**Selective PPAR**α **Modulator Pemafibrate and Sodium-Glucose Cotransporter 2 Inhibitor Tofogliflozin Combination Treatment Improved Histopathology in Experimental Mice Model of Non-Alcoholic Steatohepatitis**

**Kentaro Murakami 1,2,†, Yusuke Sasaki 1,2,†, Masato Asahiyama 2,†, Wataru Yano <sup>2</sup> , Toshiaki Takizawa <sup>2</sup> , Wakana Kamiya <sup>1</sup> , Yoshihiro Matsumura <sup>3</sup> , Motonobu Anai <sup>1</sup> , Tsuyoshi Osawa <sup>4</sup> , Jean-Charles Fruchart <sup>5</sup> , Jamila Fruchart-Najib <sup>5</sup> , Hiroyuki Aburatani <sup>6</sup> , Juro Sakai 3,7, Tatsuhiko Kodama <sup>1</sup> and Toshiya Tanaka 1,\* ,†**


**Abstract:** Ballooning degeneration of hepatocytes is a major distinguishing histological feature of non-alcoholic steatosis (NASH) progression that can lead to cirrhosis and hepatocellular carcinoma (HCC). In this study, we evaluated the effect of the selective PPARα modulator (SPPARMα) pemafibrate (Pema) and sodium-glucose cotransporter 2 (SGLT2) inhibitor tofogliflozin (Tofo) combination treatment on pathological progression in the liver of a mouse model of NASH (STAM) at two time points (onset of NASH progression and HCC survival). At both time points, the Pema and Tofo combination treatment significantly alleviated hyperglycemia and hypertriglyceridemia. The combination treatment significantly reduced ballooning degeneration of hepatocytes. RNA-seq analysis suggested that Pema and Tofo combination treatment resulted in an increase in glyceroneogenesis, triglyceride (TG) uptake, lipolysis and liberated fatty acids re-esterification into TG, lipid droplet (LD) formation, and *Cidea/Cidec* ratio along with an increased number and reduced size and area of LDs. In addition, combination treatment reduced expression levels of endoplasmic reticulum stress-related genes (*Ire1a*, *Grp78*, *Xbp1*, and *Phlda3*). Pema and Tofo treatment significantly improved survival rates and reduced the number of tumors in the liver compared to the NASH control group. These results suggest that SPPARMα and SGLT2 inhibitor combination therapy has therapeutic potential to prevent NASH-HCC progression.

**Keywords:** SPPARMα; SGLT2; ballooning; ER stress

**Citation:** Murakami, K.; Sasaki, Y.; Asahiyama, M.; Yano, W.; Takizawa, T.; Kamiya, W.; Matsumura, Y.; Anai, M.; Osawa, T.; Fruchart, J.-C.; et al. Selective PPARα Modulator Pemafibrate and Sodium-Glucose Cotransporter 2 Inhibitor Tofogliflozin Combination Treatment Improved Histopathology in Experimental Mice Model of Non-Alcoholic Steatohepatitis. *Cells* **2022**, *11*, 720. https://doi.org/ 10.3390/cells11040720

Academic Editors: Kay-Dietrich Wagner and Nicole Wagner

Received: 25 January 2022 Accepted: 15 February 2022 Published: 18 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

Non-alcoholic steatosis (NASH) is a severe form of non-alcoholic fatty liver disease (NAFLD), which is closely linked to type 2 diabetes and metabolic syndrome [1–3]. NASH is defined as the presence of steatosis, inflammation, ballooning degeneration of hepatocytes with or without fibrosis, and the eventual development of cirrhosis and hepatocellular carcinoma (HCC). In particular, higher grades of steatosis, inflammation, and ballooning degeneration are important steps in the pathogenesis of cirrhosis and HCC and are strongly associated with morbidity and mortality of liver disease [4,5]. However, the mechanism by which lipid accumulation in hepatocytes affects NASH progression is unclear. In addition, no effective therapeutic agents have been approved for treating NASH. Therefore, the development of a therapeutic approach for NASH is urgently needed.

Lipid droplets (LDs) are storage organelles that store neutral lipids such as triglycerides (TGs) and sterol esters during excess energy states and serve as a reservoir of energy supplies during the fasting state [6–8]. Importantly, not only role in maintaining lipid homeostasis but also buffering function of toxic lipid species have emerged with respect to LD biology. Dysregulated LDs homeostasis is considered to induce toxic lipid release and trigger cell death through prolonged activation of signaling pathways, such as the unfolded protein response (UPR) [9,10]. However, extensive LD accumulation in hepatocytes is not always in accordance with cellular dysfunction [11,12]. Although accumulation of LDs in hepatocytes is a prerequisite step for NASH development, changes in the composition of the lipids and proteins of LDs may play an important role in the progression from NAFLD to NASH [13,14]. Thus, the investigation of LD biogenesis and degradation, as well as the regulation of hepatic fatty acid and TG metabolism by a balance of de novo lipogenesis (DNL), glyceroneogenesis, VLDL assembly and secretion, lipolysis, and fatty acid oxidation (FAO) at the transcriptional and post-transcriptional levels, is important for understanding NASH development.

Pemafibrate (Pema) is the first clinically available selective PPARα modulator (SPPARMα); it is used to improve dyslipidemia and reduce macrovascular and microvascular complications [15–19]. We have reported that activation of PPARα by Pema induces the expression of a series of genes involved in TG hydrolysis, fatty acid uptake, fatty acid β-oxidation, and ketogenesis in the liver, supporting its ability to reduce plasma TG [20,21]. In our previous study using STAM NASH model mice, we reported that Pema treatment prevents NASH development by reducing myeloid cell recruitment without reducing hepatic TG content [22]. Therefore, we suggest that the combination of Pema and drugs that enhance the excretion or inhibit the absorption of carbohydrates and/or lipids has the potential to alleviate LD accumulation in hepatocytes and impede NASH development.

Sodium-glucose cotransporter-2 (SGLT2) inhibitors are a class of lower blood glucose drugs that increase urinary glucose excretion by inhibiting glucose reabsorption at the proximal tubule in the kidney [23–25]. Recent studies suggested that SGLT2 inhibitor treatment can reduce hepatic lipid levels and alleviate NAFLD, and has blood glucoselowering effects [26,27]. The hepatic lipid-lowering effect of SGLT2 inhibitors has been suggested, in part, based on their ability to lower circulating glucose and insulin levels, which reduces DNL. In this study, we evaluated the therapeutic potential of the combination of Pema and the SGLT2 inhibitor tofogliflozin (Tofo) in STAM NASH model mice at two time points (onset of NASH progression and HCC survival).

#### **2. Materials and Methods**

#### *2.1. Reagents*

Pema and Tofo were kindly provided by Kowa Co., Ltd. (Nagoya, Japan). Streptozotocin (STZ) was purchased from Sigma-Aldrich (St. Louis, MO, USA) and Arabic gum from Wako Pure Chemical Industries (Osaka, Japan).

#### *2.2. Animal Treatment*

#### 2.2.1. Progression Prevention Study

STAM mice were generated as previously described [22]. Pathogen-free pregnant C57BL/6J mice were obtained from CLEA Japan (Tokyo, Japan). All mice were housed in a temperature-controlled (24 ◦C) facility with a 12-h light/12-h dark cycle (08:00–20:00 h) and ad libitum access to food and water, except for the drug treatment period. Two days after birth, male mice received a subcutaneous injection of 200 µg STZ (Sigma, St. Louis, MO, USA) and were fed HFD32 (32% fat, CLEA Japan) ad libitum after 4 weeks of age of weaning. Two weeks after HFD32 feeding, mice were randomly divided into four groups: STAM control group fed HFD32 with vehicle treatment, Pema-treated group fed HFD32 with Pema (0.1 mg/kg), Tofo-treated group fed a HFD32 with Tofo (10 mg/kg), and Pema and Tofo combination (Pema 0.1 mg/kg and Tofo 10 mg/kg) for 3 weeks (6–9 weeks). Drugs were administered at 5 mL/kg body weight by oral intubation in 3% Arabic gum daily between 09:30 and 10:00 h. HFD32 was fed in a pair-feeding manner (2.3–2.8 g/mouse/day). In the drug treatment groups, animals were fed the same amount of HFD32 diet as that consumed by the control group over the preceding 24 h. In addition, normal diet (CE-2; 5% fat, CLEA Japan) fed normal group was orally administered vehicle for 3 weeks. Four hours after final administration, mice were sacrificed, and serum parameters measurement, histology, TG content determination, and gene expression analysis of liver were carried out. The study protocol was approved in accordance with the relevant guidelines and regulations of the Animal Care and Use Committee of the University of Tokyo (RAC12011, RAC170001).

#### 2.2.2. Survival Study

Male STAM mice were purchased from SMC Laboratories (Tokyo, Japan) at 5 weeks of age and were fed HFD32. C57BL/6J normal mice were purchased from Japan SLC (Shizuoka, Japan) and fed a normal diet, CE-2. All mice were housed at 23 ± 3 ◦C and 55 ± 15% relative humidity (RH) under a 12-h light/dark cycle (07:00–19:00 h) and provided with food and water ad libitum. At 6 weeks of age, STAM mice were divided into four groups based on body weight: control, Pema (0.00008% equivalent to 0.1 mg/kg), Tofo (0.015% equivalent to 10 mg/kg) [28,29], and Pema and Tofo combination (*n* = 20 each). C57BL/6J mice with normal chow were assigned to the normal group (*n* = 8). Pema and/or Tofo were mixed in the diet and administered to each group. The study protocol was approved in accordance with the relevant guidelines and regulations of the Animal Care and Use Committee of Tokyo New Drug Research Laboratories, Kowa Company, Ltd. (Tokyo, Japan).

#### *2.3. Blood Parameter*

Serum total cholesterol (TC), TG, glucose, non-esterified fatty acids (NEFA), AST, ALT, phospholipids (PL), and creatinine (CRN) levels were determined using a Labospect 003 autoanalyzer (Hitachi High-Technologies Corporation, Tokyo, Japan).

#### *2.4. Histology*

A histological study was performed as previously described [22]. For immunohistochemistry, blocking of endogenous peroxidase activity was performed using 0.03% H2O<sup>2</sup> in methanol. Obtained liver sections were treated with the anti-ER-TR7 (Abcam, Cambridge, MA, USA) antibodies overnight at 4 ◦C. After treatment with secondary antibodies, the substrate reaction was performed using 3,30 -diaminobenzidine (Dojindo, Kumamoto, Japan) solution.

According to Kleiner et al. [30], the NAFLD activity score (NAS) was calculated. Quantitative five grades assessment of Oil Red O staining was carried out by scoring of positive areas. Quantitative estimations of ER-TR7 and Sirius-red positive areas were carried out of the positive areas in five fields. Briefly, for each animal, bright field images of stained sections were captured around the central veins at 400-fold magnification using

a digital camera (DP72, Olympus, Tokyo, Japan), and quantitatively estimated using WinROOF image processing software (Mitani, Tokyo, Japan). The results were shown as the mean of five different fields in each section.

#### *2.5. RNA-Sequencing*

For genome-wide transcriptome analysis, RNA-sequencing (RNA-Seq) was performed as previously described [22]. Briefly, sequencing of the RNA libraries was carried out using 150-bp paired-end mode of the TruSeq Rapid PE Cluster Kit and TruSeq Rapid SBS kit (Illumina) on the Illumina HiSeq 2500 platform. RNA-seq reads were mapped onto the reference mouse genome (NCBI37/mm9) and transcriptome (UCSC gene), respectively, using Burrows-Wheeler Aligner. Transcript coordinates were converted to genomic positions, and then an optimal mapping result was chosen either via transcript or genome mapping by comparing the minimal edit distance to the reference. Local realignment was implemented within an in-house short read aligner with a smaller k-mer size (k = 11). Eventually, fragments per kilo base of exon per million fragments mapped (fpkm) values were calculated for each UCSC gene while considering strand-specific information.

#### *2.6. Quantitative Real-Time PCR (qPCR)*

qPCR was performed as previously described [22,31,32]. *Ppia* mRNA was used as an independent control. All primers used for qPCR are listed in Supplementary Table S1.

#### *2.7. LD Analysis*

LD evaluation was performed as previously described [22]. For hepatic LD analysis, "Image J" imaging software (https://imagej.nih.gov/ij/download.html (accessed on 7 August 2016) was applied. H&E staining images were opened with Image J software and converted into grayscale (8 bit). Then, the lipid drop areas were extracted by using the threshold (Min: 220, Max: 255). After eliminating blood vessels, LD areas were analyzed and quantified using the "Analyze particles" function. Quantified LD area data were firstly obtained by pixel, and then they were converted into µm<sup>2</sup> (1 µm = 3 pixels, determined by scale bar size). LD diameter was also analyzed. Data were shown as the mean values from three different images of each animal. A histogram was created with Microsoft Excel spreadsheet software.

#### *2.8. Statistical Analyses*

All data are presented as the mean ± SEM. Homogeneity invariance was evaluated by Bartlett's test followed by parametric or non-parametric Dunnett's multiple comparison test (two-sided). \* *p* < 0.05, \*\* *p* < 0.01. In the survival study, data were analyzed using the multiple log-rank test and Cox proportional hazard model.

#### **3. Results**

#### *3.1. Pema and Tofo Combination Prevents Ballooning Degeneration of Hepatocytes*

To investigate the effects of Pema and Tofo combination on NASH development in the STAM mouse model, each drug and combination was administered for three weeks. STAM mice showed significant hyperglycemia and hypertriglyceridemia; higher phospholipid, FGF21, and AST levels; lower body weight; and higher liver weight, compared to normal C57BL/6J mice (Table 1). Pema significantly reduced serum TG levels, but it did not alter serum glucose (Figure 1A,B), AST, and ALT levels. In addition, Pema significantly increased liver weight, which is a well-known effect of PPARα stimulation in rodents [33,34]. Tofo significantly reduced serum TG and glucose levels (Figure 1A,B). Pema and Tofo combination treatment effectively reduced serum TG and glucose levels and increased FGF21 levels.


**Table 1.** Effects of Pema, Tofo, and Pema and Tofo combination on body and liver weight, biochemical parameters in the serum, immunohistochemical analysis, and NAS.

TC: total cholesterol, PL: phosphorlipids, NEFA: non-esterified fatty acid, CRN: creatinine, AST: aspartate aminotransferase, ALT: alanine aminotransferase, NAS: NAFLD activity score. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test.

H&E staining clarified that STAM control mouse liver owned liver nodules, macroand micro-vesicular lipid accumulation, inflammatory cell infiltration, and ballooning degeneration of hepatocytes, unlike normal mouse liver (Figure 1C). Pema-treated mouse liver included less macrovesicular lipid accumulation, less ballooning degeneration, and a tendency to reduce the NAS compared to STAM control mice (Table 1). Tofo treatment reduced macrovesicular lipid accumulation and ballooning degeneration. The Pema and Tofo combination treatment significantly reduced ballooning degeneration (Figure 1D).

**Figure 1.** Pemafibrate and Tofogliflozin combination improves hypertriglyceridemia, hyperglycemia, macrovesicular steatosis, and ballooning score in STAM mice liver. (**A**) Serum triglyceride, (**B**) Serum glucose, (**C**) Representative gross morphology of liver, H&E stained, ER-TR7 stained, Siriusred stained, and Oil red O stained liver section, and (**D**) Ballooning score of normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test. **Figure 1.** Pemafibrate and Tofogliflozin combination improves hypertriglyceridemia, hyperglycemia, macrovesicular steatosis, and ballooning score in STAM mice liver. (**A**) Serum triglyceride, (**B**) Serum glucose, (**C**) Representative gross morphology of liver, H&E stained, ER-TR7 stained, Sirius-red stained, and Oil red O stained liver section, and (**D**) Ballooning score of normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test.

#### *3.2. Pema and Tofo Combination Treatment Induces Lipolysis and Re-Esterification Cycles of TG in STAM Mouse Livers 3.2. Pema and Tofo Combination Treatment Induces Lipolysis and Re-Esterification Cycles of TG in STAM Mouse Livers*

Although Pema, Tofo, and combination treatment resulted in decreased macrovesicular steatosis, this reduction was not reflected in the steatosis and Oil Red O staining scores (Table 1). To verify the effect of Pema and Tofo combination treatment on STAM mouse livers, we performed a global gene expression analysis by RNA-seq using liver tissues collected from normal, STAM control, Pema-treated, Tofo-treated, and Pema and Tofo combination-treated STAM mice. We identified 125 upregulated and 68 downregulated genes in the Pema and Tofo combination treatment compared with the STAM control group according to our stringent criteria (Supplementary Figure S1A). These genes included almost all Pema- and/or Tofo-regulated genes. In particular, the most upregulated genes by combination treatment were PPARα target genes involved in lipid metab-Although Pema, Tofo, and combination treatment resulted in decreased macrovesicular steatosis, this reduction was not reflected in the steatosis and Oil Red O staining scores (Table 1). To verify the effect of Pema and Tofo combination treatment on STAM mouse livers, we performed a global gene expression analysis by RNA-seq using liver tissues collected from normal, STAM control, Pema-treated, Tofo-treated, and Pema and Tofo combination-treated STAM mice. We identified 125 upregulated and 68 downregulated genes in the Pema and Tofo combination treatment compared with the STAM control group according to our stringent criteria (Supplementary Figure S1A). These genes included almost all Pema- and/or Tofo-regulated genes. In particular, the most upregulated genes by combination treatment were PPARα target genes involved in lipid metabolism (Supplementary Figure S1B).

olism (Supplementary Figure S1B). To understand the effect of Pema and Tofo combination treatment on STAM mouse To understand the effect of Pema and Tofo combination treatment on STAM mouse livers, lipid and carbohydrate metabolism-related gene expression levels were analyzed. We

livers, lipid and carbohydrate metabolism-related gene expression levels were analyzed. We found that the expression levels of genes related to TG hydrolysis, fatty acid uptake, fatty acid activation, fatty acid binding, peroxisomal and mitochondrial oxidation, and found that the expression levels of genes related to TG hydrolysis, fatty acid uptake, fatty acid activation, fatty acid binding, peroxisomal and mitochondrial oxidation, and ketogenesis were increased in the STAM control group than in the normal group (Supplementary Figure S2). Tofo and Pema monotherapy upregulated the expression of these genes, and the combination treatment upregulated their expression further. Importantly, the combination of Pema and Tofo dramatically increased the *Pdk4* gene expression level, indicating that it mediates the inhibition of glucose oxidation and preferential activation of FAO (Supplementary Figure S2).

Increased glucose and fructose uptake in hepatocytes accelerate glycolysis and DNL to generate TG. Especially, the glycerolipid synthesis pathway (glyceroneogenesis) and the monoacylglycerol pathway are key players in TG synthesis (Figure 2A). The STAM control mouse livers exhibited higher levels of glycolysis-related gene expression than the normal mouse livers (Supplementary Figure S3 and Figure 2B). In addition, we found that glyceroneogenesis and re-esterification of 2-monoacylglycerol were induced in STAM control livers, in addition to simultaneous TG uptake and hydrolysis. The Pema and Tofo combination treatment did not affect glycolysis-related gene expression, but it significantly induced a series of genes involved in TG synthesis from glycerol 3-phosphate and reesterification from monoacylglycerols and diacylglycerols generated by TG hydrolysis in STAM mouse livers (Figure 2B). These results suggest that the Pema and Tofo combination enhances the uptake and oxidation of fatty acids, TG synthesis from glycerol 3-phosphate, and the re-esterification of glycerol generated by TG hydrolysis in STAM mouse livers.

#### *3.3. Pema and Tofo Combination Increases Small LDs in STAM Mouse Livers*

To better understand the effect of Pema and Tofo combination treatment on steatosis in STAM mice, we measured the TG concentration in the liver (Figure 3A). The STAM control group showed a significantly increased TG content in the liver. Pema significantly increased, and Tofo and Pema and Tofo combination tended to decrease, the TG content in STAM mouse livers (Figure 3A). Because the combination of Pema and Tofo markedly improved macrovesicular steatosis (Figure 1C), we evaluated LD counts and size distributions. Pema and Tofo treatments increased the droplet number and decreased the LD area (Figure 3B,C). Furthermore, this drug combination treatment increased the percentage of cells representing small LDs (<1 µm) from 29.40% in the control to 49.65% and decreased the percentage of cells representing large LDs (>3 µm) from 39.36% in the STAM control to 7.87% (Figure 3D).

LDs consist of an inner core of neutral lipids, including TG and sterol esters, a phospholipid monolayer, and LD-associated proteins (LDAPs) [6–8]. Because LDAPs affect LD function and dynamics [6–8], we evaluated the effect of Pema and Tofo combination on LDAP expression (Figure 3E). The Pema and Tofo combination group showed increased expression of genes related to LD inner core lipid synthesis (*Agpat6*, *Dgat1*, and *Acat1*), formation (*Agpat6*, *Acsl3*, *Mettl7b*, and *Plin2*), budding (*Fitm2* and *Bscl2*), stabilization (*Plin4* and *Plin5*), lipolysis (*Pnpla2*, *Hsd17b11*, *Pcyt1a*, and *Abhd5*), expansion (*Agpat3*, *Pex3*, and *Tcp1*), and fusion (*Cidea* and *Cidec*). Among these genes, the Pema and Tofo combination induced *Cidea* expression. Recently, Sans et al. suggested that hepatic CIDEA and CIDEC correlated negatively and positively, respectively, with steatohepatitis and liver injury in mice, as well as steatosis and NASH in obese humans [35]. In addition, suppression of CIDEC has been reported to reduce LD size and stimulate lipolysis [36]. Pema, Tofo, and combination treatments induced expression of *Cidea* and *Cidec*, and combination treatment strongly induced *Cidea* gene expression, thereby increasing the *Cidea*/*Cidec* ratio. These results may contribute to the reduction in LD size and stimulation of lipolysis by combination treatment.

ketogenesis were increased in the STAM control group than in the normal group (Supplementary Figure S2). Tofo and Pema monotherapy upregulated the expression of these genes, and the combination treatment upregulated their expression further. Importantly, the combination of Pema and Tofo dramatically increased the *Pdk4* gene expression level, indicating that it mediates the inhibition of glucose oxidation and preferential activation

Increased glucose and fructose uptake in hepatocytes accelerate glycolysis and DNL to generate TG. Especially, the glycerolipid synthesis pathway (glyceroneogenesis) and the monoacylglycerol pathway are key players in TG synthesis (Figure 2A). The STAM control mouse livers exhibited higher levels of glycolysis-related gene expression than the normal mouse livers (Supplementary Figures S3 and 2B). In addition, we found that glyceroneogenesis and re-esterification of 2-monoacylglycerol were induced in STAM control livers, in addition to simultaneous TG uptake and hydrolysis. The Pema and Tofo combination treatment did not affect glycolysis-related gene expression, but it significantly induced a series of genes involved in TG synthesis from glycerol 3-phosphate and re-esterification from monoacylglycerols and diacylglycerols generated by TG hydrolysis in STAM mouse livers (Figure 2B). These results suggest that the Pema and Tofo combination enhances the uptake and oxidation of fatty acids, TG synthesis from glycerol 3-phosphate, and the re-esterification of glycerol generated by TG hydrolysis in STAM mouse

of FAO (Supplementary Figure S2).

livers.

**Figure 2.** Pemafibrate and Tofogliflozin combination induces lipolysis and fatty acid re-esterification genes expression in STAM mice liver. (**A**) Schematic representation of the glycolytic and TG synthesis pathways in the liver. (**B**) qPCR validation of glycolytic and TG metabolism-related genes expression of normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice liver. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test.

#### *3.4. Pema and Tofo Combination Inhibits the IRE1α-XBP1-PHLDA3 Pathway*

LD biogenesis and enhanced esterification of fatty acids play a key role in buffering toxic lipid species [6–8]. Several reports have indicated that fatty acid accumulation in hepatocytes can lead to cell dysfunction and cell death through endoplasmic reticulum (ER) stress [37,38]. UPR signaling is mainly driven by three sensors mediated by inositol requiring enzyme 1 α (IRE1α), protein kinase RNA-like endoplasmic reticulum (ER) kinase (PERK), and activating transcription factor 6 (ATF6) [39,40]. In addition, the luminal side of each UPR sensor interacts with chaperones of immunoglobulin-binding protein/glucose regulatory protein 78 (BiP/GRP78) [39–41]. Livers from the STAM control group showed enhanced UPR sensors (*Ire1a*, *Perk*, and *Atf6*), chaperones (*Grp78* and *Pdi1a*), antioxidant defense-regulated transcription factor (*Nrf2*), IRE1α interaction protein form apoptosis mediator complex (*Traf2*), and proapoptotic BCL-2 protein family (*Bak1* and *Bax*). Pema and Tofo combination significantly reduced *Ire1a*, *Grp78*, *Xbp1*, and *Phlda3* expression levels (Figure 4). Recently, ER stress in hepatocytes has been reported to induce PHLDA3 via

the IRE1–Xbp1s pathway, which facilitates liver injury by inhibiting Akt [42]. These data suggest that the combination of Pema and Tofo prevents liver injury by inhibiting the IRE1α-XBP1-PHLD3A pathway. *Cells* **2022**, *11*, x FOR PEER REVIEW 9 of 18

**Figure 3.** Pemafibrate and Tofogliflozin combination induces lipid droplets formation. (**A**) Liver TG content, (**B**) Lipid droplet number, (**C**) Median lipid droplet, (**D**) Lipid droplet sizes distribution, (**E**) Heatmap of hierarchical clustering of LDAPs and formation-related genes, and (**F**) *Cidea*, *Cidec*, and *Cidea*/*Cidec* ratio of control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test. **Figure 3.** Pemafibrate and Tofogliflozin combination induces lipid droplets formation. (**A**) Liver TG content, (**B**) Lipid droplet number, (**C**) Median lipid droplet, (**D**) Lipid droplet sizes distribution, (**E**) Heatmap of hierarchical clustering of LDAPs and formation-related genes, and (**F**) *Cidea*, *Cidec*, and *Cidea*/*Cidec* ratio of control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combinationtreated STAM mice. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test.

#### *3.4. Pema and Tofo Combination Inhibits the IRE1α-XBP1-PHLDA3 Pathway 3.5. Pema and Tofo Combination Improves HCC-Related Survival*

LD biogenesis and enhanced esterification of fatty acids play a key role in buffering toxic lipid species [6–8]. Several reports have indicated that fatty acid accumulation in hepatocytes can lead to cell dysfunction and cell death through endoplasmic reticulum (ER) stress [37,38]. UPR signaling is mainly driven by three sensors mediated by inositol requiring enzyme 1 α (IRE1α), protein kinase RNA-like endoplasmic reticulum (ER) kinase (PERK), and activating transcription factor 6 (ATF6) [39,40]. In addition, the luminal side of each UPR sensor interacts with chaperones of immunoglobulin-binding protein/glucose regulatory protein 78 (BiP/GRP78) [39–41]. Livers from the STAM control group showed enhanced UPR sensors (*Ire1a*, *Perk*, and *Atf6*), chaperones (*Grp78* and *Pdi1a*), antioxidant defense-regulated transcription factor (*Nrf2*), IRE1α interaction pro-Finally, to determine whether Pema and Tofo combination can prevent the progression of NASH to HCC, we treated STAM mice for 16 weeks. As observed in the 3-week drug treatment on NASH progression, the combination of Pema and Tofo treatment resulted in a significant decrease in serum TG and blood glucose levels (Figure 5A,B). Levels of serum AFP, an oncofetal protein that is used as a tumor marker, significantly increased in the STAM vehicle control group and decreased in the combination treatment group (Figure 5C). Kaplan-Meier survival curves show that the survival rate of the control group decreased to 10%. Pema in the diet showed a tendency to increase the survival rate (30%), and the combination of Pema and Tofo significantly improved the survival rates (50%) compared to

tein form apoptosis mediator complex (*Traf2*), and proapoptotic BCL-2 protein family (*Bak1* and *Bax*). Pema and Tofo combination significantly reduced *Ire1a*, *Grp78*, *Xbp1*, and

*Cells* **2022**, *11*, x FOR PEER REVIEW 10 of 18

the control group. In addition, each drug-treated group had a markedly reduced number of tumors in the liver (Figure 5E,F). ing Akt [42]. These data suggest that the combination of Pema and Tofo prevents liver injury by inhibiting the IRE1α-XBP1-PHLD3A pathway.

*Phlda3* expression levels (Figure 4). Recently, ER stress in hepatocytes has been reported to induce PHLDA3 via the IRE1–Xbp1s pathway, which facilitates liver injury by inhibit-

**Figure 4.** Pemafibrate and Tofogliflozin combination improves ER stress genes expression in STAM mice liver. qPCR validation of ER stress-related genes expression of normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice liver. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test. **Figure 4.** Pemafibrate and Tofogliflozin combination improves ER stress genes expression in STAM mice liver. qPCR validation of ER stress-related genes expression of normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice liver. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test. *Cells* **2022**, *11*, x FOR PEER REVIEW 11 of 18

**Figure 5.** Pemafibrate and Tofogliflozin combination improves survival rate in STAM mice liver. (**A**) Serum triglyceride, (**B**) Serum glucose, (**C**) Serum AFP, (**D**) Kaplan-Meier survival curves, (**E**) Number of tumors, and (**F**) Representative gross morphology of liver from normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice. Log-rank *p*value and hazard ratio were shown in the survival curve figure. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test. **Figure 5.** Pemafibrate and Tofogliflozin combination improves survival rate in STAM mice liver. (**A**) Serum triglyceride, (**B**) Serum glucose, (**C**) Serum AFP, (**D**) Kaplan-Meier survival curves, (**E**) Number of tumors, and (**F**) Representative gross morphology of liver from normal, control, pemafibrate, tofogliflozin, and pemafibrate and tofogliflozin combination-treated STAM mice. Log-rank *p*-value and hazard ratio were shown in the survival curve figure. Error bars show s.e.m. \* *p* < 0.05; \*\* *p* < 0.01: Significantly difference from STAM control group by Dunnett's multiple comparison test.

but several proposed hypotheses suggest that steatosis with additional factors, such as insulin resistance, oxidative stress, ER stress, and mitochondrial dysfunction, may be involved [1–3]. Our previous study revealed that Pema prevents NASH development without reducing the TG content in the liver [22]. We also revealed that although Pema improves macrovesicular steatosis by enhancing TG hydrolysis while simultaneously enhancing esterification of fatty acids for TG and LD biogenesis, it may not result in a sufficient TG reduction in STAM mouse livers. Based on these observations, we hypothesized that the combination of Pema with a drug that enhances the excretion of carbohydrates from the kidney via SGLT2 inhibition has the potential to improve TG accumulation and NASH development. The combination of Pema and Tofo significantly improved ballooning degeneration of hepatocytes and reduced hepatic TG accumulation. In addition, the combination of Pema and Tofo specifically reduced *Ire1a*-*Xbp1*-*Phld3a* gene expression in the NASH liver. These results suggest that the combination of SPPARMα and SGLT2 inhibitors has therapeutic potential for NASH and NASH-related HCC via reduction of ER

Hepatic TG accumulation has been suggested to play a central role in NAFLD and NASH, which can progress to cirrhosis and liver failure [1–6]. The mechanisms underly-

LDs are organelles that store neutral lipids, such as TG and sterol esters, as sources of energy and cell membrane synthesis [6–8]. When cells face excess neutral lipids, such as fatty acids and sterols, they synthesize LDs and disperse them into the cytoplasm. This process is also important for protecting cells from the toxicity associated with an excess of

**4. Discussion** 

stress-induced liver injury.

#### **4. Discussion**

Hepatic TG accumulation has been suggested to play a central role in NAFLD and NASH, which can progress to cirrhosis and liver failure [1–6]. The mechanisms underlying the pathogenesis of NASH in a subset of patients with steatosis have not been clarified, but several proposed hypotheses suggest that steatosis with additional factors, such as insulin resistance, oxidative stress, ER stress, and mitochondrial dysfunction, may be involved [1–3]. Our previous study revealed that Pema prevents NASH development without reducing the TG content in the liver [22]. We also revealed that although Pema improves macrovesicular steatosis by enhancing TG hydrolysis while simultaneously enhancing esterification of fatty acids for TG and LD biogenesis, it may not result in a sufficient TG reduction in STAM mouse livers. Based on these observations, we hypothesized that the combination of Pema with a drug that enhances the excretion of carbohydrates from the kidney via SGLT2 inhibition has the potential to improve TG accumulation and NASH development. The combination of Pema and Tofo significantly improved ballooning degeneration of hepatocytes and reduced hepatic TG accumulation. In addition, the combination of Pema and Tofo specifically reduced *Ire1a*-*Xbp1*-*Phld3a* gene expression in the NASH liver. These results suggest that the combination of SPPARMα and SGLT2 inhibitors has therapeutic potential for NASH and NASH-related HCC via reduction of ER stress-induced liver injury.

LDs are organelles that store neutral lipids, such as TG and sterol esters, as sources of energy and cell membrane synthesis [6–8]. When cells face excess neutral lipids, such as fatty acids and sterols, they synthesize LDs and disperse them into the cytoplasm. This process is also important for protecting cells from the toxicity associated with an excess of lipids such as fatty acids, glycerolipids, and sterols [6–8]. Therefore, control of LD biogenesis and consumption plays a key role in the pathogenesis of NASH. In this study, we found that the Pema and Tofo combination induced expression of genes involved in TG uptake, lipolysis, fatty acid uptake, fatty acid β-oxidation and esterification, and ketogenesis, as well as PDK4, which inhibits glucose oxidation. These results are consistent with our previous results in Pema-treated STAM mice liver [22] and were enhanced by the combination treatment used in this study. We found that Tofo improved hyperglycemia and serum TG levels and reduced TG content in the liver of STAM mice, whereas Pema reduced serum TG levels but did not reduce liver TG content. The STAM model is characterized by hyperglycemia and reduced body weight with reduced *Gck* expression, which is exclusively regulated by insulin signaling [43]. Therefore, insulinstimulated DNL gene regulation mediated through sterol regulatory element-binding protein 1c (SREBP1c) is unlikely to contribute to these effects, suggesting that carbohydrate response element-binding protein (ChREBP) signaling regulates glycolytic and DNL genes in this model. Although PPARα activation by Pema did not affect hyperglycemia, glycolytic genes, or *G6pc* expression, which is associated with ChREBP activation by fructose, Tofo tended to reduce the expression levels of these genes. These results suggest that SGLT2 inhibitors reduce the influx of the substrate for DNL and reduce TG content in the liver of STAM mice. Our transcriptome analysis also shows that PPARα activation by Pema induced FAO and ketogenesis, but in STAM mice with a high concentration of β-hydroxybutyrate in the blood stream, it led to re-esterification of fatty acids released from the TG and sterol esters by lipolysis and uptake into the TG for LD synthesis in the liver.

In addition, we found that Pema and Tofo combination significantly increased LD number, reduced LD size, and improved macrovesicular steatosis. Consistent with the increased number of LDs, Pema and Tofo combination also induced expression of genes involved in LD inner core lipid synthesis and formation (*Agpat6*, *Dgat1*, and *Acat1*), budding (*Fitm2* and *Bscl2*), and fusion (*Cidea* and *Cidec*) proteins. Although the biological role of LD diversification has not been clarified yet, increased numbers of small LDs may explain the protection against lipotoxicity [6]. LDs have been suggested to protect against lipotoxicity under a variety of stressful conditions such as lipid overload, hypoxia, oxidative stress, autophagic flux, and dysfunctional lipolysis [6–8]. In fact, DGAT1-dependent LD biogenesis

has been suggested to prevent lipotoxic mitochondrial dysfunction [44]. In addition, Becuwe et al. have reported that Fit2, encoded by *Fitm2*, is an evolutionarily conserved fatty acyl-CoA diphosphatase that maintains the ER structure, protects against ER stress, and enables normal lipid storage in LDs [45]. Furthermore, the differential expression of cell death-inducing DFF45-like effector (CIDE) family members CIDEC and CIDEA, recognized as regulators of LD growth, has been reported to be linked to NAFLD progression and liver injury, and CIDEA expression level decreases with NAFLD severity [35]. CIDEA and CIDEC are strongly expressed in brown adipocytes and white adipocytes, respectively, and are associated with the formation of multilocular small LDs (that are prone to lipolysis) and the storage form of unilocular LDs [36]. In this study, we found that *Cidea* was the most highly induced gene among LDAPs, and the *Cidea*/*Cidec* ratio was significantly increased by the Pema and Tofo combination. These results suggest that the combination of Pema and Tofo promotes fatty acid catabolism via lipolysis and β-oxidation while promoting re-esterification of excess fatty acids and LD biogenesis, thereby preventing lipotoxicity.

Hepatic steatosis has a risk of steatohepatitis, fibrosis, cirrhosis, liver failure, and HCC, and it was reported that dysregulation of microbial metabolites such as aromatic and branched-chain amino acid (AAA and BCAA) or of iron metabolism are related to liver fat accumulation and facilitate steatosis [46,47]; however, additional factors, such as insulin resistance, oxidative stress, ER stress, and mitochondrial dysfunction, are also involved in a disease progression [37,38]. Because the sequential esterification of fatty acids into a glycerol backbone to generate TG and budding as nascent LDs occur at the ER membrane, LDs are closely associated with ER homeostasis. Induction of UPR sensors (*Ire1a, Perk, and Atf6*), chaperones (*Grp78* and *Pdi1a*), antioxidant defense-regulated transcription factor (*Nrf2*), IRE1α interaction protein form apoptosis mediator complex (*Traf2*), and proapoptotic BCL-2 protein family (*Bak1* and *Bax*) genes were observed in the liver of STAM mice. Although numerous reports have indicated that the ER stress response plays a key role in NASH development, it is unknown which UPR sensor signaling contributes to the development of this disorder [39,40]. Pema and Tofo combination selectively reduced *Ire1a, Grp78, Xbp1*, and *Phlda3* expression levels in STAM mouse livers. Among the UPR sensors, IRE1 is the most evolutionarily conserved, implying that it plays a crucial role in cell fate determination under ER stress conditions. It has been indicated that IRE1 is capable of inducing cell fate by two distinct pathways through XBP1-mediated gene regulation and interaction with TNF receptor-associated factor 2 (TRAF2) to initiate the apoptosis signalregulating kinase 1 (ASK1) and c-Jun N-terminal kinase (JNK) signaling cascades [48]. However, because ASK1 is activated by stress responses to ER stress, as well as ROS and TNFα, the impact of ASK1 activation by the ER stress pathway on NASH development remains unclear. A recent study reported a positive and negative effect of hepatic ASK1 ablation on NASH development in HFD-fed mice [49]. Similarly, treatment with the ASK1 inhibitor selonsertib in patients with NASH yielded controversial results [50]. However, because the effects of ASK1 inhibitors are not limited to the liver and include effects on other tissues, further studies investigating the role of ASK1 in NASH development are warranted. Recently, several reports indicated that pleckstrin homology-like domain family A member 3 (PHLDA3) functions as an AKT inhibitor and plays a crucial role in the cell fate of cancer cells [42]. In addition, PHLDA3 overexpression causes tissue injury, and the IRE1-Xbp1 pathway induces PHLDA3 overexpression, which facilitates liver injury [51]. Therefore, these results and reports suggest that the Pema and Tofo combination prevents liver injury by inhibiting the lipotoxicity-induced IRE1α-XBP1-PHLD3A pathway, thereby controlling toxic lipid esterification, LD biogenesis, and the lipolysis cycle.

Epidemiological studies have shown that NASH is closely linked to type 2 diabetes and metabolic syndrome [1–3]. However, the STAM mouse is recognized as a type 1 diabetes-related NASH model with hyperglycemia, reduced body weight gain, and lack of insulin secretion and fatty acid mobilization from adipose tissue. In general, storage TG in hepatocytes requires both fatty acids and glycerol and has been suggested to be mainly regulated by the pool size of fatty acid [52]. Although fat accumulation in the

liver with type 1 diabetes has been reported, much less attention could be attributed to NASH prevalence of type 1 diabetes as compared to type 2 diabetes and metabolic syndrome. However, a recent report has suggested that NAFLD prevalence in patients with type 1 diabetes is considerable in meta-analysis [53], and several hypotheses have been proposed to explain the pathogenesis of liver steatosis in type 1 diabetes. These include insufficient TG secretion from the liver as VLDL, SREBP1c, and ChREBP induced DNL and conversion of sugar into fat [54]. On the other hand, the importance of circulating fatty acid influx has been suggested to contribute to increased hepatic lipid accumulation in type 2 diabetes [55], and circulating NEFA, dietary fat, and DNL have been reported to account for 59, 15, and 26% of the TG content in hepatocytes, respectively [56]. From these observations, adipose tissue-derived fatty acid influx and DNL have been suggested to play a crucial role in hepatic TG accumulation in type 2 diabetes-related NASH. In fact, it is well known that DNL is stimulated by insulin via SREBP1c activation and by influx glucose via ChREBP [57]. Insulin also activates LXRα, which in turn induces SREBP1c expression. In addition, impaired lipoprotein metabolism (VLDL export) and mitochondrial function (fatty acid entry and oxidation) have been suggested in the hepatic TG accumulation under insulin resistance [58]. Therefore, increased fatty acid influx, enhanced DNL, impaired TG secretion as VLDL, and mitochondrial dysfunction have been linked to human type 2 diabetic-related NASH. In the present study, we showed that there were no significant changes in serum NEFA in the STAM mouse compared to the normal mouse. In addition, our RNA-seq analysis indicated that impaired VLDL secretion and SREBP1c mediated DNL is unlikely to be the cause of hepatic steatosis in the STAM mouse model because VLDL assembly regulated *Mttp* was induced, insulin and SREBP1c target gene of *Gck* was reduced, and *Pck1*, which is negatively regulated by insulin, was induced. From several reports and our observations, this model may not completely reflect the human NASH liver metabolic state and may be a model in which the effect of SGLT2 inhibitors is more likely to be effective. Thus, additional studies using other NASH models with obesity and insulin resistance are warranted to evaluate the effect of the Pema and Tofo combination treatment on human NASH development.

In addition, although Pema and Tofo combination treatment significantly induced fatty acid catabolism, fatty acid re-esterification, and LD biogenesis; impeded the IRE1α-XBP1- PHLD3A pathway; and alleviated ballooning degeneration of hepatocytes, the precise underlying mechanism is still largely unknown. It is well known that hepatocytes are not equally responsible for liver metabolism, and the existence of so-called metabolic zonation based on oxygen tension has been proposed [59,60]. For example, gluconeogenesis, fatty acid β-oxidation, cholesterol synthesis, and ureagenesis are mainly considered to be performed by hepatocytes in the periportal region, where the oxygenated blood is transported via hepatic arteries, whereas glycolysis, DNL, bile acid synthesis, and xenobiotic detoxification occur in the pericentral region, which is relatively hypoxic [59,60]. Dysregulation of metabolic zonation is considered to lead to the development of lifestyle-related diseases such as obesity, diabetes, and NAFLD [61,62]. In fact, NAFLD is considered to begin with pericentral steatosis and inflammation with periportal inflammation and fibrosis considered late-occurring histological lesions [63]. However, the periportal disease has been associated with worse metabolic outcomes and more adverse hepatic fibrosis than pericentral disease [64]. In addition, interactions between hepatocytes and sinusoidal endothelial cells, Kupffer cells, and stellate cells are known to be involved in the pathogenesis of NASH [65]. Although Kupffer cells, T lymphocytes, and dendritic cells are more abundant in the periportal regions, infiltrating macrophages have been observed both in the periportal and pericentral regions [66]. Furthermore, preferential effects on the periportal and pericentral regions have been suggested for vitamin E and cysteamine, and PPARγ and FXR agonists, respectively, as per several randomized clinical trials [67–69]. Therefore, to understand the pathogenesis of NASH and the mechanism of therapeutic efficacy of the Pema and Tofo combination, it will be necessary to explore the spatial gene

expression profile of hepatocytes and non-parenchymal cells using scRNA-seq and slide seq technologies [70–72].

#### **5. Conclusions**

In conclusion, the combination of SPPARMα and SGLT2 inhibitor treatment prevented ballooning degeneration of hepatocytes and HCC progression. Our global gene expression analysis gives evidence of the liver protective effect of the combination therapy by inhibiting the lipotoxicity-induced IRE1α-XBP1-PHLD3A pathway. Taken together with our previous report that SPPARMα treatment prevents NASH development by reducing myeloid cell recruitment without reducing hepatic TG content [22], the combination of SPPARMα and SGLT2 inhibitor presents a promising new therapy for NASH. Our results presented in this report using the NASH mouse model gives reason to hope that the combination of SPPARMα and SGLT2 inhibitor will be synergistic. Therefore, this combination is much more effective in human NASH than monotherapy and could become an ideal strategy for long-term treatment for NASH-HCC progression.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11040720/s1, Supplementary Figure S1: Pemafibrate and tofogliflozin combination regulated genes in STAM mouse liver. Supplementary Figure S2: Effect of pemafibrate and tofogliflozin combination on fatty acid metabolism-related genes expression in STAM mouse liver. Supplementary Figure S3: Effect of pemafibrate and tofogliflozin combination on glucose metabolism and triglyceride synthesis-related genes expression in STAM mouse liver. Supplementary Table S1: qPCR Primer lists.

**Author Contributions:** Conceptualization, T.T. (Toshiya Tanaka), J.S. and T.K. designed the studies. T.T. (Toshiya Tanaka), K.M. and Y.S. wrote the manuscript. K.M., T.T. (Toshiaki Takizawa), Y.S., M.A. (Masato Asahiyama), W.Y. and T.T. (Toshiya Tanaka) performed experiments. W.K. assisted with the animal experiments and qPCR. M.A. (Motonobu Anai) performed the histological experiments. H.A. contributed to the RNA-sequencing data analysis. Y.M., M.A. (Motonobu Anai), T.O., J.-C.F. and J.F.-N. read and commented on the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported in part by a collaborative research fund from Kowa Co. Kowa Company, Ltd.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data are available upon request from the corresponding author.

**Conflicts of Interest:** T.K. is an advisory board member and a recipient of support from a collaborative research fund from Kowa Company, Ltd. J.-C.F., and J.F.-N. are consultants of Kowa Company, Ltd. Kowa Co. had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. K.M., Y.S., M.A. (Masato Asahiyama), W.Y. and T.T. (Toshiaki Takizawa) are employees of Kowa Company, Ltd. Rest of the authors declare no conflict of interest.

#### **References**


**Rossella Basilotta <sup>1</sup> , Marika Lanza <sup>1</sup> , Giovanna Casili <sup>1</sup> , Giulia Chisari <sup>2</sup> , Stefania Munao <sup>2</sup> , Lorenzo Colarossi <sup>2</sup> , Laura Cucinotta <sup>1</sup> , Michela Campolo <sup>1</sup> , Emanuela Esposito 1,\* and Irene Paterniti <sup>1</sup>**


**Abstract:** Glioblastoma (GB), also known as grade IV astrocytoma, represents the most aggressive form of brain tumor, characterized by extraordinary heterogeneity and high invasiveness and mortality. Thus, a great deal of interest is currently being directed to investigate a new therapeutic strategy and in recent years, the research has focused its attention on the evaluation of the anticancer effects of some drugs already in use for other diseases. This is the case of peroxisome proliferator-activated receptors (PPARs) ligands, which over the years have been revealed to possess anticancer properties. PPARs belong to the nuclear receptor superfamily and are divided into three main subtypes: PPAR-α, PPAR-β/δ, and PPAR-γ. These receptors, once activated by specific natural or synthetic ligands, translocate to the nucleus and dimerize with the retinoid X receptors (RXR), starting the signal transduction of numerous genes involved in many physiological processes. PPARs receptors are activated by specific ligands and participate principally in the preservation of homeostasis and in lipid and glucose metabolism. In fact, synthetic PPAR-α agonists, such as fibrates, are drugs currently in use for the clinical treatment of hypertriglyceridemia, while PPAR-γ agonists, including thiazolidinediones (TZDs), are known as insulin-sensitizing drugs. In this review, we will analyze the role of PPARs receptors in the progression of tumorigenesis and the action of PPARs agonists in promoting, or not, the induction of cell death in GB cells, highlighting the conflicting opinions present in the literature.

**Keywords:** glioblastoma; cancer; PPARs; brain; neuro-oncology

#### **1. Introduction**

Glioblastoma (GB) is the most common and aggressive subtype of malignant brain tumors. This tumor belongs to the large family of gliomas and is also known as grade IV astrocytoma [1]. GB originates from glial cells and astrocytes, which play supporting roles within the central nervous system (CNS) and is characterized by abnormal angiogenesis, apoptosis alteration, and invasiveness. This tumor manifests itself with nonspecific signs and symptoms, which vary according to the size and location of the tumor. Patients often present symptoms of increased intracranial pressure, including focal or progressive headache and neurologic deficits, personality changes, and seizures [2]. Among the genetic risk factors, a set of single nucleotide genetic polymorphisms (SNPs) have been identified, located on different genes (NF1, NF2, IDH1/IDH2, TERT, EGFR, CCDC26, CDKN2B, PHLDB1, TP53, RTEL1), which seem to contribute to gliomagenesis and to the development of all grades and histologies of gliomas [3,4]. The glioblastomas are divided into two broad categories: primary and secondary [5]. Primary GB accounts for 90% of total cases, is more frequent in the elderly population, and has a worse prognosis than its secondary counterpart. Primary GB onco-markers include overexpression of the epidermal growth

**Citation:** Basilotta, R.; Lanza, M.; Casili, G.; Chisari, G.; Munao, S.; Colarossi, L.; Cucinotta, L.; Campolo, M.; Esposito, E.; Paterniti, I. Potential Therapeutic Effects of PPAR Ligands in Glioblastoma. *Cells* **2022**, *11*, 621. https://doi.org/10.3390/ cells11040621

Academic Editors: Kay-Dietrich Wagner and Nicole Wagner

Received: 17 December 2021 Accepted: 8 February 2022 Published: 10 February 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

factor receptor (EGFR) and mutations in the tumor suppressor gene of phosphatase and tensin homolog (PTEN) and the telomerase reverse transcriptase promoter (TERT) [6]. Secondary GB constitutes 5% of cases and develops from astrocytomas with a lower degree of malignancy, affects younger patients and is related to mutations in isocitrate dehydrogenase 1 and 2 (IDH 1 and 2) and tumor protein 53 (P53) [7]. It is a tumor characterized by an extraordinary intra-tumoral heterogeneity which often results in the inability of traditional therapies to obtain long-term remissions. Glioblastomas, in fact, differ in phenotypic properties, including transient quiescence, self-renewal, adaptation to hypoxia, and resistance to therapy-induced DNA damage. For this reason, the development of new personalized treatment strategies for GB represents both a preclinical and clinical challenge [8]. To date, the causes and physiopathology of GB are unknown. Current therapy of choice consists of surgical resection or biopsy, followed by radiotherapy and concomitant chemotherapy with temozolamide (TMZ). TMZ, the gold standard anticancer drug for the treatment of GB, belongs to the class of orally administered alkylating agents and improves prognosis by increasing median patient survival. The approved standard therapy consists of a daily dose of 150 to 200 mg per square meter of body surface area for 5 days of each 28-day cycle. Daily therapy at a dose of 75 mg per square meter for up to seven weeks is safe; at these doses, TMZ depletes the DNA repair enzyme MGMT, resulting in tumor tissue shrinkage, an effect associated with longer survival among glioblastoma patients [9,10]. Considering the aggressiveness of GB and the low efficacy of therapeutic strategies, it is currently necessary to identify new therapeutic targets able to reduce or arrest the progression of GB. GB aggressiveness appears to be related to the presence of tumor stem cell populations called GSCs that contribute to GB malignancy by promoting tumor growth, angiogenesis, and therapeutic resistance. Unlike well-differentiated tumor cells, which show limited replicative potential, GSCs can proliferate indefinitely and spread to tissues and organs distant from the primary tumor site, becoming responsible for initiation, growth of metastases and resistance to therapy. According to some studies, the preferential overexpression of nicotinamide N-methyltransferase (NNMT) in GSC, a cytosolic enzyme involved in the biotransformation of many xenobiotics, causes the exhaustion of the methyl donor S-adenosyl methionine (SAM), with consequent hypomethylation of the GB DNA, causing the translation of the tumor towards a mesenchymal phenotype and accelerating its growth [11,12]. The main mechanism of resistance to TMZ therapy is related to the overexpression of O-6-methylguanine-DNA methyltransferase (MGMT), a gene located on chromosome 10q26 that codes for a DNA repair protein, which removes alkyl groups from the O-6 position of guanine, an important alkylation site. High levels of MGMT in tumor cells, therefore, induce the formation of a resistant phenotype by reducing the efficacy of alkylating agents such as TMZ, thus leading to therapeutic failure. MGMT in its methylated and therefore inactive form represents a molecular marker of clinical relevance, associated with the response to alkylating chemotherapy and survival of patients with GB [13–15].

Peroxisome proliferator-activated receptors (PPARs) are ligand-inducible transcription factors that belong to the superfamily of proteins called nuclear hormone receptors (Nrs), to which steroid, thyroid, and retinoid receptors also belong. Three isoforms of PPAR have been identified: PPAR-α, PPAR-β/δ, and PPAR-γ, differently expressed based on the physiological role, tissue distribution, and specificity of the ligands. Each of the isoforms activates or suppresses different genes involved primarily in the metabolism and homeostasis of fats and carbohydrates, as well as in proliferation and cell differentiation, inflammation, and cancer [16].

All three isotypes of PPAR are co-expressed in the central nervous system (CNS), but their function in this tissue is still poorly understood. Some studies show that, at the level of CNS, PPARs are involved in lipid metabolism, neuronal differentiation, and death, as well as inflammation and neurodegeneration. Observations both in vitro and in vivo show that PPAR-β/δ is the prevalent isoform in neurons of different brain areas, while PPAR-α is expressed at very low levels, predominantly in astrocytes, and appears to be involved in the neurotransmission of excitatory amino acids [17]. The expression of PPAR-γ in the brain has been extensively studied in relation to inflammation and neurodegeneration [18].

Although PPAR receptors are known for their role in lipid metabolism and glucose homeostasis, a lot of research in the literature has demonstrated the contribution of these receptors in tumors and GB biology [19,20].

Preclinical and clinical studies have shown beneficial effects of PPAR agonists against GB growth, inhibiting the invasion and motility of glioma cells and thus increasing the chance of survival [21]. The purpose of this review is to summarize the role of PPARs in GB, focusing on the antitumor action of their synthetic and natural ligands, in order to consider them as potential additional treatments to conventional therapies.

#### **2. Peroxisome Proliferator-Activated Receptors**

Peroxisome proliferator-activated receptors (PPARs) are ligand-activated transcription factors involved in various processes at the cellular level and in the regulation of lipid, carbohydrate, and amino acid metabolism. PPARs belong to the superfamily of nuclear hormone receptors (Nrs), which after interacting with specific ligands (synthetic or nonsynthetic), translocate to the nucleus where they modify their conformation and regulate gene transcription through the differential recruitment of cofactors and enzymes modifying the histone [22]. Once translocated to the nucleus, PPARs interacts with retinoid X receptors (RXR), peroxisome proliferator-activated receptor gamma-coactivators (PGC), steroid receptor coactivators (SRC), and CREB binding protein (CBP/p300), then bind to the sequences of the peroxisome proliferating receptor element (PPRE) and consequently initiate the transcription of target genes involved in various physiological processes [23]. PPARs, in fact, control gene expression involved in energy homeostasis, lipid metabolism, and adipogenesis: they represent the main receptors for dietary fats such as oleic and linolenic acids and for many lipid metabolites, for example, prostaglandin J2, 8S-hydroxyheicosatetraenoic acid (8-HETE), and oxidized phospholipids. The altered expression of PPARs is also related to the onset of many diseases such as type 2 diabetes, dyslipidemia, obesity, atherosclerosis, and metabolic syndrome [21]. PPARs are also expressed in the cardiovascular system (endothelial cells, vascular smooth muscle cells, monocytes, and macrophages), and many clinical and preclinical studies have shown the significant role of PPARs also in cardiovascular diseases [24]. Today, increasing attention has been paid to the critical role of PPARs in inflammation and cancer; numerous studies have, in fact, highlighted the overexpression of PPARs in many human solid tumors [25].

Studies based on X-ray crystallography and molecular modeling have shown that the structure of PPARs consists of six functional domains, from A to F [26]. The N-terminal portion of PPARs exhibits the ligand-independent transactivation domain (or A/B domain), also called activation function 1 (AF-1) responsible for transcriptional activation, followed by the C domain, also called DNA binding domain (DBD), involved in recognition of the DNA sequence in the promoter region of genes known as peroxisome proliferator response element (PPRE). The C-terminal of the PPAR receptor, on the other hand, contains the D domain, which thanks to its flexibility acts as a docking site for the cofactors and the E/F (or LBD) domain, which is responsible for the specificity of the ligand and for dimerization of the receptor with the retinoid X receptors (RXR). The dimerization domain is critical for the formation of heterodimers with the retinoic acid receptor α (RXRα), an important prerequisite for PPARs to bind DNA in regions containing the DNA sequence AGGTCANAGGTCA. The C-terminal also possesses the AF-2 activation domain, that after binding with the ligand, synergizes with AF-1 and undergoes conformational modifications allowing the recruitment of co-activating proteins p300, CREB binding protein (CBP), or coactivator steroid receptor 1 (SRC1), important for the transcriptional activation of their target genes (Figure 1) [27].

**Figure 1.** Schematic representation of the structure of PPAR receptor.

#### *PPARs Isoforms: Tissue Distribution and Biological Activity*

The PPAR family includes three different subtypes which differ in terms of tissue distribution, affinity for ligands, and biological activity. They are called PPAR-α, PPARβ/δ, and PPAR-γ. All isoforms participate differently in lipid homeostasis and glucose regulation (energy balance), but each of them is capable of suppressing or activating different genes [28].

Based on their tissue distribution, PPAR-α receptors are mainly expressed in brown adipose tissues, skeletal muscle, kidney, heart, liver, and intestinal mucosa and are involved in glucose metabolism and homeostasis and in the oxidation of fatty acids [29]. The PPAR-α receptor is activated by natural ligands, including saturated, monounsaturated, and polyunsaturated fatty acids and their metabolites such as 8S-HETE and 8-HEPE, leukotrienes B4 (LTB4), oxidized phospholipids, and lipolytic lipoprotein products. Among these omega-3 fatty acids, being highly polyunsaturated, oxidize easily and stimulate PPAR-α, causing a decrease in lipid levels and the elimination of triglycerides from the plasma with a consequent increase in the levels of high-density lipoprotein cholesterol (HDL) and reduction of inflammation and arteriosclerosis in the cardiovascular system. Furthermore, an important anti-inflammatory effect also derives from the inhibition of the oxidation of omega-3 fatty acids, mediated by NF-κB in a PPAR-α-dependent pathway [30]. Synthetic PPAR-α ligands are represented by fibrates (clofibrate, gemfibrozil, fenofibrate, and bezafibrate), a class of lipid-lowering drugs that are used in the treatment of hypertriglyceridemia. Through the activation of PPAR-α, they cause an increase in gene expression involved in the β-oxidation of fatty acids leading to the reduction of triglyceride-rich lipoproteins in the serum and to the increase of HDL cholesterol, slowing the progression of arteriosclerosis and reducing cardiovascular events (Figure 2) [31].

The PPAR-β/δ receptor consists of 441 amino acids, with a molecular weight of 49.9 kDa and is ubiquitously expressed in almost all tissues, including the liver, intestines, kidneys, abdominal fat, skeletal muscle, brain, and pancreas. Like the other members of the PPAR family, it mainly intervenes in the metabolism of lipids, participating in the oxidation of fatty acids, both at the level of adipose tissue, reducing adiposity, and consequently preventing the development of obesity, both at the level of skeletal muscles and heart and regulating the concentrations of cholesterol and blood glucose [32]. Moreover, many studies have revealed a large expression of PPAR-β/δ in the central nervous system (CNS), in particular at the level of neurons, astrocytes, oligodendrocytes and in microglia cells, suggesting the role of these receptors as targets for neuroinflammation and neurodegeneration [33]. Although PPAR-β/δ has a smaller binding domain (LBD) than other members of the PPAR family, it has the ability to bind many endogenous ligands, but with relatively low selectivity. Natural ligands include polyunsaturated fatty acids (arachidonic and linoleic acids) and their metabolites like prostacyclin PGI2, 13S-hydroxyoctadecadienoic acid (13S-HODE), and 15S-hydroxyheicosatetraenoic acid (15S-HETE), which could have promising applications in cardiomyopathy diabetic. Synthetic agonists (GW501516, GW0742, L-165041, and MBX-802) that have been developed and proposed as treatments for obesity and metabolic syndrome have not been used for clinical trials due to their carcinogenic effects, so none of them have been approved for clinical use to date (Figure 2) [34]. Despite controversial data on its pro-tumorigenic versus anti-tumorigenic action, few results suggest that PPAR-β/δ activation can reduce the growth of neuroectodermal tumors, including glioblastomas [35].

**Figure 2.** Classification of natural and synthetic ligands of PPARs receptors.

The PPAR-γ receptor is expressed in many human tissues and has three different isoforms: PPAR-γ1, γ2, and γ3. PPAR-γ1 is ubiquitously expressed in all human cells, PPAR-γ2 has been predominantly detected in white and brown adipose tissue, as well as in the large intestine and spleen, while PPAR-γ3 expression is limited. PPAR-γ is poorly expressed in the central nervous system (CNS), but it was found in different cell types such as neurons, astrocytes, oligodendrocytes, and microglia. Its physiological role includes the regulation of adipogenesis and the levels of adipokines such as adiponectin, TNFα, MCP-1, and resistin, and it is also involved in the energy balance and lipid biosynthesis. Thanks to the size of its binding cavity, the PPAR-γ receptor is able to bind a large variety of natural or synthetic lipophilic acids. The natural modulators of PPAR-γ are mainly unsaturated fatty acids and their metabolites, including 15- hydroxyeicosatetraenoic acid (15-HETE), 9- and 13-hydroxyoctadecadienoic acid (9/13-HODE), 15-deoxy- ∆12,14-prostaglandin J2 (15-d-∆12,14-PGJ2), and prostaglandin PGJ2, whose physiological role is still to be clarified [36]. Synthetic PPAR-γ ligands belong to the thiazolidinediones class (TZD), including troglitazone, rosiglitazone, ciglitazone, and pioglitazone, and they are known as insulin-sensitizing drugs. By activating the PPAR-γ receptor, in fact, they reduce the hepatic production of glucose and prolong the function of pancreatic cells, preventing apoptosis of β cells (Figure 2). TZD are used in the treatment of type 2 diabetes, as they increase insulin sensitivity and improve glucose control, but their use is limited by important adverse events, including the risk of bone fracture and congestive heart failure [37] (Figure 3).

**Figure 3.** A schematic representation at cellular level of PPARs signaling.

#### **3. Role of PPARs in Tumors**

Certain PPAR-related metabolic alterations, such as obesity and type 2 diabetes, have been identified as risk factors for cancer cell proliferation and thus tumor progression. Hence, research is currently focused on using PPARs as targets for cancer therapy, and a few studies have focused on understanding their role in human cancer and in the antitumor activity of their natural and synthetic agonists. Furthermore, current studies have revealed conflicting results on the role of the different isoforms of PPARs in various tumor types; most investigations have shown that PPAR-β/δ activation is linked to tumor progression, while PPAR-α and PPAR-γ are associated with an antitumor action [38]. Some in vitro studies on breast cancer cells SUM149PT and SUM1315MO2 have shown interesting results in the context of nuclear PPAR-α receptor signaling, demonstrating that its activation by clofibrate agonist suppresses the inflammatory activity of cyclooxygenase-2 (COX-2) and 5-lipoxygenase (5-LO) and determines the decrease in the secretion of prostaglandin-E2 (PGE2) and leukotrienes-B4 (LB4), effectively inhibiting cell survival and cell cycle-related kinases [39]. Another study conducted in vivo showed that the combination of clofibric acid (PPAR-α agonist) and pioglitazone (PPAR-γ agonist) reduces angiogenesis, induces apoptosis, significantly decreasing the expression of COX-2 and VEGF, through inhibition of activator protein-1 (AP-1) and suppressing the growth of solid ovarian tumors [40]. Evidence suggests that certain PPAR-α ligands, including bezafibrate and fenofibrate, may act as potential chemopreventive agents in colon carcinogenesis by reducing intestinal polyp formation in Apc-deficient mice, by inhibiting AOM/DSS-induced colon carcinogenesis. However, the exact mechanism by which PPAR-α activation suppresses colon carcinogenesis is still unclear [40]. Moreover, only a few studies support the pro-carcinogenic role of PPAR-α. For example, some investigations carried out on human breast cancer cell lines (MDA-MB-231) have shown that inhibition of PPAR-α by the antagonist GW6471 leads to an impairment of the mevalonate pathway and a substantial reduction in cholesterol and lipid droplets. This causes a consequent perturbation of lipid metabolism and cell death, influencing the pathways involved in the control of proliferation, such as the pathways involving the Rho Family and YAP/TAZ and Wnt/β-catenin signaling [41]. The conflicting results presented in the literature regarding the action of PPAR-α on tumor

progression could be due to the different evaluation conditions, cell lines used, the stage of differentiation, the cellular context, and the microenvironment.

The role of PPAR-β/δ on cell proliferation, induction of angiogenesis, and cell death has been extensively investigated. Although there are conflicting opinions on the effects of PPAR-β/δ activation for cancer progression, most reports suggest that its stimulation could have pro-tumorigenic effects [42]. PPAR-β/δ exerts proangiogenic effects, directly or indirectly modulating downstream proinflammatory or proangiogenic molecules that act on multiple cell types in the tumor microenvironment, promoting cancer progression and metastasis [43]. The increased expression of PPAR-β/δ mRNA in colon cancers has been attributed to APC-β-catenin-TCF4-mediated transcription, similar to the well-known target gene β-catenin-TCF4 CCND1, which encodes cyclin D1. This supports the hypothesis that PPAR-β/δ regulates genes that increase cell proliferation and promote colon carcinogenesis [44]. Consistent with these results, PPAR-β/δ has been shown to strongly potentiate aberrant activation of β-catenin in mouse genetic models of human CRC, with representative APC mutations and overexpression or deletion of PPAR-β/δ in intestinal epithelial cells (IEC), activating pro-invasive pathways to promote CRC tumorigenesis. These results have demonstrated that PPAR-β/δ strongly accelerates APC mutation-driven CRC progression and invasion through multiple important pro-tumorigenic pathways, including BMP7/TAK1/-catenin, PDGFRβ, AKT1, EIF4G1, and CDK1 [45]. The role of PPAR-β/δ in enhancing cell proliferation was supported by a further study performed on human liposarcoma cells (SW872, T778), in which an increase in cell proliferation was observed in response to PPAR-β/δ activation by the agonist GW0742, which appears to be caused by leptin repression, suggesting the potential therapeutic use of PPAR-β/δ antagonists for the treatment of unresectable liposarcomas [46]. PPAR-β/δ seems to be overexpressed in breast cancer cells and the elevated levels appear to correlate with greater migratory and metastatic properties. PPAR-β/δ mediates these effects by mechanisms including increased expression of antioxidant proteins such as catalase and increased AKT-mediated survival signaling after prolonged nutrient deprivation [47].

Among the three isoforms, PPAR-γ is certainly the most studied in tumors. It is, in fact, expressed in a wide variety of cancers and its role in cancer initiation/progression has long been debated. The literature suggests that PPAR-γ plays a key role in tumorigenesis as a tumor suppressor. Indeed, PPAR-γ activation by many agonists has shown antiproliferative and proapoptotic actions in colon, esophageal, thyroid, breast, lung, and prostate cancers [48–50]. The mechanism by which it induces tumor cell growth arrest appears to be related to the PPAR-γ-dependent upregulation of the tumor suppressor gene and of the homologous tensin phosphatase (PTEN), which inhibits the phosphorylation of PI3-kinase and AKT by reducing cell migration and proliferation. To the antiproliferative effects is also added the downregulation of the anti-apoptotic protein B-cells/lymphoma 2 (Bcl-2), the anti-angiogenic activity through the inhibition of VEGF and its receptors in various cells and anti-inflammatory properties through NFκB-mediated inhibition of gene transcription. Furthermore, PPAR-γ appears to hinder the formation of metastases through the inhibition of the epithelial–mesenchymal transition (EMT), a process by which epithelial cells lose their cell polarity and cell–cell adhesion, and gain migratory and invasive properties. Consistent with these findings, growing evidence suggests that PPAR-γ overexpression has important suppressive activities in colorectal cancer growth. In fact, preclinical studies currently analyze new PPAR-γ agonists, capable of inhibiting the Wnt/β-catenin pathway, acting as modulators of PPAR-γ signaling, and interfering with related pathways in order to provide new therapies for CRC [51]. PPAR-γ thiazolidinediones (TZD) ligands have been shown to counteract the stimulatory effects of leptin on breast cancer growth in vivo and in vitro models. The results show that PPAR-γ activation inhibited cell proliferation and prevented the development of leptin-induced MCF-7 tumor xenografts [52].

PPAR-γ expression was studied in patients with esophageal squamous cell carcinoma (ESCC), on which the antiproliferative effect and mechanism of action of the PPAR-γ agonist, ephatutazone, were investigated. Ephatutazone has been shown to cause a 49.6% reduction in the proliferation of xenotransplanted ESCC cells through a mechanism that involves the regulation of p21Cip1 protein levels in the nucleus by inactivating the Akt signal and dephosphorylating p21 to Thr145, without altering the transcriptional activity of p21Cip1 [53]. Another study found a significant increase in PPAR-γ expression in NSCLC cell lines by both immunohistochemistry and Western blotting. In addition, significant antitumor activity was observed both in vivo and in vitro in response to treatment with troglitazone or pioglitazone, with a correlated reduction in metastases [54].

Although many results strongly support the role of PPAR-γ as a tumor suppressor, other studies, on the contrary, argue that it plays a role as a tumor promoter. Indeed, a recent animal experiment of prostate orthotransplantation identified the involvement of PPAR-γ in the upregulation of the AKT3-PGC1α axis. PPAR-γ seems to increase the regulation of AKT3, destabilizing CRM1 and favoring the localization of PGC1α in the nucleus with a consequent increase in mitochondrial function and ATP levels. The high levels of ATP appear to be related to promoting tumor growth and metastasis [55]. Further evidence reveals that PPAR-γ activation reduces TXNIP expression in human melanoma cells (A375 and C8161), affecting the expression of proteins of particular relevance to melanoma cell invasiveness, such as integrin alpha-v/beta-3 and TIMP-2, resulting in melanoma progression to a metastatic phenotype [56].

#### *3.1. Role of PPAR-α Agonists in GB*

Conflicting results emerged regarding the role of PPAR-α in GB tumorigenesis. Studies have examined the expression of PPAR-α protein and PPAR-α mRNA in primary wild-type human IDH1 GB, arguing that their overexpression in GB is related to the degree of glioma malignancy [57]. This overexpression appears to be accompanied by the significant increase in 30-hydroxy-30-methylglutaryl-CoA reductase (HMGCR), the cholesterol biosynthesislimiting enzyme (CHO), that catalyzes the formation of mevalonate (MVA). These results indicate that PPAR-α, by regulating CHO metabolism, is involved in the strong alteration of lipid homeostasis observed in gliomas and could therefore drive the tumorigenesis process. In fact, the use of a compound derived from N-phenylsulfonyl (AA452), capable of blocking the activation of PPAR-α, has determined a strong effect on cell viability, reducing cell proliferation and migration and therefore decreasing tumor invasiveness [58]. Other evidence discusses that the expression of PPAR-α receptors is negatively correlated to the degree of malignancy of the glioma, in fact, its activation suppresses the proliferation of tumor cells, delays the cell cycle to the G1 phase and induces apoptosis and the accumulation of species reactive oxygen (ROS) in U87 cells [59,60]. The anticancer effects of the PPAR-α agonist fenofibrate have been demonstrated in several cell lines of colon, breast, endometrial, skin, medulloblastoma, and melanoma cancers [61,62]. Among the PPAR-α ligands tested in GB, fenofibrate received the most attention, due to its capacity to cross the blood brain barrier (BBB) and has an established anti-inflammatory activity and limited toxicity and a better side effect profile [63]. In order to activate the PPAR-α receptor, fenofibrate must first be converted into fenofibric acid (FA) by blood and tissue esterases. Once converted FA then binds to the PPAR-α receptor and triggers the expression of numerous metabolic enzymes involved in the oxidation of fatty acids and reduces glucose uptake by repressing the insulin-dependent glucose transporter GLUT4. This metabolic switch could explain the mechanism by which fenofibrate initiates a gradual decline in energy metabolism in cancer cells. However, the anticancer effects of fenofibrate are more pronounced than other PPAR agonists and may be due to its accumulation in the mitochondrial fraction of human GB cells, which respond with a sudden and severe inhibition of mitochondrial respiration and an immediate increase but transient glycolysis, effectively triggering an energy catastrophe in GB cells with significantly reduced toxicity in normal astrocytes [64]. The anticancer effects of fenofibrate are therefore very complex and cannot be explained simply by activation of PPAR-dependent transcription, but it is, therefore, necessary to consider PPAR-independent mechanisms. Some investigations have also confirmed that fenofibrate is a neuroprotective agent; the results obtained in vitro on high-grade glioma

(HGG) cell lines U87 and U343 (p53 wild-type), U251 and T98 (p53 mutant), confirmed the antiproliferative and pro-apoptotic effects of fenofibrate demonstrating inhibition of NF-κB expression, cyclin D1, and Akt (Figure 4). Akt is well established as a therapeutic target for HGG, and there are a number of Akt inhibitors being evaluated as an adjunct treatment for HGG [65]. The transcription factor Fork-head box O1 (FoxO1) is an Akt substrate that plays a key role in tumor suppression by promoting transcriptional activation of p27kip. A recent study showed that fenofibrate by activating the PPARα/FoxO1/p27kip pathway could actually induce the death of human U87MG glioma cells, causing cell cycle arrest in the G0/G1 phase [66]. The activation of the PPAR-α receptor by the agonist fenofibrate attenuates the signaling responses of the IGF-IR, a factor that helps to support malignant growth and invasion of glioma cells and causes the accumulation of reactive species of oxygen (ROS), loss of mitochondrial membrane potential and a deficit in ATP production, which together may explain the severe impairment of glioma cell motility [67]. A recent study demonstrated that fenofibrate modulates the expression of hypoxia-inducible factor-1 alpha (HIF-1α), an overexpressed transcription factor under hypoxic conditions in human GB samples. HIF-1α is involved in the transactivation of genes involved in the altered metabolism, causing the accumulation of metabolites in the tumor environment, thus contributing to the growth and increase of the aggressiveness of GB. Fenofibrate inhibits the expression of HIF-1α by activation of HO-1 via the AMPK pathway. Furthermore, the activation of HO-1 involves the upregulation of SIRT1, causing its translocation into the nucleus, with consequent deacetylation of HIF-1α and inhibition of transcriptional activity [68].

**Figure 4.** An exemplified overview of fenofibrate's PPAR-dependent and PPAR-independent mechanisms of action in GB tumoral cells.

A dual agonist PPAR-α/PPAR-γ, also called TZD18, inhibited cell growth and induced apoptosis in human GB T98G cells, through the activation of caspase-3 and down-regulation of the expression of Bcl-2, suggesting that TZD18 may have a therapeutic role in the treatment of human GB [69].

Moreover, PPAR-α and PPAR-γ agonists have the ability to selectively upregulate catalase expression on human astrocytes. In fact, both C6 GB cells and normal astrocytes were treated with PPAR-α and PPAR-γ agonists, showing a significant increase in catalase expression only in normal astrocytes, while, on the contrary, they failed to increase catalase expression in glioma cells. These results are promising because current data support the concept that selective manipulation of catalase gene expression and/or activity can provide greater protection of astrocytes from H2O2-induced damage and consequently can improve normal tissue survival during radiotherapy [70].

#### *3.2. Role of PPAR-β/δ in GB*

Although PPAR-β/δ agonists have been shown to cross the blood–brain barrier and modulate oxidative stress and proinflammatory responses associated with acute and chronic CNS disorders, overall, there are only a few studies evaluating the action of the PPAR-β/δ receptor in brain tumors [71]. The effects of a PPAR-β/δ ligand, namely erucic acid (EA), an omega-9 fatty acid, were investigated. EA has been shown to block the growth of C6 glioma cells and also reduce the cardiotoxicity of doxorubicin, thus suggesting that the combination of systemic EA with DOX-chemotherapy can reduce DOX concentrations in the systemic circulation, hinder toxic interactions and induce selective killing of glioma cells [72]. However, these results are too small to evaluate the role of the receptor PPAR-β/δ in brain tumors, especially considering the prevalence of the results in the literature that instead supports a procarcinogenic action in other types of tumors.

#### *3.3. Role of PPAR-γ Agonists in GB*

In addition to well-defined metabolic actions, PPAR-γ agonists exhibit various antineoplastic effects and induce cell death by apoptosis in various brain tumor cell lines. There are several possible mechanisms by which PPAR-γ agonists inhibit cell proliferation, such as induction of cell cycle arrest in the G0/G1 phase, a reduction in MYC levels upstream of the S phase transition, as well as possible down-regulation of CCND1 (cyclin D1) and associated cyclin-dependent kinases, but also the upregulation of cyclin-dependent kinase inhibitors CDKN1A, CDKN1B, and CDKN2B [21]. PPAR-γ agonists also blocking the Janus kinase/signal transducer and transcription activator (JAK/STAT) pathway, inhibit the expansion of CD133 + brain tumor stem cells (BTSC is also called tumor-initiating brain cells). In vivo and in vitro models have shown that inhibition of JAK2 (upstream regulator of STAT3) by the agonist PPAR-γ troglitazone promotes the slowing of the progression of GB disease, causing the phosphorylation of STAT3 tyrosine 705 and leading to the down-regulation of CCND1 and BCL2L1 (B-cell lymphoma extra-large protein 2) [73].

Agonists PPAR-γ, PGJ2, and rosiglitazone, have been shown to inhibit the proliferation of GB cell lines (U87-MG) through G2/M arrest and promoting the induction of programmed cell death [74]. These results are consistent with another study, in which evaluated growth inhibition and induction of apoptosis by another TZD, ciglitazone. Ciglitazone in addition to inducing cell cycle arrest, causes a significant reduction in the activity of telomerase, an enzyme that is constitutively active in most tumor cells, in human GB cell lines U-87 MG and U-118 MG [75]. Other data suggest that ciglitazone is able to induce PPAR-γ-independent apoptotic cell death in human T98G glioma cells by down-regulation of Akt and reduction of mitochondrial membrane potential (MMP), an effect that was accompanied by a down-regulation of Bcl-2 expression and an increase in Bid cleavage [76]. A study shows that the PPAR-γ receptor is an important positive regulator of the expression of CIDEA, a member of cell death-inducing DFFA-like effector (CIDE) protein family. In fact, it has been shown that PPAR-γ inhibition improves CIDEA expression, triggering glioma cell apoptosis and a decrease in HIF-1α activation, justifying further investigations aimed at evaluating the efficacy of PPAR-γ inhibitors as an effective anti-glioma therapeutic strategy [77]. Another TZD, Pioglitazone, showed anticancer efficacy on human glioma cells (U87MG, T98G, and U251MG) in vitro. Pioglitazone-induced inhibition of glioma cell proliferation and invasion occurred in a PPAR-γ-dependent manner and is in agreement with its ability to dramatically reduce β-catenin expression and transcriptional activity, resulting in decreased cell proliferation, migration, and apoptosis. These results indicate that PPAR-γ activation induces suppression of glioma cell turnover [78]. The agonist PPAR-γ Pioglitazone has also been shown to increase the functional expression of the glutamate

transporter EAAT2 in glioma cells, preventing excitotoxic damage and glutamate-mediated seizures related to glioma [79]. In addition to the antineoplastic and anticonvulsant effects of pioglitazone, there is also the demonstration of its ability to cross the blood–brain barrier after oral and intracerebral administration in a human glioma xenograft model, suggesting its possible use as an additional candidate in the current regimen for double mechanistic efficacy in subtherapeutic doses to avoid associated adverse effects [80].

Clinical studies suggest a potential protective effect of the PPAR-γ agonists pioglitazone or rosiglitazone in diabetic patients with GB. The results show that diabetic patients with GB who had been treated with PPAR-γ agonists showed an increase in median survival of 19 months compared to patients who received standard treatment only [81]. In addition, a Phase 1 clinical study conducted on patients with primary and metastatic brain tumors treated with radiotherapy highlighted the protective role of pioglitazone in the prevention of radiation-induced cognitive decline (RICD) and its good tolerability at the 45 mg dose, not showing dose-limiting toxicity (DLT), which may be suggested for efficacy studies [82] (Table 1).


**Table 1.** This table summarizes the studies that analyze the effects of PPAR ligands.

#### **4. Conclusions and Future Prospects**

In this review, a detailed analysis was carried out in order to summarize the role of the PPAR receptor family in GB, a brain tumor characterized by high aggression. The PPARs family includes three different subtypes PPAR-α, PPAR-β/δ, and PPAR-γ, which differ in terms of tissue distribution, affinity for ligands, and biological activity, and which participate differently in the maintenance of lipid and glucose homeostasis [83]. Their role in the progression and differentiation of cancer cells in different types of solid tumors is also widely studied, even if often the data present in the literature report conflicting opinions. Nevertheless, most investigations have shown that PPAR-β/δ activation is related to tumor progression, whereas PPAR-α and PPAR-γ are associated with antitumor action [84]. Based on these findings, considerable interest has been shown in PPAR ligands as potential therapeutic agents in the treatment of gliomas; however, the molecular mechanisms underlying the suppression of carcinogenesis in gliomas, determined by PPAR activation have not yet been fully elucidated. Particularly among the fibrates, a PPAR-α ligands, fenofibrate has received the most attention due to its capacity to reduce the proliferation of GB cells through both PPAR-dependent and PPAR-independent mechanisms [85]. The thiazolidinediones (TZD) class, PPAR-γ ligands known as insulin-sensitizing drugs, including troglitazone, ciglitazone, rosiglitazone, and pioglitazone have also been shown to interact with several pathways involved in the induction of cell death in GB cells [86]. On the other hand, little attention has been paid to the PPAR-β/δ ligands, probably due to the conflicting evidence in the literature regarding its pro-carcinogen action, so further studies would be needed to clarify its function in this context. Finally, there are clinical studies aimed at evaluating the efficacy and safety of these ligands in patients with GB, but further results

are certainly needed in order to be able to suggest PPAR ligands as potential treatments in the therapy of GB.

**Author Contributions:** R.B. drew up the manuscript; I.P., M.C. and E.E. were involved in the design and intellectual concept of the study; G.C. (Giulia Chisari), S.M., L.C. (Lorenzo Colarossi) and L.C. (Laura Cucinotta) performed the literature search; M.L. and G.C. (Giovanna Casili) supervised the study; I.P. and E.E. critically revised the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

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

#### **References**


## *Article* **PPAR**γ **Regulates Triclosan Induced Placental Dysfunction**

**Jing Li 1,2,\* ,†, Xiaojie Quan 1,2,†, Yue Zhang 1,2,†, Ting Yu <sup>1</sup> , Saifei Lei <sup>3</sup> , Zhenyao Huang <sup>1</sup> , Qi Wang <sup>1</sup> , Weiyi Song <sup>1</sup> , Xinxin Yang <sup>1</sup> and Pengfei Xu 3,4,\***


**Abstract:** Exposure to the antibacterial agent triclosan (TCS) is associated with abnormal placenta growth and fetal development during pregnancy. Peroxisome proliferator-activated receptor γ (PPARγ) is crucial in placenta development. However, the mechanism of PPARγ in placenta injury induced by TCS remains unknown. Herein, we demonstrated that PPARγ worked as a protector against TCS-induced toxicity. TCS inhibited cell viability, migration, and angiogenesis dose-dependently in HTR-8/SVneo and JEG-3 cells. Furthermore, TCS downregulated expression of PPARγ and its downstream viability, migration, angiogenesis-related genes *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, *MMP-9*, and upregulated inflammatory genes *p65*, *IL-6*, *IL-1β*, and *TNF-α* in vitro and in vivo. Further investigation showed that overexpression or activation (rosiglitazone) alleviated cell viability, migration, angiogenesis inhibition, and inflammatory response caused by TCS, while knockdown or inhibition (GW9662) of PPARγ had the opposite effect. Moreover, TCS caused placenta dysfunction characterized by the significant decrease in weight and size of the placenta and fetus, while PPARγ agonist rosiglitazone alleviated this damage in mice. Taken together, our results illustrated that TCS-induced placenta dysfunction, which was mediated by the PPARγ pathway. Our findings reveal that activation of PPARγ might be a promising strategy against the adverse effects of TCS exposure on the placenta and fetus.

**Keywords:** triclosan; PPARγ; placenta toxicity; cell migration; angiogenesis; inflammation

### **1. Introduction**

Triclosan (TCS) is a synthetic broad-spectrum antimicrobial and exposure mainly occurs in dermal application (soaps, hand sanitizers, toothpaste, cosmetics, antiperspirants, and bedclothes) and oral use of consumer products (water, food products) [1–3]. TCS, as a kind of exogenous biological signal termed Endocrine disruptors (EDCs), can mimic endogenous estrogenic hormones, and interferes with the maintenance of homeostasis and the regulation of developmental processes [4]. Previous epidemiological research demonstrated that TCS has been distinguished in mothers' milk (1–13.6 ng/mL), urine (2.5–107 ng/mL), and cord blood samples [5–7]. In addition, high urinary TCS levels in patients with spontaneous abortion have been reported [8]. A Denmark Odense Child Cohort study stated that median unadjusted urinary TCS was 0.88 ng/mL in pregnant women, and high maternal urinary TCS levels were associated with reduced head and abdominal circumference at birth [9]. Animal experiments suggested that the exposure of

**Citation:** Li, J.; Quan, X.; Zhang, Y.; Yu, T.; Lei, S.; Huang, Z.; Wang, Q.; Song, W.; Yang, X.; Xu, P. PPARγ Regulates Triclosan Induced Placental Dysfunction. *Cells* **2022**, *11*, 86. https://doi.org/10.3390/ cells11010086

Academic Editors: Kay-Dietrich Wagner and Nicole Wagner

Received: 9 November 2021 Accepted: 24 December 2021 Published: 28 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

pregnant mice to TCS reduced fetal body weight and viability [8]. TCS has been detected with a concentration of up to 478 ng/L and 1329 ng/g in surface waters and sediment, respectively, in China rivers [10]. Therefore, a further understanding of the toxicity of TCS is important for the inhibition of TCS pollution.

The nuclear receptors, peroxisome proliferator-activated receptors (PPARs) are activated by binding natural ligand-inducible transcription factors such as glitazones [11,12]. PPARs are responsible for metabolism and cell function and have three subtypes including PPARα, PPARβ/δ, and PPARγ [13,14]. PPARγ is highly expressed in reproductive tissues (ovary, testis, uterus, prostate, mammary gland) and the trophoblast labyrinthine zone in rodent placentas and human placentas [15]. PPAR has been suggested to be involved in regulating cell trophoblast proliferation, inflammatory reactions oxidative response, and nutrient transport to mediate placenta development [16–18]. Additionally, research has illustrated that PPARγ agonists enhanced, while PPARγ antagonists reduced, proliferative and migratory capabilities of endothelial cells [19]. The vasculature defects were observed in placentas at embryonic GD 9.5 in PPARγ-null mice [16]. Similar animal experiments implicated embryonic lethality induced by defects in placental vascularization in PPARγnull mice [20,21]. The downregulated protein level of PPARγ was associated with placental disorders [22]. PPARγ expression levels were also found to be decreased in pregnant mice or zebrafish when exposed to TCS [23,24].

The effect of prenatal TCS exposure on placental toxicity and its role in fetal growth and development are still unclear. Whether the effect of prenatal TCS exposure on placental toxicity is associated with PPARγ remains to be revealed. Our research aims to clarify the potential character of prenatal TCS exposure on placenta function and its potential mechanism of PPARγ in the process of placental exposure to TCS.

#### **2. Materials and Methods**

#### *2.1. Reagents*

TCS (CAS No. 3380-34-5, purity > 98% pure) was obtained from Sigma–Aldrich (Oakville, ON, Canada). DMEM/F12 medium and MEM medium were purchased from KeyGEN BioTECH (Jiangsu, China). RPMI 1640 medium was purchased from GIBCO (Grand Island, N.Y.). Dimethyl sulfoxide (DMSO), fetal bovine serum (FBS), the Cell Counting Kit-8 (CCK-8), and corn oil used was obtained from Vicmed (Busan, Korea) and Macklin (Shanghai, China), respectively, and were analytical grades. The pcDNA-*PPARγ* vector and *si-PPARγ* were purchased from GenePharma (Shanghai, China). Moreover, the rosiglitazone (a specific PPARγ agonist) and GW9662 (a specific PPARγ antagonist) were obtained from MedChemExpress (Shanghai, China).

#### *2.2. Cell Culture and Transfection*

Thanks to Dr. Xinru Wang from Nanjing Medical University (Nanjing, China) for sending HTR-8/SVneo and JEG-3 cells as a gift to us. HTR-8/SVneo and JEG-3 cells were seeded in DMEM/F12 and MEM medium, respectively, and supplemented with 10% FBS in a 5% CO<sup>2</sup> humidified atmosphere at 37 ◦C. Based on the manufacturer's instructions, these two cell lines were transfected with small interfering RNA (siRNA) (*si-PPARγ*; 20 nM), the control siRNA (*si-Con*; 20 nM), pcDNA-*PPARγ* (2 µg), and pcDNA 3.1 (2 µg) targeting PPARγ by Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). The medium was replaced after 4 h of cell transfection. After cell transfection for 24 h, they were incubated with triclosan for 24 h. For PPARγ activation or deactivation, cells were treated with rosiglitazone or GW9662 with 10 µM for 24 h and pre-treated for 0.5 h before incubating with TCS.

#### *2.3. Animal Treatment*

Institute of Cancer Research (ICR) mice, at 10 weeks old and weighing about 30 g, were obtained under a protocol approved by Xuzhou Medical University Animal Center. On day GD 7.5, after the vaginal plug (GD 0.5) was observed in pregnant mice, they were randomly separated into five groups (*n* = 8 per group). All pregnant mice were orally managed with

0, 10, 50, and 100 mg/kg/day TCS and from GD 7.5 to GD 17.5, while control pregnant mice accepted the same volume of corn oil. Firstly, rosiglitazone was dissolved in DMSO, and the ratio of DMSO to corn oil was 1:1000, and pregnant mice were orally administered with rosiglitazone (20 mg/kg/day) together with TCS (100 mg/kg/day). Mice placentas were isolated, put into liquid nitrogen to be quickly frozen and stored at −80 ◦C. According to manufacturers of Xuzhou Medical University Animal Ethics Committee, the entire experiment was performed in a Specified Pathogen-Free (SPF) environment (protocols 201605w025, 25 May 2016 and 202106A237, 25 June 2021).

#### *2.4. Cell Viability Assay*

HTR-8/SVneo or JEG-3 cells (5 <sup>×</sup> <sup>10</sup><sup>3</sup> cells/well) were seeded into sterile plates of 96 wells in complete medium (DMEM/F12, MEM, 10% FBS). Cells were incubated and cultured with TCS of different concentrations for 24 h. An amount of 10 µL CCK-8 (Vicmed, Busan, Korea) solution was added to each well and incubated for 0.5–4 h at 37 ◦C in a 5% CO<sup>2</sup> atmosphere. The plate was detected at 450 nm on a microplate reader Spark (TECAN, Austria, 30086376).

#### *2.5. Cell Migration Assay*

HTR-8/SVneo or JEG-3 cells were grown in 6-well cell plates overnight to obtain approximately 70% confluency. After corresponding treatment, HTR-8/SVneo and JEG-3 cells were starved in 1% serum medium. Scratch wounds were made via a sterile 10 µL pipette tip to obtain three parallel lines. Multiple photographs were taken of the wound using an inverted microscope (KS400; Carl Zeiss Imaging GmbH), and the migration distance was measured by Image J analysis software (National Institutes of Health, v1.8.0). The whole wound closure distance was calculated and the distance of newly covered cells was measured at 0 and 24 h to evaluate migration.

#### *2.6. Tube Formation Assay*

Approximately 50 µL Matrigel (10 mg/mL) (BD Biosciences, San Diego, CA, USA) was decked into 96-well plates and allowed to completely solidify at 37 ◦C for 1 h to form a gel. Approximately 5 <sup>×</sup> <sup>10</sup>3/mL HUVECs were suspended in a conditioned medium (the conditioned medium was derived from HTR-8/SVneo and JEG-3 cells and added into the wells of the Matrigel solidified plate. Tube formation in each well was monitored and imaged using an inverted microscope after the plate was incubated at 37 ◦C for 5 h. We repeated the experiment three times independently. Tube branch length was measured with six duplicates per well, while average branch length was taken from three random microscopic fields per well. Each assay was done in triplicate and quantification was conducted with Image J software (National Institutes of Health, v1.8.0).

#### *2.7. Real Time PCR (RT-PCR)*

Placental tissues, trophoblast cells, and Trizol (Vicmed, Busan, Korea) were used to isolate total RNA. RNA concentration was detected by Nanodrop 2000 (Thermo, Scientific). The reverse transcription was conducted using SYBR Green qPCR SuperMix and 500 ng of the total RNA was reverse transcribed to cDNA. The quantitative RT-PCR was carried out using SYBR Green Master Mix on a 7500 fast real-time PCR System (Applied Biosystem, Foster, CA, USA) according to the manufacturer's instructions. The mRNA levels were analyzed using the comparative cycle threshold method (2−∆∆CT), and the relative levels were normalized to the level of GAPDH. The primers were designed by Sangon Biotech (Shanghai, China), and primer sequences were listed in Supplementary Table S1.

#### *2.8. Western Blot Analysis*

Total cellular protein was extracted and lysed with RIPA lysis buffer with protease inhibitors (KeyGEN, Nanjing, China) and phosphatase inhibitors (KeyGEN, Nanjing, China). Proteins were separated with 10% SDS-PAGE gels and transferred to PVDF membranes

(Merck Millipore, MA, USA). Membranes were blocked with 5% non-fat milk for 1h and then incubated with primary antibodies of PPARγ, ANGPTL4, MMP-2, IL-1β, GAPDH, and tubulin β (Bioword, Beijing, China) overnight at 4 ◦C. After being washed five times for five minutes, the membranes were incubated with the corresponding secondary antibodies for 1h at room temperature. The protein bands were visualized with an Enhanced Chemiluminescence (ECL) detection kit (Amersham, NJ, USA). The protein bands were analyzed using Image Lab Software (Bio-Rad, CA, USA).

#### *2.9. Statistical Analysis*

The above assays were carried out three independent times and all data were carried out by GraphPad Prism 8.3 software (San Diego, CA, USA), and data were exhibited as the mean ± SEM. The comparison of the data was subjected to a one-way analysis of variance among each independent group. The *t*-test or one-way ANOVA was implemented using the SPSS 19.0 (IBM, Armonk, NY, USA). *p*-value < 0.05 and *p*-value < 0.01 were considered a statistically significant difference and higher significance, respectively.

#### **3. Results**

#### *3.1. PPARγ Is Crucial in Cell Viability Inhibition Induced by TCS*

The cell viability of HTR-8/SVneo and JEG-3 cells exposed to TCS at different concentrations was determined by the CCK-8 assay. TCS dose-dependently inhibited cell viability of HTR-8/SVneo and JEG-3 cells, significantly decreasing cell viability at 20 µM, 30 µM, or 40 µM (Figure 1A). Of interest, TCS significantly inhibited *PPARγ* mRNA expression levels in those two cell lines especially HTR-8/SVneo cells (Figure 1B).

To investigate whether PPARγ was involved in the TCS-induced inhibition of cell viability, cell viability was analyzed when PPARγ was overexpressed and activation (rosiglitazone) or knockdown and inhibition (GW9662), and then exposed to TCS. Rosiglitazone and GW9662 with the concentration of 10 µM were not toxic in HTR-8/SVneo and JEG-3 cells (Figure S1A). The efficiency of PPARγ overexpression and knockdown is shown in Figure S1B–D. PPARγ mRNA and protein expression level was increased when PPARγ was overexpressed (Figure S1B,C) but decreased after knockdown compared to the control in HTR-8/SVneo and JEG-3 cells (Figure S1B,D). The results exhibited that pcDNA-*PPARγ* and rosiglitazone alleviated TCS-elicited cell viability inhibition in HTR-8/SVneo and JEG-3 cells (Figure 1C,D). In contrast, GW9662 and *si-PPARγ* aggravated the inhibition in these two cell lines (Figure 1C,E). In addition, treatment with rosiglitazone increased the expression of PPARγ-regulated genes, such as *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9*, and decreased the expression of *p65*, *IL-6*, *IL-1β*, and *TNF-α* in vitro. On the other hand, treatment with GW9662 has an opposite trend (Figure S1E).

**Figure 1.** The viability of TCS on HTR-8/SVneo and JEG-3 cells was detected by Cell Counting Kit-8 assay for 24 h. (**A**) Cell vitality in HTR-8/SVneo and JEG-3 cells exposed to TCS. (**B**) Expression of PPARγ was determined by RT-PCR in HTR-8/SVneo and JEG-3 cells exposed to indicated TCS for 24 h. (**C**) Cell vitality was detected when exposed to TCS (40 μM for HTR-8/SVneo, 30 μM for JEG-3) in the absence or presence of rosiglitazone or GW9662 in the two cell lines. PPARγ was overexpressed (**D**) and knockdown (**E**) in the two cell lines and co-treated with TCS in HTR-8/SVneo and JEG-3 cells while cell vitality was analyzed. (**F**) HTR-8/SVneo and JEG-3 cells exposed to TCS while in the absence or presence of rosiglitazone (10 μM) and GW9662 (10 μM). HTR-8/SVneo and JEG-3 cells exposed to TCS during PPARγ overexpression (**G**) or knockdown (**H**). The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3. **Figure 1.** The viability of TCS on HTR-8/SVneo and JEG-3 cells was detected by Cell Counting Kit-8 assay for 24 h. (**A**) Cell vitality in HTR-8/SVneo and JEG-3 cells exposed to TCS. (**B**) Expression of PPARγ was determined by RT-PCR in HTR-8/SVneo and JEG-3 cells exposed to indicated TCS for 24 h. (**C**) Cell vitality was detected when exposed to TCS (40 µM for HTR-8/SVneo, 30 µM for JEG-3) in the absence or presence of rosiglitazone or GW9662 in the two cell lines. PPARγ was overexpressed (**D**) and knockdown (**E**) in the two cell lines and co-treated with TCS in HTR-8/SVneo and JEG-3 cells while cell vitality was analyzed. (**F**) HTR-8/SVneo and JEG-3 cells exposed to TCS while in the absence or presence of rosiglitazone (10 µM) and GW9662 (10 µM). HTR-8/SVneo and JEG-3 cells exposed to TCS during PPARγ overexpression (**G**) or knockdown (**H**). The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

#### *2.2. PPARγ Is Involved in TCS-Elicited Impaired Migration 3.2. PPARγ Is Involved in TCS-Elicited Impaired Migration*

The wound healing assay confirmed that the migration ability of HTR-8/SVneo and JEG-3 cells were inhibited after concentration dependent on exposure to TCS (Figure 2A). To assess the impact of TCS on cell migration influenced by PPARγ in vitro, the migration assays were conducted in the absence or presence of rosiglitazone or GW9662. Our results in Figure 2B demonstrated that treatment with rosiglitazone significantly increased cell migration, while GW9662 decreased the cell migration compared to the TCS-exposed group in HTR-8/SVneo and JEG-3 cells. PPARγ displayed overexpression and knockdown in HTR-8/SVneo and JEG-3 cells. Our results indicated that PPARγ overexpression alleviated the migration inhibition induced by TCS in HTR-8/SVneo and JEG-3 cells (Figure 2C). Moreover, PPARγ knockdown aggravated the migration inhibition induced by TCS in these two cell lines (Figure 2D). The wound healing assay confirmed that the migration ability of HTR-8/SVneo and JEG-3 cells were inhibited after concentration dependent on exposure to TCS (Figure 2A). To assess the impact of TCS on cell migration influenced by PPARγ in vitro, the migration assays were conducted in the absence or presence of rosiglitazone or GW9662. Our results in Figure 2B demonstrated that treatment with rosiglitazone significantly increased cell migration, while GW9662 decreased the cell migration compared to the TCS-exposed group in HTR-8/SVneo and JEG-3 cells. PPARγ displayed overexpression and knockdown in HTR-8/SVneo and JEG-3 cells. Our results indicated that PPARγ overexpression alleviated the migration inhibition induced by TCS in HTR-8/SVneo and JEG-3 cells (Figure 2C). Moreover, PPARγ knockdown aggravated the migration inhibition induced by TCS in these two cell lines (Figure 2D).

**Figure 2.** Role of PPARγ in migration exposed to TCS in vitro. The migration distances were measured after exposure to TCS at 0 h and 24 h. (**A**) HTR-8/SVneo and JEG-3 cell migration distance was decreased with increasing concentration of TCS. Co-treated with TCS in the absence or presence of rosiglitazone or GW9662 in HTR-8/SVneo and JEG-3 cells (**B**). The migration distance in response to PPAR overexpression (**C**) and knockdown (**D**) when exposed to TCS. Scale bar: 200 μm. The relative wound closure distances are shown on the right. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3. **Figure 2.** Role of PPARγ in migration exposed to TCS in vitro. The migration distances were measured after exposure to TCS at 0 h and 24 h. (**A**) HTR-8/SVneo and JEG-3 cell migration distance was decreased with increasing concentration of TCS. Co-treated with TCS in the absence or presence of rosiglitazone or GW9662 in HTR-8/SVneo and JEG-3 cells (**B**). The migration distance in response to PPAR overexpression (**C**) and knockdown (**D**) when exposed to TCS. Scale bar: 200 µm. The relative wound closure distances are shown on the right. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

#### *3.3. PPARγ Alleviates TCS-Elicited Angiogenesis Inhibition* An angiogenesis assay was implemented to describe the influence of TCS on tube

*Cells* **2022**, *11*, x FOR PEER REVIEW 5 of 16

*2.3. PPARγ Alleviates TCS-Elicited Angiogenesis Inhibition*

An angiogenesis assay was implemented to describe the influence of TCS on tube branch length. The HTR8/SVneo and JEG-3 cell lines were treated with different concentrations of TCS. TCS showed significantly decreased tube branch length compared to the control (DMSO) at 30 µM or 40 µM in the two cell lines, respectively (Figure 3A). The impact of PPARγ on TCS-elicited tube branch length is displayed in Figure 3. The existing results illustrated that PPARγ overexpression or rosiglitazone co-treatment alleviated the inhibition of tube branch length induced by TCS in vitro (Figure 3B,C). In contrast, PPARγ knockdown or GW9662 co-treatment aggravated the inhibition of tube branch length induced by TCS in HTR8/SVneo and JEG-3 cells (Figure 3B,D). branch length. The HTR8/SVneo and JEG-3 cell lines were treated with different concentrations of TCS. TCS showed significantly decreased tube branch length compared to the control (DMSO) at 30 μM or 40 μM in the two cell lines, respectively (Figure 3A). The impact of PPARγ on TCS-elicited tube branch length is displayed in Figure 3. The existing results illustrated that PPARγ overexpression or rosiglitazone co-treatment alleviated the inhibition of tube branch length induced by TCS in vitro (Figure 3B,C). In contrast, PPARγ knockdown or GW9662 co-treatment aggravated the inhibition of tube branch length induced by TCS in HTR8/SVneo and JEG-3 cells (Figure 3B,D).

**Figure 3.** TCS reduced HTR-8/SVneo and JEG-3 cells angiogenesis through PPARγ pathway. (**A**) TCS decreased the branch length of HTR-8/SVneo and JEG-3 cells. (**B**) Co-treated with TCS (40 μM for HTR-8/SVneo, 30 μM for JEG-3) in the absence or presence of rosiglitazone and GW9662 in these two cells. (**C**,**D**) PPARγ overexpression alleviated, while (**C**) PPARγ knockdown exacerbated. (**D**) TCS-induced cell angiogenesis inhibition of HTR-8/SVneo and JEG-3 cells. Scale bar: 400 μm. The **Figure 3.** TCS reduced HTR-8/SVneo and JEG-3 cells angiogenesis through PPARγ pathway. (**A**) TCS decreased the branch length of HTR-8/SVneo and JEG-3 cells. (**B**) Co-treated with TCS (40 µM for HTR-8/SVneo, 30 µM for JEG-3) in the absence or presence of rosiglitazone and GW9662 in these two cells. (**C**,**D**) PPARγ overexpression alleviated, while (**C**) PPARγ knockdown exacerbated. (**D**) TCS-induced cell angiogenesis inhibition of HTR-8/SVneo and JEG-3 cells. Scale bar: 400 µm. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

#### 3. *3.4. TCS Alters the Expression of PPARγ Target Genes Associated with Viability, Angiogenesis, and Migration*

data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* =

The *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* genes were determined as the target genes of PPARγ and play considerable roles in cell viability, angiogenesis, and migration [25–27]. In comparison with the control group, the levels of *HMOX1*, *ANGPTL4*, *VEGFA, MMP-2,* and *MMP-9* mRNA were significantly decreased in the TCS-exposed group (Figure 4A,B). In addition, the protein levels of *ANGPTL4* and *MMP-2* were significantly decreased in the TCS-exposed groups (Figure 4I,J). We also assessed the influence of *PPARγ* on the above genes in the absence or presence of TCS. The results revealed that rosiglitazone enhanced the expression of *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* levels, although TCS decreased them in vitro (Figure 4C,D). The *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* mRNA levels were upregulated after overexpression of PPARγ in vitro (Figure 4E,F). In contrast, PPARγ knockdown downregulated the cell viability, angiogenesis, and migration-related genes in HTR-8/SVneo and JEG-3 cells (Figure 4G,H). Therefore, TCS inhibited the PPARγ pathway, leading to the downregulation of *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9.*

#### *3.5. TCS Changes Expression Level of PPARγ-Regulated Inflammation Genes*

The effect of TCS on PPARγ-regulated inflammatory genes was investigated, and results demonstrated that TCS was able to upregulate genes involved in inflammation such as *p65*, *IL-6*, *IL-1β*, and *TNF-α* in HTR-8/SVneo and JEG-3 cells (Figure 5A,B). The protein levels of *IL-1β* were significantly increased in TCS-exposed groups (Figure 4I). However, PPARγ overexpression or rosiglitazone co-treatment significantly alleviated the level of these inflammatory cytokines elicited by TCS in vitro (Figure 5C–F). In addition, PPARγ knockdown or GW9662 co-treatment enhanced the expression levels of those genes elicited by TCS in HTR-8/SVneo and JEG-3 cells (Figure 5C–H). These results emphasized that TCS upregulated inflammatory gene expression through the PPARγ pathway.

#### *3.6. TCS Induces Placenta Dysfunction through PPARγ Pathway in Mice*

To reveal the potential toxicity of TCS on the placenta and fetal development, pregnant ICR mice were treated with TCS daily by gavage at doses of 0, 10, 50, and 100 mg/kg/day from gestation day GD7.5 to GD17.5. Uterus, placenta, and fetus were collected for analysis in our study. As the results showed, the size of uterus (Figure 6A), fetus weight (Figure 6B), and placenta (Figure 6C) in TCS-exposed mice were decreased obviously in the 50 and 100 mg/kg/day gavage group compared to the vehicle group, indicating the serious placenta toxicity of TCS. Moreover, rosiglitazone (20 mg/kg/day) administration prevented the decrease of fetus weight (Figure 6B) and placenta diameter (Figure 6C) induced by TCS.

#### *3.7. TCS Alters Expression of PPARγ-Regulated Genes in Mice Placenta*

The effects of gestational TCS exposure on viability, migration, angiogenesis, and inflammatory genes in mice placenta were analyzed. As described in Figure 7, placental *Pparγ* (Figure 7A) and *Pparγ*-regulated genes *Homx1*, *Angptl4*, *Vegfa*, *Mmp-2*, and *Mmp-9* mRNA levels (Figure 7B) were decreased, while inflammatory genes *p65*, *Il-6*, *Il-1β*, and *Tnf-α* mRNA levels (Figure 7C) were significantly elevated in TCS-exposed pregnant mice. The protein expression of ANGPTL4, MMP-2, and IL-1β were consistent with the gene expression trends in the placenta of GD17.5 mice treated with or without TCS (Figure 7D). In addition, treatment with rosiglitazone reversed the PPARγ-regulated viability, migration, angiogenesis, and inflammatory gene expression induced by TCS (Figure 7B,C).

*and Migration*

*HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9.*

*2.4. TCS Alters the Expression of PPARγ Target Genes Associated with Viability, Angiogenesis,*

The *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* genes were determined as the target genes of PPARγ and play considerable roles in cell viability, angiogenesis, and migration [25–27]. In comparison with the control group, the levels of *HMOX1*, *ANGPTL4*, *VEGFA, MMP-2,* and *MMP-9* mRNA were significantly decreased in the TCS-exposed group (Figure 4A,B). In addition, the protein levels of *ANGPTL4* and *MMP-2* were significantly decreased in the TCS-exposed groups (Figure 4I,J). We also assessed the influence of *PPARγ* on the above genes in the absence or presence of TCS. The results revealed that rosiglitazone enhanced the expression of *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* levels, although TCS decreased them in vitro (Figure 4C,D). The *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, and *MMP-9* mRNA levels were upregulated after overexpression of PPARγ in vitro (Figure 4E,F). In contrast, PPARγ knockdown downregulated the cell viability, angiogenesis, and migration-related genes in HTR-8/SVneo and JEG-3 cells (Figure 4G,H). Therefore, TCS inhibited the PPARγ pathway, leading to the downregulation of

**Figure 4.** TCS inhibited expression of PPARγ target genes related to cell vitality, migration, and angiogenesis via PPARγ pathway in vitro. (**A**,**B**) The mRNA expression was analyzed by RT-PCR exposed to the indicated dose of TCS in HTR-8/SVneo and JEG-3 cells. (**C**,**D**) HTR-8/SVneo and JEG-3 cells treated with or without rosiglitazone and GW9662 for 24 h in HTR-8/SVneo and JEG-3 cells exposed to TCS (40 µM for HTR-8/SVneo, 30 µM for JEG-3). (**E**–**H**) PPARγ was overexpressed (**E**,**F**) or displayed knockdown (**G**,**H**) and co-treated with TCS in HTR-8/SVneo and JEG-3 cells. (**I**,**J**) The protein expression of ANGPTL4 (**I**) and MMP-2 (**J**) was analyzed by Western blot in HTR-8/SVneo and JEG-3 cells exposed to TCS. The relative values of protein expression are shown below. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

*n* = 3.

way.

**Figure 4.** TCS inhibited expression of PPARγ target genes related to cell vitality, migration, and angiogenesis via PPARγ pathway in vitro. (**A**,**B**) The mRNA expression was analyzed by RT-PCR exposed to the indicated dose of TCS in HTR-8/SVneo and JEG-3 cells. (**C**,**D**) HTR-8/SVneo and JEG-3 cells treated with or without rosiglitazone and GW9662 for 24 h in HTR-8/SVneo and JEG-3 cells exposed to TCS (40 μM for HTR-8/SVneo, 30 μM for JEG-3). (**E**–**H**) PPARγ was overexpressed (**E**,**F**) or displayed knockdown (G and H) and co-treated with TCS in HTR-8/SVneo and JEG-3 cells. (**I**,**J**) The protein expression of ANGPTL4 (**I**) and MMP-2 (**J**) was analyzed by Western blot in HTR-8/SVneo and JEG-3 cells exposed to TCS. The relative values of protein expression are shown below. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group,

The effect of TCS on PPARγ-regulated inflammatory genes was investigated, and

results demonstrated that TCS was able to upregulate genes involved in inflammation such as *p65*, *IL-6*, *IL-1β*, and *TNF-α* in HTR-8/SVneo and JEG-3 cells (Figure 5A,B). The protein levels of *IL-1β* were significantly increased in TCS-exposed groups (Figure 4I). However, PPARγ overexpression or rosiglitazone co-treatment significantly alleviated the level of these inflammatory cytokines elicited by TCS in vitro (Figure 5C–F). In addition, PPARγ knockdown or GW9662 co-treatment enhanced the expression levels of those genes elicited by TCS in HTR-8/SVneo and JEG-3 cells (Figure 5C–H). These results emphasized that TCS upregulated inflammatory gene expression through the PPARγ path-

*2.5. TCS Changes Expression Level of PPARγ-Regulated Inflammation Genes*

**Figure 5.** TCS increased expression of PPARγ-regulated genes related to inflammation through PPARγ pathway in vitro. (**A**,**B**) The mRNA expression level of the inflammatory genes exposed to indicated dose of TCS in HTR-8/SVneo and JEG-3 cells. (**C**,**D**) HTR-8/SVneo and JEG-3 cells treated with or without rosiglitazone and GW9662 for 24 h in HTR-8/SVneo and JEG-3 cells cotreated with TCS (40 µM for HTR-8/SVneo, 30 µM for JEG-3). (**E**–**H**) PPARγ was overexpressed (**E**,**F**) or displayed knockdown (**G**,**H**) and co-treated with TCS in HTR-8/SVneo and JEG-3 cells. (**I**) The protein expression of IL-1β was analyzed by Western blot in HTR-8/SVneo and JEG-3 cells exposed to TCS. The relative value of protein expression is shown below. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 3.

6C) induced by TCS.

*2.6. TCS Induces Placenta Dysfunction through PPARγ Pathway in Mice*

**Figure 5.** TCS increased expression of PPARγ-regulated genes related to inflammation through PPARγ pathway in vitro. (**A**,**B**) The mRNA expression level of the inflammatory genes exposed to indicated dose of TCS in HTR-8/SVneo and JEG-3 cells. (**C**,**D**) HTR-8/SVneo and JEG-3 cells treated with or without rosiglitazone and GW9662 for 24 h in HTR-8/SVneo and JEG-3 cells co-treated with TCS (40 μM for HTR-8/SVneo, 30 μM for JEG-3). (**E**–**H**) PPARγ was overexpressed (**E**,**F**) or displayed knockdown (**G**,**H**) and co-treated with TCS in HTR-8/SVneo and JEG-3 cells. (**I**) The protein expression of IL-1β was analyzed by Western blot in HTR-8/SVneo and JEG-3 cells exposed to TCS. The relative value of protein expression is shown below. The data are shown as the means ± S.E.M. \* *p* <

To reveal the potential toxicity of TCS on the placenta and fetal development, pregnant ICR mice were treated with TCS daily by gavage at doses of 0, 10, 50, and 100 mg/kg/day from gestation day GD7.5 to GD17.5. Uterus, placenta, and fetus were collected for analysis in our study. As the results showed, the size of uterus (Figure 6A), fetus weight (Figure 6B), and placenta (Figure 6C) in TCS-exposed mice were decreased obviously in the 50 and 100 mg/kg/day gavage group compared to the vehicle group, indicating the serious placenta toxicity of TCS. Moreover, rosiglitazone (20 mg/kg/day) administration prevented the decrease of fetus weight (Figure 6B) and placenta diameter (Figure

**Figure 6.** PPARγ participated in placental and fetal development toxicity of TCS. Representative picture of uterus (**A**), fetus (**B**), and placenta (**C**) of GD17.5 mice exposed to indicated doses of TCS. The data are shown as the means ± S.E.M; \* *p* < 0.05 and \*\* *p* < 0.01, compared with control group; *p* < 0.05, compared with TCS (100 mg/kg/day) group; *n* = 8. **Figure 6.** PPARγ participated in placental and fetal development toxicity of TCS. Representative picture of uterus (**A**), fetus (**B**), and placenta (**C**) of GD17.5 mice exposed to indicated doses of TCS. The data are shown as the means ± S.E.M; \* *p* < 0.05 and \*\* *p* < 0.01, compared with control group; # *p* < 0.05, compared with TCS (100 mg/kg/day) group; *n* = 8. *Cells* **2022**, *11*, x FOR PEER REVIEW 9 of 16

#

**Figure 7.** The gene and protein expression of PPARγ-regulated genes in placentas of gestational mice exposed to TCS. (**A**) The PPARγ expression level was detected by RT-PCR in GD17.5 mice placentas. (**B**,**C**) The mRNA expression of cell growth, angiogenesis, migration (**B**), and inflammation (**C**)-related genes in the placenta of GD17.5 mice treated with indicated doses of TCS and with or without rosiglitazone. (**D**) The protein expression of ANGPTL4, MMP-2, and IL-1β were analyzed by Western blot in placenta of GD17.5 mice treated with or without TCS (100 mg/kg/day). The relative values of protein expression are shown on the right. The data are shown as the means ±

In spite of numerous developmental toxicities shown to be triggered by TCS, the

S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 8.

mechanism of TCS-elicited severe placental dysfunction has not been well elaborated. **Figure 7.** *Cont*.

**3. Discussion**

**Figure 7.** The gene and protein expression of PPARγ-regulated genes in placentas of gestational mice exposed to TCS. (**A**) The PPARγ expression level was detected by RT-PCR in GD17.5 mice placentas. (**B**,**C**) The mRNA expression of cell growth, angiogenesis, migration (**B**), and inflammation (**C**)-related genes in the placenta of GD17.5 mice treated with indicated doses of TCS and with or without rosiglitazone. (**D**) The protein expression of ANGPTL4, MMP-2, and IL-1β were analyzed by Western blot in placenta of GD17.5 mice treated with or without TCS (100 mg/kg/day). The relative values of protein expression are shown on the right. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 8. **Figure 7.** The gene and protein expression of PPARγ-regulated genes in placentas of gestational mice exposed to TCS. (**A**) The PPARγ expression level was detected by RT-PCR in GD17.5 mice placentas. (**B**,**C**) The mRNA expression of cell growth, angiogenesis, migration (**B**), and inflammation (**C**)-related genes in the placenta of GD17.5 mice treated with indicated doses of TCS and with or without rosiglitazone. (**D**) The protein expression of ANGPTL4, MMP-2, and IL-1β were analyzed by Western blot in placenta of GD17.5 mice treated with or without TCS (100 mg/kg/day). The relative values of protein expression are shown on the right. The data are shown as the means ± S.E.M. \* *p* < 0.05; \*\* *p* < 0.01; compared with the indicated group, *n* = 8.

#### **4. Discussion**

**3. Discussion**

In spite of numerous developmental toxicities shown to be triggered by TCS, the mechanism of TCS-elicited severe placental dysfunction has not been well elaborated. In spite of numerous developmental toxicities shown to be triggered by TCS, the mechanism of TCS-elicited severe placental dysfunction has not been well elaborated. Here, our research revealed that TCS dose-dependently inhibited cell growth, migration, angiogenesis in HTR-8/SVneo and JEG-3 cells, and impaired placental development in mice. The mechanism of TCS-induced effects was studied. PPARγ was partly involved in the toxicity of TCS by regulating placental cell growth, migration, angiogenesis, and inflammatory responses in vitro and in vivo.

Abnormal placental cell proliferation may induce pregnancy complications such as fetal growth retardation, miscarriage, preeclampsia, and macrosomia [28]. Previous studies have verified that TCS had cytotoxic impression, elicited apoptosis, and inhibited cell vitality of human placental trophoblasts [29,30]. Placenta angiogenesis works for placental transport, endocrine, metabolic, and immune function regulation. It is essential for embryogenesis and is involved in the reproductive cycle and wound healing [16]. The intra-placental vascular lesions may result in preeclampsia, fetal growth restriction, or birthweight decrease [31]. Previous research has reported that TCS exposure decreased fetal viability and fetal body weight through placental thrombosis [32]. Some evidence has also demonstrated that TCS stimulated vascular branch disappearance and vascular injury in zebrafish [33]. In our animal research, placental diameter and fetal weight significantly decreased in pregnant mice in the TCS high-dose group. Previous studies have illustrated that the placenta had the greatest bio-accumulation of TCS, and reduced uterine weight and abortion were observed in the TCS (600 mg/kg/day) group in rats [34]. Following TCS exposure, reduction of gravid uterine weight and the occurrence of abortion was observed in pregnant rats [34]. PPARγ-deficient mice have placental abnormalities and defects in trophoblast differentiation and vascular development [20]. In addition, rosiglitazone increased placental vascularization and trophoblast migration and invasion in villous cytotrophoblast cells (VCT) and placental explants [35]. Our results also proved that PPARγ was inhibited in the mice placentas after treatment with TCS, and rosiglitazone prevented TCS-induced placenta toxicity by activation of PPARγ. It will be interesting to know the mechanism by which TCS affects PPARγ expression. Our results demonstrated that TCS inhibited cell growth, migration, and angiogenesis and influenced placenta development through the PPARγ pathway in vitro and in vivo.

Vascular endothelial growth factors A (*VEGFA*) and Heme oxygenase 1 (*HO-1*) are key modules of the angiogenesis process [36]. Angiopoietin-like protein 4 (*ANGPTL4*) is a secretory glycoprotein member of the angiopoietin family, which participates in the regulation of migration and angiogenesis [37,38]. Furthermore, functional studies found that PPARγ targeted *ANGPTL4* in placental development and angiogenesis, mediating the survival, proliferation, migration, and invasion of HTR8/SVneo cells [39,40]. Our results indicated that PPARγ activation alleviated the decrease of *HMOX1*, *ANGPTL4,* and *VEGFA* expression caused by TCS. MMPs have been demonstrated to mediate the migration and invasion of trophoblast cells, and *MMP-2* and *MMP-9* expression levels were decreased in preeclamptic placental tissues [41,42]. Similarly, our findings suggested that PPARγ mediated TCS-induced migration inhibition via *MMP-2* and *MMP-9* expression. In this study, cell growth, migration, and angiogenesis inhibition by TCS was partly ameliorated by PPARγ elevation or activation. These findings prove that TCS affects cell viability, migration, and angiogenesis through the PPARγ pathway.

Pregnancy is a process of dynamic inflammatory phases, and the inflammatory state is present almost throughout pregnancy and towards the end [43]. Activation of PPARγ is reported to suppress inflammation through NF-kappaB and TNF-α [44]. Indeed, proinflammatory conditions have been associated with pregnancy complications such as intrauterine growth restriction and preeclampsia [45]. TCS increased the inflammatory response by promoting *TNF-a* and *IL-6* expression in HUVEC [32]. Interestingly, PPARγ has a critical role in regulating inflammatory cytokines including *TNF-α* and *IL-6,* which were linked to preterm labor, miscarriage, and preeclampsia [22]. Moreover, PPAR-γ has been proved for its anti-inflammatory effects and downregulation of the expression of proinflammatory cytokines, such as *IL-6*, *IL-8*, and *TNF-α*, in human gestational tissue [46,47]. Furthermore, activation of PPARγ by rosiglitazone reversed the LPS-mediated effects on inflammatory cytokine release and proliferation inhibition in HTR-8/SVneo cells [48]. Similarly, we demonstrated that TCS can induce the expression of inflammatory genes in two cell lines. Rosiglitazone or PPARγ overexpression alleviated PFOS-induced cell growth, migration, angiogenesis inhibition, and the release of inflammatory cytokines in HTR-8/SVneo and JEG-3 cells [49]. Previous studies indicated that decreased PPARγ expression or the inhibition of PPARγ activity led to mitochondrial fission, hyperpolarization, and increased oxidative stress [50]. The PPARγ pathway was one mechanism of triclosaninduced mitochondria-targeted effects, regulating the function of these organelles and the permeability of their membranes [51,52]. Our research showed that PPARγ activation or overexpression mitigated, while PPARγ inhibition or silence aggravated the increase of inflammatory gene expression caused by TCS. It is interesting that PPARγ prevented TCS-induced toxic phenotypes, while treatments with a PPARγ agonist or antagonist alone had no effects, suggesting some new mechanisms related to PPARγ can be initiated by TCS, which is worth more future study.

All the results in this study illustrated that TCS caused significant abnormal functions of HTR-8/SVneo and JEG-3 cells, including viability, migration, angiogenesis, and inflammation response. In addition, treatment with rosiglitazone or overexpression of PPARγ almost completely prevented the abnormal cell function changes in vitro. On the other hand, treatment with GW9662 or *si-PPARγ* aggravated the toxicity. Similarly, animal studies also indicated that rosiglitazone mitigated the decrease of placental diameter and fetal weight caused by TCS. In addition, curcumin, a natural compound and a modulator of PPARγ, has been known to have beneficial effects on pregnancy outcomes [53,54]. It will be interesting to know the effect of curcumin or other natural PPARγ modulators on TCS placental exposure in the future. Our results suggested that TCS placental exposure had adverse effects in vitro and in vivo through the PPARγ pathway (Figure 8), and activation or increased expression of PPARγ is a potential strategy to protect against the placenta toxicity induced by TCS.

TCS, which is worth more future study.

centa toxicity induced by TCS.

**Figure 8.** Schematic model depicting TCS-induced placental dysfunction and low birth-weight infants through PPARγ signaling pathways. TCS inhibited cell growth, angiogenesis, and migration and promoted inflammation of placenta via PPARγ-regulated genes *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, *MMP*-9, *p65*, *IL-6*, *IL-1β*, and *TNF-α*. **Figure 8.** Schematic model depicting TCS-induced placental dysfunction and low birth-weight infants through PPARγ signaling pathways. TCS inhibited cell growth, angiogenesis, and migration and promoted inflammation of placenta via PPARγ-regulated genes *HMOX1*, *ANGPTL4*, *VEGFA*, *MMP-2*, *MMP*-9, *p65*, *IL-6*, *IL-1β*, and *TNF-α*.

inflammatory cytokine release and proliferation inhibition in HTR-8/SVneo cells [48]. Similarly, we demonstrated that TCS can induce the expression of inflammatory genes in two cell lines. Rosiglitazone or PPARγ overexpression alleviated PFOS-induced cell growth, migration, angiogenesis inhibition, and the release of inflammatory cytokines in HTR-8/SVneo and JEG-3 cells [49]. Previous studies indicated that decreased PPARγ expression or the inhibition of PPARγ activity led to mitochondrial fission, hyperpolarization, and increased oxidative stress [50]. The PPARγ pathway was one mechanism of triclosan-induced mitochondria-targeted effects, regulating the function of these organelles and the permeability of their membranes [51,52]. Our research showed that PPARγ activation or overexpression mitigated, while PPARγ inhibition or silence aggravated the increase of inflammatory gene expression caused by TCS. It is interesting that PPARγ prevented TCSinduced toxic phenotypes, while treatments with a PPARγ agonist or antagonist alone had no effects, suggesting some new mechanisms related to PPARγ can be initiated by

All the results in this study illustrated that TCS caused significant abnormal functions of HTR-8/SVneo and JEG-3 cells, including viability, migration, angiogenesis, and inflammation response. In addition, treatment with rosiglitazone or overexpression of PPARγ almost completely prevented the abnormal cell function changes in vitro. On the other hand, treatment with GW9662 or *si-PPARγ* aggravated the toxicity. Similarly, animal studies also indicated that rosiglitazone mitigated the decrease of placental diameter and fetal weight caused by TCS. In addition, curcumin, a natural compound and a modulator of PPARγ, has been known to have beneficial effects on pregnancy outcomes [53,54]. It will be interesting to know the effect of curcumin or other natural PPARγ modulators on TCS placental exposure in the future. Our results suggested that TCS placental exposure had adverse effects in vitro and in vivo through the PPARγ pathway (Figure 8), and activation or increased expression of PPARγ is a potential strategy to protect against the pla-

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11010086/s1, Figure S1: The efficiency of PPARγ overexpression and knockdown; Table S1: RT-PCR primers.

**Author Contributions:** Conceptualization, J.L. and P.X.; methodology, X.Q.; formal analysis, J.L. and S.L.; investigation, J.L. and Q.W.; data curation, X.Q., Y.Z. and T.Y.; writing—original draft preparation, X.Q. and Z.H.; writing—review and editing, X.Q., S.L. and P.X.; supervision, J.L., X.Y. and P.X.; project administration, W.S., P.X. and J.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundation of China [No. 81703260]; the Science and Technology Department of Jiangsu Province [No. BK20160227]; the China Postdoctoral Science Foundation funded project [No. 2016M601892]; the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and Jiangsu Overseas Visiting Scholar Program for University Prominent Young and Middle-aged Teachers and Presidents, the Postgraduate Research and Practice Innovation Program of Jiangsu Province (No. KYCX21\_2725).

**Institutional Review Board Statement:** The animal study protocol was approved by the Xuzhou Medical University Animal Ethics Committee, the entire experiment was per-formed in a Specified Pathogen-Free (SPF) environment (protocols 201605w025, 25 May 2016 and 202106A237, 25 June 2021).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data sets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

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

#### **References**


### MDPI

St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Cells* Editorial Office E-mail: cells@mdpi.com www.mdpi.com/journal/cells

Academic Open Access Publishing

www.mdpi.com ISBN 978-3-0365-8111-8