HIV Neuropathogenesis in the Presence of a Disrupted Dopamine System

Graphical Abstract

Introduction

While the development of antiretroviral therapy (ART) has significantly changed the nature of HIV disease, it remains a major public health issue, affecting more than 37 million people worldwide (Lucas and Nelson 2015, Sanchez and Kaul 2017). The effectiveness of ART has transformed HIV from a terminal to a chronic illness. In the current era, people living with HIV (PLWH) have longer lifespans and often healthier outcomes than in previous decades (Cihlar and Fordyce 2016, Saylor, Dickens et al. 2016, Eggers, Arendt et al. 2017, Boender, Smit et al. 2018). However, ART can only inhibit viral replication but does not eliminate the virus. While ART ameliorates many of the sequelae that result from HIV, it does not cure the disease (Cihlar and Fordyce 2016, Ulfhammer, Eden et al. 2018). The resulting chronic infection, as well as long term ART treatment, has engendered a variety of new health issues, including many associated with HIV infection of the central nervous system (CNS) (Grabar, Weiss et al. 2006, Heaton, Clifford et al. 2010, Treisman and Soudry 2016).

Infection of the CNS occurs within days after peripheral infection (Davis, Hjelle et al. 1992, Valcour, Chalermchai et al. 2012), and can result in the development of several distinct neuropathologies, as well as HIV-associated neurocognitive disorders (HAND). Globally, the prevalence of HAND ranges from 15 – 70% of PLWH despite ART, with a higher prevalence more common in resource poor areas (Heaton, Clifford et al. 2010, Heaton, Franklin et al. 2015, Saylor, Dickens et al. 2016, Saloner and Cysique 2017, Yusuf, Hassan et al. 2017, Gascon, Vidal et al. 2018, Debalkie Animut, Sorrie et al. 2019). Many studies have shown that these neurologic sequelae, collectively known as neuroHIV, are altered by substance abuse (Arnsten, Demas et al. 2002, Martin, Pitrak et al. 2003, Lucas, Griswold et al. 2006, Baum, Rafie et al. 2009, Byrd, Fellows et al. 2011, Devlin, Gongvatana et al. 2012, Meyer, Rubin et al. 2013, Chang, Connaghan et al. 2014, Small, Milloy et al. 2016). However, the impact of distinct substances of abuse on the development of neuroHIV remains unclear, or in some cases controversial (Ferris, Mactutus et al. 2008, Byrd, Fellows et al. 2011, Chang, Connaghan et al. 2014).

While different classes of drugs have distinct mechanisms of action, all addictive substances acutely increase CNS dopamine concentrations in the mesocorticolimbic pathway (Di Chiara and Imperato 1988, Carboni, Imperato et al. 1989, Pierce and Kumaresan 2006, Kleijn, Folgering et al. 2011, Clarke, Adermark et al. 2014). This suggests that elevated CNS dopamine may be a common mechanism by which different types of drugs impact the development of neuroHIV. These effects may be most pronounced during early acute and/or intermittent use of substances of abuse, as chronic drug abuse downregulates dopamine release (Volkow, Wang et al. 1996, Volkow, Wise et al. 2017). Changes in dopamine-rich brain regions have long been associated with HIV infection within the CNS, and although these effects are often subtler in the ART era, they still persist (Larsson, Hagberg et al. 1991, Aylward, Henderer et al. 1993, Hestad, McArthur et al. 1993, Wiley, Soontornniyomkij et al. 1998, Berger and Arendt 2000, Bell, Anthony et al. 2006, Gelman, Spencer et al. 2006, Scheller, Arendt et al. 2010, Becker, Sanders et al. 2011, Kumar, Ownby et al. 2011, Wright, Pyakurel et al. 2016, Deren, Cortes et al. 2019). Defining more precisely the dopamine-associated impact of both legal and illicit drugs on the development of HIV associated comorbidities is a critical need, as the number of HIV-infected individuals with substance use disorder (SUD) continues to increase, particularly among older individuals, the fastest growing segment of the infected population (Grabar, Weiss et al. 2006, Han, Gfroerer et al. 2009, Kirk and Goetz 2009, Skalski, Sikkema et al. 2013, Deren, Cortes et al. 2019). Additionally, the specific interactions between SUD and ART, and the pathological consequences of those interactions, must also be addressed and understanding the interplay between ART and substances of abuse will be essential to improving HIV treatment in vulnerable populations. However, the study of HIV infection in the presence of a disrupted dopaminergic system is complex and technically challenging, and examining this interplay effectively will require refinements or even shifts in hypotheses, approaches, and technique. In this review, we will address these issues by briefly describing what is known about the impact of dopamine on the development of CNS infection and neuropathogenesis, examining experimental approaches, and then discussing treatment strategies that may decrease the incidence and/or severity of neuroHIV in PLWH with SUD.

Neuropathogenesis of HIV in the era of ART

Infection of the CNS occurs within 1 – 2 weeks of peripheral infection (Davis, Hjelle et al. 1992, Valcour, Chalermchai et al. 2012), primarily through the transmigration of infected mature CD14⁺CD16⁺ monocytes across the blood brain barrier (BBB) (Crowe, Zhu et al. 2003, Williams, Calderon et al. 2013, Williams, Veenstra et al. 2014). Within the brain, both infected monocytes and the monocyte-derived macrophages into which they can mature produce HIV and inflammatory mediators, infecting and activating CNS myeloid cells, including microglia and perivascular macrophages (Koenig, Gendelman et al. 1986, Minagar, Shapshak et al. 2002, Rappaport and Volsky 2015). These cells are the primary targets of HIV in the CNS, and their infection is significant to the development of neuroHIV, as well as to the establishment and maintenance of viral reservoirs in the brain (Yadav and Collman 2009, Churchill and Nath 2013, Honeycutt, Wahl et al. 2016, Castellano, Prevedel et al. 2017, Honeycutt, Thayer et al. 2017, Abreu, Veenhuis et al. 2019).

Both infected and activated myeloid cells produce inflammatory cytokines and chemokines, such as TNF-α, IL-6, IL-1β, CCL2, CXCL10 and CXCL12 (Laskin and Pendino 1995, Brabers and Nottet 2006, Arango Duque and Descoteaux 2014). This leads to recruitment and activation of additional myeloid populations, as well as other immune cells, and to BBB damage, resulting in chronic neuroinflammation (Minagar, Shapshak et al. 2002, Eugenin, Osiecki et al. 2006, Kraft-Terry, Buch et al. 2009, Saylor, Dickens et al. 2016). These cells may also release neurotoxic viral proteins, such as Tat or gp120, that exacerbate the neurotoxic environment (Sanchez and Kaul 2017). In addition, a number of studies suggest that HIV also infects astrocytes at low levels (Eugenin and Berman 2007, Churchill, Wesselingh et al. 2009, Eugenin, Clements et al. 2011), suggesting these cells may also act as a viral reservoir and contribute to neuroinflammation and CNS damage. Astrocytes are the most numerous glial cells in the CNS and modulate immune function and maintain neuronal health (Bylicky, Mueller et al. 2018). Therefore, any HIV-associated changes in astrocyte function would be important to HIV neuropathogenesis, although characterization of their specific impact is not clear and requires further study.

Prior to the introduction of ART in 1996, infection of the CNS led to a variety of neuropathologies including meningitis, microglial nodules, white matter lesions, multinucleated giant cells, myelin loss, HIV encephalitis, reactive astrogliosis, regional atrophy and significant neuronal injury and death (Navia, Cho et al. 1986, Kure, Llena et al. 1991, Hestad, McArthur et al. 1993, Everall, Hansen et al. 2005, Anthony and Bell 2008). These effects, which were found in approximately 50% of infected individuals, resulted in many symptoms of subcortical dementia, including cognitive decline, decreased mental processing and executive function, apathy, and sometimes Parkinsonian-like symptoms (Dube, Benton et al. 2005, Everall, Hansen et al. 2005, Heaton, Franklin et al. 2011). Although the percentage of PLWH who present with neurocognitive symptoms is similar with or without ART, the severity of the neurocognitive impairment is most often diminished with treatment (Bell 2004, Cysique and Brew 2009, Saylor, Dickens et al. 2016, Walker and Brown 2018). Prior to ART, extensive neurocognitive impairment was relatively common, with HIV-associated dementia (HAD) found in 15 – 20% of infected individuals. In the current era, HAD is found in fewer than 2% of PLWH (Heaton, Clifford et al. 2010, Heaton, Franklin et al. 2011, Saylor, Dickens et al. 2016).

The use of ART has also significantly increased lifespans and quality of life of PLWH (Cihlar and Fordyce 2016, Boender, Smit et al. 2018). Unfortunately, PLWH who use ART often still experience a variety of behavioral and cognitive symptoms, as well as substantial neuropathology, and the prevalence of mild and asymptomatic forms of neurocognitive impairment have increased (Valcour, Shikuma et al. 2004, Heaton, Clifford et al. 2010, Becker, Sanders et al. 2011, Heaton, Franklin et al. 2011, Saylor, Dickens et al. 2016). While ART-treated individuals have decreased neuronal death as compared to PLWH in the pre-ART era, neuronal dysfunction is still common in these individuals (Bell 2004, Everall, Hansen et al. 2005, Cysique and Brew 2009, Heaton, Franklin et al. 2011, Clifford, Samboju et al. 2017, Walker and Brown 2018). A number of recent studies also suggest that neuroHIV involves the disruption of neurocircuitry, with significant alterations in the strength and efficacy of connections, particularly in resting state functional connectivity. These effects are present irrespective of ART, and many are seen in the frontostriatal pathway, which also mediates the effects of dopamine (Itoh, Mehraein et al. 2000, Koutsilieri, ter Meulen et al. 2001, Gelman, Spencer et al. 2006, Becker, Sanders et al. 2011, Gongvatana, Harezlak et al. 2013, Ann, Jun et al. 2016, Bak, Jun et al. 2018, Walker and Brown 2018).

All of these symptoms often overlap with or exacerbate the age-related neurological co-morbidities that are increasingly prevalent in the aging, HIV-infected population in the ART era (Manfredi 2002, Grabar, Weiss et al. 2006, Valcour, Paul et al. 2011, Watkins and Treisman 2012, Clifford, Samboju et al. 2017). In the US, approximately 18% of individuals newly diagnosed with HIV are over 50 (Deren, Cortes et al. 2019), with these people exhibiting more advanced disease (Grabar, Weiss et al. 2006, Kirk and Goetz 2009). Older PLWH are also more frequently diagnosed with SUD than non-infected individuals of the same age, potentially exacerbating the development of neuroHIV through changes in the dopaminergic system (Byrd, Fellows et al. 2011, Lyons, Pitts et al. 2013, Skalski, Sikkema et al. 2013, Deren, Cortes et al. 2019, Matt and Gaskill 2019). Substance abuse may have a more extensive impact on older PLWH, as several studies show older individuals who use methamphetamine have greater impairment in more cognitive domains than non-users (Iudicello, Morgan et al. 2014, Montoya, Cattie et al. 2016, Minassian, Henry et al. 2017). Substance use disorders can also interfere with effective treatment, disrupting scheduled care and adherence to ART, and may interfere with the action of some antiretroviral medications (Becker, Thames et al. 2011, Anderson, Higgins et al. 2015, Azar, Wood et al. 2015, De Boni, Shepherd et al. 2016). These issues are further compounded by pill burden and polypharmacy, which are common in older adults, increasing medication errors, drug-drug interactions, or toxicity (Edelman, Gordon et al. 2013, Deren, Cortes et al. 2019). Thus, characterizing the specific impact of dopamine on the etiology of neuroHIV is essential to identifying and developing adjunctive therapies for PLWH with SUD, especially considering the number of older individuals in this group.

How the dopaminergic system is disrupted during HIV infection

Dopaminergic dysfunction has been associated with the development of CNS infection since early in the epidemic, and dopamine-rich brain regions appear to be particularly vulnerable to the effects of HIV (Navia, Cho et al. 1986, Kure, Llena et al. 1991, Larsson, Hagberg et al. 1991, Reyes, Faraldi et al. 1991, Aylward, Henderer et al. 1993, Hestad, McArthur et al. 1993, Fujimura, Goodkin et al. 1997, Wiley, Soontornniyomkij et al. 1998, Itoh, Mehraein et al. 2000, Valcour, Sithinamsuwan et al. 2011). Prior to ART, neuropathology was prominent in these regions, including the striatum, the substantia nigra and the prefrontal cortex (Reyes, Faraldi et al. 1991, Aylward, Henderer et al. 1993, Hestad, McArthur et al. 1993, Wiley, Soontornniyomkij et al. 1998, Itoh, Mehraein et al. 2000). These areas contained higher numbers of both infected cells and HIV RNA (Fujimura, Goodkin et al. 1997, Wiley, Soontornniyomkij et al. 1998), and exhibited the greatest amount of neuroinflammation (Navia, Cho et al. 1986, Kure, Llena et al. 1991). Individuals infected with HIV also showed aberrant dopamine metabolism in the cerebral spinal fluid (CSF) (Larsson, Hagberg et al. 1991, Kumar, Ownby et al. 2011), and HIV infection was shown to increase CSF dopamine in ART-naïve individuals (Scheller, Arendt et al. 2010). These data suggest that dopamine is involved in HIV neuropathogenesis, a hypothesis corroborated in studies of SIV infected macaques with pharmacologically elevated dopamine levels, as these animals showed both increased CNS viral load and neuropathology (Czub, Koutsilieri et al. 2001, Czub, Czub et al. 2004).

While ART has diminished HIV neuropathology in dopaminergic regions, PLWH on ART still show substantial damage in these areas (Valcour, Sithinamsuwan et al. 2011). These effects include striatal dysfunction (du Plessis, Vink et al. 2015, Ipser, Brown et al. 2015), increased microglial activation and inflammation (Tavazzi, Morrison et al. 2014, Vera, Guo et al. 2016), and neuronal damage (Gongvatana, Harezlak et al. 2013) associated with neuroinflammation. This suggests that even in individuals on ART, these CNS areas are still vulnerable (Becker, Sanders et al. 2011, Alakkas, Ellis et al. 2019, Popov, Molsberry et al. 2019). Substance abuse, which increases CNS dopamine, is associated with increased neurocognitive decline in PLWH in the ART era (Poundstone, Chaisson et al. 2001, Martin, Pitrak et al. 2003, Rippeth, Heaton et al. 2004, Ferris, Mactutus et al. 2008, Baum, Rafie et al. 2009, Meade, Conn et al. 2011, Meade, Lowen et al. 2011, Devlin, Gongvatana et al. 2012, Blackstone, Iudicello et al. 2013, Meyer, Rubin et al. 2013, Chang, Connaghan et al. 2014, Meyer, Little et al. 2014, Meade, Towe et al. 2015), suggesting that dopamine plays an important role in the development of CNS disease when viremia is significantly reduced by ART for extended time periods.

All substances of abuse, despite distinct mechanisms of action, can either directly or indirectly increase CNS dopamine (Di Chiara and Imperato 1988, Carboni, Imperato et al. 1989, Pierce and Kumaresan 2006, Kleijn, Folgering et al. 2011, Clarke, Adermark et al. 2014). The largest effects on CNS dopamine are induced by the use of stimulants, such as cocaine or methamphetamine, that act directly on the dopamine transporter to either block dopamine reuptake, as with cocaine (Jones, Garris et al. 1995), or reverse the direction of the transporter to release more dopamine, as for amphetamines (Cruickshank and Dyer 2009). Both of these drugs have deleterious effects on the CNS (Ferris, Mactutus et al. 2008, Krasnova and Cadet 2009, Kousik, Napier et al. 2012), compromising the BBB (Mahajan, Aalinkeel et al. 2008, Buch, Yao et al. 2012, Kousik, Napier et al. 2012), increasing oxidative stress (Dietrich, Mangeol et al. 2005, Ferris, Mactutus et al. 2008, Krasnova and Cadet 2009, Melo, Zanon-Moreno et al. 2010), and altering dopaminergic neurotransmission (Seiden, Fischman et al. 1976, Wu, Reith et al. 2001, Ferris, Mactutus et al. 2008). Many of these effects are similar to those associated with HIV infection in the CNS, suggesting that drug abuse and HIV infection may amplify the effects of the other, leading to increased cognitive decline in HIV infected people with SUD compared to infected non-drug users or uninfected individuals with SUD.

A large number of studies demonstrate the additive or synergistic impact of stimulants and HIV in both in vitro and in rodent systems (Turchan, Anderson et al. 2001, Nath, Hauser et al. 2002, Aksenov, Aksenova et al. 2006, Griffin, Middaugh et al. 2007, Cotto, Natarajaseenivasan et al. 2018). However, there are also studies showing that stimulants and HIV do not interact (Durvasula, Myers et al. 2000, Byrd, Fellows et al. 2011, Weed, Adams et al. 2012, Levine, Reynolds et al. 2014, Meade, Towe et al. 2015, Minassian, Henry et al. 2017), making it difficult to determine the precise impact stimulants have on HIV neuropathogenesis. This controversy is compounded by an inability to separate the effects of direct stimulant exposure from those of dopamine release induced by stimulant use. Human studies, particularly those designed to characterize molecular mechanisms by which drug-HIV interactions may occur, are needed to address these questions, but these are technically and logistically limited (Ferris, Mactutus et al. 2008, Nightingale, Winston et al. 2014, Norman and Basso 2015). Improving the integration of in vitro experiments with animal models and human studies in a clinically relevant manner is one of the primary challenges facing this field in the future.

In addition to an increased risk of SUD, individuals infected with HIV, particularly those over 50, also show an increased risk for depression, anxiety, and Parkinsonian disorders (Bing, Burnam et al. 2001, Watkins and Treisman 2012, Munjal, Ferrando et al. 2017, Matt and Gaskill 2019). This is important because these and many other age-related disorders are often treated with therapies that affect the dopaminergic system (D’Aquila, Collu et al. 2000, Manfredi 2002, Matt and Gaskill 2019). Reuptake inhibitors such as Fluoxetine (Prozac), Bupropion (Wellbutrin) or Venlafaxine (Effexor) increase extracellular dopamine (Stahl 1998, D’Aquila, Collu et al. 2000), while other medications such as Selegiline (Emsam), Aripiprazole (Abilify) or Pramipexole (Mirapex) either reduce dopamine breakdown (Stahl 1998) or directly activate dopamine receptors (Stahl 1998, D’Aquila, Collu et al. 2000). The impact of these medications on extracellular dopamine is very similar to that produced by stimulants (Matt and Gaskill 2019), so the effects of these types of drugs should also be considered and studied when examining how changes in dopamine can impact neuroHIV.

Another consideration is the impact of ART on dopamine neurotransmission. While there is currently very little information on this topic, the research that has been performed does indicate that some ART drugs impact the dopaminergic system (Matt and Gaskill 2019). Efavirenz, a non-nucleoside reverse transcriptase inhibitor that can induce neuropsychiatric complications (Silveira, Guttier et al. 2012, Mollan, Smurzynski et al. 2014), can interfere with the dopamine transporter (Gatch, Kozlenkov et al. 2013) or monoamine oxidase (Dalwadi, Kim et al. 2016, Akang 2019), and has been shown to both increase and decrease striatal dopamine in rodents, depending on the dose (Cavalcante, Chaves Filho et al. 2017). Zidovudine, a nucleoside analogue reverse transcriptase inhibitor, can interfere with the activity of D1-like receptors (Venerosi, Valanzano et al. 2005), and the antiretroviral protease inhibitor Lopinavir interferes with dopamine uptake by inhibiting plasma membrane monoamine transporters (Duan, Hu et al. 2015). These studies suggest that antiretrovirals may impact the dopaminergic system and could amplify or interfere with the effects of dopamine-inducing drugs and therapies, making it critical to expand our understanding of these interactions to treat HIV and its consequences more efficiently.

Effects of disrupted dopaminergic system on CNS immune cells

The primary targets for HIV infection in the CNS are myeloid cells, including microglia, perivascular macrophages and monocytes (Koenig, Gendelman et al. 1986, Minagar, Shapshak et al. 2002, Crowe, Zhu et al. 2003, Rappaport and Volsky 2015), although T-cells may also play a role in some circumstances (Marcondes, Burudi et al. 2001, Anthony, Crawford et al. 2003, Petito, Adkins et al. 2003, Nguyen, Soukup et al. 2010, Gaskill, Calderon et al. 2013). These cell types express all subtypes of dopamine receptors, the D1-like receptors, D1 and D5, and the D2-like receptors, D2, D3 and D4. Both myeloid cells and T-cells also express a number of other dopamine related proteins including the dopamine transporter, tyrosine hydroxylase, and monoamine oxidases (Levite, Chowers et al. 2001, McKenna, McLaughlin et al. 2002, Gaskill, Calderon et al. 2009, Gaskill, Carvallo et al. 2012, Coley, Calderon et al. 2015, Levite 2016, Nolan and Gaskill 2018, Matt and Gaskill 2019). Canonically, dopamine acts through the two subtypes of dopamine receptors by inducing opposing effects on cyclic AMP (Bacic 1991, Beaulieu and Gainetdinov 2011), but the signaling pathways and overall effects of dopamine on immune function are not well understood, particularly in myeloid cells. Recent data suggest that activation of dopamine receptors in human macrophages could act through a non-canonical pathway to regulate cell functions through changes in calcium release (Nickoloff, Mackie et al. 2019), but this area is still relatively undefined.

Dopamine has significant immunomodulatory effects in both myeloid cells and T-cells, regulating their activation, production of cytokines, chemokines and nitric oxide, chemotaxis and transmigration, and phagocytosis (Basu and Dasgupta 2000, Farber, Pannasch et al. 2005, Sarkar, Basu et al. 2010, Gaskill, Carvallo et al. 2012, Coley, Calderon et al. 2015, Yan, Jiang et al. 2015, Levite 2016, Calderon, Williams et al. 2017, Nolan and Gaskill 2018, Nolan, Muir et al. 2019). Many of these effects could specifically impact the development of neuroHIV. Dopamine concentrations induced by SUD increase cell motility and adhesion, as well as transmigration of both infected and uninfected CD14⁺CD16⁺ monocytes across the BBB that is mediated by D1-like receptors (Coley, Calderon et al. 2015, Calderon, Williams et al. 2017). Activation of D1-like dopamine receptors also increases the susceptibility of macrophages to HIV entry (Gaskill, Yano et al. 2014), potentially through modulation of calcium flux (Nickoloff, Mackie et al. 2019). Additionally, human macrophages exposed to dopamine produce greater amounts of inflammatory cytokines and chemokines such as IL-1β, IL-6, CCL2 and CXCL10 (Gaskill, Carvallo et al. 2012, Nolan, Muir et al. 2019). These effects could increase the influx of both HIV infected and inflammatory immune cells into the CNS, promote the spread of viral infection and reseeding of reservoirs within the brain, and contribute to the development of neuroinflammation and a neurotoxic environment. These effects would be greater in regions with elevated dopamine, connecting increased HIV-associated neuropathology in these regions with the effects of dopamine on immune cell function.

Although myeloid cells are the primary mediators of neuroHIV (Koenig, Gendelman et al. 1986, Minagar, Shapshak et al. 2002, Crowe, Zhu et al. 2003, Rappaport and Volsky 2015), studies suggest that T-cells can also contribute to CNS infection (Marcondes, Burudi et al. 2001, Anthony, Crawford et al. 2003, Petito, Adkins et al. 2003), especially in PLWH with SUD (Tomlinson, Simmonds et al. 1999, Anthony, Ramage et al. 2005, Gaskill, Calderon et al. 2013). Dopamine has also been shown to increase viral replication in T-cells, and to regulate a number of T-cell functions (Levite 2016), suggesting that dopamine could also impact the development of disease through its effects on this population. While the role of T-cells in HIV neuropathogenesis is not clearly defined (Gaskill, Calderon et al. 2013), they are central to systemic HIV infection (Fox and Cottler-Fox 1992, Okoye and Picker 2013) and changes in peripheral dopamine induced by substances of abuse could alter the contribution of T-cells to disease pathogenesis in non-CNS tissues. While the impact of SUD on peripheral dopamine concentrations is unknown, it needs to be examined as dopamine is important to homeostatic functions in a number of peripheral organs (Matt and Gaskill 2019). Therefore, the impact of dopamine on HIV infection in peripheral immune cells in the context of SUD is an area for future studies.

How any of these effects would be changed with ART is not well characterized. The use of stimulants, and many other types of drugs, is strongly correlated with diminished therapeutic effectiveness of ART (Ellis, Childers et al. 2003, Rippeth, Heaton et al. 2004, Meade, Conn et al. 2011, Blackstone, Iudicello et al. 2013, Levine, Reynolds et al. 2014, Dash, Balasubramaniam et al. 2015, Meade, Towe et al. 2015). This is believed to be driven by SUD associated problems with medication adherence and access to treatment, rather than specific biological interactions between the stimulants and the antiretrovirals themselves (Arnsten, Demas et al. 2002, Lucas, Griswold et al. 2006, Ferris, Mactutus et al. 2008, Byrd, Fellows et al. 2011, Meade, Conn et al. 2011, Schuster and Gonzalez 2012, Levine, Reynolds et al. 2014, Anderson, Higgins et al. 2015), although more studies are necessary to support this.

While successful ART may eliminate, or at least dramatically decrease, viral replication, the effects of antiretrovirals on myeloid cells are much less clear (Matt and Gaskill 2019). A recent study showed that in macrophages derived from a small cohort of PLWH on fully suppressive ART and from uninfected individuals, the effects of dopamine on cytokine production are similar, suggesting that ART does not change dopamine-mediated effects on cytokine production by macrophages (Nolan, Muir et al. 2019). Another recent study showed that cocaine treatment increased the expression of inflammasome associated genes in HIV-infected macrophages (Atluri, Pilakka-Kanthikeel et al. 2016), although it is not known whether cocaine induces the release of dopamine from these cells, and whether dopamine contributes to cocaine-mediated inflammasome activation in macrophages. While not associated with dopamine, experiments in the human macrophage-like THP-1 cell line showed that treatment with Stavudine, Zidovudine, Nelfinavir, and Lopinavir increased their production of IL-1β and TNF-α (Lagathu, Eustace et al. 2007). Cohort studies have also shown that while overall HIV-associated inflammation is decreased with successful ART, biomarkers of inflammation and myeloid activation are still detected in PLWH on ART when compared to uninfected individuals (Gisolf, van Praag et al. 2000, Eden, Price et al. 2007, Wada, Jacobson et al. 2015, Sereti, Krebs et al. 2017). Together, these data suggest that ART does not inhibit the impact of elevated dopamine on the development of inflammation, highlighting the need to understand the role of dopamine in neuroHIV to develop adjuvant therapies that can be included with ART for PLWH with SUD.

Defining the interaction between dopamine and HIV infection

Many studies examining the effects of either specific substances of abuse or dopamine on immune cells and viral infection have been done in vitro (Peterson, Gekker et al. 1991, Steele, Henderson et al. 2003, Mahajan, Aalinkeel et al. 2008, Gaskill, Calderon et al. 2009, Gaskill, Carvallo et al. 2012, Mantri, Pandhare Dash et al. 2012, Addai, Pandhare et al. 2015, Rao, Ande et al. 2016, Calderon, Williams et al. 2017, Cotto, Natarajaseenivasan et al. 2018, Nickoloff, Mackie et al. 2019). While these experiments are critical to define the molecular mechanism(s) underlying the effects of dopamine, some of these studies are inherently limited as they are focused on the pharmacology of the drug being studied, not specifically on dopamine (Jones, Garris et al. 1995, Ferris, Mactutus et al. 2008, Kish 2008, Cruickshank and Dyer 2009). In addition, studies using cell culture systems are not optimal due to the lack of “history” in these types of experiments. Use of cell lines, and even primary cells, makes it difficult to assess the continuous, chronic impact of infection, SUD, and/or ART on individual health, as it is not clear this can be modeled completely in these types of systems. Research demonstrates that the length of time an individual is infected with HIV prior to the use of ART can have significant impacts on long term health outcomes (Yilmaz, Price et al. 2008, Gay, Dibben et al. 2011, Shive, Biancotto et al. 2012, Funderburg and Lederman 2014, Cao, Mehraj et al. 2015, Sereti, Krebs et al. 2017, Hellmuth, Slike et al. 2019). In the brain, extended infection with HIV prior to treatment could result in irreversible damage to dopaminergic and other circuits, a kind of neurological “scar”, leaving infected individuals vulnerable to developing impairments or aberrant behaviors more easily. The use of large numbers of primary human cells from specific populations with strong demographic and epidemiological data can help to eliminate some of these deficiencies, especially if the studies include approximation of the length of time infected prior to ART. Questions about how these data can be translated to patient care will remain and need to be considered.

These studies will characterize the effect of dopamine on cellular proteins and signaling pathways involved in both CNS and peripheral myeloid cells of ART treated PLWH with SUD. This will identify novel targets for development of adjuvant therapies that inhibit the effects of SUD on neurocognitive deficits in PLWH. To optimize these studies, both in vitro and in vivo experimental systems with human cells, and rodent and non-human primate animal models should be used. Although primate models remain the gold standard in HIV research, their cost and size make them difficult to use extensively, resulting in increased use of rodent model systems. Identification of targets by most, if not all, of these model systems will add to both the rigor and translational value of the data.

Rodent systems use expression of viral proteins to simulate HIV infection, as rodents are not naturally susceptible to infection by this virus (Morrow, Wharton et al. 1987, Hatziioannou and Evans 2012). While none of these models appears to recapitulate completely the complex network of immune and neuronal cells observed in humans, these systems have been essential for evaluating the effects of HIV on the dopaminergic system in vivo, and have greatly contributed to the understanding of the impact of substance abuse and dopamine on HIV infection (Roth, Tashkin et al. 2002, Bagetta, Piccirilli et al. 2004, Griffin, Middaugh et al. 2007, Yao and Buch 2012, Kim, Lowe et al. 2015, Liu, Xu et al. 2017, Sanchez and Kaul 2017). They have been particularly valuable in examining the neuroinflammatory and neurotoxic impact of viral proteins in conjunction with methamphetamine or cocaine, and many showed a synergistic impact of these drugs with HIV proteins (Liu, Williams et al. 2000, Bagetta, Piccirilli et al. 2004, Kim, Lowe et al. 2015, Sanchez and Kaul 2017). However, expression of viral proteins in these models is, however, artificial, and can result in the release of varying amounts of viral protein in different brain regions relative to human infection (Gaskill, Miller et al. 2017). Further, as the concentrations of HIV proteins in distinct regions of the CNS are not defined (Gaskill, Miller et al. 2017), it is not clear how accurately the protein concentrations used in these studies mimic those found in the CNS of PLWH (Spudich and González-Scarano 2012).

This is significant, as exposure to non-relevant concentrations of viral protein could result in neurological alterations or damage not seen in human infection. While determining how much HIV protein is present in the CNS would improve the reliability of these model systems, expression of HIV proteins in the absence of viral infection means that the regulatory processes that control the amount of viral proteins produced and released in the human brain are not present (Gorantla, Poluektova et al. 2012, Saylor, Dickens et al. 2016). This suggests that the neuropathology in these systems may be different from what develops in humans in vivo. Future studies should focus on modeling development of neuroHIV more accurately, and on characterizing specifc effects of SUD and aberrant dopaminergic function on this disease process. However, it is important to underscore that the amount of HIV and its associated proteins in the human brain is unlikely to be consistent among individuals as well as among different demographics (Nightingale, Winston et al. 2014). Thus, human studies must clearly state which demographic or other specific group of people is being modelled.

A number of technological advances enabled the development of new rodent models. These include the creation of humanized mice, in which immunodeficient mice are implanted with human tissue or injected with human cells (Lan, Tonomura et al. 2006, Gorantla, Poluektova et al. 2012, Honeycutt, Wahl et al. 2016), and the development of a mouse tropic HIV hybrid virus, EcoHIV (Potash, Chao et al. 2005). More recently, a model has been developed in which CNS myeloid populations have been recapitulated in a mouse brain (Honeycutt, Wahl et al. 2016, Honeycutt, Thayer et al. 2017). These models provide valuable in vivo systems in which to test the effects of SUD on viral replication, disease pathogenesis, and behavior. They also show that macrophages are infected with HIV in the absence of T-cells (Honeycutt, Wahl et al. 2016, Honeycutt, Thayer et al. 2017), demonstrating the importance of these cells in the development of neuroHIV, which is largely driven by myeloid cells in humans (Koenig, Gendelman et al. 1986, Minagar, Shapshak et al. 2002, Crowe, Zhu et al. 2003, Rappaport and Volsky 2015). A caveat for models that inject virus directly into the CNS is that this can cause mechanical trauma and inflammation (Stevenson, Hawke et al. 1997, Hauss-Wegrzyniak, Lukovic et al. 1998, Cartmell, Southgate et al. 1999), potentially inducing neuronal damage independent of viral infection, confounding the results.

Another consideration, particularly when examining the immunomodulatory effects of dopamine, is that rodent cells do not always respond to immunologic stimuli in the same way as do human cells. For example, relative to human cells, rodent myeloid cells show significant species-specific differences in terms of the gene targets, TLR response, IL-1β regulation and NLRP3 inflammasome activation in response to a number of immunomodulatory stimuli (Schroder, Irvine et al. 2012, Ariffin and Sweet 2013, Yan, Jiang et al. 2015, Schaale, Peters et al. 2016, Dawson, Smith et al. 2017, Nolan 2019). Additionally, recent data indicate that dopamine may act through distinct signaling cascades in different cell types (Nickoloff, Mackie et al. 2019), and this may apply to different species as well. It is possible that these species-specific differences account for the lack of translatability of these data to humans, as the data resulting from these studies have not always proven to model the neuropathology accurately in PLWH with SUD.

Taken together, these caveats suggest that use of non-human primates to study the synergistic impact of HIV and SUD is warranted, as these animals show the greatest resemblance to the human system and can be infected with SIV (Hatziioannou and Evans 2012). Studies in primates have been used to demonstrate increases in inflammation and neuropathology in response to dopamine-altering therapeutic drugs (Czub, Koutsilieri et al. 2001, Czub, Czub et al. 2004), and alterations in dopamine-rich brain regions in response to untreated infection (Jenuwein, Scheller et al. 2004, Scheller, Sopper et al. 2005) supporting findings from human studies. SIV models have been used to examine effects of methamphetamine and other drugs of abuse on immune cells in the infected CNS, showing increased viral load, neuroinflammation, and microglial CCR5 expression (Marcondes, Flynn et al. 2010, Najera, Bustamante et al. 2016). These findings are similar to those in non-human primates treated with drugs that elevate CNS dopamine directly (Czub, Koutsilieri et al. 2001, Czub, Czub et al. 2004), suggesting that some of the effects of SUD are mediated by increases in dopamine. More studies in non-human primates, in which animals are given drugs of abuse and treated with ART, will be needed to evaluate whether drugs of abuse and dopaminergic therapies impact neuroHIV in the ART era. However, even these models do not fully replicate the complexity of infection in the human brain, and the expense and logistical challenges involving their use are significant hurdles.

While human studies have the most potential to define the broad effects of drug-induced alterations in dopamine on neuroHIV, these are limited by the inability to control the length of infection and time to ART, and to examine the more mechanistic aspects of neurologic infection. There are several additional factors that complicate these studies, particularly related to SUD in the ART era. One major issue is how studies define drug use, particularly the length of time following use that is considered current and how poly-drug use is considered (Byrd, Fellows et al. 2011, Fernandez-Serrano, Perez-Garcia et al. 2011, Norman and Basso 2015). Many of the differences among studies examining the cognitive impact of SUD in PLWH may be attributable to different definitions of substance abuse (Fernandez-Serrano, Perez-Garcia et al. 2011). Creating standard definitions of active and past drug abuse would be valuable, as they would enable better consolidation and meta-analysis of multiple studies, providing a broader understanding of the cognitive impact of drug-associated changes in dopamine concentrations in PLWH. Additionally, the length of time from initial infection to ART initiation, the type of ART, as well as other cofactors such as age, reading level, and starting CD4+ T-cell count, can all impact cognitive status independent of drug abuse (Heaton, Clifford et al. 2010, Nightingale, Winston et al. 2014, Heaton, Franklin et al. 2015). Therefore, future studies in humans should include access to detailed epidemiologic and immunologic data, with longitudinal changes in immune populations and functions being assessed.

Conclusion

Given the complexity of the problem, no single model system can accommodate all the necessary types of research into CNS infection with the level of detail required for mechanistic examination of the interactions between drugs of abuse and ART on HIV neuropathogenesis. Therefore, future studies should combine results from multiple modalities to develop a more complete picture of the impact of drug-induced changes in dopamine etiology on neuroHIV. Research that focuses on correlating results among in vitro experiments, animal models, and human studies will provide a more complete understanding of the effects of SUD and dopaminergic dysfunction in the HIV infected CNS. They will also provide a detailed understanding of the proteins and pathways involved in this disease process that is required to develop new targets and strategies to reduce both HIV infection and its consequences in the vulnerable population with SUD.

Abbreviations:

ART

antiretroviral therapy

BBB

blood brain barrier

CNS

central nervous system

CSF

cerebrospinal fluid

PLWH

people living with HIV

SUD

substance use disorder