JOURNAL OF VIROLOGY, Nov. 2002, p. 10942–10950 0022-538X/02/$04.00⫹0 DOI: 10.1128/JVI.76.21.10942–10950.2002 Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 76, No. 21
A Sensitive, Quantitative Assay for Human Immunodeficiency Virus Type 1 Integration Una O’Doherty,1* William J. Swiggard,2 Deepa Jeyakumar,1 David McGain,1 and Michael H. Malim3 Departments of Pathology and Laboratory Medicine1 and Medicine,2 University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, and Department of Infectious Diseases, Guy’s, King’s and St. Thomas’ School of Medicine, King’s College London, London, England3 Received 26 April 2002/Accepted 15 July 2002
Quantitative methods to measure human immunodeficiency virus type 1 (HIV-1) integration promise to be important tools in dissecting the mechanisms whereby latent reservoirs of provirus are established, most notably in the resting T cells of patients receiving antiretroviral therapy. Here we describe a fluorescencemonitored, nested PCR assay that is able to quantify the relatively rare integration events that occur within these cells. Following DNA extraction, a nonkinetic preamplification step is performed with primers that bind genomic Alu elements and HIV-1 gag sequences, under conditions where primers, deoxynucleoside triphosphates, and enzyme are not limiting. This is followed by a kinetic PCR that quantitates HIV-1 long terminal repeat sequences. A T-cell-based integration standard which reflects the randomness of HIV-1 integration is also described. The assay is 10 to 100 times more sensitive than previously reported quantitative Alu PCRbased integration assays. It is specific for integration events, since no proviruses are detected in cells infected either in the presence of an integrase inhibitor or with an integrase-deficient virus. This method promises to provide important new insights into the processes underlying the accumulation and persistence of latent HIV-1 reservoirs and may eventually be useful clinically in monitoring the eradication of latent virus by novel therapies.
Integration, an obligate step in retrovirus replication, is defined as the covalent insertion of viral DNA into the host cell genome. Human immunodeficiency virus type 1 (HIV-1) proviruses appear to contribute to viral persistence. This is most evident in treated patients, where the proviral load seems to be unaffected by combination antiretroviral therapy that clears viremia (10). To date, proviral DNA has proven impossible to eradicate without death of the transduced host cell. These results suggest that provirus may serve as a treatment-resistant reservoir of HIV-1. Quantitative studies of this reservoir are challenging because of the relative rarity of integration events in quiescent leukocytes. Assays for HIV-1 integration have been implemented using three main strategies, namely, inverse PCR (7), linker-primer PCR (37), and Alu PCR (2, 3, 5, 12, 34). We focus on and refine the latter method here. Alu elements are the most numerous repetitive elements in primate genomic DNA, comprising over 1 million copies per diploid cell (22, 26), or 5% of the mass of the human genome. They continue to accumulate, at a rate of about 1 insertion in every 200 live births (14), via an RNA polymerase III-dependent process called retroposition. Alu elements are randomly distributed, roughly 5,000 bp apart, and are randomly oriented. Early HIV-1 integration assays utilizing Alu repeat elements as “anchors” within genomic DNA were sensitive but were not strictly quantitative, since they lacked real-time reaction mon-
* Corresponding author. Mailing address: Transfusion Medicine Division, Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, 402C Johnson Pavilion/6060, 3610 Hamilton Walk, Philadelphia, PA 19104-6060. Phone: (215) 573-7273. Fax: (215) 573-4856. E-mail:
[email protected].
itoring and polyclonal standards. In particular, most of these studies used genomic DNA prepared from graded doses of persistently infected cell lines such as 8E5 (34), U1 (10), ACH-2 (10, 29), or mixtures of such lines (37) as their standards. The problem with these cell lines is that they are clonal and contain one (19), two (18), or two (11) nonrandom HIV-1 integration sites, respectively. Standards prepared from these lines are likely to under- or overestimate the number of integration events in nearly all types of unknowns, since the proviruses these cell lines harbor are integrated at specific, nonvariable distances from the Alu repeats within their genomes. Single integration events can be detected in these clonal DNA standards, presumably arising from integration sites relatively close to the nearest Alu element. However, in assigning numbers to these signals, an implicit assumption must be made that all integration events in the unknowns must also be detectable by PCR (7). Due to the random nature of both HIV-1 integration and Alu retroposition, each Alu-gag integration site in a nonclonal collection of infected cells will be amplified with a different efficiency. It is likely that significant (and unknown) numbers of proviruses are inserted too far away from an Alu site to be efficiently amplified. A corollary to this assertion is that the frequencies of integration within resting CD4⫹ T cells that were previously determined using clonal PCR standards are likely to have been underestimated. A superior integration standard would contain multiple unique integration sites (thousands or more), all integrated randomly at a range of distances from the nearest Alu repeat sequence. Recently, the first fully quantitative one-step kinetic Alu PCR assay for HIV-1 integration was described (5). It was shown to be capable of detecting as few as one provirus in 100
10942
VOL. 76, 2002
QUANTITATION OF HIV-1 PROVIRAL DNA
10943
FIG. 1. Preparation and characterization of the IS cell line. (A) Infection and selection strategy. (B) Equal amounts of DNA (20 g/lane) from CEM-SS cells acutely infected with wild-type HIV-1IIIB (left lane), the antibiotic-selected IS line (center lane), or uninfected CEM-ss cells (right lane) were electrophoresed on a 0.8% agarose gel and stained with ethidium bromide prior to Southern blotting. (C) Autoradiogram of the Southern blot prepared from the gel in panel B. Whereas integrated, linear, and circular forms of HIV-1 DNA were present in the acutely infected cells (left lane), only integrated forms of HIV-1 DNA were present in the IS (center lane). Uninfected cells were free of HIV-1 signals (right lane). The intensities of the HIV-1 bands in the acutely infected sample (where DNA was prepared 20 h after viral exposure) are higher than those in the IS, because spinoculation delivered such a high multiplicity of infection to the CEM-SS-R5 cells that multiple proviruses were established in the majority of cells in the sample. In contrast, following 4 weeks of drug selection, cells in the IS contained a mean (⫾ standard deviation) of 1.4 ⫾ 0.3 proviruses per cell.
target cells. The importance of appropriate standardization was emphasized, in particular the need for a standard that accurately reflected the random nature of integration. An integration standard was generated by infection of 293T cells with a high titer of virus, followed by a culture period of at least 30 days, to allow unintegrated forms of HIV-1 DNA to be eliminated (1, 16, 30). Only then was genomic DNA prepared. In order to study the efficiency of integration in resting CD4⫹ T cells, we required a fully quantitative method with improved sensitivity. In our hands, one-step PCR methods lacked the requisite sensitivity for use in primary T cells. Accordingly, a nested PCR strategy was implemented, using conditions where deoxynucleoside triphosphates (dNTPs), primers, and enzyme were not limiting in the nonkinetic preamplification of Alu-gag sequences. The preamplification is followed by a real-time PCR assay that quantitates HIV-1 long terminal repeat (LTR) sequences using molecular beacon detection (36). In parallel, we developed an alternative integration standard (IS) cell line, wherein a large number of target T cells were infected with a high-titer, randomly integrated, replication-incompetent HIV-1 construct containing an antibiotic resistance cassette. Prolonged antibiotic selection was performed prior to preparation of genomic DNA from these IS
cells. The latter step ensured that every cell in this population contained at least one provirus, and that all of the HIV-1 DNA within these cells was integrated. The large number of targets initially infected insured that these integration events were numerous, occurring at a large range of distances from genomic Alu elements. MATERIALS AND METHODS Cell lines, plasmids, and viruses. The CD4⫹ T-lymphoblastoid cell lines CEM-SS (27) and ACH-2 (11) were maintained at densities of 105 to 106 per ml and 5 ⫻ 104 to 5 ⫻ 105 per ml, respectively. The culture medium was RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 25 mM HEPES, pH 7.4, with 1⫻ penicillin-streptomycin (Invitrogen Life Technologies). In order to prepare the IS cell line, a rev-deficient, G418 resistance-conferring proviral vector, pIIIB/⌬R/N, was derived from pIIIB/⌬ 12Rev/N (25) by repairing the nef gene and U3 sequences and inserting a 5⬘-internal ribosome entry site downstream of the neomycin resistance cassette. 293T cells were then cotransfected with three plasmids, namely, (i) pIIIB/⌬R/N, (ii) pcRev, and (iii) pVSV G (pHIT/G) (20). Twenty-four hours after transfection, the supernatants containing pseudotyped virus were centrifuged at 500 ⫻ g for 10 min, treated with 30 g of RNase-free DNase I (Roche)/ml for 30 min at room temperature in the presence of 10 mM MgCl 2, and then sterile filtered through 0.2-m-pore-size syringe filters (Acrodisc, HT Tuffryn, Pall Gelman). CEM-SS cells expressing CCR5 (referred to as CEM-SS-R5 below) were prepared as described previously (20) and then were enriched by fluorescence-
10944
O’DOHERTY ET AL.
J. VIROL.
FIG. 2. PCR and standardization strategy. (A) The nonkinetic PCR preamplification uses primers that bind genomic Alu and HIV-1 gag sequences. Substrates and enzyme are not limiting. Heterogeneous amplicons of variable length are produced. (B) Kinetic PCR to quantitate HIV-1-specific sequences within the LTR. (C) First regression, for provirus content of the IS. (D) Second regression, for provirus content of the unknowns.
activated cell sorting for a population that expressed moderate levels of CCR5, approximately 90,000 molecules of CCR5/cell (data not shown). All infections were performed by spinoculation (28), as described below. Preparation of the IS cell line. CEM-SS cells were infected with pseudotyped virus (described above) by spinoculation. Briefly, 2 ⫻ 107 CEM-SS cells were mixed with 10 ml of viral stock (45 ng of p24Gag/ml) and placed into all six wells of a flat-bottom six-well tissue culture plate (1.6 ml of suspension per well). The plates were sealed in plastic bags and centrifuged in microplate carriers at 1,200 ⫻ g for 2 h at 25°C. Cells were collected and washed once with 50 ml of ice-cold culture medium (RPMI 1640 with 10% heat-inactivated fetal calf serum, 10 mM HEPES [pH 7.4], 1⫻ penicillin-streptomycin). Infected cells were then cultured for 2 days before the addition of 300 g of G418 (Geneticin; Invitrogen Life Technologies)/ml. Upon initial addition of G418, approximately 50% of the cells died. The cells were then selected with G418 for an additional 4 weeks. After selection, cells were lysed at a density of 106/ml in the following buffer: 10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.2 mM CaCl 2, 0.001% sodium dodecyl sulfate, 0.001% Triton X-100, 1 mg of proteinase K (Sigma)/ml. Lysates were digested overnight at 58°C, and then the protease was heat inactivated for 15 min at 95°C. Aliquots of IS (106/ml) were stored at ⫺80°C. Prior to assay, serial dilutions of the IS genomic DNA were made in 10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.001% sodium dodecyl sulfate, 0.001% Triton X-100 supplemented with 1 g of poly(rA)/ml, to reduce nonspecific adsorption of DNA to the walls of the reaction vessel. Southern blotting. After G418 selection, total DNA was prepared as described elsewhere (DNeasy tissue system; Qiagen) from the IS, from uninfected CEM-SS cells, and from CEM-SS cells that had been acutely infected with HIV-1IIIB (300 ng of p24Gag/ml) by spinoculation 20 h prior to harvest. After electrophoresis in a 0.8% agarose gel (loading 20 g of DNA per lane), the DNA was transferred to a charged nylon membrane (GeneScreen Plus; Perkin-Elmer). The membrane was prehybridized with PerfectHyb PLUS hybridization buffer (Sigma) and then probed with an [␣-32P]dCTP-labeled probe generated by random priming
(RadPrime kit; Invitrogen) the entire HIV-1IIIB genome. After washing, the blot was exposed to BioMax MR film (Kodak) in a BioMax TranScreen LE intensifying screen (Kodak) for 1 week at ⫺80°C. Two-step PCR amplification. The initial nonkinetic preamplification was performed on dilutions of the IS cells as well as dilutions of unknowns. It is essential to preamplify a series of dilutions of both the standards and unknowns to insure that PCR substrates are not limiting. Reactions in which these reagents are limiting lack a dose response compared to other dilutions in the standard curve. The sequences of the preamplification primers were as follows: genomic Alu forward, 5⬘-GCC TCC CAA AGT GCT GGG ATT ACA G-3⬘; and HIV-1 gag reverse, 5⬘-GCT CTC GCA CCC ATC TCT CTC C-3⬘. The reaction conditions were 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl 2, 1 mM concentration of mixed dNTPs, 50 mM KCl, 100 nM Alu forward primer, 600 nM gag reverse primer, and 0.05 U of Platinum Taq DNA polymerase (Invitrogen Life Technologies). The thermal cycler (DNA Engine PTC-200; MJ Research) was programmed to perform a 2-min hot start at 94°C, followed by 20 steps of the following: denaturation at 93°C for 0.5 min, annealing at 50°C for 1 min, and extension at 70°C for 1 min 40 s. Linear, one-way amplification was also monitored by performing the preamplification PCR with either the gag primer alone or the Alu primer alone. The second-round real-time quantitative PCR was performed using 25 l of the material from the preamplification or matched dilutions of both the nonpreamplified IS and the nonpreamplified unknowns (acutely infected cells in the presence and absence of integrase inhibitors). These were run with an HIV-1 copy number standard prepared from graded doses of ACH-2 cells. The sequences of the primers were as follows: LTR forward, 5⬘-GCC TCA ATA AAG CTT GCC TTG A-3⬘; and LTR reverse, 5⬘-TCC ACA CTG ACT AAA AGG GTC TGA-3⬘. The LTR molecular beacon probe, labeled at its 5⬘ terminus with the reporter fluorophore 6-carboxyfluorescein (FAM) and at its 3⬘ terminus with the quencher 4-(4⬘-dimethylamino-phenylazo)-benzene (DABCYL), had the following sequence: 5⬘-FAM-GCG AGT GCC CGT CTG TTG TGT GAC TCT
VOL. 76, 2002
QUANTITATION OF HIV-1 PROVIRAL DNA
10945
FIG. 3. The first regression, for provirus content of IS dilutions. An HIV-specific standard curve was generated using serial dilutions of genomic DNA from accurately counted ACH-2 cells. (A) Real-time monitoring of signals obtained from the ACH-2 standards. The known copy numbers in each dose of standard are shown over the corresponding amplification curves. 1E6 indicates 1,000,000 copies; 2E5 indicates 200,000 copies, etc. (B) Logarithmic regression to generate a line of best fit. Points outside the linear range are shown to illustrate the sigmoidal shape of the overall data set.
GGT AAC TAG CTC GC-DABCYL-3⬘. Reactions were carried out in a volume of 50 l containing 10 mM Tris-HCl (pH 8.3), 75 mM KCl, 4.25 mM MgCl2, 500 nM carboxy-X-rhodamine (ROX; Molecular Probes) as a passive reference, 1.2 mM concentration of freshly added dNTPs, 250 nM concentration of LTR forward and reverse primers; 200 nM molecular beacon probe; and 0.025 U of Platinum Taq DNA polymerase. The reactions were usually performed on a Prism 7700 sequence detection system running Sequence Detector version 1.6.3 software (Applied Biosystems). The thermal program was 2-min hot start at 95°C, followed by 40 cycles of denaturation at 95°C for 15 s and annealing and extension at 60°C for 1 min. Similar results were obtained on a DNA Engine Opticon instrument (MJ Research) running Opticon Monitor version 1.1 software with the following thermal program: 2-min hot start at 95°C, followed by 40 cycles of denaturation at 93°C for 15 s, annealing at 50°C for 15 s, a plate read, and then extension at 72°C for 1 min. To express integration as a ratio of proviruses per target cell, a kinetic PCR assay for -globin DNA was used to determine cell numbers on the same plates
used for HIV-1 quantitation. A standard curve for cellular DNA was prepared by serially diluting a CEM-SS lysate starting at 4 ⫻ 106/ml (prepared as described above). The sequences of the -globin forward and reverse primers were 5⬘-CC CTTGGACCCAGAGGTTCT-3⬘ and CGAGCACTTTCTTGCCATGA-3⬘, respectively. The -globin molecular beacon sequence was 5⬘-FAM-GCGAGCA TCTGTCCACTCCTGATGCTGTTATGGGCGCTCGC-DABCYL-3⬘. React ions were carried out in a volume of 50 l containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 3.5 mM MgCl2, 500 nM ROX, 0.8 mM concentration of mixed dNTPs, 1 M concentration of forward and reverse -globin primers, 100 nM molecular beacon probe, and 0.025 U of Platinum Taq DNA polymerase. The thermal programs were described above.
RESULTS The IS contains only integrated HIV-1 DNA. To generate the IS cell line, spinoculation was utilized to deposit large
10946
O’DOHERTY ET AL.
J. VIROL.
FIG. 4. Second regression, for provirus content in unknowns. (A) The number of copies of HIV-1 DNA in dilutions of the nonpreamplified IS are shown over the corresponding amplification curves. These numbers were determined using the ACH-2 copy number standard (Fig. 3B). Since 100% of the HIV-1 DNA in the IS is integrated (Fig. 1), these numbers reflect the total number of proviruses in each reaction. (B) The number of proviruses per reaction in the nonpreamplified IS was assigned to the kinetic PCR signals from the preamplified IS, and a second standard line was constructed. These amplifications were performed in duplicate, with consistently close reproducibility. (C) Logarithmic regression of the preamplified IS signals. As few as eight proviruses per sample could be reliably measured in this run.
numbers of replication-defective recombinant viruses onto the surfaces of a large number of susceptible CEM-SS target cells (Fig. 1A), as described previously (28). Total DNA was prepared from the IS following prolonged antibiotic selection. In addition, DNA was prepared from CEM-SS cells that had been exposed to replication-competent HIV-1IIIB by spinoculation and cultured for 20 h, and from uninfected CEM-SS cells.
Equal amounts of each DNA preparation (20 g/lane) were applied to the lanes of an agarose gel (Fig. 1B). Southern blotting revealed integrated, linear, and circular forms of HIV-1 DNA in the acutely infected cells (Fig. 1C, left lane). However, only integrated HIV-1 DNA could be detected in the IS (Fig. 1C, center lane). The signals on the blot were HIV-1 specific, since they were absent from uninfected CEM-SS cells
VOL. 76, 2002
QUANTITATION OF HIV-1 PROVIRAL DNA
TABLE 1. Intra-assay variability of the two-stage Alu PCR amplificationa IS (proviruses/reaction)
Mean amplificationb (fold, n ⫽ 6 assays)
SD (fold)
% Variation in amplification
3,600 1,200 400 133 44 15 5
12,500 27,000 61,000 79,000 84,000 91,000 86,000
1,250 1,300 3,700 1,900 13,000 17,000 19,000
10 5 6 2 15 19 22
a Six samples of each dose of IS were independently subjected to Alu-gag preamplification followed by quantitative real-time PCR amplification of HIV-1 LTR DNA. Amplification values in boldface represent the linear range of this particular assay. b Amplification factors of ⬃10,000 are usual. An additional 10-fold variation in this value can be observed, depending on the particular lot of Platinum Taq DNA polymerase used.
(Fig. 1C, right lane). To ensure that all of the HIV-1 cDNA in the IS was integrated, the IS lysate was prepared after prolonged G418 exposure, thereby expanding a subpopulation of cells that expressed the neomycin resistance phenotype efficiently. Although this selection should reduce the randomness of the integration sites represented in the selected population to some degree, there should be no selection pressure in favor of Alu-gag distances of any specific length. Standardization of the assay. The initial nonkinetic Alu-gag preamplification, with substrates and enzyme in excess, produced heterogeneous amplicons of varying length (Fig. 2A). Aliquots of these amplicons, or matched aliquots of nonpreamplified standards and unknowns, were then subjected to a quantitative kinetic PCR that detects RU5 DNA within the HIV-1 LTR and produces homogeneous amplicons 100 bp in length (Fig. 2B). To standardize this assay, one must first determine the HIV-1 copy number in the nonpreamplified IS—that is, from IS DNA that has not been subjected to the initial nonkinetic Alu-gag PCR. To do so, we used graded doses of accurately counted ACH-2 cells (with two proviruses per genome). Two separate samples of this cell suspension were counted in trypan blue dye; four hemacytometer counts were performed on each sample, and then the eight cell counts were averaged. A kinetic PCR amplification for HIV-1 LTR sequences (Fig. 2B) was performed on this ACH-2 standard and on dilutions of the nonpreamplified IS. Because all of the HIV-1 DNA in the IS is integrated (above, Fig. 1C), the known HIV copy numbers in the ACH-2 standard can be used to measure the apparent integrated HIV-1 DNA content of the IS (Fig. 2C). IS contained a mean of 1.4 ⫾ 0.3 proviruses per cell. Next, a second standard curve was constructed. The apparent integrated HIV-1 copy numbers in the nonpreamplified IS dilutions were assigned to the kinetic PCR signals from the preamplified IS dilutions (Fig. 2D). This second standard curve is used to determine the number of proviruses per reaction in preamplified dilutions of unknowns. To the extent that they are present, nonintegrated HIV sequences in the unknowns are part of the bulk HIV copy number and can be amplified linearly by gag priming in the preamplification. To correct for these, the signal contributed from unintegrated DNA is rou-
10947
tinely subtracted from the total signal by including a gag-only linear preamplification control (as elaborated below in Fig. 5). In many cases, it is revealing to express integration rates as a ratio of the mean number of proviruses per target cell. To determine this ratio, the assay can be expanded to directly detect cell numbers. Kinetic PCR is performed for -globin DNA. The standard(s) used for this assay must match the unknowns in terms of ploidy. For example, if (euploid) primary T cells are being studied, an appropriate standard would be graded doses of genomic DNA from (euploid) peripheral blood mononuclear cells, while an infected aneuploid tumor cell line should be compared to a standard prepared from matched uninfected cells. Linearity and dynamic range of the assay. In a typical kinetic PCR amplification for bulk HIV-1 copy numbers, using the ACH-2 standard (Fig. 3A), sigmoidal amplification curves were observed between 20 and 1 million HIV-1 copies per reaction, for a dynamic range spanning 4.7 logs. When the decimal logarithm of the HIV-1 copy number was plotted against the threshold cycle number between these limits, a strongly linear relationship was observed (Fig. 3B). The regression line usually has a y intercept near 40 cycles (the maximum cycle number per reaction) and a negative slope, close to ⫺3 threshold cycles per 10-fold increase in analyte concentration. Saturation is observed at copy numbers higher than 1 million copies per well (107 cells/ml). No amplification curves were observed below the limit of detection—20 copies per well in this example (Fig. 3B)—or in no-template controls (data not shown). In each run, the ACH-2 standard curve (using primers that bind the R and U5 regions of the HIV-1 LTR) was used to determine experimentally the provirus content of a dilution series of nonpreamplified IS (Fig. 4A). These numbers were then assigned to the Alu-gag signals arising from a matching dilution series of preamplified IS (Fig. 4B). The first PCR serves as a preamplification of Alu-gag signals, reducing the threshold cycle number of preamplified IS in the kinetic PCR. This shifted the IS curves to the left in Fig. 4B by about 12 cycles, corresponding to roughly 4 logs of amplification. Observed amplification factors (average of 6 performed within one assay) are listed in Table 1. In this particular assay, as few as 5 proviruses in a 50-l sample were detectable. Since each sample can contain up to 50,000 cells, 1 provirus in 10,000 cells could be reliably measured in this run. This sensitivity should be sufficient to measure integration in primary lymphocytes (7, 29). The relative uncertainty increases with increasing dilution, likely due to accumulated sampling errors. Within a run, the assay is very reproducible: duplicate reactions on the same plate containing the same master mix typically overlay each other closely (Fig. 4B). Compared to the ACH-2 standard (Fig. 3A), the dynamic range of the preamplified IS is relatively narrow, usually about 2.5 logs. These limits are attributable to the sigmoidal amplification characteristics of the nonkinetic preamplification step (13, 32). Within this range, however, the standard curve is very linear (Fig. 4C). The y intercept for this regression line typically corresponds to threshold cycles in the mid-twenties, with a slope of roughly ⫺2.5 cycles per 10-fold increase in analyte concentration. Sensitivity and specificity of the assay. To determine whether linear, one-way amplification of unintegrated HIV-1
10948
O’DOHERTY ET AL.
J. VIROL.
VOL. 76, 2002
DNA contributed a significant interfering signal, we used pIIIB, a plasmid containing a molecular clone of HIV-1IIIB without any human sequences, as a model of unintegrated HIV-1 DNA (Fig. 5A). This construct lacks an Alu repeat, so when either the gag primer alone or the Alu-gag primer pair was used in the preamplification, similar signals were obtained, corresponding to roughly a fivefold amplification. As expected, when the Alu primer was used alone in the preamplification of pIIIB, a signal was obtained which was comparable to that of a nonpreamplified plasmid sample. Signals corresponding to up to 104 copies of unintegrated DNA can be tolerated by this assay, since they appear at higher threshold cycle numbers than the most dilute of the IS-derived preamplified standard curves. The sensitivity of the assay was further examined by performing one-way and bidirectional preamplifications of the IS, a heterogeneous template that contains exclusively integrated DNA (Fig. 5B). Again, one-way amplification contributed a low-level interfering signal corresponding to about fivefold amplification over the nonpreamplified standard. In contrast, bidirectional Alu-gag preamplification resulted in a 4 log net amplification (approximately a 12-cycle reduction in threshold cycle number). Taken together, these findings indicate that the interfering signal from one-way amplification is small and can be subtracted from the total signal obtained from any given sample. Accounting for this signal becomes important at early time points following acute in vitro HIV-1 infection, when proviruses comprise only a minority of the total HIV-1 DNA present. Next, we assayed acutely infected CEM-SS-R5 cells that had been cultured for 12 h after spinoculation (Fig. 5C). Again, we found that one-way amplification with the gag primer resulted in a small increase in signal above no preamplification. In this acutely infected sample, a 2.7 log increase in signal over no preamplification was observed with bidirectional Alu-gag amplification. This relatively modest shift is due to the predominance of unintegrated HIV-1 DNA at this early time point. Specifically, this sample had a mean of 2.24 ⫾ 0.06 copies of viral HIV-1 LTR DNA per cell, but only 0.11 ⫾ 0.04 proviruses per cell (n ⫽ 3). The assay is specific for integration. When CEM-SS-R5 cells are acutely infected in the presence of the diketo acid integrase inhibitor L-731,988 (Merck) (21), no integration is detectable (Fig. 5C). That is, only a fivefold bidirectional Alu-gag amplification was seen with L-731,988 inhibition, a signal similar to that of a gag-only preamplification. The signals obtained from
QUANTITATION OF HIV-1 PROVIRAL DNA
10949
infections performed using the HIV-1YU2-derived integrase mutant D64A were also dramatically reduced, again to a level similar to that of a gag-only preamplification (Fig. 5D). DISCUSSION Here we describe an assay for HIV-1 integration based on real-time quantitative Alu-gag PCR. This assay has sufficient sensitivity and specificity to permit its use in the study of nonproductive or latent HIV-1 infection in resting primary CD4⫹ T cells. Reservoirs of such cells accumulate early in the course of HIV-1 infection in vivo (4, 6–8, 17). The virus within these cells evades both antiviral immune responses and antiretroviral therapy, preventing the cure of HIV disease by current medications (9, 31, 33). Postintegration viral latency was first documented in untreated HIV-1-infected individuals. Proviruses were estimated to be present in only 1 (7) to 5 (29) of 10,000 resting CD4⫹ T cells from both blood and lymph nodes. However, due to inherent limitations in the methods used, most importantly the lack of a polyclonal standard (see above), these frequencies may well be underestimated. Infectious virus could only be recovered from these cells if they were treated with the mitogen phytohemagglutinin in the presence of antigen-presenting cells (7). Ominously, the frequency of provirus-containing resting CD4⫹ T cells did not appear to change substantially between early and advanced HIV disease and was not influenced by effective therapy that reduced plasma viremia to undetectable levels (8, 15). This finding has not yet been confirmed using a fully quantitative integration assay. Most, but not all, of the resting cells that contained proviruses were of the memory phenotype (29). However, the fact that any of them were of the naïve phenotype casts doubt upon the dogma that full, mitotic T-cell activation is required for integration to occur. The mechanism by which postintegration latency is established and maintained is not known and, in naïve cells, may be novel. Multiple steps in the HIV-1 life cycle are inefficient or inhibited in resting CD4⫹ T cells. It is now well established that a major inefficiency exists in reverse transcription in these cells (4, 24, 35, 38, 39). A paucity of substrates for reverse transcriptase is in part responsible for this block, since the inefficiency can be partially overcome by supplementing the resting cells with exogenous deoxyribonucleosides (23). Yet additional inefficiencies must exist downstream of reverse transcription,
FIG. 5. Sensitivity and specificity of the assay. (A) The plasmid pIIIB contains a full-length molecular clone of HIV-1IIIB and no human DNA (in particular, no Alu sequences). One-way preamplification with the Alu primer alone (orange) produces a kinetic PCR signal comparable to the signal from a matched dose of nonpreamplified plasmid (green). Signals roughly fivefold higher arise from either one-way preamplification with the gag primer alone (black) or bidirectional preamplification with both Alu and gag primers. (B) One-way preamplification of the IS using either the Alu primer alone (orange) or the gag primer alone (black) produces roughly a fivefold amplification relative to a matched sample of nonpreamplified IS (green). In contrast, the kinetic PCR signal obtained from the IS following bidirectional Alu-gag preamplification (blue) is shifted roughly 12 threshold cycles to the left, corresponding to a 4-log amplification over the nonpreamplified control. (C) Twelve hours after spinoculation of CEM-SS-R5 cells with HIV-1YU-2 (1,000 ng of p24Gag/ml), proviruses comprise only a minority of HIV-1 DNA. The integration signal from the acutely infected cells (red) falls within the linear range of the preamplified IS (blue curves, corresponding to 3,300, 1,000, 330, 100, 30, and 12 proviruses per reaction). One-way preamplification with the gag primer alone (black) generates a signal that is similar to the signal obtained when the infection is performed in the presence of a 10 M dose of the integrase inhibitor L-731,988 (gray). This signal can be subtracted from the Alu-gag signal. It is produced via linear amplification of both integrated and unintegrated HIV-1 DNA. (D) Integration is not detected 12 h after spinoculation of CEM-SS-R5 cells with the HIV-1YU2-derived integrase-defective mutant virus D64A. Instead, Alu-gag preamplification results in a fivefold amplification, similar to one-way gag amplification.
10950
O’DOHERTY ET AL.
since progeny virions are not released by these latently infected cells in the absence of immune activation (4, 7, 10), even when their nucleoside pools are exogenously enhanced (23). The quantitative integration assay we have developed should prove useful in mechanistic studies to dissect these additional inefficiencies. It should also prove applicable to clinical and epidemiologic studies of viral integration and assessments of the effectiveness with which future antiretroviral treatments eradicate latent reservoirs of HIV-1. ACKNOWLEDGMENTS This study was supported in part by NIH grants K08 HL03984 (U.O.), K08 AI50458 (W.J.S.), and R01 AI46942 (M.H.M.). Additional support came from a Developmental Core Grant from the University of Pennsylvania Center for AIDS Research to W.J.S. (1P30 AI 45008 to 02), and from a Research Grant from the W. W. Smith Charitable Trust to W.J.S. (A0102). M.H.M. is an Elizabeth Glaser Scientist supported by the Elizabeth Glaser Pediatric AIDS Foundation. The integrase inhibitor L-731,988 was a generous gift of Merck Research Laboratories, West Point, Pa. The plasmids pIIIb (also known as pNG38) and pYU-2/IN (D64A, also known as pJD170) were gifts from Nathan Gaddis and Jeff Dvorin, respectively. We are grateful to Theodore Pierson for his insightful suggestions. REFERENCES 1. Bell, P., L. J. Montaner, and G. G. Maul. 2001. Accumulation and intranuclear distribution of unintegrated human immunodeficiency virus type 1 DNA. J. Virol. 75:7683–7691. 2. Benkirane, M., P. Corbeau, V. Housset, and C. Devaux. 1993. An antibody that binds the immunoglobulin CDR3-like region of the CD4 molecule inhibits provirus transcription in HIV-infected T cells. EMBO J. 12:4909– 4921. 3. Bouyac-Bertoia, M., J. D. Dvorin, R. A. M. Fouchier, Y. Jenkins, B. E. Meyer, L. I. Wu, M. Emerman, and M. H. Malim. 2001. HIV-1 infection requires a functional integrase NLS. Mol. Cell 7:1025–1035. 4. Bukrinsky, M. I., T. L. Stanwick, M. P. Dempsey, and M. Stevenson. 1991. Quiescent T lymphocytes as an inducible virus reservoir in HIV-1 infection. Science 254:423–427. 5. Butler, S. L., M. S. T. Hansen, and F. D. Bushman. 2001. A quantitative assay for HIV DNA integration in vivo. Nat. Med. 7:631–634. 6. Chun, T.-W., L. Carruth, D. Finzi, X. Shen, J. A. DiGiuseppe, H. Taylor, M. Hermankova, K. Chadwick, J. Margolick, T. C. Quinn, Y.-H. Kuo, R. Brookmeyer, M. A. Zeiger, P. Barditch-Crovo, and R. F. Siliciano. 1997. Quantification of latent tissue reservoirs and total body viral load in HIV-1 infection. Nature 387:183–188. 7. Chun, T.-W., D. Finzi, J. Margolick, K. Chadwick, D. Schwartz, and R. F. Siliciano. 1995. In vivo fate of HIV-1-infected T cells: quantitative analysis of the transition to stable latency. Nat. Med. 1:1284–1290. 8. Chun, T. W., D. Engel, M. M. Berrey, T. Shea, L. Corey, and A. S. Fauci. 1998. Early establishment of a pool of latently infected, resting CD4⫹ T cells during primary HIV-1 infection. Proc. Natl. Acad. Sci. USA 95:8869–8873. 9. Chun, T. W., and A. S. Fauci. 1999. Latent reservoirs of HIV: obstacles to the eradication of virus. Proc. Natl. Acad. Sci. USA 96:10958–10961. 10. Chun, T. W., L. Stuyver, S. B. Mizell, L. A. Ehler, J. A. Mican, M. Baseler, A. L. Lloyd, M. A. Nowak, and A. S. Fauci. 1997. Presence of an inducible HIV-1 latent reservoir during highly active antiretroviral therapy. Proc. Natl. Acad. Sci. USA 94:13193–13197. 11. Clouse, K. A., D. Powell, I. Washington, G. Poli, K. Strebel, W. Farrar, B. Barstad, J. Kovacs, A. S. Fauci, and T. M. Folks. 1989. Monokine regulation of HIV-1 expression in a chronically infected human T-cell clone. J. Immunol. 142:431–438. 12. Courcoul, M., C. Patience, F. Rey, D. Blanc, A. Harmache, J. Sire, R. Vigne, and B. Spire. 1995. Peripheral blood mononuclear cells produce normal amounts of defective Vif⫺ human immunodeficiency virus type 1 particle which are restricted for the preretrotranscription steps. J. Virol. 69:2068– 2074. 13. Crotty, P. L., R. A. Staggs, P. T. Porter, A. A. Killeen, and R. C. McGlennen. 1994. Quantitative analysis in molecular diagnostics. Hum. Pathol. 25:572– 579. 14. Deininger, P. L., and M. A. Batzer. 1999. Alu repeats and human disease. Mol. Genet. Metab. 67:183–193. 15. Dyrhol-Riise, A. M., P. Voltersvik, O. G. Berg, J. Olofsson, S. Kleivbo, and B. Asjo. 2001. Residual human immunodeficiency virus type 1 infection in lymphoid tissue during highly active antiretroviral therapy. AIDS Res. Hum. Retrovir. 17:577–586.
J. VIROL. 16. Englund, G., T. S. Theodore, E. O. Freed, A. Engelman, and M. A. Martin. 1995. Integration is required for productive infection of monocyte-derived macrophages by human immunodeficiency virus type 1. J. Virol. 69:3216– 3219. 17. Finzi, D., M. Hermankova, T. Pierson, L. M. Carruth, C. Buck, R. E. Chaisson, T. C. Quinn, K. Chadwick, J. Margolick, R. Brookmeyer, J. Gallant, M. Markowitz, D. D. Ho, D. D. Richman, and R. F. Siliciano. 1997. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science 278:1295–1300. 18. Folks, T., J. Justement, A. Kinter, C. A. Dinarello, and A. S. Fauci. 1987. Cytokine induced expression of HIV-1 in a chronically infected promonocytic cell line. Science 238:800–802. 19. Folks, T. M., D. Powell, M. Lightfoote, S. Koenig, A. S. Fauci, S. Benn, A. Rabson, D. Daugherty, H. E. Gendelman, M. D. Hoggan, S. Venkatesan, and M. A. Martin. 1986. Biological and biochemical characterization of a cloned leu-3⫺ cell surviving infection with the acquired immune deficiency syndrome retrovirus. J. Exp. Med. 164:280–290. 20. Fouchier, R. A. M., B. E. Meyer, J. H. M. Simon, U. Fischer, and M. H. Malim. 1997. HIV-1 infection of non-dividing cells: evidence that the aminoterminal basic region of the viral matrix protein is important for Gag processing but not for post-entry nuclear import. EMBO J. 16:4531–4539. 21. Hazuda, D. J., P. Felock, M. Witmer, A. Wolfe, K. Stilmock, J. A. Grobler, A. Espeseth, L. Gabryelski, W. Schleif, C. Blau, and M. D. Miller. 2000. Inhibitors of strand transfer that prevent integration and inhibit HIV-1 replication in cells. Science 287:646–650. 22. Jelinek, W. R., and C. W. Schmid. 1983. Repetitive sequences in eukaryotic DNA and their expression. Annu. Rev. Biochem. 51:813–844. 23. Korin, Y. D., and J. A. Zack. 1999. Nonproductive human immunodeficiency virus type 1 infection in nucleoside-treated G0 lymphocytes. J. Virol. 73: 6526–6532. 24. Korin, Y. D., and J. A. Zack. 1998. Progression to the G 1b phase of the cell cycle is required for completion of human immunodeficiency virus type 1 reverse transcription in T cells. J. Virol. 72:3161–3168. 25. Malim, M. H., and B. R. Cullen. 1993. Rev and the fate of pre-mRNA in the nucleus: implications for the regulation of RNA processing in eukaryotes. Mol. Cell. Biol. 13:6180–6189. 26. Mighell, A. J., A. F. Markham, and P. A. Robinson. 1997. Alu sequences. FEBS Lett. 417:1–5. 27. Nara, P. L., W. C. Hatch, N. M. Dunlop, W. G. Robey, L. O. Arthur, M. A. Gonda, and P. J. Fischinger. 1987. Simple, rapid quantitative, syncytiumforming microassay for the detection of human immunodeficiency virus neutralizing antibody. AIDS Res. Hum. Retrovir. 3:283–302. 28. O’Doherty, U., W. J. Swiggard, and M. H. Malim. 2000. Human immunodeficiency virus type 1 spinoculation enhances infection through virus binding. J. Virol. 74:10074–10080. 29. Ostrowski, M. A., T.-W. Chun, S. J. Justement, I. Motola, M. A. Spinelli, J. Adelsberger, L. A. Ehler, S. B. Mizell, C. W. Hallahan, and A. S. Fauci. 1999. Both memory and CD45RA⫹/CD62L⫹ naive CD4⫹ T cells are infected in human immunodeficiency type 1-infected individuals. J. Virol. 73:6430–6435. 30. Pauza, C. D., J. E. Galindo, and D. D. Richman. 1990. Reinfection results in accumulation of unintegrated viral DNA in cytopathic and persistent human immunodeficiency virus type 1 infection of CEM cells. J. Exp. Med. 172: 1035–1042. 31. Pierson, T., J. McArthur, and R. F. Siliciano. 2000. Reservoirs for HIV-1: mechanisms for viral persistence in the presence of antiviral immune responses and antiretroviral therapy. Annu. Rev. Immunol. 18:665–708. 32. Raeymaekers, L. 2000. Basic principles of quantitative PCR. Mol. Biotechnol. 15:115–122. 33. Schrager, L. K., and M. P. D’Souza. 1998. Cellular and anatomic reservoirs of HIV-1 in patients receiving potent antiretroviral combination therapy. JAMA 280:67–71. 34. Sonza, S., A. Maerz, N. Deacon, J. Meanger, J. Mills, and S. Crowe. 1996. Human immunodeficiency virus type 1 replication is blocked prior to reverse transcription and integration in freshly isolated peripheral blood monocytes. J. Virol. 70:3863–3869. 35. Spina, C. A., J. C. Guatelli, and D. D. Richman. 1995. Establishment of a stable, inducible form of human immunodeficiency virus type 1 DNA in quiescent CD4 lymphocytes in vitro. J. Virol. 69:2877–2988. 36. Tyagi, S., and F. R. Kramer. 1996. Molecular beacons: probes that fluoresce upon hybridization. Nat. Biotechnol. 14:303–308. 37. Vandegraaff, N., R. Kumar, C. J. Burrell, and P. Li. 2001. Kinetics of human immunodeficiency virus type 1 (HIV-1) DNA integration in acutely infected cells as determined using a novel assay for detection of integrated HIV DNA. J. Virol. 75:11253–11260. 38. Zack, J. A., S. J. Arrigo, S. R. Weitsman, A. S. Go, A. Haislip, and I. S. Y. Chen. 1990. HIV-1 entry into quiescent primary lymphocytes: molecular analysis reveals a labile, latent viral structure. Cell 61:213–222. 39. Zack, J. A., A. M. Haislip, P. Krogstad, and I. S. Y. Chen. 1992. Incompletely reverse-transcribed human immunodeficiency virus type 1 genomes in quiescent cells can function as intermediates in the retroviral life cycle. J. Virol. 66:1717–1725.
