Accelerated brain aging with opioid misuse and HIV: New insights on the role of glially derived pro-inflammation mediators and neuronal chloride homeostasis

Abstract

Opioid use disorder (OUD) has become a national crisis and contributes to the spread of human immunodeficiency virus (HIV) infection. Emerging evidence and advances in experimental models, methodology, and our understanding of disease processes at the molecular and cellular levels reveal that opioids per se can directly exacerbate the pathophysiology of neuroHIV. Despite substantial inroads, the impact of OUD on the severity, development, and prognosis of neuroHIV and HIV-associated neurocognitive disorders are not fully understood. In this review, we explore current evidence that OUD and neuroHIV interact to accelerate cognitive deficits and enhance the neurodegenerative changes typically seen with aging, through their effects on neuroinflammation. We suggest new thoughts on the processes that may underlie accelerated brain aging, including dysregulation of neuronal inhibition, and highlight findings suggesting that opioids, through actions at the μ-opioid receptor, interact with HIV in the central nervous system to promote unique structural and functional comorbid deficits not seen in either OUD or neuroHIV alone.

Introduction

Opioid use disorder (OUD) in the United States (U.S.) has reached catastrophic proportions and was declared a public health crisis in 2017 [1]. In April 2021, the U.S. Centers for Disease Control and Prevention reported for the first time that drug overdose deaths in the U.S. exceeded 100,000 during the preceding year [2]. Of these, 75,673 were overdose deaths from opioids [2]. Injection drug use increases the likelihood of contracting human immunodeficiency virus (HIV), and opioid misuse and HIV have long been described as interrelated epidemics [3,4]. In fact, the opioid crisis is thought to be a major roadblock to achieving a “cure” of eradicating HIV within the current decade [5]. Additionally, chronic misuse of opioids can exaggerate HIV neuropathology and worsen neurocognitive outcomes of people living with HIV (reviewed in [6]). There have also been considerable reports within the last decade linking exposure to opioids by themselves with the development of age-related neurodegenerative disorders [7–9]. Since infected individuals are living longer due to current improvements in HIV therapeutics, the combined neuroinflammatory effects of opioids and HIV might lead to earlier onset of histopathological/biochemical indices typical of an aging brain. We describe histopathological indices of brain aging in the context of opioids and HIV, and provide an overview of novel neurodegenerative mechanisms that might converge to predispose the comorbid brain to the risk of developing early onset age-related pathologic changes.

Histopathological indices of an aging brain

Brain aging is a term used to describe the anatomic, functional, and histological changes that occur in the brain of an individual as they age. These changes usually become more evident from 70 years of age [10] and are mostly observable in brain regions that are critical to cognition and memory function. In pathologic scenarios, age-related brain changes occur at relatively younger age ranges and at exaggerated levels, and severely impact the daily activity of affected individuals. Most common histopathological changes indicative of brain aging include Alzheimer’s disease-like pathologic inclusions, atrophy, and abnormal vascular changes [10–12].

Alzheimer’s disease-like pathological changes include extracellular aggregation of amyloid β deposits, intraneuronal tangles such as hyperphosphorylated tau, and neurofibrillary tangles, with subsequent accumulation of abnormal, dense-cored, lysosomal structures known as granulovacuolar degeneration bodies [13,14]. Although overt neuronal loss is not always observable in aging brains, subtle changes to neuronal structure such as irregularities in dendritic arborization, reduction in spine density, and alterations in receptor and ion channel trafficking, have been well-documented (reviewed in [1]). These changes contribute to the cognitive decline that is seen in aging. A substantial body of emerging research suggests that chronic inflammation can initiate pathologic Alzheimer’s-like aggregate formation and brain aging [15–18]. Aggregate formation is also promoted by oxidative stress [19] and dysregulation of calcium signaling (reviewed in [20]). Reactive gliosis, aberrant signaling between central nervous system (CNS) immune cells and neurons, cellular senescence, and epigenetic alterations also contribute to accelerated brain aging [21].

Abnormal vascular changes, such as are seen in cerebral small vessel disease, are prevalent during brain aging and are characterized by blood-brain barrier dysfunction and altered cerebral blood flow [22–24]. It is noteworthy that similar vascular changes also occur, although to a lesser extent and in a more delayed manner, during the normal aging process. However, pathologic cases are marked by alteration to the structural and functional integrity of the blood-brain barrier, which results in hypoperfusion, enhanced inflammation due to CNS infiltration of serum-derived content, edema, and eventual cognitive decline [25]. An example of this cerebral vasculature-mediated aging disorder is evident in vascular cognitive impairment and dementia (commonly referred to in the literature as VCID), a disease associated with an age-associated failure of the neurovascular unit and likely worsened by opioid misuse [26]. Changes to the integrity and function of cerebral vasculature can increase inflammation, neuronal injury, and degeneration.

The effects of opioid misuse on neuroinflammation and histopathologic indices of brain aging

Clinically, OUD is associated with cognitive impairment [27], leukoencephalopathy, brain region-specific atrophy [28], and increased hyperphosphorylated tau-containing neurofibrillary tangles are reported in comparison to age-matched controls [7,29,30]. Chronic opioid use also results in lasting cognitive and psychomotor dysfunction including opioid-induced hyperalgesia and hyperkatifeia, and altered reward/saliency processing that can endure despite sustained abstinence [31,32]. These behavioral alterations coincide with lasting structural and epigenetic (e.g., altered DNA methylation and expression of non-coding RNAs) alterations and similarly occur in brain regions implicated in saliency, reward, and motivation [33–35].

Magnetic resonance imaging of the brain in individuals with OUD generally reveals smaller mean relative volumes in specific brain regions of the cerebral cortex, including the prefrontal cortex [28] and basal ganglia [36–38], inferring a loss of neuropil and synapses. Heroin use is associated with white matter imaging hyperintensities that coincide with heightened microgliosis and inflammation [33,39,40]. In a rat model, heroin-induced reductions in axonal transport revealed by manganese-enhanced magnetic resonance imaging coincide with diminished spatial memory and ultrastructural deficits, with many outcomes worsened by prolonged heroin exposure [41]. Importantly, imaging evidence also suggests some deficits in frontocingulate function associated with OUD may improve/reverse with prolonged abstinence [42].

Repeated, nonfatal opioid overdoses can result in accelerated postmortem, age-related or Alzheimer’s disease-like degenerative changes that are [43] similar to observations in fatal opioid overdoses [38]. Histologically, as many as 90% of the postmortem brain samples of individuals with OUD exhibit edema, astrogliosis, and microgliosis especially in the hippocampus, subcortical regions, and white matter compared to non-OUD samples from the same postmortem interval [37,43,44]. The reactive microgliosis [45] is accompanied by increases in proinflammatory cytokines and inflammatory mediators, including TNF-α, IL-1β, and nitric oxide synthase [46]. Sustained, low-level inflammation may be a CNS-specific feature of OUD [47,48], since opioids in the periphery tend to suppress immune function [49,50]. Interestingly, the histopathological profile and brain regions affected can differ with morphine, heroin, oxycodone, methadone, and fentanyl misuse depending on the outcome measure [38]. Although all these opioids are preferential μ-opioid receptor (MOR) agonists, subtle differences in pharmacological profiles at MOR and non-MOR receptors, and unique biased/selective agonists effects of each at MOR [51], undoubtedly underlie differences in their neuropathological properties. For delayed heroin overdose deaths exceeding a survival period of 5 h or more, studies report neurovascular disorders, hypoxic-ischemic leukoencephalopathy, and region-specific atrophy with neuronal losses that can include the hippocampal formation, cerebellum, and olivary nucleus, as well as other areas [28,52].

In rodent models, opioids also increase microgliosis and neuroinflammation [53], including the activation of heme-oxygenase, and increase the generation of reactive nitrogen species (reviewed in [54]). Nitric oxide is a byproduct of sustained opioid exposure [54,55] and interacts with oxyradical products of HIV infection [56] promoting nitrosative damage [6]. The neuroinflammatory effects of opioids are evident with both opium derivatives (e.g., heroin and morphine) and synthetic opioids (e.g., fentanyl). Fentanyl self-administration causes region- and exposure duration-specific alterations in specific interleukins and interferon regulatory genes within the nucleus accumbens and hippocampus of rats [48]. Prolonged oxycodone self-administration (3 years) increases biomarkers associated with neurodegeneration in plasma-derived exosomes in cynomolgus monkeys [57]. Neuronal, microglial, and astroglial-derived exosomes were sampled, and revealed increases in neurofilament light chain, α-synuclein, and microRNAs indicative of neuronal damage. Increases in pathologic phosphorylated forms of tau, amyloid β, and paired helical filaments have also been observed in murine brains following repeated injections of morphine, a prototypical opioid drug with abuse potential [58,59].

The effects of HIV on neuroinflammation and histopathologic indices of brain aging

Although overt dementia following neuroHIV pathology has decreased in the era of cART, a recent study showed that virally suppressed individuals still present Alzheimer’s disease-like changes like amyloid deposits and Alzheimer’s-like symptomatology [60]. Furthermore, reduced brain volume is evident in asymptomatic HIV-infected individuals [61]. Blood-brain barrier impairment and increased neurofilament light chain levels have been observed in CSF in virally suppressed individuals [62], while hyperphosphorylated tau has also been observed in HIV-infected postmortem brain samples [63]. The HIV-1 protein Tat can directly complex with Aβ [64], as well as increase hyperphosphorylated tau, neuronal deposits of paired helical filament [58], and astrocytic amyloid deposition [8] both in vitro and in animal models. These and other pathological markers associated with cognitive aging were also identified in an infectious, humanized mouse model of HIV [65]. Related to our discussion below, the C-C motif chemokine receptor 5 (CCR5) blocker maraviroc substantially improved neuropathology in that study.

MOR and chemokine receptor interactions

There is increasing evidence that CCR5 signaling has an important role in cognitive processes and is a key site of opioid and HIV interaction. When HIV-exposed astroglia are exposed to opioids, opioids can induce synergistic increases in C-C chemokine motif ligand 5 (CCL5). We speculate that the opioid-dependent release of CCL5 by HIV-exposed astrocytes activates CCR5 and initiates a spiraling cascade involving chemokines (especially CCL2) and other inflammatory mediators in both astro- and microglia (reviewed in [6,66]). As described in later sections, this is a major factor in aging-related neuropathology. Deletion of a single CCR5 allele in otherwise normal mice improved cortical plasticity and memory related to hippocampal function, while CCR5 overexpression reduced function in the same tasks [67]. Additional studies with other CCR5 ligands were supportive of these findings [68]. Importantly, multiple methods of blocking CCR5 signaling have been shown to reduce memory impairment in animal models of CNS disorders, including stroke and traumatic brain injury [68,69]. Thus, the concept that alterations in CCR5 signaling by itself may be involved in neuropathology and age-related changes in HIV is well-founded. Mechanisms for how MOR-acting drugs might exacerbate these pathologies are more conjectural, but are based on the known abilities of MOR to interact with chemokine receptors, all of which are G-protein coupled receptors. Notably, in clinical trials assessing maraviroc as an HIV entry inhibitor, the cognitive effects have been mixed (e.g., [70,71]). The differing outcomes may reflect the fact that signaling through CCR5 can vary depending on context. To be beneficial, CCR5 blockade might depend on genetic background or could be limited or promoted depending on the duration of infection, interactions with other therapeutic agents, or comorbidities. Besides CCL5, CCR5 can also be activated by CCL3 and CCL4, which are independently regulated, and each are likely to have unique biased or agonist-selective actions at CCR5.

Opioid receptors are variably expressed by a wide variety of myeloid and lymphoid cell types and opioids have modulatory effects on immune function [49,50]. Typically, this results in immune suppression, but not exclusively [49,50]. Activation of MOR can result in increased HIV replication in infected cells in vitro, and can worsen HIV-induced neuropathology and cognitive damage through bystander inflammatory effects [6]. However, such interactions depend strongly on context, as well as the duration of co-exposure, in terms of whether they promote or inhibit HIV expression [6] (Fig. 1).

Figure 1.

Opioids can exacerbate HIV-1 CNS pathology through direct actions on phenotypically distinct subpopulations of opioid-receptor expressing glia—especially subsets of μ-opioid receptor- (MOR) expressing microglia and astroglia. The illustration above includes known and likely targets of opioid and HIV interactions. Microglia are likely infected through interactions with infiltrating, perivascular macrophages, and propagate the bulk of HIV infection in the CNS. HIV-1 also infects astroglia, but to a far lesser extent, and perhaps without the production of new virus. Infection results in the production of reactive oxygen and nitrogen species, pro-inflammatory cytokines, and the release of HIV-1 proteins such as gp120 and Tat. All of these promote inflammation and cytotoxicity in bystander neurons and glia. OUD alone can cause premature Alzheimer-like changes and morphine by itself can enhance neurotoxicity in vitro; however, opiates appear to potentiate many of the pathophysiological effects of HIV in the central nervous system of infected individuals. μ, δ and κ-Opioid receptor expression varies among phenotypically distinct subpopulations of astro- and microglia and is quite heterogeneous [98,99]—differing among brain regions, throughout ontogeny, and depending on context; many astro- and microglia subpopulations do not express opioid receptors [100–102]. Many of the neurodegenerative effects of opioid-HIV interactions are the result of direct actions on microglia and astroglia, which then lead to a positive feedback cycle of inflammatory/cytotoxic signaling between HIV-1-infected microglia and astroglia. Abbreviations: α-chemokine “C-X-C” receptor 4 (CXCR4); altered or changed (Δ); β-chemokine “C-C” receptor 5 (CCR5); blood-brain barrier (BBB); decreased (↓); fractalkine (CX3CL1); fractalkine receptor (CX3CR1); increased (↑); interferon-γ (IFN-γ); interleukin-6 (IL-6); intracellular Ca²⁺ concentration ([Ca²⁺]_i); intracellular sodium concentration ([Na⁺]_i); monocyte chemoattractant protein-1 (MCP-1 [or CCL2]); peripheral blood mononuclear cells (PBMCs); regulated upon activation, normal T-cell expressed, and secreted (RANTES [or CCL5]); Toll-like receptor (TLR). Fractalkine released by neurons (and astroglia) can be neuroprotective by limiting the neurotoxic actions of microglia (blue “┬”); red arrows suggest pro-inflammatory/cytotoxic interactions. Modified, updated, and reprinted from reference [66] an “Open Access” article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.5/).

Aside from effects on HIV replication itself, the interactive effects of HIV and OUD on HIV infection, age-related brain neuropathogenesis, and negative cognitive outcomes are likely to involve direct interactions between multiple opioid receptors and chemokine receptors. The HIV co-receptors CCR5 and C-C chemokine motif receptor 4 (CXCR4) are well-positioned to participate in interactive outcomes [72,73]. Potential mechanisms for such interactions include heterologous cross-desensitization via downstream signaling events [72], and/or more direct dimeric or heteromeric interactions between opioid receptors and chemokine receptors [72,73]. Both MOR and δ-opioid receptor activation can alter/reduce the response of monocytes to multiple CCR5 ligands (CCL3, CCL4, CCL5) via MOR-dependent protein kinase Cζ activation. The MOR-specific ligand [D-Ala²,NMe-Phe⁴,Gly-ol⁵]-enkephalin (DAMGO) has been reported to upregulate CCR5 and CXCR4 on CD14+ monocytes, enhancing replication of both CCR5- and CXCR4-tropic strains of HIV [74]. Paradoxically, κ-opioid receptor activation is reported to act in an opposite manner to MOR activation. It is reported to inhibit HIV replication in primary human microglia and CD4-positive T cells, and to reduce CCL2 production by human astroglia in vitro (reviewed in [75]). Despite multiple reports that multiple μ, δ, and κ-opioid receptor types may modulate HIV replication and chemokine receptor signaling in isolated cell types grown in vitro, a more consolidated picture emerges in the brain. There, the exacerbated aging-related, neuronal injury and synaptic losses seen with opioids in neuroHIV appear to be largely mediated via MOR-expressing astroglia and microglia [6,76]. Reciprocally, HIV-1 Tat expression is also known to affect morphine’s efficacy and potency in vivo, reducing it depending on the context [77–79]. Interactions with CCR5 clearly play a role in these outcomes since the CCR5 antagonist maraviroc can reinstate morphine antinociceptive potency and restore physical dependence in morphine-tolerant mice exposed to HIV-1 Tat [80].

Both glia and neurons can express CCR5, raising the question of whether both cell types may participate, perhaps differently, in modulating the neurotoxic effects of MOR activation. Recent work in vitro suggests that glia are more prominently involved, since loss of CCR5 from glia (but not neurons) eliminated neurotoxic HIV-1 Tat and morphine interactions [81]. Pretreatment with maraviroc had a similar effect reducing the neuronal [Ca²⁺]_i increase typically observed with Tat ± morphine exposure [81]. Selectively deleting either CCR5 [81] or MOR [76] from glia also completely protected against morphine’s ability to exacerbate Tat neurotoxicity. Unexpectedly, deleting CCR5 from glia entirely protected neurons from Tat toxicity alone. This outcome appeared to involve altered processing of the brain-derived neurotrophic factor (BDNF) precursor in a way that favored production of mature BDNF [81], a known neuroprotective agent [82,83]. In support of the neuroprotective effect of BDNF in this model, intranasally administered BDNF-conjugated clathrin triskelia reversed HIV Tat-induced deficits in hippocampal neurogenesis, synaptodendritic complexity, and learning and memory in transgenic mice [84]. Collectively, the findings suggest that loss of CCR5 may fundamentally change MOR signaling in HIV-exposed glia in a BDNF-dependent manner. Overall, the interaction of opioid and chemokine receptors, specifically MOR and CCR5, may alter the neuropathogenesis of HIV in a qualitatively unique manner not seen with either disorder alone.

Dysregulation of KCC2, Cl⁻ homeostasis, and GABAergic function: A novel pathway to neuroHIV and aging-related outcomes?

K⁺-Cl⁻ cotransporter 2 (KCC2) is the neuron-specific, major transporter responsible for maintaining [Cl⁻]_i homeostasis and γ-aminobutyric acid (GABA) inhibitory function in the adult CNS. KCC2 functions to extrude Cl⁻ from neurons, maintaining the low [Cl⁻]_i necessary for GABA type A receptor- (GABA_AR) induced membrane hyperpolarization. The functionality of KCC2 is determined by its proximity to the cell membrane and ability to oligomerize and is regulated by multiple phosphorylation sites. Serine 940 (S940) is a key residue that can be phosphorylated by protein kinase C to increase KCC2 trafficking and stability at the cell membrane. Reductions in KCC2 expression or function can increase [Cl⁻]_i, reduce the hyperpolarization normally mediated via GABA_AR, and can exaggerate the excitatory effects of glutamate [85]. Altered KCC2 function is implicated in multiple neurological disorders [86], including epilepsy, and enhancing KCC2 activity with CLP257 or its prodrug, CLP290 can alleviate key symptoms both pre-clinically and clinically, via a PP1-mediated mechanism.

KCC2 is a target of Tat, gp120, and HIV in primary human neurons in vitro and in HIV-1 Tat transgenic mice [87] and by logical extension, may be implicated in neuronal dysfunction in neuroHIV. Opioids also diminish KCC2 function and increase [Cl⁻]_i [87–89], and the combination of opioids and HIV exacerbate KCC2 and [Cl⁻]_i dysregulation experimentally [87] (Fig. 2). Both BDNF [90] and glially derived cytokines such as IL-1β [91,92], are potent regulators of KCC2. We speculate that reduced KCC2 function may underlie specific interactive effects of opioids and HIV. Loss of KCC2 activity may result from inflammation, especially excessive IL-1β, which affects dendrite and spine morphology, synaptic integrity, and excitatory/inhibitory balance [93], and is known to delay the normal developmental upregulation of KCC2 levels [91].

Figure 2. Effects of opioid- and/or HIV− induced deficits in KCC2 expression and/or function on neuronal excitability.

Our working model is that KCC2 pumps Cl⁻ out of mature neurons allowing GABA_AR activation to promote Cl⁻ entry and neuronal inhibition. HIV and opioid-induced decreases in KCC2, which increase [Cl⁻]_i result in neuronal depolarizing upon GABA_AR activation and excess excitation in the presence of glutamate agonists (not shown). NKCC1 (which pumps Cl⁻ into neurons) function tends to be constant and unaffected by extracellular factors or opioids and HIV.

Integrative BDNF, IL-1β, and KCC2 effects on indices of accelerated brain aging

In the sections above, and as summarized in Fig. 2, there are multiple pathways that intersect to drive the neuropathology and cognitive damage described as accelerated brain aging in HIV and OUD, either individually or interactively. HIV infection and/or drugs that are typical of OUD, mostly acting at glial cells, start a chain of events leading to overproduction of inflammatory factors. Among these is IL-1β, a potent regulator of KCC2 [91–94]. KCC2 maintains low [Cl⁻] inside of mature neurons, which allows Cl⁻ entry and neuronal inhibition upon GABA_AR activation. HIV and/or opioids decrease KCC2 levels, thus increasing [Cl⁻] within individual neurons. GABA_AR activation thus results in less inhibition, and in extreme situations GABA_AR activation can depolarize neurons. Overall, this drives excitotoxicity and damage in brain tissue. In addition, HIV/opiates also can independently reduce key neurotrophic factors, including mature BDNF, and increase multiple neurotoxic factors, including proBDNF. While this alone would hasten both structural and functional damage, depletion of mature BDNF also leads to disrupted processing of amyloid precursor proteins and accumulation of pathologic age-related proteins such as tau and amyloid β [95]. Amyloid β can further dysregulate Cl⁻ homeostasis by increasing levels of NKCC1 [96], which directly opposes the effects of KCC2. Of importance to the above discussion, enhancing BDNF levels can increase KCC2 levels and restore both the Cl-equilibrium potential and GABAergic signaling in hippocampal area CA1, while improving learning in the Morris water maze in an Alzheimer’s disease model [97]. BDNF may thus be protective against aging-related cognitive changes in HIV/OUD by limiting neural injury through multiple pathways.

Summary

HIV-infected individuals can now expect a relatively normal lifespan but may experience aging-related changes in brain structure and function earlier and with more severity than non-infected individuals. Brain inflammation, and deficits in BDNF, which continue even in the absence of significant viral replication, are likely contributing factors. The underlying pathology appears to be enhanced by OUD, an outcome that might involve direct interactions between opioid and chemokine receptors, especially CCR5. A more speculative concept, but one relatively well-accepted for other diseases with cognitive dysfunction, is that opioid-mediated increases in cytokine production and altered BDNF processing in HIV-infected glia cause deficits in KCC2 function, thus leading to loss of neuronal [Cl⁻]_i homeostasis. Experimental work suggests that BDNF might be an effective target to protect against combined HIV and OUD effects.