A central role for glial CCR5 in directing the neuropathological interactions of HIV-1 Tat and opiates

Human immunodeficiency virus type 1 (HIV-1) remains a global epidemic [1]. Despite significant antiretroviral suppression of HIV-1 propagation in the periphery, limited penetration of combination antiretroviral therapy (cART) drugs through the blood-brain barrier [2, 3] as well as the early viral integration into the host genome cultivates a reservoir in which low levels of viral replication can be sustained in the central nervous system (CNS) [4‐6]. Thus, HIV-1+ individuals are especially vulnerable to CNS injury, which afflicts as many as 50% of this population [7‐9]. The neurological consequences of HIV-1 infection are known collectively as HIV-associated neurocognitive disorders (HAND). HAND presents as a spectrum of deficits ranging from mild or asymptomatic cognitive disorders to severe, HIV-associated dementia and includes a variety of cognitive, behavioral, and/or motor symptoms [10]. Postmortem findings in HIV-infected individuals, even those effectively treated with cART, often include signs of prominent CNS inflammation, such as increased numbers and/or activation of microglia and astroglia, perivascular inflammation, and leukocytic infiltration resulting in marked neuronal degeneration [11‐13]. Notably, neuronal injury and alterations in signaling are not accompanied by direct viral infection of neurons [14‐16]. Instead, microglia and macrophages are the major source of productive viral infection in the CNS [17‐19]. Small numbers of astrocytes are infected and can produce toxic proteins that injure bystander neurons, but they have not been reliably shown to produce virus [20, 21]. These combined findings highlight the importance of glial impact on neurons, which is normally critical in maintaining proper neuronal activity and survival and suggests a mode of indirect injury that is a consequence of the innate CNS immune response to HIV.

HIV-1 infection and injection drug use are interlinked epidemics, due in large part to needle sharing and increased risky sexual behavior. Because heroin (diacetyl morphine) is widely abused and its active metabolite, morphine, is an opiate prescribed for pain syndromes experienced by HIV patients, we and others are interested in neurological interactions of HIV-1 and heroin/morphine. The comorbid effects of HIV and opiates are not trivial. HIV+ individuals who also abuse opiates demonstrate more severe neuropathology than those who do not, and these findings can translate to exacerbated and accelerated HAND [22‐25]. The actions of morphine in this context occur primarily through the activation of μ-opioid receptors (MORs) expressed on glial cells. MOR activation results in the potentiation of HIV-induced release of pro-inflammatory factors (e.g., TNFα, IL-1β, IL-6, CCL5, and CCL2), as well as oxidative and nitrosative stress, mitochondrial dysregulation, elevated intracellular calcium levels, and excess extracellular glutamate (via restriction of astroglial glutamate uptake), all of which promote neurotoxicity [26‐30]. Dysregulated release of inflammatory factors upon chronic exposure to viral proteins may recruit and activate more immune cells, propagating a cycle of increasing inflammation with significant downstream neuronal consequences.

One receptor vulnerable to the aforementioned dysregulation by HIV infection is C-C chemokine receptor 5 (CCR5). CCR5 is widely expressed on T lymphocytes, macrophages, microglia, and dendritic cells and plays a critical role in inducing migration of immune cells to sites of infection and injury in response to elevated levels of certain C-C-chemokine ligands (MIP-1α/CCL3, MIP-1β/CCL4, CCL5/RANTES) [31]. CCR5 and its ligands are upregulated during HIV infection, leading to excess activation of CCR5-expressing cells, including CNS microglia and astrocytes [32, 33]. A homozygous deletion of 32 base pairs in the CCR5 gene prevents its expression on the cell surface and confers improved but not absolute immunity to infection with R5-tropic strains of HIV-1 [34]. Individuals carrying the allele show slowed disease progression upon infection with HIV and less cognitive impairment [35‐38]. Furthermore, maraviroc, a CCR5 antagonist with relatively high CNS penetrance [39], reduces microglial activation in the simian immunodeficiency virus-infected model to uninfected control levels and reduces the expression of several pro-inflammatory factors [40]. cART regimens that are supplemented with maraviroc improve the neurocognitive status of HIV+ patients and reduce CSF levels of TNFα [41, 42]. Morphine can alter CCR5 expression by monocytes and activated T cells, contributing to increased viral entry and replication, and excess CNS immune activation [43, 44]. We previously demonstrated that loss of CCL5, a CCR5 ligand, prevented widespread glial activation and reduced levels of another inflammatory ligand (CCL2), suggesting CCL5 may be an upstream activating signal that promotes the expansion of downstream pro-inflammatory responses [45]. The present studies investigate whether interrupting CCR5 signaling may protect neurons against the comorbid effects of HIV-1 and opiate exposure, apart from any effect of blocking HIV entry. We demonstrate using mixed glial-neuronal co-cultures that morphine potentiates Tat-induced neuronal death and that a loss of CCR5 expression on glial cells rescues neurons from such enhanced neurotoxicity. Surprisingly, morphine completely protected against Tat neurotoxicity in cultures with CCR5-null glia even though Tat by itself was still toxic. Levels of brain-derived neurotrophic factor (BDNF) are reduced in HIV+ individuals and by glycoprotein 120 (gp120), an HIV-1 envelope protein with neurotoxic properties [46, 47]. We found that the ratio of neurotrophic BDNF to its neurotoxic precursor (proBDNF) was altered in CCR5-null glia exposed to Tat and morphine co-treatment such that the environment favored neuronal support. BDNF also rescued neurons from Tat + morphine neurotoxicity in a manner similar to the loss of glial CCR5 expression. Overall, we postulate that CCL5/CCR5 signaling is a point of convergence for opiate-Tat interactions within the inflammatory milieu of the HIV-infected CNS. Blocking CCR5 appears to enhance neuroprotection, perhaps by increasing BDNF-related neuroprotection. Inactivation or loss of CCR5 may also change heterologous interactions between MOR and CCR5 related to toxicity and protection.

Experiments were conducted in compliance with procedures reviewed and approved by the Virginia Commonwealth University Institutional Animal Care and Use Committee.

Mild-to-moderate HAND still occurs in roughly 50% of HIV-infected individuals who receive typical antiretroviral therapy, in part due to relatively low penetration of cART drugs through the blood-brain barrier but also due to the extended lifespan afforded by long-term cART treatment. Even in patients where viral replication is undetectable, the release of toxins such as HIV-1 Tat, an early viral protein required for the transactivation of HIV transcription, can persist. While Tat may have direct excitotoxic effects on neurons, particularly well-documented in purified neuronal cultures [30, 49], it also fosters an environment of chronic inflammation through effects on astroglia and microglia. This may result in indirect neurotoxicity through the production of reactive oxygen and nitrosative species and via the release of proinflammatory chemokines and cytokines [50‐52], triggering a cascade of inflammatory events that is ultimately damaging to neurons. Furthermore, Tat is present in measurable levels circulating in the blood and CSF of HIV patients [53, 54], suggesting that it is a clinically relevant surrogate for some aspects of HIV neuropathogenesis, especially in virally suppressed patients.

CCR5 is a chemokine receptor that also functions as a major co-receptor for HIV entry. It is normally expressed at low levels by glia throughout the CNS, but this expression can be upregulated by HIV-1 and by Tat [55]. CCR5 expression can also be upregulated in an additive or interactive manner with co-exposure to Tat and the preferential MOR agonist morphine [56]. We have been interested in understanding the role of the CCL5-CCR5 axis in mediating HIV and Tat-induced inflammation and neurotoxicity, with a special interest in whether moderating this chemokine system might be useful in reducing neurotoxic interactions between HIV-1 Tat and opiate exposure. We showed that intrastriatal delivery of Tat caused a local inflammation, characterized by astroglial expression of CCL2 (MCP-1) and 3-nitrotyrosine expression in microglia that was exacerbated by co-administering time-release, subcutaneous morphine pellet implants. Microglial 3-nitrotyrosine and astroglial expression of CCL2 were not seen in a CCL5 (RANTES)-deficient mouse, strongly implicating CCL5-to-CCR5 signaling in the amplification of astroglial CCL2 production and resultant macrophage/microglial activation and recruitment [45]. In a related study, we also showed that pre-treating Tat-inducible transgenic mice with maraviroc attenuated the withdrawal-mediated increase in levels of many of the cytokines that occurred in mice co-exposed to Tat and morphine and restored antinociceptive properties of morphine that were attenuated by Tat [57]. CCR5 loss is also protective against other HIV proteins, including gp120 [58], as assessed by reduced neuronal damage and microglial activation.

The present studies used both genetic (glia and neurons derived from a constitutive, CCR5-deficient mouse) and pharmacologic (CCR5 blocker maraviroc) approaches to demonstrate that CCR5 is directly involved in HIV neuropathogenic mechanisms irrespective of viral titers. That CCR5 plays a role in the inflammatory processes associated with many disease states, such as cerebral ischemia and reperfusion injury, and neuropathic pain is well-established [59, 60]. In HIV, the role of CCR5 in neurologic outcomes is more difficult to assign since CCR5 is a major co-factor for the infection process. A naturally occurring, 32 base-pair deletion in the CCR5 gene produces a nonfunctional protein, effectively rendering a knockout of the receptor. HIV+ individuals who possess one copy of this mutation show slower progression to AIDS and a reduced occurrence of HAND than those without the mutation [36, 37]. Studies in which maraviroc supplemented the antiretroviral therapy regimen in virally suppressed patients led to improvements in cognition [41, 61]. However, due to the role of CCR5 as a co-receptor for HIV entry, it is unclear if the reduced prevalence of HAND in this population is due to decreased inflammation or decreased progression of infection. In this study, we isolated the inflammatory effects of HIV-1 Tat without the confounding effects of viral infection and replication in monocytes and microglia. Since our paradigm is non-infectious, the findings establish that blocking CCR5 activity can significantly improve neuronal outcomes by mechanisms not involving reduced HIV infection.

As we hypothesized, when CCR5 was deleted from glia in mixed glial cultures, co-cultured neurons were protected against the enhanced degree of toxicity resulting from co-exposure to Tat and morphine. Surprisingly, when CCR5 was absent from glia, morphine appeared to entirely protect co-cultured neurons from the effects of Tat (Fig. 3b). These findings were further explored and confirmed in an alternative, pharmacological approach utilizing maraviroc to control the length of CCR5 blockade. When maraviroc was given to more mature co-cultures concurrent with Tat ± morphine treatments for 72 h, protection was seen against the interactive effects of Tat and morphine co-exposure. In co-cultures incubated with maraviroc for their entire developmental period in vitro, to mimic a constitutive knockout, neurons were additionally protected from Tat-induced toxicity when morphine was present (Fig. 4). Maraviroc may thus afford additional neuroprotection to individuals exposed to opioids, depending on timing and duration of maraviroc therapy. Notably, maraviroc blocked the increase in neuronal [Ca²⁺]_i seen at acute time points after exposure to Tat ± morphine (Fig. 5), suggesting that early events leading to neuronal dysfunction may be CCR5-dependent and reversible. Naloxone pretreatment revealed that both the predicted and unexpected protective effects of morphine were mediated by opioid receptors.

We explored a potential mechanism for the aforementioned protective effects against both Tat and Tat + morphine. BDNF, a neurotrophin that modulates development and survival of young neurons and is important in establishing and maintaining normal synaptic connectivity, has surfaced as a potential therapeutic target in HAND. Both microglia and astrocytes can be a major source of released BDNF [62, 63]. The mature form of BDNF (mBDNF) preferentially binds TrkB, through which it initiates PI3K and MAPK signaling pathways in neurons, as well as NF-κB-mediated activation of transcription factors [64, 65]. proBDNF binds the alternative p75NTR to initiate pro-apoptotic cascades [66]. Levels of mBDNF are reportedly reduced in HIV+ individuals and also modulated by drugs of abuse [47, 67, 68]. For example, morphine exposures in vivo appear to increase the extracellular protease tissue plasminogen activator, thus increasing cleavage of proBDNF to mBDNF. Withdrawal from morphine had the opposite effect [69]. Intracellular conversion of proBDNF to mBDNF in neurons is reduced by the HIV surface envelope protein gp120, which reduces levels of the intracellular protease furin [47]. Exogenous mBDNF can rescue neurons exposed to gp120 both in vivo and in vitro [70, 71], likely involving downregulation of CXCR4. In a prior study involving glia exposed to HIV-infective supernatant ± morphine, mBDNF levels, but not glial-derived neurotrophic factor (GDNF) levels, were significantly decreased. This reduction was reversed upon removal of HIV treatment [48]. The relatively higher proBDNF/mBDNF ratio measured in the medium of wild-type versus CCR5-null glial cultures exposed to Tat and morphine (Fig. 7c (iii)) is in line with the outcomes above since neurons co-cultured with CCR5-null glia were protected. We measured significant differences in proBDNF/mBDNF ratios at 6 h with trends at 24 h. These time points reflect prior studies that have demonstrated transient increases in BDNF mRNA and protein levels. For instance, response to cortical injury in rats involves an increase in BDNF mRNA from 1 h that begins to decline after 24 h [72, 73]. gp120 also leads to early changes in BDNF and proBDNF release from cerebellar or cortical neurons, from 1 h post-treatment [47]. As demonstrated by the persistently high percentage of neuronal survival in BDNF-treated co-cultures (Fig. 4), these rapid, transient neuroprotective signals may result in long-lasting effects even in the face of a complex milieu of secondary and tertiary factors by 72 h of treatment exposure. Activation of p75NTR, for instance, in multiple cell types has elucidated a role in immune response regulation. Injury to retinal ganglion neurons leads to the upregulation of p75NTR of nearby Müller glial cells, which can activate downstream cytokine production that is thought to eventually lead to the damage and demise of bystander neurons [74]. Our findings are further supported in other models of neuronal injury such as cerebral ischemia, in which the loss of CCR5 reduces long-term inflammatory injury, potentially through increases in BDNF levels [60, 75]. However, while we demonstrated that Tat and/or morphine treatment led to specific changes in mBDNF and proBDNF levels, shifts in p75NTR and TrkB levels may add a degree of complexity in determining the fate of the neurons, as receptor expression is responsive to injury in a time- and cell-dependent manner [76‐78].

Opiates and HIV have historically been interlinked epidemics, and injection drug use carries an increased risk of contracting HIV. Moreover, opiates are used to manage HIV-related pain syndromes, which may impact HAND symptoms in virally suppressed individuals. Cooperative effects between HIV and opiates that increase CNS inflammation are well-documented, and many involve the HIV co-receptor CCR5. For example, opiates modulate CCR5 expression in a number of CNS and immune cells, including microglia [56], astrocytes [79], and peripheral monocytes [43], and can thereby increase rates of HIV infection as well as pro-inflammatory products and immune cell recruitment. The analgesic properties of MOR activation are also reduced when there is an abundance of CCR5 ligands [80]. MOR function is also altered by Tat exposure, which lowers morphine efficacy through decreased G-protein activation [81]. Reduced MOR signaling would tend to increase the amounts of opiates used by HIV-infected individuals for effective analgesia or for illicit effects.

In earlier studies, wild-type neurons co-cultured with MOR-deficient glia were completely protected against the exaggerated neurotoxic effects of co-exposure to Tat + morphine. These result were consistent with findings that the selective MOR antagonist β-funaltrexamine prevented morphine from exacerbating Tat-induced increases in proinflammatory cytokine transcripts in astroglia [27]. Surprisingly, morphine also afforded neuroprotection against Tat [30]. Given that the present results show strikingly similar protective effects in the absence of glial CCR5 (Fig. 3b), a possible interaction between CCR5 and MOR signaling that modulates neuroprotection seems likely. Some interactions between opiates and CCR5 are likely to involve the heterologous interactions that can occur between G-protein-coupled receptors (GPCRs) and that have been shown for MOR and CCR5 [82, 83]. Heterologous desensitization may explain a reduced chemotactic effect of CCR5 by pretreatment with MOR agonists [80]. Recently, synthetic, bivalent ligands composed of maraviroc and naltrexone that target the proposed CCR5-MOR heteromers/oligomers have shown that such interactions may occur in a cell type-specific manner [84‐86]. The inability to form MOR-CCR5 heterodimers in CCR5-null (or MOR-null) glia may alter the cellular response to concurrent morphine exposure, contributing to neuroprotection. Although such a response might additionally involve contributions from κ-(KOR) or δ-(DOR) opioid receptors, morphine has lower affinity at KOR or DOR (than MOR). Prior studies also indicate that MOR is largely responsible for the morphine and Tat-induced inflammation and bystander neurotoxicity that is mediated by glia [28, 30].

This study establishes a role of glial CCR5, unrelated to infective processes, in mediating neurotoxicity due to HIV-1 Tat and the interactive effects of Tat and morphine. The differential toxic versus protective effects of morphine in the presence or absence of CCR5 hint at complex relationships that may involve heterologous interactions between CCR5 and MOR. Other results infer a role for BDNF, and perhaps an altered balance between proBDNF and mBDNF levels, in some aspects of the protection. These studies are particularly pertinent to HIV+ individuals who are virally suppressed, yet still develop mild neurocognitive deficits via ongoing, low levels of neuroinflammation in the CNS that involve CCR5 activation.

HIV antiretroviral therapy has been successful in limiting AIDS-related complications, but neurocognitive deficits due to the inflammation driven by early CNS viral penetration persist in up to 50% of HIV-infected individuals. Opiate exposure, which is common in the HIV+ population, worsens the severity of HAND. The cellular mechanism(s) that exacerbate inflammation and CNS disease remain largely unexplained. We tested the hypothesis that the CCL5-CCR5 axis plays a pivotal role in opiate exacerbation of HIV deficits using both genetic and pharmacological approaches. Loss of glial, but not neuronal, CCR5 protected striatal neurons from HIV-1 Tat and morphine co-exposure. Importantly, loss of CCR5 not only blocked the interactive effects of Tat and morphine, but also reversed the Tat toxicity seen in the cultures as long as morphine was present. These findings were confirmed with both long- and short-term exposure to the CCR5 antagonist maraviroc. We measured a shift in the ratio of proBDNF/mBDNF released from CCR5-deficient glia that favored neuroprotection, and exogenous BDNF protected neurons from HIV-1 Tat and morphine exposure in a manner that mirrored the effects of CCR5 deficiency. Our findings suggest that CCR5 is involved in processes that impact neuronal survival and function unrelated to its important role in HIV infection, at least partly by altering the balance between proBDNF and mBDNF levels. The differential toxicity of morphine/opioids in the presence or absence of CCR5 also hints at complex relationships between CCR5 and MOR that may involve heterologous interactions. Our findings are pertinent in terms of understanding and treating HIV+ individuals who develop neurocognitive deficits even though their peripheral viremia is well-controlled.