DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao...

15
Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020 SCIENCE ADVANCES | RESEARCH ARTICLE 1 of 14 DISEASES AND DISORDERS Gadofullerene inhibits the degradation of apolipoprotein B100 and boosts triglyceride transport for reversing hepatic steatosis Chen Zhou 1,2 , Mingming Zhen 1,2 *, Meilan Yu 3 , Xue Li 1,2 , Tong Yu 1,2 , Jingchao Liu 1 , Wang Jia 1,2 , Shuai Liu 1,2 , Lei Li 1,2 , Jie Li 1,2 , Zihao Sun 1,2 , Zhongpu Zhao 1,2 , Xinyao Wang 1,2 , Xiaoyan Zhang 1 , Chunru Wang 1,2 *, Chunli Bai 1,2 * Hepatic steatosis is a widespread metabolic disease characterized by excessive accumulation of triglyceride (TG) in liver. So far, effective approved drugs for hepatic steatosis are still in development, and removing the unneces- sary TG from the hepatocytes is an enormous challenge. Here, we explore a promising anti-hepatic steatosis strat- egy by boosting hepatocellular TG transport using -alanine–modified gadofullerene (GF-Ala) nanoparticles. We confirm that GF-Ala could reverse hepatic steatosis in oleic acid–induced hepatocytes, fructose-induced mice, and obesity-associated transgenic ob/ob mice. Observably, GF-Ala improves hepatomegaly and hepatic lipid accumu- lation, reduces lipid peroxidation, and repairs abnormal mitochondria. Of note, we demonstrate that GF-Ala markedly inhibits the posttranslational degradation of apolipoprotein B100 (ApoB100) and boosts hepatocellular TG transport based on their superior antioxidant property. Together, we conclude that GF-Ala could potently ameliorate hepatic TG transport and maintain hepatic metabolic homeostasis without apparent toxicity, being beneficial for treatments of hepatic steatosis and other fatty liver diseases. INTRODUCTION Hepatic steatosis is becoming one of the crucial metabolism disorders worldwide, closely associated with obesity and diabetes (12). Hepatic steatosis is the first stage of nonalcoholic fatty liver disease (NAFLD). It starts with the hepatic metabolic disorder and possibly develops into nonalcoholic steatohepatitis (NASH) and cirrhosis (34). The most typical pathological feature of hepatic steatosis is excessive lipid droplets [the main component is triglyceride (TG)] in liver, which generally cause excessive reactive oxygen species (ROS) and lipid peroxidation. These will bring further unbalance of hepatic lipid metabolism and hepatocyte damage. Because the pathogenesis of he- patic steatosis is poorly understood, effective approved drugs for he- patic steatosis are still in development, and some drugs are used in clinical trials. Drugs enhancing hepatic lipid -oxidation or reducing lipogenesis that could effectively attenuate the lipid accumulation have been shown (57). However, drugs boosting hepatocellular TG transport are rarely reported. The apolipoprotein B100 (ApoB100) is one of the essential apoli- poproteins for hepatic TG transport (8). Frequently, lipoproteins containing ApoB100, TG, and cholesterol assemble in endoplasmic reticulum (ER) to generate very low density lipoprotein (VLDL) and export from hepatocytes via vesicles (9), removing lipid from the liver to serum. The VLDL comprises ~60% TG, which is the essential way to transport hepatic TG. In hepatic steatosis, the excessive ROS generated by lipid accumulation will induce the degradation of ApoB100 in the cytoplasm and block the TG transport (1012). Thus, it offers a rational solution for hepatic steatosis treatment by reactivating lipid transport by normalizing the degradation of ApoB100. To date, the booming nanotechnology creates wide applications in the biomedicine field (1315). Functionalized fullerene or gadofullerene, a kind of unique nanomaterials containing spherical or ellipsoidal carbon cage structure, has extended -bond system with high electron affinity, showing excellent antioxidant ability and other superior properties. Some researches revealed that fullerene or gadofullerene nanomaterials were used to treat several diseases, such as tumor therapy (1617) and pulmonary fibrosis disease treatment (18). They were also reported to prevent mitochondria abnormalities (19), regu- late oxidative stress (20), inhibit cell apoptosis (2122), and decrease inflammation response (23). Our previous study revealed that gad- ofullerene nanoparticles effectively improved type 2 diabetes (24), inspiring us whether gadofullerene could be used for hepatic steatosis therapy because they both belong to metabolic disorders. In this work, we demonstrate that -alanine derivatives of gadofullerene (GF-Ala) nanoparticles reversed hepatic steatosis by boosting the hepatocellular TG transport (Fig. 1A). We reveal that GF-Ala decreases hepatic lipid accumulation in both diet-induced wild-type (WT) mice and ob/ob obese mice. In addition, GF-Ala highly reduces lipid per- oxidation and repairs abnormal mitochondria in hepatocytes suffering from hepatic steatosis. The systematic proteomics profiling and other molecular biology approaches are used to elucidate the mechanism of GF-Ala treatment. We demonstrate that GF-Ala inhibits the degra- dation of ApoB100 and boosts the hepatocellular TG transport due to its antioxidant property. It is worth mentioning that GF-Ala could be easily eliminated from the body and cause no overt toxicity toward the main organs. These exciting results indicate that GF-Ala nanoparticles could be used as a promising therapeutic agent to reverse NAFLD. RESULTS Preparation and characterization of -alanine–modified gadofullerene nanoparticles The -alanine–modified gadofullerene (GF-Ala) nanoparticles were prepared in the alkaline environment via nucleophilic attack of amino 1 Beijing National Laboratory for Molecular Sciences, Key Laboratory of Molecular Nanostructure and Nanotechnology, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China. 2 University of Chinese Academy of Sciences, Beijing 100049, China. 3 College of Life Science, Zhejiang Sci-Tech University, Hangzhou 310018, China. *Corresponding author. Email: [email protected] (M.Z.); [email protected] (C.W.); [email protected] (C.B.) Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC). on February 10, 2021 http://advances.sciencemag.org/ Downloaded from

Transcript of DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao...

Page 1: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

1 of 14

D I S E A S E S A N D D I S O R D E R S

Gadofullerene inhibits the degradation of apolipoprotein B100 and boosts triglyceride transport for reversing hepatic steatosisChen Zhou1,2, Mingming Zhen1,2*, Meilan Yu3, Xue Li1,2, Tong Yu1,2, Jingchao Liu1, Wang Jia1,2, Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, Chunru Wang1,2*, Chunli Bai1,2*

Hepatic steatosis is a widespread metabolic disease characterized by excessive accumulation of triglyceride (TG) in liver. So far, effective approved drugs for hepatic steatosis are still in development, and removing the unneces-sary TG from the hepatocytes is an enormous challenge. Here, we explore a promising anti-hepatic steatosis strat-egy by boosting hepatocellular TG transport using -alanine–modified gadofullerene (GF-Ala) nanoparticles. We confirm that GF-Ala could reverse hepatic steatosis in oleic acid–induced hepatocytes, fructose-induced mice, and obesity- associated transgenic ob/ob mice. Observably, GF-Ala improves hepatomegaly and hepatic lipid accumu-lation, reduces lipid peroxidation, and repairs abnormal mitochondria. Of note, we demonstrate that GF-Ala markedly inhibits the posttranslational degradation of apolipoprotein B100 (ApoB100) and boosts hepatocellular TG transport based on their superior antioxidant property. Together, we conclude that GF-Ala could potently ameliorate hepatic TG transport and maintain hepatic metabolic homeostasis without apparent toxicity, being beneficial for treatments of hepatic steatosis and other fatty liver diseases.

INTRODUCTIONHepatic steatosis is becoming one of the crucial metabolism disorders worldwide, closely associated with obesity and diabetes (1, 2). Hepatic steatosis is the first stage of nonalcoholic fatty liver disease (NAFLD). It starts with the hepatic metabolic disorder and possibly develops into nonalcoholic steatohepatitis (NASH) and cirrhosis (3, 4). The most typical pathological feature of hepatic steatosis is excessive lipid droplets [the main component is triglyceride (TG)] in liver, which generally cause excessive reactive oxygen species (ROS) and lipid peroxidation. These will bring further unbalance of hepatic lipid metabolism and hepatocyte damage. Because the pathogenesis of he-patic steatosis is poorly understood, effective approved drugs for he-patic steatosis are still in development, and some drugs are used in clinical trials. Drugs enhancing hepatic lipid -oxidation or reducing lipogenesis that could effectively attenuate the lipid accumulation have been shown (5–7). However, drugs boosting hepatocellular TG transport are rarely reported.

The apolipoprotein B100 (ApoB100) is one of the essential apoli-poproteins for hepatic TG transport (8). Frequently, lipoproteins containing ApoB100, TG, and cholesterol assemble in endoplasmic reticulum (ER) to generate very low density lipoprotein (VLDL) and export from hepatocytes via vesicles (9), removing lipid from the liver to serum. The VLDL comprises ~60% TG, which is the essential way to transport hepatic TG. In hepatic steatosis, the excessive ROS generated by lipid accumulation will induce the degradation of ApoB100 in the cytoplasm and block the TG transport (10–12). Thus, it offers a rational solution for hepatic steatosis treatment by reactivating lipid transport by normalizing the degradation of ApoB100.

To date, the booming nanotechnology creates wide applications in the biomedicine field (13–15). Functionalized fullerene or gadofullerene, a kind of unique nanomaterials containing spherical or ellipsoidal carbon cage structure, has extended -bond system with high electron affinity, showing excellent antioxidant ability and other superior properties. Some researches revealed that fullerene or gadofullerene nanomaterials were used to treat several diseases, such as tumor therapy (16, 17) and pulmonary fibrosis disease treatment (18). They were also reported to prevent mitochondria abnormalities (19), regu-late oxidative stress (20), inhibit cell apoptosis (21, 22), and decrease inflammation response (23). Our previous study revealed that gad-ofullerene nanoparticles effectively improved type 2 diabetes (24), inspiring us whether gadofullerene could be used for hepatic steatosis therapy because they both belong to metabolic disorders.

In this work, we demonstrate that -alanine derivatives of gadofullerene (GF-Ala) nanoparticles reversed hepatic steatosis by boosting the hepatocellular TG transport (Fig. 1A). We reveal that GF-Ala decreases hepatic lipid accumulation in both diet-induced wild-type (WT) mice and ob/ob obese mice. In addition, GF-Ala highly reduces lipid per-oxidation and repairs abnormal mitochondria in hepatocytes suffering from hepatic steatosis. The systematic proteomics profiling and other molecular biology approaches are used to elucidate the mechanism of GF-Ala treatment. We demonstrate that GF-Ala inhibits the degra-dation of ApoB100 and boosts the hepatocellular TG transport due to its antioxidant property. It is worth mentioning that GF-Ala could be easily eliminated from the body and cause no overt toxicity toward the main organs. These exciting results indicate that GF-Ala nanoparticles could be used as a promising therapeutic agent to reverse NAFLD.

RESULTSPreparation and characterization of -alanine–modified gadofullerene nanoparticlesThe -alanine–modified gadofullerene (GF-Ala) nanoparticles were prepared in the alkaline environment via nucleophilic attack of amino

1Beijing National Laboratory for Molecular Sciences, Key Laboratory of Molecular Nanostructure and Nanotechnology, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China. 2University of Chinese Academy of Sciences, Beijing 100049, China. 3College of Life Science, Zhejiang Sci-Tech University, Hangzhou 310018, China.*Corresponding author. Email: [email protected] (M.Z.); [email protected] (C.W.); [email protected] (C.B.)

Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 2: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

2 of 14

groups and hydroxyl ions (25) (Fig. 1B). After purification and freeze- drying, the structural compositions of GF-Ala were thoroughly characterized by Fourier transform infrared (FTIR) spectroscopy, elemental analysis (EA), and thermogravimetric analysis (TGA). First, the stretching vibration bands of N (O)─H (ca. 3400 to 3300 cm−1), C─H (ca. 2900 to 2800 cm−1), and C═O (1740 cm−1) in the FTIR could be ascribed to the hydroxyl and -alanine groups of GF-Ala. The bands of C═C (1570 cm−1), C─C (1410 cm−1), and C─N (O) (ca. 1320 cm−1) could be assigned to Gd@C82 carbon cages and their bonded groups (Fig. 1C). Then, the weight loss before 115°C was assigned to 10.96 weight % (wt %) bound water in GF-Ala based on the TGA result (fig. S1A). As shown in the EA results, there were 53.99 wt % C, 3.42 wt % H, and 3.04 wt % N in GF-Ala, suggesting that each carbon cage was composed of ca. 5 -alanine groups. Last, the average structure formula of GF-Ala could be described as Gd@

C82(OH)~20(NHCH2CH2COOH)~5•13H2O (calculated values: 54.01 wt % C, 3.53 wt % H, 3.25 wt % N, and 31.92 wt % O). The size and zeta potential of GF-Ala were investigated by dynamic light scattering (DLS). Hydrodynamic diameters were 89.3 nm [polydispersity in-dex (PDI), 0.227] in water as well as 86.6 nm (PDI: 0.307) in normal saline (NS) (Fig. 1D), and zeta potential (pH 7.0) was −44.0 ± 2.0 mV. In addition, GF-Ala was also highly stable without showing any aggre-gation in water, saline, and fetal bovine serum (FBS) even after cen-trifugation at 10,000 rpm for 10 min (Fig. 1E).

Hydroxyl radicals (•OH) are known as the most robust ROS in-duced by lipid accumulation in hepatic steatosis (26). As the carbon cage has an extended -bond system with high electron affinity, GF-Ala could quench •OH via directly adding with •OH by ethylenic bonds or donating hydrogen atoms to form water. Here, we tested the ability of GF-Ala to eliminate •OH in vitro by electron spin resonance

+NaOH

80°C stirring

8000

−8000

−4000

0

4000

Inte

nsity

(a.u

.)

3200 32403220 3260Magnetic field (Gauss)

-Alanine

100

95

90

75

80

85

Tran

smitt

ance

(%)

4000 100020003000Wave numbers (cm−1)

C

B

FBlank

GF-Ala 50 µMGF-Ala 25 µM

Gd@C82 powder

E

In waterIn saline

In FBS

Beforecentrifugation

Aftercentrifugation

() 20

GF-Ala

( )5

25

20

15

0

5

10N

umbe

r (%

)

1 100010010 10,000Diameter (nm)

D In water

In saline

After treatmentAfter treatment

TGTGTGGTGTGTGTGGF-Ala

Hepatic steatosis

ROSApoB100

TG

Degradation

Lipid transport

ApoB100Degradation

Lipid catabolism

Lipid catabolism AlaGF-A

Lipid transport

ROS

A

Fig. 1. Schematic depiction and characterizations of GF-Ala. (A) Anti-hepatic steatosis mechanism diagram of GF-Ala. (B) Schematic diagrams of the synthesis of GF-Ala. (C) FTIR spectroscopy of GF-Ala. (D) Hydrodynamic size distribution of GF-Ala. (E) Optical images of GF-Ala in water, NS, and FBS before and after centrifugation at 10,000 rpm for 10 min. Photo credit: Chen Zhou, Institute of Chemistry Chinese Academy of Sciences. (F) X-band EPR spectra of hydroxyl radicals captured by DMPO after treatment with GF-Ala. a.u., arbitrary unit.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 3: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

3 of 14

(ESR). The •OH were derived from H2O2 under ultraviolet (UV) light and captured by 5, 5-dimethyl-1-pyrroline N-oxide (DMPO). More than 90% of •OH were eliminated by 50 M GF-Ala (Fig. 1F). Given both [60] fullerene and gadofullerene derivatives acting as excellent radical scavengers, we also prepared -alanine–modified [60] fullerene, C60-Ala, via a similar method. Comparatively, GF-Ala exhibited a more excellent elimination of •OH than that of C60-Ala, which might be more carbon atoms and embedded gadolinium ion in GF-Ala (fig. S1B). All these results indicate that GF-Ala has narrow distribution, excellent stability, and superior antioxidant property in physiological media.

In vivo anti-hepatic steatosis treatment of GF-AlaTo investigate whether GF-Ala nanoparticles could treat hepatic steatosis in vivo, we established two animal models. One is caused by feeding with fructose in WT mice. Briefly, C57BL/6J mice (male, 6 weeks old; n = 12) were fed with drinking water containing 30% fructose and standard normal chow diet ad libitum. After 8 weeks, mice were divided into two groups randomly and treated with GF-Ala [6 mol/kg, intraperitoneally (ip)] or NS for 15 days. Another six C57BL/6J mice (male, 6 weeks old) were fed with sterile water and a normal diet for 8 weeks and treated with NS for 15 days (fig. S2A). There was no significant difference in body weight among each group (fig. S2B). The liver-to-body ratio and oil red O staining show that excessive lipids in the mice fed with fructose are remarkably reduced after GF-Ala treatment (fig. S2, C and D).

Given that it takes a long time to obtain this model and there are large intragroup differences in mice, we chose another leptin-deficient mouse model, ob/ob obese (OB) mice, commonly used for hepatic steatosis study. First, we treated OB mice with GF-Ala (6 mol/kg) for 10, 15, and 30 days (Fig. 2A). The liver of OB mice had obvious hepatomegaly and is pale and bloodless. We found that the liver’s hepatomegaly and abnormal appearances in OB mice were promi-nently improved after 10 days with GF-Ala administration (Fig. 2B). However, according to the hepatic hematoxylin and eosin (H&E) staining and oil red O staining, there were still a few lipid droplets in the liver after 10 days of treatment. When the period was pro-longed to 15 days, we detected better anti-hepatic steatosis effects of GF-Ala. Pathology results also showed that the curative effects of hepatic steatosis in 15-day treatment are similar to those in 30-day treatment (fig. S2, E and F). Therefore, we chose 15 days in the fol-lowing research, and three concentrations of GF-Ala were used: low dosage (L): 1.5 mol/kg, medial dosage (M): 3 mol/kg, and high dosage (H): 6 mol/kg (Fig. 2C).

We recorded the body weights of both WT mice and OB mice every 3 days during the 15-day treatment (fig. S2G). The average weight of WT mice treated with saline was increased about 2 g, and the OB mice gained approximately 7 g at the end of the treat-ment. The body weights of OB mice went up slowly after GF-Ala treatment. Notably, the body weights of OB mice only increased ~4 g in the GF-Ala H group, in which the weight growth rate had fallen nearly by half compared with the OB mice treated by saline (Fig. 2D). Notably, the average feed consumption in both WT mice and OB mice declined after GF-Ala treatment with concen-tration dependence (fig. S2H). Besides, the total weights of pri-mary adipose tissues (including subcutaneous, perinephric, and gonadal fat) and adipose tissue–to–body weight ratio had been significantly declined by GF-Ala treatment in a dose-dependent manner (Fig. 2, E and F).

Except for the reduction of body weight and adipose tissues, the liver-to-body weight ratio in OB mice decreased after GF-Ala treat-ment (Fig. 2G). Furthermore, the alanine transaminase (ALT) and aspartate transaminase (AST), two primary biomarkers of hepatic function, were returned to normal levels in medial and high dosage groups (Fig. 2, H and I). The hepatic pathology was used to precisely evaluate the anti-hepatic steatosis effects of GF-Ala in different dosages. H&E staining exhibited extensive hepatocellular vacuoles, lipid drop-lets, and hepatocyte damage in the liver of OB mice. Those typically pathological changes in hepatic steatosis were highly reversed by GF-Ala treatment (Fig. 2J). In oil red O staining, there was a ~90% reduction of hepatic neutral lipid content in OB GF-Ala H group compared with OB NS group (Fig. 2K). Also, the hepatic free fatty acid (FFA) and total TG were also reduced after the 15-day treat-ment (Fig. 2, L and M). Together, we strongly confirm that GF-Ala has a superior anti-hepatic steatosis effect on both fructose-induced mice and OB mice.

Because of the decline of weight gain after GF-Ala treatment, we further studied the influence of weight loss on hepatic steatosis using orlistat, an approved anti-obesity agent inhibiting lipid absorp-tion in intestine. We treated OB mice with orlistat for 15 days by oral administration (50 mg/kg, every day). The body weight gain slowed down by orlistat (fig. S3A), but no improvement on hepatic steatosis was obtained in such a short period (fig. S3B).

Quantitative proteomics studyTo gain insight into the protein interaction in GF-Ala treatment, we performed a tandem mass tag (TMT) based on the quantitative pro-teomics profiling technique. We detected 4233 proteins, among which 4222 proteins coexisted within four experiment groups, and 3945 pro-teins were identified and quantified in the Venn diagram (Fig. 3A). The WT NS, WT GF-Ala H, OB NS, and OB GF-Ala H groups were coupled pairwise to form four pairs, and the proteins with a fold change >1.2 and P < 0.05 were considered as differential proteins. There were 777 differential proteins in WT NS versus OB NS group, 314 differential proteins in OB NS versus OB GF-Ala H group, and 196 differential proteins in both two. Subsequently, we subjected these 196 differential proteins with three categories (cellular com-ponent, molecular functions, and biological process) based on the Gene Ontology (GO) annotation (Fig. 3, B to D). It was found that the differential proteins associated with ER and mitochondria might be the major executors, which mainly involve protein binding and lipid metabolism processes.

According to the pathogenesis of hepatic steatosis (27), we divided the differential proteins into three categories in heat map based on lipid homeostasis, including lipid biosynthesis, lipid catabolism, and lipid transport (Fig. 3E). We found that GF-Ala had positive im-provements on these differential protein expressions in OB mice, all of which tended to be reversed to normal levels compared with the nonhepatic steatosis controls. Moreover, GF-Ala had little impact on these differential proteins in normal mice. With the assistance of the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, we further illustrated the changes of key differential pro-teins in GF-Ala treatment.

(i) In the process of lipid biosynthesis, two vital differential pro-teins, including acetyl–coenzyme A (CoA) carboxylase 1 (ACACA) and fatty acid synthase (FASN) in fatty acid biosynthesis, were nota-bly up-regulated in OB mice, which were markedly down-regulated by GF-Ala treatment. Other differential proteins related with lipogenesis,

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 4: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

4 of 14

Adi

pose

tiss

ue w

eigh

t (g)

Adi

pose

tiss

ue-to

-bod

yw

eigh

t rat

io (%

)

Live

r-to-

body

wei

ght r

atio

(%)

ALT

(U/li

ter)

AST

(U/li

ter)

0

2

4

6

8

10

Wei

ght g

ain

(g)

*******

FE

D

0

2

4

6

8

10 *****

***

0

5

10

15

20

25 **

******

2

4

6

8

10

***

******

0

200

400

600 ******

******

0

200

400

300

100

500 ****

******

1 cm

WT NS30 days

OB NS30 days

OB GF-Ala30 days

OB GF-Ala10 days

OB GF-Ala15 days

15 days

6th week

10 daysSacrificed

Sacrificed

Sacrificed

WT 1 group

OB 4 groups

Sacrificed

30 days

+ NS

+ G

F-A

la

30 days

WT 2 groups

OB 4 groups

15 days

6th week

+ G

F-A

la

+ NSSacrificed

Sacrificed

15 days

WT NS

WT GF-Ala H

OB NS

OB GF-Ala L

OB GF-Ala M

OB GF-Ala H

C

BA

IHG

OB NSWT GF-Ala HWT NS OB GF-Ala L OB GF-Ala M OB GF-Ala H

WT NS OB GF-Ala HOB NS

0.0

0.5

1.0

1.5

Rel

ativ

e he

patic

fat c

onte

nt

******

***

Hep

atic

FFA

(mm

ol/g

prot

)

Hep

atic

TG

(mm

ol/g

prot

)

0.00

0.02

0.04

0.06

0.08

0.10 ******

0.0

0.1

0.2

0.3

0.4

0.5 **

OB NS

OB GF-Ala L

OB GF-Ala M

OB GF-Ala H

K

J

L M

Fig. 2. The anti-hepatic steatosis effect of GF-Ala in ob/ob obese mice. (A) Schematic flow diagram of anti-hepatic steatosis experiment in vivo with different thera-peutic periods. OB, ob/ob obese mice; WT, wild-type mice. (B) Photographs of liver tissues after sacrificing by GF-Ala treatment with different therapeutic periods. Photo credit: Chen Zhou, Institute of Chemistry Chinese Academy of Sciences. (C) Schematic flow diagram of anti-hepatic steatosis experiment with different dosages of GF-Ala treatment. (D) Body weight gain, (E) adipose tissue weight, (F) adipose tissue–to–body weight ratio, (G) liver-to-body weight ratio, (H) serum ALT activity, (I) and serum AST activity after GF-Ala treatment. (J) Hepatic H&E staining images after GF-Ala treatment. Scale bars, 250 m (top) and 100 m (bottom). (K) Hepatic oil red O staining images and related hepatic fat content after GF-Ala treatment. Scale bars, 250 m (top) and 100 m (bottom). Levels of (L) hepatic FFA and (M) hepatic TG after GF-Ala treatment (L, low; M, medium; and H, high). *P < 0.05, **P < 0.01, ***P < 0.001.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 5: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

5 of 14

00

0

10

01

13

42220

311

0

Cellular component

Differential proteins (%)

Molecular function

Differential proteins (%)

Biological process

Differential proteins (%)

A

DC

B

Extracellular exosomeEndoplasmic reticulum

MitochondrionCytosol

Plasma membraneEndoplasmic reticulum membrane

PeroxisomeMicrobody

Mitochondrial membraneEndoplasmic reticulum lumen

Mitochondrial inner membraneGolgi apparatus

Cytosolic ribosomeMitochondrial matrix

Lipid particle

0 5 10 15 20 25 30 35 40 45

Protein bindingSmall-molecule binding

Ion bindingOxidoreductase activity

Nucleoside phosphate bindingHydrolase activity

Transferase activityProtein dimerization activity

Cofactor bindingLipid binding

Monooxygenase activityIsomerase activity

Ligase activityLipid transporter activity

Hormone binding

0 5 10 15 20 25 30 35Response to chemical

Organonitrogen compound metabolic processCofactor metabolic process

Lipid metabolic processRegulation of biological quality

Response to oxygen-containing compoundResponse to endogenous stimulus

Lipid biosynthetic processResponse to lipid

Lipid catabolic processNucleotide metabolic process

Response to drugDefense response

Response to oxidative stressLipid localization

Fatty acid beta-oxidationIon transport

Monosaccharide metabolic processLipid transort

Peroxisome organization

0 5 10 15 20 25 30 35

Lipid catabolismLipid biosynthesis

FE

−0.5 0 0.5Log10(expression fold)

OB NS vs. WT NS

OB GF-Ala vs. OB NS

WT GF-Ala vs. WT NS

OB GF-Ala vs. WT GF-Ala

Lipid transport

WT

NS

WT GF-Ala H OB NS OB GF-Ala HEstradiol 17-beta-dehydrogenase 11Cholinephosphotransferase 1Peroxisomal 2,4-dienoyl-CoA reductaseVery-long-chain (3R)-3-hydroxyacyl-CoA dehydratase 2NADH-cytochrome b5 reductase 317-beta-hydroxysteroid dehydrogenase 13Estradiol 17 beta-dehydrogenase 5Acetyl-coenzyme A synthetase, cytoplasmicAcetyl-CoA carboxylase 1Complement C3

ATP-binding cassette subfamily D member 3

Beta-1,4 N-acetylgalactosaminyltransferase 1Thyroid hormone-inducible hepatic proteinElongation of very long chain fatty acids protein 5RegucalcinATP-citrate synthaseLanosterol 14-alpha demethylaseFatty acid synthase

ATP-binding cassette subfamily D member 1Peroxisomal multifunctional enzyme type 2Cytochrome P450 4A10Isoamyl acetate-hydrolyzing esterase 1 homologGroup XIIB secretory phospholipase A2-like proteincGMP-inhibited 3~,5~-cyclic phosphodiesterase BMedium-chain specific acyl-CoA dehydrogenase, mitochondrialATP-binding cassette sub-family D member 2Peroxisomal acyl-coenzyme A oxidase 2Acyl-CoA dehydrogenase family member 11Carboxylesterase 1DPeroxisomal bifunctional enzymeMonoglyceride lipasePerilipin-2

NPC intracellular cholesterol transporter 2Prolow-density lipoprotein receptor–related protein 1

Apolipoprotein B-100Apolipoprotein A-IIPhosphatidylcholine transfer protein

Lipid catabolism

Lipid biosynthesis

Lipid transport

Acad11Acad11Acad11Abcd3Abcd3Abcd3Abcd1Abcd1Abcd1

Hsd17b13Hsd17b13Hsd17b13

Abcd2Abcd2Abcd2

PctpPctpPctp

Hsd17b4Hsd17b4Hsd17b4

EhhadhEhhadhEhhadhAcox2Acox2Acox2

Decr2Decr2Decr2

AcadmAcadmAcadm

Acss2Acss2Acss2

Pde3bPde3bPde3b

FasnFasnFasnAcacaAcacaAcaca Akr1c6Akr1c6Akr1c6

Elovl5Elovl5Elovl5Ces1dCes1dCes1d

Plin2Plin2Plin2ApobApobApobAclyAclyAcly

Hsd17b11Hsd17b11Hsd17b11 Lrp1Lrp1Lrp1

Iah1Iah1Iah1

Npc2Npc2Npc2 Apoa2Apoa2Apoa2

C3C3C3Cyb5r3Cyb5r3Cyb5r3

RgnRgnRgn

Fig. 3. The proteomics study of GF-Ala treatment in ob/ob obese mice. (A) Venn analysis of proteins identified and quantified in four groups. (B) Annotation of GO for cellular component category. (C) Annotation of GO for molecular functions category. (D) Annotation of GO for biological process category. (E) Heat map presentation of the differential proteins classified into three categories: lipid biosynthesis, lipid catabolism, and lipid transport, in OB NS versus WT NS, OB GF-Ala versus OB NS, WT GF-Ala versus WT NS, and OB GF-Ala versus WT GF-Ala groups (fold change, >1.2 or <0.84; P < 0.05). (F) Interaction network analysis of 37 differential proteins in lipid metabolism among WT NS, OB NS, and OB GF-Ala groups provided by STRING with median confidence (P < 0.400).

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 6: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

6 of 14

such as adenosine triphosphate (ATP)–citrate synthase (ACYL), acetyl- CoA synthetase, cytoplasmic (ACSS2), thyroid hormone-inducible hepatic protein (THRSP), and elongation of very long chain fatty acids protein 5 (ELOVL5), were also increased in OB mice and de-creased by GF-Ala treatment (fig. S4).

(ii) For lipid catabolism, cytochrome P450 4A10 (CYP4A10), ATP-binding cassette subfamily member 2 (ABCD2), carnitine palmitoyltransferase IA (CPT1A), acyl-CoA dehydrogenase for medium-chain fatty acids (ACADM), and peroxisomal bifunctional enzyme (EHHADH) were all notably up-regulated in OB mice as well as down-regulated by GF-Ala treatment (fig. S5).

(iii) For lipid transport, NPC intracellular cholesterol transporter 2 (NPC2), prolow-density lipoprotein receptor–related protein 1 (LRP1), and ApoB100 were decreased in hepatic steatosis mice, and all of them were elevated after GF-Ala treatment. Perilipin 2 (Plin2), a pro-tein associated with fat storage, was with a high expression in OB mice, which was distinctively reduced after GF-Ala treatment (fig. S6).

To highlight key proteins, STRING tool was used to detect the functional relations and networks between these proteins. Here, we obtained a network of above 37 differential proteins from the STRING database using the default setting (Fig. 3F). The gray lines between nodes represent functional associations between proteins. The darker the line is, the more interactions the proteins are. There are several highly interconnected nodes among the proteins of FASN, ACACA, ACADM, CYP4A10, EHHADH, Plin2, ApoB, LRP1, and

NPC2. All these proteins and relevant proteins were verified by Western blot (WB) in the following research.

Accelerative lipid transport in GF-Ala hepatic steatosis treatmentWe sought to ascertain differential proteins selected from the above analysis. First, the protein expressions of ACACA and FASN, related to fatty acid biosynthesis, exhibited no significant change before and after treatment in OB mice (fig. S7, A and B). CPT1A and ACADM were involved in the mitochondrial fatty acid -oxidation pathway (28), and the protein expressions of both two had no visible change after the treatment in OB mice (fig. S7, C and D). Nevertheless, EHHADH, the protein of peroxisomal fatty acid -oxidation pathway, was up-regulated in OB mice and significantly reduced by GF-Ala treatment (Fig. 4A). CYP4A10, participating in microsomal fatty acid -oxidation, was also increased in OB mice and decreased by GF-Ala treatment (Fig. 4B). All these changes indicate that GF-Ala effectively regulates the lipid catabolism by reducing fatty acid peroxisomal -oxidation and microsomal -oxidation. However, these differential proteins have positive regulations in lipid catabolic process, and the decline of them after treatment suggests that they are subsequent consequences instead of driving factors during the treatment of hepatic steatosis.

For lipid transport, LRP1, NPC2, MTTP, and LPL had almost no change after GF-Ala treatment (fig. S7, E to H). However, we found

0.0

0.2

0.4

0.6

WT N

S

OB NS

OB GF-A

la

Plin

2 pr

otei

n ***

WT N

S

OB NS

OB GF-A

la0.0

0.4

0.8

CY

P4A

10 p

rote

in ******

WT N

S

OB NS

OB GF-A

la0.0

0.5

EHH

AD

H p

rote

in 1.0**

0.0

0.5

1.0

1.5

WT N

S

OB NS

OB GF-A

la

**

Apo

B10

0 pr

otei

n

β-Actin

WT N

S

OB NS

OB GF-A

la

ApoB100ApoB48

509 kDa245 kDa

43 kDa

WT N

S

OB NS

OB GF-A

la

Plin2

GAPDH

47 kDa

37 kDa

EHHADH

GAPDH

WT N

S

OB NS

OB GF-A

la

79 kDa

37 kDa

CYP4A10

GAPDH

WT N

S

OB NS

OB GF-A

la

59 kDa

37 kDa

0.0

0.5

1.0

1.5

Rel

ativ

e m

RN

A e

xpre

ssio

n ofApob

**

WT N

S

WT G

F-Ala

OB NS

OB GF-A

la0.0

0.5

1.0

1.5

WT N

S

OB NS

OB GF-A

laApo

B48

pro

tein

*

C

BA

E

D

Fig. 4. The regulation of hepatic lipid catabolism and transport by GF-Ala in ob/ob obese mice. Protein expressions of (A) CYP4A10, (B) EHHADH, (C) Plin2, and (D) ApoB in the liver after GF-Ala treatment. (E) mRNA expression of Apob in the liver after GF-Ala treatment. *P < 0.05, **P < 0.01, ***P < 0.001.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 7: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

7 of 14

that Plin2, involved in the formation of lipid droplets, was highly expressed in OB mice but observably reduced after GF-Ala treat-ment (Fig. 4C). It might be the consequence of the reduction of hepatic lipid droplets. Besides, ApoB100 and ApoB48 in liver were highly reduced in OB mice, but both of them were enhanced by GF-Ala treatment (Fig. 4D). It is known that ApoB100 takes part in removing hepatic TG from liver to serum, and conversely, ApoB48 is used to transport TG from the diet into the liver. Thus, our results reveal that GF-Ala boosts hepatic TG transport and maintains the hepatic lipid metabolic homeostasis. However, the mRNA expres-sion of Apob in OB mice was lower than that in normal mice, which was not changed by GF-Ala treatment (Fig. 4E). It suggests that GF-Ala only regulates the protein expression of ApoB100 by inhibiting its posttranslational degradation instead of promoting its produc-tion at the source.

To uncover the above anti-hepatic steatosis mechanism in detail, we established a model of hepatocellular steatosis by incubating L02 hepatocytes with oleic acid (OA) (29). First, we tracked the subcellular localizations of GF-Ala by cryopreservation transmission electron microscope (cryo-TEM). GF-Ala nanoparticles (black arrows) are mainly distributed in endosomes (E) (fig. S8A). However, the mem-brane structure is missing in some of the endosomes (red arrows),

indicating that GF-Ala could escape from endosomes to the cyto-plasm. Some of GF-Ala nanoparticles are also distributed around ER. Then, to investigate the optimum concentration of OA, L02 cells were incubated with 0.5, 1, 1.5, and 2 mM OA for 24 hours, and the proportion of living cells was identified by measuring the Hoechst 33342 fluorescence intensity through microplate reader. It shows that the quantity of L02 cells has no change within 1.5 mM OA (fig. S8B). The relative lipid content in L02 cells was determined by Nile Red staining. We found that OA treatment significantly induced lipid accumulation in L02 cells, and the lipid content in L02 cells treated with 1.5 mM OA increased almost sixfold higher than that in nor-mal L02 cells (fig. S8C). Our study reveals that GF-Ala reverses the lipid accumulation in L02 cells, and the lipid content is decreased ca. 31.3% by 0.2 mM GF-Ala (Fig. 5A). TG content in the extracellular medium was determined at different times. GF-Ala significantly in-creased TG content in the medium after 24 hours (Fig. 5B). It indi-cates that the decrease of intracellular lipid accumulation results in the increase of extracellular TG after GF-Ala treatment. In addition, we measured the protein expressions of ApoB100 in OA-induced L02 cells at different times. We found that OA could enhance intracellular ApoB100 quickly (6 hours), which was degraded over time (24 hours). GF-Ala could inhibit the degradation of ApoB100 in OA- induced

Nile RedDCF

0.0

0.1

0.3

0.4

0.2

Hep

atoc

ellu

lar M

DA

(nm

ol/m

gpro

t)

*

OA (mM)GF-Ala (mM)

0 0 1.5 1.50 0.2 0 0.2

Rel

ativ

e lip

id a

ccum

ulat

ion

0

2

4

6

8

OA (mM)GF-Ala (mM)

0 0 1.5 1.5 1.5 1.50 0.4 0 0.1 0.2 0.4

******

***

1 10 102 103 104 105 106

62.0%

59.1%

96.1%

71.1%

Control

0.2 mM GF-Ala

1.5 mM OA

1.5 mM OA0.2 mM GF-Ala

MergeDCF

ApoB100

-Actin

Incubating time 24 h 6 h 12 h 24 h 24 h 6 h 12 h 24 h1.5 mM OA − + + + − + + +

0.2 mM GF-Ala − − − − + + + +

0.0

0.5

1.0

2.0

Apo

B10

0 pr

otei

n

1.5

0

2

3

4

5

1

1.5 mM OA0.2 mM GF-Ala

Incubating time 24 h 6 h 12 h 24 h 24 h 6 h 12 h 24 h

− + + + − + + +− − − − + + + +

Rel

ativ

e m

ediu

m T

G c

onte

nt

**

509 kDa

43 kDa

CBA

ED

Fig. 5. The inhibition of ApoB100 degradation and boost of TG transport by GF-Ala in OA-induced L02 cells. (A) Lipid accumulation of L02 cells treated with OA and GF-Ala for 24 hours at different concentrations. (B) Time course of TG content in the medium with or without GF-Ala treatment. (C) Time course of ApoB100 protein ex-pression with or without GF-Ala treatment in OA-induced L02 cells. (D) Flow cytometry data (left) and confocal fluorescence images (right) of ROS (DCF) and lipid (Nile Red) in L02 cells incubated with OA, GF-Ala, or together. Scale bars, 50 m. (E) Hepatocellular MDA content of L02 cells treated with OA, GF-Ala, or together for 24 hours. *P < 0.05, **P < 0.01, ***P < 0.001.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 8: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

8 of 14

L02 cells, consequently promoting TG transport (Fig. 5C). Consid-ering that the degradation of ApoB100 is known to be closely associated with ROS and lipid peroxidation, we determined the amount of intracellular ROS and malondialdehyde (MDA) after GF-Ala treatment. The lipid accumulation in OA-induced L02 cells increases ROS, appearing as stronger fluorescent signals of 2′-7′-dichlorodihydrofluorescein (DCF) in flow cytometry. Notably, GF-Ala could reduce excessive ROS in L02 cells (Fig. 5D). The similar decline of DCF and Nile Red signals after GF-Ala treatment was also observed by confocal microscopy (Fig. 5D), which positively supported the above results. In addition, our data also show that the content of MDA, one of the typical biomarkers of lipid peroxidation, is increased in OA-induced L02 cells but observably decreased by GF-Ala (Fig. 5E). Together, we reveal that GF-Ala could effectively improve hepatocellular steatosis and inhibit the degradation of ApoB100 by reducing excessive ROS and lipid peroxidation in OA-induced L02 cells.

Improvements of lipid peroxidation and abnormal mitochondria by GF-AlaThe hepatic lipid accumulation induces overloaded lipid catabolism, causing lipid peroxidation and mitochondrial dysfunction, finally leading to further hepatocyte damages. We measured the levels of serum MDA and hepatic MDA in OB mice. Both of them were in-creased in OB NS group (serum: 5.62 ± 1.00 nmol/ml and liver: 3.55 ± 0.30 nmol/mgprot) and decreased after GF-Ala treatment in different levels (serum: 2.52 ± 0.88 nmol/ml and liver: 3.01 ± 0.24 nmol/mgprot) (Fig. 6, A and B), showing similar results to OA-induced L02 cells. We used cryo-TEM to observe the microstructure of hepatocellular mitochondria. The images displayed that hepatic mitochondria ab-normally appeared with ruptured membranes, messy cristae, lost matrix, and internal vacuoles in the OB NS group (Fig. 6C). Inversely, seldom ruptured mitochondria were observed in OB GF-Ala group, and the structural integrity of mitochondria had been improved.

Except for the morphology change of mitochondria, more indexes of mitochondrial function were analyzed in OA-induced L02 cells. The stabilization of mitochondrial membrane potential (MMP) is the prerequisite for oxidative phosphorylation, ATP production, and other mitochondrial functions. The decline of MMP indicates the mitochondrial abnormality and early cellular apoptosis. We evaluated the MMP of L02 cells by JC-1, which exhibited higher green and lower red fluorescence intensity in decreased MMP. The OA model had a higher green/red ratio than control, and GF-Ala reversed this ratio (Fig. 6D), indicating that GF-Ala could inhibit the decline of MMP in OA-induced L02 cells. The components of NADH [reduced form of nicotinamide adenine dinucleotide (NAD+)] oxidative re-spiratory chain, including the activities of mitochondrial complexes I, III, and IV, were analyzed at the same time. The activities of com-plexes III (fig. S9A) and IV (fig. S9B) remain unchanged at normal levels in both OA model and GF-Ala treatment groups. However, OA model had higher complex I activity, which catalyzed NADH to NAD+ and generated ROS in hepatocytes. Notably, GF-Ala reinstated the complex I activity of OA-induced L02 cells (Fig. 6E). Then, we analyzed the content of NADH, which is produced in cellular respi-ration and participates in the respiratory chain. Both NADH and its oxidation state, NAD+, increased in the OA model and decreased by GF-Ala treatment (Fig. 6F). It indicates that excessive lipids induce hepatocellular mitochondria to produce more NADH and NAD+. However, GF-Ala could bring NADH content back to normal and

prevent mitochondria from metabolic stress effectively. Together, we reveal that GF-Ala could improve mitochondrial abnormality in hepatocytes suffering from hepatic steatosis.

In vivo biodistribution and toxicity of GF-AlaAt last, we studied the biodistribution of GF-Ala (intraperitoneal) by labeling it with Cyanine5.5 (Cy5.5) fluorescent dye and tracked GF-Ala with an in vivo fluorescent imaging system (fig. S10A). It was observed that GF-Ala mainly distributed into the pancreas, liver, spleen, lung, and kidney. The clear majority of GF-Ala could be ex-creted from organs, except for the pancreas, 30 days after injection. After the treatment, the pathology results of main organs (heart, spleen, lung, kidney, and pancreas) show no palpable pathological lesions or changes in GF-Ala–treated mice (fig. S10B), suggesting the low toxicity of GF-Ala in vivo.

DISCUSSIONThe defining feature of hepatic steatosis is the abnormal accumulation of hepatic TG, which induces to generate excessive ROS and lipid peroxidation. It will lead to the dysfunctions of hepatic lipogenesis, lipolysis, and lipid transport, aggravating hepatic lipid accumulation. Given the pathogenesis of hepatic steatosis, researchers explored numerous strategies for treatment. For example, Aramchol (the inhibitor of stearoyl CoA desaturase), ND-630, and NDI-010976 (the inhibitors of aceyl-CoA carboxylase) were tested in fatty liver patients to inhibit de novo lipogenesis and stimulate fatty acid oxidation (30, 31). At the same time, some other medications target peroxisome proliferator–activated receptors (PPARs) to regulate gluconeogenesis, -oxidation, and the hormone fibroblast growth factor (FGF)–21 (7, 32). In addition, several antioxidants have been used for fatty liver treatment. Vitamin E, a fat-soluble antioxidant, has been applied in multiple clinical trials for hepatic steatosis therapy, but it needs a long continuous course and high dose to take effect (33).

Fullerene [60] and gadofullerene (Gd@C82) with conjugated carbon cage structure and unique properties have been widely applied in the biomedicine field. The previous studies had reported that both fullerene and gadofullerene derivatives have superior antioxidant properties (20, 34, 35). Recently, we found that gadofullerene nanopar-ticles could prominently reduce blood glucose by regulating hepatic insulin resistance and reversing pancreatic abnormalities for type 2 diabetes mellitus treatment (24). Because both diabetes and hepatic steatosis belong to metabolic disorders, here, we attempted to exploit gadofullerene nanoparticles for hepatic steatosis therapy. We used one kind of dietary mouse model and another genetic mouse model, which could reflect hepatic histopathology and pathophysiology of human hepatic steatosis with different features. The WT mice were fed with fructose for 8 weeks, showing the feature of hepatic steatosis with massive lipid accumulation observed by oil red O staining. Similarly, the OB mice, having a mutation in the leptin gene, could also develop into the hepatic steatosis. GF-Ala observably reversed hepatic steatosis in diet-induced mice and OB mice (Fig. 2 and fig. S2). The effects were strongly confirmed by the improvements of hepatomegaly, hepatic lipid accumulation, and aminotransferase abnormalities.

Inspired by their anti-hepatic steatosis, we performed a proteomic profiling study to identify the protein targets in GF-Ala treatment. We found that GF-Ala could up-regulate lipid transport and down- regulate hepatic lipid catabolism. Boosting hepatic lipid transport is

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 9: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

9 of 14

012345

Hep

atic

MD

A(n

mol

/mgp

rot)

B

0

2

4

6

8

Seru

m M

DA

(nm

ol/m

l)

A C

Con

trol

JC-1 red JC-1 green Hoechst Merge

OA

mod

elG

F-A

la tr

eatm

ent

JC-1

fluo

resc

ence

inte

nsity

gree

n/re

d ra

tio

0.0

1.0

3.0

Contro

l

OA mod

el

GF-Ala

treatm

ent

D

E

0.0

0.5

1.0

1.5

NA

D+ /N

AD

H(n

mol

/mgp

rot)

ControlOA modelGF-Ala treatment

NAD+&NADH NAD+ NADH

****

***

**

OB

GF-

Ala

OB

NS

1 2

1 2

3

3

4

45

56

I

6

F

Contro

l

OA mod

el

GF-Ala

treatm

ent

Com

plex

I ac

tivity

(U/m

gpro

t)

0

40

60

80

20

**

******

WT N

S

WT G

F-Ala

OB NS

OB GF-A

la

**

WT N

S

WT G

F-Ala

OB NS

OB GF-A

la

2.0

*

Fig. 6. Improvements in lipid peroxidation and abnormal mitochondria by GF-Ala. Contents of (A) serum MDA and (B) hepatic MDA of OB mice after GF-Ala treat-ment. (C) Hepatic mitochondrial TEM images after GF-Ala treatment (red arrow, impaired mitochondrion with ruptured membrane; yellow arrow, impaired mitochondri-on without ruptured membrane; blue arrow, normal mitochondrion). Scale bars, 500 nm. (D) Confocal fluorescence images of MMP (JC-1) in L02 cells incubated with OA and GF-Ala. Scale bars, 50 m. (E) Activities of mitochondrial complex I in OA-induced L02 cells incubated with or without GF-Ala. (F) NAD+/NADH content in OA-induced L02 cells incubated with or without GF-Ala. *P < 0.05, **P < 0.01, ***P < 0.001.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 10: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

10 of 14

known as an available strategy for anti-hepatic steatosis. Nevertheless, few studies have been reported to treat hepatic steatosis regarding this solution. Some researches focus far more on cholesterol transport, as cholesterol is closely associated with cardiovascular diseases (36, 37). The lipid or lipoid transports rely on kinds of apolipoprotein, such as apoA-II for cholesterol transport, ApoB48 for dietary TG transport, and ApoB100 for hepatic TG transport. In our study, the protein expression of ApoB100 was decreased in OB mice and up-regulated by GF-Ala (Fig. 4D). However, the mRNA expression of hepatic Apob had no change in GF-Ala treatment, suggesting that GF-Ala did not influence the source of ApoB100. There are three posttrans-lational degradations of ApoB100 protein: ER association degradation (ERAD), post-ER presecretory proteolysis (PERPP), and receptor- mediated degradation (12). The severe scarcity of lipids stimulates ERAD during the translocation of ApoB100 across the ER, but hepatic lipids are overmuch in hepatic steatosis. For receptor-mediated degra-dation, our result shows no difference in hepatic LDL receptors before and after GF-Ala treatment (fig. S7H). PERPP is a nonproteasomal pathway caused by diverse metabolic factors, including excessive intracellular ROS. We established a hepatocellular model of hepatic steatosis to investigate the degradation of ApoB100 during GF-Ala treatment. The data show that GF-Ala could decrease the excessive ROS resulting from lipid accumulation and inhibit the degradation of ApoB100 in hepatocytes, most possibly through the PERPP pathway.

Notably, the weight gain of OB mice is reduced by GF-Ala treatment (fig. S2G). Hepatic steatosis is closely related to obesity in most cases. To ascertain the effect of weight loss on anti-hepatic steatosis, we used orlistat, an approved weight loss drug, to treat hepatic steatosis in OB mice. The results reveal that only weight loss alone by orlistat treatment could not effectively improve the hepatic steatosis in 15 days (fig. S3). Another anti-hepatic steatosis treatment in fructose-induced mice also shows that GF-Ala could markedly improve hepatic ste-atosis without apparent weight loss (fig. S2, B to D). Thus, we con-sider that weight loss is not a major factor during GF-Ala treatment. We would continue to study the mechanism of food intake reduc-tion and fates of TG after removing from the liver in the following research, which might expand the applications of gadofullerene nanomaterials in obesity or other metabolic diseases.

In addition, we demonstrate that GF-Ala could not only boost hepatic TG transport but also maintain the hepatic homeostasis. Specifically, the lipid peroxidation and mitochondrial abnormality in hepatic steatosis were improved by GF-Ala (Fig. 6). In hepatocytes, the lipid accumulation accelerates cellular respiration, which generates more NADH in hepatocytes. The activity of mitochondrial complex I is also promoted in NADH oxidative respiratory chain, generating more ROS in hepatocytes. Then, the excessive ROS lead to lipid peroxi-dation, and the mitochondrial structure and function are damaged gradually. Meanwhile, the lipid transport would be stimulated by ROS at first, but continual damages inhibit lipid transport finally. As a result, the disorders of hepatocellular lipid metabolism strengthen the lipid accumulation in hepatocytes. GF-Ala nanoparticles break this vicious circle, eliminate the excessive ROS, improve the mito-chondrial abnormality, and normalize the lipid transport.

Except for therapeutic effect, toxicity is equally important to anti-hepatic steatosis agents. We find that GF-Ala could be excreted gradually 30 days after administration without pathological lesions in the main organs (fig. S10).

In conclusion, we establish an effective solution for reversing hepatic steatosis by boosting TG transport using GF-Ala nanoparticles. GF-Ala

decreases the hepatic lipid accumulation in hepatic steatosis mice within 15 days. One of the most notable findings in our study is that GF-Ala chiefly accelerates the TG outflow from hepatocytes by in-hibiting the degradation of ApoB100. We also observe that GF-Ala effectively improves the mitochondrial abnormality and maintains the hepatic lipid metabolic homeostasis. Furthermore, we certify that GF-Ala could be almost all excreted from the living body and cause no severe toxicity. Our findings provide a promising anti-hepatic ste-atosis mechanism to activate hepatic TG transport and restore lipid metabolic homeostasis by GF-Ala. It offers a new perspective for the treatment of hepatic steatosis and obesity-associated metabolic dis-eases using advanced nanomaterials.

MATERIALS AND METHODSExperimental designThis study aimed to develop an anti-hepatic steatosis strategy based on fullerene nanomaterials and studied the potential therapeutic mechanism. For this, we performed GF-Ala, a superior antioxidant, in both diet-induced WT mice and obesity-associated transgenic ob/ob obese mice. The anti-hepatic steatosis effects were confirmed in several aspects, such as hepatic morphology, weight, pathology, and blood biochemistry. Then, we focused on the hepatic lipid homeostasis in treatment via proteomics and molecular biology. Given the improve-ment in lipid transport, a relevant cellular model was built and used for mechanism study between GF-Ala and ApoB degradation.

MaterialsSolid C60 and Gd@C82 (99% purity) were purchased from Xiamen Funano Co. Ltd. (Xiamen, China). Anhydrous ethanol, methanol, NaOH, H2O2, HNO3, NaH2PO4, and Na2HPO4 were purchased from Beijing Chemical Works (Beijing, China). NaCl, -alanine, OA, glutaraldehyde, osmium tetroxide, dithiothreitol, iodoacetamide, pentobarbital sodium, urea, acetonitrile, formic acid, dithiothreitol, iodoacetamide, tetraethylammonium bromide (TEAB), Cyanine5.5 (Cy5.5), and 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) were pur-chased from Sigma-Aldrich (St. Louis, MO, USA). Nile Red, Hoechst 33342, hematoxylin, eosin, oil red O, and optimal cutting temperature (OCT) compound were purchased from Solarbio (Beijing, China). All reagents and solvents were obtained commercially and used without further purification.

Preparation of -alanine derivative gadofullereneSynthesis of -alanine derivative gadofullerene (GF-Ala) or fullerene (C60-Ala) referred to the previous study in our laboratory. In short, 50 mg of Gd@C82 and 3.6 g of -alanine were dispersed in 14% NaOH aqueous solution during ultrasonic shaking. Then, the slurry was heated to 80°C and stirred for 2 hours. After cooling to room tem-perature, the solution was centrifuged (8000 rpm, 10 min) to re-move the insoluble substance and dialyzed against ultrapure water (Millipore, Billerica, USA) to get rid of the unreacted materials. At last, GF-Ala aqueous solution was obtained after filtering through a 220-nm pore size membrane. The concentration of GF-Ala, deter-mined as Gd element concentration, was measured with an induc-tively coupled plasma source mass spectrometer (ICP-MS; NexION 300X, PerkinElmer, USA) after enough nitration. Briefly, 0.1 ml of GF-Ala solution was added into 1 ml of HNO3 and heated to 100°C for 12 hours. Then, the solution was diluted and measured by the external standard method.

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 11: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

11 of 14

Characterizations of GF-AlaThe solid GF-Ala obtained by vacuum freeze-drying was characterized by TGA and EA. GF-Ala was heated from room temperature to 600°C (5°C/min) under the nitrogen flow by using a TGA spectrometer (Shimadzu DTG-60H, Japan). EA (Flash EA 1112, Thermo Fisher Scientific, USA) was performed normally to detect carbon, hydro-gen, and nitrogen contents in GF-Ala. The FTIR spectrum of GF-Ala was performed on Nicolet iN10 MX (Thermo Fisher Scientific, USA). Hydrodynamic size, zeta potential, and PDI of GF-Ala were measured using NanoZS ZEN3600 (Malvern Instruments, Enigma Business Park, Britain).

In vitro hydroxyl radical’s eliminationForty microliters of DMPO (100 mM) and 20 l of H2O2 (100 mM) were mixed with 20 l of ultrapure water, 20 l of GF-Ala aqueous solution (100 M), and 20 l of GF-Ala aqueous solution (200 M). The mixtures were irradiated with a 500-W UV lamp for 4 min first, and then the X-band electron paramagnetic resonance (EPR) spec-tra of DMPO-OH were recorded in the dark with an EPR spectrometer (FA-200, JEOL, Japan).

Animal models and treatmentsMale obese mice (OB, 5 weeks old) and their C57BL/6J lean litter-mates (WT, 5 weeks old) were purchased from Beijing Huafukang Bioscience Co. Inc. (Beijing, China). All mice were reared in a temperature-controlled environment with a 12-hour light-dark cycle. Besides, all mice were fed on sterile water and standard normal chow diet purchased from Beijing Huafukang Bioscience Co. Inc. (Beijing, China). All the experimental protocols involving live ani-mals were reviewed and approved by the Animal Ethics Committee of Institute of Processes, Chinese Academy of Sciences. Different concentrations of GF-Ala were regulated to appropriate osmotic pressure via sodium chloride before treatment. Three animal exper-iments were performed in this work.

First, C57BL/6J mice (male, 6 weeks old; n = 12) were fed with drinking water containing 30% fructose and standard normal chow diet ad libitum. After 8 weeks, mice were divided into two groups randomly and treated with GF-Ala (6 mol/kg, ip) or NS for 15 days. Another six C57BL/6J mice (male, 6 weeks old) were fed with sterile water and a normal diet until 8 weeks old and treated with NS for 15 days. After treatment, the mice were anesthetized and sacrificed. The livers were harvested, photographed, and fixed with 4% para-formaldehyde for histopathological examination.

Then, we preliminarily studied the treatment time of GF-Ala for hepatic steatosis on OB mice. After 1 week of feeding, the OB mice were randomly divided into four groups (n = 3, each group), and one of them was treated with NS every day by intraperitoneal injec-tion (5 ml/kg) for 30 days. GF-Ala (1.2 mM) was performed on the other three groups every day by intraperitoneal injection (5 ml/kg) for 10, 15, and 30 days. Three WT mice were treated with NS every day by intraperitoneal injection (5 ml/kg) for 30 days in the mean-time. After treatment (on the 10th, 15th, and 30th day), the mice were anesthetized and sacrificed. The livers were harvested, photo-graphed, and fixed with 4% paraformaldehyde for histopathological examination.

Another 24 OB mice and 12 WT mice were prepared for re-searches on the dosage and anti-hepatic steatosis mechanism of GF-Ala. After 1 week of feeding, these OB mice and WT mice were randomly divided into four groups and two groups (n = 6, each

group). (i) Blank control group (WT NS): intraperitoneal injection of NS (5 ml/kg) every day; (ii) GF-Ala control group (WT GF-Ala H): intraperitoneal injection of GF-Ala (1.2 mM, 5 ml/kg) every day; (iii) model group (OB NS): intraperitoneal injection of NS (5 ml/kg) every day; (iv) GF-Ala L group (OB GF-Ala L): intraperitoneal in-jection of GF-Ala (0.3 mM, 5 ml/kg) every day; (v) GF-Ala M group (OB GF-Ala M): intraperitoneal injection of GF-Ala (0.6 mM, 5 ml/kg) every day; (vi) GF-Ala H group (OB GF-Ala H): intraperitoneal in-jection of GF-Ala (1.2 mM, 5 ml/kg) every day. The body weights of each group were recorded every 3 days. After 15 days of treatment, all mice were anesthetized, sacrificed, and harvested with blood, liver, adipose tissue, heart, spleen, lung, kidney, and pancreas. Blood was collected and centrifuged (1200 rpm, 15 min) to gather serum for the biochemical test. Liver and adipose tissue (perirenal, subcutaneous, and gonads) were weighted, snap-frozen, and stored at −80°C. Heart, liver, spleen, lung, kidney, and pancreas of each mouse were fixed with 4% paraformaldehyde for histopathological examination. Besides, portions of liver in WT NS, OB NS, and OB GF-Ala M groups were fixed in 2.5% glutaraldehyde for TEM observation.

Histopathology examinationH&E staining: Fixed tissues, including heart, liver, spleen, lung, kidney, and pancreas, were dehydrated through a serial alcohol gradient and embedded in paraffin wax blocks. Then, 5-m-thick sections were sliced, and standard methods were followed, including dewax-ing, hydration, staining, dehydration, and sealing.

Oil red O staining: Fixed hepatic tissues were dealt with 15 and 30% sucrose solution overnight successively. Then, we embedded hepatic tissues in OCT and sliced embedded tissues into 6- to 8-m sections by freezing microtome. Afterward, standard methods were followed, including washing, staining, differentiation, and sealing.

Both H&E staining and oil red O staining were scanned with a scanner (NanoZoomer-SQ, HAMAMATSU, Japan). Average optical density analysis of oil red O staining was performed by Image-Pro Plus 6.0 (Media Cybernetics, USA).

Biochemical analysisIn cells, MDA was extracted and measured using an MDA assay kit (Nanjing Jiancheng Bioengineering Institute, China) according to the manufacturer’s instructions. TG was measured by using TG assay kits (Nanjing Jiancheng Bioengineering Institute, China) as described.

In serum, alanine aminotransferase (ALT), aspartate amino-transferase (AST), TG, and cholesterol were all detected with an automatic biochemical analyzer (TBA-2000FR, Toshiba, Japan) with corresponding assay kits (Nanjing Jiancheng Bioengineering Institute, China). According to the manufacturer’s instructions, MDA was measured using an MDA assay kit (Nanjing Jiancheng Bioengineering Institute, China).

In liver tissues, FFA, TG, total glutathione (tGSH), and MDA were measured by using corresponding assay kits (Nanjing Jiancheng Bioengineering Institute, China) as described. All these biochemical analyses were performed following the manufacturer’s instructions.

Proteomics experiment and bioinformatics methodsThe hepatic sample (in WT NS, OB NS, and OB GF-Ala H groups in the second animal experiment) was grinded by liquid nitrogen into cell powder and then transferred to a 5-ml centrifuge tube. After that, four volumes of lysis buffer (8 M urea, 1% Protease Inhibitor Cocktail) were added to the powder, followed by sonication three

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 12: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

12 of 14

times on ice using a high-intensity ultrasonic processor (Scientz, China). The remaining debris was removed by centrifugation at 12,000g at 4°C for 10 min. Last, the supernatant was collected, and the protein concentration was determined with a BCA kit (Solarbio, China) according to the manufacturer’s instructions.

For digestion, the protein solution was reduced with 5 mM dithio-threitol for 30 min at 56°C and alkylated with 11 mM iodoacetamide for 15 min at room temperature in darkness. The protein sample was then diluted by adding 100 mM TEAB to urea concentration less than 2 M. Last, trypsin (Corning, USA) was added at 1:50 trypsin- to-protein mass ratio for the first digestion overnight and 1:100 trypsin-to-protein mass ratio for a second 4-hour digestion.

After trypsin digestion, the peptide was desalted by Strata X C18 Solid Phase Extraction (SPE) column (Phenomenex, USA) and vacuum- dried. The peptide was reconstituted in 0.5 M TEAB and processed according to the manufacturer’s protocol for TMT kit (Thermo Fisher Scientific, USA). Briefly, one unit of TMT reagent was thawed and reconstituted in acetonitrile. The peptide mixtures were then incubated for 2 hours at room temperature and pooled, desalted, and dried by vacuum centrifugation.

The tryptic peptides were fractionated into fractions by high-pH reversed-phase high-performance liquid chromatography (HPLC) using Agilent 300Extend C18 column [5 m particles, 4.6 mm in-side diameter (ID), and 250 mm length]. Briefly, peptides were first separated with a gradient of 8 to 32% acetonitrile (pH 9.0) over 60 min into 60 fractions. Then, the peptides were combined into 18 fractions and dried by vacuum centrifugation.

The tryptic peptides were dissolved in 0.1% formic acid and directly loaded onto a homemade reversed-phase analytical column (15-cm length, 75 m ID). The gradient was composed of an increase from 6 to 23% solvent B (0.1% formic acid in 98% acetonitrile) over 26 min, 23 to 35% in 8 min and climbing to 80% in 3 min, then holding at 80% for the last 3 min, all at a constant flow rate of 400 nl/min on an EASY-nLC 1000 UPLC system.

The peptides were subjected to nanospray ion (NSI) source fol-lowed by tandem mass spectrometry (MS/MS) in Q Exactive Plus (Thermo Fisher Scientific, USA) coupled online to the UPLC. The electrospray voltage applied was 2.0 kV. The mass/charge ratio (m/z) scan range was 350 to 1800 for full scan, and intact peptides were detected in the Orbitrap at a resolution of 70,000. Peptides were then selected for MS/MS using normalized collision energy (NCE) setting as 28, and the fragments were detected in the Orbitrap at a resolution of 17,500. A data-dependent procedure alternated between one MS scan followed by 20 MS/MS scans with 15.0-s dynamic exclusion. Automatic gain control (AGC) was set at 5E4. Fixed first mass was set to 100 m/z.

The resulting MS/MS data were processed using MaxQuant search engine (v.1.5.2.8). Tandem mass spectra were searched against UniProt Muridae database concatenated with reverse decoy data-base. Trypsin/P was specified as cleavage enzyme allowing up to two missing cleavages. The mass tolerance for precursor ions was set to 20 parts per million (ppm) in the first search and 5 ppm in the main search, and the mass tolerance for fragment ions was set to 0.02 Da. Carbamidomethyl on cysteine was specified as fixed mod-ification, and oxidation on Met was specified as variable modifications. False discovery rate (FDR) was adjusted to <1%, and the minimum score for peptides was set to >40.

GO annotation proteome was derived from the UniProt-GOA database (www.ebi.ac.uk/GOA/), first converting identified protein ID to UniProt ID and then mapping to GO IDs by protein ID.

If some identified proteins were not annotated by UniProt-GOA database, the InterProScan soft would be used to annotated protein’s GO functional based on the protein sequence alignment method. Then, proteins were classified by GO annotation based on three cate-gories: biological process, cellular component, and molecular function.

KEGG database was used to annotate the protein pathway, first using KEGG Automatic Annotation Server (KAAS) to annotated pro-tein’s KEGG database description and then mapping the annotation result on the KEGG pathway database using KEGG online service tools KEGG mapper (www.genome.jp/kegg/tool/map_pathway2.html). STING v.10 online database platform (http://string-db.org/) was used for network interaction analysis of differential proteins.

WB analysisAntibodies for WB were purchased from Abcam (Cambridge, UK), including anti-ApoB (ab20737), anti-CYP450A (ab140635), anti- EHHADH (ab93172), anti-Plin2 (ab52356), anti-ACACA (ab45174), anti-FASN (ab22759), anti-CPT1A (ab128568), anti-ACADM (ab110296), anti-LRP1 (ab92544), anti-NPC2 (ab114047), anti- MTTP (ab75316), anti-LPL (ab21356), and corresponding secondary anti-bodies. SDS, tris, glycine, and skim milk powder were purchased from BioDee (Beijing, China).

The proteins in livers were extracted with radioimmunoprecipi-tation assay (RIPA) lysis buffer (Beyotime, China) and quantified with a BCA protein assay kit (Beyotime, China) as described. Proteins were separated by electrophoresis. Briefly, polyacrylamide gel elec-trophoresis gels were prepared, and proteins were mixed with loading buffer with a 5-min heat at 95°C. Each sample was loaded with 20 g of total protein per mini-gel well. The gels were submerged in run-ning buffer [25 mM tris base, 190 mM glycine, 0.1% SDS (pH 8.3)]. Then, we ran the gel and stopped when the dye reached the bottom of the gel. Next, we transferred the proteins from gel to polyvinylidene fluoride membrane (Millipore, USA) by wet transfer. The transfer buffer is 48 mM tris, 39 mM glycine, 0.04% SDS, and 20% methanol. After transfer, we blocked the membrane with 5% nonfat milk for 1 hour at 4°C under agitation. Following the rinsing, we incubated with the primary antibodies at default dilution at 4°C overnight. On the next day, we rinsed the membrane and incubated with the secondary anti-bodies at default dilution at 4°C for 1 hour. After rinsing again, pro-teins were visualized by electrochemiluminescence (Millipore, USA) and imaging system (Tanon 4200, Tanon, China). The gray analysis was performed on Gel Image System version 4.00 (Tanon, China).

Quantitative real-time polymerase chain reaction measurementsTotal RNA in livers was extracted with RNA isolator total RNA extraction reagent (Vazyme, Nanjing, China), and their complemen-tary DNA (cDNA) was converted with HiScript Q RT SuperMix for polymerase chain reaction (PCR) (Vazyme, Nanjing, China). Then, the obtained cDNA was amplified with AceQ qPCR SYBR Green Master Mix (Vazyme, Nanjing, China), and relative expressions of Apob were quantified using a quantitative PCR (qPCR) detection system (CFX96, Bio-Rad, USA). Apob forward primer: 5′-TCCAG-ACAACCTCTTCCTAAAGAC-3′; Apob reverse primer: 5′-GGAT-GTCAATGTTTATTTTGTTCCT-3′.

Cell culture and OA-induced cellular model treatmentL02 cells were purchased from Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China). They were

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 13: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

13 of 14

cultured with RPMI 1640 supplemented with 10% (v/v) FBS, 1% (v/v) penicillin (100 g/ml), and 1% (v/v) streptomycin (100 g/ml) in a 5% CO2, 37°C incubator.

The OA stock solution was prepared by dissolving 0.4 mmol of OA into 40 ml of phosphate-buffered saline (PBS) solution containing 5% (w/v) fatty acid–free bovine serum albumin (BSA) (Equitech- Bio Inc., USA) and stored at −20°C. The culture medium was used to dilute OA stock solution to a certain concentration, and another 5% (w/v) BSA solution without OA was diluted in the same multiples as control.

L02 cells in 96-well plate were incubated with 0, 0.5, 1.0, 1.5, and 2.0 mM OA for 24 hours, respectively. Then, the media were re-moved, washed twice with PBS, and stained with Nile Red (excitation wavelength: 488 nm; emission wavelength: 550 nm) and Hoechst 33342 (excitation wavelength: 350 nm; emission wavelength: 461 nm) for 30 min. After that, the media were washed twice and measured with Infinite M1000 Pro (Tecan, Switzerland). The relative lipid content of each group was equal to the normalized ratio of fluores-cence intensity between Nile Red and Hoechst. GF-Ala (0, 0.1, 0.2, and 0.4 mM) was coincubated with 1.5 mM OA–induced L02 cells for 24 hours, and lipid contents were measured using the same method.

In vivo ROS eliminationL02 cells were incubated with 1.5 mM OA, 0.2 mM GF-Ala, or to-gether for 24 hours. Then, cells were collected, washed with PBS, and stained with Nile Red (excitation wavelength: 488 nm; emission wavelength: 550 nm) and DCF diacetate (DCF-DA) (excitation wavelength: 488 nm; emission wavelength: 525 nm) for 30 min. After that, the cells were washed twice and measured with a flow cytometer (Attune NxT, Invitrogen, USA). Single-dye samples were prepared for compensation regulation. The same samples were also prepared for imaging with a confocal laser scanning microscope (FV1000-IX81, Olympus, Japan).

Mitochondrial function study in cellL02 cells were incubated with 1.5 mM OA, 0.2 mM GF-Ala, or together for 24 hours. Then, the media were removed, washed twice with PBS, and orderly stained with Hoechst 33342 (excitation wave-length: 350 nm; emission wavelength: 461 nm) and JC-1 (excitation wavelength: 515 nm, emission wavelength: 529 nm; excitation wave-length: 585 nm, emission wavelength: 590 nm) (Beyotime, China). Then, samples are imaged with a confocal laser scanning microscope (FV1000-IX81, Olympus, Japan). The fluorescence intensity analysis was performed by Image-Pro Plus 6.0 (Media Cybernetics, USA).

The NAD+/NADH Kit was purchased from Beyotime (Shanghai, China), and the activities of mitochondrial complex I, III, and IV kits were purchased from Solarbio (Beijing, China). L02 cells were incubated with 1.5 mM OA, 0.2 mM GF-Ala, or together for 24 hours. Then, cells were collected and washed with PBS, and the above analysis was performed following the manufacturer’s instructions.

TG transport study in cellL02 cells were treated with 1.5 mM OA for 6, 12, or 24 hours with or without 0.2 mM GF-Ala. Then, the media were removed and washed twice with PBS, and a normal medium was added to each group. After 2 hours, the media were collected for the TG test, and cells were incubated with Hoechst 33342 to quantify the relative cell numbers.

ApoB100 degradation study in cellL02 cells were treated with 1.5 mM OA for 6, 12, or 24 hours with or without 0.2 mM GF-Ala. Then, the media were removed and washed twice with PBS. The proteins in cells were extracted with RIPA lysis buffer (Beyotime, China) and quantified with a BCA protein assay kit (Beyotime, China) as described. Then, the expression of ApoB100 protein was analyzed by WB.

In vivo fluorescence imagingGF-Ala was labeled with Cy5.5 fluorescent dye by condensation re-action under room temperature and stirring. After purification and quantitation, WT mice (C57BL/6J, n = 4) were injected with GF-Ala–Cy5.5 (1.2 mM, 5 ml/kg, ip) once. Mice were anesthetized, sac-rificed, and harvested with heart, liver, spleen, lung, kidney, pancreas, and intestines on the 1st, 5th, 12th, and 30th day. Then, an in vivo fluorescent imaging system (Spectrum CT, PerkinElmer, USA) was used to study the biodistribution of GF-Ala–Cy5.5 (excitation wave-length: 675 nm; emission wavelength: 720 nm). Blank control was also performed in the meantime.

TEM of liver tissues and L02 cellsL02 cells were incubated with 0.4 mM GF-Ala for 3 hours. Then, we collected the cells and fixed them in 2.5% glutaraldehyde. Both cells and liver tissues fixed in 2.5% glutaraldehyde were steeped in 1% osmium tetroxide for 2 hours. Then, we used a serial alcohol gradi-ent with increasing concentrations to dehydrate samples. After that, samples were embedded in Spi-pon812 resin, sectioned into ap-proximately 70-nm slices, and then stained by uranyl acetate and lead citrate. Last, we observed these slices by TEM (Tecnai Spirit 120 kV, FEI, USA).

Statistical analysisData were provided as means ± SD. Comparisons between two groups or several groups were analyzed using unpaired Student’s t test or one-way analysis of variance (ANOVA) by using IBM SPSS Statistics 19 (IBM, USA). *P < 0.05, **P < 0.01, and ***P < 0.001 were considered the statistical significance of differences.

SUPPLEMENTARY MATERIALSSupplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/37/eabc1586/DC1

View/request a protocol for this paper from Bio-protocol.

REFERENCES AND NOTES 1. Z. M. Younossi, A. B. Koenig, D. Abdelatif, Y. Fazel, L. Henry, M. Wymer, Global epidemiology

of nonalcoholic fatty liver disease-meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology 64, 73–84 (2016).

2. Y. Yu, J. Cai, Z. She, H. Li, Insights into the epidemiology, pathogenesis, and therapeutics of nonalcoholic fatty liver diseases. Adv. Sci. 6, 1801585 (2019).

3. S. L. Friedman, B. A. Neuschwander-Tetri, M. Rinella, A. J. Sanyal, Mechanisms of NAFLD development and therapeutic strategies. Nat. Med. 24, 908–922 (2018).

4. J. C. Cohen, J. D. Horton, H. H. Hobbs, Human fatty liver disease: Old questions and new insights. Science 332, 1519–1523 (2011).

5. D. W. Foster, Malonyl-CoA: The regulator of fatty acid synthesis and oxidation. J. Clin. Invest. 122, 1958–1959 (2012).

6. E. E. Cable, P. D. Finn, J. W. Stebbins, J. Hou, B. R. Ito, P. D. van Poelje, D. L. Linemeyer, M. D. Erion, Reduction of hepatic steatosis in rats and mice after treatment with a liver-targeted thyroid hormone receptor agonist. Hepatology 49, 407–417 (2009).

7. U. Risérus, D. Sprecher, T. Johnson, E. Olson, S. Hirschberg, A. Liu, Z. Fang, P. Hegde, D. Richards, L. Sarov-Blat, J. C. Strum, S. Basu, J. Cheeseman, B. A. Fielding, S. M. Humphreys, T. Danoff, N. R. Moore, P. Murgatroyd, S. O'Rahilly, P. Sutton, T. Willson, D. Hassall, K. N. Frayn, F. Karpe, Activation of peroxisome proliferator-activated receptor

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 14: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

Zhou et al., Sci. Adv. 2020; 6 : eabc1586 11 September 2020

S C I E N C E A D V A N C E S | R E S E A R C H A R T I C L E

14 of 14

(PPAR) promotes reversal of multiple metabolic abnormalities, reduces oxidative stress, and increases fatly acid oxidation in moderately obese men. Diabetes 57, 332–339 (2008).

8. G. Reyes-Soffer, B. Moon, A. Hernandez-Ono, M. Dionizovick-Dimanovski, J. Jimenez, J. Obunike, T. Thomas, C. Ngai, N. Fontanez, D. S. Donovan, W. Karmally, S. Holleran, R. Ramakrishnan, R. S. Mittleman, H. N. Ginsberg, Complex effects of inhibiting hepatic apolipoprotein B100 synthesis in humans. Sci. Transl. Med. 8, 323ra312 (2016).

9. G. F. Gibbons, Assembly and secretion of hepatic very-low-density lipoprotein. Biochem. J. 268, 1–13 (1990).

10. S. Uchiyama, T. Shimizu, T. Shirasawa, CuZn-SOD deficiency causes ApoB degradation and induces hepatic lipid accumulation by impaired lipoprotein secretion in mice. J. Biol. Chem. 281, 31713–31719 (2006).

11. M. H. Pan, A. I. Cederbaum, Y.-L. Zhang, H. N. Ginsberg, K. J. Williams, E. A. Fisher, Lipid peroxidation and oxidant stress regulate hepatic apolipoprotein B degradation and VLDL production. J. Clin. Invest. 113, 1277–1287 (2004).

12. M. Pan, V. Maitin, S. Parathath, U. Andreo, S. X. Lin, C. S. Germain, Z. Yao, F. R. Maxfield, K. J. Williams, E. A. Fisher, Presecretory oxidation, aggregation, and autophagic destruction of apoprotein-B: A pathway for late-stage quality control. Proc. Natl. Acad. Sci. U.S.A. 105, 5862–5867 (2008).

13. O. F. Khan, P. S. Kowalski, J. C. Doloff, J. K. Tsosie, V. Bakthavatchalu, C. B. Winn, J. Haupt, M. Jamiel, R. Langer, D. G. Anderson, Endothelial siRNA delivery in nonhuman primates using ionizable low-molecular weight polymeric nanoparticles. Sci. Adv. 4, eaar8409 (2018).

14. J. R. Clegg, A. S. Irani, E. W. Ander, C. M. Ludolph, A. K. Venkataraman, J. X. Zhong, N. A. Peppas, Synthetic networks with tunable responsiveness, biodegradation, and molecular recognition for precision medicine applications. Sci. Adv. 5, aax7946 (2019).

15. X.-L. Hu, Y. Zang, J. Li, G.-R. Chen, T. D. James, X.-P. He, H. Tian, Targeted multimodal theranostics via biorecognition controlled aggregation of metallic nanoparticle composites. Chem. Sci. 7, 4004–4008 (2016).

16. X. Li, M. Zhen, R. Deng, T. Yu, J. Li, Y. Zhang, T. Zou, Y. Zhou, Z. Lu, M. Guan, H. Xu, C. Shu, C. Wang, RF-assisted gadofullerene nanoparticles induces rapid tumor vascular disruption by down-expression of tumor vascular endothelial cadherin. Biomaterials 163, 142–153 (2018).

17. Y. Liu, C. Chen, P. Qian, X. Lu, B. Sun, X. Zhang, L. Wang, X. Gao, H. Li, Z. Chen, J. Tang, W. Zhang, J. Dong, R. Bai, P. E. Lobie, Q. Wu, S. Liu, H. Zhang, F. Zhao, M. S. Wicha, T. Zhu, Y. Zhao, Gd-metallofullerenol nanomaterial as non-toxic breast cancer stem cell-specific inhibitor. Nat. Commun. 6, 5988 (2015).

18. Y. Zhou, M. Zhen, H. Ma, J. Li, C. Shu, C. Wang, Inhalable gadofullerenol/[70] fullerenol as high-efficiency ROS scavengers for pulmonary fibrosis therapy. Nanomedicine 14, 1361–1369 (2018).

19. J.-J. Yin, F. Lao, P. P. Fu, W. G. Wamer, Y. Zhao, P. C. Wang, Y. Qiu, B. Sun, G. Xing, J. Dong, X.-J. Liang, C. Chen, The scavenging of reactive oxygen species and the potential for cell protection by functionalized fullerene materials. Biomaterials 30, 611–621 (2009).

20. J. Li, M. Guan, T. Wang, M. Zhen, F. Zhao, C. Shu, C. Wang, Gd@C82-(ethylenediamine)8 nanoparticle: A new high-efficiency water -soluble ROS scavenger. ACS Appl. Mater. Inter. 8, 25770–25776 (2016).

21. W. Li, L. Zhao, T. Wei, Y. Zhao, C. Chen, The inhibition of death receptor mediated apoptosis through lysosome stabilization following internalization of carboxyfullerene nanoparticles. Biomaterials 32, 4030–4041 (2011).

22. F. Lao, L. Chen, W. Li, C. Ge, Y. Qu, Q. Sun, Y. Zhao, D. Han, C. Chen, Fullerene nanoparticles selectively enter oxidation-damaged cerebral microvessel endothelial cells and inhibit JNK-related apoptosis. ACS Nano 3, 3358–3368 (2009).

23. T. Hirai, Y. Yoshioka, A. Udaka, E. Uemura, T. Ohe, H. Aoshima, J.-Q. Gao, K. Kokubo, T. Oshima, K. Nagano, K. Higashisaka, T. Mashino, Y. Tsutsumi, Potential suppressive effects of two C60 fullerene derivatives on acquired immunity. Nanoscale Res. Lett. 11, 449 (2016).

24. X. Li, M. Zhen, C. Zhou, R. Deng, T. Yu, Y. Wu, C. Shu, C. Wang, C. Bai, Gadofullerene nanoparticles reverse dysfunctions of pancreas and improve hepatic insulin resistance for type 2 diabetes mellitus treatment. ACS Nano 13, 8597–8608 (2019).

25. Y. Zhou, R. Deng, M. Zhen, J. Li, M. Guan, W. Jia, X. Li, Y. Zhang, T. Yu, T. Zou, Z. Lu, J. Guo, L. Sun, C. Shu, C. Wang, Amino acid functionalized gadofullerene nanoparticles with superior antitumor activity via destruction of tumor vasculature in vivo. Biomaterials 133, 107–118 (2017).

26. S. Satapati, B. Kucejova, J. A. G. Duarte, J. A. Fletcher, L. Reynolds, N. E. Sunny, T. He, L. A. Nair, K. A. Livingston, X. Fu, M. E. Merritt, A. D. Sherry, C. R. Malloy, J. M. Shelton, J. Lambert, E. J. Parks, I. Corbin, M. A. Magnuson, J. D. Browning, S. C. Burgess, Mitochondrial metabolism mediates oxidative stress and inflammation in fatty liver. J. Clin. Invest. 125, 4447–4462 (2015).

27. T. Hyötyläinen, L. Jerby, E. M. Petäjä, I. Mattila, S. Jäntti, P. Auvinen, A. Gastaldelli, H. Yki-Järvinen, E. Ruppin, M. Orešič, Genome-scale study reveals reduced metabolic adaptability in patients with non-alcoholic fatty liver disease. Nat. Commun. 7, 8994 (2016).

28. J. Dai, K. Liang, S. Zhao, W. Jia, Y. Liu, H. Wu, J. Lv, C. Cao, T. Chen, S. Zhuang, X. Hou, S. Zhou, X. Zhang, X.-W. Chen, Y. Huang, R.-P. Xiao, Y.-L. Wang, T. Luo, J. Xiao, C. Wang, Chemoproteomics reveals baicalin activates hepatic CPT1 to ameliorate diet-induced obesity and hepatic steatosis. Proc. Natl. Acad. Sci. U.S.A. 115, E5896–E5905 (2018).

29. M. J. Gómez-Lechón, M. T. Donato, A. Martínez-Romero, N. Jiménez, J. V. Castell, J.-E. O’Connor, A human hepatocellular in vitro model to investigate steatosis. Chem. Biol. Interact. 165, 106–116 (2007).

30. W. F. Westlin, H. Blanchette, G. Harriman, H. J. Harwood, R. Kapeller, S. Lennon, W. Miao, C. Beysen, M. Dalidd, S. Turner, K. Stiede, T. Schmalbach, NDI-010976, a potent, liver-directed, oral inhibitor of acetyl CoA carboxylase for non-alcoholic steatohepatitis: Pharmacodynamic effects on hepatic de novo lipogenesis in obese but otherwise healthy adult male volunteers. J. Hepatol. 64, S190–S191 (2016).

31. G. Harriman, J. Greenwood, S. Bhat, X. Huang, R. Wang, D. Paul, L. Tong, A. K. Saha, W. F. Westlin, R. Kapeller, H. J. Harwood Jr., Acetyl-CoA carboxylase inhibition by ND-630 reduces hepatic steatosis, improves insulin sensitivity, and modulates dyslipidemia in rats. Proc. Natl. Acad. Sci. U.S.A. 113, E1796–E1805 (2016).

32. Y. J. Lee, E. H. Ko, J. E. Kim, E. Kim, H. Lee, H. Choi, J. H. Yu, H. J. Kim, J.-K. Seong, K.-S. Kim, J.-W. Kim, Nuclear receptor PPAR-regulated monoacylglycerol O-acyltransferase 1 (MGAT1) expression is responsible for the lipid accumulation in diet-induced hepatic steatosis. Proc. Natl. Acad. Sci. U.S.A. 109, 13656–13661 (2012).

33. A. J. Sanyal, N. Chalasani, K. V. Kowdley, A. McCullough, A. M. Diehl, N. M. Bass, B. A. Neuschwander-Tetri, J. E. Lavine, J. Tonascia, A. Unalp, M. Van Natta, J. Clark, E. M. Brunt, D. E. Kleiner, J. H. Hoofnagle, P. R. Robuck; NASH CRN, Pioglitazone, vitamin E, or placebo for nonalcoholic steatohepatitis. N. Engl. J. Med. 362, 1675–1685 (2010).

34. Y. Zhang, C. Shu, M. Zhen, J. Li, T. Yu, W. Jia, X. Li, R. Deng, Y. Zhou, C. Wang, A novel bone marrow targeted gadofullerene agent protect against oxidative injury in chemotherapy. Sci. China Mater. 60, 866–880 (2017).

35. Y. Zhou, J. Li, H. Ma, M. Zhen, J. Guo, L. Wang, L. Jiang, C. Shu, C. Wang, Biocompatible [60]/[70] Fullerenols: Potent defense against oxidative injury induced by reduplicative chemotherapy. ACS Appl. Mater. Inter. 9, 35539–35547 (2017).

36. C. Liu, Q. Guo, M. Lu, Y. Li, An experimental study on amelioration of dyslipidemia-induced atherosclesis by Clematichinenoside through regulating peroxisome proliferator-activated receptor- mediated apolipoprotein A-I, A-II and C-III. Eur. J. Pharmacol. 761, 362–374 (2015).

37. G. Silbernagel, B. Genser, C. Drechsler, H. Scharnagl, T. B. Grammer, T. Stojakovic, V. Krane, E. Ritz, C. Wanner, W. Mäerz, HDL cholesterol, apolipoproteins, and cardiovascular risk in hemodialysis patients. J. Am. Soc. Nephrol. 26, 484–492 (2015).

Acknowledgments: We thank the Center for Biological Imaging (CBI), Institute of Biophysics, Chinese Academy of Science for our cryo-TEM work, and L. Wang and C. Peng for their help of making TEM sample and taking TEM images. Funding: This work is supported by the National Major Scientific Instruments and Equipments Development Project (ZDYZ2015-2) and the Key Research Program of the Chinese Academy of Sciences (QYZDJ-SSW-SLH025). Author contributions: C.Z., M.Z., C.W., and C.B. designed research. C.Z., X.L., T.Y., and J. Li synthesized and characterized the materials. C.Z., X.L., J. Liu, W.J., S.L., L.L., Z.S., Z.Z., and X.W. performed the animal experiments. M.Y., C.Z., and M.Z. performed proteomics analyses. C.Z., M.Z., and X.Z. performed the cell experiments. C.Z., M.Z., C.W., and C.B. prepared the paper. All authors have read and approved the paper. Competing interests: C.B., X.L., M.Z., C.Z., and C.W. are authors on a patent application related to this work filed by Institute of Chemistry, Chinese Academy of Sciences (note: the patent had been transferred to Beijing Funakang Biotechnology Co. Ltd.) (no. CN201710560609.8, filed on 11 July 2017). The authors declare no other competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Submitted 8 April 2020Accepted 29 July 2020Published 11 September 202010.1126/sciadv.abc1586

Citation: C. Zhou, M. Zhen, M. Yu, X. Li, T. Yu, J. Liu, W. Jia, S. Liu, L. Li, J. Li, Z. Sun, Z. Zhao, X. Wang, X. Zhang, C. Wang, C. Bai, Gadofullerene inhibits the degradation of apolipoprotein B100 and boosts triglyceride transport for reversing hepatic steatosis. Sci. Adv. 6, eabc1586 (2020).

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from

Page 15: DISEASES AND DISORDERS Copyright © 2020 Gadofullerene ...Shuai Liu1,2, Lei Li1,2, Jie Li1,2, Zihao Sun1,2, Zhongpu Zhao1,2, Xinyao Wang1,2, Xiaoyan Zhang1, ... 75 80 Transmittance

transport for reversing hepatic steatosisGadofullerene inhibits the degradation of apolipoprotein B100 and boosts triglyceride

Zhao, Xinyao Wang, Xiaoyan Zhang, Chunru Wang and Chunli BaiChen Zhou, Mingming Zhen, Meilan Yu, Xue Li, Tong Yu, Jingchao Liu, Wang Jia, Shuai Liu, Lei Li, Jie Li, Zihao Sun, Zhongpu

DOI: 10.1126/sciadv.abc1586 (37), eabc1586.6Sci Adv 

ARTICLE TOOLS http://advances.sciencemag.org/content/6/37/eabc1586

MATERIALSSUPPLEMENTARY http://advances.sciencemag.org/content/suppl/2020/09/04/6.37.eabc1586.DC1

REFERENCES

http://advances.sciencemag.org/content/6/37/eabc1586#BIBLThis article cites 37 articles, 10 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.Science AdvancesYork Avenue NW, Washington, DC 20005. The title (ISSN 2375-2548) is published by the American Association for the Advancement of Science, 1200 NewScience Advances

License 4.0 (CC BY-NC).Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on February 10, 2021

http://advances.sciencemag.org/

Dow

nloaded from