HIV Tat Protein Selectively Impairs CB1 Receptor-Mediated Presynaptic Inhibition at Excitatory But Not Inhibitory Synapses

Abstract Despite the success of antiretroviral therapy in suppressing viral load, nearly half of the 37 million people infected with HIV experience cognitive and motor impairments, collectively classified as HIV-associated neurocognitive disorders (HAND). In the CNS, HIV-infected microglia release neurotoxic agents that act indirectly to elicit excitotoxic synaptic injury. HIV trans-activator of transcription (Tat) protein is one such neurotoxin that is thought to play a major role in the neuropathogenesis of HAND. The endocannabinoid (eCB) system provides on-demand neuroprotection against excitotoxicity, and exogenous cannabinoids attenuate neurotoxicity in animal models of HAND. Whether this neuroprotective system is altered in the presence of HIV is unknown. Here, we examined the effects of Tat on the eCB system in rat primary hippocampal cultures. Using whole-cell patch-clamp electrophysiology, we measured changes in retrograde eCB signaling following exposure to Tat. Treatment with Tat significantly reduced the magnitude of depolarization-induced suppression of excitation (DSE) in a graded manner over the course of 48 h. Interestingly, Tat did not alter this form of short-term synaptic plasticity at inhibitory terminals. The Tat-induced decrease in eCB signaling resulted from impaired CB1 receptor (CB1R)-mediated presynaptic inhibition of glutamate release. This novel loss-of-function was particularly dramatic for low-efficacy agonists such as the eCB 2-arachidonoylglycerol (2-AG) and Δ9-tetrahydrocannabinol (Δ9-THC), the main psychoactive ingredient in marijuana. Our observation that HIV Tat decreases CB1R function in vitro suggests that eCB-mediated neuroprotection may be reduced in vivo; this effect of Tat may contribute to synaptodendritic injury in HAND.


Introduction
Nearly half of all HIV-positive individuals experience some degree of neurological impairment (Heaton et al., 1995;Tozzi et al., 2005;Bateup et al., 2013;Saylor et al., 2016). While antiretroviral therapy effectively suppresses viral replication, the prevalence of HIV-associated neurocognitive disorders (HAND) remains high and continues to be a significant public health burden. HAND symptoms range from subclinical cognitive impairment to debilitating dementia (Antinori et al., 2007), and their severity correlates with loss of synaptic markers (Ellis et al., 2007). Within the CNS, the hippocampus, prefrontal cortex, and striatum are particularly vulnerable to damage caused by HIV (Masliah et al., 1992;Wiley et al., 1998;Moore et al., 2006). Currently, there are no effective therapeutics to combat the neurological deficits seen in HAND patients.
HIV indirectly affects neurons, where infected microglia and macrophage release viral proteins, inflammatory cytokines, and excitotoxins resulting in synaptodendritic damage and altered network function (Ellis et al., 2007;Green et al., 2019). The HIV protein trans-activator of transcription (Tat) is a potent neurotoxin that evokes the loss of excitatory synapses Shin and Thayer, 2013), an increase in the number of inhibitory synapses (Hargus and Thayer, 2013), and ultimately neuronal death (Eugenin et al., 2007). HIV Tat is present in the cerebrospinal fluid , brain tissue (Hudson et al., 2000), and sera of HIV-infected individuals at concentrations ranging from 2 to 40 ng/ml (Xiao et al., 2000). Once viral DNA integrates into the host genome, Tat is continuously expressed, even in the presence of combined antiretroviral therapy . The titer of antibodies against Tat negatively correlates with HAND symptoms (Bachani et al., 2013), suggesting that Tat accumulation is linked to cognitive deficits.
The endocannabinoid (eCB) system regulates many physiological processes of relevance to HAND including mood, anxiety, appetite, neuroinflammation, motor control, and neuroprotection (Rodríguez de Fonseca et al., 2005). The eCB system protects against excitotoxicity and attenuates epileptiform activity (Shen et al., 1996;Marsicano and Lutz, 1999;Nagayama et al., 1999). Notably, several studies using models of HAND have shown that pharmacological interventions targeted at the eCB system protect against HIV-mediated synaptodendritic damage. Cannabinoid type 2 receptor (CB 2 R) agonists attenuate neuroinflammation in a murine model of neuroAIDS (Gorantla et al., 2010) and protect human dopaminergic neurons from toxicity elicited by an HIV envelope protein (Hu et al., 2013). Application of the eCBs anandamide and 2-arachidonoylglycerol (2-AG) reduced HIV-induced increases in [Ca 21 ] i through their actions on cannabinoid receptors in vitro (Xu et al., 2017). Similarly, pharmacological administration of exogenous cannabinoids or inhibitors of eCB metabolism attenuated neuronal damage elicited by HIV proteins (Kim et al., 2011a;Hermes et al., 2018;Zhang and Thayer, 2018). Thus, targeting the eCB system in HAND has therapeutic potential; it is less clear how the presence of HIV in the CNS affects eCB signaling.
Excitotoxic stimuli alter the eCB system. For example, febrile seizures elevate levels of the proinflammatory cytokine interleukin 1b (IL-1b ), producing a long-lasting upregulation of eCB-mediated inhibition of GABAergic transmission, resulting in a subsequent increased susceptibility to seizure (Chen et al., 2003;Feng et al., 2016). Interestingly, eCB modulation of glutamatergic transmission was not affected in this model, setting a precedent for differential modulation of eCB signaling at inhibitory versus excitatory synapses under neuroinflammatory conditions. In transgenic mice expressing HIV Tat, CB 1 Rs are upregulated (Jacobs et al., 2019); whether this increase preferentially affects GABAergic neurons is unclear.
To address this question, we examined the effects of HIV Tat on eCB-mediated retrograde signaling. Treating rat primary hippocampal cultures with Tat reduced the magnitude of depolarization-induced suppression of excitation (DSE) over the course of 48 h. Tat impaired CB 1 Rmediated presynaptic inhibition of glutamate release, but not GABA release. These results suggest that exposure to the HIV protein Tat may reduce neuroprotection mediated by the eCB system and alter the sensitivity of excitatory synaptic networks to cannabinoids.

Cell culture
The hippocampus is particularly vulnerable to the neurotoxic effects of HIV (Ellis et al., 2007); thus, we chose rat hippocampal neurons grown in primary culture for this study. Cultures were prepared as previously described (Waataja et al., 2008). Briefly, maternal Sprague Dawley rats were euthanized by CO 2 inhalation under a protocol approved by the University of Minnesota Institutional Animal Care and Use Committee in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Male and female fetuses were removed on embryonic day 17, and hippocampi were dissected and placed in Ca 21free and Mg 21 -free HEPES-buffered HBSS. Cells were dissociated by manual trituration using flame-narrowed Pasteur pipettes of decreasing aperture and resuspended in DMEM without glutamine supplemented with 10% fetal bovine serum and penicillin/streptomycin (100 units/ml and 100 mg/ml, respectively). Dissociated cells were then plated at a density of 60,000-80,000 cells per dish on either a 35-mm Petri dish with a 10-mm cover glass bottom (MatTek) or a 25-mm-round cover glass (#1) precoated with Matrigel (150 ml, 0.2 mg/ml). Neurons were grown in a humidified atmosphere of 10% CO 2 and 90% air held at 37°C in an incubator. Cells were fed on days 1 and 7 by exchanging 75% of the medium with DMEM supplemented with 10% horse serum and penicillin/streptomycin. Cultures used in this study contained a mixture of 18 6 2% neurons, 70 6 3% astrocytes, and 9 6 3% microglia (Kim et al., 2011a). Cells were grown for 12-13 d in vitro (DIV).

Electrophysiology
Synaptic currents were recorded using the whole-cell configuration of the patch-clamp technique. Pipettes were pulled using a horizontal micropipette puller (P-87, Sutter Instruments) from glass capillaries with an outer diameter of 1 mm (Sutter Instrument) and pipette resistances of 3-5 MV. Membrane potential was held at À70 mV, and monosynaptic EPSCs/IPSCs were evoked with a bipolar platinum electrode (FHC) placed near a presynaptic neuron. DSE was evoked by a 15-s depolarization to 0 mV, which was previously shown to produce 50% DSE (Roloff et al., 2010). Each cell was exposed to a single stimulus. For DSE recordings pipettes were filled with the following intracellular solution: 120 mM potassium gluconate, 15 mM KCl, 6 mM MgCl 2 , 0.2 mM EGTA, 10 mM HEPES, and 5 mM Na 2 ATP, pH 7.3 with KOH, 290 mOsm/ kg. Metabotropic suppression of excitation (MSE) was evoked by bath application of the selective group 1 metabotropic glutamate receptor (mGluR) agonist DHPG (1 mM). For MSE recordings, pipettes were filled with the following: 113 mM potassium gluconate, 15 mM KCl, 6 mM MgCl 2 , 10 mM BAPTA tetrapotassium, 10 mM HEPES, 5 mM Na 2 ATP, and 6 mM CaCl 2 , pH 7.2 with KOH, 290 mOsm/kg. Depolarization-induced suppression of inhibition (DSI) was evoked by a 2-, 5-, or 15-s depolarization to 0 mV (Ohno-Shosaku et al., 2001;Wilson and Nicoll, 2001;Isokawa and Alger, 2005). Most cells were exposed to a single stimulus. Some cells from both control and Tat-treated groups were exposed to up to three different stimulus strengths (2, 5, 15 s); there was a 6-min break between administration of serial depolarizing stimuli. For DSI recordings pipettes were filled with the following: 140 mM KCl, 0.2 mM EGTA, 10 mM HEPES, 10 mM glucose, 5 mM MgATP, and 0.3 mM Na 2 GTP, pH 7.2 with KOH, 290 mOsm/kg. All recordings were performed at room temperature in an extracellular solution composed of the following: 140 mM NaCl, 5 mM KCl, 9 mM CaCl 2 , 6 mM MgCl 2 , 5 mM glucose, and 10 mM HEPES, pH 7.4 with NaOH, 325 mOsm/kg. For EPSC recordings, 10 mM bicuculline methochloride was added to the extracellular recording solution; for IPSC recordings, 10 mM CNQX and 50 mM APV were added. Solutions were applied by a gravity-fed superfusion system. Whole-cell currents were amplified with an AxoPatch 200B (Molecular Devices), low-pass filtered at 2 kHz, and digitized at 10 kHz with a Digidata 1322A (Molecular Devices) digitizer and pClamp v.9.2 software (Molecular Devices). Voltage pulses (0.1 ms) were applied at a fixed rate of 0.5 Hz for all experiments unless otherwise specified using a Grass S44 stimulator with a SIU-5 stimulus isolation unit (Grass Instruments).

Transfection
Rat hippocampal neurons were transfected on DIV10 or DIV11 with plasmid DNA using the calcium phosphate procedure described by Hargus and Thayer (2013). Briefly, hippocampal neurons were incubated for 30 min in DMEM supplemented with 1 mM kynurenic acid, 10 mM MgCl 2 , and 5 mM HEPES to reduce neurotoxicity. A DNA/ calcium phosphate precipitate containing 1 mg of total plasmid DNA per well was prepared, allowed to form for 60 min at room temperature, and added to the culture. After a 60-min incubation period, cells were washed once with DMEM supplemented with MgCl 2 and HEPES, then returned to conditioned media that had been saved at the beginning of the procedures.

Confocal microscopy and image processing
Cultures grown in MatTek glass-bottom Petri dishes were transferred to the motorized stage of a Nikon A1 laser-scanning confocal microscope 48 h after transfection with pCMV-PKDC1ab-GFP and viewed through a 60Â oil-immersion objective. The objective was focused to the middle of the soma in the z-dimension and held constant using the autofocus feature of the microscope. PKDC1ab-GFP was excited at 488 nm, and emission was collected at 500-550 nm. After recording (1 Hz) baseline images for 60 s, DHPG (1 mM) or vehicle was added directly to the media bathing the cell.

Experimental design and statistical analysis
All data are presented as mean 6 SEM. Electrophysiology data were analyzed using pCLAMP v.9.2. Live-cell imaging data were analyzed using NIS Elements and ImageJ. To minimize the effect of variables except the independent variable, experimental controls were run in parallel in all experiments. Only experiments with proper vehicle and/or drug controls were included in statistical analyses. For all experiments, an individual sample (n = 1) is defined as a single cell from a single coverslip/Petri dish. To account for inherent variability across primary cultures from week to week, each experiment was replicated over at least three separate cultures with a sample size of at least six cells/group. Because primary neuronal cultures are derived from a mixture of male and female rat embryos, the culture preparation protocol ensures an unbiased evaluation across both sexes. Hypothesis testing was performed with Prism 8 (GraphPad Software). Data were tested for normality using Bartlett's test. Experiments with two groups and one factor were analyzed using an unpaired two-tailed Student's t test. Samples with unequal variance were analyzed with Welch's twotailed t test. Experiments with 31 groups and one factor were analyzed using a one-way ANOVA followed by Tukey's post hoc test. Experiments with 21 groups and two factors were analyzed using a two-way ANOVA. Statistical significance was defined as p , 0.05. OriginLab 2019 was used for curve-fitting decay rate and concentration-response data. Unless derived from curve fitting, confidence intervals (CIs) were determined using bootstrap resampling by uploading the same raw data used for hypothesis testing to https://www.estimationstats.com. The results of statistical analyses are presented in the figure legends and Table 1.

HIV Tat inhibits DSE in a time-dependent manner
Depolarization of a postsynaptic hippocampal neuron evokes Ca 21 influx through voltage-gated Ca 21 channels that induces production of the eCB 2-AG (Ohno-Shosaku et al., 2001; Wilson and Nicoll, 2001). 2-AG then traverses the synaptic cleft in a retrograde manner to activate presynaptic CB 1 Rs, resulting in inhibition of glutamate transmission (Straiker and Mackie, 2005). We studied this process in primary rat hippocampal cultures following exposure to the HIV Tat 1-86 (Clade B) protein. We used the whole-cell configuration of the patch-clamp technique to voltage-clamp a hippocampal neuron at À70 mV. ESPCs were evoked by a concentric bipolar stimulating electrode positioned near a presynaptic neuron. After recording a stable baseline, DSE was evoked by depolarizing the postsynaptic cell to 0 mV for 15 s ( Fig. 1A; Roloff et al., 2010). Immediately following depolarization, the EPSC amplitude was inhibited by 53 6 5% (Fig. 1A,F). DSE was completely blocked by pretreatment with the CB 1 R inverse agonist SR141716A (100 nM, n = 4). To determine whether HIV Tat affects eCB signaling, we assessed DSE in Tat-treated cultures (50 ng/ml). Treatment with Tat for 4 h did not significantly affect DSE (Fig. 1B,F). DSE was significantly reduced following 24 (11 6 6%; Fig. 1C,F) and 48-h Tat treatment (6 6 6%; Fig. 1D,F; Table 1a). EPSC amplitude was not affected by 24-h Tat treatment (control = À147 6 20 pA, Tat = À182 6 31 pA; t (24) = 0.925, p = 0.36; Table 1b). To confirm that the effect of Tat required structurally intact protein, we treated cells with heat-inactivated Tat (hi-Tat; Fig. 1E) for 24 h and assessed DSE. Exposure to hi-Tat did not affect DSE, which was comparable to control (57 6 7%; Fig. 1E,F). We conducted a subsequent Research Article: New Research analysis to determine effect size and precision. Tat treatment for 24 h was estimated to decrease DSE relative to control by 42%. Using bootstrap resampling, we calculated the 95% CI for this effect as 27-56% (Table 1a).
Tat does not affect DSI Because CB 1 Rs are also present on inhibitory terminals (Katona et al., 1999), depolarization-induced 2-AG production also inhibits GABAergic transmission (Ohno-Shosaku et al., 2001;Wilson and Nicoll, 2001). To test whether Tat (50 ng/ml, 24 h) similarly affects DSI, we recorded evoked IPSCs, then depolarized the postsynaptic neuron to 0 mV. A 15-s depolarization produced 87 6 6% inhibition of IPSC amplitude, consistent with the high density of CB 1 Rs on a subset of hippocampal interneurons (Katona et al., 1999). In cultures treated with Tat, DSI was unaffected (90 6 5%; Fig. 2A,D). To confirm that we had not saturated DSI, we reduced the stimulus strength by shortening the duration of depolarization (Ohno-Shosaku et al., 2001;Wilson and Nicoll, 2001;Isokawa and Alger, 2005). Untreated cultures depolarized to 0 mV for 5 s exhibited 75 6 7% inhibition; Tat treatment did not alter DSI evoked by a 5-s depolarization (72 6 9%; Fig. 2B,D). A 2-s depolarization in untreated cells induced 57 6 8% DSI, and Tat did not alter DSI evoked by a 2-s depolarization (58 6 7%; Fig. 2C,D; Table 1c). IPSC amplitude was not affected by 24-h Tat treatment (control = À541 6 42 pA, Tat = À638 6 91 pA; t (25) = 0.97, p = 0.34; Table 1b). Since not all inhibitory synapses contain CB 1 Rs, not all cells exhibit DSI; we found 29% of control cells responded to depolarization by exhibiting DSI (18 out of 62 cells). All synaptic pairs that failed to exhibit DSI were also insensitive to 1 mM Win-2, a potent and efficacious CB 1 R agonist, confirming a lack of presynaptic CB 1 Rs. Tat did not alter the percentage of cells that exhibited DSI, as Tat-treated cultures yielded responses in 32% of the cells (10 out of 31 cells). Overall, Tat does not alter DSI, in contrast to its marked inhibition of DSE, indicating a specificity for eCB signaling at excitatory but not inhibitory synapses. EPSC DSE : average amplitude of two responses immediately after depolarization. One-way ANOVA (F (4,25) = 17.9, p , 0.0001; n = 6 cells/group) was followed with Tukey's post hoc test (pppp , 0.001, ppppp , 0.0001 relative to control; #### p , 0.0001 relative to hi-Tat). Data are represented as mean 6 SEM.

Tat inhibits MSE
We next determined whether sensitivity to Tat was dependent on the stimulus used to evoke 2-AG synthesis. MSE is evoked by activating postsynaptic G q -coupled receptors (Straiker and Mackie, 2007). Activation of group 1 mGluRs stimulates phospholipase Cb (PLCb ) to produce diacyglycerol (DAG), the rate-limiting precursor for 2-AG synthesis (Murataeva et al., 2014). Bath application of the selective group 1 mGluR agonist DHPG (1 mM) inhibited EPSCs by 53 6 2% (Fig. 3A). Tat treatment (50 ng/ml, 24 h) significantly reduced MSE to 28 6 3% (Fig. 3B,C). We conducted a subsequent analysis to determine effect size and precision. Tat treatment for 24 h was estimated to decrease MSE relative to control by 25%. Using bootstrap resampling, we calculated the 95% CI for this effect as 19% to 31% (Table 1d). Thus, Tat impairs two forms of short-term synaptic plasticity mediated by the eCB system, both DSE and MSE.

Tat does not alter diacylglycerol (DAG) synthesis
In a previous study, Tat was shown to inhibit PLC (Vitale et al., 2013), the enzyme that produces DAG; DAG is hydrolyzed to produce 2-AG, the eCB that mediates DSE  and MSE. To determine whether Tat impairs production of DAG, cells were transfected with an expression vector (pCMV-PKDC1ab-GFP) for a fluorescent DAG bioprobe. Since DAG is a precursor for 2-AG, we used DAG as a proxy to assess changes in 2-AG synthesis. DAG production was stimulated by application of the selective group 1 mGluR agonist DHPG (1 mM). Upon stimulation, the DAG biosensor translocates from the cytosol to the membrane, where it binds DAG (Fig. 4A,B). Tat (50 ng/ml, 24 h) did not significantly alter the translocation kinetics of the DAG biosensor, as indicated by no significant change in the peak membrane-tocytosol fluorescence intensity ratio (control = 3.9 6 0.23, Tat = 3.8 6 0.15) or decay time constant (t ; control = 32 6 0.6 s, Tat = 32 6 1.3 s; Fig. 4A,C; Table 1e). While this approach does not directly assess 2-AG production, these data do indicate that the production of DAG, the precursor for 2-AG synthesis, is not altered by Tat.

Tat impairs CB 1 R function
Because the effects of Tat on eCB signaling were independent of postsynaptic stimulus and Tat did not appear to affect 2-AG synthesis, we next examined presynaptic components of the eCB system. We assessed changes in presynaptic CB 1 R function by measuring EPSC amplitude in the presence of various CB 1 R agonists. EPSCs were evoked at 0.5 Hz by stimulating the presynaptic neuron with an extracellular electrode. Exogenous 2-AG was bath applied at varying concentrations and found to inhibit EPSCs in a concentration-dependent manner (Fig. 5A,C). Following 24 h treatment with 50 ng/ml Tat, the concentration-response relationship for 2-AG-mediated inhibition of EPSCs was right-shifted (Fig. 5B,C), and the EC 50 was significantly increased from 0.39 6 0.07 mM (control) to 1.6 6 0.24 mM, indicating reduced potency of 2-AG (Table 1f). The maximal inhibition produced by 2-AG trended lower in Tat-treated cells, as indicated by a shift in the asymptote of the sigmoidal curve fit from 87% in control to 70% after Tat treatment. We examined whether this Tat-induced shift in the 2-AG concentration-response relationship was the result of an increase in 2-AG metabolism by monoacylglycerol lipase (MGL) by treating the culture with the selective, irreversible MGL inhibitor JZL184 (1 mM; Long et al., 2009) for 24 h. We first confirmed that JZL184 effectively inhibited MGL. JZL184 treatment prolonged the duration of DSE, as indicated by a markedly slowed recovery from depolarizationinduced inhibition (n = 2). However, pretreating control or Tat-treated (50 ng/ml, 24 h) cultures with JZL184 did not affect 2-AG-mediated inhibition of EPSCs ( Fig. 5D; Table 1f). These data suggest that breakdown of bath-applied 2-AG by MGL was not enhanced by Tat.
Interestingly, Tat did not alter the concentration-response relationship for Win-2-mediated inhibition of EPSCs (Fig. 6). The EC 50 for Win-2 inhibition of EPSC amplitude was 7.8 6 1.6 nM for control and 8.0 6 1.1 nM after 24 h exposure to Tat (Table 1g). Win-2 is a highly efficacious agonist for cannabinoid receptors (Shen et al., 1996;Luk et al., 2004) and has been shown to elicit near maximal effect while occupying only 7.5% of CB 1 Rs (Gifford et al., 1999). Thus, the presence of spare receptors may explain the discrepancy. If spare receptors account for the insensitivity of Win-2-mediated inhibition of EPSCs to Tat treatment relative to 2-AG-mediated presynaptic inhibition, then we would predict that the inhibition produced by D 9 -tetrahydrocannabinol (D 9 -THC), the psychoactive ingredient in marijuana and a drug that acts as a partial agonist for CB 1 Rs (Shen and Thayer, 1999;Roloff and Thayer, 2009), to be especially sensitive to Tat exposure. Presynaptic inhibition mediated by G i -coupled GPCRs is attenuated at high stimulus frequencies (Bean, 1989), which is more pronounced for low-efficacy agonists such as D 9 -THC (Roloff and Thayer, 2009); thus, to  ). A, Representative images of untreated cells (control, gray) and cells treated with 50 ng/ml Tat for 24 h (red). Scale bars = 5 mm. B, Plot shows intensity values for regions of interest in the cytosol (gray) and membrane (black) monitored over time. C, Ratio of membrane-to-cytosol fluorescence intensity plotted over time. Exponential functions were fitted to plots of peak DHPG response using a nonlinear, least-squares curve-fitting algorithm. The curve was fit by a logistic equation of the form fluorescence intensity ratio = yₒ 1 Ae -(x-xₒ)/t . Tat did not significantly alter the translocation kinetics of the DAG biosensor, as indicated by no significant change in the peak response (control = 3.9 6 0.23, Tat = 3.8 6 0.15; t (10) = 0.15, p = 0.88, t test; n = 6 cells/ group) or decay time constant (t ; t (10) = 0.31, p = 0.76; t test). Data are represented as mean ratio 6 SEM. study the effect of D 9 -THC we reduced the stimulus frequency to 0.1 Hz. Following Tat treatment, presynaptic inhibition produced by D 9 -THC was markedly reduced ( Fig.  7; Table 1h). The inhibition of EPSC amplitude at a maximally effective concentration of D 9 -THC was reduced from 55% in control to 15% after Tat treatment. Overall, Tat impairs CB 1 R function and this loss-of-function is most evident by the attenuation of the responses to lowefficacy agonists such as D 9 -THC.
Several reports have shown that Tat produces synaptodendritic injury and cell death via a pathway initiated by Tat binding to low-density lipoprotein receptor-related protein (LRP; Eugenin et al., 2007;Kim et al., 2008;Shin and Thayer, 2013). Thus, we examined whether binding to LRP is necessary for Tat-mediated impairment of eCB signaling. We found that blocking LRP by pretreating with the selective antagonist receptor-associated protein (RAP) did not prevent Tat-mediated reduction of DSE. In untreated cells DSE was 51 6 4% and was similar in the presence of 50 nM RAP (55 6 6%). Tat treatment (50 ng/ml, 24 h) reduced DSE to 14 6 12% and was not affected by RAP (12 6 7%). A twoway ANOVA found no interaction (RAP treatment Â Tat treatment; F (1,16) = 0.156, p = 0.698, n = 4-6 cells/group) while Tat treatment had a significant effect on DSE (F (1,16) = 30.94, p , 0.0001; Table 1i), indicating that Tat acts via a mechanism independent of LRP.

Tat does not affect adenosine A 1 receptor (A 1 AR)mediated presynaptic inhibition
To determine whether Tat acts selectively on CB 1 R, or whether it similarly affects other presynaptic G i/o -coupled  -2], A 1 = % inhibition in the absence of drug, A 2 = % inhibition at a maximally effective drug concentration, and p = slope factor. The following values were calculated using a nonlinear, least-squares curve fitting program: A 1 = 21.7 for control, 19.9 for Tat; A 2 = 81.1 for control, 77.0 for Tat; EC 50 = 7.8 6 1.6 nM for control, 8.0 6 1.1 nM for Tat; p = 2.5 6 1.7 for control, 2.5 6 1.2 for Tat. Tat treatment did not significantly affect the EC 50 for Win-2-mediated inhibition of EPSCs (t test; t (39) = 0.08, p = 0.93). where xₒ = EC 50 , Â = log[2-AG], A 1 = % inhibition in the absence of drug, A 2 = % inhibition at a maximally effective drug concentration, and p = slope factor. The following values were calculated using a nonlinear, least-squares curve fitting program: A 1 = 13.1 for control, 10.4 for Tat; A 2 = 87.1 for control, 70.3 for Tat; EC 50 = 0.396 0.07 mM for control, 1.6 6 0.2 mM for Tat; p = 1.6 6 0.04 for control, 1.1 6 0.1 for Tat. The EC 50 was significantly increased relative to control by treatment with Tat (t test; t (37) = 4.91, p , 0.0001), indicating reduced 2-AG potency. D, Bar graph shows the mean magnitude of 0.2 mM 2-AG-mediated inhibition of EPSCs in control (open bars) and Tat-treated groups (gray bars) in cultures treated with vehicle or 1 mM JZL184 for 24 h. A two-way ANOVA found no interaction (JZL184 treatment Â Tat treatment; F (1,20) = 0.043, p = 0.838, n = 6 cells/group). Tat treatment had a significant effect on 2-AG-induced inhibition of EPSC amplitude (F (1,20) = 59.01, p , 0.0001). Data are represented as mean 6 SEM. receptors, we examined the effects of Tat on A 1 ARs. These receptors are localized on presynaptic terminals and modulate the release of neurotransmitters, including glutamate and GABA (Dunwiddie and Masino, 2001). We assessed whether Tat impairs A 1 AR function by monitoring EPSC amplitude in the presence of a selective A 1 AR agonist N 6 -cyclopentyladenosine (CPA; 10 nM; Hargus et al., 2009;Gonçalves et al., 2015) and partial agonist capadenoson (30 nM; Albrecht-Kupper et al., 2012). CPA inhibited EPSC amplitude by 74 6 5% in control cultures and by 77 6 5% in cultures treated with 50 ng/ml Tat for 24 h (Fig. 8B). The partial agonist capadenoson inhibited EPSC amplitude by 43 6 8% in control cultures and by 41 6 6% in Tat-treated cultures ( Fig. 8C; Table 1j). Thus, Tat treatment does not affect inhibition of glutamate release initiated by selective A 1 AR agonists, indicating that A 1 AR function is not altered by Tat. These results suggest that Tat acts selectively to inhibit CB 1 R function and provides evidence that Tat exposure does not broadly suppress the function of all presynaptic G i/o -coupled receptors.

Discussion
Treating rat hippocampal cultures with the HIV protein Tat inhibited DSE and MSE, two forms of short-term plasticity mediated by the eCB system. While Tat is a neurotoxin known to contribute to synaptodendritic damage in HAND , this is the first report to show that Tat disrupts function of the eCB system. Our principal finding is that Tat impairs CB 1 R-mediated presynaptic inhibition at excitatory but not inhibitory terminals. This selective attenuation of eCB signaling may unbalance network excitability, with potentially significant effects on symptoms associated with HAND, the progression of neurological disease in people living with HIV, and the sensitivity of HIV-positive individuals to exogenous cannabinoids. Furthermore, the selective loss of eCB signaling at excitatory synapses has implications for the effects of neuroinflammatory stimuli on the eCB system.  Tat-treated groups (50 ng/ml, 24 h). Scale bars: 10 ms (horizontal bar) and 50 pA (vertical bar). C, The concentration-response curve for D 9 -THC-mediated inhibition of EPSCs shows a decrease in potency and efficacy following exposure to Tat. The curves were fit with a logistic equation of the form percentage inhibition = A 1 1 [(A 2 -A 1 )/(1 1 10 (logxₒ-x)p )], where xₒ = EC 50 , Â = log[D 9 -THC], A 1 = % inhibition in the absence of drug, A 2 = % inhibition at a maximally effective drug concentration, and p = slope factor. The following values were calculated using a nonlinear, least-squares curve fitting program: A 1 = À4.4 for control, À0.8 for Tat; A 2 = 54.6 for control, 15.3 for Tat; EC 50 = 0.04 6 0.03 mM for control, 0.27 mM for Tat; p = 1.3 6 0.8 for control, 3.4 for Tat. The marked reduction in D 9 -THC-mediated inhibition following Tat exposure limited the quality of the curve fit precluding determination of error associated with the EC 50 . To determine the significance of the Tat effect, we performed a two-way ANOVA over the 0.1-1 mM THC concentrations that were tested in both control and Tat-treated cultures and found no interaction ([D 9 -THC] Â Tat treatment; F (2,19) = 0.088, p = 0.92, n = 4-5 cells/group). Tat treatment had a significant effect on D 9 -THC-induced inhibition of EPSC amplitude (F (1,19) = 53.71, p , 0.0001). Data are represented as mean 6 SEM. Primary neuronal cultures provide an experimentally accessible system that exhibits synaptic changes that are a hallmark of HAND (Ellis et al., 2007;Eugenin et al., 2007;Kim et al., 2008Kim et al., , 2011aShin and Thayer, 2013;Raybuck et al., 2017;Hermes et al., 2018). Advantages of primary hippocampal cultures for this study include robust DSE and DSI, the ability to administer HIV Tat at specific concentrations with accurate timing, and the ease with which the culture can be treated with highly lipophilic cannabinoid receptor ligands. However, there are caveats to cell culture models. The rat primary cultures used in this study were derived from fetal neurons, and while they mature in culture, we cannot precisely match their stage of development to in vivo models. Furthermore, dissociated cultures lack the synaptic architecture found in vivo, precluding our ability to identify the specific synapses affected by HIV Tat in our model. Finally, the effects of chronic exposure to Tat produced in transgenic models cannot be replicated with short-lived primary cultures. Thus, the novel changes in the eCB system described here provide compelling evidence to extend these findings to chronic in vivo models to explore regional and developmental effects on synaptic networks.
Tat-induced attenuation of CB 1 R-mediated inhibition of glutamatergic, but not GABAergic neurotransmission, might result from several mechanisms. CB 1 Rs at inhibitory synapses are expressed at a much higher density relative to the density at excitatory terminals (Katona et al., 1999). Thus, Tat-induced loss of CB 1 R-mediated signaling at GABAergic terminals may not be sufficient to significantly impair eCB-mediated inhibition of GABA release. The importance of spare receptors in modulating eCB signaling is illustrated by comparing the failure of Tat to affect inhibition of glutamate release by the highly efficacious agonist Win-2, which elicits maximal effects while occupying only 7.5% of CB 1 Rs (Gifford et al., 1999), to the marked attenuation of the effects of the partial agonist D 9 -THC. Even when the duration of the depolarizing stimulus to evoke 2-AG production was reduced, Tat failed to affect DSI. Thus, following exposure to Tat, physiologically relevant eCB-mediated retrograde signaling remains functional at inhibitory terminals, but is impaired at excitatory synapses.
The rightward shift in the 2-AG concentration-response relationship produced by Tat treatment is reminiscent of CB 1 R desensitization (Kouznetsova et al., 2002;Sim-Selley, 2003;Lundberg et al., 2005;Daigle et al., 2008;Wu et al., 2008). Tat is a powerful excitotoxin; perhaps Tat-induced excitatory drive chronically activates the eCB system, desensitizing CB 1 Rs. The development of Tat-induced attenuation of CB 1 R function over the span of 4-48 h is consistent with gradual receptor downregulation. Alternatively, Tat can be taken up by glutamatergic neurons where it activates protein kinase C (Haughey et al., 2001), which is known to produce heterologous desensitization of CB 1 Rs (Garcia et al., 1998;Chu et al., 2010). The time course of the Tat effect also suggests that Tat does not simply bind to and occlude CB 1 Rs; direct receptor interaction would be expected to equilibrate more rapidly and also affect receptors on GABAergic terminals similarly to those on glutamatergic terminals. Tat can be internalized into neurons by binding to LRP, leading to synaptodendritic injury and cell death Eugenin et al., 2007;Kim et al., 2008;Shin and Thayer, 2013). We found that Tat-mediated impairment of eCB signaling is independent of LRP. Tat elicits a range of toxic effects on neuronal networks, some of which are mediated by neuroinflammation (Lu et al., 2011;Chivero et al., 2017;Gonek et al., 2018;Thangaraj et al., 2018), leading us to speculate that Tat may impair eCB signaling through an indirect mechanism involving non-neuronal cells, possibly via the release of inflammatory cytokines (Kim et al., 2018). The degree to which this damage results from direct or adaptive responses to the presence of Tat is unclear.
In contrast to the loss of CB 1 R function described here, CB 1 R protein increased in infralimbic cortex from transgenic mice expressing Tat (Jacobs et al., 2019). The mPFC and hippocampus display differential sensitivity to Tat (Cirino et al., 2020), suggesting the effects of Tat on the eCB system may also exhibit regional differences. Xu et al. (2016) found that Tat expression occluded the effects of CB 1 R agonists. We found DSE to be an extremely sensitive assay for Tat-induced modulation of the eCB system. Thus, the failure of Tat to occlude DSI provides strong evidence that Tat does not affect CB 1 R-mediated presynaptic inhibition at GABAergic synapses in the hippocampal culture model. Tat failed to affect DSI evoked by three different stimulus strengths; this physiological method of graded CB 1 R activation essentially produces a 2-AG concentration-response. It is possible that Tat upregulates CB 1 Rs on GABAergic terminals while also decreasing CB 1 R levels on glutamatergic terminals. Tat elicits a neuroinflammatory response (Nath et al., 1999;Pu et al., 2003), and neuroinflammation selectively upregulates CB 1 Rs on GABAergic terminals (Chen et al., 2003;Feng et al., 2016), providing a precedent for this type of change in the eCB system. Alternatively, the loss of CB 1 R function at excitatory synapses described here may be an early event initiated by Tat, akin to the initial stages of exposure, and the gain-of-function at inhibitory terminals  may result from adaptations in the eCB system that occur during prolonged Tat expression. Thus, unique changes in the synaptic network may predominate at different disease stages.
Altered eCB signaling could affect many physiological processes relevant to HAND (Rodríguez de Fonseca et al., 2005). Using brain tissue from Alzheimer's patients, autoradiography, in situ hybridization, and GTPg S binding studies have demonstrated reduced CB 1 R agonist binding, regionally discrete losses of CB 1 R mRNA expression, and less efficient coupling to G-proteins (Westlake et al., 1994;Ramírez et al., 2005). The selective loss of functional eCB signaling at excitatory terminals described here emphasizes the need to understand synapse-specific changes in the eCB system over the course of neurodegenerative disease. Antagonism of CB 1 R function elicits depression and anxiety-like behaviors (Moreira et al., 2009). If the loss of CB 1 R function on excitatory terminals we describe for hippocampal cultures also occurs in the prefrontal cortex (Rigucci et al., 2010), it might contribute to the production of these symptoms in HAND. The motor impairment produced by cannabinoids results from CB 1 R activation on both excitatory and inhibitory terminals in basal ganglia and cerebellum (Sañudo-Peña et al., 1999;Kishimoto and Kano, 2006), suggesting Tat-mediated motor effects could result from unbalanced eCB signaling (Kim et al., 2003). CB 1 Rs on glutamatergic terminals in the nucleus accumbens regulate the hedonic impact of food, while CB 1 Rs in the hypothalamus regulate the release of hormones that regulate appetite and energy expenditure (Lau et al., 2017). Thus, their loss could contribute to the loss of appetite and wasting observed in AIDS patients.
While evidence linking impaired eCB signaling to disease progression is limited, there is considerable evidence demonstrating that enhanced CB 1 R signaling protects neurological function (Aymerich et al., 2018). In HAND models, cannabinoids protect dopaminergic neurons against damage from the HIV envelope glycoprotein gp120 (Hu et al., 2013), reduce Tat-induced release of nitric oxide (Esposito et al., 2002), and diminish Tat-induced increases in neuronal [Ca 21 ] i (Xu et al., 2017). Boosting eCB tone by inhibiting eCB metabolism also attenuates HIV neurotoxicity; inhibition of MGL with the selective inhibitor JZL184 protects against excitatory synapse loss evoked by gp120 (Zhang and Thayer, 2018). Similarly, pharmacological inhibition of fatty acid amide hydrolase, the enzyme responsible for degrading the eCB anandamide, protects against Tat-mediated increases in [Ca 21 ] i and dendritic damage (Hermes et al., 2018). Thus, cannabimimetics and drugs that inhibit eCB metabolism may slow the course of HIV neurotoxicity .
This study suggests that the converse may also be true; in the presence of HIV, impaired CB 1 R function at excitatory terminals may alter the response to exogenous cannabinoids. Because efforts to legalize medicinal and recreational marijuana are increasing access (Charilaou et al., 2017;Jones et al., 2018) and cannabis use is prevalent among people living with HIV (Hartzler et al., 2017), altered response to cannabinoids could impact many patients. We found a dramatic loss of D 9 -THC efficacy at excitatory synapses in the presence of Tat. However, CB 1 R signaling at inhibitory synapses was normal in the presence of Tat. Thus, the network effects of D 9 -THC could shift to an excitatory response due to CB 1 R-mediated suppression of GABA release without a corresponding decrease in glutamate release. This prediction is consistent with the increased miniature EPSC frequency observed in mice expressing Tat (Jacobs et al., 2019).
In conclusion, we have shown in an in vitro model of HIV neurotoxicity that the HIV protein Tat, an established contributor to HIV neurotoxicity, inhibits short-term eCBmediated plasticity selectively at excitatory synapses. We speculate that this novel loss of CB 1 R function might contribute to excitotoxicity under neuroinflammatory conditions. Thus, our conclusions from this in vitro work provide a framework for future work assessing the status of the eCB system in vivo in neurodegenerative disease noting the duration of disease, the specific brain regions affected and, as shown here, the specific neurotransmitters involved.