Effect of Complement on HIV-2 Plasma Antiviral Activity Is Intratype Specific and Potent

Abstract

Human immunodeficiency virus type 2 (HIV-2)-infected individuals develop immunodeficiency with a considerable delay and transmit the virus at rates lower than HIV-1-infected persons. Conceivably, comparative studies on the immune responsiveness of HIV-1- and HIV-2-infected hosts may help to explain the differences in pathogenesis and transmission between the two types of infection. Previous studies have shown that the neutralizing antibody response is more potent and broader in HIV-2 than in HIV-1 infection. In the present study, we have examined further the function of the humoral immune response and studied the effect of complement on the antiviral activity of plasma from singly HIV-1- or HIV-2-infected individuals, as well as HIV-1/HIV-2 dually infected individuals. The neutralization and antibody-dependent complement-mediated inactivation of HIV-1 and HIV-2 isolates were tested in a plaque reduction assay using U87.CD4.CCR5 cells. The results showed that the addition of complement increased intratype antiviral activities of both HIV-1 and HIV-2 plasma samples, although the complement effect was more pronounced with HIV-2 than HIV-1 plasma. Using an area-under-the-curve (AUC)-based readout, multivariate statistical analysis confirmed that the type of HIV infection was independently associated with the magnitude of the complement effect. The analyses carried out with purified IgG indicated that the complement effect was largely exerted through the classical complement pathway involving IgG in both HIV-1 and HIV-2 infections. In summary, these findings suggest that antibody binding to HIV-2 structures facilitates the efficient use of complement and thereby may be one factor contributing to a strong antiviral activity present in HIV-2 infection.

INTRODUCTION

Intense research and efforts have been invested in the search for an effective HIV vaccine. Still, no such vaccine has been developed. According to our present understanding, a vaccine able to induce both broadly neutralizing antibodies (NAb) and cytotoxic T-lymphocyte responses against the virus would most likely represent the best strategy to pursue (1, 2). Studies on human immunodeficiency virus type 2 (HIV-2) infection are promising in that they may increase our knowledge about immune control of HIV infection. HIV-2 is known to be less transmissible and less pathogenic than HIV-1, and the majority of HIV-2-infected individuals remain asymptomatic much longer than do HIV-1-infected individuals (3–5). When matched for CD4⁺ T-cell counts, the plasma viral load in HIV-2-infected individuals is approximately 1 log lower than that observed in HIV-1-infected individuals (6).

The NAb response is more potent and broader in HIV-2 than in HIV-1 infection (7, 8). In addition, neutralization escape mutants emerge less frequently, if at all, in HIV-2 infection; this suggests that the HIV-2 envelope glycoprotein complex (Env) might play an important role in eliciting a more effective immune response (7–10). Indeed, the HIV-2 Env has been found to display multiple broadly cross-reactive epitopes and CD4 independence, both of which are characteristics that are uncommon in the HIV-1 Env (11). Furthermore, these features have been found to be correlated to the development of a potent and broad NAb response in HIV-2 infection (8, 10, 12).

In line with these observations, we recently reported on neutralizing activities (NAc) in the plasma of HIV-1- and/or HIV-2-seropositive individuals from Guinea-Bissau, a West African country with both HIV-1 and HIV-2 circulating in the general population (13). In this study, we compared, side-by-side, the breadth and potency of intra- and intertype NAc in plasma against a panel of HIV-1 and HIV-2 isolates and found that the potency of intratype NAc in HIV-2 infection was significantly higher than in HIV-1 infection (9). Interestingly, plasma from dually HIV-1- and HIV-2 (HIV-D)-infected individuals, tested for the first time, was found to display potent NAc against HIV-2 but not HIV-1, suggesting differences in the immunogenicity and/or antigenicity of the two viruses.

The antiviral effector functions of HIV-specific antibodies stretch beyond their binding to antigen and classical neutralization and include antibody-dependent cell-mediated cytotoxicity, opsonization, and the activation of complement (14, 15). The complement system is an integral part of innate immunity, providing a link to the adaptive immune responses (2, 16). Similarly to other pathogens, HIV-1 triggers a response by way of the complement system during an infection. Both neutralizing and nonneutralizing antibodies bound to the HIV-1 Env can activate the complement cascade (classical pathway). It has also been reported that HIV-1 can activate this pathway even in the acute phase of infection in the absence of HIV-1-specific antibodies through direct interaction between the Env glycoproteins gp41 and gp120 and the complement protein C1q (17). Furthermore, alternative and lectin pathways have also been implicated in the interaction of the HIV-1 Env and the complement system, in this case through an interplay between C3b and mannose-binding lectin (18–20). Thus, the role of the complement system in HIV-1 infection appears to be multifaceted and seems to be dependent on the presence of complement regulatory proteins (CRPs) in the HIV-1 envelope and on the presence of complement receptors (CRs) on the surfaces of target cells. Virus inactivation through opsonization and lysis has been reported to dominate when the expression of CRPs on the HIV-1 envelope is limited, whereas the presence of CRs on the target cells may enhance viral infectivity (21–23). With just two exceptions (24, 25), it should be noted that little is known on the role of complement in HIV-2 infection.

In this study, we focused on the role of complement in HIV-1 and HIV-2 infections by determining the antiviral effects of complement added to HIV-1- and/or HIV-2-seropositive plasma. The findings from this study suggest that the antiviral effect of antibody-dependent complement-mediated inactivation (ADCMI) is intratype specific and appears more potent in HIV-2 infection than in HIV-1 infection.

MATERIALS AND METHODS

Study population.

The study participants resided in three adjacent suburban districts, Bandim 1, Bandim 2, and Belem in Bissau, Guinea-Bissau. They were part of general population adult cohorts that have been monitored for retroviral infections since 1987 within the framework of the Bandim Health Project (26–29). The last serosurvey of HIV infection took place between 2004 and 2006. It comprised 2,548 individuals in 384 randomly selected houses; the number of houses represented a 10% sample of the total number of houses in the study area. From this screening, all individuals with dual HIV-1 and HIV-2 infection (HIV-D) (n = 7) were identified. Singly HIV-1- or HIV-2-infected individuals were randomly selected in a case-control manner from the same cohort (n = 20 in each group). These plasma samples were used in our previous study, which explored intra- and intertype neutralization of HIV-1 and HIV-2 (9). Plasma samples with a volume of ≥150 μl remaining were selected for the current study (14 HIV-1-, 17 HIV-2-, and 5 HIV-D-infected samples). The characteristics of the study participants are summarized in Table 1. While the groups were similar in regard to gender, CD4⁺ T-cell counts, human T-cell leukemia virus type 1 (HTLV-1) status, and plasma IgG levels, there were significant differences with regard to their plasma HIV load and age.

Table 1.

Characteristics of study participants

Characteristic	Value for group
	HIV-1 positive (n = 14)	HIV-2 positive (n = 17)	HIV-D positive (n = 5)	P valuea
Median (IQRb) age (yr)	33.9 (28.8–50.0)	61.6 (54.0–68.8)	33.5 (30.4–69.2)	<0.001
Gender (% women)	78	65	100	0.281
HTLV status (% positive)	21	41	0	0.154
% of participants with a viral loadc of:
<1,001 RNA copies/ml	21	70	20
1,001–10,000 RNA copies/ml	28	24	0
>10,000 RNA copies/ml	43 (data missing for 1)	6	80	0.005
Median (IQR) CD4+ T-cell count (μld)	316 (266–492)	424 (254–500)	212 (25–377)	0.266
% of participants with a CD4+ T-cell countd of:
>499 μl	21	29	0
200–499 μl	64	53	40
<200 μl	7e	6f	40e	0.159
Median (IQR) total IgG (mg/mlg)	21.0 (14.7–23.7)f	13.0 (10.8–15.4)e	17.8 (8.4–24.3)e	0.112

^aP values calculated using the Kruskal-Wallis test, comparing means over the columns, the chi-square test, or Fisher's exact test, as appropriate.

^bIQR, interquartile range.

^cThe plasma HIV-1 and HIV-2 loads were analyzed by measuring reverse transcriptase (RT) activity using the CAVIDI ExaVir load kit (Cavidi Tech AB, Uppsala, Sweden) according to the manufacturer's instructions. Both HIV-1 and HIV-2 were detected with this method and are presented as RNA copy equivalents/ml.

^dThe absolute CD4⁺ T-cell counts were determined by either flow cytometry on a FACStrak instrument (Becton Dickinson, San Jose, CA) using three two-color immunofluorescence reagents (CD45/CD14, CD3/CD4, and CD3/CD8) (Simultest; Becton Dickinson, San Jose, CA) in combination with leukocyte counts or flow cytometry on a CyFlow instrument (Partec GmbH, Münster, Germany) using a CD4% antibody kit (CyTecs GmbH, Görlitz, Germany) according to the manufacturer's instructions.

^eData missing for 1 participant.

^fData missing for 2 participants.

^gIgG levels were measured by an in-house ELISA (10) with reagents from Jackson Immunotech (Marseille, France). In brief, plates were coated overnight with AffiniPure goat anti-human IgG (20 μg/ml), alkaline phosphatase-conjugated anti-human IgG (diluted to 1:5,000) was used as a detection antibody, and purified human IgG was used as a standard.

Blood sampling and HIV-1, HIV-2, and HTLV-1 status determination.

Venous blood samples were drawn and collected in cell preparation tubes (CPT) (BD Biosciences, San Jose, CA) with sodium citrate as the anticoagulant, and plasma was separated according to the manufacturer's instructions. Plasma was then kept frozen until its use in neutralization assays, virus isolation, viral load determinations, and other laboratory analyses. Serological testing for HIV antibodies was done using the Behring Enzygnost HIV-1/HIV-2 enzyme-linked immunosorbent assay (ELISA) (Behring, Marburg, Germany). Confirmation and HIV-1 and HIV-2 diagnoses were performed by using Capillus HIV-1/HIV-2 (Cambridge Biotech Limited, Galway, Ireland) and ImmunoComb II HIV-1 and HIV-2 BiSpot RST (Origenics, Yavne, Israel). Tests for HTLV antibodies were performed using HTLV enzyme immunoassay (EIA) (Murex, Dartford, United Kingdom), and confirmation was made using the line immunoassay INNO-LIA (Innogenetics, Ghent, Belgium).

Primary HIV-1 and HIV-2 isolates.

For the analyses of the neutralizing activity of HIV-positive plasma, two HIV isolates, one HIV-1 and one HIV-2, were used. The HIV-1 circulating recombinant form (CRF)02_AG isolate, AG GB/30, originated from the same Guinea-Bissau study population from which the analyzed plasma was obtained (9, 30). The HIV-2 subtype A isolate, A WA/P1-1991, originated from West Africa (10) and was isolated by cocultivation of a patient's peripheral blood mononuclear cells (PBMC) with phytohemagglutinin (PHA)-activated blood donor PBMC as described previously (9, 10, 31). The HIV-1 subtype CRF02 (AG), GB/30, was isolated from plasma based on the method of Costa et al. (32) and was described in detail previously (30). The two isolates included in the present study were tested for coreceptor use on the U87.CD4 indicator cell lines and were found to use CCR5 for cell entry, which was our rationale for the use of U87.CD4.CCR5 cells as targets in the neutralization assays (9, 10, 33, 34). Furthermore, in order to evaluate the effect of complement independently of neutralization sensitivity, the isolates were selected to represent neutralization sensitivities similar to those of each other (Table 2). The included HIV-2 isolate was chosen based on the pattern of autologous neutralization, it being the closest to the general pattern of HIV-1 autologous neutralization (see patient 1 in reference 10).

Table 2.

Reciprocal titers^a of neutralizing activity and antibody-dependent complement-mediated inactivation in individual plasma samples against HIV-1 and HIV-2 isolates

^aGiven that reciprocal titers correspond to 1/dilution of plasma giving a 30% inhibitory concentration (IC₃₀) in the plaque reduction neutralization assay applied, and boxes are color coded as follows: white, <40 to 160; yellow, 320 to 1,280; orange, 2,560 to 20,480.

Complement.

Pooled human AB serum was the source of complement (Innovative Research, Novi, MI). Because of the labile nature of complement, aliquots of the serum pool were immediately stored at −80°C. The aliquots were thawed only once when used in the assays. To obtain a heat-inactivated equivalent complement (HIC), the complement source was incubated at 56°C for 1 h and stored at −80°C in aliquots.

Purification and quantification of plasma total IgG.

Six plasma samples, three each from HIV-1- and HIV-2-infected individuals, were selected for IgG purification. For this purpose, high-performance protein G-Sepharose (HP), prepacked into spin columns (GE Healthcare Life Sciences, Buckinghamshire, United Kingdom), was used. Briefly, plasma samples were first inactivated at 56°C for 30 min and then clarified by centrifugation at 4,000 × g for 20 min. Two hundred microliters plasma and 400 μl binding buffer (20 mM sodium phosphate [pH 7.0]) (GE Healthcare Life Sciences) were added to the equilibrated protein G-Sepharose HP, and the mixture was incubated at room temperature for 4 min in a tube rotator. Thereafter, unbound fractions were removed by centrifugation at 70 to 100 × g for 30 s and kept in a refrigerator. The columns were then washed 2 times with 600 μl of binding buffer. Neutralization buffer (30 μl, 1 M Tris-HCl [pH 9.0]) (GE Healthcare Life Sciences) was added to the bottom of fresh collection tubes, and bound IgG was eluted with elution buffer (400 μl, 0.1 M glycine-HCl [pH 2.7]) (GE Healthcare Life Sciences) and centrifuged at 70 × g for 30 s into the neutralization buffer. Elution was repeated with another 400 μl of elution buffer. IgG purification was repeated with the unbound fractions that were kept in the refrigerator. The eluted IgG was concentrated using Amicon Ultra-15 centrifugal filter units with a 30,000-Da cutoff (Merck Millipore, Billerica, MA). Purified IgGs were kept at −80°C until use. The yield of IgG was analyzed by an in-house ELISA as described previously (10, 35).

Neutralizing activity and antibody-dependent complement-mediated inactivation by plaque reduction.

Neutralizing and antibody-dependent complement-mediated inactivation (ADCMI) activities of plasma were determined using a plaque reduction assay with U87.CD4.CCR5 cells as target cells (36). U87.CD4.CCR5 cells were grown in complete Dulbecco's modified Eagle's medium (DMEM) medium, i.e., DMEM supplemented with 10% fetal calf serum (FCS), 0.1 μg/ml streptomycin, and 0.1 U/ml penicillin (Invitrogen, Carlsbad, CA) and split 1:3 to 1:6 twice a week. For the analyses of neutralization and ADCMI activities, virus stocks, heat-inactivated plasma (at 56°C for 1 h), and HIC (neutralization) or complement (ADCMI) were mixed and diluted in DMEM infection medium, i.e., complete DMEM medium supplemented with 2 μg/ml Polybrene (Sigma-Aldrich, St. Louis, MO). Following the incubation of virus-plasma-HIC and complement mixtures for 1 h at 37°C, the mixtures were distributed in triplicate (200 μl/well) to 48-well microtiter plates containing U87.CD4.CCR5 cells at a 50% confluence (seeded the day before). The final concentrations of the virus then corresponded to 20 to 40 PFU/well, the final plasma dilution started at 1:40, and the final HIC or complement dilution was 1:10. After 2 h of incubation at 37°C, 300 μl/well of fresh medium was added, and incubation was continued overnight. The day after infection, the medium was replaced with 1 ml of DMEM infection medium. Positive neutralization controls consisted of wells with virus, plasma with known NAc, and HIC or complement, negative controls consisted of wells with virus mixed with a pool of plasma from five HIV-uninfected individuals participating in the serosurvey of HIV infection in Bissau during 2004 to 2006 and HIC or complement, and cell controls consisted only of cells. The assay was terminated on day 3 by fixation with a methanol-to-acetone ratio of 1:1. The number of PFU was determined following hematoxylin staining. Plasma NAc and ADCMI were calculated by using the formula [1 − (PFU with plasma-HIC or complement/PFU without plasma-HIC or complement)] × 100, which expresses the degree of reduction in the number of PFU in the presence of plasma-HIC or complement relative to the wells without plasma-HIC or complement. In the ADCMI assays, the activities of purified IgG and plasma were compared: two plasma samples and corresponding IgG from the current cohort and four plasma samples and IgG, named AG, from another cohort (women attending the Agiubef sexual health and family planning clinic for urogenital problems in Bissau, Guinea-Bissau, from November 2006 to January 2011) were selected on the basis of the availability of plasma volume. Here, IgG levels were analyzed at a dilution that gave a final concentration corresponding to the IgG level in 1:40 diluted plasma. The cutoff for neutralization and ADCMI was based on intra-assay variation that was determined in three consecutive neutralization assays run on the same day and was repeatedly found to have a SD of <10%. The cutoff chosen corresponded to >3 SD, that is, a 30% reduction in the number of plaques (36).

Magnitudes of neutralizing activity and antibody-dependent complement-mediated inactivation.

For determining the magnitudes of NAc and ADCMI, plasma samples were titrated against HIV-1 and HIV-2 isolates. To quantify the magnitudes of NAc and ADCMI, two different methods were used, one based on log-transformed titers (37) and the other on the area under the curve (AUC). Log-transformed titers were calculated by dividing the highest dilution giving neutralization by 100 before applying a log-base-3 transformation and then adding 1: Y = [log 3(dilution/100)] + 1. All titers below the limit of detection were given a value of 33 for the sake of calculation of a neutralization and ADCMI score (37). The AUC was measured using an xy graph with the x axis showing serial plasma dilutions starting from 1:40 and the y axis showing the neutralization or ADCMI above the cutoff (30%). Analysis was performed using GraphPad Prism 5 version 5.2. The AUC for NAc and ADCMI below 30% at a 1:40 plasma dilution was 0 (38). In the AUC-based method, the effect of complement was calculated by subtracting the AUC-based magnitude (magnitude_AUC) of ADCMI from the magnitude_AUC of NAc (ΔAUC = AUC_ADCMI − AUC_NAc). Using the titer-based method, the effect of complement was measured by dividing the highest neutralizing titer of ADCMI by the highest titer of NAc, and titers of <1:40 were given a value of 1:20.

Statistical methods.

For comparisons between unrelated categorical variables, Fisher's exact test or chi-square test was used as appropriate. The differences between independent subgroups of numerical variables were assessed by the Mann-Whitney test or Kruskal-Wallis test as appropriate. Neutralizing activities with and without complement were compared using the Wilcoxon signed-rank test. These analyses were performed using SPSS and GraphPad Prism 5 version 5.2. Multivariate regression analyses including the variables of age, gender, HTLV-1 status, viral load, CD4 T-cell count, and IgG level were performed using Stata version 11. Since 6 of the 31 (19%) participants had missing information in one or more of the variables of IgG level, viral load, and/or CD4 T-cell count, multiple imputation analyses were performed.

Ethical considerations.

The study was approved by the Guinea-Bissau Government Ethics Committee and the ethical committee at the Lund University, Sweden. The study participants were counseled and then provided informed verbal consent. The participants were offered a medical examination with free essential medications. Antiretroviral treatment was not available in the country at the time of plasma sampling but is now provided to this study population.

RESULTS

Intratype antiviral activity of HIV-1 and HIV-2 plasma in the presence or absence of complement.

In order to study the impact of complement on intratype neutralization, we analyzed the antiviral activity of plasma from HIV-1- and HIV-2-infected individuals against the HIV-1 and HIV-2 isolates, respectively, in the presence or absence of complement. The magnitude of intratype neutralization was scored as the highest plasma dilution mediating neutralization. The reciprocal neutralization titers of each plasma-virus combination in the presence of heat-inactivated complement (HIC) were compared to antibody-dependent complement-mediated inactivation (ADCMI) in a heat map (Table 2). The median neutralizing titer (NT) for HIV-1 plasma was 1:40, whereas the median ADCMI titer was found to be 1:160, i.e., 4-fold higher. The increase from the median NT to the ADCMI titers of HIV-2 plasma was even larger, and median titers differed 32-fold and were 1:40 and 1:1,280, respectively (Table 2). When analyzing the performances of each individual plasma, we noted that ADCMI titers were higher than the NT for all HIV-2 plasma samples, whereas increases were absent for 3 (of 14) HIV-1 plasma samples.

Quantitative comparisons of the titers were performed in two ways, by calculating the magnitudes using logarithmic transformation (magnitude_log) and measuring the area under the curve (magnitude_AUC). Both types of comparisons showed that the magnitudes were statistically different for intratype ADCMI and NT, with P values of <0.01 and <0.001, respectively; thus, ADCMI was of higher magnitude than NT in both the HIV-1 and HIV-2 intratype activities (Fig. 1). These results suggest that the addition of complement may increase the magnitudes of intratype antiviral activities of both HIV-1 and HIV-2 plasma samples.

Fig 1.

Log- and AUC-based magnitudes of antiviral activities in HIV-1, HIV-2, and HIV-D plasma in the presence or absence of complement (C′). Shown are the median magnitudes of antiviral activities of 14 HIV-1, 17 HIV-2, and 5 HIV-D plasma samples against an HIV-1 isolate (A and B) and an HIV-2 isolate (C and D). The diameters of the circles and the areas of the squares correspond to the log-based and AUC-based magnitudes of plasma antiviral activities, respectively, in the presence (black) or absence (gray) of complement. The log-based magnitudes of 0.0, 0.2, 0.8, 1.4, and 3.3 correspond to the reciprocal titers of <40, 40, 80, 160, and approximately 1,280, respectively. **, P < 0.01; ***, P < 0.001.

After assessing the ratio of titers (ADCMI/NT) and ΔAUC (ADCMI − NT), it appeared that the addition of complement to the intratype reaction of HIV-2 plasma had a stronger effect than the addition of complement to the intratype reaction of HIV-1 plasma (P < 0.001 for both) (Fig. 2A and B, respectively). The increase in intratype antiviral activity mediated by complement added to HIV-2 plasma was found to be 4-fold higher than the effect of complement on HIV-1 plasma. These findings were confirmed by multivariate statistical analysis using the AUC-based effect of complement as an outcome, in that the type of HIV infection was independently associated with the effect of complement (P < 0.001).

Fig 2.

Titer- and AUC-based complement effects in HIV-1, HIV-2, and HIV-D plasma against HIV-1 and HIV-2 isolates. (A) Titer-based complement effects were calculated by dividing the highest titer of ADCMI by the highest titer of NAc. (B) AUC-based complement effects (ΔAUC) were calculated by subtracting the AUC-based magnitude of ADCMI from the AUC-based magnitude of NAc. ***, P < 0.001.

Intertype antiviral activities of HIV-1 and HIV-2 plasma in the presence or absence of complement.

Analyses of the effect of complement on intertype neutralization were performed in two ways. First, the magnitudes of ADCMI and NT of HIV-2 plasma for the HIV-1 isolate were compared with the magnitudes of ADCMI and NT of HIV-1 plasma for the HIV-2 isolate. Intertype neutralization by HIV-1 plasma was not improved by the addition of complement, as the median titer remained at <1:40. Similarly, the intertype antiviral activity of HIV-2 plasma, with a median titer of 1:40, remained unchanged after the addition of complement (Table 2). A quantitative comparison of the magnitudes of intertype NT and ADCMI showed no statistically significant differences (Fig. 1). These findings suggest that complement has little effect, if any, on virus inactivation in the intertype neutralization reaction.

Second, we asked whether complement would influence the specificity of neutralization reactions. Both HIV-1 and HIV-2 plasma samples, whether with or without complement, displayed higher antiviral activities against the corresponding virus type (even though HIV-2 plasma median NT values were similar) (P < 0.001 and P < 0.05, respectively, for both magnitude_log and magnitude_AUC). Accordingly, in all these comparisons, the magnitudes of intratype NT or ADCMI were significantly higher than the magnitudes of the corresponding intertype reactions (Table 2). These results demonstrate that intratype-specific neutralization can be enhanced further by the addition of complement without affecting intertype ADCMI.

HIV-1- and HIV-2-directed activities of HIV-D plasma in the presence or absence of complement.

The median neutralizing titers of HIV-D plasma against HIV-1 isolates did not differ from the titers from those who were infected singly with HIV-1 (1:40 in both cases) but were somewhat higher against HIV-2 isolates (median titer, 1:80) (Table 2). The addition of complement appeared to result in an increase in antiviral activity against both HIV-1 and HIV-2, such that median ADCMI titers were 1:80 and 1:160, respectively. When the complement effects were evaluated by AUC, median-based increases were 3.5 to 70.5 against HIV-1 and 8.3 to 78.2 against HIV-2 (Fig. 1B and D). However, the statistics did not support these differences due to the low number of HIV-D subjects included in the study. Even so, the apparent lower magnitude of ADCMI with HIV-D plasma compared to that observed with the HIV-2 plasma (Fig. 1C and D) suggests that the utilization of complement by HIV-D plasma in inactivating HIV-2 may be impaired in dual HIV-1 and HIV-2 infection.

Characterization of the complement effect.

The complement cascade can be activated by antibody-dependent (classical), antibody-independent (alternative), or lectin (14) pathways. In order to determine the pathway involved in the ADCMI, the effects of complement alone (1:10) against HIV-1 and HIV-2 isolates were tested. Complement alone did not have any antiviral effect (data not shown), thereby excluding the role of the alternative and lectin pathways. To confirm that the classical complement pathway was responsible for the antiviral effect, we performed the intratype ADCMI assay comparing the plasma and corresponding purified IgG concentrations of three HIV-1- and three HIV-2-infected individuals, where IgG concentrations corresponded with those of the plasma. The fold differences between plasma and IgG ADCMI showed little variation and were uniformly low (range, 1.4- to 2.3-fold) (Fig. 3). Thus, these results suggest that the complement effect was largely dependent on IgG concentrations in both HIV-1 and HIV-2 intratype ADCMI.

Fig 3.

Intratype antibody-dependent complement-mediated inactivation of plasma and corresponding plasma IgG concentrations in HIV-1 (A) and HIV-2 (B) plasma. For the neutralization of the isolates, plasma and purified IgG were titrated in the presence of complement (1:10 dilution). The assay cutoff was 30%, as indicated by the line.

DISCUSSION

In the present study, we have found that the effect of complement on HIV-2 plasma antiviral activity is intratype specific and potent. The results also suggest that the magnitude of the antiviral activity attributed to ADCMI is greater in the plasma of HIV-2-infected individuals than that of HIV-1-infected individuals. Furthermore, the results imply that the activation of the classical complement pathway through IgG is mainly responsible for this effect.

The reason for the intratype, but not the intertype, plasma antiviral activities being potentiated by complement is likely due to several features, including antibody binding affinity, bivalent or monovalent binding, and the accessibility and density of epitopes in the viral envelope. Bivalent antibody binding results in high-avidity interaction between the antibody fragment, antigen binding (Fab) regions, and the epitope, which in turn triggers flexing of the Fab arms and a conformational change in the fragment, the crystallizable (Fc) region (39, 40). This structural change gives access to the C1q receptor site on the IgG Fc region, which is required for efficient complement activation and subsequent ADCMI (39). HIV intertype reactive antibodies could be envisioned to be of lower affinity and to bind in a monovalent manner, excluding the potentiation of intertype ADCMI. Indeed, recent work has shown that neutralizing activity is greatly enhanced by the bivalent binding of antibodies compared to monovalent binding (41).

An intriguing finding of our studies is the difference in the capacity of complement to potentiate antiviral activities of HIV-1 and HIV-2 plasma samples in the intratype reaction. For HIV-1, it has been shown that the addition of complement to sera obtained in the chronic phase of HIV-1 infection potentiates the antiviral effect, but less so than in the acute phase of the infection (14, 16). The current study confirmed a modest potentiation of HIV-1 plasma antiviral activity during chronic infection and showed that median ADCMI and neutralizing titers differed 4-fold. In comparison, a 32-fold elevation of the antiviral effect was noted in the HIV-2 intratype reactions when complement was added. It should also be noted that the effect of complement on the intratype antiviral activities of HIV-1 and HIV-2 plasma samples was independent of the HIV-1 and HIV-2 phenotypes, since we chose to use HIV-1 and HIV-2 isolates with similar neutralization sensitivities. One explanation for the difference between HIV-1 and HIV-2 could be the number and accessibility of Env spikes present on the virion. HIV-2 spikes have been reported to be more prominent and stable after budding (42, 43), whereas the number of spikes on the HIV-1 particle drops immediately after budding and during maturation (44–47). Given the nonfunctionality of some of these spikes (48), the type of C1q-IgG binding required for the optimal complement effect can be predicted to be fairly low in the case of HIV-1. This is since a stable C1q-antibody interaction only occurs when the C1q globular heads bind to at least two IgG Fc sites that are accessible when two intra- or interspike-binding IgG molecules are in close proximity (49, 50). Another difference between the viruses is the Env configuration, where the Env of HIV-2 is more accessible and has a thinner glycan sheath in certain regions than the Env of HIV-1 (10). The coreceptor binding site of the HIV-2 spike also tends to have a more open configuration irrespective of CD4 binding than does HIV-1 (12); in addition, some HIV-2 isolates display CD4 independence, whereas this is rarely seen with HIV-1 (12, 51). Owing to these differences, it is conceivable that specific antibodies gain access more easily to the epitope on the HIV-2 spike and bind with higher affinity due to the spike arrangements and Env configuration that allow bivalent binding and better complement activation.

The difference between HIV-1 and HIV-2 in regard to complement effects may be related to affinity maturation antibodies that, in turn, are associated with differences in the levels and durations of antigenic stimulation. It is known that the magnitudes of neutralizing antibody responses in HIV-1 and HIV-2 infections are associated with elevated antigenic stimulation, i.e., a more potent plasma NAc correlates with a higher viral load in both systems (7, 8, 52). However, the plasma viral load set point is at least 1 log lower in HIV-2 than in HIV-1 infection (7–9), indicating that that level of antigenic stimulation is not the sole determinant of ADCMI. Another factor adding to this complexity is the continuous emergence of neutralization-resistant variant viruses in the HIV-1-infected host that do not allow for the optimal affinity maturation of HIV-1 antibodies.

The susceptibility of a virus to the complement system has been associated with the level of complement regulatory proteins (CRP), such as CD46, CD55, and CD59, expressed on the PBMC (53, 54) and incorporated into virions (55). For the current study, we used high-CRP-expressing PBMC to propagate our HIV-1 and HIV-2 isolates. In agreement with others, we observed that our PBMC-derived isolates were highly resistant to complement effects in the absence of plasma (14, 53, 54). Data from Sullivan et al. support the role of CRP, such that while primary isolate virions containing CD46, CD55, and CD59 were resistant to complement-mediated lysis (CML), uncultured HIV-1, purified directly from patient plasma and expressing only CD59, was highly susceptible to CML (56). In order to understand the differences between HIV-1 and HIV-2 systems in response to complement, further studies on the antiviral activity of plasma in the presence or absence of complement should be evaluated against HIV-1 and HIV-2 isolates cultivated in exclusively CD59-expressing cells. The role that complement plays in the course of HIV infection appears to be a double-edged sword. While several reports have proposed that complement can boost the effects of antibodies (14, 16, 56), other groups have demonstrated that in the presence of complement receptors (CRs) on the target cell, a pronounced complement-mediated antibody-dependent enhancement (CADE) effect of virus infection may occur (23, 57, 58). In this respect, since the U87.CD4.CCR5 cell line, which displays limited CR expression (flow cytometry analysis; data not shown), was used as the target cell, it could be argued that the potential CADE effect may have been occluded. However, a recent study reported that upon the appearance of NAb responses, CADE subsides sharply (23). In this context, it should be noted that out of all the plasma-virus combinations tested in the current study, it was only in six cases that the antiviral effect of ADCMI was less than that of the neutralizing activity (Table 2). Furthermore, following infection and during the progression of HIV-1, declines in the expressions of CR1, CR2, and CR3 on immune cells have been noted (59, 60), which may further diminish the CADE effect.

There are several methods for measuring the magnitude of antiviral NAc in the literature: based on the crude data of the highest neutralizing titer, log-based magnitude (magnitude_log) (37), and the area under the neutralization curve (AUC)-based magnitude (magnitude_AUC) (38). By evaluating plasma antiviral activities using the AUC-based readout, we found here that the type of HIV infection was independently associated with the effect of complement in the multivariate analysis, while this was not the case for the log-based magnitude method. These findings are in line with the results of Hioe et al. (38) and suggest that the magnitude_AUC assessment of antiviral activity is more comprehensive than the titer-based magnitude_log assessment. This may be explained by the inclusion of the slope of the neutralization curve, not only the highest neutralizing titer.

Taken together, our results suggest that the intratype antiviral activity of HIV-2 plasma is strongly potentiated by complement, which may be one parameter contributing to the relatively benign outcome of an HIV-2 infection.