Cartilaginous Metabolomics Reveals the Biochemical-Niche Fate Control of Bone Marrow-Derived Stem Cells

Joint disorders have become a global health issue with the growth of the aging population. Screening small active molecules targeting chondrogenic differentiation of bone marrow-derived stem cells (BMSCs) is of urgency. In this study, microfracture was employed to create a regenerative niche in rabbits (n = 9). Cartilage samples were collected four weeks post-surgery. Microfracture-caused morphological (n = 3) and metabolic (n = 6) changes were detected. Non-targeted metabolomic analysis revealed that there were 96 differentially expressed metabolites (DEMs) enriched in 70 pathways involved in anti-inflammation, lipid metabolism, signaling transduction, etc. Among the metabolites, docosapentaenoic acid 22n-3 (DPA) and ursodeoxycholic acid (UDCA) functionally facilitated cartilage defect healing, i.e., increasing the vitality and adaptation of the BMSCs, chondrogenic differentiation, and chondrocyte functionality. Our findings firstly reveal the differences in metabolomic activities between the normal and regenerated cartilages and provide a list of endogenous biomolecules potentially involved in the biochemical-niche fate control for chondrogenic differentiation of BMSCs. Ultimately, the biomolecules may serve as anti-aging supplements for chondrocyte renewal or as drug candidates for cartilage regenerative medicine.


Introduction
Age-related changes in the chondrocyte renewal in the articular cartilage predispose individuals to osteoarthritis [1,2]. Regenerative medicine aims to restore tissue functionality by harnessing the differentiation potential of stem cells in tissue replacement therapies [3]. Microfracture (MF) is a commonly used surgical method to stimulate cartilage regeneration, involving drilling on the subchondral bone to access the resident bone marrow-derived stem cells (BMSCs) [4] and creating a biochemical niche to induce chondrogenesis and sustain chondrocyte functionality [5]. However, the outcome of MF is correlated with age; i.e., greater improvement is only shown in young patients [6]. This may be due to BMSC aging, failure in chondrogenic differentiation, and dysfunction of the resident chondrocytes in old people [7,8]. Thus, the fate of the activated BMSCs plays a determinative role in the modeling and remodeling of cartilage, which are affected by both intrinsic and extrinsic factors.
The biochemical niche which is created by a variety of small molecules (metabolites) filling in the defect site in MF controls the fate of the activated resident BMSCs [9,10]. For example, omega-3 polyunsaturated fatty acids (ω-3 PUFAs), including docosapentaenoic acid 22n-3 (DPA) [11], maintain the self-renewal of embryonic stem cells [12]. The amino acids, such as tryptophan, modulate the senescence of mesenchymal stromal/stem cells (MSCs) by regulating mitochondrial integrity and function [13]. Yet, the excess reactive oxygen species (ROS), oxidative markers, impair stem cell self-renewal capacity [1,14,15]; subsequentially, the ROS-induced DNA damage causes replicative senescence in MSCs [16]. The excess ROS and related oxidative stress are correlated with the progression of osteoarthritis, characterized by the changes in the hyaline cartilage markers, i.e., an increased catabolism of proteoglycan aggrecan (ACAN) and a loss of type II collagen (COL2A1) [17]. Moreover, dietary intake of ω-3 PUFAs attenuates osteoarthritis-associated cartilage degradation [18]. Hence, we hypothesize that the metabolic cues in the regenerative niche contribute to BMSC activity and rebalance a hostile joint environment, which is of paramount importance for the fate control of BMSCs in chondrogenesis.
The aim of this study was to discover biomolecules with the potential to serve as antiaging supplements for chondrocyte renewal or as drug candidates for cartilage regenerative medicine. In this study, MF was employed in healthy rabbits to create a regenerative niche for BMSCs, which maximizes the detection of the key components in cartilage regeneration and excludes the background noise in deteriorative joints. Non-targeted metabolomics was performed to determine the metabolic differences between the normal (NOR) and regenerated (REG) cartilages. A list of cartilage regeneration-related endogenous small molecules, i.e., the differentially expressed metabolites (DEMs), were identified. Two of the metabolites, DPA and ursodeoxycholic acid (UDCA), were picked to determine their specific effects on cartilage regeneration, i.e., BMSC vitality, chondrogenic differentiation, and chondrocyte functionality, based on the chondrocyte-protective role of ω-3 PUFAs [19] and the chondrocyte-differentiation-facilitating role of cholesterol (the substrate of UDCA) [20].

Rabbit Microfracture Model
The animal study was approved by the Ethics Committee of Experimental Animals of the Affiliated Hospital of Qingdao University (No. AHQU-MAL20210419). A selfcontrolled design was used in this study. Male New Zealand white rabbits (10-12 weeks in age, 2.0-2.5 kg in weight, n = 9) were anesthetized through ear vein administration of 3% w/v pentobarbital sodium at a dosage of 1.0 mg/kg. After sterilization, a 2-3 cm anteromedial parapatellar incision was made on the left knee, and the patella was everted. An articular cartilage defect (Ø~4 × 6 mm, 2 mm in depth) was created on the trochlear groove of the left distal femur with a sterilized cranial drill bit (ØO = 2.1 mm, RWD, Shenzhen, China) ( Figure 1A). Afterward, the debris was removed and hemostasis was achieved. All rabbits were treated with gentamycin for three days after the surgery.

Tissue Collection
The rabbits were sacrificed by CO 2 asphyxiation, and the paired cartilage samples from both the left and right distal femurs (n = 9) were collected based on the reported protocol [19]. For histochemical analysis (n = 3), three pairs of trochlear grooves were kept in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 24 h and then decalcified in EDTA (14% w/v, pH 7.4) for 21 days at 4 • C until measurement. For metabolomic analysis (n = 6), six pairs of cartilage tissues were washed with PBS and then stored at −80 • C until analysis. . The microfracture (MF) surgery is conducted on the trochlear groove of the left distal femur. An articular cartilage defect (Ø ~ 4 × 6 mm, 2 mm in depth) has been created on the trochlear groove of the left distal femur with a sterilized cranial drill bit (ØO = 2.1 mm). (B). The photographs of the trochlear grooves. Compared to the normal (NOR) cartilage, the regenerated (REG) cartilage loses its transparent color and smooth and glistening appearance, replaced with a rough surface with fissures (C,D) Examples of the H&E staining. Compared to the NOR cartilage, the REG cartilage loses the highly organized structure composed of four zones, i.e., the superficial, middle, deep, and calcified zones. (E,F) Example of the Safranin-O/Fast Green staining. Compared to the NOR cartilage, the reddish-stained ACAN content in the REG cartilage was lower. The scale bars represent 40 μm and 10 μm.

Tissue Collection
The rabbits were sacrificed by CO2 asphyxiation, and the paired cartilage samples from both the left and right distal femurs (n = 9) were collected based on the reported protocol [19]. For histochemical analysis (n = 3), three pairs of trochlear grooves were kept in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 24 h and then decalcified in EDTA (14% w/v, pH 7.4) for 21 days at 4 °C until measurement. For metabolomic analysis (n = 6), six pairs of cartilage tissues were washed with PBS and then stored at −80 °C until analysis.

Histochemistry
The decalcified tissues were embedded in paraffin, and then sagittal sections were cut at 3 μm. Hematoxylin-eosin (H&E) staining was conducted to determine the cellular regularity of cartilage using a commercial kit (G1120, Solarbio, Beijing, China) according to the published procedure [20]. A commercial kit (G1371, Solarbio) was used for the Safranin-O/Fast Green staining, in which the depth of the reddish color is correlated with the content of ACAN [21], by following the company's instructions. All the slices were dehydrated, transparentized, and then mounted with neutral resin for microscopic observation (Leica, Herlev, Denmark). Images were representative results of three biological repeats. Compared to the normal (NOR) cartilage, the regenerated (REG) cartilage loses its transparent color and smooth and glistening appearance, replaced with a rough surface with fissures. (C,D) Examples of the H&E staining. Compared to the NOR cartilage, the REG cartilage loses the highly organized structure composed of four zones, i.e., the superficial, middle, deep, and calcified zones. (E,F) Example of the Safranin-O/Fast Green staining. Compared to the NOR cartilage, the reddish-stained ACAN content in the REG cartilage was lower. The scale bars represent 40 µm and 10 µm.

Histochemistry
The decalcified tissues were embedded in paraffin, and then sagittal sections were cut at 3 µm. Hematoxylin-eosin (H&E) staining was conducted to determine the cellular regularity of cartilage using a commercial kit (G1120, Solarbio, Beijing, China) according to the published procedure [20]. A commercial kit (G1371, Solarbio) was used for the Safranin-O/Fast Green staining, in which the depth of the reddish color is correlated with the content of ACAN [21], by following the company's instructions. All the slices were dehydrated, transparentized, and then mounted with neutral resin for microscopic observation (Leica, Herlev, Denmark). Images were representative results of three biological repeats.
Afterward, 150 µL of the liquid supernatant from each tube was collected, filtered by an organic phase pinhole filter (0.22 µm), and then stored at −80 • C until analysis.

Liquid Chromatography-Mass Spectrometry
Liquid chromatography-mass spectrometry (LC-MS) analysis was conducted by following a published protocol [22]. Briefly, the LC was performed using an ACQUITY UPLC HSS T3 1.8-micron column at 45 • C. The mobile phase contained water and acetonitrile with 0.1% formic acid; gradient elution was conducted at a flow rate of 0.35 mL/min. All samples were kept at 4 • C during the analysis. The injection volume was 2 µL. The MS was performed on an AB TripleTOF 6600 plus system, with electrospray ionization using both positive and negative ion modes. The full mass scan range was set at 100-1000 mass to charge ratio (m/z).

Bioinformatics Data Processing
The raw data obtained from LC-MS were analyzed using Progenesis QI v2.3 (Nonlinear Dynamics, Newcastle, UK) with main parameters of 5 ppm/10 ppm precursor tolerance, 10 ppm/20 ppm product tolerance, and 5% product ion threshold. The compounds were identified based on the RT-m/z pairs. The supervised orthogonal partial least squares discriminant analysis (OPLS-DA) analysis was employed to distinguish the metabolic profiles of the NOR and REG groups, and a volcano plot was employed to visualize the alterations of the metabolite concentrations.

Identification of the Differentially Expressed Metabolites
The metabolites were identified using Progenesis QI v2.

Pathway Enrichment Analysis
The pathway enrichment analysis for the DEMs was performed using the Kyoto Encyclopedia of Genes and Genomes database (KEGG, https://www.kegg.jp/kegg/pathway. html, assessed 22 March on 1995) through matching IDs (the KEGG ID of the corresponding DEM). A pathway with a p-value < 0.05 was considered a significant one. Passage (P) 2 of human BMSCs was purchased from Haixing Biosciences (BMHX-C106, Suzhou, China). The BMSCs were cultured in the Human BMSCs Growth Medium (HyCyte, BMHX-G101, Haixing Biosciences) supplemented with 1% penicillin/streptomycin, 1% glutamine, and 10% fetal bovine serum under the condition of 5% CO 2 in the air at 37 • C.

Chondrogenic Differentiation
The chondrogenic differentiation in monolayer culture was induced according to the previous protocol with minor modifications [23]. Briefly, P4 BMSCs were cultured using a Human BMSCs Chondrogenic Differentiation Kit (BMHX-D203R, Haixing Biosciences) for three weeks. The differentiated chondrocytes were maintained in the chondrocyte growth medium, i.e., the DMEM/F12 medium with 10% FBS and 1% penicillin/streptomycin.

Stimuli and Chemicals Administrated
The oxidative stress and DNA damage to BMSCs were created by employing hydrogen peroxide (H 2 O 2 , 3%, LIRCON, Dezhou, China) and mitomycin C (MC, GC12353, GLPBIO, Montclair, CA, USA), respectively [24,25]. The dosages of H 2 O 2 (250 µM) and MC (1 µM) were selected based on a previous study [24] and our pilot experiments (Supplementary Figure S1). The effects of UDCA (HY-13771, MCE, Monmouth Junction, NJ, USA) and DPA (GC31637, GLPBIO) on the vitality of BMSCs were examined under both control and challenge states. Various concentrations of the biomolecules (0, 5, 10, 50, and 100 µM) were added to the culture medium to determine the DEM's contribution to the vitality of BMSCs, while the biomolecules were added with the stimulus to determine the DEM's role under the H 2 O 2 and MC challenged conditions. Thereafter, an optional dosage, 50 µM for both UDCA and DPA, was used in this study for identifying the effects of small molecules on chondrogenic differentiation and chondrocyte functionality.

Cell Vitality
The cell vitality was detected using the Cell Counting Kit-8 (CCK-8, C6005, NCM Biotech, Suzhou, China) after 24 h of incubation with and without small molecules. The optical density was read at 450 nm using a multimode plate reader (PerkinElmer, Waltham, MA, USA).
The data of the CCK-8 assay were presented as mean ± standard error (SEM). Statistical analyses and graphics were conducted using GraphPad Prism version 7.0 (San Diego, CA, USA). The effects of the small molecules under control and challenge conditions were revealed by a one-way ANOVA analysis. The Dunnett test was used to partition differences among the dosages. A value of p < 0.05 was considered to be statistically significant.

Immunofluorescence
The cells were grown on glass coverslips (ØO = 14 mm). The immunofluorescence was determined according to the method previously described with some modifications [26]. Briefly, the cells were incubated with anti-ACAN (1:500, bs-1223R, Boiss, Beijing, China) overnight at 4 • C after fixation and permeabilization. The goat anti-rabbit 594 secondary antibody (1:500, bs-0295G-AF594, Boiss) was applied for 2 h at RT. The cells were counterstained with phalloidin (C1033, Beyotime, Shanghai, China) to determine the cytoskeletal arrangement [27], and the nuclei were identified after staining with DAPI (S2110, Solarbio) for 30 min at RT. The images were taken under a Leica DM 6000B microscope (Leica, Herlev, Denmark) and were representatives of three repeats.

Real-Time PCR
Total RNA was extracted using NcmZol Reagent (M5100, NCM Biotech). The concentration and quality of RNA were measured using a spectrophotometer (Nanodrop 2000c, Thermo Fisher, Waltham, MA, USA). Reverse transcription was conducted using the Reverse Transcription Reagent Pack (Applied Biosystems, Thermo Fisher). Real-time PCR analysis was carried out using Quantstudio 5 (Applied Biosystems) with the SYBR Green Master Mix (Applied Vazyme, Nanjing, China) and gene-specific primers (Table 1) of the key transcriptional factor for chondrogenic differentiation (SRY-related high mobility group-box gene 9, SOX9) and the chondrogenic indicators (ACAN and COL2A1) [21,28]. The Ct-value of β-actin mRNA was used as the internal control. The relative mRNA expression levels were analyzed using the 2−∆∆Ct method [29]. The unpaired Students' t-test was applied to determine the difference between DPA or UDCA treatment with the CON group. A value of p < 0.05 was considered to be statistically significant.

The Histomorphologic Changes in the Regenerated Cartilage
Compared to the NOR cartilage, the REG cartilage lost its transparent color and smooth and glistening appearance, replaced with a rough surface with fissures ( Figure 1B). The REG cartilage also lost the regularity seen in the H&E-stained NOR cartilage, i.e., a highly organized structure composed of four zones ( Figure 1C,D). Specifically, the thickness of the REG cartilage layer at the defect site was lower than that of the NOR group, while the chondrocytes in the REG group were more compactly organized than those in the NOR group ( Figure 1C-F). Furthermore, the shape of the chondrocytes in the NOR group was plumper and more round ( Figure 1C,E) than that of the chondrocytes in the REG group ( Figure 1D,F). Moreover, the reddish-stained ACAN content in the REG cartilage ( Figure 1E) was lower than that in the NOR cartilage ( Figure 1F).

The Metabolomic Changes in the Regenerated Cartilage
The OPLS-DA plot revealed the differences in metabolomic activities between the NOR and REG groups (Figure 2A). A total of 96 DEMs were identified (Table 2), with 31 decreased and 65 increased metabolites induced by MF ( Figure 2B). Of these, 11 DEMs were matched with the records in the Regeneration Roadmap and Age Atlas ( Figure 2C, Supplementary  Table S1). Moreover, the DEMs can be divided into six categories, including lipids and lipidlike molecules (47), organic oxygen compounds (16), organic acids and derivatives (12), heterocyclic compounds (15), organosulfur compounds (2), and hydrocarbons (4). Notably, the lipids and lipid-like molecules were accountable for half (48.96%) of the entire metabolic alterations in cartilage regeneration. Of the lipids and lipid-like molecules, the majority were increased in the REG cartilages (85.11%), including DPA ( Figure 2D, p = 0.0087) and UDCA ( Figure 2E, p = 0.005). The organic oxygen compounds, accounting for 16.67% of the DEMs, were mainly second messengers and oligosaccharides. Among the organic oxygen compounds, most of the glycolytic substances were decreased. The heterocyclic compounds, organosulfur compounds, and others accounted for the rest (21.88%) of the DEMs. and ursodeoxycholic acid (UDCA, (E)) between the NOR and REG groups are revealed. (F) The bubble diagram shows the top 20 differential metabolic pathways enriched by the DEMs. X-axis represents rich factors, and Y-axis represents pathway terms. The numbers of the involved metabolites and p-value are listed on the right side. The detailed information is presented in Supplementary Tables S1 and S2.
The enrichment analysis yielded 70 pathways related to cartilage regeneration (Supplementary Table S2), with the top 20 listed in the bubble diagram ( Figure 2F), which revealed the key information associated with the biochemical-niche control for chondrogenesis, such as the rapamycin (mTOR) signaling pathway. Moreover, the detailed information on the pathways with hits  2 and p-value < 0.05 is presented in Table 3. The enrichment analysis yielded 70 pathways related to cartilage regeneration (Supplementary Table S2), with the top 20 listed in the bubble diagram ( Figure 2F), which revealed the key information associated with the biochemical-niche control for chondrogenesis, such as the rapamycin (mTOR) signaling pathway. Moreover, the detailed information on the pathways with hits ≥ 2 and p-value < 0.05 is presented in Table 3.    The pathways listed in Table 3

Differentially Expressed Metabolites Involved in Promoting the Vitality and Adaptation of BMSCs
The facilitating effect of both DPA (p < 0.0001, Figure 3A) and UDCA (p = 0.004, Figure 3B) on the vitality of BMSCs was revealed. Specifically, 50 and 100 µM of both DPA and UDCA significantly increased the optical densities compared to the control group (0 µM). Under oxidative challenge, adding DPA and UDCA rescued the damage effect of 250 µM H 2 O 2 (p < 0.0001, Figure 3C; p = 0.002, Figure 3D). Similarly, the administration of DPA and UDCA rescued the injury effect of DNA damage caused by 1 µM MC (p < 0.0001, Figure 3E; p < 0.0001, Figure 3F).

Differentially Expressed Metabolites Involved in Promoting Chondrogenic Differentiation
After induced chondrogenic differentiation ( Figure 4A), compared to the CON group ( Figure 4B), the intensity of ACAN was increased in both the DPA ( Figure 4C) and UDCA ( Figure 4D) groups. The phalloidin staining was of high intensity, with the microfilaments easily distinguishable in the CON group ( Figure 4E), while the staining in the DPA ( Figure  4F) and UDCA ( Figure 4G) groups was almost faded; in particular, the microfilaments in the UDCA group mostly disappeared ( Figure 4G). Moreover, the ACAN signal was colocated with the nucleus in the CON group ( Figure 4B,H,K) and exhibited cytoplasmic shuttling in the DPA ( Figure 4C,I,L) and UDCA ( Figure 4D,J,M) groups. Furthermore, the quantitative analysis indicated that the administration of DPA during chondrogenic differentiation significantly elevated the mRNA expression of SOX9 ( Figure 4N), ACAN ( Figure 4O), and COL2A1 ( Figure 4P) compared to the CON group, while UDCA significantly increased the SOX9 ( Figure 4N) and ACAN ( Figure 4O) mRNA expression levels.

Differentially Expressed Metabolites Involved in Promoting Chondrogenic Differentiation
After induced chondrogenic differentiation ( Figure 4A), compared to the CON group ( Figure 4B), the intensity of ACAN was increased in both the DPA ( Figure 4C) and UDCA ( Figure 4D) groups. The phalloidin staining was of high intensity, with the microfilaments easily distinguishable in the CON group ( Figure 4E), while the staining in the DPA ( Figure 4F) and UDCA ( Figure 4G) groups was almost faded; in particular, the microfilaments in the UDCA group mostly disappeared ( Figure 4G). Moreover, the ACAN signal was co-located with the nucleus in the CON group ( Figure 4B,H,K) and exhibited cytoplasmic shuttling in the DPA ( Figure 4C,I,L) and UDCA ( Figure 4D,J,M) groups. Furthermore, the quantitative analysis indicated that the administration of DPA during chondrogenic differentiation significantly elevated the mRNA expression of SOX9 ( Figure 4N), ACAN ( Figure 4O), and COL2A1 ( Figure 4P) compared to the CON group, while UDCA significantly increased the SOX9 ( Figure 4N) and ACAN ( Figure 4O) mRNA expression levels. , and type II collagen (COL2A1, (P)) in the CON, DPA, and UDCA groups are quantified. The data are presented as mean  SEM (n = 3). ns indicates no significance (p > 0.05); * p < 0.05, ** p < 0.01, *** p < 0.001.

Differentially Expressed Metabolites Involved in Promoting Chondrocyte Functionality
The ACAN signal was detected in the nucleus and cytoplasm in the CON group (Figure 5B,H,K), while it was restricted in the cytoplasm and exhibited exocytosis (arrows) in both the DPA ( Figure 5C,I,L) and UDCA ( Figure 5D,J,M) groups. In addition, the chondrocytes in the CON group ( Figure 5E) tended to have a fibroblast-like appearance, while those in the DPA ( Figure 5F) and UDCA ( Figure 5G) groups were more likely to exhibit a polygon shape. Furthermore, the quantitative analysis indicated that the ACAN mRNA expression in the chondrocytes was increased in both DPA and UDCA groups ( Figure  5N), while the mRNA expression levels of COL2A1 ( Figure 5O) were increased in the UDCA group only. , and type II collagen (COL2A1, (P)) in the CON, DPA, and UDCA groups are quantified. The data are presented as mean ± SEM (n = 3). ns indicates no significance (p > 0.05); * p < 0.05, ** p < 0.01, *** p < 0.001.

Differentially Expressed Metabolites Involved in Promoting Chondrocyte Functionality
The ACAN signal was detected in the nucleus and cytoplasm in the CON group ( Figure 5B,H,K), while it was restricted in the cytoplasm and exhibited exocytosis (arrows) in both the DPA ( Figure 5C,I,L) and UDCA ( Figure 5D,J,M) groups. In addition, the chondrocytes in the CON group ( Figure 5E) tended to have a fibroblast-like appearance, while those in the DPA ( Figure 5F) and UDCA ( Figure 5G) groups were more likely to exhibit a polygon shape. Furthermore, the quantitative analysis indicated that the ACAN mRNA expression in the chondrocytes was increased in both DPA and UDCA groups ( Figure 5N), while the mRNA expression levels of COL2A1 ( Figure 5O) were increased in the UDCA group only.

Discussion
Currently, the common osteoarthritis therapy with nonsteroidal anti-inflammatory drugs or chondroprotective hyaluronic acid (HA) is ineffective for cartilage regeneration [30]. The pharmaceutical industry is calling for novel biomolecules covering both safety and effectiveness as anti-aging supplements or drug candidates for joint health. The metabolomics modules in the Aging Atlas and Regeneration Roadmap offer the molecular substances potentially related to the degenerative and regenerative events in the blastema of axolotls, deer antler stem cells, etc. However, orthopedic metabolic cues are lacking. In this study, the minority of identified DEMs (11/96) were matched with the records in the Regeneration Roadmap and Aging Atlas, which stresses the uniqueness and significance of the cartilage regeneration-specific metabolites. This study firstly reveals the biomolecules for cartilage regeneration, which serve each step during MF, i.e., the vitality and adaptation of BMSCs as well as chondrogenic differentiation and chondrocyte functionality.
The vitality of the activated BMSCs may be the main restriction for cartilage regeneration, especially in old people with an inadequate BMSC pool and senescent BMSCs [5]. The vitality of BMSCs is altered by the biophysical perturbation during their migration, e.g., switched from the hypoxic bone marrow niche to the oxygen-exposed defect site in MF [31]. Oxidative metabolism, e.g., the oxidative stress and the subsequent DNA damage in the defect site, has been revealed by the increase in FAPγ-adenine in the heterocyclic compounds in the REG group, exhibiting the damage effects on the vitality of BMSCs [32]. Adding DPA or UDCA increased the vitality of the BMSCs under both the control

Discussion
Currently, the common osteoarthritis therapy with nonsteroidal anti-inflammatory drugs or chondroprotective hyaluronic acid (HA) is ineffective for cartilage regeneration [30]. The pharmaceutical industry is calling for novel biomolecules covering both safety and effectiveness as anti-aging supplements or drug candidates for joint health. The metabolomics modules in the Aging Atlas and Regeneration Roadmap offer the molecular substances potentially related to the degenerative and regenerative events in the blastema of axolotls, deer antler stem cells, etc. However, orthopedic metabolic cues are lacking. In this study, the minority of identified DEMs (11/96) were matched with the records in the Regeneration Roadmap and Aging Atlas, which stresses the uniqueness and significance of the cartilage regeneration-specific metabolites. This study firstly reveals the biomolecules for cartilage regeneration, which serve each step during MF, i.e., the vitality and adaptation of BMSCs as well as chondrogenic differentiation and chondrocyte functionality.
The vitality of the activated BMSCs may be the main restriction for cartilage regeneration, especially in old people with an inadequate BMSC pool and senescent BMSCs [5]. The vitality of BMSCs is altered by the biophysical perturbation during their migration, e.g., switched from the hypoxic bone marrow niche to the oxygen-exposed defect site in MF [31]. Oxidative metabolism, e.g., the oxidative stress and the subsequent DNA damage in the defect site, has been revealed by the increase in FAPγ-adenine in the heterocyclic compounds in the REG group, exhibiting the damage effects on the vitality of BMSCs [32].
Adding DPA or UDCA increased the vitality of the BMSCs under both the control condition and oxidative challenge. Moreover, the fluctuation in the cellular metabolism directly alters the microenvironment for cartilage regeneration in situ [33]. In this study, increases in phenylalanine, tyrosine, tryptophan, 2'-O-methyladenosine (an analog of adenosine), sphingosine, and SM(d18:1/24:1(15Z)) have been identified in the REG group. Acetyl-CoA is obtained from amino acids such as phenylalanine, tyrosine, and tryptophan [34], and oxidation of acetyl-CoA accounts for ATP production [34]. It has been reported that adenosine elevates the ATP supply in MSCs [35,36]. Moreover, the sphingolipid metabolism and sphingolipid signaling pathways are associated with stem cell migration and signaling transduction [37,38]. Hence, the DEMs may regulate the vitality of the activated BMSCs via mediating acetyl-CoA synthesis, ATP supply, and signaling transduction.
The foreign supply in the defect site is critical for the renewal and functionality of chondrocytes because cartilage is an avascular tissue [39]. The elevated second messengers in the REG group may reveal the boosting of signaling transduction in the differentiating and newborn chondrocytes. The activated mTOR signaling pathway is critical for lipid biosynthesis [40] and chondrogenic differentiation [41]. Blocking mTOR signaling inhibits chondrogenic differentiation [41]. Hence, the DEMs may participate in regulating chondrogenic differentiation and chondrocyte functionality, and they may potentially prevent cartilage loss in musculoskeletal diseases via mediating mTOR signaling transduction. In addition, both DPA and UDCA promoted chondrogenic differentiation by stimulating the mRNA expression of SOX9 and ACAN and facilitating cytoskeleton rearrangement during chondrogenic differentiation via regulating F-actin depolymerization. The administration of DPA and UDCA enhanced chondrocyte functionality by promoting the extracellular processing of ACAN. The HA-bound extracellular ACAN regulates HA endocytosis [42], which may determine the chondroprotective effect of HA. Compared to DPA, UDCA exhibited higher potential in facilitating chondrogenic differentiation and chondrocyte functionality potentially due to cholesterol's facilitating role in chondrocyte differentiation [43]. The inconsistency between DPA and UDCA may be associated with the dosage effect or functioning pathway, and the related mechanism will be explored in further study.

Conclusions
Generally, the findings of this study (1) provided a list of biomolecules potentially involved in the fate control of the activated resident BMSCs in cartilage regeneration and (2) verified the functions of UDCA and DPA in promoting BMSC vitality, chondrogenic differentiation, and chondrocyte functionality. Ultimately, the small molecules may work as anti-aging supplements for joint rejuvenation. Clinical supply with the biomolecules in MF may confer a better outcome and promote the surgery to a broad base of patients, especially the elderly, by rebalancing and conquering the hostile joint environment. Additionally, combining the biomolecules with BMSC therapy may be a safe and effective strategy in stem cell-treated diseases, not limited to orthopedic disorders.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cells11192951/s1, Figure S1: The dosages of the stimuli. Various concentrations of H 2 O 2 (A) and MC (B) were added in the culture medium and co-incubated with the BMSCs for 24 h, and the cell viability was decided by CCK8 assay. The data of cell viability assay were presented as MEAN ± standard error (SEM). A one-way ANOVA analysis was employed to determine the damage effect of H 2 O 2 and MC. The Dunnett test was used to partition differences of each stimulus-treated group with the control (0 µM); Table S1: The DEMs matched with the records in Regeneration Roadmap and Aging Atlas; Table S2: The list of the potential pathways induced by MF.

Conflicts of Interest:
The authors declare no conflict of interest.