B lymphocytes express multiple TLRs that regulate their cytokine production. We investigated the effect of TLR4 and TLR9 activation on receptor activator of NF-κB ligand (RANKL) expression by rat spleen B cells. Splenocytes or purified spleen B cells from Rowett rats were cultured with TLR4 ligand Escherichia coli LPS and/or TLR9 ligand CpG-oligodeoxynucleotide (CpG-ODN) for 2 days. RANKL mRNA expression and the percentage of RANKL-positive B cells were increased in rat splenocytes challenged by E. coli LPS alone. The increases were less pronounced when cells were treated with both CpG-ODN and E. coli LPS. Microarray analysis showed that expressions of multiple cyclin-dependent kinase (CDK) pathway-related genes were up-regulated only in cells treated with both E. coli LPS and CpG-ODN. This study suggests that CpG-ODN inhibits LPS-induced RANKL expression in rat B cells via regulation of the CDK pathway.
cyclin-dependent kinase inhibitor
epidermal growth factor receptor
glyceraldehyde 3-phosphate dehydrogenase
nuclear factor kappa-light-chain-enhancer of activated B cells
protein kinase C, alpha
receptor activator of NF-κB ligand
tumor necrosis factor super family, member 6
Not only in immune defense against microbial insults, but also in bone pathogenesis, immune responses to gram-negative bacterial infections, including periodontitis, involve activated B lymphocytes [1-3]. Our studies and those of others have demonstrated that B lymphocytes produce RANKL in bone resorptive lesions of periodontal disease [3-5], and the excess RANKL shifts the balance of bone metabolism towards catabolism, resulting in pathological bone resorption . In order to design interventional strategies targeting amelioration of bone resorption in such situations, it is essential to clarify the mechanism(s) of control of B-cell-associated bone pathogenesis. While TLR signaling pathways play an important role in regulating B cell functions [7, 8], including cytokine production, phagocytosis, and apoptosis , little is known about TLR signaling in the control of B cell-mediated bone pathogenesis.
Although RANKL up-regulation is usually considered to be induced by pro-inflammatory cytokines, LPS from gram-negative bacteria can also directly increase amounts of RANKL mRNA in osteoblast- and osteoclast-lineage cells [10, 11]. Studies have demonstrated that activation of TLR2 or TLR4 results in RANKL-dependent osteoclastogenesis in rheumatoid arthritis synovium [12, 13]. In addition, upon TLR9 ligation CpG-ODN reportedly induces osteoblast osteoclastogenic activity . However, other studies have shown that activation of TLRs (specifically TLR4 and TLR9) in early osteoclast precursors results in inhibition of RANKL-induced osteoclast differentiation via IL-12 . In human osteoclast precursor cell culture models, TLR ligands inhibit RANK expression by down-regulating cell surface expression of macrophage colony-stimulating factor receptor c-Fms, thereby suppressing osteoclastogenesis . These paradoxes suggest that the role of TLRs in regulation of RANKL expression needs to be investigated and interpreted according to cell types, developmental stages and environmental factors.
Lymphocyte infiltration is a typical characteristic of progressive inflammatory lesions [17, 18]. Our previous studies demonstrated that T and B lymphocytes are the primary cellular sources of RANKL in such lesions . However, few studies have attempted to determine the relationship between multiple TLR activation and RANKL expression in B lymphocytes. As components of gram-negative bacteria, LPS and CpG-ODN trigger TLR4 and TLR9, respectively, at the site of infection; however, which of these components predominates can change during the course of an infection. Knowledge about the effect(s) of the interaction between TLR4 and TLR9 signaling on RANKL expression by B cells would be useful in understanding the pathogenesis of bone resorptive diseases and developing new treatment methods. In the current study, we investigated the effects of co-activation of TLR4 and TLR9 on RANKL production in rat spleen B cells and examined the signaling pathways involved in such effects.
MATERIALS AND METHODS
Rat strain and culture of splenocytes
Experiments were performed with inbred heterozygous normal Rowett rats (Rnu/ + , female, 2 to 3 months old) maintained under pathogen-free conditions in laminar flow cabinets. Experiments using these animals were approved by the Forsyth Institute's Internal Animal Care and Use Committee. The rats were killed in a CO2 chamber and single-cell suspensions of splenocytes obtained by dispersing spleen tissues through a 60-gauge stainless steel screen. Erythrocytes were removed by ACK lysing buffer (Lonza, Hopkinton, MA, USA). The isolated splenocytes were adjusted to 1.0 × 106/mL and added to either 96-well (200 µL/well) or 6-well plates (4 mL/well) in RPMI complete medium containing 10% FCS, 100 U/mL penicillin, 100 mg/mL streptomycin, 2 mM L-glutamine and 2.5 µg/mL amphotericin B (Hyclone, Thermo Fisher Scientific, Logan, UT, USA) and 50 µM 2-ME. The cells were cultured at 37°C in a humidified incubator with 5% CO2. E. coli LPS (strain O55:B5, Sigma–Aldrich, St Louis, MO, USA) were used as TLR4 agonist and rat stimulatory CpG-ODN (5′-GAGAACGCTCGACCTTCGAT-3′) were used as TLR9 agonist. This ODN was prepared and tested for purity by polyacrylamide gel electrophoresis (Ransom Hill Bioscience, Ramona, CA, USA). A non-stimulatory scrambled ODN (5′-GAGACCATGACCCTGTCAGT-3′) was used as control. Both ODNs had been tested previously in an athymic rat lymph node cell stimulation assay in which only addition of CpG-ODN resulted in stimulation of B cells . Cultured splenocytes were treated with various concentrations of E. coli LPS and/or CpG-ODN for the indicated times, after which they were collected for further analysis.
Reverse transcription-polymerase chain reaction
Total RNA was extracted from the cultured cells using a Purelink RNA mini kit (Life Technology, Carlsbad, CA, USA) according to the manufacturer's instructions. Isolated mRNA (0.1 µg each) was reverse transcribed into cDNA using the SuperScript II reverse transcription system in the presence of random primers (Invitrogen, Carlsbad, CA, USA). The resultant cDNA was amplified by PCR using gene-specific primer pairs with Taq DNA polymerase (Life Technology) as described by the manufacturer. The primer sequences used for amplification were as follows: TLR4: forward 5′-GGAATACCTGGACTTTCAGCAC-3′and reverse 5′-TGTTGCAGTATTCCTTTGGATG-3′ (423 bp); TLR9: forward 5′-AACAAGCTGGACCTGTACCATT-3′ and reverse 5′-GATGAATCAGGCTTCTCAGGTC-3′ (307 bp); RANKL: forward 5′-TGGAGAGCGAAGACACAGAA-3′ and reverse 5′-TGATGGTGAGGTGAGCAAAC-3′ (201 bp); GAPDH: forward 5′-TCACTGCCACTCAGAAGACTGT-3′ and reverse 5′-TTCAGCTCTGGGATGACCTT-3′ (133 bp). PCR conditions were 30 cycles at 94°C, 30 sec; 55°C, 15 sec; and 72°C, 30 sec. Amplification of the GAPDH gene was used as an internal control.
Real-time PCR was carried out in a 25 µL reaction system using a SuperScript III Platinum SYBR Green One-Step qRT-PCR Kit (Life Technology) in a Roche LightCycler 480 (Roche Diagnostics, Indianapolis, IN, USA). Each RNA sample was loaded in duplicate into plates with template amounts of 10 ng. The primers used were as follows: TLR4: forward 5′-CATGGCATTGTTCCTTTCCT-3′ and reverse 5′-TGTCATGAGGGATTTTGCTG-3′ (116 bp); TLR9: forward 5′-AGCACTCCCGTCTCAAAGAA-3′ and reverse 5′-TGACGAACATCTCTGGCTTG-3′ (106 bp); OPG: forward 5′-AATGGTCACTGGGCTGTTTC-3′ and reverse 5′-GAGGATCTTCATTCCCACCA-3′ (120 bp). The primers used for RANKL and GAPDH were the same as in RT-PCR. The real-time PCR conditions were: 50°C for 3 min, 95°C for 5 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 30 sec. Results are presented as fold changes relative to GAPDH reference.
At the termination of cell culture, splenocytes in the 96-well plates were washed with PBS followed by incubation with fluorescence-conjugated antibodies. FITC-conjugated mouse anti-rat CD45RA antibody (clone OX-33; BD Biosciences, Franklin Lakes, NJ, USA) was used to isolate B lymphocytes. For detection of RANKL-positive cells, cultured cells were stained with human OPG-Fc (a fusion protein kindly provided by Dr. Colin Dunstan, Amgen, Thousand Oaks, CA, USA) followed by PE-conjugated goat anti-human IgG (Sigma, Saint Louis, MO, USA). At least 20,000 cells were counted for each sample. Splenocytes in the 6-well plates were used for cell sorting. After staining with FITC-conjugated anti-rat CD45RA antibody, B lymphocytes were isolated individually using BD FACSAria III cell sorter/flow cytometer (BD Biosciences). The purity of the isolated B cells was routinely examined and found to be >98% at all times. For apoptotic cell detection, PE-conjugated annexin V and 7-AAD (BD Biosciences) were added to the cultured cells after the indicated times to determine cell viability. Early apoptotic cells were evaluated by the percentage of annexinV+/7-AAD− cells. At least 800,000 cells were collected in each treatment group.
Focused oligo cDNA array for gene expression profiling
An Oligo GEArray Rat Signal Transduction PathwayFinder Microarray (SA Biosciences) was used to profile the expression of 95 genes representative of 18 signal transduction pathways. Biotin-UTP labeled cRNA was synthesized from total RNA and hybridized with the array membrane. After washing, the membranes were incubated with alkaline phosphatase-conjugated streptavidin followed by CDP-Star chemiluminescent substrate. Images were analyzed by GEArray Analyzer software (SA Biosciences).
Results are presented as means ± SD. Paired Student's t-tests were used to analyze differences among groups. Results with probability values of less than 0.05 were considered statistically significant.
Increased TLR expression in rat splenocytes after stimulation with their respective agonists
To determine TLR gene expression in rat splenocytes after stimulation by E. coli LPS and CpG-ODN, mRNA transcript levels of TLR4 and TLR9, respectively, were measured using RT-PCR. As expected, it was found that E. coli LPS specifically increased TLR4 expression, whereas CpG-ODN specifically increased TLR9 expression, in a dose-dependent manner (Fig. 1a,b). However, no change in TLR4 expression was observed when the cells were treated with CpG-ODN, nor in TLR9 expression when they were treated with E. coli LPS (Fig. 1c,d). LPS-stimulated TLR4 expression was most marked at 48 hr, whereas TLR9 expression was not affected (Fig. 1c). On the other hand, TLR9 expression was dramatically increased 48 hr after stimulation with 2.5 µM CpG-ODN whereas TLR4 was undetectable throughout the observation period (Fig. 1d). The optimal time (48 hr) and dosage (LPS, 2 µg/mL; CpG-ODN, 2.5 µM) for activating TLR4 and TLR9 expression were established and used in all subsequent experiments.
RANKL expression in rat splenocytes after stimulation with TLR agonists
To determine the overall degree of expression of RANKL in cultured splenocytes, the amount of mRNA transcript of RANKL was first examined in cultured rat splenocytes by real time PCR. It was found that the amount of RANKL mRNA did not change in cells treated with CpG-ODN, but was significantly increased in cells treated with E. coli LPS (Fig. 2a). This increase was less pronounced in cells that had been treated with both CpG-ODN and E. coli LPS than in those treated with E. coli LPS alone (Fig. 2a). Flow cytometry showed that the percentage of RANKL-positive splenocytes was much higher in E. coli LPS-treated groups than in control groups. However, the percentage of RANKL-positive splenocytes was less in those treated with CpG-ODN plus E. coli LPS than in those treated with E. coli LPS alone (Fig. 2b). These results suggest that CpG-ODN effectively blocks E. coli LPS-induced RANKL expression and RANKL-positive cell production in cultured rat splenocytes.
TLR expression in rat spleen B cells after stimulation with TLR agonists
In order to determine the direct effects of TLR agonist on B cell RANKL expression and exclude the involvement of other cellular components, rat spleen B cells were isolated from splenocytes by staining with FITC-conjugated anti-CD45RA (clone OX33) followed by flow cytometry cell sorting. Purified B cells were cultured with CpG-ODN and/or E. coli LPS and the amount of mRNA transcript of TLR4 and TLR9 detected by real-time PCR. As shown in Figure 3, TLR4 mRNA transcripts were increased only in E. coli LPS-treated cells, whereas TLR9 mRNA transcripts were increased only in CpG-ODN-treated cells, indicating their respective ligand specificity. Although this finding was not statistically significant, addition of E. coli LPS appeared to attenuate CpG-ODN-induced TLR9 expression (Fig. 3), suggesting a possible interaction between TLR4 and TLR9 signaling.
RANKL expression in rat spleen B cells after stimulation with TLR agonists
RANKL expression in purified spleen B cells was further assessed using real-time PCR and flow cytometry. It was found that E. coli LPS, but not CpG-ODN, strongly induced RANKL mRNA transcript in cultured spleen B cells. Compared with cells treated with E. coli LPS alone, this effect of E. coli LPS on stimulation of RANKL expression was abolished by the addition of CpG-ODN (Fig. 4a). The results of flow cytometry were similar, demonstrating a significantly smaller percentage of RANKL-positive B cells when cells had been treated with both CpG-ODN and E. coli LPS than when they had been treated with E. coli LPS alone (Fig. 4b). Cells treated with CpG-ODN alone did not show a significant reduction in the percentage of RANKL-positive cells (Fig. 4b). Because the ratio of RANKL and OPG production in rat lymphocytes is considered an important marker of inflammatory bone resorption in experimental periodontal disease , the amount of OPG gene expression in cultured cells was also examined. It was found that splenic B cells exhibited low OPG expression and were not affected by LPS or CpG-ODN challenge (Fig. 4c).
Signal transduction pathway analysis
To identify the mechanisms underlying the antagonistic effect of TLR9 (which is activated by CpG-ODN) on TLR4-mediated RANKL production (which is activated by E. coli LPS), 95 genes involved in 18 signal transduction pathways were examined using Oligo DNA array. Compared with cells treated with CpG-ODN or E. coli LPS alone, cells treated with E. coli LPS and CpG-ODN demonstrated a unique pattern of gene expression (Fig. 5a). Three CDKIs, 1c, 2a and 2b, were up-regulated (2.54-, 2.67-, and 2.40-fold, respectively), together with up-regulation of EGFR (2.39-fold). On the other hand, TNFSF6, also known as Fas ligand, and PKCα were down-regulated by 2.38- and 2.33-fold, respectively. All these genes were not changed in LPS-treated or CpG-ODN-treated groups (less than twofold change). These results suggest that inhibition of the CDK pathway may contribute to the observed RANKL suppression in B cells (Fig. 5b).
Evaluation of cell viability
To determine cell viability after different treatments, cultured B cells were stained with PE-conjugated annexin V and 7-AAD followed by flow cytometry analysis. It was found that 24 hr after treatments, the percentage of annexin V+/7-AAD− cells in B cells treated with CpG-ODN and E. coli LPS was significantly greater than in controls (Fig. 5c), indicating an increased number of early apoptotic cells. This was not observed in B cells treated with CpG-ODN or E. coli LPS alone.
In general, periodontitis characterizes alveolar bone destructive diseases associated with gram-negative bacterial infection. Since RANKL was first identified as a cytokine that regulates osteoclast differentiation and activation , studies related to periodontal bone resorption were initially focused on RANKL expression of periodontal tissues, including osteoclasts, osteoblasts and periodontal ligament cells [22, 23]. However, using gingival tissues from patients with periodontitis, we previously found that lymphocytes are key participants in RANKL-mediated bone resorption and demonstrated a higher percentage of RANKL-positive cells in B lymphocytes than in T lymphocytes . Although it is known that various B cell subsets express multiple TLRs, including TLR4 and TLR9, the role(s) of TLR signaling on B cell-mediated bone resorption is completely unclear. Clarifying these roles is important, because it would provide new knowledge about interactions between the innate and adaptive arms of the host immune response in bone pathogenesis.
Using a cell culture system, we assessed RANKL expressions in B cells challenged with various TLR ligands. As expected, we found that TLR4 ligation with bacterial surface component LPS induced more numerous RANKL-positive cells and greater RANKL mRNA productivity in B cells. Others have also reported LPS-stimulated RANKL synthesis in T cells in gingival tissues of patients with chronic periodontitis . As bacterial infection and host immune responses advance, many cells break down and release nuclear substances before being engulfed by macrophages. At this stage of an infection, the host immune system can interact with bacterial DNA substances. In contrast with mammalian cells, in which the rate of CpG motifs is low and 80% are methylated , prokaryotic cells, such as bacteria, are characterized by enriched unmethylated CpG motifs that are the structural bases of host immune cell recognition by TLR9 . Engulfed by immune cells, bacterial DNA fragments containing CpG motifs are delivered to endoplasmic reticulum, eliciting responses through TLR9, which is predominantly expressed on plasmacytoid dendritic cells and B cells . This gives rise to the question: does TLR9 ligation through CpG affect increased RANKL expression by TLR4 activation through LPS in B cells? Interestingly, according to both RANKL-positive cell percentage and quantitative mRNA, the LPS-induced increase in RANKL expression is completely abolished by simultaneous CpG stimulation (Figs. 2, 4). Considering our previous data from an in vivo study, which demonstrated a greater degree of RANKL expression and osteoclastogenesis effect of B cells in an T-independent manner after immunization with Aggregatibacter actinomycetemcomitans (a gram-negative periodontal pathogen) , the RANKL suppressing effect of CpG-ODN is potentially significant in ameliorating B cell-mediated inflammatory bone resorption.
We have demonstrated that TLR4 mRNA transcripts are increased only in E. coli LPS-treated B cells, whereas TLR9 mRNA transcripts are increased only in CpG-ODN-treated B cells (Fig. 3). It is noted that the degree of TLR expression in purified B cells was lower that in splenocytes after the same treatment (Fig. 1). This suggests that there could be simultaneous up-regulation of TLR expressions by non-B cells (such as T cells) in cultured splenocytes after treatments. Very interestingly, the binding of CpG to TLR9 resulted not only in activation of TLR9 signaling but also induction of TLR9 mRNA expression. These results indicate that activation of TLR9 signaling could be mainly achieved by CpG-induced TLR9 up-regulation in an autocrine manner.
Although in a human study CpG-ODN treatment reportedly up-regulated TLR4 expression on activated B cells from peripheral blood , we found that in naïve rat spleen B cells TLR4 expression was unchanged after stimulation with CpG-ODN. This apparent discrepancy may be attributable to differences between human and rat B cells in TLR expression, interactions between TLRs and the B cell activation status. Indeed, interdependence of multiple TLR expression has been identified and TLR expression reportedly follows a specific timeline that may be dependent on the nature of the pathogen .
We found that CpG-ODN dramatically inhibits LPS-induced RANKL production in B cells. However, CpG alone does not inhibit RANKL expression or RANKL-positive cell formation. These results indicate that the reciprocal nature of TLR4 and TLR9 signaling within B cells may play a role in the innate immune responses to infectious diseases. This has been demonstrated by studies showing that activation of TLR9 with CpG-DNA inhibits LPS-mediated TLR4 signaling in enterocytes and such inhibition involves a mechanism that depends on the inhibitory molecule IRAK-M . Interestingly, recent studies have demonstrated that IRAK inhibition significantly reduces LTA-stimulated increases in RANKL production in periodontal ligament fibroblasts . It remains to be determined whether the TLR9-derived RANKL production inhibition in B cells observed in our study is IRAK-dependent.
Our microarray analysis showed that EGFR expression is up-regulated only in cells treated with LPS and CpG-ODN and not in cells treated with LPS or CpG alone (Fig. 5b). It has been demonstrated that EGFR suppresses TNFSF6 activity [32-34] and enhancement of CDKIs [35, 36]. Furthermore, combined TLR4 and TLR9 activation also reportedly down-regulates PKCα activity, which inhibits MAP kinase-mediated CDK activity [37, 38]. Because the CDK pathway controls the cell cycle and apoptosis [39, 40], we evaluated cell viability by detecting early apoptotic cells after treatments. B cells treated with CpG-ODN and E. coli LPS had a significantly greater percentage of annexin V+/7-AAD− cells than did control cells (Fig. 5c). These results suggest that the inhibitory effect of CpG-ODN on LPS-induced RANKL expression is mediated partially by inhibition of the CDK pathway, leading to fewer RANKL-positive B cells through cell apoptosis (Fig. 5d).
Interestingly, both CDKI and IRAK-M have been implicated as negative regulators of TLR signaling [41, 42]. Previous studies have also demonstrated that different bacterial components regulate RANKL expression via different signaling pathways. In bone marrow stromal cells, LPS regulates RANKL expression via prostaglandin E(2)  and in osteoblasts cysteine proteases induce RANKL expression through activator protein 1 signaling pathways . Further studies are warranted to fully delineate the pathway(s) that control RANKL-producing B cells. Therapeutic strategies based on control of RANKL-producing B cells, and therefore, inhibition of B cell-mediated osteoclastogenesis, may be effective in preventing and/or reducing pathological bone resorption.
Human OPG-Fc fusion protein was kindly provided by Dr. Colin Dunstan, Amgen, Thousand Oaks, CA, USA. This work was supported by NIH Grant DE-003420 and DE-021837 from the National Institute of Dental and Craniofacial Research.
The authors have no financial conflict of interest.