- Email: [email protected]
Different types of false positive anti-HIV reactions in patients on haemodialysis
Immunology Letters, 22 (1989) 35-40
Elsevier IMLET 01253
Different types of false positive anti-HIV reactions in patients on haemodialysis E s z t e r U j h e l y i 1, G e o r g e Ffist 1, G~ibor Illei 1, l~va G y o d i l, K~troly N a g y 2, Ferenc D. Toth 3, B61a Biiki 1, M a n f r e d P. D i e r i c h 1, Vlastimil M a y e r 5, G y 6 r g y G a l 6, J~inos M a k o 7 a n d S u s a n R. H o l l a n 6 INational Institute of Haematology and Blood Transfusion, Budapest, Hungary; 2Institute of Isotopes, Hungarian Academy of Sciences, Budapest, Hungary; 3Institute for Microbiology, University Medical School, Debrecen, Hungary; 4Institute for Hygiene, University Medical School, Innsbruck, Austria; 5Institute for Virology, Slovak Academy of Sciences, Bratislava, Czechoslovakia; 6Blood Bank, Szeged, Hungary; 7Haemodialysis Center, Semmelweis University Medical School, Budapest, Hungary
(Received28 February 1989; revision received6 April 1989; accepted 11 April 1989)
1. Summary Serum samples o f 589 haemodialysis patients were screened for HIV antibody by ELISA methods. O f these, 36 samples were found to be repeatedly reactive. None of the 36, however, could be confirmed by competitive enzyme immunoassays and Western blot; therefore, they were considered to be false positive. The sera could be divided in two groups. The sera o f Group 1 were designated as the usual type o f false positivity, caused most probably by anti-lymphocyte antibodies. In 19 sera, however, a special type o f false positivity was found. These sera reacted strongly with the plates coated with the supernatants o f HIV-infected cells but not with those of uninfected H9 cells. Three and two sera showed, respectively, positive immunofluorescence reaction with the HIV-infected, but not with the uninfected, H9 and CEM cells. Reactivity to HIVinfected H9 cells could be adsorbed from a part o f these samples with lesser amounts o f HIV-infected than uninfected H9 cells. This special type of false positivity was observed frequently (7/65) in patients who rejected a kidney graft. These findings suggest that this type of anti-HIV false positivity is due to
Key words: AIDS; HIV infection;Haemodialysis;Falsepositivi-
ty; ELISA Correspondence to: Dr. E. Ujhelyi,National Institute of Haema-
tologyand Blood Transfusion, 1113Budapest, P.O. Box 44, Hungary, H-1502.
antibodies reacting with cellular antigens present in HIV-infected but not in uninfected lymphocytes. Their appearance seems to be associated with the immunological activation occurring at graft rejection. 2. Introduction Multiple blood transfusions place patients on haemodialysis at increased risk of HIV infections. That is why many investigations have been performed on the incidence of HIV antibodies [1-3]. The rate of false positive enzyme-immunoassay (EIA) results in patients on haemodialysis was observed to be much higher than in blood donors in most studies [1, 4, 5]. Little is known, however, about the causes o f the development of false positive reactions in patients on haemodialysis. We have been screening patients on haemodialysis for the incidence of HIV antibodies since 1986. These measurements, together with the permanent screening of HIV seronegative haemophiliacs, have been performed as part o f the program to monitor the safety o f the nation's blood supply. Testing is repeated every year. No confirmed positive case was found; false-positivity, however, was observed to be frequent in our studies, as well. The objective of the present work was to analyse the cause of these falsepositive anti-HIV reactions. Besides the usual wellknown false positivity due to the presence o f antilymphocyte antibodies, in some sera a special type o f false positivity was observed: these sera did not react with the normal lymphocytes or with the super-
0165-2478 / 89 / $ 3.50 © 1989 ElsevierSciencePublishers B.V.(Biomedical Division)
35
natants of the cultures of these cells; they reacted strongly, however, with HIV-infected lymphocytes or the supernatants of HIV-infected cells. This special type of false positivity was observed mostly in patients who had rejected a kidney graft previously.
100/xl of adsorbed (and, as control, non-adsorbed) sera were added to 106 HIV-H9 cells, and tested by the indirect immunofluorescence assay. The results were evaluated microscopically, by counting the percentage of the stained cells.
3. Materials and Methods
3.4. Other assays
3.1. Patients tested
Anti-HLA A, B, C antibody assay was performed by the standardized NIH lymphocytotoxicity assay [6]. Adsorption of sera by pooled human platelets: 100 #1 of packed human platelets pooled from 100 donors was added to 100/~1 of serum, mixed and incubated overnight at room temperature. Adsorption was repeated with the same amount of platelets for 2 h at room temperature. Reactivity of the adsorbed and non-adsorbed sera was checked by a lymphocytotoxicity assay using separated T and B lymphocytes [7]. Antinuclear antibodies (ANF) were titrated as outlined in Ref. [8].
Three series of patients were tested: investigations on patients of series 1 (163 serum samples of 93 patients treated in Szeged) and series 2 (81 serum samples of 78 patients treated in Budapest) were performed in 1986. Patients of series 3 (418 serum samples of 418 patients treated in different dialysis centres in Budapest) were studied in 1987-1988. 3.2. H I V serological tests
Two types of enzyme immunoassay (EIA) kits were used: (1)Hg-based kits." Virgo (Electronucleonics), Vironostika (Organon Teknika), Anti-HTLVIII kit of Sorin-Biomedica consisting of HTLVIIIcoated (HTLVIII) plates and H9 supernatantcoated (H9) plates, Welcozyme (Wellcome Diagnostics). (2) CEM-based kits: Genetic (Genetics System), Elavia (Pasteur Diagnostics) containing HTLVIII-coated plates and CEM supernatantcoated (control) plates. Each kit was used strictly according to the: recommendations of the manufacturers. The following assays were used for confirmation of the EIA results: (i) Immunofluorescence test." in the experiments performed in 1986, home-made slides with HTLVIII-infected H9 cells (HTLVIII/H9) and LAV-infected CEM cells (LAV/CEM) were applied. In 1978-88 a commercial IFA kit (Virimmun, FRG) was used. (ii) Western blot assay: commercial kits (Du Pont de Nemours and Bio-Rad) were applied for these studies. 3.3. Adsorption o f sera with HIV-infected and noninfected 1t9 cells
Different amounts (0.5× 106, 1.5 × 106, and 5.0×106) of HIV-infected H9 cells (HIV-H9) or non-infected H9 cells (H9) were added to 100 ttl serum and incubated overnight at room temperature. 36
4. Results
4.1. Results o f H I V serology in series 1 and 2
In these series, anti-HIV screening was performed using the Virgo kit. Fifteen of 171 samples (8.8%) were found to be repeatedly reactive. New serum samples were taken from the patients found to be reactive at the initial screening, 3, 4, and 7 months
TABLE 1 Reactivity of the sera tested repeatedly positive at screening in other anti-HIV assays (Series 1 and 2). Test
No. o f sera tested
Positive
Indeterminate
Negative
Virgo Organon Sorin Elavia Genetic Welcozyme
15 15 15 15 15 15
15 13 10 0 0 0
0 0 0 0 0 0
0 2 5 15 15 15
IFA (CEM) IFA (H9)
9 10
2 3
l 0
6 7
WB
15
0
0
15
later (series 1) and 2 months later (series 2). These samples were also reactive in the Virgo test, although a fluctuation in the extent of reactivity could be frequently seen. The 15 repeatedly Virgo-reactive serum samples were tested by other anti-HIV screening and confirmatory assays (Table 1). The 3 ELISA assays (Virgo, Vironostika, Sorin) correlated strongly with each other. Thirteen and 10 sera were reactive in the Vironostika and Sorin kits, respectively. By contrast, no reactivity was observed in any sample in the competitive Welcozyme EIA and in two CEM-based ELISA assays (Elavia, Genetic). LAV/CEM and HTLVIII/H9 immunofluorescent assays (IFA) were performed in 9 and 10 samples, respectively. In 2 samples, a positive reaction was found in both types of IFA test. A further sample tested positive in the HTLVIII/H9 IFA, while in the LAV/CEM assay an ambiguous result was obtained. When the same 3 sera were tested by IFA with uninfected CEM or H9 cells, no reactivity was found. All the 15 samples tested negative in Western blot. 4.2. Study of the repeatedly ELISA-reactive serum samples in ELISA plates coated with supernatants of infected and non-infected 1t9 cells Eighteen sera from the 15 patients who tested Virgo-positive at the initial screening were studied using the HTLVIII and H9 plates of the Sorin kit. According to the recommendations of the manufacturer, if the HTLVIII/H90.D. ratio exceeds 3.0, the sample should be considered as true positive, while in the case of a ratio of less than 3.0 the HTLVIII reactive sera should be considered false positive. According to these criteria, 12 and 6 sera, respectively, were calculated to be true positive and false positive. On the basis of these findings, two types of false positivity could be distinguished in patients on haemodialysis: sera with a usual type of false positivity (UFP) react with the superantants of both the HIVinfected and non-infected H9 cells, whereas sera with a special type of false-positivity (SFP) react only with the supernatants of the HIV-infected H9 cells.
4.3. Incidence of UFP and SFP in transplanted and non-transplanted patients of series 3 418 patients on haemodialysis were tested for the presence of HIV antibodies using the Sorin kit. Repeated reactivity was observed in 21 samples (5.0°70). Each sample tested negative in the Welcozyme competitive EIA and in the lot of the Vironostika kit used at that time, and none of them could be confirmed by Western blot. On the basis of the results obtained with the HTLVIII and H9 plates, 14 and 7 sera were considered to be UFP and SFP, respectively. Both SFP and UFP results were highly reproducible. For example, in three repeated measurements, HTLVIII/H90.D. ratios in 3 SFP sera were 4.45-3.36-4.22, 3.87-5.17-4.22, and 5.71-4.05-4.87. Out of the 7 SFP sera, 5 were negative in Western blot, whereas in one sample an isolated p55, and in another one an isolated p66 band, was observed. Five samples tested IFA-negative, in 2 samples a weak fluorescence was observed in both the HIV-infected and uninfected H9 cells. One in seven samples showed a positive ANF reaction. Sixty-four of the 418 patients tested in series 3 were redialysed following rejection of a kidney graft. Interestingly enough, 14/21 repeatedly Sorinreactive sera, including all the seven SFP sera, were from this group of patients (Table 2). Fourteen of 64 (22070) of patients who had been grafted showed UFP or SFP, whereas in patients who had not been transplanted HIV false-positivity rarely occurred (7/354 = 2%). There were some differences between the groups of grafted patients with SFP and UFP reactions. SFP patients were on haemodialysis for a longer time (mean 5.6 (3-10) vs. 3.7 (2-6) years) and rejected the kidney graft more quickly (time from transplantation to re-haemodialysis: 7.7 (0-22) months vs. 12 (0.5-34) months) than UFP patients. 4.4. Studies on the mechanism of the special type
of faise-positivity Seven sera from series 1 and 2 found to be SFP, as well as 85 anti-HIV-negative sera, were tested for the presence of anti-HLA antibodies. All the seven SFP sera contained anti-HLA antibodies, mostly in high titres. Fifty-one anti-HIV negative samples were negative for anti-HLA antibodies as well. HLA 37
TABLE 2 Occurrence of the usual (UFP) type and special type (SFP) of false-positivity in transplanted and non-transplanted patients of haemodialysis (Series 3). Results in Sorin kit
No. o f sera tested from patients total
ever transplanted
%
Negative UFP SFP
397 14 7
50 7 7
12 50 100
Total
418
64
15
antibodies could be detected in 34 samples. In 11 samples their reactivity - measured on a lymphocyte panel of 20 donors - exceeded 50°7o. Six SFP sera were adsorbed by human platelets. Adsorption did not decrease the apparent anti-HIV reactivity o f the sera. By contrast, the same treatment fully or almost completely removed cytotoxicity of the sera against isolated T cells of all subjects tested. Reactivity for the isolated B cells decreased, but a marked cytotoxic effect could be observed in most adsorbed serum samples. Eight SFP samples were adsorbed with the same amounts o f HIV-H9 and H9 cells (Table 3). Seven
TABLE 3 Adsorption with infected (H9-HIV) or not infected (H9) H9 cells o f sera o f patients showing a special type of false-positivity. Serum
86696 86702 86705 86694 86679 86810 87123 87643
% of the FITC-labelled cells treated with sera not adsorbed
adsorbed with H9 cells
adsorbed with H9-HIV cells
90 30 100 50 20 80 90 50
60 < 10 70 30 0 20 60 30
20 < 10 70 10 0 20 15 0
Only results o f adsorption with 1.5 x 106 cells are shown. Similar results were obained using lower or higher a m o u n t s o f cells for adsorption.
38
samples reacted in the IFA assay with _ 50°/o of the H I V - H 9 cells. Adsorption with both the HIV-H9 and the H9 cells decreased the percentage of the FITC-labeled cells in each sample. In the case of four sera, the same amount of HIV-H9 cells adsorbed considerably more antibodies than the uninfected H9 cells. 5. Discussion
Our present findings, in agreement with many previous observations [1, 4, 5], indicate that repeatedly positive reactions in anti-HIV ELISA assays are more common in chronic haemodialysis patients than in blood donors. Altogether, such reactions were observed in 36 cases out of 589 patients tested (6.1°70). Neither Western blot nor competitive EIA confirmed the ELISA reactivity of the sera. Therefore the sera were considered to be false-positive. The initial screening of the patients was performed in two different H9-based kits. In the first two series o f investigations, other H9-based kits like the Vironostika kit o f Organon Teknika also showed repeated reactivity with these samples. Some repeatedly ELISA-reactive sera reacted in the indirect immunofluorescence assay as well. False-positive anti-HIV reactions can develop due to several reasons. It seems that the most common cause o f false-positivity is the presence of antibodies against antigens of the lymphocyte cell line, H9, used to produce the AIDS virus. Since in the competitive EIAs, not anti-IgG but HIV antibody conjugates are used, the anti-lymphocyte antibodies cannot interfere with the results of these assays. Lymphocyte-derived antigens can be present in the viral lysates applied for coating the ELISA plates [9-11]. In our patients, too, anti-lymphocyte antibodies and primarily those in high titres occurred more frequently in false-positive than in other sampies. However, such antibodies could also be detected in high titres in several samples which tested negative in HIV ELISA assays. Moreover, antibodies responsible for the development o f the false-positive reactions could not be removed from the sera by adsorption with platelets. This observation excludes the possible role o f the anti-HLA A, B, or C antibodies in the development of the observed falsepositivity. H9 ceils were HLA-typed by Weiss et al. [12] and
CD4+ Iwmphocwte activated bw an Immune stimulus
a p p e a r a n c e /!'f an antigen on t h e s u r f a c e of cell or
Increase In the density of a pre-existing
antigen
I transfusion of activated lymphocytes to patients on haemodlalysi5 and.,'or activation
at g r a f t
of %he I m m u n e
/
system
rejection
development of au~oantibodles
in t h e
petlents reacting with the aotlvated T ceils
and giving a special t y p e of f a l s e p o s l t l v l t y In HIV ELISR assa61s Fig. 1. Hypothetical mechanism of development of the special type of anti-HIV false-positivity in patients on haemodialysis.
found to be A1, Bw62, Bw6, Cw3, DR4, DQw3. Since anti-DR4 antibodies were found to be able to induce false-positive anti-HIV ELISA results [10], the best candidate for causing false positive results in haemodialysis patients would be anti-DR4 antibodies. In the case of samples designated in the present study as usual type of false positives (UFP) this explanation seems to be acceptable, too: these sera reacted approximately to the same extent in plates coated with the supernatant of the HIV-infected and non-infected H9 cells. The other group of sara of haemodialysis patients, designated as SFP (special type of false-positivity), however, reacted in a very different way: (1) they showed a high reactivity (O.D. value) on the plates of the Sorin kit coated with the supernatants of the HIV-infected H9 cells but did not react or showed a weak reaction in the control plates of the same kit coated by the concentrate of cellular material shed by uninjected H9 cells. (2) Only a part of the sera in which false-positive immunofluorescence reaction was observed reacted with the uninfected H9 cells in the IFA test. Interestingly enough, the same phenomenon was observed in the case of HIVinfected and uninfected CEM cells. (3) In 4/7 SFP sera, HIV-infected H9 cells adsorbed much more
anti-HIV reactivity than the uninfected H9 cells. It is remarkable that SFP occurred mainly among patients who had been transplanted and had rejected one or two kidney grafts. In series 3, all the seven sera with SFP were from such patients, while 50°70 of patients with UFP were never transplanted. Besides patients on haemodialysis, SFP was rarely (in 9 cases out of the first 800000 blood donations tested) observed among blood donors as well. Interestingly enough, all the 9 donors were females, who developed antibodies possibly during pregnancies (unpublished observations). On the basis of these findings a hypothesis could be raised (Fig. 1). According to this assumption, infection with HIV results in the appearance of an antigen on the surface of the H9 cells and other lymphocytes, which is either not present on the uninfected cells, or increases the density of a preexisting antigen. Immunological activation can induce the expression of the same antigen in lymphocytes and (considering the great number of transfusions administered to patients on haemodialysis) by chance such activated lymphocytes could have been transfused. This transfusion can lead to antibody formation against this antigen in the patients. Alternatively, immune activation which occurs at graft rejection or other type of activation can also induce the formation of autoantibodies against this antigen, which may be responsible for the special type of anti-HIV false positivity. Several recent findings indicate that HIV infection induces or increases the expression of some cellular antigens in lymphocytes [13-15]. Any of these antigens, or other unrecognized ones, could be the target of the antibodies inducing anti-HIV false positivity. Experiments aiming to characterize this antigen are underway in our laboratories. References [1] Peterman, T. A., Lang, G. R., Mikes, N. J., Solomon, S. R., Schabl¢, C. A., Feorino, P. M., Britz, J. A. and Allen, J. R. (1986) J. Am. Med. Assoc. 255, 2324-2326. [2] Goldman, M., Liesnard, C., Vanderwegheim, J.-L., Della, N., Touissaint, C., Sprecher, J., Cogniaux, J. and Thyry, L. (1986) Br. Med. J. 293, 161-162. [3] Villaneuva, S. Y., de Medina, M., Perez, G., Reddy, K. R., Jeffers, L., Watkins, E, Goldsmith, C. and Schiff, E. R. (1986) Ann. Intern. Mad. 105, 968-969. [4] Wittwer, C. T., Smith, A. M., Ash, K. O. and DeWitt, C. W. (1987) Transplantation 44, 843- 844. [5] Morikawa, K., Kuroda, M., Tofuku, Y., Uehara, H. and
39
Koshikawa, S. (1988) Nephron 50, 77-78. [6] NIAID, Staff, Research Resources Branch (1976) in: NIAID Manual of Tissue Typing Techniques (J. G. Ray, Ed.), pp. 22-24. [7] van Rood, J., van Leeuwen, A., Keuning, J. J. and Alblas, A. B. (1975) Tissue Antigens 5, 73-76. [8] Johnson, G. D. and Dorling, J. (1981) in: Techniques in Clinical Immunology (R. A. Thompson, Ed.), Blackwell, Oxford, pp. 122-125. [9] Sayers, M. H., Beatty, P. G. and Hansen, J. A. (1986) Transfusion 26, 113-115.
40
[10] Kuhnl, E, Seidl, S. and Holzberger, G. (1985) Lancet i, 1222. [11] Hunter, J. B. and Menitove, J. E. (1985) Lancet ii, 397. [12] Weiss, S. H., Mann, D. L. and Murray, C. (1985) Lancet ii, 157. [13] Sattentau, Q. J., Dalgleish, A. and Clapham, P. (1986) Science 234, 1120-1123. [14] Adahi, M., Hayami, M., Kashiwagi, N., Mizuta, T., Ohta, Y., Gill, M. J., Matheson, D. S., Tamaoki, T., Shiozawa, C. and Hakomori, S-I. (1988) J. Exp. Med. 167, 323-331. [15] Lederman, M. M., Carey, J. T., Schacter, B., Aucott, J. and Ellner, J. J. (1987) Human Immunol. 20, 279-291.
Elsevier IMLET 01253
Different types of false positive anti-HIV reactions in patients on haemodialysis E s z t e r U j h e l y i 1, G e o r g e Ffist 1, G~ibor Illei 1, l~va G y o d i l, K~troly N a g y 2, Ferenc D. Toth 3, B61a Biiki 1, M a n f r e d P. D i e r i c h 1, Vlastimil M a y e r 5, G y 6 r g y G a l 6, J~inos M a k o 7 a n d S u s a n R. H o l l a n 6 INational Institute of Haematology and Blood Transfusion, Budapest, Hungary; 2Institute of Isotopes, Hungarian Academy of Sciences, Budapest, Hungary; 3Institute for Microbiology, University Medical School, Debrecen, Hungary; 4Institute for Hygiene, University Medical School, Innsbruck, Austria; 5Institute for Virology, Slovak Academy of Sciences, Bratislava, Czechoslovakia; 6Blood Bank, Szeged, Hungary; 7Haemodialysis Center, Semmelweis University Medical School, Budapest, Hungary
(Received28 February 1989; revision received6 April 1989; accepted 11 April 1989)
1. Summary Serum samples o f 589 haemodialysis patients were screened for HIV antibody by ELISA methods. O f these, 36 samples were found to be repeatedly reactive. None of the 36, however, could be confirmed by competitive enzyme immunoassays and Western blot; therefore, they were considered to be false positive. The sera could be divided in two groups. The sera o f Group 1 were designated as the usual type o f false positivity, caused most probably by anti-lymphocyte antibodies. In 19 sera, however, a special type o f false positivity was found. These sera reacted strongly with the plates coated with the supernatants o f HIV-infected cells but not with those of uninfected H9 cells. Three and two sera showed, respectively, positive immunofluorescence reaction with the HIV-infected, but not with the uninfected, H9 and CEM cells. Reactivity to HIVinfected H9 cells could be adsorbed from a part o f these samples with lesser amounts o f HIV-infected than uninfected H9 cells. This special type of false positivity was observed frequently (7/65) in patients who rejected a kidney graft. These findings suggest that this type of anti-HIV false positivity is due to
Key words: AIDS; HIV infection;Haemodialysis;Falsepositivi-
ty; ELISA Correspondence to: Dr. E. Ujhelyi,National Institute of Haema-
tologyand Blood Transfusion, 1113Budapest, P.O. Box 44, Hungary, H-1502.
antibodies reacting with cellular antigens present in HIV-infected but not in uninfected lymphocytes. Their appearance seems to be associated with the immunological activation occurring at graft rejection. 2. Introduction Multiple blood transfusions place patients on haemodialysis at increased risk of HIV infections. That is why many investigations have been performed on the incidence of HIV antibodies [1-3]. The rate of false positive enzyme-immunoassay (EIA) results in patients on haemodialysis was observed to be much higher than in blood donors in most studies [1, 4, 5]. Little is known, however, about the causes o f the development of false positive reactions in patients on haemodialysis. We have been screening patients on haemodialysis for the incidence of HIV antibodies since 1986. These measurements, together with the permanent screening of HIV seronegative haemophiliacs, have been performed as part o f the program to monitor the safety o f the nation's blood supply. Testing is repeated every year. No confirmed positive case was found; false-positivity, however, was observed to be frequent in our studies, as well. The objective of the present work was to analyse the cause of these falsepositive anti-HIV reactions. Besides the usual wellknown false positivity due to the presence o f antilymphocyte antibodies, in some sera a special type o f false positivity was observed: these sera did not react with the normal lymphocytes or with the super-
0165-2478 / 89 / $ 3.50 © 1989 ElsevierSciencePublishers B.V.(Biomedical Division)
35
natants of the cultures of these cells; they reacted strongly, however, with HIV-infected lymphocytes or the supernatants of HIV-infected cells. This special type of false positivity was observed mostly in patients who had rejected a kidney graft previously.
100/xl of adsorbed (and, as control, non-adsorbed) sera were added to 106 HIV-H9 cells, and tested by the indirect immunofluorescence assay. The results were evaluated microscopically, by counting the percentage of the stained cells.
3. Materials and Methods
3.4. Other assays
3.1. Patients tested
Anti-HLA A, B, C antibody assay was performed by the standardized NIH lymphocytotoxicity assay [6]. Adsorption of sera by pooled human platelets: 100 #1 of packed human platelets pooled from 100 donors was added to 100/~1 of serum, mixed and incubated overnight at room temperature. Adsorption was repeated with the same amount of platelets for 2 h at room temperature. Reactivity of the adsorbed and non-adsorbed sera was checked by a lymphocytotoxicity assay using separated T and B lymphocytes [7]. Antinuclear antibodies (ANF) were titrated as outlined in Ref. [8].
Three series of patients were tested: investigations on patients of series 1 (163 serum samples of 93 patients treated in Szeged) and series 2 (81 serum samples of 78 patients treated in Budapest) were performed in 1986. Patients of series 3 (418 serum samples of 418 patients treated in different dialysis centres in Budapest) were studied in 1987-1988. 3.2. H I V serological tests
Two types of enzyme immunoassay (EIA) kits were used: (1)Hg-based kits." Virgo (Electronucleonics), Vironostika (Organon Teknika), Anti-HTLVIII kit of Sorin-Biomedica consisting of HTLVIIIcoated (HTLVIII) plates and H9 supernatantcoated (H9) plates, Welcozyme (Wellcome Diagnostics). (2) CEM-based kits: Genetic (Genetics System), Elavia (Pasteur Diagnostics) containing HTLVIII-coated plates and CEM supernatantcoated (control) plates. Each kit was used strictly according to the: recommendations of the manufacturers. The following assays were used for confirmation of the EIA results: (i) Immunofluorescence test." in the experiments performed in 1986, home-made slides with HTLVIII-infected H9 cells (HTLVIII/H9) and LAV-infected CEM cells (LAV/CEM) were applied. In 1978-88 a commercial IFA kit (Virimmun, FRG) was used. (ii) Western blot assay: commercial kits (Du Pont de Nemours and Bio-Rad) were applied for these studies. 3.3. Adsorption o f sera with HIV-infected and noninfected 1t9 cells
Different amounts (0.5× 106, 1.5 × 106, and 5.0×106) of HIV-infected H9 cells (HIV-H9) or non-infected H9 cells (H9) were added to 100 ttl serum and incubated overnight at room temperature. 36
4. Results
4.1. Results o f H I V serology in series 1 and 2
In these series, anti-HIV screening was performed using the Virgo kit. Fifteen of 171 samples (8.8%) were found to be repeatedly reactive. New serum samples were taken from the patients found to be reactive at the initial screening, 3, 4, and 7 months
TABLE 1 Reactivity of the sera tested repeatedly positive at screening in other anti-HIV assays (Series 1 and 2). Test
No. o f sera tested
Positive
Indeterminate
Negative
Virgo Organon Sorin Elavia Genetic Welcozyme
15 15 15 15 15 15
15 13 10 0 0 0
0 0 0 0 0 0
0 2 5 15 15 15
IFA (CEM) IFA (H9)
9 10
2 3
l 0
6 7
WB
15
0
0
15
later (series 1) and 2 months later (series 2). These samples were also reactive in the Virgo test, although a fluctuation in the extent of reactivity could be frequently seen. The 15 repeatedly Virgo-reactive serum samples were tested by other anti-HIV screening and confirmatory assays (Table 1). The 3 ELISA assays (Virgo, Vironostika, Sorin) correlated strongly with each other. Thirteen and 10 sera were reactive in the Vironostika and Sorin kits, respectively. By contrast, no reactivity was observed in any sample in the competitive Welcozyme EIA and in two CEM-based ELISA assays (Elavia, Genetic). LAV/CEM and HTLVIII/H9 immunofluorescent assays (IFA) were performed in 9 and 10 samples, respectively. In 2 samples, a positive reaction was found in both types of IFA test. A further sample tested positive in the HTLVIII/H9 IFA, while in the LAV/CEM assay an ambiguous result was obtained. When the same 3 sera were tested by IFA with uninfected CEM or H9 cells, no reactivity was found. All the 15 samples tested negative in Western blot. 4.2. Study of the repeatedly ELISA-reactive serum samples in ELISA plates coated with supernatants of infected and non-infected 1t9 cells Eighteen sera from the 15 patients who tested Virgo-positive at the initial screening were studied using the HTLVIII and H9 plates of the Sorin kit. According to the recommendations of the manufacturer, if the HTLVIII/H90.D. ratio exceeds 3.0, the sample should be considered as true positive, while in the case of a ratio of less than 3.0 the HTLVIII reactive sera should be considered false positive. According to these criteria, 12 and 6 sera, respectively, were calculated to be true positive and false positive. On the basis of these findings, two types of false positivity could be distinguished in patients on haemodialysis: sera with a usual type of false positivity (UFP) react with the superantants of both the HIVinfected and non-infected H9 cells, whereas sera with a special type of false-positivity (SFP) react only with the supernatants of the HIV-infected H9 cells.
4.3. Incidence of UFP and SFP in transplanted and non-transplanted patients of series 3 418 patients on haemodialysis were tested for the presence of HIV antibodies using the Sorin kit. Repeated reactivity was observed in 21 samples (5.0°70). Each sample tested negative in the Welcozyme competitive EIA and in the lot of the Vironostika kit used at that time, and none of them could be confirmed by Western blot. On the basis of the results obtained with the HTLVIII and H9 plates, 14 and 7 sera were considered to be UFP and SFP, respectively. Both SFP and UFP results were highly reproducible. For example, in three repeated measurements, HTLVIII/H90.D. ratios in 3 SFP sera were 4.45-3.36-4.22, 3.87-5.17-4.22, and 5.71-4.05-4.87. Out of the 7 SFP sera, 5 were negative in Western blot, whereas in one sample an isolated p55, and in another one an isolated p66 band, was observed. Five samples tested IFA-negative, in 2 samples a weak fluorescence was observed in both the HIV-infected and uninfected H9 cells. One in seven samples showed a positive ANF reaction. Sixty-four of the 418 patients tested in series 3 were redialysed following rejection of a kidney graft. Interestingly enough, 14/21 repeatedly Sorinreactive sera, including all the seven SFP sera, were from this group of patients (Table 2). Fourteen of 64 (22070) of patients who had been grafted showed UFP or SFP, whereas in patients who had not been transplanted HIV false-positivity rarely occurred (7/354 = 2%). There were some differences between the groups of grafted patients with SFP and UFP reactions. SFP patients were on haemodialysis for a longer time (mean 5.6 (3-10) vs. 3.7 (2-6) years) and rejected the kidney graft more quickly (time from transplantation to re-haemodialysis: 7.7 (0-22) months vs. 12 (0.5-34) months) than UFP patients. 4.4. Studies on the mechanism of the special type
of faise-positivity Seven sera from series 1 and 2 found to be SFP, as well as 85 anti-HIV-negative sera, were tested for the presence of anti-HLA antibodies. All the seven SFP sera contained anti-HLA antibodies, mostly in high titres. Fifty-one anti-HIV negative samples were negative for anti-HLA antibodies as well. HLA 37
TABLE 2 Occurrence of the usual (UFP) type and special type (SFP) of false-positivity in transplanted and non-transplanted patients of haemodialysis (Series 3). Results in Sorin kit
No. o f sera tested from patients total
ever transplanted
%
Negative UFP SFP
397 14 7
50 7 7
12 50 100
Total
418
64
15
antibodies could be detected in 34 samples. In 11 samples their reactivity - measured on a lymphocyte panel of 20 donors - exceeded 50°7o. Six SFP sera were adsorbed by human platelets. Adsorption did not decrease the apparent anti-HIV reactivity o f the sera. By contrast, the same treatment fully or almost completely removed cytotoxicity of the sera against isolated T cells of all subjects tested. Reactivity for the isolated B cells decreased, but a marked cytotoxic effect could be observed in most adsorbed serum samples. Eight SFP samples were adsorbed with the same amounts o f HIV-H9 and H9 cells (Table 3). Seven
TABLE 3 Adsorption with infected (H9-HIV) or not infected (H9) H9 cells o f sera o f patients showing a special type of false-positivity. Serum
86696 86702 86705 86694 86679 86810 87123 87643
% of the FITC-labelled cells treated with sera not adsorbed
adsorbed with H9 cells
adsorbed with H9-HIV cells
90 30 100 50 20 80 90 50
60 < 10 70 30 0 20 60 30
20 < 10 70 10 0 20 15 0
Only results o f adsorption with 1.5 x 106 cells are shown. Similar results were obained using lower or higher a m o u n t s o f cells for adsorption.
38
samples reacted in the IFA assay with _ 50°/o of the H I V - H 9 cells. Adsorption with both the HIV-H9 and the H9 cells decreased the percentage of the FITC-labeled cells in each sample. In the case of four sera, the same amount of HIV-H9 cells adsorbed considerably more antibodies than the uninfected H9 cells. 5. Discussion
Our present findings, in agreement with many previous observations [1, 4, 5], indicate that repeatedly positive reactions in anti-HIV ELISA assays are more common in chronic haemodialysis patients than in blood donors. Altogether, such reactions were observed in 36 cases out of 589 patients tested (6.1°70). Neither Western blot nor competitive EIA confirmed the ELISA reactivity of the sera. Therefore the sera were considered to be false-positive. The initial screening of the patients was performed in two different H9-based kits. In the first two series o f investigations, other H9-based kits like the Vironostika kit o f Organon Teknika also showed repeated reactivity with these samples. Some repeatedly ELISA-reactive sera reacted in the indirect immunofluorescence assay as well. False-positive anti-HIV reactions can develop due to several reasons. It seems that the most common cause o f false-positivity is the presence of antibodies against antigens of the lymphocyte cell line, H9, used to produce the AIDS virus. Since in the competitive EIAs, not anti-IgG but HIV antibody conjugates are used, the anti-lymphocyte antibodies cannot interfere with the results of these assays. Lymphocyte-derived antigens can be present in the viral lysates applied for coating the ELISA plates [9-11]. In our patients, too, anti-lymphocyte antibodies and primarily those in high titres occurred more frequently in false-positive than in other sampies. However, such antibodies could also be detected in high titres in several samples which tested negative in HIV ELISA assays. Moreover, antibodies responsible for the development o f the false-positive reactions could not be removed from the sera by adsorption with platelets. This observation excludes the possible role o f the anti-HLA A, B, or C antibodies in the development of the observed falsepositivity. H9 ceils were HLA-typed by Weiss et al. [12] and
CD4+ Iwmphocwte activated bw an Immune stimulus
a p p e a r a n c e /!'f an antigen on t h e s u r f a c e of cell or
Increase In the density of a pre-existing
antigen
I transfusion of activated lymphocytes to patients on haemodlalysi5 and.,'or activation
at g r a f t
of %he I m m u n e
/
system
rejection
development of au~oantibodles
in t h e
petlents reacting with the aotlvated T ceils
and giving a special t y p e of f a l s e p o s l t l v l t y In HIV ELISR assa61s Fig. 1. Hypothetical mechanism of development of the special type of anti-HIV false-positivity in patients on haemodialysis.
found to be A1, Bw62, Bw6, Cw3, DR4, DQw3. Since anti-DR4 antibodies were found to be able to induce false-positive anti-HIV ELISA results [10], the best candidate for causing false positive results in haemodialysis patients would be anti-DR4 antibodies. In the case of samples designated in the present study as usual type of false positives (UFP) this explanation seems to be acceptable, too: these sera reacted approximately to the same extent in plates coated with the supernatant of the HIV-infected and non-infected H9 cells. The other group of sara of haemodialysis patients, designated as SFP (special type of false-positivity), however, reacted in a very different way: (1) they showed a high reactivity (O.D. value) on the plates of the Sorin kit coated with the supernatants of the HIV-infected H9 cells but did not react or showed a weak reaction in the control plates of the same kit coated by the concentrate of cellular material shed by uninjected H9 cells. (2) Only a part of the sera in which false-positive immunofluorescence reaction was observed reacted with the uninfected H9 cells in the IFA test. Interestingly enough, the same phenomenon was observed in the case of HIVinfected and uninfected CEM cells. (3) In 4/7 SFP sera, HIV-infected H9 cells adsorbed much more
anti-HIV reactivity than the uninfected H9 cells. It is remarkable that SFP occurred mainly among patients who had been transplanted and had rejected one or two kidney grafts. In series 3, all the seven sera with SFP were from such patients, while 50°70 of patients with UFP were never transplanted. Besides patients on haemodialysis, SFP was rarely (in 9 cases out of the first 800000 blood donations tested) observed among blood donors as well. Interestingly enough, all the 9 donors were females, who developed antibodies possibly during pregnancies (unpublished observations). On the basis of these findings a hypothesis could be raised (Fig. 1). According to this assumption, infection with HIV results in the appearance of an antigen on the surface of the H9 cells and other lymphocytes, which is either not present on the uninfected cells, or increases the density of a preexisting antigen. Immunological activation can induce the expression of the same antigen in lymphocytes and (considering the great number of transfusions administered to patients on haemodialysis) by chance such activated lymphocytes could have been transfused. This transfusion can lead to antibody formation against this antigen in the patients. Alternatively, immune activation which occurs at graft rejection or other type of activation can also induce the formation of autoantibodies against this antigen, which may be responsible for the special type of anti-HIV false positivity. Several recent findings indicate that HIV infection induces or increases the expression of some cellular antigens in lymphocytes [13-15]. Any of these antigens, or other unrecognized ones, could be the target of the antibodies inducing anti-HIV false positivity. Experiments aiming to characterize this antigen are underway in our laboratories. References [1] Peterman, T. A., Lang, G. R., Mikes, N. J., Solomon, S. R., Schabl¢, C. A., Feorino, P. M., Britz, J. A. and Allen, J. R. (1986) J. Am. Med. Assoc. 255, 2324-2326. [2] Goldman, M., Liesnard, C., Vanderwegheim, J.-L., Della, N., Touissaint, C., Sprecher, J., Cogniaux, J. and Thyry, L. (1986) Br. Med. J. 293, 161-162. [3] Villaneuva, S. Y., de Medina, M., Perez, G., Reddy, K. R., Jeffers, L., Watkins, E, Goldsmith, C. and Schiff, E. R. (1986) Ann. Intern. Mad. 105, 968-969. [4] Wittwer, C. T., Smith, A. M., Ash, K. O. and DeWitt, C. W. (1987) Transplantation 44, 843- 844. [5] Morikawa, K., Kuroda, M., Tofuku, Y., Uehara, H. and
39
Koshikawa, S. (1988) Nephron 50, 77-78. [6] NIAID, Staff, Research Resources Branch (1976) in: NIAID Manual of Tissue Typing Techniques (J. G. Ray, Ed.), pp. 22-24. [7] van Rood, J., van Leeuwen, A., Keuning, J. J. and Alblas, A. B. (1975) Tissue Antigens 5, 73-76. [8] Johnson, G. D. and Dorling, J. (1981) in: Techniques in Clinical Immunology (R. A. Thompson, Ed.), Blackwell, Oxford, pp. 122-125. [9] Sayers, M. H., Beatty, P. G. and Hansen, J. A. (1986) Transfusion 26, 113-115.
40
[10] Kuhnl, E, Seidl, S. and Holzberger, G. (1985) Lancet i, 1222. [11] Hunter, J. B. and Menitove, J. E. (1985) Lancet ii, 397. [12] Weiss, S. H., Mann, D. L. and Murray, C. (1985) Lancet ii, 157. [13] Sattentau, Q. J., Dalgleish, A. and Clapham, P. (1986) Science 234, 1120-1123. [14] Adahi, M., Hayami, M., Kashiwagi, N., Mizuta, T., Ohta, Y., Gill, M. J., Matheson, D. S., Tamaoki, T., Shiozawa, C. and Hakomori, S-I. (1988) J. Exp. Med. 167, 323-331. [15] Lederman, M. M., Carey, J. T., Schacter, B., Aucott, J. and Ellner, J. J. (1987) Human Immunol. 20, 279-291.