Presence of wild-type TP53 in testicular cancer

TP53 is a tumour suppressor gene, which encodes a transcription factor with a dual role in the cellular stress response. Besides regulating its own negative feed-back regulators, MDM2 and MDM4, p53 transactivates a number of other genes that can either promote cell cycle arrest and DNA-repair, such as CDKN1A (encoding p21^cip1/waf1 [p21]) and GADD45, or apoptosis via genes belonging to the extrinsic (or death receptor) apoptosis pathway, such as TNFRSF10B (KILLER/DR5), TNFRSF6 (FAS) or the intrinsic (or mitochondrial) apoptosis pathway, via genes such as BBC3 (Puma) and PMAIP1 (Noxa) (Refs Reference Print and Loveland7, Reference Janus8, Reference Vousden and Lu9). Activation of these p53-dependent pathways is crucial in the response to chemotherapeutic drugs, such as cisplatin and a major role for p53 in determining treatment response has been described for several tumour types including testicular cancer (Refs Reference Lowe10, Reference Levine11, Reference Cheng12, Reference Wahl and Carr13, Reference Johnstone, Ruefli and Lowe14).

A relatively unique feature of testicular cancers is the high percentage of tumours that retain wild-type TP53 (Refs Reference Heidenreich15, Reference Spierings16, Reference Houldsworth17). Although many other tumour types (ranging from 10 to 70% of the tumours) harbour mutations in TP53 that compromise p53 function, less than 1% of human testicular cancers show TP53 mutations (Refs Reference Vousden and Lu9, Reference Greenblatt18, Reference Olivier19, Reference Hamroun20, Reference Petitjean21). Lack of TP53 mutations in TCs has been explained by latent transcriptional activity of p53 at basal levels, which is reversed in response to DNA damage signals (Refs Reference Lutzker and Levine22, Reference Lutzker, Mathew and Taller23). Other studies, however, suggested that p53 is functional in TC, since wild-type p53 levels correlate with expression levels of MDM2 in vitro and in vivo (Refs Reference Datta24, Reference Kersemaekers25, Reference Houldsworth26, Reference Riou27). The p53 pathway is crucial for effective tumour suppression (Refs Reference Lowe10, Reference Levine11, Reference Hanahan and Weinberg28), and it was therefore argued that TCs have defects in the p53 pathway, outside a direct TP53 mutation (Fig. 1).

Figure 1.

Presence of wild-type p53 in testicular cancer. The p53 pathway is crucial for effective tumour suppression, mainly by activation of a cell cycle arrest at the G₁/S transition. TCs are reported to have a deregulated G₁/S checkpoint, here we show a schematic overview of connected pathways, which are causative for the presence of wild-type p53 in testicular cancer. Solid black lines represent active, whereas solid grey lines represent in-active signalling. Note that if TC cells express high levels of p21 growth arrest will not occur, since p21 will be phosphorylated and therefore cytoplasmically localised.

One typical downstream effect of p53 signalling is the activation of a cell cycle arrest at the G₁/S transition. TCs are reported to have a deregulated G₁/S checkpoint (Refs Reference Bartkova29, Reference Bartkova30, Reference Bartkova31, Reference Strohmeyer32), in line with almost undetectable p21 and pRb levels in seminomas and embryonal carcinomas, whereas p27, cyclin D2 and cyclin E are overexpressed due to gene amplification (Refs Reference Datta24, Reference Bartkova29, Reference Bartkova30, Reference Bartkova31, Reference Strohmeyer32, Reference Guillou33, Reference Vaughn34). The presence of isochromosome 12 (i12p) in the majority of TC can explain the amplification of the p27 and cyclin D2 encoding genes CDKN1B and CNND2 (for further review on deregulated G₁/S checkpoint in TC see (Refs Reference Spierings, de Vries, Vellenga and de Jong35, Reference di Pietro36)). P27 is mainly cytoplasmically localised in TC cell lines due to PI3K/Akt-dependent phosphorylation of the nuclear localisation signal (NLS) domain of p27 (Ref. Reference Di Vizio37). Cytoplasmically localised p27 has lost the ability to inhibit cell cycle progression, but has gained altered tumour progression functions. In addition, p27 is associated with inhibition of apoptosis and interestingly, might actually promote G₁/S cell cycle progression in that case (Refs Reference Bagui38, Reference Blagosklonny39). A reason for the high PI3K/Akt activity in TC might be the lack of PTEN (phosphatase and tensin homologue), a negative regulator of the PI3K/Akt pathway (Ref. Reference Di Vizio37).

In a recent study, a possible mechanism in TCs explaining the low p21 expression despite the presence of wild-type TP53 in these tumours was identified (Ref. Reference Koster40). ECs share many similarities with embryonic stem cells (ES) (Refs Reference Oosterhuis and Looijenga3, Reference Josephson41), including their miRNA expression profile (Ref. Reference Josephson41) and expression of a cassette of pluripotency genes, such as OCT4 (POU5F1), NANOG and STELLAR (DPPA3P2) (Refs Reference Oosterhuis and Looijenga3, Reference Josephson41, Reference Giuliano42). In murine ES cells, it was recently described that Oct4 regulates transcription of the miR-17/106b seed family members, including miR-17-5p, miR-106a and miR-20a (Ref. Reference Marson43). The miR-106b seed family has been implied in embryonal stem cell maintenance, control of differentiation (Ref. Reference Foshay and Gallicano44) and fine-tuning of p21 expression, via inhibition of p21 translation by binding to the p21 3′-UTR (Refs Reference Koster40, Reference Ivanovska45, Reference Petrocca46, Reference Koster and de Jong47, Reference Gibcus48). In cisplatin-sensitive TC cell lines and the EC component of chemo-sensitive tumours, low levels of cytoplasmic p21 similarly correlated with high expression of Oct4 and of miRs belonging to the miR-17/106b seed family (Ref. Reference Koster40).

P53 activity can be influenced by MDM2. By binding to the transactivation domain of p53, MDM2 negatively regulates p53 activity and stability via several mechanisms such as stimulating p53 nuclear export and promoting p53 degradation through ubiquitination, (Ref. Reference Vousden and Prives49). MDM2 is a nuclear–cytoplasmatic shuttling protein (Ref. Reference Zhang and Xiong50), mainly present in the nucleus of untreated TC cells, whereas p53 was found predominantly localised to the cytoplasm of TC cells (Ref. Reference Koster51). In addition, the functionality of MDM2 is known to be strongly dependent on the expression levels of p14/Arf (p14) and MDM4 (Ref. Reference Stad52). The p14 protein directly binds and inactivates MDM2 (Refs Reference Bothner53, Reference Honda and Yasuda54, Reference Kamijo55) thus segregating MDM2 from p53 (Ref. Reference Lowe and Sherr56). However, p14 could not be detected in TC cells (Ref. Reference di Pietro57), suggesting that p53 activity is controlled by MDM2 in untreated TC cells.

Furthermore, high expression of the miR-371–373 cluster was observed in TC, which mimics mutated TP53 by targeting the tumour suppressor gene protein LATS-2, involved in the regulation of G₁–S transition in the cell cycle (Ref. Reference Voorhoeve58). Moreover, in TC, in contrast to other solid tumour types, the DNA-damage response (DDR) machinery is rarely activated spontaneously (Refs Reference Bartkova59, Reference Bartek, Lukas and Bartkova60). Therefore, no selective pressure on aberrations of the DDR genes, such as TP53, ATM/ATR and CHK2 is required for survival (Refs Reference Bartkova59, Reference Bartek, Lukas and Bartkova60).

Canonical cisplatin-induced DDR in wild-type p53 expressing cells

Intracellular, cisplatin becomes biologically active by the aquation of one or both of the two chloride groups, so that it can covalently bind to DNA, thereby forming DNA adducts (Ref. Reference Kelland61). These adducts cause inter- and intrastrand DNA cross-links and distortions in DNA, including unwinding and bending (Refs Reference Kelland61, Reference Siddik62). This process activates various signal-transduction pathways involved in the DDR leading to cell-cycle arrest, DNA repair and/or apoptosis (Refs Reference Kelland61, Reference Siddik62, Reference Rabik and Dolan63). Repair of cisplatin–DNA adducts is mainly performed through nucleotide excision repair (NER). When cisplatin-induced DNA crosslinks are not repaired using NER, DNA double-strand breaks (DSBs) may accumulate during replication. Such DNA DSBs activate the upstream DDR kinases ATM and DNA-PK (and indirectly also ATR), which in turn activate many down-stream targets, including p53, to mediate cell cycle arrest, initiate DNA repair and/or induce apoptosis (Ref. Reference Khanna and Jackson64). In response to DNA damage, a variety of post-translational modifications of p53 are induced (Refs Reference Espinosa65, Reference Murray-Zmijewski, Slee and Lu66, Reference Giaccia and Kastan67). However, conflicting data about the importance of these p53 modifications have been published (Refs Reference Toledo and Wahl68, Reference Ashcroft, Kubbutat and Vousden69). Phosphorylation of MDM2 and MDM4 by ATM, results in decreased activity and stability of the two proteins and the loss of a negative feedback on p53 (Refs Reference Toledo and Wahl68, Reference Ashcroft, Kubbutat and Vousden69, Reference Meulmeester70, Reference Cheng71, Reference Cheng and Chen72, Reference Vassilev73, Reference Marine, Dyer and Jochemsen74, Reference Ogawara75). The activation of the p53 pathway, in turn, leads to cell-cycle arrest and DNA repair, or apoptosis through the p53-dependent intrinsic and/or extrinsic apoptosis pathway (Refs Reference Print and Loveland7, Reference Janus8, Reference Vousden and Lu9, Reference Kelland61, Reference Siddik62, Reference Rabik and Dolan63).

Cisplatin-induced DDR in wild-type p53 expressing testicular cancer

Sensitivity of TC to cisplatin might be determined by efficient influx, activation and/or accessibility of the DNA to cisplatin. In turn, resistance to cisplatin could be due to the unsuccessful achievement of adequate intracellular levels of cisplatin, by a reduction in drug uptake, an increase in drug export or an increase in intracellular detoxification (Ref. Reference Koberle76). However, in TC, not much difference was observed in access of cisplatin to DNA, comparing cisplatin-sensitive and cisplatin-resistant TC cell lines to other tumour cell types (Refs Reference Masters and Koberle2, Reference Spierings, de Vries, Vellenga and de Jong35, Reference di Pietro36, Reference Koberle76, Reference Gosland77, Reference Masters78).

Testicular cancers, in contrast to other more resistant tumour types (Refs Reference Wang79, Reference Rosell80) are known to express very low levels of the essential NER components ERCC1, XPF and XPA (Refs Reference Koberle81, Reference Welsh82, Reference Honecker83, Reference Koberle84, Reference Koberle85). Low levels of ERCC1 and XPF are rate-limiting for the repair of especially interstrand crosslinks in TC cells, which could explain their cisplatin sensitivity (Ref. Reference Usanova86). In addition, a recent paper reported that TC cell lines have reduced homologous recombination activity, which correlated with their high cisplatin sensitivity (Ref. Reference Cavallo87). It is interesting that all TC cell lines could be further sensitised by a combination of cisplatin with a clinically applicable PARP inhibitor (AZD2281, olaparib) (Ref. Reference Cavallo87). A third DNA repair system involved in cisplatin-induced DNA lesions, is the mismatch repair system (MMR). Losses or defects in MMR, as reflected in reduced microsatellite stability, can confer resistance to cisplatin. In several studies, however, TCs were found to be microsatellite stable, suggesting that the mismatch repair system (MMR) system is functional in most TCs (Refs Reference Devouassoux-Shisheboran88, Reference Lothe89).

Cisplatin-induced apoptosis in testicular cancer

Treatment with cisplatin leads to rapid induction of apoptosis in TC cell lines (Fig. 2a) (Refs Reference Spierings, de Vries, Vellenga and de Jong35, Reference Chresta, Masters and Hickman90, Reference Arriola91, Reference Timmer-Bosscha92). A major role for p53 has been described in the response to chemotherapeutic drugs, such as cisplatin (Refs Reference Lowe10, Reference Levine11, Reference Cheng12, Reference Wahl and Carr13, Reference Johnstone, Ruefli and Lowe14). Several studies have shown that cisplatin can activate, via p53, the extrinsic and intrinsic apoptosis pathway in parallel in cancer cells from various tumour types, (Refs Reference Friesen, Fulda and Debatin93, Reference Petak and Houghton94, Reference Fulda95, Reference Fulda96). Cisplatin causes enhanced Fas death receptor expression and activation of the extrinsic apoptosis pathway via the Fas/Fas ligand interaction in TC cells (Ref. Reference Spierings97). Activation of the intrinsic apoptosis pathway following cisplatin treatment was also observed in these cells (Refs Reference Spierings97, Reference Gutekunst98, Reference Mueller99). Cross-talk between the extrinsic and intrinsic apoptosis pathway amplifies apoptotic signalling, since Fas signalling leads to caspase-8 induced cleavage of Bid and in turn, cleaved Bid can activate the intrinsic apoptosis pathway (Refs Reference Scaffidi100, Reference Martinou and Green101). Blocking the Fas apoptosis pathway reduced apoptosis to a large extent in TC cells at low cisplatin doses, but was less effective at higher cisplatin doses, suggesting a difference in cisplatin-concentration-dependent activation between the two apoptosis pathways in TC cells (Ref. Reference Spierings97). In contrast to other wild-type p53 expressing tumour cell types, cisplatin-sensitive TC cells display cisplatin-induced up-regulation of p53, MDM2 and the pro-apoptotic proteins Fas, and Noxa, while levels of p21 are much less affected (Refs Reference Spierings16, Reference Koster40, Reference Koster and de Jong47, Reference Koster51, Reference di Pietro57, Reference Spierings97, Reference Li102, Reference Kerley-Hamilton103, Reference Mao104). The important role of wild-type TP53 in the response to cisplatin was demonstrated in cisplatin-sensitive Tera cells. Suppression of p53 resulted in down-regulation of p53-transactivated genes implicated in the extrinsic apoptosis pathway (FAS, LRDD and PHLDA3) and the intrinsic apoptosis pathway (Puma, Noxa) (Refs Reference di Pietro57, Reference Gutekunst98, Reference Kerley-Hamilton103, Reference Mao104) and caused a concomitant reduction in cisplatin sensitivity (Refs Reference di Pietro57, Reference Gutekunst98, Reference Kerley-Hamilton103). The low expression levels of cytoplasmic p21 are causative in the sensitivity of TC cells to cisplatin and Fas-mediated apoptosis (Refs Reference Spierings16, Reference Koster40, Reference di Pietro57). Several possible anti-apoptotic functions of high cytoplasmic p21 levels have been described in other tumour types that are strongly related to Fas-induced apoptosis. Upon Fas activation, Fas is trimerised followed by Daxx recruitment to the receptor. Daxx then interacts with ASK1, which results in the activation of JNK/SAPK and p38, and concomitantly apoptosis induction. This process is blocked by cytoplasmic p21 binding to ASK1 (Refs Reference Asada105, Reference Schepers106, Reference Zhan107). In addition, following activation of Fas, CDK2 activity is induced in a caspase-dependent way leading to cross-activation of the intrinsic apoptosis pathway (Refs Reference Zhou108, Reference Choi109, Reference Jin110). Functional interaction of CDK2 with FOXO1 after DNA damage was found to be an important mechanism to induce, among others, Fas ligand expression and thus promote activation of the Fas pathway (Ref. Reference Huang111). Since p21 binds to ASK1 and CDK2 in cisplatin-resistant TC cells (Ref. Reference Koster40), it can be hypothesised that both play a role in Fas-mediated apoptosis in TC cells. Interestingly, evidence for a pro-apoptotic role of CDK2 in TC cell lines was recently demonstrated. Thus, it will be interesting to further unravel how ASK1 and CDK2 interact with the apoptosis pathways in TC cells. Two recent studies showed that the pro-apoptotic p53 target gene Noxa (signalling in the intrinsic apoptosis pathway) plays a prominent role in cisplatin-induced apoptosis in TC cells (Refs Reference Gutekunst98, Reference Grande112). High levels of Oct4 strongly correlated with high expression of Noxa, while Oct4 suppression led to a strong reduction in Noxa expression (Ref. Reference Gutekunst113). Interestingly, Noxa was found to be significantly higher expressed in tumours of TC patients with blood markers predicting a good clinical prognosis (Ref. Reference Grande112).

Figure 2.

The role of wild-type p53, Oct4, Noxa, MDM2, p21 and Fas in cisplatin sensitivity and resistance of testicular cancer. Cisplatin-induced DNA damage activates p53, which in turn transcribes MDM2, Noxa, Fas and p21, and activates the Fas apoptosis pathway and the mitochondrial apoptosis pathway. MDM2 is important for regulating p53 transactivational activity, Akt is important for dynamic shuttling of p21 from the nucleus towards the cytoplasm, while cytoplasmic p21 can block apoptosis. Oct4 regulates Noxa and miR-17/106b seed family members (miR-17/106b) that are involved in regulating p21 expression levels. (a) Cisplatin-sensitive cells are characterised by low levels of MDM2 and p21, while expressing high levels of Oct4 and consequently high levels of Noxa and miR-17/106b seed family members. The low levels of p21 and MDM2 and high levels of Noxa result in efficient induction of apoptosis, after treatment with cisplatin. (b) Cisplatin-resistant cells have lower levels of Oct4, Noxa and miR-17/106b seed family members and high amounts of MDM2 and cytoplasmic p21, resulting in moderate levels of cisplatin-induced apoptosis. Solid black lines represent active signalling after cisplatin treatment and grey lines represent less active signalling.

As discussed before, Oct4 is instrumental in regulating p21 expression levels and p21 was correlated with sensitivity to cisplatin in vivo and in vitro (Refs Reference Spierings16, Reference Koster40, Reference Wu114). Taken together, these results indicate that TC cells are primed to the pro-apoptotic functions of p53 due to high Oct4 levels and concomitantly elevated Noxa and reduced p21 protein levels (Fig. 2a).

The p53/MDM2 axis

MDM2 is an important negative feedback regulator of p53, and MDM2 activity towards p53 is regulated by phosphorylation of MDM2. ATM-mediated phosporylation of MDM2 inhibits its ability to poly-ubiquitinate p53, and is likely to be the key step in causing p53 stabilisation following DNA-damage. Alternatively, Akt can bind and phosphorylate MDM2, leading to its nuclear localisation and consequently enhanced MDM2-mediated ubiquitination and degradation of p53 (Refs Reference Toledo and Wahl68, Reference Ashcroft, Kubbutat and Vousden69, Reference Meulmeester70, Reference Cheng71, Reference Cheng and Chen72, Reference Vassilev73, Reference Marine, Dyer and Jochemsen74, Reference Ogawara75). Therefore, tight regulation of p53 by MDM2, fine-tuned through phosporylation by ATM and Akt, may be of importance for the response to cisplatin in TC. Of interest, intrinsic cisplatin-resistant TC cells showed higher expression of MDM2, both untreated and after induction with cisplatin, compared with cisplatin-sensitive TC cells (Refs Reference Koster51, Reference di Pietro57, Reference Li102). In addition, p53 function was found impeded by the interaction with MDM2 and the subsequent sequestration of p53 in the cytoplasm of cisplatin-resistant TC cells (Ref. Reference Koster51). In acquired cisplatin-resistant TC cells, a higher expression of PIK3R1 and GIGYF2 was observed compared with their parental cisplatin-sensitive TC cells (Ref. Reference Noel124). Both PIK3R1 and GIGYF2 are involved in regulating PI3K/Akt activity (Refs Reference Ajiro125, Reference Ajiro126, Reference Philp127, Reference Goswami, Ranganathan and Rangnekar128). These results indicate that enhanced activity of Akt, as seen in cisplatin-resistant cells may stimulate the inhibitory effect of MDM2 on p53, leading to a higher threshold for the ATM to activate p53. Indeed, higher cisplatin concentrations were necessary to disturb p53/MDM2 complex formation in cisplatin-resistant cells (Ref. Reference Koster51). Inhibition of MDM2, using siRNA or an MDM2/MDM4 blocking peptide led to up-regulation of p53, Fas and Puma and sensitised TC cell to cisplatin, while the increase in p21 levels was much less pronounced (Refs Reference Koster51, Reference Li102). Taken together these observations show an important role for the p53/MDM2 axis in regulating cellular responses to cisplatin in TC.

Differentiation and cisplatin resistance in testicular cancer

Expression profiling of TC tumour subtypes showed differential expression of core stem cell genes, such as Oct4, NANOG and STELLAR, comparing EC and seminoma to the other subtypes (Refs Reference Juric129, Reference Korkola130, Reference Korkola131, Reference Korkola132, Reference Sperger133). Several of these pluripotency genes were higher expressed in untreated versus cisplatin-treated tumours. Moreover, the gene expression signature predicting outcome of cisplatin-treated TC also suggested that good outcome genes are associated with immune-related functions and pathways responsible for repression of differentiation (Ref. Reference Korkola132). High expression levels of Oct4 in xenografts TC models, cell lines and tumours were also highly correlated with cisplatin response (Refs Reference Koster40, Reference Koster and de Jong47, Reference Gutekunst113, Reference Wu114, Reference Mueller134, Reference Mueller135).

Targeting the p53/MDM2 axis

In cisplatin-resistant TC cells, the p53 function was shown to be impeded through enhanced interaction with MDM2 (Ref. Reference Koster51). Therefore, interfering in the p53–MDM2 interaction with small molecules, like Nutlin-3a, provides an attractive strategy for (re)activating wild-type p53 in a non-genotoxic way. Nutlin-3a inhibits the interaction between MDM2 and p53 by binding MDM2 in the p53-binding pocket leading to the stabilisation of p53 (Ref. Reference Vassilev73). At minimally cytotoxic concentrations, Nutlin-3a sensitised wild-type p53 expressing TC cells to cisplatin (Fig. 3a) (Refs Reference Koster51, Reference Li102, Reference Bauer171). Nutlin-3a treatment, as single agent, caused a high induction of p53 and Puma protein expression and Fas membrane expression and strongly reduced survival in both cisplatin-resistant and -sensitive TC cell lines, by inducing massive apoptosis (Refs Reference Koster51, Reference Gutekunst98, Reference Li102). Combining non-toxic concentrations of Nutlin-3a with cisplatin led to a further activation of the p53 pathway and the Fas/FasL death receptor pathway and sensitised both intrinsic as well as acquired cisplatin-resistant TC cells to very low concentrations of cisplatin (Refs Reference Koster51). Blocking of the Fas/FasL death receptor pathway impaired apoptosis induction of Nutlin-3a as well as Nutlin-3a combined with cisplatin in TC cells (Refs Reference Koster51). After Nutlin-3a treatment or cisplatin treatment, p21 mRNA and protein levels remained relatively low in TC cells (Refs Reference Spierings16, Reference Koster40, Reference Koster51, Reference di Pietro57, Reference Spierings97, Reference Li102), especially compared with other tumour cell lines (Refs Reference Spierings16, Reference Li102), although, the occupancy of p53 on the CDKN1A (p21) promoter in TC cells showed similar levels (Ref. Reference Li102). Different co-factors of p53 might differentially influence the balance by facilitating and activating pro-apoptotic or anti-apoptotic promoters (Refs Reference Vousden and Prives49, Reference Meulmeester70, Reference Cheng and Chen72, Reference Gomes and Espinosa172, Reference Morachis, Murawsky and Emerson173, Reference Shikama174, Reference Samuels-Lev175). NFY is such a co-factor and was shown to inhibit CDKN1A (p21) promoter activity, whereas at the TNFRSF6 promoter it enhances transcription (Ref. Reference Gomes and Espinosa172, Reference Morachis, Murawsky and Emerson173), indicating that inhibition of the negative feedback on p53 in TC turns the balance in favour of transcription of pro-apoptotic genes, such as TNFRSF6, encoding Fas and BBC3 (Puma). This needs to be further elucidated, especially since, in contrast to TC, in most other solid tumour cells MDM2 inhibitors, as single agents, mainly induce cell cycle arrest (Refs Reference Enge176, Reference Brummelkamp177), probably through p21. Of interest, stapled peptides targeting MDM2 are in pre-clinical development and can target both the p53-MDM2 and p53-MDM4 interaction (Refs Reference Bernal178, Reference Joseph, Lane and Verma179). In melanoma, for instance, it was shown that inhibition of the p53-MDM2 interaction only had minimal apoptotic effects. In contrast, inhibiting the p53-MDM4 interaction, using a stapled peptide, restored the wild-type p53 activity in melanoma cells resulting in increased sensitivity to chemotherapy and BRAF inhibitors (Ref. Reference Gembarska180). In testicular cancer cell lines, however, the negative feedback of MDM4 on p53 appeared to be moderate and cell line-dependent (Ref. Reference Li102).

The results discussed here suggest that the activation of the complex DNA damage-induced p53 pathway, with MDM2 as feedback mechanism fine-tuning the p53 response, seems to play a crucial role in balancing the chemotherapy response of TC. Moreover, efficient activation of the p53-induced pro-apoptotic pathway by antagonising MDM2 overrules the anti-apoptotic effect of p21 in response to cisplatin-induced DNA damage in TC. Currently, three MDM2 inhibitors, i.e. RG7112, a Nutlin-3a based molecule (RO5045337 – Roche), another MDM2 inhibitor (RO5503781 – Roche) and Kevetrin (Cellceutix Pharmaceuticals), have entered phase I clinical trials in haematological malignancies and (advanced) solid tumours.

Disrupting the p53–MDM2 interaction, especially, in combination with cisplatin, enhanced Fas death receptor-mediated apoptosis in TC cells (Ref. Reference Koster51). Interestingly, apoptosis induction, after combining Nutlin-3a and cisplatin, was not only most pronounced in cisplatin-resistant TC cells but also in cisplatin-resistant cells from other tumour types (Refs Reference Koster51, Reference Mir181). Hence, targeting the p53-MDM2 interaction in combination with standard cisplatin-based treatment is a therapeutic strategy warranted to pursue. Since Fas expression is strongly induced by this combination, additional activation of the exintrinsic apoptosis pathway by administration of FasL might be an interesting approach. But apart from a hexameric form of FasL (APO010) (Ref. Reference Etter182) that had entered a phase 1 clinical trial in 2007, no Fas targeting drugs are available. Moreover, combining cisplatin with agents targeting inhibitors of the intrinsic apoptosis pathway, such as Obatoclax, ABT-199 or ABT-263 (Navitoclax), might be favourable for treatment of cisplatin-resistant/refractory TCs.

In addition, these non-genotoxic strategies might be beneficial for all metastatic TCs leading to a reduction of the standard of care chemotherapy doses, since many who are cured using cisplatin-based chemotherapy will develop serious late toxicities (Refs Reference Altena183, Reference de Haas184).

Strategies to target cytoplasmic p21 and/or (re)express Oct4

Increased levels of cytoplasmically localised p21 are a key determinant of cisplatin-resistance in TC cells (Ref. Reference Koster40). Cytoplasmic localisation of p21 is caused by PI3K/Akt-mediated phosphorylation of p21 in TC (Ref. Reference Koster40) and other cancers, such as ovarian, breast and kidney cancer (Refs Reference Zhou185, Reference Xia186, Reference Lin187, Reference Menendez, Mehmi and Lupu188, Reference Perez-Tenorio189, Reference Xia190). Cytoplasmically localised p21 can be re-localised towards the nucleus by dephosphorylation of p21, using triciribine, an Akt inhibitor currently in phase 1/2 trials, depletion of Akt using RNA interference or application of the pan PI3 K inhibitor LY294002. Relocalisation of p21 caused a loss of complex formation of p21 with pro-apoptotic CDK2 and sensitised cisplatin-resistant TC cells to cisplatin-induced apoptosis (Fig. 3b) (Ref. Reference Koster40). In other tumour types for which cisplatin is used as first-line therapy, cytoplasmic p21 levels were found to be strongly associated with poor treatment outcome (Refs Reference Xia186, Reference Lin187, Reference Xia190, Reference Jin191), arguing that in these cancers inhibition of the PI3K/Akt pathway in combination with cisplatin might be beneficial as well. Therefore, the use of PI3K/Akt inhibitors (Ref. Reference Ihle and Powis192) seems to be a therapeutically interesting approach. Whether inhibitors currently in clinical trial, such as the Akt inhibitors triciribine, perifosine and MK-2206, the PI3 K inhibitors SF1126, PX-866, BKM120 or GDC-0941 are effective, especially when combined with cisplatin, needs to be elucidated.

In addition, PI3K/Akt might be inhibited by targeting upstream tyrosine kinase membrane receptors such as KIT, ERBB2, PDGFR and VEGFRs that have been implicated in TC (Refs Reference Palumbo193, Reference Rapley194, Reference McIntyre195, Reference Goddard196, Reference Rapley197, Reference Kanetsky198, Reference Gilbert, Rapley and Shipley199, Reference McIntyre200, Reference Viglietto201, Reference Basciani202). Of interest, sunitinib, as single agent, inhibited tumour growth in an orthotopic cisplatin-sensitive and -resistant patient-derived ChC model of TC and inhibited cell growth of cisplatin-sensitive and -resistant TC cell lines (Refs Reference Castillo-Avila203, Reference Oechsle204). Disappointingly, only modest activity of sunitinib was observed in heavily pretreated TC patients (Ref. Reference Oechsle204). Sunitinib, however, synergised with cisplatin in the patient derived cisplatin-resistant ChC orthotopic model (Refs Reference Castillo-Avila203), arguing that PI3K/Akt inhibition might be especially effective when combined with cisplatin.

In addition, treatment with 5-aza sensitised an intrinsic cisplatin-resistant seminoma cell line to cisplatin, while restoring cisplatin sensitivity of acquired cisplatin-resistant EC cells (Refs Reference Wermann162, Reference Beyrouthy169). In contrast, a clinical trial with 5-aza, as single agent, did not show any activity in cisplatin pretreated TC patients (Ref. Reference Roth205). This might be due to treatment with 5-aza only, since the combination with carboplatin has shown activity in other platinum-resistant solid tumours such as ovarian cancer (Ref. Reference Fang206). Besides sensitising cisplatin-resistant TC cells to cisplatin, 5-aza treatment resulted in (re)expression of Oct4. As discussed before, Oct4 is correlated with sensitivity to cisplatin and was shown to be instrumental in lowering anti-apoptotic p21 and enhancing pro-apoptotic Noxa expression levels in TC. Therefore, we hypothesise that 5-aza-mediated (re)expression of Oct4 primes the TC cells for cisplatin-induced apoptosis among others by influencing p21 and Noxa levels.