Introduction
Rhabdomyosarcoma (RMS) is the most common childhood soft tissue sarcoma, representing approximately 50% of all sarcomas in children aged 0–14 years (McDowell
2003; O’Neill et al.
2013). Adolescents and more rarely adults may also be affected (Ferrari et al.
2003). RMS is a heterogeneous tumour that is believed to develop as a result of genetic alterations occurring in mesenchymal progenitor/stem cells, which express some markers of normal skeletal muscle but show an incompletely differentiated muscle phenotype (Merlino and Helman
1999). Alveolar RMS (ARMS) and embryonal RMS (ERMS), the two most common histological subtypes in childhood, have distinct clinicopathological features and outcomes (Coffin
1997; Parham and Barr
2013). ARMSs and ERMSs are both characterised by distinctive genetic alterations that are likely to play a decisive role in their pathogenesis (Anderson et al.
1999; Goldstein et al.
2006; Martinelli et al.
2009; Marshall and Grosveld
2012; Parham and Barr
2013; Robbins et al.
2016). ERMSs are more frequent (~ 80% of cases) and generally affect younger children (0–4 years), occurring more commonly in the neck, head and genito-urinary tract (Parham and Barr
2013). ARMSs (~ 20% of cases) usually present throughout childhood and adolescence, frequently originate in the extremities and trunk, often with regional or metastatic lymph node involvement already at diagnosis, and have high tendency to metastasize carrying a significantly worse outcome (Parham and Barr
2013). Indeed, 70% of children with localized disease survive with conventional treatment (Arndt et al.
2009), including surgery, radiotherapy and chemotherapy. Metastatic RMSs, however, are frequently resistant or present relapse after an initial response, with a 5-year event-free survival rate at about 30% (Sorensen et al.
2002; Ognjanovic et al.
2009; Wolden et al.
2015). Therefore, the outcome for high-risk RMS cases remains very poor and the discovery of innovative therapies is an absolute priority to improve therapeutic activity and reduce toxicity.
Poly(ADP-ribose) polymerases (PARPs) belong to a large family of enzymes that catalyse the formation of poly(ADP-ribose) polymers (PARylation) onto different targets and themselves, this leading to a fine modulation of different cellular processes and molecular pathways, such as DNA damage response (DDR), cellular differentiation, chromatin remodelling, transcription, cell death and mitotic progression (Helleday et al.
2007; Dungey et al.
2008; De Vos et al.
2012; Bai
2015; Brown et al.
2017). Only PARP1 and, to a lesser extent, PARP2 and PARP3 play an essential role in repairing single- or double-stranded DNA breaks (SSBs or DSBs, respectively) as well as stalled replication forks and DNA crosslinks (De Vos et al.
2012), with PARP2 being specifically able to recognize DNA gaps and flaps (Yélamos et al.
2008) and PARP3 being selectively activated by DSBs (Boehler et al.
2011). PARP1 and PARP2 are involved in fixing DNA-strand interruptions by the homologous recombination (HR) pathway (Henning and Stürzbecher
2003), whilst PARP3 acts via the nonhomologous end joining (NHEJ) repair system (Davis and Chen
2013).
PARP inhibitors (PARPi) comprehend a wide range of chemical compounds able to abrogate PARP functionality thus bringing to the accumulation of SSBs, which in turn are converted into DSBs that cells are not able to repair causing cancer cell death (Wiltshire et al.
2010). The mechanism of action of PARPi is the block of the catalytic domain of PARP enzymes, but these agents can also trap PARP proteins on the double-stranded DNA helix, this leading to cytotoxic lesions (Murai et al.
2012; D’Arcangelo et al.
2016). PARP inhibition has a potential therapeutic role as monotherapy in tumours carrying constitutive mutations in DDR genes, as well as in combination therapies for its ability to enhance the activity of anticancer drugs with genotoxic action, including DNA alkylating agents, topoisomerase II inhibitors and ionising radiation (IR) (Jorgensen
2009; Kelley et al.
2014; Lord and Ashworth
2017), since targeting similar molecular functions results in cell death. The “synthetic lethality” conferred by PARPi (Martin et al.
2008; Lord and Ashworth
2017) is not only restricted to BRCA1- and BRCA2-mutated tumours but also to neoplasia harbouring genetic alterations in other HR genes, such as ATM, RAD51, PTEN, XRCC2, etc (Bang et al.
2013; Kelley et al.
2014), this suggesting a therapeutic potential role of PARP inhibition in a wide range of human malignancies. Several clinical trials aimed at assessing for different PARPi are in progress. Olaparib (AZD2281), a selective inhibitor of PARP1 and PARP2, has been used in different solid tumours and, recently, this drug has been approved as personalized therapy (Kim et al.
2015; Goulooze et al.
2016) for patients with BRCA1/2-mutated advanced ovarian cancer, who have been treated with three or more prior lines of chemotherapy (Phase III clinical trial, NCT01874353). AZD2461, a next-generation agent able to also inhibit PARP3 activity, has been recently synthesised in order to overcome PARPi-related resistance and to be better tolerated than Olaparib (Jaspers et al.
2013; O’Connor et al.
2016; Vaidyanathan et al.
2016). An encouraging therapeutic activity has recently been reported in clinical trials with AZD2461 on refractory solid tumours (Phase I clinical trial, NCT01247168).
A limited amount of information is available about the effects and the molecular mechanisms of PARP inhibition, as monotherapy or in combination with conventional therapies, in RMS. Only very recently, Mangoni et al. have shown that pretreatment with Olaparib, Iniparib or Veliparib, three PARP1 inhibitors, is able to induce a significant radiosensitization in different soft tissue sarcoma cell lines (Mangoni et al.
2018).
In the present study, we analysed the expression of PARP1, PARP2 and PARP3 genes in a panel of RMS primary tumours and cell lines, and evaluated the biological and molecular effects of PARP inhibition in RMS in vitro models by using Olaparib or AZD2461. We tested two different doses of both PARPi and determined the minimum concentrations of each molecule able to drive a biological effect in RH30 and RD cell lines, two in vitro models of ARMS and ERMS, respectively. We also assessed the possible synergistic effects between Olaparib or AZD2461 and IR, a combination which might represent a further step towards a more effective treatment of RMS patients, especially those with metastatic disease.
Methods
Reagents and irradiation
Olaparib and AZD2461 were purchased from Selleckchem (Suffolk, UK) and were reconstituted at 10 mM using dimethyl sulfoxide (DMSO). DMSO alone was used as control in untreated cells at 0.1% (v/v) concentration.
Irradiation was carried out using an ONCOR Impression Linear Accelerator (Siemens Medical Solutions USA, Inc, Concord, CA) at a dose rate of 2 Gy (190 UM/min) or 4 Gy (380 UM/min).
Cell cultures
Human RMS cell lines, RH30 (alveolar) and RD (embryonal), were maintained as previously described (Megiorni et al.
2016). Human foetal myoblast (HFM) cells were cultured in High Glucose DMEM supplemented with 20% FBS.
Tumour samples
Seventeen RMS tumour samples, 4 ARMSs and 13 ERMSs, were obtained at diagnosis before any treatment from children admitted to the Department of Oncology at Alder Hey Children’s NHS Trust, Liverpool. ARMS1-2-4 are fusion-positive tumours, whilst ARMS3 is a fusion-negative case, as assessed by FISH analysis for PAX3/7-FOXO1 translocations. Institutional written informed consent was obtained from the patient’s parents or legal guardians. The study underwent ethical review and approval according to the local institutional guidelines (Alder Hey Children’s NHS Foundation Trust Ethics Committee, approval number 09/H1002/88).
RNA extraction and quantitative Real Time PCR (q-PCR)
Total RNA, isolated from RMS tumour biopsies and cell lines, was reverse transcribed and analysed by using quantitative Real Time PCR (q-PCR) with specific TaqMan Real-Time Gene Expression Assays (Applied Biosystems), as previously described (Megiorni et al.
2016). Human PARP1 (Hs00242302_m1), PARP2 (Hs00193931_m1) and PARP3 (Hs00193946_m1) mRNA assays were used. Samples were normalized according to GAPDH transcript levels. Expression of miR-124-3p was analysed as previously described (Megiorni et al.
2014), by using sequence-specific TaqMan MicroRNA Assays (Applied Biosystems). U6 small nuclear RNA levels were used as internal control. The amount of each mRNA or miRNA was calculated by the comparative
Ct method (Livak and Schmittgen
2001) and expressed as fold change using the StepOne v2.3 software (Applied Biosystems). Each sample was run in triplicate, in at least two independent experiments.
Cell proliferation assays
RH30 and RD cells (3 × 105) were plated in six-well plates and treated with Olaparib (1.5 and 5 µM) or AZD2461 (5 and 10 µM). After 72 h, RH30 and RD living cells were diluted in a 1:1 mixture of trypan blue (Invitrogen) and counted using the Countess II Automated Cell Counter (Invitrogen), according to the manufacturer’s instructions.
Morphological assessment
RH30 and RD cells treated with Olaparib (1.5 and 5 µM) or AZD2461 (5 and 10 µM) for 72 h were photographed with an Axio Vert.A1 microscope (Carl Zeiss Microscopy, Thornwood, NY), furnished with an AxioCam MRc5 camera (Carl Zeiss Microscopy), at 20× magnification.
Cell cycle and apoptosis analysis by flow cytometry
For the cell cycle analysis, RH30 and RD cells (3 × 10
5) were incubated in six-well cell culture plates overnight to allow cell adhesion. Cells were treated with Olaparib (1.5 and 5 µM) or AZD2461 (5 and 10 µM) for 48 h. DMSO was used as mocked control. For the cell cycle analysis of the effects induced by the PARPi and IR combination, RH30 and RD cells, pretreated for 24 h with Olaparib or AZD2461 were irradiated and incubated for additional 24 h at 37 °C. Samples were stained with Propidium Iodide (PI) solution and subjected to flow cytometry by using a BD FACSCalibur (BD Biosciences, Franklin Lakes, NJ), as previously described (Megiorni et al.
2016). FACS data were quantified by using the ModFit LT 3.0 program (Verity Software House). Experiments were performed at least twice.
Apoptosis was analysed by using PE Annexin V Apoptosis Detection Kit I (BD Biosciences), following the manufacturer’s instructions. Briefly, RH30 and RD cells (3 × 105) were seeded overnight in six-well plate and treated with Olaparib, AZD2461 or DMSO for 48 and 144 h. Approximately 2 × 105 cells were stained with Annexin V and 7-Amino-Actinomycin (7-AAD) for 15 min at RT in the dark. Fluorescence intensities of treated samples and controls were analysed by flow cytometry by using the BD CellQuest Pro software. Experiments were performed at least twice.
RH30 and RD cells (3.2 × 105) treated for 24 h with Olaparib (1.5 and 5 µM) or AZD2461 (5 and 10 µM), were irradiated at a dose of 2 or 4 Gy/min. After 4 h incubation at 37 °C, 5% CO2, 2 × 103 cells/well were seeded in 6-well plates in triplicate. Medium was replaced every 3 days and after 12 days, colonies were stained with 0.1% crystal violet for 5 min at room temperature (RT). Colonies were photographed, and then crystal violet was solubilised in 30% acetic acid in water for 15 min at RT; absorbance was measured by using the Biochrom Libra S22 UV/VIS spectrophotometer (Biochrom, Berlin, DE) at wavelength of 595 nm; 30% acetic acid in water was used as the blank. Colony formation capacity in PARPi- and/or IR-treated cells was calculated in comparison to mocked control samples (DMSO alone), arbitrarily set to 1. The results were plotted as means ± SD of two separate experiments having three determinations per assay for each experimental condition.
Protein extraction and Western blotting
Total protein extracts and Western blotting assays were performed as previously described (Megiorni et al.
2016) using the following primary antibodies: phospho (p)-AKT, AKT, cleaved caspase 3, and γH2AX (Cell Signalling Technology, Danvers, MA); Bcl2, Cdc2 phosphorylated at Thr14/Tyr15, Cdc25C, Cyclin B1, Cyclin D1 and p21 (Santa Cruz Biotechnology, Dallas, TX). Antibody against tubulin (Sigma-Aldrich) was used as a loading control.
Immunofluorescence (IF) microscopy
RH30 and RD cells (5 × 10
4), seeded onto 2% gelatine coated-glass coverslips in 24-well plates, were allowed to attach overnight and then incubated for 48 h in the presence or absence of Olaparib (5 µM) or AZD2461 (10 µM). For IF analysis of the effects induced by the PARPi and IR combination, RH30 and RD cells, pretreated for 24 h with Olaparib or AZD2461 were irradiated and incubated for additional 4 h at 37 °C. IF assays were performed as previously described (Megiorni et al.
2016) using the following primary antibodies: Cdc2, p-Cdc2, Cdc25C, Cyclin B1, Cyclin D1, RAD51 (1:20 dilution in PBS; Santa Cruz Biotechnology), and γH2AX (1:500 in PBS; Cell Signaling). All single-stained or merged images were acquired with a Zeiss ApoTome microscope (40× magnification) using the Axiovision software (Carl Zeiss, Jena, Germany). For γH2AX and RAD51, focus fluorescence intensity in the respect of cell number in each analysed field was reported.
Statistical analysis
Data are presented as means ± SD. Statistical analyses were performed by two-tailed Student’s t test and a probability (p) < 0.05 was considered statistically significant. All the experiments were done in triplicates and repeated three times unless mentioned otherwise.
Discussion
RMS is the most frequent soft tissue sarcoma in childhood (McDowell
2003; O’Neill et al.
2013). Even if the survival probability has increased to about 70% for children and adolescents with RMS (Ognjanovic et al.
2009), the 5-year survival rate for patients with relapsed or metastatic disease is approximately 40%, mainly due to the development of chemo- and radioresistances (Wolden et al.
2015). Therefore, novel more effective therapeutic strategies are a pressing need, in those advanced patients.
In the present study, RH30 and RD cell lines—two in vitro models of ARMS and ERMS subtypes, respectively—were used to evaluate the cellular and molecular responses to the PARP inhibitors Olaparib and AZD2461, as single agents or in combination with IR. PARPi have strong cytotoxic effects in tumours harbouring genetic mutations in the components of the DDR system, such as BRCA1, BRCA2, PTEN and XCCR4, due to a mechanism indicated as “synthetic lethality”, according to which the inability to correct PARPi-induced SSBs leads to fatal DNA damage and cellular death (Brown et al.
2017). This study showed that the treatment with Olaparib, a specific inhibitor of PARP1 and PARP2 enzymes, or AZD2461, a newly PARP1/2/3 competitor, used as single agents, reduces cell proliferation in both RH30 and RD cells in a dose dependent manner, with a marked arrest in the G2/M phase of the cell cycle. In accordance with the flow-cytometry data, Olaparib or AZD2461-treated cells showed morphological alterations, such as an evident cell volume enlargement, which are characteristic of a defective cell division arrest (Fig.
2a). These observations are in agreement with data very recently reported by Mangoni et al. (
2018). In the present study we examined in detail the molecular mechanisms of PARPi effects on cell cycle progression and survival. We showed that changes in cell cycle distribution are driven by the deregulation of specific regulatory markers: (1) PARPi treatment led to the downregulation of Cyclin D1 expression and to the overexpression of p21 cell cycle regulator; and (2), PARPi activated the G2/M checkpoint in RMS cells by sequestering Cdc25C in the cytoplasm compartment, promoting hyper-phosphorylation of Cdc2 (p-Cdc2) at Thr14/Tyr15, and upregulating Cyclin B1 levels. Specifically, Cyclin B1 protein resulted predominantly accumulated around the nuclear envelope, this confirming that high levels of p-Cdc2 are not able to form an active complex with Cyclin B1, a phenomenon that prevents their entrance in the nucleus and stalls the mitosis process as reported in other cancer types (Jin et al.
1996). Concerning cell survival, PARPi prolonged exposure led to apoptosis through the inhibition of AKT activation and the modulation of relative downstream molecules. Indeed, reduced levels of phospho-AKT and Bcl2 proteins, with the concomitant cleavage and activation of the caspase-3 protein were observed in PARPi-treated cells. The present findings also demonstrate that the cell survival impairment we observed is linked to the accumulation of the DNA damage, which PARPi-treated cells are unable to repair. Phosphorylation of histone H2AX on serine 139 to form γH2AX, a sensitive marker for the indirect quantification of DNA DSBs (Bonner et al.
2008), was not only evident after 48 h of drug exposure but was significantly increased after 144 h in both Olaparib and AZD2461-treated cells, this explaining the observed severe consequences on the RMS cell survival. Indeed, persistence of γH2AX signal (Löbrich et al.
2010) correlated with an inefficient reconstitution of the DNA integrity, which is essentially performed by the HR signalling molecules in the G2/M phase (Polo and Jackson
2011). Finally, the increased expression of miR-124-3p in Olaparib- or AZD2461-treated RMS cells suggests that the modulation of this miRNA is involved in the complex molecular mechanisms underlying the PARPi-mediated cytotoxicity. Interestingly, low expression of miR-124-3p has been observed in different types of human cancers, including RMS (deep sequencing data reported in Megiorni et al.
2014), and the restoration of miR-124-3p levels is able to decrease cell survival by promoting apoptosis (Wang et al.
2016; Ma et al.
2018; Zhang et al.
2018). The molecular mechanisms and pathways related to the functions of miR-124-3p in RMS models will be further investigated.
Effects on cell proliferation, apoptosis and DNA damage were more pronounced by using higher concentrations of both drugs, i.e., 5 µM Olaparib and 10 µM AZD2461, and mainly in RH30 cells. Since BRCA1/BRCA2 mutations have not been found in RMS tumours (Mendes-Pereira et al.
2009), the presence of genetic/epigenetic alterations in other DNA repair machinery components or related factors cannot be excluded. To this regard, the expression of PTEN gene, which encodes for a protein involved in DNA DSB repair, is commonly suppressed in both ARMS and ERMS tumours (Zhu et al.
2016). The importance of PTEN deficiency in the synthetic lethality driven by PARP inhibitors has been described in several cancers (Mendes-Pereira et al.
2009; Dedes et al.
2010; McEllin et al.
2010; He et al.
2015). Therefore, the higher sensitivity to Olaparib or AZD2461 observed in RH30 cells compared to RD cells might partially be explained by the more pronounced down-regulation of PTEN levels in the alveolar compared to the embryonal cell line (our preliminary data not shown), this underlying the usefulness of PARPi treatment in tumours with deregulated HR-linked proteins, such as PTEN, independently to the BRCA1/2 status alone. A different possible mechanism underlying the more conspicuous PARPi-mediated effects in ARMS cells might be related to the high levels of MYCN protein detectable in RH30 but not in RD cells (our preliminary data not shown). These findings are consistent with previous studies in neuroblastoma (NB) tumours, showing that PARPi lead to DNA damage and cell death more effectively in MYCN-amplified than in MYCN-not-amplified NB cell lines (Bridges et al.
2014; Verhagen et al.
2015). MYCN gene has been demonstrated to sustain DNA damage by delaying the resolution of DNA lesions (Venere et al.
2014), which if not properly repaired can lead to cellular death, this establishing a mechanistic link between MYCN overexpression and sensitivity to PARP inhibition.
Notably, the present study supports the possibility to combine PARPi with standard treatments, in particular with radiotherapy, in patients with RMS. Since ionising radiations induce DNA breaks that require PARP activity for proper DSB repair, PARP inhibition provides an effective tool to make cancer cells more radiosensitive. Previously, in vitro and in vivo studies have reported that a series of PARPi is able to radiosensitise different tumour models, including breast cancer, glioblastoma, neuroblastoma and lung cancer (Mueller et al.
2013; Bridges et al.
2014; Venere et al.
2014; Verhagen et al.
2015). In the present study, the synergistic antitumour activity of the combined treatment of PARPi and IR was demonstrated by the evidence that the exposure to Olaparib or AZD2461 radiosensitises RMS cells by amplifying the quantity and the duration of DNA damage induced by IR. The susceptibility of tumour cells to ionising radiations was synergistically enhanced through the increased accumulation of chromatid-type breaks, the prolonged activation of the G2/M checkpoint and the reduction of clonogenic potential in both RH30 and RD cell lines. Cell cycle regulation has been reported as an important biological mechanism affecting radiosensitivity, with cells being most sensitive to radiation during the G2/M phase (Pawlik and Keyomarsi
2004). Likewise, different drugs have been shown to promote sensitivity to radiations by inducing cells to accumulate in this phase (Pawlik and Keyomarsi
2004; Duangmano et al.
2012). Our in vitro experiments demonstrated that Olaparib or AZD2461 plus radiations significantly increase γH2AX and RAD51 expression and nuclear accumulation, this further confirming that the combined treatment has a higher sensitizing activity compared to either treatment modality. The increase of γH2AX and RAD51 foci is suggestive of the presence of stalled fork recovery sites and endogenous replication stress, both signs of the attempt by tumour cells to repair DNA lesions (Costanzo
2011; He et al.
2015). RAD51 upregulation contributes to G2 phase arrest in order to help the HR systems in fixing DNA damage, but an excess of RAD51 nuclear foci is thought to promote genomic instability for inappropriate recombination events, including translocations and other rearrangements, this in turn having deleterious effects on cell survival. Indeed, the sustained overexpression of RAD51 nuclear signals has been associated with a reduced cell growth and apoptosis, as observed in Drosophila as well as in human cell lines (Flygare et al.
2001; Yoo and McKee
2004; Klein
2008), which confirms that a balanced interaction between RAD51 and other HR factors is needed to properly repair DNA.
A further interesting finding that emerges from our data is that low concentrations of either PARPi (1.5 µM Olaparib or 5 µM AZD2461) are adequate to increase substantially the efficacy of the ionising radiations in both RH30 and RD cell lines, this having potentially important clinical implications, since a significant level of tumour cell radiotoxicity can be achieved by more tolerable concentrations of PARPi and IR in combination. However, further studies will be needed to evaluate the antitumour activity and possible toxicity of Olaparib and AZD2461 as single agents and in combination with ionising radiations in RMS xenografts.
In conclusion, the present findings demonstrate that PARPi may represent a promising therapeutic approach in RMS treatment. Furthermore, PARPi-increased sensitivity to radiations may be associated with a significant therapeutic benefit by inhibiting tumour growth and survival and by counteracting the development of radioresistance, this potentially improving clinical outcome.