Introduction
Polarity is defined as the persistent asymmetrical and ordered distribution of structures along an axis, and represents a basic property of organisms, organs, tissues and cells (1). Different types of polarity exist, like the planar cell polarity (PCP), the apical-basal polarity (ABP) and the front-rear polarity (FRP). The ABP is the most studied and best described type of cell polarity, and is a peculiar characteristic of epithelial and endothelial cells (2). It determines the formation of two cellular regions: the apical one faces the lumen and the basal one is in contact with the basement membrane. The establishment and maintenance of ABP is due to the interaction and subcellular localization of three evolutionary conserved protein complexes, CRUMBS, PAR and SCRIBBLE, that determine the domains of polarized cells, i.e. the apical, the apical-lateral and the basallateral districts, respectively. The three polarity complexes participate also in the formation and stabilization of the tight junctions (TJs) and adherent junctions (AJs), which in turn have a role in preserving the ABP and tissue homeostasis (3–6).
Over the last decades, various research has clarified the role of the polarity complexes starting from the study of the model organisms Caenorhabditis elegans and Drosophila melanogaster (7–9). The comprehension of the molecular functions of these complexes led to consider cell polarity the prerequisite for the organogenesis, and essential to ensure tissue homeostasis by controlling proliferation, asymmetric division, migration and specialized cell functions (5). Not surprisingly, both neoplastic transformation and cancer progression have been associated to loss of cell polarity/cohesiveness (LOP/C), considered a cancer hallmark (10). Moreover, LOP/C is implicated in the epithelial-mesenchymal transition (EMT), strictly connected with the metastasis development (10). Deregulated expression of ABP complex members has been described in several cancer types, in which they have been identified as either oncosuppressors or oncogenes (4).
In the present study, we focused attention on the one having the widest range of functions, i.e. the PAR (PARtitioning defective) complex (11). It is composed by a core of three proteins: PAR3, PAR6 and the atypical protein kinase C, aPKC. In mammals, paralogous genes have been described: PARD3α/β, PARD6α/β/γ, aPKCζ, and aPKCι, expressed at different levels during the embryonal development and in adult tissues (12–15). PARD3α localizes at TJs formation areas and acts as scaffold for aPKC and PARD6 to induce TJs assembly, but participates also in the regulation of other cellular processes, including migration, cell cycle and asymmetrical division by interacting with other molecules (5). PARD3β does not interact with the aPKC, although sharing the cellular localization with PARD3α, thus, it does not seem to participate in formation of TJs (3,16). Regarding the three PARD6 forms, PARD6α is present both at TJs and in the cytosol, PARD6β is localized in the cytosol and PARD6γ focuses at TJs. Some studies have described other cellular locations of PARD6α and PARD6γ, that is nucleus and centriole, respectively (17,18). Like PARD3, PARD6 acts as scaffold protein, and its principal role is to connect aPKC to PARD3α and to its downstream targets, but the different subcellular localizations of the three forms indicate further functions (17,18). Proteins aPKCζ and aPKCι are both localized in the cytoplasm and at the TJs, and phosphorylate numerous substrates, among which PARD3 (3).
Alterations in the expression levels of the PAR complex members have been reported in many types of human cancer, like the esophageal squamous cell carcinoma, the non-small cell lung carcinoma, the ovarian cancer and the breast carcinoma (19). Moreover, an oncogenic/oncosuppressor context-dependent role of PAR members emerged from functional studies (19). However, there is a lack of information on PAR gene expression and role in epithelial thyroid cancer (TC) progression.
TC represents the most common endocrine malignancy accounting for roughly 1% of all human cancers, and its incidence has been increasing over the last decades, mainly due to the improved ability to diagnose malignant transformation in small non-palpable thyroid nodules (20,21). More than 90% of thyroid carcinomas are represented by the differentiated papillary (PTC) and follicular (FTC) histotypes, whereas the invariably fatal undifferentiated (anaplastic) thyroid carcinoma (ATC) accounts for aproximately 1% of TC (22,23). Although derived from the same cell type, the epithelial thyroid tumors show specific histological features, biological behavior and degree of differentiation as a consequence of different genetic alterations (24,25). One of the most common in PTC is the BRAFV600E mutation (26,27), that associates with clinicopathological features, as advanced stage, extrathyroidal extension, lymph node and distant metastases (28). Some authors evidenced polarity alterations in TC, and it has been shown that PTC patients with tumors retaining cellular polarity had a better course than those exhibiting LOP/C features (29–31).
In the present study, we analyzed the expression of the PAR complex members in 95 PTC compared against normal tissues and 12 ATC tissues. Data from PTC patients were then correlated with clinicopathological parameters and patients' disease-free interval.
Materials and methods
Tissue samples, histology and staging of patients
Normal and matched PTC tissues were obtained from surgical specimens of 95 patients (19 males and 76 females, age range 11–83 years, median 44 years) who underwent total thyroidectomy at the Department of Surgical Sciences, 'Sapienza' University of Rome (38 patients) or at the Department of Medicine, University of Padua (57 patients); while ATC tissues were collected from surgical specimens of 12 patients (4 males and 8 females, age range 57–79 years, median 69 years) who had surgery at the Department of Medicine, University of Padua (7 patients) or at Department of Clinical and Experimental Medicine of Pisa (5 patients). All the patients gave their informed consent, and the study was approved by the local ethics committee (protocol no. 2615). Tissue samples were collected, frozen in liquid nitrogen and stored at −80°C. Of the 95 patients, 72 exhibited classical, 18 follicular, 2 tall cell and 2 oncocytic variants. The histological diagnoses were made independently by two different histopathologists according to the World Health Organization classification (32). At the time of surgery lymph node metastases were found in 39 patients. Following TNM staging, 59 patients were at stage I, 1 at stage II, 29 at stage III and 6 at stage IV. Approximately 40–50 days later all the patients underwent radioiodine therapy followed by thyroid hormone replacement therapy. To ascertain their disease-free condition, 4–5 months after intervention all the patients underwent neck ultrasound and serum Tg measurement. Recurrences were diagnosed by measurement of serum Tg levels either in basal conditions or following recombinant human TSH stimulation; FNA cytology and/or Tg determination in the FNA wash-out from lymph nodes; 131I whole body scan; histological analysis following surgical resection of the lesion. The follow-up included 79 patients (mean 57.1+36.7 months, range 5–141 months), 52 of whom were at TNM stage I. During the follow-up 16 recurrences were recorded. Regarding ATC patients, they all died from the disease (survival time range 1–25 months, median 6 months).
Determination of BRAF<sup>V600E</sup> mutation
Genomic DNA was extracted from the frozen tissues using the DNeasy Blood and Tissues kit (Qiagen, Milan, Italy) according to the manufacturer's protocol. The BRAF status of exon 15 was assessed by both direct sequencing and mutant allele-specific PCR amplification for the T to A substitution at nucleotide 1799 (V600E), using the procedure previously described (33).
Extraction and analysis of mRNA
Frozen normal and tumor thyroid tissues were homogenized with the ultra-turrax, and total RNA extracted applying the acid guanidinium thiocyanate-phenol-chloroform method (34). The first cDNA strand was synthesized from 5 μg of RNA with M-MLV reverse transcriptase and anchored oligo(dT)23 primers (Sigma Chemicals). Parallel controls for DNA contamination were carried out omitting the reverse transcriptase. The templates obtained were used for quantitative PCR amplifications of the different members of the PAR complex and three different housekeeping genes (GAPDH, RPL13A and SDHA), previously shown to be the most stable among 7 candidate reference genes (35,36), employing the LightCycler instrument (Roche Diagnostics, Mannheim, Germany), the SYBR Premix Ex Taq II (Tli RNase H Plus) (Takara, Shiga, Japan) and specific primers listed in Table I. Amplicon specificities were checked by automated DNA sequencing (Bio-Fab Research, Rome, Italy), evaluation of melting temperatures, and electrophoresis on 2% agarose gel containing ethidium bromide. Standard curves for all genes were created with 5-fold dilutions of mixed human thyroid tissue cDNA. Calculation of data for PTC was performed by the Relative Expression Software Tool (REST 2009) using a normalization factor computed as the geometric media of the 3 reference genes, as previously described (35,36). The fold change in the mRNA levels of the different PAR complex components was referred for each PTC sample to its normal counterpart. Regarding ATC samples, for which the normal matched tissues were not available, they were compared to PTC tissues through the AACt method. In the latter, the ACt of each gene was calculated using the geometric mean of the above mentioned housekeeping genes, whose expression was proven to be stable also between PTC and ATC in preliminary experiments (data not shown). The AACt were computed by subtracting the ACt of each sample the ACt of the sample showing the lowest gene expression (37).
Western blot analysis
Normal and PTC tissue samples from 12 patients were homogenized in RIPA buffer (50 mM Tris-HCl pH 7.4, 1% NP-40, 0.5% sodium deoxycholate, 150 mM sodium chloride, 1 mM EDTA, 1 mM sodium fluoride, 1 mM AEBSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate in ddH2O) by Turrax, centrifuged at 10,000 rpm for 10 min and frozen to −80°C. Protein concentrations in the cell extracts were determined by the Bradford assay (38). Aliquot of 50 μg of tissue extracts were supplemented with 5X Laemmli buffer (120 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromophenol blue) containing 5% of mercaptoethanol, heated at 95°C for 5 min, electrophoresed on polyacrylamide gels and transferred onto nitrocellulose membranes using the Bio-Rad Mini TransBlot Cell system. The membranes were then washed with TBST (50 mM Tris-HCl, pH 7.5, 150 mM NaCl and 0.05% Tween-20) and saturated with 5% low fat milk in TBST for 2 h at room temperature. Incubations with primary antibodies were performed for the identification of the different components of the PAR complex in 2.5% low fat milk in TBST at 4°C overnight. The polyclonal antibodies raised against PARD3α (1:1,000, NBP1-88861; Novus Biologicals), PARD3β (1:1,000, Ab122264; Abcam), aPKCι (1:1,000, sc-11399; Santa Cruz Biotechnology), PARD6α (1:1,000, sc-25525; Santa Cruz Biotechnology), PARD6β (1:1,000, sc-67392; Santa Cruz Biotechnology) and PARD6γ (1:2,000, orb35046; Biorbyt, Cambridge, UK) were detected with anti-rabbit horseradish peroxidase conjugated secondary antibody (1:20,000; Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Sample loadings in the different western blots were controlled with the monoclonal anti-GAPDH (1:20,000, ab8245; Abcam). The western blots were revealed by Chemiluminescence SuperSignal kit from Pierce (Rockford, IL, USA).
Statistical analysis
First, the Shapiro-Wilk test was used to check whether the mRNA data were normally distributed, and if they were not, the non-parametric Mann-Whitney U-test was used to calculate the statistical significance of differences in the expression levels of the different PAR complex components in female vs. male patients; in classical PTC variant vs. other variants; in BRAFV600E mutated vs. wild-type PTC; in metastatic (N1) vs. non-metastatic (N0) PTC; in T1-2 vs. T3-4 tumor sizes; in TNM I-II vs. ni-IV stages; in presence or absence of recurrence; in PTC patients group vs. ATC patients group. The correlation of PAR mRNAs with each other, and with patient's age was evaluated by the Spearman's Rho test. The Cox regression stepwise with backward elimination analysis was used to assess the independent association of patient's age, tumor size, histological variants, lymph node metastasis and PAR complex components mRNAs with recurrences. The impact of each PAR gene expression on disease-free interval was assessed by the Kaplan-Meier analysis combined with Mantel-Cox log-rank. For the latter, and for the Cox regression, values were classified based on the following criteria: fold change >1.2 as 'increased' ; fold change <0.8 as 'decreased'; >0.8-fold change <1.2 as 'unvaried'. All statistical analyses were carried out with the SPSS software (IBM, Armonk, NY, USA) and the results were considered significantly different at P-value of <0.05.
Results
Expression of PAR complex components in papillary (PTC) and anaplastic (ATC) thyroid cancer tissues
The analyses of mRNA levels of PAR complex components in PTC samples, compared to normal matched tissues, revealed that the expression of all transcripts was deregulated (i.e. reduced or increased) in the majority of cases (Fig. 1). In particular, aPKCι was reduced in 33/95 (34.7%) cases, unchanged in 27/95 (28.4%) cases and increased in 35/95 (36.8%) cases; PARD3α was reduced in 27/95 (28.4%) cases, unchanged in 25/95 (26.3%) cases and increased in 35/95 (36.8%) cases; PARD3p was reduced in 29/95 (30.5%) cases, unchanged in 40/95 (42.1%) cases and increased in 26/95 (27.4%) cases; PARD6α was reduced in 61/95 (64.2%) cases, unchanged in 14/95 (14.7%) cases and increased in 20/95 (21.1%) cases; PARD6β was reduced in 22/95 (23.2%) cases, unchanged in 20/95 (21.1%) cases and increased in 53/95 (55.8%) cases; and PARD6γ was reduced in 63/95 (66.3%) cases, unchanged in 15/95 (15.8%) cases and increased in 17/95 (17.9%) cases.
We next evaluated whether the expression of the PAR genes correlated with each other. As shown in Table II, with the exception of the couple PARD6α/PARD6β, a significant positive correlation was found between all the mRNAs.
The transcripts levels were investigated also in 12 ATC tissues. The results showed that the expression of all genes analyzed was significantly lower in ATC compared to PTC tissues (Fig. 2).
Twelve PTC tissue specimens were used to prepare protein extracts, and western blot experiments were carried out for the different PAR complex members. The results confirmed their deregulation (i.e. reduction or increase) in the majority of PTC tissues (Fig. 3). Unfortunately, only 12 PTC specimens were large enough to permit both RNA and proteins extraction, so their number was too small to perform any statistical analysis, including the correlation between mRNA and protein levels.
BRAF<sup>V600E</sup> mutation and PAR complex components expression in PTC tissues
To assess the effect of BRAFV600Emutation on PAR complex member expression, we analyzed the mRNA levels of each PAR gene in 76 PTC tissues for which it was possible to check the BRAF gene status. Of these, 37 (48.7%) PTC harbored the BRAFV600E mutation while 39 had the wild-type BRAF. The result of univariate analysis, reported in Table III, showed that the BRAFV600E mutation associated only with PARD6β gene transcription in PTC tissues.
Prognostic relevance of PAR complex member expression in PTC patients
No association between the expression, at the mRNA level, of PAR complex members and gender, age, tumor histology, lymph node metastasis and TNM stage was observed (Table III). However, a significant association emerged between reduced mRNA levels of aPKCι, PARD3β and PARD6γ and increased tumor size (Table III). The univariate statistical analysis evidenced also an association between higher PARD6β mRNA levels and BRAFV600E mutation. In addition, low levels of PARD3β significantly associated with tumor relapse (Table III). Notably, the Kaplan-Meier analysis demonstrated a significant correlation only for PARD3β mRNA levels with the disease-free interval (DFI) (Fig. 4). The latter is represented for patients grouped in three categories based on the PARD3β mRNA fold changes: increased, unvaried and decreased. As evident, lower PARD3β mRNA levels negatively influenced the DFI. Besides, the multivariate statistical analysis indicated PARD3β and lymph node metastasis at diagnosis as the only variables independently associated to DFI, with a hazard ratio, respectively, of 3.04 (95% CI, 1.12–8.24, P=0.029) and 8.21 (95% CI, 1.85–35.53, P=0.006) (Table IV).
Discussion
Deregulated expression of PAR complex components has been associated to cancer development and spread (19). Nevertheless, no studies have attempted to investigate their expression and role in thyroid cancer progression to date. To the best of our knowledge, only one study evidenced the homozygous deletion of PARD3α in primary ATC derived cell lines. The authors also reported that the reinstatement of PARD3α by exogenous expression was able to reduce cell proliferation, motility and invasiveness, and to restore cell-cell contacts and cell adhesion (38,39).
In this case study, we analyzed the mRNA levels of the paralogous genes encoding for PAR complex components in PTC and ATC tissues. The results indicated that the expression of all these genes was deregulated in PTC, and that all of them were significantly lower in ATC compared to PTC. These observations may suggest that the reduced expression of PAR genes may contribute to TC progression and dedifferentiation, as hypothesized in esophageal squamous cell carcinoma where reduced level of PARD3α transcripts was associated with tumor dedifferentiation (40). On the other hand, other studies reported that the overexpression of PARD3α associates with a poor prognosis in patients affected by hepatocellular carcinoma or clear cell renal carcinoma (41,42). Concerning PARD6α and PARD6β, previous studies in breast cancer tissues and cell lines demonstrated an increased expression of both genes and their ability to activate the MEK/ERK signaling pathway increasing the proliferation rate of breast epithelial cells (43,44). However, Cunliffe and colleagues (44) reported that the reduced levels of PARD6β protein associated with poorly differentiated breast cancer tissues. Regarding PARD6γ, our data are in agreement with a study on breast cancer cells suggesting for it a tumor suppressor role (45). Previous findings on ovarian and lung cancers demonstrated an increased expression of aPKCι (46,47). Although also in our PTC series aPKCι was found upregulated in 37% of cases, it was reduced in 35% of cases and further reduced in ATC samples. In interpreting these apparently conflicting data, it may be worth to consider that a possible oncogenic/oncosuppressive role of PAR components could be context dependent, as recently suggested (48–52). However, further functional studies should be performed to assess the pathogenetic role of each PAR complex component in TC progression.
We next evaluated whether the deregulated expression of the different PAR genes could be of any prognostic value. In the univariate statistical analysis, some significant associations emerged with the clinicopathological parameters of PTC patients, among which the association between higher PARD6β mRNA levels and BRAFV600E mutation. A study conducted in mice with BRAF oncogene-dependent PTC demonstrated a higher susceptibility to undergo EMT in response to TGFfi in the animals with BRAFV600E (53). It was previously demonstrated that the BRAFV600E mutation induced the TGFfl secretion by PCCl3 cell line, and that TGFR-p phosphorylates PARD6 on S345 (54,55). Once phosphorylated, PARD6 induces degradation of the GTPase RhoA leading to dissolution of TJs, which is one of the first steps of the EMT phenomenon (56). Hence, the significant association here reported between higher PARD6β mRNA levels and BRAFV600E mutation may be a sign of these molecular events. However, the most interesting significant association emerged by the univariate statistical analysis was between the low PARD3β mRNA levels and tumor relapses. This result was confirmed by the Kaplan-Meier analysis, and multivariate statistical analysis identified the downregulation of PARD3β mRNA as an independent prognostic factor for DFI. This could be of some clinical relevance since, to date, the prognosis of PTC patients still relies on clinicopathological variables such as patient's age, tumor size, histology, lymph node or distant metastasis, which are not accurate in predicting the long-term outcome (57–60). Therefore, the identification of new molecular biomarkers strictly related to the risk of PTC relapse is sorely needed.
In conclusion, the data here reported, although needing to be confirmed by means of larger case-studies, demonstrate that the expression of PAR complex members is deregulated in the majority of PTC. Besides, there is a general trend towards their reduction in ATC tissues, which could be either an effect or a cause of the loss of tissue architecture integrity occurring in undifferentiated tumors. Functional studies are required to assess the actual contribution of PAR complex alterations in TC progression and dedifferentiation. Moreover, PARD3β mRNA could represent a useful prognostic biomarker for PTC patients.