A number of the features of systemic HIV infection are important for the development of CNS HIV infection and disease. Ultimately, these stem from the fact that CD4+ T lymphocytes and, to a lesser extent, macrophages and related myeloid cells, are the main cellular targets of HIV infection (45–50). This targeted infection leads to progressive immunosuppression and also to a state of enhanced immunoactivation, with both contributing to CNS disease consequences (51–54). HIV infection is chronic and persistent, but importantly mitigated by ART. It remains, however, a major challenge to therapeutic cure efforts (55, 56), and stopping ART almost inevitably leads to a return of viremia accompanied by CSF viral rebound (57, 58).

Complications of HIV vary with the stage of systemic disease progression, most easily assessed by the blood CD4+ T lymphocyte count (38, 59, 60). AIDS is defined by the development of major OIs (and, in parallel, including HIV-associated dementia, HAD) or by a CD4+ count falling below 200 per μl (61).

CNS OIs develop when there is loss of systemic immune surveillance that allows certain organisms to escape a latent or quiescent presence in the body (e.g., JC virus or Toxoplassma gondii) or to evade defenses that would otherwise prevent systemic dissemination (e.g., Cryptococcus neoformans); this is followed by subsequent failure of these same defenses to eliminate these pathogens within the CNS. The spectrum of common CNS OIs is relatively circumscribed and involves organisms of relatively low pathogenicity that are otherwise readily contained or prevented by T-cell/macrophage defenses in the normal host. In Africa where M. tuberculosis is common in the community, HIV infection also enhances susceptibility even if this organism isn't readily classified as strictly “opportunistic” and is more common even at CD4 counts above those defining AIDS (15). However, the common CNS OIs generally occur at <200 CD4+ cells/μl (43). We emphasize this well-known susceptibility here because this also defines the susceptibility to HIV encephalitis (HIVE) and HAD which usually develops below this T-cell threshold (62), indicating that a similar level of immunosuppression is a prerequisite. In a sense HIVE might also be viewed as a CNS OI in which the same virus “creates the opportunity” through chronic systemic infection of CD4+ T cells before it can then “opportunistically” cause encephalitis. However, as discussed below, this does not apply to overall susceptibility to CNS HIV infection per se, but to “invasive” neuropathic encephalitic infection. In fact, low-grade HIV-1 meningeal infection is a common feature of systemic HIV infection that develops early in its course (37). The CNS is exposed to HIV very early in systemic infection, though it is often silent or accompanied by headache, fatigue or other unspecific symptoms. More rarely, acute encephalitis may develop during primary infection, likely involving an immunological pathogenesis (63). Over the course of chronic infection, milder neurocognitive impairment may develop and relate to low-grade forms of the viral and immunological processes that underlie HAD/HIVE, though these connections remain to be more precisely defined.

In addition to providing the background foundation and necessary level of immunosuppression for OIs and HIVE, systemic infection more directly underlies HIV CNS disease by providing both the invading virus and principal inflammatory cells that react to infection and contribute to immunopathology.

Most probably, HIV seeding of the CNS occurs via trafficking infected CD4+ T cells rather than by more direct virion penetration of the blood-brain barrier (64, 65). Infected cells entering the CNS can clonally expand and release (clonal) virus; this can then lead to further infection of susceptible cells, amplifying infection and establishing local replication (66). During later stage infection monocytes may also enter the CNS (65–70). This later CNS infection may be more compartmentalized with more notable evolution of virus populations independent of those examined in blood. Uninfected CD4+ T cells and monocytes may also enter the CNS contributing to amplified infection. This can also lead to local CNS HIV persistence after treatment, though, this has been less clearly defined, including the types of cells and anatomic locations, state of viral expression and mechanisms of replication control.

A central feature of CNS HIV infection is its changing character with systemic disease progression. This includes shifts in the relation of CSF and blood viral populations (71–73), changes in the accompanying inflammatory profiles (36) and eventual shift in the main anatomic site of productive infection from the leptomeninges to the brain in some individuals. In the earlier phase of infection when blood CD4+ T cell levels are above 200 per μl, the leptomeninges are the most conspicuous location of chronic CNS HIV-1 infection so that a clinically silent aseptic meningitis is frequent. This infection is largely “equilibrated” with CSF HIV RNA concentrations maintained at levels near 10 percent of those in blood (37, 74, 75), and CSF and blood populations are genetically similar (76), presumably because of continuous and fresh virus traffic from blood to CSF. When CD4 cells fall below 50/μl, the ratio of CSF to blood virus deceases to near 1% blood HIV RNA levels as CSF pleocytosis also diminishes, consistent with a relation between CSF WBCs and viral load (77–80). The extent of penetration of infection into the brain parenchyma at this stage is uncertain, but if present it is largely clinically silent. Whether this early type of infection and inflammation is responsible for milder cognitive impairment is still not definitively established, though often presumed.

These relationships change in those who develop HIVE that presents clinically as subacute HAD (36). This condition usually develops after blood CD4+ cells fall below 200/μl and represents an extension of infection from meninges into the brain parenchyma. White matter abnormalities are usually prominent on MRI but gray matter also is frequently affected, particularly the basal ganglia (81–83). While inflammation in those without HIVE largely involves lymphocyte-related cytokines, as CD4+ T cell counts fall, macrophage-related inflammation increases. In those with overt HIVE there is augmentation of both lymphocytic and macrophage biomarkers (36). CNS viral populations in these individuals are more compartmentalized in relation to those in blood, and exhibit macrophage tropism (76, 84). While astrocytes can be infected by HIV, this is usually considered to be non-productive with limited gene expression; hence, their role in persistence and neuropathogenesis is still uncertain (85, 86). Importantly, neurons are not infected, and thus damage to neurons is largely or exclusively by “indirect” mechanisms, meaning that they are injured from without by signals and toxins released by neighboring cells rather than from direct effects of viral genes and their products expressed within these cells (87). Likely the external toxic signals are elaborated mainly from inflammatory cells, perhaps predominantly from macrophages and other myeloid cells. Late in infection HIVE also commonly disrupts the blood-brain barrier, further contributing to neuronal injury and dysfunction (36, 88, 89).

ART has had a profound effect on preserving CNS integrity, both in preventing HAD/HIVE development and in mitigating this CNS disease after it manifests (90, 91). This effect may be in part through preservation or restoration of immunity but mainly by more directly suppressing both systemic viremia and HIV replication within the CNS. As a result, HAD incidence is now markedly reduced and confined largely to those not receiving ART.

For individuals who present with HAD, having fallen through defects in the treatment network, ART can arrest and often reverse the severity of its impact, depending on the time frame of HAD development and treatment initiation. Diagnosis should be made quickly, and treatment begun rapidly. This is a setting in which both the antiviral potency and CNS penetration of the components ART regimens are likely important (92, 93). In some of treated individuals the degree of short- and long-term recovery can be remarkable.

This term refers to situations in which the impact of ART on CNS HIV infection is relatively reduced compared to that on systemic infection, leading to CSF HIV RNA levels exceeding those of plasma (94–98). Three distinct types of CSF escape have been defined: asymptomatic, neurosymptomatic and secondary. The most important of these is neurosymptomatic CSF escape in which ART-treated individuals present with new or progressive neurological deficits (96–100). Most often, in addition to symptoms and signs of CNS injury and dysfunction, there is CSF pleocytosis, elevated CSF neurofilament light chain protein (NfL) concentration, and neuroimaging abnormalities consistent with active CNS HIV infection. Neurosymptomatic escape overlaps with pathologically-defined CD8 encephalitis (101–103). In most cases a background of reduced treatment adherence and drug resistance, at times in combination with insufficient CNS penetration of component antiviral drugs, can be identified (96). This provides further support for the need for targeted treatment of CNS, at least in some settings. Inflammation and immunopathology may be an important mechanistic component in this setting in which CD4+ T cell counts are higher than in HAD/HIVE because of the disproportionate systemic efficacy of ART that fosters CD4+ cell recovery and suppresses systemic viremia. CNS HIV isolates often exhibit drug resistance, though not always. The main avenue of treatment is changing the ART regimen to a potent antiviral drug combination that includes component drugs to which the CSF/CNS virus is susceptible and also achieve therapeutic brain concentrations.

The other two forms of CSF escape are of less clinical importance. Asymptomatic escape is an incidental finding mainly in CSF cohort studies. It is characterized by detectable CSF HIV RNA in the presence of plasma viral suppression; CSF HIV RNA levels are usually low with little or no pleocytosis. By definition these individuals lack new neurological symptoms or signs (104, 105). Secondary escape entails a disproportional increase in CSF HIV RNA in association with another CNS inflammatory process (usually another CNS infection) that provokes local HIV replication through recruitment of activated lymphocytes. Treatment of the provoking infection leads to reduction of the CSF HIV RNA elevation (79).

Despite the effectiveness of ART in suppressing systemic and CSF HIV infection, it does not cure HIV. When ART is stopped, viremia and CNS replication re-emerge (106). Because of this intractable persistence of HIV, efforts are now underway to effect a systemic cure using a variety of strategies (55). There is precedent with bone marrow transplant using an HIV-resistant donor. In one well-studied case, not only was there no evidence of viral persistence systemically but also no trace of virus in CSF (107). More broadly it remains an open issue as to whether the CNS serves as an independent viral reservoir that might require CNS-targeted cure strategies.