Human papillary thyroid cancer cell line B-CPAP and human follicular thyroid cancer cell line FTC-133 were from Professor G. Brabant (Department of Endocrinology, Christie Hospital, Manchester, UK). Human anaplastic thyroid cancer cell line Cal-62 was obtained from Leibniz Institute DSMZ, German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). B-CPAP cells were cultivated in RPMI-1640 medium, Cal-62 cells in DMEM, and FTC-133 cells in DMEM/Ham’s F12. Media were supplemented with 10% fetal calf serum (FCS), 1% penicillin-streptomycin and L-glutamine.

Cell solubilisation in the presence of protease and phosphatase inhibitors and western blot analyses were carried out as described (17).

Stock solutions and working concentrations were as follows: for MK-2206 (Selleck Chemicals), 10 mM and 1 μM, for PD0325901 (Axon Medchem) 10 mM and 2 μM, for sorafenib (Enzo Life Sciences) 10 mM and 5 μM, for U0126 (Promega) 10 mM and 20 μM, for wortmannin (Santa Cruz) 1 mM and 100 nM. All stock solutions were prepared in DMSO and stored at −20°C. In a typical experiment, cells were cultivated for three days in the absence or presence of the compound that was added once after 36 h or twice after 24 and 48 h.

To analyze cell proliferation, 500 cells in 100 μl culture medium were seeded in 96-well plates (12 wells per time-point and treatment). Drugs were added immediately after 24 as well as after 48 h in 100 μl of culture medium. Addition of pure DMSO served as a medium control. Cells were fixed in 4% formaldehyde in PBS just before treatment (T0) or after 72 h of treatment (T1). Cells from both time-points were then stained with DAPI (1 μg/ml in PBS) in order to determine microscopically the cell number in a defined area of each well.

Cell suspensions (5,000 cells/μl) in a total volume of 3 μl were seeded on defined regions in Petri dishes and allowed to adhere for two to three hours. After floating the dishes with culture medium, the adherent and confluent cells occupied a circular area. Diameters of areas (12 per dish) were determined microscopically subsequently after floating (T0) and after a 72-h culture period (T1). Compounds were added at T0 as well as 24 and 48 h later.

Cytotoxicity was analyzed with the LDH Cytotoxicity Assay Kit from Roche. Briefly, 10,000 cells (in 100 μl culture medium) were seeded per well in 96-well plates. Twenty-four hours later, cells were exposed to different drug concentrations for 24–72 h. After treatment, LDH activity was determined in the cell culture supernatants. In parallel, cells that had been treated identically were lysed in order to determine total LDH activity, a value that allows to quantify cell proliferation.

Human papillary cell line B-CPAP, anaplastic cell line CAL-62 and follicular cell line FTC-133 were maintained for 48 h in the presence or absence of serum, or for 48 h in the absence and finally for 20 min in the presence of serum. Under serum-free and serum-containing culture conditions, no considerable differences in the expression levels of pERK1/2 or pAkt were observed (Fig. 1). Stimulation of quiescent cells with a serum pulse did not lead to an increased pERK1/2 and pAkt activity, suggesting a constitutive activation of Akt and ERK1/2 in thyroid cancer cells to an almost maximal extent.

We first optimized the conditions to achieve maximal inhibition rates after drug application. Drugs were applied for 72 h either without further medium change, with one medium change after 36 h or with two medium changes after 24 and 48 h (Fig. 2). The most efficient inactivation of ERK1/2 kinases was achieved when U0126 or PD0325901 were added daily or at least once after 36 h (shown for U0126 in Fig. 2A). An almost complete inactivation of Akt over the whole three-day incubation period was obtained with a single addition of 1 μM MK-2206 in B-CPAP and with a twice repeated addition in FTC-133 cells. In contrast, in Cal-62 only an incomplete inhibition was detected, even when Mk-2206 was renewed daily (Fig. 2B) or its concentration was increased to 2 μM (not shown). Treatment with multikinase inhibitor sorafenib did not affect the activation of Akt and ERK1/2 in Cal-62 and FTC-133 cells, but strongly reduced the pERK1/2 expression level in B-CPAP cells at a concentration of 5 μM (Fig. 2C). PI3K inhibitor wortmannin efficiently inhibited pAkt expression, but due to the instability of the drug, this effect was stable only for short time periods after application (Fig. 2D, compare incubation times of 4 and 24 h). Such types of experiments also allow to analyze the presence of crosstalk between PI3K/Akt and the Ras/MAPK pathways. Indications for cross-talks of the two pathways were mostly observed after a more frequent drug application such as the upregulation of pERK1/2 upon MK-2206 application in B-CPAP cells or the downregulation of pERK1/2 upon wortmannin treatment in FTC-133 cells (Fig. 2). Based on these results, in three-day lasting functional assays, drugs were replaced every 24 h and applied at a concentration of 1 μM for PD0325901, MK-2206 and sorafenib and at 20 μM for U0126. Because of its low stability, wortmannin was not included in long-lasting experiments.

Low density thyroid carcinoma cells were allowed to proliferate in 96-well plates for three days in the absence or presence of compounds. Different sets of cells were stained with DAPI before and after the 72-h incubation period and cell numbers were determined microscopically (Fig. 3 and Table I). Perturbation of the Ras/MAPK pathway led to a strong reduction in cell number, whereby inhibition rates of 70–90% were observed for the different cell lines after U0126 treatment (Fig. 3 and Table I). For PD0325901, inhibition rates varied between 53% for FTC-113 cells and 84% for B-CPAP cells (Table I). Also with the multikinase inhibitor sorafenib proliferation of thyroid carcinoma cells could be efficiently inhibited in the range of 70 to ~90% (Table I). Disruption of the PI3K/Akt pathway with MK-2206 inhibited the proliferation rate by ~65% in B-CPAP and by ~40% in Cal-62 and FTC-133 cells.

Thyroid carcinoma cells were seeded at high density on defined areas into Petri dishes and allowed to adhere for two to three hours to establish a circular monolayer. After floating with medium, cells were cultured for three days in the presence or absence of drugs (Fig. 4A for FTC-133 cells) and the diameters of the monolayers were determined at the beginning and the end of the culture period (Table II). In general, the migration capacity of thyroid carcinoma cells was suppressed to a varying extent in the presence of the different kinase inhibitors. For B-CPAP cells, inhibition rates in the range of 30 (for U0126) up to 80% (for sorafenib) were observed. The migration capacity of FTC-133 cells was relatively robust with regard to drug treatment (inhibition rates in the range of 7–50%). Surprisingly, MEK/ERK1/2-inhibitor treated Cal-62 cells showed an increased migration rate, notwithstanding a strong inhibitory effect in the range of 45 or 80% became visible upon MK-2206 or sorafenib treatment. Eventhough statistical significance is only present for U0126-treatment, the data shown in Fig. 4B demonstrate that in more than half of the experiments performed with PD0325901 an increased migration also took place.

We thus wondered, if the stimulating effect of U0126 and PD0325901 on the migration of Cal-62 cells could be suppressed by a simultaneous inhibition of the PI3K/Akt pathway, because western blot analyses revealed that the combined treatment of thyroid carcinoma cells with U0126 and MK-2206 led to an almost complete downregulation of pAkt and pERK1/2 expression (Fig. 5). Indeed, the migration rate of U0126/MK-2206 dual-treated Cal-62 cells was substantially inhibited not only in respect to U0126- or PD0325901-single-treated, but also in comparison to untreated control cells (Table IIIA). Moreover, although sorafenib did not affect Akt activation in U0126- or PD0325901-treated Cal-62 cells (Fig. 5), its presence led to a substantial inhibition of cell migration (Table IIIA). To rule out that the inhibition observed results mainly from cytotoxic drug effects, LDH assays were carried out. After a 24-h incubation period none of the single or dual drug treatments induced a noteworthy cytotoxic effect in Cal-62 cells, although dual-treated B-CPAP cells were considerably affected (Fig. 6A). After a 72-h incubation period, however, a moderate cytotoxic effect was observed in sorafenib-treated Cal-62 cells. Therefore, it is possible that under these conditions the proliferation and migration suppressing potential of this compound is partially due to a cytotoxic effect. Thus, in a final set of experiments, we applied sorafenib at concentrations of 1 and 2 μM, i.e., at concentrations that did not induce considerable cytotoxic effects in Cal-62 cells after a 72-h incubation period neither as a single agent nor in combination with 20 μM U0126 or 2 μM PD0325901 (Fig. 6C). Under these conditions, the migration-promoting effect of U0126 was completely compensated, but an additional inhibition, as could be monitored for higher sorafenib concentrations, was not observed (Table IIIB). Altogether, these data suggest that the migration-stimulating effect of PD0325901 and U0126 can be suppressed by Akt-dependent and -independent mechanisms, even though an inactivation of Akt downstream signaling molecules cannot be completely excluded.

PI3K/Akt and Ras/MAPK signaling pathways, although first described as linear conduits were soon shown to be connected with each other in a complicated fashion that comprises cross-activation and -inhibition processes as well as negative feedback loops and pathway convergence (18). In our experiments we found evidence for the presence of cross-inhibitory circuits of PI3K/Akt on Ras/MAPK in B-CPAP and of cross-activating circuits in FTC-133 cells (Fig. 2). As such crosstalk depends inter alia on substrate availability that can vary between different cell types or cell lines it is not surprising that the interactions we observed exhibit a cell line-specific pattern that can not be generalized (Fig. 2). Kandil and colleagues (19) also recently reported that cross-activation of Akt varies between different thyroid cancer cell lines upon treatment with the MEK/ERK inhibitor AZD6244. At the present level of knowledge, it is mainly assumed that the PI3K/Akt and Ras/MAPK pathways can negatively regulate each other and, in addition, the latter one can activate the former (18). Nevertheless, and thereby underlining our data, in a few reports it has been shown, that the PI3K/Akt pathway can activate ERK1/2 signaling (20–22).

Although we have evidence for the presence of crosstalk between PI3K/Akt- and Ras/MAPK pathways in thyroid carcinoma cells, it does not illuminate the yin-yang effect we detected in Cal-62 cells treated with MEK inhibitors:

Collectively, it seems feasible that interference with a single pathway can inhibit one cellular process, here the cell’s proliferative capacity, and at the same time promote another, here the cell’s migratory potential. In our example such an effect can be interpreted by assuming the existence of an (hyper)activated Ras/ERK1/2 pathway that suppresses cell migration but can itself be suppressed by MEK inhibitors. Based on the above, one could expect that Cal-62 cells migrate with a submaximal velocity, and indeed, although we have not performed a systematic analysis, Cal-62 shows the lowest velocity amongst the three cell lines we have used, i.e., a 25 or 40% reduced migration rate in comparison to FTC-133 or B-CPAP cells, respectively. This resembles a scenario that hyperactivation of the Ras/ERK1/2 pathway can trigger a decreased proliferation rate, i.e., cell cycle arrest (23). An inhibitory effect of the MAPK cascade on cell proliferation is also active in K562 erythroleukemia cells, which differentiate into megakaryocytes upon phorbol ester stimulation, a process that is: i) accompanied by growth retardation and ERK activation, and ii) completely suppressed in the presence of MEK inhibitors (24). The aspect that Ras/ERK1/2 pathway activation not only activates (25,26), but may also suppress cell migration adds another facet to the complex pattern of MAPK signaling. To overcome the stimulatory effect of MEK inhibitors on Cal-62 cells, we have performed dual treatments to interfere with more than one signaling pathway and, thereby, successfully suppressed cell proliferation as well as migration. Such a strategy, in clinical practice known as a subtype of combination therapy (27), has been used in a number of trials to attack the fatal potential of tumor cells more efficiently. For example, in thyroid carcinoma cells, dual Ras/MAPK and PI3K/Akt/mTOR pathway inhibition leads to significant higher growth retardation even in a synergistic manner when compared to single pathway inhibition (28,29). Beside the advantage of a combinatorial treatment to increase the inhibitory action of a single drug such a strategy also decreases the risk of unfavorable yin-yang effects.