USP11 degrades KLF4 via its deubiquitinase activity in liver diseases

Abstract Krüppel‐like factor 4 (KLF4) is a zinc‐finger containing DNA‐binding transcription factor involved in tumorigenesis and acts as a tumour suppressor or an oncogene depending on the tissue. In hepatocellular carcinoma (HCC), KLF4 has been considered as a tumour suppressor, although the mechanism underlying its action remains largely unknown. In this study, we identified the ubiquitin‐specific peptidase USP11 as a KLF4‐interacting deubiquitinating enzyme using a proteomic approach. USP11 destabilizes KLF4 through the removal of K63‐dependent polyubiquitination, thereby inhibiting KLF4 expression. We also provide mechanistic insights into KLF4 degradation and show that USP11 depletion inhibits growth and chemoresistance of HCC cells by enhancing KLF4 stability. Importantly, lipid content was reduced and genes involved in fatty acid metabolism were down‐regulated in an in vitro steatosis conditions upon USP11 knockout. Finally, elevated USP11 and reduced KLF4 levels were detected both in a hepatic steatosis in vitro model and in public clinical data of non‐alcoholic fatty liver disease and HCC patients. Collectively, these findings suggest that USP11, as KLF4‐binding partner, is an important mediator of hepatic tumorigenesis that functions via degradation of KLF4 and is a potential treatment target for liver diseases.

neuroblastoma, whereas it promotes the development of breast, skin, and head and neck cancers. 4,6,7 KLF4 is commonly known as a suppressor of cell cycle progression as it induces the expression of the cell cycle inhibitor p21 Cip1/Waf1 and inhibits the cell cycle promoting genes, CCND1 and CCNB1. 8,9 Additionally, KLF4 acts as an antiapoptotic transcription factor by suppressing the p53-dependent apoptotic pathway, in particular, through inhibition of TP53 and BAX expression. 10 Overall, KLF4 is a key player in numerous physiological and pathological processes.
In hepatocellular carcinoma (HCC), the loss of KLF4 expression is closely correlated with cancer progression and reduced overall survival. 11,12 KLF4 can regulate HCC differentiation and progression, and it inhibits HCC cell migration and invasion, mediated by the hepatocyte nuclear factor, HNF6. 13 KLF4 also inhibits the oncogenic TGFβ signals via induction of Smad7. 14 These findings suggest that KLF4 could be a potential therapeutic target for liver diseases, such as fibrosis and cirrhosis, and prevent their subsequent progression to HCC. However, the underlying molecular mechanisms for modulating KLF4 expression in HCC remain poorly understood. Given KLF4 involvement in cell fate decision in tumorigenesis, its activity is critically regulated both transcriptionally and post-transcriptionally through methylation, acetylation, phosphorylation, ubiquitination and sumoylation, in a context-dependent manner. 3 KLF4 has a short half-life and turnover rate, meaning that it is regulated posttranslationally via ubiquitination. 15 A recent study revealed different elements of the ubiquitination process that are involved in the regulation of KLF4 protein stability in various cell types and cancers. In case of E3 ligases, for example Von Hippel-Lindau (pVHL) is involved in breast and colon cancer, FBXO32 in breast cancer, Cdh1-anaphasepromoting complex in lung cells, β-TrCP in stem cell self-renewal and TRAF7 in HCC. [16][17][18][19] In contrast, deubiquitinating enzymes (DUBs) can reversibly cleave ubiquitin(s) off-target proteins and rescue substrates from post-translational modification. 20 This group of enzymes contains about ~100 DUBs that are classified into six subfamilies based on sequence and domain conservation. Although DUBs are important regulatory elements of many biological processes, including protein turnover and ubiquitin recycling into monomers, the detailed mechanism of KLF4 regulation by DUBs remains unknown.
USP11 is one of the most common DUBs and belongs to the ubiquitin-specific processing protease (USP) family. It is involved in multiple signalling cascades including TGFβ, p21, p53, NF-βB and Notch signals, ultimately regulating the stability of their downstream substrates. [21][22][23] USP11 has been recognized as a tumour suppressor in lung cancer through the regulation of Mgl-1. 24 In contrast, USP11 is highly associated with tumorigenesis in other cancer types such as breast, ovarian, colorectal, pancreatic cancer, melanoma, glioma and squamous cell carcinoma, due to its effect on different signalling pathways. [25][26][27] Recently, USP11 was shown to promote HCC development, 28 but the underlying molecular mechanisms involved in this pathogenic process remain poorly understood.
In this study, we used a proteomic approach to identify KLF4interacting DUBs and firstly discovered that USP11 was responsible for deubiquitinating KLF4 in HCC cells. USP11 deubiquitinates K63dependent polyubiquitination of KLF4 and suppresses its stability.
Using HCC cells engineered to lack USP11, we clarified that loss of USP11 restrains HCC tumorigenesis by promoting KLF4 accumulation. More importantly, we demonstrated the negative correlation between KLF4 and USP11 expression in liver diseases such as nonalcoholic fatty liver disease (NAFLD) and HCC. Collectively, these findings add further evidence that USP11 is a major regulator of HCC progression via direct regulation of KLF4 expression.
Huh7 and SNU423 were purchased from Korea Cell Line Banks (KCLB).
According to the manufacturer's instructions, the growth culture medium was prepared by mixing DMEM or EMEM with 10% FBS and additives (Gibco). For the cultivation of THLE2 cells, the BEGM Bullet Kit (CC-3170) from Lonza was used. The Bullet Kit contains BEBM basal medium and supplements. The final growth medium consists of the following: BEBM supplemented with 10% FCS, bovine pituitary gland extract, hydrocortisone, epidermal growth factor (EGF), insulin, triiodothyronine, transferrin, retinoic acid, 5 ng/mL human recombinant EGF (Gibco) and 70 ng/mL ο-phosphorylethanolamine (Sigma-Aldrich).
All cells applied in this study were cultured at 37℃ in a humidified 5% CO 2 atmosphere. Polyclonal antibodies against the epitope tags

| Isolation of KLF4 interactors with a proteomic approach
For identifying KLF4 interactome, 29 HepG2 cells were transfected with Flag-tagged KLF4 expression plasmids. The transfected cells were lysed with 1 × Nonidet P-40 lysis buffer and pre-cleaned by protein A/G agarose beads. Flag-tagged KLF4 proteins were immunoprecipitated with anti-Flag Abs beads, and the immune complex was eluted from the agarose with 100 μmol/L Flag peptide (Sigma-Aldrich). The eluted proteins were digested with trypsin and characterized by mass spectrometry.

| Plasmids and lentivirus transduction
Full-length KLF4 or USP11 was PCR-amplified from human cDNA and sub-cloned to p3xFlag-CMV10 or pCMV-Myc from Addgene.

| Cell transfection and coimmunoprecipitation assay
Transient transfection was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instruction. 2 days after transfection, cells were lysed in 1 × Nonidet P-40 lysis buffer and freshly added protease inhibitor cocktail. After being pre-cleared with protein A/G agarose beads, the lysates were mixed with indicated antibodies (1 μ g) for overnight, followed by the addition of protein A/G plus-agarose beads (Santa Cruz) for the additional 2 hours at 4℃ with gentle shaking. Immunoprecipitated proteins were washed out four times with lysis buffer and boiled in 2 × SDS sample buffer and subjected to Western blot analysis.

| In vivo and in vitro deubiquitylation assay
For KLF4 deubiquitylation in vivo, 30

| Western blotting
Protein samples were boiled and separated on 8% SDS-PAGE gels followed by electro-transferring to PVDF membranes (Bio-Rad Laboratories). After blocking with 5% non-fat milk in Tris-buffered saline containing 0.1% Tween-20 for 1 hour, the membranes were incubated with specific primary antibodies overnight at 4℃. Finally, antibody-bound proteins were detected by chemiluminescence (Bio-Rad). When necessary, membranes were stripped by incubation in stripping buffer (Thermo Fisher), washed and then reprobed with other antibodies as indicated.

| Real-time quantitative RT-PCR
Total RNA from cells was isolated with RNeasy Mini Kit and was reverse transcribed with GoScript reverse transcriptase (Promega

| Cell viability and Annexin V/Propidium Iodide (PI) assay
The proliferation was determined using the Cell Counting Kit

| Colony forming (clonogenic) assay
HepG2 cells infected with WT or USP11 shRNA particles were selected for 4-6 days and then were seeded in a 6-well plate (2500 cell/well). The cells were cultured for 2 weeks, and the medium was refreshed every 2 days. The colonies were fixed and stained with crystal violet. The number of the clones in a given area was counted using ImageJ software.

| Publicly available clinical data of HCC and NAFLD
The public RNA-seq data of HCC patients were obtained from cBi-

| Statistical analysis
Statistical analysis in this study was performed using GraphPad Prism Software (GraphPad). All data were collected from two or three independent experiments, and the results were expressed as mean ± SD. One-way ANOVA, two-way ANOVA or a two-tailed Student's t test was performed to analyse the statistically significance. P values < .05 were considered as significant. *P < .05; **P < .01; ***P < .001, ****P < .0001. In addition, the correlation of mRNA expression was calculated by Pearson correlation.

| USP11 directly interacts with KLF4
To evaluate the underlying molecular mechanisms of KLF4 impaired down-regulation in tumorigenesis, we used a proteomic approach to identify KLF4-interacting DUBs in HepG2 cells. 29 Briefly, whole lysates from Flag-KLF4 overexpressing HepG2 cells were subjected to immunoprecipitation with anti-Flag Ab-conjugated agarose beads after extensive pre-cleaning. The beads were washed and eluted with Flag peptide along with the bounded proteins, which were digested with trypsin and characterized by mass spectrometry.
We found several binding USP candidates including USP5, 9X and USP11 (data not shown). Since USP11 has been previously reported as an oncogene in HCC, but its underlying molecular mechanisms in hepatic disease are not entirely understood, we decided to further explore its activity as KLF4-binding partner.
We started by assessing the interaction between KLF4 and USP11 through co-immunoprecipitation (co-IP) analysis in transfected HEK293 cells. Following IP with anti-Flag beads for Flag-KLF4, we determined the presence of Myc-USP11 ( Figure 1A).
KLF4-USP11 interaction was mainly found in the nuclear fraction ( Figure 1B,C). Immunofluorescence analysis from hepatic adenocarcinoma (Sk-Hep1) cells showed that the colocalization of both KLF4 and USP11 occurred in the nucleus ( Figure 1D) which demonstrated that the KLF4 ID was required for its physical interaction with USP11 ( Figure 1E). We also mapped the regions of USP11 that were necessary for the interaction with KLF4. USP11 has a C-terminal catalytic domain and an N-terminal regulatory region. Conducting analysis of the impact of serial deletions of USP11 structures showed that deletion of its C-terminus prevented the interaction with KLF4, whereas deletion of the N-terminus had only a minor effect on this interaction ( Figure 1F). These results demon- The results suggested that coexpression of USP11 and KLF4 significantly prevented the ubiquitination of KLF4 mediated by WT-Ub, but not from KO-Ub (Figure 2A). Moreover, mutation of a critical cysteine into serine of USP11 (C318S), which inactivated its catalytic activity, 27 completely abolished USP11 ability to catalyse KLF4 deubiquitination without affecting the KLF4-USP11 interaction ( Figure 2B). The suppression of KLF4 ubiquitination by the deubiquitinase activity of USP11 was further confirmed by using an in vitro deubiquitination assay ( Figure 2C). Incubation of ubiquitinated KLF4 with a purified glutathione S-transferase (GST)-tagged USP11 inhibited KLF4 ubiquitination. These results indicate that USP11 is a specific deubiquitinase of KLF4 and its functional catalytic domain is required for KLF4 deubiquitination.   Figure 3A). Moreover, nuclear KLF4 expression was the most affected fraction rather than the cytoplasmic KLF4 fraction ( Figure 3B).  Figure 3C). In contrast, expression of the deubiquitinase catalytically inactive USP11/CS mutant protected KLF4 from degradation, indicating that KLF4 stabilization requires ubiquitin-specific peptidase activity by USP11 ( Figure 3D). Altogether, these results indicate that USP11 is a KLF4-specific deubiquitinase that promotes KLF4 degradation.

| USP11 mediates cancer cell proliferation and tumorigenesis by promoting KLF4 instability
Next, we aimed to investigate the function of USP11 in HCC cells.
We started by knocking down USP11 expression using specific short hairpin RNAs (shRNAs) in HepG2 cells. Deletion of USP11 led to up-regulation of KLF4 expression ( Figure 4A) and longer KLF4 half-life ( Figure 4B). These results showed that USP11 may not only regulate KLF4 expression post-translationally but also at its transcriptional level. Furthermore, down-regulation of USP11 significantly suppressed HepG2 cell growth ( Figure 4C) and chemoresistance ( Figure 4D), and clonogenic assay showed that USP11 silencing greatly suppressed the colony-forming ability of these HCC cells ( Figure 4E). We also analysed the relative expression

| KLF4 level is inversely correlated with USP11 expression in liver disease
Non-alcoholic fatty liver disease (NAFLD) is often related with obesity and metabolic liver disease characterized by steatosis and lipid accumulation in liver cells. 40 Although hepatic simple steatosis is considered a benign state, it can progress to non-alcoholic steatohepatitis (NASH), which is a precursor to more serious liver diseases such as cirrhosis and HCC. 41 To investigate the role of

| D ISCUSS I ON
KLF4 is decreased or absent in HCC cells, and its overexpression has been associated with overall improved survival of patients with F I G U R E 4 USP11 deletion induces KLF4 expression and inhibits HCC cell proliferation. A, Indicated lentiviral shRNAs (sh-NC or sh-USP11) were infected into HepG2 cells. Total RNA was isolated and the levels of USP11 and KLF4 were determined by real-time quantitative PCR. B, HepG2 cells infected with the indicated lentiviral shRNAs were treated with CHX (30 μg/mL) for the indicated time.
The protein levels of KLF4, USP11 and actin were analysed by Western blotting. C, HepG2 cells were infected with the NC or USP11 shRNA and cell proliferation was monitored using CCK8 assay at the indicated time points. D, HepG2 cells infected with the indicated lentiviral shRNAs were treated sorafenib (0, 3, 6 or 12 μmol/L) during 24 h. Cell survival was measured using CellTiter-Glo (Promega). E, Anchorage-independent colony formation of HepG2 cells stably expressing indicated shRNAs was determined by soft agar assay. Photographs of Petri dishes in a representative experiment and the average number of colonies from three experiment were indicated. F, KLF4 and USP11 expression in different liver cells (T2, THLE2; H2, HepG2; H3, Hep3B; H7, Huh7; S4, SNU423) were determined by qRT-PCR analysis. Relative expression levels are normalized by internal control (GAPDH). Each experiment was repeated 3 times (n = 3). Data represent means ± SD. A and E, two-tailed Student's t test; C and D, two-way ANOVA; F, one-way ANOVA, *P < .05; **P < .01; ***P < .001; ****P < .0001 HCC. Importantly, KLF4 has been recognized as a HCC prognostic marker. 11 Huh7, SNU326 and SNU423). The targeted gene expression levels were calculated using the 2 -∆ ∆ Ct method and were normalized to GAPDH expression. F, mRNA expression levels of KLF4 and USP11 of 90 RNA-seq data with HCC. The raw data were obtained from https://www. cbiop ortal.org. G, The RSEM normalized mRNA expression levels of KLF4 and USP11 in 60 NAFLD patients by using RNA-Seq analysis. The sequencing data for NAFLD patients were downloaded from NCBI SRA (https://www.ncbi.nlm.nih.gov/sra). H, The negative correlation between KLF4 and USP11 expression levels was observed in public RNA-sequencing data from HCC patients. Pearson correlation was −0.21 (P < .05). Data represent means ± SD. A and D-H, two-tailed Student's t test; C, two-way ANOVA; *P < .05; **P < .01; ***P < .001; ****P < .0001 its incidence has increased in recent years because of lifestyle changes. 41

ACK N OWLED G EM ENTS
The authors are grateful to Dr Jung-Hwa Oh (Korea Institute of Toxicology) for technical and material assistance. This work was supported by the National Research Foundation of Korea (NRF) grant (NRF-2020R1A2C4002109).

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data available on request from the authors.