Encyclopedia of AIDS

Living Edition
| Editors: Thomas J. Hope, Douglas Richman, Mario Stevenson

HIV Reservoirs in the Central Nervous System

  • Lauren Tompkins
  • Elena Dukhovlinova
  • Ronald SwanstromEmail author
Living reference work entry
DOI: https://doi.org/10.1007/978-1-4614-9610-6_428-1


The central nervous system (CNS) is a unique anatomical compartment that is shielded from the rest of the body by the blood-brain barrier (BBB). Exchange of substances between the CNS and the blood is controlled and limited by the BBB. Human immunodeficiency virus (HIV) enters the CNS early after viral transmission and, in some cases, replicates independently of that in the blood. Viral replication in the CNS could be supported in multiple cell types including CD4+ T cells, macrophages, microglia, and potentially astrocytes or other cells lacking CD4 receptor expression. Following productive infection of permissive cells, it is plausible that viral species persist in the CNS in the form of a reservoir, or dormant state of infection that is capable of reactivation and production of new viral particles. Analyzing such a reservoir in the context of the CNS is challenging, given the fact that brain tissue can only be sampled at death. Yet, studies using cerebrospinal fluid (CSF) as a surrogate for analyzing HIV infection of the CNS have described four states of infection, which will be discussed in this section. Importantly, physiological and viral factors influence the outcome of HIV replication in the CNS, which affects the composition of a persistent viral reservoir within an individual. Future work aimed at a total cure to HIV infection will need to address all potential viral reservoirs, including those that reside within the CNS.


Human immunodeficiency virus type 1 (HIV-1) is the cause of acquired immunodeficiency syndrome (AIDS), where uncontrolled infection in the blood and lymphoid organs depletes CD4+ T cells and diminishes immune competence over time. Although antiretroviral therapy (ART) controls viral replication and disease progression, these drugs must be taken throughout life to suppress infection. A major focus of current research is to eradicate HIV-1 and cure infected patients. HIV-1 persistence in cellular and anatomical reservoirs precludes a cure; thus, efforts to characterize these reservoirs are an important part of developing a strategy for eradicating all forms of HIV-1.

In order to discuss reservoirs of HIV-1, it is important to understand how the virus utilizes different cell surface receptors for productive infection. HIV-1 infects cells that express CD4 (the viral receptor) and CCR5 or CXCR4 (viral coreceptors). Preferential infection of a specific cell type by HIV-1 (cell tropism) is defined by receptor usage. Most transmitted/founder virus strains are CCR5-utilizing and infect CD4+ T cells (R5 T cell-tropic), but over time, HIV-1 can evolve to utilize CXCR4 expressed on CD4+ T cells (X4 T cell-tropic) due to the depletion of CCR5+ CD4+ T cells from untreated infection. Additionally, HIV-1 can evolve to infect macrophages, which express CD4 and CCR5 (R5 macrophage-tropic). R5 macrophage-tropic viruses are predominantly found in the central nervous system (CNS) as an evolutionary adaptation to the paucity of T cells in the CNS. The composition of the latent reservoir could differ between individuals depending on nadir CD4, although probably all permissive cell types are infected at all times, at least at low level (reviewed in Joseph 2014b).

The central nervous system (CNS) is an anatomic compartment that harbors potentially multiple cellular reservoirs of HIV-1. The CNS is a unique anatomical site as it is relatively “immune privileged” due to the physiology of the blood–brain barrier (BBB), a selectively permeable barrier that controls communication between the CNS and the periphery. It is difficult to investigate CNS infection, as brain tissue is only available at autopsy. The cerebrospinal fluid (CSF), which baths the CNS, is a surrogate for studying CNS infection, but these studies are limited in their ability to characterize CNS tissue reservoirs of HIV-1. To understand potential CNS reservoirs of HIV-1, it is important to discuss the following: the dynamics of viral and permissive cell entry into the CNS; the conditions that favor viral replication and contribute to reservoir stability; and the potential for viral egress from cells constituting the CNS reservoir.

Defining a Viral Reservoir

Two essential criteria exist to define a viral reservoir of HIV-1 (reviewed in Blankson et al. 2002). First, a reservoir must preserve replication-competent virus in some form (i.e., viral particles or viral genomes) so that the virus can reestablish productive infection in the future. Second, a reservoir must have mechanisms of longevity. For example, reservoirs composed of virions would require escape from biochemical decay, as seen with virion particles that become trapped extracellularly in dendritic cell processes. Cell-associated viral reservoirs, however, require cell survival and escape from immune control including cytotoxic T cell (CTL) activity. Latent infection, a state with no active replication, is the best characterized cellular reservoir of HIV-1, but another type of reservoir could be composed of productively infected cells with slow turnover. The concept of a reservoir is further complicated by the detection of cells that clonally expand in vivo through transactivation of cellular growth-promoting genes by integrated viral DNA (Maldarelli et al. 2014).

A reservoir will most likely occur in cells that are normally infected with HIV-1. This virus infects cells that express the viral receptor CD4 and coreceptor (CCR5 or CXCR4). Activated CD4+ T cells are the most permissive cell type for HIV-1. Latently infected resting memory CD4+ T cells are the hallmark reservoir of HIV-1 infection and are thought to predominantly arise from productive infection of activated CD4+ T cells as they are transitioning back into a resting state (reviewed in Siliciano and Greene 2011). Latently infected T cells contain replication-competent HIV-1 genomic DNA (provirus) integrated within the human genome in the absence of ongoing viral replication. Although dormant, latently infected resting T cells can be induced to become activated and thereby transcribe integrated DNA to generate new progeny virions capable of productive infection. It is important to note that integrated DNA can be either intact or defective; thus, many infected T cells contain HIV-1 DNA that is incapable of producing functional virions (Ho et al. 2013). Latently infected cells are established early after a person becomes infected, and these cells persist even in the presence of ART-mediated viral suppression. Persistence of the latent reservoir is due in part to the virus escaping immune surveillance, as integrated viral DNA in latently infected cells is likely transcriptionally silent and does not produce antigens that signal immune attack (reviewed in Archin et al. 2014).

A second type of HIV-1 reservoir could be a productively infected cell with slow turnover. After the onset of suppressive ART, which prevents new infections without affecting previously infected cells, the blood viral load decreases rapidly (1–2 weeks) due to rapid turnover of short-lived infected T cells. For most people, the same pattern of rapid viral decay is observed in the CSF. However, occasionally the viral load decays much more slowly in the CSF than in the blood, which suggests that cells with a longer half-life than T cells can support HIV-1 replication in the CNS (reviewed in Joseph 2014b). Potential CNS cell types infected with HIV-1, including T cells as well as longer-lived cells, will be discussed later in this chapter.

The BBB and BCSFB: Barriers to HIV Infection of the CNS

The CNS, consisting of the brain and spinal cord, is an anatomical compartment isolated from the rest of the body by the blood–brain barrier (BBB) and blood-cerebrospinal fluid barrier (BCSFB). The BBB and BCSFB are semipermeable barriers that limit exchange of substances between the CNS and peripheral blood. These physiologically distinct barriers influence the composition of CNS fluid, including the CSF and CNS interstitial fluid, which differs greatly from the blood. For example, concentrations of white blood cells, albumin, and immunoglobulin G (IgG) in the CSF are all <1 % of that in the blood in spite of the fact that the water portion of the CSF is derived from the blood plasma. Key to barrier function of the BBB and BCSFB are intercellular tight junctions, which connect cerebrovascular endothelial cells (BBB) or choroid plexus epithelial cells (BCSFB). Otherwise, the BBB and BCSFB differ in physical composition and function (reviewed in Ransohoff and Engelhardt 2012).

The BBB is present along CNS blood vessels throughout the brain. Tight junction-connected BBB endothelial cells are separated from the brain parenchyma (the brain tissue proper) by two basement membranes: the endothelial and parenchymal. At the capillary level, these membranes are fused, but at all other vascular levels, these membranes separate to delineate the CSF-filled perivascular space (reviewed in Engelhardt and Ransohoff 2012). The BBB exhibits specialized function based on its position within the overall CNS vasculature: Nutrient transport occurs primarily at capillaries that lie in close proximity to neurons, whereas immune modulation occurs at the postcapillary venule (a small vessel that blood flows through after leaving the capillaries) where the perivascular space can accommodate the presence and movement of cells (reviewed in Obermeier et al. 2013).

The BCSFB resides in the choroid plexus located in each of the four brain ventricles. The choroid plexus is composed of fenestrated, permeable blood vessels (vessels containing endothelial pores) surrounded by tight junction-connected epithelial cells that are directly exposed to the CSF. Ependymal cells of the choroid plexus produce CSF from arterial blood. Newly produced CSF fills the brain ventricles, circulates around the exterior surfaces and perivascular spaces of the brain, and ultimately is reabsorbed into venous blood at the meninges. CSF flow is mediated by pulsation of the choroid plexus and action of ependymal cell cilia. Interstitial fluid of the brain parenchyma is also drained into the CSF, a fluid that acts as a surrogate for lymph by mediating immune surveillance throughout the CNS (reviewed in Spector et al. 2015).

The CSF is considered to be an immunologically active fluid as it houses T cells, B cells, and monocytes, cells which have limited access to the CNS. During physiologic conditions, T cells are primarily restricted to the CSF. However, monocytes can exit blood vessels, enter the brain, and differentiate into macrophages that primarily concentrate around CNS vasculature (perivascular macrophages) but can also be found in the meninges (meningeal macrophages) and choroid plexus (choroid plexus macrophages). CNS myeloid cells, as well as microglia (resident macrophages of the brain), have antigen-presenting capability and are considered important for immune surveillance and interaction with circulating central memory T cells (reviewed in Ransohoff and Engelhardt 2012).

Viral Entry and Cell Migration into the CNS

HIV-1 is found in the CNS very early after transmission, although the associated mechanisms of entry are not well understood (Sturdevant et al. 2015). One theory, which is the most favored, proposes trafficking of HIV-infected CD4+ T cells into the CNS as part of routine immune surveillance or in the context of neuroinflammation. Alternatively, HIV-1 virions could cross the BBB/BCSFB, likely in the setting of high blood viral load. Consistent with this hypothesis, some studies suggest that HIV-1 can specifically interfere with the production of proteins involved in the maintenance of tight junctions thereby disrupting the integrity of the BBB (reviewed in Zayyad and Spudich 2015). In contrast, the work of Price and colleagues has shown that, in the absence of neurocognitive symptoms, the CSF viral load decreases late in infection, while the viral load in the blood is on average increasing. This disparity in viral load occurs as CD4+ T cells are being depleted in the blood and white blood cell (WBC) count drops in the CSF, the latter of which suggests that T cells in the CSF are also reducing in number. These data suggest that the reduction in CSF viral load is due to a loss of CD4+ T cells that come from the blood, and thus virus enters the CNS in the form of trafficking T cells (reviewed in Price et al. 2014). Regardless of how HIV-1 enters the CNS, persistent infection in the brain requires a population of permissive cells that reside in or are trafficked to the CNS.

Peripheral immune cells, including T cells and monocytes, leave the blood and enter the CNS through a sequential mechanism of cell binding to the vascular endothelium, rolling along the endothelial lining, and extravasation into the CSF and/or CNS. First, circulating immune cells transiently bind blood vessel endothelial cells through receptor-ligand interaction involving either the selectin family of adhesion molecules or α4-integrins (reviewed in Ransohoff and Engelhardt 2012). Through these interactions, immune cells roll along the endothelium until they come into contact with chemokines that bind to their G-protein-coupled receptors on the immune cell surface. Chemokine binding induces signals that lead to increased affinity and avidity of cell surface integrins for their ligands and arrest of rolling cells. Immune cells then crawl along the endothelium by exposing protrusions that search for sites to cross, a process mediated by additional adhesion molecule interactions. Finally, the cell traverses the endothelium, either between endothelial cells or through endothelial pore-like structures, into the CSF-filled spaces of the brain. The cell then crosses the parenchymal basement membrane and underlying glia limitans, a membrane composed of astrocyte cell foot projections, to enter the parenchyma (reviewed in Engelhardt and Ransohoff 2012).

Brain pathology affects barrier function of the BBB/BCSFB, and as an associated immune response is mounted, this increases the number of CNS-infiltrating immune cells. During inflammation, CNS inflammatory cells secrete leukocyte-attracting chemokines and endothelial cells upregulate expression of adhesion molecule receptors, thereby aiding immune cell recruitment and extravasation into the CNS. Neuroinflammation can be a protective immune response to CNS tissue damage or infection, but immune pathology can also compromise the BBB/BCSFB, which alters CNS homeostasis and the proportions of immune cells in the brain (reviewed in Obermeier et al. 2013).

HIV-1 infection is associated with increased inflammation and immune activation both systemically and in the CNS. Neuroinflammation and BBB/BCSFB integrity both appear to affect the population of HIV-1 present in an infected CNS in a multifaceted, dynamic manner. Theories behind some of this complexity have been proposed based on studies comparing viral populations in the CSF, a surrogate for examining the CNS in living people, with those in the peripheral blood of the same individuals. First, in people who have very low or undetectable CSF viral load, it is likely that virus is not replicating, at least not appreciably, in the CSF or CNS but perhaps gains transient entry to the CNS. CSF virus likely comes from infected CD4+ T cells in the blood that cross the BBB/BCSFB and release viral progeny in the CNS. Second, in people who have elevated levels of HIV-1 in the CSF that is genetically similar to virus in the peripheral blood, it is likely that CSF virus also comes from migrating infected T cells but now in higher levels as part of an immune response associated with increased white blood cell count in the CSF (pleocytosis). Low-level, focal replication in the CSF or CNS could also account for increased viral load with or without pleocytosis, although without establishment of a persistent CNS infection. Third, in people who have detectable CSF virus that is genetically distinct from that in the peripheral blood, CSF virus could come from transient, clonal amplification of certain viral species in the CSF or CNS that may or may not establish persistent infection over time. The term “compartmentalized” viral replication is used to describe independent replication of HIV-1 within a given bodily compartment and is illustrated by genetic differences in the viral population between compartments, such as the CSF/CNS and blood. “Equilibrated” viral replication defines a state where HIV-1 populations are genetically similar between two compartments due to ongoing or recent intercompartmental movement of viruses (Sturdevant et al. 2015).

Roughly 30 % of acutely HIV-1 infected people have pleocytosis, which generally correlates with higher viral load in the CSF. Thus, increased viral burden in the CSF could result from an influx of infected cells in response to neuroinflammation. Consistent with this hypothesis, BBB dysfunction, indicated by an increased CSF/blood albumin ratio, often accompanies pleocytosis. A loss of barrier integrity in the context of enhanced immune cell trafficking to the CNS could allow more infected cells and/or cell-free virus to enter the CNS. Viral replication in the CSF/CNS also increases viral load in the CSF, irrespective of pleocytosis. Yet, pleocytosis occurs in a fraction of people with compartmentalized CNS replication and in an even greater proportion of people with equilibrated replication. How pleocytosis affects viral replication in the CNS is unclear, and it is plausible that pleocytosis occurs as a consequence of CNS HIV-1 infection. On the other hand, an influx of permissive cell types for infection could also promote viral replication in the CNS (Sturdevant et al. 2015; Spudich et al. 2005).

CNS Cell Types as Potential Reservoirs

CD4+ T cells

The primary target of HIV-1 infection is the CD4+ T cell; however, there are relatively few T cells in the healthy CNS. The concentration of T cells found in the CSF is less than1 % of that found in blood and even fewer, if any, are seen in the brain parenchyma. Despite the low absolute number of T cells present, the CSF has a relatively large proportion of permissive T cells; the CSF cellular composition includes primarily T cells (90 % of total CSF cells), which are mostly of memory phenotype (central and effector) and recently activated (CD69+) (reviewed in Ransohoff and Engelhardt 2012). As noted above, pro-inflammatory conditions promote immune cell influx into the CSF/CNS, thus increasing the number of potential target cells for HIV-1 replication.

The traditionally described cellular reservoir of HIV-1 is the latently infected T cell. For such a cell to contribute to a CNS reservoir of HIV-1, the cell must reside over time in the CNS. CD8+ T cells have been shown to persist in the CNS of mice infected with vesicular stomatitis virus (VSV). These cells are CD103+, which is an integrin found on tissue-resident CD8+ T cells, and expression of CD103 follows antigen recognition in the brain. Furthermore, CD103 appears to be important for retention of CD8+ T cells in the CNS, as knockdown of this molecule resulted in reduced accumulation of CNS T cells. Interestingly, CNS-resident CD8+ T cells in the brain parenchyma were shown to form clusters, some of which contained CD4+ T cells (Wakim et al. 2010). Although CNS CD4+ T cells were not analyzed thoroughly in this study, another study showed that tissue-resident CD4+ T cells in the skin are antigen-experienced and express CD103 (Watanabe et al. 2015). Taken together, these data suggest that a population of CNS-resident CD4 + CD103+ T cells could exist in the CNS and harbor HIV-1.

Some evidence exists in support of HIV-1 replication in CNS T cells. In some cases there is elevated viral load in the CSF sufficient to indicate HIV-1 replication in the CSF/CNS and rapid viral decay in the CSF upon ART initiation, suggesting that this virus was replicating in a short-lived cell, such as a T cell. Some people with rapid CSF viral decay have compartmentalized replication of HIV-1 in the CSF, and CSF virus is T cell-tropic, meaning that the virus replicates best in T cells compared to other cell types. Such individuals likely have a CNS-derived population of HIV-1 arising from infected T cells, a population that differs from T cell-tropic virus in the blood (Joseph 2014a). Alternatively, people with relatively high CSF viral load, equilibrated viral population, and evidence of pleocytosis may also have HIV-1 replicating in CNS T cells due to an increase in CSF/CNS T cell concentration but in a manner that does not result in a distinct population of virus in the CSF compared to the blood (Sturdevant et al. 2015).


Slow decay of CSF virus with ART suggests that, in this case, HIV-1 is being produced from a longer-lived cell type than a T cell. HIV-1 cell tropism depends at least in part on CD4 receptor expression density on the surface of a cell. R5 T cell-tropic virus replicates robustly in cells that express high levels of CD4 (T cells) but poorly in cells that express low levels of CD4, including macrophages, which have a similar number of cell surface CD4 molecules to T cells, but the molecules are less densely packed due to the larger surface area of macrophages (Joseph 2014a).

The ability to use low levels of CD4 for cell entry is an evolved feature of the viral envelope gene that cannot be attributed to a single mutation (Arrildt 2015). Rather, macrophage tropism likely evolves as an adaptation to the lack of CD4-rich target cells in the CNS, and the evolution of macrophage tropism appears to involve multiple genetic changes that differ between people. Pleocytosis further complicates macrophage-tropic evolution as it may alter the relative proportion of permissive cell types in the CNS thereby supporting viral replication from either T cell- or macrophage-tropic lineages. Indeed, a rhesus macaque animal model of HIV-1 infection showed that infection causes activation of bone marrow-derived monocytes and increased traffic of activated monocytes to the CNS with subsequent differentiation into CNS macrophages (Burdo et al. 2010).

The CNS is rich in macrophages that could serve as a reservoir for HIV-1. Perivascular macrophages, choroid plexus macrophages, and meningeal macrophages are all bone marrow-derived and are named for anatomical regions in which they reside. These cells could be infected by macrophage-tropic virus or with much lower efficiency by an R5 T cell-tropic virus. Perivascular macrophages are likely exposed to cell-free or cell-associated virus that crosses the BBB. Such virus could come from either the blood or CSF, depending on barrier physiology at the point of entry. Indeed, immunohistochemical staining of autopsied brain shows the presence of HIV-1 nucleic acid and protein in perivascular macrophages. Similarly, meningeal macrophages, located at the superficial brain meninges, are likely also exposed to blood or CSF virus that crosses the leptomeningeal BBB. Choroid plexus macrophages, on the other hand, are likely exposed to predominantly blood virus, as these macrophages are located in the choroid plexus stroma, which harbors fenestrated capillaries that provide blood for the production of CSF (reviewed in Joseph 2014b).

A rhesus macaque animal model of HIV-1 infection suggested that the virus could migrate between the CNS meningeal and parenchymal regions or replicate autonomously in each of them. The rapid migration of genetically homogeneous virus throughout the brain was associated with faster disease progression and widespread encephalitis. Furthermore, compartmentalized replication in the meninges versus parenchyma was associated with localized detrimental inflammation within these respective brain regions. One macaque with compartmentalized replication in the meninges versus parenchyma was suggested to have macrophage-tropic virus present in both regions, an observation that exemplifies how regional macrophages may contribute to CNS infection and disease. Collectively, these data suggest that the uncontrolled replication of HIV-1 in different brain regions, and potentially in regional macrophages, may be detrimental for local physiology through induction of pathological inflammation (Matsuda et al. 2013).


Microglia are resident macrophages of the CNS and the predominant immune cell type in the brain parenchyma. Unlike macrophages, microglia are not bone marrow-derived, rather they arise during embryonic development and are maintained throughout adulthood via local proliferation. Microglia have immune functions including phagocytic ability, inflammatory cytokine secretion, and weak antigen presentation. Studies using HIV-infected human brain tissue at autopsy show that microglia can contain HIV-1 nucleic acid and protein. Furthermore, as with monocyte-derived macrophages, HIV-1 can infect microglia in vitro, suggesting that this cell type is permissive to HIV-1 infection. Like macrophages, microglia have low surface densities of CD4, so virus capable of successfully infecting these cells is likely macrophage-tropic (reviewed in Joseph 2014b). Microglia are thought to have very long life spans, even longer than CNS bone marrow-derived macrophages, thus persistent infection of these cells could constitute a CNS reservoir of HIV-1. Alternatively, persistent replication in this cell compartment deep in the brain parenchyma could maintain an active reservoir even in the face of poorly penetrating anti-HIV-1 therapy.


Astrocytes provide mechanical and metabolic support for neurons and are the most abundant cell type in the brain. Viral DNA has been detected in astrocytes of HIV-infected people, and astrocytes can be infected at low levels in vitro (reviewed in Joseph 2014b). However, it is unclear whether astrocytes are productively infected in vivo, as they express no CD4, the viral receptor. Indeed, an analysis of macrophage-tropic HIV-1 env genes from individuals with HIV-associated dementia failed to detect CD4-independent infection (Joseph 2014a). Yet, CD4-independent infection was characterized in another study where rhesus macaques were infected with chimeric human/simian immunodeficiency virus (SHIV) that contains an R5 T cell-tropic HIV-1 env in the context of an SIV backbone. Still, only one HIV-1 env clone isolated from the CNS of a single infected macaque was able to infect CD4+ cells in vitro, so it is difficult to draw definitive conclusions from a single observation (Zhuang et al. 2014). An alternative explanation for the presence of HIV-1 nucleic acid in astrocytes is that these cells have phagocytic ability and could ingest infected T cells (reviewed in Joseph 2014b). The extent to which HIV-1 can enter cells in the absence of receptor and/or coreceptor is a poorly studied issue that deserves more attention given the number of cells without viral receptors present in a person and the concern that these alternative cell types could contribute to the reservoir.

Assessing the Contribution of CNS Reservoirs to Viremia

The defining criterion of a reservoir is that the virus is preserved in some form that allows for reestablishment of productive infection. In the case of the blood reservoir, latently infected CD4+ T cells can be isolated from the blood and used in a viral outgrowth assay to directly assess the prototypical latent reservoir (reviewed in Archin et al. 2014). Such methods are essentially moot for analyzing the CNS reservoir due to the logistics of collecting viable CNS cells postmortem in a timely manner. An alternative method of determining whether HIV-1 in the CNS can reestablish infection is to characterize virus in the CSF that emerges following ART treatment interruption (rebound virus).

HIV-1 rebound virus appears in the CSF roughly 2 weeks after the detection of virus in the blood (de Almeida 2005). Phylogenetic analysis of rebound viral populations in the CSF versus blood could be used to determine if populations in these two compartments differ, indicating that “compartmentalized” CSF rebound virus comes from the CNS and thus illustrates the presence of a CNS reservoir of HIV-1. Furthermore, if viral species previously confined to the CSF arise in the blood during rebound, then reestablishment of systemic infection would be influenced by the CNS reservoir. Although it would be difficult to prove what cell type recrudesced CNS virus originates from, the use of in vitro infectivity assays would illustrate cell tropism of rebound virus. The phenotype of CSF rebound virus likely depends on the state of viral replication in the CNS prior to initiation of ART; therefore, studies are required to characterize viral populations throughout the brain of ART-naïve as well as experienced individuals.

Although evidence greatly supports the concept of HIV-1 persistence being attributable to latency, the immune privileged CNS may represent a unique anatomical reservoir, as crosstalk with the periphery is limited due to the BBB, and many ART drugs are relatively poor at penetrating the CSF/CNS. Treatment intensification using ART drugs with optimal CNS penetrance (relative to others) does not reduce levels of HIV-1 RNA in the CSF (Yilmaz et al. 2010). These data suggest that low-level viral replication does not account for the presence of residual CSF virus. However, treatment intensification studies are limited in informing our understanding of latency in the CNS. Examination of CSF rebound virus could help fill this gap in knowledge. CNS viral persistence is further complicated by the fact that the cellular composition of the CNS is macrophage-rich with limited exposure to T cells, which reside primarily within the CSF. Thus it is possible that mechanisms of HIV-1 persistence differ between the CSF/meninges and the CNS/parenchyma and these mechanisms are affected by the presence of neuroinflammation, which alters the interaction between these bodily compartments. Finally, we do not understand the extent to which viral replication in the parenchyma is “reported” as virus in the CSF.

Recently the concept of “immune privilege” of the CNS has been confounded with the discovery of meningeal lymphatic vessels in the brain. These vessels appear to drain fluid and immune cells from the CSF and connect with deep cervical lymph nodes, which are part of the peripheral lymphatic system (Louveau 2015). The discovery of meningeal lymphatic vessels complicates our understanding of how the cerebrospinal, blood, and lymph fluids are linked. The classical mechanism of fluidic drainage from the CSF to the blood occurs across meningeal arachnoid villi, which are microscopic projections of the arachnoid mater, the middle of the three CNS membranes that enclose the parenchymal tissues and CSF (meninges). Furthermore, CSF fluid is thought to cross the cribriform plate, a paper-thin, perforated bone in the nasal cavity, to drain to the deep cervical lymph nodes (reviewed in Ransohoff and Engelhardt 2012). Meningeal lymphatic vessels comprise a potential bypass of the known CSF draining mechanisms and could contribute to HIV-1 dissemination from the CNS to the rest of the body.


Many aspects of HIV-1 infection within the CNS are poorly understood, but several key features of CNS infection can be gleaned from studies to date. It is evident that virus enters the CNS early by one or more mechanisms and that permissive or semipermissive cell types are present to support productive infection. Furthermore, the CNS is rich in long-lived macrophages and microglia, which could house virus in a similar manner to resting memory CD4+ T cells, although studies have not shown that infection of macrophages can have a persistent, dormant state. Studies using CSF rebound virus could inform our understanding of CNS viral persistence. Likewise, animal models are a valuable platform for addressing mechanistic questions of CNS infection. Defining the CNS reservoir is an important aspect of HIV-1 cure strategies and current research is aimed at clarifying our understanding of HIV-1 persistence in the CNS.



  1. Archin NM, et al. Eradicating HIV-1 infection: seeking to clear a persistent pathogen. Nat Rev Microbiol. 2014;12(11):750–64.CrossRefPubMedPubMedCentralGoogle Scholar
  2. Arrildt KT, et al. Phenotypic correlates of HIV-1 macrophage tropism. J Virol. 2015;89(22):11294–311.CrossRefPubMedPubMedCentralGoogle Scholar
  3. Blankson JN, Persaud D, Siliciano RF. The challenge of viral reservoirs in HIV-1 infection. Annu Rev Med. 2002;53:557–93.CrossRefPubMedGoogle Scholar
  4. Burdo TH, et al. Increased monocyte turnover from bone marrow correlates with severity of SIV encephalitis and CD163 levels in plasma. PLoS Pathog. 2010;6(e1000842).Google Scholar
  5. Engelhardt B, Ransohoff RM. Capture, crawl, cross: the T cell code to breach the blood–brain barriers. Trends Immunol. 2012;33(12):579–89.CrossRefPubMedGoogle Scholar
  6. Ho YC, et al. Replication-competent noninduced proviruses in the latent reservoir increase barrier to HIV-1 cure. Cell. 2013;155(3):540–51.CrossRefPubMedPubMedCentralGoogle Scholar
  7. Joseph SB, et al. Quantification of entry phenotypes of macrophage-tropic HIV-1 across a wide range of CD4 densities. J Virol. 2014a;88(4):1858–69.Google Scholar
  8. Joseph SB, et al. HIV-1 target cells in the CNS. J Neurovirol. 2014b;21(3):276–89.Google Scholar
  9. Louveau A, et al. Structural and functional features of central nervous system lymphatic vessels. Nature. 2015;523(7560):337–41.CrossRefPubMedPubMedCentralGoogle Scholar
  10. Maldarelli F, et al. HIV latency. Specific HIV integration sites are linked to clonal expansion and persistence of infected cells. Science. 2014;345(6193):179–83.CrossRefPubMedPubMedCentralGoogle Scholar
  11. Matsuda K, et al. Laser capture microdissection assessment of virus compartmentalization in the central nervous systems of macaques infected with neurovirulent simian immunodeficiency virus. J Virol. 2013;87(16):8896–908.CrossRefPubMedPubMedCentralGoogle Scholar
  12. Monteiro de Almeida S, et al. Dynamics of monocyte chemoattractant protein type one (MCP-1) and HIV viral load in human cerebrospinal fluid and plasma. J Neuroimmunol. 2005;169(1–2):144–52.CrossRefPubMedGoogle Scholar
  13. Obermeier B, Daneman R, Ransohoff RM. Development, maintenance and disruption of the blood–brain barrier. Nat Med. 2013;19(12):1584–96.CrossRefPubMedPubMedCentralGoogle Scholar
  14. Price RW, et al. Evolving character of chronic central nervous system HIV infection. Semin Neurol. 2014;34(1):7–13.CrossRefPubMedPubMedCentralGoogle Scholar
  15. Ransohoff RM, Engelhardt B. The anatomical and cellular basis of immune surveillance in the central nervous system. Nat Rev Immunol. 2012;12(9):623–35.CrossRefPubMedGoogle Scholar
  16. Siliciano RF, Greene WC. HIV latency. Cold Spring Harb Perspect Med. 2011;1(1):a007096.CrossRefPubMedPubMedCentralGoogle Scholar
  17. Spector R, et al. A balanced view of choroid plexus structure and function: focus on adult humans. Exp Neurol. 2015;267:78–86.CrossRefPubMedGoogle Scholar
  18. Spudich SS, et al. Cerebrospinal fluid HIV infection and pleocytosis: relation to systemic infection and antiretroviral treatment. BMC Infect Dis. 2005;5:98.CrossRefPubMedPubMedCentralGoogle Scholar
  19. Sturdevant CB, et al. Compartmentalized replication of R5 T cell-tropic HIV-1 in the central nervous system early in the course of infection. PLoS Pathog. 2015;11(3):e1004720.CrossRefPubMedPubMedCentralGoogle Scholar
  20. Wakim LM, Woodward-Davis A, Bevan MJ. Memory T cells persisting within the brain after local infection show functional adaptations to their tissue of residence. Proc Natl Acad Sci U S A. 2010;107(42):17872–9.CrossRefPubMedPubMedCentralGoogle Scholar
  21. Watanabe R, et al. Human skin is protected by four functionally and phenotypically discrete populations of resident and recirculating memory T cells. Sci Transl Med. 2015;7(279):279ra39.CrossRefPubMedPubMedCentralGoogle Scholar
  22. Yilmaz A, et al. Treatment intensification has no effect on the HIV-1 central nervous system infection in patients on suppressive antiretroviral therapy. J Acquir Immune Defic Syndr. 2010;55(5):590–6.CrossRefPubMedGoogle Scholar
  23. Zayyad Z, Spudich S. Neuropathogenesis of HIV: from initial neuroinvasion to HIV-associated neurocognitive disorder (HAND). Curr HIV/AIDS Rep. 2015;12(1):16–24.CrossRefPubMedPubMedCentralGoogle Scholar
  24. Zhuang K, et al. Emergence of CD4 independence envelopes and astrocyte infection in R5 simian-human immunodeficiency virus model of encephalitis. J Virol. 2014;88(15):8407–20.CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Lauren Tompkins
  • Elena Dukhovlinova
  • Ronald Swanstrom
    • 1
    Email author
  1. 1.The University of North CarolinaLineberger Cancer CenterChapel HillUSA