Five cancer cell lines derived from pediatric solid tumors were used in the present study. The neuroblastoma SH-SY5Y (ECACC 94030304), SK-N-BE(2) (ECACC 95011815), and rhabdomyosarcoma RD (ECACC 85111502) cell lines were purchased from the European Collection of Authenticated Cell Cultures (ECACC, Salisbury, United Kingdom). The medulloblastoma DAOY (ATCC HTB-186^™) and osteosarcoma Saos-2 (ATCC HTB-85^™) cell lines were obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States). All cell lines were authenticated by STR profiling and routinely tested negative for mycoplasma contamination by PCR.

All reagents for cell culture were purchased from Biosera (Nuaille, France). DAOY and Saos-2 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; low glucose, cat. no. LM-D1100) supplemented with 10% fetal calf serum (FCS; cat. no. FB-1101); RD cells were maintained in DMEM (high glucose, cat. no. LM-D1112) with 10% FCS, and SH-SY5Y and SK-N-BE(2) cells were cultured in a mixture of DMEM/F12 (1:1, cat no. LM-D1224) supplemented with 20% FCS. All media were further supplemented with 2 mM glutamine (cat. no. XC-T1715), penicillin (100 IU/ml), and streptomycin (100 μg/ml; cat. no. XC-A4122). The media used for DAOY, RD, SH-SY5Y, and SK-N-BE(2) cells also contained 1% nonessential amino acids (cat. no. XC-E1154). The cells were maintained under standard cell culture conditions at 37°C in a humidified atmosphere containing 5% CO₂ and were subcultured 1–2 times weekly.

The tyrosine kinase inhibitors SUN (cat. no. 12328), GEF (cat. no. 4765), and LAP (cat. no. 12121) were purchased from Cell Signaling Technology (Danvers, MA, United States). Thiosemicarbazones Dp44mT (cat. no. SML0186) and DpC (cat. no. SML0483) and Valspodar (VAL; cat. no. SML0572) were purchased from Sigma–Aldrich (St. Louis, MO, United States). All reagents were prepared as stock solutions in dimethyl sulfoxide (DMSO; purchased from Sigma–Aldrich) at the concentrations of 10 mM (LAP), 75 mM (SUN) or 100 mM (GEF, Dp44mT, DpC, and VAL).

Cell proliferation was evaluated after drug treatment using the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) assays. The cells were seeded at variable densities to ensure that they remained in the log growth phase during the drug treatments. For 24-hour treatment, the cells were seeded in 96-well plates at a density of 2×10⁴ cells/well (SH-SY5Y, SK-N-BE(2), Saos-2, and RD cells) or 5×10³ cells/well (DAOY cells). For 72-hour treatment, the cells were seeded at a density of 5×10³ cells/well (SH-SY5Y, SK-N-BE(2), Saos-2, and RD cells) or 8×10² cells/well (DAOY cells). After incubation with the drugs, the cells were incubated with MTT (0.5 mg/ml; cat. no. M2128, purchased from Sigma–Aldrich) for 3 h under standard cell culture conditions. Subsequently, the medium was removed, and formazan crystals were dissolved in 200 µl of DMSO. The absorbance was measured at 570 nm, and the reference absorbance was measured at 620 nm using a Sunrise absorbance reader (Tecan, Männedorf, Switzerland).

The MTT assay was used to determine the IC₅₀ values (concentration at which the cellular population was reduced by 50%) for each drug (SUN, GEF, LAP, Dp44mT, or DpC). 24 h after seeding, the medium was replaced with 200 µl of the fresh medium containing appropriate concentrations of the drugs alone. After incubation for 72 and/or 24 h under standard cell culture conditions, the cells were incubated with MTT and analyzed as described above. The IC₅₀ value was assessed using CalcuSyn software (version 2.0, Biosoft, Cambridge, United Kingdom).

The MTT assay was used to quantify the synergy between thiosemicarbazones and TKIs. Based on the initially calculated IC₅₀ values of each drug, the drug concentrations used in the combined treatment experiments corresponded to 1/8-, ¼-, ½-, 1-, 2-, 4-, and 8-fold of the IC₅₀ using methodology as reported previously (Paukovcekova et al., 2020). The cells were seeded as described above. After 24 h, the medium was replaced, and the cells were incubated with the drugs alone or in combination under standard cell culture conditions. Different experimental designs were utilized to assess the effects of the combinations of thiosemicarbazones Dp44mT and DpC with TKIs SUN, GEF, and LAP (Figure 1).

In the experiments with sequential administration of the drugs (Figure 1, Pretreatment designs), the cells were initially treated with an appropriate concentration of SUN, GEF, or LAP (Pretreatment I), SUN, GEF, or LAP with 0.2 µM VAL (Pretreatment I + VAL), or Dp44mT or DpC (Pretreatment II). After 48 h, 100 µl of the fresh medium was added, and the medium contained an appropriate concentration of Dp44mT or DpC (Pretreatment I), Dp44mT or DpC with 0.2 µM VAL (Pretreatment I + VAL), or SUN, GEF, or LAP (Pretreatment II). In the simultaneous treatment experiments (Figure 1, Simultaneous treatment), the cells were treated with an appropriate concentration of individual drugs or their combinations and incubated for another 72 h. After the drugs were incubated according to the corresponding experimental design, cell proliferation was analyzed by the MTT assays as described above.

To investigate the molecular effects of TKIs in combination with thiosemicarbazones, the cells were seeded in Petri dishes (90 mm in diameter) and allowed to adhere overnight. To reproduce the simultaneous treatment design (Figure 1), thiosemicarbazones and TKIs alone or in combination were added at the corresponding IC₅₀ concentrations. The cells were incubated under standard cell culture conditions in the presence or in the absence (control) of the indicated drugs for 72 h before being processed for immunoblotting.

CalcuSyn software (version 2.0, Biosoft, Cambridge, United Kingdom) and the Chou Talalay method were used to calculate the combination index (CI) values as described previously (Paukovcekova et al., 2020). A 1:1 ratio of the drugs was used for combination treatments, and the CI values were calculated based on the growth inhibition curves. The dose-effect relationship for each drug alone was compared to the corresponding combination to identify the synergistic (CI < 0.9), additive (CI: 0.9–1.1), or antagonistic (CI > 1.1) interactions between thiosemicarbazones and TKIs (Chou, 2006).

The relative levels of phosphorylation of 49 RTKs (Supplementary Figure S1) were assayed using a Proteome Profiler™ human phospho-RTK array kit (cat. no. ARY001B) purchased from R&D Systems (Minneapolis, MN, United States). Treated and/or untreated control cells were lysed using lysis buffer 17 and processed according to the manufacturer’s instructions. Each array was incubated with 300 µg of the whole-cell lysate. The relative levels of RTK phosphorylation were quantified using Fiji software (Schindelin et al., 2012), and analysis was performed as described previously (Neradil et al., 2019).

Whole-cell lysates of treated and untreated control cells were loaded on 10% polyacrylamide gels (10–20 µg/well), electrophoresed, and blotted on the polyvinylidene difluoride membranes (purchased from Bio–Rad Laboratories, Munich, Germany). Depending on the primary antibody, the membranes were blocked either with 5% nonfat dry milk or with bovine serum albumin (BSA; Sigma–Aldrich) in phosphate-buffered saline (PBS) containing 0.1% Tween-20 (Sigma–Aldrich) for 1 h at room temperature and then incubated at 4°C overnight with the corresponding primary antibodies listed in Table 1. Then, the membranes were incubated with the corresponding secondary antibodies (Table 1) for 1 h at room temperature. Chemiluminescence detection was performed using Amersham™ ECL™ Prime Western blotting detection reagent (purchased from GE Healthcare, Little Chalfont, United Kingdom) according to the manufacturer’s instructions. Densitometry analyses were performed using Fiji software (Schindelin et al., 2012), and the densities of protein bands of interest were normalized to that of the loading control. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and alpha tubulin were used as the loading controls. Biological replicates were normalized using the sum of all data points in a replicate as described by Degasperi et al. (2014).

Quantitative data are shown as the mean ± standard deviation (SD) of three independent experiments. The data of the MTT assays of the combination treatments and the results of densitometry were analyzed using one-way ANOVA followed by Dunnett’s post-hoc test. All statistical analyses were performed using GraphPad Prism 8.0.2 software (GraphPad Software Inc., San Diego, CA, United States), and p < 0.05 was considered statistically significant.

The drugs investigated in the present study, including the TKIs SUN, GEF, and LAP and thiosemicarbazones Dp44mT and DpC, have been demonstrated to affect the critical oncogenic signaling pathways (Pollack et al., 2010; Chen et al., 2012; Dixon et al., 2013; Fouladi et al., 2013; Lui et al., 2015b; Kovacevic et al., 2016; Menezes et al., 2017; Mudry et al., 2017; Menezes et al., 2019a; Verschuur et al., 2019; Geleta et al., 2021). Therefore, we initially assayed the basal activity of 49 key RTKs in untreated cell lines derived from pediatric solid tumors, including the SH-SY5Y and SK-N-BE(2) neuroblastoma, Saos-2 osteosarcoma, RD rhabdomyosarcoma, and DAOY medulloblastoma cell lines (Figure 2; Supplementary Figure S2).

Overall, the results of screening using the human phosphoprotein arrays revealed high phosphorylation of multiple pro-oncogenic RTKs, several of which are known to be attenuated by the drugs used in the present study. The ErbB family of RTKs are known to be targeted by GEF, LAP, and both thiosemicarbazones (Pollack et al., 2010; Fouladi et al., 2013; Kovacevic et al., 2016; Menezes et al., 2017), and EGFR was consistently identified as one of the five most phosphorylated RTKs in all tested cell lines (Figure 2). In contrast, the activation of other ErbB family members was substantially lower (Figure 2). Regarding the targets of SUN (Mudry et al., 2017; Verschuur et al., 2019), we detected relatively active PDGFRα in Saos-2 cells (Figure 2D) and PDGFRβ in both SH-SY5Y and SK-N-BE(2) neuroblastoma cells (Figures 2A,B); however, the phosphorylation of other RTKs targeted by SUN, e.g., VEGFRs or FLT3, was minimal (Figure 2). Notably, IGF-1R signaling is modulated by thiosemicarbazones (Geleta et al., 2021). The levels of phosphorylation of IGF-1R, and insulin receptor (InsR) were very high in all cell types (Figure 2). These results indicated that the activation of prosurvival RTK signaling was shared among all pediatric tumor cell lines, further supporting the rationale for the investigation of the interactions of selected RTK-targeting drugs in the present study.

Then, we determined the antiproliferative effect of thiosemicarbazones and selected TKIs used as single treatment agents and calculated the concentrations of the drugs at which cell proliferation was reduced by 50% (IC₅₀). To determine the IC₅₀ values needed for the subsequent combined treatment experiments, cell proliferation was evaluated after 24- and/or 72-hour incubation with individual drugs. In the case of TKI treatment, all tested cells were most sensitive to SUN and least sensitive to GEF after incubation for both 24 and 72 h (Table 2). Dp44mT and DpC administered for 24 h were most effective in DAOY medulloblastoma cells, whereas other cell lines showed ∼6- to 170-fold lower sensitivity (Table 2). After 72 h of incubation, SH-SY5Y and SK-N-BE(2) neuroblastoma cells showed higher sensitivity to thiosemicarbazones than other cell types (Table 2). As expected, all examined drugs demonstrated higher efficacy (corresponding to lower IC₅₀ values) after prolonged incubation. This phenomenon was particularly evident in the case of thiosemicarbazones; the corresponding IC₅₀ values for these compounds shifted from a micromolar range after 24-hour treatments to a nanomolar range after 72-hour treatments (Table 2). Furthermore, after incubation for 72 h, both Dp44mT and DpC were significantly more effective than SUN, GEF, or LAP in all cell lines (Table 2).

Based on these data, we comprehensively evaluated the interactions between thiosemicarbazones and TKIs using several combined treatment designs (Figure 1): 1) sequential treatment protocols (Pretreatment I, Pretreatment I + VAL, and Pretreatment II) and 2) a simultaneous treatment protocol (Simultaneous treatment). The results of these experiments were recalculated as the CIs under various settings to assess the effects of the combinations of Dp44mT or DpC thiosemicarbazones with selected TKIs (Figure 3; Supplementary Table S1).

Based on the studies that have demonstrated that drug resistance to TKIs is mediated by lysosomal sequestration (Krchniakova et al., 2020) and on the ability of thiosemicarbazones to release the drugs, such as doxorubicin, sequestered into lysosomes (Jansson et al., 2015; Seebacher N. et al., 2016, Seebacher et al., 2016 N. A.), we hypothesized that the Pretreatment I strategy (Figure 1, Supplementary Table S1) may be a potent approach to enhance the anticancer effects of TKIs.

To test this hypothesis, the cells were initially treated with individual TKIs for 48 h and subsequently treated with Dp44mT or DpC for the next 24 h (Figure 1). However, the results of CI analyses revealed that this sequential treatment induced predominantly antagonistic drug interactions in all tested cell lines with the exception of DAOY medulloblastoma cells (Figure 3). These results contradicted our initial hypothesis.

SH-SY5Y cells express P-glycoprotein (Pgp) (Dalzell et al., 2015), which is an extensively studied ABC transporter involved in drug resistance (Wu et al., 2011). In addition to the drug efflux activity of Pgp at the plasma membrane, Pgp at the lysosomal membranes has been implicated in trapping of TKIs (Krchniakova et al., 2020) and inducing the accumulation of thiosemicarbazones in the lysosomes (Jansson et al., 2015; Seebacher N. A. et al., 2016, Seebacher et al., 2016 N.). To determine whether Pgp activity plays a role in the effects of the tested drugs, the selective Pgp inhibitor VAL at a concentration of 0.2 µM (Jansson et al., 2015) was added to the assay (Pretreatment I + VAL design; Figure 1). The most pronounced differences between the VAL(−) and VAL(+) conditions were detected in case of LAP + Dp44mT or DpC combinations in Pgp-expressing SH-SY5Y cells, where moderate antagonistic interactions changed to synergy (Figure 3A). In other cell lines, no consistent trends in the changes in the CI values were detected in case of the VAL(−) and VAL(+) treatment strategies (Figure 3).

To further elaborate on these results, we compared effective concentrations of the tested drugs under VAL(−) and VAL(+) conditions in Pgp-expressing SH-SY5Y cells. Unlike SUN and GEF, which are able to inhibit Pgp (Kitazaki et al., 2005; Shukla et al., 2009), thiosemicarbazones are transported by Pgp (Jansson et al., 2015; Seebacher N. A. et al., 2016), and LAP has been described as both Pgp substrate and inhibitor (Dai et al., 2008; Radic-Sarikas et al., 2017). Despite reported differences in the interactions of these drugs with Pgp, no significant changes in the efficacy of the drugs were detected by comparison of VAL-treated and control SH-SY5Y cells (Supplementary Figure S3). However, Pgp is known to export a wide variety of xenobiotics, metabolites, and toxins (Aller et al., 2009), and blockade of these functions by VAL could have explained enhanced cytotoxic effects of combined treatments in SH-SY5Y cells with inhibited Pgp (Pretreatment I + VAL design; Figure 3A). The inhibition of Pgp has been shown to prevent the accumulation of Dp44mT or DpC in the lysosomes, leading to a significant reduction in the cytotoxicity of these drugs in carcinoma cells (Jansson et al., 2015); however, the results of the present study suggested that this Pgp-dependent mechanism of action was not present in pediatric tumor cells. The inhibition of Pgp did not reduce the sensitivity of the tested cells to combined treatments with thiosemicarbazones. In the case of Pgp-overexpressing SH-SY5Y cells, the blockade of the Pgp activity even potentiated the effects of the drug combinations (Figure 3). These findings in combination with the failure of thiosemicarbazones to improve the efficacy of TKIs in TKI-pretreated cells indicated that the anticancer activity of thiosemicarbazones in pediatric solid tumor cells is most likely not facilitated through a lysosomal burst.

Since the treatments with thiosemicarbazones following the TKI pretreatment (Pretreatment I design) did not result in synergistic interactions in most cell lines, we explored another sequential design, Pretreatment II (Figure 1; Supplementary Table S1). According to this strategy, the drugs were administrated sequentially in reverse order: the treatment with Dp44mT or DpC for 48 h was followed by the addition of a TKI for the next 24 h (Figure 1). Considering that the drugs used in the present study interact with Pgp (Seebacher N. A. et al., 2016; Krchniakova et al., 2020), the neuroblastoma cell lines with different levels of Pgp expression, including SH-SY5Y (high Pgp expression) and SK-N-BE(2) cells (low Pgp expression) (Bates et al., 1989; Dalzell et al., 2015), were selected for these analyses.

Interestingly, in SK-N-BE(2) cells, the treatment according to the Pretreatment II design induced a significant decrease in the CI values compared to Pretreatment I, resulting in the uniform synergistic interactions between the tested drug combinations (Figure 3B). Additive/synergistic effects were also observed in SH-SY5Y cells when SUN or GEF was combined with Dp44mT or DpC (Figure 3A). These results suggested that thiosemicarbazones apparently sensitized pediatric solid tumor cells to TKIs. However, the combinations of LAP and thiosemicarbazones remained antagonistic in SH-SY5Y cells (Figure 3A), further emphasizing the role of Pgp in drug resistance.

Overall, sequential treatment protocols induced variable and generally unsatisfactory interactions between thiosemicarbazones and TKIs across the pediatric cancer cell lines, thus excluding the therapeutic potential of this approach.

Our previous report demonstrated that Dp44mT and DpC added simultaneously with celecoxib produce synergistic effects on pediatric cancer cells (Paukovcekova et al., 2020). Therefore, we implemented a similar strategy to examine the interactions between thiosemicarbazones and TKIs in all tested cell lines. According to this Simultaneous treatment design, the drugs were added to the assay at the same time, and the cells were incubated for 72 h (Figure 1, Supplementary Table S1).

imultaneous treatment was the most effective combination strategy in all cell lines tested in the present study. The effects of thiosemicarbazones and TKIs on DAOY cells were uniformly synergistic, with minimal differences versus the effects observed in the experiments performed according to the Pretreatment I design protocol (Figure 3E). Importantly, a significant difference between the two strategies was observed in SH-SY5Y (Figure 3A), SK-N-BE(2) (Figure 3B), and Saos-2 cells (Figure 3C), resulting in the uniformly synergistic and/or additive interactions in these cells in the experiments performed according to the Simultaneous treatment protocol. Furthermore, the synergy was achieved even after combining Dp44mT or DpC with LAP in SH-SY5Y cells; this drug combination produced antagonistic effects in the experiments performed according to the sequential treatment protocols (Figure 3A). The effects on RD cells were not as uniform as the effects on other cell lines; however, this Simultaneous approach produced a decrease in the CI values in most combinations (Figure 3D).

Since this treatment design was proven to represent the most effective approach, we further analyzed whether the synergistic interactions of the drugs are reflected by an increase in the induction of apoptosis. Due to higher sensitivity to the drugs and a substantial response to combined treatments, SH-SY5Y and SK-N-BE(2) neuroblastoma cells were selected as a model and were treated with thiosemicarbazones and TKIs alone or in combination using the concentrations corresponding to the IC₅₀ values. We detected a prominent increase in caspase-3 cleavage in the cells treated with Dp44mT and DpC alone and in combination with TKIs, notably SUN or GEF (Figure 4).

Overall, the results of the experiments performed according to the Simultaneous combined treatment protocol indicated synergistic inhibition of the proliferation of pediatric cancer cells (Figure 3), and induction of apoptosis in tested neuroblastoma cells (Figure 4). Therefore, this treatment design and SH-SY5Y and SK-N-BE(2) neuroblastoma cell lines were selected to further examine the molecular mechanisms of the interactions between thiosemicarbazones and selected TKIs.

Since thiosemicarbazones have been shown to affect various RTKs and the downstream signaling pathways in the cells derived from carcinomas (Dixon et al., 2013; Kovacevic et al., 2013; Lui et al., 2015a; Liu et al., 2015; Kovacevic et al., 2016; Park et al., 2020a; Lim et al., 2020), and in neuroblastoma cells (Guo et al., 2016; Paukovcekova et al., 2020; Macsek et al., 2022), we treated SK-N-BE(2) cells with DpC to identify the potential RTK targets. Human phosphoprotein arrays were used to determine the changes in the phosphorylation of RTKs in SK-N-BE(2) cells after 24-, 48-, and 72-hour incubation with DpC (Figure 5; Supplementary Figure S4).

Thiosemicarbazones have been reported to downregulate EGFR expression and phosphorylation in pancreatic and colon cancer cells (Kovacevic et al., 2016); however, we did not detect any alterations in EGFR activity after treatment with DpC. In contrast, the phosphorylation of IGF-1R and PDGFRβ, which were activated in untreated cells, was markedly decreased during the treatment, and was almost completely abrogated after 72-hour incubation with DpC (Figure 5). The relative phosphorylation levels of other tested RTKs of the panel are shown in Supplementary Figure S4.

Considering the data of the phosphoprotein array screening and previous studies, which have described the modulation of cell signaling by thiosemicarbazones (Dixon et al., 2013; Kovacevic et al., 2016; Paukovcekova et al., 2020; Geleta et al., 2021; Macsek et al., 2022), we further focused on the signaling pathways that could have been affected by both thiosemicarbazones and TKIs selected for the present study.

The next part of the present study examined the combined effects of thiosemicarbazones and TKIs on selected RTKs that were shown to be activated in both SH-SY5Y and SK-N-BE(2) cells. These effects included EGFR expression, major sites of EGFR phosphorylation (Tyr1068 and Tyr1148) relevant to EGFR activation, expression of PDGFRβ, Tyr751 phosphorylation of PDGFRβ needed for PI3K activation, and IGF-1R expression.

As expected, TKIs inhibited the phosphorylation of the corresponding RTKs in both neuroblastoma cell lines; however, we also detected significant off-target inhibitory activity in the case of other tested RTKs (Figure 6). In SH-SY5Y cells, SUN reduced EGFR activation at Tyr1148, and GEF and LAP alone targeted pPDGFRβ (Figure 6A). In SK-N-BE(2) cells, these effects were observed only after LAP treatment (Figure 6B). TKIs did not significantly alter the total EGFR or PDGFRβ levels; however, LAP and especially GEF downregulated IGF-1R in both cell lines (Figure 6).

In agreement with the data of phosphoprotein array analysis, EGFR phosphorylation was not significantly affected in DpC-treated SK-N-BE(2) cells (Figure 6B). The studies of other authors shown EGFR inhibition and/or downregulation by thiosemicarbazones (Liu et al., 2015; Kovacevic et al., 2016; Menezes et al., 2017; Macsek et al., 2022); however, the results of the present study demonstrated a Dp44mT- and DpC-induced decrease in total EGFR only in SH-SY5Y cells (Figure 6A). In contrast, Dp44mT and DpC consistently downregulated PDGFRβ and IGF-1R in both neuroblastoma cell lines (Figure 6).

In cells treated with the drug combinations, pEGFR (Tyr1068 and Tyr1148) was uniformly decreased in both cell lines, and this decrease was detected even in cells treated with the SUN + Dp44mT/DpC combinations (Figure 6). The combination of thiosemicarbazones and TKIs significantly downregulated EGFR expression in SH-SY5Y cells (Figure 6A); however, only a partial EGFR reduction was detected in LAP + Dp44mT/DpC-treated SK-N-BE(2) cells (Figure 6B).

Similarly, inhibition of pPDGFRβ (Tyr751) was detected across all drug combinations in both neuroblastoma cell lines (Figure 6) but was not detected in SK-N-BE(2) cells treated with GEF and thiosemicarbazones (Figure 6B). The levels of the PDGFRβ protein in the cells incubated with the combinations containing SUN remained comparable to the levels in the control cells, whereas the combination of GEF or LAP with thiosemicarbazones resulted in PDGFRβ downregulation in both cell lines (Figure 6). Furthermore, IGF-1R expression was significantly decreased in the cells treated with the latter combinations (Figure 6).

These data showed that the phosphorylation and/or expression of the key pro-oncogenic RTKs was activated in SH-SY5Y and SK-N-BE(2) neuroblastoma cells; specifically EGFR, PDGFRβ, and IGF-1R were inhibited by both thiosemicarbazones and TKIs. Importantly, these effects were more pronounced in the cells treated with the drug combinations, especially after simultaneous incubation with LAP and Dp44mT/DpC, suggesting that the observed synergistic effects were at least to some extent mediated by the blockade of the activity of these crucial RTKs.

Therefore, in the next step, we aimed to determine whether these changes impair the downstream signaling pathways. Since thiosemicarbazones alone have been shown to target PI3K/AKT and MAPK signaling in carcinoma cells (Dixon et al., 2013; Kovacevic et al., 2013, 2016; Lui et al., 2015a; Menezes et al., 2017; Macsek et al., 2022), we assessed the expression and phosphorylation of AKT (Ser473), ERK1/2 (Thr202/204), and MEK1/2 (Ser201/221) after combined treatments of neuroblastoma cells (Figure 7).

Unexpected activation of these downstream kinases was detected after the treatment of SH-SY5Y cells with thiosemicarbazones (Figure 7A). Similar activation of pAKT and pERK1/2 was observed in these cells after treatment with the combinations of thiosemicarbazones with GEF or LAP (Figure 7A). Thiosemicarbazones alone did not influence AKT activation in SK-N-BE(2) cells; however, an increase in pAKT was detected when thiosemicarbazones were combined with GEF or LAP (Figure 7B). Although activation of these kinases was unexpected, other authors have shown similar effects of Dp44mT in prostate cancer cells (Dixon et al., 2013), and we have previously demonstrated the activation of AKT and induction of a stress response in DpC-treated SH-SY5Y cells (Macsek et al., 2022).

In contrast, thiosemicarbazones alone inhibited pERK1/2 and pMEK1/2 in SK-N-BE(2) cells (Figure 7B). Furthermore, a decrease in pMEK1/2 was prominent after treatment of SK-N-BE(2) cells with all drug combinations (Figure 7B). In SH-SY5Y cells, this decrease was detected only after treatment with a combination of LAP and Dp44mT or DpC (Figure 7A).

In summary, combined treatments did not uniformly impair the kinases downstream of analyzed RTKs in the tested neuroblastoma cell lines. In fact, the activation of these kinases may be attributed to other upstream receptors, which were unaffected by the treatments and/or were triggered by a stress response of the cells.

NDRG1 has been described as a potent metastasis suppressor in a number of tumors, e.g., colon, prostate, and breast cancers (Bae et al., 2013; Park et al., 2020b). NDRG1 is often upregulated after stress-inducing stimuli (Bae et al., 2013), including cellular iron depletion (Chen et al., 2012; Lane et al., 2013; Lui et al., 2015a; Kovacevic et al., 2016). In fact, NDRG1 has been identified as a target of both Dp44mT and DpC in adult and pediatric tumors (Kovacevic et al., 2016; Menezes et al., 2017; Paukovcekova et al., 2020; Macsek et al., 2022) and implicated in the downregulation of the molecules involved in signal transduction, such as the ErbB family of RTKs and several downstream kinases (Liu et al., 2012; Dixon et al., 2013; Kovacevic et al., 2016; Menezes et al., 2017; Macsek et al., 2022). Our previous studies have shown a prominent upregulation of NDRG1 in the Dp44mT- and DpC-treated cancer cell lines derived from pediatric solid tumors, including neuroblastomas (Paukovcekova et al., 2020; Macsek et al., 2022); thus, we evaluated the NDRG1 levels after combined treatments of SH-SY5Y and SK-N-BE(2) cells (Figure 8). NDRG1 was detected as two closely migrating bands at 41 and 46 kDa, as reported previously (Park et al., 2018, 2020b; Paukovcekova et al., 2020; Macsek et al., 2022).

Similar to other cell lines, Dp44mT and DpC markedly elevated NDRG1 phosphorylation and expression in both SH-SY5Y and SK-N-BE(2) neuroblastoma cells (Figure 8). SUN treatment induced pNDRG1 inhibition in SH-SY5Y cells (Figure 8A); however, it considerably increased total NDRG1 expression in SK-N-BE(2) cells. Moreover, only the lower band at 41 kDa was preferentially upregulated (Figure 8B). In contrast, LAP treatment tended to increase pNDRG1 (Figure 8), although this change was significant only in SK-N-BE(2) cells (Figure 8B).

In both cell lines, the combinations of thiosemicarbazones with SUN reduced NDRG1 phosphorylation compared with Dp44mT- and DpC-treated controls (Figure 8). However, these combined treatments induced the upregulation of total NDRG1 in SK-N-BE(2) cells, and the effect was specifically directed at the 41 kDa band (Figure 8B). In contrast, the combinations of thiosemicarbazones with LAP or GEF markedly elevated pNDRG1 in both cell lines or in SK-N-BE(2) cells, respectively (Figure 8). In SK-N-BE(2) cells, the synergy between thiosemicarbazones and both LAP and GEF was manifested as an apparent upregulation of total NDRG1 expression compared with the effects of these agents alone (Figure 8B).

These data demonstrated that the combinations of thiosemicarbazones, particularly with GEF or LAP, enhanced the efficacy of these agents in upregulating and activating the metastasis suppressor NDRG1 in neuroblastoma cells. These data were in agreement with our initial assessment that revealed strong synergistic interactions of these drugs not only in both neuroblastoma cell lines but also in the tested pediatric solid tumor cell lines in general (Figure 3).