Abstract
The source of human immunodeficiency virus type 1 (HIV-1) RNA in cerebrospinal fluid (CSF) during HIV-1 infection is uncertain. The sequence heterogeneity of HIV-1 RNA in simultaneous CSF and plasma samples was characterized for five patients at the baseline and during the first week of antiretroviral therapy by two commercial genotyping methodologies. In individual subjects, the sequences in CSF samples differed significantly from those in plasma. In contrast, the viral sequences in CSF at the baseline did not differ from the sequences in CSF during treatment. Similarly, viral sequences in plasma did not vary over this interval. This study provides evidence that HIV-1 RNA in CSF and plasma arise from distinct compartments.
Human immunodeficiency virus type 1 (HIV-1) is present in the central nervous system (CNS) during all stages of HIV infection and accounts for the high prevalence of dementia during advanced AIDS (2, 3, 11). In addition, there is evidence that the CNS is a distinct compartment for viral replication (17, 19).
Various antiretroviral regimens have been shown to lower the plasma HIV-1 RNA load to undetectable levels in treatment-naïve patients, and the CNS complications of AIDS often respond to antiretroviral therapy (1). However, control of HIV-1 replication in peripheral compartments does not ensure control in the CNS (6, 14; E. H. Gisolf, P. Portegies, R. Hoetelmans, M. E. Van der Ende, K. Brinkman, F. de Wolf, and S. A. Danner, Proc. 12th World AIDS Conf., 1999). Inadequate drug penetration into the CNS may provide a sanctuary from which resistant virus may emerge or may allow psychomotor abnormalities to develop. Although HIV-1 RNA is often present in cerebrospinal fluid (CSF), its source has not been defined (15).
The high spontaneous mutation rate of HIV-1 RNA leads to considerable in vivo sequence heterogeneity (18). Population-based sequencing approaches amplify HIV-1 RNA directly from this mixed population in plasma and are commonly used to guide antiretroviral therapy. Several commercial HIV-1 genotyping assay methods are now available (5, 9, 10). Such methodologies identify predominant viral species but not less abundant subpopulations (4).
We have previously characterized the kinetics of HIV-1 RNA decay and drug disposition in CSF and plasma among treatment-naïve adults during the initial week of three-drug therapy with stavudine, lamivudine, and nelfinavir (7). In that study, different kinetics of HIV-1 RNA decay in CSF and plasma suggested distinct sources of viral replication. To gain further insight into the possible source of virus in CSF, the present study examined the sequence heterogeneity of HIV-1 RNA in CSF compared to that in plasma by the TrueGene assay. The HIV RNA concentrations in the samples encompassed a relatively wide range. The validity of the genotyping results was confirmed by repeating the assays with CSF and plasma with the VircoGEN genotyping system.
Sampling of CSF and plasma.
Samples of CSF and plasma were collected during a prospective study that used ultraintensive CSF sampling and concomitant serial plasma sampling (7). Briefly, CSF was obtained continuously from four antiretroviral-naïve, HIV-positive adults over 48-h intervals via indwelling intrathecal catheters at the baseline (beginning 5 days before the initiation of therapy) and again during therapy (referred to as on-treatment; beginning on day 3 after the initiation of therapy). Serial plasma samples were obtained concomitantly. All subjects received stavudine (40 or 30 mg every 12 h), lamivudine (150 mg every 12 h), and nelfinavir (750 mg every 8 h). A fifth subject underwent CSF and plasma sampling only at the baseline. Samples of CSF and plasma from the baseline and day 5 of therapy (if available) were analyzed. Thus, the four fully evaluable subjects yielded eight separate genotypic assay results (i.e., baseline and day 5 CSF and plasma samples, each of which was assayed by two separate assay methods). The study was approved by the Vanderbilt University Institutional Review Board, and all subjects provided written informed consent.
HIV-1 RNA assays.
For each patient, CSF and plasma HIV-1 RNA concentrations at the baseline were based on the mean of as many as 17 separate determinations with samples obtained at 3-h intervals. Similarly, on-treatment CSF and plasma HIV-1 RNA concentrations were based on the mean of 17 separate determinations with 17 samples obtained at 3-h intervals beginning on day 3 of therapy. Quantification of HIV-1 RNA in CSF and plasma was by the Nuclisens assay (NASBA; Organon Teknika Corp., Durham, N.C.) (16).
HIV genotyping and antiretroviral resistance testing.
Sequence analysis of virus in CSF and plasma samples was performed by the TruGene assay (Visible Genetics Inc., Toronto, Ontario, Canada) (5) and the VicroGEN assay (Virco, U.K. Ltd., Cambridge, United Kingdom) (8). Extraction of RNA was performed with the QIAmp Tissue RNA kit (Qiagen Inc., Santa Clara, Calif.), according to the manufacturer's instructions. Genotyping assays used a three-step procedure: (i) PCR amplification of HIV-1 reverse transcriptase (RT) and protease regions, (ii) mutation identification by nucleotide sequencing, and (iii) antiretroviral drug resistance determination by a genomic database search (5, 10). Assays were done according to the manufacturers' instructions. The TruGene assay results include both nucleotide and amino acid sequences, while the VircoGEN assay results include only amino acid sequences. Since patients were treatment naïve, analyses were performed primarily to characterize all polymorphisms and mutations, not only those associated with drug resistance. The sequence of the reference HxB strain was used as a standard for determination of the number and sites of polymorphisms (10).
Statistical analyses.
To compare the number of mutations between pairs of sample sets, the total number of mutations for each sample was calculated, and the difference between sample sets was determined by the paired-samples Wilcoxon sign-rank test. Agreement between assay methodologies was determined by chi-square Mantel-Haenszel statistic. Relationships between HIV-1 RNA concentration and number of nucleotide changes were determined by calculating the Pearson correlation coefficient. Statistical analyses were performed with SPSS, version 9.0 (SPSS, Chicago, Ill.), and Epi Info, version 6.04b (Centers for Disease Control and Prevention, Atlanta, Ga.).
The changes in the plasma HIV-1 RNA loads from the baseline to treatment day 5 among the four fully evaluable subjects ranged from −0.80 to −1.33 log10 copies/ml (to as low as 3.54 log10 copies/ml), while the changes in the CSF HIV-1 RNA loads from the baseline to day 5 ranged from −0.38 to −1.18 log10 copies/ml (to as low as 2.54 log10 copies/ml), as shown in Table 1. The availability of these paired CSF and plasma samples obtained over 1-week intervals and representing a range of HIV-1 RNA concentrations in each patient provided the opportunity to compare the sequence heterogeneity of HIV-1 RNA in CSF and plasma. We first used the TruGene assay to compare the HIV-1 RT and protease nucleotide sequence heterogeneity in CSF and plasma. As expected, interindividual heterogeneity greatly exceeded intraindividual heterogeneity (data not shown). An analysis which included both silent nucleotide changes (i.e., changes that did not code for an amino acid change) and coding nucleotide changes (i.e., changes that altered an amino acid) in comparison to the sequence of the reference HIV-1 HxB strain demonstrated significantly more nucleotide differences for virus in CSF than for virus in plasma among the entire study population (P = 0.001) (Table 1).
TABLE 1.
Number of nucleotide mutations and HIV-1 RNA concentrations in CSF and plasmaa
| Subject | Time | Plasma
|
CSF
|
||||
|---|---|---|---|---|---|---|---|
| No. of mutations
|
HIV-1 RNA load (log10 copies/ml) | No. of mutations
|
HIV-1 RNA load (log10 copies/ml) | ||||
| Protease | RT | Protease | RT | ||||
| A | Baseline | 17 | 18 | 4.87 | 17 | 21 | 3.34 |
| Day 5 | 13 | 18 | 3.54 | 14 | 19 | 2.54 | |
| B | Baseline | 8 | 25 | 4.74 | 7 | 24 | 4.68 |
| Day 5 | 7 | 20 | 3.94 | 7 | 24 | 3.50 | |
| C | Baseline | 8 | 27 | 4.69 | 11 | 30 | 4.32 |
| Day 5 | 8 | 25 | 3.57 | 10 | 29 | 3.94 | |
| D | Baseline | 7 | 19 | 4.63 | 8 | 21 | 4.07 |
| Day 5 | 6 | 26 | 3.66 | 8 | 30 | 3.23 | |
| E | Baseline | 7 | 22 | 5.33 | 10 | 25 | 4.06 |
The sequence of the reference HIV-1 HxB strain was used as a standard for determination of the number of nucleotide changes for each sample (10).
We next determined whether the number of coding changes was affected by the treatment-induced change in the HIV-1 RNA concentration. The sequences at the baseline did not differ significantly from those at day 4 for HIV-1 RNA in both CSF and plasma (P = 0.28) (Table 1). In addition, there was no correlation between the HIV-1 RNA concentration in CSF or plasma and the number of nucleotide changes compared to the sequence of strain HxB. This analysis confirmed that the increased number of nucleotide changes identified in CSF was not an artifact due to the somewhat lower HIV-1 RNA concentrations in CSF than those in plasma.
The analyses described above involved comparison of patient HIV-1 RNA sequences to the HIV-1 RNA sequence of the HxB reference strain. We next performed four different two-way comparisons of nucleotide sequence differences between HIV-1 RNAs, from CSF and plasma samples at the baseline and on-treatment for each subject. As shown in Table 2, in every case and for both protease and RT, the number of HIV-1 RNA nucleotide differences between baseline CSF samples and baseline plasma samples exceeded those for comparisons of baseline CSF samples and on-treatment CSF samples and for comparisons of baseline plasma samples and on-treatment plasma samples. In contrast, no consistent relationship was observed when the number of nucleotide differences between baseline CSF and on-treatment CSF samples was compared to that for baseline plasma samples and on-treatment plasma samples. Similarly, no consistent relationship was observed when the number of nucleotide differences between baseline CSF and baseline plasma samples was compared to that for on-treatment CSF samples and on-treatment plasma samples. This analysis confirms that the number of HIV-1 RNA sequence differences between CSF and plasma in treatment-naïve patients exceeds the number of differences between either serial CSF samples or serial plasma samples.
TABLE 2.
Comparison of number of HIV-1 RNA nucleotide differences between CSF and plasma samples at baseline and on-treatment as determined by TruGene assay
| Gene and subject | No. of nucleotide differencesa
|
|||
|---|---|---|---|---|
| PlasmaB-CSFB | PlasmaT-CSFT | PlasmaB-plasmaT | CSFB-CSFT | |
| Protease gene | ||||
| A | 6 | 5 | 4 | 3 |
| B | 3 | 2 | 1 | 2 |
| C | 3 | 2 | 0 | 1 |
| D | 4 | 6 | 3 | 1 |
| E | 3 | |||
| RT gene | ||||
| A | 7 | 6 | 5 | 4 |
| B | 5 | 7 | 4 | 2 |
| C | 5 | 5 | 3 | 3 |
| D | 12 | 17 | 7 | 9 |
| E | 7 | |||
Subscript B, baseline; subscript T, on-treatment.
The reliability of the TruGene assay was confirmed by also analyzing the CSF and plasma samples by the VicroGen method. As anticipated, there were no RT or protease primary drug resistance mutations either at the baseline or during the first week of therapy in this treatment-naïve population. There were a number of secondary resistance mutations identified, although the clinical relevance of these mutations is uncertain. Importantly, there was complete agreement between the TruGene and VircoGEN assays in identifying possible secondary resistance-associated amino acid changes and 98.8% ± 1.0% agreement for detecting any amino change compared to the sequence of the HxB strain (Table 3). The agreement between methods was not affected by whether CSF or plasma was assayed (P = 0.91) or by whether samples were from the baseline or day 5 of therapy (P = 0.27). The close agreement between the two systems also strongly suggests that either assay may be used to identify amino acid changes in CSF as well as plasma, although a larger study which includes specimens from patients failing antiretroviral therapy would be needed to definitely compare these assays.
TABLE 3.
Concordance between TruGene and VircoGEN assays for detection of secondary resistance-related mutations and any coding mutations in HIV-1 protease or RT genome
| Subject | Time | No. (%) of matched mutations
|
|||
|---|---|---|---|---|---|
| Concordance for plasma
|
Concordance for CSF
|
||||
| Resistance related mutationa | Matched coding mutationb | Resistance related mutation | Matched coding mutation | ||
| A | Baseline | 3 (100.0) | 96 (97.0) | 3 (100.0) | 206 (99.5) |
| Day 5 | 3 (100.0) | 98 (99.0) | 3 (100.0) | 205 (99.0) | |
| B | Baseline | 1 (100.0) | 99 (100.0) | 1 (100.0) | 207 (100.0) |
| Day 5 | 1 (100.0) | 96 (97.0) | 1 (100.0) | 206 (99.5) | |
| C | Baseline | 2 (100.0) | 98 (99.0) | 2 (100.0) | 206 (99.5) |
| Day 5 | 2 (100.0) | 99 (100.0) | 2 (100.0) | 204 (98.6) | |
| D | Baseline | 1 (100.0) | 98 (99.0) | 1 (100.0) | 204 (98.6) |
| Day 5 | 1 (100.0) | 96 (97.0) | 1 (100.0) | 204 (98.6) | |
Based on 306 amino acids (99 for protease and 207 for RT) derived from 919 nucleotide base pairs covered by the genotyping testing. The sequence of the reference HIV-1 HxB strain was used as a standard for determination of codon changes (10).
These represented mutations L63P, A71T, and V77I in the protease gene (subject A); mutation I50V in the RT gene (subject B); mutations L63P and V77I in the protease gene (subject C); and mutation L63P in the protease gene (subject D). These changes were identical for CSF and plasma.
The introduction of potent antiretroviral regimens has markedly decreased AIDS-related mortality in developed countries (12, 13). However, ongoing replication of HIV-1 in the presence of suboptimal therapy selects for resistant virus and limits treatment options (10). The observation that HIV-1 RNA sequences in CSF are distinct from those in plasma provides evidence that at least some HIV-1 in CSF arises from a source other than plasma. This finding also supports the importance of controlling HIV-1 replication in the CNS as well as peripheral tissues with antiretroviral therapy.
Acknowledgments
This study was supported in part by NIH grant RR-00095 (GCRC).
REFERENCES
- 1.Albrecht H, Hoffmann C, Degen O, Stoehr A, Plettenberg A, Mertenskotter T, Eggers C, Stellbrink H J. Highly active antiretroviral therapy significantly improves the prognosis of patients with HIV-associated progressive multifocal leukoencephalopathy. AIDS. 1998;12:1149–1154. doi: 10.1097/00002030-199810000-00006. [DOI] [PubMed] [Google Scholar]
- 2.Bell J E, Brettle R P, Chiswick A, Simmonds P. HIV encephalitis, proviral load and dementia in drug users and homosexuals with AIDS. Effect of neocortical involvement. Brain. 1998;121:2043–2052. doi: 10.1093/brain/121.11.2043. [DOI] [PubMed] [Google Scholar]
- 3.Cinque P, Vago L, Ceresa D, Mainini F, Terreni M R, Vagani A, Torri W, Bossolasco S, Lazzarin A. Cerebrospinal fluid HIV-1 RNA levels: correlation with HIV encephalitis. AIDS. 1998;12:389–394. doi: 10.1097/00002030-199804000-00007. [DOI] [PubMed] [Google Scholar]
- 4.Coffin J M. HIV population dynamics in vivo: implications for genetic variation, pathogenesis, and therapy. Science. 1995;267:483–489. doi: 10.1126/science.7824947. [DOI] [PubMed] [Google Scholar]
- 5.Durant J, Clevenbergh P, Halfon P, Delgiudice P, Porsin S, Simonet P, Montagne N, Boucher C A, Schapiro J M, Dellamonica P. Drug-resistance genotyping in HIV-1 therapy: the VIRADAPT randomised controlled trial. Lancet. 1999;353:2195–2199. doi: 10.1016/s0140-6736(98)12291-2. [DOI] [PubMed] [Google Scholar]
- 6.Gisslen M, Norkrans G, Svennerholm B, Hagberg L. The effect on human immunodeficiency virus type 1 RNA levels in cerebrospinal fluid after initiation of zidovudine or didanosine. J Infect Dis. 1997;175:434–437. doi: 10.1093/infdis/175.2.434. [DOI] [PubMed] [Google Scholar]
- 7.Haas D W, Clough L, Johnson B W, Harris V L, Spearman P, Wilkinson G R, Fletcher C V, Fiscus S, Raffanti S, Donlon R, McKinsey J, Nicotera J, Schmidt D, Shoup R E, Kates R E, Lloyd R M, Larder B. Evidence of a source of HIV type 1 within the central nervous system by ultraintensive sampling of cerebrospinal fluid and plasma. AIDS Res Hum Retrovir. 2000;16:1491–1502. doi: 10.1089/088922200750006010. [DOI] [PubMed] [Google Scholar]
- 8.Harrigan P R, Hertogs K, Verbiest W, Pauwels R, Larder B, Kemp S, Bloor S, Yip B, Hogg R, Alexander C, Montaner J S. Baseline HIV drug resistance profile predicts response to ritonavir-saquinavir protease inhibitor therapy in a community setting. AIDS. 1999;13:1863–1871. doi: 10.1097/00002030-199910010-00008. [DOI] [PubMed] [Google Scholar]
- 9.Havlir D V, Hellmann N S, Petropoulos C J, Whitcomb J M, Collier A C, Hirsch M S, Tebas P, Sommadossi J P, Richman D D. Drug susceptibility in HIV infection after viral rebound in patients receiving indinavir-containing regimens. JAMA. 2000;283:229–234. doi: 10.1001/jama.283.2.229. [DOI] [PubMed] [Google Scholar]
- 10.Hirsch M S, Conway B, D'Aquila R T, Johnson V A, Brun-Vezinet F, Clotet B, Demeter L M, Hammer S M, Jacobsen D M, Kuritzkes D R, Loveday C, Mellors J W, Vella S, Richman D D. Antiretroviral drug resistance testing in adults with HIV infection: implications for clinical management. JAMA. 1998;279:1984–1991. doi: 10.1001/jama.279.24.1984. [DOI] [PubMed] [Google Scholar]
- 11.McArthur J C, McClernon D R, Cronin M F, Nance-Sproson T E, Saah A J, St Clair M, Lanier E R. Relationship between human immunodeficiency virus-associated dementia and viral load in cerebrospinal fluid and brain. Ann Neurol. 1997;42:689–698. doi: 10.1002/ana.410420504. [DOI] [PubMed] [Google Scholar]
- 12.Mocroft A, Vella S, Benfield T L, Chiesi A, Miller V, Gargalianos P, d'Arminio Monforte A, Yust I, Bruun J N, Phillips A N, Lundgren J D. Changing patterns of mortality across Europe in patients infected with HIV-1. Lancet. 1998;352:1725–1730. doi: 10.1016/s0140-6736(98)03201-2. [DOI] [PubMed] [Google Scholar]
- 13.Palella F J, Jr, Delaney K M, Moorman A C, Loveless M O, Fuhrer J, Satten G A, Aschman D J, Holmberg S D. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med. 1998;338:853–860. doi: 10.1056/NEJM199803263381301. [DOI] [PubMed] [Google Scholar]
- 14.Pialoux G, Fournier S, Moulignier A, Poveda J D, Clavel F, Dupont B. Central nervous system as a sanctuary for HIV-1 infection despite treatment with zidovudine, lamivudine and indinavir. AIDS. 1997;11:1302–1303. doi: 10.1097/00002030-199710001-00009. [DOI] [PubMed] [Google Scholar]
- 15.Price R W, Staprans S. Measuring the “viral load” in cerebrospinal fluid in human immunodeficiency virus infection: window into brain infection? Ann Neurol. 1997;42:675–8. doi: 10.1002/ana.410420502. [DOI] [PubMed] [Google Scholar]
- 16.Segondy M, Ly T D, Lapeyre M, Montes B. Evaluation of the Nuclisens HIV-1 QT assay for quantitation of human immunodeficiency virus type 1 RNA levels in plasma. J Clin Microbiol. 1998;36:3372–3374. doi: 10.1128/jcm.36.11.3372-3374.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Venturi G, Catucci M, Romano L, Corsi P, Leoncini F, Valensin P E, Zazzi M. Antiretroviral resistance mutations in human immunodeficiency virus type 1 reverse transcriptase and protease from paired cerebrospinal fluid and plasma samples. J Infect Dis. 2000;181:740–745. doi: 10.1086/315249. [DOI] [PubMed] [Google Scholar]
- 18.Wain-Hobson S. The fastest genome evolution ever described: HIV variation in situ. Curr Opin Genet Dev. 1993;3:878–883. doi: 10.1016/0959-437x(93)90008-d. [DOI] [PubMed] [Google Scholar]
- 19.Wong J K, Ignacio C C, Torriani F, Havlir D, Fitch N J, Richman D D. In vivo compartmentalization of human immunodeficiency virus: evidence from the examination of pol sequences from autopsy tissues. J Virol. 1997;71:2059–2071. doi: 10.1128/jvi.71.3.2059-2071.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]