- Email: [email protected]
neutralization index in HIV patients during highly active antiretroviral therapy (HAART)
Immunology Letters 89 (2003) 25 /30 www.elsevier.com/locate/
Significant decrease of the enhancement/neutralization index in HIV patients during highly active antiretroviral therapy (HAART) De´nes Ba´nhegyi a, Attila Ba´csi b, Ferenc D. To´th b, Zolta´n Proha´szka c,d, Anna Horva´th c, Zolta´n Beck b, Jo´zsef Ko´nya b, George Fu¨st c,d,* a Department of Immunology, St. La´szlo´ Hospital, Budapest, Hungary Institute of Medical Microbiology, Medical and Health Science Center, University of Debrecen, Debrecen, Hungary c Third Department of Internal Medicine, Faculty of Medicine, Semmelweis University, Ku´tvo ¨ lgyi u´t 4, H-1125 Budapest, Hungary d Research Group of Metabolism, Genetics and Immunology, Hungarian Academy of Sciences Budapest, Budapest, Hungary b
Received 22 November 2002; received in revised form 25 April 2003; accepted 28 April 2003
Abstract Authors studied the effect of highly active antiretroviral therapy (HAART) on balance of the antibodies that enhance or neutralize growth of HIV-1IIIB strain in MT-4 cells in the presence or absence of human complement. Sequential serum samples were collected from 28 patients in advanced stage of HIV disease before and during HAART. The balance of the enhancing and neutralizing antibodies was expressed by an index value (E/N I). Samples with an E/N I of B/0.5 (twofold decrease in virus production) were considered as neutralizing, whereas samples with an E/N I /2.0 (twofold increase in virus production) were considered as enhancing. At the beginning of HAART serum samples from eight patients enhanced, and samples from only two patients neutralized the virus in the presence of complement, median (25th /75th percentile) value of E/N I was 1.32 (0.79 /2.29). E/N I significantly (P B/0.0001) dropped to 0.37 (0.19 /0.57) during the follow-up period of 18.5 (10.5 /23.5) months under HAART. Similar changes were detected when serum samples were tested with no complement added. The E/N I values were also markedly decreased when cultures inoculated with mixtures of HIV and purified IgG prepared from serum pools taken before and during HAART, respectively, were compared. In the last samples of 20/28 patients, neutralization was measured even in the presence of complement while enhancement was found with none of these samples. These findings suggest that HAART results in disappearance of enhancing antibodies and switches the E/N I toward neutralization. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Enhancing antibodies; Neutralizing antibodies; Complement; HIV; HAART
1. Introduction According to previous studies special types of HIVspecific antibodies, the so-called enhancing antibodies that may facilitate progression of HIV infection can be detected in a part of HIV-infected patients. Two types of enhancing antibodies were described: (a) complementindependent, Fc-receptor dependent antibodies (ADE) [1], and (b) complement-dependent antibodies (C-ADE) [2]. According to the studies of Mitchell et al. [3] in accordance with the earlier observations of Robinson et al. [4,5], the primary antigenic domains responsible for
* Corresponding author. Tel./fax: /36-1-212-9351. E-mail address: [email protected] (G. Fu¨st).
the C-ADE of HIV and SIV resides in the principal immunodominant sequence of gp41. This is a conserved sequence, which*/in agreement with the site-directed mutagenesis experiments of Mitchell et al. [3] */is critical for the gp120 /gp41 interaction. Binding site for enhancing antibodies on gp41 is easily accessible after release of gp120 from gp41 or after a conformational change in gp120 following association with CD4 and co-receptors. Some studies indicate that antibody-mediated enhancement measured in HIV-1 infected individuals correlate with immunosuppression and disease [6 /9]. Recently we have reported on observations, which supported the role of C-ADE in the progression of HIV disease [10]. In this cross-sectional study we have found a highly significant positive correlation between the extent of so-called E/N index (E/N I) values
0165-2478/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0165-2478(03)00106-8
26
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
reflecting the ratio of the HIV-1 enhancement and neutralization capacity of the test sera and plasma HIV-RNA levels. C-ADE was first detected concomitantly with seroconversion. Both E/N I values and HIV1 RNA levels significantly increased during a 17-month observation period in 18 HIV-infected patients who remained untreated or were on AZT-monotherapy [10]. Recently these findings were corroborated with Subbramanian et al. [11] who detected enhancement of in vitro HIV infection in the presence of complement in 72% of serum samples of HIV patients. They also concluded that the anti-HIV-1 humoral immune response consist of a mixture of antibodies that may inhibit or enhance HIV infection and whose ratios may vary in different stages of the infection. The goal of our present work was to study the effect of the HAART on the ratio of enhancing and neutralizing antibodies. To our best knowledge, no such studies have been performed so far.
2. Materials and methods 2.1. Patients and controls Serum samples were obtained from 28 patients (22 males, six females, median age: 45.2 years, CD4/ T cell count of 3729/38 cells/ml). Each patient was treated with HAART, 23 and five patients got triple and quadruple drug regimen, respectively. Protease inhibitors were administrated to all but four patients. The patients were followed up for 18.5 (10.5 /23.5) months (median (interquartile range)), blood samples were taken at the beginning of HAART, 8.0 (5.5 /15.5) months thereafter, and at the end of the follow-up period. Serum samples were stored at /30 8C. 2.2. Preparation of purified IgG from serum pools of the HIV patients Two pools were prepared from the sera of patients 50, 153, 274 and 312, and patients 322, 387, 400 and 451 (see Table 1). Next pooled sera were divided into two aliquots, one aliquot was used for the preparation of purified IgG. IgG purification was performed by ion exchange chromatography on DEAE cellulose DE52 as described in Ref. [12]. Eluted IgG was concentrated to the original volume of serum pool by using dialysis tubing and Aquacide (Calbiochem, La Jolla, CA).
culture medium in triplicates. One hundred microliter of serum or IgG dilutions were mixed with 100 ml of fresh pooled sera from HIV-seronegative healthy persons (normal human serum, NHS). HIV-1IIIB (100 TCID50) in 100 ml culture medium was added, and incubated at 37 8C for 1 h. Thus, final concentration of human complement in the samples added to target cells was 25%. In the case of each patient’s serum sample mixtures without complement in which NHS was replaced by culture medium was also tested. In each series of experiments, control cultures infected with the virus alone were established, too. Then, 5/105 MT-4 cells in growth medium were added and incubated in this medium without change for 5 days at 37 8C. Growth of HIV in the cultures was monitored on each day by the reverse transcriptase (RT) assay as outlined by Hoffman et al. [13]. Peak RT values were observed almost exclusively at day 5. The results were measured as c.p.m., and expressed finally by an index (E/N I) value which indicated the relationship between virus growth in serum or IgG containing samples compared to that measured in cultures infected with virus alone. Index values were calculated as follows: means of triplicate peak RT (c.p.m.) values measured in cultures which infected with the mixtures of virus, patient’s serum (and NHS) at day 5 were divided by the means of triplicate peak c.p.m. values which measured in control cultures which infected with virus alone at day 5. Samples with an E/N I less than 0.5 (twofold decrease in virus production) were considered as neutralizing whereas samples with an E/N I exceeding 2.0 (twofold increase in virus production) were considered as enhancing. Variation coefficient of the method was 7.23%. Pooled heattreated (56 8C, 30 min) HIV-seronegative serum at the same dilution as the test samples (1:64) was used as a negative control at each measurement; it never influenced the growth of HIV-1 in the system. 2.4. Determination of CD4/ cell counts T cell subset determinations were performed by flow cytometry (FACScan) by using the IMK Lymphocyte kit (Becton Dickinson, Mountain View, CA). 2.5. Measurement of HIV-1 RNA level in the plasma samples of the patients HIV-1 RNA levels in the plasma samples of the patients were determined with the NASBA (Organon Teknika, Turnhout, Belgium) assay.
2.3. Measurement of in vitro HIV enhancement and neutralization
2.6. Statistical analysis
The assay was carried out as described previously [10]. Briefly, serum or IgG samples from the patients were heat-treated (56 8C, 30 min) and diluted at 1:64 in
Since E/N I values measured in the different groups did not match Gaussian distribution, non-parametric tests were used. Groups were compared using the Mann-
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
27
Table 1 Changes in the E/N I values measured in the absence and presence of complement with serum samples of 28 HAART-treated HIV patients Designation of patient
Follow-up time, months
In the absence of complement
In the presence of complement
Before HAART
After HAART
Before HAART
After HAART
50 108 125 143 153 203 274 309 312 322 337 368 373 387 389 393 400 408 411 413 415 422 446 451 463 480 484 503
14.0 10.0 6.0 46.0 19.0 21.0 19.0 18.0 34.0 33.0 20.0 28.0 21.0 24.0 23.0 31.0 24.0 20.0 16.0 11.0 10.0 7.0 10.0 10.0 18.0 10.0 11.0 11.0
0.40 0.61 0.81 0.54 0.72 0.72 0.95 0.95 1.09 1.07 1.05 0.59 0.79 0.49 0.38 1.05 1.02 0.25 0.32 0.23 0.32 0.27 0.65 0.81 0.73 0.60 0.26 0.46
0.41 0.35 0.19 0.05 0.09 0.46 0.29 0.10 0.41 0.46 0.43 0.22 0.58 0.45 0.05 0.15 0.19 0.43 0.36 0.20 0.36 0.77 0.33 0.12 0.11 0.10 0.13 0.08
2.45 0.54 0.81 1.03 3.22 0.80 3.20 1.42 3.88 3.46 1.49 1.34 0.77 2.12 0.62 1.67 3.32 0.39 1.29 0.67 1.36 0.29 0.65 2.51 1.50 1.08 0.82 1.05
0.86 0.34 0.19 0.09 0.17 0.56 0.34 0.40 0.49 0.42 0.82 0.20 0.60 0.40 0.07 0.57 0.46 0.78 0.33 0.18 0.39 0.74 0.33 0.13 0.22 0.13 0.71 0.17
Median (interquartile range)
18.5 (10.5 /23.5)
0.63 (0.39 /0.88)
0.26* (0.12 /0.42)
1.32 (0.79 /2.29)
0.37** (0.19 /0.57)
Differences between the medians of paired groups (Wilcoxon signed rank test) *P/0.0001, **P B/0.0001.
Whitney U or Wilcoxon signed rank tests. Categorical data were compared by the Fisher’s exact test. P B/0.05 was considered to be statistically significant. Statistical analysis was performed using GRAPHPAD Prism version 3.00 for Windows, (GraphPad Software, San Diego, CA, http://www.graphpad.com).
3. Results 3.1. Changes in the E/N I values in HIV patients under HAART In the samples taken at the beginning of follow-up, median (25th /75th percentile) values of E/N indices measured in the presence of complement were 1.32 (0.79/2.29) (Table 1). Out of serum samples from the 28 patients, 8 sera (28.6%) had an E/N I exceeding 2.0 that is significantly enhanced HIV-infection in vitro. By contrast, only 2/28 (7.1%) samples had a E/N I B/0.5 that is neutralized virus in the presence of complement (Table 2). Enhancement occurred in none of the samples
with no complement added, while 10/28 samples (35.7%) neutralized the virus (Table 2). E/N I measured both in the absence and presence of complement significantly (P /0.0001 and P B/0.0001, respectively) decreased during the follow-up period of 18.5 (10.5 /23.5) months under HAART. There was a significant positive correlation (Spearman correlation coefficient R /0.40, P /0.036) between the length of follow-up period and the extent of E/N indices measured in the presence of complement, whereas no such correlation (R /0.29, P/0.128) was found with samples tested with no complement added. In the last samples, neutralization (E/N I B/0.5) was measured in the absence and presence of complement in 26/28 (92.9%) and 20/28 (71.4%) patients, respectively. By contrast enhancement was found with none of the samples (Table 2). 3.2. Comparison of the enhancement/neutralization of pooled serum and purified IgG preparations In order to exclude the possibility that antiviral drugs present in the test sera affect the assay used for the
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
28
Table 2 Occurrence of serum samples that enhanced, neutralized or did not neutralize or enhance the growth of the HIV-1IIIB strain in MT-4 cells in the absence or presence of human complement in 28 HIV patients before and after HAART
No complement added Complement added a
Neutralization (E/N Ia B/0.5)
No neutralization or enhancement (E/N Ia: 0.5 /2.0)
Enhancement (E/N Ia /2.0)
Before HAART
After HAART for 19.0 (10.5 /24.0) months
Before HAART
After HAART for 18.5 (10.5 /23.5) months
Before HAART
After HAART for 18.5 (10.5 /23.5) months
10
26
18
2
0
0
2
20
18
8
8
0
Enhancement/neutralization index.
measurement of HIV enhancement/neutralization, we prepared serum pools from eight patients each taken before and during HAART. IgG were purified from these pools, diluted to the IgG concentration in the 1:64 serum dilution and used in the absence (Fig. 2A) and presence (Fig. 2B) of complement in the enhancement/ neutralization assay as described above. Only weak differences were found between the results that obtained with sera and IgG preparations, HAART resulted in a marked drop of the E/N I values in the cultures inoculated with virus /IgG complexes as well.
3.3. Dynamics of the decrease of E/N I values during HAART In both types of assays E/N I values significantly decreased in 8 months after beginning of HAART. A further significant drop in the index values occurred between 8 months and the end of follow-up (Fig. 1).
3.4. Changes in the viral load and T cells subsets Viral load significantly decreased, absolute number and percentage of CD4/ T cells significantly increased, while percentage of the CD8/ cells significantly decreased during this period (Table 3).
4. Discussion Our present findings indicate that HAART switches balance of enhancement and neutralization toward neutralization. Data obtained in the present study support our previous findings on the frequent occurrence of complement-mediated antibody-dependent HIV enhancement in the sera of untreated patients in advanced stage of HIV disease. The median (25th/75th percentile) of the E/N indices (1.32 (0.79 /2.29)) measured in the present study in samples obtained before HAART was approximately the same as in our previous study (1.35 (1.00/1.84)) performed in the sera of HIV patients mostly never treated or treated only with AZT before blood sampling [10]. Serum samples 28.6 and 19.4% of the patients in the present and previous [10] study, respectively had an E/N I exceeding 2.0 that is significantly enhanced HIV-infection in vitro. In principle our present observations can be biased by two ways. First, antiretroviral drugs present in the test samples can influence the assay system we used. We tested serum samples at a dilution of 1:64, and at this dilution residual antiretrovirals can still be present in a high enough concentration to inhibit in vitro virus infection. According to the data presented in Fig. 1, however, this assumption seems to be highly improbable. E/N I values were found to be markedly and significantly lower in the samples taken 19 months than in those taken 8 months after the beginning of HAART,
Fig. 1. Dynamics of the changes in the E/N I values measured in the absence (A) and presence (B) of complement with serum samples of 28 HAARTtreated HIV patients. P values for the comparison with Wilcoxon signed rank test are indicated in the figures.
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
29
Table 3 Changes in viral load and T cell subsets in 28 HAART-treated HIV patients* Variables
Before HAART
After HAART for 18.5 (10.5 /23.5) months
Median (interquartile range) HIV-1 RNA, copies/ml
5000 (0 /53 000)
0 (0 /358)
Mean9/S.E.M. CD4/T CD4/T CD8/T CD8/T
cells/ml cells/% cells/ml cells/%
3729/42 19.19/1.8 11419/107 55.19/2.3
P value Wilcoxon signed rank test 0.0090 Paired t test
5349/42 25.69/2.3 11719/121 48.99/2.4
B/0.0001 0.0003 0.7375 0.0007
* During a follow up of 18.5 (10.5 /23.5) months.
although the patients got the same type of HAART protocols, therefore the concentration of the antiretroviral drugs in their serum samples were approximately the same at both times. Moreover, as it is shown in Fig. 2, a marked drop of the E/N I values occurred when cultures inoculated with mixtures of HIV and purified IgG prepared from serum pools taken before and during HAART, respectively, were compared. Second, apparently it is surprising that measurements performed in our system with one single T cell adapted strain, HIV-1IIB can provide clinically meaningful results, whereas in most other systems use of primary HIV-1 isolates and assays with primary cells are essential for in vivo relevance. Recent elegant studies of Mitchell et al. [3] that are in accordance with the earlier observations of Robinson et al. [4,5] help to resolve this apparent paradoxical situation. According to these investigations the primary antigenic domains responsible for the C-ADE of HIV and SIV resides in the principal immunodominant sequence of gp41. This is a conservative sequence since according to the sitedirected mutagenesis experiments of Mitchell et al. [3] it is critical for the gp120/gp41 interaction. According to their model, binding site for enhancing antibodies on the
primary association site with gp41 is surrounded by amino-linked oligosaccharides which are likely sites for the covalent attachment of the C3dg peptide of C3, the physiological ligand of the CR2 receptor. Binding site for enhancing antibodies on gp41 may be accessible after release of gp120 from gp41 or after a conformational change in gp120 upon association with CD4. Interestingly enough HAART decreased E/N I values in the cultures infected with the serum /virus mixtures without complement, too although the extent of E/N I drop correlated to the length of HAART only in the complement-containing cultures. Therefore it can be assumed that the enhancing antibodies may exert their effect on a complement-independent way as well, the addition of complement in vitro or the presence of complement in vivo, however, markedly amplifies this enhancement. Therefore it seems probable that the apparent decrease in the E/N I values in due to the changes in the generation of enhancing and/or neutralizing antibodies that is to the drop of the production of enhancing antibodies, or an increased production of neutralizing antibodies. Changes in the titers of neutralizing antibodies in chronically infected HIV patients during
Fig. 2. Changes in the E/N I values in the absence (A) and presence (B) of complement with serum pools (SP) taken from the same 4-4 patients before (SP1A, SP2A) and after HAART (SP1B, SP2B) for 18.5 (10.5 /23.5) months as well as with IgG preparations (IgG1A, IgG1B, and IgG2A, IgG2B) purified from these pools. The IgG concentration of the sera and purified preparations was the same. No significant differences were found between serum pools and IgG preparations in the E/N I values.
30
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
HAART were studied by several groups [14 /16]. These studies were performed in limited numbers of patients and gave conflicting results. Kim et al. [15] serially tested the neutralizing antibody titres against a panel of HIV-1 primary isolates in 11 patients under HAART with sustained B/400 copies/ml viral load. Development of increased neutralizing antibodies against some isolates occurred only in roughly half of the patients. Kimura et al. [16] investigated 19 chronically infected patients who were undergoing HAART. Reconstitution of or increase in neutralization activity was observed in four of 19 patients during HAART but neutralization activity was more or less unchanged in most patients. It seems therefore improbable that the marked decrease observed in the E/N I values in our present study was the consequence of an increased neutralization activity */ at least neutralization measured by conventional methods */during HAART. By contrast, several observations, described in our previous papers [9,10] and by an other group [11] indicate that C-ADE is associated with and may potentially facilitate progression of HIV disease. Considering these observations it seems probable that drop of antigen stimulus in patients under HAART results in a markedly decreased production of the antibodies mediating C-ADE and consequently to a gradual and dramatic (more than threefold) decrease in the E/N I values. Moreover, recently it has been reported that enhancing antibodies may also contribute to the selection pressure operating on the circulating population of polymorphic HIV-1 variants [17]. Most importantly, Verrier et al. [18] demonstrated that whole virus-particle based HIV-1 vaccines in hyperimmunized rhesus macaques can play a facilitating role in the transmission of the virus and/or evolution of the disease. Therefore our present findings indicate that besides its main effect */ decrease of viral load */HAART results in an additional beneficial effect: disappearance of C-ADE and shift of the balance toward dominance of the antibodies that neutralize HIV-1 in vitro even in the presence of complement. It was described by several groups [19 / 22] that there is a drop of B cell activation and IgG levels during HAART. Our present findings indicate that the amounts of C-ADE mediating antibodies changes in parallel with the amounts of IgG-specific antibodies in the patients on HAART, similarly to the very beginning of HIV infection when the appearance of C-ADE coincides in time with the appearance of HIVspecific antibodies [10].
Acknowledgements This work was supported by the following grant: OTKA (National Research Fund) T032661 (GF).
References [1] A. Takeda, C.H. Tuazon, F.A. Ennis, Science 242 (1988) 580 / 583. [2] W.E. Robinson, Jr., D.C. Montefiori, W.M. Mitchell, Biochem. Biophys. Res. Commun. 149 (1987) 693 /699. [3] W.M. Mitchell, L. Ding, J. Gabriel, AIDS 12 (1998) 147 /156. [4] W.E. Robinson, T. Kawamura, D. Lake, Y. Kasuho, W.M. Mitchell, E.M. Hersch, J. Virol. 64 (1990) 5301 /5305. [5] W.E. Robinson, Jr., M.K. Gorny, J.Y. Xu, W.M. Mitchell, S. Zolla-Pazner, J. Virol. 65 (1991) 4169 /4177. [6] J. Homsy, M. Tateno, J.A. Levy, Lancet I (1988) 1285 /1286. [7] F.D. To´th, B. Szabo´, E. Ujhelyi, K. Pa´lo´czi, A. Horva´th, G. Fu¨st, J. Kiss, D. Ba´nhegyi, S.R. Holla´n, AIDS 5 (1991) 263 /268. [8] B. McDougall, M.H. Nymark, G. Landucci, D. Forthal, W.E. Robinson, Jr., Scand. J. Immunol. 45 (1997) 103 /111. [9] G. Fu¨st, F.D. To´th, J. Kiss, E. Ujhelyi, I. Nagy, D. Ba´nhegyi, AIDS 8 (1994) 603 /609. [10] J. Szabo´, Z. Proha´szka, F.D. To´th, A. Gyuris, J. Segesdi, D. Banhegyi, E. Ujhelyi, J. Minarovits, G. Fu¨st, AIDS 13 (1999) 1841 /1849. [11] R.A. Subbramanian, J.M. Xu, E. Toma, R. Morriset, E.A. Cohen, J. Menezes, A. Ahmad, J. Clin. Microbiol. 6 (2002) 2141 /2146. [12] L. Hudson, F.C. Hay, Practical Immunology, Blackwell Scientific Publications, Oxford, 1989, pp. 299 /300. [13] A.D. Hoffman, B. Banapour, J.A. Levy, Virology 147 (1985) 326 /335. [14] J.M. Binley, A. Trkola, T. Ketas, D. Schiller, B. Clas, S. Little, D. Richman, A. Hurley, M. Markowitz, J.P. Moore, J. Infect. Dis. 182 (2000) 945 /949. [15] J.H. Kim, J.R. Mascola, S. Ratto-Kim, T.C. VanCott, L. LoomisPrice, J.H. Cox, N.L. Michael, L. Jagodzinski, C. Hawkes, D. Mayers, B.L. Gilliam, D.C. Birx, M.L. Robb, AIDS Res. Human Retroviruses 11 (2001) 2021 /2034. [16] T. Kimura, K. Yoshimura, K. Nishihara, Y. Maeda, S. Matsumi, A. Koito, S. Matsushita, J. Infect. Dis. 185 (2002) 53 /60. [17] D. Davis, H. Trischmann, D.M. Stephens, P.J. Lachmann, J. Med. Virol. 64 (2001) 207 /216. [18] F. Verrier, C. Moog, F. Barre-Sinoussi, E. Van der Ryst, C. Spenlehauer, M. Girard, Bull. Acad. Natl. Med. 184 (2000) 67 / 84. [19] S. Yokozaki, J. Takamatsu, I. Nakano, Y. Katano, H. Toyoda, K. Hayashi, T. Hayakawa, Y. Fukuda, Blood 96 (2000) 4293 / 4299. [20] Y. Chong, H. Ikematsu, M. Yamamoto, K. Chijiwa, W. Ariyam, J. Hayashi, T. Shirai, S. Kashiwagi, Kansenshogaku Zasshi 75 (2001) 535 /540. [21] C.E. O’Sullivan, R. Peng, K.S. Cole, R.C. Montelaro, T. Sturgeon, H.B. Jenson, P.D. Ling, J. Med. Virol. 67 (2002) 320 /326. [22] M.A. Jacobson, H. Khayam-Bashi, J.N. Martin, D. Black, V. Ng, J. Acquir. Immune Def. Syndr. 14 (2002) 472 /477.
Significant decrease of the enhancement/neutralization index in HIV patients during highly active antiretroviral therapy (HAART) De´nes Ba´nhegyi a, Attila Ba´csi b, Ferenc D. To´th b, Zolta´n Proha´szka c,d, Anna Horva´th c, Zolta´n Beck b, Jo´zsef Ko´nya b, George Fu¨st c,d,* a Department of Immunology, St. La´szlo´ Hospital, Budapest, Hungary Institute of Medical Microbiology, Medical and Health Science Center, University of Debrecen, Debrecen, Hungary c Third Department of Internal Medicine, Faculty of Medicine, Semmelweis University, Ku´tvo ¨ lgyi u´t 4, H-1125 Budapest, Hungary d Research Group of Metabolism, Genetics and Immunology, Hungarian Academy of Sciences Budapest, Budapest, Hungary b
Received 22 November 2002; received in revised form 25 April 2003; accepted 28 April 2003
Abstract Authors studied the effect of highly active antiretroviral therapy (HAART) on balance of the antibodies that enhance or neutralize growth of HIV-1IIIB strain in MT-4 cells in the presence or absence of human complement. Sequential serum samples were collected from 28 patients in advanced stage of HIV disease before and during HAART. The balance of the enhancing and neutralizing antibodies was expressed by an index value (E/N I). Samples with an E/N I of B/0.5 (twofold decrease in virus production) were considered as neutralizing, whereas samples with an E/N I /2.0 (twofold increase in virus production) were considered as enhancing. At the beginning of HAART serum samples from eight patients enhanced, and samples from only two patients neutralized the virus in the presence of complement, median (25th /75th percentile) value of E/N I was 1.32 (0.79 /2.29). E/N I significantly (P B/0.0001) dropped to 0.37 (0.19 /0.57) during the follow-up period of 18.5 (10.5 /23.5) months under HAART. Similar changes were detected when serum samples were tested with no complement added. The E/N I values were also markedly decreased when cultures inoculated with mixtures of HIV and purified IgG prepared from serum pools taken before and during HAART, respectively, were compared. In the last samples of 20/28 patients, neutralization was measured even in the presence of complement while enhancement was found with none of these samples. These findings suggest that HAART results in disappearance of enhancing antibodies and switches the E/N I toward neutralization. # 2003 Elsevier Science B.V. All rights reserved. Keywords: Enhancing antibodies; Neutralizing antibodies; Complement; HIV; HAART
1. Introduction According to previous studies special types of HIVspecific antibodies, the so-called enhancing antibodies that may facilitate progression of HIV infection can be detected in a part of HIV-infected patients. Two types of enhancing antibodies were described: (a) complementindependent, Fc-receptor dependent antibodies (ADE) [1], and (b) complement-dependent antibodies (C-ADE) [2]. According to the studies of Mitchell et al. [3] in accordance with the earlier observations of Robinson et al. [4,5], the primary antigenic domains responsible for
* Corresponding author. Tel./fax: /36-1-212-9351. E-mail address: [email protected] (G. Fu¨st).
the C-ADE of HIV and SIV resides in the principal immunodominant sequence of gp41. This is a conserved sequence, which*/in agreement with the site-directed mutagenesis experiments of Mitchell et al. [3] */is critical for the gp120 /gp41 interaction. Binding site for enhancing antibodies on gp41 is easily accessible after release of gp120 from gp41 or after a conformational change in gp120 following association with CD4 and co-receptors. Some studies indicate that antibody-mediated enhancement measured in HIV-1 infected individuals correlate with immunosuppression and disease [6 /9]. Recently we have reported on observations, which supported the role of C-ADE in the progression of HIV disease [10]. In this cross-sectional study we have found a highly significant positive correlation between the extent of so-called E/N index (E/N I) values
0165-2478/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0165-2478(03)00106-8
26
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
reflecting the ratio of the HIV-1 enhancement and neutralization capacity of the test sera and plasma HIV-RNA levels. C-ADE was first detected concomitantly with seroconversion. Both E/N I values and HIV1 RNA levels significantly increased during a 17-month observation period in 18 HIV-infected patients who remained untreated or were on AZT-monotherapy [10]. Recently these findings were corroborated with Subbramanian et al. [11] who detected enhancement of in vitro HIV infection in the presence of complement in 72% of serum samples of HIV patients. They also concluded that the anti-HIV-1 humoral immune response consist of a mixture of antibodies that may inhibit or enhance HIV infection and whose ratios may vary in different stages of the infection. The goal of our present work was to study the effect of the HAART on the ratio of enhancing and neutralizing antibodies. To our best knowledge, no such studies have been performed so far.
2. Materials and methods 2.1. Patients and controls Serum samples were obtained from 28 patients (22 males, six females, median age: 45.2 years, CD4/ T cell count of 3729/38 cells/ml). Each patient was treated with HAART, 23 and five patients got triple and quadruple drug regimen, respectively. Protease inhibitors were administrated to all but four patients. The patients were followed up for 18.5 (10.5 /23.5) months (median (interquartile range)), blood samples were taken at the beginning of HAART, 8.0 (5.5 /15.5) months thereafter, and at the end of the follow-up period. Serum samples were stored at /30 8C. 2.2. Preparation of purified IgG from serum pools of the HIV patients Two pools were prepared from the sera of patients 50, 153, 274 and 312, and patients 322, 387, 400 and 451 (see Table 1). Next pooled sera were divided into two aliquots, one aliquot was used for the preparation of purified IgG. IgG purification was performed by ion exchange chromatography on DEAE cellulose DE52 as described in Ref. [12]. Eluted IgG was concentrated to the original volume of serum pool by using dialysis tubing and Aquacide (Calbiochem, La Jolla, CA).
culture medium in triplicates. One hundred microliter of serum or IgG dilutions were mixed with 100 ml of fresh pooled sera from HIV-seronegative healthy persons (normal human serum, NHS). HIV-1IIIB (100 TCID50) in 100 ml culture medium was added, and incubated at 37 8C for 1 h. Thus, final concentration of human complement in the samples added to target cells was 25%. In the case of each patient’s serum sample mixtures without complement in which NHS was replaced by culture medium was also tested. In each series of experiments, control cultures infected with the virus alone were established, too. Then, 5/105 MT-4 cells in growth medium were added and incubated in this medium without change for 5 days at 37 8C. Growth of HIV in the cultures was monitored on each day by the reverse transcriptase (RT) assay as outlined by Hoffman et al. [13]. Peak RT values were observed almost exclusively at day 5. The results were measured as c.p.m., and expressed finally by an index (E/N I) value which indicated the relationship between virus growth in serum or IgG containing samples compared to that measured in cultures infected with virus alone. Index values were calculated as follows: means of triplicate peak RT (c.p.m.) values measured in cultures which infected with the mixtures of virus, patient’s serum (and NHS) at day 5 were divided by the means of triplicate peak c.p.m. values which measured in control cultures which infected with virus alone at day 5. Samples with an E/N I less than 0.5 (twofold decrease in virus production) were considered as neutralizing whereas samples with an E/N I exceeding 2.0 (twofold increase in virus production) were considered as enhancing. Variation coefficient of the method was 7.23%. Pooled heattreated (56 8C, 30 min) HIV-seronegative serum at the same dilution as the test samples (1:64) was used as a negative control at each measurement; it never influenced the growth of HIV-1 in the system. 2.4. Determination of CD4/ cell counts T cell subset determinations were performed by flow cytometry (FACScan) by using the IMK Lymphocyte kit (Becton Dickinson, Mountain View, CA). 2.5. Measurement of HIV-1 RNA level in the plasma samples of the patients HIV-1 RNA levels in the plasma samples of the patients were determined with the NASBA (Organon Teknika, Turnhout, Belgium) assay.
2.3. Measurement of in vitro HIV enhancement and neutralization
2.6. Statistical analysis
The assay was carried out as described previously [10]. Briefly, serum or IgG samples from the patients were heat-treated (56 8C, 30 min) and diluted at 1:64 in
Since E/N I values measured in the different groups did not match Gaussian distribution, non-parametric tests were used. Groups were compared using the Mann-
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
27
Table 1 Changes in the E/N I values measured in the absence and presence of complement with serum samples of 28 HAART-treated HIV patients Designation of patient
Follow-up time, months
In the absence of complement
In the presence of complement
Before HAART
After HAART
Before HAART
After HAART
50 108 125 143 153 203 274 309 312 322 337 368 373 387 389 393 400 408 411 413 415 422 446 451 463 480 484 503
14.0 10.0 6.0 46.0 19.0 21.0 19.0 18.0 34.0 33.0 20.0 28.0 21.0 24.0 23.0 31.0 24.0 20.0 16.0 11.0 10.0 7.0 10.0 10.0 18.0 10.0 11.0 11.0
0.40 0.61 0.81 0.54 0.72 0.72 0.95 0.95 1.09 1.07 1.05 0.59 0.79 0.49 0.38 1.05 1.02 0.25 0.32 0.23 0.32 0.27 0.65 0.81 0.73 0.60 0.26 0.46
0.41 0.35 0.19 0.05 0.09 0.46 0.29 0.10 0.41 0.46 0.43 0.22 0.58 0.45 0.05 0.15 0.19 0.43 0.36 0.20 0.36 0.77 0.33 0.12 0.11 0.10 0.13 0.08
2.45 0.54 0.81 1.03 3.22 0.80 3.20 1.42 3.88 3.46 1.49 1.34 0.77 2.12 0.62 1.67 3.32 0.39 1.29 0.67 1.36 0.29 0.65 2.51 1.50 1.08 0.82 1.05
0.86 0.34 0.19 0.09 0.17 0.56 0.34 0.40 0.49 0.42 0.82 0.20 0.60 0.40 0.07 0.57 0.46 0.78 0.33 0.18 0.39 0.74 0.33 0.13 0.22 0.13 0.71 0.17
Median (interquartile range)
18.5 (10.5 /23.5)
0.63 (0.39 /0.88)
0.26* (0.12 /0.42)
1.32 (0.79 /2.29)
0.37** (0.19 /0.57)
Differences between the medians of paired groups (Wilcoxon signed rank test) *P/0.0001, **P B/0.0001.
Whitney U or Wilcoxon signed rank tests. Categorical data were compared by the Fisher’s exact test. P B/0.05 was considered to be statistically significant. Statistical analysis was performed using GRAPHPAD Prism version 3.00 for Windows, (GraphPad Software, San Diego, CA, http://www.graphpad.com).
3. Results 3.1. Changes in the E/N I values in HIV patients under HAART In the samples taken at the beginning of follow-up, median (25th /75th percentile) values of E/N indices measured in the presence of complement were 1.32 (0.79/2.29) (Table 1). Out of serum samples from the 28 patients, 8 sera (28.6%) had an E/N I exceeding 2.0 that is significantly enhanced HIV-infection in vitro. By contrast, only 2/28 (7.1%) samples had a E/N I B/0.5 that is neutralized virus in the presence of complement (Table 2). Enhancement occurred in none of the samples
with no complement added, while 10/28 samples (35.7%) neutralized the virus (Table 2). E/N I measured both in the absence and presence of complement significantly (P /0.0001 and P B/0.0001, respectively) decreased during the follow-up period of 18.5 (10.5 /23.5) months under HAART. There was a significant positive correlation (Spearman correlation coefficient R /0.40, P /0.036) between the length of follow-up period and the extent of E/N indices measured in the presence of complement, whereas no such correlation (R /0.29, P/0.128) was found with samples tested with no complement added. In the last samples, neutralization (E/N I B/0.5) was measured in the absence and presence of complement in 26/28 (92.9%) and 20/28 (71.4%) patients, respectively. By contrast enhancement was found with none of the samples (Table 2). 3.2. Comparison of the enhancement/neutralization of pooled serum and purified IgG preparations In order to exclude the possibility that antiviral drugs present in the test sera affect the assay used for the
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
28
Table 2 Occurrence of serum samples that enhanced, neutralized or did not neutralize or enhance the growth of the HIV-1IIIB strain in MT-4 cells in the absence or presence of human complement in 28 HIV patients before and after HAART
No complement added Complement added a
Neutralization (E/N Ia B/0.5)
No neutralization or enhancement (E/N Ia: 0.5 /2.0)
Enhancement (E/N Ia /2.0)
Before HAART
After HAART for 19.0 (10.5 /24.0) months
Before HAART
After HAART for 18.5 (10.5 /23.5) months
Before HAART
After HAART for 18.5 (10.5 /23.5) months
10
26
18
2
0
0
2
20
18
8
8
0
Enhancement/neutralization index.
measurement of HIV enhancement/neutralization, we prepared serum pools from eight patients each taken before and during HAART. IgG were purified from these pools, diluted to the IgG concentration in the 1:64 serum dilution and used in the absence (Fig. 2A) and presence (Fig. 2B) of complement in the enhancement/ neutralization assay as described above. Only weak differences were found between the results that obtained with sera and IgG preparations, HAART resulted in a marked drop of the E/N I values in the cultures inoculated with virus /IgG complexes as well.
3.3. Dynamics of the decrease of E/N I values during HAART In both types of assays E/N I values significantly decreased in 8 months after beginning of HAART. A further significant drop in the index values occurred between 8 months and the end of follow-up (Fig. 1).
3.4. Changes in the viral load and T cells subsets Viral load significantly decreased, absolute number and percentage of CD4/ T cells significantly increased, while percentage of the CD8/ cells significantly decreased during this period (Table 3).
4. Discussion Our present findings indicate that HAART switches balance of enhancement and neutralization toward neutralization. Data obtained in the present study support our previous findings on the frequent occurrence of complement-mediated antibody-dependent HIV enhancement in the sera of untreated patients in advanced stage of HIV disease. The median (25th/75th percentile) of the E/N indices (1.32 (0.79 /2.29)) measured in the present study in samples obtained before HAART was approximately the same as in our previous study (1.35 (1.00/1.84)) performed in the sera of HIV patients mostly never treated or treated only with AZT before blood sampling [10]. Serum samples 28.6 and 19.4% of the patients in the present and previous [10] study, respectively had an E/N I exceeding 2.0 that is significantly enhanced HIV-infection in vitro. In principle our present observations can be biased by two ways. First, antiretroviral drugs present in the test samples can influence the assay system we used. We tested serum samples at a dilution of 1:64, and at this dilution residual antiretrovirals can still be present in a high enough concentration to inhibit in vitro virus infection. According to the data presented in Fig. 1, however, this assumption seems to be highly improbable. E/N I values were found to be markedly and significantly lower in the samples taken 19 months than in those taken 8 months after the beginning of HAART,
Fig. 1. Dynamics of the changes in the E/N I values measured in the absence (A) and presence (B) of complement with serum samples of 28 HAARTtreated HIV patients. P values for the comparison with Wilcoxon signed rank test are indicated in the figures.
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
29
Table 3 Changes in viral load and T cell subsets in 28 HAART-treated HIV patients* Variables
Before HAART
After HAART for 18.5 (10.5 /23.5) months
Median (interquartile range) HIV-1 RNA, copies/ml
5000 (0 /53 000)
0 (0 /358)
Mean9/S.E.M. CD4/T CD4/T CD8/T CD8/T
cells/ml cells/% cells/ml cells/%
3729/42 19.19/1.8 11419/107 55.19/2.3
P value Wilcoxon signed rank test 0.0090 Paired t test
5349/42 25.69/2.3 11719/121 48.99/2.4
B/0.0001 0.0003 0.7375 0.0007
* During a follow up of 18.5 (10.5 /23.5) months.
although the patients got the same type of HAART protocols, therefore the concentration of the antiretroviral drugs in their serum samples were approximately the same at both times. Moreover, as it is shown in Fig. 2, a marked drop of the E/N I values occurred when cultures inoculated with mixtures of HIV and purified IgG prepared from serum pools taken before and during HAART, respectively, were compared. Second, apparently it is surprising that measurements performed in our system with one single T cell adapted strain, HIV-1IIB can provide clinically meaningful results, whereas in most other systems use of primary HIV-1 isolates and assays with primary cells are essential for in vivo relevance. Recent elegant studies of Mitchell et al. [3] that are in accordance with the earlier observations of Robinson et al. [4,5] help to resolve this apparent paradoxical situation. According to these investigations the primary antigenic domains responsible for the C-ADE of HIV and SIV resides in the principal immunodominant sequence of gp41. This is a conservative sequence since according to the sitedirected mutagenesis experiments of Mitchell et al. [3] it is critical for the gp120/gp41 interaction. According to their model, binding site for enhancing antibodies on the
primary association site with gp41 is surrounded by amino-linked oligosaccharides which are likely sites for the covalent attachment of the C3dg peptide of C3, the physiological ligand of the CR2 receptor. Binding site for enhancing antibodies on gp41 may be accessible after release of gp120 from gp41 or after a conformational change in gp120 upon association with CD4. Interestingly enough HAART decreased E/N I values in the cultures infected with the serum /virus mixtures without complement, too although the extent of E/N I drop correlated to the length of HAART only in the complement-containing cultures. Therefore it can be assumed that the enhancing antibodies may exert their effect on a complement-independent way as well, the addition of complement in vitro or the presence of complement in vivo, however, markedly amplifies this enhancement. Therefore it seems probable that the apparent decrease in the E/N I values in due to the changes in the generation of enhancing and/or neutralizing antibodies that is to the drop of the production of enhancing antibodies, or an increased production of neutralizing antibodies. Changes in the titers of neutralizing antibodies in chronically infected HIV patients during
Fig. 2. Changes in the E/N I values in the absence (A) and presence (B) of complement with serum pools (SP) taken from the same 4-4 patients before (SP1A, SP2A) and after HAART (SP1B, SP2B) for 18.5 (10.5 /23.5) months as well as with IgG preparations (IgG1A, IgG1B, and IgG2A, IgG2B) purified from these pools. The IgG concentration of the sera and purified preparations was the same. No significant differences were found between serum pools and IgG preparations in the E/N I values.
30
D. Ba´nhegyi et al. / Immunology Letters 89 (2003) 25 /30
HAART were studied by several groups [14 /16]. These studies were performed in limited numbers of patients and gave conflicting results. Kim et al. [15] serially tested the neutralizing antibody titres against a panel of HIV-1 primary isolates in 11 patients under HAART with sustained B/400 copies/ml viral load. Development of increased neutralizing antibodies against some isolates occurred only in roughly half of the patients. Kimura et al. [16] investigated 19 chronically infected patients who were undergoing HAART. Reconstitution of or increase in neutralization activity was observed in four of 19 patients during HAART but neutralization activity was more or less unchanged in most patients. It seems therefore improbable that the marked decrease observed in the E/N I values in our present study was the consequence of an increased neutralization activity */ at least neutralization measured by conventional methods */during HAART. By contrast, several observations, described in our previous papers [9,10] and by an other group [11] indicate that C-ADE is associated with and may potentially facilitate progression of HIV disease. Considering these observations it seems probable that drop of antigen stimulus in patients under HAART results in a markedly decreased production of the antibodies mediating C-ADE and consequently to a gradual and dramatic (more than threefold) decrease in the E/N I values. Moreover, recently it has been reported that enhancing antibodies may also contribute to the selection pressure operating on the circulating population of polymorphic HIV-1 variants [17]. Most importantly, Verrier et al. [18] demonstrated that whole virus-particle based HIV-1 vaccines in hyperimmunized rhesus macaques can play a facilitating role in the transmission of the virus and/or evolution of the disease. Therefore our present findings indicate that besides its main effect */ decrease of viral load */HAART results in an additional beneficial effect: disappearance of C-ADE and shift of the balance toward dominance of the antibodies that neutralize HIV-1 in vitro even in the presence of complement. It was described by several groups [19 / 22] that there is a drop of B cell activation and IgG levels during HAART. Our present findings indicate that the amounts of C-ADE mediating antibodies changes in parallel with the amounts of IgG-specific antibodies in the patients on HAART, similarly to the very beginning of HIV infection when the appearance of C-ADE coincides in time with the appearance of HIVspecific antibodies [10].
Acknowledgements This work was supported by the following grant: OTKA (National Research Fund) T032661 (GF).
References [1] A. Takeda, C.H. Tuazon, F.A. Ennis, Science 242 (1988) 580 / 583. [2] W.E. Robinson, Jr., D.C. Montefiori, W.M. Mitchell, Biochem. Biophys. Res. Commun. 149 (1987) 693 /699. [3] W.M. Mitchell, L. Ding, J. Gabriel, AIDS 12 (1998) 147 /156. [4] W.E. Robinson, T. Kawamura, D. Lake, Y. Kasuho, W.M. Mitchell, E.M. Hersch, J. Virol. 64 (1990) 5301 /5305. [5] W.E. Robinson, Jr., M.K. Gorny, J.Y. Xu, W.M. Mitchell, S. Zolla-Pazner, J. Virol. 65 (1991) 4169 /4177. [6] J. Homsy, M. Tateno, J.A. Levy, Lancet I (1988) 1285 /1286. [7] F.D. To´th, B. Szabo´, E. Ujhelyi, K. Pa´lo´czi, A. Horva´th, G. Fu¨st, J. Kiss, D. Ba´nhegyi, S.R. Holla´n, AIDS 5 (1991) 263 /268. [8] B. McDougall, M.H. Nymark, G. Landucci, D. Forthal, W.E. Robinson, Jr., Scand. J. Immunol. 45 (1997) 103 /111. [9] G. Fu¨st, F.D. To´th, J. Kiss, E. Ujhelyi, I. Nagy, D. Ba´nhegyi, AIDS 8 (1994) 603 /609. [10] J. Szabo´, Z. Proha´szka, F.D. To´th, A. Gyuris, J. Segesdi, D. Banhegyi, E. Ujhelyi, J. Minarovits, G. Fu¨st, AIDS 13 (1999) 1841 /1849. [11] R.A. Subbramanian, J.M. Xu, E. Toma, R. Morriset, E.A. Cohen, J. Menezes, A. Ahmad, J. Clin. Microbiol. 6 (2002) 2141 /2146. [12] L. Hudson, F.C. Hay, Practical Immunology, Blackwell Scientific Publications, Oxford, 1989, pp. 299 /300. [13] A.D. Hoffman, B. Banapour, J.A. Levy, Virology 147 (1985) 326 /335. [14] J.M. Binley, A. Trkola, T. Ketas, D. Schiller, B. Clas, S. Little, D. Richman, A. Hurley, M. Markowitz, J.P. Moore, J. Infect. Dis. 182 (2000) 945 /949. [15] J.H. Kim, J.R. Mascola, S. Ratto-Kim, T.C. VanCott, L. LoomisPrice, J.H. Cox, N.L. Michael, L. Jagodzinski, C. Hawkes, D. Mayers, B.L. Gilliam, D.C. Birx, M.L. Robb, AIDS Res. Human Retroviruses 11 (2001) 2021 /2034. [16] T. Kimura, K. Yoshimura, K. Nishihara, Y. Maeda, S. Matsumi, A. Koito, S. Matsushita, J. Infect. Dis. 185 (2002) 53 /60. [17] D. Davis, H. Trischmann, D.M. Stephens, P.J. Lachmann, J. Med. Virol. 64 (2001) 207 /216. [18] F. Verrier, C. Moog, F. Barre-Sinoussi, E. Van der Ryst, C. Spenlehauer, M. Girard, Bull. Acad. Natl. Med. 184 (2000) 67 / 84. [19] S. Yokozaki, J. Takamatsu, I. Nakano, Y. Katano, H. Toyoda, K. Hayashi, T. Hayakawa, Y. Fukuda, Blood 96 (2000) 4293 / 4299. [20] Y. Chong, H. Ikematsu, M. Yamamoto, K. Chijiwa, W. Ariyam, J. Hayashi, T. Shirai, S. Kashiwagi, Kansenshogaku Zasshi 75 (2001) 535 /540. [21] C.E. O’Sullivan, R. Peng, K.S. Cole, R.C. Montelaro, T. Sturgeon, H.B. Jenson, P.D. Ling, J. Med. Virol. 67 (2002) 320 /326. [22] M.A. Jacobson, H. Khayam-Bashi, J.N. Martin, D. Black, V. Ng, J. Acquir. Immune Def. Syndr. 14 (2002) 472 /477.