Surfactant Protein A Binds to HIV and Inhibits Direct Infection of CD4⁺ Cells, but Enhances Dendritic Cell-Mediated Viral Transfer¹

Surfactant protein A (SP-A)³ is a member of the collectin family of C-type lectin binding proteins that includes surfactant protein D (SP-D), mannan-binding lectin (MBL), conglutinin, and collectin 43 (1, 2). Collectins are composed of subunits of three polypeptide chains. Each polypeptide in a collectin consists of a short interchain disulfide bond-forming N-terminal domain, a collagen-like region, an α-helical coiled-coil neck, and a calcium-dependent carbohydrate recognition domain (CRD) at the C terminus that is highly conserved across species (3). Oligomerization of the collectin subunits promotes the binding of these molecules to carbohydrated-rich pathogen associated molecular patterns, with SP-A resembling a “bouquet of flowers” (4). Binding of SP-A results in the promotion of pathogen phagocytosis (5, 6). In addition, SP-A has been shown to stimulate chemotaxis (7), modulate inflammatory responses (8), and have direct anti-microbial effects (9).

Although SP-A has been shown to bind to a number of pathogens, its identification in biological fluids and at important sites for HIV transmission and disease progression such the lungs (10), amniotic fluid (11), and recently vaginal fluid and the female genitourinary tract (12), suggests that the collectin may potentially play an important immunological role in early phase, late-term, and perinatal transmission of HIV in vivo. In addition, the high level of glycosylation on the gp120 envelope protein of HIV known as a glycan shield, which is critical for its proper folding and evasion from the host immune response, would make it a prime target for the SP-A CRDs (13). The collectins SP-D and MBL have already been shown to bind to gp120 through a protein-carbohydrate interaction to inhibit HIV infectivity (14, 15).

In this study, we demonstrate that SP-A binds to HIV by targeting the high mannose oligosaccharides on the gp120 envelope protein. Infectivity experiments reveal that SP-A neutralizes both R5 and X4 strains of HIV, whereas competition experiments suggest that occlusion of the CD4 binding-site (CD4BS) on gp120 was the likely mechanism of inhibition. Experiments with immature monocyte derived dendritic cells (iMDDCs) reveal that SP-A enhances the binding of gp120 to DCs, uptake of the viral particles, and DC-mediated transfer of HIV to T cells at both a neutral pH and at a more acidic pH representative of the female vaginal tract. Thus, SP-A appears to a be a dual modulator of HIV infection by protecting CD4⁺ T cells from direct infection, but enhancing the transfer of HIV to CD4⁺ T cells by DCs.

Aldrithiol-2 (AT-2) inactivated HIV BaL particles were provided by M_r. Julian Bess of the National Cancer Institute AIDS Vaccine Program (SAIC, Frederick, MD). Infectious HIV BaL (ARP118) and HIV IIIB (ARP101.1) were obtained from the National Institute of Biological Standards and Control AIDS Reagent program. High titer stocks were generated by infecting 1 × 10⁷ pelleted PM1 cells for 90 min at 37°C/5% CO₂ with 500 μl of viral supernatant. After incubation, 10 mLs of RPMI 1640 supplemented with 50 U/ml penicillin/streptomycin (Life Technologies), 2 mM l-Glutamine (Sigma-Aldrich), and 10% (v/v) FCS (Sigma-Aldrich) (R10) was added and the virally infected cell culture was transferred to a T25 flask for growth. Aliquots (1 ml) were taken on days 5 and 7 and assayed for RT activity using the SPA Quant-T-RT assay kit (Amersham) according to the manufacturer’s instructions.

SP-A was purified from therapeutic lung lavage obtained from alveolar proteinosis patients as described previously (16). Recombinant soluble DC-SIGN was produced as described previously (17). The concentration of SP-A and DC-SIGN was determined by analysis of amino acid composition. Recombinant gp120 IIIB produced in Chinese hamster ovary (CHO) cells, biotinylated gp120 IIIB, and FITC-labeled gp120 IIIB were purchased from Immunodiagnostics. Cyanovirin (CVN) was provided by Dr. Kirk Gustafson (National Cancer Institute, Frederick, MD). Collectins used in all cell-based assays were treated for endotoxin removal, by passing the protein solutions through a 10 ml Polymixin B column (Detoxi-Gel; Pierce) in sterile PBS (pH 7.4). Remaining levels of endotoxin were assayed using a Limulus Amoebocyte Lysate kit, according to the manufacturer’s instructions (Bio-Whittaker). An endotoxin level of <10 pg/μg of protein was judged acceptable to use in cell-based assays.

The CD4 binding site directed Abs b12 and F105 were obtained from the National Institute of Biological Standards and Control AIDS Reagent program. mAb 2G12, which recognizes a mannose-dependent epitope on gp120 (18) was provided by Dr. Dennis Burton (Scripps Research Institute, San Diego, CA) and Dr. Pauline Rudd (Glycobiology Institute, Oxford University, Oxford, UK). A HRP conjugated polyclonal Ab preparation against gp120 IIIB and a HRP conjugated mouse mAb preparation directed against the V3 domain of gp120 IIIB were both purchased from Immunodiagnostics. Pooled anti-HIV Env IgG was provided by Dr. Quentin Sattentau (Dunn School of Pathology, Oxford University, Oxford, UK). Rabbit anti-CVN Ab was provided by Dr. Kirk Gustafson.

Inactivated HIV BaL (10 μg/ml), gp120 IIIB (2 μg/ml), or SP-A (2 μg/ml) was immobilized on 96-well plates (Nunc, Maxisorp) in 0.1 M sodium bicarbonate buffer (pH 9.6) at 4°C for 18h. The wells were washed with PBS and 0.05% (v/v) Tween 20 and blocked in 3% (v/v) BSA for 2 h at 37°C. Immobilization of virus was verified by incubation with purified pooled anti-HIV Env IgG, followed by peroxidase conjugated anti-human IgG, and detection using the H₂O₂-tetramethylbenzidine-based (H₂O₂-TMB) chromogenic substance according to the manufacturer’s instructions (Bio-Rad). After washing away excess BSA with PBS and 0.05% (v/v) Tween 20, the wells were incubated with increasing concentrations of SP-A (0–4 μg/ml) or AT-2 inactivated HIV BaL (0–10 μg/ml) in Tris saline calcium (20 mM Tris-HCl, 150 mM NaCl, 5 mM CaCl₂, pH 7.4) with 25% (v/v) human serum or Tris saline EDTA(TSE; 20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, pH 7.4) with 25% (v/v) human serum. For binding experiments performed at a pH of 5.0, SP-A or AT-2 inactivated HIV BaL was incubated in 20 mM Sodium Acetate, 150 mM NaCl, 25% (v/v) human serum and either 5 mM CaCl₂ or 2 mM EDTA. The wells were washed in the corresponding binding buffer, and incubated with biotinylated anti-human SP-A Ab (1 μg/ml) at 37°C for 2 h. Bound HIV particles were lysed in PBS with 1% (v/v) Triton X-100 and p24 Ag was measured by ELISA according to the manufacturer’s instructions (Immunodiagnostics). The collectin-antibody complexes or p24 Ag were detected using HRP-streptavidin and H₂O₂-TMB. The absorbance (450 nm) of individual wells was measured by a spectrophotomer (Multiscan Ascent, Labsystems; Fisher).

Biotinylated SP-A (1 mg/ml) and DC-SIGN (1 mg/ml) were immobilized on individual flow cells, all at the approximate reading of 1000 resonance units, on the same SA chip in a pH 5.0 Sodium Acetate buffer according to the manufacturer’s instructions (Biacore). To determine whether HIV could bind to the immobilized lectins, AT-2 inactivated HIV BaL particles (100 μg/ml) were passed over all of the flow cells in 10 mM HEPES (pH 7.4), 150 mM NaCl, 5 mM CaCl₂, 0.005% (v/v) surfactant P-20, 0.02% (w/v) NaN₃ (HSC) buffer or 10 mM HEPES (pH 7.4), 150 mM NaCl, 1 mM EDTA, 0.005% (v/v) surfactant P-20, 0.02% (w/v) NaN₃ (HSE) at 10 μl/min for 2 min at 25°C. The complexes were allowed to disassociate for 90 s, and bound viral particles were removed with two 20 μl washes of HSE and the chip was re-equilibrated with 20 μl of HSC buffer.

Competition assays using the SP-A/DC-SIGN chip were accomplished by flowing AT-2 inactivated HIV particles (100 μg/ml) in the presence of different concentrations of hexoses (0–500 mM). The surface plasmon resonance response obtained in the absence of any competitor was considered as 100%, and the relative binding was calculated for each competitor concentration.

To examine the effects of pH on the binding of HIV to immobilized lectins, AT-2 inactivated HIV BaL particles (100 μg/ml) were diluted into a 50 mM Sodium Acetate, 50 mM Bis-Tris, 50 mM Tris, 150 mM NaCl, 5 mM CaCl₂ buffer (Buffering capacity: pH 4.5–8.0) at various pHs (5.0, 5.5, 6.0, 6.5, 7.4). AT-2 inactivated HIV BaL particles (100 μg/ml) were passed over all of the flow cells in HSC at a particular pH at 10 μl/min for 2 min at 25°C. The complexes were allowed to disassociate for 90 s and bound viral particles were removed with two 20 μl washes of HSE and the chip was re-equilibrated with 20 μl of HSC buffer.

SP-A and SP-D were dialyzed into 50 mM HEPES, 50 mM NaCl, 2 mM EDTA and then incubated with 5 ml of Toyopearl AF-Tresyl-650M beads for 3 h at 23°C on a rotary mixer. Protein coupling was determined by OD₂₈₀ measurement of the supernatant. Uncoupled sites were consumed by incubation of beads with 100 mM Ethanolamine for 3 h at 23°C. Both SP-A and SP-D were immobilized at an approximate concentration of 1 mg/ml.

Inactivated HIV BaL (50 μg) was lysed in PBS and 1.0% (v/v) Triton X-100 for 1 h at 37°C and then incubated with 500 μl of SP-D or SP-A resin Tris saline calcium or TSE at 23°C for 2 h. Beads were pelleted and washed three times by centrifugation at 3000 rpms for 5 min and the bound proteins were eluted in 200 μl of TSE. Samples were separated on a NuPAGE 4–12% Bis-Tris gradient gel (Invitrogen) and analyzed by both Simply SafeBlue Staining (Invitrogen) and N-terminal sequencing.

Recombinant gp120 IIIB (50 μg) expressed in CHO cells was digested with 10 U of Endoglycosidase H (EndoH; New England Biolabs) for 18 h at 37°C. Controls included mock-treated gp120 (digestion buffer, no enzyme). Protein was acetone precipitated to remove free sugars and redissolved in double deionized H₂O. The extent of deglycosylation was confirmed by SDS-PAGE analysis of mock-treated and EndoH-treated gp120.

PM1 cells were infected with 3-log₁₀ TCID₅₀/ml of virus inoculum (HIV BaL or IIIB) that had been preincubated with 10 μl of serially diluted SP-A (final concentration: 1–10 μg/ml) for 1 h at 37°C. 90 min postinfection, the cells were washed extensively, resuspended in fresh R10, and aliquoted into 200 μl volumes containing 5 × 10⁴ cells/well. Infected cells were then incubated for 4–5 days before measuring extracellular p24 content from culture supernatants by ELISA (Immunodiagnostics).

SP-A (0–40 μg/ml) was incubated simultaneously with various protein competitors (biotinylated recombinant CD4, mAb b12, mAb F105, mAb 2G12, biotinylated recombinant soluble DC-SIGN or recombinant CVN) in wells coated with recombinant gp120 IIIB (2 μg/ml). Bound biotinylated CD4 and DC-SIGN was detected by streptavidin-HRP incubation for 30 min at 37°C, followed by H₂O₂-TMB. Binding of MAbs 2G12, b12, and F105 was detected using an HRP-conjugated polyclonal anti-human IgG (Sigma-Aldrich), followed by H₂O₂-TMB. CVN binding was detected using a rabbit anti-CVN Ab, followed by HRP-conjugated anti-rabbit IgG (Sigma-Aldrich) and H₂O₂-TMB.

Immature monocyte-derived DCs (iMDDCs) were generated from a highly enriched population of CD14⁺ monocytes. PBMCs were isolated from blood obtained from healthy male donors (ages: 18–40) using a Ficoll-Hypaque density gradient (Amersham Biosciences). PBMCs were washed thoroughly with sterile PBS and contaminating RBC were lysed using ACK lysis buffer. PBMCs were labeled with anti-CD14 microbeads (Miltenyi Biotec) and CD14⁺ cells were isolated using the AutoMACS (Miltenyi Biotec) magnetic cell sorter. The purity of the isolated CD14⁺ cells was >95% as assessed by flow cytometry. iMDDCs were cultured from monocytes in the presence of IL-4 (500 U/ml; R&D Systems) and GM-CSF (800 U/ml; R&D Systems), with fresh cytokines added on days 3 and 5. At day 7, the phenotype of the cultured iMDDCs was confirmed by high CD11c and HLA-DR expression, and low CD83 expression as determined by flow cytometry.

FITC-labeled gp120 IIIB was incubated in the presence of SP-A (5 μg/ml) in R10 made 5 mM CaCl₂ for 1 h at 4°C. SP-A-gp120 complexes were incubated with iMDDCs (1 × 10⁵) and incubated for 2 h at 4°C to prevent uptake of gp120. Cells were pelleted and washed in PBS with calcium and magnesium and then fixed in 3% paraformaldehyde for analysis by flow cytometry.

AT-2 inactivated HIV BaL was incubated in the presence of SP-A (5 μg/ml) in R10 made 5 mM CaCl₂ for 1 h at 37°C. Collectins complexed with HIV were then added to iMDDCs (1 × 10⁵) aliquoted into wells of a 96-well tissue culture treated microtiter plate and incubated for 2 h at 37°C. Cells were pelleted and extensively washed with RPMI and with PBS containing 10 mM EDTA to remove externally bound HIV particles. DCs incubated with AT-2 inactivated HIV BaL were lysed in PBS with 1% (v/v) Triton X-100 and viral uptake was measured by p24 ELISA according to the manufacturer’s instructions (Immunodiagnostics).

DC-mediated HIV transmission assays were performed as described previously (19). Infectious HIV BaL particles were incubated in the presence of SP-A (1–10 μg/ml) for 1 h at 37°C before incubation with iMDDCs (1 × 10⁵) for 2 h at 37°C. Cells were extensively washed to remove unbound virus, and then placed in a fresh 96-well plate for coculture with PM1 cells (1 × 10⁵/well). Supernatants were collected after 4–5 days and virus replication was monitored by p24 ELISA.

Statistical comparisons were made using the student t test with equal variance or ANOVA. Values of p for the differences between means were calculated using Excel software (Microsoft).

To determine whether SP-A can bind to HIV, we immobilized a fixed amount (10 μg/ml) of AT-2 inactivated HIV BaL particles in a microtiter plate. Inactivation using AT-2 has previously been shown to have no effect on the conformational structure of HIV particles in comparison to infectious virus (20). SP-A (0–4 μg/ml) was allowed to bind to HIV BaL and the bound collectin was detected using an anti-SP-A Ab. As shown in Fig. 1 A, SP-A bound to immobilized HIV particles in the presence of CaCl₂, but was inhibited by mannose and EDTA, suggesting involvement of the C-type lectin CRD of SP-A.

Incubation of AT-2 inactivated HIV BaL particles in microtiter wells coated with SP-A further verified the interaction between the collectin and HIV. The binding was examined at both a pH of 7.4 and 5.0 to replicate physiological environments such as the serum and lungs or the female vaginal tract. The vaginal tract is lowered to an approximate pH buffering range of 5.0–5.5 as a result of colonizing lactobacilli, which release acidic by-products (21). As shown in Fig. 1 B, immobilized SP-A binds to HIV particles at a pH of 7.4 in the presence of the calcium, but is inhibited by mannose and EDTA. At a pH of 5.0, binding of SP-A to HIV in the presence of calcium is slightly reduced, but the binding of SP-A in the presence of EDTA is significantly reduced.

To examine the interaction with greater sensitivity, we allowed HIV BaL particles to bind to immobilized SP-A in the presence of calcium or EDTA using surface plasmon resonance. BIACore assays behaved similarly to our ELISAs and indicated that SP-A binds to HIV in a calcium-dependent manner when performed at a pH of 7.4 (Fig. 1,C). We further investigated the nature of the interaction using our BIACore system by allowing SP-A to bind to HIV in the presence of increasing concentrations of hexoses (0–500 mM). As shown in Fig. 1 D, mannose and glucose were the most effective inhibitors of the interaction with significantly lower IC₅₀ values in comparison to D-galactose and N-acetyl glucosamine (GlcNAc). Collectively, these results suggest that SP-A binds to HIV using C-type lectin activity, while the strong inhibition by D-mannose suggests that the high mannose glycans on the gp120 envelope protein of HIV may potentially be ligands for SP-A.

To identify the target on HIV recognized by SP-A, we incubated 50 μg of HIV BaL viral lysate with SP-A immobilized on ToyoPearl AF-650M-Tresyl resin. The same experiments were performed using SP-D given the previously reported interaction between this collectin and gp120 (15). The bound proteins were eluted in TSE and run on SDS-PAGE under reducing conditions. As shown in Fig. 2, SP-A interacts specifically with a ∼120 kDa species in the presence of CaCl₂ (Fig. 2, lane 3) but not in the presence of EDTA. SP-D bound to a similar sized protein (Fig. 2, lane 5). N-terminal sequencing of both ∼120 kDa bands resulted in the sequence EKLWVTVYYGVPVWK, which corresponds with 100% accuracy to the known N-terminal sequence of gp120 BaL (22).

To directly verify whether SP-A can bind to gp120, we immobilized a fixed amount (2 μg/ml) of recombinant gp120 IIIB expressed in CHO cells on an ELISA plate. SP-A (0–1 μg/ml) was allowed to bind to gp120 IIIB, and the bound protein was detected by biotinylated anti-SP-A Ab. The results show that SP-A binds to gp120 IIIB in a concentration- and calcium-dependent manner (Fig. 3,A). Similarly, our ELISA data shows that SP-A immobilized to a microtiter plate can also bind to gp120 IIIB in the presence of calcium, but binding is significantly reduced in the presence of EDTA, at a pH of 7.4 and 5.0 (Fig. 3 B).

As competition with carbohydrate ligands provides only indirect evidence of carbohydrate-dependent binding, we sought to directly determine whether removal of high mannose glycans would affect binding of SP-A and SP-D to gp120. The glycosylation of gp120 IIIB synthesized in CHO cells is made up of 13 complex oligosaccharides and 11 high mannose or hybrid oligosaccharide structures (23, 24). Digestion with the enzyme Endo H selectively removes the high mannose glycans while leaving the hybrid and complex oligosaccharides intact (25). As shown in Fig. 4,A, treatment with Endo H results in a lower m.w. species (∼10–20 kDa smaller), indicating successful removal of high mannose N-linked glycans. To determine the effect of their removal, we incubated mock-treated and Endo H-treated gp120 IIIB with SP-A (2 μg/ml) or SP-D (2 μg/ml) immobilized on an ELISA plate, and then detected the bound gp120 with a monoclonal anti-gp120 IIIB Ab directed against the V3 loop. The results show that treatment with Endo H abrogates binding of SP-A and SP-D to gp120 by greater that 90% when compared with binding to mock-treated gp120 (Fig. 4 B), indicating than the high mannose glycans are the predominant targets on gp120 for SP-A.

Given that our binding studies clearly show that SP-A binds to HIV and gp120, we next asked whether SP-A could inhibit HIV infectivity. Infectious strains of HIV BaL and IIIB were incubated with increasing concentrations of SP-A before incubation with target cells. Experiments with HIV BaL were performed at a pH of 5.0, as R5 strains such as HIV BaL are typically found in early phase infection, while experiments with HIV IIIB were performed at a pH of 7.4, as viral particles found in the lungs during late-term infection have a CXCR4 receptor tropism (26). Modulation of the pH to 5.0 or 7.4 was conducted during the initial preincubation of virus with collectin and when virus-collectin complexes were incubated with target cells. All cells were washed postinfection and then resuspended in prewarmed R10, pH 7.4. At its highest concentration of 10 μg/ml, SP-A inhibited infectivity of HIV BaL and IIIB by ∼90% and ∼80%, respectively (Fig. 5,A). The effect was reproducible with different preparations of SP-A and the inhibitory effect was abrogated by preincubation of SP-A with 100 mM D-mannose (Fig. 5 A).

To determine the mechanism of neutralization, we examined whether SP-A interfered with the binding of a variety of known ligands for gp120. Using an ELISA based system, we first investigated whether SP-A could inhibit the binding of mAb 2G12 and the recombinant protein CVN. Both of these proteins neutralize HIV by targeting a specific mannose-dependent epitope (21, 27), and thus were particularly interesting given the targeting of the high mannose structures on gp120 by SP-A. As shown in Fig. 5,B, SP-A did not interfere with the binding of either 2G12 or CVN, suggesting that it did not target this particular epitope on gp120. We next tested whether SP-A could compete the binding of sCD4 to monomeric gp120 IIIB. As shown in Fig. 5 B, at a concentration of 10 μg/ml, SP-A did effectively compete the interaction by ∼70%. The blocking of the CD4BS-directed MAbs b12 and F105 to a similar degree further confirmed the targeting of this particular site on gp120 by SP-A. Thus, these results suggest that SP-A inhibits the infection of target cells by blocking the interaction of gp120 with CD4.

In addition to the role that SP-A and SP-D may have in protecting CD4⁺ cells from infection, we were also interested in investigating the effect that these collectins may have on DC-mediated transfer of HIV to CD4⁺ T cells. Therefore, we first examined whether SP-A or SP-D affects the interaction of gp120 with iMDDCs. FITC-labeled gp120 (2 μg/ml) was incubated with iMDDCs in the presence and absence of SP-A (5 μg/ml) at both a pH of 7.4 and 5.0. Experiments were performed at 4°C to prevent uptake of gp120 by DCs. As shown in Fig. 6, SP-A significantly enhanced the binding of gp120 to DCs at both pHs, but the enhancement at a pH of 5.0 was greater than 3-fold the enhancement at a pH of 7.4 as determined by mean intensity of fluorescence.

Based on reports that show that SP-A can enhance the phagocytosis of a number of pathogens (28, 29, 30), we investigated whether SP-A enhances the uptake of HIV particles by iMDDCs. AT-2 inactivated HIV BaL particles (10 μg/ml) were incubated with iMDDCs in the presence and absence of SP-A (5 μg/ml) at a pH of both 7.4 and 5.0 at 37°C. After incubation, cells were washed in 10 mM EDTA-containing buffer and then lysed and analyzed for p24 Ag content. As shown in Fig. 7 A, SP-A significantly enhanced the uptake of viral particles to ∼150% of the control at a pH of 7.4, but had a far greater effect at a pH of 5.0 with a >5-fold enhancement of viral capture in comparison to the DC control. Thus, similarly to the gp120-DC binding experiment, these results suggest that SP-A enhances the interaction of HIV with DCs.

To investigate whether SP-A affects DC-mediated transfer of CD4⁺ T cells, iMDDCs were incubated with infectious HIV BaL particles in the presence and absence of SP-A (1–10 μg/ml) at both a pH of 7.4 and 5.0 for 2h. After the incubation, DCs were washed extensively and then cocultured with PM1 cells for 5 days before analysis of culture supernatants by p24 ELISA. As shown in Fig. 7 B, SP-A significantly increases the transfer of HIV particles to PM1 cells at a pH of 7.4 in a concentration-dependent manner, with a >2.5-fold enhancement at the highest concentration of 10 μg/ml. The lowering of the pH to 5.0 significantly reduced the overall transfer of infectivity to CD4⁺ T cells, but consequently the effect of SP-A was further amplified with a >5.0-fold enhancement at 10 μg/ml. Thus, these results suggest that SP-A may have a harmful effect in vivo by augmenting HIV infection.

To understand why the effect by SP-A on gp120 binding, viral capture, and DC-mediated transfer was amplified at a pH of 5.0 relative to a pH of 7.4, we compared the HIV binding capacity of SP-A to that of the DC receptor DC-SIGN, while modulating the pH. DC-SIGN has been shown to be a key receptor in DC-mediated transfer of HIV infectivity (31, 32, 33). SP-A and recombinant soluble DC-SIGN were immobilized to individual flow cells of a BIACore chip and the binding to HIV BaL particles was measured by changes in surface plasmon resonance. As shown in Fig. 8,A, SP-A binds in a pH-dependent manner, losing ∼50% of its binding capacity at a pH of 5.0. In comparison to DC-SIGN (Fig. 8,B), which loses 100% of its binding at a pH of 5.0, the binding of SP-A was significantly stronger. The pH-sensitivity of DC-SIGN may account for the overall reduction in gp120-binding, uptake, and DC-mediated transfer at a pH of 5.0 (Figs. 6 and 7), while the retention of HIV-binding activity by SP-A would allow for preferential binding and uptake of viral particles mediated by the collectin. It may also suggest that in a low pH physiological environment, such as the female vaginal tract, SP-A may have a more important immunological role against HIV than DC-SIGN.

In this study, we investigated the interaction of SP-A and HIV. The identification of SP-A at important sites for HIV transmission and disease progression suggested that this protein could be playing an important role in various stages of the immune response against HIV. Therefore, we sought to determine whether SP-A could bind to HIV and whether this interaction may have a significant effect on viral pathogenesis.

This report clearly shows that SP-A binds to AT-2 inactivated HIV particles in a calcium-dependent manner that is inhibitable by mannose and EDTA (Fig. 1, A–C). HIV virions inactivated by AT-2 have been shown to retain conformational and functional integrity of their surface proteins (19), suggesting that SP-A would bind in a similar manner to infectious particles. The requirement of Ca²⁺ ions and inhibition by EDTA suggests involvement of the CRD of the collectin, while inhibition by various hexoses, most notably mannose, suggests that SP-A binds to a glycoconjugate on HIV through a C-type lectin interaction (Fig. 1 D). Binding assays conducted at a pH of 5.0 to replicate the physiological environment of the vaginal tract reveal enhanced binding by SP-A in the presence of EDTA relative to its binding in EDTA at a neutral pH. This could indicate that SP-A also binds to HIV through a CRD-independent mechanism as the pH is reduced, however it is more likely that the increase in acidity results in a conformational change in the CRD of SP-A which can substitute for the presence of Ca²⁺, as has previously been suggested (34).

Affinity capture and N-terminal sequencing of bound proteins from the viral lysate revealed that the primary target of SP-A is the highly glycosylated gp120 envelope protein of HIV (Fig. 2). Similar results were obtained for SP-D, which was used as a positive control in our experiments. This is consistent with previous reports that have shown that gp120 is also the primary ligand for other C-type lectins such as MBL (14) and DC-SIGN (35). Additional binding experiments revealed that SP-A could directly bind to recombinant gp120 IIIB by ELISA (Fig. 3,A) and that the interaction was inhibitable by EDTA at both a pH of 7.4 and 5.0 (Fig. 3,B). The abrogation of SP-A and SP-D binding upon removal of high mannose structures on gp120 (Fig. 4 B) was also consistent with previous reports on other C-type lectins. Binding of MBL to gp120 was inhibited by treatment with N-glycanase (14) or EndoH (36). High mannose deficient gp120 was similarly shown to be unable to bind to DC-SIGN and monocyte-derived DCs (35). Therefore, these results suggest that the SP-A binds to gp120 in a similar manner to MBL and DC-SIGN.

The binding of SP-A to gp120 in our experiments conflicts with a previous report which observed no binding of SP-A to recombinant gp120 BaL by ELISA (15). The incongruence of the results could be due to variations in assay design or strain specific differences, as SP-A has been shown to bind to different strains of IAV in a calcium-dependent and -independent manner (37, 38). However, a variety of techniques were used in this report to verify the interaction between SP-A and HIV and gp120. In addition, previous reports which indicate binding of SP-A to other viral glycoproteins, such as the F (fusion) and G (adherence) proteins of RSV, supports the notion that SP-A is likely to bind to a highly glycosylated surface protein such as gp120 (39).

Virus infectivity experiments were performed with concentrations of SP-A that are found in mucosal areas, amniotic fluid, and unconcentrated BAL fluids. Levels in BAL fluids are ∼2.82 μg/ml (40), whereas concentrations in situ in the lung in association with lung surfactant are considerably higher. SP-A levels have also been shown to be elevated to 6 μg/ml in HIV-infected BAL (41) and have been shown to be as high as 24 μg/ml in amniotic fluid at term (42), when SP-A acts as signal for parturition (43). Levels of SP-A in the vaginal tract have yet to be definitively determined, and the collectin likely makes up only a part of the milieu of defense proteins that include human β-defensins, lysozyme, lactoferrin, and human neutrophil peptides which have already been shown to inhibit HIV infection (44).

However, at these known concentrations, SP-A did inhibit infectivity of the X4 strain HIV IIIB and the R5 strain HIV BaL (Fig. 5 A). Experiments with HIV IIIB were performed at a pH of 7.4, as viral particles are found in the lungs during late-term infection usually have a CXCR4 receptor tropism (27). Assays with HIV BaL were performed at a pH representative of the vaginal tract, as R5 strains are typically found in early phase infection. Thus, these results indicate that SP-A may have a significant effect in vivo by protecting CD4⁺ cells from direct infection in a variety of physiological environments.

Our competition assay results suggest that the mechanism of neutralization by SP-A involves blocking the binding of gp120 to the cell surface molecule CD4 (Fig. 5 B), thereby preventing CD4-mediated viral fusion and entry. This was further confirmed by the inhibition of both mAb b12 and mAb F105 binding, both of which target the CD4BS. The lack of an inhibitory effect by SP-A on the binding of mAb 2G12 and CVN was surprising given its targeting of the high mannose structures on gp120 and the known affinity of the SP-A CRDs for mannose. However, this is consistent with the C-type lectin DC-SIGN, which also does not overlap the specific 2G12 or CVN epitopes, despite its mannose-dependent binding (35). The neutralization effect by SP-A was similar to that observed for the structurally homologous collectin MBL (16). When used in our competition studies, MBL also blocked the binding of gp120 to CD4 (data not shown), although not as effectively as SP-A. This would suggest that the bouquet of flowers structure of SP-A and MBL allows them to bind in a specific orientation on gp120 to occlude the CD4BS. The relatively more potent inhibition of the gp120-CD4 interaction by SP-A may explain why neutralization was achieved at a far lower concentration that those reported for MBL of ∼30–50 μg/ml (16).

The enhancement of DC-mediated transfer of infection (Fig. 7,B) suggests that SP-A would likely facilitate HIV infection through this alternate viral dissemination pathway in vivo. The localization of cells expressing SP-A in the female vaginal tract places the collectin at an important site to affect virus interaction with DCs during early phase infection (15), while the high level of SP-A in the respiratory tract would facilitate the interaction of HIV with lung DCs. Despite having been shown to have no effect on viral uptake by alveolar macrophages (45), the mechanism of SP-A-mediated enhancement of DC viral transfer that we observed most likely involves the increase in gp120 binding and viral uptake by DCs (Figs. 6 and 7 A). Although monocyte-derived dendritic cells (MDDCs) are a relatively good model for DC subsets involved in HIV infectivity in vivo (46), and the levels of p24 Ag are considerably higher with the use of PM1 cells as indicator cells in comparison to DCs alone (data not shown), it is unclear how much viral transfer is attributable to DC trans or cis-infectivity. Experiments with a replication-defective, single cycle reporter HIV would allow for the determination of whether viral replication is necessary for SP-A-mediated enhancement of HIV transfer by DCs.

The decrease in DC-mediated transfer of infectivity as the pH is reduced to 5.0 is likely the result of the loss of activity of HIV-binding cell surface receptors, such as DC-SIGN (Fig. 8), and impaired HIV infectivity (47, 48, 49). The loss of HIV binding by DC-SIGN as the pH was lowered to 5.0 is consistent with previous reports which have shown that a drop in pH results in a conformational change in the DC-SIGN CRD that alters its binding capacity (50). Our observation that SP-A is still able to bind to HIV as the pH is lowered may therefore account for the amplified effects by the collectin on gp120 binding and HIV uptake at a pH of 5.0 (Fig. 8). This enhancement on gp120 binding and HIV uptake may also be due to increased interaction between SP-A and a number of candidate receptors, such as C1qRp (51), SPR210 (52), or calreticulin and CD91 (8), each of which have been suggested to bind to the collagen region of SP-A and mediate uptake.

In summary, this study is the first to establish an interaction between SP-A and HIV. Our results suggest that SP-A may be a dual modulator of HIV infection by protecting CD4⁺ cells but enhancing the transfer of infection by dendritic cells. Therefore, we believe that our findings on this novel HIV binding factor are important in furthering our understanding of the innate immune response against HIV.

We thank Tony Willis for assistance with protein characterization and N-terminal sequencing.

The authors have no financial conflict of interest.