Following the introduction of antiretroviral therapy (ART), many HIV-1-infected patients have survived the disease for more than 20 years, leading to a change in HIV-1 epidemiology. By 2015, it is projected that half of HIV-1 patients will be 50 years old or older (1). Because of this epidemiological shift, the healthcare response to the HIV/AIDS epidemic in western countries faces new challenges. According to clinical research data, a significantly higher occurrence of dementia has been observed in aged HIV-1-infected individuals when compared with younger patients, and HIV-1-associated dementia risk in these patients is more than three times higher than in younger people (2).
Table 1. Factors hypothesized to be involved in the increase/decrease of Aβ in HIV-infected brain (further work is needed)
While the introduction of ART has reduced the occurrence of HIV-associated dementia, the prevalence of minor HIV-1-associated neurocognitive disorders (HAND) is increasing. It appears that chronic ART medication causes subtle neurodegeneration especially in hippocampal neurons (3).
Increased Amyloid Deposition in HIV-1-Infected Brains
In earlier studies, accumulation of beta-amyloid precursor protein (APP) was observed in HIV encephalitis, demonstrating that widespread axonal injury is characteristic of brains of individuals with AIDS. This suggested a pathogenetic mechanism for the neuropsychological changes in these patients (4). In parallel, an increased prevalence of amyloid plaques was found in the cortex of AIDS brains compared with age-matched, non-HIV-infected controls. This suggested that an inflammatory response in the brain to HIV-1 infection could facilitate amyloid plaque formation (5), and this study was the first to show a relationship between AIDS and amyloid plaque deposition.
Subsequently, other reports also supported increased amyloid deposition in the brains of HIV-1-infected patients (6, 7), and the patient age was found to correlate with the intracellular deposition of amyloid beta (Aβ) (7). HAND in the elderly has been correlated, in part, with early beta-amyloidosis, which showed that Aβ deposition affects the clinical outcome of HIV-1 infection (3). Due to higher survival rates after ART, a confluence of factors occurs in which aging itself, HIV-1 infection, and the secondary effects of ART all contribute to brain Aβ accumulation.
One of the potential mechanisms involved in HIV-1 progression at older ages involves HIV-1 effects on amyloid pathology of the brain. HIV-1 can increase Aβ levels (7) by increasing its synthesis (8), decreasing its degradation (6), or changing its transport mechanisms across the blood brain barrier (BBB) (9), leading to its accumulation in the brain.
There are distinct differences in Aβ deposition in Alzheimer's disease (AD) and HIV-1-infected brains. Extracellular amyloid plaques can be found in AD, while intraneuronal amyloid accumulation or perivascular diffuse amyloid depositions are more prevalent in HAND (3).
The Role of HIV-1 Proteins in Aβ Pathology
Several HIV-1 proteins have been shown to be amyloidogenic. For instance, HIV-1 Tat protein was demonstrated to inhibit the Aβ-degrading enzyme neprilysin, leading to increased levels of soluble Aβ in cell culture (6). In transgenic AD mice having HIV-1 Tat-expressing astrocytes, more neurodegeneration and Aβ deposition was observed compared with mice not expressing astrocytic Tat (10).
Besides Tat, HIV-1 gp120 protein also promotes Aβ secretion in primary rat fetal hippocampal cultures (8) and causes APP accumulation and axonal injury in corpus callosum slices (11). Because HIV-1-associated neurocognitive impairments could not be correlated with viral load, it is believed that these soluble HIV-1 factors are important in neuroinflammation and Aβ accumulation.
Neuroinflammation and HAND
Neuroinflammation is an important factor in Aβ deposition in the brain and in the pathogenesis of HAND. HIV-1 infection of the brain is mainly found in cells of the macrophage/microglia lineage (12), and monocyte/macrophages were shown to infiltrate the brain in HIV-1 encephalitis (13). Astrocytes also take part in this complicated, amplified inflammatory process (14). In addition, the subtle neurodegeneration caused by ART is associated with neuroinflammation coupled with mononuclear phagocyte activation.
HIV-1-induced inflammatory mediators such as CCL2/MCP-1, which are produced during chronic neuroinflammation, may also contribute to increased levels of Aβ in the brain (15). In our in vitro studies, HIV-1 Tat and HIV-1 itself caused an increase in activity of several inflammatory gene promoters and their expressed proteins (MCP-1, E-selectin, and IL-6) (16), which could facilitate Aβ accumulation.
Proinflammatory molecules, Aβ, and secreted HIV proteins like gp120 and Tat can all have neurotoxic effects via glutamate excitotoxicity [for review see ref. 17]. This strongly suggests that neurodegeneration and dementia in AD and AIDS have common pathogenic mechanisms, which are far from being completely understood.
ART and Aβ Accumulation in the Brain
Different ART medications have distinct properties for penetrating the BBB according to the central nervous system (CNS) penetration-effectiveness index (18). Several ART drugs reach the index of 1 and higher, indicating good penetration into the brain. ART protease inhibitors have been shown to contribute to Aβ deposition in the brain by inhibiting an Aβ-degrading enzyme (3).
In studies using ART drugs that cross the BBB, it was demonstrated that ART increases brain Aβ levels by increasing neuronal Aβ generation and inhibiting microglial phagocytosis (19). As a consequence, ART medications may contribute to the development of HAND.
The BBB is Critical in Brain Aβ Accumulation—Aβ Transporters at the BBB
The BBB consists of brain microvascular endothelial cells joined by tight junctions that prevent the unregulated exchange of substances between brain and blood. A dysfunctioning BBB with a decreased brain-to-blood clearance of Aβ could be the underlying mechanism of brain Aβ accumulation (20). Thus, the BBB has a critical role in this process. A balance between lipoprotein receptor-related protein (LRP1, which transports Aβ from the brain into the blood) and the receptor for advanced glycation end products (RAGE, which transports Aβ into the brain) in the BBB has been proposed to regulate Aβ levels in the brain (21).
While there are several reports relating Aβ to AD, the mechanisms of Aβ uptake at the BBB level in HIV-1 infection are largely unknown. Below, we will describe the evidence for LRP1, RAGE, and other transporters that have been implicated in Aβ transport at the BBB and, whenever possible, present data in the context of HIV-1 infection.
Several factors have been described as being responsible for soluble Aβ removal from the brain into the circulation such as LRP1, which is expressed in blood vessels (22). LRP1 belongs to the low-density lipoprotein receptor family and takes part in cholesterol transport and the endocytosis and transcytosis of 40 structurally diverse ligands, including Aβ, across the BBB. Soluble LRP1 (sLRP1) circulates in the plasma and acts as a key endogenous peripheral “sink” for Aβ, resulting in constant removal of Aβ from brain (23). In healthy human and mouse brain, sLRP1 normally binds >70% of circulating Aβ (23).
However, the impact of LRP1 in brain is somewhat controversial, as it has been shown to mediate bidirectional transcytosis of Aβ across the BBB in an in vitro study using mouse brain endothelial cells (24). But an in vivo study clearly demonstrated that decreasing LRP-1 levels with LRP1-antisense oligonucleotides leads to decreased Aβ clearance, accumulation in the brain, and, as a consequence, to learning/memory impairments, firmly supporting the neurovascular hypothesis for AD (25).
Our laboratory was the first to explore HIV-1-induced LRP1 level changes in an in vitro BBB model. In our experiments, HIV-1 effects on LRP1 expression were inconsistent and, overall, did not reach statistical significance compared with control (9).
The primary transporter of Aβ in the other direction, from the blood to the brain across the BBB, is the RAGE (26). When an amino group of a protein reacts with a carbonyl group of a sugar, after several chemical steps, advanced glycation end products (AGEs) are produced. The receptor for AGEs (RAGE) is a transmembrane protein that belongs to the immunoglobulin superfamily [for a review see ref. 27].
RAGE is a pattern-recognition receptor that has multiple ligands, among them Aβ. RAGE is the binding site for Aβ on endothelial cells, mediates Aβ-induced effects, and binds to different Aβ forms (28). Soluble RAGE (sRAGE) also binds different Aβ species (29). RAGE is upregulated in AD and can be found in amyloid plaques (30). At the BBB, RAGE mediates entry of circulating Aβ into the brain by a receptor-dependent transcytosis (26). Circulating Aβ can also get into the brain if the BBB is disrupted (31) via proinflammatory cytokines, possibly induced by the Aβ–RAGE interaction (26).
Our previous data indicated that exposure to HIV-1 elevates Aβ levels in brain endothelial cells, partly by increased expression of RAGE, which transports Aβ into the brain (9). However, there is limited data about the role of RAGE in HIV-1 infection. Besides the two papers from our laboratory (9, 32), there are only two publications on this subject. One of these papers found that ART-induced hyperglycemia may lead to the formation of AGEs (33). AGEs interacting with RAGE can induce inflammation and facilitate Aβ accumulation.
In addition to directly transporting Aβ into the brain, RAGE–Aβ interactions may upregulate chemokine receptor-5 (CCR5) in endothelial cells, which could facilitate HIV-1-infected T-cell migration into the brain, contributing to more neuroinflammation and Aβ deposition (34).
Besides LRP1 and RAGE, other transporters like multidrug-resistance protein 1 (MRP1) are also believed to be implicated in Aβ transport across the BBB. In a mouse model lacking MRP1, increased brain Aβ was reported (35). Currently, there are no data available about the role of MRP1 in HIV-1-induced Aβ accumulation.
P-glycoprotein (P-gp) has also been proposed as an Aβ efflux transporter at the BBB. In a transgenic mouse model of AD brain capillary P-gp, expression and transport activity were diminished, offering another possible mechanistic link to Aβ accumulation in the brain. When P-gp expression and transport activity were restored by an activator of the nuclear receptor pregnane X receptor, brain Aβ levels were reduced. This suggests that upregulation of BBB P-gp may increase Aβ clearance from the brain (36). In support of this finding, recent human in vivo data showed decreased P-gp function in AD patients (37). Decreased expression of microvascular P-gp in HIV-1 encephalitis was also reported (38, 39).
Breast cancer resistance protein (BCRP) is another candidate for a role in Aβ transport, which is believed to act by restricting apical-to-basolateral permeability of the BBB to Aβ and its transfer into the brain (40). In support of this idea, BCRP was upregulated in AD brains, possibly as a protective mechanism (41). A complete loss of microvascular BCRP occurred in HIV-1 encephalitis (42).
In a genome-wide association study, phosphatidylinositol-binding clathrin assembly protein (PICALM, also known as CALM) and apoJ (also known as clusterin) were identified as AD susceptibility genes (43) in addition to the apoE4 gene. ApoJ, a ligand for the LRP2 (or megalin) not only controls Aβ transcytosis across the BBB, facilitating an efflux from the brain to blood (44) but can also mediate re-entry of circulating Aβ into the brain (45). There are no reports about the role of PICALM in Aβ transfer across the BBB. Interestingly, increased blood apoJ levels were found in AD (46), but the underlying mechanisms are unclear. There are no data available for HIV-1 infection.
Aging by itself has been shown to alter Aβ transporter levels. For instance, RAGE levels were increased (47) while LRP1 and P-gp were decreased in aging rats (48). This effect may be important for an aging HIV-1-infected population.
The Role of Lipid Rafts in Aβ Uptake at the BBB
The plasma membrane has special microdomains called lipid rafts, which are enriched in cholesterol, sphingolipids, and saturated fatty acids. Changes in lipid rafts might play a role in different pathologies, and previous reports have claimed that lipid rafts may be involved in neurodegenerative diseases such as AD (49). In support of this possibility, Aβ was found together with the lipid raft ganglioside GM1 in the brains of AD patients (50). Moreover, the distribution of GM1 and GM2 changes in AD (51), and lipid rafts mediate endocytosis for Aβ in neurons (52).
In our previous studies, we explored whether HIV-induced Aβ accumulation in human brain endothelial cells is lipid raft-dependent, which was motivated by the findings that (a) in endothelial cells, RAGE is found in a subtype of lipid rafts called caveolae (53), (b) caveolin-1 activates RAGE (54), and (c) HIV-1 Tat mediates Ras-MAPK signaling in brain endothelial cells [results from our laboratory, (55)], and a role for caveolae in this pathway has been suggested (56). We found that HIV-1-induced accumulation of Aβ in our in vitro BBB model was lipid raft- and caveolae-dependent. Furthermore, these mechanisms involved Ras-MAPK signaling via upregulation of RAGE expression (32) (Fig. 1).
The Role of Proteoglycans in Brain Aβ Deposition
In addition to the mechanisms at the BBB explored above, there is a recent report indicating a role for agrin in brain Aβ deposition. Agrin is the major heparan sulfate proteoglycan found in amyloid plaques in AD brains. Agrin also plays a crucial role in astrocyte endfeet polarization and BBB function (57). In AD mice in which the Agrn gene was deleted from endothelial cells, Aβ levels were significantly increased in the brain. This effect could be reversed by overexpression of Agrn, suggesting that changes in brain vascular Agrn expression influence brain Aβ homeostasis (58). There are no data for the role of agrin in HIV-1-induced amyloid pathology.
Inhibition of Aβ Accumulation
Statins are drugs that effectively diminish cholesterol biosynthesis by inhibition of 3-hydroxy-3-methylglutaryl coenzyme A reductase. In HIV-infected patients, statins have been successfully introduced to improve disturbances in lipid profile. By affecting prenylation of the small GTPases, statins may also inhibit endothelial inflammatory pathways. It is also possible that statins affect membrane lipid rafts, which are key cell signaling microdomains (59). Previous studies from our lab showed that simvastatin protects against Aβ or HIV-1-induced inflammatory responses (16).
Statins are known to upregulate LRP1 on the BBB (21), and in our experiments, simvastatin not only increased LRP1 levels in control human brain endothelial cells but also in HIV-1-exposed endothelial cells. Moreover, to our surprise, simvastatin was able to reduce RAGE levels in our BBB model. Simvastatin also blocked extracellularly administered Aβ uptake/accumulation in the BBB. These results all suggest a beneficial effect of simvastatin on BBB Aβ homeostasis (9), which may be important in an environment where HIV-1 infection, chronic inflammation, and ART are present. Because defective cholesterol metabolism has been shown to cause neurodegeneration, statins are probably beneficial in the brain by decreasing neuroinflammation.
Targeting Aβ Transport Mechanisms
By decreasing RAGE levels (9), simvastatin may not only reduce Aβ transport into the brain but might also inhibit Aβ-RAGE-mediated inflammatory responses as well. Several Aβ-RAGE blockers are already being used in clinical trials for AD (60).
Besides influencing the main BBB Aβ transporters in HIV-1 infection, there are other potential mechanisms also to be explored such as the “peripheral sink” effect. Previously, in AD models, this effect was used to decrease the brain Aβ burden, with very promising results. In mice treated with sRAGE, brain Aβ decreased (26). While sRAGE could not enter the brain in these experiments, it was still able to decrease brain Aβ. These results suggest that similar concepts/treatments might be applied in HIV-1 infection in the prevention or treatment of HAND.
Using the Peripheral Sink Effect of Aβ-Degrading Enzymes
In experiments with an excellent outcome, expression of neprilysin (an Aβ-degrading enzyme) in peripheral blood was able to reduce Aβ in the brain (61). No data are available about this technique in the context of HIV-1 infection.
Creating More Specific ART Medications
Protease inhibitors used in ART were reported to reduce LRP levels (62) due to a 63% homology between the HIV1-protease and LRP. This could be one of the mechanisms by which brain Aβ accumulation occurs in ART-treated patients. Designing more specific ART protease inhibitors should diminish this side effect.
As the ART-treated HIV-1-infected population ages, brain Aβ deposition and the associated neurocognitive disorders represent an increasing problem. The pathogenic factors involved offer several therapeutical targets to potentially reduce the brain Aβ burden and/or improve the HIV-1 associated neurocognitive impairments (Table 1).
This work was supported by the National Institutes of Health, grants MH63022, MH072567, DA027569, NS39254 and by the University of Miami Developmental Center for AIDS Research grant (P30A1073961).