KCTD17-mediated Ras stabilization promotes hepatocellular carcinoma progression

Human transcriptomic analyses

All raw transcriptomic data were retrieved from the Gene Expression Omnibus under accession numbers GSE25097, GSE36376, and GSE202853. RNA-seq data and clinical information for the liver hepatocellular carcinoma (LIHC) cohort were downloaded from the Cancer Genome Atlas (TCGA) data portal. All computations were conducted using the R program (RStudio 1.4). Using the R environment for statistical analysis, gene set enrichment analysis was carried out using the clusterProfiler package [25]. All variables and signatures’ normalized enrichment score (NES), P-value, and false discovery rate (FDR) were derived in the R environment using the GTEx database as raw count data from all (371) primary tumor transcriptomes imported from TCGA-LIHC project, with the top and bottom 10% of samples based on KCTD17 expression analyzed after normalization to the log2(TMM+1) value (edgeR package). The findings of the GSEA results were visualized using R packages, including ggplot2, enrichplot, and enhanced volcano.

Animals

C57BL/6J (#664) and Rosa26-LSL-Cas9 knockin (#026175) [26] mice were purchased from the Jackson Labs. Rosa26-LSL-Cas9 KI mice were first treated with 25 mg/kg diethylnitrosamine (DEN; N0259; Sigma-Aldrich, St. Louis, MO, USA) on day 14 and then transduced with 1.5×10¹¹ genome copies per mouse of adeno-associated virus subtype 8 (AAV8) expressing either AAV8-TBG-Cre or AAV8-U6-Kctd17 sgRNA-TBG-Cre, to generate control or L-Kctd17 mice [18]. After transduction, mice were maintained on a choline-deficient, L-amino-acid-defined, high-fat diet (CDAHFD) (60 Kcal% fat and 0.1% methionine by weight, A06071302; Research Diets Inc, New Brunswick, NJ, USA) for 16–18 weeks [27].

To induce the rapid onset of mouse HCC, 10 mg of the plasmids encoding myr-Akt and N-Ras along with sleeping beauty transposase in a ratio of 25:1 was administered by hydrodynamic tail vein injection [28,29]. DEN or carbon tetrachloride (CCl₄) with fructose/palmitate/cholesterol (FPC)-enriched NASH diet-feeding to develop mouse HCC has been previously described [30].

For xenograft experiments, cells were subcutaneously injected into the left upper back of BALB/c nude mice (20 mice were randomly allocated to two groups). After 3–4 weeks, the mice were sacrificed and the tumor volume and weight were evaluated. The tumor volume was calculated using the following formula: V=L×W²/2. The mice were housed at 3–5 animals per cage, with a 12 h light/dark cycle, in a temperature-controlled environment. All animal experiments were approved by the Inha University Institutional Animal Care and Utilization Committee (IACUC).

Cell culture, constructs, transfection, and lentivirus infection

The 293T, HL-7702, HepG2, Sk-Hep1, Hep3B, or Hepa1c1c7 cell lines were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Plasmids encoding myc-tagged K-Ras or H-Ras were purchased from Sino Biological and transfected using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA) to the manufacturer’s protocol. Lentiviruses encoding empty, HA-tagged Kctd17, Kctd17 singleguide RNAs (sgRNAs), or doxycycline-inducible shKCTD17 [31] were generated by co-transfection of LentiX-293T cells (Takara, Kusatsu, Japan) with psPAX2 and pMD2.G. Viral supernatants containing 8 mL/mL of polybrene (Santa Cruz Biotechnology, Dallas, TX, USA) were applied to Hepa1c1c7 or Hep3B cells, which were then selected with puromycin (Thermo Fisher Scientific) before the expansion of single clones.

Cell proliferation, migration, and wound-healing assays

The cell proliferation reagent WST-1 (Roche, Basel, Switzerland) accessed the proliferation of HCC cells. Briefly, cells were cultured in 24- or 96-well plates at 37°C and 5% CO2. The WST-1 solution was added to each well and incubated at 37°C for 4 h prior to analysis using microplate reader (BioTek, Winooski, VT, USA) at 450 nm. For migration assays, cells were seeded into the upper chamber of 24-transwell plates with 8 mm diameter filters for 12 h, and then cells on the bottom of the filter were fixed with 4% paraformaldehyde and stained with crystal violet. For wound healing assays, cells were seeded in 24-well plates and allowed to grow to 70–80% confluence as a monolayer, which was gently scratched across the center of the well. The medium was then removed, and the wells were washed twice with PBS. Images were obtained from the same fields immediately after scratching and after 24 h or 36 h under a microscope.

RNA isolation and quantitative PCR

RNA was isolated using TRIzol (Invitrogen, Waltham, MA, USA) or NucleoSpin RNA (Clontech, Mountain View, CA, USA), and cDNA was synthesized using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA, USA), followed by quantitative RTPCR with TB Green Fast qPCR mix (Takara Bio Inc., Otsu, Japan) or 2X Universal SYBR Green Fast qPCR mix (ABclonal, Woburn, MA, USA) in a CFX Opus 96 Real-Time PCR detection system (Bio-Rad, Hercules, CA, USA).

Antibodies, Western blots and immunoprecipitation

Cells and liver tissues were lysed in modified RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.25% sodium deoxycholate, 10 mM sodium pyrophosphate, 10 mM b-glycerophosphate, 50 mM sodium fluoride, and EDTA free protease inhibitor cocktail [Roche]). For immunoprecipitation, cell lysates were incubated with primary antibody and protein A/G beads (Santa Cruz, SC-2003), and FLAG M2 or HA affinity gels (Sigma, F2426 or E6779). Protein samples were separated using SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes. Membranes were probed with antibodies against phospho-ERK1/2 (#4370), ERK1/2 (#9102), phospho-MEK1/2 (#9121), MEK1/2 (#4694), Myc-tag (#2276), HA-tag (#3724), and DYKDDDDK-tag (#14793) from Cell Signaling; Ras (05-516) from Merck; and b-actin (A5441) or a-tubulin (T5168) from Sigma.

Ubiquitination assays

Cells were collected, and lysates were prepared in 100 mL of 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 2% SDS, and then boiled at 100°C for 10 min. The lysates were diluted with 900 mL of Tris-buffered saline containing 1% NP-40. Immunoprecipitation was performed using 1 mg of cellular lysate and an anti-FLAG M2 affinity gel and then analyzed using Western blotting with the indicated antibodies.

ASO studies

Nonspecific control ASO and Kctd17 ASO-2 (5’-AAG GTA ATG ATT GTA G-3’) were synthesized by Ionis Pharmaceuticals [18,32], diluted in saline before injection, and administered by intraperitoneal injection to male C57/BL/6J mice at a dosage of 25 mg per kg body weight once weekly for 6 weeks. One week after the 1st injection of ASO, 10 mg of the plasmids encoding myr-Akt and N-Ras along with sleeping beauty transposase in a ratio of 25:1 were further administered via hydrodynamic tail vein injection for 6 weeks before the mice were euthanized [28]. C57/BL/6J mice treated with DEN and CDAHFD feeding were injected with control or Kctd17 ASO once weekly for the indicated times prior to sacrifice.

Immunohistochemistry

Following sacrifice and rapid excision of mouse livers, tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned into 7 mm thickness slides, which were then deparaffinized, rehydrated and stained for H&E or Ki-67 (MA5-14520 from Invitrogen).

Statistical analysis

All data are shown as mean±standard error of the mean. Differences between the two groups were calculated using a two-tailed Student’s t test. Analyses involving multiple groups were performed using one-way ANOVA or the Kruskal–Wallis’s test. P<0.05 was considered statistically significant.

KCTD17 is upregulated in patients with HCC and associated with poor prognosis

We observed higher expression of KCTD17 in both human and mouse HCC and adenocarcinoma cell lines (Huh7, HepG2, Sk-Hep1, Hep3B, and Hepa1c1c7) than that in normal cells (HL7702 and AML12) (Supplementary Fig. 1A, B). We assessed patients with HCC databases, which revealed higher KCTD17 levels in HCC tissues than that in adjacent non-tumor tissues (Fig. 1A). Next, we investigated the association with clinicopathological features in patients with HCC classified and analyzed according to tumornode-metastasis (TNM) stage; in this cohort, patients with advanced stage (III/IV) disease showed higher KCTD17 as compared to those with early-stage disease (Fig. 1B). These data suggested that KCTD17 expression may predict survival. To test this possibility, we conducted a Kaplan-Meier estimation analysis. Patients were stratified into high- and low-KCTD17 expression groups using optimal expression levels determined by multiple receiver operating characteristic curve analysis. Patients with HCC and high KCTD17 expression showed a poorer prognosis (Fig. 1C). Consistently, across the TNM stages, higher expression of KCTD17 was consistently associated with decreased survival (Fig. 1D). These findings underscore the potential of Kctd17 as a prognostic marker of advanced HCC and poor survival.

Figure 1.

Increased KCTD17 is associated with poor prognosis in hepatocellular carcinoma (HCC). (A) Box and whiskers plots showing expression levels of KCTD17 in non-tumor and tumor from HCC in two independent cohorts (GSE25097, non-tumor=249, tumor=268; GSE36376, non-tumor=193, tumor=240). (B) Box plots showing expression levels of KCTD17 in each stage (stage I=171, II=86, and III+IV=88) in the liver hepatocellular carcinoma (LIHC) from the Cancer Genome Atlas (TCGA) database. Significance is determined using the Wilcoxon rank sum test. (C) Survival probabilities (Kaplan–Meier curves) (left) or cumulative hazard curve (right) based on hepatic KCTD17 expression. The R package, multipleROC, is divided into two groups (i.e., high vs. low KCTD17 expression) according to the optimal gene expression level of KCTD17. (D) Kaplan–Meier curves comparing hepatic KCTD17 expression and overall survival (top) or cumulative hazard curve (bottom) in different stages of HCC. ***P<0.001, ****P<0.0001 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±standard error of the mean.

KCTD17 promotes proliferation, migration, and drug resistance in HCC

To investigate the implications of increased KCTD17 expression in HCC, we generated Kctd17 knockout (KO) or doxycycline (Dox)-inducible KCTD17 knockdown (KD) cells using sgRNA or shRNA against Kctd17 in Hepa1c1c7 or Hep3B cell lines (Supplementary Fig. 2A, B), which had moderate or high expression of KCTD17. In both experiments, cells lacking KCTD17 showed lower proliferation than that in the control cells (Fig. 2A, B). To test if observed growth inhibition may affect metastatic potential of cancer cells, we evaluated the effect of KCTD17 on tumor migration using trans-well and wound healing assays. Indeed, KCTD17 depletion impeded the migration of HCC cells (Fig. 2C–E). Consistent with the altered cellular behavior, spheroid formation assays revealed significantly smaller spheroids in cells lacking KCTD17 than those in control cells (Fig. 2F).

Figure 2.

KCTD17 deficiency attenuates hepatocellular carcinoma (HCC) cell growth and migration. (A, B) Cell proliferation of control or Kctd17 KO Hepa1c1c7 cells (A), or doxycycline (Dox)-inducible KCTD17 knockdown Hep3B cells (B). (C, D) Representative pictures and quantitation of transwell migration assay in Kctd17 KO Hepa1c1c7 cells (C), or (D) KCTD17 knockdown Hep3B cells. (E) Representative pictures and quantitation of wound width in KCTD17 knockdown Hep3B cells. (F) Representative pictures and quantitation of three-dimensional spheroid cluster assay in KCTD17 knockdown Hep3B cells. *P<0.05, **P<0.01, ***P<0.001 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±s.e.m.

We next established a gain-of-function model by overexpressing Kctd17 in Hepa1c1c7 cells (Supplementary Fig. 2C). As anticipated, Kctd17 overexpression showed the opposite results in loss-of-function experiments—increased cellular proliferation, migration, wound healing, and spheroid formation (Fig. 3A–D)—associated with increased expression of oncogenic genes such as Myc and B-Cell-specific Moloney murine leukemia virus integration site 1 (Bmi1) (Fig. 3E). These findings suggested that Kctd17 affects key HCC properties. As resistance to chemotherapy poses a major obstacle in patients with advanced HCC, given the tumorigenic effects seen above, we next investigated whether Kctd17 influences the response to two commonly used chemotherapeutic approaches in HCC, TKI (sorafenib) [33] and anti-metabolite (5-FU) [34] drug treatment. In fact, forced Kctd17 inhibited the sensitivity of HCC cells to both drugs in a dose-dependent manner (Fig. 3F and Supplementary Fig. 2D), including the effects of these agents on MAPK and Akt pathways and spheroid formation (Fig. 3G, H). Collectively, these data demonstrate that increased Kctd17 expression is likely associated with chemoresistance in HCC.

Figure 3.

Kctd17 promotes hepatocellular carcinoma (HCC) proliferation, migration and drug resistance. (A–D) Cell proliferation (A), cell migration (B), would healing ability (C), three-dimensional spheroid cluster assay (D) and gene expression of Myc and Bmi1 (E) in control or Kctd17 overexpressing Hepa1c1c7 cells. (F–H) Control or Kctd17-overexpressing cells were treated with different dosages of sorafenib for 48 h, followed by cell proliferation assay (F), Western blot (G) and three-dimensional spheroid cluster assay (H). *P<0.05, **P<0.01 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±s.e.m.

KCTD17 drives Ras signaling

To understand the mechanism of KCTD17-induced HCC, we performed functional annotation of differential expression using human datasets, which revealed significant enrichment of Ras-related signaling features (Fig. 4A–D). We observed enrichment in KRAS and ERK1/2 (extracellular signal-regulated kinase 1/2) signaling, which are features indicative of malignant characteristics, including high proliferative activity. Similar results were observed in the GSE202853 dataset (Supplementary Fig. 3A–D), further supporting the association between KCTD17 expression and activation of the Ras-related signaling pathway in HCC. We confirmed activation of Ras/MAPK (mitogen-activated protein kinase) signaling by Kctd17 upon serum stimulation (Fig. 4E). Conversely, KCTD17 depletion attenuated this activation (Fig. 4F), suggesting that KCTD17 enhances growth factor-induced Ras activation.

Figure 4.

KCTD17 drives Ras signaling. (A) Violin plot based on the expression of KCTD17 (red: 37 top expressed samples, blue: 37 bottom expressed samples). (B, C) Volcano (B) and bubble plots (C) of normalized enrichment score (NES) distribution with RAS signalingrelated gene sets marked. (D) RNA-seq datasets of low-KCTD17 enriched and high-KCTD17 enriched as compared with KRAS and ERK signature in LIHC from TCGA. (E, F) Hepa1c1c7 cells expressing control or HA/Kctd17 (E) or Hep3B cells expressing doxycycline-inducible shKCTD17 (F) were cultured in serum-free medium and then treated with serum for the indicated time, followed by evaluation of Ras/MAPK signaling using Western blot.

Kctd17 mediates Ras stabilization by Lztr1 degradation

Interestingly, we observed increased Ras protein levels with Kctd17 overexpression (Fig. 5A), without a change in Ras mRNA expression (Supplementary Fig. 4A). Kctd17 overexpression also attenuated the degradation rates of endogenous and exogenous Ras, as demonstrated by measurement in the presence of the de novo protein synthesis inhibitor cycloheximide (CHX) (Fig. 5B and Supplementary Fig. 4B). These results suggest that Kctd17 promotes MAPK signaling through Ras stabilization.

Figure 5.

Kctd17 prompts Lztr1 degradation to mediate Ras stabilization. (A) Western blots from 293T cells transfected with H/K-Ras with or without Kctd17. (B) Hepa1c1c7 cells transfected with control or Kctd17 were treated with 50 mg/mL cycloheximide (CHX) and then harvested at the indicated time points after treatment prior to Western blotting and quantitation of endogenous Ras. (C, D) Co-immunoprecipitation of Ras and Lztr1 or Kctd17 from Hepa1c1c7 cells (C) or Western blots in lysate from 293T cells (D). (E) Western blots from Hepa1c1c7 cells transfected with Kctd17 and Lztr1 with or without MG-132. (F) Co-immunoprecipitation of Lztr1 by Kctd17 from Hepa1c1c7 cells. (G) Polyubiquitination of Lztr1 by Kctd17 Myc/Ub-transfected 293T cells with MG-132. (H) Ras ubiquitination in Flag/Ub-transfected 293T cells with or without Kctd17 and Lztr1. (I) Cell proliferation in control or Kctd17 overexpressing cells with or without Lztr1/Flag from Hepa1c1c7 cells. **P<0.01, ***P<0.001 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±s.e.m.

We next sought to understand the mechanism by which Kctd17 regulates Ras stability. Previous studies have reported that leucine zipper-like transcription regulator 1 (Lztr1) promotes polyubiquitination and degradation of Ras family members by recruiting a Cul3 ubiquitin ligase complex, thereby inhibiting Ras/MAPK signaling [22-24]. Germ-line mutations in Lztr1 are associated with Noonan syndrome, a RASopathy [35]. Furthermore, Lztr1 is mutated in HCC [36] and its expression is significantly lower in normal cells than those in HCC cells [37]. Given these data, we hypothesized that Kctd17 affects Ras stability by interfering with the Lztr1-Ras interaction. Intriguingly, we observed that Kctd17 reduced Lztr1 protein levels which disrupted the interaction between Lztr1 and Ras (Fig. 5C), without altering Lztr1 mRNA expression (Supplementary Fig. 4C). Similarly, the Lztr1-mediated loss of Ras was prevented by Kctd17 overexpression (Fig. 5D), whereas the Kctd17-mediated reduction of Lztr1 was rescued by concomitant MG-132 treatment to prevent proteasomal degradation (Fig. 5E). Consistent with these data, we observed that Kctd17 interacted with Lztr1 (Fig. 5F) and increased its ubiquitination (Fig. 5G). As such, Lztr1-mediated Ras polyubiquitination was inhibited by Kctd17 under denaturing conditions (Fig. 5H). Finally, Kctd17 reversed the anti-proliferative effects of Lztr1 on hepatoma cells (Fig. 5I). Taken together, our results indicate that Kctd17 induces Lztr1 proteasomal degradation, which provokes Ras stabilization and cellular proliferation.

Kctd17 is required for HCC progression in vivo

To test repercussion of altered Kctd17 expression in HCC, we first conducted in vivo tumorigenicity assays using a xenograft mouse model. Both Kctd17 KO (Fig. 6A–D) or doxycycline-inducible KCTD17 knockdown (Fig. 6E–H) resulted in a significant decrease in tumor growth, leading to lower tumor size and weight, without significant changes in body weight (Supplementary Fig. 5A, B).

Figure 6.

Kctd17 is required for hepatocellular carcinoma (HCC) progression in vivo. (A) Experimental schematic for tumor xenograft assay using control or Kctd17 KO Hepa1c1c7 cells. (B–D) Representative pictures (B), tumor size (C) and weight (D) of control or Kctd17 KO xenografts (n=10 per group). (E) Experimental schematic for tumor xenograft assay using doxycycline-inducible KCTD17 knockdown Hep3B cells. (F–H) Representative pictures (F), tumor size (G) and weight (H) of control or doxycycline-inducible shKCTD17 xenografts (n=10 per group). (I) Experimental schematic for hepatocyte-specific Kctd17 KO mice (L-Kctd17). (J–N) Kctd17 expression in tumors and adjacent non-tumors (J), representative images of whole livers (K) and tumor volume (L), Representative images of livers stained with H&E or Ki-67 (M), Western blots from liver tumor lysates (N) of control or L-Kctd17 mice (n=7 to 9 per group), as indicated. Scale bar, 100 or 200 mm. *P<0.05, **P<0.01, ***P<0.001 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±s.e.m.

We assessed Kctd17 expression in several mouse HCC models, including carcinogen-induced (CCl₄ or DEN), dietinduced (CDAHFD feeding or fructose/palmitate/cholesterol (FPC)-enriched NASH diet-feeding) [27,30], and oncogene induction (myr-Akt/N-Ras) [38,39]. All showed higher Kctd17 mRNA expression in tumors than in non-tumors (Supplementary Fig. 6A–D). To ascertain the role of Kctd17 in HCC pathogenesis, we treated hepatocyte-specific Kctd17 (L-Kctd17) knockout mice [18] with DEN, and then fed a CDAHFD to accelerate HCC [27] (Fig. 6I, J). L-Kctd17 mice showed reduced tumor numbers, Ki-67-positive cells, lower Ras, and increased Lztr1 protein levels, compared to those in control mice, without significant changes in body weight (Fig. 6K–N, Supplementary Fig. 6E). These data demonstrate that Kctd17 functions as a tumor promoter in HCC.

Kctd17 inhibition ameliorates HCC progression

Next, to assess the potential therapeutic applications of these data, we tested the effects of liver-directed Kctd17 ASO [18] in mouse HCC. We leveraged the rapid onset of HCC development and the large number of tumors in the oncogene models (Fig. 7A). Treatment with Kctd17 ASO reduced Kctd17 expression in both tumor and adjacent non-tumor liver tissues, decreased tumor size, lowered Ras levels, and increased Lztr1 protein levels (Fig. 7B–E).

Figure 7.

Kctd17-specific ASO decreases hepatocellular carcinoma (HCC) progression. (A) Experimental schematic for oncogene-induced HCC models. After hydrodynamic injection of myr-Akt and N-Ras in WT male mice, control or Kctd17-directed ASO was administered weekly by intraperitoneal injection. (B–E) Kctd17 expression in tumors and adjacent non-tumors (B), representative images of whole liver (C), tumor size (D) and Western blots from liver tumor lysate (E) in control or Kctd17 ASO-treated mice (n=5 or 7 per group). (F) Experimental schematic for DEN/CDAHFD-induced HCC models. Mice were injected with DEN at 2 weeks of life, then started on CDAHFD diet-feeding with concurrent weekly intraperitoneal injections of control or Kctd17 ASO. (G–K) Kctd17 expression in tumors and adjacent non-tumors (G), representative images of whole mouse liver (H), tumor number (I), liver H&E and Ki-67 staining (J), and Western blots in lysate from liver tumor lysate (K), in control or Kctd17 ASO-treated mice (n=6 per group). Scale bar, 100 or 200 mm. (L) Model of Kctd17-mediated HCC progression. *P<0.05, **P<0.01, ***P<0.001 as compared to the indicated control by two-way ANOVA. All data are shown as the mean±s.e.m.

We confirmed these data in DEN/CDAHFD-induced mouse HCC (Fig. 7F). Similar to above, Kctd17 ASO reduced Kctd17 mRNA expression, tumor number, Ki-67-positive cells, and Ras protein levels, whereas it increased Lztr1 protein levels, without significant changes in body weight (Fig. 7G–K, Supplementary Fig. 6F). These data suggest that the liver-selective inhibition of Kctd17 can attenuate HCC progression in vivo, indicating its potential as a therapeutic target for the treatment of liver tumors.