Neutralization Profiles of Primary Human Immunodeficiency Virus Type 1 Isolates in the Context of Coreceptor Usage

Virus isolates.

A total of 21 primary isolates which had been passaged exclusively in PBMCs were used. These included isolates SF33 and SF2, obtained from J. Levy, University of California at San Francisco, San Francisco, Calif.; isolates CA1, CA5, CA13, CA20, VI191, VI525, VI313, and MAI, obtained from G. van der Groen, Institute of Tropical Medicine, Antwerp, Belgium; isolates 92HT593, 92HT594, 91US056, 92RW021, BK131, SM993, 92TH080, JR-FL, and 89.6, supplied by the National Institutes of Health (NIH) AIDS Research and Reference Reagent Program; and BZ167, supplied originally by J. Mascola, Walter Reed Army Institute of Research, Rockville, Md. MNp, a primary isolate of the MN strain, which had never been passaged in cell lines, was obtained from J. Sullivan, University of Massachusetts Medical School, Worcester, Mass. The clade designation and MT-2-defined phenotype of each primary isolate are shown in Table 1. All virus stocks were prepared by infecting PHA-activated human PBMCs (54). Briefly, frozen PBMCs from HIV-1-negative blood donors were thawed, stimulated with PHA (3 μg/ml; Difco Detroit, Mich.) for 3 days, centrifuged, and infected with 1 ml of virus-infected culture supernatant. After a 1-h exposure of the cells to the virus, the volume of the cell suspension was adjusted to a concentration of 2 × 10⁶ cells/ml and the culture was maintained in RPMI 1640 medium with 10% fetal bovine serum and interleukin-2 (IL-2) (20 U/ml; Boehringer Mannheim Biochemicals, Indianapolis, Ind.) at the same cell concentration. The concentration of p24 in the infected culture supernatant was checked every 3 to 4 days by a noncommercial enzyme-linked immunosorbent assay (26). The infected culture supernatant was collected at 1 to 2 weeks postinfection when the concentration of p24 was at least 100 ng/ml.

Cells.

The GHOST cell lines used herein were derived from HOS cells (24). Briefly, HOS cells were transduced with the human CD4 gene encoded by the murine leukemia virus retroviral vector, pMV7neo, to generate a CD4-positive HOS.T4 clone. HOS.T4 cells were subsequently stably cotransfected with a reporter construct consisting of the HIV-2 long terminal repeat directing the expression of humanized green fluorescent protein (GFP) and a selection construct composed of the human cytomegalovirus immediate-early (IE) promoter driving the expression of hygromycin phosphotransferase. Cells stably transfected with the GFP reporter construct were checked for sensitivity to HIV-1 Tat-mediated gene activation. One clone; clone 34, which expressed GFP strongly after Tat transactivation, was designated the parental cell line and chosen for further development. GHOST cl.34 parental cells were transduced with one of the chemokine receptors (CCR1, CCR2, CCR3, CCR5, CXCR4, Bonzo/STRL33, or BOB/gpr15) encoded on the murine leukemia virus vector, pBABEpuro. Cells expressing the CCR5 coreceptor are referred to below as GHOST-R5, those expressing the CXCR4 coreceptor are referred to as GHOST-X4, etc. About 40% of the cells from each of the GHOST cell lines were positive for CD4; 84% of the GHOST-X4 cells were positive for CXCR4; 66% of the GHOST-R5 cells were positive for CCR5. The parent GHOST cells were 3.8% positive for CXCR4 and 0.8% positive for CCR5.

GHOST cells bearing chemokine receptors were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 1% glutamine, 2% penicillin plus streptomycin, Geneticin (200 μg/ml), hygromycin (25 μg/ml), and puromycin (1 μg/ml). The cultures were maintained at 37°C in a 5% CO₂ humidified incubator. Cell monolayers, when confluent, were resuspended by using 0.25% trypsin. The cells were maintained for up to 15 passages and then replaced with fresh cells from the cryopreserved stock which had been frozen at the second or third passage.

MAbs and polyclonal Abs.

Five polyclonal serum samples and two MAbs were used in the neutralization tests. Sera F and N are from asymptomatic HIV-positive volunteers from the Veterans Affairs Medical Center (New York, N.Y.) and the University of California Los Angeles Medical Center (Los Angeles, Calif.), respectively. Serum N was provided by S. Miles (University of California Los Angeles Medical Center, Los Angeles, Calif.). Serum FDA-2 is a serum pool derived from four bleeds from an HIV-positive patient obtained from the NIH AIDS Research and Reference Reagent Program. HIVIG-Ug is the immunoglobulin G (IgG) fraction from pooled HIV-positive serum obtained in Uganda (supplied by B. Jackson, Johns Hopkins University, Baltimore, Md.); it was used at a starting concentration of 0.5 mg/ml. Pool 2 is a serum pool from 33 randomly selected HIV-positive subjects at the Veterans Affairs Medical Center, New York. An HIV-negative human serum specimen was used in the experiments as a negative control. All sera were heat inactivated at 56°C for 30 min prior to use.

Two human MAbs were used in this study, 447-52D and IgG1b12. 447-52D, produced in this laboratory, was previously described as a broadly cross-reactive V3-specific MAb (18, 19). IgG1b12, a recombinant antibody which recognizes an epitope overlapping the CD4 binding domain of the HIV-1 envelope (8), was obtained from the NIH AIDS Research and Reference Reagent Program. Stocks of both MAbs were stored in frozen aliquots as purified IgG (1 mg/ml) and used at concentrations ranging from 0.1 to 25 μg/ml.

Infectivity assay.

GHOST cells were seeded in 24-well plates (Falcon; Fisher Scientific, Springfield, N.J.) at 6 × 10⁴ cells/well/0.5 ml. On the following day, the medium was removed and the monolayers, about 70% confluent, were infected with undiluted virus stocks (100 μl/well). To each well was added DEAE-dextran to a final concentration of 8 μg/ml. The virus was allowed to adsorb overnight, after which the virus-containing medium was removed and the cell monolayers were washed once with phosphate-buffered saline. Subsequently, 1 ml of complete medium, as described above, was added per well. The day on which the virus was added was considered day 0. Cells were harvested on day 4 or 5 postinfection (p.i.). On the day of harvest, the cell monolayers were once again washed with phosphate-buffered saline, resuspended in 300 μl of 1 mM EDTA in PBS, and fixed in formaldehyde at a final concentration of 2%. The cells were then analyzed with a FACScan flow cytometer (Becton Dickenson, San Jose, Calif.). The live cells were gated on the basis of forward and side scatter. Because of the autofluorescence of uninfected GHOST cells due to basal expression of the indicator cassette, the gain on the FL 1 channel was set to bring the mean channel fluorescence of uninfected cells to <10². The number of infected cells was determined by using a scattergram of fluorescence versus forward scatter after setting the gates with uninfected cells. A total of 15,000 to 20,000 events was scored. The total number of cells was about 10⁶/well on the day of harvest. Hence the number of cells scored was approximately 1/50 of the total cells in culture. As noted above, 3.8% of the GHOST parental cells are CXCR4 positive, and hence all the coreceptor derivatives express a low but detectable level of CXCR4. To account for this, GHOST-CCR1 cells were used to establish background infectability. The mean number of fluorescent GHOST-CCR1 cells + 2 standard deviations after infection with the viruses used in this study was considered the cutoff. On this basis, >99 fluorescent cells/15,000 cells had to be present for a virus to be considered positive for infectivity in each of the GHOST cell lines tested.

GHOST cell neutralization assay.

The method for the GHOST cell neutralization assay was essentially as described above for infectivity studies except that a fixed dilution of each virus stock was used, based on predetermined infectivity titers; the virus dilution used was chosen to give about 200 to 800 fluorescent cells per 15,000 events in the absence of anti-HIV antibodies. Polyclonal Ab or MAb preparations were diluted serially in fivefold dilutions. Equal volumes of diluted Ab preparations and virus were mixed and incubated for 1 h at 37°C; 100 μl of the virus-Ab mixture was added to duplicate wells and incubated overnight in the presence of DEAE-dextran (8 μg/ml). Subsequent washes, incubations, harvest, and readout procedures were performed as described above. Neutralization assays were typically terminated 3 to 4 days p.i. Care was taken to terminate the assay before cell lysis occurred. Formation of moderate-sized syncytia did not seem to affect the flow analysis, since the forward and side scatter gates included almost all of the viable cells.

Determination of coreceptor preferences by primary isolates from different clades.

A panel of 21 primary isolates belonging to different clades and demonstrating different SI/NSI phenotypes when grown in MT-2 cells was chosen for infectivity studies (Table 1). GHOST cells expressing CCR1, CCR2, CCR3, CCR5, CXCR4, BOB, or Bonzo were infected with virus stocks in the presence of DEAE-dextran since this compound was required for optimal infectivity by some viruses. For example, while isolate SF33 was not significantly affected by the presence of DEAE-dextran, isolate CA5 showed infectivity only in its presence (data not shown). Thus, for all infectivity and neutralization assays, DEAE-dextran was included during virus adsorption.

Infection by primary isolates of GHOST cells expressing the various chemokine receptors was readily detected by flow cytometry, as depicted in Fig. 1. Of nine SI viruses tested, only three, BZ167, MNp and 92HT594, were X4, using CXCR4 but not CCR5. However, BZ167 could also use CCR3, while 92HT594 also used BOB (Fig. 1A). The other six SI viruses were dualtropic, using both CXCR4 and CCR5, or polytropic, using both major coreceptors and CCR2, CCR3, and/or Bonzo; these three coreceptors generally mediated infection with lower efficiencies than did CXCR4 and CCR5 (Fig. 1C). The 12 NSI viruses tested were R5, defined as using CCR5 but not CXCR4. Viruses CA5 and VI313 also used CCR3 and Bonzo, respectively (Fig. 1B). Again, these two coreceptors mediated infection with lower efficiency.

All the isolates, irrespective of coreceptor usage or SI/NSI phenotype, caused syncytium formation in the GHOST cells. For the majority of R5 viruses, the cytopathic effect was apparent early compared to that seen with viruses which used other coreceptors. Figures 2A and B show syncytium formation in GHOST cells by the R5 virus VI313 and by the X4/R5 virus 89.6, respectively, on day 3 p.i. Uninfected cells are shown in Fig. 2C.

Infection kinetics.

The kinetics of infection for a polytropic virus (SF33) and a dualtropic virus (92HT593), as assessed fluorocytometrically, are shown in Fig. 3A for GHOST-X4 and GHOST-R5 cells. SF33 showed a lower efficiency of infection in GHOST-R5 cells compared to that in GHOST-X4 cells, although the rates of infection in the two cell lines were parallel. Isolate 92HT593 showed similar kinetics in the two cell lines.

Ab-mediated neutralization of primary isolate infection of GHOST cells.

The neutralization assay was first performed with the polytropic clade B SF33 isolate in GHOST-X4 and GHOST-R5 cells. The HIV-positive serum pool 2 and an HIV-negative serum specimen were used at a final dilution of 1:40. The cells were harvested from days 2 to 6 p.i. The number of fluorescing GHOST-X4 or GHOST-R5 cells in wells which were infected with virus treated with normal serum was about 700 on day 6 (Fig. 3B and C). This was reduced to background levels throughout the observation period when virus was pretreated with the anti-HIV pool 2 (Fig. 3B and C). The day of harvest did not affect the degree of neutralization observed.

A checkerboard neutralization assay was also carried out with SF33 and MAb 447-52D or the HIV-positive serum F. Varying the virus input by 1 order of magnitude, resulting in 170 to 1,165 infected cells/15,000 cells in the absence of antibody, did not influence the neutralization titers observed in the assay (data not shown).

Subsequently, 10 primary isolates from clade B with different coreceptor preferences were tested in the GHOST cell neutralization assay against individual HIV-positive sera, the HIV-positive serum pool, HIVIG-Ug, and two human MAbs. Each Ab-virus combination was tested in duplicate, and each assay was repeated at least twice. There was very little variability in the duplicates, with the variation ranging from 0.3 to 10%. Neutralization curves are shown in Fig. 4 for the polytropic primary isolate SF2 with three individual immune sera in both GHOST-X4 and GHOST-R5 cells. Similar neutralization curves were used to determine the 50% neutralization titers, which are shown in Fig. 5.

Of the X4 viruses, BZ167 was neutralized best, with 50% serum neutralizing titers of 1:500 to 1:1,000 and 50% neutralizing MAb concentrations of 1 to 5 μg/ml (Fig. 5A). The primary isolate MN was also neutralized by all the Ab preparations but with lower titers. 92HT594 could not be neutralized by any of the Ab preparations tested; the lowest dilution of antiserum tested was 1:40 and the highest concentration of MAb tested was 25 μg/ml for 447-52D and 13.5 μg/ml for IgG1b12 (Fig. 5A).

Of the R5 viruses, JR-FL could be neutralized by sera N and FDA-2 at titers of 1:50 and by MAb IgG1b12 at 2 μg/ml. Only 44% neutralization was achieved with isolate 91USO56 and FDA-2 at a 1:40 dilution. Isolate CA5 could not be neutralized by any of the Ab preparations tested (Fig. 5B).

With dual- and polytropic viruses, the patterns of neutralization were essentially the same whether the viruses were assayed on GHOST-X4 or GHOST-R5 cells (Fig. 5C and D, respectively). Only SF33 showed slightly lower titers when tested on GHOST-R5 (compared to GHOST-X4) with pool 2, the three individual HIV-positive sera, and the anti-V3 MAb, 447-52D. SF33 (in GHOST-X4) and SF2 could be neutralized best, at serum titers and MAb concentrations of ∼1:1,000 and 1 to 2 μg/ml, respectively. Isolate 89.6 showed moderate neutralization, and 92HT593 showed poor neutralization.

Thus, in each of the different virus categories defined by coreceptor usage, there was a spectrum of differential susceptibilities to Ab-mediated neutralization. R5 viruses appeared to be somewhat more resistant than X4, dual-, or polytropic viruses, but this conclusion may be affected by the particular viruses used from the panel available for testing.

As shown in Fig. 5, neither 447-52D nor IgG1b12 neutralized all the isolates. MAb 447-52D neutralized 4 of 10 strains, and IgG1b12 neutralized 7 of 10 strains. When 50% neutralizing concentrations were achieved for MAb 447-52D, they ranged from <1 to 20 μg/ml; the range for IgG1b12 ranged from <0.5 to 10 μg/ml. The core epitope of MAb IgG1b12 has not been identified, and therefore it was not possible to correlate its activity with any particular virus sequence. The core epitope of MAb 447-52D was previously identified as GPXR by Pepscan analysis with overlapping peptides from the V3 sequence of the MN strain (18). Binding to the V3_MN peptide, however, accounts for only about 10% of the binding energy of the MAb (18); while neutralizing activity was associated with this V3 sequence (12), it was also correlated with the dissociation rate constant (48), which is affected by both sequence and conformation. This information suggested that the sequence at the crown of the V3 loop might not be sufficient to confer Ab binding that would lead to neutralization and that under certain conditions, conformation might overcome the need for absolute fidelity in the core epitope. To examine this, partial V3 sequences of the 10 isolates used in the neutralization experiments were analyzed; they are shown in Table 2. Of the 10 viruses, 8 contain the GPXR core epitope, but only 3 of these (BZ167, MNp, and SF2) were neutralized by MAb 447-52D. This suggests that the presence of the core epitope is not sufficient to confer neutralization sensitivity. Conversely, one of the two isolates that does not contain the core epitope (SF33) was neutralized by MAb 447-52D, suggesting that conformation-dependent structures can confer sufficient binding energy to effect neutralization.