White matter loss and oligodendrocyte dysfunction in HIV: a consequence of the infection, the antiretroviral therapy or both?

Abstract

While the severe cognitive effects of HIV-associated dementia have been reduced by combined antiretroviral therapy (cART), nearly half of HIV-positive (HIV+) patients still suffer from some form of HIV-Associated Neurocognitive Disorders (HAND). While frank neuronal loss has been dramatically reduced in HAND patients, white matter loss, including dramatic thinning of the corpus callosum, and loss of volume and structural integrity of myelin persists despite viral control by cART. It remains unclear whether changes in white matter underlie the clinical manifestation seen in patients or whether they are the result of persistent viral reservoirs, remnant damage from the acute infection, the antiretroviral compounds used to treat HIV, secondary effects due to peripheral toxicities or other associated comorbid conditions. Both HIV infection itself and its treatment with antiretroviral drugs can induce metabolic syndrome, lipodystrophy, atherosclerosis and peripheral neuropathies by increased oxidative stress, induction of the unfolded protein response and dysregulation of lipid metabolism. These virally and/or cART-induced processes can also cause myelin loss in the CNS. This review aims to highlight existing data on the contribution of white matter damage to HAND and explore the mechanisms by which HIV infection and its treatment contribute to persistence of white matter changes in people living with HIV currently on cART.

1.0. Persistence of White Matter Changes in cART-treated HIV

Approximately half of HIV+ individuals suffer from some level of cognitive, behavioral or motor abnormalities, currently classified as HIV-associated neurocognitive disorders (HAND) (Brew, 2004; Gray et al., 2003; Heaton et al., 2010, 2011; Saylor et al., 2016). However, in the era of combined antiretroviral therapy (cART), the clinical presentation has shifted from more severe forms of HIV-associated dementia (HAD), to milder presentations that are currently detected using domain deficit scoring to reveal functionally affected patients with minor neurocognitive disorder (MND) or deficits that do not impact daily living, asymptomatic neurocognitive impairment (ANI). These shifting clinical presentations are accompanied by changes in pathologic observations which have changed from prominent subcortical pathology of neuronal damage and loss, astrogliosis and microgliosis to more subtle hippocampal and cortical changes in synaptic number, neuroinflammation, and oxidative stress (González-Scarano and Martín-García, 2005; Gray et al., 2003; Heaton et al., 2011). Despite this shift, abundant clinical and pathologic studies have demonstrated that synaptodendritic damage and white matter pathologies persist and remain a potential substrate for clinical presentations seen in people living with HIV and taking cART (Ellis et al., 2007; Everall et al., 2009, 1999; Langford et al., 2003; Masliah et al., 1997; Müller-Oehring et al., 2010; Tate et al., 2010; Zheng et al., 2001). This review seeks to explore the potential contributions of white matter pathologies to HAND and the viral- and treatment-associated mechanisms that may account for damage to white matter in people living with HIV.

2.0. Pathologic Evidence for the Contribution of White Matter Damage to HAND

Prior to the introduction of antiretroviral therapy, alterations in white matter including myelin pallor, gliosis, and leukoencephalopathy were prominent in HIV-infected patients (Everall et al., 2005; Gray et al., 1996; Navia et al., 1986; Sharer et al., 1988). Now even with effective viral suppression in the cART-era, while severe forms of leukoencephalopathy have been diminished, the prevalence of white matter pathologies in patients has persisted, with the amount of white matter damage positively correlated with the degree of neurocognitive impairment observed (Everall et al., 2005; Langford et al., 2003; Müller-Oehring et al., 2010; Tate et al., 2010). In 2003, a study was performed which evaluated the presence of white matter disease in frontal cortex, cingulate cortex, corpus callosum, circle semiovale, and occipital cortex in 62 HIV⁺ individuals from the pre-cART era (1993–1995), and 89 from the cART era (1996–2001) by immunohistochemical staining of postmortem tissue (Langford et al., 2003). This report indicated that while only 11% of individuals pre-cART displayed white matter pathology, this increased to 26% in the later cART era cohort (Langford et al., 2003). Since this time as patients are living relatively normal lifespans, postmortem tissue has become less abundant, thus shifting the paradigms of testing for such abnormalities to the much less invasive modalities of MRI and diffusion tensor imaging. Using these techniques, numerous studies have revealed that continuing white matter deficits include dramatic thinning of the corpus callosum, reduction in blood flow to white matter, and loss of both structural integrity and volume from white matter structures including: the superior longitudinal fasciculus, superior corona radiata, and the internal capsule (Corrêa et al., 2015; Gongvatana et al., 2011; Hoare et al., 2010; Kelly et al., 2014; Leite et al., 2013; Pomara et al., 2001; Ragin et al., 2004; Tate et al., 2011; Wohlschlaeger et al., 2009). In 2009, a study using morphometric analysis of myelin in the corpus callosum in HIV patients indicated an increase in myelin G ratios, signifying a decrease in myelin thickness (Wohlschlaeger et al., 2009). Additionally, consistent alterations to white matter microstructure in HIV+ patients despite cART has also been demonstrated (O’Connor et al., 2017). In a recent comprehensive study by the CHARTER Group, structural MRI was performed in 253 HIV+ individuals. From this analysis, it was determined that overall white matter volume was decreased in individuals with HAD and MND compared with HIV+ patients who were neurocognitively normal (Alakkas et al., 2019). Importantly, this study found that higher levels of abnormal white matter were associated with higher risk of HAND (Alakkas et al., 2019). Despite these accumulating findings, the CHARTER Group investigators and the field still acknowledge that these findings may be a result of a legacy effect resulting from HIV infection prior to cART-intervention, antiretroviral toxicities, ongoing effects of CNS viral pools, or medical comorbidities.

Complementing the imaging findings, transcriptome analysis of frontal cortex from HIV+ patients who attained successful viral suppression through cART but were diagnosed with HAND also implicates persistent white matter changes (Borjabad et al., 2010). Among the persistently altered transcripts in cART treated patients were myelin transcription factor 1, myelin basic protein (MBP), and myelin-associated oligodendrocyte basic protein (MOBP), which are critical for oligodendrocyte maturation, myelination, and maintenance (Borjabad et al., 2010). Intriguingly, changes in these gene signatures for white matter damage persisted in cART treated patients despite reversal of pro-inflammatory gene signatures and restoration of synaptic damage signatures, although interferon signaling and cell cycle genes also remained elevated in cART-treated HIV+ patients as well (Borjabad et al., 2010). Consistent with this transcriptome study, brain lysates from HIV+ patients who were on cART and displayed signs of HAND showed significantly lower levels of MBP, with an increase in levels of 3’5’cyclic nucleotide phosphodiesterase (CNP). Since CNP is an early myelin protein whose expression begins before differentiation, the latter suggests that oligodendrocyte progenitors are attempting to replace the reduced myelin (Jensen et al., 2015). Thus, white matter pathologies remain prevalent in HIV-infected patients despite viral suppression through cART, with dysregulation of myelin mRNAs and loss of structural integrity of myelin evident by several imaging modalities.

It is still unclear whether the white matter changes observed contribute to HAND, or whether they are secondary effects resultant from ongoing neuronal toxicities and inflammatory states. Increased understanding of the molecular changes underlying white matter abnormalities and myelin phenotypes in HIV infection, antiretroviral treatment, and the complex in vivo case of HAND patients is sorely needed to delineate these possibilities.

3.0. Oligodendrocytes in Development and Disease

The myelin sheath, synthesized as the plasma membrane of the oligodendrocyte, is vital to several aspects of CNS function. Myelin insulates the neuronal axonal plasma membrane thereby reducing the transverse capacitance, increasing the resistance, and facilitating the propagation of rapid neuronal impulses. Oligodendrocytes also provide trophic support to the axons their myelin membranes surround, which is crucially important for maintaining cellular functioning and signaling at long extended distances from the neuronal soma (Edgar and Gerbern, 2004; Lappe-Siefke et al., 2003). In contrast, loss of myelin integrity can disrupt proper signaling, resulting in neuronal degeneration and neurological dysfunctions (Criste et al., 2014; Mighdoll et al., 2015).

Mature oligodendrocytes are generated from oligodendrocyte precursors cells (OPCs) following a well-characterized progression (Figure 1; Miller, 2002). In rodents, at embryonic day 12.5, OPCs originate in ventricular zones and migrate both dorsally and radially (Raff et al., 1983). OPCs transition to immature oligodendrocytes and begin to extend actin filaments to elaborate processes. Following contact with neurons, these maturing cells extend processes, upregulate expression of myelin proteins, and ultimately ensheathe neuronal axons with the lipid- and protein-based myelin membrane (Grinspan, 2002; Miller, 2002). Chief among the protein components are proteolipid protein (PLP, 17%), myelin basic protein (MBP, 8%), cyclic nucleoside phosphodiesterase (CNPase, 4%), myelin oligodendrocyte glycoprotein (MOG, 1%), and myelin-associated glycoprotein (MAG, 1%) (Jahn et al., 2009). Although myelination is generally complete in young adulthood, myelin maintenance and replacement is a continually ongoing process over the course of an individual’s lifetime, which is essential for optimal neuronal functioning and survival (McLaurin and Yong, 1995).

Figure 1.

Cartoon graphic of three stages of oligodendrocyte development, OPC, immature and mature, including factors which promote or inhibit oligodendrocyte precursor cell differentiation (above the cells), stage-specific protein markers, and transcription factors (below the cells).

Myelin is damaged or improperly formed in a variety of conditions which primarily result in motor dysfunction. Factors contributing to these diseases include genetic mutations of myelin components termed leukodystrophies, demyelinating diseases caused by inflammation such as multiple sclerosis, and hypomyelination or myelin damage caused by injury such as in perinatal white matter injury in preterm infants (Garbern, 2007; Markowitz, 2013; Volpe, 2001). Myelin deficits have now also been implicated in neurodegenerative diseases such as ALS, and neurological diseases with cognitive and behavioral consequences such as autism and schizophrenia (Chavarria-Siles et al., 2016; Kang et al., 2013; Libero et al., 2015). Although OPCs persist in the CNS in these diseases and even proliferate in response to pathology in some, they are unable to differentiate and effectively form new myelin (Nait-Oumesmar et al., 2008). It is plausible that the prevailing extracellular milieu in HIV+ patients taking cART is incompatible with oligodendrocyte persistence and/or OPC differentiation into mature oligodendrocytes, either of which would result in reduced white matter. White matter pathology is among the most consistent of pathological finding in HIV+ patients; however, very few studies have explored the affect of HIV infection or cART on oligodendrocyte maturation, function or survival.

4.1. Mechanisms of Cellular Stress due to HIV infection

HIV is known to infect cells via two co-receptors, CD4 and either CCR5 or CXCR4 (Bleul et al., 1997). Oligodendrocytes lack the CD4 receptor; thus,they are unlikely to be susceptible to HIV infection even though CD4-independent entry has been described (Bracq et al., 2018; Kaul et al., 2005; Sowinski et al., 2008). The likely primary effect of HIV on oligodendrocytes, as well as neurons, is through an indirect mechanism called the “bystander effect”, in which the damage is caused by the immune activation of HIV-infected cells. These infected macrophages, microglia, and astrocytes, as well as uninfected bystander immune responsive cells, release a myriad of cytotoxic molecules, including reactive oxygen species, nitric oxide, glutamate, and pro-inflammatory cytokines and chemokines, as well as the virus itself, which culminate in damage and dysfunction to oligodendrocytes and neurons in addition to activation of neighboring uninfected immune cells which may contribute to the secretion of toxic factors (Figure 2; González-Scarano and Martín-García, 2005). Classic cellular hallmarks in response to a neuroinflammatory environment were readily apparent in HAND patients prior to therapeutic interventions, including multinucleated giant cells, microglial nodules, astrogliosis, and myelin pallor (Alakkas et al., 2019; Glass et al., 1995; Gray et al., 1996; McArthur et al., 2003; Navia et al., 1986; Petito et al., 1986; Sharer et al., 1988). However, many of these features are less prevalent in patients who have suppressed plasma viral loads and reconstitution of immune system function due to cART. cART was anticipated to resolve HIV-mediated neurological pathologies. While explicit mechanisms remain unclear, it is thought that CNS invasion by HIV and establishment of a persistent viral reservoir perpetuates an inflammatory and excitotoxic environment and underlies continuing neurocognitive deficits despite effective peripheral viral suppression in HIV+ patients, at least in part (Anthony et al., 2005; Glass et al., 1995; González-Scarano and Martín-García, 2005; Gray et al., 2001). Alternatively, CNS side effects may be a result of antiretroviral compounds which may act directly in the CNS or indirectly by causing peripheral toxicities.

Figure 2.

Illustration of HIV-1 virus neuroinvasion and potential downstream effect on CNS-resident cells, microglia, astrocytes, oligodendrocytes, and neurons. HIV-infected monocyte derived macrophages and T-cells cross a damaged, permeable blood brain barrier. Virus particles productively infect microglia cells. Infected microglia actively replicate new viral particles and neurotoxic proteins, i.e. Tat. Astrocytes can also be infected through engulfing infected macrophage material. Infected astrocytes and microglia produce chemokines and cytokines as well as increase glutamate concentration, thus leading to excitotoxicity and cell death in neurons. The same mediators may lead to inhibition of differentiation and/or cell death in oligodendrocytes.

Several cellular stressors have been identified in the CNS of HIV+ patients. Elevated levels of inflammation can lead to an excitotoxic state, where excess extracellular glutamate chronically activates N-methyl-D-aspartate (NMDA) receptors on neurons and oligodendrocytes, raising intracellular calcium levels and causing the activation of calcium-dependent proteases calpains and caspases, ultimately leading to cell death. This mechanism of HIV-mediated toxicity has been well documented and characterized in neurons, but only studied in oligodendrocytes in relation to the HIV-Tat protein (Liu et al., 2017; O’Donnell et al., 2006; Stern et al., 2018; Zou et al., 2019, 2015). After 3 months of astrocyte-driven Tat-transgene expression, animals displayed disruption of myelin in the caudate-putamen, aberrant oligodendrocyte morphology in corpus callosum and anterior commissure, and significantly decreased levels of MBP and MAG in striatum (Zou et al., 2015). Additionally, when cells in vitro were exposed to exogenously applied Tat protein, viability of immature oligodendrocytes was significantly decreased and surviving cells displayed significant deficits in maturation as evidenced by decreased branching processes and myelin membrane extensions (Zou et al., 2015). Effects of Tat on oligodendrocytes could be ameliorated by inhibiting glutamate receptor-mediated calcium influxes, which lead to oligodendrocyte death and abnormalities, and subsequent myelin dysfunction. In a recently published follow-up to this work, the Knapp laboratory has also begun to address oligodendrocyte involvement in CNS responses to HIV infection in vivo through a transgenic mouse model expressing HIV-Tat, coupled with pharmacological manipulation of the CaMKIIβ and GSK3β signaling axis identified in their previous work (Zou et al., 2019). Together, these two studies have identified and implicated the CaMKIIβ-GSK3β signaling pathway as a major player in Tat-mediated oligodendrocyte toxicity both in vitro and in vivo. These pioneering studies have provided the groundwork that viral proteins can affect oligodendrocyte survival and extent of myelination. In addition to direct effects of viral proteins, additional models to examine the effects of factors secreted by macrophages and astroglia on oligodendrocyte differentiation are needed.

An additional stress pathway activated in HIV+ patients is the Unfolded Protein Response (UPR)/Endoplasmic Reticulum stress (ER stress); the UPR is an adaptive mechanism activated in order to cope with unfolded or misfolded proteins in the lumen of the ER (Cao and Kaufman, 2012; Ron and Walter, 2007; Zhang and Kaufman, 2008). Additionally, other physiological factors can play a role in activating ER stress, i.e. oxidative stress, calcium depletion, hypoxia, inflammation, proteasome inhibition and overload of cholesterol (Cao and Kaufman, 2012; Parker et al., 2005; Ron and Walter, 2007; Zhang and Kaufman, 2008, 2004). Through the UPR the ER increases its ability to fold proteins, decreases global protein synthesis and improves degradation of misfolded proteins (Cao and Kaufman, 2012; Ron and Walter, 2007; Zhang and Kaufman, 2008, 2004). This occurs through signaling with three ER transmembrane protein sensors: inositol-requiring kinase 1 (IRE1), protein kinase RNA-like endoplasmic reticulum kinase (PERK) and activating transcription factor 6 (ATF6) (Cao and Kaufman, 2012). Key protein targets of the ER stress response, BiP (the ER-lumenal Hsp-70 chaperone protein) and ATF6, have been shown to be upregulated in the cortex of HIV+ and HAND individuals (Akay et al., 2012; Lindl et al., 2007, 2010). Moreover, this same study showed increased immunofluorescent staining of BiP in neurons and astrocytes of HIV+ patients; however, oligodendrocytes were not examined (Lindl et al., 2007).

4.2. Mechanisms of Cellular Stress due to cART

Currently recommended cART regimens include a cocktail of nucleoside/nucleotide reverse-transcriptase inhibitors (NRTIs), non-nucleoside reverse-transcriptase inhibitors (nNRTIs), protease inhibitors (PIs), and integrase inhibitors (INSTIs) (AIDSinfo, 2016; WHO, 2016). This multi-drug approach has led to long-term viral suppression and restoration of immune system function, which underlies the reduction in morbidity and mortality caused by HIV infection since the introduction of cART (Heaton et al., 2010; McArthur et al., 2010; Sacktor and Roberston, 2014). However, it has been well documented that peripheral side effects of antiretroviral compounds are widespread and include metabolic syndrome, lipodystrophy, atherosclerosis and peripheral neuropathy, depending on the drug class and specific agent. Mechanisms for side effects of some antiretroviral drugs have also been well documented. For instance, debilitating peripheral neuropathy has been described in patients, stemming from NRTI, and to a lesser extent PI, usage (Dalakas, 2001; Ellis et al., 2008). Toxicity associated with compounds of the NRTI class are due to competition with cellular thymidine triphosphate for DNA pol-γ, resulting in premature termination of mitochondrial DNA synthesis (Dalakas, 2001). Subsequent decreases in mitochondria number and oxidative phosphorylation enzymes leads to electron transport chain uncoupling and an increase in mitochondrially-derived intracellular reactive oxygen species (ROS) culminating in activation of the apoptotic cascade and neuronal death (Dalakas, 2001). It has been clearly demonstrated that a variety of drugs from the PI and NRTI drug classes alone or in combination result in accumulation of ROS, hydrogen peroxide, and factors promoting the recruitment of monocytes in a wide range of cell populations including human adipocytes, monocytes, aortic endothelial cells, and myeloid cell lines exposed to PIs: ritonavir, saquinavir, lopinavir, indinavir, nelfinavir, and atazanavir, as well as the NRTIs: zidovudine (AZT), stavudine, and didanosine (Brandmann et al., 2012; Chandra et al., 2009; Ferraresi et al., 2006; Lagathu et al., 2007; Manda et al., 2011; Mondal et al., 2004; Touzet and Phillips, 2010).

Activation of the UPR and ER stress pathways has also been documented following PI exposure (Gupta et al., 2007; McLean et al., 2009; Parker et al., 2005; Wu et al., 2010; Zhou et al., 2006, 2005). The compounds shown to cause induction of these sensors and other downstream targets in human adipocytes, macrophages, intestinal epithelial cells, hepatocytes, and ovarian cancer cells included the PIs: lopinavir, ritonavir, saquinavir, indinavir, nelfinavir, and atazanavir (Gupta et al., 2007; McLean et al., 2009; Parker et al., 2005; Wu et al., 2010; Zhou et al., 2006, 2005). Ultimately, dysfunction through this mechanism leads to pathologies culminating in atherosclerosis, hypercholesterolemia, lipodystrophy, and insulin resistance (Carr et al., 1998; Cohen, 2005; Parker et al., 2005; Zhou et al., 2005). Chronic oxidative and ER stress play a critical role in peripheral side effects of antiretroviral drugs with implications for cellular damage resulting in the manifestation of HAND, as has also been suggested for many other neurodegenerative diseases (Mariani et al., 2005; Romano et al., 2010). Even in HIV+ patients with well-controlled viral infection, evidence for persistence of oxidative and ER stress remains (Akay et al., 2012; Blas-Garcia et al., 2011; Lindl et al., 2007).

Lipodystrophy is another known side effect of some of the antiretrovirals. Ritonavir, for example, increases in expression of the master lipid regulator transcription factor sterol regulatory element binding protein (SREBP) resulting ultimately in increased levels of cholesterol and other lipids, contributing to atherosclerosis (Pao et al., 2008; Riddle et al., 2001; Zhou et al., 2006). The effects of HIV, as well as the antiretroviral drugs that suppress it, have been studied mainly in neurons (Liu et al., 2017; O’Donnell et al., 2006; Stern et al., 2018a;) Only one study has examined the effects of select antiretroviral drugs on oligodendrocyte differentiation in culture and myelin maintenance in vivo. Among those tested, the NRTI, AZT, did not alter maturation in culture, but both PIs, ritonavir and lopinavir, significantly inhibited oligodendrocyte differentiation in culture without causing cell death (Jensen et al., 2015). In vivo, two weeks of intravenous administration of ritonavir into adult mice significantly decreased expression of myelin proteins CNP and MOG (Jensen et al., 2015). How these antiretroviral agents inhibit oligodendrocyte differentiation or myelin maintenance is not yet clear. It is possible that each drug or class of drugs will have specific mechanisms of action. These data provide evidence that at least a portion of the myelin damage present in HIV+ individuals on cART may be due to the antiretroviral drugs themselves.

5.0. Potential Mechanisms for Oligodendrocyte Dysfunction due to HIV or cART

Despite striking pathological findings in HIV-infected individuals, there is a notable lack of molecular studies detailing the cellular effects of secreted factors from HIV-infected macrophages or direct effects of antiretroviral compounds on oligodendrocyte precursors or mature oligodendrocytes. Evidence suggests that oxidative stress, inflammation, glutamate dysregulation, and ER stress are present in HAND patients (Akay et al., 2012; Gannon et al., 2017; Lindl et al., 2007). These processes are known to inhibit oligodendrocyte differentiation and/or cause oligodendrocyte death.

One mechanism of oligodendrocyte dysfunction in HIV+ patients might be through oxidative stress. Chronically high burdens of oxidative stress are abundant in HIV+ patients even following suppression of viral replication (Blas-Garcia et al., 2011; King et al., 2013). Oligodendrocytes are exceptionally susceptible to oxidative stress-mediated cell death due to low glutathione and antioxidant levels, coupled with high energy demands required to efficiently produce and maintain myelin to insulate surrounding neurons (Back et al., 2002, 1998; Folkerth et al., 2004). In rodent models of hypoxia/ischemia, oxidative stress generated by the insult led not only to immediate death of OPCs, but also to a subsequent reduction in number of mature oligodendrocytes and hypomyelination of white matter due to a halt in the oligodendrocyte differentiation process (French et al., 2009; Levison et al., 2001; Reid et al., 2012). In vitro studies have also demonstrated a maturation-stage specific vulnerability of this cell population, as mature oligodendrocytes are far more resilient than OPCs and immature oligodendrocytes due to higher glutathione and antioxidant enzyme levels (Back et al., 1998; Baud et al., 2004; Fern and Möller, 2000). Several demyelinating disorders including multiple sclerosis and periventricular leukomalacia involve oxidative stress as a contributing factor to pathology (Gilgun-Sherki et al., 2004; Haynes et al., 2003; van Horssen et al., 2008). With this in mind, it is likely that the chronically high-levels of oxidative stress still observed in cART-medicated HAND patients also contributes to the persistence of white matter pathologies in this population.

Another potential pathway contributing to oligodendrocyte dysfunction in HAND is the UPR/ER stress. Following HIV fusion and entry into the cell, one of the downstream cellular stress pathways activated in CD4+ T cells and monocytes is ER stress. Additionally, certain antiretroviral drugs are known to induce ER stress in numerous peripheral cell types. Thus, it is possible that HIV and/or antiretroviral drugs are activating ER stress in CNS resident cells and playing a role in the persistence of HAND pathology (Borsa et al., 2015; Gupta et al., 2007; McLean et al., 2009; Parker et al., 2005; Wu et al., 2010; Zhou et al., 2006, 2005). There is much evidence for ER stress response activation in other myelin disorders such as multiple sclerosis, vanishing white matter disease and Pelizaeus-Merzbacher disease (Clayton and Popko, 2016; Lin and Popko, 2009). Morphometric analysis shows that the mean surface area of myelin in a myelinating oligodendrocyte is thousands of times greater than the surface area of a typical mammalian cell (Lin and Popko, 2009). Myelinating oligodendrocytes are responsible for producing enormous quantities of plasma membrane, myelin membrane proteins, and cholesterol through the secretory pathway; thus, it is easy to imagine that oligodendrocytes are suspectible to stressors of the secretory pathway particularly the ER through the UPR (Lin and Popko, 2009; Pfeiffer et al., 1993). ER stress in oligodendrocytes is complex and studies have shown that genetic deletion of some of the component proteins are either protective or detrimental depending on the demyelinating disease (Inoue, 2017; Deslauriers et al., 2011; Southwood et al., 2002).

There are several additional signaling proteins in the brain known to inhibit oligodendrocyte differentiation and maturation. These include the bone morphogenetic proteins, Wnts, fibroblast growth factors, and notch/jagged/delta families (Figure 1; Feigenson et al., 2009; Grinspan et al., 2000; John et al., 2002; Mierzwa et al., 2013). These external signaling factors are present in early development and possibly serve to temporally regulate maturation of oligodendrocytes to match the availability of axons to myelinate. In some demyelinating diseases, such as multiple sclerosis and perinatal white matter injury, these factors are markedly increased in expression and are thought to inhibit myelination or remyelination (Ara et al., 2008; Fancy et al., 2009; John et al., 2002; Reid et al., 2012). A few studies have examined some of these pathways in the context of HAND. For example, one report demonstrated that HIV-Tat inhibited Wnt/β-catenin signaling in astrocytes (Henderson et al., 2012). Additionally, other studies have examined the role of neurotrophic factors, demonstrating decreased brain-derived neurotropic factor in HIV+ and HIV encephalitis patients (Fields et al., 2014; Mocchetti et al., 2014).

Finally, the myelin sheath is largely composed of lipid and myelin proteins, arranged in highly regulated lipid rafts that function in membrane stability and signaling (Gielen et al., 2006). Inhibition of lipid signaling compromises oligodendrocyte development and myelination (Camargo et al., 2017; Monnerie et al., 2017). Therefore, it would stand to reason that the lipodystrophy seen as a side effect of some antiretrovirals, such as ritonavir, would also affect the ability of the myelin sheath to form or to repair itself.

Going forward, much more work exploring these pathways and signaling cascades in oligodendrocytes in response to either HIV or cART exposure is necessary to have a complete understanding of oligodendrocyte involvement in HAND.

6.0. Models of WM damage in HAND

It has been difficult to reproduce the white matter deficits caused by HIV in the typical animal models used for study since HIV itself is species-limited to humans. However, the recent advent of the humanized mouse in which NOD/scid-IL-2Rgc null mice are transplanted with human CD34+ stem cells has allowed for examination of white matter in more detail, following productive peripheral infection. Despite lack of active brain infection in these animals, decreases were seen in myelin proteins MOG and MAG in whisker barrels and corpus callosum, with MAG decreases also evident in hippocampus via immunohistochemistry (Boska et al., 2014). In a subsequent study, RNA sequencing of HIV+ humanized mouse brain showed decreases in transcripts for MBP, PLP, MAG, and MOBP, and increases in LINGO1 which inhibits myelination (Li et al., 2017). Another recently developed animal model is the HIV-1 transgenic rat, which expresses the HIV provirus but harbors functional deletions of the viral proteins gag and pol (Reid et al., 2001). This model recapitulates many aspects of the human disease, including behavioral and neurological deficits, as well as astrocytic dysfunction and neuronal damage (Reid et al., 2016; Vigorito et al., 2007). In this rodent model, longitudinal diffusion tensor imaging revealed increased mean diffusivity (MD) and decreased fractional anisotropy (FA) values within the corpus callosum of transgenic animals relative to controls, suggesting demyelination and increased space between white matter tracts (Lentz et al., 2014). In a transcriptome profiling analysis of these transgenic rats, among the “neuronal survival associated genes” that were significantly downregulated were the myelin proteins: MAG, claudin1, and oligodendrocytic myelin paranodal and inner loop protein (Opalin) (Li et al., 2013). Expression of MBP, CNP, MOG, and aspartoacylase were also decreased in these animals, although to a lesser extent (Li et al., 2013).

The development of better animal models of HAND will be critical for understanding the effects of HIV and cART on CNS resident cell populations including oligodendrocytes in the absence of the confounding effects inherent in studies with human patients. Importantly, the use of animal models also allows for the independent investigation of HIV infection or cART on CNS cell populations, which will allow for the determination of which features of HAND result from viral or therapeutic means. At present investigators are still in the early stages of developing models which fully recapitulate HIV latency and the myelin deficits seen in virally suppressed HAND patients. Long term cART-related effects will be particularly important to evaluate using these models, as patients now remain on cART regimens for the entirety of their increased lifespans. Despite the relative paucity of reliable models of HIV infection that have been created thus far, it is clear from both the limited studies of present animal models combined with data from HIV+ patients, that oligodendrocyte viability and function are greatly impacted in the presence of HIV viral components and perhaps by antiretroviral agents as well.

7.0. Conclusions

Oligodendrocytes are frequently overlooked in the study of diseases; however, because myelin formation and maintenance during both during human development and adulthood are key to proper nervous system function, the consequences are profound when these processes are disrupted. In order to decrease the incidence of HAND, permitting both normal CNS function and increased compliance with cART regimens, further study of the effects of both HIV and cART used to suppress viral replication are needed to produce adjunctive therapies or alternative treatments to improve the quality of life for people living with HIV currently on cART.

Highlights.

White matter abnormalities are persistent in HIV+ individuals with HAND. This may be due to HIV reservoirs in the CNS, side effects of antiretroviral compounds or both. Some potential driving mechanisms of white matter damage include increased oxidative stress, activation of the unfolded protein response and/or dysregulation of lipid metabolism. Evidence that HIV proteins and select antiretroviral agents damage oligodendrocytes, the myelin producing glia cells in the CNS, suggests multiple mechanisms underlie observed changes in white matter in patients living with HIV.