Betulinic acid decreases lipid accumulation in adipogenesis-induced human mesenchymal stem cells with upregulation of PGC-1α and UCP-1 and post-transcriptional downregulation of adiponectin and leptin secretion

Cell culture

hMSCs were obtained from the Japanese Collection of Research Bioresources Cell Bank (JCRB Cell Bank, UE7T-13, JCRB1154, Lot #07072007). This hMSC cell line had been established for its multipotent differentiation, including osteogenesis and adipogenesis (Tsujimura, Kusamori & Nishikawa, 2019; Zhang et al., 2019). For cell proliferation, hMSCs were cultured in complete growth media (CGM: α-Minimum Essential Medium with 10% fetal bovine serum, 100 units/ml Penicillin-Streptomycin, and two mM L-glutamine) at 37 °C and 5% CO₂. The medium was changed every 3 days, and the cells were sub-cultured after about 80% confluency. For cell differentiation, hMSCs were seeded in multi-well plates. When 80% confluency was reached (within 2–3 days), the CGM was replaced with osteogenic medium (OM) or adipogenic medium (AM). BA was purchased from Sigma-Aldrich (cat#B8936) and initially dissolved in 50 mM dimethyl sulfoxide (DMSO). The final concentrations of DMSO were kept below 0.05% in all conditions.

Cell differentiation media

In this study, an OM recipe without dexamethasone was used to highlight the effect of BA. The OM is a CGM-based medium supplemented with 50 µM ascorbic acid and five mM β-glycerophosphate (Choi et al., 2016).

AM is also CGM-based with one µM dexamethasone, 10 µg/ml h-insulin, 100 µM indomethacin, and 500 µM isobutymethyxanthine (Nöth et al., 2002).

Metabolic activity of hMSCs

alamarBlue™ dye (Invitrogen) is a resazurin-based assay that can be used for continuous monitoring of mitochondrial function, and it was performed for the metabolic activity analysis of hMSCs as per manufacturer’s instructions. For metabolic tracking, the cells were seeded at 5 × 10³ cells/well in 96-well plates and treated with BA at 5, 10, 15, and 20 µM for 7 and 10 days in CGM, OM, or AM. At the designated time points, the medium was replaced with fresh CGM with 10% alamarBlue and then incubated for 2 h before absorbance reading at 570 and 600 nm. The percentage of metabolic activity was compared with the control group without BA at the selected time points for each media type (Maxson & Burg, 2008; Quent et al., 2010).

Evaluations of osteogenesis

The effects of BA on hMSC osteogenesis were evaluated in two levels, namely, alkaline phosphatase (ALP) enzymatic activity and bone matrix calcium deposition.

ALP is a critical enzyme in bone mineralization and known as an early marker of osteogenesis (Mohamadnia et al., 2007). For ALP activity assay, hMSCs were seeded at 5 × 10³ cells/well in 96-well plates and cultured in OM for 7 days. The ALP activity assay was performed as previously described (Montazerolghaem et al., 2016). Briefly, the cells were lysed in a lysis buffer, followed by three freeze-thawing cycles to fully break the cells. A total of one mg/ml p-nitrophenyl phosphate substrate in diethanolamine was added to the cell lysis at the ratio of 2:1. The reactions were incubated at 37 °C, and the absorbance was read at 405 nm every 5 min for 30 min. The total protein was measured with bicinchoninic acid assay (Walker, 1996). The enzymatic activity was calculated as described in Enzymatic Assay of Alkaline Phosphatase, Diethanolamine Assay (EC 3. 1. 3. 1):

$$\mathrm{U}\mathrm{n}\mathrm{i}\mathrm{t}/\mathrm{m}\mathrm{g}\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}={\displaystyle \frac{\mathrm{U}\mathrm{n}\mathrm{i}\mathrm{t}/\mathrm{m}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{z}\mathrm{y}\mathrm{m}\mathrm{e}}{\mathrm{m}\mathrm{g}\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}/\mathrm{m}\mathrm{l}\mathrm{e}\mathrm{n}\mathrm{z}\mathrm{y}\mathrm{m}\mathrm{e}}}$$

Calcium deposition was quantified with calcium colorimetric assay with o-cresolphthalein, which was adapted from the work of Tu et al. (2014). Cells were seeded at 2 × 10⁴ cells/well in 24-well plates and cultured in OM with 5, 10, 15, and 20 µM BA or control without BA in OM and CGM. After 14 days, the cells were washed twice with phosphate-buffered saline (PBS). Then, the accreted calcium was dissolved in 0.1 M HCl, shaken at 4 °C for 8 h, and centrifuged at 500×g for 2 min. The supernatant with dissolved calcium was then used for calcium quantification. A chromogenic reagent 190 µl (1.1 mM O-cresolphthalein complexone and 7.9 mM 8-hydroxyquinoline) was added to the 10 µl sample, and the absorbance was read at 570 nm.

To visualize the calcium nodule, we washed the cells twice with PBS, followed by 4% paraformaldehyde fixation for 10 min and staining with 2% alizarin red.

Evaluations of adipogenesis

The adipogenesis of hMSCs was analyzed with glycerol 3-phosphate dehydrogenase (GPDH) activity and intracellular lipid droplet accumulation.

Cell culture and cell lysate preparation for GPDH activity assay were performed with the same protocol as the ALP activity, but with the AM used for cell culture instead of an OM. Exactly 90 µl 334 µM b-nicotinamide adenine dinucleotide, which is the reduced form of NADH, was added to 100 µl cell lysate and then further incubated at 37 °C for 10 min before adding 10 µl four mM dihydroxyacetone phosphate. The absorbance was read at 340 nm every 5 min for 30 min. GPDH was calculated and normalized with the total protein (Sottile & Seuwen, 2001).

Oil Red O and Nile red staining were used to analyze the intracellular lipid accumulation. Oil Red O is a common oil droplet stain that can also be used to interpret the total triglyceride content by dissolving the dye for absorbance reading at 540 nm (Kanawa et al., 2013). hMSCs were cultured in an AM in 96-well plates for 21 days. Then, the cells were washed twice with PBS, followed by 4% paraformaldehyde fixation for 10 min. The oil droplets were stained with Oil Red O (1.8 mg/ml in 60% isopropanol) for 10 min. The excess dye was then removed by washing with distilled water. The stained lipid was then dissolved in 100 µl isopropanol and read at 540 nm.

For the analysis of lipid droplet size and number, the cells were cultured in 48-well plates containing the AM for 21 days and then stained with 10 µg/ml Nile red (1% acetone in non-phenol red Dulbecco’s Modified Eagle Medium) (Hogan et al., 2016). Images were captured with a fluorescence microscope (Olympus, CKX53) with the excitation wavelength at 495 nm, with five images obtained per well in triplicate and analyzed with ImageJ using function analyze the particles (Deutsch et al., 2014). Briefly, the images were binarized to black and white pixels. Then, watershedding techniques were applied to measure individual lipid droplets by size and perimeter.

Glucose uptake

The glucose uptake by hMSCs during adipogenesis was analyzed on days 10 and 14. Cells were seeded in 48-well plates at 1 × 10⁴ cell/well. After 10 or 14 days, the medium was replaced with a fresh medium, and the glucose level was measured with Accu-chek guide (Kim et al., 2018) after 4 and 8 h.

Cellular mechanisms

mRNA gene expression analysis was performed with real-time reverse-transcription polymerase chain reaction (qRT-PCR). hMSCs were seeded in 24-well plates at 2 × 10⁴ cell/well. The total RNA was extracted with Total RNA Mini Kit (Favorgen, FATRK001) and converted to cDNA with HyperScript™ RT master mix with oligo dT (GeneAll, 601–703). The quantitative PCR was performed with the Applied Biosystems™ 7,500 Real-Time PCR System using iTaq Universal SYBR Green Supermix (Biorad, 172–5,122). The PCR products were checked with melting temperature analysis. The ΔCT was calculated using GAPDH. The relative gene expressions were calculated using the 2^−ΔΔCT method with the cut-off at twofold change and P < 0.05 (Rao et al., 2013). Table 1 shows the primer sequences.

The release of adipokines, adiponectin, IL-6, leptin, and TNF-α was measured with Milliplex Map Human Adipokine Magnetic Bead Panel Kit. hMSCs were seeded at 10⁴ cells/well in 48-well plates. The culture medium was changed by half every 3 days and collected for cell-secreted adipokines on days 10, 14, and 17. The detection and analysis were performed in accordance with the manufacturer’s instructions. At least 50 beads were counted for each replicate, and the results were calculated as pg/ml by comparison with the manufacturer’s standard control (Chen et al., 2014).

Statistical analysis

All statistical analysis was performed with 27^th edition of SPSS. Data are shown as mean ± standard deviation from three biological replicates. The Shapiro–Wilk test was applied to assess normality of the distributions. In cases of a normal distribution, one-way ANOVA with Tukey multiple comparison for equal variances or Dunnett’s C multiple comparison for unequal variances were used. In cases of a non-normal distribution, the Kruskal–Wallis test with Mann Whitney U test was used to determine the level of significance between the groups. P values < 0.05 were accepted as significant.

BA increased the metabolic activity of hMSCs during adipogenesis

To track the changes in the metabolic activity of hMSCs in response to BA, we performed alamarBlue assays on days 7 and 10, proliferating phase, osteogenic differentiation, or adipogenic differentiation. The results showed that BA caused no alteration in the metabolic activity of hMSCs during the proliferating phase or osteogenic differentiation. The BA at 15, and 20 µM increased the metabolic activity of hMSCs during adipogenesis by 22% and 29%, respectively, on day 10 (Figs. 1A and 1B) without changes in total protein content (Fig. S1).

BA increased osteogenic differentiation and decreased adipogenic differentiation of hMSCs, changing calcium deposition, lipid accumulation, enzyme activities, and gene expressions

The effects of BA on hMSC osteogenesis or adipogenesis were investigated with enzyme activity assays, calcium deposition, or lipid accumulation. The results showed that ALP, the key enzyme in bone mineralization, significantly increased in hMSCs with a dose-dependent response to BA at 5–15 µM (Fig. 2A). At 15 µM, the ALP activity increased the most by 85%. Calcium accretion was also significantly increased by BA at 5, 10, and 15 µM as shown in the calcium quantification assay, with the average increase of about 4–5 µg/well or 15%–18% (Fig. 2B). With alizarin red staining, thicker and larger calcium nodules were observed in BA treatments at 5, 10, and 15 µM (Fig. 2C).

For hMSC adipogenesis, GPDH, the key enzyme in lipid biosynthesis, and its intracytoplasmical accumulation lipid were investigated. Our preliminary result showed elevated activity of GPDH in hMSCs by adipogenic induction at day 7 but not day 14 (Table S1). Day 7 was then selected as the timepoint for GPDH activity analysis. The results showed that BA at 5–20 µM significantly decreased the GPDH activity by 26%–34% (Fig. 2D). hMSC adipogenic differentiation was confirmed with Oil Red O staining (Fig. S2). At day 21, Oil Red O staining showed evident reduction in lipid accumulation with the decrease in total lipid content by 28%–45% (Figs. 2E and 2F).

BA can alter the hMSC lineage differentiations by enhancing bone formation with the increase in ALP enzymatic activity and bone matrix deposition and reducing fat cell formation with the demolishment of lipid droplet accumulation and GPDH enzyme activity.

Further, hMSC gene expression was analyzed with qRT-PCR. The effects of BA without additional differentiation supplements in the CGM were monitored on day 3. We observed that the effects of BA on hMSC early osteogenic gene expressions, namely ALP, β-catenin, and Wnt family member 3A (WNT3A) were not evident without osteogenic induction factors (Fig. 2G). However, the expression of adipogenic marker CCAAT-enhancer-binding proteins alpha (C/EBP-α) level was significantly reduced by BA alone (Fig. 2G). In addition, the remarkable downregulation of C/EBP-α, but not PPARγ-2, by BA was observed in hMSCs during adipogenic induction (Fig. 2H).

With osteogenic induction factors, BA significantly upregulated the expressions of late osteogenic markers, ALP, osteocalcin (OC), osteopontin (OPN), and RUNX2 on day 14 (Fig. 2I). However, the expression of early osteogenic marker genes, i.e., ALP, β-catenin, bone morphogenetic protein 2 (BMP2), RUNX2, and WNT3A, was not altered (Fig. 2I).

BA down-regulated C/EBP-α, upregulated IL-6 mRNA expression, but post- transcriptionally down-regulated adiponectin and leptin secretion in hMSCs

The decreased adipogenesis of hMSCs by BA with the downregulation of C/EBP-α expressions was remarkable at all time points on days 3, 7, and 14 (Figs. 2H and 3A), whereas the expressions of PPARγ-2 and fatty acid-binding protein 4 (FABP4) were not significantly altered (Fig. 3A).

Further, adipokine mRNA expressions were analyzed with qRT-PCR. The result showed that BA at 15 µM significantly increased the expression of IL-6 and leptin on day 7 (Fig. 3B). Meanwhile, the mRNA expressions of adiponectin and leptin-receptor (LEP-R), were not altered by BA on day 7 (Fig. 3B). On day 14, both IL-6 and LEP-R expressions were significantly increased by BA at 10 and 15 µM, while the expressions of adiponectin and leptin were not altered (Fig. 3B).

Further, the secreted adipokines were analyzed with Milliplex Map Human Adipokine Magnetic Bead Panel Kits. The pilot results comparing the adipokines secreted by hMSCs during adipogenic and proliferating phases showed that the secretions of adiponectin and leptin were induced tremendously after 14 days in adipogenic induction media by about 1,400-and 11-fold, respectively (Table 2). On the contrary, the IL-6 secretion of hMSCs extensively decreased with adipogenic induction by more than 100-fold. TNF-α secretion was minimal during hMSC proliferation in the lower range of less than two pg/ml and diminished during adipogenesis (Table 2). We then analyzed the effect of BA during adipogenesis on hMSC-secreted adiponectin, IL-6, and leptin. The secreted adiponectin and leptin significantly decreased on days 14 and 17 with BA treatment (Fig. 3C). The secreted IL-6 was not altered by the addition of BA during adipogenesis (Fig. 3C).

BA potentially shifted adipogenic lineage differentiation of hMSCs toward brown adipocytes with increased glucose uptake and enhancement of PPARγ coactivator one alpha (PGC-1α) and uncoupling protein 1 (UCP-1) expressions

Recent studies showed that hMSCs can give rise to the distinctive lineages of white or brown adipocytes (Park, Kim & Bae, 2014). To determine the effect of BA on brown adipocyte formation, we analyzed lipid droplet formation and glucose uptake along with the mRNA expression of brown adipocyte markers.

The number of lipid droplets stained by Nile red were aligned with the results from Oil Red O staining. BA significantly reduced the amount of lipid droplets (Figs. 4A and 4C). Moreover, lipid droplet size distribution analysis revealed that BA decreased the ratio of large lipid droplets (>3 µm²) and increased the ratio of small-sized lipid droplets (<1 µm²) (Fig. 4B). A small lipid droplet size is one of the characters that can be used to distinguish brown adipocytes from white adipocytes (Barneda et al., 2013).

Glucose uptake by hMSCs was analyzed on days 10 and 14. The results showed that BA did not alter glucose uptake during adipogenesis of hMSCs on day 10 but significantly increased the glucose uptake on day 14 (Fig. 4E). Glucose transporter 4 (GLUT4) mRNA expression was quantified on days 7 and 14. The results showed that GLUT4 expression was significantly increased by adipogenic induction on day 7 and decreased with BA treatment, while there was no change in GLUT4 expression on day 14 (Fig. 4D).

The expressions of five brown adipocyte markers, namely, bone morphogenetic protein 7 (BMP7), transcriptional co-regulator PR domain-containing 16 (PRDM16), myogenic factor 5 (MYF5), PGC-1α, and UCP-1 were analyzed. The basal expression of maker genes during adipogenesis on day 7 showed that this hMSC cell line, UE7T-13, did not express BMP7 nor PRDM16 with a cycle threshold (Ct) over 35 (data not shown). Meanwhile, MYF5, PGC-1α, and UCP-1 were normally expressed in hMSCs, BA enhanced the expressions of PGC-1α and UCP-1 on day 7 by up to 6.42 and 3.22-fold, respectively (Fig. 5). On day 14, the expression of PGC-1α was not affected by BA, and the expression of UCP-1 was decreased by BA at 10 µM.