Cerebrospinal Fluid HIV-1 Viral Load During Treatment of Cryptococcal Meningitis

Abstract

To the Editors: HIV-1 can be detected in the cerebrospinal fluid (CSF) throughout the course of HIV infection, although at lower concentrations than in blood.1 Opportunistic meningeal infections, such as Cryptococcus neoformans and Mycobacterium tuberculosis can enhance HIV replication in vitro,2-4 and during such infections, HIV-1 viral load (VL) in CSF is found to be raised,5 as a result of either this direct enhancement of replication, and/or increased HIV-1 replication due to a proinflammatory immune response and the influx of leucocytes.6 Thus, treatment of the opportunistic infection might lead to a more rapid decline of the HIV-1 VL in the compartment of infection than systemically. To examine the changes in CSF and plasma HIV-1 VL during treatment of cryptococcal meningitis, we measured the levels of HIV-1 RNA in CSF and plasma before and after 2 weeks of treatment of cryptococcal meningitis in 63 Thai patients enrolled in a trial of amphotericin B-based therapy.7 Baseline cryptococcal organism load was determined by quantitative CSF cultures, as were CSF cell count and differential and CSF and plasma cytokines [interleukin (IL)-1β, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12p70, IL-13, IL-17, G-CSF, GM-CSF, MCP-1, MIP-1β, interferon γ, and tumor necrosis factor-α]8 to investigate any association of CSF immune parameters with CSF HIV-1 RNA concentrations. CSF was available from 57 patients at baseline and 47 at day 14. Plasma was available from 62 patients at baseline and 49 at day 14. Paired CSF/plasma samples were available of 57 patients at baseline, and 46 at day 14. CSF and plasma samples were centrifuged at 1500 g for 10 minutes at 4°C and supernatants were stored at −80°C. The mean time from lumbar puncture to storage at −80°C was less than 3 hours. After collecting all samples, HIV RNA quantification in the cell-free CSF and plasma was performed with the Roche Amplicor Monitor Test (version 1.5, Roche Diagnostic Systems, Hoffman-La Roche, Basel, Switzerland). We compared median log10 HIV-1 RNA concentrations using Wilcoxon matched pairs signed-rank test and the Mann-Whitney U test. Correlation between variables was tested using Spearman rank correlation. Median log10 HIV-1 VL was significantly lower in CSF compared with plasma at baseline and at day 14 (P < 0.001 for both comparisons). There was a significant decrease in HIV-1 VL in both CSF and plasma during the first 2 weeks of treatment of cryptococcal meningitis (Table 1). However, the CSF/Plasma HIV-1 VL ratio did not change over time. Baseline CSF HIV-1 VL correlated with CSF lymphocyte count (r = 0.5, P = 0.0002). At baseline, there was a positive correlation between CSF IL-10 and CSF HIV-1 VL (r = 0.6, P < 0.0001) and a weak correlation between CSF IL-6 and CSF HIV-1 VL (r = 0.3, P = 0.03). CSF MCP-1 was negatively correlated with CSF HIV-1 VL (r = −0.5, P = 0.0005). No other CSF or plasma cytokines were associated with VL.TABLE 1: Baseline and Day 14 CSF and Plasma HIV-1 RNA ConcentrationThe patients in this study were profoundly immune-depressed with a median CD4 count of 9 × 106/L (interquartile range 6-17 × 106/L) and thus had a high probability of having systemic cryptococcosis,9 which might have caused enhanced HIV-1 replication in both plasma and CSF. Treatment with systemic antifungal therapy could reduce HIV-1 VL locally and systemically through a reduction in cryptococcal antigens, known to enhance HIV-1 replication, or alternatively, or in addition, a reduction in the inflammatory response in both systemic and localized compartments. Our results suggest this latter mechanism may play a role. From prior work we know that the CSF inflammatory response is markedly diminished by 2 weeks into antifungal therapy.8 In addition, we did not find a positive relation between the cryptococcal colony forming units (and cryptococcal antigen) and the HIV viral load in CSF. Instead we found a weak inverse correlation between baseline cryptococcal colony forming unit count and HIV-1 RNA concentrations (r = −0.3, P = 0.02). However, in prior work we have shown that baseline cryptococcal organism load is negatively correlated with CSF proinflammatory cytokine concentrations.8 Thus, it is possible a brisk proinflammatory cytokine response controls infection more effectively and so is associated with a lower baseline cryptococcal organism load, but also increases HIV-1 VL through enhanced lymphocyte influx into the CSF and enhanced HIV-1 replication. We did find a positive correlation between the number of lymphocytes and HIV-1 VL in the CSF, as was found before by Morris et al,5 although we cannot determine the relative importance of HIV-1 replication in activated lymphocytes in the CSF, as opposed transportation of HIV-1 into the CSF by these lymphocytes. We found a weak correlation between IL-6 and HIV-1 VL in the CSF, whereas there was a strong correlation between HIV-1 VL and IL-10 concentrations. The role of IL-10 in HIV-1 replicati