Secreted products of oral bacteria and biofilms impede mineralization of apical papilla stem cells in TLR-, species-, and culture-dependent fashion

Regenerative endodontics exploits the mineralization potential of stem cells from the apical papilla (SCAPs) in order to promote root maturation of permanent immature teeth. SCAPs may encounter post-disinfection residual bacteria either in planktonic or in biofilm growth mode. Bacterial components bind to Toll-like receptors (TLRs) and trigger pro-inflammatory responses. We hypothesized that biofilm-triggered TLR activation affects the mineralization potential of human SCAPs. SCAPs were challenged with conditioned media derived from standardized dual-species biofilms and planktonic bacterial cultures and their inflammatory status and mineralization capacity were studied. Bacterial products from both growth modes (planktonic vs. biofilm) compromised cell viability, proliferation and mineralization capacity of SCAPs, but in a species- and growth mode-dependent fashion. While TLR4 expression remained unaffected, TLR2 expression was upregulated coinciding with a pro-inflammatory activation of SCAPs. Moreover, TLR and its downstream TGF-β-associated kinase (TAK1) appeared to be blocking mineralization, as inhibition of these factors restored it. In conclusion, bacterial products promoted the pro-inflammatory status and inhibited mineralization of human SCAPs in a TLR-, species-, and culture-dependent fashion. TLR2 emerged as the pivotal mediator of these responses and further research is warranted towards the judicious manipulation of SCAPs in order to modify the untoward events of TLR-priming and signaling.


Methods
Isolation, culture and characterization of human SCAPs. The use of extracted teeth was approved for research purposes by the Institutional Review Board of the University Medical Center Groningen (Medisch Ethisch Toetsingscommissie, Universitair Medisch Centrum Groningen, registration number 201501165). The donors did not participate in any other part of the experimental protocol. Therefore, the study was judged as not falling under the scope of the Medical-Scientific Act for research with humans (METc 2015.584). Collection of extraction teeth was performed in accordance with the relevant guidelines and regulations after informed consent was obtained from the participants. The apical papilla was retrieved from immature impacted third molars that were extracted from young patients (age 16-18 years) who presented for scheduled tooth extraction at the Oral and Maxillofacial Surgery Department, University Medical Center Groningen, the Netherlands, for reasons not related to our experimental protocol.
Bacterial Strains and Growth Conditions. The clinical isolates Streptococcus oralis J22 and Actinomyces naeslundii T14V-J1 were used. The bacteria were streaked on blood agar plates and grown aerobically for S. oralis and anaerobically for A. naeslundii at 37 °C for 24 h. A single colony was used to inoculate 10 mL modified Brain Heart Infusion broth (37.0 g/L BHI, 1.0 g/L yeast extract, 0.02 g/L NaOH, 0.001 g/L Vitamin K1, 5 mg/L L-cysteine-HCl, pH 7.3) (BHI, Oxoid Ltd., Basingstoke, Hampshire, UK) and both bacteria were grown for 24 h under the appropriate conditions (pre-cultures).
Pre-cultures were used to inoculate 50 mL modified BHI (1:20 dilution) and grown for 16 h (main cultures). Bacteria were harvested by centrifugation (6350 × g, 10 °C, 5 min) and washed twice in sterile adhesion buffer SCiENTifiC REPORtS | (2018) 8:12529 | DOI:10.1038/s41598-018-30658-5 (0.147 g/L CaCl 2 , 0.174 g/L K 2 HPO 4 , 0.136 g/L KH 2 PO 4 , 3.728 g/L KCl, pH 6.8). The bacterial pellets were suspended in 10 mL sterile adhesion buffer and sonicated intermittently in ice-water for 3 × 10 sec at 30 W (Vibra cell model 375, Sonics and Materials Inc., Newtown, CT, USA) to break bacterial chains or aggregates. Bacteria were counted in a Bürker-Türk counting chamber (Marienfeld-Superior, Germany) and diluted in sterile adhesion buffer. The CDFF 38 was equipped with 15 sample holders and each sample holder with 5 saliva-coated hydroxyapatite (HA) discs, recessed to a depth of 250 µm. For the saliva coating of the HA disks, freeze-dried whole saliva from at least 20 healthy volunteers of both genders was used 39 . All volunteers gave their informed consent to saliva donation, in agreement with the guidelines set out by Institutional Review Board of the University Medical Center Groningen, the Netherlands. Briefly, the freeze-dried saliva was dissolved in 30 mL adhesion buffer (1.5 g/L), stirred for 2 h and centrifuged at 10,000 × g, 10 °C for 5 min. Then, the HA discs used for biofilm growth were placed in the sample holders and exposed to the reconstituted saliva under static conditions, at 4 °C for 14 h. Subsequently, the sample holders were rinsed with adhesion buffer, placed in the sterile CDFF and 200 mL of a dual-species bacterial suspension consisting of 6 × 10 8 bacteria/mL for S. oralis J22, and 2 × 10 8 bacteria/mL for A. naeslundii T14V-J1 was introduced in the CDFF. The bacterial suspension was admitted under continuous supply (flow rate 100 mL/h), while the CDFF table with the sample holders was rotating slowly (1 rpm). After 2 h, the rotation was stopped for 1 h to allow for bacterial adhesion onto the saliva-coated HA discs. Finally, rotation was resumed (8 rpm) and biofilms were grown under continuous supply of modified BHI (flow rate of 45 mL/h), at 37 °C, for 96 h in the CDFF.

Bacterial conditioned media.
After 96 h of incubation in the CDFF, 10 biofilm-carrying HA disks were first dipped three times in sterile PBS (to remove non-adherent bacteria), then transferred into 50 mL antibiotic-free cell culture medium (DMEM supplemented with 10% FBS and 2mM L-glutamine) in sterile beakers and incubated in ambient air, at 37 °C for 16 h. Next, the biofilm conditioned medium with the biofilm-carrying HA discs were sonicated in iced water, intermittently, at 30 W, for 2 min. Intermittent sonication in iced water, at 65 W, for 2 min followed for further mechanical disruption of biofilms and release of tightly bound extracellular polymeric substances (EPS) components from the biofilm matrix in the conditioned medium 40 . Collection of the conditioned media containing soluble extracellular products secreted from the organized bacterial communities as well as loosely and tightly bound EPS from the biofilm matrix was done by centrifugation (10,000 × g, 10 °C, for 5 min). Finally, the supernatant was passed through a 0.2 μm filter (Corning). Sterility of the conditioned medium was confirmed by spot plating 10 µL aliquots on blood agar plates followed by overnight incubation in aerobic and anaerobic conditions.
For the collection of conditioned media from the mono-species grown in planktonic cultures of, S. oralis J22 and A. naeslundii T14V-J1 derived from the respective main cultures were washed twice with PBS, centrifuged (6350 × g, 10 °C, 5 min) and suspended in 10 mL PBS. Next, bacteria were counted and 1 × 10 8 bacteria/mL suspended in 50 mL antibiotic-free culture medium (DMEM supplemented with 10% FBS and 2mM L-glutamine). The planktonic cultures were incubated in ambient air, at 37 °C for 16 h for S. oralis and in an anaerobic chamber, at 37 °C for 16 h for A. naeslundii. The conditioned media containing bacterial components and soluble secreted extracellular products were collected by centrifugation (10,000 × g, 10 °C, for 5 min), after an intermittent sonication in iced water, at 30 W, for 2 min. Following, the supernatants were filtered through a 0.2 μm filter. Sterility was confirmed by spot plating 10 µL aliquots on blood agar plates followed by overnight incubation in the appropriate conditions.
For the collection of conditioned media from the dual-species grown in planktonic culture the mono-species conditioned were combined. Briefly, due to the predominance of S. oralis J22 in the cultures and the lack of resemblance to the bacterial composition encountered in the CDFF dual-species biofilms (3:1 ratio for S. oralis J22 and A. naeslundii T14V-J1 respectively), collected conditioned media from the mono-species grown in planktonic cultures were combined in a ratio 3 (S. oralis):1 (A. naeslundii).
Optimization of conditioned media dilution. In order to find the maximum concentration of conditioned media capable of inducing an effect and maintaining the SCAPs population for extended time periods, a serial dilution cytotoxicity assay based on the conversion of the methyl-thiazolyl-diphenyl tetrazolium bromide (MTT) was performed. Details are provided as Supplementary Methods S3.
MTT assay (for assessing cell viability as a surrogate outcome). Human SCAPs (passage 3) were seeded onto flat bottom 96-multiwell plates at a cell density of 1 × 10 3 cells/well and incubated with 150 µL culture medium in a humidified incubator at 37 °C with 5% CO 2 for 24 h. Next, the spent media were decanted and 150 µL of the different conditioned media was added to each well. Cells cultured in culture media served as controls. The media were refreshed every 48 h. At the end-point of the assay (7 days), the wells were washed once with 150 µL PBS and 150 µL of the different conditioned media (or culture media for the control group) containing 0.5 mg/mL MTT was added. The plates were incubated for 4 h in a humidified incubator at 37 °C with 5% CO 2 . Following incubation, media were decanted, 150 µL of DMSO (Dimethyl Sulfoxide) (Sigma-Aldrich, Amsterdam, the Netherlands) was added to each well and the plates were covered with aluminium foil and agitated on an orbital shaker for 15 min. Next, the absorbance was measured at a wavelength of 570 nm (with a reference filter of 650 nm) with a Benchmark microplate reader (Bio-Rad Laboratories, Hercules, CA). The assay was repeated 3 independent times (cell cultures) with triplicate samples for each group. Methods S4. The TissueFAXS microscopy system (TissueGnostics GmbH, Vienna, Austria) was used to fully scan each well with the DAPI and Texas red filters at 10x magnification sequentially. Analysis of the captured images was carried out with Tissue Quest 4.01.0127 software (TissueGnostics GmbH, Vienna, Austria). Results were expressed as % of positive cells for Ki-67 to the total of DAPI stained cells. The assay was repeated 3 independent times (cell cultures) with triplicate samples for each group.
Mineralization. Human SCAPs (passage 3) were seeded onto 12-well plates at a cell density of 4 × 10 4 cells/ well and incubated with culture medium. Mineralizing conditioned media (mDMEM) were prepared by first diluting the different conditioned media in culture media to their optimum maximum concentration and then by adding 10 −8 M dexamethasone, 5 mM β-glycerophosphate, 50 μM ascorbic acid and 2 mM KH 2 PO4 (all reagents from Sigma, St Louis, MO, USA). Upon reaching confluency, the spent media were decanted and 1 mL from the different mineralizing conditioned media was added. Cells were cultured for 14 days with media refreshment every 3 days. Subsequently, cells were fixed with 2% PFA in PBS for 30 min, washed three times with PBS and incubated with PBS containing 1 µg/ml 4′,6-diamidino-2-phenylindole (DAPI), in room temperature and dark conditions, for 15 min. The TissueFAXS microscope system was used to fully scan each well with the DAPI filter at 10x magnification and quantification of the DAPI stained nuclei was carried out with Tissue Quest 4.01.0127 software. Next, the wells were washed three times with sterile demineralized water and incubated with 40 mM Alizarin Red-S (AR-S) (pH 4.2), in room temperature under gentle shaking for 20 min. The non-incorporated dye was aspirated and the wells were washed five times with sterile demineralized water for reduction of the unspecific staining. Finally, the wells were scanned again using the TissueFAXS microscope system and images were captured before applying an established protocol for extraction and quantification of AR-S from the stained monolayers 41 . Details of the protocol for dye extraction and quantification are provided as Supplementary Methods S5. The moles of calcium deposited per cell were extrapolated based on the principle that 1 mole of AR-S binds to 2 moles of calcium 42 . The staining was repeated 3 independent times (cell cultures) with triplicate samples for each group and during quantification triplicate absorbance read-outs were performed. Gene expression levels were determined by RT-qPCR for DSPP, ALPL, BGLAP, BMP2, BMP7, as mineralization differentiation marker genes, TNFA, IL6, IL8 as inflammation marker genes and TLR2 and TLR4, as immunophenotypical transition marker genes. All primers were purchased from Sigma-Aldrich (KiCqStart ™ Primers) and are listed in Supplementary Table 1. The detailed RT-qPCR protocol is provided as Supplementary Methods S7. The expression levels of the target genes were obtained via the ΔΔC T method after normalization using β2 microglobulin (B2M) as reference gene and as relative induction compared to non-mineralized control. The assay was repeated three times independently (cell cultures) with duplicate samples.

TLR2/4 blocking and mineralization. TLR2/4 signaling was inhibited with OxPAPC (InvivoGen, San
Diego, CA, USA), a mixture of oxidized phospholipids competing upstream with the TLR2/4 accessory proteins. Additionally, in a separate set of experiments, a pivotal downstream mediator of TLR2/4 signaling, TGF-β-associated kinase -TAK1, was also inhibited with 5Z-7-oxozeaenol (Sigma-Aldrich). Human SCAPs (passage 3) were seeded onto 48-well plates at a cell density of 2 × 10 4 cells/well and incubated with culture medium. Upon reaching confluency, the spent media were decanted and 0.5 mL from the different mineralizing conditioned media containing either 5 μM OxPAPC (reconstituted in chloroform) or 6 μM 5Z-7-oxozeaenol (reconstituted in DMSO) were added. The cells were cultured for 14 days, with the mineralizing conditioned media and the inhibitors being refreshed every 3 days. Cells incubated with the different mineralizing conditioned media containing chloroform and DMSO, served as controls. Following, an in vitro functional mineralization assay based on AR-S staining and quantification was performed as described above.
Statistical analysis. Statistical analysis was performed using SPSS 22.00 software (SPSS Inc., USA). All data were expressed as Mean ± Standard Deviation (SD). Normality of data distribution was assessed with the Shapiro-Wilk test. One-way analysis of variance (ANOVA) with a Tukey's Honest Significant Difference (Tukey's HSD) multiple comparison post-hoc test was performed to assess SCAPs cell viability, proliferation capacity, mineralization efficiency and relative gene expression (log 2 mRNA fold-change) after exposure to the various conditioned media (a = 0.05). Independent unpaired Student's t-test was used to compare differences in the mineralizing efficiency between each pair of similarly exposed SCAPs, with and without the addition of the TLR2 inhibitors (a = 0.05).  Fig. S1).

Determination of the optimal maximum concentration of conditioned media. The initial
MTT-based cytotoxicity screening of the different conditioned media yielded a concentration of 25% conditioned media as the optimal maximum concentration of exposure. SCAPs mitochondrial activity seemed to be unaffected by any concentration changes until the 25% concentration cut-off point, with any increase further causing an abrupt decline on cells' viability. The same dose-response curve applied for all different conditioned media ( Supplementary Fig. S2).
SCAPs cell viability is diversely affected, depending on the imposed stimuli. Irrespective of the origin of the conditioned media, these all significantly reduced the cell viability of SCAPs compared to non-stimulated controls (DMEM), as assessed by MTT conversion (P ≤ 0.001, Fig. 1). Conditioned media from planktonic S. oralis J22 planktonic and dual species biofilm cultures most strongly suppressed cell viability, followed by the dual species planktonic and A. naeslundii T14V-J1 cultures (Supplementary Tables 2-4).

SCAPs proliferation is reduced in a bacterial species-dependent fashion.
Irrespective of the origin of the conditioned media, these all significantly reduced the proliferation of SCAPs compared to non-stimulated controls (DMEM), as assessed by the immunocytochemical Ki-67 nuclear staining (P ≤ 0.001, except for A. naeslundii T14V-J1 where P ≤ 0.05). Conditioned media from S. oralis J22 planktonic and dual species biofilm cultures most strongly inhibited SCAPs proliferation, followed by the dual species planktonic and A. naeslundii T14V-J1 cultures (Fig. 2) (Supplementary Tables 5-7).

Mineralization of SCAPs is inhibited in a bacterial species and culture-dependent fashion.
Irrespective of the origin of the conditioned media, these all significantly reduced the normalized (i.e. on a per cell level) extracellular calcium deposited by SCAPs compared to non-stimulated mineralizing controls (mDMEM), as assessed by the colorimetric quantification of the AR-S-stained extracellular calcium deposits (P ≤ 0.001). Conditioned media from S. oralis J22 planktonic cultures most strongly inhibited mineralization, while conditioned media from A. naeslundii T14V-J1 planktonic cultures had the least influence. Interestingly, conditioned media from the dual-species biofilm (co-cultured A. naeslundii and S. oralis) had a stronger impact on the mineralization reduction of SCAPs compared to their combined planktonic media, with the latter showing an intermediate inhibitory influence on mineralization (Fig. 3A). Representative micrographs of the whole 12-well plates from each group are presented in Fig. 3B (Supplementary Tables 8-10).

Dentinogenic potential is abolished while mineralization-and inflammation-associated gene expression is regulated in a bacterial species and culture-dependent fashion.
Irrespective of the bacterial species or (co)culture conditions, their products abolished expression of DSPP in SCAPs during mineralization, while the control showed upregulated expression of DSPP (Fig. 4Ai). All assessed relevant mineralization genes (RUNX2, ALPL, BGLAP, BMP2 and BMP7) were downregulated by bacterial products in mineralizing SCAPs (Fig. 4Bi-v). Similar to the influence of bacteria on calcium deposition, conditioned media from planktonic S. oralis J22 most strongly inhibited expression of mineralizing genes, while A. naeslundii T14V-J1 conditioned media showed only little suppression. In fact, the presence of A. naeslundii in planktonic or biofilm co-cultures with S. oralis J22 suppressed the inhibitory influence of S. oralis on expression of mineralization genes (Fig. 4Bi-v).   Compared to the expression of mineralization genes, secreted products of bacteria had an inverse influence on pro-inflammatory gene expression (TNFA, IL6 and IL8). While controls showed virtually no expression of pro-inflammatory genes, these was strongly upregulated by planktonic S. oralis conditioned medium in mineralizing SCAPs, even when S. oralis was in biofilm or planktonic co-culture with A. naeslundii T14V-J1 (Fig. 4Ci-iii). A. naeslundii conditioned medium upregulated all three pro-inflammatory genes too, but to a lesser extent. The strong stimulation of pro-inflammatory gene expression from S. oralis was only marginally influenced by A. naeslundii, because conditioned media from co-cultures (planktonic and biofilm) still strongly promoted expression of TNFA, IL6 and IL8 (Fig. 4Ci-iii).
TLR2 (but not TLR4) gene expression is regulated in a bacterial species and culture-dependent fashion. The expression of TLR2, which is a specific receptor for Gram + pathogen-associated molecular patterns (PAMPs), i.e. bacterial products, was low on mineralizing SCAPs (Fig. 4Di). Incubation with conditioned media upregulated TLR2 expression, irrespective of source or culture condition. The expression pattern, however, was remarkably similar to the expression of pro-inflammatory genes (Fig. 4Ci-iii, Di), with a relatively low upregulation induced by A. naeslundii T14V-J1 conditioned medium and high upregulation by conditioned media from S. oralis monocultures or co-cultures with A. naeslundii T14V-J1, either planktonic or in biofilm (Fig. 4Di). The expression of TLR4 remained unaltered and remarkably low after incubation of SCAPs either with the control mineralizing media (MDMEM) or with the various conditioned media (Fig. 4Dii).
Overall, a correlative gene expression profile chart highlighted the gene expression trend among the groups and justified the hypothesis that the abatement of the mineralization efficiency of human SCAPs could be mediated primarily through the TLR2 signaling transduction pathway (Fig. 4E).
Detailed pairwise comparisons and statistically significant differences are presented in Supplementary  Table 13.

Bacterial activation of TLR signaling inhibits mineralization.
To corroborate the hypothesis of TLR involvement in the attenuation of the mineralization efficiency of human SCAPs, AR-S staining was performed after the continuous pharmacological inhibition of the TLR signaling, either after adding a TLR2/4 inhibitor (upstream) or a TAK1 inhibitor (downstream) during the 14-day exposure of the cells to the various bacterial conditioned media. Quantification of the extracted dye was performed, as described previously. No difference was observed in the mineralizing control groups (mDMEM) after TLR2/4 or TAK1 blockade. All experimental groups showed a significant increase in terms of extracellular amounts of deposited calcium, after prolonged upstream TLR2/4 blockade or downstream TAK1 blockade (P ≤ 0.001) (Fig. 5A). Upon visualization, the density of the stained calcium deposits had increased in all experimental groups after both upstream (TLR2/4) and downstream (TAK1) blocking of TLR signaling, compared to the corresponding controls groups without inhibitors. Microscopically, this difference was conspicuously visible in the groups exposed to the conditioned media derived from S. oralis J22 and dual-species biofilm cultures, in which mineralization was also markedly decreased before any inhibition had taken place (Fig. 5B).

Discussion
The main finding of our study was that exposure to products from endodontopathogenic bacteria (S. oralis J22 and A. naeslundii T14V-J1) inhibited mineralization of human SCAPs in a TLR-dependent fashion. This was evidenced by the re-establishment of SCAPs mineralization capacity firstly, after blocking upstream the TLR signaling pathway and secondly after blocking downstream one of the pivotal mediators of TLR activation (and also a central hub of inflammatory signaling-TAK1-). Evidence at the gene expression level pointed out clearly towards the TLR2 (and not TLR4) as being the pivotal molecule regulating this effect. A second finding was that the degree of inhibition of mineralization was highly dependent on the bacterial species and their growth mode (biofilm vs. planktonic cultures). S. oralis J22, grown either as planktonic monoculture or in dual-species biofilms with A. naeslundii T14V-J1, strongly inhibited mineralization and caused a significant upregulation of inflammatory markers in SCAPs. A. naeslundii T14V-J1 induced only a low degree upregulation of SCAPs inflammatory markers, which was coupled with a lower inhibition of mineralization as well. Allegedly, the bacterial products of A. naeslundii T14V-J1 appeared to rescue the inhibition of mineralization induced by S. oralis J22 products, while mitigating also the inflammatory phenotypical shift of SCAPs. However, this rescuing effect seemed to be highly dependent on the growth mode, as bacterial products of dual-species biofilms compromised mineralization capacity of SCAPs stronger than products from their respective mixed planktonic cultures.
The exposure of SCAPs to planktonically and biofilm-released bacterial products led to a different degree of upregulated TLR2 gene expression. A well-established fact is that specific bacterial structures derived from Gram-positive bacteria elicit TLR2-mediated responses from eukaryotic cells equipped with this recognition molecule 37 . However, this is the first report that demonstrates different TLR2-dependent dental stem cell responses after exposure to different Gram-positive species. This call for studies that investigate TLR-mediated dental stem cell behavior on a species-dependent and not only on specific bacterial structural motif level. S. oralis and A. naeslundii are clinically relevant bacterial species; both are resistant to endodontic disinfection procedures and persistently present even in previously disinfected root canals 35,36 . Although their capacity to aggregate and form robust multi-species biofilms is established 43 , the mechanisms of their impact on stem cell fate are poorly investigated. Indeed, exposure to S. oralis products had detrimental effects on the basic cell processes and the mineralizing efficiency of SCAPs. For this specific bacterial species, surface components (other than the frequently reported LTA and peptidoglycans) involved to bacterial adhesion and colonization events might have contributed to this considerably high TLR2 gene expression and the inflammatory SCAPs polarization observed 44,45 . Interestingly, A. naeslundii T14V-J1 products led only to a minor decline. Indeed, when A. naeslundii T14V-J1 was combined with S. oralis J22, irrespective of the culture or growth conditions, inhibition of mineralization seemed to be dampened. To the best of our knowledge, there is no specific evidence describing the action of A. naeslundii on SCAPs or other dental stem cells, thus making any speculations on the observed responses rather precarious. The componential characterization of the conditioned media could elucidate the intriguing results of this study and further research is warranted. However, taking as a fact that A.naeslundii T14V-J1 products impair SCAPs mineralization capacity only to a minor degree, a 'dilution effect' of the deleterious components derived from the S. oralis J22 cannot be excluded. Thus, the rescuing effect of A. naeslundii T14V-J1 observed could merely reflect a quantitative shift of the conditioned media components towards a more favorable equilibrium for the SCAPs.
Especially for the SCAPs response to the dual-species biofilm conditioned media, components of the biofilm matrix should be acknowledged as a contributing factor. The production of matrix components, such as polysaccharides and proteins vary depending on growth mode (planktonic vs biofilm) both in qualitative and quantitative terms, with biofilm outperforming the respective planktonic cultures 46,47 . Also, the presence of biofilm matrix components seems capable of triggering TLR2 signaling 48,49 . Moreover, during the phenotypic switch from planktonic to biofilm phase, the secreted or released bacterial products could change 50,51 . It has been shown that the osteogenic differentiation of human bone marrow mesenchymal stem cells is inhibited by biofilm components in a TLR2-dependent fashion 52 . Our results corroborate these findings, because biofilm-derived products compromised significantly the osteogenic differentiation of SCAPs in a TLR2-dependent fashion. Osteogenesis and dentinogenesis are governed by similar stimuli and underlying pathways, yet formation of dentin is specific for teeth and requires expression of DSPP 53 . Interestingly, it has been reported that residual mono-species Enterococcus faecalis biofilms abrogate the dentinogenic potential of SCAPs, while their osteogenic capacity is maintained 22 . Our data concur partly with these results with regard to SCAPs only concerning the dentinogenic potential, as planktonic and biofilm bacterial products led to complete loss of DSPP gene expression. However, in our study, products from S. oralis J22 and A. naeslundii T14V-J1 planktonic and biofilm cultures caused a downregulation of the mineralizing gene profile expression, with planktonic S. oralis J22 and the dual-species biofilm cultures eliciting a much more considerable effect. This indicates a species-dependent influence on osteogenic plasticity of SCAPs that likely depends on the composition and concentration of bacterial products.
Binding of ligands from oral bacteria or their biofilms, activates TRL2 in TLR2-bearing cells 54,55 and recruits adaptor protein myeloid-differentiation primary-response protein 88 (MyD88). This activates transforming growth factor-β (TGF-β)-activated kinase (TAK1), a pivotal pro-inflammatory factor, that in turn activates the major inflammatory transcription factors nuclear factor-κB (NF-κB) and p38 56 . Their activation upregulates expression of pro-inflammatory genes such as IL1B, IL6, IL8 and TNFA 57,58 . In this study, increased gene expression of TLR2, IL6, IL8 and TNFA correlated negatively with the expression of mineralizing genes and was associated with a reduced mineralizing capacity of SCAPs. Long-term exposure of SCAPs to pro-inflammatory stimuli has been shown to inhibit mineralization and reduce the expression of osteogenic and dentinogenic genes 59 . Pro-inflammatory stimulation of human DPSCs with TNFα impairs their mineralizing efficiency, that is rescued through inhibition of NF-κB, a major downstream mediator 60,61 . Our results show that inhibition of pro-inflammatory pathway mediators (TLR2/4 and TAK1) restore the mineralization capacity of SCAPs, which corroborates the findings with the DPSCs. The products of S. oralis caused a stronger pro-inflammatory activation than products of A. naeslundii, which might relate to differences in composition and concentration of products released by these bacteria. However, it still remains unclear why A. naeslundii products appeared to alleviate the pro-inflammatory stimulation by S. oralis products.
A pivotal transcription factor in osteogenic differentiation is Runx2, which is activated by mitogen-activated protein kinase (MAPK) p38 62 . Activation of p38 may occur canonically by activated TAK1, which could be activated by TLR2-bound bacterial ligands. TLR-activated p38 activates pro-inflammatory genes such as IL1B, TNFA, IL8, and IL6. Alternatively, non-canonically activated p38, e.g. by growth factors bound to receptor tyrosine kinases or by Wnt signaling, activates Runx2 that drives osteogenesis subsequently 63 . In vitro, MAPKs regulate of osteogenic differentiation of human MSCs 64 . Dental pulp stem cells which are closely related to SCAPs, also depend on growth factor-activated p38 for osteogenic differentiation 65 . Interestingly, ERK1/2 MAPK inhibits osteogenesis, while p38 MAPK promotes osteogenic differentiation of human DPSCs 66 . Our results suggest that bacterial triggering of p38-induced pro-inflammatory activation supersedes growth factor-activated p38-driven osteogenesis. The underlying mechanism is unclear, but may relate to the increased expression of TLR2 which could amplify the pro-inflammatory signal, while the osteogenic stimulus is unchanged. This causes a fate shift from osteogenesis to inflammation. This is corroborated by the rescue of osteogenesis after inhibition of TLR2 or TAK1.
In summary, this study has highlighted the consequences that biofilm-derived components bear on relevant dental pulp repair processes that are dependent on proper stem cell function and differentiation. Moreover, it has demonstrated the diverse effects that various PAMPs can have on stem cells functions. SCAPs acquired an immunological rather than a mineralizing phenotype when exposed to bacterial stimuli. The overall results suggest that the TLR2 pathway may act as a checkpoint of this immunophenotypical transition. Whether a cross-talk exists between the TLR2-induced pro-inflammatory activation and the signaling related to the mineralization of human SCAPs warrants further investigation. Our findings have clinical relevance and could facilitate the stem cell-mediated endodontic regenerative procedures. Firstly, they highlight the need for improvement of the disinfection protocols that could result in lowering the microbial burden to levels incapable of altering the stem cell functions. Secondly, they point towards the judicious manipulation of endogenously introduced human SCAPs (or exogenously implanted dental stem cells) in order to modify the untoward events of TLR-priming and signaling.
Data availability. The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.