Overexpression of Tumor Susceptibility Gene TSG101 in Human Papillary Thyroid Carcinomas

Overexpression of tumor susceptibility gene TSG101 in human papillary thyroid carcinomas

Rue-Tsuan Liu1, Chao-Cheng Huang2, Huey-Ling You4, Fong-Fu Chou3, Chih-Chi Andrew Hu4, Fang-Ping Chao4, Ching-Mei Chen2 and Jiin-Tsuey Cheng*,4

1Division of Metabolism, Chang Gung Memorial Hospital, Kaohsiung, Taiwan 833; 2Department of Pathology, Chang Gung Memorial Hospital, Kaohsiung, Taiwan 833; 3Department of Surgery, Chang Gung Memorial Hospital, Kaohsiung, Taiwan 833; 4Department of Biological Sciences, National Sun Yat-Sen University, Kaohsiung, Taiwan 804, Republic of China

Functional inactivation of tumor susceptibility gene Introduction tsg101 leads to cellular transformation and tumorigenesis in mice. While human TSG101 is located in a region It has been shown that functional inactivation of where frequent loss of heterozygosity can be detected in tsg101 in mouse NIH3T3 fibroblasts leads to cellular a variety of cancers, no genomic deletion in TSG101 transformation, and the transformed cells can form gene has been reported, casting a doubt on the role of metastatic tumors in nude mice. The neoplastic TSG101 as a classical tumor suppressor. Some studies transformation and tumourigenesis are reversible by have revealed that TSG101 is a frequent target of restoration of tsg101 function (Li and Cohen, 1996). splicing defects, which correlate with cellular stress and The human homologue TSG101 has been isolated and p53 status. Furthermore, recent reports have identified mapped to chromosome 11p, bands 15.1-15.2 (Li et al., TSG101 as a part of the MDM2/p53 regulatory 1997), a region known to exhibit a high incidence of circuitry, a well-recognized circuitry that upon deregula- loss of heterozygosity in a variety of human malig- tion results in tumorigenesis. Interestingly, nancies (Ali et al., 1987; Baffa et al., 1996; Byrne et al., overexpression of tsg101 from an adventitious promoter 1993; Lothe et al., 1993; Ludwig et al., 1991; Reeve et also leads to neoplastic transformation. On the basis of al., 1989; Shaw and Knowles, 1995; Weitzel et al., this information, we have analysed TSG101 gene 1994), suggesting that TSG101 might be a tumor expression in 20 human papillary thyroid carcinomas suppressor. Although no genomic deletion has been (PTCs) by immunohistochemistry and demonstrated that identified, aberrant transcripts can be found in various the overexpression of TSG101 protein is closely tumors and their matched normal tissues (Carney et associated with human PTCs. Further sequence analysis al., 1998; Chang et al., 1999; Lee and Feinberg, 1997; reveals no mutation in cDNA region encoding steadiness Lin et al., 1998a,b; Wang et al., 2000; Willeke et al., box in these PTC specimens, indicating that the 1998), including thyroid tumors (McIver et al., 2000). upregulation of TSG101 protein is not caused by the Abnormally spliced transcripts of TSG101 have been alteration of this region. In situ hybridization analysis found to very closely correlate with tumor grade and confirms that overexpression of TSG101 also occurs at p53 mutation in breast cancer samples. Stress condi- the transcriptional level. In addition, semi-quantitative tions, such as hypoxia, also induce splicing transcripts RT – PCR and subsequent Southern hybridization verify in primary lymphocytes (Turpin et al., 1999). In that the amounts of TSG101 transcripts are indeed lower addition, induction of p53 by g-irradiation prevents in three normal thyroid tissues than in PTC specimens. the accumulation of aberrant TSG101 transcripts in Here we report the upregulation of TSG101 expression cancer cells. Hence, while abnormal TSG101 RNA in PTC cells, providing the first evidence of the splicing is probably unrelated to oncogenic process, it association of TSG101 overexpression with human does reflect the cellular environment during cancer tumors and suggesting that upregulation of TSG101 development (Moyret-Lalle et al., 2001). steady-state level might play a role in mediating TSG101 gene is expressed ubiquitously in human tumorigenesis of human PTC. organs (Li et al., 1997). Localization of TSG101 has Oncogene (2002) 21, 4830 – 4837. doi:10.1038/sj.onc. been found to be cell cycle-dependent in synchronized 1205612 cell lines, occurring in the nucleus and Golgi complex during interphase, dispersed more generally through- Keywords: TSG101; overexpression; immunohistochem- out the cytoplasm when cells reach late S phase, and istry; in situ hybridization; RT – PCR; human papillary in mitotic spindles and centrosomes during mitosis thyroid carcinoma (Xie et al., 1998; Zhong et al., 1998). Evidence suggests that normal TSG101 expression may be stringently controlled within a narrow range (Feng et *Correspondence: J-T Cheng; E-mail: [email protected] al., 2000) since either deficiency or overexpression of Received 23 August 2001; revised 17 April 2002; accepted 26 April the tsg101 can cause neoplastic formation (Li and 2002 Cohen, 1996). How the perturbation of tsg101 TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4831 expression leads to neoplasm is currently unclear. One recent report on tsg101 knockout mice indicates that tsg101 deficiency results in p53 accumulation, growth arrest and early embryonic lethality (Ruland et al., 2001). Interestingly, TSG101 has been demonstrated to participate in a feedback control loop with MDM2 as a central part where upregulation of TSG101 inhibits ubiquitination and degradation of MDM2, thereby decreasing cellular p53 level, whereas elevation of p53 upregulates MDM2 and subsequently accel- erates the decay of TSG101 (Li et al., 2001). The equilibration between TSG101-dependent stabilization of MDM2 and MDM2-mediated proteolysis of both TSG101 and p53 determines the effect of TSG101 on MDM2/p53 feedback control loop, a well-known control loop that upon deregulation results in tumorigenesis. In this study, we screen 20 human papillary thyroid carcinomas (PTCs) by immunohistochemical staining and in situ hybridization analysis on paraffin-embedded tissues to explore the role of Figure 1 Endogenous TSG101 protein expression in SK-Hep-1 TSG101 in thyroid tumorigenesis. cells. (a) A diagrammatic illustration of the functional domains of TSG101 protein and epitopes of antisera used in this study. The numbers refer to the amino acid residues of TSG101. The in- Results active ubiquitin-conjugating domain (UBC), the proline-rich region (PRR), coiled-coil domain (CC) and steadiness box (SB) Demonstration of the specificity of anti-TSG101 are indicated. (b) Endogenously expressed TSG101 in 30 mgof SK-Hep-1 cell lysate was separated and blotted onto nitrocellu- antiserum lose membrane. It was then detected by sc-6037, a commercial The GST-TSG(1 – 291) and GST-TSG(214 – 390) fusion antiserum against mouse tsg101 protein (lane 1) or rabbit preimmune serum (lane 2); antiserum #654 (lane 3); rabbit antiserum proteins were expressed individually in E. coli strain #654 that was preabsorbed by GST protein (lane 4); and rabbit DH5a and then purified by affinity chromatography on antiserum #654 that was preabsorbed by GST-TSG (1 – 291) glutathione-conjugated agarose beads. Rabbit polyclo- fusion protein (lane 5) nal antisera, #654 and #820 (Figure 1a), were produced using these fusion proteins as immunogens. The specificity of these antisera was verified by Western features are shown in Figures 2 and 3. Intense blot analysis, confirming that these antibodies could cytoplasmic staining was seen in all the neoplastic recognize the endogenous TSG101 protein expressed in epithelial cells of each PTC specimen (Figures 2 and 3; SK-Hep-1, a human hepatocellular carcinoma cell line. Table 1). The non-neoplastic follicular cells adjacent to As shown in Figure 1b, antiserum #654 detected a carcinoma tissue revealed either negative staining or prominent band in the same position as the band that only cytoplasmic dot-staining which localized asymme- could be detected by a commercial TSG101 antibody trically by the side of nucleus, and occasionally staining (sc-6037, Santa Cruz Biotechnology, Santa Cruz, CA, in the nucleus (Figures 2c and 3d). The cytoplasmic USA) (Figure 1b; lanes 1 and 3). The specificity of dot-staining pattern was quite distinct in contrast to antiserum #654 was confirmed as no band in the same the diffuse staining patterns seen in neoplastic cells. position could be detected by preimmune serum and The non-specific background staining was greatly GST-TSG(1 – 291) fusion protein-absorbed antiserum reduced, whereas the intensity of the immunoreactivity (Figure 1b; lanes 2 and 5), while the particular protein did not undergo a significant change while replacing band at the same position could still be detected by the antiserum #654 with GST-absorbed antiserum GST protein-absorbed antiserum #654 (Figure 1b, lane (Figures 2b and 3b). However, immunoreactivity did 4). Also, the specificity of antiserum #654 was further disappear when preimmune serum was used instead of demonstrated by the fact that the immune serum, not antiserum #654 (Figures 2d and 3c). To further confirm the preimmune serum, detected endogenous TSG101 the specificity of antiserum #654, rabbit antiserum protein in SK-Hep-1 cells by immunocytochemical #820 against carboxyl terminus (amino acid residues staining (data not shown). Similar results were 214 – 390) of TSG101 protein was used to immunostain obtained with antiserum #820 (data not shown). the same specimens. Both antisera essentially gave similar staining patterns (data not shown). Detection of increased TSG101 expression in PTC specimens by immunohistochemistry Overexpression of TSG101 protein is not caused by the mutation of steadiness box Immunohistochemical analyses of 20 PTC specimens were then performed using antiserum #654. Two The mutation at C-terminal steadiness box of TSG101 representative PTC specimens with different histologic protein has been shown to cause the accumulation of

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4832

Figure 2 Immunohistochemical and in situ hybridization analyses of a classical PTC specimen. A classical PTC specimen was stained with (a) rabbit antiserum #654, the neoplastic epithelial cells (left side) revealing intense intracytoplasmic immunoreactivity; (b) rabbit antiserum #654, which was absorbed by GST protein; (c) a higher magnification of b showing negative staining of non- neoplastic follicular cells (right side); (d) preimmune serum; (e) in situ hybridization of same PTC specimen using TSG101 antisense riboprobe; (f) a higher magnification of e (Magnifications: a, b, d and e,406; c and f, 2006)

this protein at several times the wild-type protein level (Figures 2e and 3e). The speciﬁcity of hybridization (Feng et al., 2000). To detect the presence of mutation, ensured that the control experiment using sense the cDNA encompassing this region was ampliﬁed and riboprobe gave no signal on the section of the same sequenced. As summarized in Table 1, no mutation in specimen (Figure 3f). The results of immunohistochem- steadiness box could be detected in the PTCs analysed. ical staining and in situ hybridization are summarized in Table 1. Most of the cases gave consistent and unambiguous results using both assays, whereas the in Upregulation of TSG101 RNA is detected in PTC situ hybridization signals were weak in specimens 4 and specimens by in situ hybridization and semi-quantitative 7 and focal in specimen 16. Two of the samples, no. 2 RT – PCR and 9, showed no signal. The arbitrary hybridization The overexpression of TSG101 transcripts in PTC cells patterns were probably caused by poor RNA quality in of these samples was further demonstrated by in situ these samples since there was no signal in their hybridization with TSG101 antisense riboprobe adjacent normal follicular cells. To further verify the

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4833

Figure 3 TSG101 gene expression in a PTC specimen of solid/trabecular type. Immunohistochemical staining of a solid/trabecular type PTC specimen using (a) rabbit antiserum #654; (b) rabbit antiserum #654, which was absorbed by GST protein; (c) preimmune serum; (d) a higher magniﬁcation of b showing non-neoplastic follicular cells. Arrows indicated cytoplasmic dot-staining. In situ hybridization of same PTC specimen using (e) TSG101 antisense riboprobe, (f) TSG101 sense riboprobe (Magniﬁcations: a, b, c, d, e and f, 2006)

relative amount of TSG101 transcripts in tumor and The value obtained by densitometry scanning of normal thyroid tissues, semi-quantitative RT – PCR TSG101 hybridization signal was normalized to the and subsequent Southern hybridizations were value obtained from the signal of GAPDH internal performed. Limited by the availability of fresh tissue control in the corresponding sample. As indicated in samples, only three normal thyroid tissues and three Figure 4b, the relative amount of TSG101 expression PTCs samples were analysed. The PCR conditions in the tumor sample was more than twofold higher verified that the reactions were in a linear range and than that found in normal samples. not in a plateau, and were, therefore, suitable for comparing mRNA amounts. To compare the amounts of the amplified TSG101 products, we performed a Discussion quantification of the hybridization signals (Figure 4a) with a densitometric scanner (GS-700 Imaging Densit- In this report, we demonstrate overexpression of ometer, BIO-RAD Laboratories, Hercules, CA, USA). TSG101 gene in human PTCs by both immunohisto-

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4834 Table 1 TSG101 expression and cDNA sequencing analysis TSG101 overexpressiona Mutationb Case No. IHC ISH Steadiness box

1+ +7 2+ 7N 7 3+ +7 4+ +W 7 5+ +7 6+ +7 7+ +W 7 8+ +7 9+ 7N 7 10 + + 7 11 + + 7 12 + + ND 13 + + ND 14 + + 7 15 + + 7 16 + +F 7 17 + + 7 18 + + 7 19 + + 7 20 + + ND

aTSG101 expression was determined by immunohistochemistry (IHC) and in situ hybridization (ISH). Overexpression was judged by comparing the staining intensity of PTC tumor cells and the adjacent normal follicular cells (The representative cases of over- Figure 4 Semi-quantitative RT – PCR analysis of TSG101 ex- expression + were shown in Figures 2 and 3). + indicated strong pression in normal and PTC specimens. (a) Autoradiography of staining in tumor cells while negative or dot-staining in normal Southern analysis. RT – PCR products were blotted and hybri- W follicular cells. + indicated weak staining in tumor cells while dized with [32P]oligonucleotide probes speciﬁc for TSG101 and F negative in normal follicular cells. + indicated focal staining in GADPH. Normal thyroid tissues: lanes 1 – 3; PTC tissues: lanes N tumor cells and negative staining in normal cells. 7 indicated 4–6. (b) A bar graph illustrates the density ratio of TSG101/ b negative staining in both tumor and normal cells. TSG101 cDNA GAPDH hybridization signals of three normal thyroid (bars 1 – nucleotides 803 – 1263 encompassing steadiness box was ampliﬁed by 3) and PTC specimens (bars 4 – 6) that were calculated by densi- PCR using TSG-803 and TSG-R primers, and the sequence status tometric scanning of autoradiogram shown in a was determined by automatic sequencing. 7: no mutation, ND: not determined

performed nested RT – PCR on these PTC specimens following the protocol of Li et al. (1997). The chemistry and in situ hybridization. The role of truncated forms of TSG101 transcripts were detected TSG101 as an inactive homologue of ubiquitin- in a small number of samples (two out of 20 samples, conjugated enzymes (Koonin and Abagyan, 1997; data not shown). Hence, it is unlikely that the aberrant Ponting et al., 1997) has recently been demonstrated protein translated from the truncated TSG101 RNA in regulating the steady-state level of MDM2 degrada- might account for the upregulated-TSG101 protein tion, and this might account for the mechanism detected by immunohistochemical staining. However, it operated by TSG101 in the cell cycle control. remains possible that mutation in another part of Upregulation of MDM2 in cells that overexpress TSG101, which we have not sequenced might account TSG101 conversely decreases the amount of p53 (Li for the accumulation of TSG101 in PTC cells. The et al., 2001). Hence, overexpression of TSG101 might results of our in situ hybridization analysis confirm that play an oncogenic role by inactivating p53 through TSG101 was up-regulated in the RNA level, indicating MDM2 upregulation. Our findings provide the first that there might be differential RNA stability or evidence for the association of TSG101 overexpression promoter activity in normal thyroid cells and PTC with human PTCs. Since other reports have shown that cells. Further experiments are needed to clarify this MDM2 protein is upregulated in a high proportion of issue. PTC specimens (Jennings et al., 1995; Knauf et al In this study, while TSG101 expression was Abstract of Post No. P2-528. The Endocrine Society’s primarily localized in the cytoplasm of PTC tumor 83rd annual meeting, 2001), the same mechanism might cell, either negative or only dot-staining was observed likely be responsible for the TSG101 overexpression in in the cytoplasm of non-neoplastic follicular cells. mediating human PTC. Occasionally, the staining could be seen in the nucleus. The steadiness box region of TSG101 was shown to The ability of the TSG101 protein to shuttle between control the steady-state level of TSG101. Mutation in different locations during the cell cycle is consistent this region caused the accumulation of TSG101 protein with the notion that this gene product may function in within the cells. However, no mutation in the both nucleus and cytoplasm (Xie et al., 1998; Zhong et steadiness box region in our sequence analysis of 17 al., 1998). The diffused cytoplasmic localization of PTC specimens overrules this possibility. We have TSG101 protein was prominent in PTC cells in

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4835 contrast to the localized dot pattern in normal Fusion protein and antiserum production follicular cells (Figures 2c and 3d), reminiscent of the dispersed TSG101 staining pattern in cell treated with Transformed DH5a bacteria containing recombinant plasmid brefeldin A that interferes with the protein transporta- were grown in LB broth containing 200 mg/ml ampicillin at tion in the membrane compartments (Xie et al., 1998). 378C until the OD600 reached 0.5. Fusion protein expression was then induced after the addition of isopropyl-1-thio-b-D- It remains to be determined whether the diffused galactopyranoside (final concentration of 0.5 mM). Fusion TSG101 localization pattern we observed is merely protein was then purified into homogeneity by affinity the reflection of the cell cycle status of carcinoma cells chromatography on glutathione-conjugated agarose beads or due to a defect in the process of subcellular (Sigma, St. Louis, MO, USA) as previously described (Hsu et localization in neoplastic cells. al., 1998). Purified fusion protein was concentrated by using a TSG101 has been shown to repress ligand-depen- CentriconTM 30 concentrator (Amicon Inc., Beverly, MA, dent transcriptional activation by receptors such as USA) and resuspended in 200 mM phosphate buffer, pH 7.0. androgen receptor, estrogen receptor, glucocorticoid Antiserum was prepared as described in a previous receptor, vitamin D receptor, retinoid acid receptor publication (Hsu et al., 1998). For preparation of GST- and thyroid hormone receptor (Sun et al., 1999; absorbed antiserum, immunized rabbit serum (antiserum #654) was preabsorbed three times with GST protein bound Watanabe et al., 1998). In the case of glucocorticoid nitrocellulose membranes. receptor, the repression was mediated through the direct binding of TSG101 to the GR AF-1 transcriptional activation domain of glucocorticoid receptor Cell culture (Hittelman et al., 1999). Recent reports imply that Human hepatoma cell line SK-Hep-1 was routinely main- tained in Dulbecco’s modified Eagle Medium supplemented TSG101 is essential for normal membrane trafficking 1 and regulation of receptor recycling (Babst et al., with 10% fetal calf serum (Hyclone Laboratories, Inc., 2000; Bishop and Woodman, 2000). Forced expression Logan, UT, USA), and cultured at 378C in a 5% CO2 of TSG101 could suppress receptor function indirectly incubator. by interfering with the regulation of membrane trafficking and receptor recycling. Therefore, it is Western blot analysis tempting to speculate whether aberrant receptor Whole cell lysate and prestained protein molecular weight recycling control due to overexpression of TSG101 markers were electrophoresed on 12.5% SDS – PAGE gel and might resolve in abnormal cell growth in PTC cells. transferred onto nitrocellulose membrane using a semidried On the other hand, it has been shown that TSG101 blotting apparatus (Biometra, Go¨ ttingen, Germany). The could suppress transcription by direct association with membrane was blocked with blocking buffer (5% skim milk DMAP1 and DNMT1 containing transcription repres- in TBS-T; 0.8% NaCl; 0.02% KCl; 25 mM Tris-HCl, pH 7.4; sion complex (Rountree et al., 2000). It is possible 0.05% Tween 20) for 1 h. Primary antibody (1 : 2000 dilution that the deranged steady-state of TSG101 protein level in blocking buffer containing 1% NP-40) was added and interferes with the regulation normally operated to incubated at 48C overnight. The membrane was again washed four times with TBS-T and subsequently incubated with control the growth of thyroid cells through nuclear horseradish peroxidase-conjugated secondary antibody receptors. In addition, the coiled-coil domain at (1 : 4000 dilution in blocking buffer containing 1% NP-40). carboxyl terminus of TSG101 protein has been The membrane was then further washed four times with TBS- reported to interact with the cytoplasmic phosphopro- T. The immunoblot was then developed with SuperSignalTM tein, stathmin/oncoprotein 18 (Maucuer et al., 1995), CL-HRP chemiluminescent substrate according to the the latter possibly playing a role in microtubule protocol recommended by the manufacturer (Pierce, Rock- dynamics as well as cell growth and differentiation. ford, IL, USA). A further examination of the relation between TSG101 overexpression and stathmin/oncoprotein 18 Tissue samples in mediating the formation of PTC would be Twenty clinical specimens of PTCs were collected from the interesting. patients during surgery for thyroidectomy, and fixed in 10% buffered formalin and embedded in paraffin. Normal human thyroid tissues were obtained with informed consent from patients who had undergone surgical removal of parathyroid Materials and methods adenoma. Tumor tissues used for RT – PCR analysis were snap frozen in liquid nitrogen when the diagnosis of PTC was Plasmid construction confirmed. The IMAGE Consortium cDNA clones, No. 547885 and 684503, containing cDNAs (nucleotides 88 – 964 and 734 – Immunohistochemistry 1492, respectively) of human TSG101 were purchased from Research Genetics, Inc., (Huntsville, AL, USA). The cDNA The paraffin-embedded tissue blocks were sectioned in 5 mm inserts were cloned individually into appropriate pGEX4T slices and placed on slides coated with poly-L-lysine. The vectors (Pharmacia, Uppsala, Sweden) for the expression of sections were then dewaxed and incubated in a 378C oven. TSG101 protein fused in frame with glutathione S-transferase After overnight incubation, the slides were rinsed in (GST). Recombinant plasmids were then sequenced using phosphate buffered saline (PBS) and blocked for 5 min with GST gene specific primer to confirm the in-frame recombi- 3% hydrogen peroxide to deprive the endogenous peroxidase nant sequences. activity. After antigen retrieval, primary antibodies (1 : 100

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4836 dilution in PBS) were applied onto the specimens and TBS-T wash. Finally, the slides were incubated with DAB for incubated for 30 min. The sections were then washed with 5 min to develop the signal. Gill’s hematoxylin (Merck, PBS. Biotinylated secondary antibody and streptavidin- Darmstadt, Germany) was used for counterstaining before horseradish peroxidase were applied to the specimens in mounting. succession according to the manufacturer’s instructions (LSAB1 Plus Kit, DAKO, Denmark). After extensive RT – PCR and Southern blotting washing, peroxidase substrate diaminobenzidine (Sigma) was added, and the specimens were incubated for 5 min. Total RNA was extracted from clinical specimens using Thereafter, the sections were counter-stained with Mayer’s Trizol1 reagent (Life Technologies, Gaithersburg, MD, hematoxylin for 2 min, washed in running water for 5 min, USA) according to manufacturer’s instructions. First-strand dehydrated with serial ethyl alcohol and cleared with xylene. cDNA was synthesized from 100 nanogram of total RNA Finally, the slides were mounted with mounting medium. The using the AdvantageTM RT-for-PCR kit (Clontech Labora- results were observed under the microscope. tories, Palo Alto, CA, USA) in a 25 ml of reaction volume. The reverse transcription product was diluted with DEPC- treated water into a final volume of 100 ml. Then, 3 ml of the In situ hybridization first strand cDNA was mixed with 5 mlof106 PCR buffer, The plasmid pBluescript SK(7) that contained TSG101 4 ml of 2.5 mM dNTP, 1.4 mlof10mM TSG101 specific cDNA encoding amino acid residues 1 – 291 was used for the forward primer (TSG-F, 5’-AGCTCAAGAAAATGGTGTC- preparation of riboprobes. To generate anti-sense probe, CAAGTAC-3’), 1.4 mlof10mM TSG101 reverse primer plasmid DNA was linearized by EcoRI digestion and (TSG-R, 5’-CAGCTGGTATCAGAGAAGTCAGTAG-3’) transcribed by T7 RNA polymerase. For the negative control and 1 mlofTaq polymerase (TaKaRa Biomedicals, Shiga, sense probe, plasmid DNA was linearized with KpnI digestion Japan). The mixture was incubated at 958C for 5 min. Then, and then transcribed with T3 RNA polymerase using a DIG the PCR reaction was performed with denaturation at 968C RNA labeling kit (Roche, Mannheim, Germany) according for 45 s; annealing at 658C for 1 min; and extension at 728C to the manufacturer’s instructions. To perform in situ for 1 min. The complete amplification procedure was carried hybridization study, 4 mm consecutive sections of paraffin- out for 35 cycles followed by further incubation at 728C for embedded sample were processed as for immunohistochem- 10 min. In order to control for errors in input of cDNA used istry. Rehydrated sections were then rinsed twice in in PCR reaction, amplification of the ubiquitous glyceralde- diethylpyrocarbonate (DEPC)-treated water followed by hyde-3-phosphate dehydrogenase (GAPDH) cDNA was DEPC-treated PBS, and digested with 20 mg/ml proteinase performed in parallel using GAPDH specific primers K (DAKO) at 378C for 45 min. Proteinase K was then (forward: 5’-ACCACAGTCCATGCCATCAC-3’; reverse: 5’- removed by washing them two times in DEPC-treated PBS. TCCACCACCCTGTTGCTGTA-3’). To acetylate the sections, the slides were immersed in 0.1 M Aliquots of PCR product were size-fractionated by agarose triethanolamine buffer, pH 8.0, containing 0.25% acetic gel electrophoresis, transferred onto Hybond-N membrane anhydride for 10 min at room temperature and then rinsed (Amersham, Chicago, IL, USA) and prehybridized for 1 h at with 56 sodium chloride/sodium citrate (SSC) solution for 428C in ExpressHybTM hybridization solution (Clontech) and 5 min. The sections were then prehybridized with 50% then hybridized with 16107 c.p.m. of 32P-end-labeled internal deionized formamide in 56 SSC at 428C. After 1 h probes for overnight at 378C. The membrane was then washed incubation, the prehybridization solution was removed and twice with 26 SSC, 0.5% SDS at 378C for 10 min. Membrane 30 ml heat-denatured riboprobe (3 ng/ml in mRNA in situ was then autoradiographed on Kodak film XAR at 7708C. hybridization solution) was added on each section, which was The nucleotide sequences of the probes were TSG-803, 5’- then covered with coverslip and hybridized at 458C overnight TGAGATGGCGGATGAAGG-3’ for TSG101,5’-GCCACC- in a humid chamber. After hybridization, the coverslips were CAGAAGACTGTGGATGGCCCCTCC-3’ for GAPDH. removed by soaking in 26 SSC. The sections were then incubated with 20 mg/ml RNase A to digest non-hybridized probe. Stringent washing was performed by incubating the slides stepwise in 16, 0.26, 0.056 and 0.0256 SSC, Abbreviations consecutively. The endogenous peroxidase in the sections TSG101, tumor susceptibility gene 101; PTC, papillary was quenched with 3% hydrogen peroxide at room thyroid carcinoma; IHC, immunohistochemistry; ISH, in situ temperature for 5 min. Thereafter, the slides were immersed hybridization; GST, glutathione S-transferase in TBS-T followed by blocking with rabbit serum (DAKO). The sections were then incubated with 1 : 100 diluted rabbit HRP-anti-DIG antibody (DAKO) for 1 h at room tempera- Acknowledgements ture. The DAKO GenPoint kit for catalyzed reporter We thank Ms Shang-Yun Liu for the assistance in deposition was then used to amplify the signal. Briefly, the immunohistochemical analysis. This work is supported in slides were washed with TBS-T, incubated with biotinyl- part by National Science Council grants NSC 88-2314-B- tyramide solution for 15 min, washed with TBS-T, and 182A-046 (to R-T Liu), NSC 88-2314-B-110-007 and incubated with streptavidin-HRP for 15 min followed by VGHNSU88-07 (to J-T Cheng).

References

Ali IU, Lidereau R, Theillet C and Callahan R. (1987). Baffa R, Negrini M, Mandes B, Rugge M, Ranzani GN, Science, 238, 185 – 188. Hirohashi S and Croce CM. (1996). Cancer Res., 56, 268 – Babst M, Odorizzi G, Estepa EJ and Emr SD. (2000). Trafﬁc, 272. 1, 248 – 258.

Oncogene TSG101 expression in human papillary thyroid carcinomas R-T Liu et al 4837 Bishop N and Woodman P. (2000). J. Biol. Chem., 276, McIver B, Grebe SK, Wang L, Hay ID, Yokomizo A, Liu W, 11735 – 11742. Goellner JR, Grant CS, Smith DI and Eberhardt NL. Byrne JA, Simms LA, Little MH, Algar EM and Smith PJ. (2000). Clin. Endocrinol. (Oxf), 52, 749 – 757. (1993). Genes Chrom. Cancer, 8, 104 – 111. Moyret-Lalle C, Duriez C, Van Kerckhove J, Gilbert C, Carney ME, Maxwell GL, Lancaster JM, Gumbs C, Marks Wang Q and Puisieux A. (2001). Cancer Res., 61, 486 – 488. J, Berchuck A and Futreal PA. (1998). J. Soc. Gynecol. Ponting CP, Cai YD and Bork P. (1997). J. Mol. Med., 75, Investig., 5, 281 – 285. 467 – 489. ChangJG,SuTH,WeiHJ,WangJC,ChenYJ,ChangCP Reeve AE, Sih SA, Raizis AM and Feinberg AP. (1989). Mol. and Jeng CJ. (1999). Br.J.Cancer,79, 445 – 450. Cell. Biol., 9, 1799 – 1803. Feng GH, Lih CJ and Cohen SN. (2000). Cancer Res., 60, Rountree MR, Bachman KE and Baylin SB. (2000). Nat. 1736 – 1741. Genet., 3, 269 – 277. Hittelman AB, Burakov D, In˜ iguez-Lluhı´ JA, Freedman LP Ruland J, Sirard C, Elia A, MacPherson D, Wakeham A, Li and Garabedian MJ. (1999). EMBO J., 18, 5380 – 5388. L, de la Pompa JL, Cohen SN and Mak TW. (2001). Proc. Hsu SF, Chang AC, Wang JK and Cheng JT. (1998). Natl. Acad. Sci. USA, 98, 1859 – 1864. Hybridoma, 17, 251 – 255. Shaw ME and Knowles MA. (1995). Genes Chrom. Cancer, Jennings T, Bratslavsky G, Gerasimov G, Troshina K, 13, 1–8. Bronstein M, Dedov I, Alexandrova G and Figge J. (1995). Sun Z, Pan J, Hope WX, Cohen SN and Balk SP. (1999). Exp. Mol. Pathol., 62, 199 – 206. Cancer, 86, 689 – 696. Koonin EV and Abagyan RA. (1997). Nat. Genet., 16, 330 – Turpin E, Dalle B, de Roquancourt A, Plassa LF, Marty M, 331. Janin A, Beuzard Y and de The H. (1999). Oncogene, 18, Lee MP and Feinberg AP. (1997). Cancer Res., 57, 3131 – 7834 – 7837. 3134. WangNM,ChangJG,LiuTC,LinSF,PengCT,TsaiFJand Li L and Cohen SN. (1996). Cell, 85, 319 – 329. Tsai CH. (2000). Int. J. Oncol., 16, 75 – 80. Li L, Li X, Francke U and Cohen SN. (1997). Cell, 88, 143 – Watanabe M, Yanagi Y, Masuhiro Y, Yano T, Yoshikawa 154. H, Yanagisawa J and Kato S. (1998). Biochem. Biophys. Li L, Liao J, Ruland J, Mak TW and Cohen SN. (2001). Res. Commun., 245, 900 – 905. Proc.Natl.Acad.Sci.USA,98, 1619 – 1624. WeitzelJN,PatelJ,SmithDM,GoodmanA,SafaiiHand Lin PM, Liu TC, Chang JG, Chen TP and Lin SF. (1998a). Ball HG. (1994). Gynecol. Oncol., 55, 245 – 252. Br. J. Haematol., 102, 753 – 758. Willeke F, Ridder R, Bork P, Klaes R, Mechtersheimer G, Lin SY, Chen YJ and Chang JG. (1998b). J. Gastroenterol. Schwarzbach M, Zimmer D, Kloor M, Lehnert T, Hepatol., 13, 1111 – 1114. Herfarth C and von Knebel Doeberitz M. (1998). Mol. LotheRA,HastieN,HeimdalK,FossaSD,StenwigAEand Carcinog., 23, 195 – 200. Borresen AL. (1993). Genes Chrom. Cancer, 7, 96 – 101. Xie W, Li L and Cohen SN. (1998). Proc. Natl. Acad. Sci. Ludwig CU, Raeﬂe G, Dalquen P, Stulz P, Stahel R and USA, 95, 1595 – 1600. Obrecht JP. (1991). Int. J. Cancer, 49, 661 – 665. Zhong Q, Chen Y, Jones D and Lee WH. (1998). Cancer Maucuer A, Camonis JH and Sobel A. (1995). Proc. Natl. Res., 58, 2699 – 2702. Acad. Sci. USA, 92, 3100 – 3104.

Oncogene