HIV Neuropathology in the Context of Opioid Use Disorder – Clinical and Preclinical Evidence

Preclinical and Clinical Findings—a Complicated Picture
People infected with HIV (PWH) with OUD have an increased incidence of neuroHIV and CNS complications (Bell et al. 1998; Nath et al. 1999, 2000a, 2002; Anthony et al. 2008; Meyer et al. 2013; Smith et al. 2014). Injection drug use increases the probability of contracting HIV (Nath et al. 1999) and opioid drugs intrinsically alter the pathogenesis of HIV. PWH who develop intractable pain syndromes related to peripheral neuropathies often receive opioid drugs for treatment (Mirsattari et al. 1999; Denis et al. 2019). PWH who misuse opioids are more likely to undertake risky sexual behavior and are less likely to adhere to combined antiretroviral (ARV) therapy (cART) regimens (Lemons et al. 2019). Opioid receptors are widely expressed on immune cells and opioids can modulate immune function (Donahoe and Falek 1988; Plotnikoff 1988; Rouveix 1992; Adler et al. 1993; Carr and Serou 1995; Carr et al. 1996; Sheng et al. 1997; Banerjee et al. 2011; Purohit et al. 2012), which typically (but not always) result in immune suppression (Wybran et al. 1979; McDonough et al. 1980, 1981; Donahoe and Falek 1988; Donahoe et al. 1991; Falek et al. 1991; Novick et al. 1991; Chao et al. 1996a; Peterson et al. 1998; Rogers and Peterson 2003; Stein et al. 2003; Roy et al. 2006; Rittner et al. 2008). The “opiate cofactor hypothesis” proposes opioids contribute directly to the pathogenesis of acquired immune deficiency syndrome (AIDS) (Donahoe and Vlahov 1998), in part, because MOR activation can increase HIV replication in immune cells (Peterson et al. 1990, 1992, 1993, 1999; Ho et al. 2003). Furthermore, MOR and HIV co-receptors, including both CCR5 (El-Hage et al. 2013; Yuan et al. 2013; Arnatt et al. 2016) and CXCR4 (Pitcher et al. 2014) can interact via convergent downstream signaling and perhaps via direct molecular interactions (Rogers et al. 2000; Rogers and Peterson 2003; Steele et al. 2003; Chen et al. 2004; Song et al. 2011; Arnatt et al. 2016). MOR-CCR5 or CXCR4 interactions are highly contextual and can promote (Guo et al. 2002; Steele et al. 2003) or inhibit (Strazza et al. 2014) HIV expression, depending on the nature and duration of exposure (see Fig. 9; Berman et al. 2006) and cell type involved (Kim et al. 2018). Depending on the outcome measure, Tat expression reduces morphine’s efficacy and potency (Fitting et al. 2012, 2016; Hahn et al. 2016). Antagonizing CCR5 with maraviroc reinstates morphine potency in an antinociceptive assay and restores physical dependence in Tat exposed, morphine-tolerant mice (Gonek et al. 2018).
Epidemiological studies suggest OUD can increase AIDS progression (Donahoe and Vlahov 1998; Dronda et al. 2004; Meijerink et al. 2014, 2015). In the pre-cART era, opiate abuse was found to exacerbate HIV encephalitis (HIVE) (Bell et al. 1998, 2002). In Indonesian injection heroin abusers who lacked access to cART, CD4 counts (a measure of HIV progression) were reduced compared to PWH not using heroin (Meijerink et al. 2014). However, with the introduction of cART, the clinical picture has significantly changed with a 50% decline in the rate of death from AIDS, reduced incidence of opportunistic infections and HIVE, and a 40–50% decrease in the incidence of HIV-associated dementia (HAD), the most severe form of HIV-associated neurocognitive disorders (HAND) (Maschke et al. 2000; McArthur et al. 2010; Saylor et al. 2016). Nevertheless, chronic opiate exposure (which almost always is confounded by the use of other illicit and legal drugs) in PWH can worsen neuroHIV (Anthony et al. 2005; Bell et al. 2006; Anthony et al. 2008) and cognitive impairment (Rodriguez Salgado et al. 2006; Martin-Thormeyer and Paul 2009; Byrd et al. 2011; Smith et al. 2014; Martin et al. 2018; Rubin et al. 2018) despite cART, even though some studies fail to show that opioids worsen neuroHIV (Royal et al. 1991; Applebaum et al. 2010) or HAND (Martin et al. 2019). Opiate exposure in cART-treated PWH worsens CD4 counts and viral loads (Ryan et al. 2004), neuropathology (including increased tauopathy; Smith et al. 2014), CNS inflammation (Anthony et al. 2005, 2008; Smith et al. 2014), and neurocognition (Applebaum et al. 2009; Byrd et al. 2011; Meyer et al. 2013) including deficits in memory and working memory (Byrd et al. 2011). Table 1 gives an overview on reported interactive effects of HIV and opioids in some of the clinical and preclinical CNS studies referenced in this review.
Table 1 Clinical and preclinical findings
Major effects HIV pathogena ARV Opioids Outcome Model system Citation(s)
Clinical findings (human)
HIV progression and/or ARV adherence HIV cART • SUD
• Prescription opioids for pain • ↑ Viral load with SUD
• ↓ ARV adherence
• ↑ Frequency of prescription drugs with pain + SUD Human (Denis et al. 2019)
HIV cART OUD • ↓ Lasting viral suppression
• ↓ Adherence to cART for 3 years Human (Lemons et al. 2019)
HIV ARV naive Injection drug use ↓ CD4 counts Human (Meijerink et al. 2014)
HIV encephalitis (HIVE)
HIV infection CNS HIV ZDV Former drug use (+ OST) • ↑ Multinucleated giant cells
• ↑ HIV p24 Human, postmortem brain (Bell et al. 1998)
Microglial activation HIV • ARV
• ZDV OUD ↑ CD68 microglial activation only in non-OUD HIV+ PWH Human, postmortem brain (Smith et al. 2014)
HIV • ARV
• ZDV, other monotherapies Injection drug use (+ OST) ↑ Microglial activation Human (Bell et al. 2002)
HIV No info Drug use • ↑ MHC class II
• ↑ CD68 Human, postmortem brain (Anthony et al. 2005)
HIV No info OUD (44% methadone, 36% other opiates) • ↓ CD68, HLA-D in HIV and HIVE with OUD
• No effect of IDU on CD68 Human, postmortem brain (Byrd et al. 2012)
Plasma cytokines HIV cART OUD (codeine, fentanyl, morphine) ↑ sTNF-R2, not sCD14, TNF-α, sTNF-R1, in plasma Human (Ryan et al. 2004)
HIV ARV naive Reported heroin use • ↓ MIP-1α, MIP-1β, MCP-2 in blood after stimulation with LPS
• ↑ CCR5 expression in CD4 cells Human (Meijerink et al. 2015)
HIVE HIV No info OUD • ↑ Parenchymal inflammatory infiltrates
• ↑ HIV PCR amplification products Human, postmortem brain (Gosztonyi et al. 1993)
Aberrant immune responses HIV No info SUD (opioids, alcohol, marijuana, cocaine) (+ OST) • ↑ Autoantibodies and delayed hypersensitivity to neural antigens OUD only
• No HIV effect/interaction Human (Jankovic et al. 1991)
Learning-memory HIV 50-70% on cART Heroin, crack/cocaine • ↓ Total learning; ↓ Learning slope
• ↓ Delayed recall Human, female (Meyer et al. 2013)
HIV cART Reported heroin use • ↓ Recall memory
• ↓ Working memory Human (Byrd et al. 2011)
HIV No info SUD (opioids, alcohol, marijuana, cocaine) • ↓ Complex figure copy
• ↓ Delayed recall Human (Concha et al. 1997)
Neuropsychological performance cART OST (methadone) No effect of OST Human (Applebaum et al. 2010)
Cognitive function HIV cART OUD • ↓ Cognitive performance with anticholinergics, but not opioids, anxiolytics, or anticonvulsants Human (Rubin et al. 2018)
Memory
Cognitive function HIV cART SUD (alcohol, cocaine, heroin) • ↓ Working memory in HIV+
• ↓ Spatial and verbal response times in women, irrespective of HIV status
• ↑ Response time with cocaine use Human (Martin et al. 2018)
Visual and cognitive function HIV No info OUD (+ OST, methadone) • ↑ Pattern-shift visual evoked potential delay with methadone
• No HIV effect/interaction Human (Bauer 1998)
Transmission risk HIV No info OST ↓ Frequency of injection drug use Human (Kwiatkowski and Booth 2001)
HIV cART OST • ↓ Frequency of heroin injection
• ↑ On ARV Human (Pettes et al. 2010)
Motor and visual function HIV No info OST • ↓ Digital Finger-Tapping test
• ↓ Visual motor pursuit Human (Silberstein et al. 1993)
ARV adherence HIV cART OST • ↑ ARV adherence in PWH with OST vs. OUD Human (Mazhnaya et al. 2018)
PENK expression HIV Pre- and post-cART SUD • ↓ PENK in HIVE vs. HIV−
• ↓ DRD2L HIV+ vs. HIVE & HIV−
• ↓ DRD2L correlates with
↑ cognitive performance Human, post mortem brain (Gelman et al. 2012)
OPRM1 polymorphisms, splice variants HIV No info SUD C17T MOR polymorphism correlates with ↑ risk of cocaine, alcohol & tobacco (but not opiate) use Human (Crystal et al. 2012)
HIV cART No Some OPRM1 polymorphisms may alter HIV severity / response to ARV Human (Proudnikov et al. 2012)
HIV No info MOR-1K expression • ↑ MOR-1K in HIVE
• ↑ CCL2, CCL6, CCL5, but not CXCR4, CCR5 or CD4 receptor in HIVE Human, postmortem brain (Dever et al. 2014)
OPRK and PDYN polymorphisms HIV cART No Some OPRK and PDYN polymorphisms may alter HIV severity / response to ARV Human (Proudnikov et al. 2013)
Sensory Neuropathy HIV cART SUD HIV sensory neuropathy- regardless of SUD (trends, not significant) Human (Robinson-Papp et al. 2010)
Preclinical in vivo findings (animal)
HIV entry into the brain Mixture of SIV17-EFr, SHIVKU_1B, SHIV89.6P No Morphine (5 mg/kg i.m., b.i.d., ≤ 56 weeks) • ↑ CSF viral load
• ↑ Viral migration through BBB for SHIVKU Rhesus macaques (Kumar et al. 2006)
SIVmacR71/17E No Morphine (3 mg/kg i.m., q.i.d.) • ↑ CD4+ and CD8+ T cells
• ↑ CSF viral load
• ↑ Infiltration of MDMs into the brain Rhesus macaques (Bokhari et al. 2011).
Viral load and HIV progression Mixture of SIV17-EFr, SHIVKU _1B, SHIV89.6P No Morphine (5 mg/kg, i.m., t.i.d., 20 weeks) • ↑ Viral load; ↓ CD4 counts
• ↑ ROS with morphine + SIV Rhesus macaques (Perez-Casanova et al. 2007; Perez-Casanova et al. 2008)
SIV gene mutation/evolutiontat Mixture of SIV17-EFr, SHIVKU _1B, SHIV89.6P No Morphine
(5 mg/kg, i.m., t.i.d., 20–56 weeks) • ↑ Viral load; ↓ CD4 counts
• tat evolution—inverse correlation with SIV progression
• ↓ tat diversity with morphine Rhesus macaques (Noel and Kumar 2006; Noel et al. 2006b)
nef • ↑ Viral load; ↓ CD4 counts
• ↓ nef evolution; no correlation with SIV progression ± morphine (Noel et al. 2006a)
env • ↑ Viral load; ↓ CD4 counts
• ↑ env evolution (V4 region) correlates with SIV progression + morphine
• ↑ env evolution in CSF with morphine (Rivera-Amill et al. 2007, 2010b)
vpr • ↓ vpr evolution and/or Vpr R50G mutation—inverse correlation with SIV progression/mortality
• ↓ vpr evolution with morphine (Noel and Kumar 2007; Rivera et al. 2013)
Neuronal injury, survival, oxidative stress gp120 HIV-1LAV No Morphine (25 mg pellet, 5–7 days) • ↑ ROS during withdrawal
• ↓ PSD95 during chronic and withdrawal
• ↑ Sphingomyelin
• ↓ Ceramide Mouse, gp120 tgb (Bandaru et al. 2011)
HIV No Morphine (37.5 mg s.c, 5 days) ↓ neuron survival HIV tg + morphine Rat, HIV-1 tg, female (Guo et al. 2012)
SIV
HIV Tat No Morphine (3 mg/kg i.m., q.i.d., 3 weeks) • ↑ miR-29b, ↓ PDGF-B mRNA, ↑ PDGF-BB with morphine and SIV
• ↓ PDGF-B, ↓ neuron survival with CM from morphine-treated astrocytes Rhesus macaques; Ratb, primary neurons, astrocytes (Hu et al. 2012)
Synaptic transmission Tat1–86 No Morphine ex vivo (1 μM) to the bath ↓ mIPSC frequency Mouse, male and female, PFC slices, ex vivo (Xu and Fitting 2016)
SIVmacR71/17E
Tat No info • Morphine (escalating doses of 1–3 mg/kg i.m., q.i.d., 12 months)
• Morphine in vitro • SIV ↑ Synaptic protein HSPA5
• Tat ↑ HSPA5 mRNA (in vitro) Rhesus macaques;
Human,
SH-SY5Y neuroblastoma cells in vitro (Pendyala et al. 2015)
White matter effects SIVmacR71/17E No Morphine (3 mg/kg i.m., q.i.d., ≤ 59 weeks) • ↑ Focal, demyelinating lesions
• ↑ Macrophages in areas of myelin loss Rhesus macaques (Marcario et al. 2008),
CNS metabolites SIVsmm9 No info Morphine (escalating doses of 1–3 mg/kg i.m., q.i.d., ≤ 4 years) • ↑ Survival time
• ↑ Creatine in white matter (SIV + morphine only)
• ↑ Myo-inositol in putamen Rhesus macaques (Cloak et al. 2011)
Neuroinflammation Tat1–86 No Morphine (10 mg/kg i.p., b.i.d., 5 days) ↑ Iba1+ 3-NT+ microglia Mouse, Tat tg, males (Zou et al. 2011)
Chemokines Tat1–72
(25 μg intrastriatal injection) No Morphine (25 mg pellet, 5 days) • ↑ CCL2 in astrocytes is regulated by CCR5
• ↑ CCL2 in macrophages/microglia
• CCL2-knockout blocks morphine + Tat-induced glial reactivity Mouse (El-Hage et al. 2008a)
Cytokines, Chemokines HIV Tat (10 μg/kg i.v.) No Morphine (25, 75 mg pellet, 6 days) • Morphine ↑ death in Tat + bacterial infection
• ↑ TNFα, IL-6, CCL2,
• ↑ TLR2, TLR4, TLR9 Mouse, male, in vivo; microglia in vitro (Dutta et al. 2012)
MOR expression HIV-1IIIB gp120 (X4) No MOR ↑ MOR mRNA Rats, HIV-1 tg males (Chang et al. 2007)
MOR-coupling efficacy to G proteins Tat1–86 No • Morphine (acute, 10 mg/kg i.p.)
• Morphine, DAMGO (ex vivo) ↓ [35S]GTPγS binding in NAc Shell, CPu, amygdala, PFC, but not hippocampus, with morphine in Tat mice Mouse, Tat tg, males (Hahn et al. 2016)
Neuroinflammation; morphine tolerance (antinociception), physical withdrawal, reward Tat1–86 No Morphine (75 mg pellet, 5 days) • ↑ Tolerance (↓ anti-nociceptive potency and ↓ withdrawal symptoms)
• ↑ CPP and cytokines (24 h after withdrawal)
• Above effects reduced by CCR5 blockade Mouse, Tat tg, males (Gonek et al. 2018)
Neuropathy gp120 (0.2 μg), q.d. intrathecally No Morphine (3 μg, intrathecally, b.i.d., 5 days) • ↑ Mechanic allodynia
• ↑ Brd4 mRNA Rat, males, gp120 (Takahashi et al. 2018)
Morphine efficacy, potency Tat1–86 No Morphine (acute, 2–8 mg/kg s.c.) ↓ Antinociceptive potency and efficacy (tail flick) Mouse, Tat tg, males (Fitting et al. 2012)
Morphine tolerance, physical dependence Tat1–86 No Morphine (75 mg pellet, 4 days) • ↑ Antinociceptive tolerance
• ↓ Physical dependence Mouse, Tat tg, males (Fitting et al. 2016)
Locomotor function Tat1–86 No Oxycodone (0–10 mg/kg, i.p., 15 min prior behavioral assay) ↑ Locomotor activity, center entries (open field) Mouse, Tat tg, females (Salahuddin et al. 2020)
SIVmacR71/17E No Morphine (escalating doses of 1–2.5 mg/kg i.m., q.i.d., 59 weeks) ↓ Motor skill Rhesus macaques (Marcario et al. 2016)
Tat1–86 No Oxycodone (acute, 0.1–10 mg/kg, i.p.) ↑ Psychomotor effects Mouse, Tat tg, females (Paris et al. 2020)
BBB integrity Tat No Morphine (25 mg pellet, 5 days) ↑ Dextran extravasation across the blood-brain barrier Mouse, Tat tg females (Leibrand et al. 2019)
Immune cell trafficking into CNS Tat No Morphine • ↑ Infiltration of monocytes and T cells into S. pneumoniae-infected CNS with morphine
• ↑ T cell CXCR4 and CCR5 expression with morphine Mouse, CNS infection (S. pneumoniae), males (Dutta and Roy 2015)
ARV accumulation Tat DTG
ABC
3TC Morphine (2 mg/day, s.c.. osmotic pump, 5 days) ↓ Dolutegravir and abacavir, but no change in lamivudine in brains of morphine-treated animals Mouse, Tat tg females (Leibrand et al. 2019)
Circadian rhythms Tat1–86 No Morphine (25 mg pellet, last 5 days) ↓ Total wheel-running activity Mouse, Tat tg, males (Duncan et al. 2008)
aassumed Clade B, unless noted otherwise; b sex not reported; c authors reported a trend that was not significant
ABC, abacavir; ARV, antiretroviral(s); BBB, blood-brain barrier; b.i.d., twice a day; Brd4, Bromodomain-containing protein 4; CPu, caudate-putamen; CNS, central nervous system; CPP, conditioned place preference; CM, conditioned medium; CSF, cerebrospinal fluid; DAMGO [D-Ala2, N-MePhe4, Gly-ol]-enkephalin; DRD2L, type 2 dopamine receptor; DTG, dolutegravir; HIVE, HIV encephalitis (typically seen pre-cART); HSPA5, heat shock 70-kDa protein A 5; IDU, injection drug use; i.m., intramuscularly; i.p., intraperitoneal; Iba1, ionized calcium-binding adapter molecule 1; 3TC, lamivudine; MHC class II, major histocompatibility class II; mIPSC, miniature inhibitory postsynaptic currents; MOR, μ-opioid receptor; No info, information not provided or uncertain; OST, opioid substitution therapy; OUD, opioid use disorder; PFC, prefrontal cortex; PENK, preproenkephalin; q.d., once a day; q.i.d., four times a day; ROS, reactive oxygen species; s.c., subcutaneous; SUD, substance use disorder; tg, transgenic; t.i.d., three times a day; ZDV, zidovudine
For practicality, Tables 1 and 2 are limited to key studies in the CNS with emphasis on neuropathological or neuroimmune rather than psychosocial outcomes. With deference toward the excellent studies we excluded: (1) on opioid and HIV effects on peripheral blood mononuclear cells (PBMCs), or on isolated lymphocytes and monocytes, not directly related to the central nervous system or BBB; (2) on HIV or opioid and ARV interactions in the peripheral nervous system; and (3) studies not directly examining opioid-HIV interactions (irrespective of whether a positive or negative interaction was found)
Although translational, “bench-to-bedside”, research is important, reverse-translational approaches and multiple preclinical models are essential to better understand complex disease and improve established therapies (Singer 2019). Evidence suggests that HIV compartmentalizes within the CNS early during the course of the infection establishing a separate reservoir harboring “intact proviral” HIV (Churchill et al. 2016; Bruner et al. 2019) within resident neural cell populations (Bednar et al. 2015; Sturdevant et al. 2015; Veenhuis et al. 2019) and perivascular macrophages (Fischer-Smith et al. 2001; Burdo et al. 2013; Rappaport and Volsky 2015). Preclinical studies assessing opioid interactions with HIV or viral proteins permit mechanistic understanding of how particular CNS cell types, including neurons, astroglia, and microglia are affected and contribute to accentuating effects of opiates on neuroHIV, which are discussed in detail below.

Cellular and Molecular Interactions in Astroglia, Microglia, and Neurons
Prior reviews have outlined how opiate drugs likely exacerbate neuroHIV pathology in neurons and glia (Hauser et al. 2005; Dutta and Roy 2012; Hauser et al. 2012; Reddy et al. 2012; Hauser and Knapp 2014; Liu et al. 2016a; Murphy et al. 2019) including in the enteric nervous system (Galligan 2015; Meng et al. 2015). Opioid-HIV pathophysiological interactions are complex and differ depending on the timing and duration of co-exposure, the pharmacology of the opioid drug involved, the cell types and brain regions targeted, host and viral genetics, and are highly contextual (Hauser and Knapp 2014, 2018). A summary of the cellular and molecular interactions in various CNS cell types is also reviewed in detail in Table 2.
Table 2 Cellular and molecular interactions (in vitro)
Major effects HIV pathogena ARV Opioids Outcome Model system (in vitro) Citation(s)
Mixed-Glia
HIV expression HIV No • Dynorphin
• U50,488
(KOR agonists) • ↑ HIV-1 expression,
• Dynorphin (KOR agonist) ↑ TNF-α, IL-6 mRNA and protein Human fetal neural cells, HIV-infected promonocyte (U1) line (Chao et al. 1995)
HIVSF162 No • U50,488
• U69,593
• Dynorphin1–17;
(KOR agonists)
• Morphine • KOR agonists ± TNF-α differentially ↓ HIV p24 Human, primary mixed neurons and glia (Chao et al. 1998a)
Chemokines Tat1–86 No Morphine • ↑ CCL5, CCL2
• ↑ [Ca2+]i (Beclin1 dependent)
• ↓ Autophagy Mouse, primary mixed glia (Lapierre et al. 2018)
HIVSF162 (R5) No Morphine • ↑ HIV-1 Tat-induced LTR expression
• ↑ CCR5 expression (inhibited by bivalent ligand in astrocytes)
• ↑ IL-6
• ↑ CCL5 Human, primary mixed glia (El-Hage et al. 2013)
Glial restricted precursors: survival & MOR, DOR, KOR expression Tat1–72 No Morphine (acting via DOR and/or KOR) • ↑ Caspase-3 activation & ↑ cell death by Tat or morphine via DOR, KOR
• No opioid-Tat interactions Mouse, primary glial precursors (Buch et al. 2007)
MOR expression in NPCs;
NPC survival and developmental fate Tat1–72 No Morphine • MOR expressed by subsets of NPCs
• ↑ Astrocyte and immature glial death Mouse, primary mixed glia (Khurdayan et al. 2004)
MOR and CCR5 interactions Tat1–86 (from HIVIIIB) No Morphine • ↓ Neuronal survival via CCR5 activation in glia (rescued by BDNF treatment) Mouse, primary neurons and glia (Kim et al. 2018)
HIV infectivity MOR-CCR5 dimerization HIVSF162 (R5) No Morphine
CCR5-MOR bivalent ligand 1b • MOR-CCR5 bivalent ligand blocks HIV infection in astroglia, but not microglia, with morphine
• MOR-CCR5 bivalent ligand blocks the fusion of HIV gp160 and CCR5-CD4-expressing HEK cells Human, primary astrocytes and microglia; HEK-293T cells (Yuan et al. 2013; Arnatt et al. 2016)
HIV expression and maturational fate of neurons and astroglia HIVBaL (R5) No Morphine • ↑ HIV p24 and ↑ Tat mRNA levels with morphine after 21 days
• ↓ Proliferation of neural progenitors; ↑ astroglial and ↑ neuronal differentiation Human, neural progenitors (Balinang et al. 2017)
Astrocytes
HIV expression HIVSF162 (R5) No Morphine • ↑ HIV p24
• ↑ CCL2 Human, primary astrocytes (Rodriguez et al. 2017)
Toll-like receptor (TLR) expression/function • Tat1–72
• gp120 No Morphine • ↑ TLR2 with Tat, Tat + morphine, gp120
• ↓ TLR9 with Tat, morphine, gp120 Mouse, primary astrocytes (El-Hage et al. 2011a)
Chemokines Tat1–72 No Morphine • ↑ CCL5, CCL2
• ↑ IL-6
• ↑ [Ca2+]i Mouse, primary astrocytes (El-Hage et al. 2005)
Tat1–72 No Morphine • ↑ CCL2
• ↑ CCL5
• ↑ Microglial migration Mouse, primary astrocytes (El-Hage et al. 2006a)
Tat1–72 No Morphine • ↑ CCL2, ↑ IL-6, ↑ TNF-α
• ↑ [Ca2+]i
• ↑ NF-κB trafficking and transcription
• No interaction / acceleration with morphine Mouse, primary astrocytes (El-Hage et al. 2008b)
Tat No • U50,488 (KOR agonist)
• Nor-BNI (KOR antagonist) • U50,488 ↓ CCL2
• U50,488 ↓ NF-κB Human, primary astrocytes (Sheng et al. 2003)
N/A No Morphine • ↑ CCR5, CCR3, CXCR2
• ↓ IL-8, CCL4 Human, astrocytoma U87 cell line, primary astrocytes (Mahajan et al. 2002)
• Tat1–86
• gp120IIIB No Morphine Regional differences in cytokine and ROS production —differed for each insult Mouse, primary astrocytes (Fitting et al. 2010a)
Oxidative stress / damage Tat1–72 No • DPDPE
• SNC-80
(DOR agonists) DOR agonists ↓ Tat-induced oxidative stress Human derived brain cell line (SK-N-SH) (Wallace et al. 2006)
Inflammation, maturation /plasticity • Tat86
• Tat101 No Morphine ↓ β-catenin signaling and variably decreases TrkB, BDNF, and NLRP1 mRNA in fetal astrocytes b Human, U87MG and fetal astrocytes (Chen et al. 2020)
Microglia
HIV replication HIVSF162 (R5) No • Endomorphin-1
• Endomorphin-2
(MOR agonists) • ↑ HIV p24 with endomorphin-1, but not endomorphin-2
• Endomorphin-1 acts via MOR, but not DOR / KOR Human, primary microglia (Peterson et al. 1999)
HIVSF162 (R5) No Morphine ↑ HIV p24 Human, primary microglia (El-Hage et al. 2014)
HIVSF162 (R5) No • U50,488; U69,593 (KOR agonists)
• Dynorphin Al-13 ↓ HIV p24 Human, primary microglia (Chao et al. 1996b)
• HIVJR-FL (R5)
• gp120 No β-endorphin • ↑ HIV expression
• ↑ HIV p24 (14-day post infection)
• gp120 ↑ IL-1, TNF, IL-6 Human, fetal microglia (Sundar et al. 1995)
HIVSF162 No • 8-CAC, U50,488 (KOR agonists)
• Cocaine • KOR agonist ↓ p24; blocked by KOR antagonists
• KOR agonist negates cocaine-induced ↑ HIV Human, fetal brain microglia (Gekker et al. 2004)
HIVSF162 No OPRL1 antisense
Nociceptin / orphanin FQ (OPRL1 agonist) • OPRL1 antisense (and sense) ↓ p24
• Nociceptin, no effect on p24 Human, fetal brain microglia and mixed neurons/glia (Chao et al. 1998b)
HIV expression • HIVSF162
• Tat ZDV U50,488 (KOR agonist) • ↓ p24 on day 14 with U50,488
• ↓ Neurotoxicity (U50,488)
• ↓ Quinolinate by microglia Human, fetal microglia and neural cells (Chao et al. 2000)
Chemokines and Cytokines Tat1–72 No Morphine • ↑ CCR5
• ↑ CD11b, ↑ CD40
• ↑ TNF-α, ↑ IL-6, ↑ IP-10
• ↑ iNOS Mouse, BV-2 and primary microglia (Bokhari et al. 2009)
MOR signaling Tat1–72 No Morphine • ↑ MOR (intracellular)
• ↑ MOR mRNA Mouse, N9 and primary microglia (Turchan-Cholewo et al. 2008)
Oxidative Stress Tat1–72 No Morphine • ↑ ROS [O2− (DHE), ↑ HO2•, H2O2 (DCF)]
• ↑ Protein carbonyls Mouse, N9 and primary microglia (Turchan-Cholewo et al. 2009)
Glutamate release Tat1–72 No Morphine ↑ Glutamate release via ↑ xc− cystine-glutamate antiporter expression/function Mouse, primary microglia (Gupta et al. 2010)
Neurons
HIV expression HIV No Morphine ↑ HIV expression Human derived, SH-SY5Y neuroblastoma cells (Squinto et al. 1990)
Homeostasis and Injury Tat1–86 No Morphine • ↑ [Ca2+]i,
• ↑ [Na+]i
• ↓ ΔΨm (mitochondrial) instability
• ↑ Dendritic degeneration Mouse, primary neurons (Fitting et al. 2014a)
Mitochondrial inner membrane potential and ROS • Tat1–86, Tat1–72
• gp120 No Morphine ↑ ΔΨm instability and oxidative stress ↑ with Tat + morphine, ↑ neuroprotection with allopregnanolone Human, primary neurons ; mouse, striatal medium spiny neurons; mouse, striatal medium spiny neurons, SH-SY5Y neuroblastoma cells (Turchan-Cholewo et al. 2006; Paris et al. 2020)
Neuronal survival Tat1–86 No Morphine • ↓ Neuronal survival from Tat + morphine and ↓ glial CX3CL1 rescued by CX3CL
• CX3CL1 (fractalkine) regulates microglial motility Mouse, primary neurons and mixed glia (Suzuki et al. 2011)
Tat1–86 No Morphine • ↓ Proliferation
• ↑ ERK1/2 activation
• ↑ p53 and p21
• ↓ Cyclin D1 and Akt levels Human, neuronal precursors (Malik et al. 2014)
Tat1–72, Tat1–86 No Morphine • ↓ Neuronal survival
• ↑ Neuronal survival with ibudilast (AV411) (inhibiting glial NF-κB blocks Tat ± morphine neurotoxicity) Mouse, primary neurons and mixed glia (Gurwell et al. 2001; El-Hage et al. 2014)
White matter/oligodendroglial pathology
Changes in OL survival and morphology Tat1–86 No Morphine (25 mg pellet, 7 days); morphine (in vitro) • ↑ Degeneration of OLs
• ↑ TUNEL reactivity
• ↑ Caspase-3 activation Mouse, Tat tg; primary OLs (Hauser et al. 2009)
Blood-brain barrier and the neurovascular unit
BBB model integrity and function Tat1–86 No Morphine • ↑ TNF-α
• ↑ IL-8
• ↓TEER
• ↑ JAM-2 expression
• ↑ Monocyte transmigration with CCL5 Human, using primary BMVEC and primary astrocytes (Mahajan et al. 2008)
ARV accumulation Tat1–86 DTG
FTC
TFV Morphine • ↓ Intracellular ARV concentrations Human, primary astrocytes (Patel et al. 2019)
HIV-1 strain differences
Neuronal Survival Tat1–86 (clades B & C) No Morphine • ↓ Neuronal survival via MOR on mixed glia
• ↑ ROS in astrocytes
• ↑ Iba1 and 3-NT microglia with morphine Mouse, primary neurons and mixed glia (Zou et al. 2011)
• gp120IIIB
• gp120MN (X4)
• gp120ADA (R5) No Morphine ↓ Neuronal survival in presence of glia with gp120MN and transiently with gp120IIIB (X4), not R5-tropic gp120, in combination with morphine Mouse, primary neurons and mixed glia (Podhaizer et al. 2012)
Proliferation and maturational fate of neural progenitors and oligodendroglia • HIVSF162 (R5)
• HIVIIIB (X4) No Morphine • ↓ Proliferation of immature neural and OL progenitors with Tat + morphine
• ↓ NPC DNA synthesis with R5-tropic HIV + morphine
• ↑ NPC DNA synthesis with X4-tropic HIV + morphine Mouse, Tat tg; Mouse, Human, primary neural progenitors (Hahn et al. 2012)
GABA function • HIVBaL (R5)
• gp120 (ADA, MN, and IIIB)
• Tat1–86 No Morphine • Tat or morphine ↓ KCC2 levels via CCR5
• ↑ KCC2 prevents Tat and R5 HIV, gp120, but not X4, gp120 neurotoxicity ± morphine Human, primary neurons, hNPCs (Barbour et al. 2020)
Astroglial CCL5 and neuroprotection • gp120IIIB (X4)
• gp120BaL (R5) No • Morphine (10 μM)
• DAMGO • Morphine ↑ astroglial CCL5 blocking gp120BaL neurotoxicity
• Morphine (or CXCL12) does not block gp120IIIB neurotoxicity Rat, mixed neurons and glia; isolated neurons, astrocytes and microglia (Avdoshina et al. 2010)
aassumed Clade B, unless noted otherwise, b statistical findings for some results are unclear
ARV, antiretroviral(s); BMVEC, brain vascular endothelial cells; [Ca2+]i intracellular calcium concentration; 8-CAC, 8-carboxamidocyclazocine; DAMGO, D-Ala2, N-MePhe4, Gly-ol]-enkephalin; DCF, dihydro-dichlorofluorescein; DOR, δ-opioid receptor; DHE, dihydroethidium; DTG, dolutegravir; DPDPE, [D-Pen2,D-Pen5]enkephalin; FTC, emtricitabine; GABA, γ-aminobutyric acid; Iba1, ionized calcium-binding adapter molecule 1; JAM-1, junctional adhesion molecule-1; KCC2, K+-Cl− cotransporter 2; KOR, κ-opioid receptor; LTR, long terminal repeat; ΔΨm, mitochondrial inner membrane potential; MOR, μ-opioid receptor; [Na+]i, intracellular sodium concentration; nor-BNI, nor-binaltorphimine; NPCs, neural progenitor cells; OLs, oligodendroglia; ROS, reactive oxygen species; TEER, transendothelial electrical resistance; TFV, tenofovir; TUNEL, terminal deoxynucleotidyl transferase dUTP nick end labeling; ZDV, zidovudine
For practicality, the table is limited to key studies in the CNS with emphasis on neuropathological or neuroimmune rather than psychosocial outcomes. With deference toward the excellent studies we excluded: (1) on opioid and HIV effects on PBMCs, or on isolated lymphocytes and monocytes, not directly related to the central nervous system or BBB; (2) on HIV or opioid and ARV interactions in the peripheral nervous system; and (3) studies not directly examining opioid-HIV interactions (irrespective of whether a positive or negative interaction was found)

Opioid and HIV Interactive Pathology in Astroglia
Although the extent to which astroglia display productive infection is debated (Russell et al. 2017; Ko et al. 2019), there is nevertheless considerable evidence of proviral integration in the CNS of PWH (Gorry et al. 2003; Churchill et al. 2009), infectious animal models (Eugenin et al. 2011), and/or cultured human fetal astrocytes (Tornatore et al. 1994; Liu et al. 2004; Do et al. 2014; Narasipura et al. 2014; Li et al. 2015; Nath 2015; Li et al. 2020). Integrated HIV sequences have been identified in astrocytes in HIV-infected CNS tissue by laser capture microdissection (Churchill et al. 2006). Astroglia appear to infect via non-classical, CD4-independent mechanisms, that can have the appearance of virologic synapses, adding to the debate (Liu et al. 2004; Do et al. 2014; Li et al. 2015; Nath 2015; Al-Harthi et al. 2019; Li et al. 2020).
Irrespective of whether they become infected, MOR-expressing, HIV or HIV protein-exposed astrocytes release greater amounts of inflammatory cytokines and dysfunction sufficient to harm bystander neurons upon treatment with opiates (El-Hage et al. 2005, 2008b; Zou et al. 2011; El-Hage et al. 2014). MOR-expressing subsets of glia, especially microglia and astroglia, are prominent in driving the interactive opioid and HIV neuropathogenesis (Hauser et al. 2007, 2012; Hauser and Knapp 2014; Liu et al. 2016a; Chilunda et al. 2019; Murphy et al. 2019). When MOR is deleted from glia (astrocytes and microglia), morphine no longer increases the death of Tat-exposed striatal medium spiny neurons (MSNs) (Zou et al. 2011). Conversely, if MOR is deleted from MSNs, morphine exacerbates the neurotoxic effects of Tat in MSNs (Zou et al. 2011). The proinflammatory effects of Tat alone or in combination with morphine on glia are mediated through a Beclin-1-dependent autophagy pathway (Rodriguez et al. 2017; Lapierre et al. 2018). Drugs with selective glial anti-inflammatory activity (i.e., ibudilast or AV411) attenuated the deleterious effects of HIV and opiate exposure, including HIV-1 replication, cytokine release, and neurotoxicity in vitro (El-Hage et al. 2014). Thus, the observed neuronal death is largely mediated by MOR-expressing glia (Zou et al. 2011), including astroglia (El-Hage et al. 2005, 2008b) and microglia (Turchan-Cholewo et al. 2008; Bokhari et al. 2009; Turchan-Cholewo et al. 2009; Gupta et al. 2010).
The direct contributions of astrocytes to opioid and HIV interactions have been discussed previously (Dutta and Roy 2012; Hauser et al. 2012; Reddy et al. 2012; Hauser and Knapp 2014). Subsets of astroglia can express MOR, DOR and KOR (Stiene-Martin and Hauser 1991; Eriksson et al. 1992; Ruzicka et al. 1995; Gurwell et al. 1996; Hauser et al. 1996; Peterson et al. 1998; Stiene-Martin et al. 1998, 2001), as well as endogenous opioid peptides (Vilijn et al. 1988; Shinoda et al. 1989; Spruce et al. 1990; Hauser et al. 1990; Low et al. 1992). It appears that the ‘early’ events triggering the release of proinflammatory cytokines (i.e., TNF-α and IL-1β) from astroglia can be mediated by HIV Tat exposure alone (El-Hage et al. 2005, 2006a, b, 2008a). Opioids enhance HIV-1-induced inflammation later during the inflammatory cascade by exacerbating the sustained release of CCL5 from astrocytes, which subsequently triggers the release of CCL2 thereby enhancing the recruitment and activation of macrophages/microglia (El-Hage et al. 2008a) (Fig. 1). This is caused by the morphine-dependent exacerbation of Tat-induced increases in intracellular calcium concentration ([Ca2+]i) in astroglia (El-Hage et al. 2005), which accelerates the trafficking of NF-κB p65 (RelA) subunits to the nucleus and sustained CCL2, CCL5, and IL-6 transcription in astrocytes (El-Hage et al. 2008b).
Fig. 1 Opioids exacerbate HIV-1-induced CNS inflammation, in part, by augmenting CCL5-dependent increases in CCL2—key sites of opioid-HIV convergent interactions in glial inflammatory signaling cascades. Subpopulations of striatal glial fibrillary acidic protein (GFAP)-immunoreactive astrocytes in wildtype mice normally express CCR2 immunoreactivity (a-b; arrows), CCL2 (c; arrow), or μ-opioid receptor (MOR) (d; arrows) immunoreactivity (scale bars a-b = 25 μm; c-d = 15 μm). CCR2 deletion (−/−) significantly reduces HIV-1 Tat ± morphine-induced increases in GFAP+ astroglia (e) and F4/80+ macrophages/microglia (f) compared to wild type (+/+) mice at sites near (300 ± 100 μm) the site of Tat injection (*p < 0.05 vs. wild type mice) (see, El-Hage et al. 2006a). In wild-type mice, Tat ± morphine administration markedly increases the proportion of CCL2 immunoreactive astrocytes (g) or macrophages/microglia (h) [*p < 0.05 vs. other groups in wild-type or CCL5(−/−) mice; bp < 0.05 vs. vehicle- or Tat plus morphine-treated wild-type mice; #p < 0.05 vs. equivalent treatment in wild-type mice], while in CCL5 null mice, significant increases in CCL2 immunoreactivity were only seen in macrophages/microglia co-exposed to Tat and morphine (§p < 0.05 vs. vehicle injected CCL5 knockout mice) (see, El-Hage et al. 2008a). CCL5 expression in striatal GFAP-immunoreactive astrocytes (arrows) increases following Tat injections (i, j) compared to wild-type control mice (not shown) (El-Hage et al. 2008a). Opioids exacerbate HIV-1-induced CNS inflammation, in part, by increasing CCL5 and augmenting CCR5-dependent increases in CCL2 production by astrocytes resulting in the activation and recruitment of microglia/macrophages and spiraling inflammation (k). Additional factors likely mediate the proinflammatory cascade, but these are less well substantiated (?). Moreover, autocrine and reciprocal paracrine (astroglial ↔ macrophage/microglial) intercellular, feedback amplification mechanisms from macrophages/microglia are likely to be operative (dashed red arrow) (also see, Kang and Hebert 2011) and occur elsewhere within the cascade (not shown); effects of HIV-1 Tat/HIV, red arrows; sites of opioid convergence, blue arrows; pro-BDNF:mature BDNF (mBDNF) ratio (Kim et al. 2018). (a-f) Modified and reprinted with permission from El-Hage et al. (2006a). (g-k) Modified and reprinted with permission from El-Hage et al. (2008a)

Opioid and HIV Interactive Pathology in Microglia
Unlike in astrocytes, opiate and HIV interactions in microglia tend to be self-limiting (Turchan-Cholewo et al. 2009). Opiates initially trigger large increases in the production of proinflammatory cytokines (Hauser, unpublished), reactive oxygen (ROS) and nitrogen (RNS) species (Turchan-Cholewo et al. 2009), and the release of glutamate (Gupta et al. 2010) and ATP (Sorrell and Hauser 2014) extracellularly in Tat-exposed microglia. The release of glutamate is mediated by the catalytic subunit of the cystine-glutamate antiporter xc− (xCT) (Gupta et al. 2010). Interestingly, following acute increases in the release of cytokines (e.g., TNF-α; unpublished), morphine no longer increases Tat-induced cytokine levels at 24 h; instead, their levels are reduced by opiate-dependent proteasome inhibition. The proteasome inhibitor, MG115, mimics the effects of morphine in decreasing proteasome activity at 24 h and blocks TNFα, IL-6, and CCL2 release from microglia, but does not increase ROS or RNS production (Turchan-Cholewo et al. 2009). The ubiquitin proteasome system (UPS) is typically viewed as contributing to opiate tolerance and physical dependence by modulating MOR downregulation (Massaly et al. 2014; Caputi et al. 2019), rather than MOR activity constraining the UPS. Thus, while HIV-exposed, MOR-expressing microglia show a burst of ROS and proinflammatory cytokine production in response to morphine, the cytokine release collapses within 24 h seemingly because sustained opiate exposure inhibits the UPS thereby preventing degradation of the IκB subunit and nuclear translocation of NF-κB (Turchan-Cholewo et al. 2009).
While neither astroglia nor microglia alone mimic the full inflammatory profile seen with opiates and HIV in the CNS; in combination, the neuroimmune signature more accurately mimics that seen in neuroHIV. Accordingly, we have proposed that opioids promote positive feedback through separate actions in astroglia and microglia in neuroHIV—resulting in spiraling inflammation and cytotoxicity (Hauser et al. 2005, 2007).

Opioid and HIV Interactive Pathology in Neurons
Besides accentuating HIV-induced neurotoxicity via glial-mediated mechanisms, morphine appears to converge with HIV Tat to dysregulate ion homeostasis and dendritic injury through potential direct actions on neurons, even though some contributions of glia cannot be excluded in this study (Fitting et al. 2014a). Combined morphine and Tat exposure accelerates the formation of Tat-induced focal dendritic varicosities/swelling via a MOR-related mechanism that was caused by focal increases in Na+ influx and [Ca2+]i, an overload of Na+/K+-ATPase, ATP depletion, and a collapse in mitochondrial inner membrane potential (Fitting et al. 2014a). Importantly, morphine’s additive effects were mediated via a MOR-related mechanism, as the exacerbating effects of morphine were absent in neurons from MOR knockout mice, thus excluding TLR4 involvement (Fitting et al. 2014a). Further, morphine exacerbated Tat-dependent focal losses in ion homeostasis by mobilizing [Ca2+]i through ryanodine-2 (RyR2)-sensitive sites (Fitting et al. 2014a) (Fig. 2). Although morphine typically acts via MOR in an inhibitory manner by activating Gi/o-proteins (Sharma et al. 1977; Moises et al. 1994; Al-Hasani and Bruchas 2011), MOR-dependent stimulation of PI3-kinase and Ca2+ mobilization (Leopoldt et al. 1998) in neurons via the Gβγ protein subunit (Mathews et al. 2008) is presumed operative here (Fig. 2).
Fig. 2 Morphine exacerbates the excitotoxic effects of HIV Tat by mobilizing Ca2+ from ryanodine (RyR)-sensitive internal stores. (a) Tat-induced increases in [Ca2+]i were not attenuated by ryanodine, whereas ryanodine and pyruvate attenuate combined Tat and morphine-induced increases in [Ca2+]i. Nimodipine (L-type Ca2+ channel blocker) and dantrolene did not show any effects. (b) Average [Ca2+]i over 10 min indicated ryanodine significantly blocked combined Tat and morphine-induced increases in [Ca2+]i, whereas no effects were noted for nimodipine, dantrolene, or pyruvate. *p < 0.05 vs. control, #p < 0.05 vs. Tat 50 nM, §p < 0.05 vs. TM, TM: Tat 50 nM + Morphine 500 nM. (c) Summary of HIV-1 Tat and morphine interactive neuronal injury in striatal medium spiny neurons. Combined Tat and morphine promotes structural and functional defects in dendrites via α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPAR), N-methyl-D-aspartic acid receptors (NMDAR), and MOR, causing influxes of Na+ and/or Ca2+, compensatory increases in Na+/K+-dependent ATPase activity, and a rapid loss in ATP mobilization with an inability to extrude excess Na+ via Na+/K+-ATPase caused by mitochondrial hyperpolarization. Dysregulation of [Ca2+]i homeostasis by combined Tat and morphine appears to be mediated downstream of [Na+]i at the level of calcium mobilization, which in turn appears to be regulated via ryanodine (RyR)-sensitive sites, and enhanced by morphine exposure likely via MOR-dependent stimulation of PI3-kinase and Ca2+ mobilization via the Gβγ protein subunit. (a-b) Modified and reprinted with permission from Fitting et al. (2014a)
Glial-derived neuronal injury is not unidirectional. Neuronal damage-associated molecular patterns (DAMPs) and dysfunction can trigger both infected and uninfected glia to become reactive, resulting in further neuronal damage and escalating pathology. Neuronal injury can reactivate HIV in latently infected microglia (Alvarez-Carbonell et al. 2019). While the events underlying the disruption of neuronal-microglial activation that trigger the emergence of latent HIV are unclear, the induction of HIV expression appears to involve the production of DAMPs by injured neurons and can be turned “on”, e.g., by methamphetamine-induced sigma-1 (σ1) receptor activation, TNF-α and IL-1β, and TLR3 activation can be turned “off” by CX3CL1/fractalkine or glucocorticoid receptor activation (Alvarez-Carbonell et al. 2017, 2019).

Neural Systems Selectively Disrupted by Opiate and HIV Interactions

Blood-Brain Barrier and the Neurovascular Unit
Despite growing evidence on how opiates and HIV interact to impact the neuropathology of HIV, little is known about their interactive effects on the blood-brain barrier (BBB). BBB integrity and function are critical for maintaining CNS homeostasis, and mediating neuroimmune interactions with the periphery and drug delivery into the CNS. HIV and many individual HIV proteins can breakdown the BBB disrupting tight junction proteins (Dallasta et al. 1999; Boven et al. 2000; Andras et al. 2003; Mahajan et al. 2008; Banerjee et al. 2010; Gandhi et al. 2010; Xu et al. 2012; Patel et al. 2017) and decreasing transendothelial electrical resistance (TEER) (an in vitro measure of barrier integrity) (Mahajan et al. 2008; Gandhi et al. 2010; Mishra and Singh 2014; Patel et al. 2017), with resultant paracellular “leakage” of compounds/current between compromised barrier endothelial cells (Mahajan et al. 2008; Gandhi et al. 2010; Wen et al. 2011; McLane et al. 2014; Leibrand et al. 2017, 2019). Although opioids can also impair the BBB through alterations in tight junction proteins and/or increased paracellular flux (Baba et al. 1988; Mahajan et al. 2008; Wen et al. 2011; Leibrand et al. 2019), others have found that it is morphine withdrawal, not the continued exposure to morphine, that most greatly disrupts BBB integrity (Sharma and Ali 2006). In addition to perturbing paracellular dynamics, morphine may also alter the expression and/or function of drug efflux proteins, such as P-glycoprotein (P-gp). Sub-chronic and chronic morphine exposure is reported to increase P-gp expression and/or function (Aquilante et al. 2000; Mahajan et al. 2008; Yousif et al. 2008; Leibrand et al. 2019). Alternatively, other investigators report no changes in P-gp with chronic exposure (Chaves et al. 2016), while some see increases upon morphine withdrawal (Yousif et al. 2012; Chaves et al. 2016). Alterations in drug transport proteins would impact the central accumulation and efficacy of therapeutic drugs that are their substrates.
Using a primary human brain microvascular endothelial cell (BMEC) and astrocyte co-culture model, Mahajan et al. (2008) were among the first to demonstrate that co-exposure to morphine and HIV-1 Tat resulted in greater increases in TNF-α and IL-8 levels and decreases in barrier tightness (measured by TEER) than either morphine or Tat alone. Morphine and Tat co-exposure also additively increased JAM-2, while zonula occludens-1 (ZO-1) levels were decreased by morphine or by Tat individually, and occludin protein levels were decreased by morphine alone but not Tat (Mahajan et al. 2008). Using the inducible Tat transgenic mouse model, Leibrand et al. (2019), also demonstrated that HIV-1 Tat and morphine act independently to disrupt BBB integrity. In these studies, morphine, and to a lesser extent Tat, exposure increased the leakage of fluorescently labeled dextrans from the circulation into the brain (Leibrand et al. 2017, 2019) (Fig. 3). Morphine exposure decreased the penetration of select ARVs in the brain, in a region-specific manner (Leibrand et al. 2019) (Fig. 3). Morphine exposure also resulted in increased expression and function of the drug efflux transport protein, P-gp, suggesting a mechanism by which morphine decreased the ARV concentrations (Leibrand et al. 2019). This finding suggests that morphine exposure could impact the efficient delivery of any therapeutic drug that is a substrate of P-gp into the CNS. Future research should also investigate morphine’s impact on other drug transport proteins important for ARV delivery to the brain.
Fig. 3 Effects of HIV-1 Tat and morphine on BBB leakiness and on antiretroviral brain concentrations. After 14 days of Tat induction, there was a significant increase in the 10 kDa (Cascade Blue®) tracer leakage into the brain in Tat + placebo as compared to Tat − placebo mice (*p < 0.05) and in Tat − mouse brains upon exposure to morphine as compared to Tat − placebo mice (*p < 0.05) (a). There was a significant main effect of morphine, resulting in reduced integrity of the BBB and increased leakage of the higher molecular weight (40 kDa and 70 kDa) tracers in morphine-exposed groups as compared to the those groups (Tat + and Tat − together) not exposed to morphine (placebo) (#p < 0.05; significant main effect of morphine) (b, c). Data represent the fold change in mean fluorescence intensity ± SEM; n = 8 Tat−/placebo, n = 6 Tat+/placebo, n = 9 Tat−/morphine, and n = 7 Tat+/morphine mice. Additionally, morphine exposure increased horseradish peroxidase (HRP) extravasation from the vasculature into the perivascular space and/or parenchyma in the striatum (d, e). HRP antigenicity was detected by indirect immunofluorescence (red) in tissue sections counterstained with Hoechst 33342 (blue) to reveal cell nuclei and visualized by differential interference contrast (DIC)-enhanced confocal microscopy. HRP extravasation into the striatal perivascular space/parenchyma was especially prevalent in morphine-exposed mice (arrowheads; left-hand panels in e versus d). The dotted lines (············) indicate the approximate edge of the capillaries/post-capillary venules; while intermittent dotted lines (· · · · · · ·) indicate the approximate edge of a partly sectioned blood vessel that appears partially outside the plane of section. The asterisks (*) indicate white matter tracts within the striatum. Representative samples from ≥ n = 4 mice per group. All images are the same magnification. Scale bar = 10 μm. Antiretroviral tissue-to-plasma ratios in striatum (f–g). Irrespective of Tat exposure, morphine significantly reduced the levels of dolutegravir (f) and abacavir (g), but not lamivudine (h), within the striatum, as compared to placebo. (* p < 0.05; main effect for morphine). Data represent the tissue-to-plasma ratios ± SEM sampled from n = 9 Tat−/placebo, n = 9 Tat+/placebo, n = 6 Tat−/morphine, and n = 8 Tat+/morphine mice. (a–h) Modified and reprinted with permission from Leibrand et al. (2019)
HIV, HIV-1 viral proteins, and opiate-induced barrier dysfunction is associated with increased infiltration of monocyte-derived macrophages (MDMs) into the brain. Enhanced influx of peripheral (infected) macrophages into the brain can serve to replenish viral reservoirs and further promote neuroinflammation. Several studies have examined the individual impact of HIV, Tat, or morphine on monocyte adhesion or migration into the CNS (Nottet et al. 1996; Wu et al. 2000; Fischer-Smith et al. 2001; Pello et al. 2006; Williams et al. 2013a, 2014; Strazza et al. 2016; Leibrand et al. 2017; Chilunda et al. 2019). However, fewer studies have examined the combined effects of HIV/Tat and opiates. Co-exposure of HIV-1 Tat and morphine on astrocytes increases the production of chemoattractants, primarily CCL2 and CCL5, and increases microglial migration. These effects were inhibited by MOR blockade (El-Hage et al. 2006b). Co-exposure of Tat and morphine or buprenorphine to a BBB model increases monocyte transmigration in response to CCL5 and other chemokines (Mahajan et al. 2008; Jaureguiberry-Bravo et. al. 2016). In S. pneumoniae-infected mice, morphine and/or Tat exposure significantly enhances immune cell trafficking into the brain via actions at TLR2 and TLR4 (Dutta and Roy 2015).
Taken together, BBB damage from HIV and/or opiates can disrupt the homeostasis within the brain. Breakdown of paracellular processes, through decreases in tight junction proteins and increased cellular adhesion proteins, increased leakage of circulating molecules into the brain and increased monocyte/MDM adhesion and transmigration into the brain, which if infected, can serve to replenish viral reservoirs within the CNS. Furthermore, alterations in drug transport proteins within the brain can decrease ARV efficacy by decreasing drug concentrations. Collectively, these changes serve to maintain HIV infection within the brain (see Fig. 4; Tables 1 and 2).
Fig. 4 Schematic representation of the blood-brain barrier and other components of the neurovascular unit. Under normal conditions (represented above the dotted line), tight junctions are intact which restricts the leakage of paracellular, typically small hydrophilic, compounds, across the barrier and into the brain. Additionally, there is a basal expression of efflux transporters, such as P-glycoprotein (P-gp), which effluxes substrates out of the brain, serving to restrict overall accumulation within the brain. In the setting of HIV and opiate exposure (represented below the dotted line), there is a breakdown of the tight junction proteins and increased leakage of paracellular compounds into the brain. Additionally, opiate exposure increases efflux transporter expression, including P-gp and potentially breast cancer resistance protein (Bcrp), thereby restricting overall brain penetration of drugs (like many antiretroviral drugs) which are substrates for these transporters and in response to HIV and/or opioid exposure.

White Matter/Oligodendroglial Pathology
HIV can cause white matter damage (Gosztonyi et al. 1994; Langford et al. 2002; Xuan et al. 2013) even with less severe forms of HAND (Chen et al. 2009; Leite et al. 2013; Correa et al. 2015). Diffusion tensor magnetic resonance imaging (DTI) demonstrates white matter damage early in HAND (Ragin et al. 2004; Stubbe-Drger et al. 2012; Leite et al. 2013; Correa et al. 2015). White matter deficits are associated with cognitive impairment, including shortfalls in memory (Ragin et al. 2005), executive function (Correa et al. 2015), motor speed (Wu et al. 2006; Stubbe-Drger et al. 2012), and perhaps depression (Schmaal and van Velzen 2019). Preclinical studies in simian immunodeficiency virus- (SIV-) infected rhesus macaques (Marcario et al. 2008) and HIV-infected humanized mice (Boska et al. 2014) support the clinical findings. Injury to oligodendrocytes (OLs) can occur very early in the disease (see review, Liu et al. 2016b). Viral proteins, including Tat, gp120, and Nef, have been implicated in OL injury in vitro (Kimura-Kuroda et al. 1994; Bernardo et al. 1997; Radja et al. 2003; Nukuzuma et al. 2012; Zou et al. 2015), and in animal models in vivo (Radja et al. 2003; Hauser et al. 2009; Zou et al. 2015). Importantly, Tat has been detected in OLs in the brains of AIDS patients (Del Valle et al. 2000).
HIV likely damages OLs through both direct and indirect actions. OLs lack CD4, and reports of OL infection by HIV are variable (Esiri et al. 1991; Albright et al. 1996; Wohlschlaeger et al. 2009); thus, HIV infection of OLs is unlikely a major avenue of OL or white matter damage (discussed below). Alternatively, bystander damage to OLs through the production of “virotoxins” and “cellular toxins” (Nath 1999) by infected neighboring cells is more likely to be operative (Hauser et al. 2009; Zou et al. 2015; Jensen et al. 2019; Zou et al. 2019). ARVs also contribute to OL cytotoxicity (Jensen et al. 2015; Festa et al. 2019; Jensen et al. 2019). HIV-1 Tat directly induces damage in isolated OLs through α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/N-methyl-D-aspartic acid (NMDA) receptor-dependent mechanisms (Zou et al. 2015) and is also associated with abnormal Kv1.3 activity (Liu et al. 2017). Immature OLs are preferentially targeted by Tat compared to differentiated OLs (Khurdayan et al. 2004; Hahn et al. 2012; Zou et al. 2015, 2019). While the reasons why immature OLs are more susceptible to Tat are unclear, unlike mature OLs, Tat preferentially upregulates GSK-3β signaling in undifferentiated OLs by inhibiting Ca2+/calmodulin-dependent protein kinase II β (CaMKIIβ) (Zou et al. 2019).
Opioid abuse by itself can result in demyelination, leukoencephalopathy, and lesions in white matter (Offiah and Hall 2008; Eran and Barak 2009; Morales Odia et al. 2010; Bora et al. 2012; Li et al. 2013), and the degree of myelin disruption correlates with the duration of opiate dependence (Ivers et al. 2018). Chronic oxycodone exposure in rats causes some axonopathies and reduces the size of axonal fascicles, decreases myelin basic protein levels, and causes the accumulation of amyloid-β precursor protein (APP) (Fan et al. 2018). Most preclinical studies have examined the effects of opioids and opioid receptor blockade on OL maturation and/or the timing of myelination (Hauser et al. 1993; Knapp et al. 1998; Stiene-Martin et al. 2001; Sanchez et al. 2008; Knapp et al. 2009; Vestal-Laborde et al. 2014). OLs can transiently express MORs and other opioid receptor types (Knapp et al. 1998; Tryoen-Toth et al. 2000; Knapp et al. 2001; Stiene-Martin et al. 2001). Selective MOR and possibly KOR activation can directly modulate the growth of OLs in vitro (Knapp and Hauser 1996; Knapp et al. 1998, 2001).
Despite long-standing evidence of white matter damage early during the infection even in asymptomatic PWH (Price et al. 1988; Gray et al. 1996; Chen et al. 2009; Stubbe-Drger et al. 2012; Jensen et al. 2019), few studies have examined how opiate exposure affects OLs and myelin in neuroHIV (Tables 1 and 2). Increased demyelination is reported in SIV-infected rhesus macaques chronically treated with morphine (4× daily, up to 59 weeks) (Marcario et al. 2008). Specifically, morphine-treated SIV macaques developed more subtle, focal, dysmyelinating lesions, with accumulations of macrophages in areas of myelin loss (Marcario et al. 2008), as well as accompanying gliosis (Marcario et al. 2008; Rivera-Amill et al. 2010a; Bokhari et al. 2011). Morphine exposure increased degeneration of OLs in Tat+ mice, which was accompanied by elevations in caspase-3 activation and TUNEL reactivity in OLs and reversible by naloxone or naltrexone, respectively (Hauser et al. 2009). Although OLs can express MOR both in vivo (Stiene-Martin et al. 2001) and in vitro (Hauser et al. 2009), it remains unclear the extent to which MOR activation in OLs directly mediates HIV pathogenesis.

Neural Progenitors as an HIV Reservoir and Target for Opioids
Even though neural progenitors (Krathwohl and Kaiser 2004; Lawrence et al. 2004; Rothenaigner et al. 2007; Schwartz et al. 2007; Balinang et al. 2017), neuroblast cell lines (Ensoli et al. 1994; Rothenaigner et al. 2007), and/or immature astroglia (Atwood et al. 1993; Tornatore et al. 1994; Barat et al. 2018) can harbor HIV infection (reviewed by Hauser and Knapp 2014; Putatunda et al. 2019), the degree to which they are a source of active infection or serve as a latent viral reservoir (Blankson et al. 2002; Bruner et al. 2019) by retaining intact proviral DNA within incipient macroglial progeny is uncertain. In fact, spurious reports of HIV-infected adult neurons (Torres-Munoz et al. 2001; Canto-Nogues et al. 2005) may result from the retention of proviral genes that integrated into pluripotent neural progenitors or neuroblasts at earlier stages during maturation. Importantly, prolonged exposure to opioids can increase the production of HIV in human neural progenitor cells (hNPCs). Exposure of R5-tropic HIVBaL-infected hNPCs to morphine continuously for 21 d increased viral production compared to HIVBaL infection alone in vitro (Balinang et al. 2017).
Besides being able to infect hNPCs, HIV may also affect their maturation and the fate of neural stem cells. That HIV or gp120 can inhibit adult neurogenesis (Okamoto et al. 2007; Lee et al. 2013; Putatunda et al. 2018) has been the topic of past reviews (Schwartz and Major 2006; Venkatesan et al. 2007; Peng et al. 2008, 2011; Ferrell and Giunta 2014; Hauser and Knapp 2014; Putatunda et al. 2019). How HIV inhibits the self-renewal, tripotential differentiation, and survival of neural progenitors/stem cells or the genesis of adult neurons in the subgranular zone (SGZ) of the dentate gyrus is uncertain. HIV and gp120 [via actions at the same chemokine receptor(s) (Tran and Miller 2005; Li and Ransohoff 2008)] are proposed to modulate the adult neurogenesis via Notch (Fan et al. 2016), by obstructing a cell cycle checkpoint via the activation MAPK-activated protein kinase 2 and Cdc25B/C (Okamoto et al. 2007), or through signaling by platelet-derived growth factor BB (Chao et al. 2014) or BDNF (Lee et al. 2013). The extent that HIV regulates the genesis of neural progenitors within the subventricular zone of the developing CNS through similar mechanisms as in the adult SGZ of the dentate gyrus is uncertain—even though HIV disrupts the generation of neurons and glia during maturation or in adults. For example, MAPK/ERK1/2 enhances p53- and p21-dependent downregulation of cyclin D1 hindering progression through the G1 phase of the cell cycle in hNPCs (Mishra et al. 2010; Malik et al. 2014). Importantly, opioids too can affect the genesis of neurons and glia during maturation or in the adult directly via convergent pathways (Hauser and Knapp 2018; Kibaly et al. 2018) suggesting yet another site of opioid and HIV interactions in dysregulating the creation and fate of new neurons and glia.
Few studies have examined the interplay between opioids, neural progenitors and HIV/HIV proteins. Sustained exposure (4 d) to morphine (500 nM) and Tat1–72 (100 nM) decreased the viability of MOR-expressing striatal glial precursors, and to a lesser extent astrocytes, and this coincided with caspase-3 activation (Khurdayan et al. 2004). By contrast, comparably administered morphine or Tat alone was sufficient to decrease the viability of immature glia/glial progenitors in spinal cord cultures, while Tat and morphine failed to interact (Buch et al. 2007). Collectively, these findings were the first to indicate that opioid and/or Tat could enhance programmed cell death in subpopulations of glial precursors in a developmentally regulated and region-dependent manner (Khurdayan et al. 2004; Buch et al. 2007). In human glial progenitors, co-administering morphine (500 nM) increased the antiproliferative effects of Tat (12–48 h) or conditioned medium from HIV-1SF162-infected MDMs (12 h), while paradoxically reversing the antiproliferative effects from HIV-1IIIB conditioned medium (12 h) (Hahn et al. 2012). In these studies, Tat or HIV exposure reduced the proliferation of Sox2+ and Olig2+ undifferentiated glial and oligodendroglial progenitors, respectively, while progenitor viability was unchanged (Hahn et al. 2012). In human neural progenitor cells (hNPCs), sustained infection with R5-tropic HIVBaL increased the proliferation and premature differentiation of hNPCs into both neurons and astrocytes, and both measures were significantly enhanced by morphine co-exposure (Balinang et al. 2017). Importantly, immunoneutralizing antibodies (Hahn et al. 2012) or the selective antagonist, maraviroc (Balinang et al. 2017), were able to significantly attenuate the consequences of R5-tropic HIV infection on hNPC differentiation and fate confirming a direct role of CCR5 in these processes. Lastly, decreases in the proliferation of hNPCs seen with morphine and Tat are, in part, regulated by ERK1/2-dependent increases in p53 and p21 expression (Malik et al. 2014) and can be modulated by PDGF BB suggesting a possible therapeutic target (Malik et al. 2011). Thus, morphine can exaggerate R5-tropic HIV-induced alterations in the maturation and fate of human and rodent NPCs, thereby further disrupting the balance of neural cell types and CNS function.