BTK Modulates p53 Activity to Enhance Apoptotic and Senescent Responses

The tumor suppressor p53 (also known as TP53) mediates cellular responses to DNA damage (1). It is a transcription factor that prevents the transformation of cells by triggering protective mechanisms such as cell-cycle arrest or apoptosis (2, 3). Several factors cooperate to determine cell fates induced by p53 after stress, although not all of them are known (4). p53 is regulated posttranslationally through many different modifications, including phosphorylation, methylation, and acetylation (3, 5–8). Its N-terminal region has an important role in its stability, because the E3 ligase MDM2 binds to it and ubiquitinates p53, which is then targeted for proteasomal degradation (9). Different stresses lead to phosphorylation of residues in the N-terminal region, including serine 15, which disrupts the MDM2–p53 interaction and thus increases the half-life of p53 (10, 11).

An important cellular response to p53 is senescence, a permanent cell-cycle arrest with specific morphological features in which cells remain metabolically active (12). The establishment of this phenotype may be a response to stress (stress-induced premature senescence) or telomere shortening after a number of cell divisions (replicative senescence; ref. 13). Senescent cells are present in premalignant and early stages of solid cancers, after which they disappear, which suggests that the senescent barrier needs to be overcome in order to progress into malignancy (14). Senescence has also been associated with age-dependent organismal changes, and the accumulation of senescent cells with time has been shown to contribute to the functional organ impairment seen in aging (15).

Bruton's tyrosine kinase (BTK) is a nonreceptor tyrosine kinase mutated in the inherited immunodeficiency disease X-linked agammaglobulinemia (16). It is expressed in myeloid and lymphoid cells but not in T cells and it is a member of the Tec family of kinases, which play a role in B-cell receptor (BCR) signaling (17). In B cells, BTK is activated after an antigen binds to the BCR, which leads to its phosphorylation at tyrosine 551 by SRC family kinases and its autophosphorylation at tyrosine 223 (18). Although BTK is mainly located at the cell membrane, it can also be found in the nucleus (19). A pathologic BTK upregulation has been shown in B-cell malignancies such as chronic lymphocytic leukemia (CLL), Mantle cell lymphoma, and multiple myeloma (20–22). Small molecule inhibitors of BTK have been developed to treat these diseases. Ibrutinib (PCI-32765), which irreversibly binds to C481 and inhibits Y223 stimulation (23), has been approved to treat mantle lymphoma and CLL (24). Another BTK inhibitor, CGI-1746, reversibly binds to the SH3 domain and blocks phosphorylation at Y551, thus keeping the protein in its inactive form (25).

We found that BTK was induced in a screening of markers for senescent cells (26). Exploring this, here we show that BTK is involved in the p53 pathway as a modulator of its activity. We propose that, in opposition to its oncogenic role in B cells, BTK can be expressed in response to damage and contributes to the tumor suppressor mechanisms regulated by p53.

Human diploid dermal fibroblasts (HDF), HCT116 and HT1080, were maintained in DMEM supplemented with 10% FBS (Gibco) and penicillin–streptomycin (50 U/mL). EJp53 were maintained as described (27). EJp53 were a gift from Dr. S. Aaronson (Mount Sinai School of Medicine, New York, 2008), HCT116 from Dr. B. Vogelstein (Johns Hopkins University, Baltimore, 2008), and HT1080 from Dr I. Roninson (South Carolina College of Pharmacy, Charleston, 2008). HDF were obtained from ATCC (2014). All cells were expanded and stored in liquid nitrogen when received and original vials were thawed for these experiments. No further validation or authentication was performed. Chemicals: Ibrutinib (Sellechem, PCI-32765), CGI-1746 (Axon Medchem, 2018), tert-butylhydroperoxide (Merck), doxorubicin (Sigma-Aldrich), sotrastaurin (Sellechem), and dactolisib (Sellechem). Peripheral blood samples were obtained from patients with CLL after informed consent and approval from the Local Research Ethics Committee. The patients were treatment naïve and had a cell count >50 × 10⁹/l. Peripheral blood mononuclear cells (PBMC) were processed as previously described (28). An adenovirus containing p53 (Adp53), a gift of B. Vogelstein (Johns Hopkins University, Baltimore, MD), or LacZ (AdLacZ) were amplified as previously described (29). Cells were exposed to 10 μL of the appropriately diluted virus stock.

PPISURV (30) was used to correlate survival rates in cancer patients to the expression level of BTK. In each data set, samples were grouped with respect to expression rank of the gene, which reflects relative mRNA expression level and introduces no normalization bias. The “low expression” and “high expression” groups are those where expression rank of the gene is less or more than average expression rank across the data set, respectively. This separation of patients into “low” and “high” groups in the data set along with survival information is next used to find statistical differences in survival outcome. The R statistical package was used to perform survival analyses and to draw KM plots. Unadjusted P-values were generated using standard survival analysis packages.

A total of 1 × 10⁴ cells were plated in 6 cm plates a day before chemicals were added. Cells were counted every 4 days using a BIO-RAD TC20 automated cell counter, after which fresh media and chemicals were added to the cultures.

Western blots were performed as described (31). Primary antibodies used: anti-phospho-Histone 2A (ser 139, Millipore 05-636), anti-phosphoATM (S1918, Abcam ab81292), anti-phospho-p53 (ser 15, Cell Signaling 9284), anti-p53 (DO-1, Santa Cruz, sc-129), anti-BTK (D3H, Cell Signaling Technology 8547), anti-p16 (Abcam ab54210), anti-β-actin (Abcam ab8227) and anti-caspase 3 (Cell Signaling Technology 96625), anti-PARP-1 (Cell Signaling Technology 9542). An ECL detection system (Thermo Scientific) was used to visualize the results. Alternatively, an Odyssey CLx Infrared Imaging System (Li-COR) was used. For Ponceau staining, 1× Ponceau staining (Sigma; P 3504) was added to the membrane for 5 minutes with shaking.

Cells were processed as previously described (26). Coverslips were incubated overnight at 4°C with 1: 50 anti-BTK (D3H5, Cell Signaling Technology 8547) or 1:100 anti-p53-FITC (Santa Cruz sc-126). Secondary anti-rabbit (1:500) antibody (Alexa Fluor 594; Invitrogen) and DAPI (Invitrogen) were used. Slides were analyzed using a Nikon TE300 microscope.

Cells were stained as previously described (32).

Five hundred cells were split into 60 mm plates in triplicates and assays were performed as described (31).

Transfection was performed using Lipofectamine 2000 (Invitrogen) following manufacturer's protocols. Eight micrograms of empty plasmid (Mission pLKO.1 Empty vector control plasmid DNA; Sigma Aldrich SHC001) or a BTK plasmid (OriGene RG211582) were used. Medium was changed after 5 hours and cells were left for 18 to 24 hours.

shRNA against BTK (Santa Cruz sc-29841-sh) or luciferase (control, Sigma Aldrich SHC007, Mission pLKO.1puro Luciferase) were transfected into EJp53 and HCT116 using Lipofectamine 2000 following manufacturer's protocols using 1 μg of shRNA. Two μg/mL puromycin was added for 2 weeks to select for transfected cells.

RNA purification, cDNA preparation, and qRT-PCR were performed as previously described (33). Primers used: PUMA: GACCTCAACGCACAGTACGA (FWD), CACCTAATTGGGCTCCATCT (REV). p21: CACCGAGACACCACTGGAGG (FWD), GAGAAGATCAGCCGGCGTTT (REV). GAPDH: GGGAAGGTGAAGGTCGGAGT (FWD), TTGAGGTCAATGAAGGGGTCA (REV). BTK: ACAGATTCCGAGGAGAGGTGAGG (FWD), GGTCCTTCATCATATACAACCTGGAATGG (REV).

Chromatin immunoprecipitation (ChIP) assays were performed as previously described (34). A total of 3 × 10⁶ cells per sample were used. PCR was performed with 1 μL of DNA. The p53 (Ab-6, Galbichem#OP43) antibody was used. Primers used: PUMA, GCGAGACTGTGGCCTTGTGT (FWD), CGTTCCAGGGTCCACAAAGT (REV); p21, GTGGCTCTGATTGGCTTTCTG (FWD), CTGAAAACAGGCAGCCCAAG (REV); Actin, TGGCTCAGCTTTTTGGATTC (FWD), CCGAGGATTGGAGAAGCAGT (REV).

A total of 100,000 cells per well were seeded in 0.5 mL growth media in 24-well plates the day before transfection. Transfection was performed with Turbofect (Life Technologies #R0531) using manufacturer's protocol. 1.0 μg total plasmid DNA (0.3 μg luciferase construct, 0.2 μg β-galactosidase construct, and 0.5 μg empty control vector or BTK construct) was used. Twenty-four hours after transfection, media was replaced by fresh growth media and part of samples was treated with 1.0 μmol/L Doxorubicin. Twenty hours later, lysates were collected in lysis buffer and stored at −80°C overnight [10 mL of β-galactosidase stock solution, 20 mg o-nitrophenyl-β-d-galactopyranoside (ONPG)], and 35 μL β-mercaptoethanol and 100 μL β-galactosidase substrate were added to 80 μL cell lysate (per well) in a 96-well transparent plate. This was incubated at 37°C for 15 minutes. Twenty microliters of the cell lysate were transferred to a 96-well opaque white plate and read in a luminometer, using a Luciferase Kit (BioVision #K801–200) following manufacturer's instructions. Luminescence was measured at a wavelength of 450 nm with a luminometer spectrometer (PerkinElmer VICTOR X). Relative luciferase activity was calculated by first dividing the luminescence value of each sample by its B-gal value and then presenting the result as a fold difference between each sample and the control sample.

Full length of recombinant BTK (Invitrogen, PV3363) and different regions p53 of were used in this study. The GST-tagged fragments of p53 were obtained by transfecting the appropriate pGEX vectors into BL21 competent and purifying the proteins produced with glutathione HiCap matrix beads (Qiagen #30900) as previously described (6). To examine the ability of BTK to phosphorylate p53, 0.5 μg of BTK and 0.5 μg of different GST-p53 proteins were added in tube contained 25 μL of the kinase buffer [12.5 mmol/L (pH 7.5), 10 mmol/L MgCl₂, 1 mmol/L EGTA, 0.5 mmol/L Na₃VO₄, 5 mmol/L β-glycerophosphate, 0.01% Triton X-100, 2 mmol/L DTT, 200 μmol/L ATP]. Approximately, 0.37 MBq of [³²P] γ-ATP was added to each tube for 30 minutes at 30°C. Loading buffer was then added to each sample, and samples were boiled for 5 minutes and loaded into 10% polyacrylamide gel. The gel was then dried using a Model v583 Gel drier (Bio-Rad) for 1 hour and exposed to film in a phosphor screen (GH Healthcare) overnight at room temperature. Phospho-image of dried gel was then developed using a Typhoon TRIO+ Variable mode manger scanner (GE Healthcare).

A total of 40,000 cells were processed as described (35). Briefly, 170 μL of 0.6% low melting point agarose was added to the cell pellet. Eighty microliters of the mixture was placed onto the precoated slides with normal melting point agarose and covered by a coverslip. After the gel became solid, cover slips were removed and slides kept in lysis buffer (2.5 M NaCl, 0.1 M Na₂EDTA, 10 mmol/L Tris-HCl, 1% TritonX-100 at pH 10) overnight. Slides were then washed for 20 minutes with ice cold ddH₂O, protected from light, and exposed to ice cold alkaline buffer (pH > 13) for 20 minutes, protected from light. This step allows the unwinding of the DNA before applying the current at 30 v/300 mA for further 20 minutes. Slides were washed twice for 5 minutes with ddH₂O, and placed at 37°C overnight. Slides were rehydrated with ice cold ddH₂O for 30 minutes at room temperature and stained with propidium iodide (PI, 2.5 5 μg/mL) for further 20 minutes, then washed with fresh ddH₂O for 20 minutes and placed in an oven before being scored with a fluorescence microscope. Comets were visualized in the fluorescence microscope at a magnification of ×200. Images were captured and analyzed with the Komet Analysis software version 5.5 (Andor Bioimaging, Nottingham, UK). Hundred cells were analyzed per sample, 50 per duplicate slide. The percentage of DNA in the tail of the comet was calculated for each cell by the software.

All error bars represent the SD. Statistical significance (not significant, ns, P > 0.05; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001) was calculated using two-tailed unpaired t tests with Prism 6 (GraphPad) software.

In a proteomics screen (26), we identified BTK as a protein induced in cells undergoing senescence. BTK, a nonreceptor kinase involved in BCR signaling (16), is highly expressed in different types of leukemia and lymphomas (36) and this has led to BTK inhibitors being used to treat B-cell malignancies (25). To better understand the role of BTK in tumor suppressor pathways, we used EJp53, a p53-null bladder cancer cell line with a tetracycline (tet)-regulatable p53 expression system (27). We found that BTK protein levels were elevated after inducing p53 expression in these cells (Fig. 1A). Moreover, the colon cancer cell line HCT116, which has wild-type p53, also showed a p53-dependent BTK induction after being treated with DNA damaging agents (the oxidant tBH and doxorubicin), both at protein (Fig. 1B and C) and mRNA levels (Fig. 1D). This suggests a direct correlation between the induction of p53 and BTK. Using PPISURV (30), we analyzed survival rates in cancer patients and found that, despite high expression of BTK correlating with a poor prognostic in CLL, as expected (25, 36), it was a marker of good prognosis in breast and lung cancer (Fig. 1E). These data together suggest the involvement of BTK in p53-mediated responses to cellular damage and tumor suppression.

We next investigated the details of the contribution of BTK to the p53 pathway. We transfected BTK into EJp53 and observed that it elevated the levels of p53 protein induced by tet removal (Fig. 2A). This was also evident in HCT116, in which BTK expression increased p53 levels in the presence or absence of DNA damaging agents (Supplementary Fig. S1A). Of note, BTK expression was not sufficient to induce nuclear localization of p53 (Supplementary Fig. S1A), which was otherwise evident after damage. Moreover, it did not increase the levels of p53 mRNA (Fig. 2B), which indicated a potential involvement of BTK in the posttranscriptional modifications of p53. To test this possibility, we used Ibrutinib, a BTK inhibitor recently approved to treat different forms of leukemia (23), CGI-1746, a more specific BTK inhibitor (25), and an EJp53 stably expressing a shRNA against BTK (Supplementary Fig. S1B). Blocking BTK by any of these methods severely reduced the levels of p53 induced by tet removal in EJp53, both after short- and long-term inhibition (Fig. 2C and D). Reduction in p53 protein levels in the absence of BTK was also observed in EJp53 and HCT116 infected with an adenovirus containing p53 (Supplementary Fig. S1C), as well as primary malignant B cells stimulated with CD154 and IL4, a culture protocol that we observed induces p53 expression (Supplementary Fig. S1D). This shows that the effects of BTK on p53 levels can be seen in different models of p53 induction and confirms that BTK is involved in the stabilization of p53 protein levels.

We reasoned that the positive effects of BTK on p53 accumulation could be mediated by phosphorylation, the most common mechanism of p53 regulation (5). AS shown in Fig. 2E, we found that BTK was able to phosphorylate p53 in vitro and, using recombinant peptides (Supplementary Fig. S2A), we predicted that this phosphorylation takes place mainly in residues 1 to 80 of p53. This region of p53 contains serine 15 (S15), which is important in the MDM2-mediated regulation of p53 protein levels (5). Using a phospho-specific antibody, we confirmed that BTK increases S15 phosphorylation in cells in the presence or absence of DNA damage (Fig. 2F). Moreover, a S15A p53 mutant showed reduced phosphorylation by BTK in vitro (Supplementary Fig. S2B). p53 is known to be phosphorylated by ATM, ATR, DNA-PK, and PKC (37), among other kinases. Using specific inhibitors, such as dactolisib (38) and sotrastaurin (39), we found that the effects of BTK on p53 activity were independent of all these kinases (Supplementary Fig. S2C), suggesting a direct phosphorylation. This data together indicates that BTK increases the phosphorylation of p53 at S15, among other residues at the N terminus, which would result in the disruption of MDM2-p53 interactions and thus provide a mechanism for the increase in p53 stability.

Our data shows that BTK is important for the phosphorylation and stabilization of p53. This suggests that it could play a role in enhancing p53 activity. We explored this hypothesis by measuring the effect of BTK on the expression of p53 target genes. As shown in Fig. 3A, BTK increased the upregulation of the mRNA of both p21 (also known as Cdkn1a) and PUMA after treatment with doxorubicin. This was confirmed by a higher activity of the luciferase reporters in the presence of BTK (Fig. 3B). Finally, a ChIP analysis showed increased binding of p53 to the p21 and PUMA promoters after DNA damage when BTK was expressed (Fig. 3C). In these experiments, BTK did not significantly increase the activity of p53 in the absence of damage, which is consistent with the fact that although p53 can be induced by BTK expression alone, it remains in the cytosol and it is thus inactive (Supplementary Fig. S1A). Importantly, when a BTK lacking the kinase domain (40) was expressed, the increase in luciferase activity of the p21 reporter was not observed (Fig. 3D and E), indicating that the positive effects on p53 transactivation are mediated by the kinase activity of BTK.

We next studied the involvement of BTK in the cellular responses to p53. As shown in Fig. 4A, inhibition of BTK by chemical or genetic approaches increased cell proliferation in EJp53 induced to senesce. The inhibition of the permanent growth arrest was confirmed by the fact that the morphological changes associated with senescence were not as evident in cells in which BTK had been blocked (Fig. 4B) and that these cells were able to form colonies (Fig. 4C). EJp53 are a representative model of stress-induced premature senescence (41). To expand our observations, we also investigated whether BTK plays a role in replicative-induced senescence by serially passaging normal human fibroblasts in the presence of BTK inhibitors. As shown in Fig. 4D, Ibrutinib was able to delay the growth arrest of fibroblasts as they entered senescence. Moreover, expression of p16 (a well-known marker of senescence; ref. 42) was reduced in late passage cells when BTK was inhibited, similarly to what happened to the levels of p53 (Fig. 4E). This suggests that replicative senescence was also delayed in the absence of BTK. Indeed, the percentage of cells positive for senescence associated (SA)-β-gal, a widely used marker of senescence (32), was significantly lower when BTK was inhibited, and less cells showed the morphological changes typical of senescence (Fig. 4F). Collectively, these results indicate that BTK has a critical role in inducing and/or maintaining both p53-dependent stress-induced and replicative senescence.

Our results suggest a new role for BTK in tumor suppressor responses in epithelial cells. We further explored this by studying the role of BTK in DNA damage responses. We first transfected BTK into HCT116 (Fig. 5A) and measured the activation of elements of the DNA damage pathways. We found that BTK expression leads to the phosphorylation of ATM and H2AX (Fig. 5B). BTK induced ATM in normal fibroblasts as well (Supplementary Fig. S3A). This correlated with increased damage to DNA, as measured by a comet assay (Fig. 5C). Together with this, we observed that BTK expression was able to induce apoptosis in HCT116 cells, as measured by PI and Annexin V stainings (Fig. 5D) as well as cleavage of caspase 3 and PARP (Fig. 5E). Moreover, this was also observed in EJp53 cells in the absence of p53 expression (Supplementary Fig. S3B), which suggests that it is at least in part p53-independent. BTK prevented EJp53 growth (Supplementary Fig. S3C) and colony formation (Supplementary Fig. S3D), similarly to the expression of p53. Finally, inhibition of BTK by Ibrutinib in HCT116 cells treated with doxorubicin significantly reduced the amount of cell death induced by the drug (Fig. 5F), thus confirming the involvement of BTK in the response to DNA damage in these cells. These data together suggest a proapoptotic and antiproliferative role of BTK that relies on its participation in DNA damage pathways and may not depend completely on its effects on p53.

Our study reveals that BTK is a novel modulator of p53 activity and is involved in the cellular response to DNA damage. Although BTK has been implicated in oncogenic signaling in blood malignancies (16), here we show that it can also participate in tumor suppressor pathways. This is consistent with previous reports that also uncovered proapoptotic and antineoplastic properties of BTK (43–45). This suggests that BTK expression has radically different functions depending on the cellular context, which would include being part of prosurvival pathways in B cells and enhancing arrest/death signals in damaged epithelial cells. This paradox is supported by our bioinformatics analysis, which shows an opposed prognostic value of BTK levels of expression in blood malignancies and solid tumors (Fig. 1E).

Our experiments also allowed us to propose a mechanism by which BTK impacts on cell fate responses to damage. We found that following BTK expression, there is an increase in p53 phosphorylation at S15, among other residues, which would explain the BTK-mediated stabilization of p53 protein levels (5). This is independent of the activity of the kinases that most commonly phosphorylate p53 in response to damage, such as ATM, ATR, DNA-PK, or PKC. Although it would have to be further confirmed, this suggests that BTK could be directly phosphorylating p53 to activate it. Of note, despite being originally described as a tyrosine kinase, BTK is known to be able to phosphorylate serines as well (46), which would fit this hypothesis. Nevertheless, we found that BTK can increase ATM and γH2AX, consistent with a role in enhancing of DNA damage signals.

We also confirmed that the kinase function of BTK is essential for its effects on p53 accumulation. However, p53 still needs to be activated by other signals in the DNA damage pathway in order to relocate to the nucleus. All our data indicate that BTK is not sufficient to activate p53 on its own, but that it can increase its activity in the context of a DNA damage response. Indeed, we have observed that BTK expression increases the binding of p53 to the promoters of its target genes after genotoxic stress, which results in higher levels of expression of both proapoptotic and proarrest signals.

The importance of BTK in the different cellular responses to p53 is underscored by the consequences of its inhibition, because using both chemical inhibitors and RNAi we showed that p53 accumulation is compromised. The potential neoplastic effects of inhibiting BTK in clinical settings should need to be carefully considered. Large cohorts of patients are currently being given BTK inhibitors for the treatment of leukemia, which our results suggest could potentially affect p53 activity. So far there have been no reports of increased incidence of secondary malignancies but our data suggests that it is a possibility that should be studied.

Inhibition of BTK results in the partial blocking of p53-induced senescence, which shows that BTK is a novel regulator of this tumor suppressor pathway. This is especially relevant because, despite the considerable knowledge accumulated in the 50 years since Leonard Hayflick first described cellular senescence (47), the molecular pathways involved in this phenotype are still incompletely characterized. Our results show that in the absence of BTK, p53-dependent senescence can be partially avoided and cells are allowed to continue growing, probably as a direct consequence of p53 protein levels being strongly reduced. Moreover, replicative senescence, which also relies on other triggers, such as p16 (42), was also delayed in the absence of BTK, which suggests that BTK could also be contributing to this phenotype in a p53-independent manner. These results provide a rationale to use BTK inhibitors to block senescence in those situations in which it has a negative effect on the organism. For instance, they could have a therapeutic effect on diseases in which senescent cell accumulation is thought to be important, such as diabetes, osteoporosis, COPD, or neurodegenerative disorders (48). The accumulation of senescent cells strongly contributes to the progression of breast cancers (49) and this could also be avoided with the adequate regime of BTK inhibitors.

We observed that blocking BTK reduces p53-induced apoptosis as well. We show that BTK can have proapoptotic effects when overexpressed, which suggests that it could contribute to apoptotic signals in certain contexts, even in the absence of p53. Because we have shown that DNA damage signals are activated when BTK is expressed, it would be important to characterize other effects of BTK in these pathways, directly or through indirect mechanisms. We propose that the role of BTK on enhancing p53-dependent cell death is likely to be at least at two different levels: it would be important to stabilize p53 through phosphorylation and it would also contribute apoptotic signals independently.

In summary, our data presents a new side of BTK as a key modulator of p53 functions, being involved in a positive feedback loop that links its induction after p53 upregulation to its stabilization of p53. Thus, BTK emerges as a proapoptotic kinase and an important regulator of p53-mediated senescence and apoptosis. This underscores the complexity of BTK functions, which can sometimes be antagonistic and are likely to be determined by cellular context.

Salvador Macip has ownership interest (including patents) in a patent application. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M. Althubiti, S. Macip

Development of methodology: M. Althubiti, J.M. Escorsa

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Althubiti, J. Samuel, J.M. Escorsa, K.-P. Lam, G.D.D. Jones, S. Macip

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Althubiti, M. Rada, G.D.D. Jones, N.A. Barlev, S. Macip

Writing, review, and/or revision of the manuscript: M. Althubiti, M. Rada, J. Samuel, K.-P. Lam, G.D.D. Jones, N.A. Barlev, S. Macip

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Althubiti, M. Rada, K.-G. Lee

Study supervision: M. Althubiti, S. Macip

Other (conceived the study with Dr. Salvador Macip.): M. Althubiti

Other (some experimental and technical support): H. Najeeb

Other (material support: provision of plasmids): K.-G. Lee

Other (construction and provision of critical reagents, advice on experiments): K.-P. Lam

Work in S. Macip's laboratory was supported by an MRC New Blood Fellowship and an Innovation Fellowship from the University of Leicester. M. Althubiti was supported by a Saudi Government Doctoral Scholarship. M. Rada was supported by a Doctoral Studentship from the Kurdistan Regional Government (Iraq). N.A. Barlev acknowledges funding from RSCF (grant no. 14-15-00816).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.