A High-Resolution Proteomic Landscaping of Primary Human Dental Stem Cells: Identification of SHED- and PDLSC-Specific Biomarkers

Dental stem cells (DSCs) have emerged as a promising tool for basic research and clinical practice. A variety of adult stem cell (ASC) populations can be isolated from different areas within the dental tissue, which, due to their cellular and molecular characteristics, could give rise to different outcomes when used in potential applications. In this study, we performed a high-throughput molecular comparison of two primary human adult dental stem cell (hADSC) sub-populations: Stem Cells from Human Exfoliated Deciduous Teeth (SHEDs) and Periodontal Ligament Stem Cells (PDLSCs). A detailed proteomic mapping of SHEDs and PDLSCs, via employment of nano-LC tandem-mass spectrometry (MS/MS) revealed 2032 identified proteins in SHEDs and 3235 in PDLSCs. In total, 1516 proteins were expressed in both populations, while 517 were unique for SHEDs and 1721 were exclusively expressed in PDLSCs. Further analysis of the recorded proteins suggested that SHEDs predominantly expressed molecules that are involved in organizing the cytoskeletal network, cellular migration and adhesion, whereas PDLSCs are highly energy-producing cells, vastly expressing proteins that are implicated in various aspects of cell metabolism and proliferation. Applying the Rho-GDI signaling pathway as a paradigm, we propose potential biomarkers for SHEDs and for PDLSCs, reflecting their unique features, properties and engaged molecular pathways.


Introduction
Tissue engineering represents one of the most promising advances in regenerative medicine, with applications that extend as far as the restoration or regeneration of damaged tissue [1][2][3]. The field

Protein Class-Function of Molecules Identified in Both SHEDs and PDLSCs
Abundant molecules, identified in both SHEDs and PDLSCs by proteomic landscaping, were cytoskeletal proteins, as expected. This was indicated, among others, by the protein coverage, number of unique peptides and mascot score identified (Tables S1 and S2). Several members of tubulin family (α-1B, -1C, and 4A, and β-4B, -3, -2A, and -6), the main component of microtubules, were ranking high in the protein list. These hollow fibers (microtubules) serve as a skeletal system for living cells ( Figure 1B) and have the ability to shift through various formations enabling the cell to undergo mitosis or to regulate intracellular transport [35]. Moreover, α actinins (−4 and −1), actin-binding proteins residing along microfilament bundles and adherence-type junctions ( Figure 1C), were also in high abundance. Furthermore, vimentin ( Figure 1D), a type III intermediate filament that is the major cytoskeletal component of mesenchymal cells [36], was also highly expressed, providing additional evidence for the stemness character of these cells.
Through the Kyoto Encyclopedia of Genes and Genomes (KEGG) bioinformatics software, we further clustered the above data into broader categories of function, such as metabolism (n = 872), cellular processes (n = 505), localization (n = 274), biological regulation (n = 205), cellular components of organization/biogenesis (n = 186), developmental processes (n = 155) and responses to stimulus (n

Protein Class-Function of Molecules Identified in Both SHEDs and PDLSCs
Abundant molecules, identified in both SHEDs and PDLSCs by proteomic landscaping, were cytoskeletal proteins, as expected. This was indicated, among others, by the protein coverage, number of unique peptides and mascot score identified (Tables S1 and S2). Several members of tubulin family (α-1B, -1C, and 4A, and β-4B, -3, -2A, and -6), the main component of microtubules, were ranking high in the protein list. These hollow fibers (microtubules) serve as a skeletal system for living cells ( Figure 1B) and have the ability to shift through various formations enabling the cell to undergo mitosis or to regulate intracellular transport [35]. Moreover, α actinins (−4 and −1), actin-binding proteins residing along microfilament bundles and adherence-type junctions ( Figure 1C), were also in high abundance. Furthermore, vimentin ( Figure 1D), a type III intermediate filament that is the major cytoskeletal component of mesenchymal cells [36], was also highly expressed, providing additional evidence for the stemness character of these cells.
Through the Kyoto Encyclopedia of Genes and Genomes (KEGG) bioinformatics software, we further clustered the above data into broader categories of function, such as metabolism (n = 872), cellular processes (n = 505), localization (n = 274), biological regulation (n = 205), cellular components of organization/biogenesis (n = 186), developmental processes (n = 155) and responses to stimulus (n = 105) (p < 0.05). Altogether, these findings indicate the tight regulation of metabolism and biogenesis that adult dental stem cells undergo while remaining in an undifferentiated state ( Figure 2B).

Proteins Exclusively Identified in SHEDs
In total, 517 proteins were exclusively identified in SHEDs, appropriately analyzed by applying the Gene Ontology sub-routine of DAVID software and further sub-categorized based on their contribution to the "Biological Process" bioinformatics filter. According to this classification, SHEDspecific proteins seem to be implicated in: cell adhesion (n = 42), regulation of localization (n = 37), cell motility (n = 21), regulation of cell migration (n = 15), cytoskeleton organization (n = 25) and actin cytoskeleton organization (n = 17). Moreover, they were characterized as engaged in: anatomical structure development (n = 96), system development (n = 88), skeletal system development (n = 22) and regulation of cell morphogenesis (n = 11). The above data indicate that SHEDs express several proteins critically involved in cytoskeletal plasticity, motility, migration, development and morphogenesis, while maintaining their mesenchymal character ( Figure 3A).
Moreover, the proteins uniquely identified in SHEDs were suitably processed using the core analysis of Ingenuity Pathway Analysis (IPA) and a heat map was produced ( Figure 3B). Interestingly, a concrete reassembly of 137 out of the 517 total SHED-specific proteins were linked to cellular movement, migration and invasion (2.46 × 10 −3 < p < 6.15 × 10 −9 ). This is in accordance with

Proteins Exclusively Identified in SHEDs
In total, 517 proteins were exclusively identified in SHEDs, appropriately analyzed by applying the Gene Ontology sub-routine of DAVID software and further sub-categorized based on their contribution to the "Biological Process" bioinformatics filter. According to this classification, SHED-specific proteins seem to be implicated in: cell adhesion (n = 42), regulation of localization (n = 37), cell motility (n = 21), regulation of cell migration (n = 15), cytoskeleton organization (n = 25) and actin cytoskeleton organization (n = 17). Moreover, they were characterized as engaged in: anatomical structure development (n = 96), system development (n = 88), skeletal system development (n = 22) and regulation of cell morphogenesis (n = 11). The above data indicate that SHEDs express several proteins critically involved in cytoskeletal plasticity, motility, migration, development and morphogenesis, while maintaining their mesenchymal character ( Figure 3A).
Moreover, the proteins uniquely identified in SHEDs were suitably processed using the core analysis of Ingenuity Pathway Analysis (IPA) and a heat map was produced ( Figure 3B). Interestingly, a concrete reassembly of 137 out of the 517 total SHED-specific proteins were linked to cellular movement, migration and invasion (2.46 × 10 −3 < p < 6.15 × 10 −9 ). This is in accordance with the result obtained from wound-healing assays performed for SHEDs and PDLSCs ( Figure 3C), demonstrating the strong cellular capacity of SHEDs to migrate towards the artificial wound. SHEDs initially demonstrated a tendency for higher wound-healing activity than PDLSCs, but this difference was statistically insignificant ( Figure 3D). In fact, 12 h after the artificial wound was generated, both cell populations revealed 100% wound-healing proficiency, indicating their excellent proliferation and migration capacity to heal a wound.
Additionally, 134 proteins were detected to be implicated in cellular assembly and organization (2.40 × 10 −3 < p < 2.40 × 10 −8 ), and 141 ones in cellular function and maintenance (2.41 × 10 −3 < p < 2.40 × 10 −8 ). The tight regulation that SHEDs linked to their cellular morphology, assembly and organization is highlighted by the network created via IPA employment, with its core components being identified as the binding partner of Hsp70, STIP1 homology and U-box containing protein 1 (STUB1), Ras-related C3 botulinum toxin substrate 1 (RAC1) and arginine methyltransferase 6 (PRMT6) ( Figure 3E). Additionally, in regards to their cellular movement process, the network in Figure 3F illustrated the critical interactions between the core components and their binding partners that were all recognized among proteins exclusively detected in SHEDs (e.g., Gαq and Raf ( Figure 3E); trypsin and WNT5A ( Figure 3F)). These findings strongly indicate that SHEDs contain an increased number of protein molecules being implicated in the control of cellular plasticity and adaptation to extrinsic and environmental stimuli. the result obtained from wound-healing assays performed for SHEDs and PDLSCs ( Figure 3C), demonstrating the strong cellular capacity of SHEDs to migrate towards the artificial wound. SHEDs initially demonstrated a tendency for higher wound-healing activity than PDLSCs, but this difference was statistically insignificant ( Figure 3D). In fact, 12 h after the artificial wound was generated, both cell populations revealed 100% wound-healing proficiency, indicating their excellent proliferation and migration capacity to heal a wound. Additionally, 134 proteins were detected to be implicated in cellular assembly and organization (2.40 × 10 −3 < p < 2.40 × 10 −8 ), and 141 ones in cellular function and maintenance (2.41 × 10 −3 < p < 2.40 × 10 −8 ). The tight regulation that SHEDs linked to their cellular morphology, assembly and organization is highlighted by the network created via IPA employment, with its core components being identified as the binding partner of Hsp70, STIP1 homology and U-box containing protein 1 (STUB1), Ras-related C3 botulinum toxin substrate 1 (RAC1) and arginine methyltransferase 6 (PRMT6) ( Figure 3E). Additionally, in regards to their cellular movement process, the network in Figure 3F illustrated the critical interactions between the core components and their binding partners that were all recognized among proteins exclusively detected in SHEDs (e.g., Gαq and Raf ( Figure  3E); trypsin and WNT5A ( Figure 3F)). These findings strongly indicate that SHEDs contain an increased number of protein molecules being implicated in the control of cellular plasticity and adaptation to extrinsic and environmental stimuli.  Protein molecules identified by our analysis are shown in green color, whereas the ones missing from the list (but still belonging to the IPA networks) remain uncolored.

Proteins Exclusively Identified in PDLSCs
On the other hand, the majority of proteins exclusively identified in PDLSCs appeared to be mainly classified in the cellular-metabolism network ( Figure 4A), as indicated by the Gene Ontology software (p < 0.001). Bioinformatics processing, according to the "Biological Process" filter, unveiled a large number of proteins categorized in: cellular protein metabolic process (n = 303), regulation of protein metabolic process (n = 77), DNA metabolic process (n = 85), nucleoside metabolic process (n

Proteins Exclusively Identified in PDLSCs
On the other hand, the majority of proteins exclusively identified in PDLSCs appeared to be mainly classified in the cellular-metabolism network ( Figure 4A), as indicated by the Gene Ontology software (p < 0.001). Bioinformatics processing, according to the "Biological Process" filter, unveiled a large number of proteins categorized in: cellular protein metabolic process (n = 303), regulation of protein metabolic process (n = 77), DNA metabolic process (n = 85), nucleoside metabolic process (n = 16), RNA metabolic process (n = 179), mRNA metabolic process (n = 86) and non-coding (nc) RNA metabolic process (n = 52). Indicative of the intense metabolic activity of PDLSCs is the strong presence of mitochondria observed in these cells when compared to SHEDs, as demonstrated by Mitotracker Red staining ( Figure 4B). Additionally, the PDLSC-specific proteins herein recognized, and categorized in Figure 4A, were also clustered in: RNA processing, splicing and transport, as well as, RNA export from nucleus and nucleocytoplasmic transport, translation, chromatin modification and other functions that primarily involve regulatory effects, metabolism and biosynthesis. RNA and protein synthesis, and homeostasis seem to dominate PDLSCs' intracellular activities.
as, RNA export from nucleus and nucleocytoplasmic transport, translation, chromatin modification and other functions that primarily involve regulatory effects, metabolism and biosynthesis. RNA and protein synthesis, and homeostasis seem to dominate PDLSCs' intracellular activities.
The 1721 proteins, exclusively identified in PDLSCs, were further analyzed through the employment of IPA bioinformatics tool and two heat maps predominantly emerged ( Figure 4C), containing 545 proteins tightly implicated in cellular growth and proliferation (4.29 × 10 −3 < p < 2.89 × 10 −10 ), and 207 ones strongly involved in DNA replication, recombination and repair (2.33 × 10 −3 < p < 1.95 × 10 −10 ). Interestingly, one of the most concrete networks derived from IPA platform ( Figure 4D) contains as core components the centrosomal protein 55 (CEP55), ECSIT signaling integrator, required for proliferation and differentiation of embryonic stem cells [37] and protein phosphatase 4 regulatory subunit 3A (PPP4R3A) which plays a key role in gluconeogenesis of hepatic cells [38]. Moreover, the increased PDLSC cellular division observed in vitro ( Figure 4E) can be, at least, partially attributed to the central role of Rac GTPase activating protein 1 (RACGAP1) in the network of Figure 4F, a component of the centralspindlin complex which is located at the central spindle and midbody, and is highly associated with cytokinesis [39]. Overall, PDLSCs revealed a notable mitogenic advantage as indicated by growth curves and the protein molecules identified by our proteomic approach.

Potential Applications in the Biomarkers Field-A Paradigm of Different Markers on the Same Pathway
To comprehend the value of the high-throughput analysis which emerged from the successful landscaping of the proteome of SHEDs and PDLSCs, a direct comparative analysis of the identified proteins was attempted, in several canonical pathways, by suitably engaging the IPA platform. This comparison between SHED and PDLSC expressed proteins, strongly suggested that, even though they share basic molecular functions, they also exhibit differences in distinctive specialized pathways, as illustrated by an integrated heat map, embracing several canonical pathways ( Figure 5A).
To validate this scenario, the Rho-GDI signaling pathway (as reconstructed via IPA), was selected since it contains protein molecules critically implicated in various cellular processes, such as cellular migration and reorganization of cytoskeleton (mainly observed in SHEDs) and proliferation The 1721 proteins, exclusively identified in PDLSCs, were further analyzed through the employment of IPA bioinformatics tool and two heat maps predominantly emerged ( Figure 4C), containing 545 proteins tightly implicated in cellular growth and proliferation (4.29 × 10 −3 < p < 2.89 × 10 −10 ), and 207 ones strongly involved in DNA replication, recombination and repair (2.33 × 10 −3 < p < 1.95 × 10 −10 ). Interestingly, one of the most concrete networks derived from IPA platform ( Figure 4D) contains as core components the centrosomal protein 55 (CEP55), ECSIT signaling integrator, required for proliferation and differentiation of embryonic stem cells [37] and protein phosphatase 4 regulatory subunit 3A (PPP4R3A) which plays a key role in gluconeogenesis of hepatic cells [38]. Moreover, the increased PDLSC cellular division observed in vitro ( Figure 4E) can be, at least, partially attributed to the central role of Rac GTPase activating protein 1 (RACGAP1) in the network of Figure 4F, a component of the centralspindlin complex which is located at the central spindle and midbody, and is highly associated with cytokinesis [39]. Overall, PDLSCs revealed a notable mitogenic advantage as indicated by growth curves and the protein molecules identified by our proteomic approach.

Potential Applications in the Biomarkers Field-A Paradigm of Different Markers on the Same Pathway
To comprehend the value of the high-throughput analysis which emerged from the successful landscaping of the proteome of SHEDs and PDLSCs, a direct comparative analysis of the identified proteins was attempted, in several canonical pathways, by suitably engaging the IPA platform. This comparison between SHED and PDLSC expressed proteins, strongly suggested that, even though they share basic molecular functions, they also exhibit differences in distinctive specialized pathways, as illustrated by an integrated heat map, embracing several canonical pathways ( Figure 5A).
Altogether, it appears that SHEDs and PDLSCs (at least the sub-populations isolated and applied in our assays) have adopted different ways to control their respective actin-cytoskeleton networks, thus dictating the structural and morphological plasticity of dental mesenchymal stem cells during organ's development and morphogenesis. The abovementioned molecules are suggested, potential biomarkers that are associated with the cells' features, properties and prospective. These findings need further investigation to concretely identify the molecules that are representative of SHEDs and PDLSCs, and are dynamic biomarkers for different dental subpopulations linked to their cellular roles, functions, and applications. Specific experiments need to be performed, focusing on each one of the abovementioned molecules, to investigate their actual ability to affect structure, adhesion, proliferation, migration and differentiation of SHEDs and PDLSCs. The latter is a proof of the effectiveness of the potency of the cells as well as of their features.  To validate this scenario, the Rho-GDI signaling pathway (as reconstructed via IPA), was selected since it contains protein molecules critically implicated in various cellular processes, such as cellular migration and reorganization of cytoskeleton (mainly observed in SHEDs) and proliferation (predominantly found in PDLSCs). By annotating the proteins engaged in the Rho-GDI signaling pathway for both SHED (purple circle-perimeter) and PDLSC (grey circle-filling) cell types ( Figure 5B), several critical protein molecules proved to be shared between the examined sub-populations, including cadherin, Rho (GDP/GTP) and phosphatidylinositol-4-phosphate 5-kinase (PIP5K). However, SHEDs exclusively express cofilin 2, myosin light polypeptide 9, RAC1, integrin β-2 and integrin α-2 components that are all being related to actin-cytoskeleton integrity, signaling and function, and which could be recognized as potential biomarkers for the successful characterization of SHEDs, related to their distinct nature and properties. On the other hand, PDLSCs expressed exclusively the serine/threonine-protein kinase PAK6, p21-activated protein kinase-interacting protein 1 (PAK1IP1), Rho-associated protein kinase 1 (ROCK1), Wiskott-Aldrich syndrome protein family member 2 (WASF2) and protein phosphatase 1 regulatory subunit 12A (PPP1R12A) that might contribute to the elevated proliferation potency of PDLSCs.
Altogether, it appears that SHEDs and PDLSCs (at least the sub-populations isolated and applied in our assays) have adopted different ways to control their respective actin-cytoskeleton networks, thus dictating the structural and morphological plasticity of dental mesenchymal stem cells during organ's development and morphogenesis. The abovementioned molecules are suggested, potential biomarkers that are associated with the cells' features, properties and prospective. These findings need further investigation to concretely identify the molecules that are representative of SHEDs and PDLSCs, and are dynamic biomarkers for different dental sub-populations linked to their cellular roles, functions, and applications. Specific experiments need to be performed, focusing on each one of the abovementioned molecules, to investigate their actual ability to affect structure, adhesion, proliferation, migration and differentiation of SHEDs and PDLSCs. The latter is a proof of the effectiveness of the potency of the cells as well as of their features.

Discussion
Previous attempts have been made by research groups to catalogue the proteomic content of dental stem cell (DSC) sub-populations. Traditional proteomic approaches were performed, comparing cells being isolated from the same tissue however from different donors with varying health [40] age [28] and different sub-populations of cells [27]. Although 2-DE proteomic strategies provide a first insight to the proteomic landscape of the examined cells, it also exhibits a number of serious limitations, such as the inability to isolate acidic, basic and hydrophobic (membrane) proteins [29], and limited number of obtained proteins. To overcome these limitations, we applied, for the first time, the advanced gel-free nano-LC-MS/MS technology to provide the full proteome of SHEDs and PDLSCs. In total, 2032 proteins were identified in SHEDs and 3235 in PDLSCs. Overall, 1516 were commonly expressed in both cell types, while 517 were exclusively present in SHEDs and 1721 in PDLSCs.
Through thorough examination of the common protein list, several proteins that concurred within the stem-cell nature of SHEDs and PDLSCs were retrieved. As expected, at the top of our list, we re-confirmed the presence of cardinal mesenchymal markers: vimentin, a type III intermediate filament abundantly expressed in mesenchymal tissues [36] as well as an essential regulator of cell division, differentiation and apoptosis that operates in a PKC-dependent phosphorylation manner [41]; galectin-1, an abundant lectin of mesenchymal stem cells (MSCs), which is involved in immunological functions of human mesenchymal stem cells (hMSCs) [42]; and protein DJ-1, which is secreted by adult stem cells (ASCs) and induces their differentiation into osteoblasts [43]. Another important molecule identified in our samples was LASP1, a dynamic focal adhesion protein which is necessary for cell migration and survival in response to growth factors and extracellular matrix proteins, previously reported in a mesenchymal stem cell (MSC) study [44]. Mesenchymal surface markers, such as CD73 and CD105, and potential ones, such as CD29, CD13 and CD44, that were directly engaged in our in vitro experimental protocols and shown to be highly expressed, were also recognized in the nano-LC-MS/MS-mediated large inventory of our obtained proteins. Interestingly, CD166 that was found, by FACS analysis, to follow a stronger expression profile in SHEDs, compared to PDLSCs, that could be identified only in the SHED-protein list and not in the PDLSC one in a statistically significant manner; a finding that needs further investigation.
The anti-apoptotic potential of SHEDs and PDLSCs is highlighted by the presence of 53 proteins involved in the process. Among them, molecules that were identified by nano-LC-MS/MS protocols include annexin A1, heat shock 27 kDa protein 1 and heat shock 70 kDa protein 9 (mortalin). Furthermore, the list is supplemented by annexin A4, annexin A5, cofilin 1, peroxiredoxin isoform 2, 3 and 5, heat shock 60 kDa protein 1 (chaperonin) isoform, heat shock 70 kDa protein 1A and heat shock 70 kDa protein 1B. Of great interest is also the identification of heterogeneous nuclear ribonucleoprotein A2/B1, which controls the G1/S transition of human embryonic stem cells (hESCs) and, therefore, regulates their self-renewal and pluripotency [45], and elongation factor Tu, a protein that was previously proposed as a human embryonic stem cell (hESC) marker [27], which has been herein identified and previously described in PDLSCs [28].
By analyzing the commonly expressed proteins of SHEDs and PDLSCs according to their molecular functions, interesting observations have emerged regarding their metabolic and proliferative capacities. More specifically, out of the 1516 proteins, 872 ones were categorized in the cell-metabolism class, including proteins being implicated in ribosomal, proteasomal and spliceosomal complexes, or other ones involved in oxidative phosphorylation, citrate cycle (TCA cycle), ATP synthesis/binding and pyruvate/fatty-acid metabolism. Moreover, we identified a plethora of enzymes engaged in glycolysis, such as the triosephosphate isomerase 1, aldolase C fructose-bisphosphate, enolases 1 α and 2, glyceraldehyde-3-phosphate dehydrogenase, hexokinases 1 and 2, and lactate dehydrogenases A and B, suggesting that the examined cells might carry the ability to survive and proliferate under hypoxic conditions.
Besides cellular metabolism, the complexity and integrity of cytoskeletal networks in SHED and PDLSC cells are reflected on the 196 identified proteins that belong to the cytoskeleton assembly, including 140 proteins engaged in cell structure and motility. For example, we recognized several myosins (IB, IC, VI, XVIIIA, myosin heavy chain 10 and 9, and myosin light chain 12A and 6), tropomyosins (2B, 3 and 4), tubulins (α 1C and 4A, β 2A and 2C, β 3, and β 5 and 6), actins (1, α 1, and actin related protein 2 and 3), vinculin, and profilin 1 and 2, which are indispensable for biogenesis and organization of cytoskeleton, and are also implicated in differentiation-lineage commitment [28]. An observation of major significance is the identification of cofilin 1, a molecule that regulates actin cytoskeleton and is involved in the control of cell morphology and organization [46], and is abundantly expressed in human mesenchymal stem cells (hMSCs) [47]. Cofilin interacts with destrin or ADF (actin depolymerization factor), which has also been detected in our proteomic catalogue, to depolymerize actin filaments, while it (cofilin) also tethers to other actin-binding proteins to assemble stress fibers and lamellipodia formed by filamentous actin [48,49]. Finally, the identification of 48 proteins participating in focal adhesion highlights the ability of SHEDs and PDLSCs to firmly attach onto surfaces and also modify these adhesions in response to changes in the molecular composition, structure and physical forces exerted by extracellular-matrix environments [50].
The proteins exclusively identified in SHEDs are mainly related to cell migration and motility, morphogenesis and signal transduction. Cell migration is a dynamic process, in which the cytoskeleton undergoes dramatic rearrangements through polymerization and depolymerization of actin, tightly regulated by members of the Rho family of small GTPases [51]. In one of the networks created by IPA, we identified Rac1 as a key player of cellular physiology, thus supporting the suggestion that SHEDs display a higher migratory activity than PDLSCs one. Rac1 is implicated in cell-cell adhesion [52] and mesenchymal-like-driven cell migration [53], while it interacts with FH1/FH2 domain-containing protein 1 (FHOD1), also found to be expressed specifically in SHED cells. Formins act as nucleators of actin filaments and FHOD1 contributes to F-actin organization of mesenchymal cells, as well as to cell migration and invasion [54]. On the other hand, the second network, herein created by IPA, indicates that SHEDs (albeit their strong migration potency) retain the ability to firmly attach onto solid scaffolds. This is also supported by the identification of serine proteinases trypsin 1 and 3, which can degrade proteins of the extracellular matrix, but also by the expression of fibrillin 1, a major component of microfibrils in the extracellular matrix, abundantly synthesized in mesenchymal cells [55]. Additionally, 49 proteins related to cell adhesion could be exclusively recognized in SHEDs, including protein molecules such as collagen XI, XVI and XXIV, contactin 1, dystonin, integrin a2 and b2, OB-cadherin, N-cadherin, K-cadherin and cell adhesion molecule 1.
Interestingly, several proteins implicated in neuronal functions were also included in our SHED-specific proteomic catalogue. Strikingly, molecules such as N-cadherin (cadherin type 1, neuronal), with potential engagement (besides the acquisition of cell's mesenchymal character) in neuronal processes [56], and neuropilin-1, which participates in axonal fasciculation, neuronal migration, dendritic guidance and repair of the adult nervous system [57], were identified. Moreover, proteins associated with cell migration were detected, including the: (a) FEZ family zinc finger 1, a transcriptional repressor involved in axonal projection and proper termination of olfactory-sensory neurons [58]; (b) apolipoprotein A-I, which (among other functions) induces translocation of cholesterol, phospholipid and caveolin-1 to cytosol in astrocytes [59]; and (c) versican, one of the major extracellular-matrix proteins in the brain, that contains a regulatory domain of neurite growth and synaptic transmission of hippocampal neurons by EGF-R activation [60], which has also been found to interact with fibrillin 1. Tightly associated with morphogenesis, the plexin 1B, a SEMA4D receptor with a critical role in RhoA activation, actin-cytoskeleton structure and plasticity, axon guidance, and invasive growth [61], was included in our list. These are some representative examples of the long catalogue of neuron-associated proteins that characterize our SHED proteome. Their expression could be strongly associated with the neural crest-cell origin of dental pulp, the principal source of SHED-cells isolation [62].
The analysis performed for the proteins exclusively identified in PDLSCs indicated that they carry strong metabolic and proliferative activities. Many PDLSC-specific proteins are related to mitosis and cell cycle. For instance, in one of the networks created by IPA, we can foresee the central role of centrosomal protein of 55 kDa (CEP55), which is essential for centrosome duplication, cell-cycle progression and cytokinesis [63], and of small kinetochore-associated protein (SKAP), whose mouse isoform was previously shown to interact with CEP55 [64] and regulate the transition from metaphase to anaphase [65]. By searching outside of the network context, indispensable protein components of DNA replication and cell division have emerged, therefore strengthening the notion that PDLSCs exhibit higher proliferation capacity, as compared to SHEDs. PDLSDs also seem to proliferate more actively in our experimental in vitro studies, whereas, SHEDs seemed to demonstrate a settled advantage regarding their migratory capacity as compared to PDLSCs.
Remarkably, our PDLSC-specific list contains the: (a) GINS complex subunit 1 (Psf1 homolog), a subunit of GINS complex that plays an essential role in the initiation and progression of DNA replication forks [66]; (b) SPC24, a component of the essential kinetochore-associated NDC80 complex required for chromosome segregation and spindle-checkpoint activity [67]; (c) Cenp-F (or mitosin) needed for kinetochore function and chromosome segregation in mitosis, and kinetochore localization of LIS1, NDE1 and NDEL1 [68]; and (d) pericentrin, a component of the pericentriolar material of centrosome [69] associated with mitotic-spindle organization in mitotic cells [70] and operating as scaffold structure of regulatory elements that support a control mechanism of cell cycle [71]. Additionally, in our protein collection exclusively characterizing the PDLSC cells, we identified the cell cycle-checkpoint protein cyclin B1 and its binding partner cyclin-dependent kinase 1 (CDK1 or CDC2), whose complex results in the formation of maturation promoting factor (MPF), one of the key components regulating cell-cycle progression and especially the G2/M cell-cycle transition [72]. Taken together, all the above data strongly suggest that PDLSCs are characterized by comprehensive cell-cycle control, high proliferation proficiency, and, therefore, rapid growth and expansion.
A comparative study between SHED and PDLSC expressed proteins was performed through the IPA platform, embracing several canonical pathways. The Rho-GDI signaling pathway was selected and further analyzed since it contains protein molecules critically implicated in various cellular processes, such as cellular migration and reorganization of cytoskeleton (mainly observed in SHEDs) and proliferation (predominantly found in PDLSCs). Although, several proteins proved to be shared between the examined sub-populations (cadherin, Rho (GDP/GTP) and PIP5K) others were exclusively expressed in SHEDs and are all being related to actin-cytoskeleton integrity, signaling and function, and which could be recognized as potential biomarkers related to SHEDs distinct nature and properties. SHEDs solely express cofilin 2, myosin light polypeptide 9, RAC1, integrin β-2 and integrin α-2 components, while, on the other hand, PDLSCs express exclusively the serine/threonine-protein kinase PAK6, PAK1IP1, a negative regulator of PAK1 [73], ROCK1, which is required for centrosome positioning and mitotic exit [74], WASF2 and PPP1R12A and might contribute to the elevated proliferation potency of these cells. Therefore, we suggested that these molecules, among others, dictate the structural, morphological and molecular plasticity of the isolated and studied dental mesenchymal stem cells and might represent potential biomarkers associated with the cellular features, properties and prospective of these dental mesenchymal cells.

LC-MS/MS-Preparation of Samples
Protein samples of~200 µL, derived from three 75 cm 2 flasks at passage 8 for SHEDs and at passage 9 for PDLSCs, were washed and precipitated with 600 µL of acetone at room temperature, overnight. Samples were then centrifuged at 3800× g for 20 min, and a solution consisting of 8 M urea and 80 mM triethyl ammonium bicarbonate (TEAB) was applied to the pellets. After homogenization, samples were sonicated in a water bath for 30 min, and protein concentration of each sample was estimated with the Bradford assay. For each sample, 200 ng was further used for the experimental protocol applied. Briefly, samples were reduced and alkylated by incubation upon 10 mM dithiothreitol and 55 mM iodoacetamide solutions, respectively, and the digestion was performed by the application of a solution of 1 µg trypsin per 40 µg of protein for 16 h at room temperature, in order to generate peptides. The lyophilization of the peptides occurred in a vacuum concentrator and a solution of 0.1% formic acid in ddH 2 O was applied for analysis by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Three independent experiments were carried out for each cell type.

LC-MS/MS and Data Bioinformatics Analysis
The analysis of extracted peptides was performed by employing the bottom-up approach in an LTQ Orbitrap Elite instrument (Thermo Scientific, Rockford, IL, USA). The mass spectrometer (MS) was coupled to a Dionex Ultimate 3000 UPLC system and a C18 Acclaim Pepmap 15 cm column (Thermo Scientific, Rockford, IL, USA), conjugated with an Acclaim Pepmap nano-trap of 2 cm (Thermo Scientific, Rockford, IL, USA), for peptide separation. The phases were: phase A, 99.9% ddH 2 O and 0.1% formic acid, and phase B, 99.9% acetonitrile and 0.1% ddH 2 O. Samples were run at a constant flow rate of 0.3 µL per min, in a linear phase B gradient in 4 h runs, interrupted by 1 h intermediate washing steps of the columns with ddH 2 O. The Orbitrap instrument operated in a positive ion mode and the 20 most intense spectra, as measured at a 60,000 resolution, were chosen for MS/MS fragmentation, employing the higher energy collision dissociation (HCD) function. For HCD of parental ions, the collision energy value was set at 35% and the time of activation was set at 0.1 ms. Ions of m/z ≥2 were subjected to MS/MS analysis, and the Proteome Discoverer software (Thermo Scientific) and the Sequest search engine (Thermo Scientific) were used to analyze the extracted ion chromatograms (raw files). The database engaged for protein identification searches was the Homo sapiens reference proteome, exactly as downloaded from UniProt (UniProt Consortium) (version 2.16) without any further modification [75]. The limits for identification were a precursor-mass tolerance of 10 ppm and a fragment-mass tolerance of 0.05 Da. The cleavage enzyme with a maximum of 0 missed-cleavage parameter was trypsin. A false-discovery rate threshold of 0.5% was set to increase the quality and reliability of all reported protein identifications.

Immunostaining and Confocal Microscopy
To visualize cells using confocal microscopy, the examined cells were seeded on cover slips and cultured under standard conditions for 24-48 h at passage 8. They were then fixed in 4% paraformaldehyde (Sigma/Aldrich, St. Louis, MO, USA) solution in 1× PBS working solution for 10 min at room temperature. Cell permeabilization was accomplished with 0.1% Triton-X 100 solution for 10 min, followed by 0.5% Triton-X 100 (Sigma/Aldrich, St. Louis, MO, USA) for 30 min. PDL and SHED primary cells were incubated in 5% BSA blocking solution for 1 h and then in a solution containing the antibody of interest in 0.5% Triton-X 100 and 1% BSA, overnight (~16 h) at 4 • C. Washing three times with 0.5% Triton-X 100 in 1× PBS was followed by cell incubation with the secondary antibody diluted in 0.5% Triton-X 100 in 1× PBS and 1% BSA for 2 h. Next, cells were washed in 0.5% Triton-X 100 before Rhodamine-Phalloidin 1:100 (Biotium, Lab supplies) was added. The cells were washed twice with 1× PBS and VECTASHIELD ® Mounting Medium (Vector Laboratories, Inc., 30 Ingold Road, Burlingame, CA 94010 USA) was added on top to prevent rapid photo-bleaching of fluorescence. The primary antibodies used were anti-α Tubulin antibody (ab18251 rabbit 1:1000), anti-α Actinin 4 antibody (ab198608 rabbit 1:100) (Abcam, Cambridge, MA, USA), anti-Vimentin antibody (#3295 rabbit 1:100) (Cell Signaling Technology, Cambridge, MA, USA) and anti-Lamin A/C antibody (ab108595 1:1000) (Abcam). The secondary antibody used was Donkey anti-Rabbit IgG (H + L) Secondary Antibody Fluor ® 488 conjugate (Invitrogen, Waltham, MA, USA). Detection of the nucleus was achieved using 4 -6 -Diamidino-2 -phenylindole (DAPI) staining. The samples were analyzed using a Leica TCS SP5 (Leica, Wetzlar, Germany) setup with a 63× objective lens and analyzed using the Leica software, LAS AF.

Migration (Wound-Healing) Assay
For the wound-healing assay, cells at passage 6 were plated at~80-85% density in 6-well plates (~2 × 10 6 cells/well) in triplicates and incubated overnight. A 10-200 µL pipette tip was used to scratch a wound through the center of each well, and cells were immediately visualized under a light microscope; this captured image corresponded to time point 0. Cells were then incubated in a 5% CO 2 humidified environment at 37 • C, until the gap was completely filled with cells. Every 5 h from the time of the scratch, the closure of the generated wound was visualized and appropriate images were taken. The statistical analysis of the percentage closure in the wound-healing assay was performed using the Image-J software.

Growth Curves
Cells at passage 6 were seeded in 6-well plates at a density of~1.5 × 10 4 cells per well and collected at 24, 48, 72, 96, 120 and 144 h, using 0.25% trypsin. They were stained with 0.2% trypan blue and counted under a phase-contrast microscope with a hemocytometer. Three independent experiments were performed for each cell sub-population.

Mitochondria Staining
Live cells were stained with MitoTracker ® Red CMXRos dye (Cell Signaling Technology) in final concentration of 100 nM for 30 min and fixed with 4% PFA (paraformaldehyde) for 10 min. Then, they were washed 3 times with 1× PBS, followed by the immunofluorescent staining protocol.

Statistical Analysis
We performed statistical analysis by employing the Student's t-test. Significant p-value corresponds to p < 0.05. Error bars represent standard deviation from the means.

Conclusions
In this study, we performed, for the first time, a high-throughput molecular comparison of two primary human adult dental stem cell (hADSC) sub-populations, the SHEDs and PDLSCs. A detailed proteomic mapping of SHEDs and PDLSCs, via employment of nano-LC tandem-mass spectrometry (MS/MS), herein reported for the first time, revealing similarities as well as differences regarding the protein expression and the functions that these molecules are likely to be involved. In total, 2032 proteins were identified to be expressed in SHEDs and 3235 in PDLSCs. Overall, 1516 proteins were expressed in both populations, while 517 were unique for SHEDs and 1721 were exclusively expressed in PDLSCs. Further analysis of the recorded proteins suggested that SHEDs are characterized by high cytoskeletal plasticity that renders them excellent candidates for stem-cell-based therapies, during which satisfactory adhesion and migration are required. On the other hand, PDLSCs seem to carry enhanced metabolism and proliferation activities, which could be of great advantage in therapeutic applications whose engrafted cells have to intensively propagate in order to repair and/or replace the damaged tissue(s). Finally, we applied the Rho-GDI signaling pathway as paradigm, and proposed potential biomarkers for SHEDs and for PDLSCs, reflecting their unique features, properties and engaged molecular pathways.
Conclusively, our data are expected to provide a benchmark for future studies regarding the proteomic landscapes, molecular networks and functional responses to extrinsic cues of primary human adult dental stem cells.