Extracellular matrix (ECM) is a three-dimensional structure consisting of ground substance and fiber proteins (Theocharis et al., 2016). It provides physical support for the cells but also functions as regulatory factors in numerous biological events ranging from development to regeneration. ECM interacts with cells directly via cell binding domains and subsequently initiates downstream intracellular signaling (Berrier and Yamada 2007). Furthermore, ECM physical properties influence cell functions; for example, an appropriate ECM stiffness promotes adherence of specific cell population and impacts cell differentiation (Engler et al., 2006), although other external factors intervene such as topographical, geometrical, and mechanical features (Smith et al., 2018; Chu et al., 2019).

Hence, the characteristic of specific ECM and its modulations have been extensively investigated in order to control cell behaviors toward disease attenuation or healing repair. The ECM of dental pulp are rich with hyaluronan, glycosaminoglycans, and proteoglycans that are all kept together by a network of thin collagen fibrils, reticular fibrils, and fibronectin (Veis and Goldberg 2014). Due to major differences in the pulp ECM, pulp is often a non-mineralized tissue, whereas dentin is a non-collagenous ECM component. Some dentin ECM molecules function as crystal nucleators and promote crystal formation, whereas others inhibit mineralization (Goldberg and Smith 2004).

Utilization of ECM in tissue engineering has been vigorously reported. The addition of ECM to biomaterials improves their biocompatibility and enhances cell interaction (Gothard et al., 2015; Harvestine et al., 2016). The concept of decellularized ECM has been introduced as the approach for tissue engineering of complex structures (Cukierman et al., 2001). Decellularized ECM can be utilized alone or in combination with recellularization (Fu et al., 2014). Decellularized ECM contains specific signals to instruct cell response due to the distinct ECM composition and structures. To better understand this ECM, Hynes and Naba in 2012 categorized these proteins (Hynes and Naba 2012). The matrisome, the omics study of ECM proteins, had been introduced. This contains core matrisome proteins-collagens, proteoglycans, glycoproteins, and matrisome-associated proteins-regulators, secreted proteins, and affiliated proteins. Every category has been subdivided to better analyze this ECM (Naba et al., 2012).

Decellularized ECM from human dental pulp tissues supports stem cells isolated from apical papilla proliferation and differentiation in vitro (Song et al., 2017). The powder of decellularized dental pulp tissue enhanced angiogenesis (Bakhtiar et al., 2020). Those evidences support the benefits of oral tissue-derived ECM for regenerative approaches in dentistry. However, the limited amount of dental tissues is a major concern. In vitro production of ECM may circumvent this issue; stem cells combined with in vitro amplification could allow the production of high volume of oral stem cells-derived ECM. For example, dental pulp stem cell (DPSC)-derived ECM stimulates cell proliferation and cyclin D1 expression in vitro and supports dental pulp regeneration in vivo (Zhang et al., 2017). Dental pulp reacts to assaults through mineralization by building new dentine bridge in vivo or by mineralization in vitro (Schroder 1985; Zanini et al., 2012). This mineralization occurs through mobilization and differentiation of DPSCs (Gronthos et al., 2000).

Therefore, its ECM protein components in dental pulp are obvious research targets; we hypothesize that they are largely responsible for the physiological characteristics of this tissue. They may present an ECM signature associated with mineralization process. We performed a proteomics analysis to identify DPSC-derived ECM and to determine the potential matrisome proteins associated to mineralization. In addition, we will assess these ECM abilities to promote osteogenic differentiation in a cell which does not react to assaults by mineralization: the human gingival fibroblasts (GF).

The present study identifies and characterizes the composition and functional differences of ECM derived from DPSCs. In this study, we showed ECM derived from DPSCs exhibited classical ECM protein expression (type I collagen and fibronectin) whatever the culture media. Moreover, DPSC produced a diverse ECM where some matrisome proteins differ according to culture medium albeit the DPSC matrisome composition remains almost the same. Calcium and phosphate deposition within ECM confirmed the osteogenic differentiation and was associated to osteo-inductive conditions.

The aim of decellularization process was to remove cellular components from cell-derived ECM. After decellularization, we detected neither filamentous actin nor nuclei. DNA within dECM significantly lowers after decellularization (Guneta et al., 2017; Jeon et al., 2018). This confirms our results; the decellularizing processes effectively removed both cell components and mineral deposition in ECM as evidenced by negative staining with DAPI, Alizarin Red S, and Von Kossa in dECM. Moreover, the Ca and P components in EDX analysis were not prominent in both N-dECM and OM-dECM, confirming the demineralizing effect.

According to ECM classified by matrisome analysis (Naba et al., 2016), the present study identifies and quantifies the ECM proteins derived from DPSCs. We found 225 matrisome proteins showed highest abundance from glycoproteins, then regulator proteins, affiliated proteins, secreted proteins, collagens, and proteoglycan, respectively. On the other hand, 146 proteins were found in matrisome proteins of ECM derived from dentin and detected collagens, glycoprotein, affiliated proteins, regulator proteins, secreted protein, and proteoglycans, respectively (Reis et al., 2020). ECM of periodontal ligament (another oral tissue) is composed of 105 matrisome proteins, and highest detection was of regulator proteins, then glycoproteins, collagens, affiliated proteins, proteoglycans, and secreted proteins, respectively (Denes et al., 2020). Therefore, differences among ECM proteins in each matrisome component might be specific of a type of tissue and might lead to specific regulation and function from ECM proteins.

Biological network analysis is a powerful approach to achieve insights into biological systems. The interactions of different proteins regulate the mechanisms associated with physiology and pathology (Gonzalez and Kann 2012). The most significant functional enrichment in DPSCs matrisome is cellular component of collagen-containing ECM, correspondingly with PPI to collagen formation. Dental pulp is a connective tissue, which is composed of collagen and elastic fibers and contains glycoproteins such as fibronectin and laminins (Yildirim 2013). We confirmed that DPSCs produce these ECM proteins and thus behave like mesenchymal cells. Moreover, the PPI of matrisome pathways showed integrin β ₁ surface and laminin interactions. Correspondingly, these biological functions showed regulated DPSCs attachment to ECM (Zhu et al., 1998). For this reason, integrin–laminin interaction on ECM might be a major biology function of DPSCs attachment. About cell adhesion, our data confirmed the detection of fibronectin (together with integrin β ₁) which is necessary for odontoblasts differentiation (Lesot et al., 1981; Lesot et al., 1990; Saito et al., 2015). NCAM signaling for neurite out-growth pathway is important for the formation and maintenance of the nervous system of cellular processes (Panicker et al., 2003), which is found in DPSCs matrisome, especially in normal matrisome (N-ECM). It is possible that DPSCs, which originated from a cranial neural crest lineage, retain a remarkable potential for neuronal differentiation or induction and additionally express multiple factors that are suitable for neural or axonal regeneration (Luo et al., 2018).

Osteogenic differentiation/mineralization may be driven by secreted factors such as insulin growth factor (IGF) and IGF binding proteins, which were measured in the DPSCs ECM (IGFBP2, IGFBP4, IGFBP5, IGFBP5, and IGFBP7). Moreover, transforming growth factor (TGF) beta pathway appears to be regulated by some of the detected ECM proteins (Supplementary Figure S3). Last, we know that cells binding to substrate also regulate cells fate, so integrin binding and focal adhesion may also impact these cells' behavior. Otherwise, in OM-dECM, annexins may play a role in the increased ability to induce mineralization as they appeared to be enriched and are known mineralization factors (Cmoch et al., 2011; Genetos et al., 2014; Pan et al., 2015).

The present study showed that annexins are enriched in OM-ECM vs. N-ECM. We hypothesize that annexins could participate in the osteogenic inductive properties of OM-dECM derived from DPSCs. A positive role of annexins on osteogenic differentiation and ossification has been reported (Grewal et al., 2021). ANXA1, ANXA4, ANXA5, ANXA6, ANXA7, and ANXA11 were enriched in OM-ECM compared to N-ECM in the present study. Rat bone marrow MSC transfected with shRNA against annexin1 exhibited a reduction of mineralization in vitro and osteogenic marker gene expression (Pan et al., 2015). Annexin A2 or A5 knockdown in murine pre-osteoblast cell line resulted in the dramatic decrease of alkaline phosphatase enzymatic activity, while mineral deposition and osteogenic marker gene expression were not robust (Genetos et al., 2014). Annexins A1, A6, A7, and A11 were associated with S100 proteins, especially S100-A6 and S100-A10, in mineralizing matrix vesicles from osteoblasts (Cmoch et al., 2011). Consistent with the present study, annexin and S100 proteins were detected in DPSCs-derived OM-ECM; however, the S100 proteins ratio are not considered significantly different from N-ECM. These data support our hypothesis regarding the role of increased annexins in OM-ECM on its mineralization inductive properties. Therefore, the influence of annexins on ECM derived from DPSCs in mineralization requires further investigations.

On the other hand, the collagens and glycoproteins were enriched in N-ECM compared to OM-ECM. These proteins are classically associated with connective soft tissue matrix, such as elastic fibers or non-fibrillary collagens (COL6). Correspondingly, biological functions of OM-ECM showed an enrichment of inorganic substances and NABA-affiliated protein interactions, whereas N-ECM was enriched with elastic fibers and integrin interactions. We also demonstrated an over-representation of classical fiber proteins in N-ECM: fibril forming, FACIT, and beaded collagens together with some major ECM glycoproteins such as fibronectin. This confirmed the ECM proteins differences N-ECM and OM-ECM corresponded to the usual soft tissue and mineralization matrix. It highlights the ability of DPSCs to produce different ECM and the similarities between in vivo conditions and our experiments. However, it is also noteworthy that most of the ECM proteins are not differentially detected; the biological properties can be modified through few ECM changes, meaning the great impact of dental pulp biological functions.

CXCL12, also known as stromal cell-derived factor 1 (SDF1), is a chemokine protein that functions as a chemotactic factor for mesenchymal stem cells and is expressed in areas of inflammatory bone destruction, where it mediates their suppressive effect on osteoclastogenesis (Takano et al., 2014) as well as an important role in angiogenesis (Zhang and He, 2019). On the other hand, Soluble frizzled-related proteins (sFRPS) modulate Wnt signaling by direct interaction with Wnts and secreted frizzled-related protein 1 (SFRP1). They play a part in controlling cell development and differentiation in certain cell types. Wnt signaling components have positive impacts on pulp healing, dentin repair, and stemness maintenance epigenetic control (Yaemkleebbua et al., 2019; Kornsuthisopon, 2022). SFRP1 is a Wnt antagonist that acts as an ECM inhibitor of the Wnt signaling pathway, preventing osteoblast-induced osteoclastogenesis and maintaining the nonmineralized state of PDL progenitors (Hausler et al., 2004; Gopinathan et al., 2019). Therefore, the overexpression of CXCL12 and SPFR1 on N-ECM could be attributable to the regulation of DPSC homeostasis in the nonmineralized condition. Indeed, the signaling pathway regulates dental pulp homeostasis, and dentin regeneration required further investigation.

An ECM derived from cell culture is a natural biomaterial. This potential scaffold is free of cellular components, while retaining the complex network of both structure and functional protein assembled in their environment. Under chemically controlled settings, cell-derived extracellular matrices have the potential to be utilized as tissue replacements in a relatively short period of time (Ahlfors and Billiar 2007). Decellularization is completed by eliminating genetic material with DNase to reduce host immunological reactivity, as seen in tissue-derived ECMs (Crapo et al., 2011). Thus, in the future, cell culture-based ECMs could be considered as an innovative technique for direct differentiation of even somatic cells into other cell types, according to the findings of this study. Correspondingly, decellularization of natural ECM derived from cell culture showed potential in cellular biological properties such as cell proliferation and gene expression as well as cell remodeling and differentiation (Parmaksiz et al., 2020). However, dECM derived from cells as a two-dimensional (2D) structure lacks mechanical strength and stability. Hybridized decellularization with crosslinking agents, synthetic polymers, or hydrogels might be necessary to improve dECM scaffolds towards increased efficiency and higher innovation of natural biomaterial to be used in regenerative medicine (Nouri Barkestani et al., 2021).

dECM derived from tissues has shown the potential to be used as regenerative materials, for example, in pulp therapy (Bakhtiar et al., 2020). Furthermore, dentin ECM components promoted DPSC mineralization (Petridis et al., 2019). Thus, decellularized ECM derived from tissue or cells requires further investigation.

GF were selected as a model for reseeding on dECM in order to determine the regulation of DPSC-derived ECM on osteoblastic differentiation and mineralization induction. GF are the most abundant oral cell types that could be isolated from patients. The tissue availability and non-invasive harvesting procedures support the use of GF in regenerative dentistry (Fournier et al., 2013). They present less osteogenic differentiation potency compared with other dental mesenchymal stem cells (Dan et al., 2014). Therefore, GF were suitable to observe the putative dECM inductions on osteogenic differentiation and mineralization. Reseeded cells on dECM derived from DPSCs in different culture media did not affect cell viability responses. Both N-dECM and OM-dECM did not affect GF cell viability. This result is similar to the study in hADSC showing that there was no significant alteration of cell proliferation on dECM (Guneta et al., 2017). On the contrary, dECM derived from bone marrow increased periodontal ligament stem cells (PDLSCs) number compared to dECM derived from PDLSCs (Wen et al., 2019). These results imply that DPSCs ECM derived from different induction conditions do not markedly influence the cell proliferative function.

OM-dECM derived from DPSCs dramatically promoted osteogenic differentiation of GF compared to N-dECM in the presence or absence of osteogenic inductive factors. COL1A1 mRNA expression was downregulated in both N-dECM and OM-dECM at mineralization state. Col1a1 messenger RNA (mRNA) is an early marker during osteogenic differentiation appearing after mineralization process (Dacic et al., 2001). Likely, the osteogenic marker genes expression of GF in N-dECM showed upregulated ALP, RUNX2, and OCN. The upregulation of ALP and RUNX2 were expressed in GF on OM-dECM but not significantly different when cultured under different environments. Correspondingly, dECM derived from differentiation osteoblast showed more effective osteogenic differentiation than growth medium-derived dECM (Jeon et al., 2018).

The Alizarin Red S and Von Kossa staining on N-dECM and OM-dECM confirmed that OM-dECM was sufficient to induce mineralization. Interestingly, reseeded GF on dECM enhanced mineralization faster than that cultured on TCP at day 14. “Besides, mineralization, at 14 days under osteogenic induction condition, was significant in both N-dECM and OM-dECM. No differences were observed at this timepoint between conditions. However, observations along time might reveal mineralization differences between those groups” (Misof et al., 2020).

Taking all data together, the results implicate that the mineralization of GF can be induced by OM-dECM, which may be linked to specific proteins from OM-ECM. Thus, OM-dECM is mineralization-inductive.

The present study shows that matrisome proteins of DPSCs-derived ECM differ according to culture conditions, with increased core matrisome proteins (collagens, elastic fibers associated) in N-ECM, while matrisome-associated proteins (annexins) were enriched in OM-ECM. We demonstrated that ECM impacts cell behavior and differentiation. DPSCs ECM proved to be mineralization inductive (Figure 5). Our identified ECM proteins can be used as a marker for osteogenic differentiation. Further study will investigate the patterns of those proteins during osteogenic differentiation. This in vitro ECM or some of its proteins may be helpful for tissue engineering and could be used for mineralized tissue therapy or to decorate biomaterials.