Microtubule-targeting chemotherapeutics induce apoptosis in cancer cells by promoting the phosphorylation and degradation of the anti-apoptotic BCL-2 family member MCL1. The signalling cascade linking microtubule disruption to MCL1 degradation remains however to be defined. Here, we establish an in vivo screening strategy in Caenorhabditis elegans to uncover genes involved in chemotherapy-induced apoptosis. Using an RNAi-based screen, we identify three genes required for vincristine-induced apoptosis. We show that the DEP domain protein LET-99 acts upstream of the heterotrimeric G protein alpha subunit GPA-11 to control activation of the stress kinase JNK-1. The human homologue of LET-99, DEPDC1, similarly regulates vincristine-induced cell death by promoting JNK-dependent degradation of the BCL-2 family protein MCL1. Collectively, these data uncover an evolutionarily conserved mediator of anti-tubulin drug-induced apoptosis and suggest that DEPDC1 levels could be an additional determinant for therapy response upstream of MCL1.
Anti-tubulin chemotherapeutics comprise a chemically diverse group of drugs that are prescribed widely for a variety of malignancies such as lymphomas, leukaemias, non-small-cell lung cancer and breast adenocarcinomas1,2. Microtubule-targeted drugs can suppress microtubule dynamics, resulting in mitotic arrest and apoptosis3. Recently, it has been shown that anti-tubulin drug treatment results in proteasomal degradation of the anti-apoptotic BCL-2 family member MCL1 by FBW7, which acts as a recognition particle of the SCFFBW7 E3 ubiquitin ligase4,5. The interaction between MCL1 and FBW7 is triggered by direct phosphorylation of MCL1 by JNK or p38 kinases4,5. Despite these major advances, the signalling cascade linking microtubule disruption and apoptosis, a key determinant of cancer therapy6, remains elusive.
The nematode C. elegans has been a leading model organism in the discovery of the apoptotic machinery7,8. To determine whether C. elegans can be used to dissect apoptotic signalling pathways activated by chemotherapeutic drugs, we evaluated the effect of four clinically relevant chemotherapeutic compounds on the C. elegans germ line, a tissue that induces apoptosis in response to various environmental stresses8,9,10. We found that cisplatin, doxorubicin, etoposide and vincristine treatment all readily promoted germline apoptosis (Fig. 1a–d), at concentrations that did not lead to cell cycle arrest (Supplementary Fig. 1a). This was generally true for anti-tubulin drugs as nocodazole and docetaxel similarly triggered germline apoptosis (Supplementary Fig. 1c, d). Apoptosis was dependent on the core apoptotic machinery because all four drugs failed to promote apoptosis in the absence of the effector caspase CED-3 (Fig. 1e–h). Similarly, a gain-of-function mutation in the Bcl-2 homologue ced-9 blocked apoptosis in response to all four drugs, as did loss of the BH3 domain protein EGL-1 (Fig. 1e–h). The p53 homologue CEP-1 transcriptionally upregulates EGL-1 in response to DNA damage9,11,12,13. We found that apoptosis induced by cisplatin, doxorubicin and etoposide—three agents known to act by inducing different types of DNA damage—was dependent on CEP-1/p53, whereas the anti-tubulin drug vincristine triggered germline apoptosis mainly in a CEP-1/p53-independent manner (Fig. 1e–h). To further corroborate the involvement of the BH3-only domain proteins12,14 in anti-tubulin drug-triggered apoptosis, we performed time-course analyses in egl-1 and ced-13 mutants. We detected a partial apoptotic defect in either egl-1 or ced-13 single mutants (Supplementary Fig. 1h, i). egl-1 and ced-13 messenger RNA levels were both induced in response to vincristine treatment (Supplementary Fig. 1j). Importantly, egl-1;ced-13 double mutants were completely defective for vincristine-induced apoptosis, indicating that both BH3-only domain proteins contribute to anti-tubulin-induced apoptosis (Fig. 1i and Supplementary Fig. 1h–j). Taken together, these data suggest that four clinically used chemotherapeutic drugs promote germline apoptosis in C. elegans. Even though the upstream signalling cascade might be distinct for each chemotherapeutic, in all four cases, apoptosis is dependent on the core apoptotic machinery.
Whereas the pathways that mediate DNA damage-induced apoptosis have already been intensively investigated, the molecular mechanism underlying anti-tubulin-induced apoptosis remains largely unknown. Mammalian cells treated with anti-tubulin drugs can die during mitosis (mitotic catastrophe), after delayed mitosis when entering G1 or during G0/G1 phase2,15,16. In C. elegans, progression through mitosis does not seem to be necessary for vincristine-induced apoptosis, as apoptosis can be induced in non-dividing, meiotic pachytene cells17 following vincristine treatment (Fig. 1d and Supplementary Fig. 1g).
We reasoned that microtubule-associated proteins activated by microtubule perturbation might direct signal transduction to the apoptotic machinery. To test this idea, we evaluated whether RNA-interference (RNAi)-mediated knockdown of any genes previously implicated in microtubule-based processes could alter vincristine-induced apoptosis (Fig. 2a and Supplementary Table 1). We included 37 microtubule-associated genes representative for different categories of microtubule-based processes. Knockdown of 13 genes severely affected germline integrity and were excluded from further analysis (Supplementary Table 1). To distinguish between general apoptosis genes and vincristine-specific signalling molecules, we subsequently performed a secondary screen to analyse the effect of candidate genes on ionizing-radiation- and ultraviolet-C-induced apoptosis (Fig. 2a). We uncovered with this strategy three genes—let-99, par-3 and pfd-1—that were specifically required for vincristine-induced germline apoptosis (Fig. 2a–c and Supplementary Fig. 1b and Table 1). In contrast, ekl-1, a component downstream of Ras signalling18, strongly sensitized to apoptosis also in response to ionizing radiation (Supplementary Fig. 1b).
PFD-1 is a homologue of human PFDN1, a subunit of the heterohexameric prefoldin complex that chaperones nascent actin and α- and β-tubulin19. PAR-3 and LET-99 are cell polarity genes. PAR-3 is required for polarization of the zygote and contains a PDZ domain20,21. LET-99 is a DEP domain protein that has been shown to act downstream of the polarity protein PAR-3 to determine mitotic spindle position in early C. elegans embryos22,23.
We focused our subsequent efforts on LET-99, as little is known about this protein beyond its role in early embryogenesis22,23,24,25,26,27,28. Using the ok1403 deletion allele, we confirmed that vincristine-induced apoptosis is blocked in let-99(ok1403) homozygous mutants but normal in let-99(ok1403)/+ heterozygous animals (Fig. 2b). To determine whether LET-99 was also required for apoptosis in response to other anti-tubulin chemotherapeutics, we treated let-99(ok1403) animals with nocodazole and docetaxel. We found that let-99(ok1403) mutants failed to trigger germline apoptosis to nocodazole and docetaxel suggesting that LET-99 directs signal transduction in response to various anti-tubulin chemotherapeutics (Supplementary Fig. 1c, d). To assess LET-99 expression and subcellular localization, we generated transgenic animals expressing a carboxy-terminally GFP-tagged LET-99 protein under the endogenous let-99 promoter (Plet−99::let-99::gfp::let-858(3’UTR)). Whereas LET-99::GFP was localized to the cortex in the one-cell-stage embryo as previously reported23,26, we surprisingly found nuclear localization in later-stage embryos as well as in the adult germ line, where apoptosis was assessed (Fig. 2d–g). At the 16-cell embryonic stage, cortical and nuclear staining could be observed concurrently (Fig. 2f, g and Supplementary Fig. 2). We observed similar staining patterns using an anti-LET-99 polyclonal antibody (Supplementary Fig. 2). To confirm that LET-99 acts indeed in germ cells, we also assessed germline apoptosis in rrf-1 mutant animals, which exhibit a substantial deficiency in somatic RNAi, and found that let-99(RNAi); rrf-1 animals were still defective in germline apoptosis following vincristine administration (Supplementary Fig. 1f).
We next investigated how the DEP domain protein LET-99 contributes to anti-tubulin drug-induced apoptosis. DEP domains are commonly found in regulators of heterotrimeric G proteins or small Rho superfamily GTPases and have been implicated to promote interaction with either membranes or other proteins29,30. Genetic studies suggest that LET-99 regulates cytokinesis and spindle positioning in the early C. elegans embryo by controlling a pathway that includes the Gα proteins GOA-1 and GPA-16, their regulators GPR-1/2, and the adaptor protein LIN-5 (refs 22, 23, 24, 25). However, gpr-1/2(RNAi), lin-5(RNAi) and goa-1(RNAi) showed normal levels of germ cell corpses following vincristine treatment, which argues against their involvement in vincristine-induced germline apoptosis (Supplementary Fig. 3a, b). To determine whether another heterotrimeric G protein subunit might be acting downstream of LET-99, we also individually knocked down known Gα, Gβ and Gγ subunit genes (Supplementary Fig. 3a). We reasoned that RNAi directed against a gene negatively regulated by LET-99 should lead to increased levels of vincristine-induced germline apoptosis. Knockdown of gpa-11(RNAi) but none of the other G protein subunits resulted in increased vincristine-induced apoptosis in rrf-1 mutants (deficient in somatic RNAi; Supplementary Fig. 3a). We confirmed these data using the gpa-11(pk349) deletion mutants (Fig. 3a). To order let-99 and gpa-11 within the apoptotic signalling cascade, we treated let-99(ok1403); gpa-11(RNAi) or gpa-11(pk349); let-99(RNAi) animals with vincristine and found that gpa-11 was epistatic to let-99 as increased apoptotic levels were not suppressed by let-99 loss-of-function (Fig. 3a). To summarize, these data identify a heterotrimeric G protein signalling cascade in vincristine-induced apoptosis and suggest that LET-99 acts as a negative regulator of the Gα protein GPA-11 in the adult C. elegans germ line.
As little is known about Gα signalling targets in C. elegans, we further analysed the protein sequence of the DEP domain-only protein LET-99 (Supplementary Fig. 4). The DEP domain is a module of approximately 90 amino acids first identified in Drosophila Dishevelled, C. elegans EGL-10, and mammalian Pleckstrin31. Work in Drosophila has shown that the JNK homologue Basket acts downstream of the DEP domain protein Dishevelled to promote planar cell polarity32. Given that the stress kinase JNK is a known signalling molecule involved in anti-tubulin chemotherapeutic-induced apoptosis in mammals33,34, we investigated whether C. elegans jnk-1 might also function in vincristine-induced apoptosis downstream of LET-99. We found that jnk-1(gk7) and mkk-4(km23) (upstream MAPKK) mutant animals showed higher basal apoptotic levels, but failed to increase apoptosis levels following exposure to vincristine or ultraviolet-C, but not to ionizing radiation (Fig. 3b, c and Supplementary Fig. 5a–h). Moreover, increased apoptotic levels in gpa-11(pk349) mutants were suppressed by jnk-1(RNAi), suggesting that JNK-1 functions downstream of GPA-11 in this signalling pathway (Fig. 3g). Consistent with these genetic data, we observed phosphorylation of C. elegans JNK-1 following treatment with vincristine or ultraviolet-C, but not following ionizing radiation (Fig. 3d–f and Supplementary Fig. 5b). Moreover, JNK-1 phosphorylation was dependent on LET-99 (Fig. 3d) as well as on the upstream MAPKK MKK-4, but not on JKK-1, SEK-1 and MEK-1 (Fig. 3d–f and Supplementary Fig. 5c). Collectively, these data suggest that LET-99 and GPA-11 act upstream of JNK-1 to mediate vincristine-triggered apoptosis.
Previous observation of JNK activation following vincristine treatment in mammalian cells33,34 led us to question whether the signalling pathway that we detected in C. elegans might be evolutionarily conserved. On the basis of protein structure and sequence similarity, the two closely related genes DEPDC1A and DEPDC1B possibly correspond to the human homologues of C. elegans let-99 (Supplementary Fig. 4). As for LET-99 subcellular localization in the germ line, epitope-tagged DEPDC1A was mostly nuclear in HeLa and in MCF-7 cells (Fig. 4c–f). Importantly, treatment of HeLa, MCF-7 or SH-SY5Y cells with two short interfering RNAs (siRNAs) directed against DEPDC1A or DEPDC1B reduced cell death levels in response to vincristine but not to doxorubicin (Fig. 4a, b and Supplementary Fig. 3d–f). In keeping with this, DEPDC1 knockdown attenuated the apoptosis cascade as revealed by reduced cleaved caspase-3 and cleaved PARP (D214) levels following vincristine treatment (Fig. 4g, h). These data suggest an evolutionarily conserved function of LET-99/DEPDC1 in regulating vincristine-induced apoptosis.
We next investigated whether DEPDC1 was required for vincristine-induced JNK phosphorylation in mammals. We found that following vincristine treatment, siRNA targeting either DEPDC1 reduced JNK phosphorylation, whereas total JNK levels were unchanged (Fig. 4i and Supplementary Fig. 5i). Treatment of MCF-7 and HeLa cells with the SP600125 JNK inhibitor also hampered vincristine-induced cell death, confirming the role of the stress-kinase JNK in this pathway (Fig. 4j). Combining JNK inhibition with DEPDC1 knockdown did not further reduce cell death levels (Fig. 4j). Together, these data suggest that JNK participates in vincristine-triggered apoptosis in both C. elegans and mammals and that LET-99/DEPDC1 probably acts upstream of the JNK cascade in this signalling pathway.
JNK directs MCL1 recruitment to FBW7 following anti-tubulin treatment by phosphorylating the MCL1 degron motif4,5. To examine whether DEPDC1 regulates vincristine-induced apoptosis by controlling degradation of the anti-apoptotic Bcl-2 member MCL1, we assessed the effect of DEPDC1 knockdown on MCL1 degradation. Vincristine treatment led to MCL1 downregulation within 12 h in the scramble siRNA control-treated cells (Fig. 5a and Supplementary Fig. 3c). In contrast, MCL1 downregulation was attenuated in cells treated with DEPDC1 siRNA (Fig. 5a and Supplementary Fig. 3c). MCL1 downregulation could be further inhibited when combining the SP600125 JNK inhibitor with DEPDC1 siRNA (Fig. 5b). We next evaluated whether DEPDC1A overexpression was sufficient to degrade MCL1 and to sensitize to vincristine-induced cell death. First, we found that overexpression of DEPDC1A completely degraded MCL1 levels within 3 h and induced cleavage of the caspase-3 target PARP (Fig. 5d, i). Second, using a Tet-On inducible system, we found that low-copy DEPDC1A induction sensitized human cells to cell death in response to vincristine (Fig. 5e, f). Third, whereas DEPDC1 knockdown mediated resistance to vincristine-triggered cell death (Fig. 4a, b), DEPDC1 knockdown in MCL1 siRNA cells neither reduced cell death levels nor attenuated cleaved PARP (D214) levels, indicating that DEPDC1 functions upstream of MCL1 (Figs 4h and 5g, h). Collectively, we conclude that DEPDC1 regulates vincristine-induced cell death by promoting JNK-dependent degradation of the anti-apoptotic Bcl-2 member MCL1.
Anti-tubulin drugs such as vincristine have been successfully used in the clinics for the treatment of human cancer for the past 50 years. Despite their wide use, the underlying molecular mechanisms that translate microtubule disruption into their potent anti-cancer activity are still not fully understood. From a cancer therapy perspective, knowledge of apoptotic signalling cascades in cancer cells is of prime importance and should finally help to better predict patient response to therapy6.
Herein, using C. elegans as a model, we demonstrate the efficiency of an in vivo tool to systematically screen for signalling components required for apoptosis in response to a given cancer drug. As proof of principle, we dissect the apoptotic signalling cascade following vincristine treatment and describe similarities and differences between nematode and mammalian apoptosis signalling. The fact that in C. elegans arrested pachytene germ cells undergo apoptosis raises the provocative as yet not conclusively excluded possibility that anti-tubulin drugs could also act independently of their anti-mitotic activity.
In an RNAi-based screen, we unveil a general apoptosis regulator and three components specifically involved in vincristine-induced apoptosis. We show that LET-99 acts upstream of the heterotrimeric G protein signalling cascade comprising the Gα protein GPA-11 and of the stress kinase JNK-1 to orchestrate vincristine-induced apoptosis. Importantly, we provide evidence that a human candidate homologue of LET-99, DEPDC1, similarly participates in anti-tubulin drug-induced cell death, indicating an evolutionarily conserved function of LET-99/DEPDC1 (Fig. 5c). Nevertheless, we cannot rule out that other closely related DEP domain proteins such as human DEPDC4 and DEPDC7 share also some functional homology to LET-99.
Recently, the anti-apoptotic Bcl-2 family member MCL1 has been demonstrated to control the therapeutic response to anti-tubulin drugs4,5,35. Following treatment, a coordinated action of stress-activated and mitotic kinases JNK, p38 and CKII allows binding of the SCFFBW7 ubiquitin ligase, which polyubiquitylates and targets MCL1 for proteasomal degradation. We now provide evidence that human DEPDC1 is an upstream regulator of this signalling cascade. Although we cannot exclude parallel DEPDC1 targets (Fig. 5c), our findings unveil that DEPDC1 participates in JNK activation and proteasomal degradation of MCL1 in response to anti-tubulin drug treatment. It will be interesting to see whether DEDPC1 protein levels may help to predict patient response to anti-tubulin therapy upstream of FBW7 and MCL1. □
All strains were maintained and raised at 20 °C on NGM agar seeded with Escherichia coli OP50 (ref. 36). The following mutations and transgenes were used in this study: cep-1(gk138), egl-1(n1084n3082), ced-13(gk260), ced-9(n1950), ced-3(n717), jnk-1(gk7), let-99(ok1403) IV/nT1[qIs51](IV;V), par-3(e2074), sek-1(km4), mek-1(ks54), jkk-1(km2), mkk-4(km23), rrf-1(pk1417), gpa-11(pk349), opIs417 [Plet−99::let-99::gfp::let-858(3’UTR)], opIs418 [Plet−99::let-99::gfp::let-858(3’UTR)].
Low-copy transgenic lines were created by microparticle bombardment as previously described37, using a Biorad PDS-1000/He Biolistic Delivery System.
Differential interference contrast microscopy.
Worms were placed on 3% agarose pads and anaesthetized in 10 μl M9 with 5 mM levamisole (Sigma) and mounted under a coverslip for observation using a Leica DM-RA microscope equipped with differential interference contrast microscopy (Nomarski) optics.
Chemotherapeutic drug and DNA damage response assay.
Synchronized young adult worms (12 h post L4/adult moult stage) were exposed to indicated dosages of cisplatin (Sigma), doxorubicin (Sigma), etoposide (Sigma), vincristine (Sigma), X-rays or ultraviolet-C (emission peak 254 nm). An Isovolt 160 HS X-ray machine (Rich. Seifert & Co.) or Stratalinker ultraviolet crosslinker, model 1800 (Stratagene) was used to deliver the appropriate ionizing radiation and ultraviolet-C dosages. For drug treatments, 1.2 ml of the indicated drug concentrations in M9 buffer were added directly to 90 mm plates containing staged worms and OP50. Plates were allowed to dry in a hood (approximately 10 min) and then incubated at 20 °C. Germline apoptosis was quantified at the indicated time points using differential interference contrast microscopy, as previously described8. Minimal concentrations, which robustly induced germline apoptosis were chosen for subsequent studies. For RNAi experiments, synchronized L1 larvae were transferred onto plates seeded with bacteria expressing the respective RNAi clone38. Germline apoptosis was quantified as described above, starting from the 12 h post L4/adult moult stage.
Relative quantification of transcripts.
Synchronized young adults were collected for total RNA extraction followed by cDNA synthesis (SuperScript III Platinum) and quantitative real-time PCR. Transcript levels were normalized to tbp-1 and pgk-1.
The forward and reverse primer sequences were: tbp-1: 5′-TTGGATTTGAAGAAGATTGCATTG-3′, 5′-AATGACTGCTGCGAAACGTTT-3′. pgk-1: 5′-GCGATATTTATGTCAATGATGCTTTC-3′, 5′-TGAGTGCTCGACTCCAACCA-3′ egl-1: 5′-CAGGACTTCTCCTCGTGTGAAGATTC-3′, 5′-CGAAGTCATCGCACATTGCTGCTA-3′. ced-13: 5′-ACGGTGTTTGAGTTGCAAGC-3′, 5′-GCCAATATTATATTCAACCGTGTTTGAGT-3′.
Human DEPDC1 transcript levels: DEPDC1A: 5′-CTATGGAGAGTCAGGGTGTGC-3′, 5′-CGAAAAGATGTGGTAACTTCATTC-3′, DEPDC1B tr 1: 5′-GGCTCATCTACGAAGAGTCCA-3′, 5′-GGCAAAATGATGGAGCAGATA-3′, DEPDC1B tr 2: 5′-TCTTTTTGATGCTTTTGTCAGTG-3′, 5′-AGCTGTAACTTTCTCCTATTTTCAGG-3′, GAPDH: 5′-AGCCACATCGCTCAGACAC-3′, 5′-GCCCAATACGACCAAATCC-3′.
Cell cultures and viral preparation.
HEK 293T cells were cultured in IMDM (Iscove’s modified Dulbecco medium, SIGMA) supplemented with 10% dialysed fetal calf serum (Invitrogen-Gibco), 2 mM Glutamax, 100 U ml−1 penicillin/streptomycin. Viral particles were prepared as described previously39. HeLa cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, 2 mM glutamine and 100 U ml−1 penicillin/streptomycin (Invitrogen) and maintained at 37 °C in a 5% CO2 atmosphere. Calcium phosphate was used to transfect HeLa cells with mammalian expression vectors, whereas Lipofectamine-2000 (Invitrogen) was used to transfect double-stranded oligonucleotides (Ambion) designed against DEPDC1A and DEPDC1B. Scrambled siRNA was used as a control. Analyses were performed 48 h after the siRNA transfection. To induce cell death, HeLa, MCF-7 and SH-SY5Ycells were treated with 25 nM vincristine for 6, 12 or 24 h as indicated. The medium was replaced and cells cultured for the indicated time. Cell death was quantified using nuclear condensation.
The following siRNAs were used (5′ to 3′): DEPDC1A s31138: 5′-GUGGAGUAGUUAUACUACATT-3′, DEPDC1A s31137: 5′-GUCCUGAAGUUACAAGGCATT-3′, DEPDC1B s31516: 5′-GGAUAUCACUUUAUCUGCUTT-3′, DEPDC1B s31515: 5′-CUAUGAGCAUUGUUUCACATT-3′,
After the indicated experimental treatment, cells were rinsed with PBS, fixed for 20 min with 4% paraformaldehyde and permeabilized 5 min by Triton X-100. Samples were then blocked with goat serum/1% BSA/0.01% Tween-20 and incubated for 1 h or overnight with primary antibodies, washed with blocking solution, incubated with the appropriate secondary antibody for 1 h, washed again, stained with Hoechst-33342, washed and then mounted with 5% n-propylgallate in absolute glycerol.
SDS gel electrophoresis and immunoblotting.
Cells were collected in PBS and re-suspended in buffer (10 mM Tris-HCl pH 7.4, 1 mM EDTA, 0.1% Triton X-100, proteases inhibitor) followed by protein quantification with 660 nm protein assay (Pierce). Loading buffer (3.3× concentrate: 16.7% glycerol, 5% SDS, 1.3% Tris-HCl at pH 6.8, bromophenol) was added in ratio 1:3 to the samples. Proteins were heated up to 95 °C for 10 min, separated by electrophoresis and transferred onto PVDF membrane using a wet blot chamber (Bio-Rad). The membranes were saturated in PBS with 5% milk powder or 5% BSA as blocking solution at room temperature for 2 h. Primary and secondary antibodies were incubated for 2 h at room temperature in blocking solution. Worms were directly lysed in Laemmli buffer. Proteins were heated up to 95 °C for 10 min followed by protein quantification with 660 nm protein assay (Pierce). The following antibodies were used: LET-99 (provided by L. Rose, University of California, Davis, 1:5,000), MCL1 (Abcam, Ab32087, 1:1,000), phospho-JNK (Cell Signaling 9251S, 1:5,000), total JNK (56G8 Cell Signaling 9258S, 1:1,000), Myc (Abcam ab9106 and 9B11 Cell Signaling 2276S, both 1:1,000), Cleaved caspase-3 (5A1E Cell Signaling 9664S, 1:1,000), Cleaved PARP (D64E10 Cell Signaling 5625P, 1:1,000), GAPDH (Santa Cruz FL-335, 1:5,000), anti-HA.11 (16B12 Covance MMS101-P, 1:1,000), Actin (AC-74 Sigma A5316, 1:5,000), Tubulin (DM1A Sigma T9026, 1:5,000).
We thank Hengartner and Bano laboratory members for help and discussions. We thank L. Rose for the LET-99 antibody. We are grateful to L. Rose and M. Gotta for critical reading of the manuscript. This work was supported by the Swiss National Science Foundation, the Kanton of Zurich and the Josef-Steiner Foundation. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR).
Integrated supplementary information
About this article
Journal of Neuro-Oncology (2017)