Testing of the Survivin Suppressant YM155 in a Large Panel of Drug-Resistant Neuroblastoma Cell Lines

The survivin suppressant YM155 is a drug candidate for neuroblastoma. Here, we tested YM155 in 101 neuroblastoma cell lines (19 parental cell lines, 82 drug-adapted sublines). Seventy seven (77) cell lines displayed YM155 IC50s in the range of clinical YM155 concentrations. ABCB1 was an important determinant of YM155 resistance. The activity of the ABCB1 inhibitor zosuquidar ranged from being similar to that of the structurally different ABCB1 inhibitor verapamil to being 65-fold higher. ABCB1 sequence variations may be responsible for this, suggesting that the design of variant-specific ABCB1 inhibitors may be possible. Further, we showed that ABCC1 confers YM155 resistance. Previously, p53 depletion had resulted in decreased YM155 sensitivity. However, TP53-mutant cells were not generally less sensitive to YM155 than TP53 wild-type cells in this study. Finally, YM155 cross-resistance profiles differed between cells adapted to drugs as similar as cisplatin and carboplatin. In conclusion, the large cell line panel was necessary to reveal an unanticipated complexity of the YM155 response in neuroblastoma cell lines with acquired drug resistance. Novel findings include that ABCC1 mediates YM155 resistance and that YM155 cross-resistance profiles differ between cell lines adapted to drugs as similar as cisplatin and carboplatin.


Introduction
The inhibitor of apoptosis protein (IAP) survivin has multifaceted roles in cellular signalling. It is absent from most somatic cells but highly abundant in cancer cells and mediates cancer cell survival and proliferation. Elevated survivin levels have been associated with more aggressive and advanced cancer disease and lower survival rates. Hence, survivin is a potential drug target in cancer entities including neuroblastoma [1][2][3][4][5][6][7][8][9][10], the most frequent solid extracranial paediatric cancer. About half of the patients are diagnosed with high-risk disease associated with overall survival rates below 50% despite myeloablative therapy and differentiation therapy using retinoids. While many neuroblastomas respond initially well to therapy, acquired drug resistance represents a major clinical problem [11,12]. YM155 (sepantronium bromide) was introduced as a suppressor of survivin expression that displayed anti-cancer activity in pre-clinical models of different cancer entities including neuroblastoma [6][7][8]10,13]. However, DNA damage induction and Mcl-1 depletion were later suggested as additional or alternative anti-cancer mechanisms of YM155 [8,[14][15][16][17][18]. We recently confirmed that YM155 exerts its anti-neuroblastoma effects predominantly through survivin suppression [10]. YM155-induced survivin suppression proceeded DNA damage formation. Moreover, YM155 mimicked the effects of RNAi-mediated survivin depletion, whereas Mcl-1 depletion did not affect neuroblastoma cell viability [10]. Furthermore, YM155-adapted UKF-NB-3 neuroblastoma cells had developed resistance to RNAi-mediated survivin depletion [10]. We also investigated the effects of YM155 in neuroblastoma cells with acquired resistance to cisplatin, doxorubicin, or vincristine and detected in concert with previous findings reduced SLC35F2 (mediates cellular YM155 uptake) and increased ABCB1 (causes YM155 efflux) expression as drug-specific resistance mechanisms [6,10,19,20]. Our data further demonstrated that RNAi-mediated p53 depletion mediated resistance to YM155 and survivin depletion suggesting loss of p53 function to be a target-specific resistance mechanism that will affect all approaches that target survivin in neuroblastoma [10].
Here, we investigated the effects of YM155 in a larger panel of 101 neuroblastoma cell lines that focused on acquired drug resistance containing 82 drug-adapted neuroblastoma cell lines, which reflected resistance to 15 anti-cancer drugs.

Effects of YM155 on the Viability of Parental Neuroblastoma Cell Lines
YM155 was initially tested in a panel of 86 neuroblastoma cell lines consisting of 17 parental neuroblastoma cell lines and 69 drug-adapted sub-lines (Table S1). The IC 50 values in the parental cell lines ranged from 0.49 nM (UKF-NB-3) to 248 nM (LAN-6) ( Figure 1, Table S1). All parental cell lines despite from LAN-6, NB-S-124 (77 nM), and SK-N-SH (75 nM) had IC 50 values in the range of clinical achievable YM155 plasma concentrations ( Figure 1A, Table S1) that were reported to reach up to 56 nM [21][22][23].
Cancers 2020 2 of 18 neuroblastomas respond initially well to therapy, acquired drug resistance represents a major clinical problem [11,12]. YM155 (sepantronium bromide) was introduced as a suppressor of survivin expression that displayed anti-cancer activity in pre-clinical models of different cancer entities including neuroblastoma [6][7][8]10,13]. However, DNA damage induction and Mcl-1 depletion were later suggested as additional or alternative anti-cancer mechanisms of YM155 [8,[14][15][16][17][18]. We recently confirmed that YM155 exerts its anti-neuroblastoma effects predominantly through survivin suppression [10]. YM155-induced survivin suppression proceeded DNA damage formation. Moreover, YM155 mimicked the effects of RNAi-mediated survivin depletion, whereas Mcl-1 depletion did not affect neuroblastoma cell viability [10]. Furthermore, YM155-adapted UKF-NB-3 neuroblastoma cells had developed resistance to RNAi-mediated survivin depletion [10]. We also investigated the effects of YM155 in neuroblastoma cells with acquired resistance to cisplatin, doxorubicin, or vincristine and detected in concert with previous findings reduced SLC35F2 (mediates cellular YM155 uptake) and increased ABCB1 (causes YM155 efflux) expression as drugspecific resistance mechanisms [6,10,19,20]. Our data further demonstrated that RNAi-mediated p53 depletion mediated resistance to YM155 and survivin depletion suggesting loss of p53 function to be a target-specific resistance mechanism that will affect all approaches that target survivin in neuroblastoma [10].
Here, we investigated the effects of YM155 in a larger panel of 101 neuroblastoma cell lines that focused on acquired drug resistance containing 82 drug-adapted neuroblastoma cell lines, which reflected resistance to 15 anti-cancer drugs.

Effects of YM155 on the Viability of Parental Neuroblastoma Cell Lines
YM155 was initially tested in a panel of 86 neuroblastoma cell lines consisting of 17 parental neuroblastoma cell lines and 69 drug-adapted sub-lines (Table S1). The IC50 values in the parental cell lines ranged from 0.49 nM (UKF-NB-3) to 248 nM (LAN-6) ( Figure 1, Table S1). All parental cell lines despite from LAN-6, NB-S-124 (77 nM), and SK-N-SH (75 nM) had IC50 values in the range of clinical achievable YM155 plasma concentrations ( Figure 1A, Table S1) that were reported to reach up to 56 nM [21][22][23].   Table  S1. Information on the ABCB1 status of the cell lines is provided in  Table S6.

ABCB1-Expressing Neuroblastoma Cell Lines Display Low Sensitivity to YM155
In concert with previous studies [6,10], high ABCB1-expressing neuroblastoma cells generally displayed relatively low YM155 sensitivity ( Figure 1A). In a subset of these cell lines, only the high ABCB1-expressing cells were (in contrast to low ABCB1-expressing cells) sensitised to YM155 by verapamil and zosuquidar ( Figure 1B), two structurally unrelated ABCB1 inhibitors [24].

MYCN Status Does Not Influence Neuroblastoma Cell Sensitivity to YM155
MYCN amplification is a major determinant of poor disease outcome in neuroblastoma [11,12]. The YM155 IC50 ranged from 0.49 nM (UKF-NB-3) to 77 nM (NB-S-124) in MYCN-amplified cells and from 3.55 nM (SK-N-AS) to 248 nM (LAN-6) in not MYCN-amplified cells (Table S4). To exclude the effects of ABCB1, we compared the YM155 IC50 in cells with known MYCN status in the presence of the ABCB1 inhibitors verapamil and zosuquidar. In the presence of ABCB1 inhibitors, MYCNamplified and non-MYCN-amplified displayed a similar range of YM155 IC50s (Figure 2A, Table S3). SH-EP-MYCN (TET21N) cells express a tetracycline-controllable MYCN transgene. They display low MYCN levels in the presence of tetracycline antibiotics and high MYCN levels in the absence of tetracycline antibiotics [25]. SH-EP-MYCN (TET21N) cells displayed similar YM155 IC50 values in the absence or presence of doxycycline ( Figure 2B, Figure S1).   Table S1. Information on the ABCB1 status of the cell lines is provided in Table S2. (B) Effects of the ABCB1 inhibitors verapamil (5 µM) and zosuquidar (1.25 µM) on the YM155 IC 50 values in neuroblastoma cell lines characterised by high or low ABCB1 levels displayed as fold change YM155 IC 50 / YM155 IC 50 in the presence of ABCB1 inhibitor. Numerical data and the effects of the ABCB1 inhibitors alone on cell viability are presented in Table S6.

ABCB1-Expressing Neuroblastoma Cell Lines Display Low Sensitivity to YM155
In concert with previous studies [6,10], high ABCB1-expressing neuroblastoma cells generally displayed relatively low YM155 sensitivity ( Figure 1A). In a subset of these cell lines, only the high ABCB1-expressing cells were (in contrast to low ABCB1-expressing cells) sensitised to YM155 by verapamil and zosuquidar ( Figure 1B), two structurally unrelated ABCB1 inhibitors [24].

MYCN Status Does Not Influence Neuroblastoma Cell Sensitivity to YM155
MYCN amplification is a major determinant of poor disease outcome in neuroblastoma [11,12]. The YM155 IC 50 ranged from 0.49 nM (UKF-NB-3) to 77 nM (NB-S-124) in MYCN-amplified cells and from 3.55 nM (SK-N-AS) to 248 nM (LAN-6) in not MYCN-amplified cells (Table S4). To exclude the effects of ABCB1, we compared the YM155 IC 50 in cells with known MYCN status in the presence of the ABCB1 inhibitors verapamil and zosuquidar. In the presence of ABCB1 inhibitors, MYCN-amplified and non-MYCN-amplified displayed a similar range of YM155 IC 50 s ( Figure 2A, Table S3). SH-EP-MYCN (TET21N) cells express a tetracycline-controllable MYCN transgene. They display low MYCN levels in the presence of tetracycline antibiotics and high MYCN levels in the absence of tetracycline antibiotics [25]. SH-EP-MYCN (TET21N) cells displayed similar YM155 IC 50 values in the absence or presence of doxycycline ( Figure 2B, Figure S1).

TP53 Status Does Not Predict Neuroblastoma Cell Sensitivity to YM155
Previously, RNAi-mediated p53 depletion was shown to reduce the YM155 sensitivity of the neuroblastoma cell lines UKF-NB-3 and UKF-NB-6 [10]. However, the p53-null SK-N-AS cells displayed an YM155 IC50 of 3.55 nM that was further reduced to 1.01 nM and 1.31 nM by verapamil and zosuquidar, respectively ( Figure 1, Table S3). Hence, SK-N-AS belongs in the presence of ABCB1 inhibitors to the most YM155-sensitive neuroblastoma cell lines in the panel, despite the lack of functional p53.
To further investigate the relevance of the TP53 status for the neuroblastoma cell sensitivity to YM155, we determined YM155 IC50 values in a panel of 14 nutlin-3-adapted TP53-mutant

TP53 Status Does Not Predict Neuroblastoma Cell Sensitivity to YM155
Previously, RNAi-mediated p53 depletion was shown to reduce the YM155 sensitivity of the neuroblastoma cell lines UKF-NB-3 and UKF-NB-6 [10]. However, the p53-null SK-N-AS cells displayed an YM155 IC 50 of 3.55 nM that was further reduced to 1.01 nM and 1.31 nM by verapamil and zosuquidar, respectively ( Figure 1, Table S3). Hence, SK-N-AS belongs in the presence of ABCB1 inhibitors to the most YM155-sensitive neuroblastoma cell lines in the panel, despite the lack of functional p53.
To further investigate the relevance of the TP53 status for the neuroblastoma cell sensitivity to YM155, we determined YM155 IC 50  cell lines [26,27]. Our initial cell line panel included one nutlin-3-adapted neuroblastoma cell line (UKF-NB-3 r Nutlin 10µM ) that harbours a G245C loss-of-function TP53 mutation [26] and displayed 2.4-fold reduced YM155 sensitivity relative to the parental UKF-NB-3 cells (Table S1). In addition, we tested YM155 in nutlin-3-resistant, TP53 mutant sub-lines of two clonal p53 wild-type UKF-NB-3 sub-lines (UKF-NB-3clone1, UKF-NB-3clone3) and the TP53 wild-type neuroblastoma cell line UKF-NB-6 ( Figure 3, Table S5). Only four out of the 14 nutlin-3-resistant neuroblastoma cell lines displayed a > 2-fold change in the YM155 IC 50 relative to the respective parental cells, with 3.3 (UKF-NB-3clone1 r Nutlin 10µM I) being the highest fold change ( Figure 3, Table S4). These findings do not suggest the cellular TP53 status to be a good predictor of neuroblastoma cell sensitivity to YM155.

Figure 3.
Effects of YM155 on the viability of parental p53 wild-type neuroblastoma cell lines and their p53 mutant nutlin-3-adapted sub-lines. YM155 concentrations that reduce neuroblastoma cell viability by 50% (IC50, mean ± S.D., n = 3) as determined by MTT assay after a 5-day treatment period. Numerical data are presented in Table S4.

Effects of YM155 on the Viability of Neuroblastoma Cell Lines with Acquired Drug Resistance
In a panel of 69 sub-lines of the neuroblastoma cell lines IMR-5, IMR-32, NGP, NLF, SHEP, UKF-NB-2, UKF-NB-3, and UKF-NB-6 with acquired resistance to drug classes including platinum drugs, vinca alkaloids, taxanes, alkylating agents, topoisomerase I inhibitors, topoisomerase II inhibitors, and nucleoside analogues (Table S1), resistance was commonly associated with decreased YM155 sensitivity. However, 48 resistant cell lines displayed YM155 IC50 values in the range of therapeutic plasma levels (up to 56 nM) (Table S1).
The YM155 IC50 values in the drug-resistant cell lines ranged from 0.40 nM (UKF-NB-3 r GEMCI 10 ) to 21,549 nM (IMR-5 r DOCE 20 ) (Table S1). Drug class-specific differences in the YM155 resistance profiles can be observed, but the variation of the results is very large (Figure 4, Table S5).   Table S4.

Effects of YM155 on the Viability of Neuroblastoma Cell Lines with Acquired Drug Resistance
In a panel of 69 sub-lines of the neuroblastoma cell lines IMR-5, IMR-32, NGP, NLF, SHEP, UKF-NB-2, UKF-NB-3, and UKF-NB-6 with acquired resistance to drug classes including platinum drugs, vinca alkaloids, taxanes, alkylating agents, topoisomerase I inhibitors, topoisomerase II inhibitors, and nucleoside analogues (Table S1), resistance was commonly associated with decreased YM155 sensitivity. However, 48 resistant cell lines displayed YM155 IC 50 values in the range of therapeutic plasma levels (up to 56 nM) (Table S1).
For the topoisomerase II inhibitor (doxorubicin, etoposide)-and platinum drug (carboplatin, cisplatin, oxaliplatin)-adapted cell lines, we had sufficient data to perform drug-specific analyses. For six cell lines, we had doxorubicin-and etoposide-resistant sub-lines ( Figure 5A, Figure 5B, Table S1).
To determine the mean YM155 IC50 value, we again removed the UKF-NB-3 sub-lines because of the high value of UKF-NB-3 r DOX 20 . Results revealed that acquired doxorubicin resistance resulted in a generally more pronounced YM155 resistance phenotype than acquired etoposide resistance ( Figure  5A, Figure 5B, Table S1).
For the topoisomerase II inhibitor (doxorubicin, etoposide)-and platinum drug (carboplatin, cisplatin, oxaliplatin)-adapted cell lines, we had sufficient data to perform drug-specific analyses. For six cell lines, we had doxorubicin-and etoposide-resistant sub-lines ( Figure 5A,B, Table S1).
To determine the mean YM155 IC 50 value, we again removed the UKF-NB-3 sub-lines because of the high value of UKF-NB-3 r DOX 20 . Results revealed that acquired doxorubicin resistance resulted in a generally more pronounced YM155 resistance phenotype than acquired etoposide resistance ( Figure 5A,B, Table S1).  Table S1 and

Role of ABCB1 in the YM155 response of Drug-Adapted Neuroblastoma Cells
A significant amount of drug-adapted neuroblastoma cell lines displays increased ABCB1 activity [28], and ABCB1 has been previously shown to mediate YM155 resistance [6,10,19]. In  Table S1). Cisplatin and oxaliplatin resistance were associated with a lower degree of YM155 resistance than carboplatin resistance ( Figure 5C,D, Table S1).

Role of ABCB1 in the YM155 response of Drug-Adapted Neuroblastoma Cells
A significant amount of drug-adapted neuroblastoma cell lines displays increased ABCB1 activity [28], and ABCB1 has been previously shown to mediate YM155 resistance [6,10,19]. In agreement with previous results, transduction of neuroblastoma cells with ABCB1 resulted in YM155 resistance, which was reduced by siRNA-mediated ABCB1 depletion ( Figure 6, Figure S2). Further, YM155 100 nM, a concentration that did not induce survivin depletion in ABCB1-transduced UKF-NB-3 (UKF-NB-3 ABCB1 ) cells after 24h of incubation, reduced cellular survivin levels in UKF-NB-3 ABCB1 cells in the presence of the ABCB1 inhibitors verapamil and zosuquidar ( Figure 6, Figure S2). agreement with previous results, transduction of neuroblastoma cells with ABCB1 resulted in YM155 resistance, which was reduced by siRNA-mediated ABCB1 depletion ( Figure 6, Figure S2). Further, YM155 100 nM, a concentration that did not induce survivin depletion in ABCB1-transduced UKF-NB-3 (UKF-NB-3 ABCB1 ) cells after 24h of incubation, reduced cellular survivin levels in UKF-NB-3 ABCB1 cells in the presence of the ABCB1 inhibitors verapamil and zosuquidar ( Figure 6, Figure S2).  Figure S2.
Hence, we further examined the effects of verapamil and zosuquidar on YM155 sensitivity in a panel of 60 drug-adapted neuroblastoma cell lines (Table S6). Fifteen of the drug-adapted cell lines displayed an IC50 higher than 56 nM. In the presence of verapamil, only six of the drug-adapted cell  Figure S2.
Hence, we further examined the effects of verapamil and zosuquidar on YM155 sensitivity in a panel of 60 drug-adapted neuroblastoma cell lines (Table S6). Fifteen of the drug-adapted cell lines displayed an IC 50 higher than 56 nM. In the presence of verapamil, only six of the drug-adapted cell lines displayed an IC 50 higher 56 nM. In the presence of zosuquidar, only five of the drug-adapted cell lines displayed an IC 50 higher than 56 nM. The YM155 IC 50 values of eight cell lines (IMR-5 r DOX 20 , IMR-5 r VCR 10 , IMR r VINB 20 , NGP r DOX 20 , NGP r ETO 400 , UKF-NB-3 r CARBO 2000 , UKF-NB-3 r CDDP 1000 , UKF-NB-3 r VINOR 40 ) were reduced to levels below 56 nM by verapamil and zosuquidar. Notably, the fold sensitisation by ABCB1 inhibitors was low (verapamil: 1.2-fold, zosuquidar: 1.5-fold) in NGP r ETO 400 , although the YM155 IC 50 s were reduced below 56 nM. This suggests that ABCB1 expression is not a dominant YM155 resistance mechanism in NGP r ETO 400 cells (Table S6).
IMR-5 r DOCE 20 , NGP r VCR 20 , and UKF-NB-2 r CARBO 2000 were sensitised by verapamil and zosuquidar to YM155. However, the effects of zosuquidar were more pronounced resulting in YM155 IC 50 values below 56 nM, whereas the YM155 IC 50 values remained above 56 nM in the presence of verapamil (Table S6). In NLF r DOX 40 and NLF r VINB 10 cells, zosuquidar (but not verapamil) increased the YM155 IC 50 values by mechanisms that appear to be unrelated to ABCB1 (Table S6).
We selected two parental cell line/ drug-adapted subline pairs (IMR-5/IMR-5 r DOCE 20 , IMR-32/ IMR-32 r DOX 20 ) for additional confirmatory experiments. The YM155 IC 50 s in these four cell lines in the absence and presence of zosuquidar determined MTT (used in the screen, measures oxidative phosphorylation in the mitochondria) were very similar to those determined by CellTiterGlo (alternative viability assay that measures ATP production) (Table S7). Both drug-adapted sublines displayed increased ABCB1 levels ( Figure S3). Moreover, YM155 500 nM, a concentration that did not affect cellular survivin levels and PARP cleavage in IMR-5 r DOCE 20 cells after 24h incubation, caused survivin depletion and PARP cleavage in IMR-5 r DOCE 20 cells in the presence of the ABCB1 inhibitors verapamil and zosuquidar (Figure 7, Figure S3).
Cancers 2020 9 of 18 UKF-NB-3 r VINOR 40 ) were reduced to levels below 56 nM by verapamil and zosuquidar. Notably, the fold sensitisation by ABCB1 inhibitors was low (verapamil: 1.2-fold, zosuquidar: 1.5-fold) in NGP r ETO 400 , although the YM155 IC50s were reduced below 56 nM. This suggests that ABCB1 expression is not a dominant YM155 resistance mechanism in NGP r ETO 400 cells (Table S6). IMR-5 r DOCE 20 , NGP r VCR 20 , and UKF-NB-2 r CARBO 2000 were sensitised by verapamil and zosuquidar to YM155. However, the effects of zosuquidar were more pronounced resulting in YM155 IC50 values below 56 nM, whereas the YM155 IC50 values remained above 56 nM in the presence of verapamil (Table S6). In NLF r DOX 40 and NLF r VINB 10 cells, zosuquidar (but not verapamil) increased the YM155 IC50 values by mechanisms that appear to be unrelated to ABCB1 (Table S6).
We selected two parental cell line/ drug-adapted subline pairs (IMR-5/IMR-5 r DOCE 20 , IMR-32/ IMR-32 r DOX 20 ) for additional confirmatory experiments. The YM155 IC50s in these four cell lines in the absence and presence of zosuquidar determined MTT (used in the screen, measures oxidative phosphorylation in the mitochondria) were very similar to those determined by CellTiterGlo (alternative viability assay that measures ATP production) (Table S7). Both drug-adapted sublines displayed increased ABCB1 levels ( Figure S3). Moreover, YM155 500 nM, a concentration that did not affect cellular survivin levels and PARP cleavage in IMR-5 r DOCE 20 cells after 24h incubation, caused survivin depletion and PARP cleavage in IMR-5 r DOCE 20 cells in the presence of the ABCB1 inhibitors verapamil and zosuquidar (Figure 7, Figure S3).

ABCC1 Mediates Resistance to YM155
NLF r VCR 10 cells were sensitised by verapamil but not by zosuquidar to YM155 (Table S6). NLF r VCR 10 cells are characterised by ABCC1 (also known as MRP1) expression but do not express ABCB1 [29]. Since only verapamil but not zosuquidar inhibits ABCC1 [30,31], this suggests that

ABCC1 Mediates Resistance to YM155
NLF r VCR 10 cells were sensitised by verapamil but not by zosuquidar to YM155 (Table S6). NLF r VCR 10 cells are characterised by ABCC1 (also known as MRP1) expression but do not express ABCB1 [29]. Since only verapamil but not zosuquidar inhibits ABCC1 [30,31], this suggests that ABCC1 also mediates resistance to YM155. In agreement, the ABCC1 inhibitor MK571 substantially reduced YM155 sensitivity in ABCC1-expressing NLF r VCR 10 cells but not in the parental NLF cell line that does not express ABCC1 (Table S8).

Cross Resistance to YM155 is Caused by Multiple Resistance Mechanisms in Drug-Adapted Neuroblastoma Cells
Although our data indicate that ABCB1 plays an important role in the cross-resistance of drug-adapted neuroblastoma cell lines to YM155, additional mechanisms are also involved. Of the 60 drug-adapted cell lines, 35 displayed cross-resistance to YM155 (fold change YM155 IC 50 resistant sub-line/ YM155 IC 50 respective parental cell line > 2). In 24 of these drug-adapted cell lines, the YM155 IC 50 remained > 2-fold higher in the presence of verapamil compared to the YM155 IC 50 of the respective parental cell line in the presence of verapamil (Table S6). Similarly, in 19 drug-adapted cell lines, the YM155 IC 50 remained > 2-fold higher in the presence of zosuquidar compared to the YM155 IC 50 of the respective parental cell line in the presence of zosuquidar (Table S6). This included cell lines that were not sensitised by verapamil and/ or zosuquidar to YM155 and those that were sensitised to YM155 by verapamil and/ or zosuquidar but not to the level of the parental cells (Table S6).

Discrepancies in the Effects of the ABCB1 Inhibitors Verapamil and Zosuquidar on Neuroblastoma Cell Sensitivity to YM155
Both, verapamil and zosuquidar, sensitised 25 of the drug-adapted cell lines to YM155 by >2-fold. There was an overlap of 23 cell lines that were sensitised to YM155 by both ABCB1 inhibitors by > 2-fold (Table S6). Among the exceptions was the ABCC1-expressing cell line NLF r VCR 10 that has already been described above. NGP r GEMCI 20 cells were sensitised to YM155 by zosuquidar, whereas verapamil slightly increased the YM155 IC 50 value in this cell line. In addition, UKF-NB-3 r TOPO 20 cells were sensitised to YM155 by verapamil by > 2-fold (fold change 3.7) but not by zosuquidar (fold change 1.4). UKF-NB-6 r CARBO 2000 cells were sensitised by zosuquidar to YM155 by > 2-fold (fold change 2.3) but not by verapamil (fold change 1.3) (Table S6). The reasons for these differences remain unclear.
Since we had detected differences between the effects of the ABCB1 inhibitors verapamil and zosuquidar on the YM155 IC 50 values in a number of cell lines, we compared both drugs in a wider panel of 74 cell lines ( Figure 8, Table S6). In 45 cell lines, the YM155 IC 50 values were in a similar range in the presence of verapamil and zosuquidar (YM155 IC 50 in the presence of verapamil/ YM155 IC 50 in the presence of zosuquidar > 0.5 and < 2.0). This included 33 cell lines that were neither sensitised to YM155 by verapamil nor by zosuquidar and 12 cell lines that were sensitised by verapamil and zosuquidar by more than 2-fold and in a similar fashion (Table S6).
Thirteen cell lines were more strongly sensitised to YM155 by verapamil than by zosuquidar (YM155 IC 50 in the presence of verapamil/YM155 IC 50 in the presence of zosuquidar < 0.5). Only one of these cell lines (UKF-NB-6 r ETO 200 ) was sensitised by both verapamil and zosuquidar to YM155 by more than 2-fold. The remaining cell lines were either only sensitised to YM155 by more than 2-fold by verapamil and not by zosuquidar, or zosuquidar increased the YM155 IC 50 by mechanisms that do not appear to be associated with effects on ABCB1 (Table S6).  Table S6. 1 UKF-NB-3 r DOX 20 , fold change 9235; 2 IMR-5 r DOCE 20 , fold change 1581; 3 UKF-NB-3 r DOCE 10 , fold change 939.

Discussion
YM155 has been suggested as therapeutic option for the treatment of neuroblastoma including therapy-refractory disease [6,7,10]. In a larger cell line panel, mainly consisting of neuroblastoma cell lines with acquired drug resistance, we here show that 77 out of 101 tested neuroblastoma cell lines displayed YM155 IC50 values in the range of clinically achievable plasma concentrations up to 56 nM [21][22][23].
Although MYCN amplification is a major indicator of poor prognosis in neuroblastoma [11,12], the efficacy of YM155 was independent of the MYCN status. As previously shown [6,10,19], ABCB1expressing cells displayed low YM155 sensitivity and were sensitised by the ABCB1 inhibitors verapamil and zosuquidar to YM155. The role of ABCB1 expression in neuroblastoma is not clear.   Sixteen cell lines were more strongly sensitised to YM155 by zosuquidar than by verapamil (YM155 IC 50 in the presence of verapamil/ YM155 IC 50 in the presence of zosuquidar > 2.0). NGP r GEMCI 20 cells were only sensitised by zosuquidar to YM155 but not by verapamil. The other 15 of these cell lines were sensitised by more than 2-fold to YM155 by both compounds with zosuquidar exerting more pronounced effects. The relative differences between these two drugs on the YM155 sensitivity of these cell lines ranged from 2.2 (NLF r DOCE 20 , UKF-NB-3 r CARBO 2000 ) to 65.2 (UKF-NB-3 r DOCE 10 ) (Table S6).

Discussion
YM155 has been suggested as therapeutic option for the treatment of neuroblastoma including therapy-refractory disease [6,7,10]. In a larger cell line panel, mainly consisting of neuroblastoma cell lines with acquired drug resistance, we here show that 77 out of 101 tested neuroblastoma cell lines displayed YM155 IC 50 values in the range of clinically achievable plasma concentrations up to 56 nM [21][22][23].
Although MYCN amplification is a major indicator of poor prognosis in neuroblastoma [11,12], the efficacy of YM155 was independent of the MYCN status. As previously shown [6,10,19], ABCB1-expressing cells displayed low YM155 sensitivity and were sensitised by the ABCB1 inhibitors verapamil and zosuquidar to YM155. The role of ABCB1 expression in neuroblastoma is not clear. ABCB1 expression at diagnosis is commonly regarded not to be of prognostic relevance [32], although ABCB1 expression has been reported in a significant fraction of patients [6,33,34]. Notably, six (35%) out of the 17 parental neuroblastoma cell lines investigated in this study (Be(2)C, LAN-6, NB-S-124, SHEP, SK-N-AS, SK-N-SH) are characterised by ABCB1 activity, which is similar to a previous study, although there was some overlap between the cell line panels [6]. Limited information is available on ABCB1 expression as acquired drug resistance mechanism in neuroblastoma. Many drug-adapted neuroblastoma cell lines display enhanced ABCB1 activity [27,35], and drug-adapted cell lines have been shown to reflect clinically relevant resistance mechanisms [26,[36][37][38][39][40][41][42][43][44][45]. Some clinical hints may also point towards a role of ABCB1 in acquired resistance in neuroblastoma [36,46]. Hence, ABCB1 may represent an acquired resistance mechanism in neuroblastoma.
Verapamil and zosuquidar further differed in their effects on neuroblastoma cell sensitivity to YM155. Zosuquidar protected some ABCB1-negative cell lines from YM155-induced toxicity by unknown mechanisms. More strikingly, we identified cell line-specific differences in the relative potencies of verapamil and zosuquidar on YM155 activity. Twelve ABCB1-expressing cell lines were similarly sensitised to YM155 by verapamil and zosuquidar. Few cell lines were stronger sensitised by verapamil than by zosuquidar. These data are difficult to interpret with regard to ABCB1 because verapamil interacts with a broader range of transporters than the specific ABCB1 inhibitor zosuquidar [30,31,47]. However, 15 cell lines were sensitised by more than 2-fold to YM155 by verapamil and zosuquidar with zosuquidar exerting up to 65-fold (UKF-NB-3 r DOCE 10 ) more pronounced effects than verapamil. Hence, the relative effects of verapamil and zosuquidar on ABCB1-mediated YM155 transport differ in individual cell lines from similar efficacy (in 12 cell lines) to 65-fold increased zosuquidar efficacy over verapamil. Since zosuquidar is regarded as highly specific ABCB1 inhibitor [31,47], this difference seems to depend on discrepancies in the interaction with ABCB1. Notably, ABCB1 polymorphisms and mutations may substantially alter ABCB1 substrate specificity [48,49], which may explain the cell line-specific variation in the relative influence of zosuquidar and verapamil on ABCB1 function. Hence, it might be possible to design ABCB1 inhibitors that preferentially target specific ABCB1 variants. In this context, we have previously shown that certain ABCB1 inhibitors preferentially interfere with the ABCB1-mediated transport of certain ABCB1 substrates [50].
Although RNAi-mediated p53 depletion decreases neuroblastoma cell sensitivity to YM155 as previously demonstrated [10], only 4 out of 14 nutlin-3-adapted TP53-mutant neuroblastoma cell lines displayed a >2-fold increased YM155 IC 50 relative to the respective parental cell line. All nutlin-3-resistant cell lines remained sensitive to low nanomolar YM155 concentrations with IC 50 values ranging from 0.40 to 1.50 nM. This shows that the role of p53 depends on the individual cellular context and that the TP53 status on its own does not indicate neuroblastoma cell sensitivity to YM155.
These data are in accordance with initial findings that reported the activity of YM155 to be unrelated to the TP53 status [13,51].
The adaptation of neuroblastoma cell lines to drugs from different classes was associated with varying YM155 sensitivity profiles. Neuroblastoma cell adaptation to topoisomerase I inhibitors, the nucleoside analogue gemcitabine, or alkylating agents was not associated with a pronounced YM155 resistance phenotype. In contrast, neuroblastoma cell lines adapted to the taxane docetaxel or vinca alkaloids exhibited distinct cross-resistance to YM155. Topoisomerase II inhibitors and platinum drugs displayed intermediate potential to induce YM155 resistance.
For the topoisomerase II inhibitor-and platinum drug-adapted neuroblastoma cell lines, we could perform drug-specific sub-analyses. Among the topoisomerase II inhibitors, doxorubicin-resistant cells were more likely to be YM155-resistant than etoposide-resistant cells. This result may not be too surprising. Although doxorubicin and etoposide both inhibit the religation of the topoisomerase II cleavage complexes, they are structurally different compounds that differ in their exact interaction with this target and may exert additional varying effects [52][53][54]. Perhaps more strikingly, we also found differences among the carboplatin-, cisplatin-, and oxaliplatin-resistant cell lines. Carboplatin-resistant cells were more frequently YM155-resistant than cisplatin-or oxaliplatin-resistant cells. This is surprising because platinum drugs share a similar mechanism of action. In particular, the effects of carboplatin and cisplatin are thought to be much more related to each other than to oxaliplatin [55,56]. The underlying reasons remain unclear, but the findings indicate substantial gaps in our understanding of the action of these frequently used drugs that need to be filled.
Resistant neuroblastoma cell lines were established by continuous exposure to increasing drug concentrations as previously described [26] and derived from the Resistant Cancer Cell Line (RCCL) collection (www.kent.ac.uk/stms/cmp/RCCL/RCCLabout.html).
All cells were propagated in Iscove's modified Dulbecco's medium (IMDM) supplemented with 10% foetal calf serum (FCS), 100IU/ml penicillin and 100µg/ml streptomycin at 37 • C. Drug-adapted cell lines were continuously cultivated in the presence of the respective adaptation drugs but were released from the respective adaptation drugs before they were used for experiments. Cells were routinely tested for mycoplasma contamination and authenticated by short tandem repeat profiling.

RNA Interference (RNAi)
Transient depletion of ABCB1 was achieved using synthetic siRNA oligonucleotides (ON-TARGETplus SMARTpool) from Dharmacon (Lafayette, CO; USA). Non-targeting siRNA (ON-TARGETplus SMARTpool) was used as negative control. Cells were transfected by electroporation using the NEON Transfection System (Invitrogen, Darmstadt; Germany) according to the manufacturer protocol. Cells were grown to 60-80 % confluence, trypsinised, and 1.2 x 10 6 cells were re-suspended in 200 µl resuspension buffer R including 2.5 µM siRNA. The electroporation was performed using two 20 millisecond pulses of 1400 V. Subsequently, the cells were transferred into cell culture plates or flasks, containing pre-warmed cell culture medium.

Statistics
Results are expressed as mean ± S.D. of at least three experiments. Comparisons between two groups were performed using Student's t-test. Three or more groups were compared by ANOVA followed by the Student-Newman-Keuls test. P values lower than 0.05 were considered to be significant.

Conclusions
The investigation of YM155 in 101 neuroblastoma cell lines revealed complex sensitivity profiles and that larger panels of model systems are needed to uncover this complexity. Our findings confirm 1) that YM155 is a drug candidate for neuroblastoma including therapy-refractory disease with the majority of cell lines being sensitive to clinically achievable YM155 concentrations and 2) that ABCB1 is an important determinant of YM155 sensitivity. Notably, there were substantial differences in the relative efficacy of the ABCB1 inhibitors verapamil and zosuquidar in sensitising ABCB1-expressing cells to YM155. These differences may be caused by sequence variations in the transporters in the different cell lines, which suggests that it may be possible to design variant-specific ABCB1 inhibitors. Moreover, we present novel findings indicating 1) that YM155 resistance is also mediated by ABCC1 (an ABC transporter of prognostic relevance in neuroblastoma [32]), 2) that (in contrast to previous assumptions) the p53 status does not indicate YM155 sensitivity, and 3) that YM155 cross-resistance profiles differ between cell lines adapted to drugs from different classes and even between cell lines adapted to drugs as similar as cisplatin and carboplatin. YM155 has shown moderate effects in clinical trials so far [23,59,60]. An improved understanding of the complex processes underlying response to this drug may enable the identification of biomarkers and the design more effective personalised therapies.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6694/12/3/577/s1, Figure S1: Effects of MYCN expression on neuroblastoma cell sensitivity to YM155, Figure S2: Effects of YM155 in ABCB1-transduced cells, Figure S3: ABCB1 expression and YM155 activity in drug-adapted neuroblastoma cells, Table S1: YM155 concentrations that reduce the viability of neuroblastoma cell lines by 50% (IC50), Table S2: ABCB1 phenotype of cell lines, Table S3: YM155 concentrations that reduce the viability of MYCN-amplified and non-MYCN-amplified neuroblastoma cell lines by 50% (IC50), Table S4: YM155 concentrations that reduce the viability of neuroblastoma cell lines with varying p53 status by 50% (IC50), Table S5: Mean YM155 concentrations that reduce the viability of neuroblastoma cell lines with resistance to certain drug classes by 50% (IC50), Table S6: YM155 concentrations that reduce the viability of neuroblastoma cell lines by 50% (IC50) in the absence or presence of the ABCB1 inhibitors verapamil (5 µM) or zosuquidar (1.25 µM), Table S7: YM155 concentrations that reduce the viability of neuroblastoma cell lines by 50% (IC50) in the absence or presence of zosuquidar (1.25 µM) as indicated by MTT assay or CellTiterGlo after 120h of incubation, Table S8: YM155 concentrations that reduce the viability of neuroblastoma cell lines by 50% (IC50) in the absence or presence of the ABCC1 inhibitor MK571 (10 µM).