Pulpitis (toothache) is a painful inflammation of the dental pulp and is a problem of high prevalence throughout the world. It is caused primarily by oral bacteria that infect the pulp, where blood vessels and dental primary afferent (DPA) neurons are richly supplied [1]. The most common entry points of these oral bacteria are dental caries, followed by dental trauma or cracked teeth [2]. The major organisms in dental caries are Gram-positive bacteria, such as Streptococcus mutans, Lactobacillus, and Actinomyces spp. [1, 3, 4]. Conversely, the organisms that cause gingivitis (gum disease) are primarily Gram-negative bacteria, such as Porphyromonas gingivalis and Tannerella forsythia [5, 6]. This suggests that multiple bacterial characteristics, including type, structure, and where they live in the oral cavity, are critically associated with the etiology of oral infectious disease [7-9].

Pulpitis commonly accompanies the innate dental pulp immunity that is mediated by the initial sensing of the bacterial pathogens, intracellular signal transduction, and the release of by-products, such as cytokines and inflammatory mediators [1, 2]. A well-known mechanism underlying this innate immunity is the activation of pattern recognition receptors (PRRs) that recognize pathogen-associated molecular patterns (PAMPs) in odontoblasts [1, 10], pulp fibroblasts [11], and sentinel cells [1, 11]. A highly relevant class of PRRs is the family of Toll-like receptors (TLRs), of which 12 TLRs have been found in mammals [12]; therefore, TLRs play an essential role in the first-line defense of the dental pulp.

As the known cell surface TLRs in rodents are TLR1, TLR2, TLR4, TLR5, and TLR6, they can bind and respond to invading bacteria or their specific structural motifs [13, 14]. For example, TLR4 is activated by lipopolysaccharides present in the cell wall of Gram-negative bacteria, whereas TLR2, which forms heterodimers with either TLR1 or TLR6, is activated by peptidoglycan, lipoteichoic acid (LTA), or lipoproteins derived from Gram-positive bacteria [14]. Gram-positive bacteria predominate over Gram-negative bacteria on the teeth of patients diagnosed with caries and/or pulpitis [3, 4], and upregulation of TLR2 is greater than that of TLR4 in the dental pulp exposed to the oral environment [15]. Nevertheless, the contribution of TLR2 expressed in DPA neurons is not yet fully understood, while the role of TLR4 expressed in trigeminal ganglion (TG) neurons has been studied on dental pulp inflammatory pain [16-19].

Thus, in this study, we focused on the functional and molecular expression of TLR2 in DPA neurons and its change in pulpal inflammation from adult mice. We examined whether TLR2 expressed in TG neurons responds to the synthetic bacterial lipopeptides, Pam₃CSK₄ and Pam₂CSK₄. Next, we examined Tlr2 mRNA expression in individual DPA neurons and investigated the molecular changes in Tlr2, other TLRs (e.g., Tlr1, Tlr4, and Tlr6), pulpitis markers (e.g., Tnf and Il1b), and myeloid differentiation primary response gene 88 (Myd88) for TLR2 signaling pathway at 24 h following the experimental dental pulp injury in wild-type (WT) mice and TLR2 knock-out (KO) mice. We further investigated changes in TLR2 and calcitonin gene-related peptide (CGRP) protein expression as a potential pain transmitter in the dental pulp injury model. Here we first demonstrate the role of TLR2 in DPA neurons, which may contribute to pulpitis.

Animals

A total of 39 male C57BL/6J WT mice (5 to 8-week-old) (Doo Yeol Biotech, South Korea), two male transient receptor potential vanilloid 1 (TRPV1) KO mice, and four male TRPV1-ZsGreen mice (crossed the Trpv1-Cre mice [20] with Ai6 Rosa26^ZsGreen1 mice) of C57BL/6J background (5 to 7-week-old) (Jackson Laboratory, USA) were used for this study. A total of eight male and female TLR2 KO mice (5 to 7-week-old) were kindly provided by Dr. Seung Hyun Han (Seoul National University, South Korea). The mice were maintained on ad libitum standard lab chow (pellets) and water under standard conditions (23±1℃, 12 h light-dark cycle). All the surgical and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Seoul National University (protocol code: SNU-190313-6). This study also conformed to the ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines for preclinical animal studies.

Retrograde labeling of DPA neurons with DiI and primary culture of neurons

All the procedures were performed using the method described previously [21]. Briefly, the pulp of the maxillary first molar was exposed with a low-speed dental drill after induction of anesthesia with pentobarbital (60 to 70 mg/kg, intraperitoneal). The cavity was then filled with crystals of DiI (Molecular Probes, USA) and was occluded with a glass ionomer sealant (GC Fuji II, GC Corporation, Japan). The TG neurons containing DiI-labeled DPA neurons were cultured after two weeks post-labeling.

Single-cell reverse transcriptase polymerase chain reaction (scRT-PCR)

The DPA neurons were visible with DiI labeling, and they were collected by micromanipulation using a glass pipette. Each neuron was photographed for measuring neuron size (µm²) that could represent neuronal types (e.g., small to medium size for nociceptive neurons and large size for non-nociceptive neurons) before cell collection. A single neuron was placed into each tube containing reverse transcription agents (Superscript III, Invitrogen, USA). Complementary DNA (cDNA) synthesis and nested PCR amplification using nested primer pairs (listed in Table 1) were sequentially performed as previously described [21].

Retrograde labeling of DPA neurons with Fluoro-Gold (FG) before dental pulp injury

Following previously described methods [22], the enamel surfaces of the maxillary first molars on both sides were peeled using a low-speed dental drill under pentobarbital anesthesia (60 to 70 mg/kg, intraperitoneal). Then, we performed an etching with a 32% phosphoric acid etchant (Denkist, South Korea) for approximately 10 s, followed by irrigation using a sterilized saline solution. The cavity was filled with crystals of FG (Fluorochrome, USA) and was occluded with a glass ionomer sealant.

Dental pulp injury

One week post-labeling with FG, the pulp of the right maxillary first molar in each mouse was exposed using a low-speed dental drill, as previously described [23]. Dentinal debris was removed from the dental pulp surface using dental k-files (#06, MANI, INC., Japan), followed by irrigation with a sterilized saline solution. After recovering from the anesthetic, the mice were returned to their cages and were supplied with regular food and water.

Real-time RT-PCR

The mice were sacrificed at 24 h post-injury and were perfused with diethyl pyrocarbonate (DEPC) (Sigma-Aldrich, USA)-treated phosphate-buffered saline (PBS) solution (pH 7.4) to prevent blood contaminants. The TG tissues were then harvested immediately and were used for RNA isolation. Total RNA (250 ng), which was purified using an RNeasy Mini Plus Kit (Qiagen, Germany), was synthesized to cDNA using a QuantiTect Reverse Transcription kit (Qiagen, Germany) according to the manufacturer’s instructions. All the PCR amplification procedures were performed as previously described [21], and the relative expressions of the genes were calculated using the ΔΔCT method with Gapdh. The primer pairs for real-time RT-PCR are shown in Table 1.

Immunohistochemistry

At 24 h post-injury, the mice were perfused and fixed with 4% paraformaldehyde (PFA) solution, and the harvested TG tissues were transversely sectioned (thickness: 14 µm). The TG sections were incubated with a blocking solution containing 10% normal rabbit serum (NRS) in PBS containing 0.3% Triton X-100 (PBST) and then were stained with rat anti-TLR2 antibody (1:100 dilution, Abcam, Cat. #ab1186, USA) in 0.3% PBST containing 1% NRS. As a secondary antibody, rabbit anti-rat Cy3 (1:200 dilution, Jackson ImmunoResearch, USA) was used with the same diluents of the primary antibody, and NeuroTrace 640/660 Deep-Red Fluorescent Nissl Stain (1:100 dilution, Molecular Probes, USA) was followed to identify the cell bodies and their sizes (data not shown). For double immunostaining for TLR2 and CGRP, we sequentially incubated a second blocking solution containing 10% normal donkey serum (NDS) in 0.3% PBST after TLR2 immunostaining, followed by goat anti-CGRP (1:1,000 dilution, Abcam, Cat. #ab36001, USA) in 0.3% PBST containing 1% NDS and donkey anti-goat Alexa Fluor 647 (1:200 dilution, Jackson ImmunoResearch, USA) in the same diluents of the primary antibody. All images were examined using a confocal microscope (LSM700, Zeiss, Germany).

All the subsequent immunohistochemical analyses, including cross-sectional area (µm²) and fluorescence intensity of FG-labeled DPA neurons within each TG section, were performed using ImageJ (National Institutes of Health, USA). A total of 3 to 6 TG sections per mouse were randomly selected. The image of the selected sections was converted to 8-bit greyscale, and the brightness/contrast and color threshold were adjusted equally for all the images. Each neuron was determined to be either positive or negative based on its mean intensity value.

Ratiometric Ca²⁺ imaging

TG neurons were used for Ca²⁺ imaging within 24 h of the primary culture. Fura-2-based ratiometric Ca²⁺ imaging was performed as previously described [21]. Fura-2 AM (2 μM; Invitrogen, USA) was loaded for one hour at 37℃. A bath solution (containing 140 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose) and a zero Ca²⁺ bath solution (containing 140 mM NaCl, 5 mM KCl, 2 mM EGTA, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose) were adjusted to pH 7.3±0.1 with NaOH and were prepared fresh before each experiment. The 340/380 ratiometric images and [Ca²⁺]_i changes were acquired using MetaFluor software (v 7.8.13.0; Molecular Devices, USA). In all the Ca²⁺ imaging experiments, TLR2 agonist-positive neurons, displayed as Ca²⁺ transients with ≥7% of the maximal amplitude (in 50 mM KCl) and reproducibility of at least two out of three, were used for the analysis.

Chemicals

Pam₃CSK₄ (Invivogen, Cat. #tlrl-pms, USA) and Pam₂CSK₄ (Invivogen, Cat. #tlrl-pm2s-1, USA) were dissolved in the supplied endotoxin-free water (1 mg/ml). Capsaicin (Sigma-Aldrich, USA) was prepared in ethanol (10 mM) and stored at -20℃. The drugs were diluted to their final concentration with the bath solution for Ca²⁺ imaging.

Data analysis and statistics

The statistical data are presented as the mean±SEM. The differences between three groups were compared using a one-way ANOVA with a Bonferroni post hoc test, and the differences between two groups were compared using a two-tailed Student’s t-test using Prism software (version 7.00; GraphPad, USA). Differences were considered statistically significant at p values <0.05.

Pam₃CSK₄ induces intracellular Ca²⁺ ([Ca²⁺]ⁱ) responses in TG neurons

To determine the functional expression patterns of TLR2, we chose to use in our study the synthetic bacterial lipopeptide Pam₃CSK₄ (a TLR2/TLR1-specific agonist) typically used for TLR2 activation [13, 24]. Activation of TLR2 by three sequential applications of Pam₃CSK₄ (1 μg/ml; 100 s/application) elicited reproducible [Ca²⁺]_i responses in a subset of TG neurons (Fig. 1A, 1B). When we further confirmed the source of the Ca²⁺, the Pam₃CSK₄-induced [Ca²⁺]_i responses were increased even by depletion of the [Ca²⁺]_i store with a 1 μM thapsigargin pretreatment, but the Ca²⁺ response remained unchanged by the following Pam₃CSK₄ application after washout with bath solution (Fig. 1C, 1D). Conversely, the [Ca²⁺]_i responses were completely abolished by a zero Ca²⁺ bath solution (Fig. 1E, 1F) and in TLR2 KO mice (Fig. 1I). Among the 50 mM KCl-responsive TG neurons, a subset of TG neurons responded to Pam₃CSK₄ (approximately 5.7%, n=7 of 122, Fig. 1I). Based on a previous report that Pam₃CSK₄ induces nociceptive behaviors in mice and the [Ca²⁺]_i responses depending on the TRPV1 channel in dorsal root ganglion (DRG) neurons [24], we conducted Ca²⁺ imaging to examine whether Pam₃CSK₄-responding TG neurons also respond to capsaicin (CAP; a TRPV1 agonist). We confirmed that CAP (1 µM)-induced Ca²⁺ transients were elicited in the majority of the Pam₃CSK₄-responding TG neurons (82.6%, n=19 of 23 neurons, Fig. 1G, 1H), and those responding neurons were significantly reduced in TRPV1 KO mice (approximately 1.2%, n=3 of 232, Fig. 1I), suggesting that the [Ca²⁺]_i response by the TLR2 agonist is functionally associated with TRPV1.

Pam₂CSK₄ induces [Ca²⁺]ⁱ responses in TG neurons

Similar to Pam₃CSK₄, three sequential applications of Pam₂CSK₄ (a TLR2/TLR6-specific agonist) (1 μg/ml; 100 s/application) elicited reproducible [Ca²⁺]_i responses in a subset of TG neurons (Fig. 2A, 2B). The Pam₂CSK₄-responding neurons were also observed in TRPV1-positive neurons as well as TRPV1-negative neurons (Fig. 2C), and the CAP (1 µM)-induced Ca²⁺ transients were elicited in half of the Pam₂CSK₄-responding neurons (50%, n=22 of 44, Fig. 2D). Neurons responding to Pam₂CSK₄ were not observed in TLR2 KO mice (0%, n=0 of 76, Fig. 2E, 2F).

TLR2 is distinctly expressed in DPA neurons

We investigated the TLR2 expression patterns in individual DPA neurons that were retrogradely labeled with DiI using scRT-PCR (Fig. 3A). Approximately 17.1% of the DPA neurons expressed Tlr2 mRNA (n=6 of 35, Fig. 3B), with medium-sized neurons (401 to 700 µm²) accounting for the largest proportion (Fig. 3C). In the same experiments, the DPA neurons predominantly expressed Nefh (encoding NF200; approximately 94.3%, n=33 of 35, Fig. 3B), whereas Gfap expression, which would confirm satellite glial cells contamination, was not observed (n=0 of 35, Fig. 3B), consistent with previous studies [25-27].

Dental pulp injury promotes TLR2 mRNA expression and pulpal inflammation

Since significant morphological changes and pain-like behaviors can occur within 24 h of the experimental dental pulp injury [17, 23], we assessed the molecular changes to Tlr2 that take place in the TG at 24 h following dental pulp injury. The Tlr2 mRNA expression levels were significantly increased in the injured TG, compared to the non-injured TG (Fig. 4A). Tlr1 and Tlr6, which are heterodimeric receptors with Tlr2, and Tlr4 were also significantly increased in the injured TG, compared to the non-injured TG (Fig. 4B~4D). In addition, significant upregulation of Tnf (encoding TNF-α), Il1b (encoding IL-1β) mRNA, both of which are representative biomarkers for pulpitis [2], and Myd88 (encoding MYD88), which is a downstream molecule in the TLR2 signaling pathway [28], verified that 24 h post-injury was sufficient to induce pulpal inflammation in WT mice (Fig. 4E~4G). Importantly, there was no significant difference in Tnf, Il1b, and Myd88 expression levels in the injured TG, compared to the non-injured TG, in TLR2 KO mice (Fig. 4E~4G).

Expression of TLR2 in DPA neurons is upregulated by pulpal inflammation

Since non-neuronal cells like infiltrated macrophages can be contained in the injured or inflamed TG tissues [29], we examined the changes that occur in TLR2 protein expression by pulpal inflammation in the cell bodies of DPA neurons that were labeled with the retrograde neuronal tracer FG. Immunohistochemical analysis showed a significant increase of the TLR2-positive DPA neurons in the injured side, compared to the non-injured side, at 24 h post-injury (Fig. 5A, 5B), and the increase in small- to medium-sized neurons (up to 700 µm²) was greater than in large-sized neurons, suggesting that TLR2 has functional roles in nociceptive neurons associated with pain (Fig. 5C). Next, we found that both TLR2 and CGRP (pain-related neuropeptides) expression levels were upregulated in the injured DPA neurons, compared to the non-injured DPA neurons, at 24 h post-injury (Fig. 5D, 5F), and the increase was also observed in TRPV1-lineage DPA neurons, suggesting that upregulation of TLR2 in DPA neurons can contribute to pain during pulpitis (Fig. 5D, 5G). The anti-TLR2 antibody specificity used in our work was verified by a significant decrease in TLR2 immunoreactivity in TLR2 KO mice (Fig. 5E, 5F).

In this study, we present a possible role of neuronal TLR2 in pulpitis with the following findings: (1) Both Pam₃CSK₄ and Pam₂CSK₄ (TLR2-specific agonists) increased the [Ca²⁺]_i levels in a subset of TG neurons, which were responsive to capsaicin (a TRPV1 agonist), but this response was abolished entirely in TLR2 KO mice. (2) Tlr2 mRNA was expressed in a subset of DPA neurons from naïve mice. (3) Tlr1, Tlr2, Tlr4, Tlr6, Tnf, Il1b, and Myd88 mRNAs were upregulated in the injured TG, compared to the non-injured TG, at 24 h following dental pulp injury. (4) TLR2-positive DPA neurons were increased by pulpal inflammation in WT mice and TRPV1-ZsGreen mice, corresponding to the increase of CGRP protein expression.

TLR2 activation triggers downstream signal transduction pathways, usually in non-neuronal cells such as macrophages and epithelial cells, by inducing the release of Ca²⁺ from intracellular stores [30]. However, the mechanisms for this action are different in sensory neurons such as DRG neurons [24]. Wang et al. [24] has demonstrated that Pam₃CSK₄ dramatically increases the [Ca²⁺]_i concentrations, depending on the activation of downstream TRPV1 and TRPA1 channels, by confirming a significant inhibition of this response in TRPV1 KO and TRPA1 KO mice. Our Ca²⁺ imaging results confirmed that Pam₃CSK₄-induced Ca²⁺ influxes were obtained from the extracellular source (Fig. 1E, 1F), but not from the intracellular store (Fig. 1C, 1D). Moreover, TRPV1-expressing TG neurons responded to either Pam₃CSK₄ (approximately 82.6%, n=19 of 23, Fig. 1G, 1H) or Pam₂CSK₄ (50%, n=22 of 44, Fig. 2C, 2D), and Pam₃CSK₄-responding neurons were significantly reduced in TRPV1 KO mice (Fig. 1I). These results suggest that activation of the distinct TLR2 heterodimeric receptors may have different effects depending on the TRPV1 expression.

Although we primarily used Pam₃CSK₄ to verify the function of TLR2 in sensory neurons, the heterodimeric TLR2 receptors (TLR2/TLR1 or TLR2/TLR6) are distinctly activated by different agonists [24]. The Pam₃CSK₄ is the triacylated lipopeptide commonly derived from Gram-negative bacteria, while Pam₂CSK₄ is the diacylated lipopeptide derived from Gram-positive bacteria [31]. Compared to Pam₃CSK₄ actions (approximately 5.7%, n=7 of 122, Fig. 1I), the Pam₂CSK₄-induced [Ca²⁺]_i responses (approximately 8.7%, n=11 of 126, Fig. 2F) were similar in TG neurons. Moreover, some previous studies support that a single bacterial species can possess both di- and triacylated forms [32, 33]; for example, triacylated lipopeptide exists in Gram-positive bacteria, such as Staphylococcus aureus and Mycobacterium tuberculosis [34], which are also found in human caries lesions [35, 36]. Therefore, our findings suggest that both lipoprotein forms play a critical role in oral bacterial infectious diseases.

The experimental dental pulp injury model used in this study is well established and reviewed for studying neuroinflammatory interactions by allowing the molecular changes in the pulp tissue and somatic TG induced by dental injury [17, 23, 37, 38]. Using this model, we confirmed that 24 h following dental pulp injury is a sufficient amount of time to address the upregulation of (1) TLR2 expression in both the injured TG (Fig. 4A) and in (2) individual DPA neurons on the injured side (Fig. 5B) and especially (3) inflammatory marker genes, such as TNF-α and IL-1β (Fig. 4E, 4F), and MYD88, which is associated with the TLR downstream signaling pathway (Fig. 4G), in injured TG from WT mice, but not from TLR2 KO mice. Although this model cannot completely exclude the damage itself following pulpal injury, it sufficiently allows for bacterial accumulation in the exposed dental cavity [37, 39]. Accordingly, our study has limitations in demonstrating that molecular changes are caused by PAMPs but not by damage-associated molecular patterns (DAMPs). However, we showed a significant increase of both TLR2 and CGRP expression levels in TRPV1-lineage DPA neurons, which have been reported to be upregulated by bacterial infection [40], suggesting that upregulation of TLR2 can contribute to pain transmission during pulpitis (Fig. 5F, 5G).

Mammals host a complex community of commensal bacteria, with an estimated 200 to 700 species residing in the oral cavity [5, 41]. Dental caries is commonly caused by Gram-positive bacteria, such as Streptococcus mutans and Lactobacillus [3, 4], which are a major source of dental pain. Dental pain is conveyed through DPA neurons that are densely innervated in the host dental pulp, making them well-positioned to detect unwanted bacteria when infection occurs [37]. However, to date, the mechanism by which bacteria associated with dental caries elicit pain remains poorly understood [42]. One possible mechanism is a decrease in the threshold for the firing action potentials of nociceptors caused by the secretion of cytokines, such as TNF-α and IL-1β, from immune cells or nociceptor neurons during tissue inflammation [42]. In addition, recent studies have demonstrated direct interactions between nociceptors and pathogens or their signaling through toxins and metabolites [43, 44]. Therefore, further investigations using animal models of dental caries will help us understand whether activation of TLR2 mediated by Gram-positive bacteria can trigger action potential firing in DPA neurons or affect pain hypersensitivity. Moreover, a better mechanistic understanding of how bacteria act on DPA neurons can provide insight into the development of improved therapeutics for treating dental pain associated with bacterial infections. In line with this effort, TLR2 expression was not high in TG or DPA neurons from naïve mice (Fig. 1~3), while it was significantly upregulated in TG and DPA neurons in pulpal inflammation (Fig. 4, 5), suggesting that neural TLR2 plays a critical role in pulpal inflammation, and it may be a potential therapeutic target for alleviating pulpitis.

In conclusion, our results provide a better molecular understanding of pulpitis by revealing that TLR2 expression is upregulated in DPA neurons, and this study uncovers potential neural functions of TLR2 in pulpal inflammation.

Authors thank Dr. Seung Hyun Han (School of Dentistry, Seoul National University) for providing TLR2 KO mice. This research was supported by National Research Foundation (NRF) of Korea (grant number: 2017M3C7A1025602, 2018R1A5A2024418, and 2021R1A2C3003334) funded by the Korean government (Ministry of Education, Science and Technology). All authors gave their final approval and agree to be accountable for all aspects of the work.

All authors declare no conflict of interest.

List of primers used