Matrisome profile of dECM_DPSCs and dECM_GSCs

The normal ECM (N-ECM) from both cell types was decellularized. Quantitative mass spectrometry analysis was performed by comparing the triplicate of each condition. Results displayed 214 quantified proteins belonging to the matrisome proteins with at least 3 total peptides (Supplementary Table ¹). The Venn diagram presented all the quantified common matrisome proteins (169 proteins) and the unique proteins from dECM_DPSCs (15 proteins) and dECM_GSCs (30 proteins) (Fig. 1d).

The number of classified matrisome differed slightly between dECM from each cell type. A total of 85 and 93 proteins from core matrisome proteins were detected in dECM_DPSCs and dECM_GSCs, respectively (Fig. 1e). As for the matrisome-associated proteins (i.e., ECM regulators, ECM-affiliated proteins, secreted factors, and unidentified proteins), dECM_DPSCs had 99 proteins, while dECM_GSCs had 106 proteins (Fig. 1f). The major core matrisome and matrisome-associated proteins in both cell types were glycoproteins and ECM regulators, respectively.

dECM_from DPSCs and dECM_from GSCs exhibited matrisome proteins related to calcium ion binding

Metascape analysis was utilized to present the protein–protein interaction of common matrisome proteins. The overlapped matrisome proteins of dECM_DPSCs and dECM_GSCs were mainly associated with protein oxidation, regulation of basement membrane organization, and homeostasis (Fig. 2a). The gene ontology enrichment analysis was carried out with the following gene ontology (GO) sources: GO Biological Process, GO Cellular Component, and GO Molecular Functions. The main functions of the common matrisome classified by GO sources were the regulation of proteolysis, collagen-containing ECM, and calcium ion binding, respectively (Fig. 2b).

Open in a separate window

Figure 2

dECM_DPSCs and dECM_GSCs exhibited matrisome protein related to calcium ion binding. (a) Network analysis of protein–protein interaction (PPI) following the Molecular Complex Detection (MCODE) components and (b) gene ontology analysis of the overlapped matrisome proteins of dECM_DPSCs and dECM_GSCs. (c) Network analysis of PPI following the MCODE components and (d) gene ontology analysis of matrisome proteins of dECM_DPSCs. (e) Network analysis of PPI following the MCODE components and (f) gene ontology analysis of matrisome proteins of dECM_GSCs. The analyzes were performed using Metascape⁴⁹.

We next analyzed the DPSCs matrisome. The Protein–protein interaction (PPI) of dECM_DPSCs matrisome proteins was primarily associated with proteins oxidation, regulation of basement membrane organization, and ECM constituent conferring elasticity (Fig. 2c). The molecular function of these proteins was mainly calcium-ion binding, while collagen containing ECM was their main cellular component (Fig. 2d).

GSCs matrisome showed that the PPI was principally involved with ECM structural constituent conferring tensile strength, peptidyl-lysine hydroxylation, and regulation of basement membrane organization (integrin) (Fig. 2e). The highest number of proteins in each GO source (biological process, cellular component, and molecular function) was involved with supramolecular fiber organization, collagen-containing ECM, and calcium ion binding, respectively (Fig. 2f).

dECM_DPSCs’ and dECM_GSCs’ signature proteins were members of the mineralization-associated pathway and structural proteins

dECM_DPSCs had 15 signature (unique) proteins consisting of TGFB2, S100A6, WNT5A, TGFB1, S100A4, HGF, FST, S100A10, WNT5B, ZFP91, PDGFD, ANGPT1, P3H1, P3H3, and OGFOD3 (Fig. 1d). Meanwhile, dECM_GSCs demonstrated 30 signature proteins classified as collagen and non-collagen cluster. Twenty-one proteins were mainly presented in the collagen cluster. Non-collagen cluster contained FN1, TNC, MXRA5, EMILIN1, THBS1, TGFB1, PXDN, and LAMC1 (Fig. 1d). Therefore, the signature proteins of dECM_DPSCs pointed out in two key osteogenic-related pathways: TGF-β and Wnt signaling pathways.

GSCs on dECM_DPSCs expressed genes involved TGF-β, Hippo, and Wnt signaling pathways

GSCs were reseeded on either N-dECM from DPSCs or GSCs for 24 h, or on OM- dECM from DPSCs or GSCs for 24 h. The expression patterns comparing between N-dECM_DPSCs and N-dECM_GSCs, and between OM-dECM_DPSCs and OM-dECM_GSCs from RNAseq data were illustrated as heatmaps (Fig. 3a,b). The top 30 significantly upregulated and downregulated genes in N-dECM_DPSCs VS N-dECM_GSCs and OM-dECM_DPSCs VS OM-dECM_GSCs were listed in Table ​Table1.1. Selected differentially expressed genes were validated using quantitative real-time polymerase chain reaction (qPCR), which significant upregulation of bone morphogenetic protein-2 (BMP2) and periostin (POSTN) mRNA levels was found in N-dECM_DPSCs (Fig. 3c), while OM-dECM_DSPCs significantly induced lymphoid enhancer binding factor 1 (LEF1) and matrix metallopeptidase 3 (MMP3) mRNA expression, compared with dECM_GSCs from the same condition (Fig. 3d).

Open in a separate window

Figure 3

dECM_DPSCs expressed genes involved TGF-β, Hippo, and Wnt signaling pathways. The expression pattern of related genes comparing between N-dECM_DPSCs VS N-dECM_GSCs, and between OM-dECM_DPSCs VS OM-dECM_GSCs. Heatmap showed the differentially regulated genes between (a) N-dECM_DPSCs and N-dECM_GSCs (b) OM-dECM_DPSCs and OM-dECM_GSCs. Differentially expressed genes of N-dECM_DPSCs VS N-dECM_GSCs and OM-dECM_DPSCs VS OM-dECM_GSCs were validated using qPCR. The differential gene expression of (c) BMP2, POSTN, and (d) LEF1, MMP3 was confirmed. Bars indicate a significant difference between groups (*p<0.05).

Table 1

Top 30 differentially upregulated and downregulated genes between N-dECM_DPSCs and N-dECM_GSCs, and between OM-dECM_DPSCs and OM-dECM_GSCs.

Ensembl gene ID	Gene symbol	Gene name	LogFc	P value
Upregulated genes compared between N-dECM_DPSCs and N-dECM_GSCs
ENSG00000211455	STK38L	Serine/threonine kinase 38 like	0.9864	2.55E−18
ENSG00000138135	CH25H	Cholesterol 25-hydroxylase	0.9938	6.61E−09
ENSG00000175745	NR2F1	Nuclear receptor subfamily 2 group F member 1	0.9974	1.59E−12
ENSG00000144476	ACKR3	Atypical chemokine receptor 3	0.9985	5.30E−09
ENSG00000149968	MMP3	Matrix metallopeptidase 3	1.0066	2.10E−32
ENSG00000164251	F2RL1	Coagulation factor II (thrombin) receptor-like 1	1.0075	1.23E−11
ENSG00000157680	DGKI	Diacylglycerol kinase iota	1.0075	4.16E−16
ENSG00000133110	POSTN	Periostin osteoblast specific factor	1.0143	3.42E−11
ENSG00000271216	LINC01050	Long intergenic non-protein coding RNA 1050	1.0145	7.01E−14
ENSG00000135362	PRR5L	Proline rich 5 like	1.0166	1.38E−21
ENSG00000109272	PF4V1	Platelet factor 4 variant 1	1.0204	8.28E−12
ENSG00000103546	SLC6A2	Solute carrier family 6 (neurotransmitter transporter) member 2	1.0245	2.16E−09
ENSG00000148848	ADAM12	ADAM metallopeptidase domain 12	1.0258	2.24E−39
ENSG00000198535	C2CD4A	C2 calcium-dependent domain containing 4A	1.0545	4.67E−10
ENSG00000107317	PTGDS	Prostaglandin D2 synthase	1.0563	8.19E−28
ENSG00000125845	BMP2	Bone morphogenetic protein 2	1.1378	9.69E−17
ENSG00000124875	CXCL6	Chemokine (C-X-C motif) ligand 6	1.1728	1.15E−43
ENSG00000174792	C4orf26	Chromosome 4 open reading frame 26	1.1783	5.72E−12
ENSG00000102468	HTR2A	5-Hydroxytryptamine (serotonin) receptor 2A G protein-coupled	1.1802	1.61E−16
ENSG00000113361	CDH6	Cadherin 6 type 2 K-cadherin	1.2225	1.48E−23
ENSG00000188064	WNT7B	Wingless-type MMTV integration site family member 7B	1.2393	1.20E−13
ENSG00000198729	PPP1R14C	Protein phosphatase 1 regulatory (inhibitor) subunit 14C	1.275	5.23E−14
ENSG00000124225	PMEPA1	Prostate transmembrane protein androgen induced 1	1.2905	5.52E−82
ENSG00000143341	HMCN1	Hemicentin 1	1.3492	2.18E−50
ENSG00000174343	CHRNA9	Cholinergic receptor nicotinicalpha 9	1.3689	8.15E−19
ENSG00000095752	IL11	Interleukin 11	1.435	1.10E−19
ENSG00000171951	SCG2	Secretogranin II	1.5984	1.30E−53
ENSG00000166670	MMP10	Matrix metallopeptidase 10	1.6177	2.91E−99
ENSG00000145794	MEGF10	Multiple EGF-like-domains 10	1.6768	6.98E−45
ENSG00000162490	DRAXIN	Dorsal inhibitory axon guidance protein	1.7635	1.85E−30
Downregulated genes compared between N-dECM_DPSCs and N-dECM_GSCs
ENSG00000140465	CYP1A1	Cytochrome P450 family 1 subfamily A polypeptide 1	−1.9653	8.05E−37
ENSG00000168477	TNXB	Tenascin XB	−1.3258	2.39E−33
ENSG00000171345	KRT19	Keratin 19	−1.2923	1.31E−16
ENSG00000246430	LINC00968	Long intergenic non-protein coding RNA 968	−1.1767	3.28E−25
ENSG00000215853	RPTN	Repetin	−1.1412	5.01E−12
ENSG00000143631	FLG	Filaggrin	−1.1264	3.02E−15
ENSG00000107984	DKK1	Dickkopf WNT signaling pathway inhibitor 1	−1.0694	1.01E−13
ENSG00000141574	SECTM1	Secreted and transmembrane 1	−1.0547	1.82E−22
ENSG00000167641	PPP1R14A	Protein phosphatase 1 regulatory (inhibitor) subunit 14A	−1.0504	1.65E−13
ENSG00000154027	AK5	Adenylate kinase 5	−1.0411	4.00E−28
ENSG00000215182	MUC5AC	Mucin 5AC oligomeric mucus/gel-forming	−1.0376	5.50E−10
ENSG00000178882	FAM101A	Family with sequence similarity 101 member A	−1.0179	1.38E−09
ENSG00000082482	KCNK2	Potassium channel subfamily K member 2	−0.9892	1.46E−22
ENSG00000188581	KRTAP1-1	Keratin associated protein 1–1	−0.9712	1.44E−08
ENSG00000145681	HAPLN1	Hyaluronan and proteoglycan link protein 1	−0.9422	1.15E−15
ENSG00000107738	C10orf54	Chromosome 10 open reading frame 54	−0.9407	5.27E−15
ENSG00000166949	SMAD3	SMAD family member 3	−0.9237	5.21E−45
ENSG00000183287	CCBE1	Collagen and calcium binding EGF domains 1	−0.9185	2.57E−14
ENSG00000104725	NEFL	Neurofilament light polypeptide	−0.9067	7.57E−16
ENSG00000163661	PTX3	Pentraxin 3	−0.9024	1.91E−19
ENSG00000173641	HSPB7	Heat shock 27 kDa protein family member 7	−0.8888	3.16E−16
ENSG00000179314	WSCD1	WSC domain containing 1	−0.8863	5.00E−15
ENSG00000131737	KRT34	Keratin 34	−0.8826	1.97E−07
ENSG00000130600	H19	H19 imprinted maternally expressed transcript (non-protein coding)	−0.8592	1.81E−07
ENSG00000198910	L1CAM	L1 cell adhesion molecule	−0.8575	6.50E−11
ENSG00000123405	NFE2	Nuclear factor erythroid 2	−0.8327	7.18E−07
ENSG00000243137	PSG4	Pregnancy specific beta-1-glycoprotein 4	−0.8144	5.87E−12
ENSG00000185112	FAM43A	Family with sequence similarity 43 member A	−0.8117	2.74E−07
ENSG00000125965	GDF5	Growth differentiation factor 5	−0.8107	3.47E−07
ENSG00000127951	FGL2	Fibrinogen-like 2	−0.8076	1.71E−06
Upregulated genes compared between OM-dECM_DPSCs and OM-dECM_GSCs
ENSG00000223414	LINC00473	Long intergenic non-protein coding RNA 473	1.3696	4.08E−09
ENSG00000129422	MTUS1	Microtubule associated tumor suppressor 1	1.3976	3.69E−08
ENSG00000125845	BMP2	Bone morphogenetic protein 2	1.4088	6.46E−07
ENSG00000112320	SOBP	Sine oculis binding protein homolog	1.4169	8.41E−08
ENSG00000169116	PARM1	Prostate androgen-regulated mucin-like protein 1	1.4173	5.19E−11
ENSG00000145335	SNCA	Synuclein alpha (non A4 component of amyloid precursor)	1.4200	7.84E−07
ENSG00000188064	WNT7B	Wingless-type MMTV integration site family member 7B	1.4207	2.35E−06
ENSG00000124225	PMEPA1	Prostate transmembrane protein androgen induced 1	1.4216	2.32E−14
ENSG00000196843	ARID5A	AT rich interactive domain 5A (MRF1-like)	1.4296	3.64E−14
ENSG00000125430	HS3ST3B1	Heparan sulfate (glucosamine) 3-O-sulfotransferase 3B1	1.4397	1.73E−07
ENSG00000133110	POSTN	Periostin osteoblast specific factor	1.4460	2.76E−09
ENSG00000149968	MMP3	Matrix metallopeptidase 3	1.4676	1.10E−22
ENSG00000271216	LINC01050	Long intergenic non-protein coding RNA 1050	1.4700	6.97E−11
ENSG00000175745	NR2F1	Nuclear receptor subfamily 2 group F member 1	1.4759	3.72E−12
ENSG00000138135	CH25H	Cholesterol 25-hydroxylase	1.5022	3.90E−08
ENSG00000107859	PITX3	Paired-like homeodomain 3	1.5133	4.26E−07
ENSG00000145423	SFRP2	Secreted frizzled-related protein 2	1.5327	3.04E−07
ENSG00000140807	NKD1	Naked cuticle homolog 1 (Drosophila)	1.5332	1.64E−07
ENSG00000109819	PPARGC1A	Peroxisome proliferator-activated receptor gamma coactivator 1 alpha	1.5474	8.42E−22
ENSG00000180616	SSTR2	Somatostatin receptor 2	1.5520	8.61E−11
ENSG00000167874	TMEM88	Transmembrane protein 88	1.2393	1.24E−07
ENSG00000145794	MEGF10	Multiple EGF-like-domains 10	1.2750	2.58E−15
ENSG00000184995	IFNE	INTERFERON epsilon	1.2905	2.42E−11
ENSG00000054598	FOXC1	Forkhead box C1	1.3492	3.74E−22
ENSG00000150907	FOXO1	Forkhead box O1	1.3689	1.63E−12
ENSG00000120659	TNFSF11	Tumor necrosis factor (ligand) superfamily member 11	1.4350	1.72E−10
ENSG00000138795	LEF1	Lymphoid enhancer-binding factor 1	1.5984	5.15E−13
ENSG00000095752	IL11	Interleukin 11	1.6177	1.54E−22
ENSG00000198535	C2CD4A	C2 calcium-dependent domain containing 4A	1.6768	4.91E−30
ENSG00000205502	C2CD4B	C2 calcium-dependent domain containing 4B	1.7635	8.49E−39
Downregulated genes compared between OM-dECM_DPSCs and OM-dECM_GSCs
ENSG00000188581	KRTAP1-1	Keratin associated protein 1–1	−1.880	7.37E−11
ENSG00000140465	CYP1A1	Cytochrome P450 family 1 subfamily Apolypeptide 1	−1.823	2.15E−14
ENSG00000100739	BDKRB1	Bradykinin receptor B1	−1.708	2.09E−17
ENSG00000143631	FLG	Filaggrin	−1.652	8.68E−18
ENSG00000246430	LINC00968	Long intergenic non-protein coding RNA 968	−1.605	2.82E−08
ENSG00000205426	KRT81	Keratin 81	−1.482	1.86E−09
ENSG00000215853	RPTN	Repetin	−1.458	5.05E−09
ENSG00000185112	FAM43A	Family with sequence similarity 43 member A	−1.448	5.95E−19
ENSG00000130600	H19	H19 imprinted maternally expressed transcript (non-protein coding)	−1.434	1.28E−06
ENSG00000221852	KRTAP1-5	Keratin associated protein 1–5	−1.432	1.72E−06
ENSG00000127951	FGL2	Fibrinogen-like 2	−1.418	2.16E−06
ENSG00000173267	SNCG	Synuclein gamma (breast cancer-specific protein 1)	−1.416	2.52E−08
ENSG00000131737	KRT34	Keratin 34	−1.410	2.88E−07
ENSG00000183287	CCBE1	Collagen and calcium binding EGF domains 1	−1.376	2.45E−12
ENSG00000212724	KRTAP2-3	Keratin associated protein 2–3	−1.349	3.75E−06
ENSG00000014257	ACPP	Acid phosphatase prostate	−1.349	1.56E−07
ENSG00000115523	GNLY	Granulysin	−1.342	3.14E−07
ENSG00000146250	PRSS35	Protease serine 35	−1.291	2.01E−10
ENSG00000106003	LFNG	LFNG O-fucosylpeptide 3-beta-N-acetylglucosaminyltransferase	−1.274	4.71E−10
ENSG00000171346	KRT15	Keratin 15	−1.269	1.34E−06
ENSG00000166949	SMAD3	SMAD family member 3	−1.249	6.82E−16
ENSG00000184599	FAM19A3	Family with sequence similarity 19 (chemokine (C–C motif)-like) member A3	−1.248	2.62E−05
ENSG00000162591	MEGF6	Multiple EGF-like-domains 6	−1.244	2.68E−16
ENSG00000253230	LINC00599	Long intergenic non-protein coding RNA 599	−1.233	3.85E−05
ENSG00000225968	ELFN1	Extracellular leucine-rich repeat and fibronectin type III domain containing 1	−1.224	6.20E−12
ENSG00000074527	NTN4	Netrin 4	−1.223	4.45E−12
ENSG00000135480	KRT7	Keratin 7	−1.218	2.23E−05
ENSG00000169583	CLIC3	Chloride intracellular channel 3	−1.208	3.41E−05
ENSG00000105974	CAV1	Caveolin 1 caveolae protein 22 kDa	−1.207	1.29E−05
ENSG00000183671	GPR1	G protein-coupled receptor 1	−1.203	1.97E−06

Open in a separate window

Bioinformatic analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database revealed several pathways regulated by N-dECM_DPSCs and OM-dECM_DPSCs. The upregulated genes in N-dECM_DPSCs were involved in the TGF-β, Hippo, and Wnt signaling pathways (Fig. 4a), whereas the downregulated genes were categorized in the ECM receptor interaction and regulation of actin cytoskeleton (Fig. 4b). As for OM-dECM_DPSCs, the upregulated genes were related to Hippo and Wnt signaling pathways (Fig. 4c). In contrast, the downregulated genes were found in the ECM receptor interaction and TGF-β signaling pathway (Fig. 4d).

Open in a separate window

Figure 4

Top 10 Kyoto Encyclopedia of Genes and Genomes (KEGG) enriched pathways for (a) up- and (b) downregulated genes in N-dECM_DPSCs VS N-dECM_GSCs and (c) up- and (d) downregulated genes in OM-dECM_DPSCs VS OM-dECM_GSCs by over-representation analysis.

Over-representation analysis was performed, and the number of genes in each gene ontology analysis for the upregulated and downregulated genes was shown in Fig. 5a–c (for N-dECM_DPSCs VS N-dECM_GSCs) and Fig. 5d–f (for OM-dECM_DPSCs VS OM-dECM_GSCs). The differentially regulated genes were mainly associated with biological regulation, membrane, and protein binding in the categories of biological process, cellular component, and molecular function, respectively.

Open in a separate window

Figure 5

Gene ontology (GO) analyzes of the upregulated and downregulated genes comparing between (a–c) N-dECM_DPSCs and N-dECM_GSCs, and between (d–f) OM-ECM_dDPSCs and OM-dECM_GSCs. The differentially regulated genes were mainly associated with biological regulation, membrane, and protein binding in the categories of (a, d) biological process, (b, e) cellular component, (c, f) and molecular function, respectively.

Characteristics and morphological appearance of ECM and dECM derived from DPSCs and GSCs

The characteristics and morphology of ECM_DPSCs, ECM_GSCs, dECM_DPSCs, and dECM_GSCs were illustrated. OM-ECM derived from both cells exhibited an increased deposit of calcium, phosphate, and alkaline phosphatase (ALP) compared with N-ECM (Fig. 6a,b). ECMs revealed fibroblast-like cells under bright-field microscope and showed similar intricate fibrillar networks observed with scanning electron microscope (SEM) (Fig. 6a). After decellularization, all dECMs were negative for ARS, Von Kossa, and ALP stainings and showed no remaining nuclei when investigated under the microscope (Fig. 6c). SEM analysis revealed dense ECM fibers after decellularization (Fig. 6c). These results confirmed that decellularization eliminated the cells and mineral deposited contents.

Open in a separate window

Figure 6

Characterization of ECM and dECM. (a–c) Morphology, mineralization, phosphate, and alkaline phosphatase (ALP) were examined. The ultrastructure of ECM and dECM was observed using scanning electron microscopic analysis (SEM). (b) Relative Alizarin Red S (ARS) quantification of N-, OM-ECM_DPSCs and N-, OM-ECM_GSCs. Bars indicate a significant difference between groups (*p<0.05). (d) Type I collagen and fibronectin expression in ECM were determined using immunofluorescence staining. The genetic component was stained using DAPI. F-actin was visualized using phalloidin staining. e Type I collagen and fibronectin expression were determined in dECM. Scale bars: 10, 20, and 300 µm.

ECM proteins (type I collagen and fibronectin) were visualized by immunofluorescence staining. Results showed that all ECMs produced both types of ECM proteins (Fig. 6d). In addition, actin filaments and nuclei were observed (Fig. 6d). After decellularization, both ECM proteins were retained in all dECMs (Fig. 6e). Absence of actin and DAPI nuclei staining confirmed that cellular components and genetic materials were removed while ECM proteins were preserved in dECMs (Fig. 6e).

Biological responses of GSCs on dECM_DPSCs and dECM_GSCs

GSCs were reseeded on dECM_DPSCs and dECM_GSCs to determine the biological responses. MTT assay was employed to evaluate the GSCs cell viability on day 1, 3, and 7. GSCs that were seeded on ECM derived from GSCs and cultured in N-condition were employed as a control. GSCs were able to proliferate on all dECMs surfaces, suggesting that dECM_DPSCs and dECM_GSCs were biocompatible (Fig. 7a). Significant upregulation of GSC proliferation was found when reseeded on OM-dECM_DPSCs at day 3 and N- and OM-dECM_DPSCs at day 7 (Fig. 7a). Cell adhesion and spreading were examined after seeding GSCs on dECM_DPSCs and dECM_GSCs for 30 min, 24 h, 3 days, and 7 days. Phalloidin immunofluorescence staining was employed to visualize the organization of F-actin in the cytoskeleton. GSCs were adhered to all dECMs at 30 min after seeding (Fig. 7b). Cell spreading was initially observed at 24 h (Fig. 7b). GSCs spread extensively, and well-organized F-actin filaments were depicted on day 7 (Fig. 7b). Further, cell morphology observed with SEM and GSCs adhered to all dECMs and appeared a round shape at 30 min (Fig. 7c). However, only GSCs seeded on N- and OM-dECM_DPSCs exhibited slightly higher filopodial and lamellipodia extensions at the same time point (Fig. 7c). On day 3, GSCs were totally flattened and elongated on all dECM surfaces (Fig. 7c). GSCs were extensively spread and formed a monolayer of the cells that covered the entire surfaces on day 7 (Fig. 7c). These results suggested that both dECM_DPSCs and dECM_GSCs supported cell culture and growth in vitro, and OM-dECM_DPSCs exhibited a superior proliferative effect than those derived from GSCs.

Open in a separate window

Figure 7

Biological response of reseeded GSCs. (a) Cell proliferation was determined using MTT assay. (b) Cellular attachment and spreading were investigated by phalloidin immunofluorescence staining. The nuclei were counterstained using DAPI. (c) Cell morphology was observed with SEM. Scale bars: 10 and 20 µm.

Flow cytometry analysis

Surface protein expression was analyzed using flow cytometry. Single-cell suspensions were stained with fluorescence conjugated antibodies (1:50 dilution) as follows: FITC conjugated anti-human CD44 (Cat. No. 555478, BD Bioscience, USA), PE-conjugated anti-human CD105 (Cat. No. 21271054, Immuno Tools, Germany), FITC-conjugated anti-human CD90 (Cat. No. ab124527, Abcam, USA), and PerCP-conjugated anti-CD45 (Cat. No. 21810455, Abcam, USA). Mean fluorescence intensity was calculated using a FACS^Calibur flow cytometer (BD Bioscience, San Jose, CA, USA).

Osteogenic differentiation

Cells (50,000 cells/well in a 24-well plate) were cultured in an osteogenic medium consisting of growth medium supplemented with 50 µg/mL ascorbic acid (cat. no. A-4034, Sigma-Aldrich, USA), 250 nM dexamethasone (cat. no. D8893, Sigma-Aldrich, USA), and 5 mM β-glycerophosphate (cat. no. G9422, Sigma-Aldrich, USA). Osteogenic differentiation potential was elucidated using ALP, ARS, and Von Kossa staining.

For ALP staining, the cells were washed with phosphate buffer saline (PBS) and fixed with 4% formaldehyde for 10 min. Then, cells were incubated with BCIP/NBT (Roche, USA) in the dark at room temperature for 30 min. The ALP-positive cells were observed using an inverted microscope (Olympus, USA).

For the ARS staining, the cells were fixed with cold methanol for 10 min and washed with deionized water. The samples were then stained with 2% ARS solution (Sigma-Aldrich Chemical) for 3 min at room temperature with gentle agitation. The mineral deposits were solubilized with 10% cetylpyridinium chloride monohydrate in 10 mM sodium phosphate. The optical density was measured at 570 nm with a microplate reader (ELx800, BIO-TEK®, United States).

For Von Kossa staining, the cells were fixed with 4% formaldehyde in PBS and further incubated with 5% silver nitrate in sterile deionized water. The samples were exposed to ultraviolet light for 5 min at room temperature. The stained cells were examined under an inverted microscope.

Adipogenic differentiation

Cells (12,500 cells/well in a 24-well plate) were cultured in adipogenic medium comprising growth medium containing 0.1 mg/ml insulin (cat. no. 11070738 Sigma-Aldrich, USA), 1 μM dexamethasone (cat. no. D8893, Sigma-Aldrich, USA), 1 mM IBMX (cat. no. PHZ1124, Thermo Fisher Scientific, USA), and 0.2 mM indomethacin (cat. no. 53861, Sigma-Aldrich, USA) for 16 days. The intracellular lipid droplet was stained by Oil red O staining. Briefly, cells were fixed with 4% formaldehyde in PBS for 10 min, followed by incubating with 0.2% Oil Red O solution for 15 min. Lipid accumulation was examined using an inverted microscope.

Extracellular matrix production and decellularization

The culture plate was coated with 0.2% gelatin for 2 h at 37 °C. The cells were seeded on a gelatin-coated surface and divided into two groups: N-ECM and OM-ECM. In N-ECM, cells were maintained in a growth medium for 7 days and subsequently cultured in a growth medium supplemented with 50 μg/ml l-ascorbic acid for 14 days. For OM-ECM, cells were maintained in an osteogenic medium for 21 days.

Decellularization was performed using 0.5% Triton X-100 in 20 mM ammonium hydroxide and washed with a protease inhibitor in PBS. Deoxyribonuclease A at a concentration of 0.0025% in sterile PBS was added to the samples and incubated for 30 min at room temperature for DNA removal.

Protein extraction and digestion

Protein extraction was performed from dECM on day 21 using a Compartment Protein Extraction Kit (MERCKMillipore, USA). The protein pellets' solubilization and digestion were performed as previously described⁴³. In brief, dECM pellets were solubilized in a solution containing 8 M urea, 100 mM ammonium bicarbonate, and 10 mM dithiothreitol. Cysteines were alkylated by adding iodoacetamide, and samples were deglycosylated with PNGaseF (New England BioLabs, USA, 1:100 units for 1 mg sample) and subsequently digested with trypsin/LysC (Promega, USA), at a ratio of 1:10,000 enzyme: substrate. Final digestions were done using trypsin (Worthington Biochemical Corporation, USA) at a ratio of 1:1000 (enzyme: substrate), followed by a second aliquot of trypsine/LysC (Promega, USA), at a ratio of 1:10,000 (enzyme:substrate).

Mass spectrometry

Mass Spectrometry (LC–MS/MS) was performed as previously described¹⁰. In brief, chromatography was performed with an RSLCnano system (Ultimate 3000, Thermo Scientific) coupled online to a Q Exactive HF-X with a Nanospay Flex ion source (Thermo Scientific).

Peptides were trapped on a C18 column (75 μm inner diameter×2 cm: nanoViper Acclaim PepMapTM 100, Thermo Scientific) at a flow rate of 2.5 μl/min over 4 min and subsequently separated on a 50 cm×75 μm C18 column (nanoViper Acclaim PepMapTM RSLC, 2 μm, 100 Å, Thermo Scientific) at 50 °C at a flow rate of 300 nl/min over 211 min. MS full scans were performed in the ultra-high-field Orbitrap mass analyzer. Top 20 intense ions were further fragmented via high-energy collision dissociation activation. Ions with a charge range from 2+to 6+were selected for screening. Data were searched against the Homo sapiens (UP000005640) SwissProt database using Sequest HT through proteome discoverer (version 2.2). The data were subsequently processed using myProMS v3.9.3 (https://github.com/bioinfo-pf-curie/myproms) FDR calculation used Percolator. The label-free quantification was performed by peptide Extracted Ion Chromatograms (XICs) computed with MassChroQ version 2.2. To correct the XICs, median and scale normalization was applied on the total signal. For statistical analysis, a linear model was performed, and p-values were adjusted using Benjamini–Hochberg FDR procedure. Matrisome proteins database (Human Matrisome (Updated December 2022): http://matrisomeproject.mit.edu/other-resources/human-matrisome/) has been updated and used for selecting the ECM proteins out of the whole proteomics data. The mass spectrometry proteomics raw data have been deposited to the ProteomeXchange Consortium via the PRIDE (PMID: 34,723,319) partner repository with the data identifier “PXD040575” and “PXD018951” (reviewer_pxd040575@ebi.ac.uk & DYdXRjUC).

Matrisome protein–protein interaction and enrichment pathway analysis

The significant ECM proteins detected from matrisome database were analyzed using Metascape (https://metascape.org/gp/index.html#/main/step1). PPI enrichment was determined using minimum network size=3 and maximum network size=500. GO enrichment analysis, categorized as a biological process, cellular component, and molecular function, was performed. The PPI figures were created by using Cytoscape version 3.9.1.

ECM seeding experiment

GSCs were seeded at a density of 25,000 cells on dECM_DPSCs or dECM_GSCs and cultured in a growth medium for 30 min, 24 h, or 7 days. Evaluation of cell morphology, attachment, and spreading was done using SEM. Cell viability was assessed using MTT assay. For mineralization assay, GSCs were reseeded and maintained in osteogenic induction medium for 14 days.

Immunofluorescence staining

Samples were fixed with 4% formaldehyde in PBS and incubated with 0.1% Triton-X100 in PBS. Non-specific binding was blocked with horse serum (2% v/v). Samples were stained with mouse monoclonal IgG anti-type I collagen (1:200 dilution, Abcam, UK) or mouse monoclonal IgG anti-fibronectin (1:500 dilution, Invitrogen, United States) at 4 °C overnight. The secondary antibody labeled with AlexaFluor 488 was added at a 1:2000 dilution for 2 h. F-actin organization was examined using AlexaFluor 594 Phalloidin (1:1000 dilution, Invitrogen, United States). DAPI (1:500 dilution, Invitrogen, United States) was used to counterstain the nuclei. Visualization of the target protein was detected using a fluorescent microscope with an ApoTome system (Carl Zeiss, Germany).

Scanning electron microscopy

The samples were fixed with 3% glutaraldehyde in PBS for 30 min and dehydrated with a graded series of ethanol. Hexamethyldisiloxane was added for 5 min and the gold sputter-coat was performed for SEM analysis.

Cell viability test

GSCs (12,500 cells/well) were reseeded on dECM. At day 1, 3, and 7, the cells were incubated with 0.5 mg/mL MTT solution (USB Corporation) for 30 min, allowing formazan crystal formation. The precipitated crystals were solubilized using a dimethyl sulfoxide and glycine buffer. The solution was measured absorbance at 570 nm by a microplate reader (ELx800, BIO-TEK®, United States). The percentage cell number was calculated and normalized with the control.

Quantitative real-time polymerase chain reaction

Total cellular RNA was extracted using TRIzol reagent (RiboEx solution, cat. no. 301-001, GeneAll, South Korea). The cDNA was obtained by converting one microgram of total RNA using ImProm-II Reverse Transcription System (cat. no. A3800, Promega, USA). qPCR was performed using FastStart Essential DNA Green Master (Roche Diagnostic, Germany) in a CFX connect Real-Time PCR machine (Bio-Rad, Singapore). Product specificity was evaluated using melt curve analysis. The targeted mRNA expression levels were normalized to GAPDH gene. The relative expression was calculated using 2^−ΔΔCt method⁴⁴. The primer oligonucleotide sequences are shown in Supplementary Table ².

High-throughput RNA sequencing

Total RNA was extracted using a RNeasy kit (Qiagen, USA). The RNA quality was examined using an Agilent 2100 BioAnalyzer (Agilent Technologies, USA), NanoDrop (Thermo Fisher Scientific Inc.), and 1% agarose gel. Library preparation was constructed using a NEBNext® UltraTM RNA Library Prep Kit for Illumina®. The constructed library was validated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), and quantified by Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA, USA). Sequencing was performed on the illumine HiSeq platform in a 2X150bp paired-end configuration. Base-calling is performed by Illumina RTA software. Demultiplexing is performed by Illumina bcl2fastq 2.17 software based on index information and the number of reads and quality score (Q30) were counted. Data were aligned to reference genome via software HISAT2 (v2.0.1)^45,46. Differential expression analysis used the DESeq2 Bioconductor package. The sequencing data were submitted to the NCBI’s Gene Expression Omnibus (GSE226347).

This study is funded by the Thailand Science Research and Innovation Fund Chulalongkorn University (HEA663200065 to T.O.) and by a grant from the “Fondation des Gueules Cassées” and INSERM/APHP Interface contract (to BPJ. F.).