Immunoglobulin G (IgG) and IgA, but also Nonantibody Factors, Account for In Vitro Neutralization of Human Immunodeficiency Virus (HIV) Type 1 Primary Isolates by Serum and Plasma of HIV-Infected Patients

Abstract

The factors present in serum and plasma samples of human immunodeficiency virus (HIV)-infected patients that are responsible for the neutralization of four HIV type 1 (HIV-1) primary isolates in vitro have been analyzed. Purification of immunoglobulins (Ig) by affinity chromatography showed that the activities were mostly attributable to IgG and less frequently to IgA. For two samples, we have shown that the high-level and broad-spectrum inhibitory activity was essentially caused by non-Ig factors interfering with the measurement of antibody-specific neutralizing activity.

Although the role of neutralizing antibodies (nAbs) in the control and prevention of human immunodeficiency virus type 1 (HIV-1) infection has long been the subject of controversy, recent reports showing protection of macaques against pathogenic simian-human immunodeficiency virus (SHIV) infection after passive transfer of nAbs (1, 10, 15) have reinforced the idea that humoral immunity can be beneficial and should be stimulated by an efficient vaccine. However, insufficient knowledge of the mechanisms of HIV neutralization and of the epitopes targeted by Abs neutralizing a broad range of field isolates asks for further investigations. For that purpose, sera from infected patients represent a valuable source of naturally induced nAbs of great diversity. In this study, we have analyzed serum and plasma samples from infected patients for neutralizing activity against different primary isolates (PI), to select those containing nAbs with broad-spectrum activity. Purification of both immunoglobulin G (IgG) and IgA was performed to verify that neutralizing activity was antibody mediated. The inhibitory activity of the two most potent samples was shown to be caused by non-Ig factors that may interfere with the measurement of antibody-specific neutralizing activity.

Neutralizing activities of whole serum and plasma.

Peripheral blood mononuclear cells (PBMC), purified by Ficoll gradient and stimulated for 3 days with phytohemagglutinin A (PHA; Sigma), were used as target cells for the replication and study of the neutralization of different HIV-1 PI. Isolates Bx08 and Bx17 were kindly provided by H. Fleury (Bordeaux, France), 11105C was isolated from an infected individual in the Central African Republic, and Kon was provided by F. Barin (Tours, France). All PI were propagated once or twice, exclusively on PHA-stimulated PBMC, to obtain viral stocks. Large volumes of serum (100 ml) or plasma obtained by plasmapheresis (500 ml) were collected from 24 infected patients (13 serum and 11 plasma samples; approval was obtained from the Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale). These patients (19 European and 5 African) were either untreated or treated with one or two nucleoside analogs. The mean duration of infection was 7.3 years (range, 2 to 17 years), and the mean CD4 cell number was 740/mm³ (range, 147 to 1,800/mm³). The serum sample designated reference serum 2, available from the National Institutes of Health AIDS Research and Reference Reagent Program, was provided by the Agence Nationale de Recherche sur le SIDA. This serum sample is commonly used as a reference because of its broad neutralizing activity (19). A serum sample collected from an HIV-negative patient was also included in the study. All serum and plasma samples were heat inactivated before use (30 min at 56°C).

The neutralization assay combines serial dilutions of PI with serial dilutions of serum or plasma. Briefly, 25 μl of four fourfold dilutions of virus in quadruplicate wells was incubated for 1 h at 37°C, with 25 μl of serial serum-plasma dilutions, in a 96-well filtration plate (Durapor-Dv, 1.25-μm pore size; Millipore, Molsheim, France), before addition of 25 μl of PHA-stimulated PBMC (from a pool of five seronegative donors) at a concentration of 4.10⁶ PBMC/ml. After 24 h at 37°C, 100 μl of RPMI 1640 containing 10% fetal calf serum and 20 IU of interleukin-2 per ml was added to each well. Extensive washings (three washings of 200 μl of RPMI 1640 each) were performed by filtration on day 4 to remove free virus and antibodies. Cells were then cultured in complete medium (200 μl) until day 7 postinfection, at which time p24 was measured in the supernatants by enzyme-linked immunosorbent assay (ELISA) (Innotest; Innogenetics, Ghent, Belgium) to determine HIV-positive cultures. The viral titer (50% tissue culture infective dose) was determined in the presence (V_n) and in the absence (V₀) of the serum, according to the Reed and Muench method. The neutralization titer was defined as the reciprocal of the serum dilution resulting in a 90% decrease of the viral titer (V_n/V₀ = 0.1). This assay has been slightly modified from that previously described (12, 16) to achieve a 5- to 10-fold increase in sensitivity.

Monoclonal antibodies (MAbs) 2F5 and 2G12 (gifts of H. Katinger) and IgG1b12 (gift of D. Burton and P. Parren) were obtained through the National Institute of Biological Standards and Control Central Facility for AIDS Reagents. These standard PI-neutralizing antibodies (18) were included to allow comparison of the neutralization sensitivity of our four PI (Table 1). The three MAbs neutralize Bx08 at high concentrations (60 to 100 μg/ml). Bx17 and 11105C are not neutralized by 2G12 at 100 μg/ml, the highest concentration tested, and Kon is not neutralized by the three MAbs.

TABLE 1.

Neutralizing activities of MAbs and whole serum and plasma samples determined for four HIV-1 primary isolates^a

Sample or MAb	Characteristics of donor	Sample type	Neutralization of virus isolate (subtype, coreceptor use)
			Bx08 (B, R5)	Bx17 (A, R5)	11105C (A, R5)	Kon (A, X4)
HIV negative		S	—	—	—	—
2	E   M	S	—	—	—	—
3	NA  M	S	15	5	10	—
4	E   M	S	50	150	—	—
6	E   M	S	55	5	10	—
7	NA  M	S	—	25	5	—
8	E   F	S	200	—	45	—
11	E   M	S	5	10	10	—
12	E   M	S	—	—	20	—
13	E   M	S	—	—	5	—
14	E   M	S	30	5	5	—
18	E   F	S	100	25	15	120
19	E   M	S	350	400	250	300
20	E  M	S	35	—	—	—
21	E   F	P	—	—	—	—
22	E   F	P	—	5	—	—
23	CA  F	P	35	5	20	5
24	E   M	P	20	5	20	—
25	E   M	P	10	—	15	—
26	E   M	P	90	65	35	180
27	CA  F	P	—	—	10	—
28	E   M	P	5	5	10	—
29	E   F	P	10	—	20	—
32	E   M	P	—	—	—	—
33	CA  F	P	35	65	10	35
Reference serum 2		S	90	20	110	5
2F5			80	10	20	>100
2G12			100	>100	>100	>100
IgG1b12			60	100	60	>100

^aFor MAbs, neutralization values (in micrograms per milliliter) are the result of one experiment only, because of lack of material. For serum and plasma samples, the neutralizing titer is either from one experiment or the mean of two or three measurements (for high titers) (titers of >25 highlighted). E, European; CA, central African; NA, northern African; M, male; F, female; S, serum; P, plasma; —, titer of <5. 

Serum and plasma samples were analyzed for neutralizing activity of PI Bx08, Bx17, 11105C, and Kon. Of the 24 serum and plasma samples tested, 21 had a measurable neutralizing activity (Table 1). Sixteen serum and plasma samples neutralized virus Bx08, and 10 of them displayed a high rate of activity (defined as a neutralizing titer of >25). Bx17, 11105C, and Kon were, respectively, neutralized by 14, 18, and 5 of the samples tested and by 4, 3, and 4 samples with a neutralizing titer of >25. Ten serum and plasma samples, in addition to reference serum 2, had a high activity for at least one of the PI. It is noteworthy that the six samples (including reference serum 2) neutralizing Kon have an activity against the three other PI and that serum sample 19 and plasma sample 26 were able to strongly neutralize the four viruses. Thirteen serum and plasma samples were chosen to analyze activities after Ig purification.

Neutralizing activities of purified IgG and IgA fractions.

The serum and plasma samples were first fractionated by protein A-Sepharose affinity chromatography (Pharmacia) to separate unbound factors from IgG1, -2, and -4 eluted with a low-pH buffer (0.1 M glycine [pH 2.7]). Efficiency of purification was evaluated by detection of Ig isotypes in the collected fractions by specific ELISA. Briefly, 96-well plates (Nunc Maxisorp) were coated (100 μl) with anti-human IgA, IgG, and IgM (The Binding Site, Birmingham, United Kingdom) at a 1/1,000 dilution in 50 mM bicarbonate buffer (pH 9.6). Serial dilutions of the affinity column fractions were incubated for 2 h at 37°C. Bound Ig's were revealed by anti-human Ig conjugate coupled to horseradish peroxidase (Southern Biotechnology Associates) at a 1/10,000 dilution (17). Repeated experiments indicated that the elution was highly reproducible, allowing purification of 80 to 90% of total IgG (data not shown). Fractions were filtered through a 0.45-μm-pore-size filter (μSTAR; Costar), and neutralization assays were performed as for whole serum and plasma, taking into account the fivefold dilution of the fractions with respect to the original sample. The neutralizing activities of serum and plasma samples 4, 6, 8, 14, 18, 20, 23, and 33 and reference serum 2 were predominantly recovered in the IgG1-, -2-, and -4-containing fraction (Table 2), with neutralizing titers ranging from 26 to 100% of the initial value. Flowthrough fractions displayed 12.5 to 25% of the unfractionated sample neutralizing activity for serum and plasma samples 6, 8, and 33 and reference serum 2. Remarkably, for the broadly neutralizing serum sample 19 and plasma sample 26, the neutralizing activity was almost exclusively recovered in the flowthrough fraction.

TABLE 2.

Neutralizing titers of whole serum of plasma samples and corresponding protein A affinity column fractions for isolates Bx08 and Bx17^a

Sample	Bx08			Bx17
	Whole sample	Flowthrough	IgG	Whole sample	Flowthrough	IgG
2	—	—	—	—	—	—
3	10	—	—	5	—	—
4	60	—	40	160	—	50
6	40	10	25	5	—	—
8	280	35	120	—	—	—
11	5	—	—	15	—	—
14	30	—	20	5	—	—
18	75	15	20	25	—	—
19	350	200	10	560	230	—
20	35	—	20	—	ND	ND
23	30	—	15	5	ND	ND
26	90	105	—	65	75	—
33	40	—	40	65	10	50
Reference serum 2	90	15	50	25	—	25

^aValues correspond to one experiment. —, titer of <5 for whole sample or <8 for fractional samples; ND, not done; IgG, IgG1, -2, and -4. 

The residual neutralizing activity present in the flowthrough fractions can be attributed either to nonretained IgG1, -2, or -4, IgG3, IgA, or IgM or to nonantibody serum factors. To assess the involvement of IgA in this activity, further separation was achieved by Jacalin affinity chromatography (Sigma). Despite the fact that IgA2 is not retained by this lectin, this purification was performed, as the anti-HIV serum IgA response is almost completely restricted to the IgA1 subclass (7). The protein A flowthrough fractions bearing inhibitory activity (samples 6, 8, 18, 19, and 26 and reference serum 2) were loaded on Jacalin-Sepharose 4B columns, and purified IgA1 was eluted by 100 mM methyl α-d-galactopyranoside (Sigma). Purification efficiency and an absence of contaminating IgG were verified by ELISA as reported for protein A affinity chromatography, with an additional recording of IgA1 (anti-human IgA1 MAb, I-7262; Sigma; coated at a 1/500 dilution). Jacalin flowthrough and purified IgA1 fractions were filtered and evaluated for neutralizing activity (Table 3). In samples 6 and 8 and reference serum 2, neutralizing activities were associated only with IgA1. For serum sample 18, the activity detected in the protein A flowthrough fraction was not attributable to IgA1. Again, the activities of broadly neutralizing serum sample 19 and plasma sample 26 were almost entirely recovered in the flowthrough fraction.

TABLE 3.

Neutralizing titers (PI Bx08) of protein A flowthrough fraction (non-IgG1, -2, and -4) further separated by Jacalin-Sepharose chromatography^a

Sample	Titer after:
	Protein A flowthrough	Jacalin bound (IgA1)	Jacalin flowthrough
6	10	10	—
8	35	40	—
18	15	—	10
19	200	20	290
26	105	10	120
Reference serum 2	25	30	—

^a—, titer of <10. 

Neutralizing activities of the dialyzed flowthrough fractions.

To evaluate the contribution of antibodies remaining in the Jacalin-Sepharose flowthroughs to the neutralizing activity detected in these fractions, they were separated from smaller molecules by dialysis with 10-kDa (Slide-A-Lyzer; Pierce) and 50-kDa (Spectrapor; Spectrum Medical Industries) cutoff membranes. Two milliliters of flowthrough fraction was dialyzed against 1 liter of phosphate-buffered saline for 24 h at 4°C, with one replacement of the phosphate-buffered saline after 12 h. Results of neutralization assays performed with the dialyzed flowthrough fraction of plasma sample 26 are shown in Fig. 1. Dialysis with the 50-kDa membrane led to a drop of activity to under detectable levels. Neutralizing activity was, however, conserved after dialysis with the 10-kDa membrane, which allowed elimination of any drug that may have remained in the serum or plasma of treated patients. Comparable results were obtained with serum sample 19 (data not shown). Dialysis caused no loss of Ig of any class, as verified by specific ELISA, and did not affect the neutralizing activity of the protein A-Sepharose-purified IgG fraction of serum sample 8 (data not shown). The activity detected in the two strongly and broadly neutralizing serum sample 19 and plasma sample 26 flowthrough fractions can thus unequivocally be attributed to a nonantibody soluble factor(s), of a molecular mass(es) between 10 and 50 kDa.

FIG. 1.

Neutralization of Bx08, Bx17, and Kon by whole and dialyzed Jacalin flowthrough fraction of plasma sample 26. ND, not done.

Our results confirm that the majority of nAbs in the sera and plasmas of infected patients are of IgG isotype, although neutralizing IgA1 was detected in 4 of 21 neutralizing samples. A limited number of studies have reported the neutralization of primary strains of HIV-1 by serum IgA from infected patients (11) or exposed seronegative individuals (6, 14). These results show that PI-neutralizing IgA is induced during infection. However, circulating and secreted IgA originates from different compartments, and the presence of serum neutralizing IgA does not reflect mucosal IgA status. We detected neutralization of the four PI by purified Ig in 3 of 24 samples (samples 18, 23, and 33), which is close to the ±10% prevalence of broadly cross-neutralizing sera observed in other studies (3, 13). The purified IgG of these three patients neutralized PI Kon which was not neutralized by any of the three MAbs tested, suggesting the potential presence of as-yet-unidentified broadly neutralizing or synergistically acting antibodies in these samples. Interestingly, of these three samples, two were obtained from black African female patients (samples 23 and 33) (Table 1), supporting the observation made by Beirnaert et al. (3) that there may be a correlation between broadly cross-neutralizing activity, African origin, and gender. We have also shown that nonantibody inhibitory activities that may significantly affect the results of in vitro antibody-mediated neutralization studies are detected in 2 of 24 samples tested. Heat inactivation of the serum eliminated any complement-associated antiviral effect, and alpha interferon was below detectable levels in both serum sample 19 and plasma sample 26 (data not shown). Nonantibody-neutralizing activities could be due to cytokines, chemokines (e.g., RANTES, MIP-1α, MIP-1β, or SDF-1), or CD8⁺ T-cell antiviral factor, or more likely to a particularly powerful combination of some of those factors. In fact, none of these molecules can be definitively excluded after dialysis experiments, as the molecular masses of some are too close to the theoretical cutoff of the membranes, and some may be present as dimers. However, the neutralizing activity detected in our experiments is conserved at high dilutions compared to that observed either with chemokines (4, 5) or with CD8⁺ T-cell antiviral factor released from CD8⁺ T cells (2, 8). Moreover, additional experiments with three other PI as well as with the T-cell-line-adapted strain MN (data not shown) have confirmed that the non-Ig factors of samples 19 and 26 are active against both R5 and X4 strains, excluding the action of either α- or β-chemokines alone. It is no longer possible to determine whether the inhibitory factors were present in serum or plasma at the time of collection or were subsequently produced by donor cells ex vivo (9), but further experiments will be carried out to precisely define the nature of the unknown inhibitory factor(s) and to investigate the persistence of such a factor(s) in additional samples collected from the same patients. Because of the high and potent in vitro-inhibitory activity of those serum factors, we propose that, if Igs are not purified, serum and plasma samples should at least be dialyzed for performance of antibody-mediated neutralization studies.