Activation of Lysosomal Function During Dendritic Cell Maturation

See allHide authors and affiliations

Science  28 Feb 2003:
Vol. 299, Issue 5611, pp. 1400-1403
DOI: 10.1126/science.1080106


In response to a variety of stimuli, dendritic cells (DCs) transform from immature cells specialized for antigen capture into mature cells specialized for T cell stimulation. During maturation, the DCs acquire an enhanced capacity to form and accumulate peptide–MHC (major histocompatibility complex) class II complexes. Here we show that a key mechanism responsible for this alteration was the generalized activation of lysosomal function. In immature DCs, internalized antigens were slowly degraded and inefficiently used for peptide loading. Maturation induced activation of the vacuolar proton pump that enhanced lysosomal acidification and antigen proteolysis, facilitating efficient formation of peptide–MHC class II complexes. Lysosomal function in DCs thus appears to be specialized for the developmentally regulated processing of internalized antigens.

The initiation of antigen-specific immune responses requires the conversion of protein antigens into peptides that bind to MHC molecules (1). In the MHC class II (MHC II) pathway, internalized antigens encounter newly synthesized MHC II in endocytic organelles, where proteases cleave the MHC II–associated invariant (Ii) chain and create the antigenic peptides that are loaded onto MHC II assisted by the chaperone H-2M (2–8). However, peptide–MHC II complexes (pMHC) are formed under conditions that would appear largely unfavorable. Production of peptide ligands (>10 to 15 residues in length) should not be a common event because lysosomes generally produce rapid and complete degradation of internalized proteins to amino acids and dipeptides (9–11). Possibly, professional antigen-presenting cells (APCs) such as DCs have specialized mechanisms to overcome this difficulty. DC function is carefully regulated by a process of terminal differentiation termed “maturation,” a feature that contributes to their potency and versatility as APCs (12–14). In the immature state, DCs are adept at the capture of antigens by endocytosis but exhibit a limited capacity for stimulating T cells. Their MHC II molecules are generally retained in lysosomes unable to form pMHCs. Upon exposure to maturation signals, DCs adopt a mature phenotype endowed with immunogenic or tolerogenic capacities (13, 15, 16). Mature DCs can present peptides from antigens internalized hours or even days before maturation and transfer pMHCs formed in lysosomes to the plasma membrane (17–23).

Despite being a key feature of DC maturation, the mechanisms that regulate their capacity for antigen processing remain poorly understood. Conceivably, immature DCs might control access of internalized antigen to degradative compartments or regulate the proteolytic activity within those compartments. Either mechanism would slow the generation of antigen-derived peptides and their loading onto MHC II molecules by inhibiting the removal of Ii chain. Immature DCs efficiently deliver internalized antigens to MHC II–positive lysosomal structures (MIICs) (19, 20). Thus, regulation of lysosomal activity, rather than transfer to lysosomes, appears more likely. To test this possibility directly, we evaluated the degradative capacity of immature and mature DCs in living cells (24). Lysosomes in immature DCs were loaded with horseradish peroxidase (HRP) by endocytosis for 1 hour at 37°C, then cultured at 37°C for 19 hours in the presence or absence of a maturation stimulus [replating in lipopolysaccharide (LPS)]. Cell-associated HRP protein was determined with antibodies to HRP and quantified by flow cytometry (Fig. 1A). After 19 hours in the absence of LPS, immature DCs contained most (>80%) of the HRP that they had accumulated initially. In contrast, cells cultured in the presence of LPS contained substantially less HRP, ∼10% of that found initially internalized (Fig. 1A, bottom two panels; t = 19 hours). The flow cytometric data were quantified and compared with HRP levels determined on the basis of enzymatic activity, with similar results (Fig. 1B).

Figure 1

(left). Enhanced lysosomal proteolysis upon DC maturation. Synchronized cultures of immature DCs were allowed to internalize the indicated tracers for 1 hour at 37°C (24). A portion of the DCs were sampled at the initial time point (t = 0). The remaining DCs were either left resting at the immature stage or were activated by replating in growth media supplemented with LPS. (A) Enhanced degradation of HRP upon DC maturation. At the specified time points, cells were fixed and stained simultaneously for HRP and MHC II, showing that only DCs were scored for HRP content. (B) The amount of internalized HRP remaining inside DCs was quantitated by analysis of its enzymatic activity (solid bars, percentage of the activity measured at t = 0) and by flow cytometry (open bars) with anti-HRP, plotted as the mean fluorescence intensity (MFI) on the basis of the flow cytometric data in (A). (C) Enhanced degradation of HEL upon DC maturation. Immature bone marrow–derived DCs were loaded with HEL and processed as described for HRP in (A) and (B). Histograms show the quantitation of HEL retained in DCs from flow cytometric data similar to that in (A) (fig. S1). (D) Rapid degradation of BSA in both immature and mature DCs, estimated as in (B) and (C) (fig. S1). (E) Degradation of HRP, HEL, and FITC-BSA simultaneously loaded into immature DCs. Immature or LPS-matured DCs were then sampled at the indicated time points and analyzed by immunoblot.

We next monitored the fate of the well-characterized antigen hen egg lysozyme (HEL). As found for HRP, little of the internalized HEL was lost in DCs that remained in the immature stage, whereas maturation induced extensive loss of HEL (Fig. 1C and fig. S1A). The increased lysosomal proteolysis of HEL induced by maturation was accompanied by the appearance of HEL(46-61) peptide bound to the I-Ak class II molecule (fig. S1B).

To analyze the fate of a tracer less resistant to proteolysis than HEL or HRP, we used bovine serum albumin (BSA) [half-time (t 1/2) for digestion ∼5 min in macrophages (25)]. Fluorescein isothiocyanate (FITC)–labeled BSA was readily degraded by both immature and mature DCs (Fig. 1D and fig. S1A), suggesting that it was susceptible to lysosomal proteolysis even by the reduced proteolytic activity in immature DCs.

When immature DCs were simultaneously loaded with all three tracers (HEL, HRP, and FITC-BSA), they were degraded at rates similar to those observed when analyzed individually (Fig. 1E). Thus, the substrates chosen did not affect lysosomal proteolysis in immature DCs; instead, their individual degradation rates reflected inherent susceptibilities to the lysosomal proteolytic capacity of DCs at different stages of maturation. Little if any internalized tracer was released intact from DCs during maturation (see below; fig. S3, A to C), confirming that antigen loss was due to intracellular proteolysis.

Having demonstrated that immature DCs were less capable of degrading internalized antigens than mature DCs, we next investigated the mechanism underlying this pattern of regulation. The enhanced proteolytic capacity of mature DCs did not reflect increased levels of lysosomal proteases (Fig. 2A). Immature and mature DCs contained similar steady-state levels of cathepsins H, D, S, L and legumain (AEP). However, a substantial fraction of both cathepsin L and AEP in immature DCs was in the inactive proenzyme form (Fig. 2A, arrows), indicating that the activation of at least two lysosomal proteases was increased upon DC maturation. These enzymes colocalized in lysosomes with MHC II (Fig. 2B) and MHC II with antigen (fig. S1D). Thus, reduced lysosomal proteolysis and pMHC formation in immature DCs did not reflect segregation of lysosomal proteases from MHC II or internalized antigens.

Figure 2

(right). Immature and mature DCs contain similar amounts of lysosomal proteases. (A) The steady-state levels of cathepsin H, D, S, and L in immature (i) versus mature (m) DCs were analyzed by immunoblot (24). Arrows indicate the proforms of cathepsin L and legumain (AEP). The levels of the lysosomal markers Lamp-1 and H-2M are shown as loading controls. (B) Colocalization of cathepsin S, D, L, and AEP with MHC II in immature DCs analyzed by confocal microscopy. MHC II also colocalized with the HRP and BSA (fig. S1) and with Lamp-1 and H-2M (20).

We next characterized the proteolytic potential of DC lysosomes using an in vitro assay in which antigen digestion by lysosomal extracts was assayed at various times and pH values (24). Degradation was expected to be more efficient at low pH because most lysosomal hydrolases exhibit acidic pH optima (26). Degradation of BSA was very efficient by both immature and mature lysosomal extracts (Fig. 3A). A similar pattern of proteolysis was observed when ovalbumin (OVA) was used as a substrate (Fig. 3B), although digestion of OVA appeared slower and strictly required pH 4.5; little degradation was observed at pH >5.0. The degradation of both HEL (fig. S2A) and HRP (fig. S2B) followed a similar pattern. This was in marked contrast to the degradation of these proteins in vivo, where immature DCs were much less efficient than mature DCs (Fig. 1). We next monitored the digestion of tetanus toxoid (TT) as a measure of AEP activity; AEP is the sole protease capable of cleaving TT (24). TT digestion was similar in immature and mature DC lysosomal extracts (Fig. 3C), in agreement with the similar levels of total AEP observed by immunoblot. AEP activity was also highly pH dependent, with optimal activity below pH 5.0 and little if any degradation at pH 5.5 and higher. Thus, immature and mature DCs exhibit comparable capacities for lysosomal proteolysis when assayed in vitro, consistent with the observation that both cell populations contain similar levels of cathepsins and AEP. Thus, lysosomal proteolysis is attenuated in immature DCs in vivo.

Figure 3

Immature and mature DC lysosomes exhibit comparable proteolytic capacities in vitro. Microsomes from immature or mature DCs were incubated with the indicated protein substrates at 37°C in 0.1 M sodium citrate–phosphate buffer, 0.5% Triton X-100, and 2 mM dithiothreitol at the indicated pH. The protein substrates and times of incubation shown were as follows: BSA (A) (1 hour incubation), OVA (B) (5 hours), and tetanus toxoid C-fragment (C) (5 hours). Degradation at pH 4.0 was blocked in the presence of a protease inhibitor cocktail (left lane). Protein digestion was analyzed by SDS–polyacrylamide gel electrophoresis; proteins were stained with Coomassie blue.

One possible site of regulation is lysosomal acidification, a mechanism suggested by the sharp pH requirement of lysosomal proteolysis in vitro. We thus measured lysosomal pH in intact cells. Immature DCs were loaded with the pH-sensitive reporter FITC-dextran by endocytosis for 1 hour at 37°C and then chased in the presence or absence of the maturation stimulus LPS (24). Immature DCs exhibited an intralysosomal pH of ∼5.4 (Fig. 4A). As expected, lysosomal pH was increased by addition of weak bases such as NH4Cl. The effect of the weak bases was reversible, indicating that the fluorescence readings were obtained from intact, functional lysosomes.

Figure 4

Acidification of lysosomes in mature DCs. (A) Analysis of lysosomal pH in live DCs (24). Immature DCs were loaded with the pH-sensitive fluorescent dextrans by endocytosis for 1 hour at 37°C and then subjected to pulse-chase assay in the presence or absence of the maturation stimulus (LPS). (○) Values obtained from cells loaded with FITC-dextran. (•) Values obtained from DCs loaded with fluorescent dextrans of different pK a values (Oregon Green–dextran-488, Rhodol Green–dextran, or BCECF-dextran). Lanes 2, 4, and 6 depict pH values obtained with FITC-dextran after the addition of 20 mM NH4Cl. (B) Enhanced lysosomal V-ATPase activity in mature DCs. Lysosomes from immature and mature DCs were loaded with FITC-dextran (molecular weight 70,000) as in (A). DCs were then homogenized and used for in vitro acidification assays (24). (C) Immature and mature DCs contain similar levels of V-ATPase. Immunoblot analysis of the steady-state levels of V0 sector subunit a and V1 sector subunits A and E in whole-cell lysates of immature (i) and mature (m) DCs. (D) Schematic representation of regulated V-ATPase assembly. (E) DC maturation induces the recruitment of cytosolic V-ATPase subunits to membranes. Immunoblot analysis of the distribution of V1 sector subunits A (70 kD) and E (37 kD) between membrane and cytosol fractions from immature DCs (lanes 1, 2, 5, and 6) and mature DCs (lanes 3, 4, 7, and 8). (F) Cytosol from immature DCs was separated on a Superose 6 gel filtration column and analyzed by immunoblot for the content of V-ATPase subunits A and E. Molecular size markers were bovine thyroglobulin (660 kD), potato β-amylase (200 kD), and BSA (68 kD).

In contrast, the lysosomal pH values in mature DCs were consistently more acidic, measuring nearly one pH unit less than in immature DCs (pH 4.5 to 4.6) (Fig. 4B). As with immature DCs, the pH gradients in mature DCs were reversibly dissipated by NH4Cl (Fig. 4B). The pH values obtained for mature DCs were similar to those determined in macrophages and fibroblasts (Fig. 4B), in which lysosomal pH is 4.5 to 4.6 (27). Therefore, the internal pH of lysosomes in mature DCs was similar to that of other cell types, suggesting that the capacity for lysosomal acidification is restricted in immature DCs. Thus, DC maturation was accompanied by an increase in lysosomal acidification that, on the basis of in vitro degradation data, shifted lysosomal pH from one that is decidedly subpoptimal for protein digestion to one that could facilitate efficient proteolysis.

Acidification is due to the activity of the ATP-dependent vacuolar proton pump (V-ATPase). Because proton translocation is electrogenic, other factors such as the ion-permeability characteristics of the lysosomal membrane play critical roles in controlling pH (28). To determine if differences in V-ATPase activity or membrane permeability were responsible for regulating lysosomal pH in DCs, we examined the ATP-dependent acidification using an in vitro assay (24). Lysosomes from mature DCs exhibited a greater acidification capacity than immature lysosomes, consistent with the differences observed in lysosomal pH in intact cells (Fig. 4B). The use of specific ionophores to uncouple electrical from pH gradients formed after ATP addition allowed us to rule out differences in counterion conductances or proton leakage (fig. S3). Thus, the greater capacity for acidification in mature DC lysosomes was likely due to increased V-ATPase activity.

Although the V-ATPase appeared more active in mature than in immature DC lysosomes, its 37-, 70-, and 110-kD subunits were present at similar levels in both immature and mature DCs (Fig. 4C), suggesting that activation of preexisting V-ATPase subunits was responsible. The V-ATPase is only active when both of its subcomplexes or “sectors” (V1, soluble domain; and V0, integral membrane domain) are assembled onto the lysosomal membrane (Fig. 4D) (29). We therefore examined the assembly status of the V1 and V0 sectors in developing DCs. Two subunits (37 and 70 kD) of the V1 sector were enriched on the membrane fraction of mature DCs (Fig. 4E) (∼45% membrane-associated in immature DCs versus ∼80% in mature DCs). The soluble subunits were present in the cytosol of immature DCs as a high molecular weight complex (∼700 kD) (Fig. 4F), indicating their inclusion in assembled V1 sectors. Thus, it seems likely that DC maturation induces the more efficient recruitment of V1 sectors onto lysosomal membranes, resulting in a higher proportion of functional V-ATPase and increased acidification capacity in mature DC lysosomes in vitro (Fig. 4, B to F) and in intact cells (Fig. 4A). Regulated assembly of V1 and V0 sectors has been observed in yeast and insects (30, 31), and the ubiquitin pathway seems to be involved in V-ATPase assembly in yeast (32, 33).

The observation that lysosomal proteolysis is developmentally controlled provides an explanation for how immature DCs can restrict their ability to generate pMHC until after maturation, and how DCs can efficiently present antigens even when captured days before receiving a maturation signal. Low proteolytic activity in immature DCs would allow a substantial fraction of the internalized antigen to escape degradation in lysosomes, facilitating efficient conversion into pMHC after DC maturation is induced. Although we do not yet know the kinetics by which lysosomal function is activated, this mechanism may act as a clutch that is released only after a peripheral DC begins its journey to lymphoid organs, delaying the conversion of antigen to pMHC until arrival of the DC at sites of T cell encounter.

Of the various factors that may contribute to the modulation of lysosomal proteolysis, the increased intralysosomal acidification seen in mature DCs is probably the most direct because it could favor proteolysis in several ways. Enhanced acidification would provide a pH environment closer to the optimal pH of most lysosomal enzymes (pH ∼4.5). A similar increase in proteolytic capacity in concert with lysosomal acidification was described for the regulation of yolk degradation (from pH 5.5 down to pH 4.5) (34,35). Lysosomal acidification could also augment proteolysis and pMHC formation by favoring the activation and function of the proenzyme forms of GILT (36), lysosomal proteases (as we observed here for cathepsin L and legumain), and H-2M activity, all optimal at acidic pH (5).

Understanding how lysosomal function controls the formation of immunogenic complexes may contribute to elucidating how DCs might paradoxically regulate both immunity and peripheral tolerance, because the signals leading to the activation of the V-ATPase could conceivably be different from those leading to the transcriptional up-regulation of costimulatory molecules. Because cytokines can modulate the endocytic system of macrophages and perhaps DCs (37–40), controlled activation of lysosomal function during DC maturation may contribute substantially to immunogenicity, providing opportunities for pharmacological intervention.

Supporting Online Material

Materials and Methods

Figs. S1 to S3


  • * To whom correspondence should be addressed: E-mail: ira.mellman{at}


View Abstract

Navigate This Article