Acetylated-PPARγ expression is regulated by different P53 genotypes associated with the adipogenic differentiation of polyploid giant cancer cells with daughter cells

Objective: Polyploid giant cancer cells (PGCCs) with daughter cells express epithelial–mesenchymal transition (EMT)-associated proteins. Highly malignant tumor cells with EMT properties can transdifferentiate into mature tumor cells. In this study, we elucidated the potential for, and underlying mechanism of, adipogenic differentiation of PGCCs with daughter cells (PDCs). Methods: Cobalt chloride was used to induce PGCC formation in HEY (wild-type P53) and MDA-MB-231 (mutant P53) cells; these cells were then cultured in adipogenic differentiation medium. Oil red O staining was used to confirm adipogenic differentiation, and the cell cycle was detected with flow cytometry. The expression of adipogenic differentiation-associated proteins and P300 histone acetyltransferase activity were compared before and after adipogenic differentiation. Animal xenograft models were used to confirm the adipogenic differentiation of PDCs. Results: PDCs transdifferentiated into functional adipocytes. Two different cell cycle distributions were observed in PDCs after adipogenic differentiation. The expression levels of PPARγ, Ace-PPARγ, and Ace-P53 were higher in PDCs after adipogenic differentiation than in cells before adipogenic differentiation. Ace-PPARγ and FABP4 expression increased in HEY cells and decreased in MDA-MB-231 PDCs after p53 knockdown. A485 treatment increased Ace-P53, Ace-PPARγ, and FABP4 expression in HEY PDCs by inhibiting SUMOylation of P53. In MDA-MB-231 PDCs, A485 treatment decreased Ace-P53, Ace-PPARγ, and FABP4 expression. Animal experiments also confirmed the adipogenic differentiation of PDCs. Conclusions: Acetylation of P53 and PPARγ plays an important role in the adipogenic differentiation of PDCs.


Introduction
The adipogenic differentiation of human mesenchymal stem cells is typically induced with a cocktail of 3-isobutyl-1-methylxanthine (IBMX), dexamethasone, and insulin 1 . Adipocytes may be derived from not only preadipocytes and pluripotent mesenchymal stem cells, but also cancer stem cells (CSCs). One treatment for malignant tumors is induced differentiation therapy, which involves treatment with chemicals that promote the differentiation of malignant cells into normal cells. Well-differentiated and dedifferentiated liposarcoma cells can also be differentiated into adipocytes with dexamethasone, indomethacin, insulin, and IBMX. These compounds induce adipogenesis by upregulating the transcription and translation of genes involved in maintaining cancer cell stemness and adipogenic differentiation 2 . Ishay-Ronen et al. 3 have reported that combinatorial treatment with MEK inhibitors and the antidiabetic drug rosiglitazone induces the conversion of invasive and metastatic breast cancer cells into adipocytes, thereby repressing primary tumor cell invasion and metastasis in breast cancer.
We have reported that cobalt chloride (CoCl 2 ), radiation, and chemotherapy drugs induce the formation of polyploid giant cancer cells (PGCCs), and daughter cells derived from PGCCs via asymmetric division (budding and bursting) have strong invasion and infiltration abilities [4][5][6] . PGCCs with daughter cells (PDCs) have CSC properties and express multiple normal and CSC markers, including CD44, CD133, Nanog, and SOX-2 7,8 . Additionally, PDCs express epithelial-mesenchymal transition (EMT)-associated proteins, including high expression of mesenchymal markers and low expression of epithelial markers 9,10 . EMT is a de-differentiation process that is known to enhance cellular plasticity, and can be exploited therapeutically through transdifferentiation into post-mitotic and functional cells 3 . PDCs are necessary for cancer dissemination, but can be directly targeted and inhibited through a transdifferentiation approach. We previously reported that PDCs can be induced to differentiate into multiple benign lineages, such as adipocytes, bone, and cartilage 6,11,12 . However, the underlying molecular mechanism remains unclear. In this study, CoCl 2 -treated HEY and MDA-MB-231 PDCs were cultured in adipogenic differentiation medium and induced to differentiate into adipocytes. We assessed the critical transcription factors and post-translational modifications (PTMs) in adipogenic differentiation in HEY and MDA-MB-231 PDCs, including PPARγ, which has recently been demonstrated to play an important role in adipogenic differentiation 13 . In addition, recent studies have reported that fibroblast-derived cancer cells with p53 gene deletion can be induced to differentiate into adipocytes 14 . After adipogenic differentiation, Ace-P53 expression levels are regulated by different P53 genotypes. Fatty acid binding protein 4 (FABP4) is a member of the FABP family that is abundantly expressed in adipocytes. Expression of FABP4, a marker of successful adipogenic differentiation of PDCs, is associated with the acetylation of PPARγ and different P53 genotypes. Understanding the complex regulatory mechanisms in malignant tumors with EMT properties that can transdifferentiate into mature tumor cells may lead to the development of new therapies for solid tumors.

Materials and methods
Cell culture HEY and MDA-MB-231 cell lines were obtained from the American Type Culture Collection (USA), HEY cells were cultured in 1640 medium (Gibco, Thermo Fisher Scientific, Suzhou, China). MDA-MB-231 cells were cultured in Dulbecco's modified Eagle's medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, Life Technologies, New Zealand) and 1% penicillin/ streptomycin (Gibco, Life Technologies, USA). Medium supplemented with serum and antibiotics was defined as complete medium. Cells cultured in complete medium were maintained in a humidified atmosphere at 37 °C and 5% CO 2 .

Induction of PGCCs
Our previous research has described the method of CoCl 2 induction of PGCC formation 6 . Briefly, HEY and MDA-MB-231 cells were cultured in T25 cell flasks containing complete medium until they reached 80%-90% confluency. Cells were treated with 450 μM CoCl 2 (Sigma-Aldrich) for 48-72 h depending on their resistance to hypoxia. Most regular-sized cells died after CoCl 2 treatment, and only several scattered PGCCs survived. Surviving PGCCs began to produce daughter cells via asymmetric division after 10-15 days. CoCl 2 treatment was repeated 3 times to obtain sufficient PDCs for subsequent experiments.

Adipogenic differentiation of cancer cells
Control cells (2.5 × 10 5 ) and PDCs were seeded into 6-well plates and cultured in complete medium until they reached 70%-80% confluence. The medium was replaced with differentiation medium A for 24 h (StemPro Adipogenic Differentiation Kit; Cyagen Biosciences, Suzhou, China, containing 10% FBS, 0.5 mM IBMX, 4 μM insulin, 1 μM dexamethasone, and 10 mM rosiglitazone), then was replaced with differentiation medium B and incubated another 48 h (α-MEM containing 10% FBS and 4 μM insulin). The cells were repeatedly cultured in media A and B and harvested for subsequent experiments. The workflow of PGCC induction and adipogenic differentiation in HEY and MDA-MB-231 control cells and PDCs is shown in Figure 1.

Oil red O (ORO) staining
ORO staining was performed according to the manufacturer's instructions. HEY and MDA-MB-231 control cells and PDCs before and after adipogenic differentiation were fixed with 4% paraformaldehyde for 24 h at 25 °C and were then incubated with fresh ORO (Solarbio, Beijing, China) for 15 min. ORO-stained intracellular lipids were quantified on the basis of optical density at 510 nm with a microplate reader (BioTek, USA).

Western blot (WB)
Proteins were fully denatured by boiling at 100 °C for 10 min. After the concentrations were determined, the proteins were separated on a 10% sodium dodecyl sulfate-polyacrylamide gel and transferred to a polyvinylidene fluoride membrane (Beyotime, Haimen, China). The membrane was blocked with 5% skim milk at room temperature for 2 h, then incubated with primary antibodies. Detailed antibody information is provided in Supplementary Table S1. All WB assays were performed in triplicate.

Flow cytometry
To examine the DNA content and cell cycle status, we harvested control cells and PDCs before and after adipogenic differentiation. The cell pellets were fixed with 75% ethanol at 20 °C for 12 h and then permeabilized with 0.1% Triton X-100 at 26 °C for 20 min. After 30 min of RNase treatment at 37 °C, the cells were incubated with propidium iodide (50 μg/mL) at 25 °C for 15 min. The cell cycle distribution was analyzed with flow cytometry (BD FACSCalibur™, BD Biosciences). The proportions of cells in G1, S, and G2 phases were quantified with BD CellQuestä version 5.1 (BD Biosciences).

Quantitative real-time PCR (qPCR)
Total RNA was extracted with the TRIzol/chloroform method. RNA was reverse-transcribed with a high-capacity RNA to cDNA kit (TIAGEN, KR116, Beijing, China). qPCR was performed on a Roche LightCycler 480 Real-Time PCR System (Roche). The qPCR reaction system included 2 μL of cDNA diluted in nuclease-free water, 50 ng of total RNA, 25 μL of Universal PCR Master Mix (CWBIO, 0957), and 1 μL of 10 μM forward primer. Nuclease-free water was used to dilute the reverse primer to a final volume of 50 μL, and β-actin was used as the reference gene for the quality and quantity of cDNA.

A485 inhibitor treatment
A485 is a potent and selective catalytic inhibitor of P300/CBP 15 . Control cells and PDCs before and after adipogenic differentiation were cultured in 6-well plates until they reached 80% confluence. Each well was treated with 5 μM A485 (Selleck, USA) for 72 h.

Cell counting kit-8 (CCK8) assays
Cell viability was evaluated with CCK8 assays. Control cells and PDCs before and after adipogenic differentiation were seeded at a density of 2,000 cells per well in 96-well plates (3 replicate wells per group) and incubated for 12 h. Wells containing medium alone were used as controls. The cells were treated with different concentrations of A485 and cultured for various time intervals. After treatment, 10 μL CCK8 (Dojindo, Japan) was added to each well and incubated for 2 h. The optical density was measured at 450 nm with a Bio-Rad microplate reader, and the final value was calculated with the average value of the readings after subtraction of the average value of the control.

Immunocytochemical (ICC) and immunohistochemical (IHC) staining
For ICC, cells grown on coverslips were fixed with cold methanol for 30 min. After treatment with 0.3% endogenous peroxidase inhibitor (Zhongshan Inc., Beijing, China), cells were incubated with goat serum (Zhongshan Inc.) to block non-specific protein binding. The cells on coverslips were then incubated with primary antibodies, biotinylated goat anti-mouse/rabbit IgG (Zhongshan Inc.), and horseradish peroxidase-labeled streptomycin (Zhongshan Inc.). Paraffin-embedded tissue sections were deparaffinized in xylene for IHC. Antigen retrieval was performed by incubation of sections with citrate buffer (OriGene, Wuxi, China). The sections were then incubated with primary antibodies and biotinylated goat anti-rabbit IgG. The signal was detected with a labeled streptavidin-biotin system in the presence of the chromogen 3,3-diaminobenzidine or alkaline phosphatase.

Wound-healing assays
Control cells and PDCs before and after adipogenic differentiation were seeded into 12-well plates (1 × 10 5 cells per well, 3 replicate wells per group) and cultured until they reached 100% confluence. Sterile pipette tips were then used to uniformly scratch the monolayer of cells vertically to form wound tracks. After being rinsed with PBS to remove floating cells, the cells were cultured in serum-free medium. Cell migration was evaluated by imaging of the wound area at 0, 12, and, 24 h for HEY cells and at 0, 16, and, 32 h for MDA-MB-231 cells at the same scratch position. The migration area was outlined in ImageJ, and the wound-healing index was calculated with the following formula: [(wound area at 0 h) − (wound area at indicated time)]/(wound area at 0 h). A high score indicated strong migration ability.

Cell migration and invasion assays
Control cells and PDCs before and after adipogenic differentiation were washed 3 times with serum-free medium and counted with an automated cell counter (Invitrogen). Cell migration and invasion were assessed with Transwell migration and invasion assays (8 μm; Corning Inc.), respectively. For the migration assays, 5 × 10 4 cells per insert, resuspended in 200 μL serum-free medium, were seeded in the upper chamber. For the invasion assay, 2 × 10 5 cells per insert, resuspended in 200 μL serum-free medium, were seeded in the upper chamber coated with Matrigel. Medium containing 20% FBS was added to the lower chamber, and the 24-well plates were incubated for an additional 12-24 h at 37 °C. The cells were then fixed in methanol for 30 min and stained with 0.1% crystal violet for 30 min. The migration and invasion abilities were assessed by counting the number of cells per field. Images were acquired at 100× magnification, and cells in at least 5 different fields were counted. Three independent experiments were performed.

Clone formation assays
Control cells and PDCs were counted before and after adipogenic differentiation, respectively. Cell suspensions (2 mL/ well) containing 30, 60, and 120 cells were cultured in 12-well plates, and the plates were incubated for at least 2 weeks at 37 °C (when white cell clones were visible). Cell clones were washed with PBS and fixed with cold methanol for 30 min.
After staining with 0.1% crystal violet for 30 min, the number of cell clone groups per well was counted under a microscope (the number of cells in a single clone was >50), and the efficiency of colony formation was calculated with the following formula: formation efficiency = (number of clones/number of cells inoculated).

H&E staining
Tumor tissues were fixed in formalin for 24 h at room temperature and embedded in paraffin, and 4 μm-thick sections were prepared. The tissue sections were subsequently deparaffinized in xylene for 12 h at 75 °C and rehydrated with a descending ethanol series. Sections were stained with 0.2% hematoxylin (Baso, Guangzhou, China) at room temperature for 1 min and 0.5% eosin for 2 min. After staining, the sections were dehydrated and mounted on coverslips.

Statistical analysis
Statistical analyses and graphs were generated in SPSS 22 software (SPSS Inc., Chicago, USA) and GraphPad Prism software. Statistical analyses were performed with ANOVA and unpaired t-test. The number of animals used in each experiment is indicated in the figure legends. The in vivo statistical analysis was performed with the Kruskal-Wallis test and corrected for multiple comparisons. The threshold for statistical significance was set at P < 0.05. P-values represent comparisons with each control (***P < 0.001, **P < 0.005, and *P < 0.05).  for 7 days. Comparison of the cells cultured in adipogenic differentiation medium indicated that PDCs had more vacuoles in their cytoplasm (Figure 2Ba and 2Ca) than the control cells (Figure 2Bc and 2Cc). ORO staining revealed the presence of more red lipid droplets in the PDCs (Figure 2Da, 2Db and 2Ea, 2Eb) than the controls (Figure 2Dc, 2Dd and 2Ec, 2Ed). ORO staining also revealed greater intracellular lipid accumulation in PDCs (Supplementary Figure S1Aa and S1Ac) than the controls (Supplementary Figure S1Ab and S1Ad) when the cells were cultured in adipogenic differentiation medium.

Adipocyte differentiation decreases the invasion, metastasis and proliferation of control cells and PDCs
Cell invasion assays were performed with Matrigel-coated Transwell inserts. The invasive (Figure 2F and 2G) and migratory abilities (Figure 2H and 2I) of cells cultured in adipogenic differentiation medium were lower than those of cells cultured in complete medium. PDCs cultured in adipogenic differentiation medium had the lowest number of invasive and migratory cells among all cell groups (Supplementary Figure  S1B and S1C). Migration ability was also measured with wound-healing assays in HEY and MDA-MB-231 control cells and PDCs before and after adipogenic differentiation. The migration ability of PDCs was higher than that of control cells. After adipogenic differentiation, migration ability decreased in both control cells and PDCs ( Figure 2J). Significant differences were observed in wound-healing indices with or without adipogenic differentiation in the HEY control cells and PDCs at 12 and 24 h, and in the MDA-MB-231 control cells and PDCs at 16 and 32 h (Supplementary Figure S1D). To detect cell proliferative ability, we performed plate cloning assays. The number of clones formed in the HEY and MDA-MB-231 PDCs after adipogenic differentiation was significantly lower than that in the corresponding cells without adipogenic differentiation (Supplementary Figure S1E and S1F).

Adipogenic differentiation-associated protein expression in adipogenic differentiation of PDCs
In HEY and MDA-MB-231 PDCs, the expression of PPARγ and FABP4 gradually increased with increasing adipogenic differentiation time, whereas that of phosphor-PPARγ (Ser112) gradually decreased (Figure 3A and 3C). No clear change was observed in the expression of PPARγ and FABP4 in HEY and MDA-MB-231 control cells with increasing adipogenic differentiation time (Figure 3B and 3D). Statistical analysis indicated that intracellular lipid accumulation in PDCs increased with increasing culture time in adipogenic differentiation medium (Figure 3Ea and 3Ec); however, analysis of intracellular lipid accumulation did not reveal any significant change in the control cells (Figure 3Eb and 3Ed). The results of qPCR indicated that the mRNA expression levels of PPARγ (for HEY: control vs. control after adipogenic differentiation, P = 0.0116; PDCs vs. PDCs after adipogenic differentiation, P = 0.0485; for MDA-MB-231: control vs. control after adipogenic differentiation, P = 0.0028; PDCs vs. PDCs after adipogenic differentiation, P = 0.0008) and FABP4 (for HEY: control vs. control after adipogenic differentiation, P = 0.013; PDCs vs. PDCs after adipogenic differentiation, P = 0.0055; for MDA-MB-231: control vs. control after adipogenic differentiation, P = 0.0385; PDCs vs. PDCs after adipogenic differentiation, P = 0.0309) were significantly higher in HEY and MDA-MB-231 control cells and PDCs with adipogenic differentiation than in those without adipogenic differentiation (Figure 3F).  (Figure 3Ga and 3Ia) before adipogenic differentiation. No discernible difference was observed in the cell cycle between HEY (Figure 3Ha and  3Hb) and MDA-MB-231 (Figure 3Ja and 3Jb) control cells before and after adipogenic differentiation. In the first distribution, 74.45% and 58.7% of the HEY and MDA-MB-231 PDCs, respectively, were arrested in G1 phase after adipogenic differentiation, and 49.92% and 51.81% of the HEY and MDA-MB-231 PDCs, respectively, were arrested in G1 phase without adipogenic differentiation; for the second distribution, 94.33% and 87.92% of the HEY and MDA-MB-231 PDCs, respectively, were arrested in G1 phase after adipogenic differentiation (Figure 4Aa and 4Ac). The proportions of cells in S and G0/G1 phases in the HEY and MDA-MB-231 control cells increased after adipogenic differentiation (HEY control cells: 67.68%; MDA-MB-231 control cells: 59.54%; HEY control cells: 45.1%; MDA-MB-231 control cells: 50.6%) (Figure  4Ab and 4Ad). Additionally, compared with those without adipogenic differentiation, the expression levels of cyclin B1, CDK1, and cyclin D1 decreased in HEY and MDA-MB-231 PDCs after adipogenic differentiation (Figure 4B). The differences in cyclin B1, CDK1, and cyclin D1 expression in the cells before and after adipogenic differentiation were statistically significant (*P < 0.05, **P < 0.01) (Figure 4C).

Acetylation modification of PPARγ in HEY and MDA-MB-231 after adipogenic differentiation
As described above, the expression of total PPARγ gradually increased with increasing adipogenic differentiation PPARγ β-actin I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G PDCs vs. PDCs after adipogenic differentiation, P = 0.000; for MDA-MB-231: control vs. control after adipogenic differentiation, P = 0.0000; PDCs vs. PDCs after adipogenic differentiation P = 0.0000) (Figure 5B).

PPARγ interacts with P53, and P53 regulates total PPARγ and Ace-PPARγ expression
To investigate the relationship between P53 and PPARγ after adipogenic differentiation, we performed co-immunoprecipitation and WB, which confirmed that PPARγ interacted with P53 in HEY and MDA-MB-231 PDCs after adipogenic differentiation (Figure 5C-5F). P53 was knocked down in HEY and MDA-MB-231 control cells and PDCs, which were then cultured in adipogenic differentiation medium. Compared with that in cells without P53 knockdown, the expression of total PPARγ after P53 knockdown was lower in MDA-MB-231 PDCs with adipogenic differentiation (Figure 5G). After P53 knockdown, Ace-PPARγ expression increased in HEY PDCs and decreased in MDA-MB-231 PDCs undergoing adipogenic differentiation (Figure 5H and Supplementary Figure  S2A). After P53 knockdown, FABP4 expression was inhibited in MDA-MB-231 PDCs bearing mutant P53, but was promoted in HEY PGCCs and daughter cells bearing wild-type P53 ( Figure 5I); the difference in FABP4 expression was statistically significant (for HEY PDCs after adipogenic differentiation: PDCs vs. PDCs after P53 knockdown, P < 0.001; for MDA-MB-231 PDCs after adipogenic differentiation: PDCs vs. PDCs after P53 knockdown, P < 0.001) (Figure 5J). The trend in FABP4 expression was consistent with that of Ace-PPARγ during the adipogenic differentiation of PDCs. I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G I n p u t P P A R γ I g G Continued competitively performed at the same lysine site 16 . Through immunoprecipitation with anti-P53 and immunoblotting with anti-P300, the interaction between P53 and P300 was detected in the total lysates of HEY and MDA-MB-231 PDCs before and after adipogenic differentiation (Supplementary Figure S2B). P53 interacted with P300 in HEY and MDA-MB-231 PDCs after adipogenic differentiation. WB indicated that the expression of P300 increased in HEY PDCs and decreased in MDA-MB-231 PDCs cultured in adipogenic differentiation medium ( Figure 6A). The P300 expression level further increased in HEY PDCs and decreased in MDA-MB-231 PDCs when P53 was knocked down (Figure 6B).

SUMOylated P53 regulates adipogenic differentiation of HEY and MDA-MB-231 PDCs
A485 is a potent and selective catalytic inhibitor of P300/ CBP 15 , and the acetylation sites in P53 overlap with the ubiquitination or SUMOylation sites 17 . Through IP with anti-P53 and immunoblotting with anti-SUMOylation, we confirmed that P53 is modified by SUMOylation in HEY and PDCs undergoing adipogenic differentiation. SUMOylated P53 was not detected in the MDA-MB-231 PDCs (Figure 6C and 6D).
The effect of A485 on the viability of HEY and MDA-MB-231 PDCs was analyzed with CCK8 assays. On the basis of these results, 72 h pretreatment with 5 μM A485 was used to inhibit

PPARγ
In p u t P P A R γ Ig G In p u t In p u t P P A R γ P P A R γ Ig G Ig G s i P A P C P A s i P C s i P A P C P A s i P C s i P A P C P A s i P C s i P A P C P A s i P C In p u t P P A R γ Ig the HAT activity of P300 (Supplementary Figure S2C). After A485 treatment, the expression of Ace-P53, Ace-PPARγ, and FABP4 increased, whereas that of SUMOylated P53 decreased in HEY PDCs (Figure 6E). The differences in Ace-P53, Ace-PPARγ, and FABP4 were statistically significant (for Ace-PPARγ expression in HEY: PDCs vs. PDCs after A485 treatment, P = 0.0073; PDCs vs. PDCs after adipogenic differentiation and A485 treatment, P = 0.0040; for Ace-p53 expression in HEY: PDCs vs. PDCs after A485 treatment, P < 0.0001; PDCs vs. PDCs after adipogenic differentiation and A485 treatment, P < 0.0001; FABP4 expression in HEY: PDCs vs. PDCs after A485 treatment, P < 0.0001; PDCs vs. PDCs after adipogenic differentiation and A485 treatment, P < 0.001) (Figure 6Ga,  6Gc and 6Ha), whereas the expression of Ace-P53, Ace-PPARγ, and FABP4 decreased in MDA-MB-231 PDCs (Figure 6F). The differences in Ace-P53, Ace-PPARγ, and FABP4 were statistically significant (for Ace-PPARγ expression in MDA-MB-231: PDCs vs. PDCs after A485 treatment, P = 0.0025; PDCs vs. PDCs after adipogenic differentiation and A485 treatment P = 0.0083; for Ace-p53 expression in HEY: PDCs vs. PDCs after A485 treatment, P < 0.0001; PDCs vs. PDCs after adipogenic differentiation and A485 treatment P < 0.0001; FABP4 expression in HEY: PDCs vs. PDCs after A485 treatment, P = 0.004; PDCs vs. PDCs after adipogenic differentiation and A485 treatment P < 0.0001) (Figure 6Gb, 6Gd and  6Hb). SUMOylated P53 was not expressed in MDA-MB-231 PDCs. The P300 histone acetyltransferase activity assays also indicated that A485 treatment inhibited the activity of P300 and increased the expression of Ace-P53 by inhibiting the SUMOylation of P53 in HEY PDCs bearing wild-type P53, and decreased the expression of Ace-P53 in MDA-MB-231 PDCs bearing mutant P53 (Supplementary Figure S2D). to IHC and WB. Tumor growth curves showed that the average volumes of xenograft tumors inoculated with cells with adipogenic differentiation were significantly smaller than those of tumors inoculated with cells without adipogenic differentiation (Figure 7C and 7D). H&E staining was performed to observe morphological characteristics, and IHC staining was used to analyze the expression of human-specific vimentin and Ki-67. As compared with the xenografts from the CC group (Figure 7Ed and 7Hd), CA group (Figure 7Ec and 7Hc), and PC group (Figure 7Eb and 7Hb), more adipocytes were  (Figure 7Ea and 7Ha). Adipocytes in the PA group were positive for human-specific vimentin, thus confirming the human origin of the adipocytes (Figure 7Fa and 7Ia). Tumor cells in the CC (Figure 7Fd and  7Id), CA (Figure 7Fc and 7Ic), and PC (Figure 7Fb and 7Ib) groups were positive for vimentin, on the basis of IHC staining. In addition, the proportion of Ki-67-positive staining was lower in the PA group (Figure 7Ga and 7Ja) than in the CC (Figure 7Gd and 7Jd), CA (Figure 7Gc and 7Jc) and PC groups (Figure 7Gb and 7Jb). WB indicated that the expression levels of PPARγ and FABP4 in xenograft tumor tissues from HEY and MDA-MB-231 control cells and PDCs after adipogenic differentiation were higher than those in cells before adipogenic differentiation (Figure 7K), and the differences were statistically significant (for HEY: control vs. control after adipogenic differentiation, P < 0.0001; PDCs vs. PDCs after adipogenic differentiation P < 0.0001; for MDA-MB-231: control vs. control after adipogenic differentiation, P = 0.0028; PDCs vs. PDCs after adipogenic differentiation P = 0.0064) (Figure 7L).

Discussion
Differentiation therapy has long been recognized in the treatment of malignant tumors, and cancer cells with high plasticity can differentiate into more mature tumor cells; therefore, this therapy may have potential for highly malignant tumors. NCI-H446 cells can be induced to differentiate into neurons, adipocytes, and bone cells in vitro 18 . Cancer cells with homologous recombination defects, such as ovarian and breast cancer cells with BRCA1/2 mutations, can be induced to differentiate with poly ADP-ribose polymerase inhibitors 19 . Thyroid cancer cells expressing the PPARγ fusion protein can be induced to differentiate into adipocyte-like cells with pioglitazone 20 . Gupta et al. 21 have identified that the potassium ionophore salinomycin induces epithelial differentiation of tumor cells and inhibits tumor growth in human breast cancer. Cancer cell plasticity facilitates the development of therapeutic resistance and the progression of malignancy. EMT enhances cellular plasticity and can be exploited therapeutically by forcing the transdifferentiation of EMT-derived cancer cells into functional cells 3 . Daughter cells derived from PGCCs undergo EMT and exhibit strong plasticity 22,23 . In this study, we confirmed that HEY and MDA-MB-231 PDCs undergo EMT, thus obtaining a mesenchymal cell phenotype and stem cell characteristics, and can differentiate into adipocytes. The migration, invasion, and proliferation abilities of PDCs decreased, and 2 cell cycle rounds were observed in HEY and MDA-MB-231 PDCs after adipogenic differentiation. The transformation between cell proliferation and adipogenic differentiation is regulated by the cell cycle and differentiation factors 24 . Growth-arrested preadipocytes undergo several cell cycle rounds before terminally differentiating into adipocytes, thus suggesting that crosstalk exists between the cell cycle and cell proliferation. PPARγ is a critical regulator of adipogenic differentiation 23 . Protein phosphorylation is a type of PTM that regulates a wide range of signaling pathways involved in differentiation, apoptosis, proliferation, gene regulation, and metabolism [25][26][27] . PPARγ is a short-lived protein 28 that is regulated by a series of PTMs 29 , including phosphorylation and acetylation 30 . PPARγ can be phosphorylated by mitogen-activated protein kinases, cyclin-dependent kinase 5, and AMP-activated protein kinase. The phosphorylation of PPARγ decreases the expression of PPARγ mRNA and protein, and inhibits adipogenic differentiation 31,32 . The deacetylation of PPARγ by SIRT1 downregulates the expression of PPARγ and inhibits adipogenic differentiation 24 . The acetylation of the conserved lysine motif (K154/155) of PPARγ1 promotes lipid synthesis in ErbB2-positive breast cancer cells 33 . Our study confirmed that the expression levels of PPARγ and FABP4 gradually increased, and that of phospho-PPARγ (Ser112) decreased in HEY and MDA-MB-231 PDCs with increasing culture time in adipogenic differentiation medium. Helenius et al. 34 have confirmed that phosphorylation inhibits the ability of PPARγ to promote adipogenic differentiation. Phosphorylation of PPARγ decreases its transcriptional activity, promotes Figure 8 The relevant mechanism of PDCs differentiated into adipocytes. Dexamethasone, rosiglitazone, insulin, and IBMX are all potent activators of the 3′,5′-cyclic AMP (cAMP)-dependent protein kinase pathway, and cAMP regulates the phosphorylation of P300. P300 is a histone acetyltransferase transcriptional coactivator that mediates the acetylation of P53. Acetylated P53 is in an activated state, and positive feedback regulates P300. P300-P53 interacts with PPARγ and regulates the acetylation level of PPARγ. Mutant P53 is acetylated and modified by P300, and has a beneficial function in the adipogenic differentiation of PDCs, whereas wild-type P53 is ubiquitinated by P300, thus negatively regulating the adipogenic differentiation of PDCs. As a member of the nuclear-receptor superfamily, PPARγ induces growth arrest and adipogenic differentiation. In addition, the RAS/RAF/ERK signaling pathway is associated with the development and progression of malignant tumors. ubiquitination, and further limits its ability to act as a transcriptional activator 35 .
The deacetylation of PPARγ by SIRT1 downregulates PPARγ expression and inhibits adipogenic differentiation 36 . When PPARγ1 is acetylated at the conserved lysine motif (K154/155), it promotes lipid synthesis in ErbB2positive breast cancer cells 33 . In this study, we also showed that the expression of Ace-PPARγ was higher in HEY and MDA-MB-231 PDCs after adipogenic differentiation than in PDCs before adipogenic differentiation. The tumor suppressor gene P53 is involved in cell cycle control, apoptosis, and genomic stability, and P53 mutations appear in many cancers. The dysregulated expression or function of pRB or P53 is a hallmark of all cancers 37 . Although P53 is one of the most well-studied genes, its role in adipocytes remains poorly understood. Regulation of PPARγ expression by P53 depends on the P53 genotype. Our results confirmed that P53 knockdown in HEY PDCs expressing wild-type P53 increased Ace-PPARγ expression and facilitated adipogenic differentiation. In MDA-MB-231 PDCs with mutant P53, P53 knockdown inhibited adipogenic differentiation. P300 is a HAT transcriptional coactivator that is critical in several cellular processes 38 . P300-P53 regulates the acetylation level of PPARγ, and silencing P53 or P300 disrupts the formation of the P53-P300 complex [39][40][41] . In this research, the expression levels of P300 and Ace-PPARγ were associated with the expression of P53 in HEY and MDA-MB-231 cells with different p53 genotypes. The acetylation sites of P53 overlap with the ubiquitylation or SUMOylation sites 40,41 . We confirmed that P53 was modified by SUMOylation in HEY control cells and PDCs. In HEY PDCs with wild-type P53, A485 treatment inhibited the activity of P300 and increased the expression of Ace-P53 by inhibiting SUMOylation of P53. In MDA-MB-231 PDCs with mutant P53, A485 treatment decreased the activity of P300. The decreased activity of P300 inhibited the acetylation of P53, thereby decreasing adipogenic differentiation.