Endocrine-Related (2011) 18 743–757

Evidence of -induced in thyroid

Maria Grazia Vizioli, Patricia A Possik1, Eva Tarantino2, Katrin Meissl1, Maria Grazia Borrello, Claudia Miranda, Maria Chiara Anania, Sonia Pagliardini, Ettore Seregni3, Marco A Pierotti4, Silvana Pilotti2, Daniel S Peeper 1 and Angela Greco

Molecular Mechanisms Unit, Department of Experimental Oncology and Molecular Medicine, IRCCS Foundation - Istituto Nazionale dei Tumori, Via G. Amadeo, 42 20133 Milan, Italy 1Division of Molecular Genetics, The Netherlands Cancer Institute, Plesmanlaan, 121 1066CX Amsterdam, The Netherlands 2Department of Pathology, 3Nuclear Medicine Division, Department of Diagnostic Imaging and Radiotherapy and 4Scientific Directorate, IRCCS Foundation - Istituto Nazionale dei Tumori, Via G. Venezian, 1 20133 Milan, Italy (Correspondence should be addressed to A Greco; Email: [email protected])

Abstract Oncogene-induced senescence (OIS) is a growth arrest triggered by the enforced expression of cancer-promoting genes and acts as a barrier against malignant transformation in vivo. In this study, by a combination of in vitro and in vivo approaches, we investigate the role of OIS in tumours originating from the thyroid epithelium. We found that expression of different thyroid tumour-associated in primary human thyrocytes triggers senescence, as demon- strated by the presence of OIS hallmarks: changes in morphology, accumulation of SA-b-Gal and senescence-associated heterochromatic foci, and upregulation of transcription of the cyclin- dependent kinase inhibitors p16INK4a and p21CIP1. Furthermore, immunohistochemical analysis of a panel of thyroid tumours characterised by different aggressiveness showed that the expression of OIS markers such as p16INK4a, p21CIP1 and IGFBP7 is upregulated at early stages, and lost during thyroid tumour progression. Taken together, our results suggest a role of OIS in thyroid carcinogenesis. Endocrine-Related Cancer (2011) 18 743–757

Introduction a tumour measuring %1 cm in diameter and is Thyroid tumours represent the most common endo- considered the early stage of PTC. It has an indolent crine malignancy; the majority originate from thyroid course and very favourable prognosis despite the follicular cells and include differentiated follicular presence of multifocality within the thyroid and the thyroid carcinoma (FTC), papillary thyroid carcinoma synchronous nodal loco-regional spread (Sakorafas (PTC), poorly differentiated thyroid carcinomas et al. 2005). The molecular mechanisms underlying (PDTC) and undifferentiated anaplastic thyroid carci- PTC pathogenesis have been partially elucidated. nomas (ATC) (Sheils 2005). PTCs are associated with genetic alterations leading PTC accounts for w80% of all thyroid ; to the activation of the ERK1/2 signal pathway, generally it is a slow-growing tumour type with a good including point of the BRAF and RAS prognosis, although rare, aggressive forms with local genes, and rearrangements of RET and NTRK1 tyrosine invasion or distant metastasis can occur. In addition to kinase receptors, producing RET/PTC and TRK classical PTC, also known as not otherwise specified oncogenes respectively (Greco et al. 2009). Of note, (NOS), several histological variants, differing in BRAF mutations have been reported in a consistent morphological pattern as well as in prognosis, have fraction of PTMC (Lupi et al. 2007), thus supporting been described, including papillary thyroid micro- the notion that it represents an earlier stage of disease carcinomas (PTMC), solid variant (SV), PDTC and tall which eventually evolves into PTC. Nevertheless, the cell variant (TCV) (Sheils 2005). PTMC is defined as molecular mechanisms driving the evolution of PTMC

Endocrine-Related Cancer (2011) 18 743–757 Downloaded from Bioscientifica.comDOI: 10.1530/ERC-11-0240 at 09/30/2021 03:39:20AM 1351–0088/11/018–743 q 2011 Society for Endocrinology Printed in Great Britain Online version via http://www.endocrinology-journals.orgvia free access M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours in overt carcinoma or, vice versa, keeping PTMC dermal neurofibromas, murine lung adenomas, human proliferation under control, as well as the mechanisms and murine prostatic adenomas, murine pancreatic responsible for the most aggressive forms of PTC, intraductal neoplasias and murine lymphomas (Braig remain to be elucidated. et al. 2005, Chen et al. 2005, Collado et al. 2005, When expressed in primary cells, activated onco- Courtois-Cox et al. 2006, Dankort et al. 2007, Ha et al. genes can block cellular proliferation by inducing 2007). The increasing evidence of OIS involvement in senescence or (Campisi 2005, Mooi & human cancer suggests the possibility of new thera- Peeper 2006). Both mechanisms are thought to play peutic approaches based on the functional restoration important roles in suppressing tumourigenesis in vivo, of OIS in tumour cells. as they prevent the proliferation of cells at risk of In this study, we investigate the occurrence of OIS neoplastic transformation (Michaloglou et al. 2005). in thyroid carcinogenesis with in vitro and in vivo The factors that direct the choice of cancer cells into studies. We found that PTC-associated oncogenes one or the other outcome have remained unclear trigger OIS in primary human thyrocytes, as demon- (D’Adda di Fagagna 2008); most likely they are related strated by changes in cell morphology and the to the oncogenic trigger and to the balance between presence of specific OIS markers. To assess whether proapoptotic and antiapoptotic factors. this mechanism represents a barrier to thyroid cancer Oncogene-induced senescence (OIS) was first in vivo, we performed immunohistochemical analysis identified in vitro and describes a stable cell-cycle in a panel of thyroid tumours characterised by different arrest that is triggered by the aberrant proliferative aggressiveness. OIS markers were found to be signals of oncogenes (Serrano et al. 1997). Senescent upregulated in PTMCs, whereas they were pro- cells are characterised by a flat and large morphology gressively lost in PTCs and ATC. Overall, our data with vacuoles (Denoyelle et al. 2006), an increase support a role for OIS in thyroid carcinogenesis. in senescence-associated b-galactosidase activity (SA-b-Gal; Dimri et al. 1995), and alterations in structure known as senescence-associated Materials and methods heterochromatic foci (SAHF), which may repress the Cell culture expression of proliferation-promoting genes (Narita Normal thyroid samples were obtained from patients et al. 2003, Adams 2007). Moreover, senescent cells undergoing surgery at IRCCS Foundation Istituto undergo dramatic changes in pattern, Nazionale dei Tumori (Milan, Italy). All patients including, activation of , increased levels of the gave their written informed consent, and the study cyclin-dependent kinase inhibitors, p16INK4a and was approved by the independent ethics committee p21CIP1 (Beausejour et al. 2003, Lowe et al. 2004) of IRCCS Foundation Istituto Nazionale dei Tumori. and secretion of multiple factors such as extracellular Primary thyrocyte cultures were established and proteases, matrix components, growth factors, pro- maintained in nutrient mixture Ham’s F12 medium inflammatory and chemokines (Kortlever (custom made by Invitrogen) containing 5% calf serum et al. 2006, Acosta et al. 2008, Kuilman et al. 2008, and bovine hypothalamus and pituitary extracts, as Wajapeyee et al. 2008). described previously (Curcio et al. 1994). Thyrocytes OIS is a mechanism of increasing importance in expressing RET/PTC1 oncogene, obtained from infec- tumours (Kuilman et al. 2010), as several reports have tion of primary thyrocytes with RET/PTC1 retroviral demonstrated that it occurs in several diverse pre- vector, have been previously described (Borrello et al. cancerous tissues from human and mice, and acts as 2005). HEK293T cells were maintained in DMEM a barrier to invasiveness and tumourigenesis. The first F9 medium supplemented with 10% FCS. direct evidence of cellular senescence in a growth arrested human neoplasm was reported in melanocytic nevi, premalignant lesions frequently carrying the Lentivirus production and transduction BRAFV600E which rarely progress to mela- HEK293T cells (4!106) were transfected with 8 mgof noma (Michaloglou et al. 2005). The blunted pro- BRAFV600E lentiviral construct (HIV-CS-CG-Blast- liferative activity in nevi was associated with features BRAFV600E) and 3 mg of each of the helper plasmids of cellular senescence, including SA-b-Gal activity pMDLglpRRE, pHCMV-G and pRSVrev by calcium and elevated levels of p16INK4a. On the contrary, in phosphate in complete medium containing 25 mM melanoma, which is characterised by accelerated chloroquine. After 6–12 h of incubation cells were proliferation, cell senescence was extinguished. Later washed and refed with complete medium. Supernatant on, senescence markers were identified in human containing lentiviral particles was collected and frozen

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 744 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757

24 h after transfection, and subsequently added to SA-b-Gal assay thyrocytes for 6 h in the presence of 0.8 mg/ml Adherent cells were analysed for SA-b-Gal production polybrene (Sigma–Aldrich), followed by blasticidin using a Senescence-beta-Galactosidase Staining Kit selection (5 mg/ml) after 24 h. Three days later, cells (Cell Signaling, Danvers, MA, USA) following the were harvested and plated at different densities for manufacturer’s instructions. For each sample at least further analysis. The efficiency of transduction was 200 cells were scored. monitored in parallel infections using a virus expres- sing enhanced green fluorescent protein (eGFP). SAHFs formation For the analysis of SAHFs, transfected/transduced p16INK4a shRNA silencing cells growing on glass cover slips were fixed in 4% p16INK4a knockdown constructs were created in KH1 paraformaldehyde/2% sucrose, permeabilised with lentiviral vectors, carrying the GFP reporter gene, with 0.1% Triton, and stained with Prolong Gold antifade the following sense shRNA sequences: shp16.5 reagent with DAPI (Molecular Probes, Invitrogen). (TGCCCCCGGGGGAGACCCAAC) and shp16.13 Stained nuclei were observed under a fluorescent (GCCGACCCCGCCACTCTCA) and used for thyr- microscope (Nikon Eclipse E1000). For each sample ocytes transduction. Transduction efficiency was moni- at least 100 cells were scored. tored by GFP expression. Six days later, cells were re-transduced with lentiviruses encoding BRAFV600E Western blotting analysis V600E (HIV-CSCG-BRAF ) and maintained under blas- Proteins were extracted in RIPA modified buffer ticidin selection for an additional 6 days. (20 mM Tris–HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1% NonidetP-40) supplemented with Complete Nucleofection Mini EDTA-free protease Inhibitor Cocktail (Roche), 1mMNaVO and 1 mM PMSF. Protein samples were ! 6 3 4 Human primary thyrocytes (1 10 ) were resuspended quantified by Bradford’s assay with BIO-RAD Protein in 110 ml human keratinocyte nucleofector solution Assay (Bio-Rad). Supernatant obtained by incubating (Amaxa, Lonza AG Basel, Switzerland) together with cells in serum-free medium for 24 h were concentrated 1 mg of plasmid DNA, transferred to the cuvette and by centrifugation at 3200 g using AgilentSpin Con- nucleofected using a Amaxa Nucleofector II device centrators (Agilent Technologies, Inc., Wilmington, (programme T003). Immediately after nucleofection, DE, USA) and normalised to cell number. Protein and cells were resuspended in pre-heated complete culture supernatant samples were boiled in NuPAGE LDS medium and then seeded at different densities for sample buffer (Invitrogen) and separated on 4–12 or further analysis. Cells were refed 24 h later and 10% NuPAGE Novex Gel (Invitrogen) with MOPS subjected to G418 (400 mg/ml) selection. Plasmid G12V or MES running buffer respectively. Proteins were vectors carrying TRK-T3 and H-RAS oncogenes transferred onto nitrocellulose filters and immunoblotted have been previously described (Pulciani et al. 1982, with the following primary antibodies to: p16INK4a (BD Greco et al. 1995). Becton Dickinson, Franklin Lakes, NJ, USA); p21CIP1, BRAF (F7), RET (C-19), TRK (C-14), IGFBP7 (Santa BrdU incorporation Cruz Biotechnology, Inc., Santa Cruz, CA, USA); p53 (YLEM); phospho-ERK1/2 (p-ERK1/2), ERK1/2, Cells were labelled with 10 mMBrdUfor3h, b-actin (Sigma–Aldrich); RAS (Ab3; Calbiochem, San harvested and fixed in 75% ethanol for 30 min. After Diego, CA, USA); HMGA2 (kindly provided from G washing with PBS, cells were treated with RNAse A Manfioletti, University of Trieste). The immunoreactive (500 mg/ml) for 30 min, permeabilised with 5 M bands were visualised using HRP-conjugated secondary HCl/0.5% Triton for 20 min at room temperature, antibodies followed by enhanced chemiluminescence neutralised with 1 M Tris, and washed in 0.5% (GE Healthcare, Buckinghamshire, UK). Tween/PBS. Incubation with primary anti-BrdU antibody (30 min at room temperature) and secondary FITC-conjugated antibodies (30 min at room tempera- Real-time RT-PCR ture; Dako, Seattle, WA, USA) was followed by RNA was isolated with TRIzol reagent (Invitrogen) staining with propidium iodine (20 mg/ml) at 37 8C for from cells collected at days 1 and 7 postinfection and 30 min. Samples were analysed by flow cytometry purified with RNeasy purification kit (Qiagen). (Becton Dickinson, San Jose, CA, USA). Reverse transcription was performed with Superscript

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 745 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours

Reverse Transcriptase III following the manufacturer’s 2 mm thickness. Antigen retrieval was performed by instructions (Invitrogen). For each sample, 20 ng of 1 mM citrate buffer (pH 6) or 0.25 mM EDTA (pH 8) retro-transcribed RNA were amplified in PCRs carried in an autoclave at 95 8C for 6–15 min. Incubation with outintriplicateonanABIPRISM7900usingTaqMan primary antibodies was performed overnight at 4 8C gene expression assays (Applied Biosystem, Foster City, for anti-p21CIP1 and anti-IGFBP7 (Santa Cruz Bio- CA, USA). Hs00266026_A1 was used for IGFBP7 technology); for 1 h at room temperature for anti- expression; human HPRT (HPRT-Hs99999909_A1) was human Ki-67 Antigen (Mib1; DakoCytomation). used as housekeeping gene for normalisation among Sections were then incubated for 30 min with biotiny- samples. Data analysis was performed by the Sequence lated anti-goat, developed using 3,30-diaminobenzidine Detection System (SDS) 2.2.2 software (Applied (DakoCytomation, Glostrup, Denmark) as chromogen Biosystems, Foster City, CA, USA). and finally counterstained with hematoxylin. p16INK4a staining was performed by CINtec Histo- Tumour samples and immunohistochemistry logic Kit (mtm Laboratories AG, Heidelberg, Germany) following the manufacturer’s instructions. The thyroid tumour samples which were analysed were The intensity of staining for IGFBP7 was scored selected by reviewing a larger tumour series available from (K) negative to (3C) strong. The percentage of at the Department of Pathology from patients operated cells immunostained for p16INK4a and p21CIP1 ranged on at IRCCS Foundation Istituto Nazionale dei Tumori from (K) !10% to (3C) 71–100%. Ki-67 staining (Milan, Italy) from 2002 to 2009. Tumours were was scored as very low (vL, 1–5% positive cells), low classified according to the histopathological typing of (L, 6–10% positive cells), medium (M, 11–30% the World Health Organization (DeLellis et al. 2004). positive cells) and high (H, R31% positive cells). Tumour collection included: PTMC (six cases); PTC NOS (six cases); PTC SV (one case; Nikiforov et al. 2001); PDTC (two cases); ATC (one case). Clinical Results pathological features are reported in Table 1. All patients gave their written informed consent, and the PTC-associated oncogenes induce senescence study was approved by the independent ethics in primary human thyrocytes committee of IRCCS Foundation Istituto Nazionale We investigated the capability of PTC-associated dei Tumori. oncogenes, namely BRAFV600E, RET/PTC1, TRK-T3 The immunohistochemical analysis was performed and H-RASG12V, to trigger OIS in human primary on formalin-fixed, paraffin-embedded tissue sections of thyrocytes.

Table 1 Clinical–pathological information of thyroid tumour case collection

Length of Age Tumour Genetic follow-up Clinical Patient (years) Gender type lesion ITS/LNM (months) outcome

1 55 F PTMC BRAFV600E ITS, LNM 102 NED 2 36 F PTMC BRAFV600E ITS, LNM 99 NED 3 31 F PTC NOS UN ITS 60 NED 4 69 M PTC NOS UN ITS 67 NED 5 49 F PTC NOS BRAFV600E – Lost – 6 48 F PTC NOS UN ITS 64 NED 7 55 F PTC NOS UN ITS Lost – 8 57 F PTC NOS UN ITS, LNM Lost – 9 46 F PTC SV BRAFV600E ITS 87 NED 10 67 F PDTC BRAFV600E LNM 12 Progression 11 75 M PDTC ND ITS 12 Progression 12 54 F ATC ND LNM 6 Died of disease 13 40 F PTMC UN ITS 72 NED 14 53 M PTMC TRK ITS, LNM 80 NED 15 48 F PTMC UN ITS 81 NED 16 30 M PTMC UN ITS 84 NED

PTC, papillary thyroid carcinoma; PTMC, papillary thyroid microcarcinoma; NOS, not otherwise specified; SV, solid variant; PDTC, poorly differentiated thyroid carcinoma; ATC, anaplastic thyroid carcinoma; UN, unknown (indicate case negative for BRAFV600E mutations, and RET or TRK rearrangement); ND, not determined; ITS, intrathyroidal spreading; LNM, lymph node metastasis; NED, no evidence of disease.

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 746 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757

We first performed transduction of primary human SA-b-Gal activity and SAHFs. A high portion of thyrocytes with a lentiviral vector carrying BRAFV600E, thyrocytes expressing BRAFV600E exhibited SA-b-Gal the most frequent oncogene in PTC; as control, activity (85%) and SAHFs (80%), whereas in thyrocytes infected with lentivirus encoding only the thyrocytes transfected with the empty vector SA-b-Gal blasticidin resistance gene were used. After selection activity and SAHFs were detected in 8 and 3% of cells with blasticidin, cells were monitored for morphology, respectively (Fig. 1C). proliferation and the presence of senescence-specific The expression of BRAFV600E, of its downstream markers. As early as 4 days after infection, BRAFV600E- effector ERK1/2, and of proteins involved in OIS such transduced thyrocytes displayed a spindle-shaped as p16INK4a, p21CIP1, p53 and HMGA2, a structural morphology compared with cells transduced with the component of SAHFs (Narita et al.2006), was empty vector (Fig. 1A). Concomitantly, a marked investigated at days 7 and 14 postinfection in oncogene growth arrest was observed, whereas control cells and empty vector-transduced thyrocytes (Fig. 1D). continued to proliferate. Thus, as measured 7 days after BRAFV600E was expressed to similar levels at both infection, thyrocytes transduced with BRAFV600E time points, in large excess with respect to endogenous exhibited a dramatic reduction of S phase (0.47% of BRAF. Concomitantly, an increase of ERK1/2 phos- cells incorporated BrdU, vs 18.7% for control cells), phorylation, with respect to control, was detected at and they accumulated preferentially in (65% day 7, and persisted at day 14. Importantly, in cells expressing BRAFV600E vs 51% for control BRAFV600E induced expression of p16INK4a, further cells; Fig. 1B). increased at day 14, whereas p53 and p21CIP1 levels BRAFV600E-transduced thyrocytes were next ana- increased slightly only at the later time point. More- lysed for the presence of two well-known OIS markers: over, elevated expression of HMGA2 protein was

A B Vector BRAFV600E Vector BRAFV600E 18.7% 0.47% BrdU Morphology

DNA content

C Vector BRAFV600E D Day 7 Day 14 VVBB -Gal

β BRAF

SA- 8% 85% p-ERK1/2

ERK1/2

p16INK4a DAPI 3% 80% p53

p21CIP1

HMGA2 DAPI

(enlarged) Actin

Figure 1 BRAFV600E induces senescence in primary human thyrocytes. Primary human thyrocytes were infected with a lentivirus encoding BRAFV600E oncogene or control vector. (A) Representative photographs showing cell morphology at 4 days after infection (magnification 10!). (B) was monitored by BrdU labelling followed by FACS analysis 7 days after infection. The upper box identifies BrdU-positive cells (S phase), the lower left box displays G0/G1 cells and the lower right box shows G2/M cells. The percentage of cells incorporating BrdU is indicated in the upper right of each graph. (C) Cells were analysed for SA-b-Gal activity (upper panel; magnification 10!) and for SAHFs formation (middle panel; magnification 20!) 14 days after infection. The percentage of cells positive for SA-b-Gal activity or carrying SAHFs is shown. Cells enlarged to show DAPI-stained chromatin foci are indicated with arrows and shown in lower panel. (D) Expression of the indicated proteins was determined by western blot analysis at 7 and 14 days after infection. b-Actin was used as a loading control. V, empty vector; B, BRAFV600E.

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 747 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours

A VVBB endogenous BRAF, data not shown). A significant number of SA-b-Gal-positive (50%) and SAHFs- bearing (30%) cells with respect to control (10 and

Vector sh-.5 Vector sh-p16.5 Vector sh-p16.13 Vector sh-p16.13 4%, respectively) were detected (data not shown). BRAF Taken together, these results indicate that the BRAFV600E oncogene induces senescence in human p16INK4a thyrocytes, and this is associated with the activation p-ERK1/2 of the ERK1/2 pathway, p16INK4a, p21CIP1 and p53 ERK1/2 signal transduction cascades. INK4a V600E p21CIP1 To investigate the role of p16 in BRAF - mediated senescence in thyrocytes, we suppressed Actin the expression of p16INK4a using two independent lentiviral shRNA vectors (#5 and #13). Six days after B 30 infection, cells were re-transduced with a lentiviral 25 vector encoding BRAFV600E, selected for expression 20 of the oncogene, and analysed 7 days later (Fig. 2). 15 Both shRNAs could silence p16INK4a expression. 10 However, p16INK4a depletion did not affect the level 5 of ERK1/2 phosphorylation induced by BRAFV600E % BrdU-positive cells 0 nor basal p21CIP1 levels (Fig. 2A). Similarly, inhibition of p16INK4a expression did not rescue the growth V600E Vector / V Vector / B Vector / V Vector / B arrest caused by BRAF (Fig. 2B). The percentage shp16.5 / V shp16.5 / B shp16.13 / V shp16.13 / B of p16INK4a-null BRAFV600E thyrocytes positive for V600E C SA-b-Gal activity was similar to BRAF cells, β–Gal SAHF whereas the percentage of p16INK4a-null BRAFV600E thyrocytes with SAHFs was reduced by w40% Vector / V 8% 1% compared with BRAFV600E cells (Table in Fig. 2C). shp16.5 / V 10% 1% This data indicated p16INK4a is not necessary for shp16.13 / V 5% 1% the execution of BRAFV600E-induced senescence in Vector / B 85% 75% thyrocytes. shp16.5 / B 80% 50% A VB shp16.13 / B 80% 40% IGFBP7

Figure 2 Effect of p16INK4a silencing on BRAFV600E-induced senescence in human primary thyrocytes. Cells were trans- B 1.5 duced with two non-overlapping shRNA against p16INK4a (5 and 13) upon infection with BRAFV600E encoding lentivirus. The analysis of OIS markers was performed at 14 days after 1 infection with the two shp16INK4a constructs. (A) Expression of the indicated proteins was determined by western blot analysis. b-Actin was used as a loading control. (B) The percentage of 0.5 cells for BrdU incorporation is shown. (C) The percentage of IGFBP7 RQ cells positive for each indicated markers is shown in the table. V, empty vector; B, BRAFV600E. 0 maintained up to 14 days after transduction in the V BBV V600E BRAF -expressing cells, thus corroborating the Day 1 Day 7 changes detected by DAPI staining. To rule out the possibility that OIS observed in Figure 3 Detection of IGFBP7 expression in BRAFV600E- thyrocytes could be related to the abnormal transduced human primary thyrocytes. The expression of V600E IGFBP7 in primary thyrocytes infected with control lentivirus or BRAF expression levels obtained by lentiviral lentivirus-expressing BRAFV600E was monitored by western blot transduction, we performed nucleofection of thyro- analysis 7 days after infection (A) and by real-time RT-PCR analysis 1 and 7 days after infection (B). V, control vector; B, cytes, thus achieving a low level of expression of the BRAFV600E; RQ, relative quantity of IGFBP7 mRNA normalised oncoprotein (two-three times more abundant than for HPRT housekeeping gene expression.

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 748 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757

IGFBP7 is a secreted protein recently proposed as a retrovirus encoding the RET/PTC1 oncogene produced mediator of BRAFV600E-induced senescence in mela- in the course of a previous study, and shown to display nocytes (Wajapeyee et al. 2008); this role, however, an increased growth rate compared with control has been questioned by other authors (Scurr et al. uninfected thyrocytes, as determined in analyses 2010). Interestingly, our recent results implicate performed between passages 3 and 7 after transduction IGFBP7 as a tumour suppressor protein in thyroid (Borrello et al. 2005). Later on, however, by extending tumours: its expression is frequently lost in FTC and the number of passages in culture with respect to the PTC, and its restoration modulates the transformed previous study, we noted that RET/PTC1-infected features of PTC-derived tumour cells (Vizioli et al. thyrocytes, in contrast to control cells, ceased pro- 2010). To investigate the possible involvement of liferating and showed several morphological abnorm- IGFBP7 in BRAFV600E-induced thyrocytes senescence alities, which were not further investigated at that time. we analysed the expression of soluble IGFBP7 in In the course of the present work, we re-evaluated the thyrocytes 7 days after BRAFV600E transduction. previous observation, taking advantage of frozen stocks BRAFV600E-infected cells showed reduced levels of of RET/PTC1-transduced thyrocytes and relative secreted IGFBP7 protein compared with the control controls. Here again, 10–12 passages after infection, infected cells (Fig. 3A). To further characterise such a RET/PTC1-expressing cells ceased proliferating and reduction we performed real-time RT-PCR (Fig. 3B). displayed morphological features of senescence, such Also in this case, we observed a decrease in IGFBP7 as extensive vacuolisation. Moreover, SA-b-Gal and mRNA expression in senescent cells (day 7). Of SAHFs were found in 80 and 70%, respectively, of note, such a reduction was detectable as early as 1 day cells infected with RET/PTC1, but only in 10 and 2%, after infection. respectively, of control cells (Fig. 4A). Biochemical Next, to investigate the effect of RET/PTC oncogenes analysis showed a slight increase in ERK1/2 phos- in primary thyrocytes we used cells transduced with a phorylation compared with control cells. In addition, a

ABVector RET/PTC1

Vector RET/PTC1 RET Morphology

p-ERK1/2

-Gal ERK1/2 β SA- p16INK4a 10% 80%

p53

p21CIP1 DAPI

2% 75% HMGA2

Actin DAPI (enlarged)

Figure 4 Senescence programme is induced by oncogenic RET/PTC1 in primary human thyrocytes. Primary human thyrocytes transduced with retroviral vector carrying RET/PTC1 oncogene (Borrello et al. 2005) were analysed for the presence of OIS markers at passages 10–12; uninfected cells were used as control. (A) Analysis of cell morphology (upper panel; magnification 10!), SA-b- Gal activity (middle panel; magnification 10! for vector and 20! for RET/PTC1) and SAHFs formation (bottom panel; magnification 40!). Percentage of cells positive for SA-b-Gal activity and SAHFs are shown in each photograph. Cells enlarged to show DAPI- stained chromatin foci are indicated with arrows and shown in lower panel. (B) Western blot analysis of the indicated proteins in uninfected or infected cells. b-Actin was monitored as a loading control.

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 749 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours marked upregulation of p16INK4a, p53 and HMGA2, Overall, our in vitro data suggests that all the and a slight increase of p21CIP1 was observed in RET/ PTC-associated oncogenes are capable of triggering PTC1-infected cells (Fig. 4B). senescence in primary human thyrocytes, and this To introduce TRK and RAS oncogenes into thyro- is not related to the oncoprotein expression level. cytes we employed the nucleofection technology, V600E based on the evidence that BRAF -induced Analysis of OIS markers in thyroid tumour senescence in thyrocytes is detectable with this samples transfection method, as described earlier. Plasmid expression vectors carrying TRK-T3 or H-RASG12V To study the relevance of OIS in thyroid tumour pathogenesis we performed IHC analysis for the oncogenes, and the empty vector as control, were detection of OIS markers in a series of thyroid tumour transfected by nucleofection in primary human samples characterised by increasing aggressiveness, thyrocytes. Thyrocytes transfected with the TRK-T3 thus representing epithelial thyroid tumour pro- or H-RASG12V oncogenes underwent growth arrest and gression. The case collection included two PTMCs, displayed flat, enlarged morphology and massive seven PTCs (six NOS and one SV (Nikiforov et al. vacuolisation. Moreover, an increase of SA-b-Gal- 2001)), two PDTC and one ATC. Clinical–patho- positive cells was detected in TRK-T3 and H-RASG12V logical features and follow-up information for these transfected cells (60 and 70%, respectively), whereas cases are reported in Table 1. Notably, for PTMC #2 in thyrocytes transfected with empty vector SA-b-Gal three intrathyroidal spreads and a nodal metastasis activity was observed in 10% of cells. SAHFs were G12V were available. Sample #9 (SV) and sample #10 detected in 40 and 50% of TRK-T3 and H-RAS (PDTC) presented areas with papillary and solid transfected thyrocytes, respectively, but only in a small pattern of growth, thus providing a model of in situ fraction (4%) of control-transfected cells (Fig. 5A). tumour progression. The case collection was analysed Biochemical analysis was performed at days 5 and by IHC for the expression of p16INK4a and p21CIP1, 10 post-nucleofection (Fig. 5B). The expression of well-known effectors of in vitro OIS, of IGFBP7, G12V TRK-T3 and H-RAS was constant. An increase in recently proposed as a factor required for OIS in ERK1/2 phosphorylation level, with respect to control melanocytes, and of the proliferation marker Ki-67. cells, was induced by both oncoproteins at the two time The results of IHC analysis are reported in Table 2; points analysed. Similarly, an increase in p16INK4a representative pictures are shown in Figs 6 and 7. expression was observed. None of the oncogenes PTMC samples showed the highest expression of affected the levels of p21CIP1. A slight increase of p53 p16INK4a, which was upregulated in tumour areas with was observed in H-RASG12V transfected cells at day 10. respect to the adjacent non-tumoural thyroid tissue;

A B Vector TRK-T3 H-RASG12V Day 5 Day 10 Day 5 Day 10 V T V T VVRR TRK RAS Morphology p-ERK1/2

-Gal ERK1/2 β

INK4a SA- 10% 60% 70% p16 p53 p21CIP1 DAPI 4% 40% 50% Actin DAPI (enlarged)

Figure 5 Expression of TRK-T3 and H-RasG12V oncogene triggers senescence in primary human thyrocytes. Primary human thyrocytes were transfected with plasmid expression vectors encoding TRK-T3 and H-RasG12V or empty vector by nucleofection technology. (A) Cell morphology (upper panel; magnification 10!) 4 days after transfection, the appearance of SA-b-Gal (middle panel, magnification 10! for vector, 20! for TRK-T3 and H-RasG12V) and SAHFs (lower panel; magnification 40!) at 10 days after transfection are shown. Cells enlarged to show DAPI-stained chromatin foci are indicated with arrows and shown in lower panel. (B) Immunoblot analysis monitoring the expression of the indicated proteins at 5 and 10 days after transfection. b-Actin was used as a loading control. V, empty vector; T, TRK-T3; R, H-RasG12V.

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 750 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757 p16INK4a-positive cells accounted for 50–60% of total. With respect to p21CIP1, no staining was observed in Immunolabelling for p16INK4a was both cytoplasmic the non-tumoural tissue surrounding the tumour area. and nuclear. In PTMC #2 p16 INK4a was expressed at a The highest expression was detected in PTMCs, similar level in all the different tumour areas analysed, showing 30–70% of positive cells, which persisted in including nodal metastasis. The PTCs NOS samples the different lesions of PTMC #2. PTCs NOS were showed a variable extent of p16INK4a expression: one classified as negative, except for two cases that showed case was scored negative (!10% of positive cells); 11–40% of positive cells. The SV sample was scored four cases showed 11–40% of positive cells, whereas positive in papillary areas and negative in solid ones. only one case displayed more than 40% positive cells. In PDTC, expression of p16 INK4a was detected only in In general, the distribution of p16INK4a was scattered; the papillary areas of sample #10. No expression of we observed groups of positive cells within the tumour p21CIP1 was observed in the ATC sample. mass. The PTC SV sample was scored positive in the IGFBP7 showed the same trend of expression as areas with papillary growth, but negative in the solid p16INK4a and p21CIP1. In both PTMC primary tumour ones. With respect to PDTC samples the papillary samples, and in the intrathyroidal and nodal spreads of areas of case #10 were positive, whereas the solid areas PTCM #2, IGFBP7 was upregulated with respect to as well as case #11 were negative. No expression of the adjacent non-tumoural tissue. PTCs NOS showed p16INK4a was detected in the ATC sample. variable expression levels: three cases were scored as moderately positive, one case as weakly positive, and two as negative. IGFBP7 expression was detected in Table 2 Immunohistochemical analysis for p16INK4a, p21CIP1, IGFBP7 and Ki-67 in thyroid carcinomas the papillary, but not in the solid areas of the SV sample; a similar trend was observed for PDTC #10, whereas Patient Sample p16INK4a p21CIP1 IGFBP7 Ki-67 PDTC #11 showed weak expression, and no expression was detected in the ATC sample. The expression of PTMC INK4a CIP1 1 CC C CC L p16 , and IGFBP7 proteins was inversely 2 PT, RL CC CCC CC vL correlated with the proliferative index. The expression 2 ITS, RL CC CCC CC vL of the Ki-67 proliferation marker was very low or 2 ITS, Is CC CCC CC vL low in PTMCs; it ranged from low to medium in PTC 2 ITS, LL CC CCC CC vL 2 LNM CC CCC CC vL NOS and in the papillary areas of samples #9 and #10; PTC it increased to medium/high and high in PDTC and NOS ATC respectively. In Fig. 6A representative pictures of 3 CCK L PTMC, PTC NOS, PDTC and ATC are shown; Fig. 6B 4 KKC L reports the analysis of papillary and solid areas of 5 CKCCL/M 6 CCCCM PDTC sample #10. Interestingly, in PTC samples the 7 CKK L expression of OIS markers was often associated with 8 CC K CC L several structures suggestive of tumour regression SV (Fig. 6C) such as: i) areas with lymphocyte infiltration 9 Papillary CCC/CC L (left panel), which has been reported to play a protective KKK 9 Solid ND et al PDTC role in thyroid cancer (Ugolini . 2007); ii) swelling 10 Papillary CCCCCL oedematous papillae cores (middle panel) and iii) 10 Solid KKK H syncytial clusters of tumoural cells (right panel), a 11 KKC M/H frequent finding in PTC (Fulciniti et al. 2001), resulting ATC from papillae apex detachment. 12 KKK H The first conclusion that can be drawn from the IHC PTMC, papillary thyroid microcarcinoma; PTC, papillary thyroid analysis is that the expression of OIS markers is carcinoma; ATC, anaplastic thyroid carcinoma; NOS, not upregulated at early stages and then lost during thyroid otherwise specified; SV, solid variant; PT, primary tumour; ITS, intrathyroidal spread; LNM, lymph node metastasis; RL, right tumour progression. This issue is strengthened by the lobe; LL, left lobe; Is, isthmus. The percentage of cells with evidence that in samples #9 and #10, which recapitu- INK4a CIP1 staining for p16 and p21 was scored from 0 to 3: (K) late in situ tumour progression, the expression of the !10% of positive cells; (1C) 11–40%; (2C) 41–70%; (3C) 71–100%. The intensity of staining for IGFBP7 was from: (K)to three proteins analysed was detected in areas featuring (3C): (K) negative; (C) weakly positive; (CC) moderately papillary growth but not in solid ones. positive; (CCC) strongly positive. Ki-67 staining was scored The highest expression of all OIS markers analysed as very low (vL, 1–5% positive cells), low (L, 6–10% positive cells), medium (M, 11–30% positive cells) and high (H, R31% observed in PTMCs would suggest that OIS may play positive cells). ND, not done (no material available). a role in the generally indolent clinical course of this

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 751 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours

A PTMC PTC NOS PDTC ATC B Papillary Solid

N INK4a INK4a T p16 p16

CIP1 T CIP1

N

T N

N

Ki-67 IGFBP7 p21 T Ki-67 IGFBP7 p21

C p16INK4a p21CIP1

p21CIP1 p16INK4a p21CIP1 IGFBP7

Figure 6 Immunohistochemical analysis of OIS markers (p16INK4a, p21CIP1 and IGFBP7) and Ki-67 in thyroid tumour samples. (A) Representative pictures of a papillary thyroid microcarcinoma (PTMC), a PTC not otherwise specified (NOS) variant, a poorly differentiated thyroid carcinoma (PDTC) and an anaplastic thyroid carcinoma (ATC) are shown. N, normal; T, tumour. (B) Papillary and solid areas of PDTC #10. (C) PTC sample distinguished by the presence of lymphocyte infiltration (first panel); oedematous papillae cores (second panel) and multinucleated giant cells (third panel).

PTC variant by keeping proliferation under control, that all the oncogenes activated in thyroid tumours, even in the presence of intrathyroidal and nodal namely BRAFV600E, RET/PTC1, TRK-T3 and spreads, as suggested by the analysis of PTMC #2 H-RASG12V, are capable of inducing cellular senes- (Table 2 and Fig. 7). To further explore this issue cence in primary human thyrocytes, as judged by the we analysed four additional PTMCs presenting with presence of growth arrest, changes in cell morphology, intraglandular and/or nodal spreads (see Table 1); due accumulation of SA-b-Gal and chromatin modifi- to material shortage the analysis was limited to Ki-67 cations. Our conclusion is corroborated by the INK4a and p16 .InTable 3, the primary tumours showed evidence that OIS is detectable with different onco- INK4a relatively intense p16 expression and very low genes transfected with different methods and, thus, Ki-67 immunolabelling. The same pattern of expression expressed at different levels. Differences in efficiency was observed in the intraglandular and/or lymph and onset of OIS observed with the different oncogenes node tumour foci. Interestingly, in none of the PTMC can be ascribed to experimental conditions and patients was recurrence documented in the follow-up oncoprotein expression level, rather than to intrinsic period, which ranged from 72 to 102 months (Table 1). oncogenic features. Overall, our data demonstrate the presence of OIS in Shedding light on the molecular mechanisms PTMC, and suggests that this mechanism maintains the responsible for the OIS onset, the predominant view indolent behaviour of this tumour type, despite multi- is that aberrant levels of ERK1/2 activate a senescence focality within thyroid gland and loco-regional spread. programme through the critical involvement of the p16INK4a/pRB and p53/p21CIP1 networks (Lin et al. 1998). In our cellular setting, PTC-associated onco- Discussion genes trigger a marked activation of ERK1/2, We provided evidence that OIS may represent a barrier accompanied by a persistent induction of p16INK4a, to the progression of thyroid tumours. In vitro studies, whereas we failed to observe any significant upregula- using different gene transfer methods, demonstrated tion of p53 or p21CIP1, with the exception of

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 752 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757

p16INK4a p21CIP1 IGFBP7 Ki-67

PT

HL

CL

LN

Figure 7 Immunohistochemical analysis for the expression of p16INK4a, p21CIP1, IGFBP7 and Ki-67 in primary tumour (PT), homolateral (HL), controlateral (CL) and nodal (LN) spreads of papillary thyroid microcarcinoma (PTMC) #2. The Ki-67 panel of primary tumour is the same of Fig. 6A. The Ki-67-positive area in the LN sample represents lymphoid tissue.

RET/PTC1-expressing thyrocytes in which an increase Although initially considered as an in vitro pheno- of p53 was detected. We observed a reduction in menon, OIS is now viewed as an authentic cancer IGFBP7, whose role in OIS is debated (Scurr et al.2010). barrier. Several groups have reported its occurrence in The relative contribution of the p16INK4a/pRB and different in vivo lesions including human melanocytic the p53/p21CIP1 pathways in mediating a senescence nevi, human dermal neurofibromas, human and murine programme remains poorly understood. It has been prostatic and thyroid adenomas, and murine lympho- shown that p16INK4a plays a more major role than p21CIP1 mas and melanoma (Braig et al. 2005, Chen et al. in the induction of senescence in many contexts in vitro. 2005, Collado et al. 2005, Courtois-Cox et al. 2006, In some cells, p16INK4a is solely responsible for OIS Kortlever et al. 2006, Dankort et al. 2007, Ha et al. onset (Ben-Porath & Weinberg 2005), whereas in other 2007). To extend our observations to an in vivo context cell types it has been shown that p16INK4a alone is not we analysed a panel of thyroid tumour samples with sufficient to execute the OIS programme (Michaloglou increasing aggressiveness: PTMC, characterised by a et al.2005, Haferkamp et al.2009). In our experimental very indolent course; PTC NOS, a differentiated setting the inactivation of p16INK4a with shRNA did carcinoma generally associated with a good prognosis; not affect BRAFV600-induced senescence in thyrocytes, PDTC, associated with a poor prognosis; ATC, an except for a reduction of SAHF-bearing cells. This raises undifferentiated, very aggressive carcinoma with fatal the possibility that p16INK4a may cooperate with other outcome, mostly arising from a time-dependent factors in triggering senescence. dedifferentiation of differentiated carcinoma, in The upregulation of p16INK4a observed in our particular PTC. We analysed the expression of analysis is in agreement with previous studies, showing p16INK4a and p21CIP1, two universally recognised that the proliferation arrest induced in thyrocytes OIS markers, and IGFBP7, a protein with an by H-RasG12V was associated with upregulation of oncosuppressor role in thyroid tumours (Vizioli et al. p16INK4a (Jones et al. 2000). In that setting, however, 2010), and often found involved in melanocyte OIS silencing p16INK4a rescued H-RasG12V expressing (Kortlever et al. 2006, Acosta et al. 2008, Kuilman thyrocytes from senescence (Bond et al. 2004), et al. 2008, Wajapeyee et al. 2008), although its role suggesting that the OIS programme might not be in this process is controversial (Scurr et al. 2010). universal in different genetic contexts. Collectively, the three proteins were upregulated in

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 753 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours

Table 3 Immunohistochemical analysis for p16INK4a and Ki-67 OIS triggers the activation of the inflammatory network in a small series of PTMCs and evokes an innate immune response that ultimately counteracts tumour progression (Ren et al.2009). Patient Sample p16INK4a Ki-67 Analysis of a small series of PTMC showed the 13 PT, Is CC vL/L consistent upregulation of p16INK4a in association with 13 ITS, RL CC vL a very low proliferative index. Of note, p16INK4a C 13 ITS, LL vL expression and Ki-67 immunolabelling remained 14 PT, LL CC vL 14 ITS, Is ND ND unchanged in synchronous intrathyroidal and nodal CIP1 14 LNM CC vL spreads. In one PTMC sample, the analysis of p21 15 PT, RL CC vL and IGFBP7 was possible, showing the same 15 ITS, RL ND vL expression level in the primary tumour, and intrathyr- CC 16 PT, RL vL oidal and nodal foci. Notably, none of the PTMC 16 ITS, Is ND ND 16 ITS, LL C vL patients of our study experienced recurrence during the follow-up period (72–102 months). PT, primary tumour; ITS, intrathyroidal spread; LNM, lymph Expression of p16INK4a in thyroid malignancies has node metastasis; RL, right lobe; LL, left lobe; Is, isthmus. The percentage of cells with staining for p16INK4a was scored from 0 been previously investigated by several groups. Over- INK4a to 3: (K) !10% of positive cells; (1C) 11–40%; (2C) 41–70%; expression of p16 in adenomas, papillary and (3C) 71–100%. Ki-67 staining was scored as very low (vL, follicular tumours, and the loss of expression in ATC 1–5% positive cells), low (L, 6–10% positive cells), medium (M, 11–30% positive cells) and high (H, R31% positive cells). was observed in previous works (Ferru et al. 2006, Ball INK4a ND, not done (no material available). et al. 2007), suggesting that p16 expression is induced in early-stage tumours and then progressively PTMCs; the PTCs showed a variable pattern, including lost during tumour progression. Similarly, the over- loss of expression; finally, the ATC sample was scored expression of p21CIP1 in PTMC and progressive loss of negative for the all proteins. The upregulation in expression with advancing tumour grade in PTC has PTMCs and the progressive loss in PTCs and ATC been reported (Brzezinski et al. 2005). These previous support the notion that OIS may counteract oncogenic reports together with our data suggest that the indolent activity in thyroid tumours, and its escape allows course of PTMC, as well as the characteristically slow thyroid tumour progression. growth of classical PTC, may be explained by the Regarding IGFBP7, we have previously shown that persistence of a severe limit to oncogene-induced its mRNA expression is frequently downregulated in proliferation exerted by the OIS mechanism. In this PTC, and this is associated with the TC variant (Vizioli light PTMC represents, among the variants examined, et al. 2010). Our IHC results are in agreement with the the most representative model for the tumour gene expression data, as we observed a progressive loss suppressor mechanism exerted by OIS. Indeed, all of IGFBP7 protein during thyroid tumour progression. the foci examined across the thyroid and lymph nodes The trend of IGFBP7 expression was similar to that of share the same ‘frozen ’ in terms of p16INK4a and p21CIP1, suggesting that it may partici- p16INK4a, p21CIP1 and IGFBP7. pate in counteracting the aberrant proliferative signals The persistence of an OIS signal in intrathyroidal of oncogenes. However, in senescent BRAFV600E and nodal spreads might explain the favourable thyrocytes we detected a reduction of IGFBP7, at prognosis of PTMC even in the presence of multi- both the protein and mRNA level. This discrepancy focality and lymph node metastasis, two features may be explained by taking into account that the associated with lateral node recurrence, but not with induction of IGFBP7 in vivo might result from the an increased risk of mortality (Hay et al. 2008). complex crosstalk between tumour and stromal cells Moreover, the constant expression of all the OIS (Kuilman & Peeper 2009). Nevertheless, more studies markers analysed and Ki-67 labelling across the are required to assess if IGFBP7 plays a role in different lesions of the same patient suggests that thyrocyte OIS. cells have acquired the capability to invade without Further support to the suggestion that OIS counter- changes in their proliferation capability. This is in acts oncogenic activation in PTC is provided by the keeping with reports claiming invasion and prolifer- evidence that the expression of OIS markers may be ation as two distinct mechanisms (Gao et al. 2005), associated with several characteristics of tumour governed by different genes and pathways. regression, such as lymphocytic infiltration, oedema- We have identified OIS as a mechanism involved in tous papillae cores and syncytial clusters. The presence the pathogenesis of thyroid carcinoma by restraining of these features is in keeping with the notion that the evolution of PTMC to PTC, as well as the transition

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 754 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757 of well differentiated and less aggressive PTC to expression in multistep thyroid tumourigenesis. European undifferentiated and more aggressive variants. This Journal of Cancer 43 194–201. (doi:10.1016/j.ejca.2006. raises the idea that restoration of this process in tumour 08.025) cells may represent a therapeutic approach, especially Beausejour CM, Krtolica A, Galimi F, Narita M, Lowe SW, for those tumours which do not respond to standard Yaswen P & Campisi J 2003 Reversal of human cellular therapy. Moreover, a possible therapeutic strategy senescence: roles of the p53 and p16 pathways. EMBO based on OIS restoration is independent from the Journal 22 4212–4222. (doi:10.1093/emboj/cdg417) oncogenic alteration carried by the tumour, as we have Ben-Porath I & Weinberg RA 2005 The signals and pathways activating cellular senescence. International Journal of shown that all PTC-associated oncogenes are capable Biochemistry & Cell Biology 37 961–976. (doi:10.1016/j. of triggering the senescence programme. The feasi- biocel.2004.10.013) bility of enforced cellular senescence as a therapeutic Bond J, Jones C, Haughton M, Demicco C, Kipling D & approach for thyroid as well as other tumour types Wynford-Thomas D 2004 Direct evidence from siRNA- remains to be investigated, and it will require a deeper directed "knock down" that p16(INK4a) is required for understanding of the mechanisms involved in the human fibroblast senescence and for limiting ras-induced process, as well as suitable in vitro and in vivo models. epithelial cell proliferation. Experimental Cell Research 292 151–156. (doi:10.1016/j.yexcr.2003.09.021) Declaration of interest Borrello MG, Alberti L, Fischer A, Degl’Innocenti D, Ferrario C, Gariboldi M, Marchesi F, Allavena P, Greco A The authors declare that there is no conflict of interest that & Pierotti MA 2005 Induction of a proinflammatory could be perceived as prejudicing the impartiality of the programme in normal human thyrocytes by the research reported. RET/PTC1 oncogene. PNAS 102 14825–14830. (doi:10. 1073/pnas.0503039102) Funding Braig M, Lee S, Loddenkemper C, Rudolph C, Peters AH, Schlegelberger B, Stein H, Dorken B, Jenuwein T & This work was supported by AIRC grant N.4472, ACC, Schmitt CA 2005 Oncogene-induced senescence as an Ministry of Health, and Progetto Strategico Istituzionale initial barrier in lymphoma development. 436 (PSI) 5xmille – 2008. 660–665. (doi:10.1038/nature03841) Brzezinski J, Migodzinski A, Toczek A, Tazbir J & Dedecjus Acknowledgements M 2005 Patterns of cyclin E, , and p21Cip1/WAF1 immunostaining in the oncogenesis of We thank Dr Silvia Veneroni and Tissue Bank of IRCCS papillary thyroid carcinoma. Clinical Cancer Research 11 Foundation Istituto Nazionale dei Tumori for sample 1037–1043. collection; Dr Federica Perrone (Department of Pathology) Campisi J 2005 Senescent cells, tumor suppression, and for assistance in the selection of archival thyroid tumour Cell 120 material; Mrs Francesca Dominoni (Department of Path- organismal aging: good citizens, bad neighbors. ology) for performing Ki-67 immunohistochemistry; Prof. 513–522. (doi:10.1016/j.cell.2005.02.003) Guidalberto Manfioletti (University of Trieste; IT) for Chen Z, Trotman LC, Shaffer D, Lin HK, Dotan ZA, Niki M, providing anti-HMGA2; Drs Fabrizio D’Adda Di Fagagna Koutcher JA, Scher HI, Ludwig T & Pandolfi PP 2005 and Gabriele Sulli (IFOM-IEO Campus, Milan) for technical Crucial role of p53-dependent cellular senescence in advice; Dr R Marais (ICR, London, UK) for providing suppression of Pten-deficient tumorigenesis. Nature 436 BRAFV600E plasmid expression vector and Mrs Silvia Grassi 725–730. (doi:10.1038/nature03918) for secretarial assistance. Collado M, Gil J, Efeyan A, Guerra C, Schuhmacher AJ, Barradas M, Benguria A, Zaballos A, Flores JM & Serrano M 2005 Tumour biology: senescence in pre- References malignant tumours. Nature 436 642. (doi:10.1038/ Acosta JC, O’Loghlen A, Banito A, Guijarro MV, Augert A, 436642a) Raguz S, Fumagalli M, Da Costa M, Brown C & Gil J Courtois-Cox S, Genther Williams SM, Reczek EE, Johnson 2008 Chemokine signaling via the CXCR2 receptor BW, McGillicuddy LT, Johannessen CM, Hollstein PE, reinforces senescence. Cell 133 1006–1018. (doi:10.1016/ MacCollin M & Cichowski K 2006 A negative feedback j.cell.2008.03.038) signaling network underlies oncogene-induced senes- Adams PD 2007 Remodeling of chromatin structure in cence. Cancer Cell 10 459–472. (doi:10.1016/j.ccr.2006. senescent cells and its potential impact on tumor 10.003) suppression and aging. Gene 397 84–93. (doi:10.1016/j. Curcio F, Ambesi-Impiombato FS, Perrella G & Coon HG gene.2007.04.020) 1994 Long-term culture and functional characterization of Ball E, Bond J, Franc B, Demicco C & Wynford-Thomas D follicular cells from adult normal human thyroids. PNAS 2007 An immunohistochemical study of p16(INK4a) 91 9004–9008. (doi:10.1073/pnas.91.19.9004)

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 755 M G Vizioli et al.: Oncogene-induced senescence in thyroid tumours

D’Adda di Fagagna F 2008 Living on a break: cellular require the p16(INK4a) or p14ARF melanoma tumor senescence as a DNA-damage response. Nature Reviews. suppressors. Journal of Investigative Dermatology 129 Cancer 8 512–522. (doi:10.1038/nrc2440) 1983–1991. (doi:10.1038/jid.2009.5) Dankort D, Filenova E, Collado M, Serrano M, Jones K & Hay ID, Hutchinson ME, Gonzalez-Losada T, McIver B, McMahon M 2007 A new mouse model to explore the Reinalda ME, Grant CS, Thompson GB, Sebo TJ & initiation, progression, and therapy of BRAFV600E- Goellner JR 2008 Papillary thyroid microcarcinoma: a induced lung tumors. Genes and Development 21 study of 900 cases observed in a 60-year period. Surgery 379–384. (doi:10.1101/gad.1516407) 144 980–987. (doi:10.1016/j.surg.2008.08.035) DeLellis RA, Lloyd RV, Heitz PU & Eng C 2004 Pathology Jones CJ, Kipling D, Morris M, Hepburn P, Skinner J, and genetics of tumors of endocrine organs. In World Bounacer A, Wyllie FS, Ivan M, Bartek J & Bond JA Health Organization Classification of Tumours, pp 51–56. 2000 Evidence for a -independent "Clock" Lyon: IARC Press. limiting RAS oncogene-driven proliferation of human Denoyelle C, Abou-Rjaily G, Bezrookove V, Verhaegen M, thyroid epithelial cells. Molecular and Cellular Biology Johnson TM, Fullen DR, Pointer JN, Gruber SB, Su LD & 20 5690–5699. (doi:10.1128/MCB.20.15.5690-5699. Soengas MS 2006 Anti-oncogenic role of the endoplasmic 2000) reticulum differentially activated by mutations in the Kortlever RM, Higgins PJ & Bernards R 2006 Plasminogen MAPK pathway. Nature Cell Biology 8 1053–1063. activator inhibitor-1 is a critical downstream target of p53 (doi:10.1038/ncb1471) in the induction of replicative senescence. Nature Cell Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, Biology 8 877–884. (doi:10.1038/ncb1448) Medrano EE, Linskens M, Rubelj I & Campisi J 1995 A Kuilman T & Peeper DS 2009 Senescence-messaging that identifies senescent human cells in culture secretome: SMS-ing cellular stress. Nature Reviews. and in aging in vivo. PNAS 92 9363–9367. (doi:10. Cancer 9 81–94. (doi:10.1038/nrc2560) 1073/pnas.92.20.9363) Kuilman T, Michaloglou C, Vredeveld LC, Douma S, van Ferru A, Fromont G, Gibelin H, Guilhot J, Savagner F, Doorn R, Desmet CJ, Aarden LA, Mooi WJ & Peeper DS Tourani JM, Kraimps JL, Larsen CJ & Karayan-Tapon L 2008 Oncogene-induced senescence relayed by an 2006 The status of CDKN2A alpha (p16INK4A) and beta interleukin-dependent inflammatory network. Cell 133 (p14ARF) transcripts in thyroid tumour progression. 1019–1031. (doi:10.1016/j.cell.2008.03.039) British Journal of Cancer 95 1670–1677. (doi:10.1038/sj. Kuilman T, Michaloglou C, Mooi WJ & Peeper DS 2010 bjc.6603479) The essence of senescence. Genes and Development 24 Fulciniti F, Benincasa G, Vetrani A & Palombini L 2001 2463–2479. (doi:10.1101/gad.1971610) Follicular variant of papillary carcinoma: cytologic Lin AW, Barradas M, Stone JC, Van Aelst L, Serrano M & findings on FNAB samples-experience with 16 cases. Lowe SW 1998 Premature senescence involving p53 and Diagnostic Cytopathology 25 86–93. (doi:10.1002/dc. p16 is activated in response to constitutive MEK/MAPK 2010) mitogenic signaling. Genes and Development 12 Gao CF, Xie Q, Su YL, Koeman J, Khoo SK, Gustafson M, 3008–3019. (doi:10.1101/gad.12.19.3008) Knudsen BS, Hay R, Shinomiya N & Vande Woude GF Lowe SW, Cepero E & Evan G 2004 Intrinsic tumour 2005 Proliferation and invasion: plasticity in tumor suppression. Nature 432 307–315. (doi:10.1038/ cells. PNAS 102 10528–10533. (doi:10.1073/pnas. nature03098) 0504367102) Lupi C, Giannini R, Ugolini C, Proietti A, Berti P, Minuto M, Greco A, Mariani C, Miranda C, Lupas A, Pagliardini S, Materazzi G, Elisei R, Santoro M & Basolo F 2007 Pomati M & Pierotti MA 1995 The DNA rearrangement Association of BRAF V600E mutation with poor that generates the TRK-T3 oncogene involves a novel clinicopathological outcomes in 500 consecutive cases gene on 3 whose product has a potential of papillary thyroid carcinoma. Journal of Clinical coiled-coil domain. Molecular and Cellular Biology 15 Endocrinology and Metabolism 92 4085–4090. 6118–6127. (doi:10.1210/jc.2007-1179) Greco A, Borrello MG, Miranda C, Degl’Innocenti D & Michaloglou C, Vredeveld LC, Soengas MS, Denoyelle C, Pierotti MA 2009 Molecular pathology of differentiated Kuilman T, van der Horst CM, Majoor DM, Shay JW, thyroid cancer. Quarterly Journal of Nuclear Medicine Mooi WJ & Peeper DS 2005 BRAFE600-associated and Molecular Imaging 53 440–453. senescence-like arrest of human naevi. Nature Ha L, Ichikawa T, Anver M, Dickins R, Lowe S, Sharpless 436 720–724. (doi:10.1038/nature03890) NE, Krimpenfort P, DePinho RA, Bennett DC & Merlino Mooi WJ & Peeper DS 2006 Oncogene-induced cell G 2007 ARF functions as a melanoma tumor suppressor senescence – halting on the road to cancer. New England by inducing p53-independent senescence. PNAS 104 Journal of Medicine 355 1037–1046. (doi:10.1056/ 10968–10973. (doi:10.1073/pnas.0611638104) NEJMra062285) Haferkamp S, Scurr LL, Becker TM, Frausto M, Kefford RF Narita M, Nunez S, Heard E, Narita M, Lin AW, Hearn SA, & Rizos H 2009 Oncogene-induced senescence does not Spector DL, Hannon GJ & Lowe SW 2003 Rb-mediated

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access 756 www.endocrinology-journals.org Endocrine-Related Cancer (2011) 18 743–757

heterochromatin formation and silencing of E2F target Serrano M, Lin AW, McCurrach ME, Beach D & genes during cellular senescence. Cell 113 703–716. Lowe SW 1997 Oncogenic ras provokes premature cell (doi:10.1016/S0092-8674(03)00401-X) senescence associated with accumulation of p53 and Narita M, Narita M, Krizhanovsky V, Nunez S, Chicas A, p16INK4a. Cell 88 593–602. (doi:10.1016/S0092-8674 Hearn SA, Myers MP & Lowe SW 2006 A novel role for (00)81902-9) high-mobility group a proteins in cellular senescence and Sheils O 2005 Molecular classification and biomarker heterochromatin formation. Cell 126 503–514. (doi:10. discovery in papillary thyroid carcinoma. Expert Review 1016/j.cell.2006.05.052) of Molecular Diagnostics 5 927–946. (doi:10.1586/ Nikiforov YE, Erickson LA, Nikiforova MN, Caudill CM & 14737159.5.6.927) Lloyd RV 2001 Solid variant of papillary thyroid Ugolini C, Basolo F, Proietti A, Vitti P, Elisei R, Miccoli P & carcinoma: incidence, clinical-pathologic characteristics, Toniolo A 2007 Lymphocyte and immature dendritic cell molecular analysis, and biologic behavior. American infiltrates in differentiated, poorly differentiated, and Journal of Surgical Pathology 25 1478–1484. (doi:10. undifferentiated thyroid carcinoma. Thyroid 17 389–393. 1097/00000478-200112000-00002) (doi:10.1089/thy.2006.0306) Pulciani S, Santos E, Lauver AV, Long LK, Robbins KC & Vizioli MG, Sensi M, Miranda C, Cleris L, Formelli F, Barbacid M 1982 Oncogenes in human tumor cell lines: Anania MC, Pierotti MA & Greco A 2010 IGFBP7: an molecular of a transforming gene from human oncosuppressor gene in thyroid carcinogenesis. Oncogene bladder carcinoma cells. PNAS 79 2845–2849. (doi:10. 29 3835–3844. (doi:10.1038/onc.2010.136) 1073/pnas.79.9.2845) Wajapeyee N, Serra RW, Zhu X, Mahalingam M & Green Ren JL, Pan JS, Lu YP, Sun P & Han J 2009 Inflammatory signaling and cellular senescence. Cellular Signalling 21 MR 2008 Oncogenic BRAF induces senescence and 378–383. (doi:10.1016/j.cellsig.2008.10.011) apoptosis through pathways mediated by the secreted Sakorafas GH, Giotakis J & Stafyla V 2005 Papillary thyroid protein IGFBP7. Cell 132 363–374. (doi:10.1016/j.cell. microcarcinoma: a surgical perspective. Cancer Treatment 2007.12.032) Reviews 31 423–438. (doi:10.1016/j.ctrv.2005.04.009) Scurr LL, Pupo GM, Becker TM, Lai K, Schrama D, Haferkamp S, Irvine M, Scolyer RA, Mann GJ & Rizos H Received in final form 16 September 2011 2010 IGFBP7 is not required for B-RAF-induced Accepted 21 September 2011 melanocyte senescence. Cell 141 717–727. (doi:10.1016/ Made available online as an Accepted Preprint j.cell.2010.04.021) 21 September 2011

Downloaded from Bioscientifica.com at 09/30/2021 03:39:20AM via free access www.endocrinology-journals.org 757