|Home | About | Journals | Submit | Contact Us | Français|
A simplified lateral-flow assay for the detection of antibodies to HIV using magnetic-bead conjugates and multibranched peptides from both HIV-1 and HIV-2 was developed. Magnetic immunochromatography testing (MICT) uses a standard lateral-flow platform that incorporates magnetic-bead conjugates for quantitative measurement of the magnetic field distortion associated with the bound magnetic conjugate (reported as adjusted relative magnetic units [MAR]). The results of the optimized MICT assay were compared to standard enzyme immunoassay (EIA) and Western blotting (WB) results using a blinded 649-member panel of specimens from the United States, Cameroon, and West Africa. The panel was comprised of samples from individuals infected with various HIV-1 subtypes (n = 234) or HIV-2 (n = 65) and HIV-seronegative specimens (n = 350). Additionally, 13 HIV-1 seroconversion panels (total specimens = 85), a worldwide panel containing seven of the major circulating HIV-1 subtypes (n = 18), an HIV-2 panel, an HIV-1/HIV-2 mixed panel, and 100 prospective specimens were tested with completely concordant results. Assay reproducibility (observed MAR) for both intra- and interrun testing was excellent, with coefficients of variation of <12%. MICT can provide a rapid, low-cost method of determining HIV antibody status requiring no subjective interpretations.
The use of rapid HIV antibody-screening assays has permitted the global expansion of HIV testing into rural, nonlaboratory settings and has significantly increased the number of individuals that have been screened. These assays are primarily designed as lateral-flow formats that use colored detection reagents, such as colloidal gold or selenium, conjugated to HIV antigens or to proteins that bind to specific human immunoglobulins, such as protein A or G (10). The tests are rapid, inexpensive, and stable over a broad temperature range and are simple to perform, requiring no additional equipment (1). Although they are generally easy to interpret by visual inspection, there are reports of false-reactive devices, particularly in low-prevalence settings (6). Investigations to determine the sources of these problems have not identified any particular trait other than the subjective interpretation of the results (3). Diagnostic-instrument manufacturers have responded by developing lateral-flow strip readers that use reflectance, fluorescence, and magnetic measurements to provide a more precise and objective result (7, 14). Such devices could also be used to develop quantitative lateral-flow tests for a variety of diagnostic applications.
For rapid HIV testing, lateral-flow tests primarily use HIV-1 subtype B antigens from the immunodominant transmembrane region to capture HIV-specific antibodies. Current commercial assays have been shown to perform well with specimens from individuals infected with other HIV-1 subtypes, even group O (4, 23). However, how these assays perform during early seroconversion with non-subtype B infections has not been assessed, since panels for other subtypes are unavailable. Furthermore, the development of assays for HIV incidence determinations has shown that the immune responses to subtype B antigens are not equivalent across HIV subtypes (21) and that multisubtype antigens are more effective at establishing comparable incidence measurements in international cross-sectional surveys. Thus, detection of antibodies generated to a variety of HIV subtypes might be improved through the use of a broader antigenic mix (chimeric recombinant proteins or synthetic peptides) and/or a more effective antigenic presentation (multibranched peptides), both of which have proven useful in diagnostic assays for HIV and other infectious agents (13, 15, 19, 22).
The purpose of this study was to develop a quantifiable lateral-flow test for the detection and differentiation of antibodies to HIV-1 and HIV-2 using magnetic-bead markers (magnetic immunochromatography test [MICT]). In order to maximize HIV-specific antibody capture, multibranched peptides (MBP) for both HIV-1 and HIV-2 (22) were evaluated for use in a single assay that could detect and differentiate HIV infections. The assay was tested using a 649-member panel of specimens from diverse global locales, 13 HIV-1 seroconversion panels, a panel representing seven of the primary HIV-1 subtypes, an HIV-2 panel, an HIV-1/HIV-2 mixed panel, and 100 prospectively tested specimens. The results were compared to those of standard serological tests, including enzyme immunoassays (EIAs), Western immunoblot assays, and a rapid immunoassay that is licensed by the U.S. Food and Drug Administration to differentiate HIV-1 and HIV-2 infections. The MICT HIV antibody assay is compatible with available low-cost equipment, is simple to perform, and produces results in 20 min.
A blinded panel was prepared to evaluate the performance of the optimized MICT assay using specimens collected in CDC epidemiological surveys under CDC-approved protocols (IRB-1896 and IRB-1367), as well as specimens obtained from commercial sources. The panel consisted of 649 serum/plasma specimens from the United States, Cameroon, and West Africa with the following characteristics: 350 nonreactive, 234 HIV-1, and 65 HIV-2. All of the panel members were tested by EIA (Bio-Rad HIV-1/2 + O; Bio-Rad Laboratories, Hercules, CA) and Western blotting (WB) (Bio-Rad HIV-1 Western blot [Bio-Rad Laboratories] or HIV-1 Cambridge Biotech Western blot [Maxim Biomedical, Inc., Rockville, MD]), which served as the reference standard. The HIV-2 antibody-positive specimens were validated using an HIV-2-specific WB (MP Diagnostics, Singapore) and the Multispot assay (Bio-Rad), which differentiates HIV-1 and HIV-2 antibody reactivities. An additional 100 specimens submitted for reference HIV testing were prospectively tested by MICT and compared to the routine EIA/WB results. One HIV-2 panel (PRF201) (n = 10), one HIV-1/2 mixed panel (PRZ201), and a panel that included multiple HIV-1 subtypes and recombinant forms (WWRB350) (n = 18) (all from Boston Biomedica, Inc. [now SeraCare, Inc., Milford, MA]) were also included in the evaluation. The ability of the assay to detect early HIV-1 seroconversion was evaluated using 13 HIV-1 seroconversion panels (panels H, I, J, V, Y, Z, AB, AC, AF, AJ, AN, AU, and BI; SeraCare). The MICT data generated for the purchased panels were directly compared to the serological data provided by the manufacturers.
MBP were prepared for both HIV-1 and HIV-2 using an automatic peptide synthesizer (model 432; Applied Biosystems, Foster City, CA) employing standard Fmoc chemistries. Both antigens consisted of eight branches, each containing 18 amino acids selected from the immunodominant regions of the transmembrane glycoproteins from each virus plus a 4-amino-acid spacer (Fig. (Fig.1).1). In order to accommodate the variety of expressed epitopes in the gp41 region of HIV-1, more than 1 amino acid was incorporated at two locations in the synthesis so that the final HIV-1 MBP contained the peptide sequences of four of the known major HIV-1 sequence variations in this region. No modifications were made to the HIV-2 MBP. The MBP were purified by reverse-phase high-performance liquid chromatography (HPLC) and then lyophilized and stored desiccated at 20 to 25°C until they were used.
Protein A (Zymed Laboratories, Inc., South San Francisco, CA) was conjugated to superparamagnetic beads (300 nm; Ademtech, Pessac, France) using standard cross-linking chemistries, as previously described (27). Briefly, the carboxylated beads were mixed with N-hydroxy-sulfosuccinimide (sulfo-NHS) and 1-ethyl-3-[3-dimethylaminopropyl]carbodiimide (EDC) in morpholineethanesulfonic acid (MES)-buffered saline, pH 4.7, to form an amine-reactive sulfo-NHS ester, so that when protein A was added to the beads, a stable amide bond was formed between the protein and the magnetic particles. Residual active coupling sites were blocked using 5% amicase for 30 min at 37°C before final washing and storage at 4°C in buffer supplied by Ademtech. The prepared magnetic conjugates were expressed as mass per unit volume, and the conjugates prepared here contained 10 μg/μl.
The lateral-flow strips were prepared by reconstituting the HIV-1 and HIV-2 MBP in an equal mixture of 0.1% trifluoroacetic acid and 10 mM sodium citrate buffer (pH 3.0) to achieve a final MBP concentration of 1 mg/ml. Protein A was also reconstituted at 25 μg/ml using 0.1 M sodium carbonate/bicarbonate buffer and served as an assay control of the magnetic conjugate performance. All reagents were simultaneously dispensed onto nitrocellulose membrane (Hi Flow Plus HF18004; Millipore Corporation, Medford, MA) using an Isoflow Dispenser (Imagene Technology, Hanover, NH). The striped membrane was dried overnight at 32°C in a vacuum oven, followed by blocking with 10 mM phosphate buffer containing 0.08% bovine serum albumen for 5 min and drying as before. The magnetic lateral-flow assay membrane was assembled in a standard configuration on a backed adhesive card (Adhesives Research, Inc., Glen Rock, PA) using a glass fiber pad (Millipore Corp.) with no blocking agent as the sample application pad and a medium-capacity wicking pad (Millipore Corp.). Both pads overlapped the nitrocellulose membrane by ~2 mm. The assembled strip was covered with a clear plastic adhesive (Adhesives Research, Inc.) to ensure effective contact of the assay components. The completed assay preparation was cut into individual 5-mm strips using a Kinematic 2360 programmable shearer (Kinematic Automation, Twain Harte, CA), and each strip was incorporated into a unique plastic housing (MagnaBiosciences, Inc., San Diego, CA) as previously described (27).
Prior to evaluating the panel that contained a larger number of HIV antibody-positive specimens, 10 well-characterized specimens (3 HIV-1, 3 HIV-2, and 4 HIV nonreactive) were selected to be used in experimental protocols to optimize the concentrations of the MICT reagents and to develop an appropriate assay procedure. The specimens were tested in a variety of dilutions from 1:10 through 1:3,200, and the mass of the magnetic-bead conjugate was varied from 10 to 30 μg per test. After the establishment of the optimized assay protocol, an assessment of nonspecific reactivity to the HIV-1 and HIV-2 MBP antigens was performed using a subset of the established HIV-nonreactive specimen panel (n = 265) described above.
Serum and plasma specimens were diluted at 1 μl into 100 μl of 10 mM phosphate buffer (pH 7.2) containing 40% chicken serum. Two microliters of protein A magnetic beads (20-μg mass) was added to the tube, and the solution was mixed by vortexing it for 2 s. The specimen was incubated for 2 min, and then 100 μl of the mixture was pipetted into the sample port of the cassette. The reaction lines of the magnetic assays were read at 20 and 40 min using the assay development system (ADS) (MagnaBiosciences, Inc.), which quantitatively measures the magnetic field induced by the captured magnetic particles (14). The amplitude of the detected signal is directly proportional to the amount of magnetic material in the assay reaction zone. The instrument software calculates a relative magnetic unit (MAR) based on the detected magnetic signals, and this value was used to evaluate the performance characteristics of the assay.
Three well-characterized sera—one HIV-1 antibody positive, one HIV-2 antibody positive, and one HIV nonreactive—were selected to evaluate the quantitative capability of the MICT assay to detect and to differentiate HIV-specific antibodies. The reproducibility of the MICT assay was established using multiple runs (n = 10) of the three specimens in triplicate over a 10-day period. Data were analyzed over the period by averaging the quantitative results and determining the coefficients of variation (CV) for both intrarun and interrun data.
Using the 10 specimens (3 HIV-1, 3 HIV-2, and 4 nonreactive) described above, the optimal dilution of specimens was determined to be between 1:100 and 1:800. Strongly reactive HIV-1 and HIV-2 specimens could be diluted 1:5,000 with almost no decrease in detectable MAR, and they were still strongly reactive at dilutions of 1:50,000 (data not shown). Although 10 μg of the magnetic-bead conjugate allowed effective identification of HIV antibody-positive specimens, the overall assay sensitivity was improved by increasing the magnetic-bead mass to 20 μg with no impact on assay specificity (data not shown). Quantitative data for the 10 specimens tested using the optimized protocol are shown in Table Table1.1. Both HIV-1 and HIV-2 antibody-positive specimens had MAR values in excess of 200, while the four HIV-nonreactive specimens displayed no reactivity (0 MAR) at either of the HIV peptide antigen lines. Cross-reactivity of the HIV-1 and HIV-2 antibodies to the MBP in these specimens was not noted after 20 min of incubation.
Nonspecific reactivity of the HIV antibody-nonreactive specimens (n = 265) to both the HIV-1 and HIV-2 MBP was low; all of the specimens had MAR values of <10 (Fig. (Fig.22 ). Based on these data, a cutoff of 15 MAR was established that would clearly distinguish both HIV-1 and HIV-2 antibody-positive samples from HIV antibody-nonreactive specimens. This cutoff was determined using the average MAR of nonreactive specimens on the HIV-1 MBP (1.9 MAR) plus 3 standard deviations (3.5 MAR) and was used throughout the rest of the analysis.
The performance characteristics of the optimized MICT assay were evaluated using the specimen panel (n = 649) described in Materials and Methods. The data were compared to results from the standard EIA/WB testing (reference standard) (Table (Table2).2). The MICT rapid test had excellent sensitivity and specificity and correctly identified the 234 HIV-1 antibody-positive specimens, the 65 HIV-2-antibody positive specimens, and the 350 nonreactive specimens. The quantitative MAR values of the MICT assay clearly differentiated HIV-1 from HIV-2 antibody-positive specimens based on the MAR value. In fact, cross-reactivity between the two MBPs was very low, with MAR readings of antibodies to one of the HIV types normally below the cutoff of 15 MAR on the MBP to the other type (data not shown). All 10 members of the HIV-2 panel (PRF201) were also reactive to only the HIV-2 MBP and did not have any cross-reactivity to the HIV-1 MBP. The mixed panel of HIV-1 and HIV-2 specimens (PRZ 201; SeraCare, Inc.; n = 15) was also run to further validate the ability of the assay to detect and to differentiate HIV-1 and HIV-2 antibodies. The HIV-1 (n = 7) and HIV-2 (n = 6) antibody-positive specimens were detected with MAR values of >150 for HIV-1 and >300 for HIV-2. The antibody-negative and WB-indeterminate specimens in the panel had MAR values of <15 to both of the HIV MBP antigens (data not shown). The MICT assay also detected all of the members of the worldwide panel that included seven of the major group M subtypes, as well as some of the circulating recombinant forms (Table (Table3).3). The MAR values observed on the panel ranged from 157.0 for one of the subtype A specimens to over 1,800 MAR for one of the AG recombinant forms, and cross-reactivity to the HIV-2 MBP was not observed.
The ability to detect HIV-1 seroconversion was assessed using the 13 commercial panels. The MICT data were compared to data from first-, second-, and third-generation EIAs and Western blot results derived from the panel inserts that were provided by the panel supplier (Table (Table4).4). Of the 85 specimens in the collection, 43 were antibody positive by the Abbott HIVAB HIV-1/HIV-2 (rDNA) EIA, a third-generation HIV test that both captures and detects HIV antibodies using HIV-specific antigens. Only 32 of these specimens were denoted HIV antibody-positive by WB analysis. The MICT assay performed better than WB, detecting 34 of the specimens, which was fewer than were detected by the third-generation EIA but was consistent with or better than the first-generation (Abbott HIV-1) and second-generation (Genetic Systems HIV-1/2) assays, which are based on the same indirect immunoassay principles.
Samples arriving for routine screening in the HIV reference laboratory with sufficient volume remaining were further tested by the MICT assay, and the results were compared to those of the current testing algorithm of EIA/WB (Bio-Rad HIV-1/2 + O EIA and Bio-Rad HIV-1 WB). Of the 100 specimens that arrived in the laboratory, 18 were HIV-1 antibody positive, three were HIV-2 antibody positive, 60 were nonreactive, and 19 were EIA reactive-WB indeterminate by the standard algorithm. Since the ultimate status of indeterminate specimens was not known, MICT testing was not performed on those samples. The MICT results for the remaining 81 specimens were fully concordant with the standard EIA/WB testing results.
The reproducibility of the MICT assay was established using multiple runs (n = 10) of the three specimens in triplicate over a 10-day period (Table (Table5).5). MAR readings for both of the HIV-reactive specimens were high (HIV-1 average = 717; HIV-2 average = 838), while the nonreactive sera had no detectable MAR against either of the HIV branched-peptide antigens. CV for all of the measurable MAR for HIV-1 antibodies were less than 8.2% and were independent of the incubation time (20 versus 40 min). CV for detection of the antibodies to HIV-2 were slightly higher (~12%) and were also time independent.
The use of rapid HIV tests has now become routine in many areas of the world due to their simplicity, low cost, and excellent performance characteristics (8, 24). In the United States, six rapid HIV assays are now approved by the U.S. Food and Drug Administration, and four of these have been categorized as “waived” testing under CLIA 1988, where the qualifications of the testers are minimal. These traits have led to expansion of testing into a variety of nonlaboratory venues worldwide and have allowed extended outreach to at-risk populations (6, 17, 20, 25). Although some studies have shown that nonlaboratorians can effectively perform these waived tests (12), others have detected problems with HIV rapid testing, as well as with simple tests for other infectious diseases or clinical syndromes (9). One of the primary problems has been the subjective nature of the test result, which must be individually interpreted by the test performer. This particular characteristic has led to interpretive problems that could not be readily resolved despite investigation (3). The MICT assay presented here alleviates these concerns by having a quantifiable, objective result that can be used for assay interpretation. The assay retains the advantages of existing rapid HIV antibody detection assays but uses an inexpensive, stand-alone magnetic reader to detect and differentiate the presence of antibodies to HIV-1 and HIV-2 in a single device.
MBP have been used for a number of years to improve the sensitivity and specificity of diagnostic assays. The multivalent and multiepitope antigens used in the MICT assay described here permitted the detection of all of the HIV antibody-positive specimens in the test panel, which consisted of samples collected in diverse global locales and likely included members from individuals infected with various HIV-1 group M subtypes. Members of the global panel included specimens from individuals infected with subtypes A to G and some circulating recombinant forms. All were detected by the MICT assay, indicating that these antigens provided good coverage for antibodies elicited by the major group M subtypes of HIV-1. Similarly, the antigens used for the detection of HIV-2 specimens also performed well on the smaller HIV-2 antibody-positive specimens included in the panel.
One unique feature of this evaluation was the low cross-reactivity that was observed between the HIV-1 and HIV-2 MBP antigens in HIV antibody-positive specimens. Serological differentiation of HIV-1 and HIV-2 infections was initially difficult in areas where both viruses circulate due to the cross-reactivity of the antibodies elicited by the two viruses (5). However, some of the peptide-based rapid assays have been formatted to differentiate HIV infections, and some of these have been shown to be effective in this regard (2, 18). The MBPs used in the MICT assay provided excellent ability to both detect and differentiate HIV-1 and HIV-2 antibodies, with only a few cross-reactive specimens observed (data not shown).
The quantitative nature of the MICT assay could improve sensitivity and specificity through the quantitative determination of the magnetic field associated with true binding events. The MAR values determined for both MBP antigens were essentially zero for HIV antibody-negative samples, with only a few specimens having any detectable MAR. Thus, a very low cutoff value could be established, which would eliminate subjective readings and would provide an effective method to distinguish HIV antibody-positive specimens from nonreactive specimens. Furthermore, the low cutoff measure would theoretically permit detection of early antibody-reactive specimens and improve sensitivity. In the emerging antibody responses found in HIV seroconversion panels, the detection of HIV-specific antibodies by MICT was essentially identical to that detected by other indirect immunoassay methods, such as EIA and Western blotting. Since these panels have no specific time intervals during collection, it was not possible to determine if the MICT assay afforded any earlier detection of HIV-specific antibodies in these particular panels, but the fact that a few specimens were detected earlier than with WB was suggestive of improved sensitivity.
HIV antibody assays have been improved by changing the test principle from indirect immunoassays to specific antibody capture assays that use HIV antigens as both capture and detection reagents by using the multivalency of the HIV-specific antibody. This format allows the detection of multiple antibody classes and has been shown to improve early detection of HIV antibodies (Table (Table4;4; Abbott results) (16). However, in order to achieve these results, the specimen volume had to be increased. For large-scale screening, these platforms have been combined with HIV-1 p24 antigen detection, which has reduced the window for early HIV infection detection by 10 to 15 days (11, 26). A rapid HIV-1 p24 assay has already been developed using the MICT platform (27), and this assay could be combined with the MICT antibody assay described here. Modification of the HIV-1/2 antibody detection assay to the antibody sandwich method is already in development. Future plans include the combination of the antibody sandwich assay with the HIV-1 p24 antigen assay to provide a rapid, nonsubjective assay for field detection of HIV-1 and HIV-2 infections.
The findings and conclusions in this paper are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Published ahead of print on 21 April 2010.