Mutations in the kinase domain of ALK (anaplastic lymphoma kinase) have recently been shown to be important for the progression of the childhood tumour neuroblastoma. In the present study we investigate six of the putative reported constitutively active ALK mutations, in positions G1128A, I1171N, F1174L, R1192P, F1245C and R1275Q. Our analyses were performed in cell-culture-based systems with both mouse and human ALK mutant variants and subsequently in a Drosophila melanogaster model system. Our investigation addressed the transforming potential of the putative gain-of-function ALK mutations as well as their signalling potential and the ability of two ATP-competitive inhibitors, Crizotinib (PF-02341066) and NVP-TAE684, to abrogate the activity of ALK. The results of the present study indicate that all mutations tested are of an activating nature and thus are implicated in tumour initiation or progression of neuroblastoma. Importantly for neuroblastoma patients, all ALK mutations used in the present study can be blocked by the inhibitors, although some mutants exhibited higher levels of drug sensitivity than others.
Neuroblastoma is a cancer derived from neural crest cells of the sympathetic nervous system, accounting for approximately 15% of all childhood cancer . Neuroblastoma tumours show distinct biological and clinical features. A subset may spontaneously differentiate or regress with little or no therapy at all, while the majority are difficult to cure with current modalities. The gene locus of the RTK (receptor tyrosine kinase) ALK (anaplastic lymphoma kinase) has been reported to be amplified in patient samples and neuroblastoma cell lines [2–4]. More recently, germ line and somatically acquired activating point mutations of the RTK ALK have been identified both in neuroblastoma patient samples and neuroblastoma cell lines [5–9]. To date, from genetic analysis of neuroblastoma patients, 13 amino acids have been published as mutated close to or within the kinase domain of the ALK RTK.
ALK is a member of the insulin receptor superfamily of RTKs [10,11] and has been identified as a fusion partner for NPM (nucleophosmin), resulting in the oncogene NPM–ALK which has been found in a subset of ALCL (anaplastic large cell lymphoma) and other cancers (reviewed in ). So far aberrant ALK activity has been involved in the progression and maintenance of a great variety of solid and haemopoietic tumours. To date, no clinically approved treatments targeting ALK for patients with neuroblastoma are available. However, the development of c-Met/ALK-specific inhibitors such as Crizotinib (PF-02341066) , which is in different Phase II and III clinical trials (http://www.ClinicalTrials.gov identifiers NCT00932893, NCT01154140 and NCT01121588) and has recently been reported to have robust clinical effects in both NSCLC (non-small-cell-lung cancer) and inflammatory myofibroblastic tumours [14,15], provides real hope for neuroblastoma patients harbouring ALK mutations [16,17].
The aim of the present study was to investigate whether certain recently described mutations found in neuroblastoma patients are truly gain-of-function mutations and if they have the potential to be involved in disease progression. Our analyses indicate that the ALK point mutations described are constitutively active, leading to activation of important oncogenic downstream signalling in cell culture model systems which can be inhibited by ALK-specific inhibitors such as Crizotinib (PF-02341066) and NVP-TAE684. Furthermore, these mutations have transforming potential and overexpression in the Drosophila melanogaster eye results in a rough eye phenotype, confirming that they are ALK RTK ligand-independent gain-of-function mutations. The results of the present study also indicate that the various ALK mutants vary in their degree of drug sensitivity in both cell systems and in the Drosophila system. Taken together, the results of the present study suggest that ALK gain-of-function mutations play an important role in the development and progression of neuroblastoma, which may be inhibited by ALK-specific inhibitors such as Crizotinib.
MATERIALS AND METHODS
Antibodies and inhibitors
The following antibodies were used: anti-pan-ERK (extracellular-signal-regulated kinase) (1:5000 dilution) was purchased from BD Transduction Laboratories and the anti-phospho-ALK (Tyr1278), anti-phospho-ERK, anti-phospho-STAT3 (signal transducer and activator of transcription 3) (Tyr705), anti-phospho-Akt (Ser473) and anti-phospho-Akt (Thr308) (1:1000 dilution) were from Cell Signaling Technology. The activating mAbs (monoclonal antibodies) mAb46 and mAb31 have been described previously [18,19].The anti-phosphotyrosine antibody 4G10 was from Upstate Biotechnology. The HRP (horseradish peroxidase)-coupled secondary antibodies goat anti-rabbit IgG and goat anti-mouse IgG were from ThermoScientific. The ALK-specific inhibitor NVP-TAE684 has been described previously [20,21] and Crizotinib (PF-02341066)  was obtained from Pfizer.
Generation of human and mouse ALK mutant constructs in PC12 cells
The wt (wild-type) mouse ALK sequence (GenBank® accession number NM_007439.2) was PCR-amplified using the following primers: 5′-GCAAAGCCGGACTGTCTGGCATGTCGCCAC-3′ and 5′-GTTGGGCTGAGAGAAAGCCATGTTCACGTG-3′. The PCR products were subcloned into pCRIITOPO following the manufacturer's protocol (Invitrogen), and the resulting plasmid was used as a template for generation of the mouse point mutations, using the QuikChange® site-directed mutagenesis kit (Stratagene), according to the manufacturer's instructions, with the primers: G1132A (3395 G→C; 5′CCATGGCGCATTTGCGGAGGTGTATG-3′ and 5′-CATACACCTCCGCAAATGCGCCATGG-3′), I1175N (3524 T→A; 5′-CATGGAAGCTCTGATTAATAGCAAATTCAACCACC-3′ and 5′-GGTGGTTGAATTTGCTATTAATCAGAGCTTCCATG-3′), F1178L (3534 C→A; 5′-CTGATCATCAGCAAATTAAACCACCAG-3′ and 5′-CTGGTGGTTTAATTTGCTGATGATCAG-3′), R1196P (3587 G→C; 5′-CTACAAGCCCTGCCCCCCTTCATCCTGCTGGAAC-3′ and 5′-TTCCAGCAGGATGAAGGGGGGCAGGGCTTGTAG-3′), F1249C (3746 T→G; 5′-GAGGAGAATCACTGTATCCACCGGG-3′ and 5′-CCCGGTGGATACAGTGATTCTCCTC-3′) and R1279Q (3836 G→A; 5′-CTTTGGGATGGCCCAAGATATCTACAGGG-3′ and 5′-CCCTGTAGATATCTTGGGCCATCCCAAAG-3′) (underlined bases were changed by mutagenesis). The mutations generated were confirmed by sequencing from both directions. The mALK (mouse ALK) fragment containing the point mutation was digested by AflII/NgoMIV and ligated into the opened pTTPmALK vector (described in [21,23]), resulting in the respective pTTPmALK mutant plasmids (pTTP-ALKmut). Stable PC12 Tet-on clones expressing mouse pTTP-ALKmut were generated as described previously  and clones were selected and grown as described in .
A similar approach was used for the generation of mutations of hALK (human ALK). The pcDNA3 expression vector, containing the cDNA for wt human ALK (pcDNA3-hALKwt) , was used as a template for creating a 1298 bp fragment using PCR with the primers 5′-TTCTCCGGCATCATGATTGTGTA-3′ and 5′-TTGGACTGAGAGAATGCCATATT-3′ and cloned into the pCRII-TOPO vector, according to the manufacturer's protocol (Invitrogen), and the resulting plasmid was used as a template for generation of the human point mutations using the QuikChange® site-directed mutagenesis kit (Stratagene), according to manufacturer's instruction. The mutations generated and the primers used were G1128A (3383 G→C; 5′-CCATGGCGCCTTTGCGGAGGTGTATGAAG-3′ and 5′-CTTCATACACCTCCGCAAAGGCGCCATGG-3′), I1171N (3512 T→A; 5′-CATGGAAGCCCTGATCAACAGCAAATTCAACCACC-3′ and 5′-GGTGGTTGAATTTGCTGTTGATCAGGGCTTCCATG-3′), F1174L (3522 C→A; 5′-GGAAGCCCTGATCATCAGCAAATTAAACCACCAGAACA-3′ and 5′-TGTTCTGGTGGTTTAATTTGCTGATGATCAGGGCTTCC-3′), R1192P (3575 G→C; 5′-CAATCCCTGCCCCCGTTCATCCTGCTG-3′ and 5′-CAGCAGGATGAACGGGGGCAGGGATTG-3′), F1245C (3734 T→G; 5′-TTTGGAGGAAAACCACTGCATCCACCGAGACATTG-3′ and 5′-CAATGTCTCGGTGGATGCAGTGGTTTTCCTCCAAA-3′) and R1275Q (3824 G→A; 5′-CTTCGGGATGGCCCAAGACATCTACAGGG-3′ and 5′-CCCTGTAGATGTCTTGGGCCATCCCGAAG-3′) (underlined bases were changed by mutagenesis). The seven different mutated pTOPO:1298 bp plasmids were digested with BlpI/FseI and the resulting 1177 bp fragments were cloned into the BlpI/FseI site of the pcDNA3:hALKwt vector described above. All mutant constructs were verified by sequencing.
Cell lysis, immunoprecpitation and Western blot analysis
Briefly, PC12mALKmut cells were induced with doxycycline and serum-starved for 20 h. PC12mALKwt cells were additionally stimulated with 1 μg/ml of the activating mAb46 for 30 min [19,21,23]. IL (interleukin)-3-independent hALK-expressing Ba/F3 cells were treated for 3 h with Crizotinib in complete medium. Cells were washed and lysed in SDS sample buffer. Precleared cell lysates were analysed by SDS/PAGE (7.5% gel), followed by immunoblotting with the antibodies indicated. ALK downstream activation was detected by phospho-ERK, phospho-STAT3 (Tyr705) and phospho-Akt (Ser473 and Thr308), and pan-ERK was used to show equal loading. ALK phosphorylation was detected by the 4G10 and the phospho-ALK (Tyr1278) antibodies respectively. The bands for phospho-ALK and total ALK were densitometrically quantified with Chemidoc (Bio-Rad) and the level of relative ALK phosphorylation as an indicator for ALK activity was calculated. Cell lysis, immunoprecipitation and immunoblotting were performed according to the protocols described in .
Neurite outgrowth assay
PC12mALKwt and PC12mALKmut cells were seeded sparsely in six-well plates and ALK expression was induced by doxycycline in the presence or absence of the ALK-specific inhibitor NVP-TAE684 (30 and 100 nM). PC12mALKwt cells were stimulated with 1 μg/ml mAb46. Quantification of neurite outgrowth in the cells was carried out as described previously . Experiments were performed in triplicate, and each sample within an experiment was performed in duplicate. For hALK, 2×106 PC12 cells were transfected by electroporation in an Amaxa electroporator using 1.5 μg of pCDNA3-hALK, 0.5 μg of pEGFPN1 (Clontech) and 100 μl of Ingenio electroporation solution (Mirus Bio LCC). After transfection cells were transferred to DMEM (Dulbecco's modified Eagle's medium) supplemented with 7% horse serum and 3% FBS (fetal bovine serum), thereafter seeded into 24-well plates together with 1 μg/ml mAb31 . At 2 days after transfection, the fraction of GFP (green fluorescent protein)-positive and neurite-carrying cells compared with GFP-positive cells was estimated under a Zeiss Axiovert 40 CFL microscope. To be judged as a neurite-carrying cell, the neurites of the cell had to reach at least twice the length of the diameter of a normal cell body.
Cell proliferation assay and IC50 determination
Ba/F3 cell lines expressing hALKwt and hALKmut were generated by electroporating Ba/F3 cells with pCDNA3-hALK using an Amaxa electroporator. The transfected cells were then selected in RPMI medium with 10% heat-inactivated FBS, 2.5 ng/ml IL-3 (Peprotech) and 600 μg/ml G418 for at least 10 days. Cells were washed with PBS and seeded at 0.5×106 cells/ml in RPMI with 10% FBS and G418 for the generation of hALK-expressing IL-3-independent cells. In order to determine the IC50 of Crizotinib, IL-3-independent hALK-expressing Ba/F3 cells were treated with the indicated concentrations of Crizotinib for 3 days before cell viability was analysed with 55 μM Resazurin (Sigma) . The IC50 was calculated for the individual cell line and the experiment was carried out at least three times independently in triplicate.
Low passage number NIH 3T3 cells (A.T.C.C.) were transfected with Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's protocol. Briefly, 4.5×104 cells, seeded the day before into collagen-coated 12-well plates, were transfected for 6 h with 0.55 μg of pcDNA3-hALK and 1.4 μl of Lipofectamine™ 2000 in 0.3 ml of Opti-MEM. At 24 h after transfection, three-fifths of the cells from each well were transferred to wells in 12-well plates and kept in DMEM (10% FBS and 0.5 mg/ml G418 until the cells reached confluence). Thereafter cells were kept in DMEM (5% FCS and 0.25 mg/ml G418) for a further 10 days.
Generation of hALK mutant transgenic constructs for D. melanogaster
Prior to ligation of the hALK, ALKF1174L and the ALKR1275Q from pcDNA3 into the Drosophila expression vector pUAST, an 898 bp fragment preceding the translation start site was removed to increase expression efficiency in the Drosophila system (details available upon request from the authors). All three constructs were subsequently subcloned into the EcoRI-NotI site of the pUAST Drosophila vector and the resulting constructs were verified by DNA sequencing analysis. Transgenic constructs were used for the generation of transgenic fly strains (BestGene). The following stocks were used: w1118, (Bloomington, stock number 5905) and pGMR-Gal4 (Bloomington, stock number 9146). The transgenic fly strains UAS-ALKwt, UAS-ALKF1174L and UAS-ALKR1275Q were generated as described above. Expression in the eye, immunoblotting, and fluorescence and electron microscopy of wt and mutant ALK proteins was carried out as described in .
Neuroblastoma ALK mutations are ligand-independent and activate downstream targets
Our initial investigations focused on the activity of six neuroblastoma mutations which were changed at the equivalent sites in the mALK RTK, creating mALKG1132A, mALKI1175N, mALKF1178L, mALKR1196P, mALKF1249C and mALKR1279Q [5–9]. Mutations at positions 1174 and 1275 in hALK are the two most frequently found mutations in neuroblastoma patients and predicted to be of an activating nature [6,7,9,25]. The other mutants selected for the present study have been envisaged to be of oncogenic nature, but have not been examined in detail [9,26].
We used an inducible PC12 cell culture system for the clonal expression of mALK mutants in doxycycline-inducible vectors allowing us to compare the activity of these mutants in a controlled manner. PC12 cells were transfected and individual clones were selected for expression of specific mutants. As a control, wt mouse ALK RTK was used for comparison, which becomes tyrosine phosphorylated and activates/phosphorylates downstream targets, such as ERK and PKB (protein kinase B)/Akt upon stimulation with an agonist mAb, mAb46 (Figure 1, lanes 13–15). Doxycycline-induced expression of ALK mutants results in a similar ligand-independent tyrosine phosphorylation of the receptors themselves (Figures 1A and 1B, lanes 1–12). Furthermore, the relative ALK kinase activity indicated by phospho-ALK/ALK seemed to differ only slightly between the different ALK mutants, although some such as F1178L, R1196P and G1132A appeared slightly more active (Figure 1H). This was accompanied by phosphorylation of ERK and PKB/Akt (Figures 1D–1F, lanes 1–12). In addition to activation of ERK and PKB/Akt, all six mALK mutants investigated also led to the phosphorylation of STAT3 (Figure 1C). This was in contrast with mAb46-stimulated wt mALK which did not result in STAT3 activation 30 min post-stimulation (Figure 1C, compare lanes 1–12 with lanes 14 and 15). Thus it appears that all six single point mutations in the kinase domain of mALK that were expressed (mALKG1132A, mALKI1175N, mALKF1178L, mALKR1196P, mALKF1249C and mALKR1279Q) are activated in a ligand-independent manner and activate downstream targets, such as STAT3, ERK and PKB/Akt.
mALK activating point mutations are constitutively active as indicated by stimulation of different signalling pathways in PC12 Tet-on cells
Both NVP-TAE684 and Crizotinib (PF-2341066) abrogate neurite outgrowth activity of constitutively active ALK mutants
As reported previously, expression and stimulation of the wt mALK mediates neurite outgrowth in PC12 cells, which is a sensitive readout for ALK signalling activity as a result of receptor activation [21,23,27,28]. This can be completely blocked by the simultaneous addition of the ALK-specific ATP-competitive inhibitor NVP-TAE684 at a final concentration of 30 nM (Figure 2A). A small number of neurites were also observed in induced, but unstimulated, PC12 cells expressing wt mALK, which may reflect leakage of the wt mALK clone, or the presence of an unknown endogenous ligand. All six mALK mutants generated were capable of stimulating neurite outgrowth in a ligand-independent manner. Inhibition of ALK signalling upon addition of inhibitor was also observed in all six ALK mutant receptors examined (Figure 2A). In uninduced mutant clones a background of neurite outgrowth due to leakage varied between 5 and 35%. However, upon induction and expression of the mutant ALK all cells had extended neurites (Figure 2A). Interestingly, treatment with NVP-TAE684 resulted in different responses, dependent upon the particular mutant examined. Neurite outgrowth induced by expression of the mALKR1279Q mutant was completely inhibited at 30 nM NVP-TAE684, which is comparable with the response seen with activated wt mALK (Figure 2A). This is in contrast with NVP-TAE684 treatment of mALKI1175N, mALKR1196P and mALKF1249C which were only inhibited to 40–50% levels of neurite outgrowth. Treatment of clones expressing mALKG1132A and mALKF1178L with 30 nM NVP-TAE684 resulted in only a small reduction in neurite outgrowth, in agreement with the higher intrinsic activity suggested in Figure 1(H). Increasing the concentration of NVP-TAE684 to 100 nM resulted in a complete abrogation of neurite outgrowth, although this was accompanied by significant cell death (results not shown). Thus, although a reduction in neurite outgrowth can be observed upon treatment with NVP-TAE684 in PC12 cell clones expressing different ALK gain-of-function mutants, the response observed was variable. This may reflect differences in expression levels; however, the mALKR1279Q mutant clones express the mutant RTK well (Figure 1, lane 12). Therefore it is possible that the ability of NVP-TAE684 to inhibit ALK activity differentially varies between the various mutant ALK RTKs.
Crizotinib and NVP-TAE684 inhibit ALK-induced neurite formation in PC12 cells
Next we wanted to confirm that activating hALK mutations are sensitive to inhibition by ALK-specific inhibitors. Although mALK and hALK are very similar, with 87% overall homology at the protein level and within the kinase domain, they differ in very few amino acids: one major difference between mALK and hALK is at Tyr1604, which is lacking in the mALK protein and has been implicated in tumour progression . In the present study we used the ALK inhibitor Crizotinib (PF-2341066), which like NVP-TAE684 binds to the ATP pocket of ALK  and shows promise in the clinic against ALK-positive tumours [14,15]. Transient transfection of hALK in the absence of activating antibodies mediates approximately 10% of neurite outgrowth. Transfected and stimulated wt hALK receptor induced neurites in approximately 60% of all transfected cells (Figure 2B), which is in agreement with previous reports [18,21,28]. Including Crizotinib (at either 0.25 μM or 0.5 μM) or NVP-TAE684 (0.05 μM) upon transfection of wt ALK reduces the induction of neurites to 15 and 7% respectively. Moreover, transient transfection of all hALK-habouring specific mutations mediated a robust neurite outgrowth similar to that observed with stimulated wt ALK. Addition of Crizotinib and NVP-TAE684 abrogated the neurite outgrowth induced by hALK mutants in a similar manner as the stimulated wt receptor. It should be noted that PC12 cell clones expressing various mALK mutants tolerated 250 nM Crizotinib without affecting the growth rate (results not shown). Thus, in agreement with our experiments with the mALK mutants, the hALK gain-of-function mutations investigated in the present study respond effectively to the inhibitors Crizotinib and NVP-TAE684. Furthermore, the results indicate that the various ALK mutants respond differently to the inhibitors.
Crizotinib blocks both wt ALK and constitutively active ALK variants with different IC50 values in Ba/F3 cells
As NVP-TAE684 seems to have toxic effects over time, we used Crizotinib for further studies, since it does not display obvious toxicity. Crizotinib is currently in different Phase II and III clinical trials  (http://www.ClinicalTrials.gov). In order to assess the inhibition of the ALK mutants by Crizotinib, we used the Ba/F3 system . Ba/F3 cells were transfected with wt hALK and six different ALK mutants prior to selection with G418 in the presence of IL-3. After 10 days of selection and expansion, the cells were washed and seeded in IL-3-free medium with G418. The IL-3-independent cell lines obtained were subsequently analysed for ALK expression, proliferation and inhibition of ALK activity by Crizotinib. First, we confirmed that the transfected and G418-selected Ba/F3 cells indeed expressed the different ALK mutants in the presence of IL-3 (Supplementary Figure S1A at http://www.BiochemJ.org/bj/440/bj4400405add.htm). Interestingly, Ba/F3 cells expressing hALK mutants displayed a variable ability to mediate proliferation in the absence of IL-3 (Supplementary Figure 1B). For each hALK mutant we performed three independent transfections and from each transfection we isolated four IL-3-independent lines: ALKF1174L, ALKR1192P and ALKR1245C all gave rise to four IL-3-independent cell lines, whereas ALKG1128A only gave rise to three cell lines. In contrast, ALKI1171N and ALKR1275Q, although expressing ALK (Supplementary Figure S1A), were unable to substitute for IL-3 and drive proliferation in the Ba/F3 cells. Secondly, in order to assess the sensitivity of the hALK mutants to Crizotinib in Ba/F3 cells, we treated cells with increasing doses of Crizotinib. Proliferation of all four hALK mutants analysed was inhibited by Crizotinib (Figure 3A). Subsequent IC50 calculations of this inhibitor demonstrated that ALKF1174L and ALKR1192P required a significantly higher dose of Crizotinib as compared with ALKG1128A and ALKR1245C (Figure 3B). Thirdly, the inhibition of ALK activity was further investigated at the level of ALK tyrosine phosphorylation at Tyr1278, which is the first tyrosine residue of the YxxxYY motif that is necessary for the auto-activation of the ALK kinase domain and the transformation ability of NPM–ALK . Immunoblot analysis of Ba/F3 cell lysates demonstrated that Crizotinib was able to effectively reduce levels of ALK Tyr1278 phosphorylation (Figure 3C). Furthermore, ERK phosphorylation as an indicator of ALK-mediated downstream signalling was also inhibited by Crizotinib, reflecting the calculated IC50. It should be noted that Ba/F3 cells expressing hALK mutants tolerate the Crizotinib concentrations used without affecting cell viability when cells are grown in the presence of IL-3 (Supplementary Figure 1C). Two of the ALK mutations, I1171N and R1275Q, were unable to drive proliferation in Ba/F3 cells, which is contrary to previous studies . However, this may reflect our use of transient transfection protocols compared with retroviral infection employed by George et al. . Taken together, these results show that the activity of all of the ALK mutants tested can be inhibited by Crizotinib, although different doses are required for the various ALK mutants.
Crizotinib inhibits both hALKmut-mediated proliferation and phosphorylation at position Tyr1278 of ALK in Ba/F3 cells
Transformation analysis of hALK mutations
Given the results described above indicating differences in the signalling output of the different ALK mutants, together with the observation of differential sensitivity to the ALK inhibitors Crizotinib and NVP-TAE684, we decided to investigate this further. Next we examined whether activating hALK mutations (hALKG1128A, hALKI1171N, hALKF1174L, hALKR1192P, hALKF1245C and hALKR1275Q) were able to exhibit transforming potential in NIH 3T3 cells. Expression of all of the hALK mutations tested resulted in foci formation over the background monolayer (Figure 4). In contrast, neither vector control nor the wt hALK displayed foci formation. However, a dramatic difference in the ability of the various ALK mutant receptors to drive foci formation can be observed. The mutations hALKG1128A, hALKI1171N, hALKR1192P and hALKR1275Q showed rather weak foci formation when compared with hALKF1174L and hALKF1245C, which showed a robust level of foci formation. Thus all ALK mutations have the ability to transform NIH 3T3 cells, but to significantly different degrees, which is comparable with our results from the Ba/F3 cell experiments.
Transformation activity mediated by mutated hALK
Ectopic expression of hALK mutants in D. melanogaster
Transgenic Drosophila expressing two mutations, hALKF1174L and hALKR1275Q corresponding to the most common mutations found in neuroblastoma patients , were generated. These were then ectopically expressed in the Drosophila eye using the pGMR-Gal4 driver line, which directs protein expression in the developing photoreceptors of the eye. Expression of the wt hALK did not result in any obvious phenotype in adult flies and was similar to controls (Figure 5A) . Expression of hALK protein was confirmed by immunoblot analysis of developing eye discs where the protein level expressed in the eye was essentially equal among the mutants (Figure 5B). Expression of hALKF1174L and hALKR1275Q resulted in a rough eye phenotype, reflecting their robust ligand-independent activation in vivo (Figure 5A). The regular hexagonal arrangement of ommatidia, which can be seen in the control (w1118), was disorganized in both mutants, and many interommatidial bristles were also missing (Figure 5A). Similar disorganization can also be seen in developing eye discs (results not shown). Although both ALKF1174L and ALKR1275Q displayed a robust phenotype, a more severe phenotype was observed with ALKF1174L, in agreement with our previous observations in cell culture. Furthermore, treatment with NVP-TAE684 improved the rough eye phenotype of both ALK mutants, especially that seen with ALKR1275Q, whereas Crizotinib had little effect on either phenotype, even at 100 μM (Figure 6). Thus these differential responses of both mutants to the inhibitors are also in good agreement with the cell culture experiments described above.
ALK-activating mutations result in a rough eye phenotype in Drosophila
Effects of ALK inhibitors on the ALK activating mutant rough eye phenotype
The finding of ALK point mutations in neuroblastoma patient samples and neuroblastoma cell lines highlights a function for ALK in the development and the onset of neuroblastoma. Initial studies have reported the constitutive activity of some of the ALK point mutations where some mutants (F1174L and K1062M) are able to form subcutaneous tumours in nude mice . Furthermore, knockdown of ALK in neuroblastoma cell lines harbouring activating ALK point mutations inhibited cell proliferation . The present study provides further evidence for the activating nature of the ALK mutations. Six ALK mutations with the highest predictions of being oncogenic were analysed for the most important oncogenic downstream signalling pathways . Indeed, both mALK and hALK expression in PC12 cell clones led to ligand-independent phosphorylation of ERK, PKB/Akt and STAT3 (Figure 1 and results not shown). However, some differences in signalling can be observed. For instance the mALKI1175N and the mALKF1249C mutations seem to phosphorylate STAT3 to a lower extent than the mALKG1132A mutation with comparable ALK expression levels. Stimulated wt mALK does not mediate efficient phosphorylation of STAT3. However, a weak phosphorylation of STAT3 can be monitored after 24 h of stimulation of wt ALK receptor, although this is to a much significantly lower degree when compared with ALK gain-of-function mutations (results not shown). However, despite different ALK expression between the cell lines, no great differences between the relative ALK kinase activities can be observed, which might vary between cell clones. Furthermore, the finding that the neurite outgrowth of PC12 cell clones expressing mutated mALK are inhibited differently for the various mutants by NVP-TAE684 and Crizotinib suggests that the different mutations might have a different impact on the development, onset and severity of neuroblastoma.
Interestingly, in the NIH 3T3 transformation assay the mutations hALKG1128A, hALKR1192P and hALKR1275Q display a weaker ability to induce transformation. These mutations were initially identified as ALK germ line mutations . A similar trend of inhibition was observed in the neurite outgrowth assay, the mutations hALKG1128A, hALKR1192P and hALKR1275Q show less neurite outgrowth when inhibitors, such as NVP-TAE684 and Crizotinib, were included. All other mutations tested, such as hALKI1171N, hALKF1174L, hALKF1245C and hALKR1275Q, were identified as somatic mutations, although hALKR1275Q has been found both in germ-line tumour DNA samples and in somatic tumour DNA samples [7–9].
In order to assess the different responses of the human ALK mutants toward Crizotinib we created Ba/F3 cell lines expressing the various ALK mutants. Indeed, Ba/F3 cells expressing ALK mutants show sensitivity towards Crizotinib, resulting in inhibition of proliferation as well as phosphorylation at Tyr1278. Tyr1278 is the first tyrosine residue of the YxxxYY motif that is necessary for the auto-activation of the ALK kinase domain and the transformation activity of NPM–ALK, and is probably directly involved in the activation of downstream signalling pathways . However, the mutations hALKI1171N and hALKR1275Q, despite expression, were unable to substitute for IL-3 to drive proliferation in Ba/F3 cells. These mutations were able to mediate neurite outgrowth, transform NIH 3T3 cells and act as ligand-independent receptors in cell signalling experiments, suggesting that they are truly gain-of-function mutations which can be inhibited by Crizotinib and NVP-TAE684. The fact that the hALKI1171N and hALKR1275Q mutants drive transformation in NIH 3T3 cells, but not IL-3-independent proliferation in Ba/F3 cells, may reflect that these ALK mutants only display transformation potential in cells that are immortalized by other critical genetic events in neuroblastoma. We could further observe that the hALKR1275Q mutation in the different model systems employed, such as cell culture and D. melanogaster, seems to be most efficiently inhibited by Crizotinib and NVP-TAE684, suggesting promise for cancer treatment with ALK-specific inhibitors for neuroblastoma patients harbouring this mutation.
The ATP-competitive inhibitor Crizotinib (PF-02341066) shows dual specificity towards c-Met and ALK [13,22]. In order to test the specificity of Crizotinib and the other ALK-specific inhibitor NVP-TAE684 in our PC12 cell background, we treated PC12 cells with inhibitors prior to stimulation with NGF (nerve growth factor) and EGF (epidermal growth factor) respectively. Indeed, NVP-TAE684 does not inhibit NGF- or EGF-mediated signalling in PC12 cells, confirming the specificity toward ALK (Supplementary Figure S2A at http://www.BiochemJ.org/bj/440/bj4400405add.htm). However, Crizotinib inhibits NGF-, but not EGF-, mediated signalling in PC12 cells (Supplementary Figure S2). This result confirms the report by Zou et al. , where they investigated the cellular activity of Crizotinib against non-target kinases in vitro, showing that this inhibitor targets NGF-stimulated TrkA phosphorylation. As 125 nM Crizotinib inhibits ALK-mediated ERK phosphorylation, but no effect with this concentration is seen on native PC12 cells, we suggest that the effect of Crizotinib at this concentration is specific for ALK.
Taken together, the results of the present study suggest that the ALK point mutations found in a number of neuroblastoma patients are truly gain-of-function mutations and most probably play an important role in the progression of neuroblastoma. As a broad generalization, we find that somatically arising ALK mutations tend to be more aggressive than germ-line mutations. This information is of importance, since at least one previously described neuroblastoma ALK mutation (I1250T) has been reported to have defective kinase activity rather than increased activity . In addition, the ALK gain-of-function mutations can be inhibited by ALK-specific inhibitors, such as Crizotinib and NVP-TAE684. As Crizotinib is already in Phase II and III clinical trials, the present study contributes to the finding of patient-specific treatments for ALK-positive neuroblastoma.
anaplastic lymphoma kinase
Dulbecco's modified Eagle's medium
epidermal growth factor
fetal bovine serum
green fluorescent protein
nerve growth factor
protein kinase B
receptor tyrosine kinase
signal transducer and activator of transcription 3
Bengt Hallberg and Ruth Palmer designed the study and wrote the paper. Christina Schönherr, Kristina Ruuth, Yasuo Yamazaki and Therese Eriksson performed the experiments. James Christensen advised in the initial phase of the inhibitor studies. All authors discussed the results and commented on the paper prior to publication.
This work was supported by the Swedish Cancer Society [grant number BH 08-0597]; the Children's Cancer Foundation [grant numbers BH 08/084, RHP 08/074]; the Swedish Research Council [grant numbers RHP 621-2003-3399, BH 521-2009-3897]; Lions Cancer Society, Umeå; and the Association for International Cancer Research [grant number RHP 08-0177]. R.H.P. is a Swedish Cancer Foundation Research Fellow.