Centrosomal and mitotic abnormalities in cell lines derived from papillary thyroid cancer harboring specific gene alterations

Background Differentiated thyroid carcinoma offers a good model to investigate the possible correlation between specific gene mutations and chromosome instability. Papillary thyroid neoplasms are characterized by different mutually exclusive genetic alterations, some of which are associated with aneuploidy and aggressive phenotype. Results We investigated the centrosome status and mitotic abnormalities in three thyroid carcinoma-derived cell lines, each maintaining the specific, biologically relevant gene alteration harbored by the parental tumors: RET/PTC1 rearrangement in TPC1; heterozygous and homozygous BRAFV600E mutation in K1 and in B-CPAP, respectively. B-CPAP cells showed a statistically significant (P < 0.01) higher frequency of abnormal mitotic figures compared to TPC1 and K1 cells. Conclusions Our data indicate that RET/PTC1 oncogenic activity is not related to mitotic chromosome impairment and missegregation whereas, based on the consistent difference in types/frequencies of centrosome and spindle abnormalities observed between K1 and B-CPAP cells, the hetero/homozygous allelic status of BRAFV600E mutation seems to be not irrelevant in respect to chromosomal instability development.


Background
Chromosomal instability (CIN), a genetic condition that promotes a high rate of chromosome missegregation during mitosis, is a shared feature of most cancers. Aneuploidy is a distinctive trait of most human cancers and has been linked to high tumour grade, advanced stage, and poor prognosis; however, its role in neoplastic transformation and the relationship with CIN are somewhat unclear [1]. A variety of alterations have been proposed as being responsible for CIN, including defects in the spindle checkpoint, defective sister-chromatid cohesion, kinetochore assembly, upregulation of cyclins, erosion of telomeres and centrosome abnormalities [2]. The normal function and numeral integrity of centrosomes can be affected directly by the mutated products of certain proto-oncogenes and tumour-suppressor genes [3,4]. In particular, centrosome amplification (the presence in the cell of more than two centrosomes) seems to be an early event in tumourigenesis [5], and it may affect CIN in different ways, depending on the capacity to maintain a correct bipolar chromosome segregation in the presence of multipolar spindles. In addition to amplification, structural and functional defects of centrosomes may influence CIN [4].
Papillary thyroid carcinoma (PTC), accounting for approximately 80% of differentiated thyroid carcinoma (DTC), has a number of variants with specific histological characteristics. The classical form of PTC may show activation of the BRAF gene or RET/PTC variants, in 40-45% and 10-20% of cases, respectively [6]. The two alterations are usually mutually exclusive. In different PTC-derived cell lines, the alternative presence of either BRAF mutation or RET/PTC rearrangement has been confirmed [7], reinforcing the hypothesis that RET and BRAF changes are alternative oncogenic events. The majority of papillary thyroid carcinoma shows stable karyotypes, including the cases in which aneuploid karyotypes have been described [8], being chromosomal instability only observed in the more advanced clinicopathological stages [9]. Intriguingly, however, it has been demonstrated that conditional BRAF V600E expression induces chromosomal instability in thyroid PCCL3 cells [10], suggesting a possible key role of the gene in determining CIN. Moreover, the contribution of centrosome in the morphogenetic process of PTC nuclear changes has been recently suggested [11]. Nevertheless, little is known about the relationship among PTC molecular events, aneuploidy and genomic instability of these tumours.
TPC1, K1 and B-CPAP PTC-derived cell lines are characterized by the specific genetic alterations of the parental tumours from which they have been derived, namely RET/PTC1 rearrangement in TPC1 [12], BRAF V600E mutation in K1 and in B-CPAP [7]. To investigate the possible relationship between the relevant specific gene alterations and mitotic behavior of thyroid cancer cells, we studied centrosome abnormalities and mitotic spindle aberrations of these three human PTCderived cell lines and of NTHY-ORI3.1 cells derived from human follicular epithelial cells.

Fluorescence in situ hybridization
RET FISH pattern intepretation. Diploid cells that have the chromosomal rearrangement involving the RET gene will show a discrete red signal distanced from a discrete green signal, indicating the breakage of one RET allele, while the second allele will show a red/green fused signal. Depending on chromosome 10 polysomy, this pattern will change accordingly.

NTHY-ORI-3.1 cell line
Nuclei of the NTHY-ORI-3.1 cell line showed two, three or four red/green intact signals, indicating no RET rearrangement and the presence of tri-tetraploid clones, in addition to the diploid one ( Figure 1a).

PTC-derived cell lines
Most nuclei of the TPC1 cell line showed two intact red/green signals and two red and two green discrete signals, indicating the presence of RET breakage in the tetraploid clone (Figure 1b). A minor diploid clone showed one intact red/green signal and one red and one green discrete signals. Nuclei of the K1 cell line showed four intact red/green fluorescent signals, as expected for tetrasomy 10 ( Figure 1c) in a tetraploid cell line. Nuclei of the B-CPAP cell line showed red/green intact signals. According to the presence of a major clone with trisomy 10, most of the nuclei showed three intact red/green signals ( Figure 1d).
These results indicated that only TPC1 cells had RET rearrangement.

NTHY-ORI-3.1 cell line
A very low frequency of centrosome amplification (0,7%) and spindle abnormalities (1,7%) were observed in the control SV-40-immortalized cell line NTHY-ORI-3, as an effect of SV40 immortalization (Figures 2 and 3).  Fluorescence in situ hybridization. FISH with a homebrew dual color break-apart probe set for RET gene (5' labelled with Spectrum Orange and 3' labeled with Spectrum Green). Cells with unrearranged RET show intact red/green signals (one fused signal for each RET gene); cells with rearranged RET show probe signal split into its 5' (red signal) and 3' (green signal) portions (two discrete red and green signals for each rearranged RET gene). NTHY-ORI 3-1 representative nuclei with four (a, top) and two (a, bottom) copies of intact RET; TPC1 representative nucleus with two copies of intact RET (two red/green signals) and two copies of rearranged RET (two discrete red and green signals) (b); K1 representative nucleus with four copies of intact RET (c); B-CPAP repesentative nucleus with three copies of intact RET (d). Nuclei are counterstained with 4',6diamidino-2-phenylindole.

Discussion
A number of gene alterations, such as point mutations in RAS and BRAF genes, point mutations or amplification of PIK3CA, and fusion genes involving RET, NTRK1 and PPARg are known to frequently occur in differentiated thyroid carcinoma [6,13], and are correlated to different morphological subtypes. Moreover, they are characterized by different chromosome patterns, defining specific cytogenetic subgroups which are  often correlated with different histopathological features [14].
Common mutations found in the papillary histologic subtype are point mutations of the BRAF and RAS genes as well as RET/PTC rearrangements, which are considered molecular markers of diagnostic and prognostic significance. TPC1, K1, and B-CPAP are among the most frequently used papillary thyroid carcinoma-derived cell lines for in vitro investigations of thyroid oncogenesis. Although tumour derived cell lines are considered to be the outcome of adaptation and in vitro evolution leading to a common undifferentiated phenotype [15], all these cell lines maintained the biologically relevant oncogenic events specific to the different subtypes of their parental PTCs. TPC1 and K1 cell lines, bearing RET/PTC1 and heterozygous BRAF V600E mutation respectively, had an extremely low number of aberrant cells. No statistically significant differences were found in centrosome and spindle alterations, as well as multipolar mitoses, compared to follicular thyroid cell-derived NTHY-ORI-3.1 cell line, which indeed is a SV-40 immortalized cell line and a population of abnormal mitotic cells can be expected. In contrast, the B-CPAP cell line, bearing a homozygous BRAF V600E mutation, showed remarkable alterations: several mitotic abnormalities such as metaphase chromosome misalignments, multipolar figures, and c-metaphases were observed. Intriguingly, these types of alterations were barely observed in the K1 cell line harbouring heterozygous BRAF V600E mutation. The consistent difference in types/frequencies of centrosome and spindle abnormalities observed between K1 and B-CPAP cells might suggest that the allelic status of BRAF V600E mutation is not irrelevant in respect to CIN development, and that the hetero/homozygous mutational status of the gene may influence at different rate the onset of aneuploidy in PTC cells. B-CPAP cells show a mitotic machinery deeply impaired, with   centrosome amplification, acentrosomal spindle poles and chromosome misalignment, giving rise to an increased mitotic instability, whereas BRAF V600E heterozygous K1 cells maintain correct centrosome features and spindle polarity in the majority of cells. In fact, B-CPAP cell line has a less stable karyotype compared to K1. Moreover, as expected, because of centrosome amplification, B-CPAP showed a significantly higher frequency of multipolar mitoses compared to K1. A small proportion of these mitoses showed acentrosomal spindle poles, which are reminiscent of spindle formation involving nucleation through a chromatin-dependent spindle assembly pathway [16]. Very recently, a link between the BRAF V600E oncogene and chromosome instability in melanoma has been suggested [17]. More than 65% of cutaneous melanoma [18] and approximately 45% of PTC [6] share the same BRAF V600E mutation, however, differently from melanoma in which aneuploid karyotypes are frequent [8] and homozygous BRAF V600E mutation is described [19], PTC usually has heterozygous mutation [20]. To the best of our knowledge, in vivo homozygous BRAF V600E mutation in thyroid tumours has not been yet described, and it is likely that homozygous cell lines have acquired the second mutation during their establishment. Nevertheless, as B-CPAP is the only of the three PTCderived cell lines harboring a mutation in TP53, which has been closely related to centrosome duplication and genomic integrity maintenance [4,21], the possibility that TP53 mutation could be involved in the observed enhanced instability, can not be ruled out.
Chromosome number in K1 and TPC1 cell lines is maintained rather stable in culture. However, K1 cells karyotype is characterized by both structural and numerical chromosome changes, whereas TPC1 karyotype is characterized by structural changes [15]. Indeed, the small population of K1 cells bearing mitotic abnormalities shows mainly ana-telophase laggings, compared to the small population of mitotic abnormal TPC1 cells (P < 0.05). As lagging chromosomes at anaphase represent a potential source of aneuploidy, the presence of this small abnormal ana-telophase population may explain the presence of clonal chromosome numerical changes in K1 karyotype. Intriguingly, BRAF V600E positive PTC may harbour aneuploidy [9], whereas RET/PTC positive carcinoma very rarely shows numerical chromosome changes -and indeed RET/PTC rearrangements have been reported in non-malignant thyreocytes [22]. A direct ascertainment of the mitotic behavior and centrosome status of PTC with BRAF V600E mutation in primary PTCs might favourably confirm this view.

Conclusions
Our data suggest that difference in spindle abnormalities and possibly in centrosome amplification may depend on the BRAF V600E heterozygous/homozygous mutational status. Whether the proneness to accumulate numerical chromosome changes only in the more rare advanced forms of PTC might depend on the gain of a second BRAF V600E mutation, and/or accumulation of other gene alterations, remains to be clarify.
The degree and type of genetic instability in cancer is emerging as an important feature, also considered as a possible ancillary and integrative parameter in tumour classification. Besides the prognostic relevance of chromosomal instability, CIN could also contribute to the ability of cancer cells to acquire chemoresistance, generating occasionally cells with the capacity to grow more efficiently in adverse environments. In this view, investigation of the correlation between neoplasia gene-specific mutational status and chromosomal instability could provide better targets for gene-specific therapies.

Cell lines
The PTC-derived TPC1 and B-CPAP cell lines were kindly provided by Dr. Fusco (Medical School, University Federico II of Naples, Naples, Italy), the PTC-derived K1 cell line was purchased from Health Protection Agency Culture Collections [23]. TPC1, K1 and B-CPAP cell lines were chosen according to their reported molecular and cytogenetic features: the TPC1 cell line originally showed a near-diploid karyotype and contains a RET/PTC1 rearrangement [7], the B-CPAP cell line has an aneuploid karyotype (chromosome mode 72) [24] and BRAF V600E homozygous mutation [7], the K1 cell line had a near tetraploid karyotype [25] and BRAF V600E heterozygous mutation [7]. We confirmed by FISH that the RET rearrangement was present only in TPC1 cells. TaqMan Real-Time polymerase chain reaction assay confirmed the presence of the BRAF V600E mutation in B-CPAP and K1 cells. Karyotyping revealed tetraploidization of TPC1 (a phenomenon already reported in this cell line) [25] and maintenance of the described B-CPAP [23] and K1 [25] chromosome patterns. K1 cells are also characterized by PIK3CA amplification [26], which is absent in TPC1, B-CPAP, and NTHY-ORI 3-1 cell lines. For comparison, a commercially available [23] NTHY-ORI 3-1 cell line was used [27]. This SV40-transfected cell line, obtained from human normal thyreocytes, bears wild-type RET and BRAF genes and retains some morphologic and physiological characteristics of normal thyreocytes, without evidence of malignant transformation [27]. TP53 mutation was present in B-CPAP [7] cells. All cell lines lack RAS mutations [7]. The cell lines were maintained in DMEM/ F12 (Sigma-Aldrich, Milan, Italy) [except NTHY-ORI 3-1 that was grown in RPMI 1640 (Gibco-BRL. Life Technologies, Milan. Italy)] supplemented with 10% fetal bovine serum (Gibco-BRL), at 37°C in humidified 5% CO2.

Immunofluorescence staining
Cells for immunofluorescence were cultured directly on flaskette glass slides (NUNC A/S, Roskilde, Denmark). The cells were fixed with methanol:acetone (1:1) for 10 min at -20°C, followed by 10 min methanol at -20°C. Centrosomes and mitotic spindle were labeled with a rabbit polyclonal anti-γ-tubulin antibody (Sigma-Aldrich) and with a mouse anti-β-tubulin antibody (Sigma-Aldrich), respectively. After 1 h incubation with the antibodies diluted 1:200 in PBS/1% BSA at 37°C and 3 × 5 min washing in PBS, the slides were incubated with fluorescein isothiocyanate (FITC)-coupled anti-rabbit secondary antibody (Sigma-Aldrich) and TRITC-conjugated anti-mouse secondary antibody (Sigma-Aldrich) for 30 min at 37°C. After 3 × 5 min washing in PBS, the cells were counterstained with DAPI. Centrosome structural abnormalities were considered to be present when a diameter of at least twice that of normal centrosomes in lymphocytes and in NTHY-ORI 3-1 cells was observed; amplification was recorded when there were more than two centrosomes per cell [28]. At least 200 consecutive cells per sample were examined. Centrioles were labeled with a rabbit anti-centrin 2 antibody (dilution 1:200) (Santa Cruz Biotechnology, Santa Cruz (CA), USA), applying the same procedure described above. Preparations were observed using an epifluorescence microscope (Provis AX70, Olympus, Milan, Italy), and images were acquired with a digital CCD camera (C4742 Orca II, Hamamatsu, Japan) driven by Cytovision (Applied Imaging Corp., Santa Clara, CA, USA). The DAPI, FITC and TRITC images were acquired with selective single-bandpass filters at 1000× optical magnification.

Giemsa staining
The cells grown on slides were washed twice with PBS, fixed with methanol:acetic acid (3:1) for 30 min at 4°C, and stained in 4% buffered Giemsa stain (pH 7.2) for 25 min. At least 200 consecutive mitotic figures per sample were examined by light microscopy, and images were acquired with a CCD camera at 1000× optical magnification.

Statistics
Frequency distributions of all variables were calculated, and chi-square analyses were used for categorical comparisons. A p-value < 0.01 was considered statistically significant.