Journal of Medical Virology 38:49-53 (1992)

Evaluation of a New Generation Synthetic Peptide Combination Assay for Detection of Antibodies to HIV-I, HIV-2, HTLV-I, and HTLV-I1 Simultaneously

Xiang Zhang, Niel T. Constantine, Jaya Bansal, Johnny D. Callahan, and Vincent C. Marsiglia Department of Pathology, University of Maryland School of Medicine (X.Z. , N.T.C., J.B., J.D.C.) and BDC Laborator.y, Baltimore City Health Department ( V.C.M.1, Baltimore, Maryland

A new generation combination test (Detect-Plus, IAF BioChem, Montreal, Canada) based on syn- thetic peptides for HIV-1, HIV-2, HTLV-I, and HTLV-II was compared with three routine com- mercial screening assays and confirmatory as- says to determine its sensitivity and specificity and to evaluate it as a substitute screening method. Samples from 356 sexually transmitted disease (STD) patients were tested by the four screening tests. All initially reactive samples were retested in duplicate by the corresponding EIA and repeatedly reactive samples were con- firmed by Western blots for HIV-1, HIV-2, and HTLV-1/11. The confirmed positives detected by each screening assay were HIV-1 (23/356,6.46%), HIV-2 (1 1/356,3.09%), and HTLV-1/11 (5/356,1.4%). The new generation Detect-Plus test produced only two results (2/356, 0.56%) that were pre- sumed to be false-positives in comparison to the screening tests, but the OD/CO values were just slightly high (1.5 and 1.9). There were no false- negative results, indicating that the sensitivity of the new combination test was excellent (100%). Compared with routine retroviral EIA assays, the test is easy to perform-the total time require- ment is only 2 hr and there is no need for incuba- tion equipment. The OD/CO values were very high when samples were positive, making even visual interpretation possible. We conclude that this new combination assay is an excellent screening method for detection of antibodies to the human retroviruses, and may be particularly useful for screening blood for transfusion and in epidemiological investigations. rk) 1992 Wiley-Liss, Inc.

~~

0 1992 WILEY-LISS, INC.

KEY WORDS combination test, retroviruses, STD, EIA, commercial screen- ing assay, HIV, HTLV, synthetic peptide test

INTRODUCTION Assays to test for antibodies to human immunodefi-

ciency viruses (HIV) and human T-cell lymphotropic viruses (HTLV) are widely available in number and variety. HIV and HTLV tests, licensed by the Food and Drug Administration (FDA), for detecting antibodies to the retroviruses most often use partially purified, dis- rupted virus derived from infected human cell lines [Petricciani, 1985; Gallo et al., 1986; MMWR, 19881. Few FDA-licensed recombinant and synthetic peptide- based tests exist for use in screening for antibodies to HIV. Because fdse-positive results occur, viral lysate assays may yield a problematic number due to nonspe- cific reactions when testing blood donors [MMWR, 19891. Busch et al. [1991] showed that when using a licensed EIA to test for HIV-1 antibodies, 92% of sam- ples that initially reacted on EIA and 79% of repeatedly reactive donor samples were determined to be the re- sult of nonspecific reactions when compared to confir- matory testing and clinical follow-up. Concern for transmission-related infection by the HIVs and the HTLVs has dictated the need for exquisitely sensitive screening tests. Therefore, interest in the development of sensitive serologic assays to detect retroviral infec- tion for blood bank use has been increased dramatically [Gout e t al., 1990; Fretz et a]., 19911. Currently, blood banks in the United States are required by the FDA to screen donated blood for antibodies to HIV-1 and HTLV-1/11, By June 1992, the FDA will be recommend- ing that blood be additionally tested for antibodies to HIV-2. Therefore, unless combination assays are used, a t least three separate tests will be needed in an effort to help prevent blood-borne transmission by the human retroviruses.

Recently, recombinant DNA-derived and synthetic peptide antigens have been used to develop sensitive and specific HIV-1 and HIV-2 assays [Thorn et al.,

Accepted for publication February 21, 1992. Address reprint requests to Dr. Neil T. Constantine at The

University of Maryland, Department of Pathology, Room 758, 10 S. Pine St., Baltimore, MD 21201.

50 Zhang et al.

1987; Gnann et al., 19871. These reagents contain the immunodominant regions of selected viral antigens which are not contaminated with human cellular pro- teins; thus these antigens theoretically should result in more sensitive and specific assays than those based on viral lysates. In addition, combination assays using re- combinant or synthetic peptide antigens have been used outside the United States for simultaneous detec- tion of HIV-1 and HIV-2 infections [Berry et al., 1991; Callahan et al., 1990; Constantine et al., 19891. Re- cently, the first combination EIA (Genetic Systems, Inc.) to detect HIV-IIHIV-2 has been licensed by the FDA. This combination assay is based on combined vi- ral lysate antigens from both viruses.

In this report, we describe an evaluation of a new generation combination test ( Detect-Plus, IAF Bio- Chem, Inc., Montreal, Canada). The Detect-Plus test is a solid phase enzyme immunoassay (EIA) which uti- lizes a mixture of synthetic peptides for the detection of antibodies to any or all of the retroviruses that cause human disease: HIV-1, HIV-2, HTLV-I, and HTLV-11. We evaluated the sensitivity and specificity of this as- say, and comment on its applicability for use as a sub- stitute screening method.

MATERIALS AND METHODS Serum samples from a total of 356 individuals were

included in the evaluation. All samples were collected from a clinic for sexually transmitted diseases in Balti- more, Maryland. Testing was performed blindly on coded samples. Each sample was tested by FDA-11- censed EIAs for detection of antibodies to HIV-1, HIV-2 (Genetic Systems, Inc.), and HTLV-1/11 (Cellular Prod- ucts, Inc.), and by the new generation IAF BioChem, Inc., combination test (Detect-Plus), according to the protocols established by the manufacturers. Specimens reactive by any test were retested in duplicate before confirmation by Western blots for HIV-1 (Bio-Rad), HIV-2 (Genetic Systems), and HTLV-1/11 (Bio-Rad). The criterion for a sample to be considered as reactive by EIA was an optical density/cutoff ratio (OD/CO) of > 1 .O. Criteria for positivity by Western blots included, for HIV-1, reactivity to any two of p24, gp41, or gpl20/ 160 as recommended by ASTPHLDKDC; for HIV-2, reactivity to any two envelope antigens ( W.H.O. crite- rion); and for HTLV-1/11, reactivity to one gag (p19 or p24) and one envelope antigen (gp21 or gp46) as recom- mended by W.H.O. Infection by HTLV-I and HTLV-I1 could not be differentiated using these ELISA and Western blot assays.

Samples that exhibited discrepant results between any of the assays were retested by all methods to ensure that technical error had not occurred and to determine if a result were possibly misclassified by the homolo- gous assay. To evaluate the accuracy of the new combi- nation test and the other three routine EIA tests, the Western blot results were considered as the reference. Sensitivity and test efficiency of the Detect-Plus test were calculated in reference to the results of the other

three screening tests, while specificity was determined based on results of the Western blot assays. Test indices were caIcuIated as foltows

True-positives True-positives + false-negatives Sensitivity = x 100.

True-negatives True-negatives + false-positives Specificity = x 100.

True-positives + true-negatives Efficiency = True-positives + false positives x 100.

+ true negatives + false negatives

RESULTS Of the 356 sera tested, there were 82 repeatedly reac-

tive results by the screening assays combined, repre- senting 24 (6.7%) samples that were confirmed positive by at least one of the Western blot assays. Twenty-eight (7.9%) were repeatedly reactive by the Detect-Plus test, ofwhich 24 (86%) of these were subsequently confirmed as positive either by the HIV-1, HIV-2, or HTLV-1/11 Western blot assays; two produced indeterminate re- sults by the Western blot. This yielded a specificity for the Detect-Plus test of 98.8% (332/336) in comparison with Western blot tests, and 99.4% (330/332) when the two indeterminate results were not included in the cal- culation (true status unknown). Twenty-three of 25 HIV-1 ELISA repeatedly reactive samples were con- firmed by the HIV-1 Western blot, and were reactive by the Detect-Plus test. The two samples which could not be confirmed by the HIV-1 Western blot did produce indeterminant profiles (reactivities to only p24 and p51 on one sample, and high background on the other) and were also reactive by the Detect-Plus test. Eleven of the HIV-2 EIA reactive sera were confirmed by the HIV-2 Western blot. All HIV-2 EIA repeatedly reactive, West- ern blot confirmed samples were also among the group of confirmed HIV-1 positives. Similarly, four of the five HTLV-1/11 confirmed positive samples were also con- firmed positive for HIV-1 by Western blot (dual reac- tors), while one was positive only by the HTLV-1/11 Western blot assay. All five of these were reactive by the Detect-Plus test. With the exception of one sample which was confirmed positive for HTLV-1/11, all HTLV EIA reactive sera were also HIV-1 EIA reactive. One of the 5 non-confirmed HTLV reactive samples exhibited an HTLV-1/11 indeterminant Western blot profile (reac- tivity to p24 only) and was among the HIV-1 and HIV-2 confirmed positives. This latter sample was strongly reactive by the Detect-Plus assay.

Table I indicates the percentages of initially reac- tives to repeatedly reactive samples for each of the four screening tests. Table I1 compares the EIA methods with regard to their specificities and test efficiencies. Two samples which were repeatedly reactive by the Detect-Plus test, but non-reactive by the other three screening assays, could not be confirmed by any one of

Combination Assay for HIV-1, HIV-2, HTLV-I, HTLV-I1

TABLE I. Comparison of Number of EIA Reactive Results Between Initial and Repeat Testing

by Each Test* HIV-1 HIV-2 HTLV-1/11 Detect-Plus

Initial 40 21 17 38 Repeat 25 19 10 28 Percent 63 90 59 74 *n = 356.

51

units to be tested for evidence of Chagas disease in the near future, and possibly other agents [Goldsmith, 19901. The necessity to perform all of these tests re- quires significant reagent cost and an enormous amount of technologist time.

Combination tests capable of detecting antibodies to more than one of these agents are available [Constan- tine et al., 19911. Most common are those to detect in- fection by HIV-1 and HIV-2 simultaneously [Constan- tine et al., 1989; Lacroix et’ al., 19891. In the past 2 years, at least 5 commercial companies have marketed these tests, but only two commercial HIV-1/HIV-2 com- bination EIA have become FDA-licensed and thus available for clinical use in the United States. Most combination assays are based on recombinant or syn- thetic peptide antigens and are available in dot-blot or EIA methods. Some are capable of differentiating and identifying the specific agent [Brolden et al., 19911, while others only demonstrate the presence of antibod- ies to one or the other agent. More recently, Pharmacia, Inc. has introduced a combination test designed to de- tect infection by HIV and hepatitis B surface antigen at the same time, but this has not been widely adopted. The use of combination assays for detection of HIV-1/ HIV-2 has found utility in geographic locations where both viruses are present (i.e., West Africa), but less so in other areas. However, HIV-2 infections are spread- ing and have been detected unexpectedly in certain parts of the world, including East Africa and the Per- sian Gulf region [Constantine et al., 19901. At least 36 countries have reported the presence of HIV-2 infection [DeCook and Brun-Vezinet, 19891, and i t is clear that some HIV-2 infected sera are not detected by HIV-1 screening assays [Onorato et al., 1991; Constantine et al., 19891.

The application of a combination assay capable of detecting infection by four retroviruses simultaneously provides an excellent means to increase the efficiency of the testing process, while being cost effective. However, the sensitivity of the test must not be compromised to achieve these goals since it is the screening assays that must have the ability to protect the blood supply. In addition, the test must not produce a significant num- ber of false-positive results, in order to maintain cost effectiveness and testing efficiency. Finally, the test must be easy to perform and be adaptable for large- scale screening.

IAF BioChem’s Detect-Plus combination assay is based on chemically synthesized peptides having from 23 to 39 amino acid residues, which correspond to re- gions of the gp120 and gp41 of HIV-1, gp36 of HIV-2, and the gp46 and p19 of HTLV-I and HTLV-I1 viral components. The gp41 and gp36 peptides each include the two natural cysteine residues and they have been made in cyclic form. When these peptides are used in their cyclic rather than the linear form a gain in sensi- tivity occurs [Lacroix et al., 19891. The main advantage of the Detect-Plus test is its ability to simultaneously detect antibodies to the four retroviruses that cause human disease. This efficiency can have a major impact

TABLE 11. Comparison of Assay Performances HIV-1 HIV-2 HTLV-1/11 Detect-Plus

EIA + 25a 19 10 28a WB + 23 11 5 24 Percent of 92% 58% 50% 86%

confirmed Specificity 99.4% 97.7% 98.6% 99.4%b Efficiency 99.4% 97.8% 98.6% 98.9% a T w ~ samples produced indeterminant results by HIV Western blot. bSamples producing indeterminate results were not included in the calculation.

TABLE 111. Comparison of Average OD and O D E 0 of Reactive Samples by Each Screening Method

HIV-1 HIV-2 HTLV-1/11 Detect-Plus Average OD 1.7 1.2 0.9 >3.0a Average 5.7 6.0 2.6 >10.0b on/co aMaximum value recorded by spectrophotometer. bBased on average OD of 3.0, actual value much higher (see text).

the Western blot assays. The OD values of these two samples were 0.58 and 0.46, and the O D E 0 were 1.9 and 1.5 respectively. In contrast, most (86%) of samples that were reactive by the Detect-Plus test had OD val- ues >3.0, the maximum value that could be recorded by the spectrophotometer. If 3.0 were used as the OD for these samples, the mean OD/CO for all samples was >lo . It is reasonable to assume that the mean OD/CO for the Detect-Plus test is actually much higher than 10. Table 111 indicates a comparison of the OD and OD/CO values of samples which were reactive accord- ing to each of the screening assays. For samples which were repeatedly reactive by each of the screening as- says but which were not confirmed by Western blot, the mean O D E 0 values were generally low: HIV-1 = 2.7, HIV-2 = 2.8, HTLV-1/11 = 1.4, Detect-Plus = 1.7. Table IV compares the operational characteristics and cost of each assay.

DISCUSSION Currently, it is required for blood banks to screen

potential blood donors for evidence of a variety of infec- tious agents in order to protect the blood supply. Among the tests required are those for syphilis, hepatitis B and hepatitis C, HIV-1, and HTLV-1/11, In addition, the FDA is recommending that by June of 1992 all units should be tested for HIV-2. There is also discussion for

52 Zhang et al.

TABLE IV. Characteristics of Each EIA Screening Assay HIV-1 HIV-2 HTLV-1/11 Detect-Plus

Antigen Viral lysate Viral lysate Viral lysate Synthetic peptides Total time 2 hr 3 hr 3 hr 1.5 hr Incubation 37" c 37" c 37" c Room temperature Cost per test 1 3 2 3

(dollars)a aCosts provided by manufacturers, but are dependent on quantity purchased.

on decreasing the time required for testing samples prior to transfusion and in epidemiological surveys. Al- though the test does not identify which particular anti- bodies are present, its utility as a general screening method is evident. Another advantage is the lower number of false-positive reactions. By assuming that errors can potentially occur with each test performed, the use of three tests should be expected to produce three times the number of errors. Therefore, false-posi- tives due to technical reasons (pipeting, carry-over, sample dilution, washing, etc.) should be decreased by two-thirds since only one and not three tests will be performed (currently only one screening test is re- quired to detect antibodies to both HTLV-I and HTLV- 11). The occurrence of biologic false-positives should be decreased also since the test is based on synthetic pep- tides rather than viral lysates, which contain contami- nating cellular proteins. In fact, many investigators consider synthetic peptide-based tests to be used as con- firmatory rather than as screening tests. Another sig- nificant advantage of the Detect-Plus test is that all incubations are performed at room temperature, obvi- ating the need for water baths or electrical incubators, and decreasing the necessity to transport plates from the bench to other areas of the laboratory. OD readings and OD/CO ratios are usually high with truly positive samples (excellent resolution), suggesting that visually interpreted results may be possible. Finally, the cost of $3 per test is quite inexpensive in comparison to the cost required to perform three separate retroviral as- says, perhaps one-half to one-third less. Also, technolo- gist time, and the cost of expendables (tips, tubes, etc.) would be decreased by the same, resulting in additional savings.

In the present study, the Detect-Plus combination test has exhibited an excellent sensitivity (100%) and a quite acceptable level of false-positive results (specific- ity of 99.4%) in comparison with other widely accepted screening assays. Even if the two samples that pro- duced repeatedly reactive results by the HIV-1 and De- tect-Plus tests and that were indeterminant by the HIV-1 Western blot had been included in the calcula- tion, the specificity of' the Detect-Plus would be 98.8%. Our investigation did, however, focus on a population a t high risk for retroviral infection, where reactive samples were most likely from truly infected individu- als. However, regardless of the population tested, the sensitivity of the test would not be affected.

Although the Detect-Plus test produced two definite false-positive results, the resultant specificity was higher than two of the routine screening tests. In addi- tion, the ODiCO ratios were low, which is typical of false-positive reactions. These two specimens subse- quently were tested by the manufacturer and our re- sults were confirmed. The percentage of repeat reac- tives following initial reaction by the Detect-Plus test was slightly higher but similar to two of the other screening assays, and higher than the rate reported recently by three HIV-1i2 combination assays [Berry et al., 19911. The Detect-Plus test also exhibited an excel- lent ability to identify samples that subsequently were confirmed by one of the Western blots (percentage of confirmed results). OD and ODIC0 ratios of confirmed positive samples were much higher by the Detect-Plus test when compared to the other three screening tests, indicating better resolution of truly reactive samples. Although i t is not recommended, visual interpretations of results could be accomplished in situations where electricity was not available.

Many of our samples exhibited reactivity to more than one of the retroviruses by screening and confirma- tory assays. Although these reactions may indicate dual infections, which would be consistent with this high-risk population, cross-reactivity to the retrovi- ruses is likely. In either case, the Detect-Plus test de- tected all as repeatedly reactive, the purpose for which i t was developed.

In conclusion, the new generation IAF BioChem's Detect-Plus combination assay is capable of detecting antibodies to four retroviruses simultaneously, exhib- its an excellent sensitivity, produces a low number of false-positive results, and is cost and time effective. It possesses the performance and operational characteris- tics required for use as a substitute screening method in blood banks and for epidemiological investigation. Pre- viously, we have reported on the use of combination assays to detect antibodies to HIV-1 and HIV-2 simul- taneously [Constantine et al., 1989; Callahan et al., 19901, and now suggest that the use of combination assays capable of detecting antibodies to all retrovi- ruses can be a n effective means of screening for retrovi- ral infections.

REFERENCES Berry NJ, Salker R, Contreras M, Barbara JAJ, Tedder RS (1991): A

comparison of four enzyme immunoassays for the simultaneous

Combination Assay for HIV-1, HIV-2, HTLV-I, HTLV-I1

detection of HIV-1- and HIV-2-specific antibody. Journal of Viro- logical Methods 34:91-100.

Brolden PA, Ruden U, Ouattara AS, The G, Solver E, Trojnar J, Wahren B (1991): Specific synthetic peptides for detection of and discrimination between HIV-1 and HIV-2 infection. Journal of Ac- quired Immune Deficiency Syndromes 4:952-958.

Busch MP, Amad Ze, McHugh TM, Chien D, Polito AJ (1991): Reliable confirmation and quantitation of human immunodeficiency virus type 1 antibody using a recombinant-antigen immunoblot assay. Transfusion 3:129-137.

Callahan JD, Constantine NT, Watts DM (1990): More support for the use of HIV combination assays. Journal of Virological Methods 28:217-2 18.

Constantine NT, Callahan JD, Watts DM (1989): Time for HIV-11 HIV-2 combination tests? Journal of Virological Methods 26:219- 222.

Constantine NT, Sheba F, Fox E, Woody JN, Abbatte EA (1990): Serological evidence for HIV-2 in East Africa. International Jour- nal of Sexually Transmitted Diseases and Acquired Immune Defi- ciency Syndrome 1:53-54.

Constantine NT, Callahan JD, Watts DM (1991): HIV Testing and Quality Control: A Guide for Laboratory Personnel. Family Health International, Durham, North Carolina. ISBN 0-939704-07-2.

DeCook KM, Brun-Vezinet F (1989): Epidemiology of HIV-2 infection. Acquired Immunodeficiency Syndrome 3:S89395.

Fretz C, Jaulmes D, Fournel JJ, Jullien AM, Gout 0, Gluckman JC, The GD (1991): HTLV-I transmission and myelopathy induced by blood transfusion. Transfusion 31:379.

Gallo D, Diggs JL, Shell GR, Dailey PJ, Hoffman MN, Riggs J L (1986): Comparison of detection of antibody to the acquired im- mune deficiency syndrome virus by enzyme immunoassay, irnmu- nofluorescence, and Western blot methods. Journal of Clinical Mi- crobiology 23:1049-1051.

53

Gnann JW, McCormick JB, Mitchell S, Neison JA, Oldstone MBA (1987): Synthetic peptide immunoassay distinguishes HIV type 1 and HIV type 2 infections. Science 237:1346-1349.

Goldsmith MF (1990): Ways to assess test costs vary. Journal of the American Medical Association 263:1750.

Gout 0, Baulac M, Gessain A, Semah F, Saal F, Peries J , Cabrol C, Fretz CF, Laplane D, Sigaux F, The GD (1990): Rapid development of myelo-pathy after HTLV-I infection acquired by transfusion during cardiac transplantation. The New England Journal of Med- icine 322:383388.

Lacroix M, Chalifour R, Stern M, Dionne G (1989): Comparative per- formance of cyclic versus linear peptides in an ELISA for HIV-1 and HIV-2 specific antibodies. V International Conference on AIDS, Montreal, Abstract W.B.P. 110.

MMWR (1988): “Update: Serologic Testing for Antibody to Human Immunodeficiency Virus.” Vol. 36, pp 833-840.

MMWR (1989): “Interpretation and Use of the Western Blot Assay for Serodiagnosis of Human Immunodeficiency Virus Type 1 Infec- tion.” Vol. 38 ISuppl71, pp 1-7.

Onorato I, Schable C, Sprvill C, McCray G , Allen D, George R (1991): Warning of the next epidemic: Surveillance HIV-2 infection in high-risk populations, United States, 198%1990. VII Interna- tional Conference on AIDS, Florence Italy, Abstract MC 3068.

Petricciani J C (1985): Licensed tests for antibody to human T-lympho- tropic virus type 111. Sensitivity and specificity. Annals of Internal Medicine 103:726729.

Thorn RM, Beltz GA, Hung CH, Fallis BF, Winkle S, Cheng KL, Marciani DJ (1987): Enzyme immunoassay using a novel recombi- nant polypeptide to detect human immunodeficiency virus env antibody. Journal of Clinical Microbiology 251207-1212.