The Mannose Receptor Mediates Induction of IFN-α in Peripheral Blood Dendritic Cells by Enveloped RNA and DNA Viruses¹

Interferons, especially IFN-α and -β, play an important and indispensable role in resistance to viral infection (1, 2). These important mediators are produced by cells in response to stimulation with numerous agents, including viruses, bacteria, synthetic nucleic acids (e.g., polyinosinic acid:polycytidylic acid), and small molecules such as the antiviral compound, imiquimod. Monocytes produce IFN upon infection with the paramyxovirus, Sendai virus, as well as after incubation with polyinosinic acid:polycytidylic acid (3, 4, 5). In contrast, most enveloped viruses, including herpesviruses (e.g., herpes simplex virus type 1 (HSV-1),³ CMV, retroviruses (e.g., HIV-1), and rhabdoviruses (e.g., vesicular stomatitis virus (VSV)), stimulate IFN-α production primarily by rare (∼1/1000) cells in peripheral blood that are negative for lineage-specific markers of T cells, B cells, NK cells, and monocytes (4, 6, 7, 8). These cells, which have been termed natural IFN-α-producing cells, or NIPC, copurify with peripheral blood dendritic cells (DC) using FACS (9). Most recently, Cederblad et al. demonstrated using intracellular flow cytometry for IFN-α that these IFN-α-producing cells (IPC) appear identical to immature DC (10).

Several studies have sought to define the mechanisms involved in the stimulation of IFN-α expression in PBMC by enveloped viruses. Lebon reported that Abs to HSV-1 glycoprotein D inhibit the induction of IFN-α, and we observed that sera from HSV seropositive, but not seronegative, donors inhibited IFN production induced by this virus, suggesting that interaction of HSV-1 envelope proteins with IPC is necessary to stimulate IFN-α synthesis (11, 12). In addition, chloroquine, a lysosomotropic agent, inhibited the induction of IFN-α, suggesting that a receptor that recycles through the lysosomal/endosomal compartment may be involved in the induction mechanism (4, 12). Besides live HSV, UV-irradiated HSV and HSV-infected, glutaraldehyde-fixed fibroblasts are potent IFN inducers of IFN-α (11, 13, 14). Furthermore, liposomes derived from Sendai virus envelope are reported to stimulate IFN-α production, and soluble, recombinant HIV-1 gp120 has also been reported to induce IFN-α (15, 16). Although other laboratories, including our own, have not confirmed the results of Capobianchi that soluble recombinant gp120 is able to induce high levels of IFN-α production (in our hands gp120 is a very weak inducer), we have been able to block the induction of IFN-α by PBMC in response to HIV-1-infected H9 cells by either Ab to CD4 or gp120, thus suggesting a role for the viral glycoprotein in induction of IFN (Bocarsly et al., unpublished studies). Similar results were reported by Francis and Meltzer (17).

Although envelope glycoproteins seem necessary for induction of IFN-α by viruses, the nature of their interaction with the IPC is unknown. Studies by Charley et al. using the porcine coronavirus, transmissible gastroenteritis virus, however, indicate an important role for glycosylation of envelope proteins. Transmissible gastroenteritis virus treated with glycosidases or mutated at N-linked glycosylation sites of glycoprotein M lost its IFN-α-inducing potential (18, 19). This requirement for properly glycosylated viral envelope glycoprotein lead us to postulate that a lectin receptor on DC may be involved in the virus-IPC interaction that leads to the induction of IFN-α expression.

Mononuclear phagocytes express numerous receptors with carbohydrate binding ability. One of the best characterized of these receptors, the mannose receptor (MR), was initially identified on alveolar macrophages (20). Subsequently, this receptor was also observed on many tissue-differentiated macrophages, but it is not expressed on freshly isolated peripheral blood monocytes (21). The MR binds several monosaccharides, including fucose, N-acetylglucosamine, and mannose, with high affinity (20, 22). Several studies have shown that this receptor has many functions: it internalizes ligands via receptor-mediated endocytosis, transduces a glycoprotein-induced mitogenic signal to alveolar smooth muscle cells, and may mediate the induction of TNF-α produc-tion, lysosomal enzyme secretion, and cytotoxicity by macrophages (20, 21).

In addition to the MR, a galactose/N-acetylgalactosamine binding receptor has been cloned and found to be expressed by rat, murine, and, most recently, human macrophages (23, 24). A murine receptor with similar binding specificity and sequence homology to the MR, DEC205, has also been cloned, and this receptor appears to be expressed primarily on murine DC (25). Lanzavecchia et al. have reported the expression of the MR on monocyte-derived cultured human DC (26). They showed that this receptor plays an important role in Ag capture and transport of this captured Ag to the lysosomal/endosomal compartment for degradation and association with MHC class II. Recently, Noorman et al. demonstrated MR expression on branched DC in the dermis, but not on Langerhans cells in the epidermis (27). They suggest that MR expression may therefore be specific to immature DC and tissue macrophages. Whether peripheral blood DC express the MR, among which are cells with a more immature phenotype that corresponds to the phenotype of the NIPC, has not been reported.

To examine the role of a lectin on DC in the stimulation of IFN-α synthesis, we first assessed the ability of monosaccharides and mannan to inhibit viral stimulation of IFN production in mononuclear cells. Since the MR is reported to be on culture-derived DC, we also examined the ability of a polyclonal anti-MR antiserum to block the IFN synthesis induced by several viruses, and we assessed MR expression on freshly isolated DC. This study demonstrates that the MR is expressed by peripheral blood DC that produce IFN-α, and it plays a significant role in the induction of IFN-α synthesis by different viruses. The inhibition pattern of the monosaccharides examined also indicates that an additional receptor(s) with a galactose/N-acetylgalactosamine specificity may also be involved in this IFN induction.

d-Glucose (Glc), d-galactose (Gal), d-mannose (Man), N-acetylglucosamine (Glc-N-Ac), N-acetylgalactosamine (Gal-N-Ac), l-fucose (Fuc), and mannan (isolated from Saccharomyces cerevisiae) were purchased from Sigma (St. Louis, MO) and used at the concentrations indicated in the experiments. The goat anti-human MR antiserum (α-MR) was a gift from Dr. Philip Stahl (Washington University, St. Louis, MO). The normal goat serum (NGS) control was obtained from Atlanta Biologics (Norcross, GA). The α-MR and NGS were used at the dilutions indicated in the experiments. Phycoerythrin (PE)-conjugated mAb to the MR was obtained from PharMingen (San Diego, CA).

HSV-1 strain 2931 and VSV (originally obtained from Dr. Nicholas Ponzio of New Jersey Medical School) were grown and titrated by plaque forming assay on Vero cells as previously described (28). Sendai virus, strain Sendai/Cantell, was obtained from the American Type Culture Collection (Manassas, VA), grown in 10-day-old embryonated chicken eggs, and titrated by hemagglutination assay using chicken RBC. Allantoic fluid was used in all IFN inductions using Sendai virus. All virus stocks were stored at −70°C until use.

GM-0459A (National Institute of General Medical Sciences Human Genetic Mutant Cell Line Repository, Camden, NJ), a primary fibroblast cell line trisomic for chromosome 21, were grown in DMEM (JRH Biosciences, Lenexa, KS) supplemented with 15% FCS (HyClone, Logan, UT), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (DMEM, 15%). Vero cells (originally obtained from American Type Culture Collection) were grown in DMEM-10% FCS. H9 cells persistently infected with HIV-1 strain IIIB (HTLV-IIIB/H9; obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Disease, National Institutes of Health (Bethesda, MD), from Dr. Robert Gallo) were grown in RPMI 1640 (JRH Biosciences) supplemented with 10% FCS, l-glutamate, penicillin, and streptomycin (RPMI, 10%).

ELISPOT assays, used to determine the frequency of IPC, were performed as previously described (29, 30). Briefly, PBMC were isolated by Ficoll-Hypaque density centrifugation of heparinized venous blood obtained from healthy volunteers with informed consent. PBMC were suspended in RPMI-10% FCS and incubated with viruses at the concentrations described in the experiments for 6 to 8 h at 37°C in a 5% CO₂ incubator. Ninety-six-well microtiter plates with a nitrocellulose membrane bottom (Millititer HA plates, Millipore, Bedford, MA) were coated with ammonium sulfate-precipitated bovine anti-human IFN-α antiserum and fixed with 0.05% glutaraldehyde, and free aldehyde groups were blocked with 3% glycine in PBS. Stimulated or mock-stimulated PBMC were washed and added undiluted or in serial threefold dilutions to the Millititer plates and standard 96-well microtiter plates. The cells were incubated in the plates for 12 to 16 h at 37°C to allow production and capture of the IFN-α. Plates were developed using a murine anti-human IFN-α mAb (293 mAb, provided by Drs. Brita Cederblad and Gunnar Alm, Uppsala, Sweden) that cross-reacts with multiple IFN-α subtypes. This was followed by incubation with goat antiserum to murine IgG conjugated to horseradish peroxidase (The Jackson Laboratory, Bar Harbor, ME) and horseradish peroxidase substrate 3,3′-diaminobenzidine tetrahydrochloride (Sigma). The resulting reddish brown spots were enumerated using a dissecting microscope, and the frequencies of IPC were calculated and expressed as IPC per 10⁴ PBMC.

IFN bioassays were performed on supernatants harvested from standard microtiter plates using a cytopathic effect reduction assay with GM-0459A cells and VSV as a challenge virus as previously described (28). IFN-α reference standard (National Institute of Allergy and Infectious Disease standard G-023-901-527) was used at 100 IU/ml.

DC were isolated using a blood DC isolation kit (Miltenyi Biotec, Auburn, CA). Briefly, PBMC were resuspended in PBS with 0.5% BSA and 2 mM EDTA (Sigma) and were then labeled with haptenized murine Abs to CD3, CD11b, and CD16 followed by anti-hapten Ab attached to diamagnetic beads. The fraction of cells labeled with magnetic beads (monocytes, NK cells, and T cells) was separated from the nonmagnetic fraction (B cells and DC) using a magnetic separation column. The combined B cell and DC-enriched fraction was then further separated into a B cell and a DC-enriched fraction by labeling with anti-CD4 conjugated to diamagnetic beads. The CD4⁺ DC were then positively selected on a miniMACS RS⁺ magnetic separation column, and the nonmagnetic B cells were collected in the flowthrough. DC purity was assessed by flow cytometry, with DC defined as HLA-DR⁺ cells lacking expression of CD3, CD14, CD16, and CD19. Isolated DC were then used for experiments as described in Results.

PBMC were sorted into monocyte, NK cell, and T cell-enriched fractions; B cell- and DC-enriched fractions; B cell-enriched fractions; and DC-enriched fractions using the Miltenyi Biotec Blood DC Isolation Kit as described above. Total RNA was then extracted from a fixed number of cells for each population (1 × 10⁵ to 2 × 10⁵) using 0.8 ml of RNAzol (Tel-Test, Friendswood, TX) and 10 μl of microcarrier-GT (Tel-Test). The RNA pellets were dried and resuspended in exactly 10 μl of diethylpyrocarbonate-treated water. cDNA was generated from 2 μl of extracted RNA using Moloney murine leukemia virus reverse transcriptase and the specific downstream antisense primer. RT-PCR was performed according to the manufacturer’s instructions provided with the Perkin-Elmer/Cetus GeneAmp RNA PCR kit (Perkin-Elmer/Cetus, Norwalk, CT). The sequences of the MR-specific primers used were: sense, 5′-CCCTTCCTTGACTAATCC-3′; and antisense, 5′-ACCTCACCCTCC-ACTTATC-3′. This primer set was generated using the GCG software package (Genetics Computer Group, Madison, WI), and the primers were synthesized by the New Jersey Medical School Molecular Resource Facility (Newark, NJ). The sequences of the β-actin-specific primers used were: sense, 5′-GTGGGGCGCCCCAGGCACCA-3′; and antisense, 5′-GTCCTTAATGTCACGCACGATTTC-3′. This primer set was synthesized by Operon Technologies (Alameda, CA). The PCR reactions for the MR and β-actin were performed in an 50-μl total reaction volume using the Applied Biosystems 9600 thermocycler (Applied Biosystems, Foster City, CA) and AmpliTaq Gold. cDNA was first kept at 94°C for 10 min to activateAmpliTaq Gold, and amplification was conducted to 40 cycles of 1 min at 94°C, 1 min at 53°C, and 1 min at 72°C per cycle. Forty microliters of each RT-PCR reaction product was then separated on a 2% agarose gel in TBE buffer, and the products (272 kDa for the MR, 520 kDa for β-actin) were visualized using ethidium bromide staining and a fluorimager (Molecular Devices, Sunnyvale, CA). The PCR product generated with the MR primers was sequenced using an ABI 373 Automated Sequencer (Applied Biosystems) and was found to be 99% homologous with the human MR (data not shown).

Monoclonal mouse anti-human IFN-α (293, a gift from Dr. G. Alm and B. Cederblad, Uppsala, Sweden) was purified from ascites fluid using Ultralink-immobilized protein A (Pierce, Rockford, IL) according to the manufacturer’s instructions. Biotinylation of 293 was performed using a previously described method (31). Briefly, a 1 mg/ml solution of 293 in sodium borate buffer (0.1 M, pH 8.8) was prepared by performing six buffer exchanges (∼10-fold dilution for each) using a Centricon 50 spin concentrator (Amicon, Beverly, MA). The Ab was then reacted with 250 μg of N-hydroxysuccinimide biotin ester (Sigma) dissolved in DMSO (10 mg/ml stock) for 4 h at room temperature. Twenty microliters of 1 M NH₄Cl was added, and the Ab was incubated for another 10 min at room temperature. The free biotin ester was then removed by performing another six buffer exchanges (∼10-fold dilution for each) using PBS and a Centricon 50 concentrator. Ab was then stored at 4°C with 0.1% sodium azide until use.

PBMC were prepared for intracellular detection of IFN-α by flow cytometry using a modification of methods previously described by Svensson et al. (10). Briefly, PBMC at 2 × 10⁶ cells/ml were stimulated with HSV at 1 × 10⁶ pfu/ml or mock stimulated for 5 h at 37°C in 5% CO₂. The cells were then washed with PBS and labeled with Ab to the desired surface determinants or the appropriate isotype control at matched concentrations for 30 min at 4°C in 50 μl of PBS with 2% heat-inactivated human serum. The cells were again washed, and then fixed overnight at 4°C in 1% paraformaldehyde in PBS. Following fixation, the cells were washed twice in 0.1% BSA in PBS and then permeabilized for 1 h at room temperature in 0.1% BSA and 0.1% Tween-20 in PBS. The cells were pelleted and resuspended in the remaining volume (∼50 μl) of permeabilization buffer. Biotinylated 293 mAb to IFN-α was added to a final concentration of 1 μg/ml, and the cells were incubated for 30 min at room temperature. The cells were washed twice and resuspended in the remaining volume, and 10 μl of streptavidin conjugated to Quantum Red (QR; Sigma) was added. The cells were incubated for another 30 min, then washed once and resuspended in 300 μl of buffer. For all steps following fixation, the buffer used was 0.1% BSA and 0.1% Tween-20 in PBS. In some experiments, nonlabeled 293 Ab was used for intracellular staining of IFN-α. In these cases, a secondary PE-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) was used for analysis.

PBMC, stimulated and labeled as described above, were analyzed using the FACScan flow cytometer and CellQuest Analysis software (Becton Dickinson, San Jose, CA). PBMC were first gated according to scatter to include lymphocytes and monocytes within a single region (region 1), and then data were acquired for cell surface fluorescence (FITC and PE) and intracellular IFN-α (QR). Cells labeled with only a single fluorochrome were used to adjust compensation and eliminate overlap among the fluorescence measured by the different channels. IPC were then defined according to the QR fluorescence using mock-stimulated cells as a control. IFN-α-positive cells were then gated within a second region (region 2), and the cells within regions 1 and 2 were analyzed for surface fluorescence. The percentages of cells expressing the different surface Ags were then determined by comparison with the appropriate isotype control using the histogram subtraction method provided with the Becton Dickinson CellQuest software.

Experiments involving monosaccharides and mannan were performed using a randomized block design. The data were analyzed using a two-way ANOVA to account for subject to subject variation, and Student’s t test was used for post-hoc comparisons with adjustment of the p value by a Bonferonni correction. Statistical tests were performed using the JMP statistical software package (SAS Institute, Cary, NC). Curve fitting using the spline method was performed using the Deltagraph 4.0 graphing package (DeltaPoint, Monterey, CA). Statistical analysis of flow cytometric data was conducted by Kolmogorov-Smirnov statistics using Becton Dickinson CellQuest software.

We assessed the abilities of six different monosaccharides to reduce the frequency of HSV-induced IPC in a dose-dependent manner (Fig. 1,A). In this representative experiment, most sugars had inhibitory effects at high concentrations (> 50 mM). However, Fuc had the greatest inhibitory effect (IC₅₀ = 25 mM), followed by Gal-N-Ac and Glc-N-Ac (IC₅₀ ≅ 37 mM). Since the specificity of monosaccharide/lectin interaction is reduced at high concentrations, we used the sugars in subsequent experiments at a 50-mM concentration. This concentration was within the linear range of inhibition for all sugars and corresponded to the approximate IC₅₀ of the Glc and Gal. As shown in Figure 1,B, Fuc was the most inhibitory monosaccharide reducing the frequency of HSV responsive IPC by 70%, followed by Gal-N-Ac (60%), Glc-N-Ac (49%), and mannose (31%). The monosaccharides also decreased the amount of total IFN produced in response to HSV in a pattern reflecting the frequency changes; the inhibition was reversible, and cell viability after the 6-h induction period was >99% by trypan blue exclusion (data not shown). In contrast to HSV, only Fuc inhibited the induction of IFN by Sendai virus, yielding a fivefold reduction (Fig. 1 C). Since no receptor with specificity for both Glc-N-Ac/Man and Gal-N-Ac has been identified, these results suggest either a novel receptor with both mannose and galactose specificities, or that more than one receptor with different specificities are involved in the stimulation of IFN-α synthesis by HSV and HIV. Furthermore, the sensitivity of Sendai virus to only Fuc suggests that Sendai virus stimulates IFN synthesis via a unique receptor or mechanism.

Since the results of the monosaccharide inhibition studies supported a role for lectin receptors in the induction of IFN synthesis by enveloped viruses, the effect of the yeast cell wall polysaccharide, mannan, on HSV-induced IFN production in PBMC was examined. Mannans from a variety of sources are reported to be good ligands for mannose binding proteins such as the human MR and serum mannose binding protein (21, 32). In data from five donors (Fig. 2,A), mannan achieved a mean inhibition of approximately 35% at 1 mg/ml; however, this inhibition was not quite statistically significant (p = 0.0578), most likely due to the large variability in the data. This donor variability is further exhibited by the results of two donors shown in Figure 2 B in which mannan reduced the IPC frequency by 40% in one donor and by >90% in another. The lack of complete inhibition by mannan in most cases further supports that two or more lectin receptors with different monosaccharide specificities are involved in the induction process.

To further delineate the receptors involved in recognition and induction of IFN-α expression by different enveloped viruses, the ability of an antiserum specific for the human MR (α-MR) to block virus-induced IFN production was assessed. Figure 3,A shows the dose-dependent effects of anti-MR on the frequency of HSV-responsive IPC. The antiserum significantly reduced the IPC frequency from the 1/50 to 1/200 dilution in a dose-dependent manner (60–80%) as well as the total IFN produced at the 1/50 dilution (Fig. 3,B), supporting a role for the MR in the stimulation of IFN-α synthesis. Since HIV and VSV display sensitivities to monosaccharide inhibition that are similar to that of HSV but that differ from that of Sendai virus, we examined the blocking effects of α-MR on HIV, VSV, and Sendai virus IFN induction. Results from representative experiments shown in Figure 4 demonstrate that IFN induction by HIV and VSV is blocked by anti-MR similar to HSV, whereas Sendai virus-induced IFN production was unaffected by α-MR. These data further substantiate a common role for the MR in IFN-α induction by different enveloped viruses in PBMC. It also supports prior studies indicating a separate IFN induction mechanism for Sendai virus that may correlate with the production of IFN-α predominantly by a CD14⁺ monocyte that lacks expression of the mannose and Gal/Gal-N-Ac receptors.

Although a prior study by Sallusto et al. demonstrated MR expression on cultured DC derived from CD14⁺ monocytes (26), this receptor has not been reported on DC freshly isolated from blood. RT-PCR and flow cytometry were therefore used to investigate whether freshly isolated peripheral blood DC also express the MR. We have used the Miltenyi dendritic cell isolation kit for these studies. This kit uses negative depletion of T cells, NK cells, and monocytes followed by positive selection of CD4-expressing cells. Typical yields for this procedure are 5 × 10⁵ enriched DC from 10⁸ cells. DC are identified as cells that are MHC class II (HLA-DR) positive but negative for lineage-specific markers of T cells, B cells, NK cells, and monocytes. As expected, the enriched DC are immature, in that they do not express costimulatory molecules or CD83, but up-regulation of both CD80 and CD83 are seen after overnight culture at 37°C (A. Izaguirre and P. Fitzgerald-Bocarsly, in preparation). Wright-Giemsa staining of cytospins of the freshly isolated, enriched DC revealed a homogeneous population of cells that are monocytoid in appearance and size and that lack surface dendritic projections or abundant granules (data not shown), a morphology identical with that previously described for peripheral blood DC and IPC (9, 10, 33).

RT-PCR with primers specific for the MR was performed using PBMC and cells from each of the steps in the DC isolation protocol. A low level expression of MR mRNA was observed in unfractionated PBMC (Fig. 5,A). A similar level of expression was observed in cells enriched for T cells, NK cells, and monocytes, but depleted of B cells and DC, and in cells enriched for B cells and DC (∼10-fold enriched for DC). The cell population highly enriched for DC displayed much higher expression compared with the other cell fractions (Fig. 5 A, lane 5), demonstrating that the MR is primarily expressed on freshly isolated DC. The low level expression observed in the DC-depleted fraction (i.e., T cell-, NK cell-, and monocyte-enriched fraction) may signify contamination of this fraction by DC, since the T cell-, NK cell-, and monocyte-enriched fraction is often only partially depleted of IPC responding to HSV (data not shown).

MR expression was also assessed at the protein level. At least 35% of enriched DC expressed the MR as determined by flow cytometry using mAb specific for the MR (Fig. 5 D) at levels similar to those seen with cultured macrophages (27).

While DC express the MR, only a fraction of these DC may be IPC. We therefore assessed MR expression on IPC using intracellular flow cytometry for IFN-α combined with surface analysis for MR expression or the binding of an appropriate isotypic control (10). For the intracellular flow studies we used intact PBMC that were stimulated with HSV for 5 h at 37°C and then analyzed the phenotype of the cells positive for intracellular IFN-α as has been previously described by Svensson et al. (10). We chose to use PBMC for these studies to get an accurate phenotypic profile of the IPC. Moreover, work by Cederblad et al. has suggested that for optimal production of IFN-α, the HSV-responsive IPC require cytokines or growth factors that are produced by other cells (34), a finding we have confirmed in our laboratory. Figure 6 demonstrates that IPC can be detected by flow cytometry, and they have high forward light scatter, but low side scatter as previously reported by Svensson et al. using this technique (10), a scatter profile that is characteristic of DC. Typically, 0.1 to 0.25% of the PBMC were found to express intracellular IFN-α following stimulation with HSV-1 (for Fig. 6, 614 cells were positive for IFN-α from a total of 350,000 analyzed cells (0.18%)), numbers consistent with those obtained by ELISPOT determination for HSV-responsive IPC (10, 30). The IFN labeling was blocked by an excess of exogenous rIFN-α2 (Fig. 6, E and F), demonstrating that this technique is specific for IFN-α. In addition, these cells express HLA-DR and CD4, but have low to absent expression of CD8 and CD33 (Fig. 7), similar to the pattern previously described on DC (35) and HSV-stimulated IPC (10). Examination of the MR expression of the IPC (Fig. 8) demonstrated that approximately 50% of IPC express the MR (p < 0.001 compared with isotype control), but the MR is absent from monocytes and lymphocytes (Fig. 8, D and E, respectively) as previously reported (21). These data therefore provide the first evidence that, similar to culture-derived DC (26), peripheral blood DC and IFN-α-producing DC express the MR.

Adaptive humoral and cellular immune responses are essential for the clearance and ultimate resolution of viral infections. Numerous studies, however, clearly show the importance of innate, nonspecific host defense mechanisms in viral resistance (36, 37, 38). The IFN system was one of the first antiviral resistance mechanisms identified. Unlike adaptive immunity, which uses remarkably specific receptors and circulating factors to recognize highly variable Ags such as proteins, the receptors and factors of innate immunity react more broadly to less variable structures. One of the most fundamental modes of recognition in the innate immune system is carbohydrate recognition using lectins. Lectin receptors exist on most, if not all, cells of the mononuclear phagocyte system mediating endocytosis and phagocytosis (21, 39). Recently, culture-derived DC were also shown to express a lectin receptor, the MR, that they use to capture and concentrate Ags for processing and presentation (26).

In the current study, we have begun to explore the role of lectin receptors in viral stimulation of IFN-α production by NIPC. We observed that the frequency of IPC responding to HSV could be decreased to various degrees by monosaccharides known to block mannose and galactose binding receptors. Large quantities of sugar were required to achieve appreciable inhibition, but these quantities were most likely necessary due to the low affinity binding (K_d = ∼10⁻³ M) of monosaccharides for the carbohydrate recognition domains of lectins compared with the high functional avidity that occurs with binding to polysaccharide ligands (40). In addition, these concentrations were similar to those used to study lectin receptors on NK cells (41, 42). No decrease in cell viability occurred, and the monosaccharide inhibition was reversible, arguing against a direct cytotoxic effect of these high monosaccharide concentrations.

Our results with the polysaccharide, S. cerevisiae mannan, demonstrate that this substance can also block the induction of IFN-α by HSV. However, the inhibition observed with mannan was highly variable, with complete inhibition rarely observed. These results as well as the inhibition observed with monosaccharides that block both mannose and galactose binding lectins support a DC system with heterogeneity in the expression of two receptors with differing specificities. Nevertheless, the possibility that a receptor exists on DC that binds ligands with both specificities cannot be excluded.

The studies with monosaccharides and mannan lead us to hypothesize that viruses may induce IFN through interaction with the MR. We evaluated this possibility and demonstrated that the MR is expressed on a portion of freshly isolated DC, and approximately half of the cells among PBMC that stained positively for IFN-α (following stimulation with HSV-1) expressed the MR, a result that was highly significant. Furthermore, polyclonal antiserum to the MR was found to significantly decrease the frequency of IPC responding to HSV-1 as well as the amount of IFN-α generated during the viral induction. In future studies, it will be interesting to isolate and compare the MR-expressing and nonexpressing IPC.

Since VSV and HIV stimulate the same population of IPC as HSV (4), we hypothesized that a common receptor may also be involved in induction of IFN-α by these viruses. We evaluated the ability of the anti-MR antiserum to interfere with IFN induction by VSV, HIV, and Sendai virus. As hypothesized, the antiserum decreased the frequency of IPC when induced with VSV and HIV. Although CD4 and chemokine receptors are known to be the main receptors for HIV, the binding of gp120 to a mannose binding lectin has previously been reported: HIV was shown to bind in a CD4-independent manner to a cell-associated mannose binding lectin in human placenta (43). To date, we have only used HIV-infected H9 cells in our studies of blocking of IFN-α production by anti-MR Abs. Steinman’s group has reported the interaction of HIV-1 with dendritic cells occurring via multiple chemokine coreceptors (44). It will be interesting to determine whether IFN induction by both T cell- and monocyte-tropic HIV strains are equally blocked by the anti-MR Ab and whether chemokine receptors are also involved in this process.

Interestingly, the anti-MR antiserum failed to inhibit IFN-α induction by Sendai virus. This finding is consistent with previously reported differences between IPC responding to Sendai virus and most other enveloped viruses, in that Sendai virus induces a high frequency IPC response that is predominantly monocytic compared with the low frequency DC response observed with most viruses (4, 5). This correlates well with the results of the MR expression studies: DC express the MR, but monocytes do not. Recently, we have also observed that two compounds, chloroquine and bafilomycin A1, inhibit the stimulation of IFN-α production by viruses that stimulate DC, but they do not affect IFN-α induction by Sendai virus (our unpublished data). Both drugs raise endosomal pH by independent mechanisms (45, 46), and their inhibitory effects may relate to interruption of normal lysosomal functioning or MR recycling through this compartment of the cell. Since Sendai virus infection can occur by extracellular envelope fusion and its induction of IFN is not inhibited by blocking the MR, these results may reflect the unique ability of this virus to stimulate monocytes.

The MR plays a role in the induction of IFN-α production by DC; however, some evidence suggests that additional mechanisms act in combination with or independently of the MR. Most obvious is the inhibition observed with Gal-N-Ac. This sugar poorly blocks the MR, but is an excellent ligand for lectins with Gal/Gal-N-Ac specificity (47). Recently, a human receptor that binds Gal and Gal-N-Ac has been cloned in humans that is expressed on macrophages (23), and it will be interesting to examine its expression on DC.

The precise function of the MR in the induction of IFN-α remains to be elucidated. One plausible scheme involves binding of the virus using the MR or other receptors. Following binding, the virus is internalized, resulting in either productive or nonproductive infection. The stimulus for IFN-α synthesis may then be derived from the viral nucleic acid. This model is supported by evidence that IFN induction correlates with virulence in some systems, and both dsRNA and DNA containing certain palindromic sequences can induce IFN (48). Some evidence, however, suggests that viruses may be capable of inducing IFN by simply binding to a surface receptor that can transduce a signal for IFN-α synthesis independent of infection (15, 49). Again, the MR may be a candidate receptor given its ability to transduce a mitogenic signal in alveolar smooth muscle cells in response to the appropriate ligand (50).

Finally, lectin receptors may serve as a universal, nonspecific means by which DC recognize many different viral and nonviral pathogens. Following binding and internalization, these pathogens would be shuttled to the MHC class II-rich endosomal compartment of DC for Ag processing and ultimately presentation (26), and some of these pathogens may then induce the production of IFN-α and other mediators by the DC. The IFN may serve as a local antiviral substance protecting neighboring cells and activating NK cells and macrophages (38, 51). However, the IFN-α would also serve as an important modulator of T cell activation and maintenance of T cell memory during the presentation process. The interaction of virus with lectin receptors on DC may therefore serve as a critical link between the innate and adaptive immune responses to the myriad of pathogens encountered.

We gratefully acknowledge Dr. Philip Stahl (Washington University, St. Louis, MO) for the gift of anti-MR Abs and for helpful discussions, Drs. Brita Cedarblad and Gunnar Alm (Uppsala, Sweden) for the mAbs to IFN-α, and Wellcome-Glaxo for providing polyclonal antiserum to IFN-α. We also acknowledge Drs. Sheela Amrute and Robert Donnelly of the New Jersey Medical School for the sequencing of the MR RT-PCR product, and Dana Stein of New Jersey Medical School for assistance with flow cytometry.