Phenotypic and Functional Maturation of Dendritic Cells Mediated by Heparan Sulfate¹

Dendritic cells regulate immunity (1), particularly primary immune responses (2). This property depends, critically, on functional maturation and migration of the cells. The dendritic cells residing in peripheral tissues are generally immature phenotypically and functionally. Immature dendritic cells do not induce primary immune responses, because they do not express the requisite costimulatory molecules, nor do they express antigenic peptides as stable complexes with MHC molecules. Immune responses, rather, are thought to begin when dendritic cells are driven by various agonists to mature and to migrate to regional lymph nodes. A central question, then, is what is the nature of the stimulus that induces maturation of the dendritic cell.

Dendritic cells can be induced into mature immunostimulatory APC by microbial components (3), such as LPS (4) and dsRNA (5); by cytokines, such as TNF-α, IL-1, and type I IFN (4, 6, 7, 8); and by ligation of CD40 (9). Maturation of dendritic cells can also be induced by ligation of Fcγ receptors (10), CD43 (11), and CD44 (12), by cell adhesion (13, 14), and by prostaglandins (15, 16). These stimuli require the local presence of microorganisms or activated leukocytes. How primary immune reactions, as might be seen in response to viral and tumor Ags and to transplants, are induced in the absence of these inflammatory mediators is unknown. Gallucci et al. (17) call these mediators of dendritic cell activation, natural adjuvants; however, the identity of such natural adjuvants is, likewise, unknown. We postulated that one component of normal tissues, which might mediate this effect, is heparan sulfate.

Heparan sulfate is a linear copolymer, comprised of repeating hexuronic acid and hexosamine residues, which is found ubiquitously distributed on cell surfaces and extracellular matrices. As a component of proteoglycans, the heparan sulfate polymers are covalently attached to a protein core (18). Structural modifications of the saccharide residues, especially sulfation, confer manifold biological activities, including regulation of cell adhesion, proliferation, development, anticoagulant, and chemical mediator functions (18, 19, 20, 21). Injury of tissues (22) or exposure of endothelial cells and, perhaps, other cells to activated complement or to neutrophils or platelets causes rapid cleavage and shedding of heparan sulfate proteoglycans and glycosaminoglycan fragments (23, 24, 25, 26). The release of heparan sulfate is postulated to be mediated by proteolytic cleavage of the protein core or by endoglycolytic cleavage of the heparan sulfate chains (25, 27, 28, 29).

We and others have proposed that the release of heparan sulfate proteoglycans may play an important role in the regulation and manifestation of immune responses in the local inflammatory sites (21, 30). Consistent with this hypothesis, we previously found that soluble heparan sulfate delivers signals to macrophages, including activation of tyrosine kinase and protein kinase C, calcium influx, elevation of inositol phosphate, and NF-κB translocation (31). These events cause murine macrophages to release IL-1, IL-6, and PGE₂ (32), and to up-regulate MHC class II and CD86 (33), profoundly increasing the ability of the macrophages to modify immune responses.

Because the conditions associated with maturation and migration of dendritic cells might be accompanied by degradation of heparan sulfate proteoglycans in peripheral tissues, we hypothesized that soluble heparan sulfate might be one factor promoting the maturation of dendritic cells. To test this idea, we stimulated murine bone marrow-derived dendritic cells with soluble heparan sulfate, without exogenous cytokines, analyzing the phenotypic and functional changes that ensued. We report, in this study, that interaction with soluble heparan sulfate induces maturation of dendritic cells, suggesting that release of heparan sulfate proteoglycans may provide an endogenous stimulus to induce cellular immune response.

FITC-conjugated mAbs, HL3 (anti-CD11c), HM40-3 (anti-CD40), RA3-6B2 (anti-CD45R/B220), 3E2 (anti-CD54: ICAM-1), 16-10A1 (anti-CD80: B7-1), GL1 (anti-CD86: B7-2), RB6-8C5 (anti-Ly-6G: Gr-1), G235-2356 (hamster IgG isotype standard), and R35-95 (rat IgG isotype standard); biotin-conjugated Abs, AF6-88.5 (anti-H-2K^b), AF6-120.1 (anti-I-A^b), PE-conjugated streptavidin, and 2.4G2 (anti-CD16/CD32); neutralizing Abs, 35F5 (anti-CD121a: IL-1R type I/p80), MP5-20F3 (anti-IL-6), G281-2626 (anti-TNF-α), and R3-34 (isotype control for rat IgG1); recombinant murine TNF-α and IL-6 were purchased from PharMingen (San Diego, CA). FITC-conjugated anti-macrophage F4/80 Ag mAb was from Caltag (Burlingame, CA). Neutralizing Abs contained ≤0.01 ng endotoxin/μg protein. Heparan sulfate (from bovine kidney), Limulus amebocyte assay (E-TOXATE), and LPS from Escherichia coli were from Sigma (St. Louis, MO). An endotoxin filter (END-X) and endotoxin removal resin (END-X B15) were from Associates of Cape Cod (Woods Hole, MA). Cytokine ELISA kits for murine IL-1β, IL-6, and TNF-α, and murine rIL-1β were purchased from R&D Systems (Minneapolis, MN).

Dendritic cells were generated from murine bone marrow cells, as described by Inaba et al. (34, 35), with minor modifications. Briefly, bone marrow was flushed from the long bones of C57BL/6 mice (Charles River, Wilmington, MA) and depleted of red cells with ammonium chloride. Cells were plated in six-well culture plates (10⁶ cells/ml, 3 ml/well) in RPMI 1640 supplemented with 5% heat-inactivated FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 5 × 10⁻⁵ M 2-ME, 10 mM HEPES (pH 7.4) (Life Technologies, Gaithersburg, MD), and 3.3 ng/ml of murine rGM-CSF at 37°C, 5% CO₂. At day 3 of culture, floating cells were gently removed and fresh medium was added. At day 6 or day 7 of culture, nonadherent cells and loosely adherent proliferating dendritic cell aggregates were harvested for analysis or stimulation, or, in some experiments, replated in 60-mm petri dishes (10⁶ cells/ml, 5 ml/dish). At day 10 of culture, nonadherent cells (dendritic cells) were removed for analysis. Less than 1% of cells were B220⁺, and less than 10% were Gr-1⁺ analyzed by flow cytometry (not shown). Dendritic cells were prepared by cytospin (Shandon, Pittsburgh, PA) for morphological analysis with Wright-Giemsa stain (EM Diagnostic Systems, Gibbstown, NJ).

In certain experiments, CD11c⁺ dendritic cells were isolated from cultured bone marrow cells using anti-CD11c (N418) microbeads and a magnetic cell sorting system (Vario MACS; Miltenyi Biotec, Auburn, CA). Purity of the selected cell fraction was >92%.

Heparan sulfate dissolved in distilled water, or distilled water alone, as control, was added to cultures of isolated dendritic cells in six-well plates (10⁶ cells/ml, 3 ml/well). Heparan sulfate was treated with LPS-binding protein, conjugated to a filter or resin to remove endotoxin (36). In certain experiments, heparan sulfate was pretreated with HNO₂ at pH 1.5 to depolymerize the glycosaminoglycan chains (37). Neutralizing Abs against TNF-α, IL-1R, and IL-6 were used at ∼10-fold the concentrations needed for 50% inhibition, as reported in the manufacturer’s instructions.

Splenocytes, isolated from female BALB/c mice at 5–8 wk of age (Charles River), were passed over nylon wool columns to enrich T cells. Nylon wool nonadherent splenocytes (4 × 10⁵/well) were cocultured with irradiated (15 Gy) dendritic cells isolated from C57BL/6 mice, as described above. Mixed cultures were conducted in RPMI 1640 supplemented with 1% mouse serum (from BALB/c mice), 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 5 × 10⁻⁵ M 2-ME, and 10 mM HEPES (pH 7.4) in 96-well round-bottom plates at 37°C, 5% CO₂ in air. Cell proliferation was estimated based on uptake of [³H]thymidine (ICN, Irvine, CA). For this purpose, cells were pulsed with 5 μCi/ml of [³H]thymidine for 18 h, with the radioactivity of harvested cells, then measured by a liquid scintillation counter (Wallac Oy, Turku, Finland).

Flow cytometric analysis was performed, as described by Kodaira et al. (38). Briefly, 5 × 10⁵ cells were incubated in staining buffer (PBS with 2% FBS and 0.1% sodium azide) containing anti-CD16/CD32 Ab (PharMingen) to block nonspecific binding of Igs for 5 min on ice. Cells were then stained, as indicated, and analyzed by FACScan, using CellQuest software (Becton Dickinson, San Jose, CA). FACStar cell sorter was used for cell sorting (Becton Dickinson).

Endocytosis was quantitated, as described by Lutz et al. (39) and Sallusto et al. (4). Briefly, 2 × 10⁵ cells were equilibrated at 37°C or 0°C for 10 min and then pulsed with fluorescein-conjugated dextran (40,000 m.w.; Molecular Probes, Eugene, OR) at a concentration of 1 mg/ml. After different times in incubation, at 37°C or 0°C, cold staining buffer was added to stop the reaction. Cells were washed three times and stained with biotin-conjugated anti-I-A^b Abs and streptavidin-PE, then analyzed by FACScan. Nonspecific binding of dextran to dendritic cells, determined by incubation of dendritic cells with fluorescein-conjugated dextran at 0°C, was subtracted. The medium used in the culture, to stimulate dendritic cells with heparan sulfate, was supplemented with GM-CSF, because the ability of dendritic cells to capture Ag is lost if dendritic cells are cultured without GM-CSF (data not shown) (4).

Culture supernatants were analyzed by ELISA. OD at 450 nm of duplicate samples was determined and corrected by a microplate reader, with readings at 570 nm (PowerWave_x with KC4 software; Bio-Tek, Winooski, VT). The minimum detection levels were as follows: IL-1β, 3 pg/ml; IL-6, 3.1 pg/ml; and TNF-α, 5.1 pg/ml.

The Mann-Whitney U test was used for statistical analysis. A p value of less than 0.005 was considered to indicate statistical significance.

After 6 days of culture with GM-CSF, the nonadherent fraction of murine bone marrow cells consisted mainly (∼70%) of immature dendritic cells with a phenotype of CD11c⁺, MHC class II^intermediate, CD40^low, CD54^low (ICAM-1), and CD86^low (B7-2) (Fig. 1), consistent with previous reports (34, 40). After 10 days of culture with GM-CSF, 80–90% of cells exhibited a relatively mature dendritic cell phenotype based on expression of CD11c, I-A, CD40, CD54, CD80, and CD86 (Fig. 1). While expression of MHC class II, CD40, and CD86 can be detected at day 10, the expression of these molecules is well below maximum levels (seen in stimulated dendritic cells, as illustrated in Fig. 2).

To test whether heparan sulfate influences the maturation of dendritic cells, immature dendritic cells from day 6 murine bone marrow cultures were incubated for 48 h, with heparan sulfate without exogenous cytokines. Analysis of the heparan sulfate-treated cells revealed up-regulation of I-A on dendritic cells. Increased expression of CD40 and CD86 (B7-2), and, to a lesser extent, up-regulation of CD54 (ICAM-1) and CD80 (B7-1), were detected on the I-A⁺ cells (Fig. 2,a). Not only was the intensity of I-A, CD40, and CD86 increased, but the fraction of cells expressing these molecules was increased by heparan sulfate in a dose-dependent fashion. Increase of I-A⁺/CD40⁺ and I-A⁺/CD86⁺ cell fractions and the levels of expression of the costimulatory molecules induced by heparan sulfate was comparable with changes induced by LPS (Fig. 2, a and b). More than 98% of sorted I-A⁺/CD86⁺ cells exhibited morphological characteristics of dendritic cells, such as extended cell processes and large size (not shown). Although up-regulation of I-A and slight increase of CD54 were observed in control cells, increase in expression of CD40 and CD86 was not observed without stimulation of dendritic cells with heparan sulfate or LPS. The reason for up-regulation of I-A and CD54 is unclear; however, the control cells may be partially stimulated by mechanical manipulation during transfer between culture plates (17). More than 98% of sorted I-A⁺/CD86⁺ cells exhibited morphological characteristics of dendritic cells, such as extended cell processes and large size (not shown). The maturation of dendritic cells was not the result of lipid or protein contaminants of heparan sulfate because: 1) only a minimal level of endotoxin was detected by Limulus amebocyte assay in heparan sulfate solution; 2) absorbing the solution with LPS-binding protein to remove endotoxin did not affect the maturation; 3) the same results were obtained using heparan sulfate solution, which had been boiled for 10 min to denature any contaminating proteins; and 4) up-regulation of CD40 and CD86 did not occur if heparan sulfate was subjected to deaminative cleavage with HNO₂. In contrast, up-regulation of CD40 and CD86 was observed in dendritic cells stimulated with LPS that had been treated with HNO₂ (data not shown).

Maturation of dendritic cells exposed to heparan sulfate could reflect the direct action of heparan sulfate on dendritic cells or a response of dendritic cells to cytokines released by other cells. For example, we previously showed that macrophages, stimulated with heparan sulfate, release IL-1 and IL-6 (31). To rule out the possibility that maturation of dendritic cells is induced by cytokines released from macrophages or other cells, the responses of highly purified dendritic cells were tested. CD11c⁺ dendritic cells (>92%) also up-regulated CD40, CD54, CD80, and CD86 after stimulation with heparan sulfate for 48 h (Table I). The same mature phenotype was observed if immature dendritic cells were stimulated with heparan sulfate in GM-CSF-supplemented medium, indicating that exogenous GM-CSF did not alter the effects of heparan sulfate (not shown). Therefore, cytokines released from cells other than dendritic cells are not necessary for the maturation of dendritic cells induced by heparan sulfate.

We next asked whether stimulation of dendritic cells with heparan sulfate changes the ability of dendritic cells to capture Ag quantitated based on uptake of fluorescein-conjugated dextran (41, 42). As shown in Fig. 3, dendritic cells from day 6 bone marrow cultures, treated for 48 h with heparan sulfate, had remarkably low uptake of Ag, compared with untreated, immature dendritic cells (bone marrow culture at day 6) or dendritic cells treated without heparan sulfate. These data suggest that heparan sulfate induced the mature metabolic state of dendritic cells.

To test whether heparan sulfate induces maturation of dendritic cells to fully functional APC, dendritic cells from day 6 cultures that had been incubated with heparan sulfate for 48 h were tested for their capacity to stimulate allogeneic T cells. As Fig. 4 shows, heparan sulfate-treated dendritic cells stimulated proliferative responses more effectively than control dendritic cells, and as well as LPS-treated cells did. Augmentation of proliferative responses was most evident under suboptimal conditions in which the number of stimulator cells was limited. The addition of GM-CSF to heparan sulfate further augmented proliferative responses induced by dendritic cells under suboptimal conditions (not shown). Viability of the stimulator cells incubated with and without heparan sulfate was 95.2 ± 0.9% and 90.2 ± 2.1% (mean ± SE of three experiments), respectively, suggesting that the effects of heparan sulfate were not due to differences in viability of dendritic cells.

We next analyzed whether or not heparan sulfate might induce release of cytokines by dendritic cells and whether or not such cytokines might, in turn, mediate the maturation of dendritic cells. Significant amounts of TNF-α, IL-1β, and IL-6 were detected in the culture supernatant of purified CD11c⁺ dendritic cells stimulated with heparan sulfate for 48 h (Fig. 5). However, as seen from results in Table I, the up-regulation of CD40 and CD86 after stimulation with heparan sulfate was not blocked by neutralizing Abs against TNF-α, IL-1R type I, or IL-6, nor did various combinations of neutralizing Abs prevent phenotypic maturation (not shown). The neutralizing activities of anti-TNF-α and anti-IL-1R Abs at the indicated concentrations were confirmed in the same model systems as shown in Table I. The maturation of dendritic cells induced by TNF-α (10–20 ng/ml) or IL-1β (0.2–10 ng/ml) was completely blocked by the neutralizing Abs, but not by isotype control. The presence of IL-6 alone, at the concentration shown in Fig. 5 (40 ng/ml), did not induce the up-regulation of CD40 and CD86 (not shown). Furthermore, the up-regulation of CD40 and CD86 induced by a mixture of TNF-α, IL-1β, and IL-6 was abrogated at 92% by a simultaneous presence of the three neutralizing Abs. Therefore, TNF-α, IL-1β, and IL-6, released from dendritic cells, do not directly contribute to maturation of dendritic cells induced by heparan sulfate, suggesting that signals, provided by heparan sulfate itself, or the other factors induced by heparan sulfate, cause maturation.

We report, in this study, that heparan sulfate induces phenotypic and functional maturation of dendritic cells. This finding suggests that metabolism of natural components of cell surfaces and extracellular matrices might provide a sufficient stimulus for the induction of immune responses. As a result of our findings, we offer a model that might explain the induction of primary immune responses directed against tumor cells or virus-infected cells, as well as in cases of autoimmunity, that arise in the absence of the exogenous stimuli that would otherwise be needed to induce the maturation of dendritic cells. Our model is consistent with Gallucci et al. (17), who found that dendritic cells could be activated by the products of injured or necrotic cells.

The major requirement for the effects of heparan sulfate on dendritic cells may be solubilization of heparan sulfate proteoglycans from normal tissues. Solubilization of heparan sulfate might be brought about by enzymatic cleavage of the protein core, the glycosaminoglycan chains, or, in the case of glypican, the lipid anchor (25, 27, 28, 29). For example, in injured or inflamed tissue, heparanase released from platelets and stabilized by acidic pH could mediate cleavage of heparanase sulfate (25, 30). The activation of complement to induce ischemia or tissue damage might cause the activation of proteases that cleave heparan sulfate proteoglycans from vascular endothelium (23). It should be noted that cleavage of heparan sulfate following activation of complement is nearly immediate, and the metabolites can be assumed to be the products of normal tissues. Because heparanase is also produced by some tumors (27, 43, 44), heparan sulfate fragments might be released in the neighborhood of invasive or metastasizing tumors. It has been reported that tumor cells undergoing necrosis provide signals for maturation of dendritic cells (17, 45), although detailed mechanisms on the induction of maturation are not completely clarified. Heparan sulfate, released from necrotic tumor, might well be one of the maturation factors that initiates tumor immunity through dendritic cell maturation.

Although it is still unclear how heparan sulfate provides signals to APCs, direct binding of heparan sulfate to cell surface receptors is one possibility (31). We have reported that signaling via GPI-linked receptor was at least partially involved in activation of murine peritoneal macrophages by heparan sulfate, leading to activation of protein kinase C and NF-κB translocation (31).

Peptidoglycan and LPS signal cells through a toll-like receptor that has been reported as a 70-kDa protein on the surface of mouse lymphocytes (46, 47). This receptor may also be shared by heparan sulfate. These reports suggest the possibility of receptor-mediated activation of APCs by heparan sulfate. On the other hand, heparan sulfate might also modify stimulation of dendritic cells by growth factors, chemokines, or cytokines. Binding of heparan sulfate to some growth factors, such as fibroblast growth factors, induces structural modifications (48) or protects the factors from degradation by proteases (22). Heparan sulfate might also provide a bridge between components of the extracellular matrix and components of cell surfaces (49), leading to induction of maturation of dendritic cells (13, 14).

The metabolism of heparan sulfate proteoglycan and action of soluble fragments on dendritic cells might also contribute to chronic immune reactions. For example, in rheumatoid arthritis, autoreactive T cells are induced and maintained by local dendritic cells, with this ongoing interaction possibly leading to formation of local lymphoid tissue (50). Because rheumatoid synovium contains elevated levels of cleaved glycosaminoglycans (51, 52), it is not unlikely that heparan sulfate stimulates dendritic cells to present self Ags to autoreactive T cells, perhaps under control of endogenous TNF-α (53), and that cytokines produced by the stimulated dendritic cells contribute to local inflammation and tissue injury.