Hepatic miR-20b promotes nonalcoholic fatty liver disease by suppressing PPARα

Background Non-alcoholic fatty liver disease (NAFLD) is associated with hepatic metabolic reprogramming that leads to excessive lipid accumulation and imbalances in lipid metabolism in the liver. Although nuclear receptors (NRs) play a crucial role in hepatic metabolic reprogramming, the underlying mechanisms of NR regulation in NAFLD remain largely unclear. Methods Using network analysis and RNA-seq to determine the correlation between NRs and microRNA in NAFLD patients, we revealed that miR-20b specifically targets PPARα. miR-20b mimic and anti-miR-20b were administered to hepatocytes as well as high fat diet (HFD)- or methionine-deficient diet (MCD)-fed mice to verify the specific function of miR-20b in NAFLD. We tested the inhibition of the therapeutic effect of a PPARα agonist, fenofibrate, by miR-20b. Results We revealed that miR-20b specifically targets PPARα through miRNA regulatory network analysis of nuclear receptor genes in NAFLD. The expression of miR-20b was upregulated in free fatty acid (FA)-treated hepatocytes and the livers of both obesity-induced mice and NAFLD patients. Overexpression of miR-20b significantly increased hepatic lipid accumulation and triglyceride levels. Furthermore, miR-20b significantly reduced FA oxidation and mitochondrial biogenesis by targeting PPARα. In miR-20b-introduced mice, the effect of fenofibrate to ameliorate hepatic steatosis was significantly suppressed. Finally, inhibition of miR-20b significantly increased FA oxidation and uptake, resulting in improved insulin sensitivity and a decrease in NAFLD progression. Conclusions Taken together, our results demonstrate that the novel miR-20b targets PPARα, plays a significant role in hepatic lipid metabolism, and present an opportunity for the development of novel therapeutics for NAFLD. Funding This research was funded by Korea Mouse Phenotyping Project (2016M3A9D5A01952411), the National Research Foundation of Korea (NRF) grant funded by the Korea government (2020R1F1A1061267, 2018R1A5A1024340), the Future-leading Project Research Fund (1.210034.01) of UNIST and the National Research Foundation of Korea (NRF) grant funded by the Korea government (2020R1I1A1A01074940).


Introduction
decreased by miR-20b in human liver cells and mouse primary hepatocytes. Consistent with the RNA-146 seq data, the expression change of PPARA at both the protein and mRNA levels with miR-20b 147 transfection was the most distinct compared to the control ( Figure 2H, I). Moreover, among candidate 148 targets, only PPARA was selected as an overlapped predicted target of miR-20b between various 149 miRNA target prediction programs, including miRDB, picTAR, TargetSCAN, and miRmap ( Figure 2J, 150 including citrate and succinate, also decreased ( Figure 3E). 171 Enforced expression of miR-20b in HepG2 cells under both basal and OA treatments decreased the 172 expression of PPARGC1A and SIRT1, which are involved in mitochondrial biogenesis ( Figure 3F). 173 The copy number of two mtDNA genes, VIPR1 and MT-ATP6, was decreased by miR-20b 174 overexpression following OA treatment ( Figure 3G). Consistently, mitochondrial function that was 175 analyzed via OCR (oxygen consumption rate) was reduced by miR-20b under both basal and OA 176 treatment conditions compared to the control ( Figure 3H). In particular, the basal respiration and 177 maximal respiratory capacity were significantly suppressed by miR-20b ( Figure 3I). Furthermore, the 178 level of ATP production, FA uptake, and FA oxidation was reduced in miR-20b overexpressed cells 179 compared with that in the control under both basal and OA-treated conditions ( Figure 3J-L). 180 To further clarify the role of miR-20b in hepatic steatosis, miR-20b inhibitor (anti-miR-20b), which 181 silences miR-20b, was delivered into HepG2 cells and primary hepatocytes with OA treatment (

miR-20b promotes hepatic steatosis in HFD-fed mice 199
To confirm the in vivo roles of miR-20b in obesity model mice, we introduced miR-20b using an 200 adenovirus-associated vector (AAV), referred to as AAV-miR-20b, into C57BL/6 mice that had been 201 fed a normal chow diet (NCD) or a high-fat diet (HFD). Administration of AAV-miR-20b led to high 202 expression levels of miR-20b in the livers of NCD-and HFD-fed mice compared to AAV-Control 203 injection ( Figure 4A). However, the expression level of miR-20b was not changed in peripheral tissues 204 including white and brown adipose tissues except in muscle (Katayama et  supplement 1). Consequently, AAV-miR-20b injected mice exhibited a reduction in the PPARα protein 206 levels compared with AAV-Control injected mice on both NCD and HFD ( Figure 4B). 207 Alterations in body weight were not detected in NCD-fed mice after AAV-miR-20b administration; 208 however, AAV-miR-20b led to a significant increase in the body weight of HFD-induced obese mice 209 ( Figure 4C). The ratio of fat mass to body weight in AAV-miR-20b administration HFD-fed mice was 210 higher than that in AAV-Control treated mice ( Figure 4D and Figure 4-figure supplement 2); however, 211 the ratio of lean mass to body weight showed no significant differences ( Figure 4E). Consistently, 212 AAV-miR-20b administration increased liver weight and steatosis in HFD-fed mice ( Figure 4F, G). The 213 hepatic TG level, serum activities of aspartate aminotransferase (AST) and alanine aminotransferase 214 (ALT), markers of liver injury, were significantly increased with AAV-miR-20b administration compared 215 with AAV-Control administration in HFD-fed mice ( Figure 4H-J). 216 Additionally, we observed that delivery of AAV-miR-20b to HFD-fed mice significantly impaired 217 glucose tolerance and insulin sensitivity compared to the AAV-Control ( Figure 4K, L). Fasting glucose, 218 insulin, and homeostasis model assessment of insulin resistance (HOMA-IR) levels were also 219 increased in AAV-miR-20b administrated HFD-fed mice ( Figure 4M-O). We observed that genes 220 involved in FA β-oxidation and FA uptake pathways were downregulated by AAV-miR-20b compared 221 to AAV-Control in both NCD-and HFD-fed mice, whereas lipogenesis genes were not altered in AAV-222 miR-20b administrated mice ( Figure 4P-R). These results suggest that miR-20b could aggravate 223 NAFLD by dysregulating lipid metabolism in a HFD-induced obesity model. 224

225
Inhibition of miR-20b alleviates hepatic steatosis in HFD-fed mice. 226 Next, we introduced anti-miR-20b into HFD-fed mice. Administration of AAV-anti-miR-20b led to 227 decrease of miR-20b in the livers of NCD-and HFD-fed mice compared to AAV-Control injection 228 ( Figure 5A). AAV-anti-miR-20b significantly increased PPARα expression in the livers of both NCD-229 and HFD-fed mice ( Figure 5B). Administration of AAV-anti-miR-20b in HFD-fed mice reduced the 230 body weight compared to that of AAV-Control administrated mice ( Figure 5C). We further determined 231 that alterations in body weight were highly associated with fat mass loss ( Figure 5D and Figure 5-232 figure supplement 1). While the ratio of lean mass to body weight of AAV-anti-miR-20b administrated 233 HFD-fed mice was increased, the lean mass was comparable to that of the control ( Figure 5E). We 234 next observed that AAV-anti-miR-20b administration reduced liver weight and hepatic steatosis in 235 HFD-fed mice than in AAV-Control mice ( Figure 5F). H&E and Oil Red O staining demonstrated that 236 delivery of AAV-anti-miR-20b significantly attenuated the size and number of lipid droplets in the liver 237 compared to AAV-Control administration in HFD-fed mice ( Figure 5G). In accordance with histological 238 changes, metabolic parameters were reduced in AAV-anti-miR-20b administrated mice compared 239 with the AAV-Control administrated mice ( Figure 5H-J). Furthermore, AAV-anti-miR-20b significantly 240 improved glucose tolerance ( Figure 5K) and insulin sensitivity ( Figure 5L) compared to the AAV-241 Control in HFD-fed mice. Consistently, we determined that both fasting glucose and fasting insulin Next, we confirmed that the regulation of FA β-oxidation and mitochondrial function by miR-20b is 249 primarily mediated through the reduction of PPARα. Transfection of miR-20b into HepG2 cells 250 reduced the expression and activity of PPARα, but co-transfected PPARα expression vector restored 251 them ( Figure 6A, B). Furthermore, the decreased expression of genes involved in lipid metabolism, 252 such as FA β-oxidation and FA uptake by miR-20b, was significantly restored by the forced 253 expression of PPARα ( Figure 6C-E). Next, we tested whether the effect of anti-miR-20b was inhibited 254 by the suppression of PPARα. The increased expression and activity of PPARα by anti-miR-20b was 255 reduced by siRNA targeting PPARα (siPPARA) ( Figure 6F, G). The increased expression of genes by 256 anti-miR-20b was also suppressed by siPPARA ( Figure 6H-J). In addition, fenofibrate, a PPARα 257 agonist, increased the expression of PPARα and its transcriptional activity in HepG2 cells transfected 258 with miR-20b, but could not restore as much on its own effects ( Figure 6K, L). Interestingly, 259 fenofibrate treatment increased the expression of genes involved in FA β-oxidation and FA uptake 260 which are regulated by PPARα, but could not overcome the inhibitory effect of miR-20b ( Figure 6M-O). 261 Taken together, these results indicate that the contribution of miR-20b to hepatic steatosis is 262 mediated by direct inhibition of PPARα and is important for the treatment of NAFLD. 263 The effects of fenofibrate are limited in miR-20b-introduced mice. 265 Next, we tested whether NAFLD treatment with fenofibrate was affected by miR-20b expression in 266 vivo. Administration of AAV-miR-20b led to elevated hepatic levels of miR-20b compared to AAV-267 Control injection in HFD-fed mice, and the level was slightly decreased by fenofibrate treatment 268 ( Figure 7A). Interestingly, we observed that administration of AAV-Control with fenofibrate increased 269 the level of PPARα; however, fenofibrate could not restore the reduced PPARα expression by AAV-270 miR-20b ( Figure 7B). Administration of fenofibrate reduced the body and liver weights of AAV-Control 271 injected mice; however, AAV-miR-20b injected mice exhibited no significant differences by fenofibrate 272 ( Figure 7C, F). The ratio of fat to body weight also displayed no alterations between AAV-miR-20b 273 and AAV-miR-20b with fenofibrate ( Figure 7D). While the ratio of lean mass to body weight was 274 increased by fenofibrate in AAV-miR-20b injected mice, the lean mass was comparable ( Figure 7E). 275 H&E staining, Oil Red O staining, and hepatic TG levels demonstrated that fenofibrate significantly 276 attenuated lipid accumulation in the liver of HFD-fed mice, but the effect of fenofibrate was 277 suppressed by AAV-miR-20b ( Figure 7G, H). Serum AST and ALT levels were decreased by 278 fenofibrate, but this benefit was did not detected in AAV-miR-20b injected mice ( Figure 7I, J). We 279 further observed that blood glucose tolerance and insulin sensitivity were improved by fenofibrate; 280 however, AAV-miR-20b offset the improvement by fenofibrate ( Figure 7K, L). Fasting glucose, fasting 281 insulin, and HOMA-IR levels were markedly decreased by fenofibrate in HFD-fed mice ( Figure 7M-O). 282 In AAV-miR-20b injected mice, fenofibrate did not reduce fasting insulin levels, but decreased fasting 283 glucose and HOMA-IR levels. Fenofibrate also did not restore the suppressed expression of genes 284 regulating FA β-oxidation in AAV-miR-20b-injected mice ( Figure Figure 8A, B). This implies that miR-20b is able to set up an NR 292 transcription program similar to that of NASH and liver fibrosis. To test this hypothesis, AAV-Control or 293 AAV-anti-miR-20b was administered to C57BL/6 mice placed on a methionine/choline-deficient diet 294 (MCD), which is the most widely used diet to induce NAFLD/NASH. Administration of AAV-anti-miR-295 20b led to decrease of miR-20b in the livers of NCD-and MCD-fed mice compared to AAV-Control 296 injection ( Figure 8C). We observed that the expression of PPARα was increased in MCD-fed mice 297 and administration of AAV-anti-miR-20b displayed an elevation of PPARα, both at the mRNA and 298 protein levels ( Figure 8D, E). We next observed that AAV-anti-miR-20b administration significantly 299 reduced hepatic steatosis in MCD-fed mice than in AAV-Control mice ( Figure 8F). Liver sections 300 clearly showed a decrease in both lipid accumulation and fibrosis with AAV-anti-miR-20b 301 administration in MCD-fed mice ( Figure 8G). Consistently, AAV-anti-miR-20b administration 302 decreased the levels of hepatic TG, AST, and ALT activity compared to AAV-Control injection (  were highly elevated in T2DM/NAFLD patients compared to those in T2DM patients (Ye et al., 2018). 317 However, the molecular mechanism through which miR-20b regulates NAFLD progression remains 318 unknown. In this study, we demonstrated that miR-20b promotes NAFLD progression by modulating 319 lipid metabolism, including FA β-oxidation and FA uptake, as well as ATP production by mitochondrial 320 biogenesis. Our data clearly showed the regulatory mechanism of PPARα by miR-20b, and miR-20b 321 may serve as a novel biological marker in NAFLD. 322 A previous study demonstrated that upregulated miR-20b levels in obesity-induced metabolic 323 disorders such as T2DM were considered to prevent several targets, such as STAT3, CD36, and 324 PTEN, which are involved in glucose and lipid homeostasis. The miR-20b/STAT axis, which is 325 involved in the insulin signaling pathway, alters glycogen synthesis in human skeletal muscle miR-20b can be a promising target in liver cancer development. Indeed, we observed that the level of 367 miR-20b was increased in NAFLD patients, but even robustly increased in the NASH stage. 368 Furthermore, we observed that the hepatic function of miR-20b dramatically regulates the genes 369 involved in inflammation and fibrosis by directly repressing PPARα in MCD-fed mice. Thus, our study 370 strongly suggested that miR-20b regulates the pathogenesis of NAFLD, but might also be relevant in 371 the development of severe stages of liver fibrosis and even in HCC. 372 Our present results strongly suggest that miR-20b may be a druggable target in NAFLD patients.    GGTGATGCTTTTGTTCTTACTGG  TGTCTCCACAGATCTTGCATG  RORC  TGGTGCTGGTTAGGATGTG  GGAGTGGGAGAAGTCAAAGATG  THRB  CATCAAAACTGTCACCGAAGC  TCCAAGTCAACCTTTCCACC  NRBP1  GTTCCACCCAGCATTGTTTG  CAGGGATTTCAGCCAGTACG  NAGK  TATTTCCAGGTGCCAGATCG  CTGAAGATATAGCGGGAAAGGG  USP46  ATACACCAAGCTGTCTTACCG  ATATAATGCCCACGATTAGGACC  ITGB8  CGTCTCATCTCGCTCTTGATAG  TTCTCTGAAAGTTGGCCTAGTG  BMPR2  GGCTGACTGGAAATAGACTGG  CACAGTCCCTCAAGTTCACAG  ZNFX1  CCGAGGATTGTCATAGTGGAAG  AGATCATACACGTTGGCACTG  EPHA5  AGATTGAGGCAGTGAATGGAG  GCCAAGACAAAGAGATGCTG  E2F1  TCTCCGAGGACACTGACAG  ATCACCATAACCATCTGCTCTG  TBP  CCACTCACAGACTCTCACAAC  CTGCGGTACAATCCCAGAACT  B2M TGCTGTCTCCATGTTTGATGTATCT TCTCTGCTCCCCACCTCTAAGT

Hepatocyte isolation 503
Briefly, mice were anesthetized with isoflurane, and 24-gauge needle was inserted into the portal 504 vein. Then the inferior vena cava was cut, and the mouse liver was perfused sequentially with solution 505 I (142 μM NaCl, 6.7 μM KCl, 10 μM HEPES, and 2.5 mM EGTA), and solution II (66.7 mM NaCl, 6.7 506 mM KCl, 50mM HEPES, 4.8 mM CaCl 2 ·2H 2 O, and 0.01 % Type IV collagenase (Sigma-Aldrich, St. 507 Louis, MO)). After digestion, the liver was disrupted over a 70-μm cell strainer, and cell suspension 508 was spun at 50 x g for 5min at 4 °C. The supernatant was gently aspirated and the cells were 509 resuspended in M199 with EBSS (M199/EBSS) medium and gently mixed with equal volume of 510 Percoll working solution (48.6 % Percoll). The cell suspension was spun at 100 x g for 5 min at 4 °C, 511 and the pellet washed once with M199/EBSS. After viable cells were counted with trypan blue, the 512 isolated hepatocytes were seed in M199/EBSS medium supplemented with 10 % FBS, 1 % 513 penicillin/streptomycin, and 10 nM dexamethasone. 514 515

Metabolic analysis 526
Mice were fasted overnight (18 h) before intraperitoneal injection of D-glucose (2 g/kg body weight) 527 for glucose tolerance test. For insulin tolerance test, mice were fasted for 4 h before intraperitoneal 528 injection of insulin (0.75 U/kg body weight). Every glucose was examined with tail-vein blood at 529 indicated intervals after injection using a glucometer. For analyzing metabolic parameters, insulin 530

Histological analysis 536
Liver tissues were isolated from mice and immediately fixed with 4% formalin (Sigma-Aldrich, St. 537 Louis, MO). Histological changes of lipid droplets were examined by H&E staining and Oil Red O 538 staining. As counterstain, Mayer's hematoxylin was used for every slide. Liver fibrosis was further 539 examined by Sirius red with liver section. Images were obtained with Olympus BX53 microscope and 540 DP26 camera. 541 542

Figure 9. E2F1 is upregulated in both NAFLD patients and mice model 887
The expression of E2F1 was analyzed by quantitative RT-PCR. Hepatic E2F1 expression levels of 888 steatosis or NASH patients were normalized to those of normal patients. *P < 0.05 and **P < 0.01 vs 889 normal patients (A). E2F1 expression levels from HepG2 cells (B) and Huh-7 cells (C) treated with OA 890 for 24 h were normalized to no treatment (NT). Hepatic E2f1 expression levels from C57BL/6J mice 891 fed a HFD (D) and a MCD (E) were normalized to NCD. Values represent means ± SEM (n = 3-5). 892 *P < 0.05, **P < 0.01, ***P < 0.001 vs NT in cells or NCD-fed mice, respectively. 893