| |
Maraviroc Concentrations in Cerebrospinal Fluid in HIV-Infected Patients: "MVC achieves concentrations within the EC90 range in CSF. All patients with undetectable plasma viral load although receiving nucleoside-sparing regimens including new drugs showed viral suppression in CSF."
|
| |
| |
JAIDS Journal of Acquired Immune Deficiency Syndromes:
15 December 2010
"In this study, MVC plasma concentrations in all patients were in the expected range; that is, several times higher than the median serum-adjusted EC90. Moreover, MVC CSF concentrations were above this value in most patients......Median MVC concentrations in plasma were 124.75 (7.3-517) ng/mL. In all except one, CSF sample-receiving an erroneous MVC dose while taking concomitantly nevirapine-MVC concentrations [2.58 (<0.5-7.22) ng/mL] were within the EC90 range (0.06-10.70). Median MVC CSF: plasma ratio was 0.022 (0.004-0.17), and when the free MVC plasma concentration was used, 0.094 (2.58-27.44). CSF viral load was <40 copies per milliliter in all 9 patients with undetectable plasma viral load.......Although the upper limit of the known EC90 for wild-type virus is higher than the values obtained, it should be taken into account that almost all the drug in CSF is protein unbound and therefore, active. Thus, the minimal drug concentration needed to inhibit 90% of the virus in CSF would likely be lower. Furthermore, the use of triple ARV therapy with CSF-penetrating drugs would assure sufficient antiviral activity in this reservoir."
"Several reported findings suggest that MVC has potentially good activity within the CNS: a relatively low protein binding of 76%, evidence that HIV strains infecting macrophages or microglia likely use CCR5,17,18 and some access to CSF, demonstrated in a rat model.19 In addition, in a small study including 7 patients, the reported data about MVC concentrations in CSF were similar to ours, with a median of 3.63 (1.83-12.2) ng/mL and a CSF:plasma ratio of 0.03 (0.01-0.10).13.........Our observations provide additional evidence suggesting a role for MVC in the treatment of HIV-1 in the CNS......Knowledge is still very limited in this field, however, and many unanswered questions remain, such as the clinical significance of drug penetration into CSF and the number of penetrating drugs needed to reach viral suppression. In addition, the ACTG 736 study paradoxically found a poorer neurocognitive course in patients receiving an ARV regimen-containing drugs with good CNS penetration, which could indicate neurotoxicity as a potential cause.....Despite these controversial data, the available evidence suggests that CSF penetration of ARV drugs may improve neurocognitive disturbances and prevent long-term development of neurocognitive impairment in patients with a chronic disease, such as HIV-1 and viral dissemination into the CNS in the initial phases of infection.1 Our data suggest that MVC may contribute to inhibit HIV-1 replication in CNS, whereas combined nucleoside-sparing regimens, including new ARV drugs, seem to be locally active."
Tiraboschi, Juan Manuel MD; Niubo, Jordi PhD; Curto, Jordi MSc; Podzamczer, Daniel PhD
From the HIV Unit, Infectious Disease Service, Hospital Universitari de Bellvitge, L'Hospitalet de Llobregat, Barcelona, Spain.
Received for publication February 25, 2010; accepted June 17, 2010.
This study was funded by Pfizer.
Presented as a poster and themed discussion presentation in CROI 2010 (Poster 612).
J.M.T. and D.P. have received research grants and honoraria for conferences and/or advisories from several pharmaceutical companies manufacturers of antiretroviral drugs. No conflict of interest is declared by J.C. and J.N.
Correspondence to: D. Podzamczer, MD, HIV Unit, Infectious Disease Service, Hospital Universitari de Bellvitge, L'Hospitalet de Llobregat, Barcelona, Spain (e-mail: dpodzamczer@bellvitgehospital.cat).
Abstract
Objective: To determine maraviroc (MVC) concentrations in cerebrospinal fluid (CSF) in HIV-infected patients.
Methods: Twelve CCR5+ HIV-1 adult antiretroviral-experienced patients receiving MVC-containing regimens for at least 1 month were enrolled. Both CSF and blood samples were taken around 12 hours after the last MVC dose. liquid chromatography tandem mass spectrometry was used to determine MVC concentrations, and HIV-1 viral load was determined by real-time polymerase chain reaction, (LOD, 40 copies/mL).
Results: Twelve blood and 12 CSF samples were collected. Median CD4 count was 281(120-759) cells per microliter, and median HIV-1 viral load was <40 copies per milliliter. Median time on MVC was 13.5 weeks (4-60). Nucleoside analogues (tenofovir/didanosine) were given in only 1 case. Median MVC concentrations in plasma were 124.75 (7.3-517) ng/mL. In all except one, CSF sample-receiving an erroneous MVC dose while taking concomitantly nevirapine-MVC concentrations [2.58 (<0.5-7.22) ng/mL] were within the EC90 range (0.06-10.70). Median MVC CSF: plasma ratio was 0.022 (0.004-0.17), and when the free MVC plasma concentration was used, 0.094 (2.58-27.44). CSF viral load was <40 copies per milliliter in all 9 patients with undetectable plasma viral load.
Conclusions: MVC achieves concentrations within the EC90 range in CSF. All patients with undetectable plasma viral load although receiving nucleoside-sparing regimens including new drugs showed viral suppression in CSF.
BACKGROUND
The use of combined antiretroviral therapy has converted HIV-1 into a chronic infection with lengthy survival and a dramatic decrease in opportunistic diseases and HIV-1-related complications, such as HIV encephalopathy. However, subclinical or mild neurocognitive impairment seems to be a frequent event in HIV-1-infected patients despite the use of this treatment.1 Patients with symptomatic HIV encephalopathy have increased viral replication, and it has been suggested that even low-level viral replication within the central nervous system (CNS) while on combined antiretroviral therapy can promote local immune activation, and inflammatory response, and subsequent brain damage.1,2 Moreover, the CNS may act as a separate viral reservoir, where HIV-1 strains with a resistance pattern different from that of plasma HIV-1 strains, may be found.3 Good penetration of antiretroviral (ARV) drugs to the cerebrospinal fluid (CSF), a suggested surrogate CNS marker, has been associated with a decrease in viral replication and an improvement in neurocognitive alterations assessed by neuropsychological tests.4
CSF penetration can vary between the different ARV compounds. Most nucleoside analogues (with the exception of tenofovir) and the nonnucleoside reverse transcriptase inhibitor nevirapine (NVP) penetrate well, as has been shown in previous studies: CSF: plasma ratio of 0.3-1.3 for zidovudine,5 0.3-0.42 for abacavir, 0.11 for lamivudine, and 0.28-0.45 for NVP.6 The ratio with efavirenz is 0.01,7 whereas penetration is variable in the case of protease inhibitors, achieving good concentrations with indinavir5, fair concentrations with darunavir8 and lopinavir9, and lower concentrations with atazanavir (0.0021),10 although it has been suggested that these levels would suffice to inhibit viral replication in some cases.1,6,11,12
New ARV drugs approved over the last years have led to the design of effective and well-tolerated regimens for naive and experienced patients, but little data are available on their CSF penetration.8,13,14 This information is important because a growing number of patients are being treated with nucleoside-sparing ARV regimens, particularly after previous failures.
Maraviroc (MVC) is the first CCR5 antagonist approved for HIV treatment. We present a series of HIV-infected ARV-experienced patients receiving an MVC-containing nucleoside-sparing regimen in whom MVC concentrations and viral loads were determined in plasma and CSF.
PATIENTS AND METHODS
Twelve asymptomatic adult HIV-infected ARV pretreated patients were enrolled in our HIV outpatient unit. All of them had been taking MVC for at least 4 weeks as part of an ARV regimen. MVC dosage was set according to the recommendations stated in the European Medicines Agency scientific report.15 In each patient, blood samples were obtained by peripheral venous puncture and CSF samples by lumbar puncture. All samples were centrifuged and frozen at -70°C until analysis. Both blood and CSF samples were taken approximately 12 hours after the previous MVC dose, to obtain the lowest MVC concentrations.
The study was approved by the hospital ethics committee and the Spanish Drug Agency, and patients gave written informed consent to participate.
Total MVC concentrations in plasma and CSF samples were analyzed by liquid chromatography tandem mass spectrometry, using positive turbo ion spray mode16 (Tandem Labs-New Jersey, West Trenton, NJ). The ion transition monitored was 514.1/389.1. The reverse-phase chromatography calibration range was 0.5-500 ng/mL in Li heparin plasma samples. The extraction procedure is based on protein precipitation with acetonitrile, using 50 μL of plasma. The internal standard was the stable isotope label for MVC. HIV-1 viral load was quantified with a real-time polymerase chain reaction technique (Abbot Molecular Inc, Des Plaines, IL), performed according to the manufacturer's recommendations. The limit of detection of the method was 40 copies per milliliter.
RESULTS
All patients were men, and the mean age was 48.1 years. Median HIV-1 viral load was <40 (<40 to 1777) copies per milliliter, and median CD4 count was 281 (120-759) cells per microliter. Patients were receiving MVC for a median period of 13.5 (4-60) weeks. The pharmacokinetic and virologic data, and the ARV regimens received, are summarized in Table 1. Raltegravir (RGV) was given in 92% of cases, darunavir in 67%, and etravirine (ETV) in 42%. Only 1 patient received nucleoside/nucleotide analogues tenofovir/didanosine, and one other received NVP. According to the recommendations,15 this last patient should have received 300 mg twice a day of MVC in combination with RGV and NVP, but he erroneously took 150 mg twice a day.
In all except 1 CSF sample (the NVP patient), MVC concentrations exceeded the median serum-adjusted EC90 of 0.57 ng/mL (0.06-10.7).15 The median MVC CSF: plasma ratio was 0.022 (0.004-0.17) (Fig. 1) and the MVC CSF: estimated free plasma concentration (22% of the total MVC concentration) ratio was 0.094. No correlations were found between MVC plasma and CSF concentrations (correlation coefficient 0.084, P = 0.795).
HIV-1 viral load was undetectable in both plasma and CSF in 9 patients and detectable in 3 patients. These 3 patients had been receiving 3 or 4 drug MVC-containing regimens for only 4-5 weeks, with a very good initial virological response (viral load reduction of 1.9-2.8 log), but with persistently detectable levels. Later follow-up confirmed the virological response, and undetectable plasma viral load was achieved in all 3 patients.
DISCUSSION
In this study, MVC plasma concentrations in all patients were in the expected range; that is, several times higher than the median serum-adjusted EC90. Moreover, MVC CSF concentrations were above this value in most patients. Although the upper limit of the known EC90 for wild-type virus is higher than the values obtained, it should be taken into account that almost all the drug in CSF is protein unbound and therefore, active. Thus, the minimal drug concentration needed to inhibit 90% of the virus in CSF would likely be lower. Furthermore, the use of triple ARV therapy with CSF-penetrating drugs would assure sufficient antiviral activity in this reservoir.
Several reported findings suggest that MVC has potentially good activity within the CNS: a relatively low protein binding of 76%, evidence that HIV strains infecting macrophages or microglia likely use CCR5,17,18 and some access to CSF, demonstrated in a rat model.19 In addition, in a small study including 7 patients, the reported data about MVC concentrations in CSF were similar to ours, with a median of 3.63 (1.83-12.2) ng/mL and a CSF:plasma ratio of 0.03 (0.01-0.10).13
The fact that MVC CSF concentrations were several times lower than the protein-unbound percentage of plasma concentrations suggests that other mechanisms such as P-glycoprotein (PGP) may counteract in part the penetration of the drug by simple passive diffusion.20
Our observations provide additional evidence suggesting a role for MVC in the treatment of HIV-1 in the CNS. Moreover, the fact that almost all patients received a nucleoside-sparing regimen reinforces the probability of reaching high enough antiviral activity in this compartment by combining drugs with at least some CSF penetration. In 1 study in 16 patients,14 RGV exceeded the EC95 levels reported to inhibit wild-type strains in 50% of the 25 CSF specimens studied, whereas in another study,8 darunavir concentrations were detectable in all 7 CSF samples assessed, and most of them exceeded or were in the same range as the levels needed to inhibit replication of the wild-type virus. To our knowledge, no data regarding ETV concentrations in CSF are available.
Concomitant use of CYP3A inhibitors and/or inducers with MVC may alter the concentrations of this drug. NVP is a potent CYP inducer, but there are no existing recommendations for MVC dose adjustment when it is used with NVP. Unfortunately, the potential consequences of an interaction between these drugs on MVC CSF concentrations could not be evaluated in the patient taking NVP because he did not receive the correct dose. On the other hand, MVC concentrations in patients concomitantly taking PIs, which are potent CYP3A4 inhibitors (4 of them received the CYP3A-inducer ETV), were within the expected ranges (Table 1).
MVC is a substrate of PGP and MVC concentrations seem to be largely influenced by PGP. In our patients, median MVC plasma concentrations were 5 times higher and CSF concentrations 1.5 times higher in patients receiving a PGP inhibitor (data not shown). Furthermore, MVC CSF concentrations were more than 4-fold lower than the estimated plasma protein-free concentrations, suggesting that PGP could be involved in efflux of the drug from CSF.
Knowledge is still very limited in this field, however, and many unanswered questions remain, such as the clinical significance of drug penetration into CSF and the number of penetrating drugs needed to reach viral suppression. In addition, the ACTG 736 study paradoxically found a poorer neurocognitive course in patients receiving an ARV regimen-containing drugs with good CNS penetration, which could indicate neurotoxicity as a potential cause. It has also been suggested that in patients with CSF viral load <40 copies per milliliter, low-level viral replication of 2 copies per milliliter may be associated with continuous immune activation and subsequent brain damage.1-4,8,11-14,17-19,21
Despite these controversial data, the available evidence suggests that CSF penetration of ARV drugs may improve neurocognitive disturbances and prevent long-term development of neurocognitive impairment in patients with a chronic disease, such as HIV-1 and viral dissemination into the CNS in the initial phases of infection.1 Our data suggest that MVC may contribute to inhibit HIV-1 replication in CNS, whereas combined nucleoside-sparing regimens, including new ARV drugs, seem to be locally active.
| |
| |
| |
|
|
|
