Oncogene (2007) 26, 2747–2758 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE CEACAM1 impedes thyroid cancer growth but promotes invasiveness:a putative mechanism for early metastases

W Liu1,2, W Wei1,2, D Winer1,2, A-M Bamberger3, C Bamberger4, C Wagener5, S Ezzat2,6 and SL Asa1,2

1Department of Pathology, University Health Network and Toronto Medical Laboratories, Toronto, Ontario, Canada; 2Ontario Cancer Institute and University of Toronto, Toronto, Ontario, Canada; 3Department of Pathology, University Hospital Hamburg Eppendorf; 4Department of Internal Medicine, University Hospital Hamburg Eppendorf; 5Department of Clinical Chemistry, University Hospital Hamburg Eppendorf, Toronto, Ontario, Canada and 6Department of Medicine, Mount Sinai Hospital, Toronto, Ontario, Canada

CEACAM1, also known as biliary glycoprotein (BGP), Introduction CD66a, pp120 and C-CAM1, is a member of the CEA immunoglobulin superfamily. CEACAM1 is a putative Thyroid carcinomas derived from follicular epithelial tumor suppressor based on diminished expression in some cells represent a model of malignant transformation solid neoplasms such as colorectal carcinoma. However, (Kondo et al., 2006). They comprise a broad spectrum of CEACAM1 is overexpressed in some tumors such as non- neoplastic phenotypes ranging from benign adenomas small cell lung cancer. To clarify the mechanism of action and well-differentiated carcinoma to poorly differen- of this cell adhesion molecule, we studied thyroid tiated thyroid carcinoma (PDTC) and the rare but carcinoma that has a spectrum of morphologies and rapidly lethal anaplastic (or undifferentiated) thyroid variable behavior allowing separation of proliferation carcinoma (ATC) (DeLellis et al., 2004; Kondo et al., from invasion and metastasis. CEACAM1 is expressed in 2006). Well-differentiated carcinoma is one of the most thyroid carcinoma cell lines derived from tumors that rapidly increasing malignancies with papillary thyroid exhibit aggressive behavior. Introduction of CEACAM1 carcinoma (PTC) more common than follicular thyroid into endogenously deficient WRO cells resulted in reduced carcinoma (FTC) in most parts of the world (DeLellis cell cycle progression associated with p21 upregulation et al., 2004). Several genetic molecular abnormalities and diminished Rb phosphorylation. Forced CEACAM1 have been associated with human thyroid carcinoma expression enhanced cell–matrix adhesion and migration (Fagin, 1992; Kondo et al., 2006). Of these, rearrange- and promoted tumor invasiveness. Conversely, small ments of the RET tyrosine kinase (RET/PTC) and interfering RNA (siRNA)-mediated downregulation of activating mutations of BRAF, a mediator of the CEACAM1 expression in MRO cells accelerated cell tyrosine kinase-RAS signaling pathway, are the most cycle progression and significantly enhanced tumor size in common. However, RET rearrangements represent early xenografted mice. CEACAM1 is not appreciably ex- events in thyroid tumorigenesis and are equally repre- pressed in normal thyroid tissue or benign thyroid tumors. sented in incidental micropapillary carcinomas (Piersanti In a human thyroid tissue array, CEACAM1 reactivity et al., 2003) and in clinically significant PTC (Sugg et al., was associated with metastatic spread but not with 1998). Moreover, mouse models carrying these genetic increased tumor size. These findings identify CEACAM1 alterations develop thyroid tumors that do not typically as a unique mediator that restricts tumor growth whereas display invasive or metastatic behavior (Jhiang et al., increasing metastatic potential. Our data highlight a 1996; Powell et al., 1998). Whereas BRAF mutations are complex repertoire of actions providing a putative associated with clinical PTC, they do not consistently mechanism underlying the spectrum of biologic behaviors predict invasion and/or metastasis. Ras mutations have associated with thyroid cancer. been identified in rare well-differentiated thyroid carci- Oncogene (2007) 26, 2747–2758. doi:10.1038/sj.onc.1210077; nomas but are more common in poorly differentiated published online 23 October 2006 tumors that behave aggressively (Ezzat et al., 1996; Garcia-Rostan et al., 2001). Additional genetic events Keywords: CEACAM1; thyroid cancer; cell adhesion; tumor have been identified in PDTC and ATC (Kondo et al., invasion; metastases 2006). However, a major clinical dilemma arises from the diagnosis of well-differentiated thyroid carcinoma, since only a small percentage of patients will develop Correspondence: Dr SL Asa, Department of Pathology, University metastases, but there is no reliable marker to separate of Toronto, 610University Avenue, Toronto, Ontario, Canada those that need more aggressive management from the M5G-2M9. E-mail: [email protected] low-risk cases that will not manifest further sequelae. Received 12 May 2006; revised 31 August 2006; accepted 14 September Thus, the identification of factors involved in modifying 2006; published online 23 October 2006 thyroid cancer behavior remains an important goal. CEACAM1 in thyroid cancer WLiuet al 2748 Many cancers display altered expression patterns of lines representative of the different types of thyroid cell–matrix and cell–cell adhesion molecules. These carcinomas. Using gain- and loss-of-function ap- changes serve to promote tumor progression and proaches in cell lines devoid and endowed with metastatic growth presumably through distinct interac- endogenous CEACAM1, respectively, we examined the tions with surrounding extracellular matrix, control of functional impact of this putative adhesion factor on cell proliferation and migration (Ruoslahti and Obrink, several parameters of growth associated with thyroid 1996). Given the spectrum of biologic behavior dis- cancer behavior. played by thyroid carcinomas, we examined putative cell adhesion molecules in modulating thyroid carcinoma behavior. -related Cell Adhesion Results Molecule 1 (CEACAM1), also known as biliary glycoprotein (BGP), CD66a and pp120, the human CEACAM1 expression in human thyroid carcinoma homolog of the rat C-CAM, is a member of the human cell lines carcinoembryonic antigen family, which belongs to the CEACAM1 expression with two isoforms consistent immunoglobulin superfamily of cell adhesion molecules. with the L-CEACAM1 and S-CEACAM1 were detected Located at 19q13.2 within the cluster of in MRO (follicular) and ARO (anaplastic) thyroid CEA-related (Olsen et al., 1994), CEACAM1 is cancer cells (Figure 1a). No signal or a faintly detectable the most conserved member with one known in signal of the predicted size was identified by reverse humans (Beauchemin et al., 1999). This gene gives rise transcriptase–polymerase chain reaction (RT–PCR) in to alternatively spliced mRNA species with isoforms the less aggressive TPC-1, NPA and WRO cell lines. The differing mostly in their cytoplasmic domains. expression of CEACAM1 was investigated by CEACAM1 is an example of a transmembrane Western blotting. As shown in Figure 1b, CEACAM1 protein widely expressed by epithelial and endothelial protein corresponded with the RT–PCR data, and the cells, and most types of hematopoietic cells tissues (Prall predicted 130kDa protein product of CEACAM1 was et al., 1996; Obrink, 1997; Singer et al., 2000; Plunkett detected in MRO and ARO cells. Expression of and Ellis, 2002). CEACAM1 shows both homophilic CEACAM1 protein was undetectable in TPC-1, NPA adhesion Ocklind, 1984 1446/id as well as heterophilic and WRO cells. Immunolocalization confirmed CEA- binding to other CEA family members in vitro (Oikawa CAM1 protein expression in MRO and ARO cells et al., 1992) and is believed to be an important player in (Figure 1c). maintenance of normal tissue architecture. Several pathogens have also been shown to engage CEACAM1 Functional impact of CEACAM1 on thyroid carcinoma in promoting their invasion and colonization of human cell growth mucosa (Muenzner et al., 2005). CEACAM1-mediated To examine the impact of CEACAM1 on thyroid cancer signaling also promotes network formation on base- cell proliferation in vitro and tumorous growth in vivo, ment membrane and increases endothelial cell motility we established stable lines of CEACAM1-deficient (Muller et al., 2005). WRO cells with forced expression of CEACAM1. We In the current studies, we examined the expression of also examined the impact of attenuation of endogenous CEACAM1 in primary thyroid tumors and clonal cell CEACAM1 expression in MRO cells using stable small

Figure 1 CEACAM1 is expressed in human thyroid carcinoma-derived cell lines. (a) RNA extracted from human thyroid carcinoma- derived cell lines was PCR-amplified using primers for CEACAM1. Products consistent with CEACAM1-L (408 bp) and the shorter CEACAM1-S (355 bp) form were identified in ARO anaplastic and MRO follicular carcinoma cells. The lower panel represents the PGK-1 control. Each lane is followed by the same reaction with omission of the reverse transcriptase. (b) Western immunoblotting of corresponding cell lysates demonstrates CEACAM1 expression in the ARO and MRO cell lines. The actin control is shown below. (c) CEACAM1 expression in ARO and MRO cells was localized to the cell membrane in ARO and MRO cells using immunohistochemistry.

Oncogene CEACAM1 in thyroid cancer WLiuet al 2749 interfering (siRNA)-mediated gene silencing. CEA- (Figure 2c). No significant impact on p27 levels was CAM1 expression was confirmed by Western immuno- evident in response to CEACAM1 manipulation blotting and by immunocytochemistry (Figure 2a). (Figure 2d). To investigate the effect of CEACAM1 on cell proliferation, synchronized WRO cells were examined by flow cytometry for assessment of cell cycle progres- CEACAM1 restores thyroid carcinoma cell–matrix sion. Overexpression of CEACAM1 inhibited the G1/ adhesion S phase transition significantly decreasing the percentage CEACAM1 has been recognized as an intercellular of cell population in S phase (28.372.0vs control adhesion molecule (Obrink, 1997). However, its binding 47.670.4; P ¼ 0.009) and induced G0/G1 phase arrest to extracellular matrix components has not clearly been (61.872.6 vs control 25.770.8; P ¼ 0.002). These cell defined. To investigate the impact of CEACAM1 on cycle changes were accompanied by p21 protein cell–matrix adhesion, we examined the effect of forced accumulation in WRO cells with forced expression of CEACAM1 expression on adhesiveness to a number of CEACAM1 and by p21 protein decrease in MRO extracellular matrix components. CEACAM1 expres- cells expressing CEACAM1 siRNA (Figure 2b) and sion was associated with an overall increase in adhesion a parallel reciprocal effect on Rb phosphorylation to extracellular matrices, an effect that was most evident

Figure 2 CEACAM1 impacts different parameters of thyroid cancer cell growth. (a) Forced expression of CEACAM1 in deficient WRO cells (left) and downregulation using stable siRNA in CEACAM1-expressing MRO cells (right) was established in multiple stable clones. Western immunoblotting and immunohistochemistry of pelleted cells grown in vitro and cells xenografted in vivo confirms sustained CEACAM1 expression status. Synchronized cells were examined for cell cycle control elements including p21 (b), phospho-Rb (c), and p27 (d) by Western blotting of total cell lysates and by immunohistochemistry. WRO control cells were compared with WRO cells with forced CEACAM1 expression and MRO cells with endogenous CEACAM1 were compared with MRO cells expressing CEACAM1 siRNA. In each figure, representative immunohistochemistry is shown over corresponding bar graphs that represent densitometric analyses of expression relative to actin levels derived from three separate experiments; Po0.05 compared to control is denoted by an (*).

Oncogene CEACAM1 in thyroid cancer WLiuet al 2750 when cells were allowed to adhere to a collagen I, CEACAM1 promotes thyroid carcinoma cell invasion collagen IV or laminin matrix (Figure 3a). To investigate the effect of CEACAM1 on cell invasion, we examined independent clones of WRO cells with CEACAM1 promotes thyroid carcinoma cell migration forced CEACAM1 expression using matrigel invasion Given the enhanced cell adhesiveness afforded by assay. These studies demonstrated that CEACAM1 CEACAM1, we investigated the impact of CEACAM1 unequivocally promotes cell invasion (Figure 3d). The on cell migration on collagen IV and laminin surfaces. impact of CEACAM1 under-expression could not be CEACAM1 expression was associated with measurably examined as MRO cells failed to invade through increased migration on both surfaces as depicted in Matrigel. The latter finding suggests that CEACAM1 Figures 3b and c. does not enhance invasion in general.

CEACAM1 modulates tumorous growth in xenografted mice To study the impact of CEACAM1 on tumorous growth in vivo, we used a flank thyroid cancer xenograft model as previously described (Liu et al., 2005). Tumor volume was significantly reduced in WRO cells with forced CEACAM1 expression compared to the control CEA- CAM1-negative cells (Figure 4a). Conversely, CEA- CAM1 silencing in MRO cells resulted in a measurable increase in tumor volume (Figure 4b). However, there was a striking change in tissue invasiveness attributable to CEACAM1 that could not be appreciated from tumor volume assessment alone. Forced expression resulted in smaller but more infiltrative WRO tumor nodules whereas control cells formed larger expansile but noninvasive nodules (Figure 4c); the converse was true of MRO cells where silencing of CEACAM1 resulted in rounder bulging nodules with reduced invasiveness. Immunohistochemical evaluation con- firmed altered CEACAM1 expression in xenografts (Figure 2a), as well as changes in p21 (Figure 4d) but with no impact on p27 (Figure 4d) as identified in vitro.

Figure 3 CEACAM1 increases cell adhesion, migration, and invasion in thyroid cancer cells. (a) Effect of CEACAM1 on thyroid cancer cell-matrix adhesion. WRO control cells devoid of CEACAM1 and WRO cells expressing CEACAM1 were allowed to attach to collagen I, IV, fibronectin and laminin matrix-coated plates. Cells were washed and the number of adherent cells counted with the aid of imaging software as detailed under Materials and methods. The bar graph represents the mean7s.e.m. of three separate experiments, each performed in triplicate. Po0.05 compared to control is denoted by an (*). (b and c) Effect of CEACAM1 on thyroid cancer cell migration. Wound healing assay demonstrates that forced CEACAM1 expression significantly stimulates cell migration on both collagen IV- and laminin-coated plates as indicated. (b) Phase-contract micrographs (5 Â )of scratched monolayers of control and CEACAM1overexpressing WRO cells on collagen IV- and laminin-coated plates at 0, 6 and 12 h. (c) The migrating rate was determined by measuring cell migrating areas on collagen IV- (left) and laminin-coated (right) plates. Data from experiments performed in triplicate are expressed as the mean7s.e.m. of three independent experiments. Asterisks indicate a significant difference from control cells. (d) Effect of CEACAM1 on thyroid carcinoma cell invasion. CEACAM1- expressing WRO and control negative cells were allowed to invade through Matrigel as detailed under Materials and methods. Overexpression of CEACAM1 significantly promotes thyroid cancer cell invasion. Po0.05 compared to control is denoted by an (*).

Oncogene CEACAM1 in thyroid cancer WLiuet al 2751

Figure 4 CEACAM1 attenuates thyroid carcinoma tumor size in xenografted mice but promotes invasion. (a) Flank xenograft experiments were performed using 5 Â 106 WRO cells stably expressing CEACAM1 or their empty vector-transfected control cells (two clones of each) in the flank of 6-week-old SCID mice. The indicated tumor volumes represent the mean7s.e.m. of three independent experiments each with five animals in each group. (b) Flank xenografts introduced 5 Â 106 MRO cells with CEACAM1 downregulation by siRNA; control cells were transfected with scrambled sequence (two clones of each). For all experiments, the indicated tumor volumes represent the mean7s.e.m. of three independent experiments each with five animals in each group. The (*) denotes Po0.05. (c) Control WRO cells grow as round expansile nodules under the muscle of the skin in this xenograft (left); in contrast, WRO cells with forced expression of CEACAM1 result in smaller but more infiltrative tumor nodules that invade through the muscle and into the dermis and epidermis (right). (d) Protein expression of p21 and p27 in mouse xenografts was examined by immunohistochemistry.

CEACAM1 is preferentially expressed in metastasizing (n ¼ 98); P ¼ 0.29. More strikingly, CEACAM1 was human thyroid carcinomas expressed more frequently in metastatic PTC (21/35 or In a human thyroid tissue array, CEACAM1 was not 60%) compared to primary lesions (43/141 or 30%) with identified in normal thyroid or benign neoplasms a statistical significance of P ¼ 0.0016. CEACAM1- (Table 1). Instead, it was preferentially expressed with positive PTCs also showed an increased tendency for a membrane apical localization in a subset of dif- extra-thyroidal extension (ETE): 7/43 ¼ 16% of CEA- ferentiated thyroid malignancies (Figure 5). Primary CAM1-positive vs only 7/98 or 7% of CEACAM1- papillary carcinomas with strong CEACAM1 expres- negative tumors (P ¼ 0.13). Dysregulated cytoplasmic sion tended to be smaller in diameter: 2.1571.37 cm staining was seen in poorly differentiated and anaplastic (n ¼ 19) compared with negative tumors 2.5571.52 cm carcinomas.

Oncogene CEACAM1 in thyroid cancer WLiuet al 2752 Table 1 Summary of characteristics of primary human thyroid specimens Mean age Mean nodule CEACAM1 CEACAM1 No. M F (years)7s.d. Size (cm)7s.d. Weak (%) Strong (%)

Benign samples Follicular adenoma 36 7 29 46713 2.971.8 1 (3) 0(0) Chronic thyroiditis 18 1 17 46712 — 4 (22) 7 (39) Hurthle cell adenoma 7 1 6 47719 2.370.8 0 (0) 0 (0) Normal thyroid 23 7 16 46715 — 0(0) 0(0) Degenerative nodule 6 3 3 55715 1.970.6 0 (0) 0 (0) Nodular hyperplasia 4 1 3 5078 — 0(0) 0(0)

Total benign samples 94

Malignant samples Primary lesions Papillary Ca 141 29 112 43713 2.471.5 24 (17) 19 (13) Follicular Ca 7 1 6 51712 4.371.4 2 (29) 0(0) Medullary Ca 4 2 2 607201.8 70.9 1 (25) 1 (25) Hurthle cell Ca 5 1 4 517103.6 71.2 1 (20) 0 (0) Insular/anaplastic 4 4 066 74 7.575.2 2 (50) 0 (0) Nodal metastases Papillary Ca 35 13 22 41715 — 14 (40) 7 (20) Medullary Ca 3 2 1 4876 — 0(0) 1 (33) Insular/anaplastic 4 3 1 54712 — 2 (50) 2 (50)

Total malignant samples 203

Figure 5 CEACAM1 is preferentially expressed in metastatic primary human thyroid carcinomas. (a) No expression of CEACAM1 was found in normal thyroid tissue, but blood vessels were positive, serving as an internal control. (b) CEACAM1 was expressed with a luminal or apical staining pattern (inset) in papillary carcinomas with lymph node metastases.

Discussion to attenuate tumor size while promoting tissue invasive- ness. The findings unmask a potential novel mechanism The present study was performed to investigate the for identifying metastasizing thyroid carcinomas. expression pattern and functional role of the cell adhesion Cell adhesion molecules are not only involved in molecule CEACAM1 in human thyroid carcinoma. We generating and maintaining normal tissue architecture show here that CEACAM1 retards cell cycle progression and cellular polarity, but also play an important role in

Oncogene CEACAM1 in thyroid cancer WLiuet al 2753 tumor invasion and progression (Hirohashi and Kanai, et al., 2001) impacting cell proliferation and differentia- 2003). Compared with normal tissues, malignant tumors tion. CEACAM1 phosphorylation has been reported in are characterized by disruption of tissue architecture colon and prostate cancers at tyrosine (Tyr-488) and and deranged differentiation. It has been postulated that serine (Ser-503) residues (Izzi et al., 1999; Estrera et al., changes in cell–cell and cell–matrix interactions could 2001; Fournes et al., 2001). Point mutation of Tyr-488 account for the ability of cancer cells to cross tissue was sufficient to reverse the in vivo tumor growth boundaries and to disseminate to distant sites. The loss inhibition while substitution or deletion of Ser-503 leads of cell–cell binding that closely correlates with dif- to reversal of tumor inhibition (Izzi et al., 1999). ferentiation and the invasive potential of malignant Although phosphorylation of CEACAM1 appears to tumors is accompanied by alteration in expression of cell be critical in the control of proliferative functions, adhesion molecules (Pignatelli and Vessey, 1994). mutation of Tyr-488 alone has been described not to Dysregulated CEACAM1 protein expression in parti- abolish tumor suppressive actions (Estrera et al., 2001). cular has been noted in several forms of solid neoplasia Transfection of the cytoplasmic domain alone was including breast (Obrink, 1997; Riethdorf et al., 1997; sufficient to cause tumor suppression (Estrera et al., Plunkett and Ellis, 2002), colorectal (Neumaier et al., 1999) consistent with a putative tumor suppressive role 1993; Brummer et al., 1995), prostate (Kleinerman et al., for CEACAM1-L. 1995; Luo et al., 1999), bladder (Kleinerman et al., 1996), Thyroid cancer represents a unique model to dissect endometrial (Bamberger et al., 1998) and hepatocellular the proliferative and metastatic functions of CEACAM1 carcinomas (Takanishi et al., 1997). More frequently, that may help explain some of the previous paradoxical however, downregulation has been observed compared features noted in other malignancies. Well-differentiated with adjacent non-malignant surrounding structures. thyroid carcinoma is endowed with growth promoting This pattern has suggested a tumor suppressive role for properties that give rise to a clinically palpable nodule. CEACAM1. Indeed introduction into tumor cells Only less than 20% of these nodules manifest metastatic derived from these carcinomas has been shown to slow spread, usually to local lymph nodes. The growth and tumor progression. For example, human prostate and metastatic potentials are not necessarily associated, as colon cancer cells show diminished tumorigenesis when exemplified by occult papillary microcarcinomas that transfected with CEACAM1 (Hsieh et al., 1995; Kunath metastasize. There are currently no histopathological et al., 1995; Luo et al., 1999; Busch et al., 2002). and few immunohistochemical or molecular markers In contrast to the putative tumor suppressive func- that can distinguish these two properties of this tions, CEACAM1 is not expressed in several normal neoplasm. This situation has resulted in a clinical epithelial tissues (Prall et al., 1996), including thyroid as dilemma for the rational therapy of low-risk thyroid shown by our current studies. It has also been found to carcinoma. Our data illustrate a unique expression be upregulated in primary lung carcinomas when pattern for CEACAM1 and a mechanism of action that compared with adjacent normal lung tissue (Ohwada may clarify this discrepancy. While not appreciably et al., 1994; Wang et al., 2000). This upregulation expressed in normal human thyroid tissue and rarely in appears to be associated with increased metastatic benign tumors, CEACAM1 is significantly upregulated potential. Similar upregulation has been noted in gastric in thyroid carcinomas, particularly in metastatic tumors. adenocarcinoma (Kinugasa et al., 1998) and cutaneous Moreover, there is a clear dissociation between CEA- malignant melanoma (Thies et al., 2002). These latter CAM1-mediated metastatic potential and growth that observations have been interpreted as evidence for an may clarify the mechanism of spread in small tumors. oncogenic function for CEACAM1. To investigate the role of CEACAM1 in thyroid The CEACAM1 gene contains nine exons encoding a carcinomas, we chose to use WRO cells, which are transmembrane protein. Exon 1 encodes the leader usually devoid of CEACAM1 expression, to compare sequence, exons 2–5 encode a variable immunoglobulin- the effects of CEACAM1 in control cells and clones like N terminal domain followed by up to three repeats with forced CEACAM1 expression. We also used MRO of immunoglobulin constant-like A or B domains cells with strong endogenous CEACAM1 protein followed by a transmembrane domain encoded by exon expression to examine the effects of silencing of 6 and an alternative cytoplasmic domain (exon 7–9) expression of this protein. In vitro, FACS analysis (Barnett et al., 1993). Insertion of exon 7, comprised of revealed that forced expression of CEACAM1 caused a 53 bp, results in the translation of a longer cytoplasmic significant G0/G1 phase arrest. Moreover, CEACAM1 domain of 71–73 amino acids (CEACAM1-L). The expression resulted in enhanced cell–matrix adhesion as shorter form (CEACAM1-S) of CEACAM1 includes a well as increased cell invasion. The combined impact of cytoplasmic tail of only 10residues. Structural and these properties was associated with diminished tumor functional analyses have suggested that the adhesive progression in xenografted mice. In a xenograft model, activity of CEACAM1 is mediated by the first extra- CEACAM1 expression resulted in diminished tumor cellular while the cytoplasmic growth but increased tumor invasiveness. Conversely, domain is necessary and sufficient for growth regulatory downregulation of CEACAM1 significantly stimulated functions (Fournes et al., 2001). MRO cell tumor growth with reduced invasiveness. CEACAM1 also participates in signal transduction These findings clearly define CEACAM1 as an impor- via the cytoplasmic domain that undergoes phospho- tant inhibitor of cell proliferation and tumor growth but rylation (Izzi et al., 1999; Estrera et al., 2001; Fournes as a mediator of invasion.

Oncogene CEACAM1 in thyroid cancer WLiuet al 2754 Cell cycle progression is regulated by cyclins, cyclin- Branchburg, NJ, USA). The reaction mixture was incubated dependent kinases (CDKs) and CDK inhibitors at 251C for 10min, 48 1C for 30min and 95 1C for 5 min. The (CDKIs). The CDKIs p21 and p27 are implicated in synthesized cDNA was used for PCR amplification or stored thyroid neoplasia (Khoo et al., 2002a, b; Kondo et al., at À201C for further analysis. 2006). To further explore the mechanism of cell cycle PCR primers were designed to amplify the long (CEA- CAM1-L) (408-bp) and short (CEACAM1-S) (355-bp) regulation by CEACAM1, we examined the effect of forms of CEACAM1 cDNA. The forward primer, CEACAM1 on p21 and p27 expression both in cultured 50-GGTTGCTCTGATAGCAGTAG-30,issituatedinthetrans- cells and in mouse xenografts using immunohistoche- membrane domain (exon 6), and the reverse primer, mistry or Western immunoblotting. The results showed 50-AGCCTGGAGATGCCTATTAG-30, is situated within the that CEACAM1 significantly enhanced p21 but not p27 30 untranslated region. The PCR reaction was performed in a expression and conversely downregulation of CEA- 20 ml volume containing 4.0 ml RT reaction mixture of 0.5 mM CAM1 decreased p21 but not p27 expression. These of each primer, 1 Â PCR buffer, 3.0m M MgCl2 and 2.5 U data are in partial agreement with a recent report where AmpliTaq DNA polymerase (PE Applied Biosystems, Foster CEACAM1 depletion was shown to be associated with City, CA, USA). Reaction conditions included initial dena- 1 1 diminished p21 and p27 in mouse colonic epithelium turation at 94 C for 2 min, followed by 35 cycles at 94 C for 1 min, annealing at 551C for 1 min and extension at 721C for (Leung et al., 2006). 1 min, followed by a 10min final extension at 72 1C. PCR In summary, we have demonstrated here that products were separated on 2.0% agarose gels and visualized CEACAM1, a member of the human carcinoembryonic by ethidium bromide staining. antigen family, is principally expressed in aggressive The quality of the total RNA and RT reaction were thyroid carcinomas. However, instead of promoting determined by PCR for the house-keeping gene PGK-1 (sense growth, we show that CEACAM1 functions to retard strand: 50-CAGTTTGGAGCTCCTGGAAG-30 and anti- several parameters of neoplastic growth and to enhance sense strand: 50-TGCAAATTCAGGGTGCAGTG-30). The cell–matrix adhesion. This finding is consistent with the PCR reaction contained 2.0 ml of cDNA, 0.8 mM of each expression of CEACAM1 in extravillous intermediate primer, 1 Â PCR buffer, 2.0m M MgCl2, and 2.5 U AmpliTaq trophoblast where it has been functionally implicated in DNA polymerase (PE Applied Biosystems, Foster City, CA, USA) in a final volume of 20 ml using the following conditions: mediating trophoblast/endometrial interactions during 951C for 5 min, 34 cycles of 941C for 30s, 57 1C for 30s, and trophoblastic invasion of the endometrium (Bamberger 721C for 1 min and a 7 min final extension at 721C. et al., 2000, 2006), and with its correlation with invasion in cutaneous malignant melanoma (Thies et al., 2002; Ebrahimnejad et al., 2004). These findings can be Protein isolation and Western immunoblotting Cells were lysed in lysis buffer (0.5% sodium deoxycholate, reconciled by a model in which upregulation of 0.1% sodium dodecyl sulfate (SDS), 1% Nonidet P-40 and 1 Â CEACAM1 during cell transformation may reflect a PBS) containing proteinase inhibitors (100 mg/ml phenyl- mechanism serving to impede further growth at the methylsulfonyl fluoride (PMSF), 69 mg/ml aprotinin (Sigma, expense of promoting metastatic behavior. St Louis, MO, USA), and 1 mM sodium orthovanadate. Total cell lysates were incubated on ice for 30min, followed by micro-centrifugation at 10 000 g for 10min at 4 1C. Protein Materials and methods concentrations of the supernatants were determined by the Bio-Rad protein assay. Cell lines and cell culture Equal amounts of protein (50 mg) were mixed with 2 Â SDS TPC-1 cells (a gift from Dr SM Jhiang, Ohio State University, sample buffer, boiled for 4 min and separated by 10% SDS– Columbus, OH, USA), derived from a well-differentiated polyacrylamide gel electrophoresis, and transferred onto papillary thyroid carcinoma with a ret/PTC gene rearrange- nitrocellulose membranes (Bio-Rad laboratories, Hercules, ment, were cultured in Dulbecco’s modified Eagle’s medium CA, USA). Non-specific binding was blocked with 5% nonfat (DMEM) supplemented with 5% FBS, 2 mmol/l L-glutamine. milk in 1 Â TBST (tris-buffered saline with 0.1% Tween-20). NPA cells, a papillary carcinoma cell line with homozygous Primary antibodies and antisera were directed against CEA- BRAF mutations, WRO and MRO follicular carcinoma cell CAM1 (monoclonal 4D1/C2) (Stoffel et al., 1993) (1:1000), lines and ARO anaplastic thyroid carcinoma cell line were p21 (polyclonal 1:500, BD Biosciences, Mississauga, Canada), obtained from Dr J Fagin (University of Cincinnati, Cincin- p27 (monoclonal 1:1000, Transduction Laboratories, Lexing- nati, OH, USA) and were maintained in RPMI 1640 ton, KY, USA), phospho-Rb (Ser780, polyclonal 1:1000, supplemented with 10% FBS, 2 mmol/l L-glutamine, 1 mmol/ Cell Signaling Technology, Beverley, MA, USA) or actin l sodium pyruvate, and 1 Â nonessential amino acid (Sigma- (monoclonal 1:500, Sigma, St Louis, MO, USA). After Aldrich, Irvine, UK). Cells were all cultured in a standard washing three times each for 10min in 1 Â TBST, blots were exposed to the secondary antibody (anti-mouse or anti-rabbit humidified incubator at 371C in a 5% CO2 atmosphere. Cell viability was assessed using Trypan blue exclusion before and IgG-HRP, Santa Cruz Biotechnology Inc., Santa Cruz, CA, after all experiments. USA) at a dilution of 1:2000 and visualized using ECL chemiluminescence detection system (Amersham, Buckinghamshire, UK). Band intensities were quantified by RNA extraction and analysis densitometry. Total RNA was isolated from cells with TriZol reagent (Invitrogen Corp., Carlsbad, CA, USA) according to the manufacturer’s instructions. Approximately 1.5 mg of total CEACAM1 silencing or overexpression in human thyroid RNA from each sample was used to conduct reverse carcinoma cells transcription reaction in a 30 ml volume using TaqMan reverse The hairpin siRNAs of human CEACAM1 were designed transcription reagents (Applied Biosystems Inc., using Ambion software according to the manufacturer’s

Oncogene CEACAM1 in thyroid cancer WLiuet al 2755 suggestions. The pSilencer 2.1-U6 neo-vector was linearized Migration assay with BamH1 and HindIII to facilitate directional cloning. The Control and CEACAM1-transfected WRO (0.5 Â 106) cells CEACAM1 siRNA target sequence was 50-AACCGT were plated on collagen IV- and laminin-coated six-well plates CAAATTGTAGGATAT-30 (nucleotide positions 226–246). (BD Biosciences). After reaching confluence, cells were gently The double-stranded oligonucleotide templates containing the scratched with a standard 200 ml tip, rinsed three times, and sequence of the 21-mer siRNA were synthesized as follows: returned to cell incubator after growth medium was replaced. forward strand: 50-GATCCGCCGTCAAATTGTAGGATAT Cells were permitted to migrate into the area of clearing and TCTGCAGGAATATCCTACAATTTGACGGTTTTTTGG photomicrographied at 0, 6 and 12 h using an Axiovert 200 M AAA-30 and reverse strand: 50-AGCTTTTCCAAAAAACCG phase-contrast microscope (Zeiss Axiovert S100TV, Leipzid, TCAAATTGTAGGATATTCCTGCAGAATATCCTACAA Germany). Migrating areas were analysed using Image-Pro TTTGACGGCG-30. The expression vector pSilencer 2.1-U6 Plus Software. neo-siCEACAM1 was constructed according to the manufac- turer’s instructions. Ligation of the vector and oligo was Invasion assay performed with 1 ml of vector, 1 mlofoligo,2mlof10Â T4 DNA BioCoat Matrigel Invasion Chambers (BD Biosciences) were ligase buffer, 3 ml of T4 DNA ligase for 5 min at room allowed to reach room temperature. Warm (371C) RPMI 1640 temperature. Three microlitres of ligation reaction was then medium was added to the interior of the inserts (0.5 ml/each) used to transform DH5a Esherichia coli. Clones containing the and bottom of the wells (0.5 ml/well). Rehydration for 2 h in oligo inserts were identified by restriction enzyme digestion humidified culture incubator (371C, 5% CO2 atmosphere) was and by DNA sequencing. Plasmid DNA for transfection was followed by careful removal of media without disturbing the prepared using the QIAGEN Maxi Kit (QIAGEN, Mississauga, layer of Matrigel Matrix on the membrane. Canada). After trypsinization, an equal number (2.0 Â 104) of WRO WRO cells which did not endogenously express CEACAM1 cells expressing CEACAM1 and control cells in 0.5 ml of were selected for stable forced expression of CEACAM1-L serum-free RPMI 1640medium containing 0.2%BSA were using the pcDNA 3.1(À)/Zeo-L-CEACAM1 expression vector plated in each insert. The growth medium containing 5% FBS as described previously (Briese et al., 2005). Empty vector was used as a chemoattractant in the bottom well. served as a control. Conversely, MRO cells which did express Following 24 h of incubation, non-invading cells were endogenous CEACAM1 were transfected with CEACAM1 removed from the upper surface of the membrane by siRNA following the Lipofectamine protocol as recommended scrubbing with a cotton swab. Cells on the lower surface of by the manufacturer (Invitrogen). A circular pSilencer2.1-U6 the membrane were stained with Diff-Quik Stain (BD neo-vector that expresses a hairpin siRNA without homology Biosciences). Staining was accomplished by sequentially to any known sequences in the human, mouse and rat genomes transferring the inserts through three solutions. After air- was used as a negative control. Cell populations that stably drying, the membrane was removed from the insert and express the siRNA for sustained knockdown experiments were invading cells photographed under the microscope for identified by neomycin selection. Specific forced expression quantitation. and reduction of targeted CEACAM1 expression was con- firmed by Western blotting analysis with the anti-CEACAM1 antibody 4D1/C2 (Stoffel et al., 1993). Mouse xenograft studies To assess tumor formation, 5 Â 106 cells were injected subcutaneously into the flank of 6-week-old SCID mice. Cell cycle analysis Tumor volume was monitored every 5 days as previously For cell cycle assessment cells were starved in serum-free described (Liu et al., 2005). Mice were killed 21 days following growth medium for 24 h before being exposed to growth tumor cell implantation and the tumors were excised, weighed medium containing 10% FBS for another 24 h. After and measured. Excised tumor tissue was fixed in formalin and trypsinization, cells were washed with D-PBS (Ca2 þ -Mg2 þ embedded in paraffin for microscopy and immunohisto- -free) and fixed with cold 80% ethanol for 1 h on ice. Fixed chemical staining. Animal handling protocol was approved by cells were washed with staining buffer (0.2% Triton X-100 and the Ontario Cancer Institute Animal Care and Utilization 1mM/l EDTA, pH 8.0, in PBS) and resuspended in the staining Committee. buffer containing 50 mg/ml DNase-free RNase A (Sigma) and 50 mg/ml propidium iodide for 1 h. Cell-cycle analysis was Human thyroid tumor specimens performed by fluorescence-activated cell sorting (FACS; Becton Dickinson, San Jose, CA, USA) using Cellquest Formalin-fixed, paraffin-embedded tissues from 230patients analysis and specific S phase was analysed using Modfit who underwent thyroid surgery at the University Health DNA Analysis (Verity Software House Inc., Topsham, ME, Network (UHN) Hospitals between 2001 and 2004 were used USA). for tissue array construction. From the surgical specimens of these 230patients, a total of 297 distinct samples were obtained, including 94 benign samples and 203 malignant Adhesion assay samples (Table 1). Each patient was represented by one tumor After trypsinization, an equal number (1 Â 105) of WRO cells sample, with 67 additional samples taken from these patients expressing CEACAM1 and control cells were plated in each for the complete (n ¼ 230 þ 67 ¼ 297) array assessment. The well of 24-well plates coated with collagen I, collagen IV, additional samples were lymph node metastases of 35 patients fibronectin or laminin (BD Biosciences, Bedford, MA, USA) with primary tumors in the array, and multiple lymph node and incubated for 1 h, then rinsed with PBS four times. The metastases of two patients, two different neoplastic nodules in remaining cells in each well were fixed in 10% buffered seven patients with multifocal disease, normal thyroid tissue in formalin for 20min at room temperature, washed in PBS and 20patients (five benign, 15 malignant), foci of nodular the entire field counted with a stereomicroscope. All images hyperplasia in two patients, and anaplastic metastasis from a were recorded and analysed using Image-Pro Plus Software primary papillary carcinoma. Data on tumor characteristics (version 4.5; Media Cybernetics Inc., Silversprings, MD, and prognostic factors were obtained from the surgical USA). pathology reports. All patient samples and related data

Oncogene CEACAM1 in thyroid cancer WLiuet al 2756 analyses were performed in accordance with research ethics Labeling Reagent (ID Labs Inc.). Color development was approval at the University Health Network in Toronto. performed with freshly prepared NovaRed solution (Vector Labs Inc.) and counterstained with Mayer’s hematoxylin. Thyroid tissue array Finally, sections were dehydrated through graded alcohols, The regions of interest were carefully identified on H&E- cleared in xylene and mounted in Permount (Fisher, Ontario, stained slides, and then marked directly on the corresponding Canada). donor blocks. One mm diameter cylindrical cores were taken from each donor block and arrayed at 2.0mm intervals (from core center to center) using a manual tissue arrayer (Beecher Tissue array scoring Instruments, Silver Spring, MD, USA). Three cores for each Positive staining was based on the distribution and intensity of desired sample (benign or malignant) were arrayed, yielding a distinct membrane or cytoplasmic signals in each core using the following criteria. The percentage of cells staining was total of 297 Â 3 or 891 cores. Ten recipient blocks were constructed. Normal thyroid controls were represented in all scored as 0 (0%), 1 (1–50%) or 2 (51–100%), and then blocks. averaged for each sample to form a distribution score. Next, the intensity of signal was scored in each core as 0(no signal), 1 (weak), 2 (moderate) or 3 (marked) and then averaged for Immunohistochemistry each sample to form an intensity score. The total of the 2 þ 2 þ Cultured cells were washed in cold D-PBS (Ca -Mg -free) distribution score and intensity score was summed for each three times, gently scraped off and centrifuged into pellets that sample into a total score (TS), which was indicated as negative were coated in 2% bactoagar until solidified. These pellets (TSp2), weakly positive (2oTSp3) or positive (TS>3). were fixed in 10% formalin, and embedded in paraffin. In samples where one of the three cores was not Four-micron-thick sections cut from cell pellets, harvested interpretable (6/297 samples ¼ 2.0%, due to either insufficient xenografted tumors or paraffin-embedded tissue microarray cells or loss of the core during array slide processing), the blocks were de-waxed in five changes of xylene and rehydrated distribution and intensity scores were based on the average of through graded alcohols into water. Sections were then the remaining two cores. All cores were scored by one microwave-heated in 10m M citrate buffer at pH 6.0inside a investigator (DW) and verified by a second independent pressure cooker as previously optimized. Endogenous per- investigator (SLA). oxidase and biotin activities were blocked, respectively, using 3% hydrogen peroxide and Vector’s avidin/biotin blocking kit (Vector Laboratories Inc., Burlingame, CA, USA). Sections Statistical analysis were treated for 10min with protein blocker (ID Labs Inc., Data were assessed by Student’s t-test. Differences in two London, ON, Canada) and then incubated overnight with proportions were calculated using the Fisher’s exact test. primary antibody or antiserum. The primary reactions were as Significance level for all tests was assigned at Po0.05. follows: anti-CEACAM1 mouse monoclonal antibody (4D1/ C2) at 1:200; monoclonal anti-MIB-1 antibody (Immunotech, Marseilles, France) at 1:600; p21 (BD Biosciences) at 1:300; Acknowledgements p27 (BD Transduction Laboratories) at 1:1000. This incuba- tion was followed by 30min each with biotinylated linking This work was supported by the Toronto Medical Labora- reagent (ID Labs Inc.) and HRP-conjugated Ultra Streptavidin tories and the Rita Banach Thyroid Cancer Research Fund.

References

Bamberger AM, Minas V, Kalantaridou SN, Radde J, colocalized with the adhesion molecule CEACAM1 in Sadeghian H, Loning T et al. (2006). Corticotropin-releasing the extravillous trophoblast of the human placenta and hormone modulates human trophoblast invasion through enhances invasion of CEACAM1-expressing placental cells. carcinoembryonic antigen-related cell adhesion molecule-1 J Clin Endocrinol Metab 90: 5407–5413. regulation. Am J Pathol 168: 141–150. Brummer J, Neumaier M, Gopfert C, Wagener C. (1995). Bamberger AM, Riethdorf L, Nollau P, Naumann M, Association of pp60c-src with biliary glycoprotein (CD66a), Erdmann I, Gotze J et al. (1998). Dysregulated expression an adhesion molecule of the carcinoembryonic antigen of CD66a (BGP, C-CAM), an adhesion molecule of family downregulated in colorectal carcinomas. Oncogene the CEA family, in endometrial cancer. Am J Pathol 152: 11: 1649–1655. 1401–1406. Busch C, Hanssen TA, Wagener C, OBrink B. (2002). Down- Bamberger AM, Sudahl S, Loning T, Wagener C, Bamberger CM, regulation of CEACAM1 in human prostate cancer: Drakakis P et al. (2000). The adhesion molecule CEACAM1 correlation with loss of cell polarity, increased proliferation (CD66a, C-CAM, BGP) is specifically expressed by the rate, and Gleason grade 3 to 4 transition. Hum Pathol 33: extravillous intermediate trophoblast. Am J Pathol 156: 290–298. 1165–1170. DeLellis RA, Lloyd RV, Heitz PU, Eng C. (2004). Pathology Barnett TR, Drake L, Pickle W. (1993). Human biliary and Genetics of Tumours of Endocrine Organs, WHO glycoprotein gene: characterization of a family of novel Classification of Tumours. IARC Press: Lyons, France. alternatively spliced RNAs and their expressed . Mol Ebrahimnejad A, Streichert T, Nollau P, Horst AK, Wagener C, Cell Biol 13: 1273–1282. Bamberger AM et al. (2004). CEACAM1 enhances invasion Beauchemin N, Draber P, Dveksler G, Gold P, Gray-Owen S, and migration of melanocytic and melanoma cells. Am J Grunert F et al. (1999). Redefined nomenclature for Pathol 165: 1781–1787. members of the carcinoembryonic antigen family. Exp Cell Estrera VT, Chen DT, Luo W, Hixson DC, Lin SH. (2001). Res 252: 243–249. Signal transduction by the CEACAM1 tumor suppressor. Briese J, Oberndorfer M, Patschenik C, Schulte HM, Phosphorylation of serine 503 is required for growth- Makrigiannakis A, Loning T et al. (2005). Osteopontin is inhibitory activity. J Biol Chem 276: 15547–15553.

Oncogene CEACAM1 in thyroid cancer WLiuet al 2757 Estrera VT, Luo W, Phan D, Earley K, Hixson DC, Lin SH. Leung N, Turbide C, Olson M, Marcus V, Jothy S, (1999). The cytoplasmic domain of C-CAM1 tumor Beauchemin N. (2006). Deletion of the carcinoembryonic suppressor is necessary and sufficient for suppressing the antigen-related cell adhesion molecule 1 (Ceacam1) gene tumorigenicity of prostate cancer cells. Biochem Biophys Res contributes to colon tumor progression in a murine model of Commun 263: 797–803. carcinogenesis. Oncogene 25: 5527–5536. Ezzat S, Zheng L, Kolenda J, Safarian A, Freeman JL, Liu W, Asa SL, Ezzat S. (2005). 1{alpha},25-dihydroxyvitamin Asa SL. (1996). Prevalence of activating ras mutations D3 targets PTEN-dependent fibronectin expression to in morphologically characterized thyroid nodules. Thyroid 6: restore thyroid cancer cell adhesiveness. Mol Endocrinol 409–416. 19: 2349–2357. Fagin JA. (1992). Genetic basis of endocrine disease 3. Luo W, Tapolsky M, Earley K, Wood CG, Wilson DR, Molecular defects in thyroid gland neoplasia. J Clin Logothetis CJ et al. (1999). Tumor-suppressive activity of Endocrinol Metab 75: 1398–1400. CD66a in prostate cancer. Cancer Gene Ther 6: 313–321. Fournes B, Sadekova S, Turbide C, Letourneau S, Muenzner P, Rohde M, Kneitz S, Hauck CR. (2005). Beauchemin N. (2001). The CEACAM1-L Ser503 residue is CEACAM engagement by human pathogens enhances cell crucial for inhibition of colon cancer cell tumorigenicity. adhesion and counteracts bacteria-induced detachment of Oncogene 20: 219–230. epithelial cells. J Cell Biol 170: 825–836. Garcia-Rostan G, Camp RL, Herrero A, Carcangiu ML, Muller MM, Singer BB, Klaile E, Obrink B, Lucka L. (2005). Rimm DL, Tallini G. (2001). Beta-catenin dysregulation Transmembrane CEACAM1 affects integrin-dependent in thyroid neoplasms: down-regulation, aberrant nuclear signaling and regulates extracellular matrix protein-specific expression, and CTNNB1 exon 3 mutations are markers morphology and migration of endothelial cells. Blood 105: for aggressive tumor phenotypes and poor prognosis. Am J 3925–3934. Pathol 158: 987–996. Neumaier M, Paululat S, Chan A, Matthaes P, Wagener C. Hirohashi S, Kanai Y. (2003). Cell adhesion system and (1993). Biliary glycoprotein, a potential human cell adhesion human cancer morphogenesis. Cancer Sci 94: 575–581. molecule, is down-regulated in colorectal carcinomas. Proc Hsieh JT, Luo W, Song W, Wang Y, Kleinerman DI, Van NT Natl Acad Sci USA 90: 10744–10748. et al. (1995). Tumor suppressive role of an androgen- Obrink B. (1997). CEA adhesion molecules: multifunctional regulated epithelial cell adhesion molecule (C-CAM) in proteins with signal-regulatory properties. Curr Opin Cell prostate carcinoma cell revealed by sense and antisense Biol 9: 616–626. approaches. Cancer Res 55: 190–197. Ohwada A, Takahashi H, Nagaoka I, Kira S. (1994). Biliary Izzi L, Turbide C, Houde C, Kunath T, Beauchemin N. (1999). glycoprotein mRNA expression is increased in primary lung cis-Determinants in the cytoplasmic domain of CEACAM1 cancer, especially in squamous cell carcinoma. Am J Respir responsible for its tumor inhibitory function. Oncogene 18: Cell Mol Biol 11: 214–220. 5563–5572. Oikawa S, Kuroki M, Matsuoka Y, Kosaki G, Nakazato H. Jhiang SM, Sagartz JE, Tong Q, Parker-Thronburg J, Capen (1992). Homotypic and heterotypic Ca(++)-independent CC, Cho JY et al. (1996). Targeted expression of the ret/ cell adhesion activities of biliary glycoprotein, a PTC1 oncogene induces papillary thyroid carcinomas. member of carcinoembryonic antigen family, expressed Endocrinology 137: 375–378. on CHO cell surface. Biochem Biophys Res Commun 186: Khoo ML, Beasley NJ, Ezzat S, Freeman JL, Asa SL. (2002a). 881–887. Overexpression of cyclin D1 and underexpression of p27 Olsen A, Teglund S, Nelson D, Gordon L, Copeland A, predict lymph node metastases in papillary thyroid carcino- Georgescu A et al. (1994). Gene organization of the ma. J Clin Endocrinol Metab 87: 1814–1818. pregnancy-specific glycoprotein region on human chromo- Khoo ML, Freeman JL, Witterick IJ, Irish JC, Rotstein LE, some 19: assembly and analysis of a 700-kb cosmid contig Gullane PJ et al. (2002b). Underexpression of p27/Kip in spanning the region. Genomics 23: 659–668. thyroid papillary microcarcinomas with gross metastatic Piersanti M, Ezzat S, Asa SL. (2003). Controversies in disease. Arch Otolaryngol Head Neck Surg 128: 253–257. papillary microcarcinoma of the thyroid. Endocr Pathol 14: Kinugasa T, Kuroki M, Takeo H, Matsuo Y, Ohshima K, 183–191. Yamashita Y et al. (1998). Expression of four CEA family Pignatelli M, Vessey CJ. (1994). Adhesion molecules: antigens (CEA, NCA, BGP and CGM2) in normal and novel molecular tools in tumor pathology. Hum Pathol 25: cancerous gastric epithelial cells: up-regulation of BGP and 849–856. CGM2 in carcinomas. Int J Cancer 76: 148–153. Plunkett TA, Ellis PA. (2002). CEACAM1: a marker Kleinerman DI, Dinney CP, Zhang WW, Lin SH, Van NT, with a difference or more of the same? J Clin Oncol 20: Hsieh JT. (1996). Suppression of human bladder cancer 4273–4275. growth by increased expression of C-CAM1 gene in an Powell Jr DJ, Russell J, Nibu K, Li G, Rhee E, Liao M et al. orthotopic model. Cancer Res 56: 3431–3435. (1998). The RET/PTC3 oncogene: metastatic solid-type Kleinerman DI, Troncoso P, Lin SH, Pisters LL, Sherwood ER, papillary carcinomas in murine thyroids. Cancer Res 58: Brooks T et al. (1995). Consistent expression of an epithelial 5523–5528. cell adhesion molecule (C-CAM) during human prostate Prall F, Nollau P, Neumaier M, Haubeck HD, Drzeniek Z, development and loss of expression in prostate cancer: Helmchen U et al. (1996). CD66a (BGP), an adhesion implication as a tumor suppressor. Cancer Res 55: molecule of the carcinoembryonic antigen family, is ex- 1215–1220. pressed in epithelium, endothelium, and myeloid cells in a Kondo T, Ezzat S, Asa SL. (2006). Pathogenetic mechanisms wide range of normal human tissues. J Histochem Cytochem in thyroid follicular-cell neoplasia. Nat Rev Cancer 6: 44: 35–41. 292–306. Riethdorf L, Lisboa BW, Henkel U, Naumann M, Wagener C, Kunath T, Ordonez-Garcia C, Turbide C, Beauchemin N. Loning T. (1997). Differential expression of CD66a (BGP), a (1995). Inhibition of colonic tumor cell growth by biliary cell adhesion molecule of the carcinoembryonic antigen glycoprotein. Oncogene 11: 2375–2382. family, in benign, premalignant, and malignant lesions

Oncogene CEACAM1 in thyroid cancer WLiuet al 2758 of the human mammary gland. J Histochem Cytochem 45: Sugg SL, Ezzat S, Rosen IB, Freeman J, Asa SL. (1998). 957–963. Distinct multiple ret/PTC gene rearrangements in multifocal Ruoslahti E, Obrink B. (1996). Common principles in cell papillary thyroid neoplasia. J Clin Endocrinol Metab 83: adhesion. Exp Cell Res 227: 1–11. 4116–4122. Singer BB, Scheffrahn I, Obrink B. (2000). The tumor growth- Takanishi K, Miyazaki M, Ohtsuka M, Nakajima N. (1997). inhibiting cell adhesion molecule CEACAM1 (C-CAM) is Inverse relationship between P-glycoprotein expression differently expressed in proliferating and quiescent epithelial and its proliferative activity in hepatocellular carcinoma. cells and regulates cell proliferation. Cancer Res 60: Oncology 54: 231–237. 1236–1244. Thies A, Moll I, Berger J, Wagener C, Brummer J, Schulze HJ Stoffel A, Neumaier M, Gaida FJ, Fenger U, Drzeniek Z, et al. (2002). CEACAM1 expression in cutaneous malignant Haubeck HD et al. (1993). Monoclonal, anti-domain and melanoma predicts the development of metastatic disease. anti-peptide antibodies assign the molecular weight 160 000 J Clin Oncol 20: 2530–2536. granulocyte membrane antigen of the CD66 cluster to a Wang L, Lin SH, Wu WG, Kemp BL, Walsh GL, Hong WK mRNA species encoded by the biliary glycoprotein gene, et al. (2000). C-CAM1, a candidate tumor suppressor gene, a member of the carcinoembryonic antigen gene family. is abnormally expressed in primary lung cancers. Clin Cancer J Immunol 150: 4978–4984. Res 6: 2988–2993.

Oncogene