Loss of ATOH8 Increases Stem Cell Features of Hepatocellular Carcinoma Cells

Accepted Manuscript Loss of ATOH8 Increases Stem Cell Features of Hepatocellular Carcinoma Cells Yangyang Song, Guangjin Pan, Leilei Chen, Stephanie Ma, Tingting Zeng, Tim Hon Man Chan, Lei Li, Qizhou Lian, Raymond Chow, Xiujuan Cai, Yan Li, Yan Li, Ming Liu, Yun Li, Yinghui Zhu, Nathalie Wong, Yun-Fei Yuan, Duanqing Pei, Xin-Yuan Guan

PII: DOI: Reference:

S0016-5085(15)00868-9 10.1053/j.gastro.2015.06.010 YGAST 59843

To appear in: Gastroenterology Accepted Date: 13 June 2015 Please cite this article as: Song Y, Pan G, Chen L, Ma S, Zeng T, Chan THM, Li L, Lian Q, Chow R, Cai X, Li Y, Li Y, Liu M, Li Y, Zhu Y, Wong N, Yuan Y-F, Pei D, Guan X-Y, Loss of ATOH8 Increases Stem Cell Features of Hepatocellular Carcinoma Cells, Gastroenterology (2015), doi: 10.1053/ j.gastro.2015.06.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. All studies published in Gastroenterology are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.

ACCEPTED MANUSCRIPT Loss of ATOH8 Increases Stem Cell Features of Hepatocellular Carcinoma Cells Yangyang Song,1, 6 Guangjin Pan,8 Leilei Chen,6,7 Stephanie Ma,2, 4, 5 Tingting Zeng,9 Tim Hon Man Chan,6 Lei Li, 9 Qizhou Lian,3 Raymond Chow, 1, 4, 5 Xiujuan Cai,8 Yan Li,1, 4, 5

Duanqing Pei,8 Xin-Yuan Guan1, 4, 5, 9,*

1

Clinical Oncology,

2

Anatomy and

3

Medicine,

4

Centre for Cancer

SC

Departments of

RI PT

Yan Li,9 Ming Liu, 1, 4, 5 Yun Li,1, 4, 5 Yinghui Zhu,9 Nathalie Wong,10 Yun-Fei Yuan,9

M AN U

Research, 5State Key Laboratory of Liver Research, Li Ka Shing Faculty of Medicine, The University of Hong Kong, Hong Kong, China; 6Cancer Science Institute of Singapore and 7

Department of Anatomy, National University of Singapore, Singapore; 8Key Laboratory

of Regenerative Biology，Guangzhou Institutes of Biomedicine and Health, Chinese

TE D

Academy of Sciences, Guangzhou, China; 9State Key Laboratory of Oncology in Southern China, Sun Yat-sen University Cancer Center, Guangzhou, China;

10

Department of

AC C

China.

EP

Anatomical and Cellular Pathology, The Chinese University of Hong Kong, Hong Kong,

*Correspondence: Professor Xin-Yuan Guan, Department of Clinical Oncology, The University of Hong Kong, Room 56, 10/F, Laboratory Block, 21 Sassoon Road, Hong Kong. Tel: (852) 3917-9782; Fax: (852) 2816-9126; E-mail: [email protected]. Funding This work was generously supported by grants from NSFC/RGC Joint Research Scheme (N_HKU712/12), China National Basic Research Program (2012CB967001), China 1

ACCEPTED MANUSCRIPT National Key Sci-Tech Special Project of Infectious Diseases (2013ZX10002-011), Hong Kong RGC Collaborative Research Funds (C7027-14G and C7038-14G), RGC Theme-based Research Scheme fund (T12-403/11), RGC GRF (767313), National Natural

RI PT

Science Foundation of China (81272416 and 31261160494), National Basic Research Program of China (2012CB966802 and 2011CB965204), the "Strategic Priority Research Program" of the Chinese Academy of Sciences (XDA01020401), National S&T Major

SC

Special Project on Major New Drug Innovation (2011ZX09102-010), Ministry of Science

M AN U

and Technology International Technology Cooperation Program (2012DFH30050), Bureau of Science and Technology of Guangzhou Municipality (2010U1-E00521) and Key Laboratory of Regenerative Biology (2008DP173344).

TE D

Conflicts of interest The authors disclose no conflicts.

EP

Author contributions

AC C

Y. S. (Department of Clinical Oncology, The University of Hong Kong, Hong Kong) and X-Y. G. initiated and designed the study. Y. S. wrote the manuscript with input from X-Y.G. Y. S., G. P. (Key Laboratory of Regenerative Biology，Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou, China) and L.C. designed the experiments and interpreted the results. Y. S. performed all cancer related experiments with assistance from S.M., T. Z., T.H.M.C., L. L., R. C., Yun Li , Yan Li, M.L., Y. Z. and N. W. iPS technology was provided by G. P. and D. P. And the experiments 2

ACCEPTED MANUSCRIPT related to iPSCs were performed by Y. S. with assistance from X. C. and Q-Z. L. Illumina transcriptome sequencing was performed by L.C. (Cancer Science Institute of Singapore, National University of Singapore, Singapore). HCC clinical samples and the relevant

RI PT

clinical information were provided by Y.-F.Y. and extracted by L.C., Y. S., Z. T., Z. Y. and Yan Li (State Key Laboratory of Oncology in Southern China, Sun Yat-sen University

SC

Cancer Center, Guangzhou, China). X.-Y.G. supervised the project.

M AN U

Abbreviations used in this paper: 5-Aza, 5-azacytidine; 5-FU, 5-fluorourail; AFP, Alpha-fetoprotein; AJCC, American Joint Committee on Cancer; ALB, Albumin; ATOH8, Atonal Homolog 8; AP, Alkaline phosphatase; ATRA, all-trans retinoic acid; bp, base pair; c-MYC, c-mycmyelocytomatosis viral oncogene homolo; cDNA, complementary DNA;

TE D

CDDP, Cisplatin; CK8, Cytokeratin 8; CK18, Cytokeratin 18; CpG, Cytosine phosphate guanine; CSCs, cancer stem cells; EMSA, Electrophoretic mobility shift assay; EB, Embryonic body; ESCs, embryonic stem cells; FACS, fluorescence-activated cell sorting;

EP

GATA4, GATA binding protein 4; H&E, Hematoxylin and eosin; HCC, hepatocellular

AC C

carcinoma; IF, Immunofluorescence; IHC, immunohistochemistry; iPSCs, Induced pluripotent stem cells; MOI, Multiplicities of Infection; mRNA, messenger RNA; non-CSCs, Non-cancer stem cells; OCT4, Octamer-binding transcription factor 4; qRT-PCR, quantitative real-time polymerase chain reaction; shRNA, short hairpin RNA; RNAi,

RNA

interference;

RNA-Seq,

RNA

sequencing;

RT-PCR,

Reverse

transcription-polymerase chain reaction; SNP, Single nucleotide polymorphism; SOX1, Sex determining region Y-box 1, SOX2, Sex determining region Y-box 2; SOX17; Sex 3

ACCEPTED MANUSCRIPT determining region Y-box 17; SSEA4, Stage-specific embryonic antigen 4; TBX1, T-box protein 1; TRA-1-60, Tumor rejection antigen-1-60; TRA-1-81, Tumor rejection

AC C

EP

TE D

M AN U

SC

RI PT

antigen-1-81.

4

ACCEPTED MANUSCRIPT BACKGROUND & AIMS: Levels of atonal homolog 8 (ATOH8) are reduced in 48% of hepatitis B virus-associated hepatocellular carcinomas (HCCs). ATOH8 downregulation is associated with loss of tumor differentiation, indicating an effect mediated by cancer stem cells. We investigated the effects of loss of ATOH8 in human hepatocellular carcinoma

RI PT

(HCC) cells and cell lines.

METHODS: HCC and adjacent non-tumor tissues were collected, from 2001 through 2012, from 242 patients undergoing hepatectomy at Sun Yat-Sen University Cancer Center

SC

in China; 83% of HCCs were associated with HBV infection. CD133+ cells were isolated from tumor tissues by flow cytometry. Experiments were performed in HBV-positive and

M AN U

HBV-negative HCC cell lines, the immortalized liver cell line LO2, and 8 other HCC cell lines. ATOH8 was expressed from lentiviral vectors in PLC8024 and Huh7 cells; levels were knocked down with small interfering RNAs in QSG7701 cells. Cells carrying empty vectors were used as controls. Gene regulation by ATOH8 was assessed in mobility shift and luciferase reporter assays. Cells were analyzed in proliferation, foci formation, and

TE D

colony formation assays. The tumorigenic and chemo-resistant potential of cells were investigated by assessing growth of xenograft tumors in immunocompromised mice. Metastatic features of cells were assessed in Matrigel invasion assays and wound healing

EP

analyses.

RESULTS: Levels of ATOH8 mRNA were reduced by more than 4-fold, compared to

AC C

non-tumor tissues, in 118/242 HCC samples (48.8%). Patients with tumor reductions in ATOH8 had significantly shorter times of disease-free survival (mean, 41.4 months) than patients with normal tissue levels (mean, 52.6 months). ATOH8 expression was reduced in HepG2, Huh7, PLC8024 and CRL8064 HCC cells, as well as CD133+ cells isolated from human HCC samples. Transgenic expression of ATOH8 in HCC cell lines significantly reduced proliferation and foci colony formation, as well as their invasive and migratory abilities. Transgenic expression of ATOH8 reduced the ability of HBV-positive PLC8024 cells to form tumors in mice, compared to control cells. Cells with ATOH8 knockdown formed xenograft tumors more rapidly, in more mice, than control cells. ATOH8 repressed 5

ACCEPTED MANUSCRIPT transcription of stem-cell associated genes including OCT4, NANOG, and CD133. Knockdown of ATOH8 in CD133-negative QSG7701 cells caused them to express CD133; acquire self-renewal, differentiation, chemo-resistance properties; form more xenograft tumors in mice; and generate induced pluripotent stem cells (based on staining for alkaline

RI PT

phosphatase and their ability to form embryoid bodies and teratomas). Alternatively, expression of ATOH8 in PLC8024 and Huh7 cells significantly reduced the numbers of cells expressing CD133, and increased the chemo-sensitivity of Huh7 cells to

SC

5-fluorouracil (5-FU) and cisplatin, in vitro and in mice.

CONCLUSIONS: ATOH8 appears to be a tumor suppressor that induces stem-cell

M AN U

features and chemoresistance in HCC cells. Strategies to restore its levels and activities might be developed to treat patients with liver cancer.

KEYWORDS: cancers stem cell, iPSCs, tumor progression, chemotherapy drug

AC C

EP

TE D

resistance

6

ACCEPTED MANUSCRIPT Hepatocellular carcinoma (HCC) is one of the most common solid tumors in the world with extremely poor prognosis.

1

It is particularly prevalent in China, Southeast and

Eastern of Asia and sub-Saharan Africa.2 In China, the development of HCC is closely associated with hepatitis B viral (HBV) infection. Despite advances in the detection and

RI PT

treatment of HCC, the mortality rate remains high because current therapies are limited by the advanced stage of the disease. Recent advances in cancer stem cell (CSC) studies have associated the poor prognosis of cancer with CSCs.3 CSCs were first described in leukemia,4 and have been reported in many solid tumors, including breast cancer,5 brain

SC

cancer 6 and HCC.7 CSC theory suggests that tumors are organized within a hierarchy of mixed tumor cells, with the CSCs at the apex.8-12 CSCs possess self-renewal and

M AN U

differentiation capabilities and are believed to give rise to tumor heterogeneity.

In CSC study, one important unclear question is whether this hierarchical structure is reversible, i.e., whether a non-CSC can be reprogrammed into a CSC? Induced pluripotent

TE D

stem cell (iPS) technology has demonstrated that non-pluripotent adult fibroblasts can be reprogrammed into iPSCs induced by SKOM (SOX2, KLF4, OCT4 and c-MYC).13

EP

Therefore, it is logical to hypothesize that a mature cancer cell can be reprogrammed into a CSC through specific genetic and/or epigenetic alterations. Several recent studies

AC C

suggested that cancer cell dedifferentiation could be manipulated in cell lines by controlling the expression of the stemness-associated genes Nanog,14Oct4,15 and TGFβ1.16 In the present study, we characterized a stemness regulator, Atonal Homolog 8 (ATOH8), which was frequently down-regulated in HCC. ATOH8 belongs to a group of basic-helix-loop-helix (bHLH) transcription factors and contains 321 amino acids with a bHLH domain that typically binds to a consensus sequence (CANNTG) E-box.17-19 bHLH transcription factors are involved in the regulation of many developmental processes, 7

ACCEPTED MANUSCRIPT including cardiovascular development,20 hematopoiesis,21 skeletal muscle development,22 neurogenesis,23 and embryogenesis.24 Here we demonstrated that knockdown of ATOH8 could mark the potential CSCs in HCC and induce CD133- cells into CD133+ cells, which

RI PT

possessed CSC properties including abilities of self-renewal, differentiation and chemo-resistance.

SC

Materials and Methods

M AN U

Clinical Specimens and Cell Lines

Fresh human liver tumor and adjacent non-tumor tissue specimens were collected between 2001 and 2012 from 242 HCC patients undergoing hepatectomy at Sun Yat-Sen University Cancer Center (Guangzhou, China). Over 83% HCC patients are associated

TE D

with HBV infection. HBV-positive HCC cell lines QGY7703 and PLC0824, HBV-negative HCC cell lines QSG7701, BEL7402, SMMC7721 and immortalized liver cell line LO2 were obtained from the Institute of Virology, Chinese Academy of Medical

EP

Sciences (Beijing, China). Huh7, CRL8064 and hepatoma cell line HepG2 were purchased

AC C

from American type culture collection (ATCC, Manassas, Virginia, USA). The SF002, A153, G2F1 and hNF1 cell lines were obtained from the Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou, China.

Cell Culture. All cancer cells and human fibroblasts were maintained in high-glucose Dulbecco’s Modified Eagle Medium (DMEM, Hyclone) supplemented with 10% fetal bovine serum 8

ACCEPTED MANUSCRIPT (FBS, Gibco), 0.1mM MEM Non-Essential Amino Acids (NEAA) (Invitrogen), 1% L-glutamine (Invitrogen). iPSC colonies were routinely maintained on feeders in Knockout Serum Replacement (KSR, Invitrogen) medium or on Matrigel (BD Biosciences)

RI PT

in mTeSR1 (Stemcell Technologies) on 24-well plates. All cells were incubated in 5% CO2 humidified chamber at 37°C.

SC

Interactions between ATOH8 and Stemness-Associated Genes

M AN U

The regulatory mechanism of the ATOH8 and stemness-associated genes interaction was investigated by EMSA and luciferase reporter assay (Supplementary Materials and Methods).

TE D

FACS

CD133+ cells were sorted by FACS with PE-conjugated anti-human CD133/2 antibody (MiltenyiBiotec). The isotype control mouse IgG1b-PE (MiltenyiBiotec) was served as a

EP

control. Only the top 2%, corresponding to the most brightly stained cells, or the bottom

AC C

2%, corresponding to the most dimLy stained cells, were selected as CD133 positive and negative populations, respectively.

In Vitro and in Vivo Functional Assays In vitro tumorigenic abilities were assessed by XTT cell proliferation assay, foci

formation assay, and colony formation assay in soft agar. In vitro metastatic abilities were assessed by invasion assay and wound healing assay. In vivo tumorigenic ability and 9

ACCEPTED MANUSCRIPT chemo-resistant assays were investigated in xenograft mice models (Supplementary Materials and Methods).

RI PT

Characterization of iPSCs iPSCs were characterized by AP staining, IF, EB assays and teratoma formation assays

SC

(Supplementary Materials and Methods).

M AN U

Results ATOH8 is Frequently Down-Regulated in HCC

Our previous high-throughput RNA-Seq in three pairs of HCCs identified 102 down-regulated and 187 up-regulated genes in all tested tumor tissues compared to

TE D

corresponding non-tumor tissues (Supplementary Figure 1A).25 One of the down-regulated genes, ATOH8 (Supplementary Figure 1B), was further characterized. Down-regulation (defined as >4-fold change) of ATOH8 was detected in 118/242 (48.8%) of HCC tissues

EP

by qRT-PCR, compared with adjacent non-tumor tissues. The average level of ATOH8

AC C

expression in tumor tissues was significantly lower than that in non-tumor tissues (0.34 versus 1.18, P<0.001, paired Student’s t test; Figure 1A). Clinical association study found that down-regulation of ATOH8 was significantly associated with poor differentiation (P=0.01) and high serum AFP level (P=0.03) (Supplementary Table 1). Kaplan-Meier survival analysis found that patients with ATOH8 down-regulation displayed a worse disease-free survival (DFS) (estimated mean=41.4 months) compared to patients without ATOH8 down-regulation (estimated mean=52.6 months) (log-rank=4.631, P=0.031; 10

ACCEPTED MANUSCRIPT Figure 1B). Down-regulation of ATOH8 in protein level was also observed in 49/104 (47.1%) of HCC tumor tissues by immunohistochemistry (IHC) staining (Figure 1C). Expression of ATOH8 in 1 immortalized liver cell lines (LO2) and 8 HCC cell lines was

RI PT

detected by RT-PCR and western blot analysis. Compared with LO2 cells, down-regulation of ATOH8 was detected in HepG2, Huh7, PLC8024 and CRL8064 cells

SC

(Figure 1D).

M AN U

ATOH8 Possess a Tumor Suppressive Function in HCC

To investigate the tumor suppressive role of ATOH8, full-length human ATOH8 was cloned into a lentiviral vector, and then stably transfected into HCC cell lines PLC8024 and Huh7 (Figure 1E). To further confirm the tumor suppressive function of ATOH8, in

TE D

vitro assays was performed to assess the tumorigenicity of ATOH8-expressing QSG7701 cells in which ATOH8 expression was silenced with two siRNAs (si22 and si45) targeted against ATOH8 (Figure 1E). Both in vitro and in vivo functional assays were applied to

EP

investigate the tumor suppressive role of ATOH8. Functional assays indicated that ATOH8

AC C

significantly inhibit the tumor cell growth rate by XTT proliferation assay found (Figure 1F). Foci formation and soft agar assays found ATOH8 to significantly reduce the frequency of colony formation on solid plates (Figure 2A) and in soft agar (Figure 2B), indicating that ATOH8 inhibited tumor growth in both anchorage-dependent and independent manners. We also studied the effect of ATOH8 down-regulation on HCC tumor invasion and metastasis. Matrigel invasion and wound healing assays found ATOH8 to efficiently suppress the invasive (Figure 2C) and migratory (Figure 2D) abilities, 11

ACCEPTED MANUSCRIPT respectively. In addition, ATOH8 expression was silenced in QSG7701 cell line with two siRNAs (si22 and si45) targeted against ATOH8 (Figure 1E). As expected, the tumorigenic abilities of QSG7701, including tumor growth rate, foci formation, cell migratory and

RI PT

invasive ability were significantly increased when ATOH8 was knocked down (Figure 1E and Figure 2A-D).

For in vivo tumor formation assay, tumor formation was observed in 3/5 and 5/5 of

SC

mice injected with ATOH8-transfected and empty vector-transfected PLC8024 cells,

M AN U

respectively. Similar results were observed in 2/5 and 3/5 of mice injected with ATOH8-transfected and control Huh7 cells. In addition, the tumor sizes and tumor volumes of tumors induced by empty vector-transfected cells were significantly larger than ATOH8 overexpressed cells (Figure 2E, left). Furthermore, shRNAs (sh4 and sh7)

TE D

against ATOH8 were cloned into the lentiviral vector and then stably transfected into QSG7701, respectively. Tumor formation was observed in all SCID mice (n=5) injected with sh4- and sh7-transfected cells within 7 weeks, whereas tumor formation was only

EP

observed in 2/5 SCID mice injected with control cells. The tumor sizes and tumor volumes

AC C

were significantly larger in tumors induced by ATOH8-silenced QSG7701 cells compared with tumors induced by control QSG7701 cells (Figure 2E, right). These findings demonstrate that ATOH8 could suppress both the in vitro and in vivo tumorigenicity of HCC cells.

ATOH8 Represses Transcriptional Activity of Stemness-Associated Genes. As bHLH transcription factors often regulate genes associated with embryogenesis and 12

ACCEPTED MANUSCRIPT developmental processes, MatInspector Professional software (Genomatix, Munich, Germany)26, 27 was applied to investigate the potential regulatory effect of ATOH8 on stemness-associated genes (OCT4, NANOG, SOX2), differentiation-associated marker Results showed that the above genes all

RI PT

AFP,28 and CSC marker CD133 (Figure 3A).

contained at least one E-box motif within their promoter regions. qRT-PCR was used to confirm whether ATOH8 can regulate these genes at the transcriptional level, and results

SC

showed that these genes were significantly down-regulated in ATOH8-transfected cells

M AN U

compared with empty vector-transfected cells (Figure 3B). The repressive effect of ATOH8 on these genes was further extended in QSG7701 cells following ATOH8 silencing by si22 and si45. Results suggested that the expression levels of these genes were significantly increased compared with scrambled control cells when ATOH8 was

TE D

silenced (Figure 3C). Western blot analysis showed that the expression of these genes was decreased when ATOH8 was introduced into cells and increased when ATOH8 was

EP

silenced (Figure 3D).

13

, the

AC C

Since OCT4 and NANOG play key roles in the reprogramming of stem cells

repressive effect of ATOH8 on them was further studied by a luciferase reporter assay. The results showed that the luciferase activities of pGL3-OCT4, pGL3-NANOG and pGL3-CD133 were significantly decreased in cells co-transfected with pLenti6-ATOH8 (P<0.05), but not with pLenti6-vec (Figure 3E). Electrophoretic mobility shift assay (EMSA) was then used to determine whether ATOH8 could directly bind to the E-box in the promoter regions of OCT4 and NANOG. Results found that ATOH8 was able to 13

ACCEPTED MANUSCRIPT specifically bind to the biotin-labeled probe (Figure 3F), confirming the direct binding of ATOH8 to E-box motif of the target genes.

RI PT

Absent Expression of ATOH8 is Preferentially Observed in CD133+ CSC Subset We next studied the frequency and distribution of ATOH8- cells by IHC using a tissue microarray (TMA) containing 104 HCCs, and results showed that ATOH8- cells were

SC

observed in 81/104 (77.9%) of non-tumor tissues and all tumor tissues (Figure 4A). The

M AN U

average frequency in tumor tissues (2.52%, ranged 0.1-5.7%) was significantly higher (P<0.001) than that in non-tumor tissues (0.38%, ranged 0.1-2.2%, Figure 4A). As the expression level of ATOH8 was extremely low in CD133+ cells, we hypothesized that ATOH8- cells might be CD133+ CSCs in HCC. Double IF staining with ATOH8 (green)

TE D

and CD133 (red) was performed in cryopreserved tissue sections from 5 HCC clinical samples. Results demonstrated that absent expression of ATOH8 was preferentially detected in most CD133+ cells (Figure 4B). qRT-PCR also showed that the expression

EP

level of ATOH8 was significantly higher in CD133+ cells compared with CD133- cells

AC C

(Figure 4C). The negative correlation of ATOH8 and CD133 expressions was confirmed by Western blot analysis in clinical HCC samples (Figure 4D).

Knockdown of ATOH8 Increases CD133+ CSC Subset Because knockdown of ATOH8 led to an up-regulation of CD133, we next studied if ATOH8 silencing could induce the population of CD133+ cells in HCC cell lines. When ATOH8 was introduced into PLC8024 and Huh7 cells, the subpopulation of CD133+ cells 14

ACCEPTED MANUSCRIPT decreased significantly (Figure 4E). In QSG7701 and BEL7402 cells, ATOH8 depletion increased the proportion of CD133+ cells (Figure 4F), respectively. Interestingly,

RI PT

QSG7701 has been considered as CD133- cell line.12, 29

CD133+ Cells Induced by ATOH8 Silencing Possess Cancer Stem Cell-Like Properties

SC

To answer the question whether CD133+ cells induced by ATOH8 depletion possess

M AN U

cancer stem cell-like properties, ATOH8 was transiently silenced by si22 in QSG7701 cells and ATOH8-induced CD133+ cells were isolated by fluorescence-activated cell sorting (FACS) for functional studies. Compared to CD133- cells, sorted CD133+ cells possessed stronger tumorigenicity, as demonstrated by their increased ability to form foci (P<0.001,

TE D

Figure 5A), colonies in soft agar (P<0.001, Figure 5B), and tumor formation in SCID mice with fewer cells (P<0.001, Figure 5C). Xenograft tumor formation was observed in 4/5 and 2/5 animals when 1×105 and 1×104 of sorted CD133+ cells were subcutaneously

EP

injected into SCID mice, respectively. No tumor formation was observed when the same

AC C

numbers of CD133- cells were injected; while tumor formation was only observed in 2/5 of SCID mice when 4×106 of CD133- cells were injected (Figure 5C and Supplementary Table 2).

Spheroid formation assay found sorted CD133+ cells to possess a stronger ability to self-renew, as compared with CD133- cells (Figure 5D). RT-PCR results showed that expressions of stemness-associated genes were considerably higher in CD133+ cells 15

ACCEPTED MANUSCRIPT induced by ATOH8 silencing compared with CD133- cells (Figure 5E). The differentiation ability of sorted CD133+ and CD133- cells was induced by treatment with 4µM ATRA for 6 days. Following ATRA treatment, the expression of mature hepatocyte genes, including

RI PT

CK8,30 CK18,30 albumin,31 and ATOH8 were increased in CD133+ cells but not in CD133- cells; while the expression of AFP was attenuated in CD133+ cells but not in CD133- cells (Figure 5E). Furthermore, the potential chemo-sensitivity of sorted CD133+

SC

cells was tested by detecting the apoptotic cell population (pre-G1 cells detected as the M4

M AN U

subpopulation, Figure 5F). Flow cytometry results found sorted CD133+ cells to display a stronger ability to resist cisplatin (CDDP) treatment compared to CD133- cells (final concentration at 4µg/mL for 48 hours).

TE D

ATOH8 Depletion Increases the Efficiency of iPSCs Generation Because ATOH8 represses expression of many stemness-associated genes, we further tested whether ATOH8 depletion could increase the efficiency of iPSC generation.

EP

qRT-PCR result found that expression of ATOH8 could be detected in 4 parental

AC C

fibroblasts but not in corresponding iPSCs (Figure 6A). IF staining demonstrated that ATOH8 (green) expressed in fibroblasts, but not in corresponding iPSCs (Figure 6B).

To address this issue, fibroblasts (SF002) at passage 8 were infected by lentiviruses producing SKOM with or without shRNA targeting ATOH8. The frequency of AP-positive colony was significantly increased in fibroblasts treated with ATOH8 depletion (sh4: 4.18±0.19‰; sh7: 2.37±0.15‰) compared with fibroblasts treated without ATOH8 16

ACCEPTED MANUSCRIPT depletion (0.46±0.1‰, P<0.001), suggesting that ATOH8 depletion could enhance the efficiency of iPSC generation (Figure 6C). Eight AP-positive colonies from each of SKOM/sh4- and SKOM/sh7-treated cells and 5 colonies from SKOM/Ctl cells were

RI PT

further characterized. Results showed that 6 SKOM/sh4 colonies, 7 SKOM/sh7 colonies and 2 SKOM/Ctl colonies were passaged as stable iPSCs, respectively.

SC

These stable iPSCs had an ESC-like phenotype and displayed pluripotent ability. IF

M AN U

examination revealed that the expression of human ESCs makers (OCT4, SSEA4, TRA-1-60 and TRA-1-81) was detected in both ESCs (H1) and iPSCs (SF002-C-1, SF002-sh4-8 and SF002-sh7-2) but not in the original fibroblasts (SF002) (Figure 6D). We next evaluated the pluripotency of iPSCs and ESCs with an embryoid body (EB) assay.

TE D

qRT-PCR showed that the stemness-related genes were significantly down-regulated (P<0.01), while early markers of the 3 germ layers were significantly enhanced (P<0.01) in all tested iPSCs compared with EBs (Figure 6E). Finally, teratoma formation in nude

EP

mice was performed by subcutaneously injecting SF002-C-1, SF002-sh4-8 and

AC C

SF002-sh7-2 cells, respectively. The induced teratomas contained complex differentiated structures, such as bone, gastrointestinal epithelium and neural tube-like tissue, and were induced in all 3 tested cells (Figure 6F).

ATOH8 Increases the Chemo-Sensitivity of HCC Cells Since ATOH8 depletion enhanced CSCs population and the stemness of HCC cells, we next investigated whether ATOH8 could increase the chemo-sensitivity of HCC cells by 17

ACCEPTED MANUSCRIPT treating Huh7-ATOH8 and Huh7-Vec cells with 5-fluorouracil (5-FU) and CDDP. XTT assay showed that the cell viability was significantly decreased in Huh7-ATOH8 cells. The IC50 of 5-FU was decreased from 8.54 to 5.50, and the IC50 of CDDP was decreased from

RI PT

5.22 to 4.28 compared with Huh7-Vec cells (Figure 7A). To further explore whether ATOH8 could inhibit the chemo-resistance of HCC cells, a chemo-resistant xenograft model was established. Xenograft tumors were generated by subcutaneously injecting into

SC

nude mice a 1mm3 tumor cube that was induced by Huh7 cells. When the xenograft

M AN U

tumors reached approximately 5mm in diameter (approximately 1 week), the nude mice were divided into 6 groups (n=5-6 per group) and treated with an empty vector (LV-Vec), lentiviral-ATOH8 (LV-ATOH8), 5-FU+LV-Vec, CDDP+LV-Vec; 5-FU+ATOH8, or CDDP+ATOH8 (Figure 7B). Compared with the control group, the tumor growth rates and

TE D

tumor volumes of the ATOH8-, 5-FU- and CDDP-treated groups were significantly decreased. When mice were treated with the combination of ATOH8 and 5-FU or CDDP, the inhibitory effect on tumor growth was much stronger than that with any individual

EP

treatment (Figure 7C). Histological study confirmed that a significantly greater extent of

AC C

necrosis was observed in the combination treatment with ATOH8/chemotherapeutics, compared with any individual treatment (Figure 7D). IHC result showed that the expression of ATOH8 was induced by LV-ATOH8 (Supplementary Figure 2). These data demonstrated that virus-mediated ATOH8 overexpression dramatically inhibited the in vivo tumor growth and increased the chemo-sensitivity of HCC cells to 5-FU and CDDP.

18

ACCEPTED MANUSCRIPT Discussion ATOH8 belongs to bHLH family of proteins which essentially regulate cell differentiation.32

The hallmark of this family is a bHLH domain, which directly contacts

RI PT

DNA sequence harboring an E-box motif (CANNTG). Here we reported that ATOH8 was frequently down-regulated in primary HCCs, which was significantly associated with poor prognosis, poor differentiation and high serum AFP, indicating expression level of

SC

ATOH8 might reflect materialness of cancer cells. Our mechanistic studies demonstrated

M AN U

that the physical binding of ATOH8 to the E-box motif within NANOG and OCT4 gene could repress transcription of NANOG and OCT4, which are transcription factors required to maintain the pluripotency and self-renewal of ESCs. Moreover, ATOH8- cells were found to possess cancer stem cell properties and only present in an average 2.52% of HCC

TE D

cells in clinical HCC tissues, supporting the current opinion that CSCs consist a very small population within tumor tissues.33

EP

Interestingly, we found that overexpression of ATOH8 could decrease CD133+ cell

AC C

population while ATOH8 depletion could increase CD133+ cell population in HCC cell lines. A CD133+ cell population is thought to be a CSCs population in brain tumor,34 melanoma,35 and HCC tissues.7 Our next question is whether ATOH8- cells are CD133+ CSCs? Double IF staining confirmed that ATOH8- cells were CD133+ CSCs in both cell lines and clinical specimens. In addition, ATOH8 was absently expressed in iPSCs. Currently, CSCs in HCC are mainly defined by surface markers such as CD133,7 CD90,36 and CD24,37 although their effect on stemness has not been explored. Our finding 19

ACCEPTED MANUSCRIPT suggested that absent expression of ATOH8 might be a better marker to detect CSCs in HCC.

RI PT

To test whether CD133+ cells induced by ATOH8 depletion were CSCs, functional assays were applied to study CSC properties of these CD133+ cells. The results demonstrated that CD133+ cells induced by ATOH8 depletion possessed CSC properties,

SC

including stronger tumorigenicity and abilities of self-renewal, differentiation and

M AN U

chemo-resistance. This finding suggested that knockdown of ATOH8 might reprogram a non-CSC into a CSC. This finding suggested that knockdown of ATOH8 might reprogram a non-CSC into a CSC. In CSC study, whether a non-CSC can be reprogrammed into a CSC is a very important issue needs to be addressed. Several studies have indicated that

TE D

overexpression of certain stemness-associated gene, such as TGFβ1, Nanog and Oct4, can induce cancer cells to dedifferentiate into CSC-like cells. As a bHLH family member, ATOH1 is one of the essential genes required for secretory intestinal epithelial cells (IECs) The depletion of another family member Atoh7 also can initiate the

EP

differentiation.38

AC C

differentiation of the retina into neurons.39

More importantly, ATOH8 has also been

proved to be involved in the differentiation of ESCs to endothelial cells.40

In the present

study, we provide more solid evidences that ATOH8 depletion is necessary for CSCs maintenance whereas ATOH8 silencing stimulate HCC cells with non-CSCs to CSCs reprogramming.

To further confirm whether ATOH8 has effect on cell reprogram, shATOH8 was 20

ACCEPTED MANUSCRIPT co-transfected with SKOM into human fibroblasts to determine if ATOH8 depletion could increase the frequency of iPSCs generation. The results demonstrated that ATOH8 depletion dramatically enhanced the number of AP-positive colonies, which could be

RI PT

passaged as stable iPSCs induced by SKOM. The ESC-like phenotype and pluripotent ability of these AP-positive colonies was confirmed with EB assays and teratoma formation assays. These findings implied that a non-CSC could be reprogrammed into a

SC

CSC under some circumstances, such as down-regulation of ATOH8. This hypothesis has

M AN U

been also supported by a recent study, in which non-CSCs of human basal breast cancers are plastic cell populations that readily switch from a non-CSC to CSC state.41

The potential therapeutic value of ATOH8 in HCC treatment was also investigated by

TE D

testing whether introduction of ATOH8 could increase the chemosensitivity of HCC cells. Both in vitro and in vivo assays demonstrated that the combination of ATOH8 and 5-FU or CDDP significantly inhibited tumor growth in nude mice compared with only 5-FU or

EP

CDDP treatment, suggesting that ATOH8 expression in HCC has a great therapeutic

AC C

potential in HCC treatment, especially to increase the chemosensitivity of tumor cells to chemotherapy. In the present study, we noted that over 80% of HCC cases are HBV associated. It is unclear whether ATOH8 has the same effect on other HCCs without HBV infection. In summary, our findings indicate that ATOH8 plays a very important role in the regulation of cancer stemness and its down-regulation can reprogram non-CSCs into CSCs.

21

ACCEPTED MANUSCRIPT Figure legends Figure 1. Down-regulation of ATOH8 in HCC. (A) ATOH8 expression in HCC and matched non-tumor liver specimens from 242 HCC patients. The dot plots represent the

RI PT

ATOH8 expression in tumor and non-tumor samples from 242 HCCs. The P-value shown was calculated by paired Student’s t test. (B) Kaplan Meier analysis of the DFS rate of HCC patients in correlation with (-) and without ATOH8 down-regulation (+). (C)

SC

Representative images of ATOH8 expression (brown nuclear staining) in tumor tissue and

M AN U

adjacent non-tumor liver tissue detected by IHC (magnification, 400×). (D) The ATOH8 expression in 10 cell lines was evaluated by RT-PCR and western blot analysis. 18S rRNA and β-actin were used as loading controls, respectively. (E) ATOH8 was overexpressed by Lenti-virus system in PLC8024 and Huh7, and QSG7701 cells were transiently transfected

TE D

with siRNAs (si22 and si45) against ATOH8 or scramble siRNA as control, respectively. The expression of ATOH8 was detected by Western blot analysis, and β-actin was used as

AC C

Student’s t test).

EP

a loading control. (F) The cell growth rates were detected by XTT assay (**, P < 0.01,

Figure 2. The tumor suppressive effect of ATOH8 in HCC cells. Representative images of (A) foci formation and (B) colony formation in soft agar induced by PLC8024-Vec, PLC8024-ATOH8, Huh7-Vec, Huh7-ATOH8, QSG7701-si22, QSG7701-si45 and QSG7701-Ctl cells. The numbers of foci and colonies were calculated and are depicted in the bar chart. The values indicate the mean standard deviation of 3 independent experiments (**, P < 0.01, Student’s t test). (C) Matrigel invasion assay was performed to

22

ACCEPTED MANUSCRIPT compare cell invasion ability between ATOH8- and vector-transfected PLC8024 and Huh7 cells (up) / between Ctl- and siRNAs against ATOH8 in QSG7701 (down). Invaded cells were fixed and stained with crystal violet (magnification, 200×). The number of invaded

RI PT

cells was calculated and is depicted in bar chart. All data are shown as the mean ±SEM of 3 independent experiments (**, P < 0.001, Student’s t test). (D) A wound-healing assay showed that cell motility was suppressed by ATOH8 in the PLC8024 and Huh7 cells (up)

SC

and was increased after ATOH8 knockdown in QSG7701 (down). Microscopic images

M AN U

were acquired at 0 and 48, 60 or 72 hours after scratching the surface of a confluent layer of cells (magnification, 50×). (E) Representative images are shown for the tumors derived from SCID mice induced by PLC8024-Vec and PLC8024-ATOH8 cells (left), Huh7-Vec and Huh7-ATOH8 cells (left) , or stable sh-ATOH8 (sh4 or sh7) transfected QSG7701 cells

TE D

and control cells (right) after subcutaneous injection (5 mice per group). Growth curves of tumors derived from the indicated cell lines over 4 weeks are summarized in the bottom

EP

(**, P < 0.001, Student’s t test).

AC C

Figure 3. ATOH8 repressed transcription of stemness-associated genes. (A) The promoter of many stemness-associated genes contains ATOH8 binding domains. qRT-PCR was performed to compare expressions of stemness-associated genes between ATOH8- and vector-transfected cells (B) as well as between ATOH8-silenced and control QSG7701 cells (C). The relative expression level of tested genes (defined as fold change) was normalized to the endogenous 18S rRNA. (D) Western blot analysis was performed to investigate the effect of ATOH8 on the expression of stemness-associated genes at the 23

ACCEPTED MANUSCRIPT protein level in ATOH8-transfected cells and the ATOH8-silenced cells. β-actin was used as a loading control. (E) A luciferase reporter assay was performed to test the repressing effect of ATOH8. The pGL3-control vector was used as the positive control, and the

RI PT

pGL3-enhancer vector was used as the negative control. pRL-TK was used as an internal control. (F) EMSA was used to detect the interaction between ATOH8 and OCT4 or

M AN U

extract protein (NE) added with biotin-labeled probe.

SC

NANOG DNA probes/ mutant probe. Supershift band could be only detected in nuclear

Figure 4. Absent expression of ATOH8 in CSCs. (A) Representative IHC images of ATOH8 expression in tumor and their adjacent non-tumor tissues (magnification, 630×). Arrows denote the ATOH8-negative cells. Frequency of ATOH8- cells was compared

TE D

between HCC tumor tissues and their paired non-tumor liver tissues from 104 HCC patients. The P-value shown was calculated by paired Student’s t test. (B) Double IF staining of ATOH8 (green) and CD133 (red) was performed in cryopreserved tissue

EP

sections (magnification, 630×). DAPI (blue) was used for nuclei counterstain. Arrows

AC C

denote CD133+/ATOH8- cells. (C) qRT-PCR was performed to compare the expression level of ATOH8 between sorted CD133+ and CD133- cells in PLC8024, Huh7 and QSG7701 cell lines. 18S rRNA was used as an internal control. (D) Expressions of ATOH8 and CD133 in 6 representative HCC cases detected by Western blot analysis. GAPDH was used as a loading control. T, tumor tissue; N, adjacent non-tumor tissue. (E) Flow cytometry histogram analysis showed the introduction of ATOH8 could reduce CD133+ cells (red), compared to parental PLC8024 and Huh7 cells (green). Purple 24

ACCEPTED MANUSCRIPT represents the isotype control. (F) Flow cytometry histogram analysis showed the enrichment of CD133+ cells (red) in ATOH8 depleted cells, compared to parental

RI PT

QSG7701 and BEL7402 cells (green). Purple represents the isotype control.

Figure 5. CD133+ cells induced by ATOH8 depletion possessed properties of CSC. Quantification of foci formation (A) and colony formation in soft agar (B) induced by the

SC

sorted CD133+ cells generated by ATOH8 silencing in the QSG7701 cells. (C) Xenograft

M AN U

tumors derived from serial subcutaneous injections of sorted CD133+ induced by ATOH8 depletion or CD133- cells in the QSG7701 cell line. (D) In the sphere formation assay, the self-renewal ability was enhanced in sorted CD133+ cells induced by ATOH8 silencing in the QSG7701 cell line in three passages. (E) RT-PCR was performed to compare

TE D

expression of stemness-related genes between sorted CD133+ cells induced by ATOH8 silencing and CD133- cells in the QSG7701 cell line. Cell differentiation assay was performed by treating CD133+ and CD133- cells with ATRA, and the expression of

EP

differentiation-associated genes was analyzed by RT-PCR. 18S rRNA was used as a

AC C

loading control. (F) Sorted CD133+ cells induced by ATOH8 silencing and CD133- cells were treated with CDDP. M4 was counted as apoptotic cells. All data are shown as the mean ±SEM of 3 independent experiments (*, P < 0.05; **, P < 0.001, Student’s t tests).

Figure 6. ATOH8 enhanced efficiency of the iPSCs generation. (A) qRT-PCR analysis showed that the expression of ATOH8 was extremely low in 4 iPSCs, compared to their parental fibroblasts. (B) Representative of IF staining of ATOH8 expression (green color) 25

ACCEPTED MANUSCRIPT in A153 and SF002 cell lines, and their generated iPSCs (magnification, 630×). DAPI (blue) was used for nuclei counterstain. (C) Calculation of the increase in the iPSC generation efficiency due to ATOH8 silencing based on the number of AP-positive

RI PT

colonies per 10,000 infected cells seeded on feeders in a 10cm dish. (D) IF staining for the expression of the human embryonic stem cells (ESCs) markers. DAPI (blue) was used for nuclei counterstain. ES cell (H1, positive control), iPSC induced by SKOM (SF002-C-1),

SC

iPSCs induced by SKOM plus ATOH8 silencing (SF002-sh4-8 and SF002-sh7-2), and

M AN U

parental fibroblasts (SF002, negative control) were tested. (magnification, 400×) (E) qRT-PCR analysis of pluripotent genes and lineage-specific markers corresponding to the 3 germ layers at Day 8 after attaching the EBs for another 8 days on gelatin-coated dishes. ESC markers (1, NANOG; 2, OCT4), endoderm lineage markers (3, AFP; 4, GATA4; 5,

TE D

SOX17), endoderm lineage marker (6, TBX1) and mesoderm lineage markers (7, PAX6; 8, SOX1) were tested. (F) Teratomas were produced by the indicated iPSCs clones and fixed and embedded in paraffin (magnification, 400×). The teratomas were sectioned and

EP

stained with hematoxylin/eosin for analysis. All data are shown as the mean ±SEM of 3

AC C

independent experiments (*, P < 0.05; **, P < 0.001, Student’s t test).

Figure 7. ATOH8 increased the chemosensitivity of HCC cells. (A) XTT assay was used to compare the cell viabilities between ATOH8- and vector-transfected cells after treatment with 5-FU and CDDP, respectively. The data are shown as the mean ±SEM of 3 independent experiments (**, P < 0.01, student t test). (B) Schematic representation of xenograft tumorigenesis and treatment with ATOH8, 5-FU and CDDP in nude mice. A 1 26

ACCEPTED MANUSCRIPT mm3 xenograft piece was subcutaneously implanted in a 3-4 weeks old male nude mouse individually. When the tumor reached approximately 5mm in diameter, the nude mice were divided into 6 groups (5-6 mice per group). (C) Representative images of xenograft

RI PT

tumors in nude mice after different treatment. The tumor growth curves of each group of mice are summarized (*, P < 0.05; **, P < 0.001, student t test). (D) Xenograft were fixed and embedded in paraffin and stained with hematoxylin/eosin for analysis tumor necrosis

AC C

EP

TE D

M AN U

SC

induced by ATOH8/chemotherapeutics (magnification, 200×).

27

ACCEPTED MANUSCRIPT Reference 1.

Thorgeirsson SS, Grisham JW. Molecular pathogenesis of human hepatocellular carcinoma. Nat Genet 2002;31:339-46. Badvie S. Hepatocellular carcinoma. Postgrad Med J 2000;76:4-11.

3.

Suetsugi A, Nagaki M, Aoki H, et al. Characterization of CD133(+) hepatocellular

RI PT

2.

carcinoma cells as cancer stem/progenitor cells. Biochemical and Biophysical Research Communications 2006;351:820-824. 4.

Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy

5.

SC

that originates from a primitive hematopoietic cell. Nat Med 1997;3:730-7.

Al-Hajj M, Wicha MS, Benito-Hernandez A, et al. Prospective identification of

6.

M AN U

tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 2003;100:3983-8. Hemmati HD, Nakano I, Lazareff JA, et al. Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci U S A 2003;100:15178-83. 7.

Ma S, Chan KW, Hu L, et al. Identification and characterization of tumorigenic liver cancer stem/progenitor cells. Gastroenterology 2007;132:2542-2556. Visvader JE, Lindeman GJ. Cancer stem cells in solid tumours: accumulating

TE D

8.

evidence and unresolved questions. Nature reviews. Cancer 2008;8:755-68. 9.

Pardal R, Clarke MF, Morrison SJ. Applying the principles of stem-cell biology to

10.

EP

cancer. Nature reviews. Cancer 2003;3:895-902. Ma S, Lee TK, Zheng BJ, et al. CD133+ HCC cancer stem cells confer chemoresistance by preferential expression of the Akt/PKB survival pathway.

11.

AC C

Oncogene 2008;27:1749-58.

Dalerba P, Clarke MF. Cancer stem cells and tumor metastasis: First steps into uncharted territory. Cell Stem Cell 2007;1:241-242.

12.

Ma S, Tang KH, Chan YP, et al. miR-130b Promotes CD133(+) liver tumor-initiating cell growth and self-renewal via tumor protein 53-induced nuclear protein 1. Cell Stem Cell 2010;7:694-707.

13.

Takahashi K, Tanabe K, Ohnuki M, et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 2007;131:861-872.

14.

Moon JH, Kwon S, Jun EK, et al. Nanog-induced dedifferentiation of p53-deficient 28

ACCEPTED MANUSCRIPT mouse astrocytes into brain cancer stem-like cells. Biochemical and Biophysical Research Communications 2011;412:175-181. 15.

Kumar SM, Liu S, Lu H, et al. Acquired cancer stem cell phenotypes through Oct4-mediated dedifferentiation. Oncogene 2012;31:4898-4911. Zhang HX, Wu HT, Zheng JH, et al. Transforming Growth Factor beta 1 Signal is

RI PT

16.

Crucial for Dedifferentiation of Cancer Cells to Cancer Stem Cells in Osteosarcoma. Stem Cells 2013;31:433-446. 17.

Chaudhary J, Skinner MK. Basic helix-loop-helix proteins can act at the E-box

SC

within the serum response element of the c-fos promoter to influence hormone-induced promoter activation in Sertoli cells. Molecular Endocrinology

18.

M AN U

1999;13:774-786.

Ejarque M, Altirriba J, Gomis R, et al. Characterization of the transcriptional activity of the basic helix-loop-helix (bHLH) transcription factor Atoh8. Biochim Biophys Acta 2013;1829:1175-1183.

19.

Patel N, Varghese J, Masaratana P, et al. The transcription factor ATOH8 is

TE D

regulated by erythropoietic activity and regulates HAMP transcription and cellular pSMAD1,5,8 levels. Br J Haematol 2013. 20.

Fischer A, Leimeister C, Winkler C, et al. Hey bHLH factors in cardiovascular

21.

EP

development. Cold Spring Harb Symp Quant Biol 2002;67:63-70. Porcher C, Liao EC, Fujiwara Y, et al. Specification of hematopoietic and vascular development by the bHLH transcription factor SCL without direct DNA binding.

AC C

Development 1999;126:4603-4615.

22.

Balakrishnan-Renuka A, Morosan-Puopolo G, Yusuf F, et al. ATOH8, a regulator of skeletal myogenesis in the hypaxial myotome of the trunk. Histochem Cell Biol 2014;141:289-300.

23.

Lee SK, Pfaff SL. Synchronization of neurogenesis and motor neuron specification by direct coupling of bHLH and homeodomain transcription factors. Neuron 2003;38:731-745.

24.

Chung S, Chavez C, Andrew DJ. Trachealess (Trh) regulates all tracheal genes during Drosophila embryogenesis. Developmental Biology 2011;360:160-172. 29

ACCEPTED MANUSCRIPT 25.

Chen L, Li Y, Lin CH, et al. Recoding RNA editing of AZIN1 predisposes to hepatocellular carcinoma. Nat Med 2013;19:209-16.

26.

Quandt K, Frech K, Karas H, et al. MatInd and MatInspector: New fast and versatile tools for detection of consensus matches in nucleotide sequence data.

27.

RI PT

Nucleic Acids Research 1995;23:4878-4884. Cartharius K, Frech K, Grote K, et al. Matlnspector and beyond: promoter analysis based on transcription factor binding sites. Bioinformatics 2005;21:2933-2942. 28.

Kuhlmann WD, Peschke P. Hepatic progenitor cells, stem cells, and AFP

29.

SC

expression in models of liver injury. Int J Exp Pathol 2006;87:343-59.

Chai S, Tong M, Ng KY, et al. Regulatory role of miR-142-3p on the functional

30.

M AN U

hepatic cancer stem cell marker CD133. Oncotarget 2014;5:5725-35. Zhang Z, Gong Y, Guo Y, et al. Direct transdifferentiation of spermatogonial stem cells to morphological, phenotypic and functional hepatocyte-like cells via the ERK1/2 and Smad2/3 signaling pathways and the inactivation of cyclin A, cyclin B and cyclin E. Cell Commun Signal 2013;11:67.

Cereghini S, Blumenfeld M, Yaniv M. A liver-specific factor essential for albumin

TE D

31.

transcription differs between differentiated and dedifferentiated rat hepatoma cells. Genes Dev 1988;2:957-74.

Yin C, Evason KJ, Maher JJ, et al. The basic helix-loop-helix transcription factor,

EP

32.

heart and neural crest derivatives expressed transcript 2, marks hepatic stellate cells in zebrafish: analysis of stellate cell entry into the developing liver.

33.

AC C

Hepatology 2012;56:1958-70.

Bao B, Ahmad A, Azmi AS, et al. Overview of cancer stem cells (CSCs) and mechanisms of their regulation: implications for cancer therapy. Curr Protoc Pharmacol 2013;Chapter 14:Unit 14 25.

34.

Singh SK, Clarke ID, Terasaki M, et al. Identification of a cancer stem cell in human brain tumors. Cancer Res 2003;63:5821-8.

35.

Welte Y, Davies C, Schafer R, et al. Patient derived cell culture and isolation of CD133(+) putative cancer stem cells from melanoma. J Vis Exp 2013:e50200.

36.

Yang ZF, Ho DW, Ng MN, et al. Significance of CD90+ cancer stem cells in 30

ACCEPTED MANUSCRIPT human liver cancer. Cancer Cell 2008;13:153-66. 37.

Lee TKW, Castilho A, Cheung VCH, et al. CD24(+) Liver Tumor-Initiating Cells Drive Self-Renewal and Tumor Initiation through STAT3-Mediated NANOG Regulation. Cell Stem Cell 2011;9:50-63. Gerbe F, van Es JH, Makrini L, et al. Distinct ATOH1 and Neurog3 requirements

RI PT

38.

define tuft cells as a new secretory cell type in the intestinal epithelium. J Cell Biol 2011;192:767-80. 39.

Kiyama T, Li H, Gupta M, et al. Distinct neurogenic potential in the retinal margin

Fang F, Wasserman SM, Torres-Vazquez J, et al. The role of Hath6, a newly identified

shear-stress-responsive

transcription

factor,

M AN U

40.

SC

and the pars plana of mammalian eye. J Neurosci 2012;32:12797-807.

in

endothelial

cell

differentiation and function. J Cell Sci 2014;127:1428-40. 41.

Chaffer CL, Marjanovic ND, Lee T, et al. Poised Chromatin at the ZEB1 Promoter Enables Breast Cancer Cell Plasticity and Enhances Tumorigenicity. Cell 2013;154:61-74.

AC C

EP

TE D

Author names in bold designate shared co-first authors

31

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Supplementary Materials and Methods RNA extraction and qRT-PCR. Total RNA was extracted with TRIzol reagent (Invitrogen) and 1µg total RNA was

RI PT

used for cDNA synthesize using an RT-PCR kit (Roche Applied Science, Indianapolis, IN, USA). qRT-PCR was performed using an ABI Prism 7900 System (Applied Biosystems, Carlsbad, CA, USA) with the SYBR Green PCR master mix (Applied

SC

Biosystems) and corresponding primers (Supplemental Table 1). Data was processed

M AN U

using the ABI SDS v2.3 software (Applied Biosystems). For HCC cell lines and clinical samples (clinical HCC and matched non-tumor samples), the relative -∆CT

expression level (defined as “fold change”) of the target gene is given by 2 target

(∆CT=∆CT

18S

-∆CT

) and normalized to the fold change detected in the

TE D

corresponding control cells/non-tumor tissue, which was defined as 1.0.

Establishment of ATOH8 Overexpression and Knockdown Cells.

EP

Full-length human ATOH8 was amplified by PCR (primers are listed in

AC C

Supplemental Table 3). The purified PCR product was cloned into the pLenti6/V5-TOPO vector (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The 293FT virus producing cell line was co-transfected with the ATOH8 construct and optimized with a packaging mix (Invitrogen) using Lipofectamine 2000 (Invitrogen). The viral supernatant was collected and stably transfected into PLC8024 and Huh7 cells selected by 6µM Blasticidin (Invitrogen). Empty vector transfected cells were used as controls. 1

ACCEPTED MANUSCRIPT Based on the ATOH8 sequence (NM_032827.6), two siRNAs were designed using siRNA Target Finder (Ambion, Austin, TX, USA): si22, 5′-GGUGCCGUGCUACU CAUAUTT-3′; si45, 5′-AGUUCCUACUCGUCAAUUUTT-3′. The ATOH8-specific

RI PT

shRNA expression vectors and the scrambled ineffective shRNA cassette in the psiHIV-H1 plasmid were purchased from GeneCopoeia (Rockville, MD). The sequences

of

the

2

shRNAs

against

ATOH8

are

as

follows:

sh4,

SC

AGTTCCTACTCGTCAATTTTT, and sh7, GGTGCCGTGCTACTCATAT. For

M AN U

ATOH8 knockdown experiment, siRNAs were transiently transfected into QSG7701usingLipofectamine 2000 (Invitrogen). For stable cell line, we first co-transfected the 293FT (Invitrogen) virus producing cell line with psiHIV– shATOH8 constructs or the scrambled shRNA construct and an optimized packaging

TE D

mix (GeneCopoeia) using Lipofectamine 2000.The viral supernatant was transfected into QSG7701 cells and stable clones were selected by 3µg/mL puromycin

EP

(Sigma-Aldrich).

AC C

In Vitro Functional Assays XTT proliferation assay, foci formation assay, colony formation in soft agar,

invasion assay, and wound healing assay were used to test the in vitro functional roles of ATOH8. For XTT cell proliferation assay, 1,000 cells were seeded per well in 96-well plates and grown for 5-6 days. A 50µL volume of the XTT working solution provided by the kit (Roche Applied Science) was added to the culture wells. The cells were incubated at 37°C in the dark for 4hr, and the absorbency was measured with a 2

ACCEPTED MANUSCRIPT scanning multi-well spectrophotometer (Tecan Sunrise, Tecan Trading AG, Switzerland) at a test wavelength of 492 nm and a reference wavelength of 630nm.Triplicate independent experiments were performed. For the foci formation

RI PT

assay, 1×103 cells per well were seeded in 6-well plates. After culture for 7-14 days, colonies consisting of more than 50 cells were counted and stained using crystal violet (Sigma-Aldrich). Triplicate independent experiments were performed. For colony

SC

formation in soft agar, 5×103 cells in 0.4% Bacto agar (Sigma-Aldrich) were seeded

M AN U

on top of a solidified layer of 0.6% Bacto agar in 6-well plates. Triplicate independent experiments were performed. The surviving colonies (>50 cells per colony) were counted after 3-4 weeks, and the data are expressed as the mean ±SD of triplicate wells within the same experiment. Invasion assay was performed using 24-well

TE D

BioCoat Matrigel Invasion Chambers (BD Biosciences, San Jose, CA, USA) according to the manufacturer’s instructions. Briefly, 1×105 cells in DMEM without FBS were added to the top chamber, and 10% FBS in DMEM was added to the

EP

bottom chamber as an attractant. After 24 hours of incubation, the invaded cells were

AC C

fixed and stained with crystal violet. The number of cells was counted in 10 fields under a 20× objective lens and imaged using SPOT imaging software (Nikon, Japan). Cell migration was assessed by wound-healing assay measuring the movement of cells into a scraped a cellular area created with a 200µL pipette tip. Wound closure was observed after 48 and 60 hours and photographed under a 5× objective lens (Nikon).

3

ACCEPTED MANUSCRIPT Tumor Formation in SCID Mice The in vivo tumor formation was performed by subcutaneously injecting 4×106 of ATOH8-overexpressed and empty vector cells into the right or left dorsal flank of

RI PT

3-4weeks old male SCID mice, respectively. Injecting 4×106 of QSG7701-shATOH8 cells and scramble QSG7701-Ctl cells subcutaneously into the right or left dorsal flank of 4-week-old male SCID mice, respectively, carried out a similar experiment.

SC

Tumor formation in SCID mice was monitored over 4 weeks. To test

M AN U

tumorigenicity of CD133+ cells induced by ATOH8 depletion, 1×105 or 1×104 of sorted CD133+ cells, and 4×106, 1×106, 1×105 or 1×104 CD133- cells were injected into the right or left dorsal flank of 3-4weeks old male SCID mice, respectively. Tumor formation in the mice was monitored over 4 weeks. The tumor volume was

TE D

measured weekly and calculated using the formula V=0.5×W2×L (V, volume; L, Length; W, Width). The mice were euthanized when the tumor volume exceeded 1cm3, and the tumors were excised and embedded in paraffin. All animal experiments were

EP

approved by and performed in accordance with the Committee of the Use of Live

AC C

Animals in Teaching and Research at the University of Hong Kong.

Western Blot Analysis and Antibodies Western blot analyses were performed according to the standard protocol with

antibodies for ATOH8 and β-actin from Abcam (Cambridge, MA, USA), OCT4, SOX2 and GAPDH from Santa Cruz Biotechnology (Santa Cruz, CA, USA), NANOG from R&D System (Minneapolis, MN, USA), AFP from Sigma-Aldrich and 4

ACCEPTED MANUSCRIPT CD133 from MiltenyiBiotec (BergischGladbach, Germany).

IHC and TMA

RI PT

The paraffin-embedded tissue blocks were sectioned for IHC staining. In brief, sections/TMA were deparaffinized and rehydrated by immersing the slides in xylene for 10 minutes 3 times and in 100% ethanol for 5 minutes 3 times, followed by

SC

incubation in 95%, 85%, 75% and 50% ethanol for 5 minutes each at room

M AN U

temperature. The endogenous peroxidase activity was blocked with 3% hydrogen peroxide (H2O2) for 30minutes. For antigen retrieval, the slides were immersed in 10mM citrate buffer (pH 6.0) and boiled for 15minutes in a microwave oven. Nonspecific binding was blocked with 5% normal goat serum for 30minutes.The

TE D

sections were incubated in a 1:100 dilution of a monoclonalATOH8 antibody (Abcam, Cambridge, UK) at 4°C overnight in a humidified chamber, and then sequentially incubated with biotinylated goat antibody to mouse IgG (1:100, Santa Cruz, Dallas,

EP

Texas) for 30minutes and streptavidin-peroxidase conjugate for 30minutes at room

AC C

temperature. A 3,5-diaminobenzidine (DAB) Substrate Kit (Dako, Carpinteria, CA, USA) was used for color development followed by Mayer’s hematoxylin counterstaining.

IF Cell slides/cryopreserved tissue sections were used to do IF. For cell slides, cells were plated on glass slide at 50% confluence, and the slides were washed with 5

ACCEPTED MANUSCRIPT ice-cold PBS and then fixed with 4% paraformaldehyde for 30 minutes at room temperature. Then, the cell slides/cryopreserved tissue sections were permeabilized in PBS containing 0.1% Triton X-100 (Sigma-Aldrich) for 30 minutes at 4°C, followed

RI PT

by blocking with 5% BSA/PBS for 30 minutes at room temperature. The slides were incubated with antibodies against ATOH8 (1:100, Abcam), CD133 (1:10, MiltenyiBiotec), OCT4 (1:100, Santa Cruz), SSEA-4 (1:100, Invitrogen), TRA-1-60

SC

(1:100, Millipore, Billerica, MA, USA) or TRA-1-81 (1:100, Millipore) overnight at

M AN U

4°C.The corresponding secondary antibodies (Invitrogen) were applied at 1:400, followed by incubation at room temperature for 1hour. Finally, the cells were washed and mounted with mounting medium containing DAPI (Invitrogen).

TE D

Dual-Luciferase Reporter Assay

Approximately 2kbof the promoter regions of OCT4, NANOG and CD133 genes were PCR amplified (primers are listed in Supplemental Table 3) and cloned into the

EP

pGL3-enhancer vector (Promega, Madison, WI), respectively. The luciferase reporter

AC C

constructs [10:10:1 mixture of OCT4/NANOG/CD133 luciferase constructs, pLenti6-ATOH8 and the Renilla luciferase plasmid (pRL-TK) (Promega)] were transfected into cells using Lipofectamine2000 (Invitrogen) according to the manufacturer’s instructions.pGL3-controland pGL3-enhancer vectors (Promega) were used as the positive and negative controls, respectively. pRL-TK was used as an internal control. After 48 hours, the Dual-Luciferase assay kit (Promega) was used to measure the firefly and Renilla luciferase activities according to the manufacturer’s 6

ACCEPTED MANUSCRIPT instructions. The results were normalized to the Renilla activities.

EMSA

RI PT

DNA fragments containing HLH binding motifs from promoter regions of OCT4 and SOX2 that end-labeled with biotin were synthetized by Invitrogen. The probe sequences and mutated probe sequences were described as below. OCT4-seq:

AGGCTCTGCACATCCAAGCCGTCTGGAATCACTCCCACACC

M AN U

OCT4-mutant:

SC

AGGCTCTGCACATCCCAGCTGTCTGGAATCACTCCCACACCTCCATGTTC;

TCCATGTTC; NANOG-seq: GGTCTTGAACTCCTGATCTCAGATGATGCACC TGGCTCGGCCTCGCAAAGTGCTG;

NANOG-mutant:

GGTCTTGAACTC

CTGATCTAAGACGATGAACCCGGCTCGGCCTCGCAAAGTGCTG.

Nuclear

TE D

extract protein was prepared by the NucBuster Protein Extraction kit (Novagen, Madison, WI, USA). The EMSA was applied with 10µg of nuclear extracts and 50ng of biotin-labeled or unlabeled probes by the Band shift kit (Amersham Pharmacia

EP

Biotech, Piscataway, NJ, USA) and rabbit anti-ATOH8 antibody (1:100, Abcam). The

AC C

reaction for binding was running on the non-denaturing polyacrylamidegel and transferred tonylon membrane (Amersham Pharmacia Biotech). Protein-DNA complexes were cross-linked for 3minutes by UV crosslinker oven.

Spheroid Formation Assay A total of 500 single HCC cells were plated onto 24-well polyHEMA-coated plates (Sigma-Aldrich). The cells were cultured for 3-4 weeks in DMEM/F12 medium 7

ACCEPTED MANUSCRIPT (Invitrogen) supplemented with 4µg/mL insulin (Sigma-Aldrich), B27 (1:50, GIBCO, Grand Island, NY, USA), 20ng/mL EGF (Sigma-Aldrich) and 20ng/mL basic FGF (Sigma-Aldrich). For serial passage of primary spheres, the primary spheres were

RI PT

collected, subsequently dissociated with trypsin and resuspended in DMEM/F12 medium with the above supplements. The surviving colonies (>50 cells per colony) were counted after 4-6 weeks, and the data are expressed as the mean±SD of triplicate

M AN U

Differentiation of Cancer Cells

SC

wells within the same experiment.

Flow-sorted CD133+ and CD133- cells were grown in medium containing ATRA (Sigma-Aldrich) 1. Briefly, 2 ml 1 × 106/m1 Flow-sorted CD133+ and CD133- 7701-sh

TE D

ATOH8 cells in 6-well plates were incubated with 4µM ATRA for 72 hours 2. ATRA was suspended in DMSO to make a stock ATRA concentration of 10 mM, which was then diluted into media to the desired concentration. For control experiments, DMSO

EP

was diluted into standard media at similar concentrations. All experiments were

AC C

handled with minimal exposure to ambient light or air. Then, RNA was extracted to detect the expression of differentiation-associated genes by RT-PCR with corresponding primers (Supplemental Table 1).

Chemoresistance Assay Chemotherapy-induced cytotoxicity was determined by XTT cell proliferation assay (Roche Applied Science). Briefly, 1×104 cells/well were seeded in 96-well 8

ACCEPTED MANUSCRIPT plates, allowed to attach overnight, and then chemotherapeutic agents 5-FU and CDDP were added in complete culture medium, at various concentrations. XTT assay was performed according to manufacturer’s instructions. The relative number of

RI PT

viable cells as compared to the number of cells without drug treatment was expressed as percent cell viability. All results in the study were based on at least three parallel measurements each time and each measurement was repeated in up to two

M AN U

SC

independent experiments.

Mouse Xenograft Tumor Treatment Model

First, a xenograft tumor was induced by subcutaneously injecting 4×106 of Huh7 cells in a 3-4weeks old male nude mouse. When the tumor volume reached >1cm3

TE D

(about 4 weeks), the tumor was removed and cut into ~1mm3 pieces and subcutaneously implanted into 3-4weeks old male nude mice. When tumors reached approximately 5mm in diameter, the nude mice were divided into 6 groups (5-6 mice

EP

per group). Group 1 received an intratumoral injection of LV-Vec at a dose of 2×107

AC C

MOI in 20µLof PBS once per week for 4 weeks. Group 2 received an intratumoral injection of LV-ATOH8 at a dose of 2×107 MOI in 20µL of PBS once per week for 4 weeks. Group 3 received an intratumoral injection of LV-Vec once per week for 4 weeks, plus intraperitoneal injection of CDDP (5mg/kg body weight) twice per week for 4 weeks. Group 4 received an intratumoral injection of LV-ATOH8 once per week for 4 weeks, plus an intraperitoneal injection of CDDP twice per week for 4 weeks. Group 5 received an intratumoral injection of LV-Vec once per week for 4 weeks, plus 9

ACCEPTED MANUSCRIPT an intraperitoneal injection of 5-FU (40mg/kg body weight) twice per week for 3 weeks. Group 6 received an intratumoral injection of LV-ATOH8 once per week for 4 weeks, plus an intraperitoneal injection of 5-FU twice per week for 3 weeks. The

RI PT

tumor volume was measured weekly and calculated using the formula V=0.5×W2×L (V, volume; L, Length; W, Width). All animal experiments were approved by and

M AN U

and Research at the University of Hong Kong.

SC

performed in accordance with the Committee of the Use of Live Animals in Teaching

Generation of iPSC

Cells were transduced with a cocktail composed of a psin-Lentiviral vector encoding SKOM together with the shRNA of ATOH8.At Day 2 post-infection, the

TE D

medium was changed to DMEM/F12 (Invitrogen) with 20% dFBS (Hyclone, Omaha Nebraska) and Vc (sodium L-ascorbate, 50mg/mL; Sigma-Aldrich). At Day 6, the infected cells were seeded on feeders (mitotically inactivated-murine embryonic

EP

fibroblasts) and cultured in standard human ESC medium consisting DMEM/F12 plus

AC C

20% knockout serum replacement (KSR, GIBCO), 50IU/mL penicillin/streptomycin (Hyclone), 1mM glutamine (Invitrogen), 0.1mM non-essential amino acids (Invitrogen),

0.1mM

beta-mercaptoethanol

(Invitrogen)

and

8ng/mL bFGF

(Sigma-Aldrich). VPA (1mM; Merck, Darmstadt, Germany) was added from Day 7– 20. Human ESC-like colonies appeared around Day 25-28 post-infection and were picked manually around Day 30.Picked iPSC colonies were routinely maintained on feeders in human ESC medium or on Matrigel (BD Biosciences) in mTeSR1 10

ACCEPTED MANUSCRIPT (Stemcell Technologies, Vancouver, BC, Canada) on 24-well plates.

Characterization of iPSC.

RI PT

iPSCs were characterized by AP staining, IF, EB assays and teratoma formation assays. For AP staining assay, the medium was gently aspirated, and the cells were rinsed with 0.25mL of PBS. For AP Staining, iPSCs were fixed in 4%

SC

paraformaldehyde (Sigma-Aldrich), and the plates were washed twice with 0.5mL of

M AN U

TBST (Tris-buffered saline plus 0.05% Tween-20, Sigma-Aldrich). Fresh AP staining solution was prepared with 4.5µL nitro blue tetrazolium (50mg/mL, Sigma-Aldrich), 3.5µL 5-bromo-4-chloro-3-indolyl phosphate (50mg/mL, Sigma-Aldrich) in 100mM Tris-HCl (Sigma-Aldrich), 100mM NaCl (Sigma-Aldrich) and 50mM MgCl2

TE D

(Sigma-Aldrich), and the plates were incubated with fresh AP staining solution in the dark for 15minutes at room temperature. The solution was aspirated, and the plates were washed with PBS. IF was detected using a LEICA (Wetzlar, Germany)

EP

DMI6000B microscope. The cells were plated on glass slide at 50% confluence, and

AC C

treated as described above. For EB assay, iPSCs/ESCs were detached with Dispase (Invitrogen), and the suspended cells were plated onto low-adherence dishes for 8 days in human ESC medium without bFGF (Sigma-Aldrich). The EBs were then plated onto gelatin-coated dishes in the same medium for another 8 days before the RNA was extracted. To quantify gene expression in the EB assays, equal amounts of cDNA were synthesized using an RT-PCR kit (Roche Applied Science) and used for qRT-PCR analysis with corresponding primers of three germ lines (Supplementary 11

ACCEPTED MANUSCRIPT Table 1). For teratoma formation assay, iPSCs were subcutaneously injected into 4-week-old male nude mice. The tumors were excised after 6-8 weeks, fixed and embedded in paraffin. Tumor sections were then stained with hematoxylin/eosin for

RI PT

histology analysis.

Statistical Analyses

SC

The SPSS statistical package for Windows (v20, SPSS) and Microsoft Excel

M AN U

(Microsoft Office 2010) were used for data analysis. Chi-square test was used for clinicopathological correlation study. The amount of ATOH8 silencing in the tumors and the matched non-tumor tissues was compared by paired student’s t test. Kaplan-Meier plots and log-rank tests were used for the disease-free survival analysis.

TE D

The unpaired two tailed Student’s t test was used to compare the difference between tested and control cells in number of foci, colonies formed in soft agar, migratory and invasive cells, the tumor volume and the relative expression levels of the target genes.

AC C

EP

A P value of < 0.05 was considered to be statistically significant.

Supplementary References

1.

Tallman MS, Andersen JW, Schiffer CA, et al. All-trans-retinoic acid in acute promyelocytic leukemia. N Engl J Med 1997;337:1021-8.

2.

Park SH, Lim JS, Jang KL. All-trans retinoic acid induces cellular senescence via upregulation of p16, p21, and p27. Cancer Lett 2011;310:232-9.

12

ACCEPTED MANUSCRIPT Supplementary Table 1. Correlation between the Expression of ATOH8 with Clinicopathological Features in 242 HCCs.

18 (52.9%) 106 (51%)

199 42

96 (48.2%) 22 (52.4%)

103 (51.8%) 20 (47.6%)

0.375

38 191

16 (42.1%) 94 (49.2%)

22 (57.9%) 97 (50.8%)

0.267

133 94

57 (42.9%) 53 (56.4%)

SC

RI PT

0.489

76 (57.1%) 41 (43.6%)

0.03*

39 (46.4%) 74 (49.7%)

45 (53.6%) 75 (50.3%)

0.368

13 (61.9%) 98 (47.1%)

8 (38.1%) 110 (52.9%)

0.144

30 (50%) 82 (48%)

30 (50%) 89 (52%)

0.451

51 (47.7%) 12 (46.2%)

56 (52.3%) 14 (53.8%)

0.533

180 60

85 (47.2%) 32 (53.3%)

95 (52.8%) 28 (46.7%)

0.251

145 72

63 (43.4%) 44 (61.1%)

82 (56.6%) 28 (38.9%)

0.01*

130 101

65 (50%) 48 (47.5%) 41.4 (31.7-48.3)

65 (50%) 53 (52.5%) 52.6 (45.1-60.0)

21 208

TE D

60 171 107 26

0.405 0.031*

Partial data is not available, and statistic was based on available data.

b

The cases showing more than 4 folds in non-tumor than their matched tumor samples are classified as

“Down-regulated” Group. c

P-value

16 (47.1%) 102 (49%)

84 149

AC C

a

ATOH8 Expression Down-regulated No change

34 208

EP

Gender Female Male Age ≤60 >60 Serum HBsAga Negative Positive Serum AFP (ng/ml) a ≤400 >400 Tumor size (cm) a ≤5 >5 Cirrhosisa Absent Present Tumor encapsulation Absent Present Thrombus Absent Present Vascular. Invasiona Absent Present Cell differentiationa Well differentiated Poorly differentiated Recurrencea Absent Present Mean DFSc (months)

Number

M AN U

Clinical features

Disease-free survival time

* Statistical significance (P<0.05) is shown in bold italic.

ACCEPTED MANUSCRIPT Supplementary Table 2. Engraftment Rates of CD133+ and CD133- Cells Sorted from ATOH8 Knockdown HCC Cell Line QSG7701 in SCID Mice.

4,000,000 1,000,000 100,000 10,000 4,000,000 1,000,000 100,000 10,000

4/5 2/5 2/5 1/5 0/5 0/5

70 90 90 120 -

AC C

EP

TE D

M AN U

QSG7701-CD133-

Latency (days)c

RI PT

QSG7701-CD133+

Tumor incidenceb

SC

Cell numbers injected

Cell typea

ACCEPTED MANUSCRIPT Supplementary Table 3: Primers Sequence for qRT-PCR

AFP CD133 ALB GATA4 SOX17 TBX1 PAX6 SOX1 CK8

AC C

CK18

SC

NANOG

M AN U

SOX2

TE D

OCT4

Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse Forward Reverse

EP

ATOH8

18S

Sequence (5'-3') CAGGTGCCGTGCTACTCATA AGTCACTCCTTGCGCTTCTT CGACCATCTGCCGCTTTGAG CCCCCTGTCCCCCATTCCTA TGGACAGTTACGCGCACAT CGAGTAGGACATGCTGTAGGT AAGGTCCCGGTCAAGAAACAG CTTCTGCGTCACACCATTGC AGACTGAAAACCCTCTTGAATGC GTCCTCACTGAGTTGGCAACA TGGATGCAGAACTTGACAACGT ATACCTGCTACGACAGTCGTGGT GCTCAGTATCTTCAGCAGTGTCC AGGTTTGGGTTGTCATCTTTGT CAGAAAACGGAAGCCCAA TTGCTGGAGTTGCTGGAAG ACGGAATTTGAACAGTAT CAGGATAGTTGCAGTAAT AGCGAGAAATATGCCGAGG TTCGCGAAGGGATTGCT TTGCTTGGGAAATCCGAG TGCCCGTTCAACATCCTT TTTCCCCTCGCTTTCTCA TGCAGGCTGAATTCGGTT CAGAAGTCCTACAAGGTGTCCA CTCTGGTTGACCGTAACTGCG TCGCAAATACTGTGGACAATGC TTGGCGAGGTCCTGAGATTTG CTCTTAGCTGAGTGTCCCGC CTGATCGTCTTCGAACCTCC

RI PT

Name

ACCEPTED MANUSCRIPT Supplementary Table 4: Primers for LOH Sequence (5'→3')

AT-LOH1-F

CGCCATTCTCCTGCCTCA

AT-LOH1-R

AACACTGCGGTCTCCCCAT

AT-LOH2-F

TTAGTGGGAAAGGTGTTGG

AT-LOH2-R

ATCTGTAGCCTCAGGTAAAGTC

AT-LOH3-F

CTGAAATCACCTCAAATCTAAC

AT-LOH3-R

CCAAAAAAAGGGAAGGAC

AC C

EP

TE D

M AN U

SC

RI PT

Name

ACCEPTED MANUSCRIPT Supplementary Table 5: Primers for cloning Sequence (5'-3')

psiHIV-H1-F

CCGACAACCACTACCTGA

psiHIV-H1-R

CGTGAAGAATGTGCGAGAC

pLenti6-CMV Forward

CGCAAATGGGCGGTAGGCGTG

pLenti6-V5 Reverse

ACCGAGGAGAGGGTTAGGGAT

pLenti6-ATOH8-F

CACCGCCATGAAGCACATCCCG

pLenti6-ATOH8-R

AGTGACTCCTTGCGCTTCTTG

CD133-Pgl3-Sma1-F

TTCCCCCGGGCCCCGCAGAGTCCCTTAC

CD133-Pgl3-Xho1-R

GCCGCTCGAGGCATTGGCAAATCAAAACTGT

Oct4-Pgl3-Sma1-F

AACCCCCGGGCGAGGATCAACCCAGCCC

Oct-Pgl3-Xho1-R

GCCGCTCGAGAGTATCGGGATGGGAATGCC

TE D

M AN U

SC

RI PT

Name

AACCCCCGGGTCTCCTGTCTCAGCCTCC

Nanog-Pgl3-Xho1-R

GCCGCTCGAGTCCTTCCTATTCCCAAAC

AC C

EP

Nanog-Pgl3-Sma1-F

ACCEPTED MANUSCRIPT Supplementary Figure Legends Supplementary Figure 1. RNA-Seq analysis of 3 pairs of HCC samples. (A) Venn diagram of down-regulated genes and up-regulatedgenesin 3 pairs HCC tumor and

pairs of HCC tumor and non-tumor samples.

RI PT

adjusted non-tumor specimens. (B) Numbers of ATOH8 gene reads were aligned in 3

SC

Supplementary Figure 2. Expression of ATOH8 in xenograft tumors in nude mice was detected by IHC. The tumors induced by Huh7-Vec or Huh7-ATOH8 cells were

M AN U

treated with5-FU and CDDP, respectively. After resection, tumors were fixed, embedded in paraffin, sectioned and stained with ATOH8 antibody (magnification,

AC C

EP

TE D

630×).

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT