- Email: [email protected]
A quantitative PCR method for the assay of HIV-1 provirus load in peripheral blood mononuclear cells
Journal of Virological Methods 83 (1999) 11 – 20 www.elsevier.com/locate/jviromet
A quantitative PCR method for the assay of HIV-1 provirus load in peripheral blood mononuclear cells Julie M. Bennett *, Steve Kaye, Neil Berry 1, Richard S. Tedder Department of Virology, Royal Free and Uni6ersity College London Medical School, 46 Cle6eland Street, London W1P 6DB, UK Received 30 April 1999; received in revised form 4 June 1999; accepted 11 June 1999
Abstract The use of high activity antiretroviral therapies (HAART) to treat HIV-infected patients frequently results in the long-term suppression of plasma virus RNA loads below levels detectable by current assays. The measurement of provirus DNA load in peripheral blood mononuclear cells provides a means of continuing to monitor the efficacy of treatment and the decline in reservoirs of latent virus. A quantitative PCR assay was developed for HIV-1 provirus using a three-point internal calibrator system to give high reproducibility and accuracy at the low copy numbers of provirus seen in clinical samples. Provirus DNA copies are related to cell number in the samples using a fluorescent dye-binding assay for measurement of input DNA. The assay agreed closely with an end-point dilution PCR and gave accurate quantification of extracts from an HIV-1 infected continuous cell line containing known provirus copy numbers. The inclusion of a second primer set in the LTR region of the HIV-1 genome, optimised to non-clade-B virus strains improved the detection and quantification of samples from patients infected with genetically divergent virus strains. Application of the assay to clinical trial patients showed no relationship between changes in provirus DNA loads and plasma virus RNA and changes in provirus load over 24 weeks were small. © 1999 Elsevier Science B.V. All rights reserved. Keywords: HIV-1; Provirus load; Quantitative PCR
1. Introduction The measurement of HIV-1 cell-free virus RNA load in plasma has become an important diagnos* Corresponding author. Tel.: +44-171-504-9490; fax: + 44-171-580-5896. E-mail address: [email protected] (J.M. Bennett) 1 Present address: Division of Retrovirology, National Institute for Biological Standards and Controls, Potters Bar, EN6 3QG, UK.
tic method for determining disease prognosis (Mellors et al., 1996) and assessing the efficacy of antiretroviral therapy in infected patients (Saag et al., 1996). In our experience and that of others, the use of high activity antiretroviral combination therapies (HAART) often results in plasma virus RNA loads falling below the level of detectability, even when using high sensitivity methods which can detect levels of 50 RNA copies/ml plasma or less (Collins et al., 1997; Sun et al., 1998).
0166-0934/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 0 9 3 4 ( 9 9 ) 0 0 0 9 6 - 8
12
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
To continue monitoring the efficacy of drug therapy in such patients, the assay of provirus DNA loads in peripheral blood mononuclear cells (PBMCs) may allow the continuing clearance of virus to be assessed. Studies of the dynamics of HIV-1 replication in infected patients and the rates of clearance of cell-free virus and integrated provirus have shown that although cell-free virus levels fall rapidly within a few days of commencing therapy, the decline of provirus loads proceeds much more slowly, and detectable provirus persists for months or years despite continuance of therapy (Finzi et al., 1997; Wong et al., 1997). To allow further studies of changes in provirus loads in both individual infected patients and patients enrolled in clinical trials of antiretroviral therapy, a method was developed for quantification of proviral genomes based on PCR amplification. The assay uses an internal calibration system described previously (Van Gemen et al., 1994) in which a standard curve is generated for each PCR amplified sample from three calibrators added to the reaction in the ratio 100:10:1. This system gives a high degree of reproducibility and accuracy in quantifying the low copy numbers of HIV-1 provirus found in PBMC samples. Quantitative detection of PCR products is made using enzymelabelled probes and a chemiluminescent substrate (Whitby and Garson, 1995) and sample input DNA is assayed by a fluorescent dye-binding method.
2. Methods and materials
2.1. Patient samples PBMCs were fractionated from whole blood samples collected into EDTA on Ficoll-Paque or from CPT collection tubes (Becton-Dickinson) and stored at − 70°C. Samples were collected from individual patients before and after undergoing various antiretroviral therapies, and from study subjects enrolled in the MRC Quattro Trial of four reverse transcriptase inhibitors used in combination (Kitchen, 1998).
2.2. DNA preparation DNA was extracted from the PBMCs from 1–2 ml blood using spin columns (QIAamp blood kit, Qiagen) and the final DNA extract was eluted into 50 ml of nuclease-free water. Routinely, 3 ml of this extract was added to a 50 ml PCR reaction volume. The concentration of DNA in the extract was estimated using a fluorometer (Ascent FL, Labsystems) after reacting the extract with Hoechst dye H33258. The DNA extract was diluted 1:50 in Tris/EDTA/NaCl (TEN) buffer and 150 ml reacted with 50 ml H33258 diluted to 5 mg/ml in TEN buffer for 5 min before reading (excitation 355 nm, emission 460 nm). Concentrations of DNA in the test samples were compared to a standard curve (32 ng/ml to 32 mg/ml in a halflog10 dilution series) made from calf thymus DNA (Sigma). To relate HIV-1 target copy number detected in the QPCR to input PBMC number it was assumed that a single PBMC yielded 6 pg DNA.
2.3. Preparation of oligonucleotide calibrators Three internal calibrators were prepared by nested PCR to incorporate unique probe binding sites and HIV-1 specific primer binding sites in a PCR product that was the same length (142 bp) as the PCR product amplified from the test samples. The method of preparation of the calibrators is shown in Fig. 1. A fX174 target (used as a random DNA sequence) was amplified in a nested PCR as shown, to incorporate the HIV-1 primer binding sites and internal probe sequence in a product of similar size and GC/AT ratio as the HIV-1 target sequence. The first round PCR reaction mix contained buffer, nucleotide and primer concentrations as shown below in Section 2.4 and approximately 1 pg fX174/Hae III molecular weight markers (Promega Life Sciences) as the PCR target. Reaction conditions were: four cycles 94°C 1 min, 37°C, 1 min, 72°C 1 min, six cycles 94°C 1 min, 50°C, 1 min, 72°C 1 min. Second round PCR reaction mix and conditions were identical to the first round other than the primers and the target which was a 1 ml transfer from the
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
first round reaction. Following PCR generation the calibrators were purified by agarose gel electrophoresis and further purified from the excised band with Geneclean (Anachem). After quantification (see below) and pooling in the ratio 100:10:1 copies of Ca, Cb and Cc respectively, the calibrators were stored at − 20°C in TE buffer containing 1 mg/ml herring sperm DNA at 1000× the working concentration. Calibrators were quantified by end-point dilution as described below, and by parallel line titration with DNA extracted from the 8E5 cell line, which is considered to contain a single proviral HIV-1 DNA copy per cell (Folks et al., 1986).
13
2.4. Quantitati6e PCR method DNA extracts (3 ml) were added to 50 ml PCR reactions and amplification was carried out in a hot-start PCR using AmpliTaq Gold (PE Biosystems) in order to reduce the generation of nonspecific products. Reaction mixes contained 10 mM Tris/HCl, pH 8.3, 50 mM KCl, 100 nM each of primers IEP1B, IEP2A-bio, LTRNB1, LTRNB2-bio (Table 1), 1.25 U AmpliTaq Gold, 200 mM each dNTP, 1.5 mM MgCl2 and 1 ml calibrator mix diluted to contain 1000, 100 and 10 DNA copies of Ca, Cb and Cc respectively. Reactions were amplified in a Perkin-Elmer 2400 thermal cycler. Amplification conditions were: one cycle 95°C, 12 min, four cycles 94°C 1 min, 50°C
Fig. 1. Synthesis of internal calibrators. Calibrators Ca, Cb and Cc were synthesised in a nested PCR, using fX174 DNA as a random internal sequence. The first round product generated from primers Cca, Ccb or Ccc (sense) and CC2 (antisense) contains the binding site for the assay antisense PCR primer (IEP2B), the probe binding site (Ca, Cb or Cc) and a random 10 base pair sequence at the 5%-end. The second round product, generated from primers CC1 (sense) and CC2 (antisense) introduces the primer binding site for the PCR sense primer (IEP1A).
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
14 Table 1 Oligonucleotide sequences Designation
Function
Sequence (5% to 3%)
Binding region
IEP1B
AGTTGGAGGACATCAAGCAGCCATGCAAAT
IEL1-AP
QPCR (sense) QPCR sense) QPCR (sense) QPCR sense) Target
LTR-AP
Target probe
Ca-AP Cb-AP Cc-AP Cca
Calibrator Calibrator Calibrator Calibrator
HIV-1 gag gene bases 568–597 consensus B HIV-1 gag gene bases 709–683 consensus B HIV-1 LTR bases 16–37 HIV-1 ELI HIV-1 LTR bases 168–196 HIV-1 ELI HIV-1 gag gene bases 630–651 consensus B HIV-1 LTR bases 124–148 HIV-1 ELI Probe binding site Ca Probe binding site Cb Probe binding site Cc fX174 bases 3151–3160
CCb
Calibrator
CCc
Calibrator
CC1
Calibrator
CC2
Calibrator
IEP2A-bio LTRNB1 LTRNB2-bio
primer
primer (anti-bioTGCTATGTCACTTCCCCTTGGTTCTCT primer
ACCAGRTYTGAGCCTGGGAGCT
primer (anti-bioCCTGTTCGGGCGCCACTGCTAGAGATTTT probe
APAATGGGATAGATTGCATCCAGT APACTCTGGTAACTAGATCCCT
probe APACAGTGTAGATAGATGACAGTC probe APATGCAAGGTCGCATATGAGTAA probe APATAAGCACGTGACTGAGTATGA synthesis ATTGATCATTACAGTGTAGATAGATGACAGTCGCTAAAGCTG synthesis ATTGATCATTATGCAAGGTCGCATATGAGTAAGCTAAAGCTG synthesis ATTGATCATTATAAGCACGTGACTGAGTATGAGCTAAAGCTG synthesis AGTTGGAGGACATCAAGCAGCCATGCAAATATTGATCATT synthesis TGCTATGTCACTTCCCCTTGGTTCTCTGCAGAAGTGC
1 min, 72°C 1 min, 34 cycles 94°C 1 min, 55°C 1 min, 72°C 1 min, one cycle 72°C 7 min and held at 4°C indefinitely. Amplified target and calibrator sequences were quantified in a chemiluminescent, microtitre hybridisation assay as previously described (Whitby and Garson, 1995). Briefly, for each sample 5 ml of PCR product diluted to 100 ml in a casein/Tween 20 buffer were captured on five streptavidin-coated wells, and denatured with 0.15 M NaOH. The second strand was washed away and the biotinylated captured strand (antisense) hybridised with alkaline phosphatase-labelled probes specific for gag or LTR PCR product amplified from the HIV-1 target or for Ca, Cb or Cc sequence amplified from the calibrators. Captured enzyme label was quantified with Lumiphos substrate (Lumigen) and counted in a Topcount luminometer (Canberra Packard). For each sample, a standard curve was generated from the three co-amplified calibrators by linear
fX174 bases 3151–3160 fX174 bases 3151–3160 Random sequence ATTGATCATT fX174 bases 3194–3203
regression and the proviral copy number generated by the gag and LTR primers calculated from this curve. Values from 1 to 10 copies/PCR and from 1000 to 10 000 copies/PCR were derived by extrapolation of the standard curve.
2.5. End-point dilution PCR Quantification of the synthesised calibrators and of unknown test samples for comparison with the newly developed method were made by nested PCR amplification of replicates of targets diluted to end-point as described by Simmonds et al. (1990).
2.6. Plasma 6irus RNA load Plasma virus loads were assayed using the Roche Amplicor Monitor assay (version 1.0) using the standard protocol with a detection limit of 400 copies/ml.
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
3. Results
3.1. Determination of DNA input by fluorometry Comparison of DNA estimation by the fluorometric assay with UV adsorption at 260 nm for extracts from 20 PBMC samples gave a correlation coefficient of R =0.95 (P B 0.001) between the two methods. The mean DNA concentration of the 20 extracts by the two methods was 94.3 and 87.9 mg/ml by fluorometry and spectrophotometry, respectively.
3.2. Titration of calibrators by end-point dilution The synthesised calibrators (Ca, Cb and Cc) were titrated by end-point dilution PCR in a half-log dilution series from 10 − 3 to 10 − 6, four replicates/dilution. PCR products were detected as described in Section 2.4. The assays showed that calibrators Ca, Cb and Cc contained 106, 105.5 and 105.5 copies/ml respectively, and thus were diluted as follows for use: Calibrator Ca 1:103 to give 1000 copies/ml Calibrator Cb 1:103.5 to give 100 copies/ml Calibrator Cc 1:104.5 to give 10 copies/ml
15
Inter-assay coefficients of variation were close to 10% for samples 1 to 4 in which mean input copy numbers per PCR reaction were greater than 50 (1.72 log10). Sample 5 showed a much higher level of variation which has probably resulted from the implicit sampling error at a mean input copy of 7 (0.84 log10) per PCR reaction.
3.5. Comparison of QPCR assay with end-point dilution PCR To establish assay performance with clinical samples, a comparison of provirus quantification between end-point dilution PCR and the present method was made on 20 samples. Results are shown in Fig. 4. The correlation between the two methods was R= 0.81 (PB 0.001). Copy numbers showed good equivalence with the intercept of the line of best fit being at two copies/reaction and the slope of the line being 1.12.
3.6. Changes in pro6irus load during antiretro6iral therapy To demonstrate the changes in provirus DNA load that accompany changes in plasma virus
3.3. Generation of standard cur6e using DNA extracted from 8E5 cell line A DNA extract from the 8E5 cell line was made in water containing 500 ng/ml DNA extracted from an uninfected T-cell line (CEM) and the dilutions assayed in triplicate to generate the standard curve shown in Fig. 2. An example of the internal standard curve generated for each individual reaction is shown in Fig. 3. Note that the higher copy number indicated by the LTR primer set is a consequence of the provirus genome having two copies of the LTR and hence the copy number is halved to give the final result based on this primer set.
3.4. Assay reproducibility Assay reproducibility was determined by assaying five DNA extracts in triplicate in three separate assay runs. Results are presented in Table 2.
Fig. 2. Standard curve generated from dilutions of DNA extracted from 8E5 cell line. Each point represents the mean of three assays.
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
16
Fig. 3. Example of the internal standard curve generated in each reaction. A line of best fit was produced from the luminescence signals from the three calibrators in each reaction and the signals from the gag and LTR primer sets compared to this line to determine the test proviral DNA copies/reaction. LTR signals are higher as each provirus contains two LTR copies.
RNA load, serial samples over a 24- to 28-week period from subjects participating in the MRC Quattro trial were assayed. Results and therapies received are shown in Fig. 5. Patient 1, taking quadruple AZT/3TC/ddC/loviride therapy showed a slow decrease in virus load over 20 weeks with a nadir of 0.7 log10 copies/106 PBMC below baseline, before returning towards baseline after 20 weeks of therapy. The reason for the rebound in provirus load is unclear as plasma
Fig. 4. Comparison of QPCR assay and end-point dilution PCR. DNA extracted from PBMC samples from 20 patients were assayed by end-point dilution PCR and the current QPCR assay.
virus RNA load remained approximately 3 log10 copies/ml below baseline throughout the sampling period. Patient 2, taking AZT/3TC dual therapy showed a similar slow decline in provirus load, reaching 0.5 log10 copies/106 PBMC below baseline at 28 weeks. Abrupt increases in RNA load of 0.7 log10 copies/ml between weeks 4 and 10 and 1.4 log10 copies/ml between weeks 24 and 26 were not reflected in similar changes in provirus load. Patients 3 and 4, taking cyclical four-drug therapy showed slight increases in provirus load of 0.2 and 0.3 log10 copies/106 PBMC respectively over 28
Table 2 Assay reproducibility Run 1
Run 2
Run 3
Replicate
1
2
3
1
2
3
1
2
3
Sample Sample Sample Sample Sample
2.02 1.76 1.69 2.08 1.40
1.84 1.93 2.01 2.25 Neg
1.50 1.89 1.68 2.10 0.78
1.72 1.84 1.66 1.66 1.00
1.72 1.95 1.78 2.02 1.30
2.08 1.78 1.43 2.26 1.30
1.92 2.14 1.84 1.54 0.00
1.67 2.19 1.46 2.03 1.00
2.14 1.76 1.90 2.26 0.00
1 2 3 4 5
* Log copies/PCR reaction.
Mean
S.D.
C.V. (%)
1.85* 1.91 1.72 2.02 0.84
0.20 0.15 0.18 0.24 0.52
11 8 10 12 62
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
weeks. Again, larger changes in RNA load in patient 4 were not reflected in provirus load changes.
3.7. Assay of pro6irus load in patients infected with genetically di6ergent (African) 6irus strains Table 3 gives the results of the provirus load assays with the gag and LTR primer sets in samples from seven patients infected in Africa. The LTR loads have been adjusted to allow for the presence of two LTR copies in the HIV-1 proviral genome. Of the seven samples four showed close agreement (B0.5 log difference) between the results with the gag and LTR primer sets, whereas three showed \ 2.0 log higher loads with the LTR primer set. One sample (sample 2)
17
gave no detectable amplification with the gag primers, whereas the LTR primers indicated a provirus load of 137 copies/reaction. Two other samples (3 and 4) were shown to have high copy numbers using the LTR primers (3538 and 1030 copies/reaction respectively) despite being assayed as having less than 10 copies/reaction with the gag primers.
4. Discussion Previous studies of HIV-1 provirus loads have observed prevalences of infection of CD4+ ve lymphocytes between 1:10 and 1:100 000 depending on the disease stage of the patient (Simmonds et al., 1990; Wood et al., 1993). As the maximum
Fig. 5. Changes in RNA and DNA loads over 24 weeks of antiretroviral combination therapy. Open symbols show changes in provirus load (assayed as DNA copies/106 PBMC), closed symbols changes in plasma virus RNA load (assayed as RNA copies/ml plasma). Therapies were either a combination of zidovudine, lamivudine, loviride and zalcitabine (patient 1), the same drugs taken as monotherapy cyclically, 8 weeks each (patients 3 and 4) or zidovudine and lamivudine alone (patient 2).
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
18
Table 3 Comparison of LTR and gag primer set performance with African virus strains Sample
1 2 3 4 5 6 7
Provirus copies (log10 copies)/reaction gag primer set
LTR primer set
237 0 2 6 298 628 218
200 137 3538 1030 279 961 543
(2.37) (–) (0.30) (0.78) (2.47) (2.80) (2.34)
(2.30) (2.14) (3.55) (3.10) (2.45) (2.98) (2.73)
sample input into a PCR reaction is approximately 1 mg DNA, equivalent to the DNA from 2× 105 cells, the number of HIV-1 provirus copies/reaction will generally fall in the range 1 – 10 000 for untreated patients, and is likely to be less after prolonged antiretroviral therapy. The assay described has been designed to cover this range and has used a three-point internal calibration system to improve assay reproducibility and accuracy, since any assay working at this low input copy number can potentially be subject to high sampling error. The use of three calibrators means that variability in the quantification of each individual calibrator will tend to be minimised when a line of best fit is added to the calibrator signals to generate the internal standard curve. Routinely, any calibrator regression line with an R 2-value less than 0.98 was rejected and the assay repeated. The assay we developed uses an internal calibration curve from 10 to 1000 copies/reaction. Extrapolation of this curve allowed quantification of target inputs up to 10 000 copies/reaction (data not shown) and detection between one and 10 copies/reaction. Clearly, results below 10 copies/ reaction can only be treated as semi-quantitative as sampling errors are particularly high, although qualitative provirus detection at this level may still be clinically useful. Assay dilutions of DNA extracts from the 8E5 cell line, shown in Fig. 2 indicate that the assay gives a linear response over the range 10–1000 input copies and replicate as-
says of samples containing a nominal single provirus DNA input copy show that the assay is sensitive down to this level. Comparison of the assay with an end-point dilution PCR method gave good correlation between the two procedures, especially when it is noted that end-point dilution does not give continuously variable quantification. Assay reproducibility, described in Table 2, gave levels of variation which are comparable to those seen with assays for plasma virus RNA (Schuurman et al., 1996) with S.D. values for the provirus load assay in the range 0.15–0.52 log10 copies. The denominator used in the assay was input DNA as measured by a fluorometric dye-binding method. Assuming a DNA content of 6 pg/cell, the provirus copies/reaction can be related to the cell equivalents of DNA added to the reaction to give copies/cell. The cellular DNA input in the assay was derived from fractionated PBMCs, and hence, to determine provirus copies/ml of blood or copies/106 CD4+ ve lymphocytes, it is necessary to know the CD4 + ve and total lymphocyte counts of the patient. As with plasma virus RNA load assays, the quantitative detection of non-clade-B virus strains by PCR can be perturbed due to primer/target mismatching. The provirus load assay described addressed this problem by including a second set of primers targeted at the LTR region of the HIV-1 genome which were optimised for the detection of genetically divergent virus strains (Berry et al., 1998). As shown in Table 3, the differences in load as determined by the two primer sets can be substantial, with loads derived by the LTR primer set being up to 1000-fold (3 log10 higher) than with the primer set targeted at the gag primer set, and in one case the gag primers failed to result in any amplification of the target. Previous studies have shown that primer mismatching can result in false negative PCR reactions (Arnold et al., 1995) and can underestimate RNA virus loads in plasma virus load assays (Alaeus et al., 1997). The changes that were seen in provirus load during antiretroviral therapy occurred slowly compared to the rapid changes in plasma virus
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
RNA load. As described previously, therapies that result in a rapid and sustained reduction in RNA load, such as patients 1 and 2 in the present study, show a slow decline in provirus DNA load (Perelson et al., 1997; Izopet et al., 1998). The probable explanation of this slow decline is that the virus-associated RNA seen in plasma is a result of virus production in lymphoid tissues, where the virus is turning over rapidly, with a short lifecycle of approximately 1 – 1.5 days (Ho et al., 1995; Wei et al., 1995). In contrast, the provirus DNA seen in PBMCs may represent cells in which there is a much longer virological latent period, or defective provirus which is unable to complete a replicative cycle (Coffin, 1995). Hence, the decline in the levels of provirus seen in PBMCs is limited by the length of the latent period, or by the lifespan of the latently infected cell. The latent period of replication competent provirus is related to the frequency with which the latently infected lymphocytes are activated by immune stimulation. It has been shown that the virus can remain latent in resting CD4+ ve memory T-cells for months and years, but is still replication competent, and can yield infectious virions if the cells are stimulated in vitro (Finzi et al., 1997; Wong et al., 1997). Studies of the comparative rates of evolution of drug resistance mutations in plasma virus RNA and PBMC provirus DNA, in which the appearance of the mutations is delayed in DNA, support the hypothesis that provirus DNA in PBMCs is not representative of the infected cell population as a whole, but consists of an infected cell population which has a longer latent period (Kaye et al., 1995; Wei et al., 1995). The existence of long lived reservoirs of HIV makes the monitoring of provirus loads an important tool for research into the dynamics of HIV-1 replication and virus clearance in patients taking antiretroviral therapy. The use of provirus load monitoring in clinical diagnostic situations allows the continuing monitoring of the efficacy of therapy when RNA loads have become undetectable, although the slow rates of change in PBMC provirus load probably make frequent monitoring unnecessary. Unlike plasma RNA loads which have a prognostic value in predicting disease pro-
19
gression (Mellors et al., 1996), the relationship between PBMC provirus load and disease progression is less well defined. HIV-1 provirus load, measured as copies/106 PBMCs has been shown to increase with disease progression and to be directly proportional to plasma RNA load and inversely proportional to CD4+ cell count (Simmonds et al., 1990; Wood et al., 1993). However, the relationship between absolute provirus loads in asymptomatic patients and disease prognosis has yet to be fully established.
Acknowledgements This work was supported in part by a grant from the Medical Research Council. Permission to use samples collected during the MRC Quattro Trial was granted by the MRC Quattro Steering Committee.
References Alaeus, A., Lidman, K., Sonnerborg, A., Albert, J., 1997. Subtype-specific problems with quantification of plasma HIV-1 RNA. AIDS 11, 859 – 865. Arnold, C., Barlow, K.L., Kaye, S., Loveday, C., Balfe, P., Clewley, J.P., 1995. HIV type 1 sequence subtype G transmission from mother to infant: failure of variant sequence species to amplify in Roche Amplicor test. AIDS Res. Hum. Retroviruses 11, 999 – 1001. Berry, N., Ariyoshi, C., Jaffar, S., Sabally, S., Corrah, T., Tedder, R., Whittle, H., 1998. Low peripheral blood viral HIV-2 RNA in individuals with high CD4 percentage differentiates HIV-2 from HIV-1 infection. J. Hum. Virol. 1, 457 – 468. Coffin, J.M., 1995. HIV population dynamics in vivo: implications for genetic variation, pathogenesis and therapy. Science 267, 483 – 489. Collins, M.L., Irvine, B., Tyner, D., Fine, E., Zayati, C., Chang, C., Horn, T., Ahle, D., Detmer, J., Shen, L.P., Kolberg, J., Bushnell, S., Urdea, M.S., Ho, D.D., 1997. A branched DNA signal amplification assay for quantification of nucleic acid targets below 100 molecules/ml. Nucleic Acids Res. 25, 2979 – 2984. Finzi, D., Hermankova, M., Pierson, T., Carruth, L.M., Buck, C., Chaisson, R.E., Quinn, T.C., Chadwick, K., Margolick, J., Brookmeyer, R., Gallant, J., Markowitz, M., Ho, D.D., Richman, D.D., Siliciano, R.F., 1997. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science 278, 1295 – 1300.
20
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
Folks, T.M., Powell, D., Lightfoot, M., Koenig, S., Fauci, A.S., Benn, S., Robson, A., Dougherty, D., Gendelman, H.E., Hoggan, M.D., Sundararajan, V., Martin, M.A., 1986. Biological and biochemical characterisation of a cloned Leu-3-cell surviving infection with the acquired immune deficiency syndrome virus. J. Exp. Med. 164, 280 – 290. Ho, D.D., Neumann, A.U., Perelson, A.S., Chen, W., Leonard, J.M., Markowitz, M., 1995. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373, 123 – 126. Izopet, J., Tamalet, C., Pasquier, C., Sandres, K., Marchou, B., Massip, P., Puel, J., 1998. Quantification of HIV-1 proviral DNA by a standardized colorimetric PCR-based assay. J. Med. Virol. 54, 54–59. Kaye, S., Comber, E., Tenant-Flowers, M., Loveday, C., 1995. The appearance of drug resistance-associated point mutations in HIV type 1 plasma RNA precedes their appearance in proviral DNA. AIDS Res. Hum. Retroviruses 11, 1221 – 1225. Kitchen, V., 1998. A randomised trial comparing regimens of four RT inhibitors given together or cyclically with a standard two drug regimen: the Quattro trial. Abstract 701, 5th Conference on Retroviruses and Opportunistic Infections, Chicago, USA. Mellors, J.W., Rinaldo, C.R.J., Gupta, P., White, R.M., Todd, J.A., Kingsley, L.A., 1996. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 272, 1167 – 1170. Perelson, A.S., Essunger, P., Cao, Y., Vesanen, M., Hurley, A., Saksela, K., Markowitz, M., Ho, D.D., 1997. Decay characteristics of HIV-1 infected compartments during combination therapy. Nature 387, 188–191. Saag, M.S., Holodniy, M., Kuritzkes, D.R., O’Brien, W.A., Coombs, R., Poscher, M.E., Jacobsen, D.M., Shaw, G.M., Richman, D.D., Volberding, P.A., 1996. HIV viral load markers in clinical practice. Nat. Med. 2, 625–629.
.
Schuurman, R., Descamps, D., Weverling, G.J., Kaye, S., Tijnagel, J., Williams, I., Van Leeuwen, R., Tedder, R., Boucher, C.A.B., Brun-Vezinet, F., Loveday, C., 1996. Multicenter comparison of three commercial methods for quantification of human immunodeficiency virus type 1 RNA in plasma. J. Clin. Microbiol. 34, 3016 – 3022. Simmonds, P., Balfe, P., Peutherer, J.F., Ludlam, C.A., Bishop, J.O., Leigh Brown, A.J., 1990. Human immunodeficiency virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol. 64, 864 – 872. Sun, R., Ku, J., Jayakar, H., Kuo, J.-C., Brambilla, D., Herman, S., Rosenstraus, M., Spadoro, J., 1998. Ultrasensitive reverse transcription-PCR assay for quantitation of human immunodeficiency virus type 1 RNA in plasma. J. Clin. Microbiol. 36, 2964 – 2969. Van Gemen, B., van Beuningen, R., Nabbe, A., van Strijp, D., Jurriaans, S., Lens, P., Kievits, T., 1994. A one-tube quantitative HIV-1 RNA NASBA nucleic acid amplification assay using electrochemiluminescent (ECL) labelled probes. J. Virol. Methods 49, 157 – 168. Wei, X., Ghosh, S.K., Taylor, M.E., Johnson, V.A., Emini, E.A., Deutsch, P., Lifson, J.D., Bonhoeffer, S., Nowak, M.A., Hahn, B.H., Saag, M.S., Shaw, G.M., 1995. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373, 117 – 122. Whitby, K., Garson, J.A., 1995. Optimisation and evaluation of a quantitative chemiluminescent polymerase chain reaction assay for hepatitis C virus RNA. J. Virol. Methods 51, 75 – 88. Wong, J.K., Hezareh, M., Gunthard, H.F., Havlir, D.V., Ignacio, C.C., Spina, C.A., Richman, D.D., 1997. Recovery of replication competent HIV despite prolonged suppression of plasma viraemia. Science 278, 1291 – 1295. Wood, R., Dong, H., Katzenstein, D.A., Merigan, T.C., 1993. Quantification and comparison of HIV-1 provirus load in peripheral blood mononuclear cells and isolated CD4 + T cells. J. Acquir. Immune Defic. Syndr. 6, 237 – 240.
A quantitative PCR method for the assay of HIV-1 provirus load in peripheral blood mononuclear cells Julie M. Bennett *, Steve Kaye, Neil Berry 1, Richard S. Tedder Department of Virology, Royal Free and Uni6ersity College London Medical School, 46 Cle6eland Street, London W1P 6DB, UK Received 30 April 1999; received in revised form 4 June 1999; accepted 11 June 1999
Abstract The use of high activity antiretroviral therapies (HAART) to treat HIV-infected patients frequently results in the long-term suppression of plasma virus RNA loads below levels detectable by current assays. The measurement of provirus DNA load in peripheral blood mononuclear cells provides a means of continuing to monitor the efficacy of treatment and the decline in reservoirs of latent virus. A quantitative PCR assay was developed for HIV-1 provirus using a three-point internal calibrator system to give high reproducibility and accuracy at the low copy numbers of provirus seen in clinical samples. Provirus DNA copies are related to cell number in the samples using a fluorescent dye-binding assay for measurement of input DNA. The assay agreed closely with an end-point dilution PCR and gave accurate quantification of extracts from an HIV-1 infected continuous cell line containing known provirus copy numbers. The inclusion of a second primer set in the LTR region of the HIV-1 genome, optimised to non-clade-B virus strains improved the detection and quantification of samples from patients infected with genetically divergent virus strains. Application of the assay to clinical trial patients showed no relationship between changes in provirus DNA loads and plasma virus RNA and changes in provirus load over 24 weeks were small. © 1999 Elsevier Science B.V. All rights reserved. Keywords: HIV-1; Provirus load; Quantitative PCR
1. Introduction The measurement of HIV-1 cell-free virus RNA load in plasma has become an important diagnos* Corresponding author. Tel.: +44-171-504-9490; fax: + 44-171-580-5896. E-mail address: [email protected] (J.M. Bennett) 1 Present address: Division of Retrovirology, National Institute for Biological Standards and Controls, Potters Bar, EN6 3QG, UK.
tic method for determining disease prognosis (Mellors et al., 1996) and assessing the efficacy of antiretroviral therapy in infected patients (Saag et al., 1996). In our experience and that of others, the use of high activity antiretroviral combination therapies (HAART) often results in plasma virus RNA loads falling below the level of detectability, even when using high sensitivity methods which can detect levels of 50 RNA copies/ml plasma or less (Collins et al., 1997; Sun et al., 1998).
0166-0934/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 6 - 0 9 3 4 ( 9 9 ) 0 0 0 9 6 - 8
12
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
To continue monitoring the efficacy of drug therapy in such patients, the assay of provirus DNA loads in peripheral blood mononuclear cells (PBMCs) may allow the continuing clearance of virus to be assessed. Studies of the dynamics of HIV-1 replication in infected patients and the rates of clearance of cell-free virus and integrated provirus have shown that although cell-free virus levels fall rapidly within a few days of commencing therapy, the decline of provirus loads proceeds much more slowly, and detectable provirus persists for months or years despite continuance of therapy (Finzi et al., 1997; Wong et al., 1997). To allow further studies of changes in provirus loads in both individual infected patients and patients enrolled in clinical trials of antiretroviral therapy, a method was developed for quantification of proviral genomes based on PCR amplification. The assay uses an internal calibration system described previously (Van Gemen et al., 1994) in which a standard curve is generated for each PCR amplified sample from three calibrators added to the reaction in the ratio 100:10:1. This system gives a high degree of reproducibility and accuracy in quantifying the low copy numbers of HIV-1 provirus found in PBMC samples. Quantitative detection of PCR products is made using enzymelabelled probes and a chemiluminescent substrate (Whitby and Garson, 1995) and sample input DNA is assayed by a fluorescent dye-binding method.
2. Methods and materials
2.1. Patient samples PBMCs were fractionated from whole blood samples collected into EDTA on Ficoll-Paque or from CPT collection tubes (Becton-Dickinson) and stored at − 70°C. Samples were collected from individual patients before and after undergoing various antiretroviral therapies, and from study subjects enrolled in the MRC Quattro Trial of four reverse transcriptase inhibitors used in combination (Kitchen, 1998).
2.2. DNA preparation DNA was extracted from the PBMCs from 1–2 ml blood using spin columns (QIAamp blood kit, Qiagen) and the final DNA extract was eluted into 50 ml of nuclease-free water. Routinely, 3 ml of this extract was added to a 50 ml PCR reaction volume. The concentration of DNA in the extract was estimated using a fluorometer (Ascent FL, Labsystems) after reacting the extract with Hoechst dye H33258. The DNA extract was diluted 1:50 in Tris/EDTA/NaCl (TEN) buffer and 150 ml reacted with 50 ml H33258 diluted to 5 mg/ml in TEN buffer for 5 min before reading (excitation 355 nm, emission 460 nm). Concentrations of DNA in the test samples were compared to a standard curve (32 ng/ml to 32 mg/ml in a halflog10 dilution series) made from calf thymus DNA (Sigma). To relate HIV-1 target copy number detected in the QPCR to input PBMC number it was assumed that a single PBMC yielded 6 pg DNA.
2.3. Preparation of oligonucleotide calibrators Three internal calibrators were prepared by nested PCR to incorporate unique probe binding sites and HIV-1 specific primer binding sites in a PCR product that was the same length (142 bp) as the PCR product amplified from the test samples. The method of preparation of the calibrators is shown in Fig. 1. A fX174 target (used as a random DNA sequence) was amplified in a nested PCR as shown, to incorporate the HIV-1 primer binding sites and internal probe sequence in a product of similar size and GC/AT ratio as the HIV-1 target sequence. The first round PCR reaction mix contained buffer, nucleotide and primer concentrations as shown below in Section 2.4 and approximately 1 pg fX174/Hae III molecular weight markers (Promega Life Sciences) as the PCR target. Reaction conditions were: four cycles 94°C 1 min, 37°C, 1 min, 72°C 1 min, six cycles 94°C 1 min, 50°C, 1 min, 72°C 1 min. Second round PCR reaction mix and conditions were identical to the first round other than the primers and the target which was a 1 ml transfer from the
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
first round reaction. Following PCR generation the calibrators were purified by agarose gel electrophoresis and further purified from the excised band with Geneclean (Anachem). After quantification (see below) and pooling in the ratio 100:10:1 copies of Ca, Cb and Cc respectively, the calibrators were stored at − 20°C in TE buffer containing 1 mg/ml herring sperm DNA at 1000× the working concentration. Calibrators were quantified by end-point dilution as described below, and by parallel line titration with DNA extracted from the 8E5 cell line, which is considered to contain a single proviral HIV-1 DNA copy per cell (Folks et al., 1986).
13
2.4. Quantitati6e PCR method DNA extracts (3 ml) were added to 50 ml PCR reactions and amplification was carried out in a hot-start PCR using AmpliTaq Gold (PE Biosystems) in order to reduce the generation of nonspecific products. Reaction mixes contained 10 mM Tris/HCl, pH 8.3, 50 mM KCl, 100 nM each of primers IEP1B, IEP2A-bio, LTRNB1, LTRNB2-bio (Table 1), 1.25 U AmpliTaq Gold, 200 mM each dNTP, 1.5 mM MgCl2 and 1 ml calibrator mix diluted to contain 1000, 100 and 10 DNA copies of Ca, Cb and Cc respectively. Reactions were amplified in a Perkin-Elmer 2400 thermal cycler. Amplification conditions were: one cycle 95°C, 12 min, four cycles 94°C 1 min, 50°C
Fig. 1. Synthesis of internal calibrators. Calibrators Ca, Cb and Cc were synthesised in a nested PCR, using fX174 DNA as a random internal sequence. The first round product generated from primers Cca, Ccb or Ccc (sense) and CC2 (antisense) contains the binding site for the assay antisense PCR primer (IEP2B), the probe binding site (Ca, Cb or Cc) and a random 10 base pair sequence at the 5%-end. The second round product, generated from primers CC1 (sense) and CC2 (antisense) introduces the primer binding site for the PCR sense primer (IEP1A).
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
14 Table 1 Oligonucleotide sequences Designation
Function
Sequence (5% to 3%)
Binding region
IEP1B
AGTTGGAGGACATCAAGCAGCCATGCAAAT
IEL1-AP
QPCR (sense) QPCR sense) QPCR (sense) QPCR sense) Target
LTR-AP
Target probe
Ca-AP Cb-AP Cc-AP Cca
Calibrator Calibrator Calibrator Calibrator
HIV-1 gag gene bases 568–597 consensus B HIV-1 gag gene bases 709–683 consensus B HIV-1 LTR bases 16–37 HIV-1 ELI HIV-1 LTR bases 168–196 HIV-1 ELI HIV-1 gag gene bases 630–651 consensus B HIV-1 LTR bases 124–148 HIV-1 ELI Probe binding site Ca Probe binding site Cb Probe binding site Cc fX174 bases 3151–3160
CCb
Calibrator
CCc
Calibrator
CC1
Calibrator
CC2
Calibrator
IEP2A-bio LTRNB1 LTRNB2-bio
primer
primer (anti-bioTGCTATGTCACTTCCCCTTGGTTCTCT primer
ACCAGRTYTGAGCCTGGGAGCT
primer (anti-bioCCTGTTCGGGCGCCACTGCTAGAGATTTT probe
APAATGGGATAGATTGCATCCAGT APACTCTGGTAACTAGATCCCT
probe APACAGTGTAGATAGATGACAGTC probe APATGCAAGGTCGCATATGAGTAA probe APATAAGCACGTGACTGAGTATGA synthesis ATTGATCATTACAGTGTAGATAGATGACAGTCGCTAAAGCTG synthesis ATTGATCATTATGCAAGGTCGCATATGAGTAAGCTAAAGCTG synthesis ATTGATCATTATAAGCACGTGACTGAGTATGAGCTAAAGCTG synthesis AGTTGGAGGACATCAAGCAGCCATGCAAATATTGATCATT synthesis TGCTATGTCACTTCCCCTTGGTTCTCTGCAGAAGTGC
1 min, 72°C 1 min, 34 cycles 94°C 1 min, 55°C 1 min, 72°C 1 min, one cycle 72°C 7 min and held at 4°C indefinitely. Amplified target and calibrator sequences were quantified in a chemiluminescent, microtitre hybridisation assay as previously described (Whitby and Garson, 1995). Briefly, for each sample 5 ml of PCR product diluted to 100 ml in a casein/Tween 20 buffer were captured on five streptavidin-coated wells, and denatured with 0.15 M NaOH. The second strand was washed away and the biotinylated captured strand (antisense) hybridised with alkaline phosphatase-labelled probes specific for gag or LTR PCR product amplified from the HIV-1 target or for Ca, Cb or Cc sequence amplified from the calibrators. Captured enzyme label was quantified with Lumiphos substrate (Lumigen) and counted in a Topcount luminometer (Canberra Packard). For each sample, a standard curve was generated from the three co-amplified calibrators by linear
fX174 bases 3151–3160 fX174 bases 3151–3160 Random sequence ATTGATCATT fX174 bases 3194–3203
regression and the proviral copy number generated by the gag and LTR primers calculated from this curve. Values from 1 to 10 copies/PCR and from 1000 to 10 000 copies/PCR were derived by extrapolation of the standard curve.
2.5. End-point dilution PCR Quantification of the synthesised calibrators and of unknown test samples for comparison with the newly developed method were made by nested PCR amplification of replicates of targets diluted to end-point as described by Simmonds et al. (1990).
2.6. Plasma 6irus RNA load Plasma virus loads were assayed using the Roche Amplicor Monitor assay (version 1.0) using the standard protocol with a detection limit of 400 copies/ml.
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
3. Results
3.1. Determination of DNA input by fluorometry Comparison of DNA estimation by the fluorometric assay with UV adsorption at 260 nm for extracts from 20 PBMC samples gave a correlation coefficient of R =0.95 (P B 0.001) between the two methods. The mean DNA concentration of the 20 extracts by the two methods was 94.3 and 87.9 mg/ml by fluorometry and spectrophotometry, respectively.
3.2. Titration of calibrators by end-point dilution The synthesised calibrators (Ca, Cb and Cc) were titrated by end-point dilution PCR in a half-log dilution series from 10 − 3 to 10 − 6, four replicates/dilution. PCR products were detected as described in Section 2.4. The assays showed that calibrators Ca, Cb and Cc contained 106, 105.5 and 105.5 copies/ml respectively, and thus were diluted as follows for use: Calibrator Ca 1:103 to give 1000 copies/ml Calibrator Cb 1:103.5 to give 100 copies/ml Calibrator Cc 1:104.5 to give 10 copies/ml
15
Inter-assay coefficients of variation were close to 10% for samples 1 to 4 in which mean input copy numbers per PCR reaction were greater than 50 (1.72 log10). Sample 5 showed a much higher level of variation which has probably resulted from the implicit sampling error at a mean input copy of 7 (0.84 log10) per PCR reaction.
3.5. Comparison of QPCR assay with end-point dilution PCR To establish assay performance with clinical samples, a comparison of provirus quantification between end-point dilution PCR and the present method was made on 20 samples. Results are shown in Fig. 4. The correlation between the two methods was R= 0.81 (PB 0.001). Copy numbers showed good equivalence with the intercept of the line of best fit being at two copies/reaction and the slope of the line being 1.12.
3.6. Changes in pro6irus load during antiretro6iral therapy To demonstrate the changes in provirus DNA load that accompany changes in plasma virus
3.3. Generation of standard cur6e using DNA extracted from 8E5 cell line A DNA extract from the 8E5 cell line was made in water containing 500 ng/ml DNA extracted from an uninfected T-cell line (CEM) and the dilutions assayed in triplicate to generate the standard curve shown in Fig. 2. An example of the internal standard curve generated for each individual reaction is shown in Fig. 3. Note that the higher copy number indicated by the LTR primer set is a consequence of the provirus genome having two copies of the LTR and hence the copy number is halved to give the final result based on this primer set.
3.4. Assay reproducibility Assay reproducibility was determined by assaying five DNA extracts in triplicate in three separate assay runs. Results are presented in Table 2.
Fig. 2. Standard curve generated from dilutions of DNA extracted from 8E5 cell line. Each point represents the mean of three assays.
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
16
Fig. 3. Example of the internal standard curve generated in each reaction. A line of best fit was produced from the luminescence signals from the three calibrators in each reaction and the signals from the gag and LTR primer sets compared to this line to determine the test proviral DNA copies/reaction. LTR signals are higher as each provirus contains two LTR copies.
RNA load, serial samples over a 24- to 28-week period from subjects participating in the MRC Quattro trial were assayed. Results and therapies received are shown in Fig. 5. Patient 1, taking quadruple AZT/3TC/ddC/loviride therapy showed a slow decrease in virus load over 20 weeks with a nadir of 0.7 log10 copies/106 PBMC below baseline, before returning towards baseline after 20 weeks of therapy. The reason for the rebound in provirus load is unclear as plasma
Fig. 4. Comparison of QPCR assay and end-point dilution PCR. DNA extracted from PBMC samples from 20 patients were assayed by end-point dilution PCR and the current QPCR assay.
virus RNA load remained approximately 3 log10 copies/ml below baseline throughout the sampling period. Patient 2, taking AZT/3TC dual therapy showed a similar slow decline in provirus load, reaching 0.5 log10 copies/106 PBMC below baseline at 28 weeks. Abrupt increases in RNA load of 0.7 log10 copies/ml between weeks 4 and 10 and 1.4 log10 copies/ml between weeks 24 and 26 were not reflected in similar changes in provirus load. Patients 3 and 4, taking cyclical four-drug therapy showed slight increases in provirus load of 0.2 and 0.3 log10 copies/106 PBMC respectively over 28
Table 2 Assay reproducibility Run 1
Run 2
Run 3
Replicate
1
2
3
1
2
3
1
2
3
Sample Sample Sample Sample Sample
2.02 1.76 1.69 2.08 1.40
1.84 1.93 2.01 2.25 Neg
1.50 1.89 1.68 2.10 0.78
1.72 1.84 1.66 1.66 1.00
1.72 1.95 1.78 2.02 1.30
2.08 1.78 1.43 2.26 1.30
1.92 2.14 1.84 1.54 0.00
1.67 2.19 1.46 2.03 1.00
2.14 1.76 1.90 2.26 0.00
1 2 3 4 5
* Log copies/PCR reaction.
Mean
S.D.
C.V. (%)
1.85* 1.91 1.72 2.02 0.84
0.20 0.15 0.18 0.24 0.52
11 8 10 12 62
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
weeks. Again, larger changes in RNA load in patient 4 were not reflected in provirus load changes.
3.7. Assay of pro6irus load in patients infected with genetically di6ergent (African) 6irus strains Table 3 gives the results of the provirus load assays with the gag and LTR primer sets in samples from seven patients infected in Africa. The LTR loads have been adjusted to allow for the presence of two LTR copies in the HIV-1 proviral genome. Of the seven samples four showed close agreement (B0.5 log difference) between the results with the gag and LTR primer sets, whereas three showed \ 2.0 log higher loads with the LTR primer set. One sample (sample 2)
17
gave no detectable amplification with the gag primers, whereas the LTR primers indicated a provirus load of 137 copies/reaction. Two other samples (3 and 4) were shown to have high copy numbers using the LTR primers (3538 and 1030 copies/reaction respectively) despite being assayed as having less than 10 copies/reaction with the gag primers.
4. Discussion Previous studies of HIV-1 provirus loads have observed prevalences of infection of CD4+ ve lymphocytes between 1:10 and 1:100 000 depending on the disease stage of the patient (Simmonds et al., 1990; Wood et al., 1993). As the maximum
Fig. 5. Changes in RNA and DNA loads over 24 weeks of antiretroviral combination therapy. Open symbols show changes in provirus load (assayed as DNA copies/106 PBMC), closed symbols changes in plasma virus RNA load (assayed as RNA copies/ml plasma). Therapies were either a combination of zidovudine, lamivudine, loviride and zalcitabine (patient 1), the same drugs taken as monotherapy cyclically, 8 weeks each (patients 3 and 4) or zidovudine and lamivudine alone (patient 2).
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
18
Table 3 Comparison of LTR and gag primer set performance with African virus strains Sample
1 2 3 4 5 6 7
Provirus copies (log10 copies)/reaction gag primer set
LTR primer set
237 0 2 6 298 628 218
200 137 3538 1030 279 961 543
(2.37) (–) (0.30) (0.78) (2.47) (2.80) (2.34)
(2.30) (2.14) (3.55) (3.10) (2.45) (2.98) (2.73)
sample input into a PCR reaction is approximately 1 mg DNA, equivalent to the DNA from 2× 105 cells, the number of HIV-1 provirus copies/reaction will generally fall in the range 1 – 10 000 for untreated patients, and is likely to be less after prolonged antiretroviral therapy. The assay described has been designed to cover this range and has used a three-point internal calibration system to improve assay reproducibility and accuracy, since any assay working at this low input copy number can potentially be subject to high sampling error. The use of three calibrators means that variability in the quantification of each individual calibrator will tend to be minimised when a line of best fit is added to the calibrator signals to generate the internal standard curve. Routinely, any calibrator regression line with an R 2-value less than 0.98 was rejected and the assay repeated. The assay we developed uses an internal calibration curve from 10 to 1000 copies/reaction. Extrapolation of this curve allowed quantification of target inputs up to 10 000 copies/reaction (data not shown) and detection between one and 10 copies/reaction. Clearly, results below 10 copies/ reaction can only be treated as semi-quantitative as sampling errors are particularly high, although qualitative provirus detection at this level may still be clinically useful. Assay dilutions of DNA extracts from the 8E5 cell line, shown in Fig. 2 indicate that the assay gives a linear response over the range 10–1000 input copies and replicate as-
says of samples containing a nominal single provirus DNA input copy show that the assay is sensitive down to this level. Comparison of the assay with an end-point dilution PCR method gave good correlation between the two procedures, especially when it is noted that end-point dilution does not give continuously variable quantification. Assay reproducibility, described in Table 2, gave levels of variation which are comparable to those seen with assays for plasma virus RNA (Schuurman et al., 1996) with S.D. values for the provirus load assay in the range 0.15–0.52 log10 copies. The denominator used in the assay was input DNA as measured by a fluorometric dye-binding method. Assuming a DNA content of 6 pg/cell, the provirus copies/reaction can be related to the cell equivalents of DNA added to the reaction to give copies/cell. The cellular DNA input in the assay was derived from fractionated PBMCs, and hence, to determine provirus copies/ml of blood or copies/106 CD4+ ve lymphocytes, it is necessary to know the CD4 + ve and total lymphocyte counts of the patient. As with plasma virus RNA load assays, the quantitative detection of non-clade-B virus strains by PCR can be perturbed due to primer/target mismatching. The provirus load assay described addressed this problem by including a second set of primers targeted at the LTR region of the HIV-1 genome which were optimised for the detection of genetically divergent virus strains (Berry et al., 1998). As shown in Table 3, the differences in load as determined by the two primer sets can be substantial, with loads derived by the LTR primer set being up to 1000-fold (3 log10 higher) than with the primer set targeted at the gag primer set, and in one case the gag primers failed to result in any amplification of the target. Previous studies have shown that primer mismatching can result in false negative PCR reactions (Arnold et al., 1995) and can underestimate RNA virus loads in plasma virus load assays (Alaeus et al., 1997). The changes that were seen in provirus load during antiretroviral therapy occurred slowly compared to the rapid changes in plasma virus
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
RNA load. As described previously, therapies that result in a rapid and sustained reduction in RNA load, such as patients 1 and 2 in the present study, show a slow decline in provirus DNA load (Perelson et al., 1997; Izopet et al., 1998). The probable explanation of this slow decline is that the virus-associated RNA seen in plasma is a result of virus production in lymphoid tissues, where the virus is turning over rapidly, with a short lifecycle of approximately 1 – 1.5 days (Ho et al., 1995; Wei et al., 1995). In contrast, the provirus DNA seen in PBMCs may represent cells in which there is a much longer virological latent period, or defective provirus which is unable to complete a replicative cycle (Coffin, 1995). Hence, the decline in the levels of provirus seen in PBMCs is limited by the length of the latent period, or by the lifespan of the latently infected cell. The latent period of replication competent provirus is related to the frequency with which the latently infected lymphocytes are activated by immune stimulation. It has been shown that the virus can remain latent in resting CD4+ ve memory T-cells for months and years, but is still replication competent, and can yield infectious virions if the cells are stimulated in vitro (Finzi et al., 1997; Wong et al., 1997). Studies of the comparative rates of evolution of drug resistance mutations in plasma virus RNA and PBMC provirus DNA, in which the appearance of the mutations is delayed in DNA, support the hypothesis that provirus DNA in PBMCs is not representative of the infected cell population as a whole, but consists of an infected cell population which has a longer latent period (Kaye et al., 1995; Wei et al., 1995). The existence of long lived reservoirs of HIV makes the monitoring of provirus loads an important tool for research into the dynamics of HIV-1 replication and virus clearance in patients taking antiretroviral therapy. The use of provirus load monitoring in clinical diagnostic situations allows the continuing monitoring of the efficacy of therapy when RNA loads have become undetectable, although the slow rates of change in PBMC provirus load probably make frequent monitoring unnecessary. Unlike plasma RNA loads which have a prognostic value in predicting disease pro-
19
gression (Mellors et al., 1996), the relationship between PBMC provirus load and disease progression is less well defined. HIV-1 provirus load, measured as copies/106 PBMCs has been shown to increase with disease progression and to be directly proportional to plasma RNA load and inversely proportional to CD4+ cell count (Simmonds et al., 1990; Wood et al., 1993). However, the relationship between absolute provirus loads in asymptomatic patients and disease prognosis has yet to be fully established.
Acknowledgements This work was supported in part by a grant from the Medical Research Council. Permission to use samples collected during the MRC Quattro Trial was granted by the MRC Quattro Steering Committee.
References Alaeus, A., Lidman, K., Sonnerborg, A., Albert, J., 1997. Subtype-specific problems with quantification of plasma HIV-1 RNA. AIDS 11, 859 – 865. Arnold, C., Barlow, K.L., Kaye, S., Loveday, C., Balfe, P., Clewley, J.P., 1995. HIV type 1 sequence subtype G transmission from mother to infant: failure of variant sequence species to amplify in Roche Amplicor test. AIDS Res. Hum. Retroviruses 11, 999 – 1001. Berry, N., Ariyoshi, C., Jaffar, S., Sabally, S., Corrah, T., Tedder, R., Whittle, H., 1998. Low peripheral blood viral HIV-2 RNA in individuals with high CD4 percentage differentiates HIV-2 from HIV-1 infection. J. Hum. Virol. 1, 457 – 468. Coffin, J.M., 1995. HIV population dynamics in vivo: implications for genetic variation, pathogenesis and therapy. Science 267, 483 – 489. Collins, M.L., Irvine, B., Tyner, D., Fine, E., Zayati, C., Chang, C., Horn, T., Ahle, D., Detmer, J., Shen, L.P., Kolberg, J., Bushnell, S., Urdea, M.S., Ho, D.D., 1997. A branched DNA signal amplification assay for quantification of nucleic acid targets below 100 molecules/ml. Nucleic Acids Res. 25, 2979 – 2984. Finzi, D., Hermankova, M., Pierson, T., Carruth, L.M., Buck, C., Chaisson, R.E., Quinn, T.C., Chadwick, K., Margolick, J., Brookmeyer, R., Gallant, J., Markowitz, M., Ho, D.D., Richman, D.D., Siliciano, R.F., 1997. Identification of a reservoir for HIV-1 in patients on highly active antiretroviral therapy. Science 278, 1295 – 1300.
20
J.M. Bennett et al. / Journal of Virological Methods 83 (1999) 11–20
Folks, T.M., Powell, D., Lightfoot, M., Koenig, S., Fauci, A.S., Benn, S., Robson, A., Dougherty, D., Gendelman, H.E., Hoggan, M.D., Sundararajan, V., Martin, M.A., 1986. Biological and biochemical characterisation of a cloned Leu-3-cell surviving infection with the acquired immune deficiency syndrome virus. J. Exp. Med. 164, 280 – 290. Ho, D.D., Neumann, A.U., Perelson, A.S., Chen, W., Leonard, J.M., Markowitz, M., 1995. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373, 123 – 126. Izopet, J., Tamalet, C., Pasquier, C., Sandres, K., Marchou, B., Massip, P., Puel, J., 1998. Quantification of HIV-1 proviral DNA by a standardized colorimetric PCR-based assay. J. Med. Virol. 54, 54–59. Kaye, S., Comber, E., Tenant-Flowers, M., Loveday, C., 1995. The appearance of drug resistance-associated point mutations in HIV type 1 plasma RNA precedes their appearance in proviral DNA. AIDS Res. Hum. Retroviruses 11, 1221 – 1225. Kitchen, V., 1998. A randomised trial comparing regimens of four RT inhibitors given together or cyclically with a standard two drug regimen: the Quattro trial. Abstract 701, 5th Conference on Retroviruses and Opportunistic Infections, Chicago, USA. Mellors, J.W., Rinaldo, C.R.J., Gupta, P., White, R.M., Todd, J.A., Kingsley, L.A., 1996. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 272, 1167 – 1170. Perelson, A.S., Essunger, P., Cao, Y., Vesanen, M., Hurley, A., Saksela, K., Markowitz, M., Ho, D.D., 1997. Decay characteristics of HIV-1 infected compartments during combination therapy. Nature 387, 188–191. Saag, M.S., Holodniy, M., Kuritzkes, D.R., O’Brien, W.A., Coombs, R., Poscher, M.E., Jacobsen, D.M., Shaw, G.M., Richman, D.D., Volberding, P.A., 1996. HIV viral load markers in clinical practice. Nat. Med. 2, 625–629.
.
Schuurman, R., Descamps, D., Weverling, G.J., Kaye, S., Tijnagel, J., Williams, I., Van Leeuwen, R., Tedder, R., Boucher, C.A.B., Brun-Vezinet, F., Loveday, C., 1996. Multicenter comparison of three commercial methods for quantification of human immunodeficiency virus type 1 RNA in plasma. J. Clin. Microbiol. 34, 3016 – 3022. Simmonds, P., Balfe, P., Peutherer, J.F., Ludlam, C.A., Bishop, J.O., Leigh Brown, A.J., 1990. Human immunodeficiency virus-infected individuals contain provirus in small numbers of peripheral mononuclear cells and at low copy numbers. J. Virol. 64, 864 – 872. Sun, R., Ku, J., Jayakar, H., Kuo, J.-C., Brambilla, D., Herman, S., Rosenstraus, M., Spadoro, J., 1998. Ultrasensitive reverse transcription-PCR assay for quantitation of human immunodeficiency virus type 1 RNA in plasma. J. Clin. Microbiol. 36, 2964 – 2969. Van Gemen, B., van Beuningen, R., Nabbe, A., van Strijp, D., Jurriaans, S., Lens, P., Kievits, T., 1994. A one-tube quantitative HIV-1 RNA NASBA nucleic acid amplification assay using electrochemiluminescent (ECL) labelled probes. J. Virol. Methods 49, 157 – 168. Wei, X., Ghosh, S.K., Taylor, M.E., Johnson, V.A., Emini, E.A., Deutsch, P., Lifson, J.D., Bonhoeffer, S., Nowak, M.A., Hahn, B.H., Saag, M.S., Shaw, G.M., 1995. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373, 117 – 122. Whitby, K., Garson, J.A., 1995. Optimisation and evaluation of a quantitative chemiluminescent polymerase chain reaction assay for hepatitis C virus RNA. J. Virol. Methods 51, 75 – 88. Wong, J.K., Hezareh, M., Gunthard, H.F., Havlir, D.V., Ignacio, C.C., Spina, C.A., Richman, D.D., 1997. Recovery of replication competent HIV despite prolonged suppression of plasma viraemia. Science 278, 1291 – 1295. Wood, R., Dong, H., Katzenstein, D.A., Merigan, T.C., 1993. Quantification and comparison of HIV-1 provirus load in peripheral blood mononuclear cells and isolated CD4 + T cells. J. Acquir. Immune Defic. Syndr. 6, 237 – 240.