Oncogene (2011) 30, 2057–2069 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE H regulated by the thyroid hormone receptors associate with tumor invasion in human hepatoma cells

S-M Wu1, Y-H Huang2, C-T Yeh3, M-M Tsai1,4, C-H Liao1, W-L Cheng1, W-J Chen5 and K-H Lin1

1Department of Biochemistry, School of Medicine, Chang-Gung University, Taoyuan, Taiwan; 2Medical Research Center, Chang Gung Memorial Hospital, Taoyuan, Taiwan; 3Liver Research Unit, Chang-Gung Medical Center, Taipei, Taiwan; 4Department of Nursing, Chang-Gung Institute of Technology, Taoyuan, Taiwan and 5First Cardiovascular Division, Chang Gung Memorial Hospital, Taoyuan, Taiwan

0 Thyroid hormone, 3, 3 , 5-triiodo-L-thyronine (T3), Introduction mediates cell growth, development and differentiation by 0 binding to its nuclear receptors (TRs). The role of TRs in Thyroid hormone (3, 3 , 5-triiodo-L-thyronine; T3)isa cancer is still undefined. Notably, hyperthyroxinemia has potent mediator of many physiological processes in- been reported to influence the rate of colon cancer in an cluding embryonic development, cell differentiation, experimental model of carcinogenesis in rats. Previous metabolism and the regulation of cell proliferation microarray analysis revealed that cathepsin H (CTSH) is (Huang et al., 2008). Two TR , THRA and THRB, upregulated by T3 in HepG2-TR cells. We verified that have been identified and mapped to human chromo- mRNA and expression of CTSH are induced by somes 17 and 3, respectively (Lazar, 1993; Yen, 2001). T3 in HepG2-TR cells and in thyroidectomized rats The liver is a typical target organ for T3. Equal amounts following administration of T3. The possible thyroid of TRa1 and TRb1 are expressed in human hormone-responsive elements of the CTSH promoter hepatocytes (Chamba et al., 1996). However, the localized to the nucleotides –2038 to –1966 and –1565 mechanisms involved in the regulation of liver-specific to –1501 regions. An in vitro functional assay showed that genes by TRa1 and TRb1 have not been elucidated. CTSH can increase metastasis. J7 cells overexpressing Evidence that TRs might be involved in carcinogen- CTSH were inoculated into severe combined immune- esis came from studies on v-erbA protein (Thormeyer deficient mice and these J7-CTSH mice displayed a and Baniahmad, 1999), thus, v-erbA antagonizes TR greater metastatic potential than did J7-control mice. The function via competition for response elements or clinicopathologic significance of CTSH expression in coactivators and ligand-dependent transcriptional acti- hepatocellular carcinoma (HCC) was also investigated. vation (Yen et al., 1994). That TRs can have an The CTSH overexpressing in HCC was associated with important role in tumor transformation was further the presence of microvascular invasion (P ¼ 0.037). The verified by aberrant expression and mutations of TRs in microvascular invasion characteristic is closely related to various human cancers including pituitary tumors our in vitro characterization of CTSH function. Our (Ando et al., 2001), hepatocellular carcinomas (HCCs) results show that T3-mediated upregulation of CTSH led (Lin et al., 2001), thyroid cancers (Puzianowska- to matrix metallopeptidase or extracellular signal-regu- Kuznicka et al., 2002) and renal clear cell carcinomas lated kinase activation and increased cell migration. This (Rosen and Privalsky, 2009). Transgenic mice (TRbPV/PV), study demonstrated that CTSH overexpression in a subset harboring a PV mutation originally identified in a hepatoma may be TR dependent and suggests that this patient with thyroid hormone resistance, spontaneously overexpression has an important role in hepatoma developed thyroid cancer (Suzuki et al., 2002). This progression. study showed that TRb could act as a tumor suppressor. Oncogene (2011) 30, 2057–2069; doi:10.1038/onc.2010.585; Recently, Zhu et al. (2010) reported that mice devoid of published online 10 January 2011 functional TRs (TRa1À/ÀTRb1À/À) spontaneously devel- Keywords: cysteine ; receptors; migration; op follicular thyroid cancer and metastases in the lung. hepatocellular carcinoma TRs could function as tumor suppressors in such mice. Besides their suppressor role, TRs enhanced tumor progression in other studies. Thyroid hormones stimu- late the proliferation of several cancer cell types including those from the pituitary (Barrera-Hernandez Correspondence: Dr K-H Lin, Department of Biochemistry, School of et al., 1999), gliomas (Davis et al., 2006), breast cancers Medicine, Chang-Gung University, 259 Wen-Hwa 1st Road, Taoyuan (Hall et al., 2008) and prostate cancers (Tsui et al., 333, Taiwan. 2008). Thyroid hormone and the TRa1 receptor control E-mail: [email protected] Received 14 June 2010; revised 23 November 2010; accepted 26 epithelial cell proliferation and the expression of November 2010; published online 10 January 2011 components of the Wnt pathway (Plateroti et al., 2006). Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2058 Various proteases are implicated in cancer progres- et al., 2008a), SPON2 (Liao et al., 2010), MATA1A (Wu sion by degrading the basement membrane and extra- et al., 2010), SULT2A1 (Huang et al., 2006) and cellular matrix in a cascade-like manner (Mohamed and AKR1B1 (Liao et al., 2009). Supplementary Table S1 Sloane, 2006; Wilson and Singh, 2008). Cysteine lists several cancer-related genes, including those encod- are a family of lysosomal proteases that are ing cathepsin proteases positively regulated by T3. often upregulated in various human cancers, and these We first used oligonucleotide microarrays to identify proteases have been implicated in distinct tumorigenic the genes induced by T3 in HepG2-TRa cells. Upregula- processes, such as angiogenesis, proliferation, apoptosis tion of cathepsin cysteine proteases was identified and and invasion (Berdowska, 2004; Joyce and Hanahan, the result was verified by quantitative reverse transcrip- 2004; Gocheva and Joyce, 2007). Cysteine cathepsin tion–PCR. Several human cysteine cathepsins (CTSH, proteases have recently emerged as that are CTSB, CTSS and CTSL) were induced by 3.2–8.6 times active in cancer development, and cysteine cathepsin after the incubation of HepG2-TRa1#1 cells with 10 nM inhibitors have been proposed as anticancer agents T3 for 72 h (Supplementary Figure S1). However, the (Palermo and Joyce, 2008). There is increasing evidence mRNA levels of cysteine proteases, such as CTSV, were that the expression of cathepsin H (CTSH) is increased not induced by T3. in disease states including breast carcinoma (Gabrijelcic Northern blot analysis was used to assess further the et al., 1992), melanoma (Kos et al., 1997), glioma response of CTSH mRNA expression to the exogenous (Sivaparvathi et al., 1996), and colorectal and prostate addition of T3. A 1.47-kb CTSH transcript was carcinoma (del Re et al., 2000). identified in HepG2-TRa1#1 cells (Figure 1a). The Although several individual target genes have been CTSH mRNA levels were induced by 3–4 times after studied in the liver, a large-scale profile of the target incubation of HepG2-TRa1#1 cells with 10 nM T3 for genes regulated by thyroid hormone has not been 48 h. A similar induction was observed in HepG2- undertaken. CTSH expression was induced by T3 TRa1#2 and -TRb1 cells. As a control, HepG2 cells in HepG2-TRa cells identified from oligonucleotide were transfected with the empty vector, yielding a cell microarrays. CTSH and other proteases may be line expressing the Neo protein (HepG2-Neo cells). involved in the destruction of extracellular matrix components, leading to cancer proliferation, migration and metastasis (Tsushima et al., 1991). However, T3 upregulates expression of CTSH protein in HepG2-TR elucidation of the underlying mechanisms by which T3 cells regulated requires further study. Previous reports have The effect of TRs on the degree of CTSH protein shown that TRs are potent inducer of tumorigenesis expression in HepG2 isogenic cell lines was studied (Plateroti et al., 2006; Chen et al., 2008a; Kress et al., (Figure 1b). CTSH is synthesized as an inactive 2009). In this study, we report that the T3/TR controls proenzyme (41 kDa, upper arrow), which is then the transcription of the CTSH gene and that T3- processed proteolytically to an active single chain, mediated cell migration appears to involve CTSH. mature form (28 kDa, lower arrow) within the endo- somes/lysosomes (Turk et al., 2001). The incubation results indicated that T3 increased the amount of 28- kDa CTSH protein in the HepG2-TRa1#1, -TRa1#2 Results and -TRb1 cells. The CTSH protein levels increased after incubation of HepG2-TRa1#1, -TRa1#2 and -TRb1 Cathepsin mRNA is regulated by T3 in the HepG2-TR cell line cells with 1 or 10 nM T3 for 24–72 h. These results show We used isogenic HepG2 cell lines that consistently that the effect of T3 on the CTSH protein level in TRa1- express wild-type TRa1 (HepG2-TRa1#1, #2) and overexpressing cells is time and dose dependent. TRb1 (HepG2-TRb1). As a control, HepG2 cells were Immunoblot analysis showed that incubation of control transfected with the empty vector, yielding a cell line HepG2-Neo cells, which did not express detectable expressing the Neo protein (HepG2-Neo cells). In the endogenous TR proteins, in 10 nM T3 for 24–72 h did not four HepG2 cell lines investigated in this study, HepG2- increase the CTSH protein level significantly. However, TRa1#1, -TRa1#2, -TRb1 and -Neo, overexpression of the CTSH proenzyme was affected marginally by T3. TR protein was approximately 10-, 3.3- and 2.5-fold To analyze whether cathepsin is secreted into the that of the HepG2-Neo control cell line (data not culture medium, the HepG2-TR stable cell line was shown). Previously, complementary DNA-microarrays incubated in serum-free medium, and the medium was were performed to identify the target genes regulated collected at 24 and 48 h. Analysis using anti-CTSH by T3 in a hepatoma cell line-overexpressing TRa1 antibodies revealed a major extracellular product with (HepG2-TRa1) (Shih et al., 2004). That study demon- an apparent molecular mass of 41 kDa (Supplementary strated that 148 of the 7597 genes represented were Figure S2). This product was upregulated markedly in positively regulated by T3, including those encoding HepG2-TR cells incubated in media containing varying fibrinogen, fibronectin, transferrin and several other levels of T3 for 24 and 48 h. CTSH protein level did not coagulation factor system components. Recently, oligo- increase significantly in HepG2-Neo cells incubated in nucleotide microarrays were carried out to identify more 10 nM T3 for 24–48 h. TR-target genes. Some of the results have been Similar results were observed in the other HCC cell published including the identification of FURIN (Chen line, Huh7 cells, which expressed detectable endogenous

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2059

Figure 1 Effect of T3 on the amount of cysteine proteases and CTSH mRNA and protein in HepG2-TR cell lines. HepG2 cell lines that consistently express wild-type TRa1 (HepG2-TRa1#1, #2) and TRb1 (HepG2-TRb1) were used. As a control, HepG2 cells were transfected with the empty vector, yielding a cell line expressing the Neo protein (HepG2-Neo cells). (a) The CTSH mRNA expression level in four TR stable lines. HepG2-TRa1orb1 cell lines were incubated for 24–72 h in the absence or presence of T3, after which total RNA was isolated and subjected (20 mg per lane) to Northern blot analysis with 32P-labeled CTSH and 18S RNA probes. Positions of the 1.5-kb CTSH and 18S RNAs are indicated. Student’s t-test. **Po0.01; *Po0.05, T3-depleted (Td) vs 1 or 10 nM T3 treated. (b) Three TR stable lines and Neo cells were incubated in Td medium in the absence or presence of T3 for 24–72 h, after which cell lysates (20 mg protein) were subjected to immunoblot analysis with antibodies to CTSH. The pro-form (upper arrow) and active form (lower arrow) of CTSH band is shown. Actin was used as an internal control. (c) Quantitative real-time PCR analysis of CTSH mRNA in Huh7 cells after 1–10 nM T3 for 24–48 h treatment. (d) Western blot analysis of total cell lysates or extracellular forms of CTSH in Huh7 cells. Huh7 cells were incubated in serum-free medium and 1–10 nM T3 for 24 and 48 h on 10 cm dishes. The culture medium was collected and 20 mg of media were subjected to immunoblot analysis with CTSH antibodies and loading control by using coomassie blue stained. Data are means±s.e. of values from three independent experiments.

TR proteins: 10 or 100 nM T3 increased CTSH mRNA create a luciferase-based reporter plasmid. A series of and protein expression by 1.5–2 times in Huh7 cells at pA3TK-luc reporter constructs containing progressively 48 h (Figures 1c and d). The induction of extracellular shorter fragments of CTSH 50-flanking DNA were CTSH protein was also observed in Huh7 cells co-transfected with pcDNA-TRa1 in HepG2 cells (Figure 1d). (Figure 2a). Increases of 3.67 and 8.16 times in the— 2723 to þ 52 bp CTSH promoter activity were stimu- lated by 10 and 100 nM T3, respectively. A bioinfor- T3 induces expression of CTSH at the transcriptional matics search revealed putative TREs in the CTSH level promoter. Therefore, serially deleted mutants were also Promoter activity assays were performed to determine constructed from the CTSH fragment to identify the whether the regulation of CTSH expression by T3 region targeted by TR. Treatment with 10 or 100 nM T3 occurs at the transcriptional level. The nucleotide À2723 increased the activity of the CTSH deletion mutants to þ 52 upstream region of CTSH was inserted (À2038 to À1801, À1739 to À1501, À1685 to À1501 and upstream to a minimal thymidine kinase promoter to À1565 to À1415) compared with the T3-free condition.

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2060

Figure 2 Regulation of CTSH expression by T3 at the transcriptional level. (a) Schematic representation of CTSH promoter with potential TRE-binding sites indicated by oval shapes. The numbers beside each construct indicate the at which the site starts based on the transcription start site ( þ 1). A series of pA3TK-luc reporter constructs containing progressively smaller amounts of CTSH 50-flanking DNA (1 mg per well) were co-transfected with pcDNA-TRa1#1 in HepG2 cells. After transfection, cells were harvested after T3 treatment for 24 h. Data were normalized to co-transfected b-galactosidase, and fold-activation was calculated relative to each pA3TK-luc control. Results are presented as means s.d. of data from three independent experiments performed in triplicate. (b) In vitro-translated TRa1orTRb1 incubated with wild-type or mutated TRE of CTSH gene promoter À2038 to À1966 along with TNT in vitro-translated RXRa in a final volume of 20 ml. Antibody C4 is a mouse monoclonal antibody against TR proteins, and MOPC21 is a nonspecific monoclonal antibody. The position of probe complexed with TR and RXR is indicated by the arrows. SS: TR complexes supershifted by C4 or RXRa antibody. The wild-type or mutated TRE of CTSH gene promoter À1565 to À1501 were amplified and labeled with [a-32P-dCTP] as probe.

The possible TREs located to the À2038 to À1801 and the two possible TREs of the CTSH promoter À1565 to À1415 constructs. Transfection analysis (Figure 2b). Reactions performed with [a-32P]-labeled showed that two possible TREs (À2038 to À1801 and À2038 to À1966 (lanes 1–16) and À1565 to À1501 (lanes À1565 to À1415) in the upstream region of human 17–32) fragments of in vitro-translated TRa1/b1 and CTSH were responsible for T3-induced human CTSH RXRa proteins produced one predominant band (lanes gene transcription. Mutations in the two TRE sequences 4, 11, 20 and 27). This band was supershifted by TR of the CTSH promoter-caused loss of the TR/T3- antibody (lanes 5, 12, 21 and 28) but not by control stimulated CTSH promoter activity. These results antibody (MOPC21; lanes 6, 13, 22 and 29). This band indicate that the two possible TREs of the CTSH was competed out by adding a 2- or 10-times molar promoter are located in the À2038 to À1801 and À1565 excess of cold probe (compare lanes 7 and 8; 14 and 15; to À1415 regions. 23 and 24; 30 and 31). However, mutations in the TREs We used an electrophoretic mobility shift assay in the –2038 to –1966 and –1565 to –1501 regions caused (EMSA) to investigate whether TRs bind directly to loss of the TR-binding activity (lanes 9, 16, 25 and 32).

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2061 the CTSH promoter in vivo (Figure 3b). Both TRa1/b1 and RXRa were clearly recruited to the TRE-, whereas control immunoglobulin G produced only background levels. The positive control (human furin gene containing TRE) showed detectable bands using antibodies to RXRa or TRs. However, no band was detected using a primer set for the negative control (human glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene). The ChIP assay revealed that the TRs and RXRa complex bound to the CTSH promoter.

T3 induces CTSH mRNA and protein expression in vivo Two groups of 6-week-old male Sprague–Dawley rats (n ¼ 6) underwent surgical thyroidectomies (TX) to determine the in vivo response of CTSH to T3 treatment. One group (three rats per group) was injected with T3 daily for 2 weeks (TX þ T3), the second group did not receive T3 injections (TX) and the third group included the control sham-operated rats (n ¼ 3). The rats were killed at the end of the experiment, and T3 and TSH concentrations were measured in the serum. The livers were removed and cell lysates were prepared for western blot analysis. The serum T3 level in the TX group was about 0.016 times (7.6 vs 489 ng/dl) that in the group that underwent T3 treatment (TX þ T3). The TSH level in the TX group was 226 times (3.16 vs 0.014 mg/ml) that in the T3-treated group. The T3 and TSH serum Figure 3 Identification of the TR proteins that are associated with levels in the sham group were around 40 ng/dl and the TRE regions of CTSH gene promoter in vivo.(a) Nuclear 2.23 mg/ml, respectively. extracts (NE) were isolated from HepG2-TRa#1 cells, and 5 mgof Western blot analysis of cell lysates demonstrated that protein were incubated with [a-32P-dCTP] labeled probes men- tioned in Figure 2. A upstream region À2264 to À2054 in the CTSH protein levels were much higher in the rats CTSH promoter gene devoid of the TRE region was a nonspecific administered T3 than the TX group. The expression of (ns) probe. (b) ChIP assay as described in Materials and methods CTSH mRNA and protein was higher in the TX þ T3 or section. DNA corresponding to TRE of CTSH promoter region sham group than in the TX group (Figures 4a and b). was analyzed by PCR with the indicated primers. Furin or glyceraldehyde 3-phosphate dehydrogenase (GAPDH) promoter Further, we verified the in vivo response of the CTSH region was used as a positive or negative control, respectively. gene to T3 showing that CTSH mRNA and proteins Another positive control was used chromatin before immunopre- were upregulated in thyroidectomized rats after T3 cipitation (input) and negative controls were immunoprecipitates treatment. These results help elucidate the regulation with MOPC21 antibody. ChIP assays were also performed in of CTSH expression by thyroid hormones. The common HepG2-TRa or b cells. proteolytic degradation by lysosomal cysteine protei- nases could be a general process (Dickinson, 2002). TR and RXR proteins form a complex with TREs located Besides their role inside lysosomes, cysteine proteases in the CTSH promoter degrade proteins outside lysosomes. It is required to TR proteins also bound to the two TREs of the CTSH clarify the other role of CTSH regulated by T3 in normal promoter when HepG2-TRa nuclear extracts were used liver in the future. (Figure 3a). Briefly, one predominant band was detected The clinicopathologic significance of CTSH expres- (lanes 2 and 10) when using [a-32P]-labeled –2038 to sion in HCC was also investigated. A total of 93 patients –1966 and –1565 to –1501 fragments. This band was with HCC were enrolled consecutively in this study. An specifically supershifted by antibody to the TR (lanes 4 equal amount (30 mg) of protein from each specimen was and 12) or RXRa (lanes 5 and 13) but not by the control loaded for electrophoresis and western blot analysis. antibody (lanes 3 and 11). The band could be competed Actin was used as an equal loading control. CTSH out by adding a 2- or 10-times molar excess of specific proteins were detected in most of the cancerous tissues. cold probe (compare lanes 6 and 7, and 14 and 15) but The results from 15 representative paired HCC speci- not by nonspecific cold probe (lane 8 and 16). The mens are shown in Figure 4c, which shows increased promoter assay and EMSA findings suggest that TRs CTSH expression and concomitantly increased most activate CTSH gene transcription by binding directly to of the TR proteins expression in HCC tissues. The the TRE sites located in the À2038 to À1966 and À1565 percentage of all paired samples with upregulated CTSH to À1501 regions. relative to the matched noncancerous adjacent tissues In the chromatin immunoprecipitation (ChIP) assay, (B1 cm distant from HCC) was 64.5% (60/93) in the the TR proteins associated with the two TRE regions of cancerous tissues. Both TRa1 and TRb1 expression was

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2062

Figure 4 The CTSH was regulated by TR in thyroidectomized rats and human HCC specimens. (a) Induction of CTSH expression by thyroid hormone in rat liver. The protein expression of CTSH in normal (N, sham-operated), TX, or TX þ T3 male Sprague–Dawley rat liver was determined by western blot analysis. Student’s t-test. **Po0.01, TX þ T3 or N vs TX treated. (b) The mRNA expression of CTSH was determined by quantitative reverse transcription (Q-RT)–PCR. (c) TR and CTSH protein overexpression in 15 representative pairs of human HCC specimens. The band intensities were quantified and normalized to actin. The normalized tumor/ non-tumor (T/N) ratio is shown in the right panel.

increased in 32.3% (30/93) of the cancerous tissues. Migration was measured for 24 h with medium contain- Analysis of the CTSH tumor/normal ratio showed no ing 1% (control) or 10% serum (Figure 5b). Cells correlation with disease-free or overall survival. Multi- traversing the membrane were stained blue with crystal ple regression analysis was performed using a stepwise violet (Figure 5b, lower panel). The number of model (SPSS version 13.0). The stepping method criteria transmigrating cells was quantified in three or more was set as probability ¼ 0.05 for entry and 0.10 for separate experiments. Migration was significantly great- removal. The T/N ratio of CTSH was associated with er in the CTSH-expressing cells. Invasion of J7- being male (b ¼À0.559; 95% confidence interval CTSH#1, -CTSH#2 or -CTSH-pooled cells was 2–3 (CI) ¼À0.929–0.188; P ¼ 0.004); albumin level (b ¼ 0.281; times higher than in J7-control cells when assessed using 95% confidence interval ¼ 0.017–0.545; P ¼ 0.011); and the crystal violet staining. The number of invasive cells was presence of microvascular invasion (b ¼ 0.481; 95% con- quantified (Figure 5c). These gain-of-function studies fidence interval ¼ 0.112–0.849; P ¼ 0.037). The microvas- showed that CTSH can increase the metastatic potential cular invasion characteristic in HCC is closely related to of hepatoma cells. our in vitro characterization of CTSH function. The T3 can induce CTSH protein expression in Huh7 cells, overexpression of CTSH is not correlated to the hepatitis which express endogenous TR proteins. To determine B virus-caused HCC. whether CTSH is necessary for migration and invasion in hepatoma cells, we reduced CTSH expression in Huh7 cells. Huh7 cells were transfected with a selectable vector CTSH overexpression promotes cell migration and expressing CTSH small interfering RNA or a control invasion vector and then selected with puromycin. Stable clones To determine whether CTSH overexpression increases resistant to puromycin were selected, and the levels of the the metastatic potential in hepatoma cells, we used the cellular and extracellular forms of CTSH protein were parental HCC cell line J7, which expresses the lowest measured (Figure 5d). Huh7-green fluorescent protein - levels of CTSH of the hepatoma cell lines. J7 cells with KD control, Huh7-CTSH-KD#1, -KD#2 and -KD- stable integration of the CTSH gene were selected in pooled cells, which express very low levels of CTSH, media containing G418. Stable clones resistant to G418 showed less migration and invasion compared with green were selected, and the levels of cellular and extracellular fluorescent protein-KD cells. Knockdown of CTSH with forms of CTSH protein were measured (Figure 5a). In a small interfering RNA led to a 42–54% decrease in Transwell migration assay, J7 cells stably expressing migration and 34–43% decrease in invasion (Figures 5e CTSH in medium containing 1% serum were added to and f). These knockdown studies show the functional role the upper surface of the chamber and allowed to adhere. of CTSH in invasion and metastasis.

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2063

Figure 5 Functional assay of CTSH in hepatoma cells. (a) Overexpression of CTSH promotes migration and invasion. In all, 20 mgof media were subjected to immunoblot analysis with CTSH antibodies. Bottom was shown as loading control using coomassie blue stained. Overexpression of CTSH increases (b) migration and (c) invasion. J7 cells stably expresses CTSH (CTSH#1, #2 clones) and J7 cells stably expresses control vector (control) migration over a time period of 24 h was quantified with a Transwell migration assay and an invasion assay through a Matrigel-coated membrane. The mean migration or invasion of J7-vector cells was assigned a value of 1. Student’s t test. **Po0.01; *Po0.05, CTSH#1, #2 or pool vs vector. (d) Two Huh7 CTSH knockdown (KD) stable lines (CTSH- KD#1, #2), one pool (CTSH-KD pool) stable cells and one Huh7 luciferase knockdown control cell line (KD-control) were used. Knockdown of CTSH in Huh7 cells was confirmed by western blot in cell lysate or conditioned medium. Knockdown of CTSH reduces (e) migration and (f) invasion. Cell image was shown in the bottom of b, c, e and f.

CTSH overexpression induced by T3 enhances migration control mice. CTSH protein expression, shown as a To determine whether the in vitro results can be applied dark-brown color in immunohistochemical staining, was to the in vivo condition, severe combined immunodefi- markedly stronger in J7-CTSH lung tumors than in cient (SCID) mice were used. Mice injected with the J7-control tumors. To verify further whether CTSH is J7-control and J7-CTSH SCID mice developed multiple involved in T3-mediated migration, we reduced CTSH macroscopic tumor nodules in the lung (Figure 6a) as expression in HepG2-TRa#1 and Huh7 cells using shown by hematoxylin and eosin staining. The meta- CTSH small interfering RNA, which knocks down static index was 2.3 times higher in the J7-CTSH SCID CTSH expression. CTSH small interfering RNA de- mice than in the mice injected with J7-control cells. The creased the T3-mediated migration in HepG2-TRa#1 tumor size relative to the average tumor size (per cm2 and Huh7 cells (Figure 6b). To verify whether antibody lung section in the J7-CTSH/J7-control mice) was about against CTSH prevents the cellular response, similar 4.8 times larger in J7-CTSH SCID mice than in J7- migration experiments were performed. Anti-CTSH

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2064

Figure 6 Mediated overexpression of CTSH by thyroid hormone promotes hepatocyte migration. (a) J7-control or J7-CSTH SCID mice (n ¼ 4 per group) were established. All analyses were carried out 7 weeks after inoculation of tumor cells. Immunohistochemical staining of metastasized J7-control or J7-CSTH cells in lung sections are displayed. The metastasis index (fold, density of tumor 2 numbers in J7-CSTH or J7-control per cm area) in lung is shown in the lower panel. (b) Effect of T3-mediated migration activity in HepG2-TRa#1 and Huh7 cells, which knocked down CTSH expression. HepG2-TRa or Huh7 cells were transient transfected with either control (luc short hairpin RNA (shRNA)) or CTSH shRNA by electroporation. After puromycin selection, stable clones were pooled for migration assay. The cell lines were added to the upper chamber of Transwell units and incubated in the absence or presence of T3 (10 nM) for 24 h. The number of cells that transversed the filter to the lower chamber was then determined and expressed as the relative migration activity. (c) Effect of anti-green fluorescent protein (GFP) or anti-CTSH antibody treatment on migration activity in HepG2-TRa#1 and Huh7 cells. The cell lines were added to the upper chamber of Transwell units and incubated in the absence or presence of T3 (10 nM) and anti-GFP or anti-CTSH antibody (10 mg/ml) for 24 h. (d) Two different control cell clones (J7-control#1 and J7-control#2) that stably express neo proteins and two J7-CTSH stable clones were used for MMP assays. J7-control or J7-CSTH cells (2 Â 106) were plated in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum. After 24-h incubation and subsequent washing, the medium was replaced with serum-free medium. Medium was collected for MMP detection. Gelatin zymography was performed; the position of proenzyme and active form of MMPs is shown to the left. (e) HepG2-TRa#1 and J7- derived cells were treated in serum-free medium and were incubated in T3 (0, 1 and 10 nM) for 24 or 48 h. Medium was collected for western blot analysis of MMP3. (f) J7-control or J7-CSTH cells were cultured in serum-free medium for 24 h. Cell lysates were western blotted with anti-ERK and anti-phospho-ERK antibodies. Actin was used as an internal control. (g) HepG2-TRa#1 or Huh7 cells were stimulated to migrate T3 by in the presence or absence of indicated MEK inhibitors (U0126). The relative migration means that in response to T3 ( þ or – inhibitor) normalized to dimethylsulphoxide (DMSO) treated, T3-deprived controls. Data are means±s.e. of values from three independent experiments.

antibody but not anti-green fluorescent protein antibody To determine whether MMPs are potential down- reduced the T3-mediated cell migration in HepG2- stream mediators of CTSH, both pro-MMP-9 (97 kDa) TRa#1 and Huh7 cells (Figure 6c). and pro-MMP-2 (72 kDa) were studied in J7-derived

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2065 cells. Gelatin zymography revealed that CTSH-induced low levels of endogenous TR proteins, CTSH was not MMP-2 activity and cleaved pro-MMP-2 to MMP-2 significantly induced by T3 in HepG2-neo cells. How- (Figure 6d). Western blot analysis of pro-MMP-3 and ever, another hepatoma cell line, Huh7, which express MMP-3 protein expression demonstrated an increase in relatively high levels of TR proteins and T3 can induce the active form of MMP-3 in T3-stimulated HepG2- CTSH proteins expression in these cells. The gain-of- TRa#1 cells. Similarly, MMP3 was activated in J7- function studies showed that CTSH can increase the CTSH cells (Figure 6e). Phosphorylation of extracellular metastatic potential of hepatoma cells. Our results show signal-regulated kinase (ERK) regulates the association that CTSH overexpression in some HCCs may be TR of focal adhesion kinase, and paxillin is involved in focal dependent and that this overexpression might have an adhesion. ERK might also have a role in cell migration important role in HCC progression. Taken together, by inhibiting the ability of integrins to bind to their these results suggest that T3 regulates CTSH gene extracellular matrix ligands (Teranishi et al., 2009). To expression, thereby enhancing cancer invasiveness. study whether the CTSH-mediated signaling involves Cathepsins belong to the class of proteases in the phosphorylation, p-ERK and total ERK levels were complex interplay between tumor and stroma. A measured. ERK was phosphorylated in J7 cells, which majority of cathepsins are lysosomal proteases, but overexpress CTSH (Figure 6f). We also tested the effect some are secreted or membrane associated. The of U0126, an inhibitor of mitogen-activated protein intracellular active form (28-kDa) of CTSH is auto- kinase (MAPK) kinase (MEK) activation and MEK catalyzed from the 41-kDa pro-form at acidic pH catalytic activity (Figure 6g). Cells were pretreated with conditions and localized within the intercellular com- U0126 or dimethylsulphoxide for 15 min before treat- partments (Turk et al., 2000; Dickinson, 2002). The ment with 10 nM T3 for an additional 24 h. U0126 secreted CTSH exists predominantly in a pro-form in decreased T3-mediated migration in HepG2-TRa#1 and the medium. Here, we showed an increase in both of the Huh7 cells. These data are consistent with the role of 28- or 41-kDa form of CTSH in HepG2-TR cells after ERK in T3-stimulated migration. Taken together, the T3 treatment. However, the trafficking or sorting results show that T3-mediated upregulation of CTSH led mechanisms of CTSH need to be clarified. to MMP or ERK activation and increased cell migration. Schraufstatter et al. reported that interleukin-8- induced transactivation of the epidermal growth factor receptor is mediated by CXCR2 and involves CTSB, and that this pathway is important for the migratory Discussion and tumorigenic effects of interleukin-8 (Schraufstatter et al., 2003). We used oligonucleotide microarrays and This study characterized CTSH, which was identified found that the expression of interleukin-8 and CTSB previously in microarray screening for T3-responsive was induced by T3 in HepG2-TR cells (Supplementary genes in HepG2-TR cells. However, the transcriptional Table S1). Using HepG2-TR and Huh7 cells, which regulation of the human CTSH gene was not elucidated express endogenous TR, we also found that T3 enhances in previous studies. We attempted to elucidate the the metastatic ability of hepatoma cell lines. These molecular mechanism of CTSH regulation by T3 in results show that T3 enhances the migratory and isogenic HepG2 cell lines. Our study indicates that T3- invasive properties of hepatoma cell lines. induces CTSH expression at both the mRNA and A previous study indicated that T3, acting through protein levels in HepG2 cell lines. Similar results were TRs, inhibits the expression of NM23-H1 and promotes observed in Huh7 cells, which expressed detectable tumor metastasis (Lin et al., 2000). Thyroid hormone endogenous TR proteins and in thyroidectomized rats. and TRa1 control epithelial cell proliferation and the Further study demonstrated that T3 upregulates CTSH expression of components of the Wnt pathway (Plateroti at the transcriptional level. The promoter assay and et al., 2006). The expression of the b-catenin and sFRP2 EMSA findings suggest that TR activates CTSH gene gene is regulated directly by TRa1, and the Wnt/ transcription by binding directly to the two TRE sites b-catenin signaling networks contribute to intestinal located in the À2038 to À1966 and À1565 to À1501 tumorigenesis (Kress et al., 2009). Chiloeches et al. regions. The ChIP assay revealed that the TRs and (2008) reported that ERK but not c-Jun N-terminal RXRa complex bound to the CTSH promoter. kinase or p38-MAPK is activated on exposure to T3 in Although our study investigated T3-induced CTSH rat pituitary tumor (GH4C1) cells. In general, ERK has expression in human hepatoma cell lines, similar results been implicated in the migration of numerous cells have been obtained in animal studies. (Huang et al., 2004). In our study, T3-mediated ERK Although the role of CTSH in tumor progression is activation was observed in HepG2-TRa1 cells and in not understood well, a possible function of CTSH in parallel with the phosphorylation of ERK in J7 CTSH- tumor progression is its ability to degrade fibrinogen overexpressing cells. Treatment of HepG2-TRa1or and fibronectin. CTSH and other proteases may be Huh7 cells with U0126 inhibited the migration in involved in the destruction of extracellular matrix response to T3. We demonstrated previously that T3 components, leading to cancer proliferation, migration regulates furin gene expression via a newly identified and metastasis (Tsushima et al., 1991). Similar induction mechanism or in cooperation with transforming growth by T3 of extracellular CTSH protein was observed in factor-b to enhance tumor metastasis in vitro and in vivo HepG2-TR and Huh7 cells. As HepG2 cells express very (Chen et al., 2008a). T3 and transforming growth factor-b

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2066 regulate furin gene expression directly, thereby enhan- enhance the metastatic potential of hepatoma cells. T3 cing the invasiveness of liver cancer cells in SCID mice also stimulates MMP2 and MMP3 activation, in parallel models. with the CTSH-mediated migration, by increasing The role of TRs in cancer progression appears to be MMP activity. The studies discussed above and this multifaceted. Recent report showed that hypothyroid- report shows that TR is a potential tumor promoter in ism was associated with a high risk of HCC in women tumorigenesis. CTSH and furin or other proteases may (Hassan et al., 2009); however, why hypothyroidism be involved in liver cancer progression. Our current data causes HCC is not clear. Martinez-Iglesias et al. (2009) extend our previous results showing that T3 has a role in reported that TRb1 might act as a potent suppressor of cancer progression. Although CTSH expression is tumor invasiveness and metastasis in hepatocarcinoma upregulated in some types of HCC, its significance (SK-hep1) and breast cancer (MDA-MB-468) cells. The requires further study. results indicate that the re-expression of TRb1in hepatoma and breast cancer cell lines that have lost TR expression leads to tumor growth retardation, Materials and methods partial mesenchyme-to-epithelium transition and sup- pression of metastasis in nude mice. Interestingly, the Cell culture results in that model indicate that TRs can have diverse The human hepatoma cell lines (HepG2 and its isogenic lines, roles at different stages of tumorigenesis. TR-aÀ/ÀTR-bÀ/À Huh7, and J7) were routinely grown in Dulbecco’s modified double knockout mice, which are vulnerable to skin Eagle’s medium supplemented with 10% (v/v) fetal bovine carcinogenesis, showed that TRs deficiency seemed to serum. Serum was depleted of T3 by AG 1-X8 resin (Bio-Rad, inhibit benign tumor formation at early stages of Hercules, CA, USA) as described elsewhere (Samuels et al., carcinogenesis but increased malignant transformation 1979). Cells were cultured at 37 1C in a humidified atmosphere at later stages (Martinez-Iglesias et al., 2009). The of 95% air and 5% CO2. Both HepG2 and Huh7 cell lines are well differentiated; J7 is intermediately differentiated. T3 was thyroid hormone status seems to have a dual effect on purchased from Sigma-Aldrich, St Louis, MO, USA. For T tumor development. A recent report demonstrated that 3 induction, cells were exposed to 1 and 10 nM T3 for 24, 48 and the expression levels of several components of the Wnt 72 h, respectively, after 24 h cultured in the T3-depleted-medium. pathway were directly regulated by TRa1 in an animal model, allowing increased expression of b-catenin/Tcf4 Quantitative reverse transcription–PCR target genes and stimulation of cell proliferation (Kress Real-time quantitative reverse transcription–PCR was con- et al., 2010). However, TRa1 when specifically over- ducted in a 15 ml reaction mixture containing 25 nM forward expressed in the intestinal epithelium is unable to induce and reverse primers, 1 Â SYBER Green reaction mix (Applied cancer development. Conversely, the investigators Biosystems, Carlsbad, CA, USA), and varying quantities of observed that TRa1 accelerated tumorigenesis in mice template as described previously (Shih et al., 2004). The with a Wnt-activated Apc þ /1638N genetic background. sequences of primers used in quantitative real-time PCR listed These divergent actions of TRs in HCC development in Supplementary Table 2. possibly result from varying tissue- and stage-specific microenvironments. Northern blot analysis Previously, we reported that TRs might have a Total RNA was extracted from the cells with TRIzol Reagent suppressor role by reducing PTTG1 and Sp1 expression (Life Technologies Inc., Carlsbad, CA, USA). Equal amounts (Chen et al., 2008b). Knockdown of these genes of total RNA (20 mg) were analyzed on a 1.2% agarose- inhibited HepG2-TRa1 cell growth in parallel with formaldehyde gel. Separated RNA molecules were then T -mediated HepG2-TRa1-induced cell growth inhibi- transferred to a nitrocellulose membrane and subjected to 3 Northern blot analysis as described previously (Huang et al., tion. However, furin and CTSH expression levels were 2006). Intensities of the CTSH mRNA bands on blots were positively regulated by T in parallel with T -mediated 3 3 quantified, and the degree of the T3-induced increase in CTSH HepG2-TRa1 cell migration. These contradictory re- transcripts were determined at each point. The signals were sults in cancer progression suggest that the final roles of assayed for CTSH mRNA expression level after normalizion TRs in liver cancer might depend on reciprocal actions with 18S rRNA using FujiFilm image gauge V3.46 analysis on tumor cells and on other cellular regulators on the software (Fuji Film, Tokyo, Japan). tumor milieu. Tumor-infiltrating cells orchestrate the microenvironment and provide growth factors, proan- Immunoblot analysis giogenic factors and proteases, as well as adhesion Cells were lysed in lysis buffer (25 mM Tris-HCl pH 7.4, molecules that facilitate tumor cell proliferation, angio- 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% genesis and migratory activity (Nelson and Ganss, sodium dodecyl sulfate) containing inhibitors (apro- 2006). Interestingly, one of the TRb1 interacting tinin, leupeptin and PMSF) on ice, and centrifuged at partners, transforming growth factor-b, works both as 10 000 Â g at 4 1C for 30 min. The supernatant was transferred a tumor suppressor and a tumor promoter (Yang and to a new tube. Total cell lysates (20 mg per lane) were fractioned by SDS–polyacrylamide gel electrophoresis on a Moses, 2008). Therefore, further studies are necessary to 10% gel. Separated proteins were transferred to a nitrocellu- investigate what factors might mediate the switching of lose membrane and visualized by chemiluminescence using an the functions of TRs in the microenvironment of HCCs. ECL detection kit (Amersham Inc., Piscataway, NJ, USA) as Here, we found that CTSH is also regulated directly described previously (Liao et al., 2009). The antibodies used by T3 in human hepatoma cell lines and that it can were goat polyclonal antibodies to CTSH (Santa Cruz

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2067 Biotechnology Inc., Santa Cruz, CA, USA). The intensities of and MOPC21 mouse monoclonal antibody (ICN Biomedicals, the immunoreactive bands were quantified using Image Gauge Aurora, OH, USA) as a negative control for EMSAs. software (Fuji Film). Chromatin immunoprecipitation Animals ChIP assay was performed as described previously (Liao et al., Twelve male Sprague–Dawley rats underwent TX at 6 weeks 2010). The upstream TRE region of the human CTSH gene 0 of age according to method used in previous reports (Yang was amplified with the forward primer 5 -AAAGAGAGGGA 0 et al., 1999). Before removing the thyroid gland, the anterior AGGGCACAG-3 at nucleotide À2038 and reverse primer 0 0 and posterior thyroid arteries were ligated, and the external 5 -AGTGATCCTCCTGCCTCA-3 at nucleotide À1966. A parathyroid glands were carefully isolated during removal of proximal TRE region was amplified with the forward primer 0 0 the thyroid gland. Sham operations were performed under the 5 -AGTGAAGAACAGGGCATC-3 at À1565 and reverse 0 0 same conditions except that the thyroid, parathyroid glands primer 5 -AGAGG GTGATTCAGACCA-3 at À1501. The and the blood supply to the glands were kept intact. Each rat TRE region of furin promoter was amplified with forward 0 0 was given 1% calcium lactate in drinking water after surgery. primer 5 -TACTAGCGGTTTTACGGGCG-3 and reverse 0 0 Two weeks after surgery, each rat was intraperitoneal injection primer 5 -TCGAACAGGAGCAGAGAGCGA-3 for positive control. The 166-bp pair product of glyceraldehyde 3- with T3 at 10 mg per 100 g body weight or a control vehicle (2.5 mM NaOH in phosphate-buffered saline) daily for 2 phosphate dehydrogenase promoter without TRE was ampli- 0 0 further weeks. Rats were killed at the end of the experiment, fied with forward primer 5 -CTCCAAAGACCCACTGCG-3 and reverse primer 50-CCACTTGTCCTC AGGCCTAG-30 and their serum was used for T3 and TSH determination. The expression levels of mRNA and proteins in the liver were for a negative control. C4 mouse monoclonal antibody to TRs analyzed by quantitative reverse transcription–PCR and western and rabbit polyclonal antibody to RXRa (Santa Cruz) were blot, respectively. All animal experiments in this study were used for ChIP analysis. performed in accordance with National Institutes of Health guidelines and the Chang-Gung Institutional Animal Care and Zymography Use Committee Guide for Care and Use of Laboratory Animals. In all, 20 mg of concentrated medium was diluted in 50 mM Tris-HCl (pH 7.4) without a reducing agent and separated by 10% SDS–polyacrylamide gel electrophoresis in the presence Transient transfection and luciferase assays of 1 mg/ml gelatin as described (Chen et al., 2008b). To clone the CTSH 50-flanking region and promoter activity assay, fragments of the CTSH promoter (nucleotides À2723 to þ 52) were amplified using PCR according to the published Migration and invasion assays nucleotide sequence and then inserted into the pA3TK-Luc Hepatoma J7 cells were transfected with the pcDNA-CTSH vector. Transient transfection in HepG2 cells was performed expression vector, and the J7 stable line-expressing CTSH was by the lipofectamine transfection method using as described identified by G418 selection. To establish CTSH knockdown, (Wu et al., 2010). We used vector NTI v7 software (Invitrogen we used the stable hepatoma Huh7 cell line and the clone Inc., Carlsbad, CA, USA) for possible TRE sequences analysis TRCN0000003699 of short hairpin RNA targeting CTSH and CTSH gene promoter sequence in Supplementary Figure S3. from the National RNA Interference Core Facility (Institute of Molecular Biology, Academia Sinica). Cell migration was studied using a modified two-chamber migration assay (Falcon Effect of knocked-down CTSH expression BD, Franklin Lakes, NJ, USA) with a pore size of 8 mm as The clone TRCN0000003699 of short hairpin RNA targeting described (Huang et al., 2010). CTSH was purchased from the National RNA Interference Core Facility (Institute of Molecular Biology, Academia Murine models of tumor progression and metastasis Sinica, Taipei, Taiwan). Electroporation was performed in SCID mice were used to determine the invasive ability of the BTX ECM 600 electroporator (BTX Inc., San Diego, CA, CTSH-overexpressing J7 cells. J7-vector and J7-CTSH cells USA). Cells (70–80% confluent) were trypsinized, centrifuged (each 1 Â 107) were injected intravenously. All animals were and suspended in 400 ml of Dulbecco’s modified Eagle’s medium killed in the seventh week after tumor inoculation, whereupon containing 2.5% FBS. Subsequently, 15 mg short hairpin RNA the liver and lungs were removed. All animal experiments were plasmid was added and transferred to 4 mm gap electroporation performed in accordance with United States National In- cuvettes. Cells were plated in 10-cm dishes after pulsing. After stitutes of Health guidelines and the Chang-Gung Institutional 16 h, medium was changed with Dulbecco’s modified Eagle’s Animal Care and Use Committee Guide for Care and Use of medium containing puromycin for stable clone selection. Laboratory Animals. Formalin-fixed and paraffin-embedded tissues from the Electrophoretic mobility shift assay livers or lungs of SCID mice were examined by immunohis- The wild-type CTSH gene probe, which encompassed nucleo- tochemistry using a polyclonal antibody to CTSH (Proteintech tides À2038 to À1966 and À1565 to À1501 of the CTSH Group, Inc., Chicago, IL, USA) following the avidin–biotin promoter, was used for EMSA probes. The normal TRE complex method as described previously (Chen et al., 2008a). sequence 50-AGGGCACAGAGGCTCA-30 of the CTSH Positive staining, indicating cancer cells, appeared as a dark- promoter À2038 to À1966 and mutant sequence 50-GGGGGG brown color showing CTSH immunoreactivity. CAGAGGGGGG-30 were labeled with [a-32P-dCTP] by PCR reaction following gel purification. The normal TRE sequence Human HCC specimens 50-AGGGCATCACTGGGCA-30 of the CTSH promoter With informed consent, 93 patients with HCC diagnosed À1565 to À1501 and mutant sequence 50-GGGGGGTCACG between 2000 and 2003 were selected consecutively for this GGGGG-30 were also labeled with [a-32P-dCTP]. EMSA was study. All samples of HCC tissues with paired adjacent normal performed as described previously (Lin et al., 2000). C4 mouse liver tissues (B1 cm distant from the HCC biopsies) were monoclonal antibody to TRs (sc-740, Santa Cruz Biotechnology obtained during surgical resection from the Chang Gung Inc.), rabbit polyclonal anti-body to RXR (sc-553, Santa Cruz) Memorial Hospital medical research center for western blot

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2068 analysis. HCC and adjacent noncancerous liver tissue samples Conflict of interest were confirmed by histopathology. The study protocol was approved by the Medical Ethics and Human Clinical Trial The authors declare no conflict of interest. Committee at Chang-Gung Memorial Hospital. Acknowledgements Statistical analysis Values are expressed as mean±s.e. of at least three indepen- This work was supported by grants from Chang-Gung dent experiments. Statistical analysis of data was carried out University, Taoyuan, Taiwan (CMRPD 34013, NMRP using Student’s t-test. **Po0.01; *Po0.05. Po0.05 was 140511) and from the National Science Council of the considered statistically significant. Republic of China (NSC 94-2320-B-182-052).

References

Ando S, Sarlis NJ, Krishnan J, Feng X, Refetoff S, Zhang MQ et al. Huang YH, Liao CH, Chen RN, Liao CJ, Lin KH. (2010). Human (2001). Aberrant alternative splicing of thyroid hormone receptor in testicular orphan receptor 4 enhances thyroid hormone receptor a TSH-secreting pituitary tumor is a mechanism for hormone signaling. J Cell Physiol 222: 347–356. resistance. Mol Endocrinol 15: 1529–1538. Huang YH, Tsai MM, Lin KH. (2008). Thyroid hormone dependent Barrera-Hernandez G, Park KS, Dace A, Zhan Q, Cheng SY. (1999). regulation of target genes and their physiological significance. Thyroid hormone-induced cell proliferation in GC cells is mediated Chang Gung Med J 31: 325–334. by changes in G1 cyclin/cyclin-dependent kinase levels and activity. Joyce JA, Hanahan D. (2004). Multiple roles for cysteine cathepsins in Endocrinology 140: 5267–5274. cancer. Cell Cycle 3: 1516–1619. Berdowska I. (2004). Cysteine proteases as disease markers. Clin Chim Kos J, Stabuc B, Schweiger A, Krasovec M, Cimerman N, Kopitar- Acta 342: 41–69. Jerala N et al. (1997). Cathepsins B, H, and L and their inhibitors Chamba A, Neuberger J, Strain A, Hopkins J, Sheppard MC, stefin A and cystatin C in sera of melanoma patients. Clin Cancer Franklyn JA. (1996). Expression and function of thyroid hormone Res 3: 1815–1822. receptor variants in normal and chronically diseased human liver. Kress E, Rezza A, Nadjar J, Samarut J, Plateroti M. (2009). The J Clin Endocrinol Metab 81: 360–367. frizzled-related sFRP2 gene is a target of thyroid hormone receptor Chen RN, Huang YH, Lin YC, Yeh CT, Liang Y, Chen SL et al. alpha1 and activates beta-catenin signaling in mouse intestine. J Biol (2008a). Thyroid hormone promotes cell invasion through activa- Chem 284: 1234–1241. tion of furin expression in human hepatoma cell lines. Endocrinology Kress E, Skah S, Sirakov M, Nadjar J, Gadot N, Scoazec JY et al. 149: 3817–3831. (2010). Cooperation between the thyroid hormone receptor Chen RN, Huang YH, Yeh CT, Liao CH, Lin KH. (2008b). Thyroid TRalpha1 and the WNT pathway in the induction of intestinal hormone receptors suppress pituitary tumor transforming gene 1 tumorigenesis. Gastroenterology 138: 1863–1874. activity in hepatoma. Cancer Res 68: 1697–1706. Lazar MA. (1993). Thyroid hormone receptors: multiple forms, Chiloeches A, Sanchez-Pacheco A, Gil-Araujo B, Aranda A, Lasa M. multiple possibilities. Endocr Rev 14: 184–193. (2008). Thyroid hormone-mediated activation of the ERK/dual Liao CH, Yeh SC, Huang YH, Chen RN, Tsai MM, Chen WJ et al. specificity phosphatase 1 pathway augments the apoptosis of (2010). Positive regulation of spondin 2 by thyroid hormone is GH4C1 cells by down-regulating nuclear factor-kappaB activity. associated with cell migration and invasion. Endocr Relat Cancer 17: Mol Endocrinol 22: 2466–2480. 99–111. Davis FB, Tang HY, Shih A, Keating T, Lansing L, Hercbergs A et al. Liao CS, Tai PJ, Huang YH, Chen RN, Wu SM, Kuo LW et al. (2006). Acting via a cell surface receptor, thyroid hormone is a (2009). Regulation of AKR1B1 by thyroid hormone and its growth factor for glioma cells. Cancer Res 66: 7270–7275. receptors. Mol Cell Endocrinol 307: 109–117. del Re EC, Shuja S, Cai J, Murnane MJ. (2000). Alterations in Lin KH, Shieh HY, Hsu HC. (2000). Negative regulation of the cathepsin H activity and protein patterns in human colorectal antimetastatic gene Nm23-H1 by thyroid hormone receptors. carcinomas. Br J Cancer 82: 1317–1326. Endocrinology 141: 2540–2547. Dickinson DP. (2002). Cysteine peptidases of mammals: their Lin KH, Wu YH, Chen SL. (2001). Impaired interaction of mutant biological roles and potential effects in the oral cavity and other thyroid hormone receptors associated with human hepatocellular tissues in health and disease. Crit Rev Oral Biol Med 13: 238–275. carcinoma with transcriptional coregulators. Endocrinology 142: Gabrijelcic D, Svetic B, Spaic D, Skrk J, Budihna M, Dolenc I et al. 653–662. (1992). Cathepsins B, H and L in human breast carcinoma. Eur J Martinez-Iglesias O, Garcia-Silva S, Regadera J, Aranda A. (2009). Clin Chem Clin Biochem 30: 69–74. Hypothyroidism enhances tumor invasiveness and metastasis Gocheva V, Joyce JA. (2007). Cysteine cathepsins and the cutting edge development. PLoS One 4: e6428. of cancer invasion. Cell Cycle 6: 60–64. Mohamed MM, Sloane BF. (2006). Cysteine cathepsins: multifunc- Hall LC, Salazar EP, Kane SR, Liu N. (2008). Effects of thyroid tional enzymes in cancer. Nat Rev Cancer 6: 764–775. hormones on human breast cancer cell proliferation. J Steroid Nelson D, Ganss R. (2006). Tumor growth or regression: powered by Biochem Mol Biol 109: 57–66. inflammation. J Leukoc Biol 80: 685–690. Hassan MM, Kaseb A, Li D, Patt YZ, Vauthey JN, Thomas MB et al. Palermo C, Joyce JA. (2008). Cysteine cathepsin proteases as (2009). Association between hypothyroidism and hepatocellular pharmacological targets in cancer. Trends Pharmacol Sci 29: 22–28. carcinoma: a case-control study in the United States. Hepatology 49: Plateroti M, Kress E, Mori JI, Samarut J. (2006). Thyroid 1563–1570. hormone receptor alpha1 directly controls transcription of the Huang C, Jacobson K, Schaller MD. (2004). MAP kinases and cell beta-catenin gene in intestinal epithelial cells. Mol Cell Biol 26: migration. J Cell Sci 117: 4619–4628. 3204–3214. Huang YH, Lee CY, Tai PJ, Yen CC, Liao CY, Chen WJ et al. (2006). Puzianowska-Kuznicka M, Krystyniak A, Madej A, Cheng SY, Indirect regulation of human dehydroepiandrosterone sulfotrans- Nauman J. (2002). Functionally impaired TR mutants are ferase family 1A member 2 by thyroid hormones. Endocrinology present in thyroid papillary cancer. J Clin Endocrinol Metab 87: 147: 2481–2489. 1120–1128.

Oncogene Thyroid hormone receptor regulates cathepsin H S-M Wu et al 2069 Rosen MD, Privalsky ML. (2009). Thyroid hormone receptor carcinoma cells by downregulation of the B-cell translocation gene mutations found in renal clear cell carcinomas alter corepressor 2. Prostate 68: 610–619. release and reveal helix 12 as key determinant of corepressor Tsushima H, Ueki A, Matsuoka Y, Mihara H, Hopsu-Havu VK. specificity. Mol Endocrinol 23: 1183–1192. (1991). Characterization of a cathepsin-H-like from a Samuels HH, Stanley F, Casanova J. (1979). Depletion of L-3,5,30- human melanoma cell line. Int J Cancer 48: 726–732. triiodothyronine and L-thyroxine in euthyroid calf serum for use in Turk B, Turk D, Turk V. (2000). Lysosomal cysteine proteases: more cell culture studies of the action of thyroid hormone. Endocrinology than scavengers. Biochim Biophys Acta 1477: 98–111. 105: 80–85. Turk V, Turk B, Turk D. (2001). Lysosomal cysteine proteases: facts Schraufstatter IU, Trieu K, Zhao M, Rose DM, Terkeltaub RA, and opportunities. EMBO J 20: 4629–4633. Burger M. (2003). IL-8-mediated cell migration in endothelial cells Wilson TJ, Singh RK. (2008). Proteases as modulators of tumor- depends on activity and transactivation of the stromal interaction: primary tumors to bone metastases. Biochim epidermal growth factor receptor. J Immunol 171: 6714–6722. Biophys Acta 1785: 85–95. Shih CH, Chen SL, Yen CC, Huang YH, Chen CD, Lee YS et al. Wu SM, Huang YH, Lu YH, Chien LF, Yeh CT, Tsai MM et al. (2004). Thyroid hormone receptor-dependent transcriptional (2010). Thyroid hormone receptor-mediated regulation of the regulation of fibrinogen and coagulation proteins. Endocrinology methionine adenosyltransferase 1 gene is associated with cell 145: 2804–2814. invasion in hepatoma cell lines. Cell Mol Life Sci 67: 1831–1843. Sivaparvathi M, Sawaya R, Gokaslan ZL, Chintala SK, Rao JS. Yang H, Yuan P, Wu V, Tache Y. (1999). Feedback regulation of (1996). Expression and the role of cathepsin H in human glioma thyrotropin-releasing hormone gene expression by thyroid hormone progression and invasion. Cancer Lett 104: 121–126. in the caudal raphe nuclei in rats. Endocrinology 140: 43–49. Suzuki H, Willingham MC, Cheng SY. (2002). Mice with a mutation Yang L, Moses HL. (2008). Transforming growth factor beta: tumor in the thyroid hormone receptor beta gene spontaneously develop suppressor or promoter? Are host immune cells the answer? Cancer thyroid carcinoma: a mouse model of thyroid carcinogenesis. Res 68: 9107–9111. Thyroid 12: 963–969. Yen PM. (2001). Physiological and molecular basis of thyroid Teranishi S, Kimura K, Nishida T. (2009). Role of formation of an hormone action. Physiol Rev 81: 1097–1142. ERK-FAK-paxillin complex in migration of human corneal Yen PM, Ikeda M, Brubaker JH, Forgione M, Sugawara A, Chin epithelial cells during wound closure in vitro. Invest Ophthalmol WW. (1994). Roles of v-erbA homodimers and heterodimers in Vis Sci 50: 5646–5652. mediating dominant negative activity by v-erbA. J Biol Chem 269: Thormeyer D, Baniahmad A. (1999). The v-erbA oncogene (review). 903–909. Int J Mol Med 4: 351–358. Zhu XG, Zhao L, Willingham MC, Cheng SY. (2010). Thyroid Tsui KH, Hsieh WC, Lin MH, Chang PL, Juang HH. (2008). hormone receptors are tumor suppressors in a mouse model of Triiodothyronine modulates cell proliferation of human prostatic metastatic follicular thyroid carcinoma. Oncogene 29: 1909–1919.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene