Antiretroviral Drug Penetration into the CNS Compartment
Andrea Calcagno
Unit of Infectious Diseases
Department of Medical Sciences, University of Torino Torino, Italy
Email: andrea.calcagno@unito.it
Definition
The central nervous system is reached and infected a few hours after HIV infection. Viral replication occurs in perivascular macrophages, microglia and, although restricted, in astrocytes: neuronal damage is believed to be a consequence of neurotoxins productions by the aforementioned cells of the immune system. Such cells are reached by antiretroviral drugs either directly (crossing blood-brain and blood-cerebrospinal fluid barriers) or through cerebral extracellular fluid (drained into cerebrospinal fluid); for several drugs cerebrospinal fluid concentrations has been shown to reflect cerebral interstitial fluid concentrations. The penetration of several compounds into the central nervous system has been shown to be highly variable and to depend on drugs’ (molecular weight, lipophilicity, ionization, plasma protein binding, transport mechanisms) and patients’ characteristics (age, blood flow, blood brain barrier permeability). Although the exact amount of drug necessary to inhibit viral replication in the central nervous system is currently unknown, antiretrovirals have been ranked according to their concentration/penetration effectiveness: higher scores have been repeatedly associated with lower cerebrospinal fluid HIV viral loads.
Main text
(1) Introduction
HIV RNA has been recovered from the cerebrospinal fluid (CSF) few days after systemic infection and central nervous system (CNS) involvement parallels this event. [Valcour, 2012] The presence of viral replication in perivascular macrophages, microglia and astrocytes (restricted) is eventually associated with neuronal damage due to persistent immune activation and cytokines production. Inflammatory cytokines and chemokines have been found to be abnormally elevated in HIV-positive patients and they have been linked to the alteration of blood brain barrier (BBB). [Gannon, 2011]. An impaired BBB might facilitate the penetration of HIV-infected monocytes thus increasing the viral biomass in the CNS. Furthermore CNS has been recognized as a site of compartmentalized viral replication: HIV has been recognized to harbour different resistance associated mutations in the CSF as compared to plasma. [Canestri 2010] Approximately 10% of treated patients have detectable HIV RNA in the CSF despite plasma viral control; the clinical and immunological consequences of this event in the absence of resistance-associated mutations are currently unclear. [Eden, 2010]
The clinical endpoint of untreated CNS infection is the appearance of HIV-associated dementia (HAD). With the introduction of highly active antiretroviral treatment (HAART) the incidence of dementia significantly declined; nevertheless cognitive impairment (asymptomatic and moderate according to the impact on everyday life and globally defined as HIV-associated neurocognitive disorders, HAND) remains highly prevalent. [Clifford 2013] Furthermore 19-22.7% of subjects may progress to symptomatic neurocognitive impairment along time. [Heaton, 2015]
To be efficacious drugs must reach adequate concentrations at the site of action: in the case of CNS infection by HIV the targets are macrophages, microglia and astrocytes within the brain parenchyma. After intestinal absorption orally administered antiretrovirals (ARVs) (the vast majority of available drugs, with the exception of intravenous zidovudine and subcutaneous enfuvirtide) are transported by plasma proteins in the bloodstream and distributed to organs and tissues. The CNS receives approximately 14% of cardiac output but two anatomical barriers can be found that prevent the free passage of drugs into the brain: the blood brain barrier and the blood CSF barrier. The first one is characterized by endothelial cells connected by tight junctions and by the presence of astrocytes end feet: several substances are restricted from crossing the BBB. [Varatharajan and Thomas, 2009] Nevertheless in some areas in the brain (hypothalamus, area postrema, subfornical organ) tight junctions are not present and direct diffusion is possible.
The study of antiretrovirals pharmacokinetics in the CNS has several methodological obstacles. For instance data on tissue and intracellular drug concentrations are limited and most of the knowledge derives from cerebrospinal fluid measurements. Cerebrospinal fluid is believed to be produced by filtration from blood plasma (two thirds) and from brain extracellular fluid (one third). Several animal studies have suggested that cerebrospinal fluid is a surrogate reliable marker for most of the studied drugs although a significant variability in tissue levels prediction was observed. For instance data in non-human primates suggest a good correlation between zidovudine CSF and brain parenchyma concentrations, while data for other ARVs are scarce.
(3) Factors affecting antiretroviral penetration into the CNS
Patients’ related factors include age (older age may affect the passage of several drugs into the CNS due to reduced blood efflux, permissive BBB and altered CSF flow) and BBB permeability. BBB impairment is often observed in HIV-positive patients: a permissive barrier may allow the passage
of both drugs and plasma proteins thus increasing the CSF total concentration but reducing the free drug levels thus questioning its clinical impact; however it has been shown to be significant for tenofovir, emtricitabine and raltegravir.
Four chemical characteristics that affect drug passage have been identified: molecular weight (the smaller the higher), lipophilicity (the higher the higher, measured as octanol water distribution coefficient, LogP), ionization (the higher the lower) and plasma protein binding (the lower the higher). Several data confirm the effect of such drug related features: a near direct linear correlation has been shown between plasma protein binding and CSF to plasma drug concentration ratios.
Several transporting proteins have been shown to be expressed at the BBB and at the blood/cerebrospinal fluid barrier (BCB) and to be involved in drug transport into the CNS: p-glycoprotein (P-gp), Organic Anion Transporter 1, 2 and 3 (OAT1,2 and 3), Breast Cancer Resistance Protein (BCRP) and others. Single nucleotide polymorphisms affecting such enzymes may affect drugs penetration into the CNS although data showing such influence are very limited.
(4) Antiretroviral penetration in the CNS
Antiretrovirals’ molecular weight, plasma protein binding, CSF to plasma ratios and concentration/penetration effectiveness score (CPE, discussed below) are shown in Table 1. [Yilmaz 2012, Eisfled, 2013]
NRTIs are small, poorly bound, hydrophilic molecules reaching very variable CSF to plasma ratios. Tenofovir is ionized at physiological pH and this limits its uptake by membrane transporters. NRTIs are transported by Organic Anion Transporters (OATs) that have been showed to be present at the choroid plexus (OAT1 and OAT3); the modulation of their activity (either by other drugs such as probenecid or by genetic polymorphisms in the encoding genes) may be relevant for zidovudine, stavudine, lamivudine and tenofovir
passage. With the exception of didanosine (whose CSF exposure has been found to be undetectable or very low) the other NRTIs have been associated with therapeutic CSF concentrations.Animal data suggested a good CSF passage (through the blood CSF barrier and OATs-independent) but a poor penetration into deep brain tissue; CSFtenofovir levels in humans are low and often undetectable.
NNRTIs are small, lipophilic, highly protein bound (with the exception of nevirapine) compounds. Efavirenz CSF levels are low (very close to the limit of detection of the instruments thus suggesting undetectable concentrations) but frequently above 50% inhibitory concentrations (IC50). While data on rilpivirine and on etravirine are still limited, nevirapine high CSF to plasma ratios has been confirmed: the compound properties as well as the in vivo data suggest that nevirapine is one of the ARVs with the best CSF penetration. PIs are large, lipophilic, highly protein-bound (with the exception of indinavir) compounds with CSF to plasma ratios around 1%; they have been recognized as substrate of p-glycoprotein as well as OAT1A2 and this may limit drug accumulation into the CNS (as well as into other key tissues such as lymph nodes). The comparison among the three nowadays most used protease inhibitors favours darunavir and lopinavir given atazanavir low or undetectable CSF concentrations.
Enfuvirtide is a synthetic 36 amino acid oligopeptide (interacting with viral gp41) with a very large molecular weight: a single study and a case report confirmed CSF very low or undetectable samples in the majority of patients.
Maraviroc is a small, lipophilic, intermediately protein-bound compound that targets the human co-receptor CCR5 and that is effective in preventing R5-tropic HIV viruses entry. It is substrate of both cytochrome P450 3A4 and p-glycoprotein and drug to drug interaction, potentially affecting CSF concentrations, have been reported.The available data have been
obtained with twice-daily different dosages (150 mg with PIs, 300 mg with NRTIs and nevirapine and 600 mg with efavirenz or etravirine): CSF concentrations were detectable and approximately 2-3% of plasma levels.
Integrase inhibitors are the latest ARV drug class and they are somehow heterogeneous: while all are small molecules and highly bound to plasma proteins, lipophilicity varies considerably (raltegravir is hydrophilic while elvitegravir is lipophilic). So far no data has been released on elvitegravir CSF exposure (a study is currently ongoing: ClinicalTrials.gov Identifier: NCT02251236) while the single unpublished study reported low dolutegravir CSF to plasma ratios (0.4%) but CSF concentrations above IC50 in all samples. Given
raltegravir peculiar pharmacokinetics properties such as very wide inter and intra-individual variability and an unclear concentration-dependant efficacy, its effectiveness in the CNS is still unknown.
(5) Target levels
The study of the pharmacodynamic effect of ARVs in the CNS is complicated by the absence of a clear pharmacodynamic target. The optimal marker would be the inhibition of HIV tissue replication in the whole brain parenchyma, although such marker is not currently available. Several other markers of CNS efficacy are currently being studied: CSF biomarkers, magnetic resonance imaging, astrocytes targeted positron emission tomography, electro-encephalographic rhythms. CSF HIV RNA and neurocognitive testing are currently used in clinical practice.
The use of CSF HIV RNA as a marker of antiviral activity is the most commonly used marker since it is dramatically affected by HAART introduction and since the decrease in CSF replication parallels cognitive improvement in patients with HAD. Nevertheless HIV replication may differ in different brain areas not accurately reflecting CSF levels. Furthermore commercial kits for
measuring HIV RNA have not been validated in the CSF and the threshold is unclear: second generation methods can quantify as low as 20 copies/mL; very sensitive experimental techniques (quantifying 0.3-1 copies/mL) have been assessed showing residual CSF HIV RNA in a high proportion of subjects. This low level HIV RNA in the CSF did not change after treatment intensification. However subjects having the lowest viral load in the cerebrospinal fluid had the lowest levels of biomarkers of immune activation (such as neopterine, produced by macrophage-derived cells and believed to be a sensitive marker of CNS immune activation). [Dahl, 2014]
Cognitive function is tested and monitored with neurocognitive tests; nevertheless complete testing is time-consuming and it may influenced by the choice of the control group and by learning effect (patients repeating slightly-modified tests may perform better).
Given the inaccessibility of in vivo brain tissue CSF inhibitory concentrations (IC50, IC90 and IC95) have been used to compare the adequacy of ARVs exposure: this concentrations represent the level at which 50%, 90% or 95% of viral replication is inhibited in vitro using wild-type viruses. However this in vitro protein-free values have significantly variable values and the same drug has been judged to reach optimal or insufficient concentrations in different studies when compared to different thresholds. Using standardized ICs, 95% inhibitory quotients (as CSF exposure divided by IC95) has been described for several antiretrovirals: they represent how many times a drug overcome
the target level in a single patient. Patients with higher inhibitory quotients and detectable NRTIs had a better CSF viral control. [Calcagno, 2015]
(6) The CPE score
The CNS Penetration Effectiveness score (CPE score) has been proposed by a large collaborative study group in the USA (the CHARTER group): in the revised 2010 version ARVs were scored 1 to 4 (where 4 is the most neuro-effective drug, Table 1) according to drug characteristics,
pharmacokinetics and pharmacodynamic properties. In the last available analysis several other demographic (race), clinical (current depression, current adherence to medications) and virological variables (plasma HIV RNA and duration of current HAART) have been added to CPE strata (<5, 5-10, ≥ 10) thus creating a “Cerebrospinal Fluid HIV Risk score”. [Hammond, 2014]
While in the original study it nicely correlated with the prevalence of CSF HIV RNA above 50 copies/mL several other investigator have studied the composite CPE (obtained adding single drug scores to obtain a treatment score). Most of the studies found a lower CSF HIV RNA with higher CPE score while the effect on immune-activation, MRI cerebral metabolites levels and neurocognitive testing were less concordant among studies. Furthermore only one study (out of three) found a correlation with CSF escape and CPE cut offs are still debated.
Some limitations of the CPE score must be highlighted: the limited amount of evidence regarding PD data and regarding drugs standard dosages, the absence of a clear cut off, the validation in patients receiving triple therapies and with fully sensitive viruses.As an example a CPE corrected for plasma resistance associated mutations was a better predictor (compared to standard CPE) of HAND in a cross-sectional study. For these reasons some authors prefer not to use the aggregate CPE but they suggest that treatment optimization in patients with CNS diseases may include drugs with individual elevated CPE score.
The CPE score may be a useful tool for choosing neuro-active antiretroviral drugs although with some limitations. Nevertheless a recent review using rigorous methods found that neuro-HAART (i.e. a combination including ARVs with high CPE scores) was effective in improving neurocognitive function and decreasing CSF viral load (although only two of the included studies were adequately powered): this confirms the possible optimization of CNS treatment and calls for prospective, randomized, adequately powered studies. [Cysique, 2011] The only randomized controlled trial trying to answer this question was prematurely interrupted for slow accrual (326 patients screened and 59 enrolled): CNS-targeted HAART was not associated with either
virological nor neurocognitive improvements although in patients with baseline suppressed viral load a trend for improved cognitive performances over time was observed. [Ellis, 2014]
(7) Efficacy in monocytes, macrophages and astrocytes
In vitro data suggest that the endogenous nucleoside pool in resting macrophages is smaller than the one in activated lymphocytes and therefore that the effective phosphorylated NRTI levels required to inhibit HIV replication is lower. Some authors used in vitro effective concentration in acutely infected macrophages (EC50) to calculate a “monocyte efficacy score”: the obtained composite score
was associated with neurocognitive performances and with the presence of HAND or minor motor cognitive disorders. [Shikuma, 2012] Recent data challenging infected astrocytes with several NRTIs, NNRTIs and raltegravir found that some drugs (zidovudine, lamivudine and stavudine) may have inadequate inhibitory activity in astrocytes, with 90% inhibitory concentrations (EC90)
exceeding those achievable in the CSF.
These preliminary observations warrant further studies on the differential efficacy of ARVs according to target cells. Furthermore the repeated association between HIV reservoir size (measured as PBMC- or monocyte-associated quantitative HIV DNA) and the prevalence of HAND support the implementation of specific drug strategies in selected patients (those with low CD4+ cells nadir, high HIV RNA zenith and high cumulative viremia for instance). [Valcour, 2013]
Conclusion
The infection of central nervous system immune cells by HIV is associated with severe long-term consequences. With the widespread use of highly active antiretroviral therapy HIV-associated dementia is currently a rare AIDS-defining condition; less severe forms are however still common
with multifactorial pathogenesis. Incomplete antiretroviral penetration into the central nervous system might be one of the reasons for compartmentalized infections and residual indirect neuronal damage. Several data has now associated lower central nervous system exposure (either measuring cerebrospinal fluid concentrations, ranking drugs according to the CPE or monocyte activity scores) with detectable cerebrospinal fluid HIV RNA.
Cross References
1. Macrophages in Immune pathogenesis 2. neuro-AIDS, Immunepathogenesis of
3. Co-morbidity: drugs of abuse, including alcohol & behavior
4. Global NeuroAIDS 5. HAND biomarkers
6. HIV Neurocognitive Diagnosis, Natural History and Treatment 7. Overview of HIV CNS infection
8. HIV Reservoirs in the Central Nervous System
References
Calcagno A, Simiele M, Alberione MC, et al. Cerebrospinal fluid inhibitory quotients of antiretroviral drugs in HIV-infected patients are associated with compartmental viral control. Clin Infect Dis. 2015;60(2):311-7.
Canestri A, Lescure FX, Jaureguiberry S, et al. Discordance between cerebral spinal fluid and plasma HIV replication in patients with neurological symptoms who are receiving suppressive antiretroviral therapy. Clin Infect Dis. 2010;50(5):773-8.
Clifford DB, Ances BM. HIV-associated neurocognitive disorder. Lancet Infect Dis. 2013;13(11):976-86.
Cysique LA, Waters EK, Brew BJ. Central nervous system antiretroviral efficacy in HIV infection: a qualitative and quantitative review and implications for future research. BMC Neurol. 2011;11:148.
Dahl V, Peterson J, Fuchs D, et al. Low levels of HIV-1 RNA detected in the cerebrospinal fluid after up to 10 years of suppressive therapy are associated with local immune activation. AIDS. 2014;28(15):2251-8.
Edén A, Fuchs D, Hagberg L, et al. HIV-1 viral escape in cerebrospinal fluid of subjects on suppressive antiretroviral treatment. J Infect Dis. 2010;202(12):1819-25.
Eisfeld C, Reichelt D, Evers S, Husstedt I. CSF penetration by antiretroviral drugs. CNS Drugs. 2013;27(1):31-55.
Ellis RJ, Letendre S, Vaida F, et al. Randomized Trial of Central Nervous System-Targeted Antiretrovirals for HIV-Associated Neurocognitive Disorder. Clin Infect Dis. 2014;58(7):1015-22.
Garvey LJ, Pavese N, Politis M, et al. Increased microglia activation in neurologically asymptomatic HIV-infected patients receiving effective ART. AIDS. 2014;28(1):67-72.
Hammond ER, Crum RM, Treisman GJ, et al. The cerebrospinal fluid HIV risk score for assessing central nervous system activity in persons with HIV. Am J Epidemiol. 2014;180(3):297-307.
Heaton RK, Franklin DR Jr, Deutsch R, et al. Neurocognitive change in the era of HIV combination antiretroviral therapy: the longitudinal CHARTER study. Clin Infect Dis. 2015;60(3):473-80.
Shikuma CM, Nakamoto B, Shiramizu B, et al. Antiretroviral monocyte efficacy score linked to cognitive impairment in HIV. Antivir Ther. 2012;17(7):1233-42.
Valcour V, Chalermchai T, Sailasuta N, et al. Central nervous system viral invasion and inflammation during acute HIV infection. J Infect Dis. 2012;206(2):275-82.
Valcour VG, Ananworanich J, Agsalda M, et al. HIV DNA reservoir increases risk for cognitive disorders in cART-naïve patients. PLoS One. 2013;8(7):e70164.
Varatharajan L, Thomas SA. The transport of anti-HIV drugs across blood-CNS interfaces: summary of current knowledge and recommendations for further research. Antiviral Res. 2009;82(2):A99-109.
Yilmaz A, Price RW, Gisslén M. Antiretroviral drug treatment of CNS HIV-1 infection. J Antimicrob Chemother. 2012;67(2):299-311.
