Activation of c-Abl Kinase Potentiates the Anti-myeloma Drug Lenalidomide by Promoting DDA1 Protein Recruitment to the CRL4 Ubiquitin Ligase*

Cullin-RING ligase 4 (CRL4), a complex of Cul4 and DDB1, regulates the cell cycle, DNA damage repair, and chromatin replication by targeting a variety of substrates for ubiquitination. CRL4 is also hijacked by viral proteins or thalidomide-derived compounds to degrade host restriction factors. Here we report that the c-Abl non-receptor kinase phosphorylates DDB1 at residue Tyr-316 to recruit a small regulatory protein, DDA1, leading to increased substrate ubiquitination. Pharmacological inhibition or genetic ablation of the Abl-DDB1-DDA1 axis decreases the ubiquitination of CRL4 substrates, including IKZF1 and IKZF3, in lenalidomide-treated multiple myeloma cells. Importantly, panobinostat, a recently approved anti-myeloma drug, and dexamethasone enhance lenalidomide-induced substrate degradation and cytotoxicity by activating c-Abl, therefore providing a mechanism underlying their combination with lenalidomide to treat multiple myeloma.

Additionally, immunomodulatory drugs (IMiDs), including thalidomide and its derivatives lenalidomide and pomalidomide, can repurpose CRL4 to target and destroy IKZF1 and IKZF3 (17,18), two lymphoid transcription factors essential for multiple myeloma (MM) cell survival. Similarly, lenalidomide induces ubiquitination and degradation of casein kinase 1A1 (CK1␣), which accounts for the clinical efficacy of lenalidomide in myelodysplastic syndrome (MDS) with deletion of chromosome 5q (del(5q)) (19). Crystal structure studies reveal that the glutarimide moiety of lenalidomide directly inserts into a hydrophobic pocket of cereblon (CRBN), a CRL4 substrate receptor, and that the exposed chemical moiety, together with CRBN, creates a new surface for receiving substrates (20 -22). Based on these results, novel chemicals conjugating glutarimide to other protein-interacting chemical structures are designed to control protein stability and thus enable targeting of previously intractable proteins (23). CRBN by itself mediates the ubiquitination of several putative substrates, including the large conductance, Ca 2ϩ -and voltage-activated K ϩ (BK) channels (24), MEIS2 (20) and glutamine synthetase (25), suggesting that the CRL4 CRBN ubiquitin ligase activity is diverted by IMiDs or its derived chemicals from its physiological functions.
Human DDA1 (DDB1-and DET1-associated protein 1) is a small (12-kDa) DDB1-interacting protein that co-purifies with the Cul4-DDB1 complex (26,27). DDA1 associates with some DDB1-and Cul4-associated factors and is assembled into CRL4 complexes (28), but how this process is regulated remains elusive. Here we present evidence that the c-Abl non-receptor kinase phosphorylates DDB1 at residue Tyr-316 to recruit DDA1, leading to increased substrate ubiquitination. Furthermore, the newly approved anti-myeloma drug LBH589 (panobinostat) (29), a histone deacetylase (HDAC) inhibitor, stimulates c-Abl kinase activity to promote lenalidomide-dependent ubiquitination and degradation of IKZF transcription factors in MM cells, justifying their combination in treating MM patients for better clinical benefits.

Results
Inhibition of c-Abl Kinase Desensitizes MM Cells to Lenalidomide-IMiDs target the CRL4 CRBN ubiquitin ligase to induce the ubiquitination and degradation of IKZF1 and IKZF3 to block MM cell proliferation (17,18). c-Abl was reported to bind and phosphorylate DDB1 (30), an adaptor protein of the CRL4 complex, but the exact role of c-Abl in regulating CRL4 is not clear. We first tested whether c-Abl kinase affects the cytotoxicity of lenalidomide in MM cells. In our initial experiments, imatinib (Gleevec), an inhibitor of Abl family kinases, ameliorated the anti-proliferative activity of lenalidomide in several MM cell lines, including MM1S, MM1R, and U266 ( Fig. 1, A-C). Deletion of Abl1, which encodes c-Abl, by CRISPR-Cas9 genome editing via lentiviral vectors similarly desensitized the cells to lenalidomide-induced cell death (Fig. 1, D and E). These results show that the activity of c-Abl kinase is required for optimal lenalidomide-induced inhibition of MM cell proliferation.
Activation of c-Abl Enhances Lenalidomide-induced Cytotoxicity-c-Abl can be activated by multiple stress signals, such as cell adhesion (31) and DNA damage (32). Trichostatin A, a pan-histone deacetylase inhibitor, was reported to activate c-Abl by inducing chromatin distortion (33). LBH589 (panobinostat), a pan-histone deacetylase inhibitor, has recently been approved to treat relapsed MM patients in combination with bortezomib and dexamethasone (29). We first generated an MM1S cell line stably expressing c-Abl by lentiviral vectors and treated these cells with increasing concentrations of LBH589. Using c-Abl self-phosphorylation (Tyr(P)-245) as a readout, we confirmed that LBH589 activated c-Abl kinase in a dose-dependent manner in these MM cells (Fig. 1F). To test whether LBH589 would affect lenalidomide cytotoxicity in MM cells by activating c-Abl, we treated MM1S cells with lenalidomide and LBH589 in the presence or absence of imatinib. Proliferation of MM1S cells was better inhibited by combination treatment with LBH589 and lenalidomide, but the enhanced inhibition could be abolished by imatinib treatment (Fig. 1G).
Dexamethasone is a glucocorticoid approved to treat MM in combination with IMiDs and proteasome inhibitors (34,35). Little is known about the mechanism underlying how dexamethasone enhances the anti-myeloma activity of these drugs. We found that imatinib similarly reversed the dexamethasonedependent increase of lenalidomide-induced MM cell proliferation arrest (Fig. 3A). We conclude that LBH589 and dexamethasone increase the cytotoxic effect of lenalidomide on MM cells by activating the c-Abl kinase.
Activation of c-Abl Promotes Lenalidomide-dependent Substrate Degradation-We next sought to determine whether c-Abl enhanced lenalidomide cytotoxicity by stimulating the turnover of the lenalidomide substrates IKZF1 and IKZF3. LBH589 increased lenalidomide-induced degradation of IKZF1/ IKZF3 and loss of viability in lenalidomide-sensitive MM cell lines, including MM1S cells ( Fig. 2A), OPM2 cells (Fig. 2B), and U266 cells (Fig. 2C), with no significant effect on the abundance of CRBN protein (Fig. 2, A-C). LBH589 did not alter IKZF1/ IKZF3 levels or cause significant cytotoxicity in these sensitive cell lines as a single agent (Fig. 2, A-C, second lanes). We generated pomalidomide-resistant MM1S-P5000 cells by continuous culture of MM1S cells in medium containing gradually increasing concentrations of pomalidomide. The final pomalidomide concentration in medium was 5 M to maintain the growth of these cells. MM1S-P5000 cells are also lenalidomideresistant, likely because of the loss of CRBN expression ( Fig.  2A). We found that LBH589 did not alter IKZF1/IKZF3 levels

c-Abl Activates CRL4
or cause obvious cytotoxicity in MM1S-P5000 cells as a single agent or even in combination with lenalidomide ( Fig. 2A), suggesting that LBH589-induced cytotoxicity is dependent on lenalidomide. Deletion of c-Abl mitigated the decrease of IKZF1/IKZF3 protein levels and the arrest of cell proliferation in cells treated with LBH589 and lenalidomide (Fig. 2D). Interestingly, treatment with MLN4924, a neddylation inhibitor that inactivates all CRLs, including CRL4 CRBN (36), completely rescued the decrease in IKZF1/IKZF3 protein abundance after treatment with either lenalidomide alone or a combination of LBH589 and lenalidomide in these three MM cell lines (Fig. 2E). As LBH589 activated c-Abl, we then examined whether c-Abl regulates the lenalidomide-dependent ubiquitination of IKZF1/ IKZF3. IKZF1/IKZF3 ubiquitination was detectable in HEK 293T cells only when the cells were treated with lenalidomide, and expression of exogenous wild-type c-Abl, but not the kinase-defective c-Abl (Abl-KR), enhanced the levels of IKZF3 and IKZF1 ubiquitination (Fig. 2, F and G). These data suggest that LBH589 stimulates the ubiquitination and proteasomal degradation of IKZF1/IKZF3 in the presence of lenalidomide.
c-Abl Phosphorylates Multiple Tyrosines of DDB1 Proteinc-Abl was reported to bind and phosphorylate DDB1 (30) and promote the ubiquitination and proteasomal degradation of DDB2 in cells exposed to UV damage (37). We confirmed that c-Abl kinase phosphorylated DDB1 in a kinase-dependent manner (Fig. 4A). However, which sites of DDB1 are phosphor-

c-Abl Activates CRL4
MARCH 3, 2017 • VOLUME 292 • NUMBER 9 JOURNAL OF BIOLOGICAL CHEMISTRY 3685 ylated is not clear. There are three potential c-Abl target sites in the primary sequence of DDB1 protein with the consensus YXXP sequence (Fig. 4B): Tyr-182 and Tyr-718 on the exposed surface and Tyr-678 inside the DDB1-Cul4 binding scaffold (27). Mutation at Tyr-182 or Tyr-718, but not Tyr-678, reduced the phosphorylation of DDB1 (Fig. 4C). These results suggest that c-Abl phosphorylates diverse tyrosines of DDB1 protein.
c-Abl Phosphorylates DDB1 to Recruit DDA1-Interestingly, we found that c-Abl kinase promoted the interaction of DDB1 with DDA1 (Fig. 5A), a small protein (12 kDa) with no obvious motif and likely a positive regulator of the CRL4 ligase (28). The kinase activity of c-Abl is essential for the increased DDB1-DDA1 interaction because treating cells with imatinib or expressing kinase-defective Abl-KR abolished this effect (Fig.  5A). A conserved cluster of residues (Tyr-316/Asp-318/Asn-319) on one of the three seven-bladed WD40 ␤ propellers (BPA, BPB, and BPC) of DDB1 (27) is required for DDA1 interaction (26). Conversion of Tyr-316 of DDB1 to Phe completely abrogated the interaction of DDB1 and DDA1, which also failed to be enhanced by c-Abl expression (Fig. 5B). However, the total tyrosine phosphorylation of DDB1 was only marginally reduced by this mutation (Fig. 5B), consistent with DDB1 harboring additional sites (Tyr-182 and Tyr-718) phosphorylated by c-Abl. Besides, c-Abl-induced recruitment of DDA1 was independent of phosphorylation at these two sites (Fig. 5C). Triple mutations (TMs) in Tyr-316/Tyr-182/Tyr-718 of DDB1 exhib-

c-Abl Activates CRL4
ited a more severe phosphorylation defect than the double mutations (DMs, Tyr-182/Tyr-718), indicating that all three tyrosine residues are phosphorylated (Fig. 5C). c-Abl phosphorylated DDB1 mostly in the cytoplasm, enriched DDA1 in the nucleus, and promoted the assembly of the DDB1-DDA1 complex in both the cytoplasm and nucleus (Fig. 5D). These data demonstrate that DDB1 is phosphorylated by c-Abl at Tyr-316 to recruit DDA1.
Lenalidomide-dependent IKZF3 Degradation Requires DDB1 to Recruit DDA1-To address the role of DDA1 in lenalidomide-induced substrate turnover, we examined the ubiquitination and stability of IKZF3 protein overexpressed in HEK 293T cells with knockdown of DDA1 expression. Expression of c-Abl increased the levels of ubiquitinated IKZF3, whereas knocking down DDA1 reduced both the basal and Abl-induced IKZF3 ubiquitination levels (Fig. 6, A and B). Similarly, lenalidomidedependent destabilization of IKZF3 was abrogated by knockdown of DDA1 expression but enhanced by overexpression of c-Abl with intact kinase activity (Fig. 6, C and D). Expression of c-Abl without lenalidomide treatment did not alter IKZF3 stability (Fig. 6C). Put together, these results suggest that DDA1 is recruited to DDB1 at phosphorylated Tyr-316 to promote lenalidomide-induced ubiquitination and degradation of IKZF3.

Discussion
In summary, we have uncovered a mechanism underlying clinically approved combination treatment options with lenalidomide for MM (Fig. 7). Other than their inherent cytotoxicity, small-molecule drugs such as LBH589 and dexamethasone can activate the c-Abl kinase, which phosphorylates DDB1 at Tyr-316 to recruit DDA1, leading to increased substrate ubiquitination.
CRL4 assembles with a subset of WD40 proteins to target a variety of substrates to regulate the cell cycle, chromatin functions, and DNA damage repair (1-3). c-Abl-induced activation of CRL4 would presumably affect the ubiquitination of all CRL4 substrates. Therefore we checked the ubiquitination of histone H3 (10,11) and CK1␣ (19), substrates of CRL4, and found that c-Abl increased ubiquitination of the two proteins (Fig. 8, A and  B). These data suggest that c-Abl kinase can increase the ubiquitination of diverse substrates of CRL4 E3 ligase. In this way, it would be interesting to explore the functional significance of c-Abl activation in regulating the turnover of the various CRL4 substrates in physiological or pathological processes.
DDA1 is a small protein without an obvious motif that copurifies with the Cul4-DDB1 complex (26,27) and has been suggested to be a positive regulator of the CRL4 ligase (28),

c-Abl Activates CRL4
MARCH 3, 2017 • VOLUME 292 • NUMBER 9 but how DDA1 is assembled into CRL4 is not clear. Here we show evidence that c-Abl kinase phosphorylates DDB1 at Tyr-316 to recruit DDA1. Next, we would like to explore how DDA1 recruitment accelerates the ubiquitination of various substrates.
Lenalidomide is empowered by the activated CRL4 CRBN ubiquitin ligase to target its substrates more effectively. Therefore, any intervention that activates Abl kinases would, in principle, potentiate the activities of lenalidomide or other IMiDs that engage CRL4 CRBN to treat MM (17,18) or del(5q) MDS (19). For example, DNA damage induced by ionizing radiation or cisplatin can activate c-Abl (32), begging a closer examination of these agents in DDA1-CRL4 complex assembly, leading to CRL4 activation and justifying clinical investigation into the combination of IMiDs with conventional radiotherapy or chemotherapy in treating MM and del(5q) MDS patients. Additionally, the development of new drugs to hijack the CRL4 CRBN ubiquitin ligase to degrade previously intractable proteins should take into account the critical role of Abl kinase activation in achieving better targeting efficacy (23).

Experimental Procedures
Cell Lines-HEK 293T, 293FT, HeLa, MM1S, MM1R, OPM2, and U266 cell lines were obtained from the ATCC. Cells were cultured in media recommended by the ATCC and maintained at 37°C in 5% CO 2 . Pomalidomide-resistant MM1S-P5000 cells were generated by culturing MM1S cells in medium containing gradually increasing concentrations of pomalidomide. The final pomalidomide concentration in medium was 5 M to maintain the growth of these cells.
cDNA and Plasmids-pMT21 expression vectors carrying a C-terminal Myc-tagged c-Abl and Abl-KR were gifts from Stephen Goff (Columbia University, New York, NY). The FLAG-ub expression vector was a kind gift from Zongping

c-Abl Activates CRL4
Xia (Zhejiang University). The full-length cDNAs (Human ORFeome Collection) of DDA1, CRBN, Abl1, and Abl1-KR were amplified and subcloned into the pXF6H expression vector (3ϫFLAG) from Xinhua Feng (Zhejiang University) between the EcoRI and BamHI restriction sites and named FLAG-DDA1, FLAG-CRBN, FLAG-Abl, and FLAG-Abl-KR plasmids. The full-length human DDB1 was amplified and subcloned into the pXF4H expression vector (2ϫHA) from Xinhua Feng (Zhejiang University) between the XbaI and HindIII restriction sites and named HA-DDB1 plasmid. Site-directed mutagenesis to generate HA-DDB1 (Y316F), HA-DDB1 (Y182F), HA-DDB1 (Y678F), HA-DDB1 (Y718F), HA-DDB1-DM (Y182F/Y718F), and HA-DDB1-TM (Y182F/Y718F/ Y316F) mutant expression vectors were carried out utilizing the QuikChange method (Agilent Stratagene). Human IKZF3 and IKZF1 cDNA clones were amplified with an HA tag in the C terminus and then cloned into the pXF4H expression vector between the ClaI and EcoRI sites and named IKZF3-HA and IKZF1-HA plasmids. The human CRBN cDNA clone was amplified with a Myc tag in the N terminus and then cloned into the pXF4H expression vector between the ClaI and EcoRI sites and named Myc-CRBN plasmid.
Compounds and Antibodies-MG132 (S2619), MLN4924 (S7109), LBH589 (S1030), Dexamethasone (S1322), imatinib (S2475), and lenalidomide (S1029) were purchased from Selleck Chemicals. All compounds were dissolved in DMSO. Anti- Immunoblot Analysis-All Protein lysates were resolved by standard SDS-PAGE and transferred to PVDF membranes. Blots were blocked in 5% nonfat milk in Tris-buffered saline containing 0.1% Tween 20 (TBS-T) and incubated with primary antibodies in 5% BSA at 4°C overnight. Blots were washed three times with 1ϫ TBS-T the next day and then incubated with secondary antibodies for 1 h at room temperature. Detection was performed with ECL Western blotting detection reagent (Thermo Scientific).
Co-immunoprecipitation Assays-Cells were transfected with the indicated plasmids and lysed in NETN lysis buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, and 1 mM Na 3 VO 4 supplemented with complete protease inhibitors (Roche)). After centrifugation at 13,000 rpm for 15 min, supernatants were subjected to immunoprecipitation by incubation with anti-HA affinity gel for either 4 h or overnight at 4°C. Samples were washed three times with lysis buffer before being separated by SDS-PAGE and analyzed by immunoblot.
In Vivo Ubiquitination Assay-Plasmids were transfected into HEK 293T cells as indicated. Cell lysates were prepared for ubiquitination assays as described previously (7). For ubiquitination of histone H3, histones were extracted and purified as described elsewhere (38) and then subjected to the ubiquitination assay.
Preparation of Nuclear and Cytoplasmic Extracts-HeLa cells were transfected as indicated, and the fractions were prepared as described previously (30).
Protein Degradation Analysis and Protein Quantification-Cells were treated with cycloheximide (CHX, 100 g/ml), and cells were harvested for lysate preparation using NETN lysis buffer at the indicated time points and subjected to immunoblot analysis. For quantification, each band of the WB for IKZF3 was quantified with Gel-Pro Analyzer 4.0 software (Media Cybernetics) and normalized to the corresponding actin band. The relative IKZF3 protein level at the starting time (0 h) in each group was set to 1. The relative IKZF3 protein levels at other time points were compared with that at the starting time. Data points were adjusted to a one-exponential decay curve using GraphPad Prism software (GraphPad), and the protein degradation rate (t1 ⁄ 2 ) was expressed as the time for degradation of 50% of the protein.
siRNA Sequences and Transfection-The siRNA sequences for DDA1 were CCUCAUAGGAGCCGAUGUAdTdT and GCGAGUACCCGUCUGAACAdTdT. Cells were transfected using Lipofectamine RNAiMax (Invitrogen). Lipofectamine 2000 (Invitrogen) was used for transfection with vectors and siRNAs simultaneously.
Real-time RT-PCR-Total RNA was extracted using TRIzol reagent (Life Technologies) and reverse-transcribed into cDNA using the PrimeScript RT Master Mix Kit (RR036A, TaKaRa) according to the protocol of the manufacturer. Real-time PCR was performed in triplicate using Power SYBR Green PCR Master Mix (4367659, Life Technologies) and an Applied Biosystems 7500 real-time PCR system. PCR primers were as follows: DDA1-F, GTACCCGTCTGAACAGATCATCG; DDA1-R, GGCAGCGTTCTTTTTGTCCC; GAPDH-F, CGACCACTT-TGTCAAGCTCA; and GAPDH-R, TTACTCCTTGGA-GGCCATGT.

JOURNAL OF BIOLOGICAL CHEMISTRY 3689
Viability Assay-MM cells were seeded in 96-well plates and treated with compounds at the indicated concentration and time. Proliferation rates were analyzed using Cell Counting Kit 8 ( Dojindo Laboratories, Kumamoto, Japan) according to the protocol of the manufacturer.
Statistical Analysis-The data presented were acquired from a minimum of two independent experiments unless otherwise indicated. Statistical significance in assays with identical cell lines was assessed with Student's t test (two-tailed). GraphPad Prism 5 (GraphPad) was used for all statistical analyses.
Author Contributions-S. G. and C. G. performed the experiments. T. S., J. L., and X. L. provided technical assistance. S. G., Z. C., and Y. C. analyzed the data. S. G. and Y. C. designed the study and wrote the manuscript.