Impaired Phagosomal Maturation in Neutrophils Leads to Periodontitis in Lysosomal-Associated Membrane Protein-2 Knockout Mice¹

Periodontitis is an infectious disease that is one of the most widespread diseases worldwide (1). It is estimated to affect up to 15% of the adult dentate population (2). Periodontitis is an inflammatory disease of the supporting tissues of the teeth leading to resorption of alveolar bone and eventually tooth loss. The disease is characterized by a constant interaction between pathogenic bacteria and the host defense mechanisms. In health, host immune responses are sufficient to hold in check the pathogenic potential of both the normal resident microbial flora and exogenous microbial pathogens. Complex inflammatory and immune reactions are involved in the progression of periodontitis. Polymorphonuclear leukocytes (PMNs)⁴ and circulating neutrophils constitute the first defense barrier against the oral bacterial challenge in the periodontium (3). They are rapidly recruited from the blood to the site at risk, and then phagocytose and kill the intruders. Neutrophils may release proinflammatory mediators that amplify the local inflammatory reaction, further promoting leukocyte and platelet recruitment. Quantitative or qualitative abnormalities of the PMNs may, therefore, have an effect on the accumulation of plaque in the supra and subgingival regions. Malfunctioning of PMNs including a disturbed adhesion to the endothelium, chemotaxis, detoxification of bacterial products, phagocytosis, or degranulation have been associated with early onset periodontitis (4, 5, 6, 7).

Phagosome-lysosome fusion bestows on the phagocytic vacuole the lytic properties for efficient removal of internalized pathogens. Lysosomes play a crucial role in the oxygen-independent killing of bacteria, which is believed to be an important killing mechanism in the oxygen-deprived periodontal pocket (8).

The limiting membrane of the lysosomal compartment is thought to be of importance for phagosome maturation (9). Lysosome-associated membrane protein (LAMP)-2 is a highly glycosylated protein. It is an abundant and important constituent of the lysosomal membrane involved in lysosomal biogenesis and late steps of autophagy and phagocytosis (10, 11, 12, 13, 14). LAMP-2-deficient mice exhibit elevated postnatal mortality, and the surviving mice are of reduced weight. Autophagic vacuoles accumulate due to an impaired proteolysis of long-lived proteins in a number of tissues, especially in myocytes, cardiomyocytes, and hepatocytes (10). In the latter cells, we observed an elevated secretion of lysosomal enzymes, impaired processing of cathepsin D, and an abnormal retention of mannose-6-phosphate receptors in autophagic vacuoles (13).

We report that LAMP-2-deficient mice show an increased susceptibility to periodontitis. PMNs isolated from these mice show an impaired fusion between phagosomes and lysosomes leading to a decreased bacterial killing capacity. This impairment is the likely cause of the development of severe periodontitis.

LAMP-2-deficient mice were generated as previously described (10). The mice were maintained in a conventional facility. For antibiotic treatment selected mice received amoxicillin in the drinking water before birth (the pregnant mothers) and immediately after birth (in a concentration of 5 mg/L).

Mouse primary neutrophil granulocytes were isolated from 2- to 4-mo-old mice. Mice were injected i.p. with 1 ml of sterile 4% Brewer’s thioglycolate solution (Difco/BD Biosciences). After 4 h the mice were killed, and cells were recovered by peritoneal lavage using 5–10 ml of ice-cold PBS/0.02% EDTA (w/v). The number of viable cells was checked by staining the cells with 0.4% trypan blue (Invitrogen Life Technologies). The neutrophil yield was determined by flow cytometry after staining with PE-conjugated anti-Gr-1 Ab (Miltenyi Biotec) and with FITC-conjugated F4/80 Ab (Serotec) after three washes in PBS/0.5% BSA/0.02% NaN₃. Before staining, nonspecific binding of Abs was blocked with the anti-Fc receptor Ab 2.4G2 (BD Pharmingen). Flow cytometric measurements were performed using FACScan (BD Biosciences). Unless stated otherwise, the cells were cultured in RPMI 1640 containing 10% FCS and penicillin/streptomycin (Invitrogen Life Technologies) for 2 h to adhere to the surface. Before their use for the experiments nonadherent cells were washed away, which resulted in >95% pure PMN cultures.

The following Abs were used in this study: rabbit antiserum against mouse cathepsin D (15); rat anti-mouse LAMP-1 and rat anti-mouse LAMP-2 (Developmental Studies Hybridoma Bank); rabbit anti-LC3 from I. Tanida and T. Ueno (Juntendo University, Tokyo, Japan); anti-lactoferrin (Upstate Biotechnology); anti-myeloperoxidase (Dianova), anti-β-tubulin (E7; Developmental Studies Hybridoma Bank); and rabbit anti-DNP (ICN Biomedicals). Alexa Fluor-conjugated secondary Abs were from Molecular Probes (Invitrogen Life Technologies).

Mice were killed between 7 wk and 12 mo after birth. Upper and lower jaws were fixed in 4% paraformaldehyde and 1% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4). Jaws were micrographed and visualized by high resolution MicroCT. Following demineralization in EDTA, jaws were postfixed in 1% OsO₄ and embedded in epoxy resin. Sections were cut parallel to the longitudinal axis of the molars and stained with methylene blue. One midsagittal section of each molar block (four per animal) was subjected to image analysis using the Leica-Qwin-Pro software (objectives ×6.3 and ×10). To this end, the area between the first and second molars, the standard interproximal area, was assessed for microbial plaque, infiltrated interdental epithelium, supraalveolar connective tissue, alveolar bone, and periodontal ligament. In addition, connective tissue attachment and bone levels were determined distal to the first molar and mesial to the second molar in each molar block (four per animal). The connective tissue attachment level was defined as the distance from cemento-enamel junction (CEJ) to the apical termination of the junctional epithelium. Bone level was defined as the distance between CEJ and bone crest. Data were analyzed by Student’s t test. Differences were considered significant at p < 0.05 (two-tailed).

After dissection of the molar blocks but before fixation, supragingival plaque samples were taken from the interdental areas (lingual aspect) of the molars using paper points. The samples were cultured aerobically and anaerobically according to routine laboratory procedures for detection of periodontal pathogens.

Latex beads (3 μm; Sigma-Aldrich) were opsonized in RPMI 1640 medium containing 10% human serum for 30 min at 37°C and were washed three times with medium. The 1–5 × 10⁵ peritoneal PMNs per well were cultured in a 24-well plate with RPMI 1640 medium (without antibiotics and FCS) to adhere to the surface. Opsonized beads (0.05% solid) in serum-free RPMI 1640 were then added. Phagocytosis was synchronized by spinning at 300 × g for 1 min. To induce internalization, cells were incubated at 37°C. After different time periods the cells were fixed with 4% paraformaldehyde in PBS for 20–30 min at room temperature and used for immunofluorescence. Before permeabilizing the cells external beads were labeled with fluorophore-conjugated goat anti-human IgG (1:500) in PBS for 1 h at room temperature.

A. actinomycetemcomitans bacteria were cocultured with peritoneal neutrophils isolated from wild-type and LAMP-2-deficient mice (ratio 1:50) at 37°C and 5% CO₂ for 2 h. Extracellular bacteria were eliminated by washing with PBS and gentamicin treatment. After different time points the cells were fixed with 4% paraformaldehyde in PBS and examined microscopically.

PMNs were cultured on coverslips for 2 h and fixed with 4% paraformaldehyde in PBS for 20–30 min at room temperature. Cells were permeabilized in PBS/0.2% saponin. Primary and secondary Abs were diluted in 3% BSA (Sigma-Aldrich) in PBS/0.2% saponin and added to the cells for 1 h. Goat anti-rabbit, anti-rat, or anti-mouse Abs conjugated to Alexa Fluor 488 or 594 (Molecular Probes) were used. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole; Sigma-Aldrich). Acidic compartments were labeled by incubating living cells in RPMI 1640 containing 20 mM HEPES (pH 7.4), and 0.1 mM DAMP (N-(3-((2,4-DNP)aminopropyl)-N-(3-aminopropyl)methylamine; Molecular Probes). The coverslips were mounted with Mowiol (Calbiochem) containing the anti-fading reagent DABCO (1,4 diazobicyclo-(2.2.2) octane; Sigma-Aldrich), and viewed with an Axiovert 200M fluorescence microscope (Zeiss) with or without an Apotome device for optical sectioning.

For electron microscopy, ultrathin sections were cut from the interdental region between the first and second mandibular and maxillary molars. Sections were stained with uranyl acetate and lead citrate and examined in a Philips EM 420. Isolated PMNs were fixed in 4% formaldehyde and 1% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) and processed for LX-112 embedding. For autophagic vacuole quantification, ultrathin sections of PMNs were scored under the microscope. The number of autophagic vacuole profiles was counted in at least 100 cell profiles per phenotype.

A total of 1 × 10⁸ Escherichia coli cells were opsonized in RPMI 1640 medium containing 10% human serum for 30 min at 37°C and were washed three times with medium. Peritoneal PMNs (5 × 10⁵ per well) were cultured in a 24-well plate with RPMI 1640 medium (without antibiotics and FCS) to adhere to the surface before 5 × 10⁶ E. coli were added. Phagocytosis was synchronized by spinning at 720 × g for 2 min. Cells were incubated at 37°C for 1 h. Phagocytosis was stopped by putting the cells on ice. Extracellular bacteria were eliminated by washing with PBS and incubation with 100 μg/ml gentamicin for 30 min. The PMNs were now cultured in RPMI 1640 containing 0.1% FCS. After various time periods the medium was removed, and 0.1% BSA in H₂O_dest was added to lyse the PMNs. The plates were frozen at −80°C before warm RPMI 1640 (without antibiotics and FCS) was added and the plates were thawed fast by incubation at 37°C. This procedure did not affect the viability of E. coli (data not shown). The 100 μl of the lysate was plated on Luria-Bertani agar plates incubated at 37°C overnight, and CFUs were counted.

A. actinomycetemcomitans (American Type Culture Collection No. 29522) was grown in tryptic soy broth (Sigma-Aldrich) with 0.1% yeast extract (BD Biosciences) and 2.5% glucose (Merck) or on Columbia blood agar plates (Oxoid) at 37°C in 5% CO₂. A total of 1 × 10⁶ peritoneal neutrophils were incubated with opsonized A. actinomycetemcomitans (ratio 1:50) in PBS at 37°C and 5% CO₂ for 120 or 240 min. Phagocytosis was synchronized by spinning the mixture at 720 × g for 2 min. Serial dilutions of the supernatant were plated on blood agar plates and cultured for 48 h, and the number of CFU was counted. The percentage of viable bacteria in each sample was then determined by comparing the number of CFU from the control sample without added neutrophils (100% viability) to the number of CFU obtained for A. actinomycetemcomitans incubated with neutrophils.

The assay was performed according to Pick and Mizel (16). Peritoneal PMNs (1 × 10⁵ per well) were cultured in a 96-well plate with RPMI 1640 medium (without antibiotics and FCS) for 2 h to allow adherence before nonadherent cells were removed by shaking the plates and washing the wells three times with 0.1-ml volumes of warm phenol red-free HBSS. A total of 2 × 10⁷ opsonized E. coli bacteria XL1 blue or 0.1 mg of zymosan was added in 100 μl of phenol red solution (140 mM NaCl, 10 mM NaH₂PO₄, 5.5 mM sucrose, 0.56 mM phenol rot (pH 7.0), sterile filtered 0.22 μm) containing 19 U/ml HRP. After 1 h at 37°C, the reaction was stopped by adding 10 μl of 1 M NaOH per well. The absorbance was measured in a microtiter plate reader at 600 nm.

We have reported earlier that LAMP-2 deficiency caused an increased postnatal lethality, a reduced weight development, and a massive accumulation of autophagic vacuoles in numerous tissues (10, 13). Because we needed to raise the LAMP-2-deficient mice after the weaning period with liquid food, we wanted to study the reasons for this impaired feeding behavior. We observed that all LAMP-2 knockout animals, which were housed in a conventional breeding facility, exhibited early onset natural periodontitis with overt migration of molars (Fig. 1, B and C), alveolar bone loss and furcation involvement (Fig. 1 E). Increased mobility of molar teeth was diagnosed in several animals at the day of sacrifice.

Microscopically, loss of connective tissue attachment level and alveolar bone crest resorption was already evident at the age of 7 wk after birth (Fig. 1,G). In the knockout animals, all molar surfaces exposed to the oral cavity proved to be covered with a layer of microbial plaque. Because the first molars erupt around day 14, this response implies that periodontitis had developed within a time period of 1 mo. Neither plaque nor signs of periodontitis could be diagnosed in the knockout animals that were supplied with amoxicillin in the drinking water (Fig. 1 M). In none of the wild-type or other transgenic animals kept in the same conventional breeding facility were signs of periodontitis and plaque development observed.

All interdental areas in the knockout mice exhibited massive amounts of microbial plaque in the region occlusal to the level of the gingiva and within the sulcus area (Fig. 1, G and H). However, neither at the light microscopical level nor at the electron microscopical level were microorganisms observed within the gingival tissues, except (very occasionally) within the very superficial layers of the sulcular epithelium (Fig. 1 I). In the older animals the biofilm did not only occupy the crevicular domain but had grown out to cover the free dental surfaces up to the level of the occlusal plane. Microorganisms were sometimes found within the dentinal tubules of the molar cusp regions free of enamel (data not shown). Overt caries lesions, however, were not detected.

Upon culturing, it appeared that none of the classical periodontal pathogens that are commonly found in the human and are associated with natural periodontitis in humans had nested. In particular, there was no colonization of A. actinomycetemcomitans, Porphyromonas gingivalis, Prevotella intermedia, or Tannerella forsytia. The dominant flora in all animals consisted of facultative anaerobic bacteria with relatively high percentage of Gram-negative rods with a slightly increased number of Actinomyces species. In histological sections the bacterial morphotypes observed were classified as cocci, rods, and filaments.

The dominant infiltrating cell type in the inflammatory lesions was the PMN (Fig. 1,H). Many of them had infiltrated the junctional and sulcular epithelia. Also within the gingival crevice (outside the gingival tissue) numerous PMNs were observed, many of them loaded with phagocytosed bacteria (Fig. 1,J). Signs of phagocytosis by PMNs or by any other cell type (e.g., macrophage-like cells) were not found within the tissue, either at the light microscopical level or at the electron microscopical level. No plasma cells were noted within the connective tissue and only a few cells belonging to the lymphocytic lineage were identified. The epithelium in the interdental region had lost its normal appearance and was characterized by extensive proliferation, widened intercellular spaces, and loss of keratinization (Fig. 1, G and H). Many small blood vessels were observed throughout the interdental epithelium and they showed fenestrations. PMNs (without showing signs of bacterial phagocytosis) were abundantly present in contact with the blood vessel wall, within the collagenous fiber framework of the gingiva, and within the epithelium. The area occupied by (infiltrated) epithelium was significantly larger for LAMP-2 knockout mice compared with wild-type mice (p < 0.001) (Fig. 1,K). Supraalveolar connective tissue was significantly more extensive in knockout mice compared with wild-type animals (p < 0.05) (Fig. 1,F). The surface area of alveolar bone in the interproximal area of the wild-type animals was larger compared with that of the knockout mice (p < 0.001) (Fig. 1,K). The surface area of periodontal ligament gave higher values for the knockout than the wild-type mice (p < 0.05) (Fig. 1 K).

Analysis of histomorphometric parameters showed considerable loss of connective tissue attachment (the distance between the CEJ to the apical termination of junctional epithelium) in knockout animals (Fig. 1,L). Also the alveolar bone crest was displaced in the apical direction (Fig. 1,L), whereas the total molar root length (CEJ-apex) was about the same (Fig. 1 L). Plaque growth and extent of connective tissue attachment loss did not show statistically significant differences between upper and lower molar regions (data not shown).

The periodontal pathology may have been caused by PMNs able to phagocytose pathogens but unable to efficiently kill the microorganisms. To investigate this, we isolated PMNs from wild-type and this LAMP-2-deficient mice. After attachment of these cells they were incubated for 1 h with E. coli cells. After 1 h phagocytosis, extracellular bacteria were killed by incubation for 30 min with gentamicin. Cells were lysed and plated on bacterial plates to estimate CFUs after 0, 30, and 60 min of subsequent incubation (Fig. 2,A). Wild-type PMNs were able to phagocytose and kill the ingested bacteria as expected. LAMP-2-deficient PMNs were significantly less effective in killing bacteria. Even 60 min after incubation, the number of viable bacteria was comparable to the number of bacteria at 0 min of wild-type PMNs (Fig. 2,A). To also analyze whether periodontally relevant pathogens are susceptible to phagocytotic killing, we isolated PMNs from wild-type and LAMP-2-deficient mice and incubated these cells with A. actinomycetemcomitans (ratio 1:50) at 37°C for 120 or 240 min, respectively. We also observed that these bacteria were less efficiently killed by LAMP-2-deficient PMNs (Fig. 2,B), and the bacteria were present intracellularly in a higher number in LAMP-2-deficient PMNs (Fig. 2, C–E). These data suggest that the phagosomal killing capacity is impaired in LAMP-2 lacking PMNs. To elucidate whether oxidative or nonoxidative killing pathways were affected we analyzed the capacity of wild-type and LAMP-2 knockout PMNs to produce oxygen radicals, which are known to be involved in bacterial killing by these cells (17, 18). PMNs from both genotypes retained the capacity to react after addition of E. coli or zymosan by producing similar amounts of oxygen radicals measured by the production of H₂O₂ (Fig. 3) and reactive oxygen species (data not shown). These data suggest that impairment of the nonoxidative, lysosomal killing pathway is mainly responsible for the reduced killing in LAMP-2-deficient PMNs.

Using electron microscopy, we observed that PMNs within the gingival crevice and PMNs isolated from LAMP-2-deficient mice were characterized by an accumulation of autophagic vacuoles (Figs. 1,J and 4,B), in contrast to wild-type cells (Fig. 4,A). Few early autophagic vacuoles were detected in wild-type PMNs. In agreement with our earlier results with PMNs in vivo (10), the LAMP-2-deficient PMNs showed a prominent accumulation of late autophagic vacuoles, containing partially degraded cytoplasmic material (Fig. 4,C). Because LC3 was shown to be a specific marker for early autophagic vacuoles (19), we performed an immunocytochemical analysis using Abs against this autophagosome marker. The staining was diffuse in wild-type cells, whereas the LAMP-2-deficient cells showed a punctuate staining (Fig. 4, D and E). This result is consistent with autophagic accumulation in LAMP-2-deficient PMNs.

The localization of LAMP-2 in vesicular structures of wild-type PMNs was easily detectable (Fig. 4,F). As expected no LAMP-2 expression was observed in LAMP-2-deficient cells (Fig. 4,F, inset). We next analyzed the lysosomal compartment using Abs against the related major lysosomal membrane protein LAMP-1. In wild-type PMNs lysosomes were distributed within the entire cytosol and also close to the plasma membrane (Fig. 4,G). Importantly, in LAMP-2-deficient PMNs lysosomes appeared enlarged and clustered in the center of the cell (Fig. 4,H). To monitor the distribution and the presence of primary or secondary granules, which are essential to supply the phagosome with bactericidal substances (20), we stained PMNs with Abs against myeloperoxidase (Fig. 4, I and J), a constituent of primary granules, and against lactoferrin, a constituent of secondary granules (Fig. 4, K and L). We did not observe differences in the distribution, number and size of primary granules. Secondary, lactoferrin-containing granules were less frequently found close to the plasma membrane and showed a reduced colocalization with LAMP-1 (Fig. 4 L) in LAMP-2 knockout cells, which may suggest an impaired function and/or transport of these vesicles.

Our data suggested that impaired nonoxidative killing processes may be the cause for the reduced bacterial killing capacity of LAMP-2 knockout PMNs. To analyze this further, we made use of the ability of PMNs to spontaneously phagocytose latex beads in culture. The ingested latex bead-phagosomes can be easily monitored microscopically (Fig. 5). Electron microscopy confirmed that LAMP-2-deficient PMNs were capable of phagocytosing such beads with similar efficiency as wild-type cells (Fig. 5, A and B). To analyze the maturation defects in more detail, we stained the cells after 1 h of phagocytosis with proteins localized in the lysosomal compartment. Whereas the majority of phagosomes in wild-type cells recruited LAMP-1 (Fig. 5,C), and cathepsin D (Fig. 5,E) there was a significant decrease in the number of LAMP-2-deficient phagosomes positive for either LAMP-1 (Fig. 5,D) or cathepsin D (Fig. 5,F). Under the given experimental conditions, 60–70% of wild-type latex bead phagosomes were positive for lysosomal markers. In contrast, only ∼30% of LAMP-2-deficient phagosomes contained lysosomal components (Fig. 5, G and H). In agreement with these results, we also observed a decreased acidification in the LAMP-2 knockout phagosomes using immunolabeling of the acidotropic compound DAMP (Fig. 6).

Lysosomes play a very important role in the oxygen-independent killing of bacteria, which is believed to be an important mechanism in the oxygen-deprived periodontal pocket (8). Naturally occurring periodontal disease normally does not happen in laboratory mice (21). We now demonstrate that in mice lacking LAMP-2 periodontitis represents a very striking phenomenon. The early onset periodontal disease noted in the present study was clearly related to an accumulation of plaque. No evidence was found (in any of the specimens) for invasion of microorganisms into the deeper tissues of the gingiva. Thus tissue damage and attachment loss were not likely to be directly caused by bacterial invasion but more likely the result of inflammation.

We recorded ample evidence of plaque formation in the interproximal areas of the knockout animals. In several cases plaque had overgrown even the occlusal surfaces. Although our microbiological data were not very detailed, our observations suggest a mixed flora in the oral cavity of the LAMP-2 knockout animals. Human periodontopathogens were not found.

Despite the fact that knockout mice were housed under the same conditions as their wild-type littermates, in the latter animals bacterial accumulations were never noticed. The prevention of plaque accumulation and tissue breakdown in the mice receiving antibiotic treatment proves that periodontitis in LAMP-2 knockout mice is bacteria-related. Our observations also prove that LAMP-2 associated cell functions are pivotal in self-cleaning properties of the oral cavity. Although we cannot exclude the possibility that the observed periodontitis was, at least partially, due to systemic effects arising from LAMP-2 deficiency, the most likely reason for the huge outgrowth of dental plaque in the LAMP-2 knockout mice is that bacterial killing by PMNs (analyzed in this study: E. coli and A. actinomycetemcomitans) was severely hindered.

In humans, disorders of neutrophil function are commonly associated with severe periodontal destruction (7, 22, 23). Of course other immune cells might have been affected as well by the deficiency. In this respect, dendritic cells are of interest, which capture exogenous Ags for eventual processing in endosomes-lysosomes (24). LAMP-2a facilitates MHC class II presentation of cytoplasmic Ags. Decreased display of cytoplasmic epitopes via class II molecules was observed in cells with diminished expression of LAMP-2 (25).

An impaired function of macrophages in bacterial killing in LAMP-2-deficient mice is unlikely because these cells were only seldom observed within the affected tissue. We could also recently show that phagosomal maturation is not affected in LAMP-1- or LAMP-2-deficient macrophages (14, 26).

PMNs are the most abundant immune cells in the inflammatory gingivial sites of patients with periodontitis, and their pathogenic role in this setting has been suggested (27). The role of PMNs in innate immunity and the specific role of the lysosomal compartment in these cells is underscored by congenital defects such as Chediak-Higashi syndrome (4) and Papillon-Lefevre syndrome (28) in which lysosomal secretion events and lysosomal proteolysis is impaired, respectively. Our observations suggest that LAMP-2-associated functions in PMNs are pivotal in the self-cleaning properties of the oral cavity. They help to orchestrate the natural defense against oral biofilm formation.

Whereas in LAMP-2-deficient PMNs the distribution and fusion of primary granules with latex bead phagosomes is apparently unaffected (data not shown), the cellular localization of lactoferrin-positive granules was changed and these granules showed a reduced colocalization with lysosomal markers, suggesting a disturbed biogenesis, traffic, or function of a subset of granules. Both types of granules contribute to the killing of bacteria (29) and it is likely that LAMP-2 contributes to the fusion of granules with phagosomes.

Phagosome-lysosome fusion is essential for efficient degradation of internalized pathogens. Fusion with lysosomes results in delivery of an assortment of luminal and membrane proteins to phagosomes. In a recent study we showed that in LAMP-1 and LAMP-2 double-deficient fibroblasts, phagosomes acquired the early endosome markers Rab5 and PI3 phosphate, but failed to recruit Rab7 and did not fuse with lysosomes. We attributed the deficiency to impaired organellar motility toward the cell center (14). We proposed that LAMPs might directly or indirectly assist the movement of phagosomes toward the cell center. Elimination of lysosome transport by disruption of microtubules (30) or by interference with RILP or dynein function (31) impairs contact and fusion between lysosomes and phagosomes. In contrast to fibroblasts, where either LAMP-1 or LAMP-2 is required for lysosomal motility and phagosomal maturation, in PMNs the lack of LAMP-2 only seems to be enough to interfere with the maturation of phagosomes. As already indicated by the accumulation of autophagic vacuoles in LAMP-2-, but not in LAMP-1-deficient PMNs (32), the lack of LAMP-2 cannot be compensated by LAMP-1. Similar to the successful maturation of autophagic vacuoles, the destruction of phagocytosed bacteria requires the subsequent fusion with early and late endosomes and lysosomes. We show that the later maturation events are disturbed in LAMP-2-deficient PMNs, leading to impaired bacterial clearance and development of periodontal disease in LAMP-2 knockout mice. Further analyses in PMNs and LAMP-deficient cells will be required to determine the molecular mechanisms for the impaired maturation and disturbed lysosomal motility.

In conclusion, our data indicate that LAMP-2 is critically required for the maturation process of phagosomes in PMNs. A lack of LAMP-2 leads to a reduced maturation and to a reduced bactericidal activity. This functional defect may be directly associated with the increased susceptibility of LAMP-2 knockout mice to develop periodontal disease.

We thank Martijn van Steenbergen (University of Amsterdam), Arja Strandell (University of Helsinki), and Marlies Rusch (University of Kiel) for assistance in the completion of this project.

The authors have no financial conflict of interest.