Pregnane steroidogenesis is altered by HIV-1 Tat and morphine: Physiological allopregnanolone is protective against neurotoxic and psychomotor effects

Pregnane steroids, particularly allopregnanolone (AlloP), are neuroprotective in response to central insult. While unexplored in vivo, AlloP may confer protection against the neurological dysfunction associated with human immunodeficiency virus type 1 (HIV-1). The HIV-1 regulatory protein, trans-activator of transcription (Tat), is neurotoxic and its expression in mice increases anxiety-like behavior; an effect that can be ameliorated by progesterone, but not when 5α-reduction is blocked. Given that Tat's neurotoxic effects involve mitochondrial dysfunction and can be worsened with opioid exposure, we hypothesized that Tat and/or combined morphine would perturb steroidogenesis in mice, promoting neuronal death, and that exogenous AlloP would rescue these effects. Like other models of neural injury, conditionally inducing HIV-1 Tat in transgenic mice significantly increased the central synthesis of pregnenolone and progesterone's 5α-reduced metabolites, including AlloP, while decreasing central deoxycorticosterone (independent of changes in plasma). Morphine significantly increased brain and plasma concentrations of several steroids (including progesterone, deoxycorticosterone, corticosterone, and their metabolites) likely via activation of the hypothalamic-pituitary-adrenal stress axis. Tat, but not morphine, caused glucocorticoid resistance in primary splenocytes. In neurons, Tat depolarized mitochondrial membrane potential and increased cell death. Physiological concentrations of AlloP (0.1, 1, or 10 nM) reversed these effects. High-concentration AlloP (100 nM) was neurotoxic in combination with morphine. Tat induction in transgenic mice potentiated the psychomotor effects of acute morphine, while exogenous AlloP (1.0 mg/kg, but not 0.5 mg/kg) was ameliorative. Data demonstrate that steroidogenesis is altered by HIV-1 Tat or morphine and that physiological AlloP attenuates resulting neurotoxic and psychomotor effects.

Despite treatment with combined antiretroviral therapeutics, a considerable proportion of patients infected with HIV-1 present with endocrine dysfunction. Approximately 25% of HIV + patients experience dysfunction of the hypothalamic-pituitary-gonadal (HPG; e.g., hypogonadism) and/or hypothalamic-pituitary-adrenal (HPA) axes (e.g., elevated basal corticosterone with adrenal insufficiency in response to HPA activation; Chrousos and Zapanti, 2014;George and Bhangoo, 2013;Gomes et al., 2017;Lachâtre et al., 2017;Mirza et al., 2018;Wong et al., 2017). In the era of combined antiretroviral therapy, HPG and HPA disruption typically involve dysfunction within the CNS, rather than dysfunction at peripheral steroid sources such as the gonads or adrenals (Bons et al., 2013;Chrousos and Zapanti, 2014;Freda and Bilezikian, 1999;Rochira and Guaraldi, 2014;Mirza et al., 2018). In support, others have found evidence of neurosteroidogenic dysregulation in post-mortem HIV + brain tissue, cultured human fetal neurons, and the brains of cats infected with feline immunodeficiency virus (Maingat et al., 2013). Maintaining HPG/HPA function may improve neurological outcomes.
Restoration of physiological steroid content may ameliorate neurological deficits caused by HIV-1 Tat and/or combined opioid exposure. Estradiol or pregnane steroids [progesterone (PROG) or the 3αhydroxy/5α-reduced PROG metabolite, allopregnanolone (AlloP)] rescue Tat-mediated neurotoxicity in vitro (Adams et al., , 2012Kendall et al., 2005;Paris et al., 2016;Salahuddin et al., 2019;Wallace et al., 2006). Moreover, transgenic expression of HIV-1 Tat in male or female mice increases anxiety-and depression-like behavior in vivo (McLaughlin et al., 2017;Paris et al., 2014d;Salahuddin et al., 2019). In females, these effects coincide with Tat-mediated elevations in circulating corticosterone and are exacerbated by exposure to the clinically-used opioid, oxycodone (Salahuddin et al., 2019). Exogenous PROG can ameliorate Tat-mediated anxiety-like effects in mice (Paris et al., 2014c(Paris et al., , 2016) and this appears to be dependent on metabolism to AlloP (Paris et al., 2016). Thus, the PROG metabolite, AlloP, may confer protection against the CNS-targeted effects of HIV-1 Tat, but additional evidence is needed to confirm this and to assess its effects on Tat and opioid interactions.
Given that mitochondrial function is both a target for HIV-1 Tat and a rate-limiting step for neurosteroidogenesis to occur, we hypothesized that conditional expression of an HIV-1 tat transgene in mice would alter central and/or circulating steroid content and that morphine would further influence these effects. We expected this to co-occur with neuronal mitochondrial dysfunction and neuron cell death. Lastly, we anticipated the pregnane steroid, AlloP (previously found to exert neuroprotection in cell culture), would ameliorate Tat-mediated mitochondrial dysfunction, cell death, and the potentiation of morphine's behavioral effects.

Materials and methods
The use of mice in these studies was pre-approved by the Institutional Animal Care and Use Committees at Virginia Commonwealth University and the University of Mississippi. All experiments were conducted in accordance with ethical guidelines defined by the National Institutes of Health .

Animals and housing
Steroid analysis and behavioral studies utilized adult male mice that expressed (or did not express) a GFAP-driven, doxycycline-inducible, HIV-1 IIIB tat 1-86 transgene (N = 355) as previously described . Tat(−) and Tat(+) mice were generated in the vivaria at Virginia Commonwealth University (MCV campus; for steroid analyses and glucocorticoid insensitivity assay) and at the University of Mississippi (for behavioral analyses). Primary cell culture analyses utilized offspring acquired from timed-mated, C57BL/6J females (N = 10; obtained from the Jackson Laboratory, Bar Harbor, ME). All mice were~70 days of age at the time of testing, were housed 4-5/cage (Tat-transgenic mice) or were singly-housed (timed-mated C57BL/6J mice), and were maintained in a temperature-and humidity-controlled room on a 12:12 h light/dark cycle with ad libitum access to food and water.

Chemicals used in vivo
Mice were treated with vehicle (saline 0.9%) or morphine (30 mg/ kg, i.p.; #M8777, Sigma-Aldrich) as well as vehicle or AlloP (0.5 or 1 mg/kg, s.c., dissolved in EtOH and diluted 1:10 in oil; #P8887, Sigma-Aldrich). To induce HIV-1 tat 1-86 transgene expression (or not), Tat(+) and Tat(−) control mice were placed on doxycycline chow (6 g/kg; Dox Diet #2018, Harlan Laboratories, Madison, WI) for 4 weeks in order to assess changes in steroidogenesis. To assess the influence of Tat and morphine on glucocorticoid insensitivity, Tat transgene expression was induced for 8 weeks by doxycycline chow, and mice were administered saline or ramping doses of morphine (10-40 mg/kg, s.c. BID; increasing by 10 mg/kg every 2 days). To assess the combined influence of Tat and AlloP on morphine-mediated psychomotor behavior, transgene expression was induced via injection of doxycycline (30 mg/kg, i.p.; #14422, Cayman Chemical, Ann Arbor, MI) for 5 days with an additional two days of doxycycline washout prior to behavioral testing (to minimize any potential non-specific behavioral effects of doxycycline). Systemic AlloP (0.5 or 1 mg/kg, s.c.) was concurrently-administered for the full duration of the behavioral testing paradigm (8 days in total). AlloP dosing fell within a range that is expected to produce physiological concentrations in the brain (Gabriel et al., 2004) and has previously demonstrated stepwise, dose-dependent changes on affective behavior (Reddy and Kulkarni, 1996). The chosen morphine concentration is found to enhance AlloP in the rodent brain and circulation when administered acutely (Concas et al., 2006;Porcu et al., 2014).

Primary Tat-transgenic mouse splenocytes
Primary splenocyte cultures were prepared as previously described (Avitsur et al., 2001;Kinsey et al., 2007) and were used to assess the ability of prolonged HIV-1 Tat and morphine to alter glucocorticoidinduced inhibition of peripheral immune cell viability. Briefly, mice were sacrificed 24 h after administration of the final dose of morphine (or vehicle) and spleens were harvested. Tissues were passed through a 70 μm pore nylon mesh (Greiner Bio-One) with 10 mL of RPMI media (#22400089, Life Technologies) to prepare a cell suspension. Cells were lysed with a hypotonic salt solution (0.16 M NH 4 Cl, 10 mM KHCO 3 , 0.13 mM EDTA) and suspended in RPMI media (#22400089, Life Technologies), supplemented with 10% heat-inactivated fetal bovine serum (FBS; #SH30071.03, Thermo Scientific), 7.5% NaHCO 3 (#25080094, Life Technologies), and antibiotic/antimycotic mixture (Life Technologies), centrifuged, and filtered through a 70 μm pore nylon mesh (Greiner Bio-One). Duplicate splenocyte samples were seeded onto 96-well plates with corticosterone and stimulated with LPS for 48 h.

GC/MS/MS
A panel of steroids was identified and quantified simultaneously in individual tissues by GC/MS/MS as previously described (Liere et al., 2000;Zhu et al., 2017). Briefly, steroids were first extracted from hemisected whole brains and plasma with 10 vol of methanol (MeOH) and the following internal standards were added to the extracts for steroid quantification: 2 ng of 2 H 6 -5α-dihydroprogesterone (DHP; CDN Isotopes, Sainte Foy la Grande, France) for 5α/β-DHP, 2 ng of epietiocholanolone for pregnenolone (PREG), 20α-dihydropregnenolone (20α-DHPREG), 3α/β5α/β-tetrahydroprogesterone (THP), 5α20α-THP, 3α/β5α/β20α-hexahydroprogesterone (HHP) and 3α/β5α/β-tetrahydrodeoxycorticosterone (THDOC), 2 ng of 13 C 3 -progesterone (P) for P, 2 ng of 19-norPROG for 20α-DHP, 16α-hydroxyprogesterone (16α-OHP), 17α-hydroxyprogesterone (17αOHP), 5α/β-DHDOC (dihydrodeoxycorticosterone), 5 ng of 13 C 3 -deoxycorticosterone (DOC) for DOC, 10 ng of 2 H 8 -corticosterone for corticosterone, 5α-dihydrocorticosterone (5α-DHB) and 3α/β5α/β-tetrahydrocorticosterone (THB). Samples were purified and fractionated by solid-phase extraction with the recycling procedure (Liere et al., 2004). Briefly, the extracts were dissolved in 1 ml MeOH and applied to the C18 cartridge (500 mg, 6 ml, International Sorbent Technology, IST), followed by 5 ml of MeOH/H 2 O (85/15, vol/vol). The flow-through, containing the free steroids, was collected and dried. After a previous re-conditioning of the same cartridge with 5 ml H 2 O, the dried samples were dissolved in MeOH/H 2 O (2/8, vol/vol) and re-applied. The cartridge was then washed with 5 ml H 2 O and 5 ml MeOH/H 2 O (1/1, vol/vol) and unconjugated steroids were eluted with 5 ml MeOH/H 2 O (9/1, vol/vol). The fraction containing the unconjugated steroids was then filtered and further purified and fractionated by high-performance liquid chromatography (HPLC). The HPLC system is composed of a WPS-3000SL analytical autosampler and a LPG-3400SD quaternary pump gradient coupled with a SR-3000 fraction collector (Thermo Scientific). The HPLC separation was achieved with a Lichrosorb Diol column (25 cm, 4.6 mm, 5 μm) in a thermostated block at 30°C. The column was equilibrated in a solvent system of 90% heptane and 10% of a mixture composed of heptane/isopropanol (85/15, vol/vol). Elution was performed at a flow-rate of 1 ml/min, first 90% heptane and 10% of heptane/isopropanol (85/15, vol/vol) for 15 min, then with a linear gradient to 100% acetone in 2 min. The column was washed with acetone for 15 min. Two fractions were collected from the HPLC system: 5α/β-DHP were eluted in the first HPLC fraction (3-15 min) and was next silylated with 50 μl of a mixture N-methyl-N-trimethylsilyltrifluoroacetamide/ammonium iodide/dithioerythritol (1000:2:5 vol/wt/wt) for 15 min at 70°C. The second fraction (15-29 min) containing PREG and its derivatives, P, DOC, corticosterone and their precursors and their reduced and hydroxylated metabolites were derivatized with 25 μl of HFBA and 25 μl of anhydrous acetone for 1 h at 20°C. Both fractions were dried under a stream of nitrogen and resuspended in heptane. GC/MS/MS analysis of the biological extracts was performed using an AI 1310 autosampler, a Trace 1310 gas chromatograph (GC), and a TSQ 8000 mass spectrometer (MS) (Thermo Scientific). Injection was performed in the splitless mode at 250°C (1 min of splitless time) and the temperature of the gas chromatograph oven was initially maintained at 80°C for 1 min and ramped between 80 and 200°C at 20°C/min, then ramped to 300°C at 5°C/min and finally ramped to 350°C at 30°C/min. The helium carrier gas flow was maintained constant at 1 ml/min during the analysis. The transfer line and ionization chamber temperatures were 300°C and 200°C, respectively. Electron impact ionization was used for mass spectrometry with ionization energy of 70 eV and GC/MS/MS analyses were performed in multiple reaction monitoring (MRM) mode with Argon as the collision gas. GC/MS/MS signals were evaluated using a computer workstation by means of the software Excalibur®, release 3.0 (Thermo Scientific). Identification of steroids was supported by their retention time and two or three transitions and quantification was performed according to the more abundant transition with a previously established calibration curve. The range of the limit of detection was roughly 0.5-20 pg according to the steroid structure. The GC/MS/MS analytical procedure was fully validated in terms of accuracy, reproducibility and linearity in mouse brains (Liere et al., 2000;Zhu et al., 2017). Concentrations of 5β-DHP, 5β-DHDOC, 3α,5β-THDOC, and 3α,5β-THB were below the limit of quantification in all samples.

Splenocyte proliferation assay
Glucocorticoid insensitivity was assessed using a CellTiter 96®AQ ueous Non-Radioactive Cell Proliferation Assay kit (#G5421, Promega, Madison, WI) per manufacturer instructions. Briefly, the proliferation assay was performed 48 h after primary Tat(−) and Tat (+) splenocytes were seeded on 96-well plates in duplicate and incubated with LPS and corticosterone. Media was replaced with the tetrazolium (MTS) and phenazine methosulfate (PMS) solution (1:20) and incubated (37°C; 5% CO 2 ) for 4 h and read at 490 nm using a PHERAstar FS microplate reader (BMG Labtech, Cary, NC). Viability of each sample was expressed as the mean optical density (OD) of duplicates. Corticosterone resistance was assessed by calculating the percentage of viable cells in each sample treated with a given dose of corticosterone compared with the same sample treated with vehicle [(OD of corticosterone treated sample/OD of non-treated sample) * 100].

Mitochondrial membrane potential (Δψ m )
Mitochondrial membrane potential (Δψ m ) was assessed using a JC-10 kit per manufacturer recommendations (#22800, AAT Bioquest, Sunnyvale, CA). Briefly, JC-10 (a JC-1 analogue with improved H 2 O solubility) reversibly forms aggregates when Δψ m is maintained; when depolarized, JC-10 falls into its monomeric state. Excited JC-10 aggregates and monomers display distinct emission spectra allowing for a ratiometric assessment of Δψ m (ex/em aggregate : 490/590 nm; ex/ em monomer : 490/525 nm). Experimental treatments were applied to differentiated SH-SY5Y cells seeded on 96-well plates or primary, striatal, medium spiny neurons seeded on 24 well plates and incubated (37°C; 5% CO 2 ) for 30 min, followed by the addition of JC-10 and another incubation (37°C; 5% CO 2 ) of 30 min in the dark. Fluorescence was assessed in SH-SY5Y cells using a PHERAstar FS microplate reader (BMG Labtech) at 5, 15, and 30 min later. Primary neurons were imaged 15 min later on a Zeiss Axio Observer Z1 microscope with Zeiss Axiovision 4.8 software (Carl Zeiss). For both methods, each plate represented one observation (n) in statistical analyses. 96-well plates had four technical replicates per treatment group and 24-well plates had two technical replicates. For imaging, fields within each well were randomly chosen and at least 200 cells were counted per condition. Δψ m was calculated as the monomer:aggregate ratio (presented as percent change from control values).

Live/dead assay
Neuron viability was assessed using a LIVE/DEAD® Viability/ Cytotoxicity Kit (#L-7013, Molecular Probes, Eugene, OR) per manufacturer instructions and as described (Salahuddin et al., 2019). Briefly, experimental treatments were applied to primary medium spiny neurons or SH-SY5Y cells seeded on 24-well plates and live/dead assay was performed 20 h later. Prior work utilizing time-lapse microscopy (0-60 h) identified the 20 h time-point as the earliest time when Tatand AlloP-treated cells significantly diverged on the measure of viability (Paris et al., 2016). A working solution of DEAD Red (ethidium homodimer-2; ex/em: 535/624 nm) and SYTO 10 (ex/em: 495/ 520 nm) was prepared by diluting stocks in Hank's Balanced Salt Solution (HBSS; 1:500 dilution). Cell media was replaced with the DEAD Red/SYTO 10 working solution and incubated (37°C; 5% CO 2 ) for 15 min in the dark. Cells were washed 3 times with HBSS, fixed with 4% paraformaldehyde, and imaged as described for 24-well plates above. Viability was assessed by calculating the proportion of necrotic cells [# DEAD Red + cells/# Total Cells) * 100].

Locomotor activity
Mice were acclimated to the testing room with 70 dB white noise for 30 min prior to drug administration. Mice were administered saline (0.9%, i.p.) or morphine (30 mg/kg, i.p.) and, 30 min later, were allowed to freely explore an open field apparatus (40 cm × 40 cm × 40 cm; Stoelting Co., Wood Dale, IL) for 5 min. Behavior was tracked and digitally-encoded via ANY-maze animal tracking software (Stoelting Co.). Horizontal and vertical (i.e., rearing) locomotor behaviors were assessed as indices of psychomotor activity (Paris et al., 2016;Salahuddin et al., 2019).

Statistical analyses
All datasets were determined to be normally-distributed with equal variance per Kolmogorov-Smirnov and subsequent Bartlett tests, respectively. Steroid content for each analyte was assessed via separate Student's t-tests ( Fig. 1) or two-way ANOVAs ( Fig. 2) with drug manipulation (saline or morphine) and genotype [(Tat(−) or Tat(+)] as between-subjects factors. Mitochondrial membrane potential across time in SH-SY5Y cells was assessed via repeated measures ANOVA with treatment condition as the between-subjects factors and time as the within-subjects factor. Glucocorticoid insensitivity assays were analyzed via three-way ANOVAs with drug condition (vehicle or morphine), Tat condition (vehicle or Tat), and corticosterone concentration as factors. Remaining analyses for Δψ m , live/dead analyses, and behavioral dependent measures were assessed via three-way ANOVAs with drug condition (vehicle-or morphine-exposed), AlloP condition (vehicle-or AlloP-exposed), and Tat condition (control or Tat-exposed) as factors. Fisher's Protected Least Significant Difference post-hoc tests determined group differences following main effects. Interactions were delineated via simple main effects and main effect contrasts with alpha controlled for multiple comparisons. For post-hoc contrasts, all possible pairwise comparisons were made with the exception of repeated measures analyses conducted in Fig. 4B, wherein a priori planned contrasts where performed between all groups and control-(hatched line) or Tat-(red open circle) treatment groups following a significant interaction. All analyses were considered significant when p < 0.05.

Exposure to HIV-1 Tat, but not morphine, induces glucocorticoid resistance
Alterations in central and circulating steroid expression prompted further examination of the effects of HIV-1 Tat and morphine on peripheral glucocorticoid activity. Splenocytes were harvested from Tat (−) and Tat(+) mice after 2 weeks of ramping morphine or saline exposure and incubated with LPS and corticosterone for 48 h. Cell viability was used to assess the ability of corticosterone to suppress mitogen-stimulated splenocyte proliferation and survival. An inhibition of the ability of corticosterone to reduce splenocyte viability indicates glucocorticoid resistance. HIV-1 Tat significantly interacted with corticosterone concentration such that a greater proportion of Tat(+) splenocytes were resistant to 5 μM corticosterone-induced suppression of proliferation compared to Tat(−) splenocytes [F(4,25) = 10.24, p < 0.0001] (Fig. 3), indicating a resistance to the inhibitory effects of corticosterone. Tat did not significantly interact with morphine to influence glucocorticoid resistance.

Physiological, but not high, AlloP concentrations rescue Tat/morphinedysregulated Δψ m
Given that alterations in central and circulating steroid synthesis were observed, we next assessed the influence of HIV-1 Tat and morphine on mitochondria which are known targets for Tat-mediated dysfunction and a rate-limiting source of steroidogenesis. We assessed the influence of Tat (100 nM) and/or morphine (500 nM) on mitochondrial membrane potential (Δψ m ) via JC-10 using differentiated human neuroblastoma cells (SH-SY5Y) and primary, striatal, medium spiny neurons obtained from C57BL/6J mice. Central elevations of 5α-reduced progestogens, such as those seen here in response to Tat or morphine, are observed in models of brain injury and are proposed to underlie an endogenous mechanism of neural protection (Schumacher   Zhu et al., 2017). As such, we further assessed the potentially protective role that 5α-reduced pregnane steroids may play on Δψ m following a Tat/morphine challenge by pretreating cells with a concentration-curve of AlloP (0-100 nM).
In C57BL/6J primary neurons, HIV-1 Tat significantly interacted with AlloP to influence Δψ m [F(2,36) = 6.17, p = 0.005] (Fig. 4D and E). Tat significantly depolarized Δψ m compared to controls (p = 0.009) and pretreatment with AlloP (10 or 100 nM) significantly attenuated this effect (p = 0.0003-0.002; Fig. 4D and E). AlloP also significantly interacted with morphine [F(2,36) = 3.28, p < 0.05] such that AlloP abolished Tat or morphine-mediated depolarization at 10 nM, but this protection was lost when the highest AlloP concentration (100 nM) was co-administered with morphine (p > 0.05; Fig. 4D and E). Morphine and Tat did not significantly interact on their own. As such, Tat's contribution to the collapse of Δψ m can be ameliorated by low-to-high physiological concentrations of AlloP. Morphine did not appear to contribute to Tat-mediated effects on Δψ m among isolated medium spiny neurons, but did interact with the highest concentration of AlloP to promote mitochondrial membrane depolarization.

Physiological, but not high, AlloP rescues Tat-induced cell death
The capacity for AlloP to ameliorate Tat-mediated depolarization of Δψ m prompted further examination of AlloP's direct protective capacity on long-term neuron viability. Cells were incubated with treatments for 20 h (a previously-established timepoint at which PROG or AlloP ameliorated Tat-mediated cell death; Salahuddin et al., 2019;Paris et al., 2016). The direct neurotoxic effects of Tat (i.e., in the absence of glial contribution), the concentration-dependent protective effects of AlloP, and their interactions with morphine were sought.

HIV-1 Tat potentiates, and AlloP attenuates, morphine's psychomotor effects
To assess the functional behavioral interactions between HIV-1 Tat, opiates, and exogenous neurosteroid, Tat(+) mice were pretreated with vehicle or AlloP (0.5 or 1.0 mg/kg/d for 8 days, s.c.) while Tat was induced via doxycycline (30 mg/kg/d for 5 days, i.p.; allowing for 2 days of doxycycline-washout prior to the testing day). Multiple negative controls were included. To address the potential non-specific effects of doxycycline, some mice were administered saline (5 d, i.p.) as uninduced controls. To account for potential leaky transgene expression which can be observed in rtTA-transgenic mice, Tat(−) controls were commensurately treated using the high AlloP concentration (1.0 mg/ kg). On day 8, mice received an acute dose of vehicle or morphine (30 mg/kg, i.p.) and were assessed for psychostimulated locomotion 30 min later (Fig. 6A).

Discussion
The hypotheses that exposure to Tat and morphine would alter central or circulating steroid synthesis, directly promote neuronal mitochondrial dysfunction, and neuron cell death were partly upheld. Tat and morphine exerted separate effects to elevate steroid content, with Tat increasing central pregnenolone and 5α-reduced progestogens (while decreasing deoxycorticosterone) independent of changes in plasma. Morphine largely increased both central and circulating pregnane steroids, the glucocorticoid corticosterone, and their 5α-reduced metabolites consistent with activation of the HPA stress axis. In cell culture studies, Tat and morphine also exerted separate effects with Tat promoting glucocorticoid insensitivity, depolarizing Δψ m , and facilitating neuronal death, while morphine hyperpolarized Δψ m and did not alter cell viability on its own. However, Tat and morphine did interact in vivo, with Tat potentiating the psychomotor effects of acute morphine in mice. Consistent with this, Tat exposure has been observed to potentiate acute or sensitized psychostimulant-induced locomotion and reward-related behavior of rodents (Harrod et al., 2008;Kesby et al., 2017;Mediouni et al., 2015aMediouni et al., , 2015bParis et al., 2014a), as well as acute opioid-induced locomotion and behavioral disinhibition (Salahuddin et al., 2019). Notably, Tat's effects to potentiate psychostimulant-reward and opioid-mediated behavioral disinhibition could be attenuated concomitant with cyclical elevations in circulating hormones (Paris et al., 2014b;Salahuddin et al., 2019). The hypothesis that physiological concentrations of AlloP could ameliorate Tat-mediated neuronal insults and resulting changes in behavior was likewise upheld. However, an intriguing interaction was observed between highconcentration AlloP (100 nM) and morphine to depolarize Δψ m and promote cell death. No untoward interactions between AlloP and morphine were observed in vivo; rather, AlloP (1.0 mg/kg) attenuated Tat's capacity to potentiate the physiological effects of acute morphine. These findings support the notion that central steroidogenesis is influenced by HIV-1 Tat or morphine and exogenous AlloP may exert several therapeutic benefits that are of interest for present and future work.
This study is the first to comprehensively profile central and circulating pregnane steroids in an animal model of neuroHIV. In other animal models of CNS insult, including ischemic stroke and traumatic brain injury, increased neurosteroidogenesis is observed and may reflect a local adaptive neuroprotective response (Guennoun et al., 2015;Liu et al., 2012;Schumacher et al., 2014;Zhu et al., 2017). In support, the 18 kDa translocator protein, which mediates mitochondrial cholesterol import for steroidogenesis, is highly expressed in microglia and astrocytes and is markedly upregulated following CNS injury and neuroinflammation (Rupprecht et al., 2010). Exogenous AlloP can improve experimental outcomes, including reductions to infarct volume, gliosis, cytokine upregulation, and improvements to post-recovery cognitive performance (Djebaili et al., 2005;Melcangi and Panzica, 2014;Sayeed et al., 2009). In vitro, AlloP is neuroprotective against hypoxic or excitotoxic insults (Frank and Sagratella, 2000). These effects are likely due in part to its pharmacodynamic profile as a potent positive allosteric modulator of GABA A receptors [and direct agonist in high concentration (Callachan et al., 1987)] and as an antagonist of Ltype Ca 2+ channels (Earl and Tietz, 2011;Hu et al., 2007), readily restoring tonic inhibition within neural networks. As such, AlloP's protective effects against Tat may be partly due to its capacity to rapidly reinstate excitatory/inhibitory tone.  Tat-transgenic mice (n = 14-16/group) were treated with vehicle or allopregnanolone (AlloP; 0.5 or 1.0 mg/kg, s.c.) for 8 days prior to testing and concurrent vehicle or doxycycline (30 mg/kg, i.p.) for 5 days prior to testing (with a 2-day doxycycline washout period prior to testing day). (B,D) Horizontal and (C,E) vertical (rearing) motor behavior were assessed 30 min after an acute injection of saline or morphine (30 mg/kg, i.p.) in Tat(+) and Tat(−) mice. * indicates a significant increase with Tat induction compared to respective uninduced control; † indicates a significant difference with morphine compared to respective saline-administered control; ∧ indicates a significant decrease compared to morphine/oil/Tat-induced mice; ‡ indicates a significant difference for saline/AlloP (0.5 mg/kg)treated mice compared to all other groups (irrespective of Tat condition); (three-way ANOVA in panels B,C; p < 0.05).
HIV-1 Tat and morphine independently altered steroidogenesis in vivo. In addition to local neurosteroidogenesis, Tat may influence central steroid biosynthesis via alterations in cholesterol homeostasis. RNA deep sequencing reveals Tat-mediated dysregulation of genes encoding for lipid and cholesterol homeostasis in rat hippocampal neurons (Mohseni Ahooyi et al., 2018). Moreover, exposure to Tat significantly increased free and total cholesterol and cholesteryl ester within cultured rat neurons, the latter of which was further increased by co-application of cocaine (Mohseni Ahooyi et al., 2018). Opposing effects for Tat (with or without cocaine) to reduce intra-and extracellular cholesterol in astrocytes have also been observed (Cotto et al., 2018), supporting the notion of dynamic changes in central cholesterol bioavailability following exposure to Tat and psychostimulants. Given that Tat increased splenocyte glucocorticoid resistance, it is reasonable to assume that the opioid-driven HPA response would be modified in an opioid-dependent organism.
Opiates are also demonstrated to influence steroidogenesis in rodents, albeit largely via HPA activation and with a bias towards synthesis of 5α-reduced metabolites, which can differ among species, throughout ontogeny, and depend on the nature and duration of opioid exposure (Bunce et al., 2015;Donegan and Bancos, 2018;Taylor et al., 1997). In support, acute or chronic morphine increases central and systemic 5α-reduction and AlloP formation in rodents (Concas et al., 2006;Porcu et al., 2014), effects that could be blocked by the 5α-reductase inhibitor, finasteride (Moradi-Azani et al., 2011;Verdi and Ahmadiani, 2007). While there are notable species-related differences in opioid activation of the HPA axis (Al-Hashimi et al., 2013;Vuong et al., 2010), morphine is observed to induce steroidogenesis in rodents via peripheral sources, particularly the adrenals (Concas et al., 2006;El Daly, 1996). Adrenally-derived corticosterone likely contributed to the great increase presently observed in the brain. High central glucocorticoid content is associated with neuronal injury and death (de Kloet et al., 2015;Sapolsky, 2015). The extent to which these effects may translate to opioid-dependent patients is not known; however, opioid dependence may reduce HPA responsivity and/or obviate diurnal ACTH/cortisol rhythmicity in people (Facchinetti et al., 1984;Vuong et al., 2010). Pharmacologically-inducing withdrawal in people can produce rebound HPA activation, adding complexity (Culpepper-Morgan et al., 1992Kreek, 2007). When abused, the initially pleasurable effects of opioids dissipate and their continued use is largely motivated by a desire to escape the aversive effects of addiction (George et al., 2012;Koob, 2020;Koob and Kreek, 2007). Thus, while Tat-mediated changes in steroidogenesis may involve dysregulation of cholesterol homeostasis, interactions with opioid drugs of abuse may be expected to influence steroid-sensitive behaviors and are likely to differ with dependence vs. acute exposure.
HIV-1 Tat is known to disrupt mitochondrial function and maintaining AlloP at physiological levels may ameliorate mitochondrial deficits. Congruent with the current findings, Tat can depolarize mitochondrial membranes, induce opening of the mitochondrial permeability transition pore (mPTP), drive reactive oxygen species formation, and reduce ATP [reviewed in (Fields and Ellis, 2019;Rozzi et al., 2017)]. Conversely, AlloP or synthetic neurosteroid analogues maintain Δψ m , improve mitochondrial oxygen consumption rate and glycolysis, reduce opening of the mPTP, and preserve ATP synthesis in response to insult (Grimm et al., 2014;Lejri et al., 2017;Leśkiewicz et al., 2008a,b;Sayeed et al., 2009;Waters et al., 1997). However, protective effects of AlloP are not seen at high, pharmacological concentrations (Leśkiewicz et al., 2008a,b;Sayeed et al., 2009). On its own, morphine hyperpolarized Δψ m which is consistent with past observations in primary mouse neurons (Fitting et al., 2014b) and with reduced mPTP opening and Cyt-c release in rat heart tissue (Chen et al., 2016). Physiological AlloP concentrations (e.g., 10 nM) may restore cellular and mitochondrial homeostasis. Thus, increased Tat-induced neurosteroidogenesis may be part of an adaptative cerebroprotective response, perhaps via the enhancement of AlloP concentrations in the 10 nM range that are mito-and neuroprotective. On the contrary, the large increases of corticosterone and AlloP (≂ 50 nM) that occur with morphine treatment may contribute to mitochondrial and neuronal dysfunction.
The mito/neurotoxic interactions between morphine and highconcentration AlloP (100 nM) are intriguing. In low-to-physiological concentrations, AlloP mediates phasic inhibition via activation of synaptic GABA A Cl − channels; at high concentrations, AlloP activates δsubunit-containing, extrasynaptic GABA A receptors mediating tonic or continuing Cl − conductance [reviewed in (Reddy and Estes, 2016)]. In response to an excitotoxin, such as Tat, increased phasic or tonic Cl − influx may be beneficial, restoring ion homeostasis. However, morphine-dependence is demonstrated to downregulate KCC2 in several CNS cell types leading to increased intracellular Cl − concentrations (Ferrini et al., 2013;Gagnon et al., 2013;Taylor et al., 2016), which may promote Cl − efflux in response to a potent agonist, such as AlloP. Moreover, AlloP exhibits high potency and is a particularly efficacious positive allosteric modulator of δ-containing GABA A receptors (Carver and Reddy, 2016). Under these circumstances, high or supra-physiological AlloP and morphine may increase cell excitability. Whether supra-physiological concentrations of AlloP would produce toxic interactions with morphine in vivo is not known; however, benzodiazepines represent another class of positive GABA A allosteric modulators that can exacerbate respiratory depression and many of the pathophysiological effects of opioids seen clinically (Dowell et al., 2016;Jones et al., 2012;White and Irvine, 1999). Future work will further investigate the influence of opioid dependence on AlloP-mediated Tat protection.
It should be noted that toxic interactions between opioids and Tat largely depend on glial μ-opioid receptors (Fitting et al., 2014b;Zou et al., 2011) and interactions with chemokine receptors (Festa and Meucci, 2012;Kim et al., 2018). The present findings utilized isolated neuron cultures to assess the direct neurotoxic relationship that may exist between HIV-1 Tat and morphine, as well as the potential for AlloP rescue. However, AlloP is demonstrated to influence immune responding, quiescing activated microglia (Ahmad et al., 2005;Müller and Kerschbaum, 2006) and directly or indirectly downregulating cytokine production (Ghezzi et al., 2000); immune factors that play a critical role in the progression of HIV (Thaney and Kaul, 2018). It will be important to parse the influence of neurotoxic bystander effects in these systems in the future.

Conclusions
These findings demonstrate the capacity for HIV-1 Tat or morphine to influence central and circulating pregnane steroidogenesis. Administration of exogenous AlloP in physiological concentrations ameliorates Tat-mediated mitochondrial dysfunction and neurotoxicity, even in combination with morphine. In mice, Tat potentiates, and AlloP reverses, morphine's psychomotor effects. Thus, adjunctive therapeutics that can maintain or enhance pregnane steroidogenesis may be beneficial for neuroHIV symptomology among opioid-using and opioidnaïve populations.