Articles

Detection of antibodies to human immunodeficiency virus type 2 (HIV-2) in blood donor sera using United States assay

methods for anti-HIV type 1

K. SAZAMA, I.K. KURAMOTO, P.V. HOLLAND, A-M. COUROUCB, D. GALLO, AND C.V. HANSON

Twelve serum samples from French blood donors that were uniformly reactive in tests for antibody to human immunodeficiency virus type 2 (anti-HIV-2) also were reactive in 92 to 100 percent of tests with three anti-HIV type 1 (anti-HIV-1) enzyme- linked immunoassays currently in widespread use for donor screening in the United States. Supplemental tests for anti-HIV-1 on these anti-HIV-2-reactive samples dif- fered in their responses. All samples reacted in a licensed anti-HIV-1 Western blot, but there was an atypical band near the p41 position, which could be a clue to the fact that this result was a cross-reaction with anti-HIV-2. A recombinant immunoblot gave an indeterminate result for anti-HIV-1 in all 12 samples. A local immunofluo- rescence assay for anti-HIV-1 reacted with 92 percent of the samples, but a com- mercial one detected only 58 percent. TRANSFUSION 1992;32:39&401.

Abbrevlatlons: anti-HIV-1 = antlbodles to human Immunodeflclency vlrus type 1; antl- HIV-2 = antl-HIV type 2; CDC = Centers for Dlsease Control; EIA = enzyme Immunoas- say; IFA = Immunofluorescence assay: RlBA = recomblnant lmmunoblot assay.

AT THE TIME OF THE licensure of a specific test for antibodies to human immunodeficiency virus type 2 (anti- HIV-2) in July 1990, blood banking professionals rec- ommended against requiring such testing. There was lit- tle likelihood of identifying anti-HIV-Zreactive donors in the United States (US) who would not be detected by testing for antibodies to HIV type 1 (anti-HIV-1) alone, because of the high frequency of cross-reactivity with anti-HIV-2 in anti-HIV-1 testing. 1-3 However, in 1988, at a time when only 612 cases of HIV-2 infection had been reported ~ o r l d w i d e , ~ the rate of detection of anti- HIV-2 in asymptomatic persons (such as blood donors) by anti-HIV-1 assays was reported to vary from 48 to over 90 percent.5 Because of this wide range in the re- ported ability of anti-HIV-1 testing methods to detect anti-HIV-2, additional data derived by using current techniques may be useful in reinforcing the decision to depend on anti-HIV-1 testing along with other exclu- sionary methods to avoid HIV-2 transmission by blood donors in the US.

From the Sacramento Medical Foundation Center for Blood Re- search, Sacramento, California; the Institut National de Transfusion Sanguine, Paris, France; and the Viral & Rickettsia1 Disease Labo- ratory, Division of Laboratories, California Department of Health Services, Berkeley, California.

Received for publication January 3,1991; revision received Novem- ber 20, 1991, and accepted November 25, 1991.

We tested 12 well-characterized6 anti-HIV-Zreactive serum samples obtained from asymptomatic French blood donors, who were qualified to donate at the time the sample was taken, by using 12 currently available US anti-HIV-1 and anti-HIV-2 testing methods. Seven methods were for detection of anti-HIV-1 and five for anti-HIV-2.

Materials and Methods Blood donor sera

The 12 liquid samples were identified by and sent from the National Institute of Blood Transfusion (Institut National de Transfusion Sanguine, Paris, France,) via air shipment in 5- mL tubes to the Center for Blood Research, Sacramento, Cal- ifornia. The sdmples had been thawed and refrozen several times prior to their receipt in California, and they were stored in the frozen state (-70°C) until assayed.

Enzyme immunoassay The three anti-HIV-1 enzyme immunoassays (EIAs), two

from Abbott Laboratories (North Chicago, IL) and one from Genetic Systems (Seattle, WA), and the two anti-HIV-2 assays (Genetic Systems; United Biomedical, Lake Success, NY) are listed in Table 1. The tests were performed according to the directions of the manufacturers.

Western blot The two anti-HIV-1 Western blot kits, one using viral lysate

(DuPont, Wilmington, DE) and one using HIV-1 recombinant

398

TRANSFUSION 1992-Vol. 32. No. 5 DETECTING ANTI-HIV-2 IN BLOOD DONORS 399

Table 1. Comparison of detection of anti-HN-2-reactive donor sera by currently avellable anti-HN-1 and anti-HN-2 testing systems

Testing method and manufacturer

Positive sera Indeterminate sera Negative sera

Number Percent Number Percent Number Percent

Anti-HIV-1 testing systems Enzyme-linked immunoassay

Abbott 3A11 Abbott 2A95 Genetic Systems

In-houset Electro-Nucleonics

Western blotS DuPont

lmmunofluorescence assay

12 100 0 0 0 0 11 92 0 0 1. 8 11 92 0 0 1* 8

11 92 0 0 1. 8 7 58 0 0 5 42

12 100 0 0 0 0 Recombinant lmmunoblot assay

Chiron 0 0 12c 100 0 0

Anti-HIV-2 testing systems

Genetic Systems United Biomedical

In-house+

Enzyme-linked lmmunoassay

lmmunofluorescence assay

12 100 0 0 0 0 12 100 0 0 0 0

12 100 0 0 0 0 Western blot

In-house' 12 100 0 0 0 0 Genetic Systems 12 100 0 0 0 0

Not the same donor sample. t Viral 8 Rickettsia1 Disease Laboratory, California Department of Health Services. * Using Centers for Disease Control criterla for Interpretation (see text). 8 All reacted at p24 and p31; none reacted at gp4l or gp120.

immunoblot antigens (RIBA, Chiron Corporation, Emeryville, CA), and the anti-HIV-2 Western blot kit (Genetic Systems) were used according to the manufacturers' directions. An in- house Western blot, using HIV-2 viral lysate (Genetic Sys- tems), was used as described previously.'

For the DuPont Western blot, the presence of any two bands at p24, gp41, or gp120/160 was considered positive. The pres- ence of any bands other than the ones defining a positive re- action were considered to be indeterminate. Absence of all bands was interpreted as negative.

For the Chiron anti-HIV-1 RIBA, interpretation was based on the following criteria. 1) Two bands of 1 + or greater reac- tivity, neither band being p24 or p31, defined a positive re- action. 2) The presence of any band that reacted at 1+ or greater, but did not meet the criteria for a positive reaction was indeterminate. 3) Reactivity of less than 1 + on all bands, including no reactivity, was a negative result.

We interpreted the anti-HIV-2 Western blots, both the Ge- netics Systems and the in-house methods, as follows. l ) A positive reaction required the presence of gp140 or gp36 and p26. 2) An indeterminate reaction resulted from the presence of any viral or nonviral bands not meeting the criteria for positivity. 3) When no bands were present, the test was negative.

Immunofluorescence assay The anti-HIV-1 immunofluorescence assay (IFA) kit (Elec-

tro-Nucleonics, Columbia, MD) was used according to the manufacturer's directions. We performed two in-house IFAs as described previously,l using HIV-1- or HIV-Zinfected cells (in France, called human T-lymphotropic virus type IIIB and

lymphadenopathy-aiated virus type 2, respectively) on slides for anti-HIV-1 and anti-HIV-2 determinations, respectively.

Results All 12 anti-HIV-Zreactive samples also reacted when tested

by two anti-HIV-1 methods widely used in the US (Abbott second-generation [3A11] EIA and DuPont Western blot), as well as by all five anti-HIV-2 methods tested (Genetic Sys- tems, EIA and Western blot; United Biomedical, EIA; and the in-house anti-HIV-2 IFA and Western blot). Three of the anti- HIV-1 testing systems detected 92 percent (11/12) of these samples (two commercially available EIAs and the in-house IFA [Table 11); a fourth system, the Electro-Nucleonics anti- HIV-1 FA, detected 58 percent (7/12) of these anti-HIV-2- reactive samples.

None of the 12 sera were defined as positive by the Chiron anti-HIV-1 RIBA (Table 1) according to the criteria provided for this RIBA, however, all samples showed bands at p24 and p31 and would be considered indeterminate. Thus, the RIBA results may indicate that these anti-HIV-2 samples are not typ- ical anti-HIV-1 sera and need further evaluation, especially with anti-HIV-2 assays.

Although the Centers for Disease Control (CDC) criteria8 for a positive Western blot using HIV-1 viral lysate were met when these 12 sera containing anti-HIV-2 were tested, the blot pattern differed from that Seen with sera containing anti-HIV-1. The samples with anti-HIV-2 reacted at neither gp120 nor gp41 with the HlV-1 proteins on the blot; the band close to the gp41 position may reflect antibody to p40, a precursor of core pro- teins (Fig. 1). Thus, the presence of anti-HIV-2 was suspected

400 SAZAMA ET AL. TRANSFUSION

Vol. 32, No. 5-1992

FIG. 1. Reactivity of anti-HIV-2-reactive blood donor sera (4-1 1) compared with anti-HIV-1 controls (1-3: strong, weak, negative) to HIV-1 viral lysate proteins. Test sera 4 through 11 show atypical bands near the gp41 position.

on the anti-HIV-1 Western blot by the unusual band pattern. Current CDC Western blot criteria do not permit distinction between anti-HIV-1 and anti-HIV-2. The interpretation is de- signed to establish a serodiagnosis of HIV infection and not to differentiate between HIV-1 and HIV-2 antibodies.

Discussion Because the rate of anti-HIV-2 seroreactivity is not

yet calculable in US blood donors (No instances were discovered after screening of 28,140 samples, derived from at least 20,000,000 blood donors, that were found to be repeatably reactive on anti-HIV-1 EIA.9), the effect that adding anti-HIV-2 or even combination anti-HIV-l/ 2 testing will have on blood safety cannot be accurately predicted. The first US case of acquired immune defi- ciency syndrome due to HIV-2 was reported in 1988 and was not attributed to blood transfusion.’O Since then, only 18 adults have been found to be anti-HIV-2 pos- itive, and only one of those persons was identified through a blood donation.” To date, no transfusion transmission has been found, probably because the currently used anti- HIV-1 screening assays are detecting and excluding all, or nearly all, HIV-Zinfected donors. The few cases of HIV infection identified to date in transfusion recipients have been due to anti-HIV-1 that was not detected by the current anti-HIV-1 EIAs.12

There are no current published data regarding the sen- sitivity of current US-licensed anti-HIV EIAs in detect-

ing possible HIV-2 infection in US blood donors. The study reported here, though derived from a small number of samples, provides information obtained by testing sera from French blood donors who underwent a process sim- ilar to that for US donors, namely, donor interviewing and actual blood donation. These data, admittedly biased by the preselection of anti-HIV-Zreactive samples with seroreactivity on anti-HIV-1 French test kits, is intended to provide some factual basis in support of the current policy of not requiring the addition of separate anti-HIV-2 testing in the We showed that at least one anti-HIV-1 screening test currently in widespread use was apparently capable of detecting all 12 anti-HIV- 2-reactive samples (all confirmed positive by the li- censed anti-HIV-1 Western blot)13 and that other screen- ing procedures detected more than 92 percent. Use of an anti-HIV-1/2 combination test may increase the sen- sitivity of existing assays, but the additional number of donors (infected with HIV-2) detected by this assay may be minuscule.

Our data, though based on only 12 samples, are the only data specific for blood donors using currently avail- able US-licensed screening and confirmatory testing sys- tems. Other recently published studies that tested anti- HIV-Zreactive sera either included no blood donors14J5 or did not clearly differentiate blood donors from other populations in their study.16 George et a1.,16 who studied 55 seroreactive samples, detected 85 to 100 percent of HIV-Zreactive samples from 13 apparently healthy peo- ple (blood donors, clients of prenatal clinics), a rate inexplicably higher than that for the remaining 42 “ill” persons in their study.

A note of caution should be sounded regarding the interpretation of anti-HIV-1 supplemental assays when one encounters or suspects an anti-HIV-Zreactive sam- ple. For those blood centers relying on “confirmation” by the Electro-Nucleonics IFA procedure, up to 42 per- cent of EIA anti-HIV-2-seroreactive samples may not be detected. Additionally, the Chiron anti-HIV-1 RIBA (in its present configuration and using current criteria) may not be sufficient for verification of anti-HIV-2, since bands to only p24 and p31 were found (and bands to gp41 or gp120 are necessary to define a sample as anti- HIV-positive), which resulted in an indeterminate inter- pretation. Using a conventional Western blot prepared from HIV-1 viral lysate material, the anti-HIV-Zcon- taining sera tested met the CDC criteria for confirmation but gave atypical bands, which suggests that the samples actually contained anti-HIV-2 and not anti-HIV-1. Sus- pected anti-HIV-Zreactive samples should, nonetheless, be confirmed with a bona fide anti-HIV-Zspecific sup- plemental test.

The uniform detection of anti-HIV-2 in all anti-HIV-2 test systems studied strengthens the anticipated reliabil- ity of combined anti-HIV-1/2 testing methods currently

TRANSFUSION 1992-Val. 32, No. 5 DETECTING ANTI-HIV-2 IN BLOOD DONORS 40 1

undergoing clinical trial. A combined assay would elim- inate the need for a separate test for anti-HIV-2 and could add a (very small) incremental margin of safety to the health care system in the United States.17

References 1. AABB Technical Bulletin 90-3. Screening for HIV-2 infection.

Arlington: American Association of Blood Banks, April 30,1990. 2. Licensure of a screening test for antibodies to HIV-2. Meeting of

the Blood Products Advisory Committee, March 15, 1990. Be- thesda: Food and Drug Administration, 1990.

3. Foucault C, Lopez 0, Jourdan G, Fournel JJ, Perret P, Gluckman JC. Double HIV-1 and HIV-2 seropositivity among blood donors (letter). Lancet 1987;2:165-6.

4. Horsburgh CR, Holmberg SD. The global distribution of human immunodeficiency virus type 2 (HIV-2) infection. Transfusion

5 . Denis F, Leonard G, Sangare A, et al. Comparison of 10 enzyme immunoassays for detection of antibody to human immunodefi- ciency virus type 2 in West African sera. J Clin Microbiol 1988;26: 1000-4.

6. Courouce AM, and the Retrovirus Study Group of the French Society of Blood Transfusion: Barin F. Baudelot J, Chamaret S, et al. HIV 2 infection among blood donors and other subjects in France. Transfusion 1989;29:368-70.

7. Gallo D, Diggs J, Shell G, Dailey P, Hoffman M, Riggs J. Com- parison of detection of antibody to the acquired immune deficiency syndrome virus by enzyme immunoassay, immunofluorescence, and Western blot methods. J Clin Microbiol 1986;23:1049-51.

8. Interpretation and use of the Western blot assay for serodiagnosis of human immunodeficiency virus type 1 infections. Morbid Mor- tal Weekly Rep 1989;38(S-7):1-7.

9. Surveillance for HIV-2 infection in blood donors-United States, 1987-1989. Morbid Mortal Weekly Rep 1990;39:829-31.

10. AIDS due to HIV-2 infection-New Jersey. Morbid Mortal WmY, Rep 1988;37:33-5.

1988;28: 192-5.

11.

12.

13.

14.

15.

16.

17.

Update: HIV-2 infection-United States. Morbid Mortal Weekly Rep 1989;38:572-74, 579-80. Ward JW, Holmberg SD, Allen JR, et al. Transmission of human immunodeficiency virus (HIV) by blood transfusions screened as negative for HIV antibody. N Engl J Med 1988;318:473-8. Pau CP, Granade TC, Parekh B, et al. Misidentification of HIV-2 proteins by western blots (letter). Lancet 1991;337:616-7. Schumacher RT, Howard J, Ayers L, et al. Cross-reactivity of anti-HIV-2 positive in serum in U.S. FDA licensed screening tests for anti-HIV-1 (abstract). Sixth International Conference on AIDS: AIDS in the nineties, from science to policy, vol3. San Francisco, University of California, 1990:245. Rubsamen-Waigmann H. Briesen HV. Maniar JK, Rao PK, Scholtz C, Pfeutzner A. Spread of HIV-2 in India (letter). Lancet

George JR, Rayfeld MA, Phillips S, et al. Efficacies of US Food and Drug Administration-licensed HIV-I-screening enzyme immu- noassays for detecting antibodies to HIV-2. AIDS 1990,4321-6. Revised recommendations for the prevention of human immuno- deficiency virus (HIV) transmission by blood and blood prod- ucts-Section I, Parts A & B. Bethesda: Food and Drug Administration, December 5, 1990.

1991;337:550-1.

Kathleen Sazama, MD, JD, Director and Associate Medical Direc- tor, Center for Blood Research, Sacramento Medical Foundation, 1625 Stockton Boulevard, Sacramento, CA 95816-7089. [Reprint requests]

I. Ken Kuramoto, MT(ASCP), Research Coordinator, Center for Blood Research, Sacramento Medical Foundation.

Paul V. Holland, MD, Medical Director and CEO, Sacramento Medical Foundation.

Anne-Marie Couroud, PhD, Director, Institut National de Trans- fusion Sanguine, Paris, France.

Dana Gallo, Public Health Microbiologist, Viral & Rickettsial Dis- ease Laboratory, Division of Laboratories, California State Department of Health, Berkeley, California.

Carl V. Hanson, PhD, Director, Viral & Rickettsial Disease Laboratory.