Dendritic Cell Maturation Results in Pronounced Changes in Glycan Expression Affecting Recognition by Siglecs and Galectins¹ Free

Dendritic cells (DC)⁴ are the most potent APC in the immune system. They reside as immature DC (iDC) in the peripheral tissues, where they sense for pathogens (1). Pathogen recognition often results in the activation of iDC via TLR present on their membrane or in intracellular compartments (2). The interaction of TLRs with their ligands elicits a complex signaling cascade leading to the migration of DC to the neighboring lymph nodes. In these lymph nodes, the DC arrive as mature DC (mDC), ready to interact with a naive lymphocyte carrying the appropriate TCR. mDC are characterized by the expression of high levels of MHC class II, costimulatory molecules, chemokines, and cytokines, in contrast to iDC, which show low expression of these molecules and high levels of Ag-uptake receptors. Thus, DC suffer a dramatic change in phenotype and functionality upon maturation. This change is mediated by the modulation of a wide array of molecules and ensures the development of a potent and specific immune response. As a result, mDC have a limited Ag uptake and processing capacity, whereas Ag presentation and T cell costimulation are promoted (3).

To limit Ag uptake, C-type lectin receptor (CLR) expression on the cell surface of DC is down-regulated during maturation. CLRs constitute an important family of pattern recognition receptors expressed on DC (4) and are well-characterized as Ag-uptake receptors for glycosylated structures (5) found on pathogens. Except for the recognition of pathogens, CLR have been implicated in several other functions, such as cell migration (6) and intercellular communication (7). One of the best studied CLRs is the DC-specific ICAM-3-grabbing nonintegrin, also known as DC-SIGN (8). Besides a pattern recognition receptor for HIV-1 (9), CMV (10), Schistosoma mansoni (11), and other pathogens (5, 12), DC-SIGN has been shown to also recognize glycan epitopes on endogenous ligands, such as ICAM-2 on endothelial cells (6), ICAM-3 on T cells (8), or Mac1 on neutrophils (13). Interaction of DC-SIGN with its ligands regulates DC precursor migration into peripheral tissues, stabilizes the DC-T cell contact surface in the immunological synapse (14), and allows neutrophils to induce DC maturation, respectively. Another CLR involved in intercellular communication is the macrophage galactose-type C-type lectin, which binds to specific CD45 glycoforms on T cells (7) to down-regulate effector T cell functions. Other members of the lectin superfamily that play an important role in the immune system are siglecs (15) and galectins (16). The sialic acid-binding Ig superfamily lectins (siglecs) are a class of Ig superfamily proteins which show binding activity to specific glycan structures containing sialic acid (17). Many siglecs have molecular features of inhibitory receptors, including conserved tyrosine-based motifs. This is the case of siglec-2 (CD22), involved in the control of the BCR signaling (18), and the siglec-3 (CD33) related siglecs, known to relay inhibitory signals or to inhibit activatory pathways when cross-linked with activating receptors (19, 20, 21, 22). Galectins have a binding specificity toward β-galactose-containing glycoconjugates (23). To date, 15 members have been characterized and although all galectins require β-galactosides for binding, several structural and functional differences have been described. In the immune system, galectins have been shown to operate at different levels in innate and adaptive immune responses by modulating cell survival and cell activation or by influencing the Th1/Th2 cytokine balance (24).

Classically, CLRs, siglecs, and galectins on immune cells have been studied considering that regulation of their function is achieved by the modulated expression of the receptor. However, their ligands may also be regulated, adding another layer of complexity to the system. Importantly, whereas information about the expression of these receptors is widely available, only recently has the regulation of the glycosylation in cells of the immune system received more attention. Recent studies have shown that activation by cytokines results in profound changes in the glycan profile of T cells (25) and on endothelial cells (26). In the case of DC, maturation has been shown to be accompanied by a decrease in α2,6-sialylation (27), however, a more extensive biochemical and functional characterization of this important immune cell is still lacking.

In this study, we have investigated the changes in the glycosylation of DC that are associated with maturation with regards to the glycosylation machinery. We evaluated changes in glycosylation-related gene expression by quantitative real-time RT-PCR and a glycosylation-related gene expression microarray. Mass spectrometric (MS) data were used as well as lectin binding to DC to confirm the data of the RT-PCR and microarray. In this study, we show that maturation of DC results in large changes in the expression of glycosylation-related genes, involving fucosyltransferases, galactosyltransferases, and sialyltransferases. This results in a high expression of LacNAc structures, sialylated glycans, and Lewis structures. We further demonstrated that increased expression of these glycans result in binding epitopes for siglecs and galectins.

Monocytes were isolated from buffy coats of healthy blood donors (Sanquin) through Ficoll gradient centrifugation and positive selection of CD14⁺ cells using MACS sorting (Miltenyi Biotec). Isolated monocytes were cultured in RPMI 1640 (PAA Laboratories) supplemented with 10% FCS (BioWhittaker), 10,000 U/ml penicillin (BioWhittaker), 10,000 U/ml streptomycin (BioWhittaker), and 10,000 U/ml glutamine (Sigma-Aldrich) in the presence of IL-4 (500 U/ml; Schering-Plough) and GM-CSF (800 U/ml; Schering-Plough) for 7 days (28). Maturation was induced 6 days after isolation by incubation of iDC with 10 ng/ml LPS (from Salmonella typhosa; Sigma-Aldrich) for 24 h. DC maturation was assessed by flow cytometric determination of the maturation markers MHC class II, CD80, CD83, and CD86.

mRNA from iDC and mDC was specifically isolated by capture of poly(A⁺) RNA in streptavidin-coated tubes using a mRNA Capture kit (Roche). cDNA was synthesized using a Reverse Transcription System kit (Promega) following the manufacturer’s guidelines. Cells (0.5 × 10⁶) were washed twice with ice-cold PBS, pelleted, and lysed in 500 μl of lysis buffer. Lysates were incubated with biotin-labeled oligo(dT)₂₀ for 5 min at 37°C and then 50 μl of the mix was transferred to streptavidin-coated tubes and incubated for 5 min at 37°C. After washing three times with 250 μl of washing buffer, 30 μl of the reverse transcription mix (5 mM MgCl₂, 1× reverse transcription buffer, 1 mM dNTP, 0.4 U of recombinant RNasin RNase inhibitor, 0.4 U of reverse transcriptase, 0.5 μg of random hexamers in nuclease-free water) were added to the tubes and incubated for 10 min at room temperature followed by 45 min at 42°C. To inactivate AMV reverse transcriptase and separate mRNA from the streptavidin-biotin complex, samples were heated at 99°C for 5 min, transferred to microcentrifuge tubes and incubated in ice for 5 min, diluted 1/2 in nuclease-free water, and stored at −20°C until analysis.

Oligonucleotides were designed by using the computer software Primer Express 2.0 (Applied Biosystems), synthesized by Invitrogen Life Technologies, and are published elsewhere. PCR were performed with the SYBR Green method in an ABI 7900HT sequence detection system (Applied Biosystems). The reactions were set on a 96-well-plate by mixing 4 μl of the two times concentrated SYBR Green Master Mix (Applied Biosystems) with 2 μl of a oligonucleotide solution containing 5 nM/μl of both oligonucleotides and 2 μl of a cDNA solution corresponding to 1/100 of the cDNA synthesis product. The thermal profile for all the reactions was 2 min at 50°C, followed by 10 min at 95°C and then 40 cycles of 15 s at 95°C and 1 min at 60°C. The housekeeping gene GAPDH was used as endogenous reference (29). To calculate the relative abundance of the genes, the formula 100 × 2^{(Ct GAPDH-Ct glycosyltransferase)} was used, where Ct is the cycle threshold. In this formula, the Ct value is defined as the number of PCR cycles at which the SYBR-green fluorescent signal exceeds the threshold of 0.2 relative units (26).

iDC and mDC were incubated with FITC- or biotin-labeled lectins (10 μg/ml) for 45 min at room temperature in TSM (20 mM Tris (pH 7.4), 150 mM NaCl, 1 mM CaCl₂, and 2 mM MgCl₂) supplemented with 0.5% albumin from bovine serum (Fluka Biochemika). After washing the cells with TSM, FITC-labeled streptavidin was added to the cells incubated with biotin-labeled lectins for 30 min at room temperature. After another washing step, cells were analyzed by immunofluorescence by collecting data for 10⁴ cells per histogram (FACSCalibur; BD Biosciences). Corresponding negative controls were performed using the FITC-labeled streptavidin alone.

Core C of the Consortium for Functional Glycomics performed experiments to profile the N-glycans of iDC and mDC. The procedure was performed as previously described (25). N-glycans were released from extracted glycoproteins of cell preparations by peptide N-glycosidase (PNGase F) digestion and O-glycans were chemically released by reductive elimination from the glycopeptides remaining after the release of N-glycans. Released glycans were permethylated using the sodium hydroxide procedure, and purified on a Sep-Pak C 18 cartridge, as previously described (30). Derivatized glycan samples were dissolved in methanol/water 8:2 (v/v) and mixed in a 1:1 ratio with 10 mg/ml 2,5-dihydroxybenzoic acid in 80:20 (v/v) methanol/water. A total of 0.5- to 1-μl aliquots were spotted onto a target plate and dried under vacuum. MS spectra were obtained using a Voyager DE STR MALDI-TOF (Applied Biosystems) mass spectrometer in the reflectron mode with delayed extraction. Peaks observed in the MS spectra were selected for further MS/MS. MS/MS data were acquired using a 4800 MALDI TOF/TOF (Applied Biosystems) mass spectrometer. The potential difference between the source acceleration voltage and the collision cell was set to 1 kV and argon was used as collision gas. The 4700 Calibration Standard kit, calmix (Applied Biosystems), was used as the external calibrant for the MS mode and [Glu¹]fibrinopeptide B human (Sigma-Aldrich) was used as an external calibrant for the MS/MS mode.

RNA from iDC and mDC of three different donors was extracted using TRIzol reagent according to the manufacturer’s protocol (Invitrogen Life Technologies). Total RNA was treated with DNase (Ambion), and purified using the RNeasy kit (Qiagen). The RNA was amplified and biotin labeled with the Bioarray High Yield RNA transcript labeling kit (Enzo Life Sciences). Hybridization and scanning of the glycogene-chip v3, designed for the Consortium for Functional Glycemics, were performed according to Affymetrix’s recommended protocols. The expression data obtained with this chip array contain the expression of ∼2000 human and mouse transcripts relevant to glycosylation. A complete description for the GLYCov3 array is available at www.functionalglycomics.org/static/consortium/recources/resourcecoree.shtml.

Siglec-Fc chimeras were used as tissue culture supernatants derived from stably transfected Chinese hamster ovary cells at 10 μg/ml and were incubated for 45 min on room temperature with FITC-labeled anti-human-Fc Abs 1:200 (Jackson ImmunoResearch) in TSM supplemented with 0.5% albumin. After incubation, flow cytometric analysis using a FACSCalibur flow cytometer and CellQuest software (BD Biosciences) was performed. As a negative control, cells were treated with 10 μl of Vibrio cholerae neuraminidase (Boehringer) in 100 μl of PBS for 60 min at 37°C.

Galectins-3, -4, and -8 were produced as recombinant proteins in Escherichia coli and FITC labeled as previously described (31, 32). DC were incubated in HBBS (Invitrogen Life Technologies) medium containing 0.5 M lactose 30 min at room temperature to remove membrane-bound endogenous galectins. Subsequently, cells were incubated with the FITC-labeled galectins in HBBS (Invitrogen Life Technologies) supplemented with 1% BSA and 2 mM 2-ME. Cells were washed with TSM and immunofluorescence analyzed using a FACSCalibur flow cytometer and CellQuest software (BD Biosciences).

iDC are located in the peripheral tissues, where they establish multiple interactions with surrounding cells, many of which are expected to be carbohydrate mediated. Although much is known about the lectins expressed by iDC, their glycans and glycosylation machinery have been poorly described. The glycosylation machinery consists of a set of enzymes, transporters, regulatory molecules, cofactors, sugar donors, and other molecules present mainly in the endoplasmic reticulum and the Golgi apparatus that are implicated in the biosynthesis of glycans. The expression level of the different components of the glycosylation machinery has been reported to correlate with the array of glycans expressed by a certain cell (26, 33) and, therefore, can be used to establish a prediction of the glycan profile. To account for interindividual variability, microarray analysis was performed on RNA extracted from monocyte-derived iDC generated according to standard methods (28) from three unrelated healthy donors. Raw data files for each of the experiments performed are available at the Consortium for Functional Glycomics website (www.functionalglycomics.org/fg).

The transcript levels of 147 of the 258 glycosylation-related genes measured were detected as present, according to the microarray internal controls (34) (Fig. 1, supplementary table Ia).⁵ The transcripts marked as present were ranked and clustered according to their relative expression level in high (first quartile), low (third quartile), and moderately expressed (rest), as shown in Fig. 1. To facilitate analysis, the transcripts were subdivided according to their function in HKG (involved in the synthesis of core structures, sugar donors, or transporters), Term (glycosylation related-genes implicated in the terminal modification of glycans), and GAG (involved in the synthesis of glycosaminoglycans). As expected, HKG transcripts showed the highest expression levels, constituting the majority (80%) of the first quartile, while GAG and Term were widely distributed over the first quartile and the moderately expressed gene cluster. Expression of genes involved in glycosaminoglycan biosynthesis was markedly reduced as compared with terminal glycosyltransferases, with only two genes standing up, NDST2 and CSGlcAT, involved in the synthesis of heparan and chondroitin sulfate. Interestingly, six transcripts encoding for enzymes involved in terminal modifications of glycans were found in this cluster, namely: ST6Gal-1, MGAT4B, β3GnT-5, β4GalT-1, and ST3Gal V. The high transcript levels found for these genes predicts an abundance of triantennary N-glycans (MGAT4B), type 2 chains (β4GalT-1), α2,6-sialylated structures (ST6Gal-1), and glycolipids of the lacto- (β3GnT-5), and ganglioseries (ST3Gal-V). The range of expression levels for moderately expressed transcripts (42–206, coefficient of variation = 5.5%) was narrower compared with that of highly expressed transcripts (206–1075, coefficient of variation = 10.2%), indicating that the regulation of the glycosylation machinery may be more sensitive to subtle changes in the levels of moderately expressed genes. The transcripts found in the intermediate cluster (Fig. 1) suggest the presence of α3-fucosylated and 6-O-sulfated structures, as well as α2,3-sialylated and polysialylated glycans. The relative expression levels of glycosylation-related genes involved in the terminal modification of glycans was further confirmed by quantitative real-time PCR using a set of primers (26) to assay for 75 transcripts, including Gal-, GlcNAc-, Fuc-, sialyl-, and sulfotransferases (data not shown).

Two pools of iDC from at least five unrelated donors were subjected to glycan profiling by MALDI-TOF MS analysis. Portions of representative MALDI-TOF MS profiles of iDC with the most likely structures based on glycan compositions, MS/MS fragmentation patterns and gas chromatography-MS linkage analysis, are shown in Tables I and II. The MALDI-TOF MS spectra of N-glycans showed high mannose (m/z 1580.4, 1784.5, 1988.7, 2192.9, and 2397.0) and complex type glycans, the latter comprising bi-, tri-, and tetra-antennary structures which are mono-, di-, tri-, and tetrasialylated. Some of the complex N-glycans also carry fucose (compatible with (s)Le^X) and/or poly-N-acetyllactosamine (m/z 2432.0–5215.7, Table I), as would be predicted from the gene expression data. The O-glycan profile (Table II) demonstrated that the most abundant glycan species is sialylated core 1 (m/z 895.5). Rigorous MS/MS analyses indicated that the sialic acid can be present either attached to the Gal or GalNAc residue of the core 1 structure (data not shown). Disialylated core 1 structures (m/z 1256.8) as well as unsialylated and sialylated core 2 glycans are also present (m/z 983.6, 1344.8, and 1706.0).

Upon Ag capture, DC undergo a maturation process characterized by the up-regulation of molecules involved in migration, Ag presentation, and costimulation (3). Activation of iDC with the TLR4 ligand LPS has been classically used as a model for the study of DC maturation (35). To investigate whether the maturation process also affects the glycosylation of DC, we assessed the expression of glycosylation-related genes in DC incubated in the presence or absence of LPS (10 ng/ml, 24 h).

An initial screening by a glycogene-oriented microarray showed that the expression levels of the majority of HKG transcripts were not significantly affected by maturation, while a large group of GAG transcripts was down-regulated and many Term transcripts were up-regulated (Fig. 2,A), indicating that GAG biosynthesis might be turned down during maturation while dramatic changes might occur on the terminal modifications of N- and O-glycans. This was confirmed by gene expression profiling by quantitative real-time PCR performed on a set of glycosylation-related genes selected from the list of up- and down-regulated Term transcripts in the glycogene microarray. Based on the modulation of the different Term transcripts (Fig. 2,B), the maturation process is predicted to be accompanied by an increase in poly-N-acetyllactosamine chains (up-regulation of β3GnT-2 and β4GalT-4), decrease in N-glycan branching (down-regulation of MGAT-4A/B), increase in α3-fucosylation (FUT4), increase in α2,3 sialylation (ST3Gal-4/6) and α2,8-sialylation (ST8Sia-4), a decrease in α2,6-sialylation (ST6Gal-1), and an increase in GlcNAc 6-O sulfation (GST-5). Sialylated core 1 O-glycans are expected to increase their abundance with respect to core 2 O-glycans due to the concomitant up-regulation of core 1 β3GalT and ST3Gal-1/2 and the down-regulation of C2GnT-1 (Fig. 2 B).

Several other models have been used in the literature to induce DC maturation, based on the targeting of different TLR (36). To examine whether the modulation of glycosylation-related genes observed was purely dependent on TLR4 signaling or was due to a more general maturation-dependent effect, the expression of a set of 75 glycosylation-related genes was assessed in poly I:C (TLR3) or flagellin (TLR5) treated DC. Whereas LPS induced the strongest modulatory response in the expression of glycosylation-related transcripts, maturation induced by TL3 or TLR5 had comparable effects (data not shown).

Two pools of mDC from at least five unrelated donors were subjected to glycan profiling by MALDI-TOF MS analysis and compared with the profile obtained from iDC (Tables I and II). The N-glycan profile of mDC comprised of high mannose (m/z 1580.7, 1785.0, 1989.2, 2193.4, and 2397.7) and complex bi-, tri- and tetra-antennary glycans, which were sialylated, fucosylated, and contained poly-N-acetyllactosamine structures (m/z 2070.8 to 5575.1). Studies have demonstrated that relative quantitation based on signal intensities of permethylated glycans analyzed by MALDI-TOF MS is a reliable method, especially when comparing signals over a small mass range within the same spectrum (37). The mDC N-glycan profile exhibited an increase in antennal fucosylation relative to the iDC. This is exemplified by comparing the relative intensities of fucose-containing glycan signals, for example, the ratio of m/z 2606.9 to 2780.3, which correspond to a mono- (NeuAcFucHex₅HexNAc₄) and difucosylated (NeuAcFuc₂Hex₅HexNAc₄) biantennary structure, respectively (Table I), was altered from 7:1 in iDC to 2:1 in mDC. Also, the ratio of signals at m/z 3056.4 (monofucosylated triantennary N-glycan) to 3230.6 (difucosylated triantennary N-glycans) was altered from 10:1 in iDC to 3:1 in mDC. This observation was further supported by an increase of 3,4-linked GlcNAc in mDC compared with iDC in the gas chromatography-MS linkage analyses on partially methylated alditol acetates (data not shown). Therefore, an increase in the expression of FUT4 is consistent with the increase in α3-fucosylation observed from these data. In addition, an increase in N-glycan structures with poly-N-acetyllactosamine chains is also observed (Table I). The abundance of signals above m/z 3956.4, which from compositions must contain at least one LacNAc repeat, has increased in the mDC compared with iDC. For example, there is a relative increase in the flanking signals (e.g., m/z 4678.1, NeuAc₃FucHex₈HexNAc₇ and 4767.2, NeuAc₂FucHex₉HexNAc₈) of m/z 4591.0, NeuAc₂Hex₈HexNAc₇.

The 2- to 3-fold decrease in the expression of MGAT4A/B could not be correlated with a decrease in N-glycan branching, because the majority of the complex-type N-glycans species observed in iDC and mDC were of the tetra-antennary type (Table I). This apparent paradox could be explained by the high basal levels observed for MGAT4A/B, among the highest for the group of branching GlcNAc-transferases (supplementary table Ia). Also, the enzymes they encode for have been recently shown to have the highest kinetic efficiency for the triantennary N-glycan product of the GnT-II and GnT-V reactions, which indicates that the preceding branch formations on the Manα1-6 arm by other GnTs promote the actions of both GnT-IV enzymes (38).

In the O-glycan spectrum (Table II), the sialic acid on the signal at m/z 895.5 is exclusively attached to the Gal residue of the core 1 O-glycan structure (data not shown). This result is consistent with the increased expression of ST3Gal-1 and ST3Gal-2 and a decrease in ST6Gal-1 expression in mDC compared with iDC.

Both the glycosylation-related gene expression profile and the glycan profile demonstrated an increase in poly-N-acetyllactosamine-elongated glycans (see Fig. 4,A) and α2,3/α2,8 sialylated structures (Fig. 3,A), suggesting that binding of galactose- and sialic acid-specific lectins, such as galectins and siglecs, might be affected during maturation. The binding of sialoadhesin (siglec-1), CD22 (siglec-2), siglecs-3, -5, -7, -9, and -10 and galectins-3, -4, and -8 to iDC and mDC was tested by flow cytometry. Only sialoadhesin, CD22 and siglec-7 bound with high affinity to iDC and the binding could be prevented by neuraminidase treatment (Fig. 3,B). As expected, the binding of the α2,3-specific siglec sialoadhesin increased with maturation (Fig. 3,B), consistent with the staining obtained with the plant lectin Maackia amurensis agglutinin (MAA), although to a lower extent (Fig. 3,C). The presence and maturation-dependent increase of polysialic acid could be confirmed by using the specific Ab 735 (Fig. 3,D), and correlates nicely with an increased binding of the α2,8-specific siglec-7. Surprisingly, the binding of CD22 also increased with maturation (Fig. 3,B), correlating with an increased number of α2,6-linked sialic acids on mDC, as shown by the plant lectin Sambucus nigra agglutinin (SNA) (Fig. 3 C).

LacNAc and poly-N-acetyllactosamine (Fig. 4,A) have been reported as ligands for several galectins (39). To test whether the up-regulation of these structures on mDC would lead to an increased interaction with galectin, binding of galectin 3, 4, and 8 was tested by flow cytometry. Only galectins 3 and 8 showed binding to iDC, which could be blocked by lactose (Fig. 4,B) and increased during maturation (Fig. 4,B). The increase in galectin binding correlated with an increased binding of the β-galactoside-specific lectin Ricinus communis agglutinin (RCA) (Fig. 4 C).

We have investigated the array of glycans expressed on iDC, its regulation during DC maturation, and the functional consequences for the binding of several carbohydrate recognition molecules of importance in the immune system. The data showed the presence of abundant amounts of mono-, di-, and trisialylated tri- and tetra-antennary N-glycans, with some glycan species decorated with Lewis-type fucose or elongated with poly-N-acetyllactosamine, while the dominant O-glycan structure was the sialyl-T Ag. Maturation resulted in the modulation of the expression levels of several glycosylation-related genes, mainly affecting the synthesis of i-type elongated poly-N-acetyllactosamine chains, and the terminal capping of glycans by sialic acid and fucose, which correlated with the changes observed in the glycan profiles. These glycan changes are expected to play an important role in the biology of the DC, because their up-regulation correlated with an increased binding of their receptors, the carbohydrate recognition molecules galectin-3 and -8, as well as sialoadhesin, CD22, and -7.

Galectins are animal lectins recognizing β-galactose that comprise a family of 15 members, all containing conserved carbohydrate-recognition domains (CRD) of ∼130 aa responsible for carbohydrate binding (40, 41). Two types of galectins exist, those with one CRD (galectin-1, -2, -3, -5, -7, -10, -11, -13, -14, and -15), or with two homologous CRDs separated by a linker (galectin-4, -6, -8, -9, and -12). Galectin-3 is unique in that it contains a long N-terminal region rich in Pro and Gly that is involved in oligomerization (42). Although all galectins bind β-galactose, their fine specificity varies (31, 39, 41). Galectin-1 prefers terminal LacNAc in its core disaccharide binding site, although α2,3-sialylation on the Gal may be tolerated, with equal or slightly decreased affinity (43). Galectin-3 prefers to bind Galβ1-4GlcNAc and the affinity may increase with extensions on the Gal by another LacNAc or by α-linked Gal(NAc). The N-terminal CRD of galectin-8 prefers α2,3-sialylated galactosides as found in GM3 and GD1a, but in intact galectin-8 the two CRDs can cooperate to give high-affinity cell surface binding via LacNAc residues in N-linked glycans (31, 32). The CRDs of galectin-4 bind to the 3-O-sulfated glycolipids SM4, SM3, and SB1a, but also to some glycoproteins (44). iDC express relatively high levels of the transcripts for the enzymes β4GalT-1, -4, and iGnT, previously described to be involved in the expression of poly-N-acetyllactosamine chains (45), which are clearly detected by MS analyses of the N-glycans. This observation could explain the high binding of galectin-3 and -8 to iDC. The binding of galectin-3 and -8 increases during maturation, correlating with an increase in the expression of poly-N-acetyllactosamine elongated glycans possibly due to the up-regulation of β4GalT-4 and other enzymes potentially involved in the synthesis of these structures, such as β4GalT-5 (46) and β3GnT-2 (47). Binding of other galectins to monocyte-derived DC has been shown previously to induce maturation, giving rise to mDC with a strong proinflammatory capacity, as demonstrated for galectin-1 (48, 49) and -9 (50). Interaction of galectin-3 with its cell-associated ligands has been associated with a multitude of functional effects that include the modulation of cell adhesion, cell activation, chemoattraction, cell growth, or apoptosis (51), while galectin-8 has been associated with adhesion to the extracellular matrix and integrin-related reorganization of the cytoskeleton (52, 53).

The modulation of the expression of different sialyltransferases during maturation has also a clear repercussion on siglec binding. Siglecs are a family of type 1 membrane proteins with variable numbers of Ig domains and one N-terminal V-set domain where sialic acid binding is mediated via well-characterized molecular interactions (54). Siglecs can be divided into two subsets: the CD33-related siglecs (siglec-3, -5, -6, -7, -8, -9, -10, and -11) and a group comprising sialoadhesin, myelin-associated glycoprotein (siglec-4) and CD22, which are more distantly related. Strikingly, all siglecs except siglec-4 are expressed by cells of the immune system and most of them have two or more ITIMs. Although the function of the B cell-restricted CD22 has been analyzed in detail (18), the functions of the other siglecs remain poorly understood (54). The binding of CD22, and specially sialoadhesin, and siglec-7 to iDC was very strong and could, in the three cases, be up-regulated during maturation. The preferred ligand of CD22 is NeuAcα2-6Galβ1-4GlcNAc, but CD22 has also been shown to interact less strongly with other glycans capped with α2,6-linked sialic acid (55). The binding does not differ greatly for several sialylated proteins suggesting that the presence and density of the carbohydrate, but not the protein backbone, determine binding (18). Interestingly, the increase in the binding of CD22, or the α2,6-specific plant lectin SNA, did not correlate with the expression of the enzyme involved in the synthesis of its glycan, ST6Gal-1, which decreased dramatically. This could be explained by the maturation-triggered mobilization of molecules stored in granules, a phenomenon that has been previously described on DC (56, 57), and needs to be further investigated. The interaction between DC and B lymphocytes has been shown to result in multiple effects, such as the inhibition of B cell apoptosis (58), blockage of the production of IgE (59), or Ag presentation (60).

In contrast to CD22, the macrophage siglec sialoadhesin prefers α2,3-linked sialic acid (61), as found on the sialomucins P-selectin glycoprotein ligand 1 and CD43 (62), both expressed on DC (63, 64). As shown by MALDI-TOF mass spectrometry, the O-glycan NeuAcα2-3Galβ1-3GalNAc (sialyl-T Ag) is highly abundant on both iDC and mDC and correlates with a high siglec-1 binding. Macrophages and DC coexist in the spleen, tonsils, and lymph nodes (65, 66, 67). Interaction of siglec-1 with its ligand on DC may serve two purposes, either to enhance intercellular contact or to induce or modulate signaling on DC.

Siglec-7 is a CD33-related siglec containing intracellular inhibitory motifs that is expressed on NK cells and to a lesser extent, monocytes, macrophages, DC, and a subset of CD8⁺ T cells (68). In vitro-binding studies have revealed that siglec-7 shows a marked preference for glycoconjugates bearing α(2,8)-linked disialic acid (GD3 and GT1b) and branched α(2,6)-linked sialic acid (69). The function of siglec-7 on NK cells appear to be related with a down-modulation of the NK killing activity (70), therefore trans interactions between NK cells and DC could serve this function. Siglec-7 is also expressed on DC, and cis interactions could thus exist and be involved in regulating DC activation, as speculated for the cis interactions on NK cells.

In conclusion, our results provide a detailed description of the glycans expressed on DC, their regulation during maturation, and some functional repercussions with regards to interactions in trans with receptors expressed on other immune cells or in cis on DC. DC appear to be rich in galactosylated and sialylated structures, which are up-regulated during maturation, resulting in an enhanced binding of the lectins galectin-3 and -8, and sialoadhesin, CD22, and siglec-7. Further research is expected to shed more light on the functional consequences of these interactions.

The mAb 735 was provided by Dr. M. Muhlenhoff (Medizinische Hochschule Hannover, Hannover, Germany). Galectins-3 and -4 were produced by B. Kahl-Knutson, and galectin-8 was produced by S. Carlsson. We also thank G. Kraal, W. van Dijk, and T. K. van den Berg for critical reading of the manuscript and helpful suggestions. The glycan analyses were performed by the Analytical Glycotechnology Core of the Consortium for Functional Glycomics (GM62116).

The authors have no financial conflict of interest.