Morphine Disrupts Macrophage Functions Even During HIV Infection

Abstract

HIV-associated Neurocognitive Impairment (HIV-NCI) is a debilitating comorbidity that reduces quality of life in 15–40% of people with HIV (PWH) taking antiretroviral therapy (ART). Opioid use has been shown to increase neurocognitive deficits in PWH. Monocyte derived macrophages (MDM) harbor HIV in the CNS even in PWH on ART. We hypothesized that morphine, a metabolite of heroin, further dysregulates functional processes in MDM to increase neuropathogenesis. We found that, in uninfected and HIV-infected primary human MDM, morphine activates these cells by increasing phagocytosis and upregulating reactive oxygen species (ROS). Effects of morphine on phagocytosis were dependent on μ-opioid receptor activity and were mediated, in part, by inhibited lysosomal degradation of phagocytized substrates. All results persisted when cells were treated with both morphine and a commonly prescribed ART cocktail, suggesting minimal impact of ART during opioid exposure. We then performed mass spectrometry in HIV-infected MDM treated with or without morphine to determine proteomic changes that suggest additional mechanisms by which opioids affect macrophage homeostasis. Using downstream pathway analyses, we found that morphine dysregulates ER quality control and extracellular matrix invasion. Our data indicate that morphine enhances inflammatory functions and impacts additional cellular processes in HIV-infected MDM to potentially increases neuropathogenesis in PWH using opioids.

Graphical Abstract

Introduction

According to the UNAIDS report, in 2020 nearly 38 million people worldwide were living with HIV[1]. During that year, 90% of people with HIV (PWH) had undetectable plasma viral loads due to successful treatment with antiretroviral therapy (ART)[1]. Despite viral suppression, many PWH experience neurocognitive dysfunction termed HIV-associated Neurocognitive Impairment (HIV-NCI)[2]. HIV-NCI is categorized by a spectrum of deficits, including recall memory, motor abilities, and other cognitive abilities that reduce quality of life and increase overall mortality[3]. Approximately 15–40% of PWH have some form of HIV-NCI, and this comorbidity persists in people who are virally suppressed on ART[3–5]. There are no therapies that specifically alleviate these symptoms.

Opioid use increased dramatically in the United States over the past two decades. According to the 2021 National Survey on Drug Use and Health, in 2020 2.7 million Americans had an opioid use disorder (OUD). Approximately 10 million people misused prescription opioid drugs, and over 900,000 people injected heroin in the past year[6]. Heroin is metabolized rapidly into morphine once taken[7]. The annual HIV Surveillance Report from the CDC published in 2020 indicated that nearly 4,000 injection drug users were newly diagnosed with HIV[8]. These people must take ART for the rest of their lives to maintain peripheral and CNS viral suppression[9]. It has been documented since the mid-2000s that opioid use reduces neurocognitive abilities in both PWH and in uninfected individuals[10, 11]. Having HIV increases the onset or extent to which learning and motor abilities are impaired in people actively taking opioids[12]. Some studies found altered immune function in people using opioids, such as heroin, including diminished inflammatory cytokines in response to microbial lipopolysaccharide (LPS) [13, 14]. However, the mechanisms by which opioid exposure contributes to HIV neuropathogenesis in the ART era are incompletely characterized.

Examining the impact of morphine and ART drugs on cellular reservoirs for HIV in the CNS may be critical to developing targeted therapies to reduce neurocognitive deficits in PWH using opioids. After transmigration across the blood-brain barrier (BBB), HIV-infected monocytes can differentiate into long-lived monocyte derived macrophages (MDM) that reside in perivascular regions[15]. HIV-infected MDM release virus that infects other MDM and resident microglia to perpetuate viral seeding[15]. These cell types were shown, in post-mortem tissue studies, to harbor HIV RNA, DNA, and p24 protein in the brains of people who took ART long-term. Limited penetration of ART across the BBB and into myeloid cell reservoirs also perpetuates CNS viral persistence[16].

HIV infection of MDM causes functional dysregulation that contributes to chronic neuroinflammation and CNS damage. The HIV protein, Tat, decreases phagocytic uptake of extracellular materials[17]. The metabolic demands of infected MDM to support active viral replication increase oxidative phosphorylation and production of reactive oxygen species (ROS)[18]. These ROS can causes oxidative damage to healthy lipids and nucleic acids that can trigger cell toxicity in neurons and astrocytes[19, 20]. ROS formation in macrophages is also necessary for killing phagocytized microbes and processing of neurotoxic, inflammatory cytokines including IL-1β, by NLRP3 inflammasomes[21]. Some studies demonstrated increased aerobic respiration when cells are treated with various ART compounds. Additional characterization of the metabolic features of virally suppressed macrophages may be important for improving ART[22, 23].

We hypothesized that exposure of MDM to morphine, even in the presence of ART, causes functional changes that contribute to long-term CNS damage in people with HIV-NCI who use opioids. MDM express all three main subtypes of the opioid receptor, μ, κ, and δ (MOR, KOR, DOR, respectively) isoforms, suggesting that opioids can modulate their functions[24–26]. We cultured and treated uninfected and HIV-infected primary human MDM with morphine, in the presence or absence of ART, for 24 hours, and examined functional changes. Morphine and/or ART increased phagocytosis and enhanced ROS levels regardless of HIV infection. The impact of morphine on phagocytosis was mediated by MOR activity, and morphine decreased degradation of phagocytized substrates by the lysosome. Mass spectrometry was performed on lysates from HIV-infected MDM treated with or without morphine for 24 h to identify proteomic changes that may mediate functional changes. Using pathway analyses, we found that morphine enriches for proteins involved in other pathogenic process. These included Vascular endothelial growth Receptor 2 (VEGFR2) signaling that enhances cell migration, quality control in the endoplasmic reticulum (ER), and translation inhibition caused by viral infection. Western blotting confirmed changes in proteins of these pathways. Specific proteins identified by our analyses might represent novel targets in people with HIV-NCI using opioids to mitigate neurocognitive deficits.

Materials and Methods

General culture methods, HIV infection, and key reagents

We obtained leukopaks from the New York Blood Center and isolated peripheral blood mononuclear cells (PBMC) by Ficoll gradient centrifugation. These leukopaks were deidentified, and we had no information about the donors prior to processing other than blood type. For experiments involving proteomics and Western blotting, PBMC were cultured directly into MDM for 6 days as described previously[27, 28]. We also negatively isolated monocytes from collected PBMC by MojoSort Pan Monocyte Isolation (Biolegend Cat# 480060, San Diego, CA) according to manufacturer’s protocols for phagocytosis and ROS experiments. In both culture strategies, monocytes are not bound to any antibodies or other compounds that may alter their differentiation. Our prior studies of autophagy in uninfected and infected cells treated with morphine demonstrated that MDM respond to opioids similarly using either culture strategy[29].

MDM were infected or not for 24 h with 20 ng/mL HIV_ADA, a CCR5-tropic viral strain adapted from a person with HIV (NIH, Bethesda, MD). Plates were then washed with culture medium containing 10% fetal bovine serum (FBS, Gibco Cat#16000044, Waltham, MA), 5% human AB serum (Corning Cat# 35–060-CI, Corning, NY), 1% penicillin/streptomycin (Gibco Cat# 15140122), 1% glutamine (Gibco Cat# 35050061), and 1% 1M HEPES (Teknova Cat# H1030, Hollister, CA) in Dulbecco’s Modified Eagle Medium (DMEM, Gibco Cat# 11995073), and cultured for 2–3 additional days to propagate infection. We then treated uninfected or infected MDM with 100 nM morphine (MOR, Sigma-Aldrich Cat#6211–15-0, St. Louis, MO) and/or ART (tenofovir 15 μM (Cat# 10199), emtricitabine 15 μM (Cat# 10071), raltegravir 1 μM (Cat# 11680)) for 24 h. ART compounds were obtained from the NIH AIDS Reagent Program (NIH). Chosen morphine concentrations were based on pharmacokinetic studies of average CSF morphine levels in people using heroin[7, 29, 30]. ART concentrations reflect those tested commonly in primary human myeloid cells to achieve viral suppression and examine potential functional changes[31–33]. All drugs were in water.

Measurement of HIV infection

Two hundred μL supernatants were collected from 24-well plates used for microscopy-based phagocytosis assays in infected cells to measure HIV by capsid p24 levels, determined by a p24 alphaLISA (Cat# AL291C, Perkin-Elmer, Waltham, MA). Cells were untreated controls (Untx) or treated with morphine and/or ART for 24 h as described above.

Phagocytosis experiments

For studies in uninfected cells, MDM were cultured on black-bordered, clear-bottom 96-well plates at 100,000 cells/well. MDM were untreated (Untx) or treated with morphine and/or ART for 24 h. In the last 4 h of treatment, untreated and treated cells received 10 ng/mL LPS (Sigma-Aldrich Cat# L2630, St. Louis, MO). After treatments, cells were incubated in carboxylate-modified polystyrene 2.0 μm-diameter latex bead solution at 30 × 10⁶ beads/mL in culture media with 2% FBS for 2 h to allow phagocytosis to occur (Sigma-Aldrich Cat# L4530, St. Louis, MO). Wells were washed, and fluorescent signal outside the cells was quenched in 0.2% trypan blue prior to signal detection using a SpectraMax M5 fluorometer with excitation at 470 nm and emission 505 nm (Molecular Devices, San Jose, CA). Samples were run in sextuplet per treatment group, and values from negative control wells containing beads alone were subtracted as background fluorescence. Normalized values were averaged, and the mean for untreated cells was set to 1.0 in each experiment as a baseline to compare the effects of treatments on baseline and LPS-mediated phagocytosis.

A similar 96-well protocol was used to assess phagocytosis of heat-killed fluorescently labeled E. coli according to the manufacturer’s protocol (Vybrant™ Phagocytosis Kit Cat# V6694, Invitrogen, Carlsbad, CA). Two percent FBS was added to the phagocytosis solution. Trypan blue quenched wells were excited at 494 nm with emission at 518 nm by fluorometer, and values were normalized as described for latex bead phagocytosis. The same assay was used to determine the roles of opioid receptors, MOR and KOR, during phagocytosis. MDM were cultured and treated as described on 96-well plates. Some cells were pre-treated for 30 min with the MOR antagonist, 10 μM naloxone (Nal, Sigma-Aldrich Cat#N7758, St. Louis, MO), or the KOR antagonist, 10 μM nor-Binaltorphimine (Nor-B, abcam Cat#ab120078, Cambridge, UK). MDM were then treated or not with morphine for 24 h as described, including naloxone and Nor-B pre-treated cells. Either opioid antagonist was present or not throughout the 24 h treatment. Phagocytosis experiments were conducted and analyzed as described above.

For studies in HIV-infected MDM, 100,000 cells/well were plated in 24-well plates on glass coverslips. MDM were infected with HIV as described above and then treated or not with morphine and/or ART for 24 h. In the last 2 h of treatment, 1 μM Beta-Amyloid (1–42), HiLyte™ Fluor 488 peptide in phenol-free culture media was added (Anaspec Cat# AS-60479–01, Fremont, CA). This solution was previously agitated at 37 °C for a minimum of one hour. Cells were washed with phosphate buffered saline (PBS), and coverslips were mounted onto glass slides with ProLong Diamond Antifade Mountant with DAPI (Invitrogen Cat# P36962, Carlsbad, CA). After drying, cells were imaged on a Zeiss AxioObserver CLEM, and fluorescent green signal per cell was quantified in ImageJ (NIH). Imaging and quantification were performed by an observer blinded to treatment condition. Cytochalasin D and appropriate DMSO control were tested to demonstrate successful phagocytic suppression upon inhibiting the actin cytoskeleton[34, 35]. Fifty to 75 cells were quantified per treatment, and median fluorescence in untreated cells (HIV Untx) was set to 1.0 in each experiment. Effects of treatment were determined by fold change relative to control.

We performed the same phagocytosis experiments by microscopy in uninfected and HIV-infected MDM to study degradation of phagocytized materials. Prior to the completion of 24 h treatment with morphine and/or ART, 200 μM leupeptin (Sigma-Aldrich Cat# 108975, St. Louis, MO) were added to a set of untreated and treated cells for 2 h to halt lysosomal degradation. This concentration of leupeptin was tested in prior studies to ensure blockage[27, 35]. Fluorescently labeled E. coli were then added for 2 h, and leupeptin treated conditions were co-incubated with E. coli and leupeptin. Coverslips were processed, cells were imaged, and phagocytosis was measured in all 8 treatment conditions as described above. Signal among all conditions was compared, and intracellular lysosomal degradation, or flux, was calculated for each treatment condition by dividing the signal during leupeptin blockage (Leup) by the signal without blockage (Baseline). The same calculations were used to calculate flux of autophagosomes in prior studies[27, 35]. These flux values were then compared to evaluate for differences between groups. Increased flux indicates higher lysosomal degradation, while lower flux indicates reduced degradation.

Reactive oxygen species measurement

For reactive oxygen species (ROS) assessment, uninfected or HIV-infected MDM were cultured as in the fluorometric plate reader phagocytosis assay on 96-well plates. Uninfected or HIV-infected cells were treated with morphine and/or ART for 24 h as above. Ten μM CM-H₂DCFDA (Invitrogen Cat# C6287, Carlsbad, CA) was added for 1 h to stain total ROS followed with incubation in PBS, 0.03% H₂O₂ (uninfected cells) or 30 μM Carbonyl cyanide 3-chlorophenylhydrazone (CCCP, Sigma-Aldrich Cat# 2759, St. Louis, MO) (infected cells) for 1 h. Fluorescence was read by fluorometric plate reader with excitation of 485 nm and emission 520 nm. Negative control wells containing cells without dye were used as background signal subtracted from all treatment groups in each experiment.

Mass spectrometry analysis

HIV-infected MDM from four blood donors were left untreated or treated with morphine for 24 h. Plates were washed with cold PBS three times and lysed with radioimmunoprecipitation (RIPA) buffer. Protein concentrations in the lysate were determined by the Bradford method. Twenty μl of protein were dissolved in 5 mM DTT for 30 min followed by incubation in 20 mM iodoacetamide in the dark. Proteins were then digested using trypsin in the ratio of 1:20 enzyme:sample ratio at room temperature overnight. Digested peptides were eluted, dried, separated, and acquired by a Dionex Ultimate 3000 nano-HPLC coupled online with an Orbitrap Fusion Lumos mass spectrometer (both Thermo Scientific). Spectra were identified using Proteome Discoverer software (v2.4, Thermo Scientific), using SEQUEST as a search engine, and the human protein database downloaded from SwissProt. Our search included methionine oxidation and acetylation of protein N-terminus as dynamic post-translational modifications, up to 2 missed cleavages for trypsin digestion, and a 1% false discovery rate (FDR) as peptide and protein filters. Protein abundances were log₂ transformed, normalized, and missing values were imputed as described previously[36, 37].

We tested log₂ protein levels for statistical significance by paired t test across the four experiments to compare untreated and morphine-treated samples. These values were assumed to be distributed normally. Proteins with an average fold change of ≤ 0.8 (log₂(−0.32)) or ≥ 1.2 (log₂(0.26)) and p-value < 0.1 (log₂(3.32)) with morphine treatment were considered differentially expressed. These cutoffs were based on experiments in primary human monocytes and macrophages that demonstrated that a consistent increase occurs with fold changes ≥ 1.2 relative to control, and a consistent decrease occurs with fold change ≤ 0.8. Upregulated and downregulated proteins in response to morphine were assessed by Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) ifor significant interactions (String Consortium, Switzerland). Using these outputs, we identified cellular and disease related processes enriched with morphine, as determined by Kyoto Encyclopedia of Genes and Genomes (KEGG pathway, Kyoto, Japan), Gene Ontology, WikiPathway, Diseases database (Novo Nordisk Foundation Center for Protein Research, Denmark), and UniProt keyword[38–43].

Western blotting

Lysates in RIPA buffer were collected from uninfected or HIV-infected MDM treated with or without morphine for 24 h. Fifty μg protein per lysate were subjected to gel electrophoresis on 16% gels under reducing conditions and transferred onto either nitrocellulose or polyvinylidene fluoride (PVDF) membranes overnight at 4 °C. Total protein, a validated loading control in our primary human MDM system, was stained and developed by the Li-Cor method, which detects signal in the dynamic range, using an Fc Odyssey (Li-Cor, Lincoln, NE, USA). Optical density (OD) of total protein was quantified per lane for loading control normalization.

Membranes were blocked in 5% nonfat milk in Tris-buffered saline with Tween (TBS-T) for a minimum of 2 h prior to incubation overnight at 4 °C in primary antibody solutions in 5% bovine serum albumin. After washes in TBS-T, appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies were diluted in milk, and membranes were incubated for 1 h. Membranes were washed again in TBS-T, and chemiluminescent signal was developed using SignalWest Femto Chemiluminescent Substrate and Luminol/Enhancer following the manufacturer’s protocol (Cat# 34096, Thermo Fisher Scientific). OD for the appropriate protein was quantified and normalized to the total protein OD in each lane. Signal from the HIV-infected untreated sample (HIV Untx) was set to 1.0 in each experiment, to which effects of treatments were measured by fold change. Primary antibodies used are those for Binding-Immunoglobulin Protein (BiP, 1:1,000, 980 ng/mL, Cell Signaling Technology Cat# 3177, Danvers, MA), Matrix metalloproteinase 14(MMP14)/ membrane type 1 matrix metalloproteinase (MT1-MMP, 1:1,000, 10 μg/mL, Cell Signaling Cat# 13130), Superoxide dismutase 2 (SOD2, 1:1,500, 66.7 ng/mL, Cell Signaling Cat# 13194), and Cathepsin D (1:10,000, ~17.7 ng/mL, Abcam Cat# ab75852, Cambridge, UK). These antibody concentrations were optimized prior to beginning the experiments. Secondary antibody for tested proteins was HRP-fixed goat anti-rabbit antibody (65.7 ng/mL, 1:1,000 (but 1:10,000 for Cathepsin D), Cell Signaling Technology Cat# 7074).

Statistical analyses for functional assays

Results from phagocytosis assays were analyzed by fold change relative untreated cells set to 1.0 due to variability in raw data among primary human cells from different individuals. All quantitative data were analyzed in Prism v.9.0.1 (GraphPad, San Diego, CA). Fold changes were analyzed for normality by Shapiro-Wilk test with a cutoff of p = 0.05. Appropriate one-sample t tests or Wilcoxon signed rank tests were used to analyze individual fold changes for statistical significance relative to control with ɑ = 0.05. Comparisons involving more than two groups were analyzed for significance by one-way analysis of variance (ANOVA), which were followed up by appropriate multiple comparison Dunnett’s or Turkey tests if the data were normally distributed. If not normally distributed, a Friedman test was performed.

Results

Morphine and ART increase phagocytosis and reduce lysosomal degradation of phagocytized materials

Phagocytosis is a major function of macrophages that contributes to HIV pathogenesis in the CNS[44]. We determined in uninfected MDM whether morphine and ART for 24 hours affect phagocytosis of carboxylate-modified, fluorescently labeled latex beads. These are negatively charged particles similar to apoptotic cellular debris in the CNS[45]. By fluorometric plate reader assay, morphine or ART significantly increased phagocytosis relative to control (1.35-fold for ART, 1.28-fold for morphine) (Figure 1A). Early on in HIV infection, disruption of the gut-blood barrier causes leakage of enteric bacterial products into circulation[46]. This process is termed microbial translocation[46]. Thus, LPS is increased systemically in PWH, which prompted us to examine the impact of LPS as well as of morphine and/or ART on phagocytosis in the presence of LPS. Added in the last 4 h of treatment, LPS significantly increased phagocytosis (Figure 1A). Morphine and/or ART with LPS had no additional impact on phagocytosis relative to the LPS alone control, indicating no interactive effect with LPS (Figure 1B).

Figure 1. Morphine increases phagocytosis in uninfected primary human monocyte derived macrophages (MDM) by binding μ-opioid receptors.

Primary MDM were cultured for 6 days with M-CSF on 96-well plates. MDM were untreated (Untx) or treated with morphine (100 nM, MOR) and/or ART (15 μM tenofovir, 15 μM emtricitabine, 1 μM raltegravir) for 24 h prior to phagocytosis of fluorescent latex beads (2.0 μm diameter) or E. coli for 2 h. (A) Fluorescent signal from internalized latex beads in uninfected MDM normalized to the untreated control set to 1.0 in each experiment. (B) Phagocytosis normalized to the LPS (10 ng/mL for 4 h) alone control set to 1.0 in each experiment. (C) Baseline phagocytosis of heat-killed E. coli. (D) E. coli phagocytosis normalized to the LPS alone set to 1.0. N = 9–10, *p < 0.05 **p < 0.01 one-sample t test relative to control. ^##p < 0.01 Wilcoxon signed rank test relative to control. Data are presented as mean ± SEM. (E) Uninfected MDM were pre-treated for 30 min with μ-opioid receptor antagonist, 10 μM naloxone (Nal), or κ-opioid receptor antagonist, 10 μM nor-binaltorphimine (Nor-B), and then treated or not with morphine for 24 h. E. coli phagocytosis assay was performed as in (C-D), including during LPS exposure for 4 h. Fold change in phagocytosis presented elative to Untx set to 1.0. (F) Fold change in phagocytosis relative to LPS alone set to 1.0. N = 4, *p < 0.05 **p < 0.01, one-way ANOVA. ns = not significant. Data are presented as mean ± SEM.

We next assessed how morphine and ART impact phagocytosis of E. coli, a naturally occurring phagocytic substrate[47]. Morphine or ART increased phagocytosis of fluorescently labeled E. coli, 1.76-fold or 1.63-fold, respectively, higher than control. Morphine + ART significantly increased phagocytosis by 1.97-fold. LPS alone enhanced phagocytosis (Figure 1C). In contrast to the latex assays, morphine and/or ART also significantly increased phagocytosis in the presence of LPS relative to the LPS alone control (Figure 1D). Macrophages have specific receptors that recognize LPS on the surface of E. coli, whereas the mechanisms of latex bead uptake are not clearly defined[44, 45]. Results from both assays suggest that morphine and/or ART may further activate MDM by increasing baseline and LPS-mediated phagocytosis.

Involvement of the opioid receptors, MOR and KOR, in the effects of morphine on phagocytosis was assessed using the same fluorometric plate reader E. coli. assay. Some cells were pre-treated with the MOR anatagonist, naloxone (Nal), and others were pre-treated with the KOR antagonist, nor-binaltorphimine (Nor-B), for 30 min. Pre-treated cells were then treated with morphine or not for 24 h followed by the phagocytosis assay. Nal or Nor-B did not significantly change phagocytosis relative to control in the baseline condition (Figure 1E) or to the LPS condition (Figure 1F). During morphine treatment, both Nal and Nor-B appeared to decrease phagocytois relative to morphine alone, although only the difference between morphine and Nal+morphine was statistically significant (Figure 1E). This difference between morphine and Nal+morphine was maintained even in the presence of LPS. These results indicate that MOR is the primary receptor by which morphine mediates its effect on phagocytosis, although there may some contribution by KOR.

We next examined the impact of morphine and/or ART on phagocytosis in HIV-infected MDM. Phagocytosis of aggregated, 1–42 amyloid-β peptide, which accumulates to mediate neurodegeneration in the context of HIV infection as well as in other CNS pathologies, was assessed [48, 49]. We used fluorescence light microscopy to measure phagocytic accumulation of fluorescently labeled 1–42 amyloid-β in response to morphine and/or ART exposure for 24 h (Figure 2A). All drug conditions significantly increased phagocytosis of amyloid-β (Figure 2B). Prior studies suggested that morphine and certain ART cocktails can impair lysosomal function in myeloid cells [69, 70]. Thus, we hypothesized that reduced lysosomal degradation increases accumulation of phagocytized materials. To test this, we performed the same phagocytosis assay in HIV-infected MDM by microscopy. In the last 2 hours of morphine and ART treatment, we added the lysosomal protease inhibitor, leupeptin (L or Leup), to half of treatment conditions. We then added the fluorescent E. coli for 2 h, and leupeptin was added again to the cells pre-treated with leupeptin. This caused significant accumulation of phagocytized E. coli in the control untreated condition (Figure 2D). Lysosomal degradation of E. coli, termed flux, was calculated for each of the four drug treatment conditions. We also collected supernatants from HIV-infected MDM and analyzed them by alphaLISA for HIV p24 to confirm active infection.

Figure 2. Morphine and ART decrease lysosomal degradation of phagocytized materials in HIV-infected MDM.

Primary MDM were cultured and infected with HIV on glass coverslips in 24-well plated followed by treatment with morphine and/or ART for 24 h. Fifty to 75 cells were analyzed per treatment for quantification. (A) Representative images of infected, treated MDM after phagocytosis of aggregated 1–42 amyloid-β for 2 h. (B) Fold change in fluorescent signal of accumulated amyloid-β after phagocytosis relative to HIV-infected, untreated (HIV Untx) cells set to 1.0. (C) HIV p24 levels measured by alphaLISA from culture supernatants of infected MDM treated or not for 24 h during phagocytosis. (D) 2 h prior to 24 h treatment completion, some MDM were treated with 200 μM leupeptin (L), which was also present throughout phagocytosis of fluorescent E. coli for an additional 2 h. Representative images of each condition are shown. (E) Fold change in phagocytosis relative to HIV Untx set to 1.0. (F) Degradation of phagocyotized E. coli calculated as signal with leupeptin divided by signal without leupeptin. N = 3–9, *p < 0.05 **p < 0.01 one-sample t test; ^†p < 0.05 one-way ANOVA. Data are presented as mean ± SEM. Scale bar is 15 μm.

HIV p24 was detected in media from MDM treated with morphine and ART for 24 h but not from uninfected cells. There were no significant differences in HIV p24 levels across all four treatment conditions, suggesting that the effects we observed occur independent of infection (Figure 2C). Similar p24 levels in ART-treated conditions and controls were expected given that our treatment time was shorter than a full HIV replication cycle of 1–2 days [35]. Morphine + ART significantly increased baseline E. coli signal in HIV-infected MDM, supporting data from our amyloid-β studies. There was a trend toward increased signal in morphine-alone treated cells (Figure 2E). Morphine alone also significantly reduced the difference between signal with leupeptin blockage signal and without blockage. This led to significantly reduced lysosomal degradation, or flux, compared to the HIV-infected untreated control (Figure 2F). There was a trend toward decreased flux in MOR+ART treated cells, as well, but not in ART-alone treated cells. Our results suggest that increased phagocytic signal in response to morphine is mediated, at least in part, by reduced intracellular degradation of phagocytized substrates by the lysosome.

Morphine and ART increase ROS in uninfected and HIV-infected MDM

ROS released by macrophages injure other cell types in the parenchyma such as neurons and astrocytes[19, 50]. We tested the impact of morphine and ART on ROS levels in uninfected and HIV-infected MDM by staining untreated and treated cells with a DCF-based dye that couples to total ROS. Fluorescent signal was quantified by fluorometry. In uninfected MDM, morphine with or without ART significantly increased ROS (Figure 3A). We used 0.03% H₂O₂ as a technical positive control. Peroxide itself is an ROS, so we used it in this assay to ensure optimal ROS detection (Figure 3A)[51]. This condition also models higher oxidative stress present in the CNS of PWH with have neurocognitive impairment[19, 23]. When morphine and ART treated MDM were exposed to H₂O_2, there were significantly higher levels of ROS relative to untreated cells that received only H₂O₂. Morphine + ART treated MDM also had significantly higher ROS than MDM treated with ART alone (Figure 3B). This indicates that morphine increases ROS in MDM even in the context of higher baseline ROS.

Figure 3. Morphine and ART increase reactive oxygen species (ROS) in MDM.

(A) ROS in uninfected MDM treated with morphine and/or ART for 24 h relative to the untreated (Untx) control. Peroxide (H₂O₂) for 1 hour was a positive control. (B) ROS in MDM treated with morphine and/or ART for 24 h and stimulated with peroxide relative to untreated cells stimulated with peroxide (Untx+H₂O₂). (C) ROS in untreated HIV-infected MDM (HIV Untx) or treated with morphine and/or ART for 24 h. CCCP for 1 hour was a positive control. (D) ROS in HIV-infected MDM treated with morphine and/or ART for 24 h and stimulated with CCCP relative to untreated cells stimulated with CCCP (HIV CCCP). N = 4–9. *p < 0.05 by one-way ANOVA compared to Untx or HIV Untx. ^✝ p < 0.05 by Friedman test compared to Untx+H₂O₂. ^‡ p < 0.05 compared to ART + H₂O₂ by Friedman test. ^§p < 0.05 compared to HIV ART by one-way ANOVA. ^¶p < 0.05 compared to HIV MOR by one-way ANOVA. Data are presented as mean ± SEM.

We additionally measured ROS levels in HIV-infected MDM treated with morphine and/or ART for 24 h. Infected MDM had overall higher ROS in every treatment condition, as expected (Figure 3C). Morphine, with or without ART, significantly increased ROS levels relative to control (Figure 3C). Interestingly, ROS levels in morphine + ART treated cells were significantly higher than in cells treated with either ART or morphine alone. This indicates a possible additive effect of morphine and ART that did not occur in uninfected cells. Instead of using H₂O₂ in these experiments, we directly stimulated mitochondrial ROS production with 30 μM CCCP, which uncouples the electron transport chain from oxidative phosphorylation.[52] Unlike peroxide treatment, treatment with CCCP models how cells respond to a factor that increases a specific type of oxidative stress during mitochondrial dysfunction. Treatment with CCCP significantly increased ROS in HIV-infected MDM (Figure 3C). When morphine and ART treated MDM were stimulated with CCCP, there was no significant difference in ROS with any treatment relative to CCCP alone (Figure 3D). These results, together with our uninfected cell data, suggest that morphine increases ROS through mechanisms independent of acute mitochondrial stress.

HIV-infected MDM exposed to morphine have altered protein expression

To determine potential mechanisms and additional pathogenic functions of macrophages in the CNS of PWH impacted by opioids, we analyzed the proteomes of HIV-infected MDM treated with or without 100 nM morphine for 24 h by mass spectrometry, n = 4 independent experiments. With average fold change (FC) cutoffs of ≥ 1.2 or ≤ 0.8 (log₂ FC ≥ −0.32 or log₂ FC ≥ 0.26) and p-value cutoff of 0.05 (log₂(4.32)), there were 20 upregulated and 31 downregulated proteins (Supplemental Table S1). We also analyzed proteins with the same fold change cutoffs and a p-value cutoff of 0.1 (log₂(3.32)). By this method, there were 60 upregulated and 73 downregulated proteins (Figure 4A). Using the latter set of differentially expressed proteins, we performed downstream analysis by KEGG pathways, Gene Ontology keywords, WikiPathways, UniProt keywords, and Diseases pathways[39–43]. We also used the STRING database to create an interactome (Supplemental Figure S3).

Figure 4. Morphine for 24 h induces differential protein expression in HIV-infected MDM.

(A) Levels of peptide reads by mass spectrometry were normalized and mapped to the human and HIV proteomes. Differential expression, as determined by paired t test of upregulated proteins, is shown in red, fold change ≥ 1.2 relative to HIV Untx, p < 0.1 (n = 4). Differential expression of downregulated proteins is shown in blue, fold change ≤ 0.8 relative to HIV Untx, p < 0.1 (n = 4). (B) STRING database, Gene Ontology, Wiki Pathway, KEGG Pathway, UniProt Keyword, and Disease-gene associations analyses of differentially expressed proteins, FDR < 0.05. (C) Representative Western blot for MMP14 (also called MT1-MMP) involved in VEGFR2 signaling and quantifications relative to HIV Untx set to 1.0 in each experiment. (D) Representative Western blot for BiP involved in ER quality control and quantifications relative to HIV Untx set to 1.0 in each experiment. *p < 0.05 **p < 0.01 one-sample t test, n = 10. Data in C and D are presented as mean ± SEM.

With an FDR < 0.05, we found that morphine treatment enriches for proteins involved in metabolic pathways and metabolic diseases (Figure 4B). For instance, Mitochondrial contact site and cristae organizing system (MICOS) complex subunit, MIC26, maintains mitochondrial structure, Nicotinamde adenine dinucleotide + hydrogen (NADH) dehydrogenase [ubiquinone] flavoprotein 1 participates in oxidative phosphorylation, and Delta-1-pyrroline-5-carboxylate synthetase is important for amino acid synthesis[53–55] (Supplemental Table S1). Changes in metabolism may mediate the increased ROS that we measured in response to morphine treatment (Figure 3). Morphine also significantly altered levels of three ER proteins, calnexin, protein sel-1 homolog 1, and F-box protein 6, which participate in quality control (Supplemental Table S1)[56–58]. Alterations in ER protein composition and quality may dysregulate proteostatic processes to contribute to amyloidosis, which was also enriched (Figure 4B). There were also changes in four different eukaryotic initiation factors that cause inhibited host cell translation by SARS-CoV-2, suggesting that morphine can modify host-viral interactions in human MDM (Figure 4B, Supplemental Table S1)[59]. Morphine also enriched for proteins involved in VEGFA-VEGFR2 signaling, which is associated with endothelial cell proliferation and cancer metastasis (Figure 4B)[60]. Proteins in this pathway may impact cell migration and phagocytic uptake of extracellular substrates as a novel mechanism of opioid induced dysregulation macrophage functions during HIV neuropathogenesis.

Validation of proteomics by Western blotting

We confirmed by Western blotting that morphine treatment alters expression of certain proteins that participate in the pathways identified by our proteomics analyses. Because we found that VEGFA-VEGFR2 signaling was upregulated, we quantified MMP14, a matrix metalloproteinase in this pathway that was increased. MMP14 breaks down collagen and activates other key MMPs such as MMP2[61]. Proteolysis degrades the extracellular matrix to facilitate cell migration[62]. This could facilitate movement of HIV-infected macrophages within the CNS to disseminate viral particles, toxic viral proteins, inflammatory mediators, and ROS. MMP14 is not secreted as a pro-enzyme that requires downstream activation and is active once it is membrane-bound[63]. We assessed levels of MMP14 in uninfected untreated cells, HIV-infected untreated cells (HIV Untx), and HIV-infected cells treated with morphine for 24 h. Relative to HIV-infected cells, uninfected MDM had a trend toward decreased MMP14. Morphine treatment significantly increased MMP14 levels by 1.35-fold compared to infected control, which aligned with our mass spectrometry data (Figure 4C, Supplementary Table S1).

Our proteomic pathway analyses indicated that ER quality control was significantly impacted by morphine treatment (Figure 4B). We measured levels of the chaperone, BiP, which is involved extensively in regulating misfolded proteins in the ER[56]. Prior studies demonstrate that regulation of proteostasis within the HIV-infected CNS is a critical to neuropathogenesis[48, 49, 64]. By mass spectrometry, BiP was one of the top most abundant proteins detected across all samples (Supplementalery Materials). As a result, it could be difficult to detect to a potential increase in BiP. Given the importance of BiP as a marker of ER stress, we also determined changes in BiP by Western blotting[65]. This method could detect greater variation in BiP levels among different experiments, perhaps due to recognition of an epitope that was not detected by mass spectrometry, with an increased sample size. HIV infection had no impact on BiP expression relative to uninfected samples. Morphine treatment of infected MDM, however, significantly increased BiP relative to HIV Untx (Figure 4D). This aligns with our analyses suggesting that morphine dysregulates ER quality control (Figure 4B).

Discussion

We demonstrated that morphine and ART alter vital functions of monocyte derived macrophages (MDM), which contribute to neuroinflammation and long-term CNS injury in people with HIV-NCI[50]. In uninfected cells, morphine or ART increased phagocytosis of negatively charged latex beads, and morphine and/or ART increased phagocytosis of heat-killed E. coli (Figure 1A–1D). These effects were depenent on MOR activity, and KOR may contributre as well (Figure 1E, 1F). In HIV-infected MDM, morphine and/or ART increased phagocytosis of amyloid-β (Figure 2B). Prior studies in murine macrophages demonstrated that morphine decreases F_c-mediated phagocytosis from 1 to 18 hours of treatment at several concentrations[34, 66]. One study in human MDM, however, did find that F_c-mediated phagocytosis returned to baseline after 24 h morphine treatment[67]. Our assays also did not specifically measure F_c-mediated phagocytosis but rather those mediated by pattern recognition receptors or non-specific engulfment. Morphine and ART may modify phagocytosis depending on the timing and mode of substrate recognition.

ART alone also increased phagocytosis in both uninfected and HIV-infected MDM (Figure 1). We tested a combination of tenofovir, emtricitabine, and raltegravir because it represents a commonly prescribed regimen to treat HIV infection[9]. The off-target effects in MDM of these three drugs outside the context of HIV replication are not extensively characterized and may contribute to comorbidities associated with long-term HIV infection. Tenofovir and emtricitabine can inhibit DNA polymerase Ɣ, which helps replicate mitochondrial DNA[9]. This may cause downstream changes in metabolism that modify activation. M1 macrophages, considered more inflammatory and respond robustly to foreign pathogens, rely more on anaerobic metabolism through the HIF-1ɑ pathway[68]. Our results indicate that these regularly prescribed drugs may have off-target effects on cell types that are targets for HIV infection.

During phagocytosis, extracellular substrates are recognized and internalized into vesicles that mature through fusion with lysosomes. Lysosomal enzymes facilitate macromolecule degradation and nutrient recycle[44]. The only substrates we tested that cannot be degraded by lysosomes were latex beads[45]. Thus, morphine and ART likely increase substrate uptake. We also assessed phagocytosis of E. coli by microscopy in HIV-infected cells treated with morphine and/or ART. To some cells, we added leupeptin, which inhibits lysosomal proteases and degradation of phagocytized E. coli, during phagocytosis. Using these fluorescence intensities, we calculated lower lysosomal degradation in morphine-treated cells compared to control (Figure 2F). Studies in primary human microglia treated for 24 h with 500 nM morphine (5-fold higher than in our studies) demonstrated decreased degradative capacity[69]. Other studies of rat microglia treated with tenofovir, emtricitabine, and dolutegravir showed similarly compromised lysosomes[70]. Preliminary data from our experiments in uninfected cells indicate the same defect in substrate degradation that suggests disrupted lysosomal function (Supplementary Figure S2). Interestingly, ART alone did not appear to impact lysosomal degradation, suggesting that increasd phagocytosis may be mediated by more so by increased uptake (Figure 2F).

By proteomics, there was a potential increase (average fold change: 1.31, p = 0.058) in vinculin with morphine (Supplementary Table S1). This factor regulates actin rearrangements essential for vesicular internalization[71]. Morphine also decreased phospholipase C-Ɣ 2 and phospholipase C-δ 1 (Supplementary Table S1). These factors are activated by G_q-associated G-protein coupled receptors, among other signaling modes, to increase intracellular calcium[72]. Future studies will examine how signaling changes caused by morphine may contribute to decreased activation of the IP₃ pathway by phospholipase C[72]. Mediators of these cascades may be novel targets to ameliorate macrophage phagocytic activation in PWH using opioids.

Morphine significantly increased ROS (Figure 3). However, in our study we measured total ROS and not mitochondrial ROS, specifically. Metabolic pathways and disease of metabolism were enriched with morphine treatment of HIV-infected MDM by proteomics. The pathways that relate most closely to ROS levels are those involved in aerobic respiration(Supplementary Figure S3). Protein components of NADH dehydrogenase (part of complex I) were shown to decrease in PBMC from PWH stably suppressed on ART[54]. Expression of HIV Tat was shown to inhibit cytochrome c oxidase (part of complex IV) in the brain[73]. Changes in ETC proteins could be caused by compromise of mitochondrial structural integrity. Morphine increased expression of MICOS subunit MIC26 and Sorting and assembly machinery component 50 homolog (SAMM50), which maintain cristae morphology essential for proper ETC function and shunting of metabolites (Supplementary Figure S3)[74, 75]. Other potential sources of increased ROS include phagocyte oxidative bursts and reduced ROS neutralization[21, 76].

Morphine significantly increased expression of CD38, a key enzyme that creates Nicotinic acid adenine dinucleotide phosphate (NAADP) using NAD⁺ (Figure 4A). This is implicated in metabolic aging since cells continuously need NAD⁺ reserves to support glycolysis and the citric acid cycle[77]. As such, morphine may enhance neurocognitive deficits in opioid users with HIV-NCI by accelerating aging processes. CD38 on macrophages also binds CD31 on T cells to promote inflammatory cytokine release, and therefore, is a major marker of the M1 macrophage subtype[78]. It was also shown to activate macrophages to enhance phagocytosis and increase autophagy[77]. Thus, CD38 may mediate macrophage activation that dysregulates phagocytosis, ROS, and autophagy after morphine treatment.

By proteomics, we found that morphine modifies other pathways that may contribute to functional dysregulation in the context of HIV. One finding was increased VEGFA/VEGFR2 signaling, which is important for tumor cell invasion, angiogenesis, and metastasis[60]. Some studies indicate that VEGFR2 is expressed on macrophages in the context of disease processes such as tumor development[60, 61]. VEGFR2 signaling increases expression and activation of enzymes that degrade the extracellular matrix, including A Disintegrin and metalloproteinase domain-containing protein 10 (ADAM10) and MMP14, which were both increased significantly by morphine (Supplementary Table S1)[62]. MMP14 is also a marker of advanced cancer disease and higher mortality when expressed by tumor-associated macrophages[61]. Our studies suggest future examination of post-mortem tissue from opioid users with HIV-NCI to correlate macrophage expression of MMP14, and other metalloproteinases, with known markers of CNS damage. As a control, we measured by Western blotting levels of two proteins that were unchanged by mass spectrometry. Consistent with the proteomics, in HIV-infected MDM, morphine did not change expression of the mitochondrial superoxide dismutase or cathepsin D (Supplementary Table S1, Supplementary Figure S1).

Regulation of ER quality control was also disrupted by morphine treatment (Figure 4B). There was a significant decrease in sel-1 homolog 1, which mediates the degradation of ER proteins in the ER-associated degradation (ERAD) pathway[58]. F-box only protein 6, which is specific to degradation of luminal ER proteins, was also differentially expressed[56]. Through these mechanisms, morphine may cause significant retention of ER proteins that are misfolded and dysfunctional. This is compounded by an increase in calnexin, which also retains incorrectly folded proteins in the ER (Supplementary Table S1)[57]. Additionally, we found by Western blotting that morphine significantly increases ER stress marker, BiP (Figure 4D)[57]. Morphine significantly increased levels of Reticulophagy regulator 3, which mediates selective autophagic degradation of ER components, termed ER-phagy, to maintain homeostasis (Supplementary Table S1)[57]. Future studies will examine the functional impact of dampening ER stress by increasing levels of key chaperones[70].

Overall, our data suggest that morphine and ART treatment for 24 h activates macrophages regardless of HIV infection, as indicated by enhanced phagocytosis and higher ROS. Our proteomics data demonstrate ER stress and cell migration as additional pathways that may contribute to this activation. We did not assess differential protein expression in the context of ART since ROS and phagocytosis were not as significantly impacted by ART treatment alone in infected cells (Figure 2, Figure 3). ART adherence can also be affected significantly in PWH using substances, underscoring the need to study the impact of opioids alone on HIV-infected reservoir cells[11]. Future studies will address the effects of opioids on proteomic changes during concomitant ART treatment to model PWH taking opioids who also take daily ART as prescribed.

Like our culture system, HIV-infected macrophages in the CNS exist as a heterogeneous population. Some cells harbor actively replicating virus, while others do not harbor HIV but are exposed to viral proteins and inflammatory mediators [15]. Our system reflects opioid treatment of both HIV-harboring and HIV-exposed cells in PWH taking opioids. We examined HIV-infected MDM treated with or without morphine for 24 h because it models what occurs in PWH taking opioids for one day. This treatment timing models a person with HIV who then develops opioid use disorder and takes opioids and ART daily. In future studies, we will determine whether these effects persist with more chronic exposure to morphine by treating MDM daily for 7 days. This will contribute to understanding of how opioids influence macrophage homeostasis in PWH using these drugs on a longer-term basis. Specific proteins identified by our proteomics may be important targets for interventional strategies to mitigate neurocognitive decline in opioid users with HIV-NCI.

Supplementary Material

1

Supplementary Table S1: Differentially expressed proteins with morphine treatment of HIV-infected MDM that were statistically significant.

Supplementary Figure S1: Western blotting and quantifications of proteins found to be unchanged by proteomic analysis.

Supplementary Figure S2: Microscopy-based analysis of lysosomal degradation of phagocytized E. coli in uninfected MDM.

Supplementary Figure S3: Interactome generated from the STRING database demonstrating significantly enriched pahways by proteomic analysis.

Abbreviations

HIV

Human Immunodeficiency Virus

PWH

People with HIV

CNS

Central Nervous System

CDC

Center for Disease Control

HIV-NCI

HIV-associated Neurocognitive Impairment

ART

Antiretroviral Therapy

MDM

Monocyte Derived Macrophage

BBB

Blood-Brain-Barrier

ROS

Reactive oxygen species

ER

Endoplasmic Reticulum

OUD

Opioid Use Disorder

MOR

Morphine or μ-opioid receptor

KOR

κ-opioid oeceptor

DOR

δ-opioid receptor

PBMC

Peripheral Blood Mononuclear Cells

Untx

Untreated

LPS

Lipopolysaccharide

ETC

Electron Transport Chain

MICOS

Mitochondrial contact site and Cristae Organizing System

ERAD

ER-Associated Degradation

CCCP

Carbonyl Cyanide M-chlorophenyl Hydrazone

L or Leup

Leupeptin

Nal

Naloxone

Nor-B

nor-Binaltorphimine

STRING

Search Tool for the Retrieval of Interacting Genes/Proteins

KEGG

Kyoto Encyclopedia of Genes and Genomes

PVDF

Polyvinylidene fluoride

VEGF

Vascular endothelial growth receptor

DMEM

Dulbecco’s Modified Eagle Medium

BiP

Binding-immunoglobulin Protein

MMP

Matrix Metalloproteinase

NADH

Nicotinamde adenine dinucleotide + hydrogen

NAADP

Nicotinic acid adenine dinucleotide phosphate

ADAM

A Disintegrin and metalloproteinase domain-containing protein