Papillary thyroid carcinoma (PTC) is a well-differentiated endocrine malignant tumor that develops from thyroid follicular epithelium. The tumor represents the most common type of endocrine malignancy; however, its tumorigenesis is not fully elucidated. The aim of this study was to address the functional role of the sorting nexin (SNX) family in PTC because of recent experimental evidence suggesting that the SNX family members actively control endocytotic transportation as well as cell fate. Expression profiles of SNX family members of PTC showed a significant quantity of transcripts of SNX5. Further immunohistochemical analysis with an SNX5-specific monoclonal antibody established in this study consistently demonstrated the preferential expression of SNX5 in PTC (94.2%, 113/120 cases) as indicated by studies on 440 cases of various tumors. In contrast, other major carcinomas originating from the lung (2.6%, 1/38 cases), breast (5.1%, 2/39 cases), and intestine (4.2%, 1/24 cases) scarcely expressed SNX5. When we investigated models of murine thyroid tumors induced by the administration of carcinogens, high expression of Snx5 was also observed in well-differentiated thyroid tumors, further implying that the tumorigenesis of the thyroid gland was tightly associated with the abundance of SNX5/Snx5. Moreover epithelial cells expressing excess SNX5 showed high levels of Caspase-2 of an initiator caspase. Collectively these findings suggest that the evaluation of SNX5 expression would support pathological diagnosis of primary and secondary PTC. (Cancer Sci 2012; 103: 1356–1362)
The thyroid gland, controlling energy production and many metabolic pathways, is the most common site for the development of malignant tumors among a variety of endocrine organs. The proportion of malignant thyroid tumors has steadily increased over the last three decades.[2-4] Most thyroid tumors originate from thyroid follicular epithelial cells known as thyrocytes and exhibit various histopathological subtypes, of which papillary thyroid carcinoma (PTC) comprises the predominant subtype, with a female:male ratio of about 3:1. While PTC generally has a favorable prognosis, the tumor can potentially metastasize to regional lymph nodes, the lung and other organs.[5-8] Pre-existing benign thyroid lesions and ionizing radiation are known risk factors, and gene alterations such as BRAF and RAS point mutations, and RET/PTC and TRK gene rearrangements have been reported in PTC.[9-13] Gene regulatory factors making critical contributions during the development of thyrocytes are of diagnostic value for PTC in the pathologic laboratory, including thyroid transcription factors including TTF-1 and TTF-2, a hematopoietically expressed homeobox (HHEX), and paired box gene-8 (PAX8).[14-17] However, the etiology of the tumor development has not been fully clarified.
Thyrocytes synthesize the thyroid hormones through a multiple intracellular process coordinated by thyroid-stimulating hormone. During the process, a transcytotic pathway of thyrocytes plays an important role as suggested by anatomic examinations and other studies.[18-20] Once an iodinated glycoprotein of thyroglobulin is synthesized, endosomes convey it into a luminal area surrounded by thyrocytes. Then the thyrocytes, if required, retrieve thyroglobulin molecules from the primary colloidal storage and liberate triiodothyronine (T3) and thyroxin (T4) through a lysosomal pathway using recycling endosomes.[21-23] Recent studies on the membrane-associated traffic system of endosomes have revealed a unique role of sorting nexin (SNX) retromer family members.[24-26] The SNX family has the capacity to bind phopsphatidylinositol phosphate of the lipid bilayer of endosomes through their signature moieties of the Phox-homology (PX) domain. Once such a membrane is recognized, various effector domains of SNX molecules characterize the subsequent process of the membrane compartment. Within various such domains, a Bin/Amphiphysin/Rvs167 (BAR) domain helps SNX molecules (called SNX-BAR) form a banana-shaped structure fitting the curvature of small vesicles, whose domain is also shared by non-SNX molecules involved in Alzheimer's disease and diabetes mellitus.[27-30] In addition to these functions, accumulating evidence reveals a more fundamental function of such SNX-BARs, regulating signal transduction and growth activities, to control epithelial cell integrity.[31-35] These facts led us to hypothesize a possible function of the SNX family in the tumor biology related to the development of PTC.
In this study we first demonstrated preferential expression of SNX5 of an SNX-BAR molecule in PTC as assessed by immunohistochemistry on 440 tumor cases. Murine models of PTC showing Snx5 overexpression in the tumors further supported importance of SNX5 in the pathogenesis of PTC. More interestingly Caspase-2 as an initiator caspase would be under the control of SNX5, implying that an SNX5-Caspase-2 axis might have a pivotal role of the development of PTC.
Materials and Methods
Tissues and cell culture
Thyroid cancer tissues were obtained from patients undergoing thyroidectomy in Sapporo Medical University Hospital and Muroran City General Hospital in Japan. All human materials analyzed in this study were obtained with informed consent and the approval of the institutional review board in each hospital. For primary culture of tumor cells, tissues were minced into small pieces and dispersed in PBS containing 0.7 mg/mL Blendzyme 3 (Roche, Basel, Switzerland) and 0.4 mg/mL DNase I (Sigma-Aldrich, St. Louis, MO, USA) as described previously. After washing the cells three times with PBS, cells were cultured in RPMI1640 (Sigma-Aldrich) supplemented with 10% heat-inactivated fetal calf serum, 50 μg/mL streptomycin and 100 units/mL penicillin. Human embryo kidney (HEK) 293 cells and 8505c PTC cells were obtained from RIKEN Bioresource Center (Japan) and maintained in modified DMEM (Sigma-Aldrich) supplemented with the same reagents as described above. All cells were cultured at 37°C in a humidified atmosphere in 5% CO2.
Reverse transcription-polymerase chain reaction (RT-PCR) was conducted to detect transcripts as previously reported. Primer pairs were summarized in Table S1. The PCR cycling conditions were as follows: 95°C for 1 min; 60°C for 1 min; 72°C for 1 min with 25 cycles. Quantitative RT-PCR was performed as described in the manufacturer's protocol for Assays-on-Demand Gene Expression products (Applied Biosystems, Foster City, CA, USA). To compare the levels of transcripts, the ∆∆CT method was used to analyze triplicate specimens according to the manufacturer's instructions.
Antibodies, immunohistochemistry, and immunoblotting
A mouse anti-human SNX5 monoclonal antibody (clone 48C2; IgG2a subclass) was established per standard procedures by immunizing mice with recombinant SNX5 protein produced in bacterial BL21 cells containing the pET expression vector (Merck KGaA, Darmstadt, Germany) harboring human SNX5 cDNA, which was initially obtained from human epidermal HaCaT cells. A mouse anti-enhanced green fluorescent protein (EGFP) mAb (JL-8; Clontech, Mountain View, CA, USA) was used for detecting EGFP-tagged proteins in immunoblot analysis. A mouse anti-TTF-1 mAb (SPT24; Nichirei, Tokyo, Japan) and rabbit anti-thyroglobulin pAb (DAKO, Copenhagen, Denmark) were used for immunohistochemistry. For studying cells by immunoblotting, antibody sampler kits were used to analyze molecules regulating apoptosis, DNA repair or cell cycle (BD Biosciences, San Diego, CA, USA). For detecting Caspase-2, antibodies of mouse mAb (clone 35) were used provided with the sampler kit as well as rabbit pAb (poly6340) purchased from Biolegend (San Diego, CA, USA). Immunofluorescence, immunohistochemistry and immunoblotting were performed as previously described.[36, 37] Immunofluorescent signals were detected under an immunofluorescence microscope (IX71; Olympus, Tokyo, Japan) or confocal laser microscope (R2100AG2; Bio-rad, Hercules, CA, USA). To obtain concordant results regarding the immunohistochemical expression of SNX5, the slides were examined on a multiheaded microscope by three investigators. The staining profile of SNX5 of tissue sections was graded in accordance with positive-staining areas as follows: less than 10% areas; (−), 10–50% areas; (+), over 50% areas; (++).
Animal models of thyroid tumors
Thyroid carcinomas were chemically induced in BALB/c female mice 6 weeks of age as described previously. In brief, tumors were initiated with a single subcutaneous injection of N-bis(2-hydroxypropyl)-nitrosamine (DHPN; Toronto Research Chemicals, Toronto, Canada) at 2800 mg/kg body weight. One week later, drinking water containing 0.1% sulfadimethoxine (SDM; Sigma-Aldrich) was provided ad libitum for up to 12 weeks. Tissue specimens including normal thyroid gland and tumors around the trachea were obtained using forceps and scissors for microsurgery under a binocular wide-field dissecting microscope. All of the experiments using mice were performed in accordance with the institutional guidelines for the care and use of animals.
Cell transformation and cell proliferation assay
Expression vectors of pCMV-HA and pEGFP-C2 (Clontech) were used to transform HEK 293 cells. Transformation of cells was performed with LF2000 (Invitrogen, Carlsbad, CA, USA) following the manufacturer's protocol. For retrovirus-mediated gene transfer into 8505c cells, pLVSIN-CMV-puro vector harboring SNX5 cDNA was transfected into Lenti-X 293T cells using Lenti-X HTX Packaging System as described in the manufacturer's protocol (Takara, Tokyo, Japan). After transfection, cells were maintained in complete medium containing 1 μg/mL puromycin (Sigma-Aldrich). Growth activities of cells were investigated using a premix WST-1 cell proliferation assay system following the manufacturer's instructions (Takara).
Statistical significance was determined using the unpaired t-test and P-values of less than 0.05 were considered significant. Values were expressed as the mean ± standard deviation (SD).
High expression of SNX5 in papillary thyroid carcinoma
Prior to starting this study, we conducted RT-PCR analysis to investigate which SNX family members were dominantly expressed in PTC. Examinations of tissue specimens from three PTC cases revealed that the transcripts of certain types of SNXs were indeed detected at various levels (Fig. 1a). Among those we examined, transcripts of SNX5 were most abundantly presented in the tumors. To a lesser extent, transcripts of SNX1, SNX2, SNX6, SNX9, SNX12, SNX13, SNX18, SNX19, SNX22, SNX23, and SNX24 were observed at moderate levels. We also performed quantitative RT-PCR analysis on normal and tumor tissue areas of PTC. The results demonstrated that the transcripts of SNX5 in the tumor lesions were 3.2-fold increased compared with those of normal thyroid areas (Fig. 1b).
To further determine the expression profile of SNX5 in tumor tissues, we established a mouse anti-human SNX5 monoclonal antibody (clone 48C2) specifically reacting to a part of the N-terminal domain of SNX5, but not other SNXs including SNX4, SNX6 (most similar to SNX5), and SNX8 (Fig. S1a,b). By using this mAb, we could certainly detect SNX5 in primary culture cells of PTC (Fig. S1c). We then performed immunohistochemistry with this mAb on tissue sections from a total of 267 cases of various thyroid tumors as summarized in Table 1. Indeed PTC, featuring papillary structure with empty-appearing nuclei, preferentially presented SNX5 (94.2%, 113/120 cases positive), like TTF-1 and thyroglobulin (Fig. 2). The sensitivity of SNX5 was 95.2% (100/105 cases; data not shown), very close to the values of TTF-1 and thyroglobulin (both indicating 100%, 105/105 cases). Interestingly, when we investigated the metastatic regions of PTC to the lymph nodes or lung, the expression of SNX5 was seemingly observed (Fig. 3b,c). Tumor tissues of composite-type carcinoma with PTC and poorly differentiated carcinoma demonstrated SNX5 expression only in the region of PTC (Fig. 3d). These findings suggested that investigation of the expression profile of SNX5 would be useful to define primary and secondary lesions of PTC. Normal thyrocytes very faintly expressed SNX5 (Figs 1b and 3a), implying that the amount of SNX5 would be enhanced during the tumorigenesis of PTC and other thyroid-origin tumors might show possible expression of SNX5 as well. In fact some other malignant thyroid tumors expressed SNX5. However, the positive rates were lower than those of PTC, ranging from 0% for medullary carcinoma (0/4 cases) and undifferentiated carcinoma (0/5 cases) to 41.2% for follicular carcinoma (7/17 cases). Benign tumors and tumor-like regions such as follicular adenoma and nodular hyperplasia (adenomatous goiter) also showed SNX5 expression with 28.6% (6/21 cases) and 67.9% (19/28 cases) positive rates, respectively. We further studied 173 cases of major malignant tumors not originated from thyroid as summarized in Table 2. As a result, we found very low positive rates of SNX5 in tumors that emerged in the lung (2.6%, 1/38 cases), breast (5.1%, 2/39 cases), colon (4.2%, 1/24 cases), liver (0%, 0/11 cases), kidney (0%, 0/21 cases), prostate (9.1%, 1/11 cases), ovary (14.3%, 2/14 cases), and uterus (6.7%, 1/15 cases). Our findings included other malignancies such as squamous cell carcinomas and lymphomas that occurred in a variety of organs, which showed negative expression of SNX5 (Table S2). Collectively these suggested the characteristic predominance of SNX5 in PTC, especially in the case of differential diagnosis from those of other tissue origin. So far examination of the expression profile of SNX5 would be beneficial for thyroid tumors with papillary lesions, while we failed to find any significant relationship between expression profile of SNX5 and tumor stages in PTC cases in this study (data not shown).
Table 1. Expression of SNX5 in thyroid tumors
Table 2. Expression of SNX5 in nonthyroid tumors
Clear cell carcinoma
High expression of Snx5 in chemically induced murine papillary thyroid carcinoma
Next we investigated murine models of PTC induced by the administration of chemical reagents. After injection of DHPN as an initiator, mice were given free access to drinking water containing SDM as a promoter for consecutive weeks. PTC was eventually emerged 12 weeks after the administration of SDM (Fig. 4a). When the tumor lesions were examined by quantitative RT-PCR, the levels of transcripts of Snx5 of the tumors were found to be about twofold those of normal thyroid tissue (Fig. 4b). This suggested that high expression of SNX5 might be a prerequisite for the tumorigenesis of PTC.
Induction of Caspase-2 by SNX5
Following these studies, we tried to determine the functional role of SNX5 in PTC cells. To do this we initially investigated HEK293 cells that expressed SNX5 at high levels (HEK293-SNX5 cells) and mock control cells (HEK293-control cells), because HEK293 cells are often used to study a fundamental role of a molecule in concern. When examined molecules regulating DNA repair, cell cycle and apoptosis, we unexpectedly found a novel role of SNX5 in the regulation of Caspase-2 (Fig. 5a,[39-41]). We next established and examined transformants of 8505c PTC cells that expressed SNX5 at high levels (8505c-SNX5 cells) and mock control cells (8505c-control cells). Like HEK293-SNX5 cells, 8505c-SNX5 cells showed upregulation of Caspase-2 (Fig. 5b). Immunohistochemical analysis on PTC tissue sections further demonstrated the presence of Caspase-2 (Fig. 5d), suggesting that SNX5 would induce Caspase-2 in PTC cells. Conversely, when we investigated 8505c transformed cells, overexpression of SNX5 conferred growth advantage (Fig. 5c). Therefore it was possible to consider that the action of Caspase-2 relating to apoptosis might probably be abrogated in the tumor cells.
Here we report a unique role of SNX5 frequently presented in PTC. TTF-1, like Galectin-3 and other markers, is often used for the pathological diagnosis of PTC, although their expression is also noted in most other malignancies with papillary structures such as breast, intestine, and lung carcinomas.[42-44] Thyroglobulin is tissue-specific to the thyroid gland, but it should not be limited to PTC. As noted in this study, adenocarcinoma originated from tissues other than thyroid scarcely presented SNX5, which could be used to define primary and secondary PTC. Currently we do not know the mechanism of the upregulation of SNX5 in PTC, while our results indicated that SNX-BAR molecules, including SNX5, might have a cardinal role in the maintenance of thyrocyte function. This was expected from experimental evidence that SNX-BAR molecules such as SNX1, SNX2, SNX4, SNX5 and SNX6 operate to transfer small cargos between endosomes and the trans-Golgi network.[24-29] Regarding tumor biology, SNX2 is highly presented in tumor cells as a chimeric molecule with ABL1; however, the SNX5 gene is localized on chromosome 20p11, whose locus is believed unlikely to be altered in the majority of PTC.[45, 46]
A further surprising finding of this study was that SNX5 could control Caspase-2. Caspases, a family of cysteine-dependent aspartate-direct proteases, play critical roles in the initiation and execution of cell death.[47, 48] Our results demonstrated that PTC and parts of well-differentiated tumors certainly expressed SNX5. In contrast, poorly differentiated thyroid carcinomas including undifferentiated carcinoma did not. These findings probably indicate some correlation of the expression profile of SNX5 and the malignant potential of thyroid tumors. It is reported that functional loss of SNX1 may affect alteration of a cell regulatory mechanism, eventually leading to malignant progression, implying its tumor suppressor activity in certain tumors.[31, 35] Therefore, as an SNX-BAR, SNX5 might also have a tumor-suppressive function similar to SNX1. In particular, Caspase-2 controlled by SNX5 would have clinical relevance, comprising the slow growth potential of PTC.
While the manner of the action of SNX5 in the accumulation of Caspase-2 remains unknown, SNX-BAR molecules have been suggested to play multiple roles to preserve cellular integrity.[32-34, 49] Considering this, together with our experimental results, excess amounts of SNX-BAR molecules can provoke unique functions in cells, where saturation for binding to the corresponding curvatures of endosomes eventually results in the emergence of a “free form“ of SNX-BAR molecules from the endosome binding. In this regard SNX-BARs can act not only as transporters of vesicles, but also as possible sensors of the number or quality of the vesicles in cells. Thyrocytes are probably regulated by SNX-BARs, which might represent the number of intracellular loading units of endosomes with a traditional sorting function and eventually monitor cellular activities. It is well recognized that abnormalities of intra-cytoplasmic transfer of vacuoles such as endosomes occur in many tumor cells. SNX6, which has the ultimate function of metabolism of a p27kip1 tumor suppressor as indicated by cell-transformation experiments, is presented in PTC like SNX5.[49, 50] So far further investigations will provide clues to fully illustrate the functional significance of the SNX5-Caspase-2 pathway in the tumorigenesis of PTC.
We thank Mr Kim Barrymore for help with the manuscript. This study was supported in part by grants from the Japanese Society for the Promotion of Science to S. Ichimiya (No. 23590405) and T. Kikuchi (No. 22790350) and the Suhara Memorial Foundation to S. Ichimiya.