JAIDS Journal of Acquired Immune Deficiency Syndromes:
HIV-Producing T Cells in Cerebrospinal Fluid
Neuenburg, Jutta K MD*†; Sinclair, Elizabeth PhD‡; Nilsson, Annelie BA†; Kreis, Christophe PhD*; Bacchetti, Peter PhD§; Price, Richard W MD†; Grant, Robert M MD, MPH*
From the *Gladstone Institute of Virology and Immunology, San Francisco, CA; †Department of Neurology, University of California, San Francisco, CA; ‡Department of Medicine, University of California, San Francisco, CA; and §Department of Epidemiology and Biostatistics, University of California, San Francisco, CA.
Received for publication April 7, 2004;
accepted June 10, 2004.
Supported by National Institutes of Health grants RO1 NS37660, RO1 MH62701, and RR-00083 and the Gladstone Institute for Virology and Immunology.
Reprints: Robert M. Grant, Gladstone Institute of Virology and Immunology, PO Box 419100, San Francisco, CA (e-mail: firstname.lastname@example.org).
In HIV-1–infected subjects, the magnitude of HIV-1 viral load in cerebrospinal fluid (CSF) correlates with the CSF white cell count. To determine whether HIV-1–producing T cells appear in CSF and whether their percentage and number correlate with viral load in CSF, we developed a flow cytometric assay that detects HIV-1–producing T cells by identifying intracellular p24 HIV-1 antigen. We found that most CSF T cells were not HIV-1 producing, even when cell-free viral load in CSF was high. Most activated T cells in CSF were also not HIV-1 producing, but the activated CD38+ CD4 T-cell fraction in CSF was independently associated with the fraction of HIV-1–producing T cells in CSF. We conclude that HIV-1–producing T cells appear in CSF and that their percentage and number correlate with cell-free viral load in CSF, even though the CSF total white cell count remains the best predictor for CSF viral load. In HIV-1 infection, CSF white cell counts seem to contain a large number of uninfected cells. White cell counts and viral load in CSF may result from systemic inflammation and immune activation.
HIV-1 infection causes significant injury to the brain, which is reflected in HIV-1–associated neurocognitive deficits and HIV-1 encephalopathy.1–4 Cerebrospinal fluid (CSF), which surrounds brain tissue, may reflect cellular events in brain tissue. The magnitude of HIV-1 viral load in cell-free CSF correlates with the number of CSF white cells,5–8 which we found to be mostly T cells. Among T cells, CD4+ cells can be productively infected with HIV-1. It is unknown whether HIV-1–producing T cells appear in CSF and, if so, whether the percentage and number of HIV-1–producing T cells in the CSF correlate with cell-free viral load. Alternatively, viral load and increased white cell counts in the CSF may result from systemic inflammation and immune activation.
Although antiretroviral therapy reduces viral load and immune activation in CSF and blood and ameliorates the clinical manifestations of HIV-1 infection,9–14 there is evidence of persistent HIV-1–specific brain pathology in the era of antiretroviral therapy.15–18 In living subjects, HIV-1 brain disease can lead to a loss of brain white matter density detectable in magnetic resonance imaging and spectroscopy.19–22 A loss of brain white matter density can also be seen in other systemic diseases in which immune activation is present, for example, multiple sclerosis,23 lupus erythematosus, and Behçet disease.
Even though we are not examining brain tissue in this study, a study of CSF T cells can give us clues regarding the inflammatory process of the nervous system during HIV-1 infection. The viral load in blood and the magnitude of T-cell activation predict clinical and immunologic injury during HIV-1 infection.24 Less information is available about the viral load and magnitude of T-cell activation as well as their consequences in the CSF and brain tissue.
Elsewhere, we describe the use of multiparameter flow cytometry to characterize T cells in CSF using cross-sectional samples of HIV-infected and -uninfected subjects.25 (J. K. Neuenburg et al, submitted for publication). We now describe a subgroup of these study subjects, some of whom have been followed over time. Their specimens were assessed for intracellular expression of p24 HIV-1 capsid antigen. The presence of intracellular HIV-1 p24 capsid antigen inside T cells characterizes an HIV-1–producing cell. To determine whether HIV-1–producing T cells appear in CSF and, if so, whether their percentage and number in the CSF correlate with cell-free viral load, we developed a novel flow cytometric assay that detects intracellular HIV-1 p24 capsid antigen and allows simultaneous detection of surface markers. We explored the relationships between the frequency and number of HIV-1–producing T cells in CSF and several clinically relevant parameters of HIV-1 disease, including blood CD4 T-cell count, activation levels of CD4+ T cells in CSF and blood, and cell-free CSF and blood viral load.
Subjects and Study Population
Study subjects were seen at the San Francisco General Hospital (SFGH) General Clinical Research Center (GCRC) and gave written informed consent. The study protocol was approved by the University of California, San Francisco Committee on Human Research. Subjects were recruited from clinical studies at the SFGH Neurology Clinic through hospital flyers and newspaper ads. The inclusion criteria were HIV-1 infection documented by detection of HIV-1 RNA in blood for the HIV-1 group and HIV-1 seronegativity in blood for the control group. Lumbar punctures and blood sampling were performed as described elsewhere.26 In this study, 10 to 12 mL of CSF was used for the p24 assay and 8 to 12 mL of CSF was used for surface staining of activation markers.
Viral Load Testing
Plasma and CSF viral loads were determined by the ultrasensitive and standard Roche Amplicor reverse transcriptase–polymerase chain reaction (RT-PCR; Roche, Branchberg, NJ) using a lower limit of detection of 20 and 400 copies/mL, respectively. If viral load was above the dynamic range of the assay, the specimen was diluted and retested.
1. Cell Preparation
A new protocol was developed to compare peripheral blood mononuclear cells (PBMCs) and CSF cells. PBMCs were separated from blood with Histopaque (Sigma, Irvine, UK), washed twice with phosphate-buffered saline (PBS), and resuspended in PBS supplemented with 2% fetal bovine serum. Because anti-p24 antibody stains monocytes and granulocytes nonspecifically in peripheral blood and because monocytes and granulocytes both carry the marker CD11b, PBMCs were incubated with CD11b magnetic beads (Miltenyi Biotech, Auburn, CA) and separated into bead-positive and bead-negative fractions with a magnetic bead sorter (Miltenyi Biotech). This treatment made blood more comparable with CSF, which normally has few of these cell types. CSF specimens were centrifuged at 400 g for 15 minutes. The supernatant was removed, and the cells were resuspended in a residual volume (∼100 μL) of CSF. Because CSF cells are sensitive to permeabilization, ideal concentrations and treatment times with different permeabilizing agents were determined in preliminary experiments.
2. Surface and Cytoplasmic Staining Procedures
The CSF sample and 100 μL of resuspended bead-negative PBMCs were incubated with monoclonal antibodies and isotype controls at room temperature in the dark for 20 minutes and then washed with PBS. The antibodies used were anti-CD4 phycoerythrin (PE), anti-CD14 peridinin chlorophyll protein (PerCP), and anti-CD3 allophycocyanin (APC), all from BD Biosciences (San Jose, CA). After staining, cells were permeabilized with 0.1% Tween 20 (Sigma) for 15 minutes, washed, and treated for 20 minutes with mouse IgG (1:10) to block Fc-receptors and decrease nonspecific staining. After intracellular staining with fluorescein isothiocyanate (FITC)–labeled anti-p24 antibody (1:100 ratio; Beckman Coulter, Miami, FL) at room temperature in the dark for 15 minutes, cells were washed and resuspended in PBS and analyzed with a FACSCalibur flow cytometer and Cell Quest software. At least 300,000 events were accumulated for each blood sample. Isotype controls were included with each sample. The event counts for each CSF sample varied, but at least 10,000 events were accumulated for each CSF sample, which was analyzed entirely to maximize event count.
3. Analysis of Flow Cytometric Data
Forward and side scatter plots were examined, and a live lymphocyte gate was drawn. CD14+ cells were gated out to remove remaining monocytes. We then examined the CD3+p24+ population, which was considered to represent HIV-1–producing T cells. The p24+ population in CSF usually stained with higher intensity than in blood. We therefore concentrated our studies in this report on HIV-1–producing T cells in the CSF. CD4 expression of p24+ T cells was assayed but was not used to define HIV-1–producing T cells, because HIV-1–infected T cells have been shown to downregulate CD4 in vivo.27 The HIV-1–producing T-cell population (p24+CD3+CD14−) was quantified as a percent of the total T-cell population (CD3+CD14−). HIV-1–seronegative samples and isotype controls were used to determine background staining and showed similar results.
Surface Staining for Activation Marker CD38
Blood (100 μL) containing monoclonal antibodies was incubated at room temperature in the dark for 20 minutes. Erythrocytes were lysed with ImmunoPrep reagents and a Multi-Q-Prep Workstation (Beckman Coulter) according to the manufacturer’s instructions. A mixture of anti-CD38 PE (BD Biosciences), anti-CD4 PE–Texas red (ECD; Beckman Coulter) was added to cells. After staining, cells were washed, resuspended in 400 μL of 1% paraformaldehyde, and analyzed on an Epics XL flow cytometer with System II Software (Beckman Coulter). Ten thousand events were accumulated for each sample. Forward and side scatter plots were examined, a live lymphocyte gate was drawn, and CD4+ subpopulations were defined. CD38+ subsets were defined by quadrant gates and quantified as percent of CD4+ cells. CSF specimens were centrifuged at 400 g for 5 minutes. The supernatant was decanted, and cells were resuspended in the residual volume of CSF (300–600 μL) and stained as blood, except that the lysis step was omitted. The events counted for each CSF sample varied. To maximize event count, we used the entire CSF sample.
Quantitative Polymerase Chain Reaction for Viral DNA
Cells were lysed with DNAzol BD reagent (BD Biosciences) in the presence of 100 μg of proteinase K (Ambion, Austin, TX). After incubation at 60°C for 10 minutes, the lysates were shredded (QIAshredder; Qiagen, Valencia, CA) and 5 μg of carrier was added (linear polyacrylamide [LPA]; Ambion). The manufacturer’s conditions for mini-isolation of DNA were followed (DNAzol; BD Biosciences). The yields of viral DNA were at least 90% when mixed with 600 pg of cellular DNA (equivalent to 100 cells). Cellular DNA in the extracted specimens was measured by quantitation of rRNA gene copies by real-time PCR according to the manufacturer’s instructions (ribosomal RNA control reagents kit; Applied Biosystems, ABI, Foster City, CA). A conversion factor of 6 pg of DNA per PBMC was used.28 To quantify HIV DNA in cellular DNA from sorted cells, primers and a probe were selected to bind to conserved segments of HIV-1 gag listed with nucleotide positions: probe PrGAG-1448-5′VIC-CAGGGCCTATTGCAC-3′ minor grove binder (ABI), forward primer FGAG-1401-5′-CATCAATGAGGAAGCTG CAGA-3′, and reverse primer RGAG-1476-5′-CTATGTCACTTCCCCTTGGTTCT-3′. Standards were constructed for absolute quantification of viral DNA using the HIV-1 plasmid control DNA diluted in 10 ng of HIV-1–negative human placental DNA (GeneAmpliomer; ABI). Up to 200 ng of HIV-1–negative human placental DNA mixed with HIV-1 control reagents did not interfere with the assay. All specimens were analyzed in duplicate and at 2 dilutions. HIV-1 gag amplification reactions were performed in 50-μL volumes containing 5 μL of extracted DNA, 1× Taqman PCR buffer (ABI), 400 μmol of each nucleotide, ng of LPA, 1.25 units of Amplitaq Gold polymerase (ABI), 150 nmol of each gag primer, and 50 nmol of gag probe. The cycling profile was 95°C for 10 minutes, followed by 40 cycles at 95°C for 30 seconds and then at 62°C for 30 seconds. Fluorescence readings taken at the annealing step of the reaction were analyzed with ABI Prism software. Results are expressed as copies of HIV-1 DNA per number of cells.
Regression models included random subject effects, which account for possible correlations among multiple observations from the same patient. All models of outcome variables measured in CSF were adjusted for CSF white cell count and antiretroviral therapy. Models of outcome variables measured in blood were adjusted for antiretroviral therapy. Residuals were examined for normality. For models with nonnormal residuals, 95% confidence intervals (CIs) and P values were obtained by the bias-corrected accelerated bootstrap method29; P values were defined as 1 minus the highest confidence level for which the interval still excluded 0, reported as P < 0.001 if all 2000 bootstrap estimates had the same sign.
We assessed intracellular p24 HIV-1 capsid antigen in T cells in 37 paired CSF/blood specimens from 13 HIV-1–seropositive subjects and in 3 paired CSF/blood specimens from HIV-1–seronegative controls (Table 1). All paired samples were collected at the same time point. T cells were defined by CD3 (T-cell surface marker) and the absence of CD14 (monocyte surface marker), resulting in the phenotype CD3+/CD14−. Thirty-three of the 37 paired samples that were assessed for p24 expression were also evaluated for activation marker CD38 on CD4+ and CD8+ T cells the same day. The median viral load was higher in blood than in CSF (20,887 and 3193 copies/mL, respectively). The median CD4+ T-cell count in HIV-1–seropositive subjects was 302 cells/μL. CSF white count correlated strongly with CSF viral load (Fig. 1A).
The patients on therapy were largely chosen because they had suffered virologic drug failure during at least 1 time point. Viremic subjects were selected because they would have higher levels of CSF white cell counts, which facilitates the flow cytometric characterization of cells. Eight of 13 subjects who supplied 37 paired specimens (CSF and blood) were taking antiretroviral therapy during at least 1 time point. Only 1 subject was treatment naive. Of the 37 pairs of specimens, 19 (51%) were collected during combination antiretroviral therapy, including at least 3 drugs involving at least 2 classes (eg, nucleoside analogue reverse transcriptase inhibitors, nonnucleoside reverse transcriptase inhibitors, protease inhibitors). Because of this preselection, therapy did not completely suppress viral load at most time points: 17 (84.2%) of the 19 blood specimens and 16 (89.5%) of the 19 CSF specimens taken during therapy had viral loads >50 copies/mL. The high viral loads in these subjects likely reflect drug resistance as a result of exposure to prior therapy or nonadherence. Accordingly, antiretroviral therapy was only weakly associated with lower viral load in this group. None of the study subjects suffered from opportunistic central nervous system (CNS) diseases at the time their CSF and blood samples were taken.
Intracellular HIV p24 Capsid Protein Can Be Detected in T Cells Containing Viral DNA
We developed an intracellular flow cytometric assay to detect HIV-1–producing T cells by detecting intracellular p24 HIV-1 capsid antigen. We removed monocytes and remaining granulocytes in blood after Ficoll testing using CD11b magnetic beads. We identified T cells in CSF and blood with CD3 (T-cell marker) and the absence of CD14 (monocyte marker). The p24+CD3+CD14− T-cell population was considered to represent HIV-1–producing T cells. Analysis of CSF and blood from uninfected controls showed that the median false-positive rate of p24+CD3+CD14− cells was 0.20% for blood and 0.25% for CSF. p24 staining above background was detected in 22 (59.5%) of 37 blood specimens and in 35 (94.6%) of 37 CSF specimens. We therefore concentrated on the p24 percentage data for CSF in this report, also because the p24 staining intensity for CSF cells was higher.
To confirm that p24+ cells were infected with HIV-1, we sorted p24+ and p24− cell populations from 1 patient with a FACSVantage cell sorter and measured total cell-associated viral DNA in each population using a quantitative kinetic PCR-based assay. CD4 expression on p24+ T cells varied, possibly because of CD4 downregulation mediated by HIV-1 infection. The HIV-1–infected sample used for sorting had a low p24 percentage of 0.24% (0.04% above background) in blood. CD3+CD14− T cells were sorted into 3 populations: (1) p24+CD4+, (2) p24+CD4−, and (3) p24− cells. By PCR analysis, HIV gag DNA was detected in p24+CD3+CD14− cells in the CD4+ and CD4− fractions, ranging from 0.5 to 1.4 copies/cell. In the p24−CD3+CD14− fraction, HIV gag DNA was undetectable.
Percentage of HIV-1–Producing T Cells Is Higher in Cerebrospinal Fluid Than in Blood
HIV-1–producing T cells were found in a median of 0.28% of T cells in blood and 0.66% of T cells in CSF. The percentage of T cells that were HIV-1 producing was higher in CSF than in blood (median difference = 0.27%; 95% CI: 0.20%–0.52%; N = 37; P < 0.0001), as indicated by the majority of CSF data points seen above the identity line in Figure 1B. There was a trend toward correlation of HIV-1–producing T cells in both compartments (P = 0.12), suggesting trafficking of cells between compartments or systemic regulation of cells for both compartments. This trend linking CSF and blood HIV-1–producing T-cell percentages was unchanged after controlling for CSF white cell count and therapy.
Cerebrospinal Fluid Viral Load Correlates Strongly With the Concentration But Not the Percentage of HIV-1–Producing T Cells in Cerebrospinal Fluid
The concentration of HIV-1–producing T cells in CSF was estimated by multiplying the CSF white cell count per microliter and the CSF p24 percentage. This estimated cell concentration correlated strongly with viral load in CSF (P = 0.0006), but the percentage of HIV-1–producing T cells in CSF alone (p24 percentage) correlated only weakly (P = 0.11) with cell-free viral load. In regression analysis, with each doubling of the HIV-1–producing T-cell count, CSF viral load was estimated to be 0.26 log10 copy/mL higher. For example, if the HIV-1–producing T-cell count increased from 15 to 30 cells/μL, viral load would increase from 1840 to 3322 copies/μL. An analysis of CSF p24 percentage, HIV-1–producing cell count, and CSF white count variables (all controlled for therapy) indicated that CSF white count was the most statistically significant independent correlate of cell-free viral load in the CSF (P < 0.0001). Therefore, unexpectedly, the analysis of p24 percentage left the association almost unchanged and did not add additional predictive value to the CSF white count.
Low CD4 T-Cell Count in Blood Is Associated With a Higher Percentage of HIV-1–Producing T Cells in Cerebrospinal Fluid
The CD4 T-cell count in blood is a marker of HIV-1 disease stage and predicts the risk of clinical disease progression.30 A low CD4 T-cell count in this study was associated with a high CSF p24 percentage (estimate = −0.38, 95% CI: −0.75 to −0.01; P = 0.046), even after controlling for CSF white count and therapy (estimate = −0.39, 95% CI: −0.78–0.00; P = 0.049) and for CSF viral load and blood viral load (estimate= −0.43, 95% CI: −0.88–0.02; P = 0.059). For each doubling of the CD4 T-cell count in blood, the percentage of HIV-1–producing T cells was estimated to decrease 0.39% (controlled for therapy and white cell count). Antiretroviral therapy was associated with a decrease in the percentage of HIV-1–producing T cells. Unexpectedly, however, the relationship between CD4 and CSF p24 percentage was stronger in treated subjects (see Fig. 1C).
Viral Load and HIV-1–Producing T Cells Correlate With CD4+ T-Cell Activation in the Cerebrospinal Fluid Compartment But Not in the Blood Compartment
CD38 is a cell surface marker of cell activation on CD4+ T cells. CD4+ T-cell activation in CSF correlated to cell-free viral load in CSF and blood and with the percentage of HIV-1–producing T cells in CSF. In contrast, activated CD4+ T cells in blood did not appear to correlate with cell-free viral load or HIV-1–producing T cells in the same fluid. CSF p24 percentage was associated with CD4+ T-cell activation after adjustment of the effects of total CSF white cell count and treatment in multivariate analysis (P = 0.052; see Fig. 1D).
Composition of Cerebrospinal Fluid White Cell Count
HIV-1–infected subjects with higher CSF white cell counts tended to have lower percentages of CSF monocytes, indicating that the observed increase in white cell counts predominantly involved T cells. Granulocytosis in CSF was not observed.
The magnitude of HIV-1 viral load in cell-free CSF correlates with the number of white cells in CSF.5–8 To determine whether HIV-1–producing T cells appear in CSF and, if so, whether the percentage and number of infected cells in the CSF correlate with cell-free viral load, we developed a flow cytometric assay that detects intracellular HIV-1 p24 capsid antigen as a measure of HIV-1–producing T cells. Cell-free HIV-1 viral load in the blood has been associated with the rate of clinical progression in untreated subjects.30,31 In treated subjects, the change in plasma viral load relative to pretreatment baseline is a strong predictor of CD4 T-cell count decline. The biologic relevance of cell-free measures of viral load has been questioned, however.32 Indeed, HIV-1 requires host cells for replication and probably for transmission as well. Most of viral replication seems to occur in lymphoid tissues in which T cells and macrophages are closely approximated in a manner that appears to greatly enhance viral production.33–36 For these reasons, we have been interested in measuring viral expression in cells of tissues. CSF, which surrounds brain tissue, may reflect cellular events in brain tissue. The flow cytometry–based assay described here proved to be a sensitive and specific measurement of p24 HIV-1 capsid antigen expression in T cells, allowing us to identify HIV-1–producing T cells. Previous attempts to identify HIV-1–infected cells in fresh PBMC samples have failed because of abundant and highly variable background staining,37 which seems to be caused, in part, by nonspecific staining of monocytes and granulocytes with anti–HIV-1 p24 antibodies (data not shown). In contrast, anti-p24 staining in cultured cells has been successful,38,39 probably because of the following facts. In a stimulated culture, the percentages of HIV-1–producing cells are higher than in vivo; therefore, a high background is not as critical. Second, background staining in culture is reduced because of the adherence of monocytes to the culture dish and the relative absence of granulocytes after the Ficoll step. Anti-p24 staining has also been demonstrated in cell lines,40,41 where percentages of infected cells are also higher and the background staining is minimal because of the absence of monocytes and granulocytes. Our variation of the intracellular staining protocol, which includes removal of monocytes and granulocytes by magnetic bead separation and removing remaining monocytes from analysis by gating out of CD14+ cells, enables us to reliably detect HIV-1–producing T cells that might be present at very low percentages. The capacity to identify and quantify virus expression in specific T cells by flow cytometry may prove to be useful for research that aims to characterize target cells for HIV-1 and to evaluate viral and T-cell dynamics in tissues.
We found that HIV-1–producing T cells are uncommon in CSF. Typically, less than 1% of T cells were HIV-1 producing, with the highest value of 37 specimens being 2.39%. As reported earlier in a different set of patients,5,26 we found a strong relationship between CSF white cell count (which consists mainly of T cells) and CSF viral load (see Fig. 1A). We now find that CSF viral load is weakly associated with the percentage of T cells in CSF that express viral p24 antigen (P = 0.11). CSF viral load is strongly associated with the calculated number of HIV-1–producing cells, however, expressed as cell numbers per microliter of CSF (P = 0.0006). Hence, the strong association between numbers of HIV-1-producing cells and cell-free viral load in CSF is driven by the association between HIV-1 viral load and total CSF T-cell number, and the link between viral load and inflammatory cells in the CSF involves infected as well as uninfected cells. CD4+ T cells and CD4+/CD14+ cells from the monocyte/macrophage lineage can be readily infected with HIV-1, and their turnover is increased in HIV-1 infection.42 CD8+ T cells and B cells are not readily infected with HIV-1, but the turnover of these uninfected cell types is also increased in HIV-1 infection.42–47 The abundance of activated cells that do not produce HIV-1 in CSF may reflect increased proliferation or decreased destruction of uninfected cells in CSF compared with blood.
Indeed, we found that the relationship between viral load, HIV-1–producing cells, and CD4+ T-cell activation were different in CSF compared with blood. Viral load, HIV-1–producing cells and the level of CD4+ T-cell activation were associated with each other in CSF, in contrast to blood, where an association was not evident (see Fig. 1D). Interactions between viral replication and T cells may be different in the 2 tissue compartments. Extensive cell-to-cell contact in blood and lymphoid tissues may increase viral replication while concurrently enhancing loss of activated CD4+ T cells because of increased apoptosis, sequestration, or effector-cell killing. In contrast to blood, cell concentrations are much lower in CSF (1–5 cells per microliter of fluid), which may decrease cell-to-cell contact and allow activated CD4+ T cells to circulate, remain uninfected, or stay infected for longer periods of time. In this way, the linkage between viral replication and T-cell destruction that occurs in hematologic tissues may not occur in the CSF compartment. Therefore, in CSF, virus and target cells (activated CD4) might be high at the same time, whereas in blood, when virus is high, target cells might be low, and vice versa, because of the consumption of activated CD4 target cells in blood and not in CSF. Unlinking of viral replication and T-cell destruction in CSF may also contribute to the higher levels of T-cell activation that we observed in CSF (compared with blood). The mechanisms that allow activated CD4+ T cells in CSF to remain uninfected or stay infected for longer periods of time despite cocirculation with HIV-1–producing T cells and functional CD8 effector cells are not known.
Low blood CD4 T-cell count is a strong predictor of the percentage of HIV-1–producing T cells in CSF. This relationship remained after controlling for (1) CSF cell count, (2) antiretroviral therapy, (3) blood viral load, or (4) CSF viral load. Low CD4+ T-cell counts have been associated with high plasma viral load and high T-cell activation in blood, linking late-stage HIV-1 disease and inflammation.24,48 The increased percentage of HIV-1–producing T cells in CSF during late-stage HIV-1 disease may reflect increased systemic levels of viral replication or increased trafficking of infected cells into tissues. Cell trafficking from blood to CSF is consistent with our finding that all T-cell subsets (CD4 and CD8, activated and nonactivated) were correlated between blood and CSF.25 The fact that percentages of HIV-1–producing T cells in blood and CSF (P = 0.12) are somewhat linked is consistent with trafficking between blood and CSF or systemic effects that affect both fluids. Systemic determinants might regulate T-cell biology in blood and CSF, or upregulation of signals in the choroid plexus might lead to trafficking of certain cells types to the CSF.49,50
HIV-1–producing T cells in the CSF may contribute to HIV-1–specific brain pathology, which also occurs in later disease stages and continues to be an autopsy finding even in the era of highly active antiretroviral treatment (HAART).15–18 Interestingly, the relationship between low CD4+ T-cell count and higher frequency of HIV-1–producing T cells was stronger in treated subjects (see Fig. 1C), who were mostly suffering virological drug failure, than in untreated patients. Although clinical dementia is markedly improved during antiretroviral therapy, the persistence of HIV-1–specific pathologic changes during the HAART era raises the concern that clinical progression may accumulate over long periods of time among people surviving with HIV-1. Long-term clinical evaluation is required to address this concern.
We conclude that HIV-1–producing T cells appear in CSF and that their percentage and number correlate with cell-free viral load in CSF, even though the CSF total white cell count remains the best predictor for CSF viral load. Most CSF T cells were not HIV-1 producing, even when cell-free viral load in CSF was high. Most activated T cells in CSF were also not HIV-1 producing, but the activated CD38+ CD4 T-cell fraction in CSF was independently associated with the fraction of HIV-1–producing T cells in CSF. In HIV-1 infection, CSF white cell counts seem to contain a large number of uninfected cells. White cell counts and viral load in CSF may both result from systemic inflammation and immune activation.
The authors thank all study participants. The authors also thank the Gladstone Virology Core Laboratory for performing the viral load assays and Sarah Mays for computer assistance.
1. Hinkin CH, Castellon SA, Atkinson JH, et al. Neuropsychiatric aspects of HIV infection among older adults. J Clin Epidemiol. 2001;54(Suppl 1):S44–S52.
2. Everall IP, Heaton RK, Marcotte TD, et al. Cortical synaptic density is reduced in mild to moderate human immunodeficiency virus neurocognitive disorder. HNRC Group. HIV Neurobehavioral Research Center. Brain Pathol. 1999;9:209–217.
3. Castellon SA, Hinkin CH, Wood S, et al. Apathy, depression, and cognitive performance in HIV-1 infection. J Neuropsychiatry Clin Neurosci. 1998;10:320–329.
4. Lawrence DM, Major EO. HIV-1 and the brain: connections between HIV-1-associated dementia, neuropathology and neuroimmunology. Microbes Infect. 2002;4:301–308.
5. Price RW, Paxinos EE, Grant RM, et al. Cerebrospinal fluid response to structured treatment interruption after virological failure. AIDS. 2001;15:1251–1259.
6. Martin C, Albert J, Hansson P, et al. Cerebrospinal fluid mononuclear cell counts influence CSF HIV-1 RNA levels. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;17:214–219.
7. Garcia F, Niebla G, Romeu J, et al. Cerebrospinal fluid HIV-1 RNA levels in asymptomatic patients with early stage chronic HIV-1 infection: support for the hypothesis of local virus replication. AIDS. 1999;13:1491–1496.
8. Gisslen M, Fuchs D, Svennerholm B, et al. Cerebrospinal fluid viral load, intrathecal immunoactivation, and cerebrospinal fluid monocytic cell count in HIV-1 infection. J Acquir Immune Defic Syndr. 1999;21:271–276.
9. Brodt HR, Kamps BS, Gute P, et al. Changing incidence of AIDS-defining illnesses in the era of antiretroviral combination therapy [comments]. AIDS. 1997;11:1731–1738.
10. Brew BJ, Dore G. Decreasing incidence of CNS AIDS defining events associated with antiretroviral therapy. Neurology. 2000;55:1424.
11. Autran B, Carcelaint G, Li TS, et al. Restoration of the immune system with anti-retroviral therapy. Immunol Lett. 1999;66:207–211.
12. Cauda R, Tacconelli E, Tumbarello M, et al. Role of protease inhibitors in preventing recurrent oral candidosis in patients with HIV infection: a prospective case-control study [comments]. J Acquir Immune Defic Syndr. 1999;21:20–25.
13. Holtzer CD, Jacobson MA, Hadley WK, et al. Decline in the rate of specific opportunistic infections at San Francisco General Hospital, 1994–1997. AIDS. 1998;12:1931–1933.
14. Ives NJ, Gazzard BG, Easterbrook PJ. The changing pattern of AIDS-defining illnesses with the introduction of highly active antiretroviral therapy (HAART) in a London clinic. J Infect. 2001;42:134–139.
15. Masliah E, DeTeresa RM, Mallory ME, et al. Changes in pathological findings at autopsy in AIDS cases for the last 15 years. AIDS. 2000;14:69–74.
16. Dore GJ, Correll PK, Li Y, et al. Changes to AIDS dementia complex in the era of highly active antiretroviral therapy. AIDS. 1999;13:1249–1253.
17. Neuenburg JK, Brodt HR, Herndier BG, et al. HIV-related neuropathology, 1985 to 1999: rising prevalence of HIV encephalopathy in the era of highly active antiretroviral therapy. J Acquir Immune Defic Syndr. 2002;31:171–177.
18. Gray F, Adle-Biassette H, Chretien F, et al. Neuropathology and neurodegeneration in human immunodeficiency virus infection. Pathogenesis of HIV-induced lesions of the brain, correlations with HIV-associated disorders and modifications according to treatments. Clin Neuropathol. 2001;20:146–155.
19. Meyerhoff DJ, Bloomer C, Cardenas V, et al. Elevated subcortical choline metabolites in cognitively and clinically asymptomatic HIV+ patients. Neurology. 1999;52:995–1003.
20. Laubenberger J, Haussinger D, Bayer S, et al. HIV-related metabolic abnormalities in the brain: depiction with proton MR spectroscopy with short echo times. Radiology. 1996;199:805–810.
21. Barker PB, Lee RR, McArthur JC. AIDS dementia complex: evaluation with proton MR spectroscopic imaging. Radiology. 1995;195:58–64.
22. Paley M, Cozzone PJ, Alonso J, et al. A multicenter proton magnetic resonance spectroscopy study of neurological complications of AIDS. AIDS Res Hum Retroviruses. 1996;12:213–222.
23. Gartner S, Liu Y. Insights into the role of immune activation in HIV neuropathogenesis. J Neurovirol. 2002;8:69–75.
24. Giorgi JV, Lyles RH, Matud JL, et al. Predictive value of immunologic and virologic markers after long or short duration of HIV-1 infection. J Acquir Immune Defic Syndr. 2002;29:346–355.
25. Price RW, Neuenburg JK, Nilsson A, et al. Cerebrospinal fluid T lymphocytes: expression of phenotypic activation markers in HIV Infection. Presented at the 8th Conference on Retroviruses and Opportunistic Infections, Chicago, February 2001.
26. Staprans S, Marlowe N, Glidden D, et al. Time course of cerebrospinal fluid responses to antiretroviral therapy: evidence for variable compartmentalization of infection. AIDS. 1999;13:1051–1061.
27. Piguet V, Chen YL, Mangasarian A, et al. Mechanism of Nef-induced CD4 endocytosis: Nef connects CD4 with the mu chain of adaptor complexes. Embo J. 1998;17:2472–2481.
28. Christopherson C, Kidane Y, Conway B, et al. PCR-based assay to quantify human immunodeficiency virus type 1 DNA in peripheral blood mononuclear cells. J Clin Microbiol. 2000;38:630–634.
29. Efron B, Tibshirani RJ. An Introduction to the Bootstrap. London: Chapman and Hall; 1993.
30. Mellors JW, Munoz A, Giorgi JV, et al. Plasma viral load and CD4+ lymphocytes as prognostic markers of HIV-1 infection. Ann Intern Med. 1997;126:946–954.
31. Mellors JW, Rinaldo CR Jr, Gupta P, et al. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science. May. 1996;272:1167–1170.
32. Wain-Hobson S. AIDS. Virological mayhem. Nature. 1995;373:102.
33. Frankel SS, Wenig BM, Burke AP, et al. Replication of HIV-1 in dendritic cell-derived syncytia at the mucosal surface of the adenoid. Science. 1996;272:115–117.
34. Cameron P, Pope M, Granelli-Piperno A, et al. Dendritic cells and the replication of HIV-1. J Leukoc Biol. 1996;59:158–171.
35. Pope M, Gezelter S, Gallo N, et al. Low levels of HIV-1 infection in cutaneous dendritic cells promote extensive viral replication upon binding to memory CD4+ T cells. J Exp Med. 1995;182:2045–2056.
36. Pope M, Betjes MG, Romani N, et al. Dendritic cell-T cell conjugates that migrate from normal human skin are an explosive site of infection for HIV-1. Adv Exp Med Biol. 1995;378:457–460.
37. Cameron PU, Hunter SD, Jolley D, et al. Specificity of binding of HIV-1 anti-p24 antibodies to CD4+ lymphocytes from HIV-infected subjects. Cytometry. 1998;33:83–88.
38. Mascola JR, Louder MK, Winter C, et al. Human immunodeficiency virus type 1 neutralization measured by flow cytometric quantitation of single-round infection of primary human T cells. J Virol. 2002;76:4810–4821.
39. Darden JM, Polonis VR, deSouza MS, et al. A flow cytometric method for measuring neutralization of HIV-1 subtype B and E primary isolates. Cytometry. 2000;40:141–150.
40. Cory JM, Ohlsson-Wilhelm BM, Steck ME, et al. Kinetics of infected cell appearance as a determinant of number of human immunodeficiency virus-1 infectious units. AIDS Res Hum Retroviruses. 1989;5:97–106.
41. Lee SW, Gallardo HF, Gaspar O, et al. Inhibition of HIV-1 in CEM cells by a potent TAR decoy. Gene Ther. 1995;2:377–384.
42. McCune JM, Hanley MB, Cesar D, et al. Factors influencing T-cell turnover in HIV-1-seropositive patients. J Clin Invest. 2000;105(Suppl:R1–R8.
43. Carbonari M, Pesce AM, Cibati M, et al. Death of bystander cells by a novel pathway involving early mitochondrial damage in human immunodeficiency virus-related lymphadenopathy. Blood. 1997;90:209–216.
44. Ullrich CK, Groopman JE, Ganju RK. HIV-1 gp120- and gp160-induced apoptosis in cultured endothelial cells is mediated by caspases. Blood. 2000;96:1438–1442.
45. Oyaizu N, Adachi Y, Hashimoto F, et al. Monocytes express Fas ligand upon CD4 cross-linking and induce CD4+ T cells apoptosis: a possible mechanism of bystander cell death in HIV infection. J Immunol. 1997;158:2456–2463.
46. Suzuki S, Tobiume M, Kameoka M, et al. Exposure of normal monocyte-derived dendritic cells to human immunodeficiency virus type-1 particles leads to the induction of apoptosis in co-cultured CD4+ as well as CD8+ T cells. Microbiol Immunol. 2000;44:111–121.
47. Larsson M, Fonteneau JF, Lirvall M, et al. Activation of HIV-1 specific CD4 and CD8 T cells by human dendritic cells: roles for cross-presentation and non-infectious HIV-1 virus. AIDS. 2002;16:1319–1329.
48. Tilling R, Kinloch S, Goh LE, et al. Parallel decline of CD8+/CD38++ T cells and viraemia in response to quadruple highly active antiretroviral therapy in primary HIV infection. AIDS. 2002;16:589–596.
49. Kivisakk P, Mahad DJ, Callahan MK, et al. Human cerebrospinal fluid central memory CD4+ T cells: evidence for trafficking through choroid plexus and meninges via P-selectin. Proceedings of the National Academy of Sciences. 2003;100:8389–8394.
50. Harouse JM, Wroblewska Z, Laughlin MA, et al. Human choroid plexus cells can be latently infected with human immunodeficiency virus. Ann Neurol. 1989;25:406–411.
This article has been cited 1 time(s).
Plos OneCellular Composition of Cerebrospinal Fluid in HIV-1 Infected and Uninfected SubjectsPlos One
cerebrospinal fluid; p24 HIV-1 capsid antigen; cerebrospinal fluid white cell count; HIV-1; T cells; activation marker CD38; immune activation; flow cytometry
© 2004 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.