Previous Article | Next Article ![]()
Journal of Virology, December 2008, p. 12069-12081, Vol. 82, No. 24
0022-538X/08/$08.00+0 doi:10.1128/JVI.01379-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
,
Karolin Hijazi,3
Willie Koh,1
Vanessa Tack,2
Agnieszka Szynol,4
Charles Kelly,3
Áine McKnight,1,
Theo Verrips,4
Hans de Haard,2,4 and
Robin A. Weiss1*
MRC/UCL Centre for Medical Molecular Virology, Division of Infection and Immunity, University College London, 46 Cleveland Street, London W1T 4JF, United Kingdom,1 Ablynx NV, Technologiepark 4, B-9052 Ghent, Belgium,2 King's College London, Dental Institute, Oral Immunology, Tower Wing, Guy's Hospital, London SE1 9RT, United Kingdom,3 Department of Molecular and Cellular Biology, Cellular Architecture & Dynamics, University of Utrecht, H.R. Kruytgebouw, Padualaan 8, 3584 CH Utrecht, The Netherlands4
Received 2 July 2008/ Accepted 26 September 2008
|
|
|---|
|
|
|---|
HIV-1 entry into target cells is mediated by the viral envelope spike, which consists of homotrimers of the surface glycoprotein, gp120, noncovalently bound to the transmembrane glycoprotein, gp41 (89, 91, 95). In addition to the functional spikes, there is also evidence for the presence of nonfunctional derivatives, such as gp41 stumps and gp120/gp41 monomers, on the viral surface (54, 66). Most variants of HIV-1 enter cells through attachment of the envelope spike to the main cellular receptor CD4 (15, 39), which triggers a conformational change allowing interaction with a cellular coreceptor, typically CCR5 or CXCR4 (53), eventually leading to fusion of virus and cell membranes. Potent entry inhibitors can target various stages of this process (51). Neutralizing monoclonal antibodies (MAbs) can act as HIV-1 entry inhibitors by targeting epitopes on the functional spike (61). HIV-1 has, however, evolved a number of ways to evade the humoral immune response, including variable regions, carbohydrate shields, extreme diversity, and conformational and entropic masking, and the neutralizing antibody response in HIV-1 infection is therefore in general rather weak and narrow (63, 89).
Many MAbs to HIV-1 envelope have been isolated from animals such as mice postimmunization and from humans following HIV-1 infection. Of these, only a handful have been found to be broadly neutralizing across HIV-1 subtypes (8), and have been the result of HIV-1 infection rather than immunization. Two of these are directed against gp120: MAb b12, which binds to an epitope that overlaps a subset of the CD4-binding site (CD4bs) of gp120 (3, 10, 11, 68, 94), and MAb 2G12, which recognizes a carbohydrate motif (9, 70, 72, 80). Two broadly neutralizing MAbs, 4E10 and 2F5, recognize gp41 (9, 56, 77, 96). MAbs X5 (55), which recognizes an epitope on gp120 that is better-exposed after CD4 binding, and m14 (92), which competes with CD4 for binding to gp120, also display some neutralizing activity across HIV-1 subtypes, as do a minority of MAbs to the V3 region of gp120 (30, 31).
All of the broadly neutralizing MAbs reported to date are from individuals infected with HIV-1 of subtype B, which is dominant in Europe and North America. Phylogenetically, HIV-1 is classified into groups M, N, and O, with group M accounting for over 99% of infections and being the most variable, with extraordinary diversity in the envelope sequence between isolates. Group M is divided into subtypes A to D, F to H, J, and K, plus a number of circulating recombinant forms (CRFs), with subtype C currently infecting more people than any other subtype (34, 47). MAbs to HIV-1 envelope, whether obtained from natural HIV-1 infection or from immunization, can help in defining neutralizing and nonneutralizing epitopes on HIV-1 envelope proteins of various subtypes. We employed the nonconventional immune system of camelids to generate novel antienvelope antibodies.
In addition to conventional antibodies, members of the Camelidae family (camels, dromedaries, and llamas) produce antibodies without light chains, so-called heavy chain antibodies (33). The antigen-binding properties of these heavy chain antibodies are provided by one single fragment, the variable region of the heavy chain, which has been termed the VHH, or nanobody. Despite their small size of approximately 14 kDa, these VHH have characteristics in terms of affinity and specificity similar to those of conventional antibodies (81). They display complementarity-determining regions (CDRs) of which on average the CDR2 and CDR3 are longer than the corresponding CDRs of conventional antibodies (85). Furthermore, they have been shown to have a preference for cleft recognition and for binding into active sites (16, 41). The VHH domain can be easily cloned and expressed to high levels in bacteria and yeast (26, 27). This notion, together with advantageous characteristics in terms of stability and solubility (20, 64, 81), has led to successful development of camelid VHH in a number of applications against a range of biological targets (2, 13, 14, 19, 21, 57, 58, 69, 83, 84), including neutralization of rotavirus (28, 60). We hypothesized that the small size of VHH in combination with their protruding CDR3 loops and their preference for cleft recognition may allow them to recognize conserved motifs on gp120 that are occluded from conventional antibodies.
Here, we describe the identification of a set of llama VHH that can inhibit binding of soluble CD4 (sCD4) to HIV-1 envelope proteins and neutralize HIV-1 primary isolates of subtypes B and C. These VHH may be useful as tools for HIV-1 vaccine design and could possibly be developed as candidate HIV-1 microbicides.
|
|
|---|
Recombinant HIV-1 envelope proteins. Recombinant gp120 from HIV-1 CN54 (subtype B'/C; CRF07_BC) was kindly provided by I. Jones, Reading University, Reading, United Kingdom, through the European Microbicides Project. Recombinant gp140 from HIV-1 92UG037 (subtype A) was kindly provided by S. Jeffs, Imperial College London, London, United Kingdom. Recombinant gp120 from HIV-1 IIIB (EVA607) was obtained from the CFAR, NIBSC. Recombinant gp120 derived from HIV-1 92UG037 and 92BR025 (subtype C) was expressed and purified as detailed in File S1 in the supplemental material.
Viruses. The sources of HIV-1 peripheral blood mononuclear cell (PBMC) isolates, HIV-1 replication-competent molecular clones, and env clones used to prepare pseudotyped viruses and to express recombinant gp120 are detailed in File S1 of the supplemental material.
Cells. TZM-bl cells (18, 65, 86) were obtained through the NIH AIDS Research and Reference Reagent Program from J. C. Kappes, X. Wu, and Tranzyme, Inc., and cultured in Dulbecco's modified Eagle medium (Invitrogen, Paisley, United Kingdom) containing 10% (vol/vol) fetal calf serum (FCS). NP2 glioma cells (76), expressing the HIV-1 cellular receptor CD4 and either of the coreceptors CXCR4 (NP2/CD4/CXCR4) or CCR5 (NP2/CD4/CCR5), were cultured in Dulbecco's modified Eagle's medium containing 5% (vol/vol) FCS. Phytohemagglutinin-stimulated PBMCs were obtained from blood donors and cultured in RPMI 1640 supplemented with 10% (vol/vol) FCS and 20 U of interleukin-2 (Roche, Lewes, United Kingdom) per ml.
Immunization of Llama glama, evaluation of antibody response, and construction of VHH phage libraries. Two llamas (numbered L40 and L44) were immunized with recombinant gp120 derived from HIV-1 CN54. Immunizations and VHH library construction were carried out as described previously (17). In brief, the llamas received six intramuscular injections at weekly intervals. Each injection consisted of a freshly prepared 4.5-ml water-in-oil emulsion prepared by vigorously mixing 2 volume units of antigen (50 or 100 µg) with 2.5 volume units of the adjuvant Stimune (CEDI Diagnostics, Lelystad, The Netherlands). The antienvelope immune response in sera was verified in an enzyme-linked immunosorbent assay (ELISA) against immobilized recombinant gp120. The neutralization activities of serum or plasma samples from day zero (preimmunization) and days 28, 39, and 43 (postimmunization) were evaluated in TZM-bl cells. Total RNA was isolated from peripheral blood lymphocytes and lymph node biopsies collected postimmunization (on days 39 and 43), and cDNA was prepared. The VHH repertoire was amplified and cloned into the pAX050 phagemid vector. To obtain recombinant bacteriophages expressing the VHH as fusion proteins with the bacteriophage gene III product, transformed TG1 Escherichia coli cells were grown to logarithmic phase and then infected with helper phage M13KO7. The phage particles were precipitated with polyethylene glycol to remove free VHH.
Selection of anti-CD4bs VHH through panning on gp120 followed by competitive elution with sCD4 and subsequent isolation and screening of individual VHH. Phage expressing the cloned VHH repertoire was incubated with immobilized gp120 and eluted with sCD4 to enrich for VHH targeting the CD4bs in two subsequent rounds of selection. Individual VHH were isolated and screened for binding to recombinant gp120 and ability to neutralize HIV-1. The selection and screening procedures are detailed in File S1 of the supplemental material, as are the procedures for expression and purification of selected VHH.
HIV-1 neutralization assays. VHH neutralization activities were evaluated in three different neutralization assays. In the initial screening of VHH, their neutralization ability was assessed in NP2/CD4/CCR5 (or CXCR4) glioma cells against concentration-matched irrelevant VHH, followed by detection of infection by HIV-1 p24 immunostaining, as described previously (1). In order to enable high-throughput neutralization screening and characterization of VHH, neutralization was subsequently measured using 200 50% tissue culture infective doses of virus in the TZM-bl cell-based assay developed by Derdeyn et al. (18), Wei et al. (86), and Li et al. (43), with Bright-Glo luciferase reagent (Promega, Southampton, United Kingdom). The neutralization activity of each VHH and MAb b12 was assayed in duplicate and on a minimum of two separate occasions, apart from the CRF07_BC pseudovirus panel, against which the VHH and MAb b12 were assayed only once, in duplicate. No virus inactivation was observed with a negative control VHH or with a pseudovirus bearing a rabies virus G-protein envelope (87). VHH and MAb b12 50% inhibitory concentration (IC50) and IC90 titers were calculated using the XLFit4 software (ID Business Solutions, Guildford, United Kingdom).
In addition, VHH neutralization of HIV-1 was assayed in phytohemagglutinin-stimulated PBMCs. Serial dilutions of VHH (and MAb b12 in parallel) were incubated with virus for 1 h at 37°C and added to 105 cells in 96-well round-well plates. After 24 h at 37°C the cells were washed twice in growth medium. Cell culture supernatant was collected on days 0, 1, 3, 5, and 8 after infection and assayed for HIV-1 p24 antigen using an antigen capture assay kit (SAIC-Frederick, Frederick, MD) according to the manufacturer's instructions. Neutralization was measured as the reduction in p24 content in test wells compared to that in virus control wells. All VHH, including a negative control VHH (and MAb b12), were assayed in triplicate and in PBMCs from two separate donors.
VHH binding to recombinant envelope proteins in an ELISA, competition ELISA, and surface plasmon resonance assay. VHH binding to recombinant envelope proteins and ability to compete with sCD4, anti-gp120 MAbs, and each other was assayed in an ELISA, as detailed in File S1 of the supplemental material. VHH binding to gp120 and competition with sCD4 was assayed in surface plasmon resonance experiments carried out using BIAcore (Uppsala, Sweden), as detailed in File S1 of the supplemental material.
|
|
|---|
In order to select VHH targeting the CD4bs, we employed a directed selection strategy, as outlined in Fig. 1A. VHH were selected by panning the phage display libraries on immobilized gp120, either directly coated onto the plate or captured by the D7324 antibody, followed by a competitive elution using excess sCD4, at two different concentrations and with varied lengths of elution. Selections where a larger number of clones were eluted by sCD4, compared to elution with bovine serum albumin (BSA), were taken forward to a second round of CD4bs-targeted selection (Fig. 1B). In general, no correlation was observed between the number of clones eluted by sCD4 and the concentration of sCD4 or length of incubation with sCD4. After two rounds of panning, selections where enrichment of clones eluted by sCD4 could be observed were chosen and individual VHH were expressed and screened for binding to recombinant gp120 in an ELISA and/or in vitro neutralization assay of HIV-1.
![]() View larger version (46K): [in a new window] |
FIG. 1. (A) Schematic overview of the strategy for isolation of llama VHH targeting the CD4bs of gp120. (B) Titration of eluted phage onto E. coli TG1 cells. Phage bound to gp120 was eluted using sCD4. As a control, elution with BSA was performed in parallel, as was a general elution by low-pH shock using glycine. Shown is a representative titration of eluted phage, where more clones were eluted by sCD4 than by BSA and from which individual VHH were isolated and expressed.
|
Only 1 out of 96 clones selected using HIV-1 CN54 gp120 was found to bind to CN54 gp120 in the ELISA (data not shown). This VHH was derived from llama L40. It reduced infection of HIV-1 CN54 by approximately 50% at 100 µg/ml and displayed 0 or less than 50% neutralization with other strains (data not shown). Due to the limited neutralizing activity, this VHH was not further characterized.
Panning on IIIB gp120 was more successful. Out of 48 clones picked for analysis, 30 were confirmed to bind to IIIB gp120 in the ELISA, and 24 out of the 30 were observed to neutralize HIV-1 IIIB (>90% reduction of infection using VHH in E. coli periplasmic extracts). These VHH were all derived from llama L44. Fingerprint analysis and sequencing of the 24 clones revealed some identical VHH, leaving 8 different clones that could be separated into two distinct groups. The first group consisted of three clones with 8 to 14 amino acid differences in the framework 1, 2, and 3 regions as well as in CDR1, CDR2, and CDR3. They displayed a CDR3 of 18 amino acids. Based on superior neutralizing activities against the limited panel of HIV-1 isolates (data not shown), two of these VHH, designated A12 and D7, were chosen for further characterization. The second group of VHH consisted of five clones with one to three amino acid differences only, in framework regions 1 and 3. These displayed a shorter CDR3 of 10 amino acids. All five clones were shown to have similar neutralization profiles (data not shown), and one representative clone, designated C8, was chosen for further characterization. Thus, in total, three VHH (A12, D7, and C8) were selected for further characterization (see File S2 and Table S1 in the supplemental material). The amino acid sequences of these VHH are shown in File S2 and Fig. S1 of the supplemental material.
In order to allow for selection of VHH recognizing motifs conserved among HIV-1 subtypes, the phage libraries were also panned against recombinant gp120 derived from subtype A (92UG037) and subtype C (92BR025) gp120, in addition to IIIB gp120, alternating the antigen in different combinations in two subsequent rounds of panning. Enrichment of clones eluted by sCD4 was only observed in selections using the phage library from llama L44. Approximately 700 individual VHH were isolated, expressed, and screened for neutralization of HIV-1 as well as for binding to gp120 in the ELISA. VHH clones selected using 92BR025 gp120 were screened for the ability to neutralize HIV-1 92BR025, whereas VHH selected using 92UG037 gp120 were screened against HIV-1 92UG037. In addition, all VHH were screened for the ability to neutralize HIV-1 IIIB. Out of the 700 clones isolated, 43 were observed to bind to gp120 in the ELISA and to neutralize HIV-1. The results from the neutralization and binding screening showed complete correlation, with the exception that the clones selected using subtype A 92UG037 gp120 were not able to neutralize 92UG037 virus, despite being able to neutralize HIV-1 IIIB. The relatively low frequency of VHH able to bind to gp120 and to neutralize HIV-1 (43 out of 700, corresponding to 6%) probably reflects the nature of the competitive elution method, as a significant number of clones were nonspecifically eluted by BSA, as can be seen in Fig. 1B. Sequence analysis of the 43 neutralizing clones revealed that they were all either identical to VHH A12 (21 clones), which had already been selected from the same library through panning on HIV-1 IIIB gp120 alone, as described above, or identical to A12 with 1 amino acid difference in the framework 1 region (22 clones). Further characterization of the latter VHH revealed a neutralization profile similar to that of A12 (data not shown).
Characterization of VHH neutralization properties in TZM-bl cells. For high-throughput characterization of the neutralization activities of VHH A12, D7, and C8, we employed the TZM-bl cell-based neutralization assay (18, 43, 86). The neutralization potencies were assayed against HIV-1 of subtypes A, B, C, D, CRF07_BC, and CRF02_AG using either PBMC-propagated primary isolates, T-cell-line-adapted viruses, or recombinant replication-competent chimeric viruses as well as envelope-pseudotyped viruses expressing envelopes cloned straight from plasma of infected individuals or from PBMC-propagated primary isolates. The lowest VHH concentrations required to achieve 50% and 90% reductions of infectivity (IC50 and IC90) compared to a virus control were determined. A negative control VHH was tested in parallel, as well as a pseudovirus bearing the rabies virus G-protein (87). No nonspecific virus inactivation was observed.
The results of the neutralization characterization are summarized in Fig. 2 and 3. In summary, VHH A12 generally showed the most broad and potent neutralizing activity of the VHH, being able to neutralize HIV-1 primary isolates of subtypes B, C, and CRF07_BC, but not the other subtypes tested. Overall, it neutralized 27 out of 65 viruses (42%), with IC50s in the range of <0.003 to 38 µg/ml (Fig. 2 and 3). VHH D7 showed a similar but slightly less cross-reactive neutralization profile compared to that of A12, neutralizing 31% of viruses on an IC50 level. VHH C8 neutralized 35% of viruses (IC50 level). It seemed less potent than A12 and D7 against subtype B viruses but was as reactive as VHH A12 against subtype C viruses, although it neutralized a different pattern of viruses (Fig. 2 and 3).
![]() View larger version (46K): [in a new window] |
FIG. 2. VHH and MAb b12 IC50 titers against HIV-1 in TZM-bl cells. VHH and MAb b12 neutralization activity was assessed in the indicated viruses, as described in the text. Rabies virus CVS-11 is pseudotyped with rabies virus G-protein from strain CVS-11 (87). Abbreviations for virus types: TCLA, T-cell-line-adapted isolate; PBMC, PBMC-propagated primary isolate; MC, molecular clone; PV, envelope pseudotyped virus. The column labeled Tier indicates whether the virus is classified as suitable for tier 1, 2, or 3 assessment of neutralizing antibodies (46). , IC50 >50 µg/ml; nd, not determined; n/a, not applicable. To aid comprehension, the titers have been shaded, with darker colors indicating more potent neutralization.
|
![]() View larger version (16K): [in a new window] |
FIG. 3. Percentage of HIV-1 isolates neutralized by the VHH and by MAb b12, according to HIV-1 subtype. Virus neutralization was assayed in TZM-bl cells as described in the text. Shown is the percentage of viruses neutralized with an IC50 and IC90 of 50 µg/ml.
|
50%) the subtype A, A/G, or D viruses included in the study (Fig. 2). It should be noted that the VHH are approximately 10 times smaller than MAb b12 but that MAb b12, on the other hand, has two antigen-recognizing domains per molecule, making it impossible to accurately estimate relative neutralization potencies. To be able to make a direct comparison, the VHH would need to be presented in the context of a complete camelid immunoglobulin. VHH neutralization of HIV-1 in primary cells. The ability of the VHH to inhibit HIV-1 SF162 infection of primary cells was tested in a PBMC-based assay. VHH and MAb b12 IC50, IC75, and IC90 titers against SF162 in PBMCs from two different donors, scored on day 5 postinfection, are shown in Table 1. VHH A12 was able to inhibit infection of SF162 in PBMCs from both donors, with IC75s of 35 and 44 µg/ml, respectively. In contrast, VHH C8 did not inhibit SF162 infection in primary cells, which is concordant with results obtained in the TZM-bl assay. For VHH D7, the results were less clear, as this VHH inhibited infection in PBMCs from only one of the two donors.
|
View this table: [in a new window] |
TABLE 1. VHH and MAb b12 neutralization of HIV-1 SF162 in PBMCsa
|
VHH bind to recombinant IIIB gp120 with high affinities. VHH affinity for gp120 was determined using surface plasmon resonance techniques. The kinetic data are summarized in Table 2. In summary, VHH A12 and D7 showed affinities in the high picomolar range for recombinant IIIB gp120, with equilibrium dissociation constants of 100 and 97 pM, respectively, approaching the affinity ceiling that has been suggested for antibodies generated through in vivo maturation (5, 24, 78). VHH C8 had a faster off-rate, leading to a more-than-eightfold-higher KD.
|
View this table: [in a new window] |
TABLE 2. VHH kinetic constants and affinities for IIIB gp120a
|
VHH bid to recombinant envelope proteins of subtypes A, B, and C in the ELISA. VHH binding to HIV-1 gp120 was evaluated in an ELISA. Cross-subtype reactivity was determined using gp120 or gp140 from HIV-1 of subtypes A, B, and C. Concentration-dependent binding of the selected VHH to gp120 and gp140 is shown in Fig. 4A. VHH A12, D7, and C8 bound equally well to HIV-1 IIIB gp120. In contrast, A12 and C8 showed stronger binding to HIV-1 92UG037 gp140 than did D7. VHH A12 and C8 were also able to bind well to recombinant HIV-1 92BR025 gp120, whereas D7 bound with a weaker signal. In spite of their ability to neutralize HIV-1 strains of subtypes B and C, none of VHH A12, D7, and C8 was found to bind well to the protein against which it had been raised, CN54 gp120, in the ELISA (data not shown). This contradictory observation might to some extent be explained by doubts regarding the structural integrity of the immunogen, as it did not show good binding to sCD4, nor was it able to mediate infection when expressed on the surface of virus particles (data not shown); however, these VHH had been selected through panning on IIIB gp120, as opposed to on CN54 gp120.
![]() View larger version (20K): [in a new window] |
FIG. 4. (A) VHH binding to recombinant envelope proteins derived from HIV-1 92UG037 (subtype A), IIIB (subtype B), and 92BR025 (subtype C) in the ELISA. Recombinant envelope proteins were captured by immobilized antibody D7324. Serial dilutions of VHH A12, D7, C8, and a negative control VHH were then added, and binding was detected as described in the text. (B) Dose-dependent competition of VHH A12, D7, and C8 with sCD4 for binding to recombinant envelope proteins in the ELISA. Threefold serial dilutions of VHH were preincubated with IIIB gp120 or 92UG037 gp140 and subsequently incubated with sCD4 precoated on microtiter plates. Envelope protein binding to sCD4 was detected as described in the text. Chemiluminescence was measured, and background subtracted luminescence readings (in relative light units [RLU]) were plotted against VHH concentration. Data points represent the means and bars show the standard deviations of duplicate reactions.
|
To further confirm that the selected VHH recognize an epitope overlapping the CD4bs, or at least bind in a way that sterically hinders binding of sCD4, VHH D7 and C8 were assessed for their ability to inhibit binding of sCD4 to gp120 in a BIAcore surface plasmon resonance assay. Surface plasmon resonance enables testing whether or not the VHH can inhibit binding of sCD4 to CN54 gp120 (the immunogen), since binding of fluid-phase (only) CN54 gp120 to immobilized sCD4 can be observed in the BIAcore assay but not in the ELISA. Recombinant sCD4 was captured by an anti-CD4 antibody immobilized on the sensor chip. Titrations of VHH were preincubated with gp120 and then injected onto the chip. Both VHH D7 and C8 could completely inhibit binding of sCD4 to CN54 gp120, even at equimolar concentrations of VHH and gp120 (see File S2 and Fig. S3 in the supplemental material). Both VHH also inhibited sCD4 binding to IIIB gp120 (data not shown). This finding confirms the observations made in the ELISAs, i.e., that the VHH compete with sCD4 for binding to gp120, either through binding to an epitope that partly overlaps with the CD4bs or by binding in a way that sterically hinders binding of sCD4; alternatively, VHH binding locks the gp120 in a conformation that hampers interaction with sCD4. Again, this finding may indicate that the VHH inhibit HIV-1 infection by interacting with gp120 prior to its engagement to CD4.
(ii) VHH compete with anti-CD4bs MAbs for binding to recombinant gp120. To further map the epitopes of the selected VHH, they were tested for their ability to compete with anti-CD4bs MAbs b12, 654-D, and GP68. Antibody b12 has been shown to bind to an epitope that overlaps a subset of the CD4bs (94). Human monoclonal antibodies 654-D and GP68 have been shown to compete with sCD4 for binding to recombinant gp120 but can only neutralize some T-cell-line-adapted isolates and not primary isolates (29, 37, 40, 74).
VHH A12, D7, and C8 were all found to compete with MAb b12 for binding to recombinant gp120 (Fig. 5A). VHH A12 inhibited MAb b12 binding to IIIB gp120 at slightly lower concentrations than the other VHH. This observation may indicate that VHH A12 recognizes an epitope that overlaps with the b12 epitope to a greater extent, which could be in line with the neutralization results, where VHH A12 showed a broader neutralizing ability compared to the other VHH (Fig. 2), indicating that it binds to an epitope that is more conserved. VHH A12, D7, and C8 were all found to compete with anti-CD4bs antibodies 654-D and GP68, further confirming that the VHH recognize CD4bs-related epitopes (Fig. 5A).
![]() View larger version (39K): [in a new window] |
FIG. 5. VHH and anti-gp120 MAb cross-competition analysis. (A) Dose-dependent competition of VHH A12, D7, and C8 with anti-CD4bs MAbs b12, 654-D, and GP68 for binding to recombinant IIIB gp120 in the ELISA. Serial dilutions of VHH were preincubated with IIIB gp120. Envelope protein binding to human anti-CD4bs MAbs b12, 654-D, and GP68 was detected as described in the text. (B) VHH competition with anti-gp120 MAbs 2G12 (carbohydrate motif), 17b (CD4i), and 447-52D (V3), as well as MAb b12 (CD4bs), for binding to recombinant gp120 in the ELISA. Wells of microtiter plates were coated with VHH A12, D7, and C8, as indicated above each graph. Serial dilutions of each MAb were preincubated with IIIB gp120 and subsequently incubated with the immobilized VHH. Envelope protein binding to VHH was detected as described in the text. (C) VHH competition with each other and with sCD4 for binding to recombinant gp120 in the ELISA. Wells of microtiter plates were coated with VHH A12, D7, and C8, as indicated above each graph. Serial dilutions of each VHH and sCD4 were preincubated with IIIB gp120 and subsequently incubated with the immobilized VHH. Envelope protein binding to VHH was detected as described in the text. Background-subtracted luminescence readings were plotted against MAb concentration. Data points represent the means and bars show the standard deviations of duplicate reactions.
|
(iii) VHH compete with each other for binding to recombinant gp120. To investigate whether the VHH bind to epitopes that are related, VHH A12, D7, and C8 were tested for their abilities to compete with each other for binding to gp120 (Fig. 5C). Soluble CD4 was also included in the experiment. Each of the VHH was able to inhibit binding of IIIB gp120 to itself and to the other two VHH, indicating that either the VHH bind to epitopes that overlap or that steric hindrance inhibits simultaneous binding. VHH C8 required higher concentrations than VHH A12 or D7 to completely inhibit gp120 binding to VHH A12 and D7, as well as to itself. This finding is in concordance with the higher affinity for IIIB gp120 observed for VHH A12 and D7 compared to VHH C8 (Table 2). VHH A12 and D7 showed similar inhibition curves, in line with their similar affinities for IIIB gp120. Soluble CD4 could also completely inhibit VHH-gp120 binding. This result supports the observation reported in Fig. 4B, where the VHH were found to potently inhibit sCD4 binding to gp120. On a molar level, sCD4 was as potent as VHH A12 and D7 at inhibiting VHH-gp120 binding (Fig. 5C).
|
|
|---|
The neutralizing activities of the VHH were evaluated against a large panel of HIV-1 isolates from subtypes A, B, C, D, CRF02_AG, and CRF07_BC. The most potent and broadly reactive VHH was A12. A12 was able to neutralize 59% of the subtype B viruses and 48% of the subtype C viruses included in the study on an IC50 level, compared to 77% and 56% for the broadly neutralizing anti-CD4bs MAb b12. The difference in neutralization breadth between VHH A12 and MAb b12 was even more pronounced at an IC90 level, indicating that VHH A12 is less broad in its cross-subtype neutralization ability than MAb b12. It should also be noted that MAb b12 was more reactive against the well-characterized subtype B and C reference panels of envelope pseudotyped viruses, which has been classified as suitable for tier 2 assessment of neutralizing antibodies.
Apart from the autologous virus HIV-1 CN54, the VHH were not able to neutralize any of the 12 CRF07_BC viruses tested, despite being derived from a llama immunized with the CRF07_BC isolate CN54. Taking the relatively broad neutralization properties of the VHH against viruses from subtypes B and C, this finding is a little surprising. Further studies are needed to elucidate the extent to which these VHH-resistant CRF07_BC envelope clones differ from the CN54 envelope.
The neutralization ability of MAb b12 has been tested in many studies. In this study, MAb b12 was tested in parallel to the VHH against a panel of 65 viruses and was able to neutralize 54% of the viruses on an IC50 level and 40% on an IC90 level. These findings are in concordance with a previous study carried out by Binley et al. (8), in which a small panel of MAbs was tested against an extensive panel of 90 viruses in a high-throughput pseudovirus-based neutralization assay and in which MAb b12 was shown to neutralize 50% and 34% of viruses on an IC50 and IC90 level, respectively. Furthermore, MAb b12 was used in the characterization of the subtype B and C reference panels of envelope pseudotyped viruses described by Li et al. (43, 44), which were included in this study. Neutralization data obtained in this study for MAb b12 against these reference panels are concordant with the data previously published.
The VHH were found to neutralize a different spectrum of viruses compared to MAb b12. For example, MAb b12 was able to neutralize 12 subtype C or CRF07_BC viruses that VHH A12 was not able to neutralize (IC50 level). At the same time, VHH A12 was able to neutralize eight subtype C or CRF07_BC viruses that b12 was unable to neutralize (IC50 level). These differential neutralization reactivities may indicate that the VHH and MAb b12 recognize different but perhaps overlapping epitopes on gp120, despite competing with each other for binding to gp120 in the ELISA. Additional studies, such as alanine scanning of gp120, as well as structural studies are needed to fully characterize the precise epitopes to which the VHH bind and to determine whether the VHH and MAb b12 epitopes are related. Determining the epitopes of these VHH may provide additional information about vulnerable sites in proximity to the CD4bs of gp120, which may be of value for HIV-1 immunogen design.
Given the unique properties attributed to VHH, in terms of solubility, thermal and chemical stability, and high expression levels leading to a low production cost (20, 26, 64, 81), neutralizing VHH might prove useful in a number of applications, for example, as candidate HIV-1 microbicides as well as antiretroviral drugs or prophylactics. Topical application of MAb b12 has been shown to protect macaques from infection after vaginal challenge with SHIV, which supports the potential use of antibodies for topical prevention of HIV-1 transmission (82). Further studies are needed to establish whether the identified VHH would have preventive effects in vivo. Antienvelope VHH may also be of use in characterizing epitopes on HIV-1 envelope proteins. Attempts to design an HIV-1 immunogen that can elicit a broadly neutralizing humoral immune response have so far failed (61), although a recent study reported that extensively cross-reactive antibodies (albeit at a low titer) were induced by immunization of rabbits with a recombinant gp140 (93). It is possible that a set of neutralizing and nonneutralizing VHH may become useful in the design of such an immunogen.
The stability and solubility of VHH allow for engineering of multivalent and bispecific molecules (23). Engineered multivalent VHH to other antigens have previously been shown to exhibit better binding and neutralizing properties than their monovalent counterparts (13). It is possible that introduction of multivalency will increase the potency of the identified anti-gp120 VHH. Construction and evaluation of such multivalent VHH, using the VHH identified in this study, are ongoing but have so far not led to increased neutralization potency or breadth (data not shown). Selecting for VHH recognizing other epitopes and linking them to the CD4bs-targeted VHH described in this study, thus producing a bispecific VHH, might also result in more potent binding and neutralization properties.
The results obtained in this study confirm previous findings that antibody binding to recombinant envelope proteins does not necessarily correlate with the ability to neutralize virus (25, 35, 50, 54, 62, 66, 71, 73), as the structures of recombinant envelope glycoproteins are likely to be different from the structures of the envelope glycoproteins in the context of the functional spike (61, 91, 95). The ability of the VHH to bind to gp120 and gp140 from different subtypes indicates that they bind to a motif that is conserved and accessible on recombinant, soluble envelope proteins. The lack of ability of the VHH to neutralize some of the corresponding viruses indicates that the epitopes recognized by the VHH on recombinant envelope proteins are either not accessible or not present in the context of the functional envelope spike. The fact that the VHH can inhibit binding of sCD4 to recombinant 92UG037 gp140, despite not being able to neutralize 92UG037 virus, suggests that recombinant 92UG037 gp140 is not a good representative of the functional spike, and therefore such an antigen may not be suitable as an immunogen or for panning of VHH libraries. Interestingly, VHH binding to recombinant 92BR025 gp120 seems to correlate with the ability to neutralize 92BR025 virus. VHH A12 and C8 were able to bind well to recombinant 92BR025 gp120 in the ELISA, whereas D7 bound with a weaker signal (Fig. 4A). This observation correlates with the corresponding neutralization data, where A12 and C8 could neutralize HIV-1 92BR025 with IC50 titers of 0.2 and 2 µg/ml, respectively, whereas the IC50 for D7 was >50 µg/ml (Fig. 2). This finding may indicate that recombinant 92BR025 gp120 represents a good mimic of the 92BR025 gp120 structure in the functional spike. This notion is, however, speculative, and further studies are needed to evaluate the antigenic properties of this gp120. Immunizing llamas and selecting for anti-CD4bs VHH using an antigen that is a better representation of the functional envelope spike might result in VHH with broader cross-subtype neutralizing activities and lower inhibitory doses. Again, identification of such an envelope antigen is one of the major challenges in screening for agents that can act as HIV-1 entry inhibitors as well as in HIV-1 vaccine design (61).
The immunogen CN54 gp120 showed poor binding to sCD4 in biochemical assays and did not mediate infection when expressed on viruses. It was chosen as the immunogen due to its availability before these properties were known. It is possible that immunizing with an immunogen with better CD4-binding properties would result in more potent VHH. Previous studies have shown that purified trimeric envelope proteins are somewhat better at eliciting cross-subtype-neutralizing antibodies than monomeric recombinant gp120 (6, 7, 22, 32, 38, 90). We are currently screening llamas immunized with recombinant trimeric gp140.
The phage libraries were panned on HIV-1 IIIB gp120 alone, subtype A gp120 alone, or subtype C gp120 alone, as well as on the same antigens alternated in various combinations in two subsequent rounds of panning, always using the competitive elution with sCD4. Interestingly, with this method of selection, we could select for VHH that could bind to envelope proteins of subtypes A, B, and C, regardless of which of the above antigens that were used in the panning. However, when selected using HIV-1 IIIB gp120 only, a range of VHH were isolated that could be grouped into two families, whereas when the subtype A and C envelopes were included in the panning, only the most potent and broadly reactive VHH in the set were selected. This finding suggests that alternating the antigen in the panning procedure enables selection of VHH with better cross-reactive properties.
In conclusion, we have characterized three cross-reactive VHH that can neutralize several HIV-1 primary isolates of subtypes B and C. To our knowledge, this is the first description of broadly neutralizing MAbs to HIV-1 envelope which were derived from an immunized animal, as all previously reported broadly neutralizing anti-HIV-1 MAbs have been a result of natural infection rather than immunization (61). These VHH compete with sCD4 and anti-CD4bs MAbs for binding to recombinant gp120 and gp140. The results indicate that heavy chain antibody fragments have a possible use as potent HIV-1 entry inhibitors. Since VHH are stable and can be produced at a relatively low cost, they may be considered for applications such as HIV-1 microbicide development. Antienvelope VHH might also prove of use in defining neutralizing and nonneutralizing epitopes on HIV-1 envelope proteins, with implications for HIV-1 vaccine design.
This work was supported by the European Commission 6th Frame Work Programme as part of the European Microbicides Project and the European Vaccines and Microbicides Enterprise, the Bill & Melinda Gates Foundation, as part of the Collaboration for AIDS Vaccine Discovery (CAVD grant 38637), and the United Kingdom Medical Research Council.
Published ahead of print on 8 October 2008. ![]()
Supplemental material for this article may be found at http://jvi.asm.org/. ![]()
Present address: Virology and Chemotherapy Section, Catholic University of Leuven, Minderbroedersstraat 10 Blok x, Bus 1030, 3000 Leuven, Belgium. ![]()
Present address: Centre for Infectious Disease, Institute of Cell and Molecular Science, Barts and The London School of Medicine and Dentistry, The Blizard Building, 4 Newark Street, London E1 2AT, United Kingdom. ![]()
|
|
|---|
monoclonal antibody, in HIV-1 infected cells. J. Virol. Methods 139:17-23.[CrossRef][Medline]
1
2 mannose residues on the outer face of gp120. J. Virol. 76:7306-7321.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»