A critical role for Macrophage-derived Cysteinyl-Leukotrienes in HIV-1 induced neuronal injury

Macrophages (


Introduction
The neurological manifestations of HIV infection, also known as HIVassociated neurocognitive disorder (HAND), have been estimated to occur in up to 50 % of the HIV-positive population, despite effective virological suppression (Heaton et al., 2011;Sacktor et al., 2002).Treatment with combination antiretroviral therapy (cART) has made great advances in eliminating HIV replication in infected individuals, lengthening lifespans of HIV patients, and decreasing the severity of cognitive impairment and brain pathology; however it has not reduced the prevalence of HIV-associated neurocognitive disorders (HAND) (Antiretroviral Cohort and C., 2017;Tozzi et al., 2001;Yuan and Kaul, 2021).While the underlying pathological mechanism is poorly understood, low-grade chronic inflammation is considered a major impediment to neurocognitive health even in cART treated HIV-infected individuals (Massanella et al., 2016;Spadaro et al., 2017).
Macrophages (MΦ) and microglia are the major target of HIV infection in the brain and play a key role in HIV neurotoxicity (Giulian et al., 1990;Kaul et al., 2001).We have previously shown that activation of p38α mitogen-activated protein kinase (MAPK) is required for stimulation of neurotoxicity in macrophages (MΦ) and microglia by HIV-1 envelope glycoprotein gp120 (Medders et al., 2010).Our present study sought potential neurotoxic pathways downstream of p38α MAPK induced by HIV-1 or HIVgp120 in MΦ, and found a pivotal role of the cysteinyl leukotriene pathway.
Leukotrienes (LTs) are pro-inflammatory lipid mediators and originate from arachidonic acid (AA) through enzymatic modification by the 5-lipoxygenase (5-LO) pathway (Fig. 1).The expression of 5-LO and its activating protein FLAP occurs primarily in cells of the myeloid lineage, including monocytes, MΦ and microglia (Bertin et al., 2012;Duroudier et al., 2009).Upon stimulation of the immune cells, cytosolic phospholipase A2 (cPLA2) releases AA from cellular membranes in a Ca 2+dependent process.AA translocates to the nuclear membrane where FLAP-activated 5-LO converts it to LTA4, which in turn is further processed by one of two enzymes.The zinc-dependent LTA 4 hydrolase (LTA4H) generates LTB 4 whereas the glutathione S-transferase LTC 4 synthase (LTC4S) produces a cysteinyl-LT, LTC4.Both LTB 4 and LTC 4 are released from the cells, and LTC 4 can subsequently be proteolytically converted to LTD 4 and LTE 4 .These latter three constitute the cysteinyl LTs (CysLTs) which interact with five known receptors: CYSLTR1, CYSLTR2, GPR17, P2Y12, and GPR99 (Bertin et al., 2012;Duroudier et al., 2009).
CYSLTR1 and CYSLTR2 can be expressed by MΦ, microglia, astrocytes and neurons in a context-dependent fashion (Bertin et al., 2012;Fang et al., 2006;Okubo et al., 2010;Zhao et al., 2011).CysLTs have been implicated in the pathophysiology of inflammatory diseases (Montuschi, 2010;Savari et al., 2014).However, the role of CysLTs in the central nervous system (CNS) and NeuroHIV/HAND is largely unknown.We show here that inhibition of p38 MAPK reduces the expression of LTC4S and that pre-incubation of neuronal cultures with CysLT receptor inhibitor montelukast (MTLK) prevents neuronal death caused by neurotoxins released from MΦ stimulated with viral envelope protein or infected with HIV-1.Moreover, we find components of the CysLT pathway are differentially regulated in brains of HIV-infected individuals.Utilizing a transgenic mouse model of HIV-associated brain injury (HIVgp120tg mice) (Maung et al., 2014;Toggas et al., 1994) we show that genetic ablation of LTC4S (Ltc4sKO) or CYSLTR1 (Cysltr1KO) prevents injury of neuronal dendrites and pre-synaptic terminals in the rodent cortex and hippocampus, and ameliorates gp120induced neurocognitive impairment.Altogether, our study suggests a critical role for the CysLT pathway in HIV-associated brain injury and neurocognitive impairment.

Reagents
Recombinant viral envelope glycoprotein gp120 from different HIV-1 strains, SF2 (dual-tropic) or SF162 and BaL (CCR5-preferring) and viral stock of HIV-1 BaL, were obtained from the NIH AIDS Research and Reference Reagent Program.A specific p38 MAPK inhibitor, SB203580 was obtained from Calbiochem (San Diego, CA).Montelukast, a specific cysteinyl leukotriene 1 receptor (CYSLTR1) antagonist, was purchased from Cayman Chemical (Ann Arbor, MI).Bacterial lipopolysaccharide (LPS) of E. coli 0111:B4 was purchased from Sigma (St. Louis, MO).The recombinant HIV gp120s were reconstituted in 0.1 % bovine serum albumin (BSA) at 100 X the final concentration and controls received BSA vehicle alone (0.001 % final concentration).LPS was reconstituted in sterile water for tissue culture at 1,000 X the final concentration (1 μg/ ml).SB203580 and montelukast were dissolved in DMSO at 1,000 X the final concentration (10 µM and 1 µM, respectively).SB203580 was added to the cultures for 15 min prior to treatment with 200 pM or 1 nM gp120 of the indicated HIV-1 strain.Montelukast was pre-incubated in cerebrocortical cultures for 30 min prior to treatment with neurotoxic macrophage supernatants.

Preparation of cerebrocortical, mixed neuronal-glial cell cultures
The preparation of mixed neuronal/glial cerebrocortical cultures from embryos of Sprague-Dawley rats has previously been described by our group (Kaul and Lipton, 1999;Kaul et al., 2007).In brief, cells were cultured in 35 mm dishes with poly-L-lysine-coated glass coverslips (1.87 x 10 6 cells per dish) or poly-L-lysine-coated clear-bottom 96 well plates for imaging (BD Falcon; 0.087 x 10 6 cells per well).Generally, cerebrocortical cell cultures were used after 15-17 days in vitro when the majority of neurons were considered to be fully differentiated and susceptible to excitotoxic injury by N-methyl-D-aspartate (NMDA) (Kaul et al., 2007;Medders et al., 2010).The mature cerebrocortical cell cultures from rat consisted of ~ 30 % neurons, ~70 % astrocytes and ~ 0.1 to 1 % microglia (Kaul et al., 2007).Importantly, rodents express homologues of CCR5, CXCR4 and CD4, which are capable of interacting with HIV-1 envelope gp120 (Kaul and Lipton, 1999;Kaul et al., 2007;Maung et al., 2014).For some experiments in which the contribution of microglia to neurotoxicity was investigated, microglia were depleted by pre-treatment of the cultures for 16 h with 7.5 mM L-leucine methyl ester (LME) (Kaul and Lipton, 1999;Medders et al., 2010).All experiments involving animals were approved by the Institutional Animal Care & Use Committees of the Sanford-Burnham Prebys Medical Discovery Institute, The Scripps Research Institute and the University of California Riverside.

siRNA nucleofection
Mononuclear THP-1 cells were transfected at a density of 1 X 10 6 cells/ml using MAPK14 (p38α MAPK) siRNA (Ambion, ID# 1217; "sip38α-1") or non-targeting siRNA duplexes (Ambion/Applied Biosystems, Austin, TX) and the Amaxa nucleofector kits (Walkersville, MD) according to the manufacturer's instructions and as previously described (Medders et al., 2010).At 48 h after nucleofection with 2 µg siRNA, THP-1 cells were exposed for 24 h with 200 pM HIVgp120 SF2 or gp120 SF162 .Conditioned media (CM) from these incubations were then collected and stored at − 80 • C until use in neurotoxicity assays or analysis for secreted CysLT.The mononuclear cells were lysed as described below for Western blotting analysis.

Isolation and preparation of monocyte-derived macrophages
The preparation of human primary monocyte-derived macrophages (MΦ) using a Ficoll gradient was performed as previously described (Medders et al., 2010).In brief, buffy coats or white blood cell concentrates were obtained through the San Diego Blood Bank (San Diego, CA) from healthy volunteers whose identity and sex is anonymous.Peripheral blood mononuclear cells were isolated using density gradients of Ficoll-paque (ρ 1.073; GE Health Life Sciences, Piscataway, NJ) and transferred into 75 cm 2 cell culture flasks.After allowing adherence in RPMI 1640 containing 2 mM glutamine, antibiotics (P/S) and 10 % human AB serum (RPMI-ABS) at 37 • C, 6 % CO 2 , in humidified atmosphere, non-adherent cells were removed by rigorous washing with warm RPMI 1640 (37 • C) without supplements.Isolated, adherent monocytes were differentiated for 5-7 days in the human AB serumcontaining medium.For harvest and further experimentation the cells were detached by treatment with phosphate buffered saline (PBS) containing EDTA (Sigma, 0.2 g/l) for 5 to 10 min at 37 • C and scraping with a rubber policeman.After washing three times in PBS cells were reseeded at 0.25 x 10 6 cells / 0.5 mL medium per well in 24 well plates.Cell viability was monitored with Trypan Blue exclusion and usually exceeded 95 % (Medders et al., 2010).

HIV-1 infection of macrophages
Human MΦ (0.25 x 10 6 cells / well) were inoculated with the M− tropic HIV-1 strain BaL at multiplicities of infection (MOI) ranging from 0.1 to 1.The virus was obtained from the NIH AIDS Research and Reference Reagent Program and propagated in PBMCs using standard procedures (Biggs et al., 1995).Mock infected cells were used to generate un-infected controls.Control cells received media changes and washes in the absence of virus.CM of HIV-infected and un-infected MΦ derived from six different donors were collected and HIV p24 was determined using a commercially available ELISA (Aalto Bio Reagents Ltd., Dublin, IRE) following the suppliers instructions to confirm productive infection.

Mononuclear cell-mediated neurotoxicity
Human mononuclear THP-1 cells or primary MΦ were incubated for up to 4 days in the presence or absence of recombinant viral envelope gp120 (1 nM) of HIV strains BaL and SF162 in the same media used to grow rat cerebrocortical cells in order to generate conditioned cell culture media (CM).Controls without gp120 received 0.001 % BSA (vehicle).Following the conditioning and stimulation period, cell-free supernatants were transferred at a final 10 % concentration to the mixed rat neuronal-glial cultures depleted of microglia as readout for neurotoxicity (Giulian et al., 1990;Giulian et al., 1993).HIV-infected and un-infected MΦ were transferred between 6 and 15 days post infection (dpi.)into conditioned media from rat cerebrocortical cultures for 24 h.Afterwards, the cell-free supernatants were transferred into microglia-depleted cerebrocortical neurons and astrocytes at a final 10 or 50 % concentration.If not stated otherwise, neuronal injury and survival was assessed after 72 h of incubation, as described hereafter.In addition, MΦ and monocytic THP-1 cells were collected as described below for isolation of RNA or protein extraction and immunoblotting.

Cysteinyl leukotriene ELISA
Cysteinyl leukotriene concentrations in supernatants of stimulated mononuclear THP-1 cells or primary MΦ or HIV infected macrophages were determined using a Cysteinyl Leukotriene ELISA Kit (#500390; Cayman Chem., Ann Arbor, MI) following the supplier's instructions.

Animals
HIVgp120tg mice were kindly provided by Dr. Lennart Mucke (Gladstone Institute of Neurological Disease, University of California, San Francisco, CA (Toggas et al., 1994).Two founder lines of HIVgp120tg mice (L1 and L2) were characterized in terms of CNS gene expression and the role of CCR5 in brain injury associated with the expression of the viral envelope protein (Maung et al., 2014).The latter L2 line (Ccr5WT) was selected for this study for its higher expression of the gp120 transgene, more pronounced neuropathology in cortex and previously established behavioral impairment (Maung et al., 2014).Mice deficient in functional LTC4S and CYSLTR1 were kindly provided by Dr. Yoshihide Kanaoka (Brigham and Women's Hospital, Boston MA).Ltc4sKO and Cysltr1KO mice, respectively, were crossbred with HIVgp120tg mice and the F3 generations of HIVgp120tg het -Ltc4s het and HIVgp120tg het -Cysltr1 het mice were used to generate the respective experimental genotypes.All experimental procedures and protocols involving animals were in accordance with NIH and ARRIVE guidelines and approved by the Institutional Animal Care and Use Committee of the Sanford Burnham Prebys Medical Discovery Institute, The Scripps Research Institute, and the University of California, Riverside.

Behavioral testing
The LTC4S cohort consisting of: Wild-type (male: 11, female: 12), GP120 (male: 10, female:10), Ltc4sKO (male: 12, female: 10), and Ltc4sKO/GP120 (male: 13, 9) were tested at 7-8 months of age.CYSLTR1 cohort consisting of: Wild-type (male: 5, female: 7), GP120 (male: 4, female: 7), Cysltr1KO (male: 5, female: 7), and Cysltr1KO/ GP120 (male: 5, female: 7) mice were tested at 12-13 months of age.Testing was performed by the Scripps Research Institute's (TSRI) Animal Models Core Facility.The behavioral test battery was designed to examine cognitive functionality as well as general activity, anxiety-like behavior and visual ability.The behavioral test battery was designed to examine cognitive behavior as well as general activity, anxiety and visual ability.The order of testing (same as the order of test descriptions below) was planned to reduce the potential for confounds due to previous experience.Behavioral testing occurred between 9:00 AM and 12:00 PM (active phase) with 5-7 days between tests.
Light/dark transfer test: This procedure has been used to assess anxiety-like behavior in mice by capitalizing on the conflict between exploration of a novel environment and the avoidance of a brightly lit open field (Crawley, 1999).The apparatus is a rectangular box made of Plexiglas divided by a partition into two environments.One compartment (14.5 x 27 x 26.5 cm) is dark (8-16 lx) and the other compartment (28.5 x 27 x 26.5 cm) is highly illuminated (400-600 lx) by a 60 W light source located above it.The compartments are connected by an opening (7.5 x 7.5 cm) located at floor level in the center of the partition.The time spent in the light compartment was used as a predictor of anxietylike behavior, i.e. a greater amount of time in the light compartment was indicative of decreased anxiety-like behavior.Mice were placed in the dark compartment to start the 5-min test.
Optomotor vision test: This test allows for assessment of visual ability and consists of a stationary elevated platform surrounded by a drum with black and white striped walls.Each mouse was placed on the platform to habituate for 1 min and then the drum rotated at 2 rpm in one direction for 1 min, was stopped for 30 sec, and then rotated in the other direction for 1 min.The number of head tracks (15-degree movements at speed of drum) was recorded.Blind mice do not track the moving stripes.
Locomotor activity test: Test was measured in polycarbonate cages (42 x 22 x 20 cm) placed into frames (25.5 x 47 cm) mounted with two levels of photocell beams at 2 and 7 cm above the bottom of the cage (San Diego Instruments, San Diego, CA).These two sets of beams allowed for the recording of both horizontal (locomotion) and vertical (rearing) behavior.A thin layer of bedding material was applied to the bottom of the cage.Mice were tested for 120 min and data were collected in 5 min intervals.
Barnes maze test: The Barnes maze is a spatial learning and memory test sensitive to impaired hippocampal function (Barnes, 1979;Paylor et al., 2001).It consists of an opaque Plexiglas disc 75 cm in diameter elevated 58 cm above the floor by a tripod.Twenty holes, 5 cm in diameter, are located 5 cm from the perimeter, and a black Plexiglas escape box (19 x 8 x 7 cm) is placed under one of the holes.Distinct spatial cues are located all around the maze and are kept constant throughout the study.
On the first day of testing, a training session was performed, which consisted of placing the mouse in the escape box and leaving it there for one minute.One minute later, the first session was started whereby the mouse was placed in the middle of the maze in a 10 cm high cone-shaped silver start chamber.After 10 s the start chamber was removed, a buzzer (80 dB) and a light (400 lx) were turned on, and the mouse was set free to explore the maze.Sessions ended when the mouse entered the escape tunnel or after 3 min elapsed.When the mouse entered the escape tunnel, the buzzer was turned off and the mouse was allowed to remain in the dark for one minute.If the mouse did not enter the tunnel by itself, it was gently put in the escape box for one minute.The tunnel was always located underneath the same hole (stable within the spatial environment), which was randomly determined for each mouse.Mice were tested once a day for 6 days for the acquisition portion of the study.
A probe test was performed on the day following the final acquisition trial.For this test, the escape tunnel was removed, and the mouse was allowed to freely explore the maze for 3 min.The time spent in each quadrant was determined and the percent time spent in the target quadrant (the one originally containing the escape box) was compared with the average percent time in the other three quadrants.This is a direct test of spatial memory as there is no potential for local cues to be used in the mouse's behavioral decision.
Each acquisition session was videotaped and scored by an experimenter blind to the genotype of the mouse.Measures recorded included the latency to escape the maze and the number of errors made per session.Errors were defined as nose pokes and head deflections over any hole that did not have the tunnel beneath it.The probe data was analyzed using Noldus Ethovision software (Leesburg, VA) to determine time spent in each quadrant of the maze as well as to assess activity measures.We have previously published all these behavioral protocols (Maung et al., 2014;Ojeda-Juarez et al., 2020).

Histological analysis, immunofluorescence staining and deconvolution microscopy
Brain tissue collection, sectioning, immunofluorescence staining of in vitro cell cultures and brain sections, fluorescence and deconvolution microscopy and analysis was performed as described earlier using a computer-supported Zeiss Axiovert M200 microscope controlled by Slidebook software (Intelligent Imaging Innovations; Denver, CO) (Maung et al., 2014;Ojeda-Juarez et al., 2020).In brief, mice were anesthetized with isoflurane and transcardially perfused with 0.9 % N.Y.Yuan et al. saline.The CYSLTR1 mouse cohort for histology consisted of 6 nonbehaviorally tested 6-month old mice (n = 6: 3 male, 3 female per genotype).The LTC4S mouse cohort for histology consisted of 12 of the behaviorally tested cohort (n = 12: 6 male, 6 female per genotype).Brains were removed and immersion fixed with 4 % paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) for 72 h at 4 C. Neuropathological assessment of 40 µm thick sagittal brain sections was performed as previously published with minor modifications (Maung et al., 2014;Ojeda-Juarez et al., 2020).Three independent sections for each mouse were stained.Brain sections were immunostained with primary antibodies against mouse synaptophysin (Dako, Carpinteria, CA 93013, USA, cat# M7315, 1:50) and mouse MAP2 (Sigma, cat # M4403, 1:200) as neuronal markers, or GFAP (Dako, Carpinteria, CA 93013, USA, cat# Z0334, 1:500) and Iba1 (Wako, Richmond, VA 23237, USA, cat# 01919741, 1:125) as astrocytic and microglial markers, respectively.For control purposes, primary antibodies were replaced with mouse IgG1 (MOPC21, Sigma, cat# M9269, 1:500), or secondary antibody alone.Secondary antibodies were goat anti-mouse-Rhodamine Red (Jackson ImmunoResearch, cat # 115295146, 1:100) and goat anti-rabbit Alexa Fluor 488 (Thermo Fisher Scientific, cat# A11034, 1:400).Nuclear DNA was stained with Hoechst (H) 33,342 (Invitrogen, cat# H1399, 1:150).Immunolabelled brain sections were mounted with fluorescenceprotecting mounting medium (Vector laboratories, Burlingame, CA 94010, USA, cat# H1000) onto glass slides and covered with coverslips.To assess neuronal injury and glial activation in murine brain slices, images were evaluated for coded samples using two different approaches: direct quantitative analysis of fluorescence intensities or by deconvolution-supported volumetric estimation of neuropil area.For direct quantitative analysis of fluorescence intensities of GFAP (Layers 1-5) in the frontal cerebral cortex or the hippocampus (CA1), 2D images of each brain section were taken using a 10x objective and the sum of fluorescence intensities (SFLI) values were derived from the area of interest.The SFLI values were normalized for the measured area and adjusted for background by subtracting values obtained from sections incubated with secondary antibodies.Deconvolution microscopy was performed for MAP2 and SYP by taking 3D images from five fields per brain section with 0.5 μm steps along the Z-axis using a 40x objective.
The Z-stack images were deconvolved and used for threshold segmentation to estimate the percentage of the neuropil occupied by MAP2 + and SYP + presynaptic terminals in the frontal cortex (Layer 3) or the hippocampus (CA1).To quantify Iba1 + microglial cells, cell bodies were counted on the lateral side of the sagittal sections of the cerebral cortex (Layer 1-5) or the hippocampus (CA1).Slidebook software (Intelligent Imaging Innovations, Denver, CO) was used for controlling image acquisition and analysis for both approaches.

RNA extraction, microarray and qRT-PCR
RNA was isolated from monocytic THP-1 cells treated for 4 or 24 h with HIV/gp120 (1 nM), LPS (1 µg/ml), vehicle control (0.001 % BSA) in the presence or absence of the p38 MAPK inhibitor SB203580 (15 min pretreatment with 10 µM SB203580 or 0.01 % DMSO vehicle control).RNA was isolated from three independent replicates per experimental condition using RNeasy kits commercially available from Qiagen (Valencia, CA).For each sample ~ 2 x 10 6 cells were collected to obtain 500 ng -2 µg RNA.RNA preparations with an A260/A280 ratio of ~ 1.8-2.0(NanoDrop system, Thermo Scientific, Rockford, IL) was provided to the microarray/QPCR core facility at Sanford-Burnham Medical Research Institute and RNA microarrays was performed using the Human-HT12 v3 Expression BeadChips from Illumina (San Diego, CA).The manufacturer's BeadArray Reader was used to collect primary data.Further analysis comprised three stages: First, expression intensities were calculated for each gene probed on the array for all hybridizations using Illumina's Beadstudio#3 software.Second, intensity values were quality controlled and normalized.Quality control was carried out with Illumina Beadstudio and the detection p-value set to ≤ 0.05 as a cut-off.
All arrays were normalized using the /normalize.quantiles/routine of the Affymetrix software in Bioconductor.This procedure accounted for any potential variation in hybridization intensity between the individual arrays.Finally, the normalized data were imported into GeneSpring and genes significantly differentially expressed between groups were determined on the basis of modified t-tests (cut-off p = 0.05).Robustness and reproducibility of the hybridization signal for a given gene determines the minimum fold change (FC) that can be detected and the overall lowest cut-off was set at FC ≥ 1.36.All analysis was performed in the Bioinformatics Core, Sanford-Burnham Prebys Medical Discovery Institute, La Jolla CA.
Murine brain mRNA extraction was performed following terminal anesthetization and cardiac perfusion with 0.9 % saline solution.The RNA for the CYSLTR1 mouse cohort consisted of 6 non-behaviorally tested 6-month old mice (n = 6: 3 males, 3 females per genotype).The LTC4S mouse cohort for histology consisted of 6 of the behaviorally tested cohort (n = 6: 3 males, 3 females per genotype).Total brain and cortex RNA were isolated using the Qiagen RNeasy Lipid Tissue Midi Kit (Qiagen, cat# 75144) according to the manufacturer's instructions.500 ng of RNA was next synthesized into cDNA by SuperScript II reverse transcriptase (Thermo Fisher Scientific, cat# 18064014) as per manufacturer's instructions.cDNA was then treated with RNase H (Thermo Fisher Scientific, cat# 18021071) for 20 min at 37 • C and stored at − 20 • C until use.Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using Power PCR SYBR Green Master Mix (Thermo Fisher Scientific, cat# 4368708) on a QuantStudio 6 flex Real-Time PCR System (Applied Biosystems/ Thermo Fisher Scientific, Carlsbad, CA) and the primers listed in Table 1.The ΔΔCt values were calculated as previously described (Maung et al., 2014;Ojeda-Juarez et al., 2020).
RNA samples from the National NeuroAIDS Tissue Consortium (NNTC) were prepared by Dr. Benjamin Gelman's laboratory (UTMB Galveston, TX).Highly active antiretroviral therapy (HAART)-era HIV positive subjects assessed in this study were selected on the basis of participants who underwent autopsy with neuropathological evaluation, with non-infected age and demographically matched individuals for control.Post-mortem tissue samples were collected from the middle frontal gyrus matter (neocortex) of HAART-era HIV positive patients and non-infected age matched individuals.At the time of analysis, the known demographic composition of the total HIV positive cohort from which all downstream analysis was performed was: race categorized as 71 % White, 24 % Black, 1 % Asian, 4 % Other race; Ethnicity categorized as 15 % Hispanic, 85 % Not Hispanic; Gender categorized as 80 % male, 20 % female; 38 % of all participants reporting a history of any drug abuse; the average age at the time of autopsy was 45 years old; the average post-mortem interval was 14 h; treatment status of the individuals was unknown but categorized as post HAART-era indicating that at the time of autopsy, highly active antiretroviral treatment was available on the market.For the analysis, RNA from the brains of a total of 145 individuals was used to determine expression levels of CYSLTR1 (n = 137) and LTC4S (n = 126; Supplementary Table S4, see below for determination of outliers).All samples were coded, and the investigators were blind to the HIV status and HIV brain damage diagnostic of the patients during the qRT-PCR experiments.QRT-PCR was performed as previously described (Maung et al., 2014;Ojeda-Juarez et al., 2020).Briefly, reverse transcription was performed using SuperScript II reverse transcriptase (Invitrogen, USA) following the manufacturer's instructions.QRT-PCR was performed using Power PCR SYBR Green mastermix (Applied Biosystems, USA) on an QuantStudio 6 flex Real-Time PCR System (Applied Biosystems/ Thermo Fisher Scientific, Carlsbad, CA).For primers see Table 2.The results obtained were analyzed using QuantStudio Real-Time PCR Software v1.3 (Applied Biosystems/ Thermo Fisher Scientific, Carlsbad, CA).Levels of mRNA of LTC4S and CYSLTR1 in the brain were expressed as relative levels of the internal control, β-ACTIN (Actb).
The RNA microarray data of human monocytic THP-1 cells stimulated with HIV gp120 or LPS in the presence or absence of pharmacological inhibitor of p38 MAPK SB203580 generated in this study have been deposited in the GEO database under accession number GSE53712 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE53712).The whole brain RNA microarray of HIVgp120tg mice expressing or lacking CCR5 are available under accession number GSE47029 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=jhgxnkcscag istw&acc=GSE47029).

Statistical analysis
Data for cell culture experiments are expressed as mean values ± SEM for three-nine independent experiments.Statistical analysis applied one-way analysis of variance (ANOVA) followed by Fisher's.
PLSD post hoc test for multiple comparisons unless stated otherwise using the StatView software package (version 5.0.1,SAS Institute Inc.) or GraphPad Prism 8 (GraphPad Software) as indicated with the significance level set at p ≤ 0.05.Graphs are visualized using GraphPad Prism 8. Analysis of histopathological data, ELISA, and mRNA expression were performed using Prism software (GraphPad Software, Inc., CA, USA) while behavioral data were evaluated using StatView (SAS Institute, Cary, NC).Comparisons of two groups were made by unpaired twotailed Student's t-test, whereas multiple groups were compared by oneway ANOVA followed by Fisher's PLSD post hoc test.p-values ≤ 0.05 were considered statistically significant (p > 0.05 (ns), p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****)).Images of immunoblotting and cellular immunofluorescence are representative samples of the at least three independent experiments performed.No statistical methods were employed to predetermine the numbers of human autopsy samples.After relative gene expression levels for the genes of interest were assessed in a blinded fashion, samples were grouped into noninfected controls, HIV + individuals with no brain pathology, and HIV + individuals with brain pathology (microglial nodules or HIV encephalitis) (n = 66, 56, and 23 respectively).Outliers for CYSLTR1 and LTC4S were determined independently using GraphPad Prism ROUT (Q/FDR = 1 %) (Motulsky and Brown, 2006).Excluding outliers, our final analysis for CYSLTR1 included 137 individuals: 62 noninfected controls, 52 HIV + individuals with no brain pathology, and 23 HIV + individuals with brain pathology (microglial nodules or HIV encephalitis).Analysis for LTC4S included 126 individuals: 55 noninfected controls, 51 HIV + individuals with no brain pathology, and 20 HIV + individuals with brain pathology.The demographics of these groups are described in Supplementary Table S4.Additional NNTC figures for these genes include these same subsets of HIV + individuals, among whom were evaluated for Neurocognitive Dysfunction and/or HAND and those who had quantifiable levels of HIV DNA/RNA in the frontal neocortex and CD68 mRNA in the frontal neocortex.HIV RNA and DNA levels were logarithm transformed using (log 10 x + 200) where x is copies of HIV RNA per gram, and 200 represents the observed threshold of HIV RNA detection of the assay.Levels of CD68 in the samples were provided by Dr. Benjamin Gelman.Demographic and virological characteristics of the cohort from which the samples were derived have previously been described (Kovacsics et al., 2017).Statistical analysis was performed using GraphPad Prism 8 using one-way ANOVA followed by Fisher's PLSD post hoc test for multiple comparisons unless stated otherwise.p-values ≤ 0.05 were considered statistically significant (p > 0.05 (ns), p ≤ 0.05 (*), p ≤ 0.01 (**), and p ≤ 0.0001 (****)).Correlations shown in the supplementary material were also analyzed using GraphPad Prism software with correlation matrices and computation of Pearson's correlation coefficients (r) with two-tailed p value with 95 % confidence interval.

Pharmacological inhibition of p38 MAPK with SB203580 and bacterial lipopolysaccharide cause more pronounced changes in gene expression in monocyte/macrophages than HIVgp120
In order to investigate cellular mechanisms contributing to MΦ neurotoxicity downstream of p38 MAPK, we exposed monocytic THP-1 cells for 4 and 24 h to HIVgp120 (strain SF162; 1 nM), bacterial lipopolysaccharide (LPS) (1 µg/ml), the p38 MAPK inhibitor pSB203580 (SB, 10 µM), combinations of SB203580 with HIVgp120 or LPS, or vehicle control (0.001 % BSA, 0.1 % DMSO).The concentrations of HIVgp120 and SB203580 were based on our previous studies demonstrating the neurotoxicity of the viral envelope protein and the prevention of the toxicity by inhibition of p38 MAPK at those concentrations, respectively (Kaul et al., 2007;Medders et al., 2010).LPS was included as a well-established positive control for induction of proinflammatory signaling.Following the treatments, we isolated RNA and performed a microarray (Illumina HumanHT-12 v4 Expression BeadChip kit) as described in the Methods Section (Fig. 2a-g).A multigroup comparison revealed that many genes were strongly differentially regulated after treatments with SB203580 or LPS alone or in combinations but not HIVgp120 alone (Table S1).In fact, of the genes identified as differentially regulated by HIVgp120 using an unadjusted threshold p-value of ≤ 0.01, 219 and 287 at 4 and 24 h, respectively (Fig. S1b and Fig. 2b), none reached that threshold of significance after adjusting for multiple comparisons (Table S2).In contrast, SB + HIVgp120 altered at 4 and 24 h expression of 1,593 and 2,374 genes, SB203580 alone 1,019 and 1,132 genes, LPS altered 5,400 and 3,574 genes, and SB + LPS 5,109 and 3,811, respectively (Fig. 2b, c, Fig. S1a, b, Table S2).Using an adjusted threshold p-value of ≤ 0.01, SB + HIVgp120 significantly altered expression of 14 and 330 genes, SB203580 alone 3 and 133 genes, LPS 2,190 and 1,708 genes, and SB + LPS 1,610 and 1,656 at 4 and 24 h, respectively (Table S2).The 10 most up-and down-regulated genes in SB + HIVgp120 and SB + LPS samples, respectively, are shown in Fig. 2d, e and Fig. S1c, d.While the results for stimulation with LPS confirm the ability of the THP-1 cells to respond to an inflammatory stimulus and pharmacological blockade of p38 MAPK modifies that response, inhibition of the baseline activity of the kinase also causes significant changes of the cells' gene expression (Fig. 2b-e and Fig. S1ad).Moreover, HIVgp120 significantly affects the transcriptome of SB203580-inhibited but not control/DMSO-treated THP-1 cells (Table S2, Fig. 2d, f and Fig. S1e).Hence, the data suggest that changes of gene expression associated with induction of neurotoxicity in MΦ by HIVgp120 are fewer and less pronounced compared to those caused by LPS or SB203580.

Inhibition of p38 MAPK modifies expression of genes implicated in HIV-1 infection of macrophages and down-regulates cysteinyl leukotriene synthase LTC4S
Next we checked, if blockade of p38 MAPK in the presence and absence of HIVgp120 affected the expression of genes that have previously been linked by others to HIV infection of microglia and MΦ and HIV-associated dementia (HAD) (Albright and Gonzalez-Scarano, 2004).Comparing the published gene lists with our microarray data showed that inhibition of p38 MAPK in the presence and absence of HIVgp120 caused significant changes in the expression of several genes implicated in HIV infection (Table S3, Fig. 1f, Fig. S1e).Of note, genes of the proinflammatory cytokine families (i.e.interleukin (IL)-1β and tumor necrosis factor (TNF)-α, β-chemokines (i.e.CCL2, − 3, − 4, and − 5), regulators of proteases and cytokines (i.e.tissue inhibitor of metalloproteinases (TIMP) 3, α2-macroglobulin (A2M)) and genes involved in AA metabolism (prostaglandin endoperoxidase synthase (PTGS) 1/cyclooxygenase (COX)-1) were up-regulated by exposure to SB203580 alone and in combination with HIVgp120.However, one gene of AA metabolism, leukotriene C 4 synthase (LTC4S) was down-regulated by inhibition of p38 MAPK in the presence and absence of HIVgp120 (Fig. 1f, g, Fig. S1e).The decrease of LTC4S appeared noteworthy because an earlier microarray study of microglia and MΦ detected significant increases in the expression of leukotriene-synthesizing enzymes upon activation (Albright and Gonzalez-Scarano, 2004), and another in vitro study reported HIV-infected MΦ produce LTD 4 (Genis et al., 1992).Therefore, we next tested whether the reduction in gene expression by p38 MAPK inhibition led to a decrease at the protein level.In order to confirm the effect of p38 MAPK activity on LTC4S and CysLT production, we knocked down the kinase in monocytic THP-1 cells using siRNA two days before stimulation with HIVgp120 for 24 h as described in the Fig. 2. LPS and inhibition of p38 MAPK with SB203580 but not HIVgp120 trigger pronounced differential gene expression in monocyte/macrophages. a) Experimental design: In three independent experiments monocytic THP-1 cells were exposed for 4 or 24 hrs to LPS (1 µg/ml) or recombinant HIV gp120 (1 nM) in the presence and absence of pharmacological blockade of p38MAPK activity by SB203580 (10 µM).Subsequently, cellular RNA was isolated and analyzed by microarray.b & c) Venn diagram of differentially expressed genes at 24 hrs following exposure to HIVgp120 (b) and LPS (c) +/-SB203580.d & e) Heatmaps of the genes most upand downregulated at 24 hrs due to SB203580 in the presence of HIVgp120 (d) and LPS (e).f) Heatmap of genes implicated in neuroHIV and differentially regulated by exposure to HIV gp120 with and w/o SB203580 and SB203580 alone at 24 hrs.g) Comparison of microarray signals for a subset of inflammation-related genes differentially regulated by pharmacological blockade of p38MAPK.Note the down-regulation of LTC4S in panels f and g. methods section and previously published (Medders et al., 2010).Western blotting analysis of the cells confirmed the knock-down of p38 MAPK and revealed that it also resulted in a down-regulation of LTC4S, which due to variability did not reach significance in the densitometry assessment (Fig. 3a, b).There also appeared to be a further reduction in LTC4S protein after the addition of HIVgp120 for 24 h with reduction in p38 MAPK, an observation that is in line with the effect of pharmacological inhibition of the kinase.The downstream substrate of p38 MAPK, HSP27, however was not affected by the knock-down of its upstream kinase.On the other hand, HIVgp120 increased the release of CysLTs by MΦ and the knock-down of p38 MAPK suppressed that increase of CysLT production (Fig. 3c).This finding is in line with the observation that blockade or knock-down of p38 MAPK leads to a reduced cellular LTC4S expression as was found in the microarray.The finding that HIVgp120 appears to slightly reduce LTC4S in monocytic THP-1 cells while increasing the release of CysLTs suggests a potential feedback mechanism (Fig. 2g, Fig. 3a-c).Since we showed earlier that knock-down of p38 MAPK also prevents neurotoxicity of HIVgp120, these findings suggested the possibility that CysLTs play a role in HIV-induced brain injury (Medders et al., 2010).

Cysteinyl leukotriene receptor antagonist, montelukast, protects neurons from both HIVgp120 and HIV-1 induced macrophage neurotoxicity without affecting infection
We next sought to determine, if pharmacological blockade of CysLT receptors might prevent HIVgp120-induced neurotoxicity and neuronal loss.Montelukast (Singulair®, MTLK), a specific CYSLTR1 antagonist (1 µM) (Tahan et al., 2008), was added to microglia-depleted neuronalastroglial cell cultures 30 min prior to incubation for three days with conditioned media (CM) from THP-1 cells treated with HIVgp120 (THP1 gp120 CM) or without (THP-1 Ctl CM) for 24 h.Two strains of CCR5 preferring gp120 were examined: the M− tropic gp120 BaL and gp120 SF162 .After the incubation, the cerebrocortical cells were fixed and neurons were immunofluorescence-labeled in combination with nuclear DNA staining by Hoechst33342.The microscopic quantification of postsynaptic microtubule-associated protein (MAP2)+/NeuN + neurons indicate that pretreatment with MTLK abrogates the neuronal injury observed after treatment with neurotoxic THP-1 conditioned media (Fig. 4a, b).
Next, we inoculated human monocyte-derived MФ (0.25 x 10 6 cells / well) with the M− tropic HIV-1 strain BaL at multiplicities of infection (MOI) of 0.1 as described in the methods section.MФ CM of HIV-infected (MФ HIV CM) and mock/un-infected MФ (MФ Control CM) was generated in the same media used for the cerebrocortical cell cultures.Microglia-depleted cerebrocortical neuroglial cells were pre-treated with MTLK or vehicle control before MФ CM of HIV-infected and un-infected MФ derived from three different donors was transferred into the neuronal-astrocytic cell cultures at 50 % final concentration.After three days of incubation the cerebrocortical cells were fixed and neurons labeled with MAP2 in combination with nuclear DNA staining by Hoechst33342.The quantification of MAP2 + neurons by microscopy showed that 50 % of MФ HIV CM caused significant neuronal loss whereas CM of un-infected MФ does not compromise neuronal survival.Additionally, inhibition of CYSLTR1 by MTLK protected neurons against HIV-associated MФ toxicity (Fig. 4c, d).In a separate experiment, MФ CM of HIV-infected and un-infected MФ was transferred into the cell cultures at 10 % final concentration in the presence and absence of MTLK.After three days of incubation the cerebrocortical cells were fixed and neurons and pre-synaptic terminals labeled with MAP2 and presynaptic marker protein synaptophysin (SYP), respectively, in combination with nuclear DNA staining by Hoechst33342.The 10 % concentration of MФ HIV CM suffices to cause significant synaptic injury without cell loss.The quantification of immuno-labeled synaptophysin indicated that inhibition of CYSLTR1 by MTLK prevented the loss of neuronal pre-synaptic terminals that is associated with HIV neurotoxicity (Fig. 4e, f).
Next, CM of HIV-infected and un-infected MФ derived from six different donors were collected and HIV p24 was determined using a commercially available ELISA (Aalto Bio Reagents Ltd., Dublin, IRE) to confirm productive infection.Quantification of HIV p24 levels at 9 dpi in CM of infected MФ pre-incubated with either p38 MAPK inhibitor SB203580 (10 µM) or Montelukast (1 µM) revealed that blocking of p38 MAPK decreases viral replication which is in line with the prevention of increased CysLT production.In contrast, CYSLTR1 blockade did not alter viral titers produced by infected MФ (Fig. 4g).
CysLT levels were measured and monitored in CM of infected and uninfected MФ over a period of 24-30 dpi.Compared to uninfected controls, HIV-infected MФ produced significantly higher levels of CysLTs (Fig. 4h).However, HIV-1 induced increases in CysLT production above baseline was largely prevented by inhibition of p38 MAPK activity with SB203580.

Expression of CYSLTR1 and LTC4S is altered in brains of HIV patients
Since our data suggested a potential role of CysLTs in HIV-associated neuronal injury, we performed quantitative reverse transcription polymerase chain reaction (qRT-PCR) using RNA extracted from the middle frontal gyrus matter (neocortex) of 79 HAART-era HIV + individuals and 66 non-infected age-and demographically matched control patients.These human RNA samples were provided by the National NeuroAIDS Tissue Consortium (NNTC) and the demographic and virological information and neuropathological evaluation of the patients in this particular autopsy cohort have been established in several previously published reports that used this patient cohort (Gelman et al., 2013;Ojeda-Juarez et al., 2020).HIV + samples were separately evaluated based on the presence and absence of HIV associated brain pathology, defined as patients diagnosed with either HIV encephalitis (HIVE) or microglial nodule encephalitis (Gelman et al., 2013).HIV + individuals displaying no brain pathology had significantly lower levels of CYSLTR1 than either non-infected controls or HIV + individuals with brain pathology.In contrast, HIV + patients with HIV-associated brain pathology had significantly higher levels of CYSLTR1 than both noninfected and infected patients without neuropathology (Fig. 5a).LTC4S expression in HIV + individuals with brain pathology was significantly lower compared to non-infected controls (Fig. 5b).
Additionally, when the two HIV + groups are merged, LTC4S is significantly lower among individuals with HIV (Fig. S2a).Among these HIV + individuals who were neurocognitively assessed and non-infected control group with either no assessment or were found to be neurocognitively normal, CYSLTR1 was not significant (Fig. S2b) but LTC4S was significantly lower among HIV + patients with neurocognitive dysfunction (Fig. S2c).Further separating the type of neurocognitive dysfunction into HIV-associated (HAND) vs neuropsychological impairment due to non-HIV-related or undetermined causes (NPI-O), shows no significant difference for CYSLTR1 (Fig. S2d) but reveals significantly lower levels of LTC4S mRNA in HIV + individuals with HAND compared to both HIV + without HAND and non-infected individuals (Fig. S2e).This latter finding was reminiscent of the observation that HIVgp120 lowered LTC4S in monocytic THP-1 cells while increasing the release of CysLTs suggesting a potential feedback mechanism between increased CysLT concentrations acting on CysLTRs and LTC4S expression (Fig. 2g, Fig. 3a, b).Within the HIV + group LTC4S correlates with CD68 mRNA expression (Fig. S2f).CYSLTR1 expression on the other hand correlates with HIV DNA (Fig. S2g) and RNA (Fig. S2h) which is in line with signs of pathology in the HIV + brain.

Cysteinyl leukotriene associated gene expression increases in brain tissue of HIVgp120tg mice over time
HIVgp120tg mice manifest hallmark neuropathological features observed in NeuroHIV brains, such as decreased synaptic and dendritic density, increased numbers of activated microglia and pronounced astrocytosis (Toggas et al., 1994).HIVgp120tg mice also develop significant physiological and behavioral changes, such as reduced spatial retention at 12 months of age (D'hooge et al., 1999;Krucker et al., 1998;Maung et al., 2014).Our earlier CNS gene expression study comparing HIVgp120tg and control (WT) mice revealed a significant overlap for differentially expressed genes between HIVgp120tg mice and human HIV encephalitis (HIVE) patients, including alterations in the leukotriene pathway (Gelman et al., 2012;Maung et al., 2014;Yuan et al., 2022).HIVgp120tg mice display increased expression of both Flap and Cysltr1 over time compared to WT controls (1.5 to 16 and 6 to 16 months, respectively; Fig. 6a, detection in microarray) and Ltc4s, Flap and Cysltr1 at 12 months as detected by qRT-PCR (Fig. 6b, e, g).

Genetic deletion of Ltc4s or Cysltr1 in HIVgp120tg mice affects expression of other components of the CysLT pathway
In order to further assess the potential role of the CysLT system in HIV neurotoxicity in vivo, we cross-bred HIVgp120tg separately with Ltc4sKO and Cysltr1KO mice.HIVgp120tg mice deficient in Ltc4s or Cysltr1 are viable and fertile.Knockout was confirmed via qRT-PCR of mRNA extracted from murine cerebrocortical brain tissue (Fig. 6b, i).The expression of the gp120 transgene results in marked increases of Montelukast prevents HIVgp120-induced neurotoxicity from human monocytic THP-1 cells.Conditioned media (CM) was collected from THP-1 cells treated with gp120 of HIV-1 SF162 and BaL (each at 1 nM) or 0.001 % BSA-vehicle (Control) for 4 d.Microglia-depleted cerebrocortical cells were then pre-treated with 1 µM montelukast for 30 min prior to incubation with THP-1 CM for 3 d.Neuronal survival was assessed using fluorescence microscopy and cell counting after fixation, permeabilization and immunolabeling of neurons for MAP2 in combination with nuclear DNA staining by H33342.THP-1 Control CM-treated cerebrocortical cells were set as baseline of 100 % for neuronal survival (32.8 % ± 2.9 % neurons [mean ± s.e.m], n = 9 experiments).b) Fluorescence microscopy of cerebrocortical cells exposed to cell-free CM from THP-1 cells +/-montelukast as described above.c -f) Conditioned media from HIV-1 infected macrophages (MΦ HIV CM; 50 % (c, d) or 10 % (e, f)) are neurotoxic but blockade of CYSLTR1 rescues neuronal survival (c, d) and presynaptic terminals (e, f).Microglia-depleted cerebrocortical cells were pre-treated with 1 µM montelukast for 30 min prior to incubation with MΦ HIV CM or MΦ control CM (CM of mock-infected MΦ) for 3 d and evaluated for neuronal survival as described above.Cerebrocortical cells exposed to 10 % MΦ HIV CM showed no loss of neurons but quantification of immunofluorescence staining for Synaptophysin indicated loss of presynaptic terminals which was prevented by Montelukast.g) Blockade of active p38 MAPK (SB203580, 10 µM) but not of CYSLTR1 (Montelukast, 1 µM) inhibits viral replication in HIV-1 infected human MΦ.Cells were cultured in the presence of the inhibitors of DMSO vehicle during the infection and HIV-1 p24 was measured in cell-free CM at 9 dpi.h) HIV-1 infection of MΦ increases release of cysteinyl leukotrienes (CysLT) and the increase is inhibited by blockade of p38 MAPK.HIV infected and MΦ were cultured in the presence of absence of 10 µM SB203580.Controls received DMSO vehicle.CysLT in cell-free MΦ CM collected 24-30 days post infection (dpi) were quantified using an ELISA.Values in graphs represent mean ± s.e.m. of three independent experiments each with 6-8 replicates per condition.** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to control values (ANOVA with Tukey's HSD post hoc test).nearly all of the leukotriene pathway related constituents in the 12months Ltc4sKO/WT cohort, with the exception of Cysltr2 (n.s.), (Fig. 6c-h).This trend holds true in the younger 6-months Cysltr1KO/ WT cohort but was only statistically significant in the increase of Cysltr1 (Fig. 6j-o).Ltc4sKO alone resulted in only minor changes in the expression of CysLT pathway components including decreases in 5-Lo and Cysltr2 expression (Fig. 6d, h).Likewise, Cysltr1KO alone resulted in no significant alterations in the expression of pathway components.HIVgp120tg mice with the Ltc4sKO (Ltc4sKO GP120) led to reductions in the gp120-induced increases in 5-Lo, Flap, Gpr17, and Cysltr1 (Fig. 6d, e, f, g).In contrast, Cysltr1KO in the HIVgp120tg mice (Cysltr1KO GP120) led to increased levels of gp120-induced gene expression for cPla2, 5-Lo, and Flap (Fig. 6k, l, m).Transcript levels of HIVgp120 transgene for both Ltc4sKO GP120 and Cysltr1KO GP120 were not significantly different from those in the HIVgp120tg/GP120 animals but were significantly different from one another as the Ltc4sKO GP120 mice expressed lower levels of the HIVgp120 transgene than the Cysltr1KO GP120 (Fig. 6p).3.7.Ltc4sKO protects HIVgp120tg mice from neuronal injury in the cortex and hippocampus and reveals a sex differential effect in Iba1 + microglia Quantification of immunofluorescence staining in HIVgp120tg mice using microscopy shows a marked decrease in MAP2 and Synaptophysin while exhibiting a significant increase in both GFAP and the number of activated microglia which is in line with previously published findings from our lab and others (Fig. 7a-t).Genetic knockout of the Ltc4s prevents gp120-induced loss of postsynaptic MAP2 throughout the cortex layer (L) 3 and the hippocampal CA1 regions (Fig. 7a, b, e, f).Similarly, Ltc4s knockout protected against gp120-induced loss of presynaptic SYP (Fig. 7c, d, g, h) with levels in the hippocampus showing significantly higher expression than in the WT controls.Interestingly, Ltc4sKO animals exhibited an increase in GFAP protein in both the cortex and the hippocampus.However, Ltc4sKO GP120 animals had significantly lower levels of cerebrocortical GFAP than GP120 mice (Fig. 7i, j), suggesting that Ltc4s may play a role in astrocytic cell activation.Intriguingly, Ltc4sKO had clear sex differences in the Iba1 + microglial cell counts in both the cortex and the hippocampus.Ltc4sKO in females alone led to a significant increase in the number of Iba1 + microglia, which was unaltered in the presence of the microglial activating proinflammatory gp120 viral protein (Fig. 7m -p).Quite the opposite, Ltc4sKO in male mice resulted in a suppression in the number of cortical and hippocampal Iba1 + cells despite the presence of gp120 (Fig. 7q-t).

Cysltr1KO protects HIVgp120tg mice from neuronal injury in cortex and hippocampus and reveals a sex differential effect in the GFAP astrocytic cell marker
Similar to the Ltc4sKO, absence of Cysltr1 prevented gp120-induced loss of MAP2 + neuropil (Fig. 8a, b, e, f) and rescued gp120-induced SYP loss throughout the cortex L3 and hippocampal CA1 regions (Fig. 8c, d,  g, h).Interestingly, HIVgp120-expression appeared to slightly boost SYP + neuropil of the Cysltr1KO mice in the cerebral cortex (Fig. 8c -d).Cysltr1KO in the hippocampus had decreased levels of presynaptic SYP staining compared to the WT controls which was boosted back to control levels in the presence of the neuroinflammatory HIVgp120 transgene in Cysltr1KO GP120 mice (Fig. 8g-h).Similar to the Ltc4sKO, Cysltr1KO also resulted in a sex differential effect but it was exhibited in the quantitative expression of astrocytic GFAP rather than the microglial Iba1 + cell numbers.Female Cysltr1KO mice have a significant increase of GFAP in both the cortex and the hippocampus which is further exacerbated in the cortex by HIVgp120 expression (Fig. 8i, j, k, l).However, in male mice, Cysltr1KO does not appear to play a role in astrocytic GFAP levels in either the absence or presence of HIVgp120 (Fig. 8m-p).Cysltr1 knockout did not alter the gp120-induced increase in microglial numbers in either the cortex or hippocampus.However, Cysltr1KO mice without the gp120 transgene had a decrease in the baseline number of microglia in the cortex (Fig. 8q-t).

Knockout of Ltc4s or Cysltr1 rescues HIVgp120-induced spatial memory impairment and reveals a role for the CysLT pathway in memory performance
Assessing spatial memory performance in the Barnes maze probe test found WT controls and Ltc4sKO GP120 mice spent in contrast to GP120 animals equally and significantly more time in the target quadrant than in any other quadrants (p < 0.01), indicating a rescue of spatial memory function in Ltc4s-deficient GP120 mice (Fig. 9a).However, Ltc4sKO mice failed to differentiate the target quadrant from the other quadrants, indicating the need of CysLTs for normal spatial memory function in the absence of neurotoxicity of HIVgp120.Similarly, in contrast to GP120 animals, the WT controls (p < 0.05), Cysltr1KO (p < 0.05), and Cysltr1KO GP120 (p < 0.01) mice spent more time in the target quadrant than in any other quadrants, indicating a rescue of spatial memory function in Cysltr1-deficient GP120 mice (Fig. 9b).However, the difference in time spent between target and other quadrants for Cysltr1KO mice was only significant because the performance was less variable than that of GP120 mice, indicating that Cysltr1 may also contribute to normal spatial memory function in the absence of HIVgp120.No significant differences were observed between genotypes in errors and latencies except for a shorter latency in Ltc4sKO compared to their respective WT mice (Fig. S3a-d).Other behavioral tests indicated that mice of all genotypes had intact vision (head tracks in optomotor test) and there were no significant differences in locomotor activity (Fig. S3e-h).Light-dark transitions and time in light as a measure of anxiety showed only minor differences between genotypes (Fig. S3i-l).In contrast to the neuropathological analyses, none of the behavioral tests indicated any sex-dependent effects.

Discussion
HIVgp120, in contrast to bacterial LPS and the pharmacological inhibition of p38 MAPK, alters gene expression of MФ to such a subtle degree that none of the changes observed in a microarray experiment reach statistical significance upon correction for multiple comparisons.The same amount of HIVgp120 concentration is however able to induce neurotoxicity (Medders et al., 2010).Thus, HIVgp120 appears to be able to induce neurotoxicity of MФ in association with subtle effects on the cells' gene expression compared to LPS or inhibition of p38 MAPK.The finding also suggests the possibility that the toxicity causing mechanism relies, at least in part, on constitutively expressed host genes.
We showed previously that inhibition of p38 MAPK abrogates HIVgp120-induced neurotoxicity of MФ (Medders et al., 2010) but the downstream factors executing the toxicity remained to be identified.
Here we show that blockade of the kinase significantly changes the gene expression profile in the absence and presence of HIVgp120 or LPS.The induction of pro-inflammatory cytokine or chemokine production from MФ has been linked to p38 MAPK activation in previous studies (Kang et al., 2006).Activation of p38 MAPK was increased in HIVgp120 treated monocytic cells, but our microarray study revealed it did not significantly increase cytokine or chemokine production.Paradoxically, the suppression of p38 MAPK activity in monocytic cells via pharmacological inhibitor SB203580, and SB203580 plus HIVgp120 (SB + gp120), was able to significantly increase the expression of several cytokines, chemokines, and growth factors.While this finding might suggest that the pattern of cytokine and chemokine production observed after p38 MAPK inhibition was not associated with neurotoxicity, the increase observed for mRNA of multiple factors remains to be confirmed at protein level.
Inhibition of p38 MAPK activity in monocytes/MФ reduces HIVgp120 neurotoxicity either by pharmacological blockade, p38α siRNA knockdown, or heterologous overexpression of dominant negative p38α MAPK mutant (Medders et al., 2010).Therefore, we inspected the differential gene expression for potential protective or toxicitydisrupting mechanisms induced by SB203580 alone or in combination with HIVgp120.One gene, leukotriene C 4 synthase (LTC4S), was significantly reduced in MФ after treatment with SB203580 alone or combination of SB + gp120, and slightly changed after 24 h of HIVgp120 alone.Leukotrienes are lipid mediators of inflammation and metabolites of arachidonic acid, a lipid previously implicated in HIV-1 associated neurodegeneration (Basselin et al., 2011;Yuan et al., 2022).Other groups have linked activation of p38 MAPK to leukotriene biosynthesis from monocytic cells with the potential for neuronal injury (Okubo et al., 2010;Werz et al., 2000).LTC4S is involved in the conversion of LTA 4 into LTC 4 within MФ, which is then released from the cells and further converted into LTD 4 and LTE 4 .All three CysLTs have been implicated in mediating excitotoxic or ischemic neuronal injury (Lecca et al., 2008;Zhang et al., 2006).Our study is the first to show that expression of LTC4S, a key enzyme required for all downstream CysLT biosynthesis, is down-regulated by p38 MAPK inhibition, both at the  mRNA and protein level thus providing a potential explanation for neuroprotection.
Importantly, an increase in CysLT receptor expression has been reported in injured neurons within the brain and addition of CysLT receptor antagonists (pranlukast and montelukast) protected neurons from excitotoxic injury (Yu et al., 2005a;Yu et al., 2005b;Zhang et al., 2006).In line with those reports, our study shows that inhibition of CYSLTR1 with the specific antagonist montelukast in neuronal cultures protected against MФ neurotoxicity induced by HIVgp120 and also HIV-1 infection (Fig. 2a-f).This observation supports the notion that CysLTs play a significant role in HIV neurotoxicity, and the envelope protein suffices to trigger the CysLT-dependent pathway to neuronal injury.Whether or not other viral proteins can initiate the same neurotoxic mechanism remains to be elucidated.
HIV-1 infected and uninfected human MФ produce CysLTs.However, HIV-1 infected MФ produce higher levels of CysLTs compared to uninfected cells over time (Fig. 2h, p ≤ 0.0001).This indicates that HIV-1 infection stimulates CysLT production in human MФ.Additionally, inhibition with p38 MAPK using SB203580 in infected monocytederived MФ significantly decreases HIV-1 induced increases in CysLT production (p < 0.01).Notably, the production of CysLTs by monocytic THP-1 cells in response to HIVgp120 was on average comparably higher than that observed in HIV-infected primary MΦ.While we cannot directly compare the intensity of stimulus provided by envelope protein exposure for 4 days and HIV infection for 6-15 days, we also cannot exclude that monocytic THP-1 cells may not completely reflect primary MΦ derived from different donors.CysLTs while critical for neurotoxicity are presumably not the only essential factor in HIV-(or gp120-) induced neurotoxic supernatants (Ellis et al., 2007;Gartner, 2000;Gonzalez-Scarano and Martin-Garcia, 2005;Kaul et al., 2001;Saylor et al., 2016).However, a slight yet critical increase in CysLTs in combination with other factors that HIV-infected macrophages release may suffice to bring about neuronal injury.Also, whether the inhibition of increased CysLT production is the lone neuroprotective effect brought about by blockade of p38 MAPK in MФ is yet unknown.It is conceivable that changes in gene expression triggered by p38 MAPK inhibition, such as we observed here, result in the disruption of other toxicity producing pathways or release of other protective factors.
Interestingly, measurements of viral titer through HIVp24 reveals that there are inherent differences in how protection is achieved by p38 MAPK inhibition using SB203580 versus CYSLTR1 blockade using montelukast (Fig. 2g).Whereas p38 MAPK inhibition decreases HIVp24 (p ≤ 0.0001), CYSLTR1 inhibition does not; as SB203580, but not montelukast, diminished viral replication in HIV-1 infected MФ.Thus, the protective effect of CYSLTR1 inhibition cannot be explained by reduction of HIV-1 infection, which is also in line with montelukast's protection against MФ toxicity induced by recombinant viral gp120 in the absence of an infectious virus.
The qRT-PCR results from post-mortem samples of non-infected controls and HIV-positive patients with or without neuropathology reveals that HIV-positive patients differ in their relative levels of CYSLTR1 with higher expression in cases with hallmarks of neuropathology (Fig. 2i, p ≤ 0.0001).Decreased expression of LTC4S mRNA among HIV + individuals with brain pathology (Fig. 2j, p ≤ 0.05) as well as the overall trend downwards among HIV + individuals (Fig. S2) suggests overall alterations in the leukotriene pathway in HIV infection.The reduced expression compared to uninfected controls was a resemblance of the effect HIVgp120 exerted on monocytic THP-1 cells (Fig. 1g-i).It remains to be elucidated if the reduction of LTC4S in human HIVinfected brains can occur as a negative feedback response to increased CysLT production as our observation with HIVgp120-stimulated THP-1 cells suggests (Fig. 1j).However, an earlier microarray study of human microglia and MΦ detected a significant increase in the expression of LTC4S upon inflammatory activation, such as it may occur in HIVinfection (Albright and Gonzalez-Scarano, 2004).That is in line with our observation of elevated mRNA levels for the enzyme in 8-13 months-old HIVgp120tg mouse brains.Gene expression in brains of HIVpositive patients and controls of the NNTC cohort are more variable than in the model systems.The overall down-regulation of LTC4S could also be due to cART having a similar effect to what inhibition of p38 MAPK achieved in monocytic THP-1 cells, where the expression of the enzyme was reduced, and in HIV-infected MΦ, where release of CysLT and HIVp24 was diminished.Analysis of both CYSLTR1 and LTC4S in correlation to neurocognitive domains demonstrates lower expression levels of LTC4S specifically among HIV + individuals with HAND (Fig. S2).However, due to the overall lack of representation of HIV + patients who did not exhibit HAND in the evaluated sample population, further investigation on these neurocognitive trends is necessary.While these human samples of the NNTC cohort provide a valuable crosssectional view, the HIVgp120tg mice allowed a longitudinal analysis showing that beginning at 3 months the gene expression of two leukotriene pathway associated genes Flap and Cysltr1 in the gp120-exposed brains appeared to steadily increase with age until 12 months of age whereupon Flap levels began to decrease (Fig. 3a).
Two additional CysLT receptors, CysLTR2 and GPR17 have also been implicated in neuronal injury and protection and were therefore evaluated at RNA level in the HIVgp120tg knockout mouse model (Eriksson et al., 2018;Huang et al., 2008;Lecca et al., 2008;Marschallinger et al., 2015; Zhao et al., 2011).Ltc4sKO decreased Cysltr2 expression and prevented HIVgp120-induced increases in GPR17.Cysltr1KO alone did not significantly alter the gene expression of Cysltr2 or Gpr17.However, HIVgp120 expression in Cysltr1KO mice led to a significant increase in Gpr17 mRNA expression.It has been posited that both CysLTR2 and GPR17 reduce CYSLTR1 function by direct physical interactions (Jiang et al., 2007).Thus, the increased Gpr17 expression may be a result of the absence of Cysltr1 and the corresponding receptor-to-receptor regulation during chronic inflammation brought on by HIVgp120 resulting in neuronal protection.In contrast, it cannot be discounted that some amount of the neuroprotection observed in the Ltc4sKO may in part be due to the decrease in Gpr17 as this receptor has also been implicated in brain injury in the presence of Cysltr1 (Lecca et al., 2008).However, in the absence of Ltc4s and without CysLTs, there is presumably no injurious signaling of CYSLTR1 or GPR17 occurring and thus no regulation by CYSLTR2 interference may be required to ensure neuronal protection (Jiang et al., 2007).Increased transcript levels of the enzymes involved in leukotriene synthesis (Ltc4s, Flap, cPla2, and 5-Lo) in the absence of Cysltr1 may be the result of compensatory activation due to the major role that CYSLTR1 plays in the detection of CysLTs and fits with a feedback regulation of the pathway.Likewise, the increase in mRNA of HIVgp120-induced leukotriene pathway constituents caused by Ltc4sKO (Cysltr1, Flap, cPla2, and 5-Lo) may be a feedback mechanism due to the lack of available CysLT ligands.
Knockout of both Ltc4s and Cysltr1 led to complete rescue of neuronal loss in cerebrocortical and hippocampal regions of HIVgp120tg mice and a rescue of spatial memory.However, impaired performance of Ltc4s-and Cysltr1KO mice in the absence of HIVgp120 and the decrease in baseline levels of synaptophysin-positive neuropil in the hippocampus of Cysltr1KO suggests that the functional role of CysLTs may not be confined to the diseased or injured brain but also include general neuronal health of presynaptic terminals and memory function.Genetic ablation of the receptor led to complete protection of MAP2 + and Synaptophysin + neuropil in the cortex and the hippocampus; with levels beyond those observed in WT of Synaptophysin in Cysltr1KO GP120 cortex (Fig. 5c, g).Decreases in the baseline expression of Synaptophysin in Cysltr1KO hippocampus may be suggestive of a regiospecific developmental role for CYSLTR1 in neuronal presynaptic terminals as this decrease was not observed in the cortex or montelukast exposed cerebrocortical cell cultures.
Interestingly, the two knockout models most strongly diverge in the sex differences which emerge in the glial cell signatures.The genetic ablation of Ltc4s significantly disrupts the numbers of Iba1 + microglial cell observed in both the cortex and hippocampus of male and female mice but in opposing directions.Ltc4sKO in female mice results in an observed increase in the basal levels of microglia while male mice have the same number as their wild-type controls.Expression of the HIVgp120 transgene fails to elicit an increase of Iba1 microgliosis in both sexes of Ltc4sKO (Fig. 4m-t).In contrast, Cysltr1KO appears to have little effect on the HIVgp120-induced increases in microgliosis.However, the basal level of microglia in the cortex of Cysltr1KO mice is lower than those found in WT controls.While both male and female Ltc4sKO mice have increased levels of basal GFAP, only female Cysltr1KO mice exhibited increases in GFAP.The two knockout models again diverge as the Ltc4sKO appears to decrease cortical HIVgp120-induced astrocytosis while the Cysltr1KO increases it.It has been shown by other groups that Cysltr1 plays a role in microglial activation (Yu et al., 2014) and astrocyte migration (Huang et al., 2012).In addition, there has been evidence that pharmaceutical modulation of 5-LO, FLAP, CYSLTR1, and CYSLTR1/CYSLTR2 prevents astrocyte proliferation in the mouse model of oxygen-glucose deprivation (Huang et al., 2008).Correspondingly, Ltc4sKO in our model increased basal GFAP expression yet decreased HIVgp120-induced activation in the cortex.Likewise, Ltc4sKO strongly inhibits HIVgp120-induced increases in microgliosis, though in females it cannot be discounted that this may be due to the microglial cell population reaching a maximum density.In contrast to the Ltc4sKO, knockout of Cysltr1 in our model unexpectedly resulted in increased astrocytosis in the cortex of female mice in the absence and presence of the viral gp120 transgene.These findings suggest that both Ltc4s and Cysltr1 play a central role in glial homeostasis in a sex-dependent manner.The diverging histological findings for GFAP and Iba1 markers suggests that LTC4S and CYSLTR1 modulate astrocytic and microglial function contingent on gender.The Cysltr1 gene is located on the X chromosome in both mice and humans.Studies of the sex differences in the production of CysLTs have revealed that formation of leukotrienes and 5-hydroperoxyeicosatetraenoic acid in female blood is significantly higher than male blood when stimulated with LPS or Ca 2+ ionophore A23187 (Pergola et al., 2008).Several studies have observed that females produce higher LT levels and respond to LT modifiers with higher efficacy than males (Pace et al., 2017;Rossi et al., 2019).This sexual dimorphism may be the result of testosterone's ability to suppress LT production (Pace et al., 2017).Thus, it cannot be discounted that pharmaceutical targeting of leukotrienes would have a greater effect in those with higher leukotriene levels.
Our results suggest that CysLT synthesis and CYSLTR1 have a major contributing role in HIV-induced neuronal injury, while concurrently modulating the surrounding glial environment.Despite these differences, both knockout models prove to rescue the mice from neuronal damage and loss of function, regardless of gender.Challenging the HIVgp120tg mice via the Barnes maze test reveals HIVgp120-induced cognitive impairment while knockout of the both Ltc4s and Cysltr1 gene led to complete rescue of this phenotype.Our finding does not exclude that other pathways are contributing to gp120 neurotoxicity.However, it should be noted that in the absence of the neuroinflammatory HIVgp120 transgene, knockout of these genes compromises memory function.Taking these findings in combination with the results from post-mortem samples of HIV patients appear to indicate that individuals may have better outcomes (no neuropathology, lower HIV DNA/RNA levels in frontal neocortex) when CYSLTR1 levels are lower.It is well reported that in a healthy brain, leukotriene levels are low or undetectable but injury can result in a marked elevation in the brain (Farias et al., 2009;Schuhmann et al., 2003).Interestingly our results suggest that though the relative levels of LTC4S and resultingly the produced CysLTs may be similar among all HIV + individuals; the expression of the downstream receptor of the ligands, CYSLTR1, may be more indicative of brain health outcomes.Thus, it may be beneficial to have lower expression of the receptor in a neuroinflammatory condition and pharmaceutical intervention targeting CYSLTR1 may be a valuable option to consider when attempting to prevent HIV encephalitic damage.

Conclusions
In summation, we observed that blockade of CYSLTR1 downstream of p38 MAPK and LTC4S abrogates neurotoxicity of HIV-infected and HIVgp120-stimulated MΦ thus identifying the CysLT pathway as a critical pathological player.Moreover, genetic ablation of either Ltc4s or Cysltr1 in HIVgp120tg mice prevents neurotoxicity and cognitive deficits.This protection is seemingly independent of glial activation, however the leukotriene pathway appears to play a significant yet sexually dimorphic role in astrocytes and microglia.Memory testing results further indicate that LTC4S and CYSLTR1 play a role not only in neuronal injury and survival but also behavioral and cognitive performance.Of note, the drop in hippocampal SYP + neuropil in Cysltr1KO mice as well as the impaired performance of memory of the Ltc4sKO and Cysltr1KO mice indicate that it may not be advisable to completely suppress the CysLT pathway in the absence of neuroinflammatory disease.Finally, post-mortem brain samples of HIV-positive patients with neuropathology had significantly higher levels of CYSLTR1 than patients with no neuropathology, suggesting inflammatory leukotriene receptor signaling in the neuropathogenesis of HIV encephalitis and HAND.These findings also suggest that the CysLT pathway and resulting

Fig. 3 .
Fig. 3. Knock-down of p38 MAPK with siRNA diminishes expression of LTC4S and release of CysLTs in HIVgp120 stimulated monocyte/macrophages. a) Westernblot (WB) of cellular lysates from THP-1 cells treated with siRNA to knockdown p38α MAPK (sip38) or scrambled, non-targeting siRNA (NT).b) Quantification of three independent experiments by WB of which one representative is shown in D. c) Quantification of CysLT by ELISA in tissue culture supernatants of THP-1 cells treated with siRNA for p38 (sip38) or non-targeting (NT) before exposure to recombinant HIVgp120 of two viral strains (SF162 and SF2).Note the down-regulation of LTC4S in panels a and b.Error bars represent the S.E.M. of three independent experiments.* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to control values (ANOVA with Tukey's HSD post hoc test).

Fig. 4 .
Fig.4.(Previous page) Inhibition of CYSLTR1 with montelukast prevents HIV-1 induced macrophage neurotoxicity but not HIV infection of macrophages.a) Montelukast prevents HIVgp120-induced neurotoxicity from human monocytic THP-1 cells.Conditioned media (CM) was collected from THP-1 cells treated with gp120 of HIV-1 SF162 and BaL (each at 1 nM) or 0.001 % BSA-vehicle (Control) for 4 d.Microglia-depleted cerebrocortical cells were then pre-treated with 1 µM montelukast for 30 min prior to incubation with THP-1 CM for 3 d.Neuronal survival was assessed using fluorescence microscopy and cell counting after fixation, permeabilization and immunolabeling of neurons for MAP2 in combination with nuclear DNA staining by H33342.THP-1 Control CM-treated cerebrocortical cells were set as baseline of 100 % for neuronal survival (32.8 % ± 2.9 % neurons [mean ± s.e.m], n = 9 experiments).b) Fluorescence microscopy of cerebrocortical cells exposed to cell-free CM from THP-1 cells +/-montelukast as described above.c -f) Conditioned media from HIV-1 infected macrophages (MΦ HIV CM; 50 % (c, d) or 10 % (e, f)) are neurotoxic but blockade of CYSLTR1 rescues neuronal survival (c, d) and presynaptic terminals (e, f).Microglia-depleted cerebrocortical cells were pre-treated with 1 µM montelukast for 30 min prior to incubation with MΦ HIV CM or MΦ control CM (CM of mock-infected MΦ) for 3 d and evaluated for neuronal survival as described above.Cerebrocortical cells exposed to 10 % MΦ HIV CM showed no loss of neurons but quantification of immunofluorescence staining for Synaptophysin indicated loss of presynaptic terminals which was prevented by Montelukast.g) Blockade of active p38 MAPK (SB203580, 10 µM) but not of CYSLTR1 (Montelukast, 1 µM) inhibits viral replication in HIV-1 infected human MΦ.Cells were cultured in the presence of the inhibitors of DMSO vehicle during the infection and HIV-1 p24 was measured in cell-free CM at 9 dpi.h) HIV-1 infection of MΦ increases release of cysteinyl leukotrienes (CysLT) and the increase is inhibited by blockade of p38 MAPK.HIV infected and MΦ were cultured in the presence of absence of 10 µM SB203580.Controls received DMSO vehicle.CysLT in cell-free MΦ CM collected 24-30 days post infection (dpi) were quantified using an ELISA.Values in graphs represent mean ± s.e.m. of three independent experiments each with 6-8 replicates per condition.** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to control values (ANOVA with Tukey's HSD post hoc test).

Fig. 5 .
Fig. 5. CYSLTR1 and LTC4S are differentially regulated in prefrontal cerebral cortex of HIV-infected individuals with and without brain pathology compared to uninfected controls.a) CYSLTR1 and b) LTC4S are detectable in human prefrontal cerebral cortex RNA samples of HIV-infected individuals with and without brain pathology and uninfected controls.QRT-PCR analysis was performed as described in Methods.Values in graphs represent mean ± s.e.m. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 compared to control values (ANOVA with Tukey's HSD post hoc test).

Table 1
QRT-PCR primers for mouse genes.
Fwd: AGG TCG GTG TGA ACG GAT TTG Rev: TGT AGA CCA TGT AGT TGA GGT CA N.Y.Yuan et al.
Fwd: CAT GTA CGT TGC TAT CCA GGC Rev: CTC CTT AAT GTC ACG CAC GAT