Osteoimmunolgy involves the interaction of the immune system with skeletal elements. This interaction can lead to the formation of osseous lesions. To investigate how the acquired immune response could contribute to osteolytic lesions, we injected the periodontal pathogen Porphyromonas gingivalis adjacent to calvarial bone with or without prior immunization against the bacterium. Activation of the acquired immune response increased osteoclastogenesis and decreased coupled bone formation. The latter was accompanied by an increase in nuclear translocation of the transcription factor FOXO1 in vivo, increased apoptosis of bone-lining cells measured by the TUNEL assay and number of activated caspase-3 positive cells and a decrease in bone lining cell density. Further studies were conducted with MC3T3 osteoblastic cells. Apoptosis and increased FOXO1 DNA binding activity were induced when a combination of cytokines was tested, IL-β, TNF-α, and IFN-γ. Knockdown of FOXO1 by small interfering RNA significantly reduced cytokine stimulated apoptosis, cleaved caspase-3/7 activity and decreased mRNA levels of the proapoptotic genes, TNF-α, FADD, and caspase-3, -8, and -9. These results indicate that activation of the acquired immunity by a periodontal pathogen reduces the coupling of bone formation and resorption. This may occur by enhancing bone lining cell apoptosis through a mechanism that involves increased FOXO1 activation. These studies give insight into inflammatory bone diseases such as periodontal disease and arthritis were the formation of lytic lesions occurs in conjunction with deficient bone formation and activation of an acquired immune response.
Osteoimmunolgy is a recently emerged interdisciplinary field that encompasses both osteology and immunology (1). The rapid emergence of this field can be attributed to pathologic importance of accelerated bone loss seen in several inflammatory bone diseases such as multiple myeloma, rheumatoid arthritis, diabetes mellitus, lupus erythematosus, and periodontal diseases (2). Recently, it has been shown that the immune and skeletal systems not only respond to common cytokines but also have similar signaling molecules, transcription factors, and membrane receptors (3).
In physiological conditions, bone resorption is followed by osteoblast-mediated bone formation in a process termed “coupling” (4). This occurs in discrete bone metabolic units (4). Failure to form bone following resorption is a principal cause of osteoporosis, a skeletal disorder characterized by low bone mass and increased skeletal fragility (5). The importance of the uncoupling phenomenon in osteoporosis is illustrated by its treatment with intermittent parathyroid hormone (PTH)3 injection (6). Although PTH initially causes bone resorption it represents an anabolic treatment for osteoporosis because it enhances coupled bone formation resulting in a net increase in bone mass (6).
A number of inflammatory disorders affect bone by reducing bone mass. They include osteoarthritis, rheumatoid arthritis, periodontal disease, and lytic lesions from cancers that target bone such as multiple myeloma (7). These disorders are characterized by increased accumulation of inflammatory cells and cytokines such as TNF-α, IL-1β, IFN-γ, and RANKL, which are known to directly or indirectly promote recruitment of osteoclast progenitors and the formation and activity of osteoclasts thereby stimulating bone resorption (2, 8). The osteolytic lesions are typically characterized by increased activity of osteoclasts and decreased activity of osteoblasts (7, 9).
The mechanisms by which leukocytes contribute to bone resorption have received considerable attention while those responsible for “uncoupling” are not as well understood. It is possible that the inflammatory events associated with bone resorption also contribute to reduced bone formation. In many lytic lesions, cells of both the acquired and innate immune response are present. T and B lymphocytes and monocytes are thought to contribute to the pathogenesis of erosive arthritis and several cytokines have been implicated including TNF-α, IFN-γ, IL-6, and IL-1β (1, 2, 3). Recent evidence from animal studies suggests that both innate and acquired arms of the immune response participate in periodontal disease (10). Moreover, human studies also indicate that cytokines associated with both arms of the immune response are elevated in gingival crevicular fluid or gingival tissue of individuals with periodontal disease (11, 12).
Osteoblast survival is thought to be an important aspect of bone coupling (13, 14, 15, 16). Osteoblast cell density is dependent on the number of osteoblasts generated from mesenchymal progenitors minus the cells removed by apoptosis (17). Biological factors that stimulate bone formation such as bone morphogenetic proteins, growth factors, and intermittent PTH release reduce osteoblast apoptosis and lead to enhanced bone formation (13, 15, 16). Conversely, increased apoptosis of osteoblastic cells is also linked to a decrease in bone formation. Mice with genetic deletion of PTHrP have increased numbers of apoptotic osteoblastic cells and fewer osteoblasts (18, 19). The periosteal cells lining the bone surface is an important source of osteoblast precursors (20, 21). Α possible mechanism through which immune response could interfere with bone coupling is by the production of factors that induce apoptosis of periosteal or osteoblastic cells. Though this concept has received relatively scant attention, studies of multiple myeloma, a lytic bone disease, support this notion where uncoupling and increased osteoblast apoptosis is associated with increased TNF-α, IFN-γ, and IL-1β (22).
Apoptosis is influenced by transcription factors. FOXO1 is a transcription factor that belongs to the forkhead-O family (FOXO1, FOXO3, and FOXO4) that regulates cell death, inhibits cell cycle progression, and modulates the response to oxidative stress (23, 24, 25, 26, 27). In addition to being proapoptotic, FOXO1 stimulates blood vessel organization in endothelial cells (28), muscle wasting in myocytes (29), and inhibition of adipocyte differentiation in fat tissue (30). The role of FOXO1 in osteoblast apoptosis has not been previously studied.
The goal of studies undertaken here was to investigate whether activation of the acquired immune response increases uncoupling of bone formation in response to a periodontal pathogen. The results indicate that activation of the acquired immune response in immunized mice causes enhanced uncoupling and that this occurs in the context of increased FOXO1 nuclear translocation, greater apoptosis of bone lining cells and decreased cell numbers. In vitro, a combination of innate and acquired immune cytokines induced apoptosis and caspase activity in osteoblastic cells that is significantly reduced by FOXO1 knockdown with small interfering RNA (siRNA). siRNA also reduced expression of proapoptotic genes induced by these cytokines. Taken together, the data suggest that acquired immune response to a periodontal pathogen may contribute to lytic bone diseases by promoting apoptosis of bone lining cells through a mechanism that involves FOXO1 activation.
Materials and Methods
Eight-week-old CD-1 mice were purchased from Charles River Laboratories. Mice were immunized and subsequently challenged with P. gingivalis as described below. Anesthesia was achieved with ketamine (80 mg/kg of body weight) and xylazine (10 mg/kg) in sterile PBS delivered i.p. All animal protocols were approved by the Institutional Animal Care and Use Committee of Boston University Medical Center. Animals had free access to food and water and were maintained under a 14-h on/10-h off light cycle. For all experiments, measurements were made by a blinded examiner and confirmed with a second blinded examiner. Unless stated n = 6 per group.
Bacterial inoculation and experimental design
Broth-grown P. gingivalis strain 381 in logarithmic growth phase was collected and suspended in sterile PBS. For immunization, bacteria were fixed with 1% paraformaldehyde for 4 h just before injection. An inoculum of 2.5 × 108 bacteria in 50 μl of sterile PBS was injected s.c. into the dorsal dermis of animals once weekly for three consecutive weeks. Animals undergoing this protocol represent the immunized group in this study, and we have previously shown that this protocol results in activation of the acquired immune response 1 wk after the third inoculation (31). Control animals were sham immunized with an equal volume of sterile PBS and are referred to as the nonimmunized group. Four weeks after the first inoculation, both the immunized and the nonimmunized animals were inoculated with live P. gingivalis in the scalp as described previously (8, 32). In brief, mice were inoculated by supraperiosteal injection at the midline of the scalp between the ears with P. gingivalis (2 × 108 bacteria), and animals were subsequently euthanized 5 or 8 days later for histologic analysis. In some cases, vehicle alone, sterile PBS, was inoculated into the scalp.
Preparation of histologic specimens and histomorphometric analysis
Calvariae with intact soft tissue were prepared for fixation, embedded in paraffin, and sectioned as previously described (8). To assess the bone lining cell density, 5-μm H&E-stained sections were examined at ×1,000 magnification in the area of greatest inflammation in 7 to 10 fields, which were typically between the coronal and the occipital sutures. The number of periosteal cells was determined in each field, and the results were expressed as mean values per mm bone length. Similarly the number of osteoclasts were counted in TRAP-stained sections from tissue obtained 5 days after injection of bacteria into the scalp using Image ProPlus software (Media Cybernetics) as previously described (8). The percent eroded bone surface was measured between and including the coronal and occipital sutures as described (8). Bone formation was also assessed as previously described by quantifying the amount of bone formed adjacent to a reversal line in TRAP stained sections (32, 33). To assess the effect of immunization on bone coupling a ratio of bone resorption and bone formation was calculated for each animal and the data is presented as mean ± SEM.
In vivo apoptosis assay
Apoptotic bone-lining cells were examined by in situ terminal deoxynucleotidyltransferase-mediated dUTP (TdT)-biotin nick end labeling (TUNEL) assay using a DeadEnd fluorometric kit (Promega) following the manufacturer’s instructions. After staining, the slide was mounted in mounting medium containing nuclear stain 4′, 6-diamidino-2-phenylindole (Vector Laboratories). The number of TUNEL-positive periosteal cells between the occipital and coronal sutures was counted at ×1,000 magnification. Apoptotic periosteal cells were counted in the two cell layers adjacent to calvarial bone. Fluorescent and phase contrast images (original magnification ×400) were digitally captured. The data presented are mean TUNEL-positive cells ± SEM. The number of activated caspase-3 immunopositive bone lining cells was also measured. Caspase-3 immunostaining was conducted with 1/100 dilution of specific Ab for cleaved caspase-3 (Cell Signaling Technology) followed by incubation with a biotinylated anti-rabbit secondary Ab (Chemicon) and avidin-biotin complex, which was localized with 3,3′-diaminobenzidine tetrahydrochloride as a chromogen (Vector Laboratories). Immunopositive cells were counted as described above at ×1,000 magnification. There was no nonspecific binding determined by incubation with matched control Ab.
FOXO1 nuclear translocation
FOXO1 nuclear translocation was detected by confocal laser scanning microscopy (Axiovert-100M, Carl Zeiss). Primary Ab to FOXO1 (Santa Cruz Biotechnology) was detected by a Cy5-tagged secondary Ab with propidium iodide nuclear stain in the mounting medium. The length of calvarial bone between the occipital and coronal sutures was scanned for the FOXO1 presence of FOXO1 in the nuclear compartment by comparing Cy5 images with FOXO1 immunostaining, nuclear stain with propidium iodide, and a corresponding merged image. Matched control Ab was used as a negative control.
In vitro experiments
Murine osteoblastic MC3T3 cells were purchased from American Type Culture Collection and cultured as described (34). Cells were cultured in DMEM plus 10% FBS and changed to low serum (medium supplemented with 0.25% FBS) at least 24 h before stimulation with recombinant murine TNF-α, IL-1β, and IFN-γ (5 ng/ml each, 48 h) (R&D System). Nuclear and cytoplasmic proteins were isolated from cell lysates and protein concentration measured as described previously (35). The nuclear fraction was assayed for FOXO1 DNA binding activity (Active Motif) following the manufacturer’s instructions. The cytoplasmic fraction was tested for cleaved caspase-3/7 activity measured with a Caspase-Glo 3/7 kit from Promega following the manufacturer’s instructions. For apoptosis, assays cells were cultured in DMEM plus 0.25% FBS and transfected with or without murine FOXO1 specific siRNA (5 nM in HiPerfect transfection reagent) or scrambled siRNA (Qiagen) and then incubated with cytokines TNF-α, IL-1β, and IFN-γ (5 ng/ml each, 48 h). Apoptosis was measured by cytoplasmic histone-associated DNA by ELISA using a kit from Roche Diagnostics (36). Assays were conducted three times to obtain mean values ± SEM. The data are shown as percent of the group with highest value so that the three experiments could be combined.
Histologic sections were examined under blind conditions by one examiner, and measurements were independently confirmed by a second examiner. The mouse was the unit of measurement and for each data point there were six mice per group unless stated otherwise. Data are presented as means ± SEMs. Statistical analysis between immunized and nonimmunized groups for a given parameter was established by Student’s t test, at p < 0.05. Three separate experiments were conducted for real-time PCR and the data from three experiments was combined and expressed as the mean ± SEM. Statistical analysis was assessed by ANOVA with Sheffe’s posthoc test with significance set at p < 0.05.
Activation of the acquired immune response enhances bone resorption induced by P. gingivalis
We previously demonstrated that immunization of mice with P. gingivalis leads to activation of the immune response 4 wk later as assessed systemically by Ab titer (37) or by the increased production of IFN-γ when immunized mice were challenged by injection into the scalp (31). To examine the impact of the acquired immune response on bone resorption stimulated by a periodontal pathogen, the impact of P. gingivalis on osteoclast numbers was measured 5 days later (Fig. 1,a, Supplemental Fig. 1).4 In the absence of bacterial stimulation few if any osteoclasts were observed (data not shown), while injection of P. gingivalis stimulated osteoclastogenesis. The number of osteoclasts was 2.3-fold higher in immunized compared with nonimmunized mice (p < 0.05). The functional effect of increased osteoclast numbers was evaluated by measuring the eroded bone surface. The percent eroded bone surface resulting from injection of P. gingivalis was increased by 1.6-fold when the acquired immune response was induced by immunization compared with when it was not (Fig. 1 b).
The impact of the acquired immune response on bone formation was assessed by using the reversal line as a landmark in TRAP stained sections obtained 8 days following injection of bacteria (Supplemental Fig. 2), a time point when bone formation is at a high level in this model (8). The amount of bone formation following injection of bacteria was 3.3-fold higher in the nonimmunized compared with immunized mice following injection of P. gingivalis adjacent to the calvarial bone (Fig. 1,c). The degree of coupling was then assessed by dividing the amount of bone formed by the percent eroded bone surface or osteoclast numbers. The amount of coupling was 6.5-fold higher in the nonimmunized compared with immunized mice when assessed by eroded bone surface (Fig. 1,d). When coupling was calculated in terms of bone formation divided by osteoclast numbers it was 8.9-fold greater in the nonimmunized group compared with immunized group (Fig. 1,e). This represents an ∼85–90% reduction in bone coupling following bacterial injection when the acquired immune response is stimulated (Fig. 1, d and e).
The acquired immune response enhances bone lining cell apoptosis, increases FOXO1 nuclear translocation, and reduces periosteal cell density
The cells that participate in calvarial bone formation are thought to be derived from the bone-lining cells largely in the periosteum (38). Bacterial infection may limit reparative bone formation through a mechanism that involves apoptosis (14). The TUNEL assay was performed 5 days after injection of P. gingivalis to determine whether bacteria enhanced apoptosis of periosteal bone cells and whether it was increased by bacterial activation of the acquired immune response. There was little or no apoptosis detected in the absence of bacterial stimulation (data not shown). In nonimmunized mice P. gingivalis induced apoptosis in periosteal cells, which was enhanced 2.8-fold by stimulating the acquired immune response (p < 0.05) (Fig. 2,a). The results of the TUNEL assay are supported by results measuring cleaved caspase-3, because activation of caspase-3 is associated with apoptosis of osteoblasts (39). The number of cleaved capase-3 immunopositive bone-lining cells following supraperiosteal injection of P. gingivalis was increased 2.3-fold in immunized compared with nonimmunized mice (p < 0.05) (Fig. 2 b).
We recently showed that in fibroblasts the anti-apoptotic bias of the cell is overcome by induction of FOXO1 and its target genes (35). To establish whether FOXO1 levels were induced by P. gingivalis and enhanced by the acquired immune response, we measured the nuclear translocation of FOXO1 in periosteal cells by laser scanning confocal microscopy. Nuclear translocation reflects the activation status because FOXO1 is rapidly transported to the nuclear compartment in the activated state and upon deactivation is quickly transported out of the nucleus (26). Confocal microscopy avoids pitfalls associated with standard immunofluorescence because the images are taken at a focal plane that bisects the nuclear compartment. An example of a periosteal cell lining bone that exhibits FOXO1 nuclear translocation is evident in Fig. 2,c. Periosteal cells that were positive for FOXO1 nuclear translocation typically exhibit increased cytoplasmic expression in addition to nuclear translocation consistent with an overall increase in expression. Conversely, cells without FOXO1 nuclear translocation typically exhibited only faint FOXO1 immunofluorescence in the cytoplasm indicating a low basal level of expression as indicated. Matched control IgG was negative. Quantitative analysis of periosteal cells with FOXO1 nuclear translocation demonstrated that the number was increased more than 3.6-fold in the immunized group compared with nonimmunized group (p < 0.05) (Fig. 2 d).
The ultimate effect of increased bone lining cell apoptosis was evaluated by determining the bone lining cell density 5 and 8 days postinoculation. The bone lining cell density was significantly reduced in immunized group when compared with nonimmunized group in both the time points. Five days after the inoculation of P. gingivalis, the bone lining cell density was reduced by 23% in immunized group when compared with nonimmunized group (Fig. 3,a) (p < 0.05). By day 8 there was an even greater reduction, 37%, when the acquired immune response was activated (Fig. 3 b) (p < 0.05).
Cytokines induced FOXO1 and its role in osteoblastic cell apoptosis
To further investigate the impact of FOXO1, in vitro assays were conducted with murine MC3T3 cells of the osteoblastic lineage. When tested alone, TNF did not induce increased FOXO1 DNA binding activity at doses that did in fibroblastic cells (35, 36). However, when TNF-α was combined with IL-1β and IFN-γ, there was increased FOXO1 DNA binding (Fig. 4,a; p < 0.05). IL-1β and IFN-γ together had no effect (p > 0.05). To test the impact of FOXO1 knockdown with siRNA, DNA binding activity was measured. Silencing FOXO1 reduced FOXO1 DNA binding activity in cytokine stimulated cells by ∼60% compared with scrambled siRNA. The combination of TNF-α, IL-1β, and IFN-γ also increased FOXO1 mRNA levels by 1.5-fold when compared with unstimulated, TNF-α, or IL-1β/IFN-γ stimulated cells (p < 0.05) (Fig. 4 b). Knockdown of FOXO1 by siRNA reduced FOXO1 mRNA levels also by ∼60% in cytokine-stimulated MC3T3 cells vs scrambled siRNA agreeing well with the level of knockdown in DNA binding activity.
The role of innate and acquired immune cytokines in apoptosis of osteoblastic MC3T3 cells was investigated further in vitro. TNF-α alone and IL-1β and IFN-γ together had almost no effect on apoptosis. However, the stimulation with a combination of IL-1β, TNF-α, and IFN-γ induced apoptosis by ∼2.5-fold compared with unstimulated, TNF-α alone, and IL-1β alone stimulated cells (p < 0.05). The role of FOXO1 in apoptosis was assessed by knockdown with RNA interference. FOXO1 siRNA significantly reduced apoptosis of MC3T3 cells by ∼60% in cells stimulated with TNF-α, IL-1β, and IFN-γ compared with scrambled siRNA (p > 0.05; Fig. 5,a). Similarly, the combination of cytokines stimulated caspase-3/7 activity by 2-fold compared with untreated cells. Knockdown of FOXO1 reduced caspase-3/7 activity in cytokine stimulated cells by ∼60% (p < 0.05) while scrambled siRNA had no effect (p > 0.05; Fig. 5 b).
To further explore mechanisms by which FOXO1 may contribute to osteoblast apoptosis, several genes of interest such as TNF-α, FADD, and caspase-3, -8, and -9 were examined by real time PCR (Fig. 6). Cells were stimulated by a combination of TNF-α, IL-1β, and IFN-γ. mRNA levels of TNF-α were increased 17-fold, FADD by 2.2-fold, caspase-3, 2.2-fold, caspase-8, 1.7-fold, and caspase-9, 1.6-fold compared with unstimulated cells. Knockdown of FOXO1 siRNA significantly reduced the mRNA levels of each in cytokine-stimulated cells. TNF-α was reduced by 52%, FADD by 61%, caspase-3 by 64%, caspase-8 by 53%, and caspase-9 was reduced by 42% when compared with cells treated with scrambled siRNA (p < 0.05) (Fig. 6). The scrambled siRNA did not have a significant effect compared with cells not transfected with siRNA (p > 0.05).
The objective of the studies described here was to investigate further the impact of adding the acquired immune response to P. gingivalis-host interactions in a well-controlled model (31). The murine calvarial model used in this study is especially suited for examining the impact of host-bacterial interactions on bone and allows for delivery of a bacterial inoculum of known size that induces a well-defined spatiotemporal sequence of events (40, 41). We report in this study that immunization significantly enhanced P. gingivalis-induced osteoclast numbers and eroded bone surface. We also show, for the first time, that acquired immunity potentiates the apoptotic effect of P. gingivalis on periosteal bone lining cells as measured by the TUNEL assay and number of cleaved caspase-3 immunopositive cells. The increased apoptosis is consistent with higher mRNA levels of proapoptotic genes TNF-α, fas ligand and caspase 3 that are induced by P. gingivalis in immunized mice compared with nonimmunized mice in this model (31). The greater loss of bone-lining cells may play a critical role in reparative bone formation following bone loss because it may deplete the precursor pool from which mature osteoblasts are generated and lead to reduced bone formation (33). This is supported by findings that when the acquired immune response was activated, there was a significant reduction in the number of periosteal cells and reduced bone formation. The latter occurred even in the presence of greater bone resorption, indicating that a major effect of the acquired immune response was on osseous coupling.
To gain insight into mechanisms by which the acquired immune response could enhance osteoblastic cells apoptosis, we examined FOXO1 nuclear translocation in periosteal cells in vivo and in MC3T3 cells in vitro. MC3T3 cells are considered preosteoblastic and were derived from the periosteal lining of murine calvaria (42). Immunization by s.c. injection of the periodontal pathogen P. gingivalis led to significantly enhanced FOXO1 nuclear translocation in periosteal cells when P. gingivalis was injected adjacent to calvarial bone. We then determined whether FOXO1 DNA binding activity was stimulated in MC3T3 cells by cytokines. Because we previously determined the immunization of mice against P. gingivalis significantly increased the expression of TNF-α, IL-1β, and IFN-γ, these cytokines were tested in combination. The combination was needed to significantly increase apoptosis consistent with results previously reported (43) and to increase caspase-3/7 activity. When FOXO1 was knocked down, there was an equivalent reduction in the capacity of the cytokines to stimulate apoptosis and caspase-3/7 activity. The knockdown of FOXO1 also resulted in significant inhibition of proapoptotic genes such as TNF-α, FADD, and caspase-3, -8, and -9.
Immune mediated inflammatory bone disease involves bone destruction and formation of lytic lesions and currently the mechanism of their pathogenesis has not been clearly identified (44). Results presented in this study indicate that a significant contribution of the acquired immune response to formation of lytic lesions is the inhibition of coupling through diminished bone formation. The observation that cytokines of both innate and acquired immune system are required to induce osteoblast apoptosis can be explained by the fact osteoblasts are more resistant to apoptotic stimuli compared with many other cell types (45).
FOXO1 has been recently shown to play an important role in oxidative stress by inducing the expression of genes such as superoxide dismutase that reduce oxidative damage (46). Thus, the role of FOXO1 as a proapoptotic factor at first glance appears paradoxical. However, it has been proposed that under conditions where activation is high or is prolonged, FOXO1 then functions as a proapoptotic factor (47). This may be the case when the acquired immune response is activated by bacterial infection. Thus, under conditions of a strong host response that is prolonged, such as in periodontal disease where infection leads to both acquired and innate immune-mediated bone loss, arthritis, or multiple myeloma, both arms of the immune response may cause high and prolonged levels of FOXO1 activation, apoptosis of bone cells and reduced coupling. Chronic exposure to P. gingivalis in periodontal lesions may exacerbate periodontal disease progression by inducing an acquired immune response that limits reparative bone formation. Similarly an acquired immune response to P. gingivalis could potentially aggravate a systemic diseases where P. gingivalis has been implicated such as cardiovascular disease.
We thank Alicia Ruff for administrative support in preparing this manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work is supported by Grants DE018307 and DE017332 from National Institutes of Dental and Craniofacial Research.
Abbreviations used in this paper: PTH, parathyroid hormone; siRNA, small interfering RNA; TUNEL, terminal deoxynucleotidyltransferase-mediated dUTP (TdT)-biotin nick end labeling.
The online version of this article contains supplementary material.