Individualised Multimodal Treatment Strategies for Anaplastic and Poorly Differentiated Thyroid Cancer

The prognosis of anaplastic (ATC) and poorly differentiated thyroid cancer (PDTC) is poor, due to their radioiodine refractoriness (RAI-R), high metastatic potential and current lack of effective treatment strategies. We aimed to examine the efficacy of the tyrosine kinase inhibitors (TKIs) sorafenib and selumetinib and the histone deacetylase inhibitor (HDACI) panobinostat in patient-derived tumor tissue (PDTT) of ATCs/PDTCs, the expression of sodium iodide symporter (NIS) and radioiodine up-take (RAI-U). High Mobility Group AT-Hook 2 (HMGA2) and associated miRNAs expression was correlated with the clinical course of the patients. Inhibitory effects of panobinostat, sorafenib and selumetinib were measured by real time cell analyser xCELLigence in five PDTTs and human foreskin fibroblasts (HF) used as control. Expression of NIS, HMGA2 and associated miRNAs hsa-let-7f-5p, hsa-let-7b-5p, hsa-miR-146b-5p and hsa-miR-146b-3p was performed by RT-qPCR and Western blot. RAI-U was performed by Gamma Counter with I-131. Panobinostat showed the strongest cytotoxic effect (10 nM) in all PDTTs and HF and caused a significant over-expression of NIS transcript. TKIs were able to up-regulate NIS transcript in patient 5 and in HF. RAI-U was up-regulated after 24 h of treatment with TKIs and panobinostat in all PDTT and HF, except in patient 5. Selumetinib caused a significant suppression of HMGA2 in PDTT 1, 2, 4, 5 and HF; whereas sorafenib caused no change of HMGA2 expression. Panobinostat suppressed significantly HMGA2 in PDTT 2, 4 and HF. The expression of miRNAs hsa-let-7f-5p, has-let-7b-5p hsa-miR-146b-5p and hsa-miR-146b-3p was modulated heterogeneously. NIS protein level was over-expressed in three PDTTs (patients 1, 3 and 4) after 24 h of treatment with selumetinib, sorafenib and in particular with panobinostat. HF showed a stable NIS protein level after treatment. Panobinostat showed the strongest cytotoxicity in all treated PDTTs at the lowest dosage in comparison with TKI. All three compounds were able to modulate differently NIS, HMGA2 and related miRNAs. These factors represent valuable markers in PDTT for new treatment strategies for patients suffering from ATC/PDTC. Thus, the establishment of PDTT could be a useful tool to test the efficacy of compounds and to develop new and individualised multimodal treatment options for PDTCs and ATCs.


Introduction
Although poorly differentiated thyroid carcinoma (PDTC) and anaplastic thyroid carcinoma (ATC) account only for a substantial portion of the morbidity and mortality associated with thyroid cancer (TC), both entities have unfavorable prognosis because of iodine/radioiodine refractoriness (RAI-R) and high metastatic potential.
Up to now, aggressive multimodal treatment based on surgical resection in combination with radio-and/or chemotherapy is recommended [1], but despite various attempts to improve the outcome of ATC and PDTC, the clinical course did not change significantly within the last decades [1,2].
The cellular origin of ATCs and PDTCs and the understanding of molecular mechanisms underlying thyroid carcinogenesis gained important implications for the therapeutic and follow-up strategies [3] and, in the last years, the tyrosine kinase inhibitors (TKIs) and histone deacetylase inhibitors (HDACIs) have shown promising results for the treatment of thyroid cancer [4].
Every individual ATC and PDTC has shown a different sensitivity to treatment with many different systemic compounds [5,6]. The resistance to the current available therapies is mainly due to the various epigenetic alterations occurring in these tumors; thus, playing a decisive role in tumorigenesis and metastasis, which characterise the aggressive clinical course of PDTCs and ATCs. Only a deeper understanding of the individual epigenetic dysregulations of individual ATCs and PDTCs may lead to the development of novel epigenetics-based therapeutic strategies [7]. Giordano et al. described the genomic landscape of 496 papillary thyroid carcinomas (PTCs) in 2014 and showed that the constitutive activation of the Mitogen-activated protein kinase (MAPK) cascade, mediated by oncogenic v-Raf murine sarcoma viral oncogene homolog B (BRAF) V600E mutation, down-regulates the expression of genes involved in iodine metabolism; thus resulting in RAI-R [8]. Nevertheless, the expression, modulation and function of the sodium iodide symporter (NIS; SLC5A5) in radioiodine refractory ATCs and PDTCs remain unclear. Some authors stated that the dysfunction of NIS is mainly due to suppression or loss of NIS gene [9]. Another mechanism causing dysfunction may be a distorted alignment at the cytosolic membrane site [10]. Furthermore, it has been shown that NIS is a target of miRNAs belonging to Let7 family. MiRNAs regulate gene expression by blocking translation or by inducing the degradation of their target mRNA. Damanakis et al. demonstrated the expression of SLC5A5 (NIS) gene in ATCs and showed an inverse correlation between SLC5A5 and the tumor suppressor microRNA (miRNA) hsa-let-7f-5p in thyroid cancer [11]. Hsa-let-7f-5p is inversely correlated with NIS expression due to its high affinity for the 3 UTR of NIS transcript [11][12][13]. De-regulation/suppression of Let-7 family members act in several types of cancer, including differentiated thyroid cancer (DTC) [14][15][16]. However, little is known regarding the distinct function of hsa-let-7f in TC. Among them, hsa-let-7f is described as critical for proper regulation, growth and differentiation of thyroid cells. In particular, hsa-let-7f-5p was reported to exert its tumor suppressor role by reducing cell proliferation and inducing thyroid differentiation markers [17]. Onco-miRNA hsa-miR-146b-5p is significantly over-expressed in papillary thyroid cancer (PTC) and associated with tumor migration, invasion and EMT (epithelial-mesenchymal transition) [12,[18][19][20]. The over-expression of hsa-miR-146b-5p can be promoted by RET proto-oncogene (RET/PTC3) and BRAF activation [18].
High Mobility Group AT-Hook 2 (HMGA2), a well-known non-transcription factor with oncogenic properties, is highly expressed in several tumors [21,22], including TC [23]. It is responsible for supporting the activity of proliferation-related transcription factors, e.g., E2F1, leading to proliferation and survival of tumor cells [24].
HMGA2 down-regulation, mediated by epigenetic modifications that imply the over-expression of Let-7 family miRNAs, leads to cell cycle block and cell death [25].
In this study, we aimed to examine the efficacy of two different TKIs, sorafenib and selumetinib and one HDACI, panobinostat, by using patient-derived tumor tissue (PDTT) of ATCs/PDTCs. We aimed to highlight the ability of these compounds to induce re-differentiation via the over-expression of NIS and associated radioiodine uptake (RAI-U). Furthermore, HMGA2 and associated miRNAs expression were correlated with the clinical course of the patients.

Preparation of Patient-Derived Human Tumor Tissue (PDTT)
The tumor tissue resected from 5 patients who underwent surgery was immediately collected in sterile Phosphate-Buffer saline (PBS) without Ca 2+ and Mg + (L1825 Biochrom, Berlin, Germany). The tissue was washed (3×) with sterile PBS in order to remove any tissue debris and blood. Afterwards, the tissue was cut in small pieces with a sterile scalpel (Feather, Osaka, Japan). The small pieces were rinsed through a cell strainer (352350 BD Labware, Franklin Lakes, NJ, USA) and washed with Roswell Park Memorial Institute 1640 (RPMI1640) Medium (FG1215 Biochrom, Berlin, Germany). The cell suspension was centrifuged at 1500 rpm for 8 min at room temperature. The pellet was suspended with complete growth medium RPMI 1640 (Biochrom) supplemented with 10% fetal bovine serum (FBS, Biochrom) and 10 U/mL penicillin and 100 µg/mL streptomycin (Biochrom) and the suspension was pipetted in a cell culture 6-well plate (83.3920 Sarstedt, Nümbrecht, Germany). After 2 h, the adherence of the cells was checked. The medium was changed regularly every second day. The cells were then trypsinized and transferred into 25 cm 2 flasks.

Real-Time Cell Viability Analysis
The xCELLigence RTCA SP system (Roche Applied Science, Mannheim, Germany) was used for real-time and time-dependent analysis of the cellular response of human thyroid cancer cells following incubation with 1 to 100 µM selumetinib and sorafenib and 1 to 100 nM panobinostat. To perform this analysis, 3000 cells were seeded in 150 µL complete growth medium per well in a 96-well E-plate (OLS, Bremen, Germany) and incubated with the previously mentioned compounds. Cell index, which indicates attachment and adherence of cells to the plate's electrode, was measured continuously for the following 130 h using the cell impedance detection system. Data analysis was performed using the RTCA Software v1.2.1 (OLS) for calculation of the temporal dynamics of cellular attachment (i.e., viability).

RNA Isolation and Quantitative Real-Time RT-PCR
Cells were seeded in 25 cm 2 cell culture flasks (0.5 × 10 6 cells/flask) and treated with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat for 48 h. Total RNA, including short RNA was isolated by use of miRNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.
For the amplification of SLC5A5 and HMGA2, lysates were reverse transcribed using the iScript cDNA Syntesis Kit (Bio-Rad). PCR was run with the SsoFast Eva Green Supermix (BioRad) on CFX96 cycler. GAPDH was amplified as reference gene. Primer Assays for SLC5A5 (PPH10926A) and GAPDH (QT01192646) were purchased from Qiagen. HMGA2 (10025636) was purchased from BioRad.
Data Analysis: Results were analysed using CFX Manager (BioRad) and Rest 2008. Significance was calculated using the t-test for paired samples. p < 0.05 was regarded as significant.

In Vitro Radioiodine Up-Take
For the analysis of radioiodine uptake (RAI-U), the cells were seeded in 6-well plates at a density of 4 × 10 5 cells/well. Immediately after treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat, 1 Mbq I-131 was added to the cells and the plates were incubated for 24 and 48 h. Subsequently, the cells were washed with PBS and trypsinized for 5 min. The suspension was rinsed with 3 mL PBS, collected and centrifuged at 900 rpm for 5 min. The supernatant was discarded and the pellet was suspended in 5 mL PBS. The suspension was once more centrifuged. The supernatant was discarded and the cells were processed into a Gamma counter for the measurement of the retained radioactive I-131. The untreated cells were used as control for the measurement.

Statistical Analysis
Data were collected using Excel (Microsoft Office 2010, Redmond, WA, USA). Significance was calculated using the t-test for paired samples. p < 0.05 was regarded as significant.

Ethical Approval
The study was conducted under the approval of the ethic committee of the University Hospital of Marburg (No. 92/15) and all the patients signed out the inform consent.

Results
Five PDTTs were established from five tumor tissues of patients affected by ATC and PDTC. The patients were divided into two classes based on the tumor origin and the resected tissue. Primary tumor was collected from patients affected by ATC; lymph node metastases were resected from PDTC affected patients. Shown below are Table 1 for ATC patients and Table 2 for PDTC patients.

In Vitro Results of the First Patient
The first patient was affected by ATC and survived only two weeks after the first diagnosis. He received palliative treatment only and was characterised by BRAFV600 mutation and RAI-R (Table 1). No individualised therapy was established. The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions. 3.1.1. Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib Treatment with selumetinib showed no cytotoxicity in the PDTT isolated from this patient. Sorafenib caused a reduction of cell viability at 100 µM. Interestingly, panobinostat treatment caused a significant reduction of cell viability with 10 and 100 nM. The reduction of cell viability after prolonged treatment with 100 µM sorafenib and 10-100 nM panobinostat was nearly 100% ( Figure 1A).

In Vitro Results of the First Patient
The first patient was affected by ATC and survived only two weeks after the first diagnosis. He received palliative treatment only and was characterised by BRAFV600 mutation and RAI-R (Table 1). No individualised therapy was established. The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions. 3.1.1. Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib Treatment with selumetinib showed no cytotoxicity in the PDTT isolated from this patient. Sorafenib caused a reduction of cell viability at 100 µM. Interestingly, panobinostat treatment caused a significant reduction of cell viability with 10 and 100 nM. The reduction of cell viability after prolonged treatment with 100 µM sorafenib and 10-100 nM panobinostat was nearly 100% ( Figure 1A).

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
The function of NIS protein was monitored by RAI-U measurement after administration of the three compounds. Radioiodine retention was up-regulated after 24 h of treatment with 10 µM

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
The function of NIS protein was monitored by RAI-U measurement after administration of the three compounds. Radioiodine retention was up-regulated after 24 h of treatment with 10 µM selumetinib/10 µM sorafenib/10 nM panobinostat ( Figure 1B). After 48 h of treatment, the cells of this PDTT showed no increase of RAI-U in comparison with the untreated cells.

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
Treatment with 10 nM panobinostat induced a significant over-expression of SLC5A5 transcript, whereas 10 µM sorafenib determined a significant down-regulation of it and selumetinib did not cause any change of SLC5A5 expression ( Figure 2A). HMGA2 transcript was significantly down-regulated after treatment with 10 µM selumetinib. The administration of sorafenib and panobinostat did not alter the expression of HMGA2 in comparison with untreated cells (Figure 2A).

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
Treatment with 10 nM panobinostat induced a significant over-expression of SLC5A5 transcript, whereas 10 µM sorafenib determined a significant down-regulation of it and selumetinib did not cause any change of SLC5A5 expression (Figure 2A). HMGA2 transcript was significantly down-regulated after treatment with 10 µM selumetinib. The administration of sorafenib and panobinostat did not alter the expression of HMGA2 in comparison with untreated cells (Figure 2A).

Expression of NIS Protein Level
Additionally, NIS protein level was analysed after 48 h of treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat. As shown in Figure 2C, the untreated cells showed no detectable band corresponding to NIS protein. The treatment with the three compounds caused a strong significant re-expression of NIS protein as confirmed by densitometry analysis ( Figure 2D).

Expression of NIS Protein Level
Additionally, NIS protein level was analysed after 48 h of treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat. As shown in Figure 2C, the untreated cells showed no detectable band corresponding to NIS protein. The treatment with the three compounds caused a strong significant re-expression of NIS protein as confirmed by densitometry analysis ( Figure 2D).
Despite the TKIs were not showing a significant effect on the modulation of NIS transcript, the protein level was over-expressed after treatment.
Here, it could be observed that TKI were able to increase the radioiodine up-take and the protein level of NIS without any inverse correlation with has-let7f-5p. HMGA2 was significantly suppressed only after treatment with 10 µM selumetinib.
The higher potential of panobinostat in terms of benefit for patients affected by PDTC/ATC could be attributed to its low dosage and stronger cytotoxic effect that involve the suppression of the regulatory miRNA hsa-let7f-5p and the over-expression of NIS at transcript and protein level. The other miRNAs were over-expressed after treatment with TKI and panobinostat.

In Vitro Results of the Second Patient
The second patient was affected by ATC and survived for eight weeks after the first diagnosis. He received palliative treatment only and was characterized by RAI-R. No mutation of BRAF was detected (Table 1). He received palliative treatment and no individualized therapy was established. The treatment with TKIs and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
The PDTT obtained from the tissue of this patient showed a significant reduction of cell viability with 100 µM selumetinib and sorafenib ( Figure 3A). Panobinostat showed a significant effect at 10 and 100 nM that determined a reduction of the growth curve to zero values. Despite the TKIs were not showing a significant effect on the modulation of NIS transcript, the protein level was over-expressed after treatment.
Here, it could be observed that TKI were able to increase the radioiodine up-take and the protein level of NIS without any inverse correlation with has-let7f-5p. HMGA2 was significantly suppressed only after treatment with 10 µM selumetinib.
The higher potential of panobinostat in terms of benefit for patients affected by PDTC/ATC could be attributed to its low dosage and stronger cytotoxic effect that involve the suppression of the regulatory miRNA hsa-let7f-5p and the over-expression of NIS at transcript and protein level. The other miRNAs were over-expressed after treatment with TKI and panobinostat.

In Vitro Results of the Second Patient
The second patient was affected by ATC and survived for eight weeks after the first diagnosis. He received palliative treatment only and was characterized by RAI-R. No mutation of BRAF was detected (Table 1). He received palliative treatment and no individualized therapy was established. The treatment with TKIs and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
The PDTT obtained from the tissue of this patient showed a significant reduction of cell viability with 100 µM selumetinib and sorafenib ( Figure 3A). Panobinostat showed a significant effect at 10 and 100 nM that determined a reduction of the growth curve to zero values.

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
An increase of NIS monitored via RAI-U was observed after 24 h of treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat ( Figure 3B). After 48 h of treatment, the cells showed no change of RAI-U in comparison with untreated cells.

Expression of NIS Protein Level
The protein NIS was normally expressed in untreated cells. 48 h of treatment with 10 µM selumetinib, sorafenib and 10 nM panobinostat caused a significant reduction of NIS protein level as shown in Figure 4C, as confirmed by densitometry ( Figure 4D).
Here, panobinostat showed the highest efficacy with the lowest dosage (10 and 100 nM). Furthermore, panobinostat alone caused an over-expression of NIS transcript together with the suppression of HMGA2 and hsa-let7f-5p. 10 µM selumetinib was able to down-regulate HMGA2 and hsa-let7f-5p. 10 µM sorafenib showed cytotoxic effects and the suppression of hsa-let7f-5p only. TKI

Expression of NIS Protein Level
The protein NIS was normally expressed in untreated cells. 48 h of treatment with 10 µM selumetinib, sorafenib and 10 nM panobinostat caused a significant reduction of NIS protein level as shown in Figure 4C, as confirmed by densitometry ( Figure 4D).
Here, panobinostat showed the highest efficacy with the lowest dosage (10 and 100 nM). Furthermore, panobinostat alone caused an over-expression of NIS transcript together with the suppression of HMGA2 and hsa-let7f-5p. 10 µM selumetinib was able to down-regulate HMGA2 and hsa-let7f-5p. 10 µM sorafenib showed cytotoxic effects and the suppression of hsa-let7f-5p only. TKI and panobinostat increased the radioiodine up-take, whereas NIS protein level was down-regulated. The expression of the other miRNAs was stable after treatment with the three compounds.

In Vitro Results of the Third Patient
The third patient was affected by ATC and he is still alive after 282 weeks from his first diagnosis. After initial treatment with docetaxel and cisplatin, his therapy was interrupted and an individualised therapy with sorafenib was established. Treatment with sorafenib was characterised by multiple adverse effects and stopped. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
In this case, sorafenib and panobinostat were the most effective compounds. Treatment with 10-100 nM panobinostat and 100 µM sorafenib resulted in nearly 100% reduction of cell viability. As shown in Figure 5A, a significant reduction of cell viability could also be observed after treatment with 100 µM selumetinib. and panobinostat increased the radioiodine up-take, whereas NIS protein level was down-regulated. The expression of the other miRNAs was stable after treatment with the three compounds.

In Vitro Results of the Third Patient
The third patient was affected by ATC and he is still alive after 282 weeks from his first diagnosis. After initial treatment with docetaxel and cisplatin, his therapy was interrupted and an individualised therapy with sorafenib was established. Treatment with sorafenib was characterised by multiple adverse effects and stopped. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
In this case, sorafenib and panobinostat were the most effective compounds. Treatment with 10-100 nM panobinostat and 100 µM sorafenib resulted in nearly 100% reduction of cell viability. As shown in Figure 5A, a significant reduction of cell viability could also be observed after treatment with 100 µM selumetinib.

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
The activity of NIS, in terms of RAI-U, was only significantly up-regulated after 24 h of treatment with 10 µM selumetinib ( Figure 5B). After 48 h of treatment, no increase of RAI-U was observed.

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
Alone the treatment with 10 nM panobinostat induced a significant over-expression of SLC5A5 ( Figure 6A). Neither treatment with panobinostat nor selumetinib nor sorafenib caused a significant change of HMGA2 transcript level.
Panobinostat, selumetinib and sorafenib caused a significant down-regulation of hsa-let7f-5p ( Figure 6B). Hsa-let7b-5p was stably expressed in comparison with untreated cells. After treatment with every single compound, an up-regulation of hsa-miR-146b-5p was observed. Furthermore, selumetinib and panobinostat induced an over-expression of hsa-miR-146b-3p. The activity of NIS, in terms of RAI-U, was only significantly up-regulated after 24 h of treatment with 10 µM selumetinib ( Figure 5B). After 48 h of treatment, no increase of RAI-U was observed.

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
Alone the treatment with 10 nM panobinostat induced a significant over-expression of SLC5A5 ( Figure 6A). Neither treatment with panobinostat nor selumetinib nor sorafenib caused a significant change of HMGA2 transcript level.

Expression of NIS Protein Level
Treatment with 10 µM sorafenib and especially with 10 nM panobinostat was able to significantly increase the protein level of NIS as shown in Figure 6C; the treatment with 10 µM selumetinib caused no change of NIS protein. Figure 6D shows the densitometry analysis of NIS protein.
Once again, it could be observed the highest efficacy of panobinostat able to reduce cell viability at lower concentrations than TKI. Furthermore, panobinostat was able to up-regulate NIS transcript and protein level. This effect could be inversely correlated with the suppression of hsa-let7f-5p. In Figure 6. Effects of selumetinib, sorafenib and panobinostat on mRNAs, miRNAs and NIS protein expression in PDTT of patient 3 (ATC). RT-qPCR of NIS, HMGA2 (A), hsa-let7f-5p, hsa-let7b-5p, hsa-miR-146b-5p and hsa-miR-146b-3p (B). The expression was normalised to GAPDH (A) and RNU6 (B) set at 1.0. Shown are means of experiments performed in triplicates ± SEM. * p < 0.05 regarded as significant. Western blot detection of NIS protein in PDTT cells treated with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat (C). Beta-actin was detected as quality loading control; (D) densitometry of NIS protein level normalised to beta-actin. Shown are means of experiments performed in triplicates ± SEM. * p < 0.05 regarded as significant.

Expression of NIS Protein Level
Treatment with 10 µM sorafenib and especially with 10 nM panobinostat was able to significantly increase the protein level of NIS as shown in Figure 6C; the treatment with 10 µM selumetinib caused no change of NIS protein. Figure 6D shows the densitometry analysis of NIS protein.
Once again, it could be observed the highest efficacy of panobinostat able to reduce cell viability at lower concentrations than TKI. Furthermore, panobinostat was able to up-regulate NIS transcript and protein level. This effect could be inversely correlated with the suppression of hsa-let7f-5p. In addition, 10 µM selumetinib caused an increase of RAI-U and a down-regulation of hsa-let7f-5p. The miRNAs mi-146b were over-expressed after treatment with the three single compounds.

In Vitro Results of the Fourth Patient
The fourth patient was affected by PDTC and the overall survival was 396 weeks after his first diagnosis. Radiotherapy was applied three times between 2010 and 2011. The patient was then treated with carboplatin and etoposide. The second line treatment was based on sorafenib and it was characterised by adverse events. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
Selumetinib and sorafenib treatment caused a slight reduction of PDTT cell viability at 10 µM ( Figure 7A). The highest concentration (100 µM) determined a significant reduction of the growth curve with values near to zero. Panobinostat showed a significant effect with 10 and 100 nM. These low concentrations reduced the cell viability down to zero. addition, 10 µM selumetinib caused an increase of RAI-U and a down-regulation of hsa-let7f-5p. The miRNAs mi-146b were over-expressed after treatment with the three single compounds.

In Vitro Results of the Fourth Patient
The fourth patient was affected by PDTC and the overall survival was 396 weeks after his first diagnosis. Radiotherapy was applied three times between 2010 and 2011. The patient was then treated with carboplatin and etoposide. The second line treatment was based on sorafenib and it was characterised by adverse events. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKI and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
Selumetinib and sorafenib treatment caused a slight reduction of PDTT cell viability at 10 µM ( Figure 7A). The highest concentration (100 µM) determined a significant reduction of the growth curve with values near to zero. Panobinostat showed a significant effect with 10 and 100 nM. These low concentrations reduced the cell viability down to zero.

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
The function of NIS protein was monitored by RAI-U measurement after administration of the tree different compounds. RAI-U was not significantly up-regulated after 24 h of treatment with 10 µM selumetinib/10 µM sorafenib/10 nM panobinostat ( Figure 7B). After 48 h of treatment, no variation of the radioiodine up-take was observed.

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
The function of NIS protein was monitored by RAI-U measurement after administration of the tree different compounds. RAI-U was not significantly up-regulated after 24 h of treatment with 10 µM selumetinib/10 µM sorafenib/10 nM panobinostat ( Figure 7B). After 48 h of treatment, no variation of the radioiodine up-take was observed.

Expression of NIS Protein Level
Additionally, NIS protein level was analysed after 48 h of treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat. As shown in Figure 8C, the treatment with 10 µM selumetinib did not cause any change in NIS protein level in comparison with untreated cells. In

Expression of NIS Protein Level
Additionally, NIS protein level was analysed after 48 h of treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat. As shown in Figure 8C, the treatment with 10 µM selumetinib did not cause any change in NIS protein level in comparison with untreated cells. In addition, 10 µM sorafenib and 10 nM panobinostat, especially caused a significant increase of NIS protein, further confirmed by densitometry ( Figure 8D).

In Vitro Results of the Fifth Patient
The fifth patient was affected by PDTC and the overall survival was 174 weeks after his first diagnosis. The patient received once radiotherapy. Afterwards, the patient was treated with sorafenib and then with lenvatinib. Both treatments were interrupted because of severe adverse events. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKIs and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
In the PDTT cells of the fifth patient case, selumetinib caused a reduction of cell viability at 100 µM only. Treatment with sorafenib was able to reduce significantly the cell viability at 1 µM already. In addition, 10 and 100 nM panobinostat showed, once again, the strongest efficacy in cell viability reduction ( Figure 9A). addition, 10 µM sorafenib and 10 nM panobinostat, especially caused a significant increase of NIS protein, further confirmed by densitometry ( Figure 8D).

In Vitro Results of the Fifth Patient
The fifth patient was affected by PDTC and the overall survival was 174 weeks after his first diagnosis. The patient received once radiotherapy. Afterwards, the patient was treated with sorafenib and then with lenvatinib. Both treatments were interrupted because of severe adverse events. RAI-R was diagnosed. No mutation of BRAF was detected ( Table 1). The treatment with TKIs and panobinostat was applied at the cells derived from his resected tumor after they have been kept proliferating in in vitro culture conditions.

Effects on Cell Viability of the Individual Tumor Cells Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
In the PDTT cells of the fifth patient case, selumetinib caused a reduction of cell viability at 100 µM only. Treatment with sorafenib was able to reduce significantly the cell viability at 1 µM already. In addition, 10 and 100 nM panobinostat showed, once again, the strongest efficacy in cell viability reduction ( Figure 9A).

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
As shown in Figure 9B, the treatment with the three compounds showed an unchanged radioiodine up-take after 24 h and 48 h in comparison with untreated cells.

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
All three compounds were able to cause an over-expression of NIS transcript ( Figure 10A). Sorafenib and panobinostat caused an over-expression of HMGA2 transcript, whereas selumetinib down-regulated its expression ( Figure 10A).

Radioiodine Up-Take after Treatment with Sorafenib, Panobinostat and Selumetinib
As shown in Figure 9B, the treatment with the three compounds showed an unchanged radioiodine up-take after 24 h and 48 h in comparison with untreated cells.

Expression of SLC5A5 (NIS), HMGA2 and miRNAs
All three compounds were able to cause an over-expression of NIS transcript ( Figure 10A). Sorafenib and panobinostat caused an over-expression of HMGA2 transcript, whereas selumetinib down-regulated its expression ( Figure 10A).

Expression of NIS Protein Level
NIS protein level was significantly reduced after treatment with TKIs and HDACI ( Figure 10C) and confirmed by densitometry analysis ( Figure 10D).
Here, it could be observed a significant cytotoxic effect exerted by treatment with sorafenib and panobinostat. All three compounds were able to increase the expression of NIS transcript.
However, no increase of RAI-U was observed and NIS protein level decreased after treatment with the three compounds. The miRNAs were up-regulated after treatment, thus could not be inversely correlated with NIS over-expression. Additionally, the onco-miRNAs 146b were over-expressed after treatment with the three compounds.

Expression of NIS Protein Level
NIS protein level was significantly reduced after treatment with TKIs and HDACI ( Figure 10C) and confirmed by densitometry analysis ( Figure 10D).
Here, it could be observed a significant cytotoxic effect exerted by treatment with sorafenib and panobinostat. All three compounds were able to increase the expression of NIS transcript.
However, no increase of RAI-U was observed and NIS protein level decreased after treatment with the three compounds. The miRNAs were up-regulated after treatment, thus could not be inversely correlated with NIS over-expression. Additionally, the onco-miRNAs 146b were over-expressed after treatment with the three compounds.

Effects on Cell Viability of Human Fibroblast Caused by Treatment with Sorafenib, Panobinostat and Selumetinib
Human fibroblasts (HF), used as control cells, were affected, in general, by treatment with TKIs and HDACI. In particular, cell viability was reduced by 1 to 100 µM selumetinib, 100 µM sorafenib and 10 to 100 nM panobinostat ( Figure 11A). Human fibroblasts (HF), used as control cells, were affected, in general, by treatment with TKIs and HDACI. In particular, cell viability was reduced by 1 to 100 µM selumetinib, 100 µM sorafenib and 10 to 100 nM panobinostat ( Figure 11A). The transcript level of NIS was significantly over-expressed after treatment with the three compounds ( Figure 12A). HMGA2 was significantly down-regulated after treatment with 10 µM selumetinib and 10 nM panobinostat ( Figure 12A). Furthermore, it could be observed that 48 h treatment with panobinostat induced the over-expression of has-let7f-5p, has-miR-146b-5p and hsa-miR-146b-3p. In addition, 10 µM sorafenib reduced significantly the expression of hsa-let7b-5p ( Figure 12B). The transcript level of NIS was significantly over-expressed after treatment with the three compounds ( Figure 12A). HMGA2 was significantly down-regulated after treatment with 10 µM selumetinib and 10 nM panobinostat ( Figure 12A). Furthermore, it could be observed that 48 h treatment with panobinostat induced the over-expression of has-let7f-5p, has-miR-146b-5p and hsa-miR-146b-3p. In addition, 10 µM sorafenib reduced significantly the expression of hsa-let7b-5p ( Figure 12B).

Expression of NIS Protein Level
The protein level of NIS was significantly up-regulated after treatment with 10 µM selumetinib and stable after treatment with 10 µM sorafenib and with 10 nM panobinostat ( Figure 12C). The protein expression was analysed densitometrically ( Figure 12D).
The human fibroblast, used as control, showed a reduction of cell viability mediated by TKIs and panobinostat. Additionally, these cells had an increase of RAI-U independently of the administered compound together with the over-expression of NIS transcript. Even HMGA2 was down-regulated after treatment with 10 µM selumetinib and 10 nM panobinostat. MiRNAs were generally stable or up-regulated after treatment, except for hsa-let7b-5p that was down-regulated by 10 µM sorafenib.
We could find evidence that the effect of different compounds had a similar effect in normal cells used as control for the proposed study.

Discussion
This study first focused on the evaluation of variable responses of PDTTs derived from PDTCs and ATCs to the treatment with two different TKIs, selumetinib and sorafenib, and one HDACI, panobinostat. The study focused on the expression of NIS, HMGA2, associated miRNAs and RAI-U modulation after treatment.
Our findings revealed that every PDTT enrolled in the study showed variable individual sensitivity to the treatment with TKIs and HDACI.

Expression of NIS Protein Level
The protein level of NIS was significantly up-regulated after treatment with 10 µM selumetinib and stable after treatment with 10 µM sorafenib and with 10 nM panobinostat ( Figure 12C). The protein expression was analysed densitometrically ( Figure 12D).
The human fibroblast, used as control, showed a reduction of cell viability mediated by TKIs and panobinostat. Additionally, these cells had an increase of RAI-U independently of the administered compound together with the over-expression of NIS transcript. Even HMGA2 was down-regulated after treatment with 10 µM selumetinib and 10 nM panobinostat. MiRNAs were generally stable or up-regulated after treatment, except for hsa-let7b-5p that was down-regulated by 10 µM sorafenib.
We could find evidence that the effect of different compounds had a similar effect in normal cells used as control for the proposed study.

Discussion
This study first focused on the evaluation of variable responses of PDTTs derived from PDTCs and ATCs to the treatment with two different TKIs, selumetinib and sorafenib, and one HDACI, panobinostat. The study focused on the expression of NIS, HMGA2, associated miRNAs and RAI-U modulation after treatment.
Our findings revealed that every PDTT enrolled in the study showed variable individual sensitivity to the treatment with TKIs and HDACI.
Here, we could confirm-to our knowledge for the first time-the high potency of panobinostat in comparison with TKIs for the treatment of ATCs-and PDTCs-derived PDTT cells. This finding also confirms previous studies showing panobinostat as an effective treatment option for aggressive thyroid cancer [26,27]. In particular, panobinostat showed the strongest cytotoxic effect in all PDTTs independently of their tumor type. Except for the first patient-derived tumor tissue culture, panobinostat showed already an overall cytotoxic effect at 10 nM. The response to sorafenib and selumetinib was quite different in all PDTT cells. Their cytotoxic effects were detected only at higher concentrations than panobinostat. In addition, 100 µM sorafenib was cytotoxic in all PDTTs, whereas 100 µM selumetinib caused a reduction of cell viability in the PDTT cells of patient 2, 3 and 4. Interestingly, the three compounds caused a significant reduction of cell viability in human fibroblasts used as control.
It is already known that treatment with panobinostat may lead to NIS RNA and protein expression and to a relevant (125) radioiodine up-take in ATC [28,29]. Hong et al. even stated that HDACIs can induce the RAI-U in patients affected by RAI-R thyroid cancer [30].
In our study, all PDTT cells showed a strong significant over-expression of NIS transcript after treatment with panobinostat. In one PDTC (patient 5) only, all three compounds induced an up-regulation of NIS. In all the other cases neither selumetinib nor sorafenib were able to up-regulate the NIS transcript.
The different sensitivity of the PDTT cells to the treatment with different compounds highlights the importance of using primary tumor cells to individualise the most efficacious therapy for patients with these thyroid cancers and to exclude the ineffective compounds. Moreover, this study revealed the poor efficacy of TKIs in terms of NIS restoration/over-expression at transcript level. Furthermore, the function of NIS protein was analysed by RAI-U measurement after administration of the three different compounds. The present study could confirm that the treatment with selumetinib, sorafenib and panobinostat was able to increase the RAI-U in thyroid cancer cells as previously shown [31]. However, we could observe quite different reactions-three PDTTs (patients 1, 2 (ATC) and 4 (PDTC)) showed an increased RAI-U after treatment with all three compounds, one PDTC (patient 5) showed no RAI-U and one ATC-derived PDTT (patient 3) showed a significant up-regulation of NIS activity after treatment with 10 µM selumetinib. In contrast with the finding of Ho et al. that BRAFV600E-mutated cell lines show a better RAI-U after the treatment with selumetinib [32], it could be here observed that the RAI-U was not correlated with the mutation status of BRAFV600E of the PDTT cancer cells.
As previously reported, NIS expression could be modulated by the expression of miRNAs belonging to Let7 family. In particular, NIS transcript in silico analysis revealed NIS as validated target of hsa-let7f-5p [23]. Four PDTTs (patients 1, 2, 3 and 4) showed a down-regulation of hsa-let7f-5p after treatment with panobinostat; only one PDTC (patient 5) did not show any change. As previously shown, hsa-let7f-5p was found to be up-regulated in all histopathological subgroups of TC [17]. The possibility to suppress its expression in order to restore NIS transcript could represent a valuable target for the treatment of ATC and PDTC.
The stable/up-regulated expression of hsa-let7b-5p after treatment could be inversely correlated with the down-regulation of HMGA2, a non-transcription factor with well-known oncogenic properties as previously shown in liver cancer cell lines [25]. Thus, showing that the treatment with TKIs and HDACI can re-establish an inverse correlation between hsa-let7b-5p and HMGA2, which was found to be absent in ATC samples of patients collected in the previous study [23].
The oncogene HMGA2 was stably expressed or down-regulated after the treatment with panobinostat, selumetinib and sorafenib in all cases included in the study. Only in one case of PDTC (patient 5), sorafenib and panobinostat induced an up-regulation of HMGA2.
Interestingly, in our study, it could be found that these miRNAs were stable and or up-regulated in four PDTTs (patients 1, 2, 3 and 5) derived from ATC and PDTC, whereas patient 4 (PDTT derived from PDTC) showed a down-regulation of both hsa-miR-146b-5p and hsa-miR-146b-3p after treatment with 10 µM selumetinib. It becomes clear that these miRNAs could play a key role in the cell death mechanisms induced by treatment with TKIs and panobinostat. Further investigations are needed to clarify the mechanisms regulated by hsa-miR-146b in ATC and PDTC.
Finally, NIS protein level was found regulated by treatment with selumetinib, sorafenib and panobinostat. In particular, patient 1-derived PDTT cells, which have no basal level of NIS, showed an over-expression of it after treatment with 10 µM selumetinib, 10 µM sorafenib and 10 nM panobinostat. Furthermore, patients 3 and 4 PDTT cells showed an over-expression of NIS protein after treatment with 10 µM sorafenib and 10 nM panobinostat. The level of NIS protein was down-regulated after treatment with the three compounds in PDTT from patients 2 and 5.
In terms of correlation between NIS transcript expression, radioiodine up-take and expression of NIS protein after treatment with TKI and panobinostat, it could be found a discrepancy between all PDTT used in the study. Even though NIS transcript was generally up-regulated after treatment, its activity in terms of radioiodine up-take and its protein level were differently modulated. Only the PDTTs derived from patients 1 and 4 showed an up-regulation of NIS transcript, activity and protein level after treatment, especially, with 10 nM panobinostat. The other patients showed a stable radioiodine up-take or even a suppression of NIS protein level that could not be functionally correlated with the significant over-expression of NIS transcript. Here, we could only hypothesize that some patients could express a non-functional/distorted NIS protein, as previously reported [10]; and/or that NIS protein membrane localization and activity is dependent on its glycosylation status as previously described in thyroid cancer [10]. Further studies should focus on the activity of NIS promoter as it has been previously shown in breast cancer cells [33]. Additionally, the post-translational modifications occurring at NIS protein should be further investigated to better understand its function and its subcellular localization.
Nevertheless, there are some limitations referring the method of analysing PDTT of ATCs/PDTCs. Concerning the clinical course of the patients enrolled in this study, it has been observed that the patients suffering from ATC are mainly older people with multiple comorbidities. Furthermore, it is well known that ATCs are very aggressive tumors with early invasion of surrounding tissue. To our knowledge, the procedure of preparing PDTT including the testing of different compounds takes about six weeks-this might be a long time for patients suffering of ATC; patients 1 and 2 died after less than six weeks. Moreover, it is a time-and money-consuming method to prepare PDTT. In our opinion, it is a promising method for selected patients suffering from ATCs and especially PDTCs with no more treatment options. Considering the clinical course of the patients suffering from PDTC, they obviously need individual and personalised treatment options suitable for their individual tumors. Additionally concerning the PDTCs, they might change their tumor behaviour during the progression of the disease and develop resistances after radioiodine therapy, which might require alternative treatment options. Looking at patient 5, suffering from PDTC, sorafenib was an effective drug at the beginning, but, in the further clinical course, an elevation of the tumor marker thyroglobuline (TG) was recorded, thus providing evidence of a tumor recurrence. An additional treatment with lenvatinib caused again a decline of TG.

Conclusions
In conclusion, a successful treatment strategy might be on in vitro testing of patient-derived tumor tissue, since every ATC-or PDTC-derived PDTT shows different response to the treatment with certain HDACIs and/or TKIs. Considering the clinical course of the patients described in this study, the treatment with TKIs and HDACIs might be associated with multiple adverse events, which might require a dose reduction or even the stop of a therapy. Patient 3, for example, needed a dose reduction and furthermore caused even the termination of the therapy with sorafenib due to increasing polyneuropathy. In this case, the need of alternative treatment options-for example in terms of HDACIs-would be required. Especially, HDACI treatment of ATC/PDTC seems to be promising, thus leading to a better prognosis in patients affected by these rare malignancies.