Functional role of ALK-related signal cascades on modulation of epithelial-mesenchymal transition and apoptosis in uterine carcinosarcoma

Full-length cDNA of human ALK, c-myc, and N-myc (Open Biosystems, Huntsville, AL, USA) were subcloned into pcDNA3.1 (Invitrogen, Carlsbad, CA, USA). The human ALK promoter between −2056 and +30 bp and the human N-myc promoter encompassing −221 to +1312 bp (where +1 represents the transcription start site) were amplified by polymerase chain reaction (PCR) and were subcloned into the pGL-3B vector (Promega, Madison, WT, USA). The human Twist 1 promoter (GenBank accession number NG008114) between −1085 to +350 bp was also cloned using similar procedures. A series of 5’-truncated promoter constructs of ALK and Twist1 genes were generated by PCR-based methods. Site-directed mutagenesis in putative E1- and E2-boxes in the ALK promoter region was also carried out using the PrimeS-TAR Mutagenesis Basal kit (Takara Bio, Shiga, Japan). The pGL3B-Snail (containing the −1109/+36 sequence), pGL3B-Slug (−2125/-235 bp), pcDNA3.1-Sox2, pcDNA3.1-Sox3, pcDNA3.1-Sox4, pcDNA-Sox5, pcDNA3.1-Sox6, pcDNA3.1-Sox7, pcDNA3.1-Sox9, pcDNA3.1-Sox11, pcDNA3.1-Sox17, pcDNA3.1-mouse p65, and pNF-κB were also employed as described previously [29, 30]. Two sets of short hairpin oligonucleotides directed against ALK were designed using the siDirect version 2 software. Single-stranded ALK oligonucleotides were first annealed and then cloned into BamHI-EcoRV sites of the RNAi-Ready pSIREN-RetroQ vector (Takara, Shiga, Japan), according to the manufacturer’s instructions. The sequences of PCR primers used in this study are listed in Table 1.

The Em Ca cell lines, Ishikawa and Hec251 cells, were maintained in Eagle’s MEM with 10% bovine calf serum. The full-length ALK expression plasmid or empty vector was transfected into Hec251 cells, and the stable overexpressing clones were established as described previously [31].

Anti-ALK, anti-phospho-Akt at serine (Ser) 473 (pAkt), anti-Akt, anti-Slug, anti-Snail, and anti-cleaved caspase 3 antibodies were purchased from Cell Signaling (Danvers, MA, USA). Anti-Sox11 and anti-β-actin antibodies and doxorubicin were obtained from Sigma-Aldrich Chemicals (St. Louis, MO, USA). Anti-N-myc, anti-Twist1, and anti-Histone H1 antibodies were from Abcam (Cambridge, MA, USA). Anti-NF-κB/p65, anti-p27^kip1, and anti-bax antibodies were from BD Biosciences (San Jose, CA, USA). Anti-bcl-2 and anti-p21^waf1 antibodies were from Dako (Glostrup, Denmark). Anti-cyclin A antibody was from Novocastra (Newcastle, UK). Recombinant human tumor necrosis factor (TNF)-α, transforming growth factor (TGF)-β1, and hepatocyte growth factor (HGF) were purchased from R&D Systems (Minneapolis, MN, USA).

Transfection was carried out using LipofectAMINE PLUS (Invitrogen), in duplicate or triplicate as described previously [26‐28]. Luciferase activity was assayed as described previously [29‐31]. The two siRNAs against NF-κB/p65 or the negative control were transfected using the siPort NeoFx transfection agent (Ambion, Austin, TX, USA), according to the manufacturers’ instructions.

cDNA was synthesized from 2 μg of total RNA. For quantitative analysis, real-time RT-PCR was carried out using the Power SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) with specific primers (Table 1). Fluorescent signals were detected using the ABI 7500 real-time PCR system, and data were analyzed using the associated ABI 7500 System SDS software (Applied Biosystems). Analysis of the GAPDH gene was also applied as internal control, as described previously [29‐31].

Total cellular proteins were isolated using RIPA buffer [20 mM Tris–HCl (pH7.2), 1% Nonidet p-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate]. The nuclear fraction was prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce Biotech., Rockford, IL, USA). Aliquots of the proteins were resolved by SDS-PAGE, transferred to PVDF membranes, and probed with primary antibodies coupled to the ECL detection system (Amersham Pharmacia Biotechnology, Tokyo, Japan). The intensity of individual signals was measured using ImageJ software version 1.41 (NIH, Bethesda, MD, USA).

Cells were fixed using 70% alcohol and stained with propidium iodide (Sigma-Aldrich) for cell cycle analysis. The prepared cells were analyzed by flow cytometry using BD FACS Calibur (BD Biosciences) and CellQuest Pro software (BD Biosciences).

ChIP analysis was performed using the EpiXplore ChIP assay kit (Clontech Laboratory, Mountain View, CA, USA). Briefly, cells were cross-linked with formaldehyde after transient transfection of pcDNA3.1-mouse p65. Cell lysates were sonicated to shear DNA to lengths between 200 and 1000 bp, and then precipitated overnight using anti-NF-κB/p65 antibody or mouse IgG as negative control, along with magnetic beads. After proteinase K digestion, DNA was extracted and analyzed by PCR. ChIP analysis was conducted with a reduction in the number of cycles from 30 to 25, using four specific primer sets (Table 1).

Hec251 cells stably overexpressing full-length ALK were incubated with anti-ALK antibody. FITC-labeled anti-rabbit IgG (Molecular Probes, Eugene, OR, USA) was used as secondary antibody as described previously [26‐28].

We reviewed cases of comprehensively staged high-grade endometrial adenocarcinomas from the patient records of Kitasato University Hospital for the period from 1997 to 2015. According to the criteria of the 2014 World Health Organization classification, [32] tumors were designated as UCS if they had evidence of both malignant epithelial (endometrioid, serous, or clear cell components) and mesenchymal (homologous or heterologous) elements. Endometrioid adenocarcinomas with spindle elements and hyalinized stroma were specifically excluded. Finally, a total of 27 UCSs were investigated (Table 2). Of these, 20 cases had endometrioid components and 7 cases contained non-endometrioid epithelial components, including serous and clear types, while 21 and 6 cases showed homologous and heterologous mesenchymal elements, respectively. All tissues were routinely fixed in 10% formalin and processed for embedding in paraffin wax. Approval for this study was given by the Ethics Committee of the Kitasato University School of Medicine (B14–35).

IHC was performed using a combination of the microwave oven heating and polymer immunocomplex (Envision, Dako) methods. For immunohistochemical detection of ALK, the ALK iAEP kit (Nichirei Biosciences, Tokyo, Japan) was applied. Lung carcinoma tissues with ALK overexpression due to a gene abnormality were used as positive control.

For evaluation of IHC findings, scoring of nuclear or cytoplasmic immunoreactivity was performed, on the basis of the percentage of immunopositive cells and the immunointensity with multiplication of the values of the two parameters, as described previously [29‐31]. ALK immunopositive cells located in the carcinomatous, sarcomatous, or both components were defined as epithelial, stromal, or mixed type, respectively.

Riboprobes for ALK containing nucleotides 3946 to 4633 of the ALK gene were generated by in vitro transcription using full length ALK cDNA, and ISH assays were performed using the GenPoint Tyramide Signal Amplification System (Dako) as described previously [33]. ISH signal score were classified into four levels, as follows: −, none; 1+, fewer than 10% positive cells; 2+, 10–30%; 3+, more than 30%. Samples with a score of either 1+, 2+, or 3+ were considered as positive and—was considered as negative.

Representative images of IHC findings for ALK are illustrated in Fig. 1a. Cytoplasmic immunoreaction was mainly observed in both carcinomatous and sarcomatous components of UCSs. ALK immunopositivity was evident in 11 (40.7%) of 27 UCS cases, including 3 (11.1%) of the epithelial type, 7 (25.9%) of the stromal type, and one (3.7%) of the mixed type (Table 2). In 21 UCS cases, positive signals for ALK mRNA were detected in 15 (71.4%) cases, including 4 (19%) of epithelial type, 5 (23.8%) of stromal type, and 7 (33.3%) of mixed type (Fig. 1b and Table 2). The observed ALK mRNA signals tended to be positively associated with the IHC score, although it did not reach statistical significance (Fig. 1c). Finally, FISH assay revealed no rearrangement or amplification of the ALK locus in 5 UCS cases with strong immunoreactivity (Fig. 1d and Table 2). These findings indicated that overexpression of full-length ALK at both mRNA and protein levels was frequently observed in UCSs.

Since some Sox genes, as well as ALK, are essential for development of general neuronal properties, [34] we first examined the association between Sox factors and ALK expression using Ishikawa cells. Transient transfection of the longest ALK promoter constructs (Fig. 2a), along with nine Sox factors, revealed that Sox11, as well as Sox4 and Sox7, resulted in increased activity of the ALK promoter, in contrast to the inhibition by Sox5, Sox6, and Sox9 (Fig. 2b). Using a series of 5’-truncated promoter constructs (Fig. 2a), we found that deletion from −2056 to −416 bp had little effect on induction of the promoter activity by Sox11, and the shortest construct (−146/+30 bp), which lacks putative Sox-binding sites, still preserved the responsiveness to Sox11 activation (Fig. 2c).

Transient transfection of N-myc, but not c-myc, resulted in activation of the ALK promoter, in particular the shortest reporter constructs (−146/+30 bp) (Fig. 2d). ChIP assay also revealed that overexpression of N-myc caused its recruitment to the region from −126 to +12 bp within the ALK promoter (Additional file 1: Figure S1A). However, although four nucleotide alterations in E-boxes, which are binding sites for N-myc, were introduced in the shortest construct, changes in ALK promoter activity were relatively minor (Additional file 1: Figure S1B and C). Transfection of Sox11 also resulted in an increase in N-myc mRNA expression, along with activation of its promoter (Fig. 2e), although cooperation of Sox11 and N-myc for induction of ALK promoter activity was not observed (Fig. 2f). These findings suggest that both Sox11 and N-myc serve as positive transcriptional regulators for the ALK gene in Em Ca cells, probably through associations with the basic transcriptional machinery at the promoter.

The investigation of ALK signaling in UCSs was carried out using two Em Ca, but not UCS, cell lines, since we focused on an association between the functional roles of ALK with induction of EMT features in the carcinomatous components of UCSs. In addition, UCS cell lines are in general very rare as compared to Em Ca cells.

To examine whether ALK expression is closely linked to induction of EMT properties in Em Ca cells, two independent Hec251 cell lines stably overexpressing full-length ALK (H251-ALK#8 and #16) with strong cytoplasmic immunoreaction were established (Fig. 3a). These two independent stable clones showed high proliferation rates, particularly in the exponential growth phase, along with decreased amounts of p21^waf1, but not cyclin A and p27^kip1 (Additional file 2: Figure S2A and B). H251-ALK#16 cells treated with TGF-β1 and HGF, known as EMT inducers, demonstrated a dramatically altered morphology toward a fibroblast-like appearance after 6 days as compared to mock-treated cells, along with stabilization of exogenous full-length ALK and increased expression of pAkt, nuclear p65, as well as Twist1, but not Snail and Slug (Fig. 3b). NF-κB activity as determined by a pNF-κB reporter construct was also increased in H251-ALK#16 cells treated with TGF-β1 or HGF as compared to that of the mock cells (Additional file 2: Figure S2C). In addition, transient transfection of ALK induced increases in pAkt and nuclear p65 expression, but these effects were inhibited by cotransfection of the shRNAs against ALK in Ishikawa cells (Fig. 3c).

Given that cytokines including NF-κB effectively and reproducibly induce EMT, [35] we next examined whether p65 can affect expression of Snail, Slug, and Twist1, all of which are EMT-related genes. Treatment of Ishikawa cells with TNF-α resulted in dramatically increased expression of Twist1 as compared to Snail and Slug, along with stabilization of nuclear p65 (Fig. 3d). Transient transfection of p65 resulted in a considerable increase in Twist1 expression at both mRNA and protein levels, along with increased activity of its promoter. However, such associations were relatively minor for Snail and Slug (Fig. 3e).

Next, analysis of an approximately 1000 bp fragment upstream of the transcription start site in the Twist1 gene revealed six potential NF-κB/p65-binding elements (5’-GGRNNYYCC-3’) (Fig. 4a). Using a series of 5’-truncated promoter constructs (Fig. 4a), we found that deletion from −1086 to −101 bp had little effect on induction of the promoter activity by p65, whereas the −101/+101 bp deletion appeared to have prevented binding of p65 and reduced the promoter activity to a very low level (Fig. 4b). Similar changes in the Twist1 promoter were also observed by TNF-α treatment (Additional file 2: Figure S2D). ChIP assay also revealed that increased amount of p65 caused its recruitment to the region of −101 to +62 bp within the promoter lacking putative NF-κB-binding sites (Fig. 4c). Finally, knockdown of endogenous p65 resulted in a decrease in Twist1 expression in H251-ALK#16 cells (Fig. 4d), although pNF-κB reporter activity was not altered by overexpression of ALK (Additional file 2: Figure S2E). These findings suggest that ALK indirectly contributes to NF-κB/p65-meditaed Twist1 expression.

Since it is known that Akt has potential anti-apoptotic function, [36] we examined the contribution of the ALK/Akt axis to susceptibility to apoptosis in UCSs. Treatment of H251-ALK#16 with doxorubicin resulted in a reduction in the quantity of apoptotic cells as compared to mock-treated cells (Fig. 5a), along with stabilization of exogenous full-length ALK (Fig. 5b). In addition, the expression ratio of pAkt relative to total (t) Akt progressively increased in a dose-dependent manner in the doxorubicin-treated H251-ALK#16 cells as compared to the mock cells. Similar findings were also observed in the expression ratio of bcl2 relative to bax (Fig. 5c). These data indicate that overexpression of full-length ALK abrogates susceptibility to apoptosis through alteration in expression of Akt and bcl2 in Em Ca cells.

To confirm the above findings, immunohistochemical analyses for ALK and its related molecules were carried out using clinical UCS samples. Representative IHC findings for ALK and its related molecules are illustrated in Fig. 6a. Distinct nuclear staining for Sox11, N-myc, pAkt, and Twist1 and cytoplasmic staining for bcl2 were observed in both carcinomatous and sarcomatous components, but there were no differences in the IHC scores for these markers between the two components, with the exception of Twist1 scores (Fig. 6b).

As shown in Table 3, there were positive correlations among ALK, Sox11, N-myc, and Twist1 scores in UCS tissues. The bcl2 and pAkt scores were positively correlated with ALK, Sox11, and N-myc scores, and Sox11, N-myc, and Twist1 scores, respectively. These findings supports the in vitro results that show the existence of a Sox11/N-myc/ALK axis and an association of ALK with EMT and apoptotic features through Twist1 and bcl2 expression in Ishikawa and H251-ALK#16 cells.