Single-Copy Quantification of HIV-1 in Clinical Samples

Abstract

HIV replication in humans proceeds with substantial viral RNA levels in plasma. Antiretroviral therapy results in suppression but not eradication of HIV infection. Continuous therapy is essential for durable clinical responses. Discontinuing antiretroviral therapy results in prompt rebound in viremia. The source of HIV during suppressive therapy and mechanisms of persistence remain uncertain. Sensitive assays for HIV have been useful in quantifying viremia in response to antiretroviral therapy and in experimental studies of drug intensification, drug simplification, and potential anatomic sanctuary site investigations. As clinical eradication strategies move forward, robust, sensitive quantitative assays for HIV at low levels represent essential laboratory support modalities. Here we describe in detail an assay for HIV-1 RNA with single-copy sensitivity.

1. Introduction

Quantitative nucleic acid detection assays have become an important and useful application of real-time PCR techniques, resulting in dramatic advances in molecular diagnostics. Real-time assays have been developed to detect a wide variety of targets, including mutations in neoplastic cells, organisms with pathogenic potential for use in clinical medicine [1–6]. HIV RNA detection in plasma has enabled a new understanding of HIV pathogenesis, and has since revolutionized clinical care of HIV infected individuals [7–9]. A variety of HIV detection methods have been developed for commercial use with progressively more sensitive lower limits of detection of 50, 40, or 20 copies/ml [10–15]. These assays have been validated for clinical use, and excellent clinical studies demonstrate the clinical benefit of suppression <50 copy limit.

In the course administering antiretroviral therapy and measuring infectious HIV or levels of HIV RNA in plasma, it became clear that ARV therapy was effective in suppressing, but not eliminating HIV infection [16, 17]. Interruption of antiretroviral therapy results in prompt and brisk rebound in viremia in the majority of patients [18]. The source of HIV during suppressive antiretroviral therapy is unknown, but may be the result of active cycles of HIV replication or production from long lived reservoirs of cells with integrated proviruses [19]. Distinguishing between production from active spreading replication and production from reservoirs of infected cells is critical in determining future efforts to eliminate or functionally cure HIV infection. If persistent infection is the result of ongoing cycles of HIV infection, then the development of new antiretrovirals capable of efficiently penetrating sanctuary sites of HIV replication will be an essential component of HIV therapy. Alternatively, if HIV persists because of production from long-lived reservoirs of cells with integrated proviruses, then new strategies to eliminate such reservoirs will be essential.

To investigate the source of HIV during suppressive therapy, new assays to quantitate HIV at levels below that detected by commercial assays were essential. Several assays detect HIV in the range of 6–12 copies/ml have been used to demonstrate that HIV infected individuals undergoing suppressive antiretroviral therapy remain persistently viremic, although all patients were not detectably viremic at all time points, suggesting that episodic HIV production took place, or that HIV was present at levels lower than the lower limit of detection of these ultrasensitive assays [16, 20]. In order to investigate HIV viremia at these relatively low levels during suppressive antiretroviral therapy, we applied real-time PCR technology to develop a new assay for HIV RNA capable of detecting a single copy of HIV RNA/ml plasma [21]. Clinical studies of low level viremia with single-copy assays have identified median level of persistent viremia is approximately a single copy [22]; level of viremia during suppressive therapy remains stable after a slow decline [23] is likely determined by a number of factors including pretherapy level of HIV, but not CD4 count, or increase in CD4 count during therapy [22]. Single-copy assays have been useful in determining the effects of antiretroviral drug intensification on the level of viremia in virally suppressed patients, and a number of studies have revealed no effect of intensification on the levels of viremia, suggesting that persistent virus in plasma is not derived from the new cycles of ongoing replication [24–27]. Using analogous PCR technology, additional studies identified an increase in the presence of 2-LTR circles in a subset of suppressed patients intensified with raltegravir [28] suggesting the presence of ongoing cycles of infection. Single-copy assays have also been useful in investigating HIV RNA in compartments, such as the cerebrospinal fluid (CSF) [29]; and in studies of HIV recombination [30], drug switching, drug simplification [31, 32], and reservoir activation [33].

Several features ensured sensitive, robust detection of HIV RNA:

1. Use of a well-conserved region of HIV for quantification. Requirements for primer and probe hybridization in real-time assays require high degree of sequence identity with few or no nucleotide mismatches. The region chosen for amplification described here is within HIV gag gene. The single-copy assay was initially developed for subtype B HIV, and application to non-B subtypes requires specific attention to sequence variation.

2. Concentration of HIV from plasma prior to extraction and amplification. During persistent infection, HIV is present at relatively low copy numbers ranging from <0.2 to 40 copies/ml. We concentrate HIV using ultracentrifugation. In initial studies we determined that we could obtain reproducibly amplifiable material from as much as 10 ml of EDTA plasma. Use of heparinized plasma or serum results in variable loss of starting material. In initial experiments we determined that the amplification signal in plasma was entirely pelletable by ultracentrifugation as described below; An assay of the supernatants revealed no HIV RNA, indicating that the signal is derived from particle associated (and not free) HIV RNA.

3. Use of an internal control to ensure sample recovery. Singlecopy assay requires a number of manipulations prior to amplification. Samples must be concentrated, and starting material is typically 7–10 ml of plasma. RNA extraction and cDNA synthesis require additional manipulations. In order to monitor whether sample loss has occurred during preparation, we add an internal standard to the plasma prior to any manipulation. RCAS is an RNA containing avian retrovirus with physical properties similar to HIV. After centrifugation and extraction, the avian retrovirus is amplified from an aliquot of an independent real-time amplification for RCAS. If sample loss or other error occurs during the centrifugation, extraction, or cDNA synthesis procedures, it will be reflected in a loss of RCAS copy number.

4. Controlled cDNA synthesis. In order to enumerate viral RNA, it must first be quantitatively converted into cDNA in preparation for PCR amplification. To ensure that cDNA synthesis yields only a single DNA molecule from each RNA, we employ reverse transcriptase that does not have functional RNase H, and thus no degradation of RNA occurs after first round synthesis, and multiple DNA copies are not generated during cDNA synthesis.

5. Standardization of HIV standards. The amplification profiles must be compared to a “gold standard” viral stock with known copy number. For this purpose we initially standardized the single-copy assay using samples obtained from the AIDS Clinical Trial Group (ACTG) Viral Quality Assessment (VQA) program, which maintains HIV stocks with known copy number determined independently using electron microscopic particle counting techniques [21].

6. Use of control amplifications to rule out artifactual amplification. During initial sample processing, the presence of any extraneous cellular material containing HIV DNA can subsequently be amplified as part of the real-time amplification. To ensure that the final HIV signal we observe is from RNA and not DNA, we carry out a series of “no-RT controls” (NoRT). Samples are processed as described, but RT is omitted from the cDNA synthesis step. Therefore, any amplification of HIV observed during real-time amplification must be the result of contaminating DNA, and not from HIV RNA. Similarly, additional control amplifications containing only buffer and no sample material (do template controls, NTC) are included to determine whether observed amplification signal is derived from contamination in any of the reagents used in the assay (see Note 1).

2. Materials

2.1. Preparation of Standards for HIV-1 RNA Quantification

Plasmid pQP1, which contains a 1,420-bp fragment of HIV-1 HXB2 gag (SacI to BglII) downstream of the T7 promoter, was linearized with EcoRI, purified with a PCR purification kit (Qiagen, Valencia, Calif.), and transcribed with T7 RNA polymerase by using the RiboMax Express Large Scale RNA production system (Promega, Madison, WI) The template DNA was degraded with 5 U of RNase-free DNase, and the RNA transcripts were purified twice with an RNeasy kit (Qiagen). The RNA was quantified spectrophotometrically at 260 nm, diluted to 10⁶ copies/μl, divided into aliquots, and stored at −80 °C. Diluted HIV-1 transcripts were used to generate a standard curve for each single-copy assay. In general transcripts used for this assay are relatively stable for > 1 year when stored at −80 °C at high copy number (>10⁸ copies/μl).

2.2. Preparation of Standards for RCAS Quantification

A region of RSV gag was amplified from RCASBP(A)gfp plasmid (a gift of S. Hughes, National Cancer Institute) by using primers 1661F (CATT GACT GCTTTAGGCAGA) and 2215R (AACAGCGCGGTGATATAC) The resulting 554-bp PCR product was cloned into the vector pPCR-Script Amp (Stratagene, La Jolla, CA). The plasmids were purified on columns with a Qiagen kit and linearized with EcoRI. The resulting DNA template was quantified and then transcribed with T3 RNA polymerase by using the MAXIscript T3 kit (Ambion, Austin, TX). The RNA product was diluted and stored at −80 °C. Diluted RCAS transcripts were used to generate a standard curve for each RCAS real-time PCR assay.

2.3. RCAS Internal Virion Standard

High-titer RCAS stocks were obtained by transfecting DF-1 cells with RCASBP(A)gfp plasmid and harvesting the supernatant at the peak of production, as determined by measuring viral RT activity. RCAS stocks were diluted to 1.5 × 10⁶ copies/ml of viral RNA (as measured by the real-time RT-PCR assay) with RPMI 1640 medium containing 5 % fetal calf serum. A total of 200 μl of the diluted RCAS stock was added to each plasma sample as an internal control for virion recovery following centrifugation, RNA extraction, and determination of RT-PCR efficiency.

3. Methods

3.1. Real-Time PCR Detection of HIV RNA

3.1.1. Viral RNA Extraction

All procedures are carried out at room temperature unless otherwise noted. For samples expected to have <1,000 copies of HIV-1 RNA per ml, 7 ml of patient plasma to which 200 μl of RCAS stock (300,000 copies of RNA) had been added was ultracentrifuged at 170,000 ×g for 30 min at 4 °C in a Sorvall T-1270 rotor. For samples expected to have HIV-1 RNA concentrations >1,000 copies/ml, a smaller volume of plasma (typically 1–2 ml) was diluted to 7 ml with Tris-buffered saline and processed in the same manner described above for 7 ml of plasma. After ultracentrifugation, the supernatant was removed and the virion pellet was treated with 100 μl of 5 mM Tris-HCl (pH 8.0) containing 10 μl of proteinase K for 30 min at 55 °C. The pellet was then treated with 325 μl of 5.8 M guanidinium isothiocyanate containing 10 μl of glycogen, and the lysate was transferred to a microcentrifuge tube. After the addition of 500 μl of 100 % isopropanol, the lysate was centrifuged at 21,000 ×g for 15 min. The pellet was washed with 70 % ethanol, and the nucleic acids were resuspended in 55 μl of 5 mM Tris-HCl (pH 8.0) containing 1 μM dithiothreitol and 1,000 U of an RNase inhibitor (RNase out; Invitrogen, Carlsbad, CA). Control experiments showed that the recovery of HIV-1 and RCAS RNA was independent of the plasma volume over the range of volumes used (1–7 ml). For each sample, three separate 10-μl aliquots of the resuspended nucleic acids were used for the real-time PCR for HIV-1 and two separate 5-μl aliquots were used for the real-time PCR for RCAS.

3.1.2. Real-Time PCR Assay

The real-time assay involved a two-step, two-enzyme RT-PCR protocol.

3.1.3. cDNA Synthesis

Reverse transcription reactions (30 μl) were performed in 96-well plates and contained the following components at the indicated final concentrations or amounts in sterile molecular-grade water: random hexamers (0.15 μg/reaction; Promega), MgCl₂ (5 mM), deoxynucleoside triphosphates (0.5 mM), RNase out (20 U), dithiothreitol (0.67 mM), Taqman buffer A (Applied Biosystems, Foster City, cA) diluted to 1×, and RT (20 U of Superscript II RT; Invitrogen). After 15 min at 25 °C, the reverse transcription reaction mixture was incubated at 42 °C for 40 min. Following completion of the reverse transcription step, the reaction mixture was heated to 85 °C for 10 min and then held at 25 °C for 30 min, after which the plate was cooled to 4 °C.

3.1.4. Real-Time PCR

Twenty microliters of the PCR mixture was added to the cDNA reaction products (final volume, 50 μl) containing the following components at the indicated final amounts or concentrations in sterile molecular-grade water: PCR buffer II (Applied Biosystems) diluted to 1×, MgCl₂ (4 mM), AmpliTaq Gold (1.25 U, Applied Biosystems), and the primer-probe set for HIV-1 quantification designed to bind to a conserved region of gag: primers 6 F (5’-CATGTTTTCAGCATTATCAGAAGGA-3’) and 84R (5’-TGCTTGATGTCCCCCCACT-3’) (600 nM) and probe 5’ FAM-CCACCCCACAAGATTTAAACACCAT GCTAA-Q 3’ (100 nM), where FAM indicates a reporter 6-carboxyfluorescein group and Q indicates a 6-carboxytetramethylrhodamine group quencher conjugated through a linker arm nucleotide. The primer-probe set used for internal standard quantification was selected to bind to a conserved region of RSV gag: primers 1849F (5’-GTCAATAGAGAGAGAGGGATGGACAAA-3’) and 1896R (5’-TCCACAAGTGTAGCAGAGCCC-3’) (600 nM) and probe 5’FAM-TGGGTCGGGTGGTCGTGCC-Q 3’ (100 nM).

Following thermal activation of the AmpliTaq Gold (95 °C, 10 min), 45 cycles of PCR amplification (with each cycle consisting of 95 °C for 15 s and 60 °C for 1 min) were performed. For each run, two standard curves, one from diluted HIV-1 transcripts and one from diluted RCAS transcripts, were generated. For each experiment, new dilutions were prepared from thawed single-use aliquots of transcript stocks stored at −80 °C. A series of controls were added to each plate as described above, including no-template controls and no-RT controls. A typical plate setup (Fig. 1) can assay samples from seven patients.

Fig. 1.

Typical run results of patient samples (red) plotted on standard curve of HIV standards (black). Data demonstrate that the cycle threshold for real-time amplification (y axis) is a linear function of the log₁₀ value of the RNA copy number (x axis). Correlation coefficients for standard curves are typically >0.99

Threshold cycle (C_t) values from the real-time run were plotted as a function of the input transcript copy number, and linear regression was performed with LightCycler 480 Detection software (Roche Diagnostics). For each specimen, three replicate reactions were performed for HIV-1 quantification, with 18 % of the total RNA extracted from the original sample used in each reaction. Two replicate reactions were performed for internal standard RCAS quantification, with 9 % of the total RNA extracted from the original sample used in each reaction. One reaction mixture for each specimen was processed and amplified without the addition of RT as a control to detect HIV-1 DNA in the source specimen. Additional controls lacking RNA templates were run for both the HIV-1 and the RCAS reactions to test for contamination with the PCR product during sample processing or assay setup. The numbers of copies of HIV-1 and RCAS RNA in the test samples were calculated by interpolation of the experimentally determined C_t value for the test sample by using the transcript-derived linear regression as a standard curve and were rounded to the nearest integer value. For HIV-1 quantification, the calculated number of copy equivalents per reaction mixture was expressed as the number of copies per milliliter of the starting plasma sample. Assay acceptability was contingent on the R² value for the HIV-1 and RCAS linear regressions, which was >0.95, and the average measured copy equivalent per reaction mixture for the RCAS internal standard was greater than 15,000 copies (55 % recovery). Acceptance of the results for samples with RCAS levels greater than 15,000 copies ensured that the measured HIV-1 RNA levels were within a factor of 2 of the actual HIV-1 RNA levels. RCAS levels were >15,000 copies for 95 % of the plasma samples assayed. Negative and positive (12.5 copies/reaction mixture) plasma sample controls were included with each assay run. A typical analysis yields a standard curve and amplifications from experimental samples (Fig. 2). At high dilution (<10 copies) it should be noted that detection by any assay is limited by Poisson statistics, which has strong implications for assessing HIV viremia at low levels. Consider estimating the observed number of copies (x) given the true known number of copies (μ); the probability of observing x given μ is represented as p(x;μ) and is given by the Poisson relation: $p(x;\mu )=\frac{{\mu }^x{e}^{-\mu }}{x!}$. This probability distribution has the useful property that the variance of μ = μ itself; the Poisson distribution is also skewed at low μ such that the absolute reproducible precise possible at high viral RNA levels is not possible at one to five copies. Thus the lowest standards (1, 3 copy standards) are expected to yield a result of 0 copies in a significant proportion of the replicates (see Note 2).

Fig. 2.

A typical 96-well plate setup for single-copy assay. Well arrangements include the presence of no-template controls (NTC), no-RT controls (NoRT), HIV standards from 0.3 to 10⁶ copies/ml, internal controls (RCAS), and samples