HIV viral load testing

Abstract

Methods of testing HIV viral load are described. The methods comprise detecting HIV viral RNA in a sample of leukocyte-depleted blood. Such methods can be carried out on low-volume samples obtained without the need for venipuncture or a centrifuge. The methods are particularly suited for HIV viral load testing in resource-limited settings. Methods for monitoring HIV infection are also described, as well as kits for carrying out the methods.

Claims

1. A method of testing HIV viral load, comprising detecting HIV viral RNA in a sample of leukocyte-depleted blood, wherein the sample of leukocyte-depleted blood is a whole blood sample that has been selectively depleted of leukocytes or a diluted whole blood sample that has been selectively depleted of leukocytes, and wherein the method is carried out without preparing a plasma sample.

2. The method according to claim 1, wherein the sample of leukocyte-depleted blood has been depleted of more than 99.9% of the leukocytes present in the whole blood sample, or the diluted whole blood sample, from which the leukocyte-depleted sample was obtained.

3. The method according to claim 1, further comprising selectively depleting a whole blood sample, or the diluted whole blood sample, of leukocytes prior to said detecting.

4. The method according to claim 3, wherein the whole blood sample is depleted of at least 99.9% of the number of leukocytes present in the whole blood sample, or the diluted whole blood sample is depleted of more than 99.9% of the number of leukocytes present in the diluted whole blood sample.

5. The method according to claim 3, further comprising diluting the whole blood sample prior to leukodepletion.

6. The method according to claim 3, wherein the whole blood sample, or the diluted whole blood sample, is selectively depleted of leukocytes by filtering the sample through a leukoreduction filter.

7. The method according to claim 6, wherein a pressure differential is applied across the leukoreduction filter to cause the whole blood sample, or the diluted whole blood sample, to pass through the filter.

8. The method according to claim 6, wherein the thickness of the leukoreduction filter is selected such that at least a 99.9% reduction in the number of leukocytes from the whole blood sample, or the diluted whole blood sample, is obtained by filtering the sample through the leukoreduction filter.

9. The method according to claim 1, wherein the number of copies of HIV viral RNA/ml of the sample of leukocyte-depleted blood is determined.

10. The method according to claim 1, in which HIV viral RNA in the sample of leukocyte-depleted blood is detected by extracting nucleic acid from the sample of leukocyte-depleted blood, and detecting HIV viral RNA present in the extracted nucleic acid.

11. The method according to claim 1, wherein HIV viral RNA is detected by reverse transcription of the HIV viral RNA, and subsequent isothermal nucleic acid amplification of a product of the reverse transcription.

12. The method according to claim 1, comprising detecting HIV-1 viral RNA in the sample of leukocyte-depleted blood.

13. A method of testing HIV viral load, comprising the steps of: selectively depleting a whole blood sample, or a diluted whole blood sample, of leukocytes to provide a sample of leukocyte-depleted blood, wherein the whole blood sample is a low volume whole blood sample up to 500 l; and detecting HIV viral RNA in the sample of leukocyte-depleted blood; wherein the method is carried out without preparing a plasma sample.

14. The method according to claim 13, further comprising obtaining the whole blood sample from a subject by finger prick or heel prick.

15. A method of testing HIV viral load, comprising the steps of: selectively depleting a whole blood sample, or a diluted whole blood sample, of leukocytes to provide a sample of leukocyte-depleted blood, wherein the diluted whole blood sample is diluted at least 1-in-2 with an isotonic solution; and detecting HIV viral RNA in the sample of leukocyte-depleted blood; wherein the method is carried out without preparing a plasma sample.

16. The method according to claim 15, wherein the diluted whole blood sample is diluted at least 1-in-2 with phosphate buffered saline.

17. A method of testing HIV viral load, comprising the steps of: rehydrating a dried sample of leukocyte-depleted blood to provide a sample of leukocyte-depleted blood, wherein the dried sample of leukocyte-depleted blood is a whole blood sample that has been selectively depleted of leukocytes or a diluted whole blood sample that has been selectively depleted of leukocytes; and detecting HIV viral RNA in the rehydrated sample of leukocyte-depleted blood; wherein the method is carried out without preparing a plasma sample.

18. A method of testing HIV viral load, comprising the steps of: drying a sample of leukocyte-depleted blood, wherein the sample of leukocyte-depleted blood is a whole blood sample that has been selectively depleted of leukocytes or a diluted whole blood sample that has been selectively depleted of leukocytes; rehydrating the dried sample of leukocyte-depleted blood; and detecting HIV viral RNA in the rehydrated sample of leukocyte-depleted blood; wherein the method is carried out without preparing a plasma sample.

Description

(1) Embodiments of the invention are described in the following examples with reference to the accompanying drawings in which:

(2) FIG. 1 shows leukocyte removal by different filter compositions from whole blood samples spiked with 100,000 8E5 cells/ml, tested by RT and CD45 qPCR. Error bars represent the standard deviation;

(3) FIG. 2 shows 8E5 cell removal by different filter compositions from whole blood samples spiked with 100,000 8E5 cells/ml, tested by HIV-1 qPCR. Error bars represent the standard deviation;

(4) FIG. 3 shows leukocyte removal by filters of 6-10 layers from whole blood samples spiked with 100,000 8E5 cells/ml, tested by RT and CD45 qPCR. Error bars represent the standard deviation;

(5) FIG. 4 shows leukocyte removal by centrifugation or filtration from whole blood samples spiked with 100,000 8E5 cells/ml, tested by RT and CD45 qPCR. Error bars represent the standard deviation;

(6) FIG. 5 shows 8E5 cell removal by centrifugation or filtration from whole blood samples spiked with 100,000 8E5 cells/ml, tested by HIV-1 PCR. Error bars represent the standard deviation;

(7) FIG. 6 shows free HIV-1 particle detection in filtered or unfiltered spiked whole blood samples at a range of input concentrations, tested by HIV-1 RT-qPCR. Error bars represent the standard deviation;

(8) FIG. 7 shows average signal strengths of test and control lines on test strips at a range of input HIV concentrations;

(9) FIG. 8 shows the percentage of test strips with a test line present at a range of input HIV concentrations. In all cases, n=8; and

(10) FIG. 9 shows whether positive or negative results were obtained on test strips for a range of input HIV concentrations. Filled markers represent data points which fall outside of the accepted range of 30.3 log.sub.10 copies/ml.

EXAMPLES

(11) Materials and Methods

(12) Whole blood was spiked with cells of the 8E5 cell line containing one copy of proviral HIV DNA, or with culture supernatant of HIV-1 LAI (RNA, subtype B).

(13) The efficacy of whole blood filtration for nucleated cells was assessed in three different ways:

(14) i) White cell count pre- and post-filtration using a sensitive flow cytometric assay. This method counts white cells present in the whole blood as well as the spiked 8E5 cells;

(15) ii) Relative CD45+ cell amounts were assessed using the Taqman Gene Expression PCR assay (Life Technologies, primer/probe set Hs04189704_m1). The reverse transcription (RT) step required for this assay was completed using the SuperScript III First Strand Synthesis System (Invitrogen) according to the manufacturer's specifications. The CD45 PCR was completed using the Stratagene Mx3000 instrument. This method detects white cells present in the whole blood as well as the spiked 8E5 cells;

(16) iii) Quantification of proviral DNA against a plasmid-derived standard. This was performed by real-time PCR (qPCR) using the Stratagene Mx3000 as previously reported (Candotti et al., J Virol Methods 2004; 118: 39-47: Multiplex real-time quantitative RT-PCR assay for hepatitis B virus, hepatitis C virus and human immunodeficiency virus type 1); HIV-1 RNA was quantified using real-time RT-PCR as previously described (Candotti et al. 2004). The reference used was a secondary standard calibrated using the Artus HIVirus-1 RG RT-PCR kit in copies/ml.

(17) All nucleic acid extractions for PCR and RT-PCR reactions were performed with the High Pure Viral Nucleic Acid kit (Roche), according to the manufacturer's instructions.

(18) Where filtered samples were tested alongside plasma as a comparison to the leukoreduction levels normally employed for viral load testing, plasma was generated by centrifugation at 4,000 rcf (relative centrifugal force) for 15 minutes.

Example 1

(19) Comparison of Different Leukoreduction Filters

(20) The efficacy of removal of leukocytes from whole blood spiked with 8E5 cells using three different leukoreduction filters (a Macopharma1 filter, containing five layers, a Macopharma2 filter, containing six layers, and a Pall filter) was compared. Leukocyte removal by each filter was assessed using the three different methods described above: (i) white blood cell count by flow cytometric analysis; (ii) white blood cell quantification by RT and CD45 qPCR; and (iii) qPCR for HIV-1 proviral DNA.

(21) The results as assessed by white blood cell (WBC) count by flow cytometry are shown in Table 1 below. The results as assessed by RT and CD45 qPCR are shown in FIG. 1, and the results as assessed by HIV-1 qPCR are shown in FIG. 2. Table 2 records the percentage leukocyte removal from the results shown in FIGS. 1 and 2.

(22) TABLE-US-00001 TABLE 1 Leukocyte removal from whole blood as assessed by flow cytometry Log reduction % reduction WBC Log compared to compared to count WBC Diluted Diluted Sample (cells/l) Count Whole Blood Whole Blood Diluted whole blood 723.00 2.86 Macopharma1 97.30 1.99 0.87 86.54 filtered blood Macopharma2 0.80 0.10 2.95 99.89 filtered blood Pall filtered blood 4.61 0.66 2.20 99.36 Centrifuged plasma 2.46 0.39 2.47 99.66

(23) TABLE-US-00002 TABLE 2 Percentage leukocyte removal from whole blood as assessed by RT and CD45 qPCR, and HIV-1 qPCR Log reduction in Log reduction in leukocytes from % Leukocyte removal leukocytes from % Leukocyte removal diluted whole blood from diluted whole diluted whole blood from diluted whole as assessed by blood as assessed as assessed by blood as assessed Filter RT and CD45 qPCR by RT and CD45 qPCR HIV-1 qPCR by HIV-1 qPCR Macopharma1 1.27 94.53 1.12 92.38 Macopharma2 2.77 98.83 1.99 98.97 Pall 2.61 98.80 1.54 97.14

(24) The results as assessed by flow cytometry show that 86.54% of leukocytes were removed using the Macopharma1 filter, 99.89% of leukocytes were removed using the Macopharma2 filter, and 99.66% of were removed using the Pall filter. The results as assessed by CD45 qPCR show that 94.53% of leukocytes were removed using the Macopharma1 filter, 99.83% of leukocytes were removed using the Macopharma2 filter, and 98.80% of leukocytes were removed using the Pall filter. The results as assessed by HIV-1 qPCR show that 92.38% of leukocytes were removed using the Macopharma1 filter, 98.97% of leukocytes were removed using the Macopharma2 filter, and 97.14% of leukocytes were removed using the Pall filter.

(25) The highest level of leukocyte removal was achieved using the Macopharma2 filter material. The lowest level of leukocyte removal was achieved using the Macopharma1 filter material, and the Pall filter performed at an intermediate level. The Macopharma2 filter was used for subsequent experiments.

Example 2

(26) Optimisation of Filter Thickness

(27) The effect of the number of layers of filter material used for leukoreduction was determined. The Macopharma2 filter material used in Example 1 contained six layers. Since leukoreduction filters rely primarily on depth filtration to deplete the cell content, it was expected that increasing the number of layers of filter material would improve the efficacy of leukoreduction.

(28) Whole blood samples spiked with 100,000 8E5 cells/ml were filtered using six, eight, or ten layers of the selected Macopharma filter. The efficacy of leukoreduction was determined by RT and CD45 qPCR, as described above. The results are shown in FIG. 3. Table 3 records the percentage leukocyte removal from the results shown in FIG. 3.

(29) TABLE-US-00003 TABLE 3 Percentage CD45+ cell removal from samples filtered using different numbers of layers of filter material Log reduction in Percentage leukocyte leukocytes from diluted removal from diluted Number of whole blood as assessed whole blood as assessed layers by RT and CD45 qPCR by RT and CD45 qPCR 6 2.55 99.72 8 3.04 99.91 10 3.02 99.90

(30) The results show that a filter containing eight layers performs better than a filter with six layers. Greater than 3 log reduction (>99.9%) of white blood cells was obtained using eight layers of the Macopharma2 filter. However there was no significant difference in performance between filters of eight and ten layers.

(31) The dead volume of each filter was determined by inserting the filter into a pre-weighed tube and centrifuging at 13,000 rpm for one minute. The tube was then re-weighed and the difference used to calculate the volume of liquid collected (given that the diluted blood has approximately the same density as water). The dead volume of the filter increases with the number of layers, as shown in Table 4.

(32) TABLE-US-00004 TABLE 4 Dead volume of different filter thicknesses, measured by liquid weight Number of layers Average Dead volume (l) 6 39.2 8 50.8 10 77.2

(33) Given the small starting volume of sample, a filter comprised of eight layers was selected because it achieves the desired reduction of 3 log white blood cells, with a dead volume of only 50 l.

Example 3

(34) Optimisation of Sample Dilution

(35) In view of the relatively small volumes of whole blood that can be collected without venipuncture, the effect of dilution of a small volume of a whole blood sample on the efficacy of leukoreduction was determined.

(36) 100 ul of whole blood (un-spiked) was diluted 1:2, 1:3, or 1:4 with phosphate buffered saline (PBS). The diluted samples were filtered using the selected Macopharma2 leukoreduction filter, and the level of white blood cell removal was assessed by RT and CD45 qPCR, as described above. No significant difference in the effectiveness of leukoreduction was observed using different sample dilutions.

(37) The eight-layer Macopharma2 filter configuration was assessed for optimal sample dilution to pass sample through the filter, so as to achieve the optimal level of white blood cell removal when using an input of 100 l of whole blood sample. A dilution of 100 l of whole blood in 200 l of phosphate buffered saline (PBS) (i.e. a dilution factor of 1:2) was found to be sufficient to allow blood to pass through the filter without diluting the sample excessively.

Example 4

(38) Optimisation of Filtration Pressure

(39) Samples of diluted whole blood were filtered through the selected Macopharma2 filter. The volume of air used was constant (1 ml), but the speed was varied. The level of white blood cell removal at the different pump speeds was assessed by RT and CD45 qPCR, as described above. The results are shown in Table 5.

(40) TABLE-US-00005 TABLE 5 CD45+ cell removal from samples filtered under different pressures. Percentage Log reduction in leukocyte CD45 PCR leukocytes from reduction from Quantification diluted whole diluted whole Pump speed (l/s) (cells/ml) blood blood Not filtered (diluted 1,657,652 blood sample) 25 1,129 3.17 99.93 50 1,198 3.14 99.93 100 2,082 2.90 99.87 200 3,150 2.72 99.81

(41) As shown in Table 5, at a lower pump speed (and, therefore, at a lower pressure), there is a slightly higher level of removal of white blood cells. However, we have found that with decreasing pump speed there is an increase in filter dead volume. Thus, there may need to be a compromise between the efficacy of removal of white blood cells and the volume of sample lost to the filter.

(42) The eight-layer Macopharma2 filter configuration was assessed for optimal pump speed to pass sample through the filter, so as to achieve the optimal level of white blood cell removal when using an input of 100 l of whole blood sample. This was assessed by RT and CD45 qPCR as detailed above. A pump speed of 50 l/s was found to be optimal for 100 l whole blood sample diluted in 200 l PBS.

Example 5

(43) Assessment of Leukodepletion of Spiked Whole Blood

(44) The leukodepletion performance of the eight-layer Macopharma2 filter was assessed. Whole blood samples spiked with 100,000 8E5 cells/ml were filtered, and leukodepletion of the filtered samples was assessed using the three different methods described above: (i) white blood cell count by flow cytometric analysis; (ii) white blood cell quantification by RT and CD45 qPCR; and (iii) qPCR for HIV-1 proviral DNA. Leukodepletion of the filtered samples was compared to plasma samples (i.e. samples in which the cells have been removed from whole blood by centrifugation) since these are conventionally used for HIV viral load tests. The results obtained are shown in Table 6 below, and in FIGS. 4 and 5. Table 7 records the percentage leukocyte removal from the results shown in FIGS. 4 and 5.

(45) TABLE-US-00006 TABLE 6 Leukoreduction by centrifuqation or filtration from whole blood samples spiked with 100,000 8E5 cells/ml, assessed by flow cytometry. Average Log Log difference Percentage WBC count WBC Average from whole reduction from Sample (cells/l) count WBC count blood whole blood Diluted whole 1448.00 1510.50 3.18 blood 1573.00 Diluted plasma 0.77 1.31 0.12 3.06 99.91 1.85 Macopharma2 (8 1.08 0.50 0.30 3.48 99.97 layers) filtered 0.15 blood 0.26

(46) TABLE-US-00007 TABLE 7 Percentage leukocyte removal from whole blood as assessed by RT and CD45 qPCR, and HIV-1 qPCR Log reduction of Log reduction of leukocytes from % Leukocyte removal leukocytes from % Leukocyte removal diluted whole blood from diluted whole diluted whole blood from diluted whole as assessed by RT blood as assessed as assessed by blood as assessed Sample and CD45 qPCR by RT and CD45 qPCR HIV-1 qPCR by HIV-1 qPCR Plasma 2.47 99.66 1.70 98.02 Macopharma2 3.15 99.93 1.86 98.63 (8 layers) filtered blood

(47) The results as assessed by flow cytometry, and by RT and CD45 qPCR, show that filtration using eight layers of the selected Macopharma2 filter achieved greater than 3 log leukoreduction. By all methods of assessment, filtration of the whole blood sample achieved a greater level of leukoreduction than centrifugation.

Example 6

(48) Assessment of Retention of HIV Particles by Leukoreduction Filter

(49) This example describes an assessment of the degree to which HIV particles are retained by a leukoreduction filter when spiked whole blood samples are passed through the filter.

(50) It is important that the filter does not retain any of the free HIV particles in circulation so that the viral load of the sample is not underestimated. To determine the level of retention of viral particles by the Macopharma2 filter, blood was spiked with various concentrations of culture supernatant of HIV-1 LAI (subtype B). The amount of HIV in the diluted blood samples before and after filtration was assessed by HIV-1 RT-qPCR. The results are shown in FIG. 6. Table 8 records the HIV concentration in copies/ml from the results shown in FIG. 6.

(51) TABLE-US-00008 TABLE 8 Amount of HIV particles in blood samples before and after filtration HIV (c/ml) in Macopharma2 HIV (c/ml) in non- HIV-1 input (c/ml) (8 layers) filtered blood filtered blood 10,000 5,368 5,347 5,000 2,868 3,318 1,000 358 336 500 133 142

(52) The results show that there is no significant retention of HIV particles on the filter.

Example 7

(53) Viral Load Testing of Spiked Blood Samples for Use in Resource-Limited Settings

(54) In resource-limited settings, it is necessary to determine whether HIV viral load is below or above 1000 copies/ml. This Example describes a method for determining whether the HIV viral load of a sample is above or below 1000 copies/ml.

(55) Blood samples were spiked with different concentrations of the World Health Organisation (WHO) 3rd International Standard for HIV-1 RNA (subtype B) (National Institute for Biological Standards and Control, NIBSC). The NIBSC viral stock is quantified in International Units (IU). There are approximately 1.7 copies of HIV per 1 IU.

(56) 100 l whole blood samples were diluted in 200 l phosphate buffered saline (PBS), and filtered through eight layers of the Macopharma2 filter material at a pump speed of 50 l air per second.

(57) Viral RNA was extracted, amplified by isothermal nucleic acid amplification, and the amplification products were detected by rapid visual detection with a dipstick, using a simple amplification-based assay (SAMBA) method similar to the method described in Lee et al., Journal of Infectious Diseases 2010; 201(S1):S65-S71.

(58) The results are shown in FIGS. 7 and 8. The results shown in FIG. 7 demonstrate that a good average signal strength was obtained for test samples for which the HIV input was 800 or 1000 IU/ml. The results shown in FIG. 8 demonstrate that a positive test signal was obtained for 75% of the samples for which the HIV input was 800 or 1000 IU/ml. It is concluded that the viral load test procedure described in this Example provides reliable results for determining whether or not a test sample has at least 1000 copies/ml HIV viral RNA, the cut-off for virologic failure in resource-limited settings.

Example 8

(59) Viral Load Testing of Spiked Blood Samples for Use in Resource-Limited Settings

(60) Blood samples were spiked with plasma samples from four patients collected in Namibia (HIV-1 subtype C) (each of which has a very high viral load; >30,000 copies/ml), and diluted in five doubling dilutions, with final viral load ranging between 120 and 6100 copies/ml.

(61) 120 l whole blood samples were diluted in 240 l phosphate buffered saline (PBS), and filtered through eight layers of the Macopharma2 filter material at a pump speed of 50 l air per second.

(62) Viral RNA was extracted, amplified by isothermal nucleic acid amplification, and the amplification products were detected by rapid visual detection with a dipstick, using a SAMBA method similar to the method described in Lee et al., Journal of Infectious Diseases 2010; 201(S1):S65-S71.

(63) The results are shown in FIG. 9, and demonstrate that within the acceptable range given for viral load tests (30.3 log.sub.10 copies/ml) the procedure is 98.75% accurate.

Example 9

(64) Viral Load Testing of Clinical Samples

(65) 207 whole blood samples from infected individuals, most of whom were undergoing treatment, were leukodepleted by filtration through a leukodepletion filter. Viral RNA was extracted, amplified by isothermal nucleic acid amplification, and the amplification products were detected by rapid visual detection with a dipstick, using a SAMBA method similar to the method described in Lee et al., Journal of Infectious Diseases 2010; 201(S1):S65-S71. The assay was designed to monitor HIV viral load with a cut-off of 1,000 copies/ml.

(66) In parallel, the Roche COBAS AmpliPrep/COBAS TaqMan HIV-1 test was used to quantify the HIV viral copy in plasma, obtained by centrifugation at 2,200 g for 5 minutes. Testing by the SAMBA and Roche tests was completed on the same day for each patient sample.

(67) Plasma samples were stored frozen at 80 C. for discordant analysis using the Abbott RealTime HIV-1 test. Discordant samples for which there was insufficient sample to complete repeat testing were removed from the final analysis.

(68) Table 9 below shows the distribution of the 198 samples remaining for the final analysis. Given that quantitation by the Roche COBAS AmpliPrep/COBAS TaqMan HIV-1 test has an accuracy of 0.3 log, the comparative data were stratified into 4 categories: undetectable (no HIV target present); between 500 and 2,000 copies (i.e. within the 30.3 log 10 accuracy of the Roche test); <500 and >2,000 copies/ml.

(69) TABLE-US-00009 TABLE 9 Distribution of HIV viral load measurement by SAMBA test compared to Roche COBAS AmpliPrep/COBAS TaqMan test Viral load measurement by Roche CAP/CTM (copies/ml) undetectable <500 500-2,000 >2,000 SAMBA >1,000 0 8 6 24 (copies/ml) <1,000 114 43 3 0

Overall Concordance: 96%

(70) It can be seen in this patient population that the leukodepletion filter was effective in removing the CD4+ cells, since the overall concordance between the Roche test, using plasma, and the SAMBA test, using leukodepleted whole blood, was 96% (190/198).