Homozygous MMP13-deficient mice (Mmp13 ^−/− ) on a C57BL/6 background were gifted by Dr. D’Armiento. Male WT and Mmp-13 ^−/− mice were characterized and housed as described previously (Nakatani et al., 2016). Experimental (Mmp13 ^−/− ) and control (WT) mice were sacrificed by CO₂ narcosis at a range of time-points (1 day postnatal to 3 months). Animal weights were recorded before experimentation and subsequently. All experiments followed protocols approved by the New York University Institutional Animal Care and Use Committee (IACUC). All the animal work was complied with the Animal Research: Reporting In Vivo Experiments (ARRIVE) guidelines.

Mouse dental pulp cells (DPCs) were isolated from molar and incisor teeth of Mmp13 ^−/− , and WT was homogenized with TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, United States) and 6 mm zirconium oxide (Thomas Scientific, Swedesboro, NJ, United States) using the bead-based homogenizer (BeadBug, Thomas Scientific) and incubated in PBS containing 3 mg/ml collagenase A (Sigma-Aldrich, St. Louis, MO, United States) and 2.5 mg/ml trypsin (Sigma Aldrich) for 30 min at 37°C. Human dental pulp cells (HDPCs) were purchased from Lonza (Morristown, NJ, United States) and were recently authenticated and verified to be contamination-free. Mouse DPCs and HDPCs were cultured in alpha-MEM with 20% fetal bovine serum, 2 mM L-glutamine, and 100 U/ml penicillin–streptomycin (Life Technologies, Grand Island, NY, United States) at 37°C in 5% CO₂. The colony-forming unit fibroblast (CFU-F) assay followed a modification of a previous protocol (Chou et al., 2009). A total of 2.0 × 10³ cells were cultured in a six-well plate for 14 days, followed by Giemsa staining, colony counting, and analysis. For cell proliferation, 5.0 × 10⁴ cells were cultured in serum-free media in a 12-well plate. After 1-day, the media was changed to include serum (day-0), and cell numbers were counted using a hemocytometer daily until day-5. The Vybrant succinyl dehydrogenase (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) Cell Proliferation Assay Kit (Molecular Probes, Eugene, OR, United States) was used according to manufacturer’s protocol. For DPC differentiation, mouse or human DPCs were cultured with differentiation media containing alpha-MEM (as aforementioned) plus 2 mM L-glutamine, 10 nM dexamethasone, 50 μg/ml L-ascorbic acid, and 10 mM beta-glycerophosphate (Sigma-Aldrich, St. Louis, MO, United States) for up to 21 days. Beta-glycerophosphate was only used for the last 3–4 days of culture. After differentiation, the cells were fixed by 4% paraformaldehyde, stained with Alizarin red solution, washed three times with distilled water, and microscopically analyzed.

Mouse-mandibles were dissected for micro-computed tomography (µCT) assessment, while the remainder of the head was fixed in 10% buffered formalin at 4°C for 24-h, prior to decalcification in 10% EDTA. The fixed-tissue was dehydrated through ascending concentrations of ethanol, paraffin-embedded, and serially sectioned (5 μm). Thereafter, the comparative sections were deparaffinized, hydrated, and stained with hematoxylin and eosin (H&E) (Sigma-Aldrich) prior to morphological analysis at 10-day, 3-week, 6-week, and 3-month using a Zeiss Axio (Carl Zeiss, Jena, Thuringia, Germany) light microscope.

To detect the expression and distribution of the mineralization-related markers; amelogenin, Axin2, nestin, dentin sialoprotein (DSP), proteases (MMP8, MMP9, and MMP13), and class-II HDAC (4 and 5), IHC was carried out in similar sections using the Envision + horseradish peroxidase (HRP) staining system (Dako, Agilent, Cork, Ireland) according to the manufacturer’s instructions. Briefly, deparaffinized and hydrated sections were rinsed in phosphate-buffered-saline (PBS) prior to endogenous peroxidase activity being blocked with a 0.3% hydrogen peroxide solution for 20 min (Dual Endogenous Enzyme Block, Dako). The sections were rinsed and blocked using a 1% bovine serum albumin (BSA) solution with 2% goat serum (Santa Cruz Biotechnology, Heildelberg, Germany), and 0.05% Tween-20 (Bio-Rad, Hertfordshire, United Kingdom) for one-hour at room temperature. The sections were incubated with primary antibody in a TBS solution with 1% BSA overnight at 4°C in a humidifying chamber. MMP13-expression was analyzed in WT mice using anti-MMP13 [Abcam, Cambridge, United Kingdom; Cat. number ab39012; Batch number GR157514; Dilution used 1:100; (Soundia et al., 2018)] at 1, 3, 6, and 9-day post-natal. The sections from Mmp13 ^−/− and WT molars at postnatal day 3, day 10, or 3 months were incubated with anti-amelogenin [Abcam; Cat. number ab153915; Batch number GR114097; Dilution used 1:200; (Pandya et al., 2017)], anti-Axin2 [Abcam; Cat. number ab32197; Batch number GR3363772; Dilution used 1:200; (Zhang et al., 2020)], anti-DSP [Millipore, Temecula, CA, United States; Cat. number MABT37; Lot number 2844589; Dilution used 1:200; (Baba et al., 2004)], anti-nestin [Millipore; Cat. number MAB353; Lot number 2370131; Dilution used 1:200; (Hiraga et al., 2010)], anti-MMP8 [Abcam; Cat. number ab53017; Batch number GR38794; Dilution used 1:100; (Zhu et al., 2019)] anti-MMP9 [Abcam; Cat. number ab38898; Batch number 573145; Dilution used 1:100; (Soundia et al., 2018)], anti-HDAC4 [Abcam; ab12172; Batch number GR285278; Dilution used 1:200; (Wein et al., 2016)], and anti-HDAC5 [Abcam; Cat. number ab55403; Batch number GR36969; Dilution used 1:100; (Wanek et al., 2018)] antibodies or PBS replacing primary antibodies as a negative control. The positive control was demonstrated by antibody-expression in tissues previously shown to have a high expression. After washing, the sections were placed in HRP-labeled polymer conjugated to goat anti-rabbit and anti-mouse secondary antibodies (Envision+, Dako) for 30 min, prior to the completion of staining with a 5-min incubation with 3,3′-diaminobenzidine (DAB) chromagen solution, washing in distilled water and counterstaining with hematoxylin (Sigma-Aldrich). As before, a Zeiss Axio (Carl Zeiss) light microscope was used.

The mandibles of Mmp13 ^−/− and WT mice were fixed in 70% ethanol and prepared for high-resolution µCT (SkyScan 1172, Bruker, Kontich, Belgium) of incisor and first molar teeth. A three-dimensional analysis was carried out to determine total volume, enamel volume, dentin volume, pulp volume, and total mineral density (TMD). The samples were scanned using a 10-MP digital detector, 10W of energy (70 kV and 142 mA), and a pixel size of 7.5microns, exposure 850 ms/frame rotation step 0.3° with ×10 frame averaging, 0.5 mm aluminum filter, and scan rotation of 180°. After scanning, the radiographs were reconstructed using NRecon software (version 1.7.3.0; Bruker). Reconstruction was conducted with NRecon using GPU acceleration. Gaussian smoothing was applied with a 2-voxel radius, ring artifact and beam hardening corrections were applied in reconstruction. Ring artefact reduction set to 7 pixels. Beam hardening correction was set to 40%. CTAn software (CTAn Micro-CT software, Bruker) was used to generate 2-D images for color density and 3-D images for CT volume.

Total RNA from cultured pulp cells was isolated using the TRIzol (Thermo Fisher Scientific) method, and reverse transcribed to complementary DNA (cDNA) with TaqMan Reverse Transcription Reagents (Thermo Fisher Scientific) according to the manufacturer’s instructions. The sequences were amplified by adding complementary DNA to the PCR mixture containing each primer (listed Supplementary Table S1) and Platinum SYBR Green qPCR SuperMix uracil-DNA glycosylase (UDG) (Thermo Fisher Scientific). The reactions were pre-incubated at 50°C for 2 min for decontamination of deoxyuridine (dU)-containing DNA by UDG and then incubated at 95°C for 2 min to inactivate UDG and activate Taq. The PCR program continued 46 cycles of denaturation at 95°C for 15-sec, annealing at 60°C for 30-sec, and elongation at 72°C for 30-sec. All data were normalized using the Ct value of beta-actin gene expression from the same sample.

The Illumina HiSeq 2500 system (Illumina, Inc., San Diego, CA, United States) was used to analyze the transcript profiles of pulp tissue isolated from the incisors of Mmp13 ^−/− and WT mice at 3 months. The RNAseq analysis was performed on three independent biological samples for each genotype (n = 3) and run on a 2 × 150 bp configuration and single index per lane in a paired end experimental design. The pulpal tissue was homogenized [T10 basic S2-Ultra-Turrax tissue disrupter (IKA, Staufen, Germany)], and the total RNA was extracted using the TRIzol reagent (Thermo Fisher Scientific, Wilmington, DE, United States) and quantified spectrophotometrically (Nanodrop 2000, Thermo Fisher Scientific). RNA-seq was performed according to the manufacturer’s instructions, and raw data were exported to Strand-NGS ver4.0 (Strand Life Sciences, Bengaluru, India) for subsequent analysis. Genes that were differentially expressed (>2.0 fold) in the MMP13KO group relative to the WT control were identified after passing a t-test (p < 0.05) and post hoc test (Storey with Bootstrapping) with a corrected q-value of 0.05. Genes in the expression data sets were first “ranked” based on Log2 values from highest to lowest for both groups at both time points, prior to hierarchical clustering being used to group gene expression in each condition using the default settings in Strand-NGS. Gene Ontology (GO) was evaluated using Go-Elite (http://www.genmapp.org/go_elite) (Zambon et al., 2012), which is designed to identify a minimal non-redundant set of biological ontology terms or pathways to describe a particular set of genes or metabolites. The subsequent pathway analysis was undertaken using Pathvisio (http://www.pathvisio.org/ - version 3.3.0+) (Kutmon et al., 2015), which uses an over-representation analysis, only reporting on GO terms and pathways with a z score >2, a permutation p < 0.01, and three or more regulated genes for the pathway. This data was then linked and illustrated using WikiPathways (Slenter et al., 2018). RNA-sequencing data followed the Minimum Information About a Sequencing Experiment (MINSEQE) guidelines and have been submitted to the Gene Expression Omnibus (GEO), accession number: GSE178898.

The analysis was carried out on cell lysates from DPCs isolated from 3-month old mouse molars (Mmp13 ^−/− ) and cultured in (proliferation, 5-day culture) mineralization media for 14 days (differentiation). The protein lysates were prepared by incubating the cells for 30 min at 4°C in extraction buffer, comprising RIPA buffer, phenylmethyl sulfonyl fluoride (PMSF), and halt protease inhibitor (all from Thermo Fisher Scientific, Waltham, MA, United States), prior to centrifugation at 13,000 rpm for 10 min at 4°C. Total protein was quantified by a Bradford assay (Bio-Rad Laboratories GmbH, München, Germany) using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), before 50–100 µg proteins were used for sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) on a 4%–15% polyacrylamide gel (Mini-PROTEAN, Bio-Rad) and the migrated proteins were transferred to the polyvinylidene difluoride (PVDF) membrane. To block non-specific protein binding, the membrane was incubated in 5% non-fat dry milk at room temperature for 1 h, prior to overnight incubation at 4°C with primary antibodies raised against HDAC4 (Cell signaling technology; Cat. number 15164; Dilution used 1.1,000), HDAC5 (Cell signaling technology; Cat. number 20458; Dilution used 1.1,000), HDAC6 (Cell signaling technology; Cat. number 7612; Dilution used 1.1,000), Axin2 (Abcam; Cat. number ab32197; Dilution used 1.1,000), phospho-β-catenin (Cell Signaling Technology; Cat. number 9561; Dilution used 1:2,000), β-catenin (Cell Signaling; Cat. number 8480; Dilution used 1:1,000) phospho-p38 (Cell Signaling Technology; Cat. number 9211; Dilution used 1:1,000), p38 (Cell Signaling Technology; Cat. Number 9212; Dilution used 1:1,000), and β-actin (Cell Signaling Technology; Cat. number 4967; Dilution used 1:1,000) in 1:1,000 solution in 2.5% skimmed milk. After washing, the membrane was incubated with 2ry antibodies at room temperature for 2 h, developed using an enhanced chemiluminescence (ECL) detection kit (Clarity Western, Bio-Rad), and then detected using the Chemidoc Touch Imaging System (Bio-Rad). The protein expression was evaluated using Fluor Chem R Hy8300 (ProteinSimple) with β-actin used as a loading control to normalize the data. In total, three independent experiments (n = 3) were carried out for each HDAC target and repeated in triplicate. The quantitative results were obtained using the image analyzing software ImageJ.

The Student’s t-test was used for quantitative analysis of µCT, CFU-F, MTT assay, and qRT-PCR data (p < 0.05). The results are expressed as mean ± standard deviation. The presence or absence of dystrophic pulp calcification in Mmp13 ^−/− and WT was determined using contingency tables and a Fisher’s exact test (p < 0.05). Data analysis used IBM SPSS (v25, Dublin, Ireland) (p < 0.05). Differential expression in RNAseq experimentation was identified after passing a t-test (p < 0.05) and post hoc test (Storey with Bootstrapping) with a corrected q-value of 0.05. The number of independent experiments or animals is listed in the results and figures.

MMP13-expression was analyzed in developing (1, 3, 6, and 9-day) and adult (3-month) WT maxillary first-molars by immunohistochemistry (IHC) and qRT-PCR. MMP13 was detectable 1-day postnatally with strong staining evident in preodontoblasts, odontoblasts, ameloblasts, and the root sheath of Hertwig, compared with the unstained control ( Figure 1A). MMP13-expression persisted at 3, 6, and 9-day with increased detection in the odontoblast, ameloblast layer, and mineralized-dentin (Figure 1B), while exhibiting comparatively low expression in the central pulp and perivascular area at all-time points (Figure 1B). MMP13-expression in 3-month mice was evident in the mineralized-dentin than that in the unstained control (Figures 1Bdii,diii), while expression remained high at 3-month in constantly remodeling tissues such as oral mucosa and alveolar bone (Figure 1Bdi). MMP13-expression was reduced in the odontoblast and predentin of adult molars (Figure 1Bd) than that in developing teeth (Figure 1B), suggesting a role for MMP13 in odontoblast-activity and differentiation. Mmp13 gene-expression was also significantly increased (p < 0.05) in mineralizing-hDPC cultures and differentiated-hDPCs compared with untreated DPC controls (Figure 1C), while MMP13-ablation eliminated Mmp13 gene expression (Figure 1D).

MMP13 was highly-expressed in alveolar bone and oral mucosa, positively confirming tissues reported to have a high MMP13-expression (Uitto et al., 1998) (Figure 1Aa).

The maxillary first molar sections from 3-week, 6-week, and 3-month male Mmp13 ^−/− and WT mice demonstrated abnormal dentin morphology (Figure 2A). The mineralization front in Mmp13 ^−/− was irregular (Figures 2Ab,d,f), the predentin layer was thin and lacked definition, and the odontoblast layer was uneven, less populated, and less well-organized than the regular, organized palisade effect evident in WT samples (Figure 2A). The ordered tubular structure of WT dentin was phenotypically altered and replaced by a disorganized tubular structure at 3-weeks, 6-weeks, and 3-month ( Figure 2A ). The μCT analysis of 3-month-old incisors revealed that the enamel volume was reduced by 38.3% and dentin volume by 29.8% (p < 0.0001) in Mmp13 ^−/− samples, with no significant change in body weight and total pulp volume (Figure 3A) or evidence of occlusal wear reported in MMP9-knockout (KO) samples (Figure 3C) (Yuan et al., 2017), indicating the teeth to be smaller. The reduction was also evident in first molars of 3-month-old mice, revealing significant enamel volume reductions of 16% (p = 0003) and dentin 13% (p = 0.045) ( Figure 3A ). The enamel (14.5% down) and dentin (3.5% down) TMD was significantly reduced (p = 0.009; p = 0.048 respectively) in mandibular incisors but not molars (enamel p = 0.189; dentin p = 0.693) (Figures 3A–C).

Pulp tissue responds to injury and irritation by localized inflammation and the production of tertiary dentine, which forms beneath the area of challenge (Lesot et al., 1994; Smith, 2002). There are two types of tertiary dentin formed depending on the severity of the irritating stimulus; mild irritation as a result of cusp damage induces an upregulation of existing odontoblast activity to form reactionary dentine, while stronger stimuli result in odontoblast death and the recruitment of dental pulp progenitor cells, which differentiate into odontoblast-like cells to form reparative dentine (Lesot et al., 1994). In this study, we were interested in the role of MMP13 not only in tooth development, but also in the production of tertiary dentin at the cusp tips of mouse molars. In this study, significant reactionary tertiary dentin was visible at injured and damaged cusp tips in 6-week and 3-month WT mouse molars, but not the Mmp13 ^−/− equivalents despite thinner dentin (Figures 3B,C). Dystrophic pulpal-calcification was present more frequently in serial sections of maxillary first molars in Mmp13 ^−/− than that in WT (Figure 2C) at 3-week (present 2 out of 5 Mmp13 ^−/− ; 0/5 WT biological samples [p = 0.429]), and significantly at 6-week (5/5 Mmp13 ^−/− ; 0/5 WT biological samples [p = 0.008]) and 3-month (8/9 Mmp13 ^−/− ; 1/9 WT biological-samples [p = 0.003]) (Figure 3C). The dystrophic mineralized-tissue did not exhibit a tubular structure indicative of dentin, and did not stain positively for nestin (Figure 4Cd); however, the marker HDAC5 was expressed in the mineralized tissue (Figure 3Cc), perhaps suggesting a cemental or bony origin (Bradley et al., 2015).

IHC showed that in vivo expression of amelogenin increased in Mmp13 ^−/− molar samples at postnatal 10-day compared with WT (Figure 4A), while the gene expression of ameloblastin (>5-fold) and MMP20 (>3-fold) was significantly upregulated at 3-month in DPCs from Mmp13 ^−/− incisors (Figures 4Ae,f). DSP expression reduced in Mmp13 ^−/− molar dentin and the odontoblast layer at 10-day postnatal, and 3-month samples (Figure 4B), and the preodontoblast-marker nestin was progressively reduced in expression in odontoblasts of Mmp13 ^−/− at 6-weeks and 3-month (Figure 4C). qRT-PCR highlighted the significant attenuation of nestin at 3-month in pulp tissue from Mmp13 ^−/− mouse incisors (Figure 4Ce), and a (>10-fold) decrease in expression in 21-day Mmp13 ^−/− molar-cultures compared with WT (Figures 4Cf,H). The expression of other mineralization markers, dentin matrix acidic phosphoprotein (Dmp-1), Runt-related transcription factor 2 (Runx2), and osterix (Osx) was significantly reduced in DPC cultures from Mmp13 ^−/− molars at 21-day (Figure 4H). The Alizarin red staining of DPC cultures from both genotypes demonstrated less calcific deposits at 14-day in knockout-samples (Figure 4G).

DPC proliferation increased after MMP13-loss. The CFU-f assay demonstrated a significant increase in colony formation (>3-fold; p < 0.05), which was supported by increased Giemsa staining in MMP13-null cultures (Figure 4D). The cell number was markedly increased at time points up to 5 days (p < 0.05) in MMP13KO cultures, with cell doubling time reduced by >50% compared with WT controls (Figure 4E); furthermore, the MTT assay demonstrated a significant increase in cell metabolic activity in MMP13-null cultures (Figure 4F). qRT-PCR showed no difference in the expression of Klf-4 in DPC Mmp13 ^−/− cultures at 21 day (Figure 4H); however, it was significantly upregulated in tissue taken from 3-month old incisor teeth (Figure 7C). Nifc was significantly downregulated at 21-day (Figure 4H). Taken together, these results suggest that MMP13 increases dentinogenic differentiation and attenuates DPC proliferation processes.

As individual MMPs can synergize and regulate other MMPs activity (Ortega et al., 2003), the effect of MMP13-loss on two highly-expressed dental-MMPs was investigated (Sulkala et al., 2007; Yuan et al., 2017), with the loss of MMP9 previously shown to affect the tooth structure. IHC revealed a high MMP8-expression in the oral mucosa of 10-day teeth, but low dentin–pulp expression. There was an increased expression in the odontoblast layer, predentin, and pulp tissue of Mmp13 ^−/− molars at 6-week and 3-month (Figure 5A), which was supported by a significant increase in Mmp8 expression at 3-weeks and 3-month in pulp tissue from Mmp13-null incisors (Figure 5Ag). MMP9 expression decreased in the odontoblast, predentin layer, and pulp of WT compared with Mmp13 ^−/− samples at 3-week, with reduced expression in predentin and odontoblast was progressively more evident in WT at 6-week and 3-month (Figure 5B). The mRNA expression of Mmp9 was increased at 6-weeks and 3-months in MMP13-ablated samples (Figure 5Bg).

MMP13 modulation by the mineralization-associated HDAC-4 (Shimizu et al., 2010) and -5 (Bradley et al., 2015) has been shown in osteoblastic cells. Specifically in DPC cultures, the application of HDACi stimulated pro-mineralization responses by increasing MMP13 expression (Duncan et al., 2016), and broadly the role of acetylation in dentin–pulp hard tissue formation (Yamauchi et al., 2020) has been highlighted. In the current study, the class II HDAC, 4 and 5, expressions significantly increased in mineralizing HDPCs at every time point up to 17-day compared with non-mineralizing control cultures (Figure 6A). Furthermore, IHC-analysis highlighted the high HDAC4-expression in developing odontoblasts, ameloblasts, alveolar bone, dentin, and pulp at post-natal 3-day WT samples compared with unstained control, while in adult teeth, HDAC4-expression was greatly reduced, but was higher in Mmp13 ^−/− samples compared with WT (Figure 6B). In vivo HDAC5-expression was generally low in WT pulp and the odontoblast layer, but high in alveolar bone (Figure 6Cb), with notable expression in the odontoblast and predentin layer of 3-month Mmp13 ^−/− molars compared with controls (Figure 6C). HDAC5’s high expression in the alveolar bone agreed with a previously reported positive control (Figure 6Cb) (Bradley et al., 2015). An Increased Hdac5 mRNA expression was evident at 3-month in pulp tissue from Mmp13 ^−/− incisors compared with WT (Figure 6D). The Western blot (WB) analysis of HDAC4 expression in cultured pulp tissue revealed that expression was stable under proliferating (5-day culture) and differentiating (14-day culture) conditions in Mmp13 ^−/− samples; however, HDAC4 expression decreased in the differentiating cells of WT pulp tissue under mineralizing compared to proliferating conditions (Figure 6E). HDAC5 increased in differentiating DPC samples and in MMP13KO samples in both proliferating and differentiating conditions compared to WT, supporting data in (C). The Class IIb HDAC6, showed a significant reduction in expression under mineralizing conditions with a slight increase in MMP13KO samples compared with WT (Figure 6E). It is clear that selected mineralization-associated class II HDACs are altered in expression in mineralizing conditions (Huynh et al., 2016), and also the novel finding suggests that MMP13 influences their expression in the dentin–pulp complex.

To determine the molecular mechanisms whereby MMP13 alters mineralization effects in vitro and in vivo, we performed RNAseq analysis. Based on initial qRT-PCR experimentation on pulp tissue isolated from WT and Mmp13 ^−/− (Figure 4H), we hypothesized that MMP13-loss would reprogram gene expression during mineralization, altering the expression of a range of genes, while decreasing the expression of specific mineralization-associated markers. Subsequently, we isolated mRNA from the incisor teeth of WT and Mmp13 ^−/− mice at 3-months and showed that of the 45,796 genes analyzed, MMP13-loss increased the expression > two-fold of 3,352 transcripts and suppressed the expression of 1,668 genes, representing alteration of 10.96% of the oligonucleotides (Table 1). The top 20 most up- and downregulated genes at 3-month are listed in (Supplementary Tables S2, S3). The observed gene expression patterns demonstrate that although MMP13-loss induced transcriptional change, more genes were induced than suppressed and the expression of selected genes associated with energy and metabolism (Cartpt 365.4-fold up (Takatani et al., 2021)) and development (Duxf3, 136.6-fold up (Chen and Zhang, 2019)) was dramatically upregulated (Supplementary Table S2), while downregulation was at a more modest level (Supplementary Table S3). Notably, similar patterns of differential gene expression have been reported in other MMP13KO high-throughput transcriptomic studies albeit using microarray technology in different tissues (Toriseva et al., 2012). Interestingly, in support of the hypothesis that showed the absence of MMP13 reduced the differentiation of odontoblasts and increased cell turnover, selected mineralization-associated transcripts were seen to be significantly downregulated (Coll-1, 4.2-fold down; TGFβ1, 3.45-fold down), while other markers of proliferation were upregulated (Pcna, 2.7-fold up; Ki67 4.95-fold up). The downregulation of the odontoblast-marker nestin, and the upregulation of several Mmps (-8, -9 and -20), identified in the RNAseq was validated to be of similar magnitude in subsequent qRT-PR experimentation (Table 2). In order to further interrogate mechanistic interactions, pathway analysis was investigated.

To elucidate the biological processes and identify potential pathways that are affected by MMP13 in primary murine dental pulp tissue at 3-month, the RNAseq gene expression dataset was subjected to pathway analysis using Go-Elite and Pathvisio analysis software (Duncan et al., 2016). Bioinformatics analysis using strict redundancy controls revealed the significant upregulation (Supplementary Table S4) and downregulation (Supplementary Table S5) of multiple pathways related to the host of metabolic and immunological pathways. The subsequent analysis focused on pathways that have previously been identified as being central to developmental and repair processes in the odontoblast cell and dentin–pulp complex in general. A significant downregulation of the Wnt-signaling and pluripotency pathway (<0.001) and the MAPK pathway (<0.003) was specifically noted as Wnt has previously been strongly linked to the odontoblast differentiation (Kawata et al., 2021) and tooth development (Kim et al., 2021), while upregulating of the MAPK pathway has been linked to DPSC differentiation (Lv et al., 2016) and tooth development (Chen et al., 2018). The pathway illustration (Figure 7B) highlighted the dysregulation of several key markers in the Wnt signaling and pluripotency pathway with significant increases in the pluripotency markers Nanog (15.9-fold up) and Oct3/4 (7.7-fold up), findings that were validated in subsequent qRT-PCR (Figures 7B,C). Furthermore, RNAseq at 3-month revealed that mineralization markers associated with DPC differentiation were decreased in expression in Mmp13 ^−/− including the odontoblast-responsive Wnt gene Axin2 (2.5-fold down), Smad3 (2.15-fold down), and TGFβ1 (3.45-fold down) (Figure 7B); findings that were significantly verified by qRT-PCR at 3 and 6 weeks (Figure 7D). From this group of genes, Axin2 was further evaluated by IHC as it is expressed in differentiating odontoblasts (Bae et al., 2015), and has previously been implicated with Wnt/β-catenin signaling in odontoblast development and maturation (Lohi et al., 2010; Babb et al., 2017). Furthermore, Axin2 is considered a direct target of canonical Wnt signaling (Kim and Simmer 2007). In this study, Axin was reduced in expression in Mmp13 ^−/− teeth in vivo at 10-day, 3-week and 3-month in the odontoblast and predentin central pulp tissue compared with WT and negative controls (Figure 7A), confirming the interaction between MMP13 and Wnt signaling in the control of DPC differentiation evidenced in the RNAseq pathway analysis (Figure 7B). The links between MMP13 expression and Wnt/β-catenin signaling have been reported elsewhere in mechanically stretched chondrocytes (Song et al., 2021), and between MMP9, 13, and Wnt in maxillary-expansion techniques (Guerrero et al., 2020). Further analysis revealed that although β-catenin expression was similar between WT and MMP13 samples, phospho-β-catenin protein expression was increased in DPCs from MMP13KO cultured in proliferating and markedly in differentiating conditions, which fits with phosphorylation-dependent degradation of β-catenin is a key step in turning off Wnt signals (Figure 7E). Although the MAPK pathway was significantly downregulated (Supplementary Table S5), the protein expression of p38 and phosphorylated-p38 was increased in MMP13KO under mineralizing conditions, a finding linked to enhanced odontogenic differentiation in vitro (Du et al., 2020) with phosphorylation linked to odontoblast-stimulation in vivo (Simon et al., 2010) (Figure 7E). Despite increases in phosphorylated p38, there was reduced dentine and altered dentin structure in this study, which suggests the absence of MMP13 in tooth development and repair may dysregulate the normal odontoblast activity, which has been demonstrated with other MMPs in tooth development (Yuan et al., 2017).