing receptor expression or activity that impacts these behavioral measures (Yadav-Samudrala et al., 2024). However, this needs to be further explored in a chronic paradigm as well as in Tg26 transgenic mice. Future studies could also explore cell-type (e.g., microglia) specific receptor levels.

Additionally, female mice showed lower plasma levels of Δ⁹-THC compared to male mice after this chronic THC administration paradigm, which could be due to sex differences in THC metabolism caused by differing levels of cytochrome P450 (CYP) and UDP-glucuronosyltransferase (UGT) enzymes. CYP and UGT enzymes play an important role in the metabolism of xenobiotics, including THC, in both humans and rodents (Bardhi et al., 2022; Renaud et al., 2011). Important CYP and UGT enzymes involved in THC metabolism include human CYP2C9, CYP3A4, and UGT1A9 enzymes, which are homologs and/or orthologs to mouse Cyp2c29, Cyp3a11, and Ugt1a9 enzymes, respectively (Li et al., 2009; Patilea-Vrana et al., 2019; Smith & Gruber, 2023; Wang et al., 2020; Xu et al., 2020). Cyp2c29, Cyp2a11, and Ugt1a9 tend to show higher expression or activity in female mice, suggesting the increased metabolism and clearance of Δ⁹-THC, especially into its primary metabolite, 11-OH-THC (Izumi et al., 2013; Löfgren et al., 2009; Lu et al., 2013; Renaud et al., 2011). This may explain the lower Δ⁹-THC plasma levels seen in female mice; however, lower plasma levels in female mice were also seen for 11-OH-THC, and no significant differences between male and female mice were noted in our study for 11-COOH-THC metabolites. Thus, more research is needed to assess plasma levels of THC metabolites in mice after chronic drug administration to better understand drug metabolism and elimination in mice. Furthermore, the current study did not assess THC and its metabolites after the acute THC challenge. However, previous publications from our lab have studied THC and its metabolites after acute THC administration in Tat transgenic mice (Yadav-Samudrala et al., 2024). This study showed sex- and genotype-dependent differences in plasma metabolite levels, especially for THC-COOH. Therefore, in future studies, it may be beneficial to explore how these differences change with an acute THC administration after chronic THC history, as performed in the current study.

Furthermore, while the Tg26(+/–) genotype resulted in increased antinociception, Tg26(+/–) female mice showed reduced antinociceptive effects after the acute THC challenge, suggesting that antinociceptive effects of THC may be dependent on HIV status, particularly for female mice. HIV-1 proteins such as those expressed in the Tg26(+/–) mouse model may interfere with analgesia induced by cannabinoid receptor agonists due to interactions between HIV-1 coreceptors and cannabinoid receptors (Palma et al., 2018). CC and C-X-C chemokine receptors (i.e., CCR3, CCR5, CCR2b, CXCR4) are types of chemokine receptors that act as HIV-1 coreceptors and are activated by HIV-1 protein-induced cytokines and chemokines such as C-C and C-X-C motif ligands (e.g., CCL3/MIP-1α, CCL3/MIP-1β, CCL5/RANTES, CXCL12/SDF-1α; Arimont et al., 2017; Connor et al., 1997; Guyon, 2014; Modi et al., 2006; Nickoloff-Bybel et al., 2021; Olivetta et al., 2003; Planes et al., 2018; Wu & Yoder, 2009). Previous literature posits that upon the activation of HIV-1 coreceptors like CXCR4, heterologous desensitization occurs between the coreceptor and CB₁Rs that may lead to the downregulation or desensitization of CB₁Rs, therefore attenuating the effects of cannabinoid receptor agonists like THC (Palma et al., 2018). Moreover, once these cytokines and chemokines bind to their respective coreceptors, the coreceptors undergo dimerization, leading to the activation of the JAK/STAT (Janus kinase/signal transducers and activators of transcription) pathway which regulates immune function, inflammation, and apoptosis, among other functions (Guyon, 2014; Rusek et al., 2023).

Furthermore, this dysregulation or atypical activation of the JAK/STAT pathway is commonly observed in a variety of neurodegenerative pathologies, including HIV-1 infection, and JAK/STAT pathway signaling promotes neuroinflammation by stimulating proinflammatory responses in microglia and astrocytes (Reece et al., 2022; Rusek et al., 2023; Yan et al., 2024). These microglial and astrocytic responses include the release of inflammatory factors and further production of both proinflammatory and anti-inflammatory cytokines and chemokines (Rusek et al., 2023; Simon et al., 2021; Yan et al., 2024). As such, it is surprising that we did not see a pronounced effect of the Tg26(+/–) genotype on the increased expression of cytokines and chemokines in our study, except for CCL3/MIP-1α (Supplementary Table S4). Due to the long-term central nervous system exposure of HIV-1 proteins since birth in this Tg26 mouse model, immune tolerance might be at play. Previous work done in our lab has shown that immune tolerance may develop after the course of prolonged exposure to the HIV-1 transactivator of transcription (Tat) protein. The Tat protein increases the efficacy of HIV viral transcription and contributes to neuropathology, neuroinflammation, and neurotoxicity by disrupting the blood brain barrier and endolysosome function; thus, short-term Tat exposure increases oxidative damage and impairs glutamatergic neurotransmission, thereby negatively impacting the neural environment (Clark et al., 2017; Hermes et al., 2020). However, it is notable that long-term Tat exposure demonstrated only minor effects on cytokine levels and no longer significantly increased microglial reactivity to chronic insult, indicating possible immune tolerance (Hermes et al., 2020). Similarly, long-term exposure to multiple HIV-1 proteins in the Tg26 mouse model may result in reduced reactivity to these proteins and therefore, no significant increase in cytokines and chemokines, which are involved in the immune response to HIV-1 protein insult.

While we only saw subtle effects of the Tg26(+/−) genotype on cytokines and chemokines in homogenate tissue (Supplementary Table S4), it should be noted that we saw a pronounced genotype effect in immunohistochemistry experiments colocalizing Iba1-labeled microglia and CCL3/MIP-1α-labeled cells, especially in the basolateral nucleus of the amygdala and the dorsal striatum; this suggests a cell-type specific mechanism of cytokine and chemokine release. The continued release of inflammatory markers via JAK/STAT pathway signaling results in prolonged exposure to inflammatory stimuli, causing issues such as microglial priming and leading to dystrophy in microglia, which may explain why the influence of the Tg26(+/−) genotype is more prominent in our microglia-specific immunohistochemistry data (Lima et al., 2022; Rauf et al., 2022). Depending on the environment, microglia can transition between basal conditions and reactive conditions, and they often exist in disease-associated states during neurodegenerative disease (Paolicelli et al., 2022; Soysa et al., 2022). Depending on various factors such as age, health, and microglial location, microglia often undertake various transcriptional signatures that contribute to a specific response or phenotype, many of which have varying degrees of proinflammatory and anti-inflammatory effects (Orihuela et al., 2016; Paolicelli et al., 2022). During sustained disease states, such as with HIV-1, despite their waning activity, fragmented microglia, often referred to as “dystrophic” microglia, continue to contribute to the neuroinflammatory response by producing chemokines such as CCL3/MIP-1α to recruit inflammatory cells; the upregulation of these factors, as seen in the current study, can lead to dysregulated neutrophil infiltration and the progression of neurodegeneration (Johnson et al., 2011). In order to provide more insight into the interplay between microglial function and the increase in inflammatory factors such as CCL3/MIP-1α, future experiments can quantify CCL3/MIP-1α levels outside of Iba1+ cells or co-stain CCL3/MIP-1α with other markers such as ferritin light chain (FTL). FTL stores intracellular iron, and high FTL levels are often implicated in the “dystrophic” microglial state due to possible dysregulation of iron homeostasis (Shahidehpour et al., 2021). Our data show that regardless of microglial count, Tg26(+/−) mice had higher co-occurrence with CCL3/MIP-1α-labeled cells in some brain regions, suggesting that this is not simply an effect of having more microglia in those brain regions. This CCL3/MIP-1α and FTL co-staining may lend credence to the idea that the microglia that are present in these regions are operating in a fragmented state that is contributing to the recruitment of inflammatory cells.

While cytokine and chemokine levels fluctuate with microglial function, the decrease in both proinflammatory and anti-inflammatory cytokines and chemokines that was observed in the current study occurred particularly in animals with a chronic THC history. This may be due in part to the influence of cannabinoids on CB₂R activity, which in turn mediates the JAK/STAT pathway, particularly in astrocytes and microglial cells (Ehrhart et al., 2005; Jeong et al., 2025; Rusek et al., 2023; Simon et al., 2021; Yan et al., 2024). In previous literature, cannabis has been shown to inhibit the JAK/STAT signaling pathway in T-cells via the activation of CB₂Rs, and CB₂R agonists have been shown to inhibit the JAK/STAT signaling pathway induced by cytokines and chemokines in microglial cells (Ehrhart et al., 2005; Jeong et al., 2025). This suggests that cannabinoids such as THC can improve the inflammatory response in T-cells and microglia and limit the continued release of both proinflammatory and anti-inflammatory cytokines and chemokines, as seen in the current study (Supplementary Table S4). Additionally, cannabis and cannabinoids such as THC have been shown to downregulate several proinflammatory cytokines and chemokines, including IL-12(p70), IFN-γ, and CCL3/MIP-1α, which showed reduced expression after chronic THC administration in our study (Blumstein et al., 2014; Graczyk et al., 2021; Klein et al., 2000). They have also been shown, albeit uncommonly, to downregulate anti-inflammatory cytokines and chemokines such as IL-10, which was also reduced following chronic THC administration in our study (Tan et al., 2024; Zaiachuk et al., 2023). These cytokines and chemokines contribute to changes in microglial proliferation, activity, and functional state, with IL-12(p70), IFN-γ, and CCL3/MIP-1α being associated with microglia responding with a more proinflammatory reactive phenotype, and IL-10 being associated with microglia responding with a more phagocytic reactive phenotype (Afridi et al., 2020; Gunalp et al., 2023; Laffer et al., 2019; Li et al., 2021; Paolicelli et al., 2022; Wang et al., 2008). Accordingly, IL-10-, TGF-β-, and glucocorticoid-responsive microglia often contribute to phagocytosis, debris clearance, matrix remodeling, tissue repair, and immunoregulation (Paolicelli et al., 2022; Wendimu & Hooks, 2022). Alternatively, lipopolysaccharide (LPS)- and IFN-γ-responsive microglia often release reactive oxygen and nitrogen species (RONS), along with proinflammatory cytokines and chemokines, such as IL-12(p70) and CCL3/MIP-1α, which contribute to the neuroinflammatory response, antigen presentation, and neurotoxicity, leading to tissue injury and behavioral deficits (Saylor et al., 2016; Wang et al., 2017; Wendimu & Hooks, 2022). Although microglial response states are highly specific to different developmental and immune circumstances, these changes in cytokine and chemokine expression may alter the induction of various microglial profiles, thus impacting the neuroimmune state.

In fact, if THC downregulates IFN-γ expression, as seen in our study, it may reduce IFN-γ-responsive microglia induction, thus limiting the release of RONS and proinflammatory cytokines and chemokines, including IL-12(p70) and CCL3/MIP-1α; this is in accordance with the findings of our study, which also noted the decreased expression of IL-12(p70) and CCL3/MIP-1α. Over the course of neurodegeneration associated with disease such as neuroHIV, the immune-complex microglial balance tips towards more detrimental reactivity due to regulatory T-cell suppression and increased IFN-γ production, leading to sustained neuroinflammation, “dystrophic” microglial activity, and impaired neuronal function (Masrori et al., 2022).

It is expected, then, that cannabinoid inhibition of the JAK/STAT pathway and the subsequent decrease in cytokines, chemokines, and inflammatory markers should ameliorate prolonged exposure to inflammatory stimuli, thereby limiting dystrophic microglial function; however, this was not seen in any of the ROIs in our immunohistochemistry experiments. Interestingly, we did observe that microglial counts were higher in the BLNa and the DS, but not in the cPAG, in animals with a chronic THC history compared to their chronic vehicle history counterparts; this was especially true for female mice. This can be explained by the current body of literature, which shows that there are brain-region dependent sex differences in microglial density, development, recruitment, and activity (Mouton et al., 2002; Lynch, 2022; Ocanas et al., 2023). However, because THC is meant to improve the inflammatory response, it is surprising that its administration increased microglial counts.

Since the current study was performed using a chronic administration paradigm, it is possible that chronic administration of THC simply resulted in the continued activation and proliferation of microglia in a sex-dependent manner, despite the fact the most of the existing literature focuses on how microglial activity is affected by chronic and subchronic prenatal or adolescent THC exposure (Cutando et al., 2013; Freels et al., 2023; Hasegawa et al., 2023; Pham et al., 2025; Zamberletti et al., 2015). One study that explored reactive microglia counts in the mPFC of adult rats after chronic adolescent THC exposure showed that female rats had more reactive microglia compared to male rats (Freels et al., 2023). Another study, which looked at microglial populations and activity in the amygdala after prenatal THC exposure, found an overall decrease in microglial phagocytic activity and microglial populations during development, but no significant differences between male mice and female mice (Pham et al., 2025). Finally, another study found that subchronic THC treatment during adolescence caused microglial apoptosis in the mPFC, but only in male mice and not in female mice (Hasegawa et al., 2023). Taken together, these studies might indicate that microglial functions and populations are impacted differentially in male and female rodents in a brain region- and sex-dependent manner. These differential effects may occur in part due to THC and microglial cannabinoid receptor interactions. Literature proposes that one common mechanism by which THC interacts with microglia is through CB₂Rs expressed on microglia; interestingly, chronic THC exposure may contribute to the modulation or dysfunction of these microglial CB₂Rs, thus exerting effects on microglial function (Zamberletti et al., 2015). Further research is needed to understand how these interactions differ based on sex and brain region, and also how chronic THC administration differentially impacts microglial characteristics in adult mice.

Other interactions between the endocannabinoid system, cannabinoids, and HIV viral proteins may also explain some of the varying effects of THC in HIV. THC and other CB₁R agonists suppress L-type and T-type voltage-gated calcium channels while CB₂R agonists suppress T-type voltage-gated calcium channels, both of which are important in neurotransmitter release, synaptic plasticity, and cell regulation; as such, irregular intracellular calcium mediation is often associated with neuronal degeneration, such as that seen in neuroHIV (Qian et al., 2017). Interestingly, HIV viral proteins such as Tat act as L-type calcium channel activators, suggesting that CB₁R and CB₂R agonists may counteract HIV viral protein activity via modulation of these activated calcium channels (Hu, 2016; Khodr et al., 2017; Martin, 2022; Qian et al., 2017). Additionally, Tat has been associated with presynaptic stimulation of glutamate release, leading to excitotoxicity, while CB₁R stimulation (e.g., via agonists such as THC) limits glutamate-mediated synaptic excitation, counteracting excitotoxicity (Starr et al., 2021). Alternatively, HIV accessory proteins Nef and Vpu are implicated in the downregulation of CD28, with Vpu also being implicated in the suppression of the NF-κB transcription factor; THC is known to suppress NF-κB binding activity and CD28 expression, which indicates that THC administration may enhance the effects of these proteins and change the HIV immune response (Langer et al., 2019; Ngaotepprutaram et al., 2015; Pawlak et al., 2017; Sido et al., 2015). THC and other cannabinoids can also alter the immune response to HIV viral proteins in a dose-dependent manner (Starr et al., 2021). Literature shows that T-cell response to the gp120 HIV viral protein changes based on cannabinoid levels and intracellular calcium levels (Chen et al., 2015; Starr et al., 2021). Since gp120 and other viral proteins trigger an inflammatory immune response, including the activation of the JAK/STAT pathway and cytokine and chemokine release, it is possible that THC administration alters this chain of responses and regulates subsequent inflammation (Chen et al., 2015; Guyon, 2014; Rusek et al., 2023; Starr et al., 2021). Therefore, the literature suggests that THC may interact differentially with multiple HIV viral proteins and cause changes in neuroinflammatory and behavioral responses to HIV.

While this study aims to contribute to the progress of HIV-1 therapies, there are also several potential shortcomings that must be considered. To begin with, cannabis remains an illicit substance in multiple countries or states, which would likely result in some obstacles throughout attempts to implement cannabis-based therapies. This is because of its potential for abuse, which is due to the THC compound in particular, the focus of this study (Kesner & Lovinger, 2021; Romero-Sandoval et al., 2018). As such, THC therapies should be implemented at mild to moderate dosages to avoid aversive effects and maximize beneficial effects (Kesner & Lovinger, 2021; Lichenstein, 2022). The frequency of doses should also be monitored since tolerance may result from the chronic maintenance of tissue THC levels, especially as reduced sensitivity to THC effects was observed in female mice in the current study (Kesner & Lovinger, 2021; Kubilius et al., 2018). It is true that all the animals in the current study received multiple acute injections, and while the use of multiple acute drug injections has been conducted in previous studies and seems to be the standard in pharmacological studies, it is worth noting when interpreting the neurochemical and neurobiological analyses due to possible differences in acute cannabinoid signaling (Kasten et al., 2019; Marusich & Wiley, 2023; Yadav-Samudrala et al., 2024). Regardless, given these findings, it would be beneficial to further explore how HIV-1 status impacts the development of insensitivity and tolerance differentially in females and males. Also, while literature on cannabinoid cessation in rodents states that peak effects of THC withdrawal from a very high dose (30 and 36 mg/kg) occur 1.5 to 3 days (12 to 96 hours) after cessation, effects of cessation may be different in mice that are treated for longer than one week, as was the case in our study (Paronis et al., 2022). This is another important factor to consider when interpreting these results. It is also notable that dose response curves for THC may differ for both rodent and human males and females, resulting in sex differences in THC sensitivity (Fogel et al., 2016; Moore et al., 2024). Thus, in future studies, it could be useful to utilize different doses based on sex.

Furthermore, the use of CBD rather than THC may be a productive alternative since it does not possess the same psychoactive properties as THC while still providing many similar anxiolytic and anti-inflammatory benefits, in addition to neuroprotective effects in HIV-1 infection (Brown & Winterstein, 2019; DeMarino et al., 2022; Yadav-Samudrala et al., 2024; Yndart Arias et al., 2023). Moreover, future studies could investigate the effects of administering combined ratios of cannabinoids such as THC and CBD or even combined cannabis compounds such as phytocannabinoids, terpenes, or flavonoids. When administered together, CBD has been shown to reduce some of the aversive effects of THC while potentiating other effects and increasing the overall efficacy of the treatment (Freeman et al., 2019). Cannabis compounds such as terpenes may improve the efficacy of cannabinoids by increasing blood-brain barrier permeability, although further research is needed to determine the extent of the benefit of combining compounds (Andre et al., 2024). Finally, in the present study, THC was administered via subcutaneous and intraperitoneal injections which, although shown to be translationally relevant, may be less so than oral gavage or aerosolized routes of administration since humans typically consume edibles or smoke cannabis (Marusich & Wiley, 2023). Regardless, this study supplements previous research that shows the benefits of THC and cannabinoids as adjunct therapies for HIV-1 symptom management (e.g., pain, anxiety, etc.) while also highlighting that a history of drug exposure may render THC therapies less effective, particularly for females.

Highlights.

• Chronic THC history significantly affects behavior, especially for female mice

• HIV status may influence reduced sensitivity to some chronic THC effects in female mice

• THC therapies for HIV-1 may be less effective in females with a history of drug use

Data Availability:

The data that support the findings of this study are available from the corresponding authors upon request.