Journal Pre-proof Interplay between fatty acids, SCD, mTORC1 and YAP/TAZ in promoting hepatocellular carcinoma

Nonalcoholic fatty liver disease (NAFLD) has reached pandemic proportions with one of its most consequential complications being hepatocellular carcinoma (HCC). NAFLD-related HCC is becoming the leading indication for liver transplantation in the USA. Given the scarcity of available organs, early detection and prevention remain key in prevention and management of the disease. Over the years, the YAP/TAZ pathway emerged as a key signal transduction pathway in the pathogenesis of HCC. In this review, we explore the interplay between the YAP/TAZ pathway as a point of convergence in HCC pathogenesis. We review the evidence of how lipid reprogramming and key lipid pathways, saturated and monounsaturated fatty acids (through the rate limiting enzyme stearoyl Co-A desaturase), the mevalonic acid pathway (the role of statins) and mTORC1 all play critical roles in intricate and complex network that tightly regulate the YAP/TAZ pro-oncogenic pathway. patients (OB-NAFLD; n=102) and healthy controls (n=35) 46 . Compared to AV-HCC and OB-NAFLD, NAFLD-HCC patients had higher triglycerides (47:0 and 45:1) and phosphatidylcholine (16:0/17:0, 18:2/0:0, 0:0/18:2) levels but lower monounsaturated fatty acids (MUFAs), polyunsaturated fatty acids (PUFAs) and linoleic acid, after adjusting for age, sex and body mass index (BMI) 46 . The authors propose that the lower MUFA and PUFA pool may be related to an increase uptake of these lipid species by the HCCs from the serum to sustain their growth and proliferation. Consistent with these findings, Muir et al ., identified that stearoyl-Co-A activation, at least in part through actin remodeling 68 and modulation of the STRIPAK multi-subunit complex 60 . A recent study found that the endoplasmic reticulum (ER) protein Nogo-B, is highly expressed in both murine and human NAFLD-associated HCCs and leads to lysophosphatidic acid (LPA)-enhanced YAP oncogenic activity 93 . LPA is known to stimulate YAP via G protein-coupled receptors that promote Rho activation 94 . Another study concluded that YAP directly interacts with SREBP1c and SREBP2 on the promoters of the FASN and HMG CoA reductase, thereby stimulating their transcription and promoting hepatocyte lipogenesis and cholesterol synthesis 95 . Several studies linked mevalonate metabolism and YAP/TAZ activity and identified lipophilic statins as YAP/TAZ inhibitors in a variety of cell types 96-98 but the

Introduction: Primary liver cancer is the second leading cause of cancer death worldwide 1 . Hepatocellular carcinoma (HCC), which comprises 90% of the cases, commonly presents in patients with hepatic fibrosis associated with chronic hepatitis B virus (HBV) or hepatitis C virus (HCV), excessive alcohol consumption, or metabolic syndrome (MetS). Error! Hyperlink reference not valid.As described below, the genetic landscape of HCC is complex and includes mutations in the Wnt/β-catenin (CTNNB1) pathway, chromatin remodeling, telomere maintenance and inactivation of p53.
Nonalcoholic fatty liver disease associated HCC burden: In the USA, liver cancer incidence rates have more than tripled since 1980, while the cancer-associated death rates have more than doubled during this time.
In parallel, non-alcoholic fatty liver disease (NAFLD), a common condition in patients with obesity and type 2 diabetes (T2D), has reached pandemic proportions and continues to rise 2 . In the era of direct-acting antiviral therapy for HCV and improved HBV treatment, NAFLD-HCC is predicted to become the leading indication for liver transplantation (LT) 3 . As LT remains a scarce resource with recent evidence suggesting that HCCs are transplanted less often following changes in United States Organ Procurement Network policy 4 , understanding its underlying mechanisms is of upmost importance to develop and implement chemoprevention and early detection programs, as well as tailor treatment approaches.
NAFLD and HCC are complex diseases that result from genetic and environmental interactions 5 . Many clinical potentially modifiable risk factors have been associated with NAFLD-HCC development, including features of the MetS, as defined by the clustering of T2D, hypertension, dyslipidemia and obesity, where each condition plays an additive or synergistic role in the pathogenesis of NAFLD-HCC 6,7 . Although statins are commonly prescribed for the treatment of dyslipidemia and the MetS, their use has been shown to be protective in all etiologies of HCC 8,9 , including in non-alcoholic steatohepatitis (NASH), the more severe form in the NAFLD spectrum 10,11 .
The mechanisms involved in NAFLD-related HCC pathogenesis are therefore of major fundamental and translational importance but remain incompletely understood. In a large Veteran population, the estimated incidence of HCC in the patients with NAFLD is 0.21/1000 person-years (PY), with the highest incidence occurring in the sub-population of patients with cirrhosis (10.6/1000 PY) 7 . Simon et al. demonstrated similar findings in a European population-based cohort of biopsy-proven NAFLD and NASH, where HCC occurred at J o u r n a l P r e -p r o o f an incidence of 6.2/1000 PY for patients with underlying cirrhosis (95% CI 4.2-8.8/1000 PY) 6 . The incidence of HCC was lower for patients with biopsy-proven simple steatosis (n=5,939) at 0.8/1000 PY, and NASH without fibrosis (n=1,050) at 1.2/1000 PY 6 . The differences in incidence and disease progression likely reflect the heterogeneity in the NAFLD population, and potentially non-modifiable genetic differences, and their geneenvironment interactions. There is evidence from previous genome wide association studies (GWAS) demonstrating a predominance of NAFLD in Hispanic patients, followed by Caucasians and African Americans 12 , which have been further substantiated in the HCC population 13 . Unlike viral and ethanol-associated causes of HCC, NAFLD patients represent a distinct high-risk population where 20-30% of all HCCs occur in the absence of cirrhosis 14,15 . Although current guidance does not recommend HCC screening in patients without cirrhosis, the estimated disease burden globally 16 and health care-associated costs 17 , prompt the need for further research in this population, as outlined by recent professional society recommendations 18 .

Mutations in HCC development:
Heritability of NAFLD and NASH has been suggested based on early observations of the clustering of disease in families and twins 19,20 . In NAFLD-related HCC, teasing out inherited genetic contributions from chronic fibrosis/cirrhosis versus NAFLD-specific causes using GWAS remains a challenge and is still being investigated. Independently of its genetic predisposing factors, HCC often results from an accumulation of somatic mutations that lead to tumor initiation and progression. Studies have demonstrated that each HCC has about 40-60 somatic mutations, several of which converge on key molecular pathways 21 . Many mutations affect telomere maintenance (TERT or telomere reverse transcriptase activation); WNT activation (mutation in CTNNB1 which encodes β-catenin, APC and AXIN1); the tumor suppressor p53 gene (TP53); chromatin remodeling (including in the AT-rich interaction domain 1A or ARID1A); activation of the PI3K/mechanistic target of rapamycin (mTORC1); and activation of receptor tyrosine kinases [21][22][23][24][25] . The majority of mutations identified have been shown not to be actionable therapeutically 22,26 . Of the multiple pathways identified, certain ones deserve special attention because of their common occurrence. For instance, β-catenin activation through point mutations of CTNNB1 and mutation of TP53 occur in up to 50% HCC cases 25,27,28 .
HCCs can be sub-classified into proliferation and non-proliferation classes with some overlaps in the mutational landscape, including in β-catenin activation 22,29 . An emerging area of interest is the interplay among lipid J o u r n a l P r e -p r o o f metabolism, mTORC1 and the Hippo pathway in HCC. These represent major foci of this article, which are discussed in subsequent sections.

The lipid environment and NAFLD-related HCC:
It is widely accepted that growing cells require unsaturated fatty acids and cholesterol as essential building blocks for biogenesis of cellular membranes. Early on investigators observed that cancer cells also synthesize lipids 30 . Lipids, and specifically fatty acids, have been shown to have an integral role in changing the tumor environment (TME) to meet the demands of rapid cell proliferation observed in cancer cells, where oxygen and nutrients are scarce 31 . This hallmark of cancer cells has been termed 'metabolic reprogramming' 32 , and has been linked to clinical aggressiveness of many tumors 33 . This is supported by a large body of evidence that suppression of lipogenic pathways decreases cell growth in both in vitro and in vivo models 34,35 . Understanding tumor lipid metabolism and the crosstalk between the lipidome and the pathways that drive HCC has therefore advanced the targeting of lipid pathways in cancer and linoleic acid, after adjusting for age, sex and body mass index (BMI) 46 . The authors propose that the lower MUFA and PUFA pool may be related to an increase uptake of these lipid species by the HCCs from the serum to sustain their growth and proliferation. Consistent with these findings, Muir et al., identified that stearoyl-Co-A J o u r n a l P r e -p r o o f desaturase (SCD), the rate-limiting enzyme in the conversion of saturated fatty acids (SFA) to MUFAs, which is preferentially found on the endoplasmic reticulum, was upregulated in human NAFLD-related HCC samples and Pten-null NASH-HCC murine models 47 . While unsaturated fatty acids are required for cellular proliferation and membrane biogenesis, fatty acids also play important signaling functions that remain much less understood. These data suggest that the TME, supported by these other cell types, may play an integral role in HCC growth and aggressiveness. However, detailed human cell-type studies are lacking and have only recently been able to be identified with the advent of single cell and single nucleus RNA sequencing 54 .

SCD
As shown in figure 1, SCD expression is regulated at multiple levels 55 , including by the sterol regulatory element-binding protein 1 (SREBP1), the master transcription regulator of lipid biosynthesis, which has higher expression in HCC through mammalian target of rapamycin (mTOR) signaling. mTOR functions as a catalytic subunit in two distinct multi-protein complexes, the mechanistic target of rapamycin 1 (mTORC1), characterized by the subunit, Raptor (regulatory-associated protein of mTORC1) and mTORC2, characterized by the subunit Rictor (Raptor-independent companion of mTORC2). The heterodimer of the tumor suppressor TSC1 and TSC2 represses mTORC1 activity by acting as the GTPase activating protein for Rheb, a potent activator of mTORC1 in its GTP-bound state (Fig. 1).
Using an interactive open-access database (www.proteinatlas.org/pathology), we found that higher expression of Rheb is significantly associated with poor survival in HCC ( Fig. 2A). In the presence of amino acids, activated Akt and/or ERK/p90RSK phosphorylate and uncouple TSC1/TSC2 from J o u r n a l P r e -p r o o f Rheb, leading to Rheb-GTP accumulation and mTORC1 activation at lysosomal membranes, which then promotes cell growth 56 . Ragulator complex proteins LAMTOR 1, 3 and 5, which are involved in amino acid sensing and mTORC1 activation are also significantly associated with unfavorable prognosis (survival) in HCC ( Fig. 2B-D). The association of Rheb and LAMTORS1, 3 and 5 with unfavorable prognosis emphasizes the importance of mTORC1 function in HCC development. In turn, mTORC1 increases lipogenesis through regulation of SREBP1 at several levels, including trafficking, processing, and transcription 57 Error! Hyperlink reference not valid. that culminate in its nuclear localization and transcriptional activation. In turn, nuclear SREBP1 induces the expression of lipogenesis genes, including SCD. Lats1/2, in complex with its regulatory protein MOB1/2 (Fig. 1). In turn, LATS1/2 phosphorylate YAP and TAZ.
Structurally, YAP and TAZ share nearly half of the overall amino acid sequence, have very similar topology and highly conserved residues located within a consensus sequence phosphorylated by Lats1/2 (HXRXXS). The phosphorylation of YAP by Lats1/2 at Ser-127 and Ser-397 (and equivalent residues in TAZ) restricts its cellular localization to the cytoplasm and reduces its protein stability. When the Hippo pathway is not functional, YAP and TAZ are dephosphorylated and translocate to the nucleus where they bind and activate a number of transcription factors, primarily the TEA-domain DNA-binding transcription factors (TEAD1-4). When the Hippo pathway is active, LATS1/2 phosphorylates YAP and TAZ at multiple sites, thereby restricting their cellular localization to the cytoplasm, thus preventing their interaction with TEADs. An increased expression of SAV1, a protein necessary for the function of Mst1/2 upstream in the Hippo tumor-suppressive pathway, is associated with favorable prognosis in HCC (Fig. 2E), whereas, reciprocally, the expression of BIRC5 and TEAD2 has J o u r n a l P r e -p r o o f been associated with poor overall survival in HCC (Fig. 2, F, G). BIRC5, which encodes survivin, a potent inhibitor of apoptosis and a downstream target of YAP 59 . In addition, it is emerging that the Mobs are a family of proteins with positive and negative effects on the Hippo pathway 60 . Accordingly, high expression of Mob 4, a component of the multi-subunit protein complex STRIPAK (Striatin-interacting phosphatase and kinase) that negatively regulates the Hippo pathway 60, 61 , is associated with unfavorable prognosis in HCC (Fig. 2H).
In line with the associations identified above, several studies in preclinical models indicate that YAP/TAZ play a major role in promoting HCC. An early study demonstrated that YAP overexpression in the liver induces rapid HCC development, implying a potent oncogenic role of this protein 62 . Further studies using liver-specific knock out of Mst1/2 or Sav1 demonstrated that the Hippo signaling pathway is a critical regulator of mammalian liver growth and a potent suppressor of liver tumor formation 63 . Reciprocally, heterozygous knock out of YAP in mouse models, suppressed the development of HCC-like tumors 64 . Subsequent multiple studies substantiated that the tumor suppressive Hippo pathway plays a critical role in liver regeneration 65 and drug resistance 66,67 .
Furthermore, YAP promotes mTORC1 activation via downregulation of PTEN and increase in amino acid uptake 68 . In turn, mTORC1 leads to YAP/TAZ accumulation through inhibition of GSK3β 69 , thus establishing a positive feedback loop (Fig. 1). A fundamental crosstalk of the Hippo-YAP pathway with multiple metabolic pathways has been reviewed recently 70 . The Hippo YAP/TAZ signaling pathway has also been implicated in changing the TME in pancreatic adenocarcinoma, including in macrophages, highlighting the multi-faceted role and complex interplay of the Hippo-YAP pathway between cells in cancer development and progression 71 .
Emerging evidence supports the importance of lipid regulation of the Hippo pathway. In endothelial cells, palmitic acid, the most common of the SFA, induces expression of the Hippo pathway kinase Mst1 leading to YAP phosphorylation and nuclear export, therefore inhibiting nuclear YAP function 72 76,77 . Using a series of biochemical and genetic approaches, Chan et al. demonstrate that TEADs are autopalmitoylated at conserved cysteine residues, which in turn indirectly regulate YAP and TAZ by changing the conformation of TEAD proteins and therefore affecting their downstream association with these co-activators 78 . It was subsequently shown that S-palmityoylation is key in the homeostasis of TEAD protein levels in cells, thus providing another layer of regulation of the Hippo pathway 79 . FASN is key in the synthesis of plalmitoyl-CoA. Thus, it is conceivable that the FASN benefits observed in early clinical trials for non-small cell carcinoma, breast cancer and sarcoma (see "the lipid environment and NAFLD-related HCC"), may be partially due to inhibition of this autophosphorylation loop, which in turn dampens the YAP/TAZ oncogenic pathway. Although this concept is starting to be investigated in HCC in vitro cell culture models 80 , it remains an understudied area of HCC development of progression, especially in the context of NAFLD and NASH where intrahepatic lipid content is high. TEAD palmitoylation is a topic attracting interest as inhibitors of this critical modification are being developed 79, 81 that could offer novel therapeutic approaches in HCC and other cancers YAP, TAZ and HCC: Although YAP has been investigated and linked to many malignancies, its role in HCC pathogenesis is only emerging [82][83][84][85][86] . In a predominant viral hepatitis cohort, patients with high YAP expression in HCCs (n=177) were more likely to have poorly differentiated HCCs, which was associated with >2-fold increase in HCC-specific death 87 . Consistent with the mechanism of action of YAP, most tumors (62% had YAP expression) showed nuclear YAP localization 87 . A recent study uncovered that YAP and TAZ have overlapping and distinct roles in the development of liver carcinogenesis 88 Error! Hyperlink reference not valid.. Several signaling pathways interact and cross-talk in cancer development to sustain tumor growth, including with YAP.
Molecular subclassifications of HCC have been described 21,29 . One commonly mutated pathway that has been shown to converge on the YAP/TAZ pathway is the Wnt-β-catenin signaling pathway which affects 30-50% of all cases 21,50,84 . Mechanistic studies linking YAP/TAZ to Wnt-β-catenin signaling pathway suggest that this regulation occurs, at least in part, indirectly 84 through SCD and the MUFA/SFA intracellular pool 50 . Lai and colleagues elegantly demonstrate this in vivo and in vitro models where SCD was over-expressed both in activated hepatic stellate cells and hepatocytes. The Wnt-β-catenin in turn increased SREBP1, which stabilized J o u r n a l P r e -p r o o f LRP5 and LRP6 via MUFA intracellular pools and promoted liver fibrosis and HCC development 50 . These data further support TME cross-talk with hepatocytes for HCC pathogenesis.
Although several studies have focused on the role of YAP in HCC, TAZ may play a more important role in the transition from steatosis to NASH and fibrosis, which are precursors to HCC development. TAZ protein was found to be higher in human liver samples and a NASH murine models (methionine/choline deficiency-MCD-or fructose, palmitate, cholesterol-FPC-), compared to controls, with its silencing leading to reversal of inflammation, cell death and even fibrosis (through stellate cell activation) 89 . The crosstalk between these pathways and convergence on the YAP pathway need further exploration.  60 . A recent study found that the endoplasmic reticulum (ER) protein Nogo-B, is highly expressed in both murine and human NAFLD-associated HCCs and leads to lysophosphatidic acid (LPA)enhanced YAP oncogenic activity 93 . LPA is known to stimulate YAP via G protein-coupled receptors that promote Rho activation 94 . Another study concluded that YAP directly interacts with SREBP1c and SREBP2 on the promoters of the FASN and HMG CoA reductase, thereby stimulating their transcription and promoting hepatocyte lipogenesis and cholesterol synthesis 95 . Several studies linked mevalonate metabolism and YAP/TAZ activity and identified lipophilic statins as YAP/TAZ inhibitors in a variety of cell types [96][97][98] but the J o u r n a l P r e -p r o o f impact of statins on YAP localization, phosphorylation and activity in HCC cells remains unknown. The possibility that statins impair actin remodeling and/or STRIPAK activity in response to multiple agonists and growth factors that induce Rho activation (i.e. conversion of Rho-GDP to Rho-GTP) are intriguing molecular mechanisms in the context of HCC and other cancers but remain untested. Interestingly, it has been identified that the ATPbinding cassette (ABC) A1 transporter (ABCA1) is a target gene of p53 99 . Activation of ABCA1 results in reverse cholesterol transport, which leads to decreased SREBP2, the master transcription factor of cholesterol biosynthesis, which in turn leads to repression of the mevalonic acid pathway 99 .
Conclusions: Treatment of advanced and unresectable HCC has greatly changed over the recent years with the advent of immunotherapy and advances in locoregional therapies 100 . For decades, Sorafenib, a multi-protein kinase inhibitor, was first-line for treatment of advanced HCC. Its long-standing use informed patterns of resistance and how it affected HCC recurrence. In the context of this discussion, it is of interest that several recent studies identified YAP as a major mediator of HCC resistance to Sorafenib 101,102 . Similarly, Hippo signaling has been recently been shown to be a target of Regorafenib, which has a similar mechanism of action to Sorafenib, resistance in HCC using a CRISPR-Cas9 screen 66 .
In the light of the signaling network discussed herein, the exciting possibility of synergism between statins, SCD inhibitors, mTORC1 inhibitors (including rapamycin and its analogues) and kinase inhibitors in dampening HCC cell proliferation and migration and preventing drug resistance through inhibition of YAP/TAZ is a novel therapeutic approach that has not been examined in HCC. Additionally, the use of lipid modulators including statins as chemoprevention methods for NAFLD-related HCC should be further explored to tackle the large healthcare and economic burdens anticipated in the near future. Finally, understanding the HCC subclassifications through these different pathways may help explain the heterogeneity of HCC development and treatment responses, and would advance the role of precision oncology, especially in NAFLD-HCC, which has distinct features compared to its viral counterparts.
J o u r n a l P r e -p r o o f