Open Full Page

Published OnlineFirst May 7, 2019; DOI: 10.1158/1535-7163.MCT-18-1129

Cancer Biology and Translational Studies Molecular Cancer Therapeutics Acquired Resistance to BET-PROTACs (Proteolysis-Targeting Chimeras) Caused by Genomic Alterations in Core Components of E3 Ligase Complexes Lu Zhang1, Bridget Riley-Gillis2, Priyanka Vijay2, and Yu Shen1

Abstract

Proteolysis-targeting chimeras (PROTAC) are bifunction- tions, it would be important to understand whether and al molecules that hijack endogenous E3 ubiquitin ligases to how resistance to these novel agents may emerge. Using induce ubiquitination and subsequent degradation of pro- BET-PROTACs as a model system, we demonstrate that tein of interest. Recently, it has been shown that PROTACs resistance to both VHL- and CRBN-based PROTACs can with robust in vitro and in vivo activities and, in some cases, occur in cancer cells following chronic treatment. However, drug-like pharmaceutical properties can be generated using unlike what was often observed for many targeted thera- small-molecule ligands for the E3 ligases VHL and CRBN. peutics, resistance to BET-PROTACs did not result from These ﬁndings stoked tremendous enthusiasm on using secondary mutations that affect compound binding to the PROTACs for therapeutics development. Innate and target. In contrast, acquired resistance to both VHL- and acquired drug resistance often underlies therapeutic fail- CRBN-based BET-PROTACs was primarily caused by geno- ures, particularly for cancer therapy. With the PROTAC mic alterations that compromise core components of the technology progressing rapidly toward therapeutic applica- relevant E3 ligase complexes.

Introduction of these small-molecule PROTACs (5–7). BET-PROTACs trigger rapid and prolonged degradation of BET proteins with exception- Ubiquitination-mediated proteolysis is a central mechanism of al potencies, and exhibit robust antitumor activities in preclinical protein homeostasis in cells. E3 ubiquitin ligases bind to deﬁned models of leukemia, lymphoma, prostate cancer, and triple- substrates and mediate ubiquitination and degradation of these negative breast cancers (5, 6, 10, 11). proteins. Cullin-RING ubiquitin ligases (CRL) are the largest Compared with traditional small-molecule inhibitors, family of E3 ubiquitin ligases. CRLs are multicomponent com- PROTACs offer several advantages. For example, PROTACs can plexes that minimally consist of a cullin, a RING ﬁnger protein, exert more rapid, potent, and durable inhibition of targets, and a substrate recognition subunit (1). Proteolysis-targeting abolish the scaffolding function of a protein, and turn a non- chimeras (PROTAC) are bifunctional molecules that hijack functional binder into functional degraders. In addition, the endogenous E3 ubiquitin ligase to cause ubiquitination and stringent conformational requirement and potential linker inter- subsequently degradation of proteins of interest (2). VHL and action between target protein and E3 ligase within ternary com- Cereblon (CRBN) are the substrate recognition subunit of the plex may allow PROTACs to offer an additional layer of selectivity cullin2 (CUL2)-containing VHL–CRL complex and the cullin4- over small-molecule inhibitors (12–15). These unique properties containing CRBN–CRL complex, respectively (3, 4). Recently, it make PROTACs a promising modality for the development of has been shown that PROTACs with robust in vitro and in vivo next-generation therapeutics. However, innate and acquired drug activities and, in some cases, drug-like pharmaceutical properties resistance is a common cause of therapeutic failure, particularly can be generated using small-molecule ligands for the E3 ligases for cancer therapy. With the rapid advancement of the PROTAC VHL and CRBN (5–9). PROTACs targeting the bromodomain and technology toward therapeutic applications, it would be impor- extraterminal domain proteins (BET-PROTAC) are the prototype tant to understand whether and how drug resistance to these novel agents may emerge. In this study, using BET-PROTACs with both VHL and CRBN ligands as a model system, we interrogate 1 2 Oncology Discovery, AbbVie Inc., North Chicago, Illinois. Genomic Research potential mechanisms of acquired resistance to PROTACs in Center, AbbVie Inc., North Chicago, Illinois. cancer cell lines. Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). Corresponding Authors: Lu Zhang, AbbVie Inc., North Chicago, IL, 60064. Materials and Methods Phone: 847-938-2212; Fax: 847-935-0014; E-mail: [email protected]; and Yu Shen, [email protected] Cells, compounds, and antibodies SKM1, MV4:11, LNCaP, and OVCAR8 cell lines were purchased Mol Cancer Ther 2019;18:1302–11 from ATCC or DSMZ and maintained by a Core Cell Line Facility. doi: 10.1158/1535-7163.MCT-18-1129 All cell lines were tested for Mycoplasma using MycoAlert Detection 2019 American Association for Cancer Research. Kit (Lonza) and authenticated using the Gene Print10 STR Kit

1302 Mol Cancer Ther; 18(7) July 2019

Defects in E3 Ligase Complex Cause Resistance to BET-PROTACs

(Promega) and maintained in RPMI1640 medium with 10% FBS and MultiQC (17). SNV and indel variant calls were imported (Gibco). ABBV-075, ARV-771, and ARV-825 were synthesized at into the VarSeq Software (Golden Helix) for filtering and AbbVie by the methods according to (5, 6, 16). All antibodies were annotation. purchased from commercial sources as follows: antibodies Somatic copy number calling was performed using VarScan against BRD2/3/4 from Bethyl; antibodies against c-MYC, PARP, v2.4.2 using tumor/normal mpileup from samtools v1.7 as and VHL from Cell Signaling Technology; antibody against CRBN input (18). Parameters used were minimum segment size of from Thermo Fisher Scientific; antibody against CUL2 from 100 bp, minimum coverage of 20 reads, minimum mapping Abcam; and antibody against b-actin from Sigma. O1R- and quality of 20, and minimum base quality of 20. LOH in deleted O3R-overexpressing cells were created by infection with pLOC regions were also visually confirmed by plotting minor allele or pLOC-CUL2 or pLOC-CRBN lentiviral particle (Dharmacon) frequencies from germline variant calling with Sentieon joint in the presence of 10 mg/mL of polybrene. Cells were selected with HaplotypeCaller (Sentieon Inc.). 10 mg/mL of blasticidin. Western blot analysis was performed to confirm expression of these proteins in the cells. RNA sequencing Each cell line was sampled in triplicate and mRNA library Cell viability and caspase 3/7 activity assay preparation from total RNA was conducted following the man- Cells were seeded in 96-well plates and incubated at 37Cin ufacturer's protocol for the Illumina TruSeq mRNA preparation an atmosphere of 5% CO2. Compounds were added at a kit. Briefly, 1 mg of total RNA was purified by using poly-T oligos series of dilution after overnight incubation. After 3 days attached to magnetic beads then fragmented by divalent cations incubation, Caspase-Glo3/7 luminescent and CellTiter-Glo under elevated temperature. The fragmented RNA underwent first Assay (Promega) were performed according to the manufac- strand synthesis using reverse transcriptase and random primers. turer's instruction. Luminescence signal from each well was Second strand synthesis created the cDNA fragments using DNA measured using the EnSpire Luminometer (PerkinElmer), and polymerase I and RNaseH. The cDNA fragments then went the data were analyzed using the GraphPad Prism Software through end repair, adenylation of the 30 ends, and ligation of (GraphPad Software Inc.). adapters. The cDNA library was enriched using 15 cycles of PCR and purified. Final libraries were assessed using the Agilent qPCR Bioanalyzer and Qubit (Thermo Fisher Scientific) assay methods Total RNA was harvested using the RNeasy Plus Mini Kit then sequenced on an Illumina NextSeq sequencer using 2 75 (Qiagen) and cDNA was made with the High-Capacity cDNA bp read length. The raw sequencing data is deposited at SRA Reverse Transcription Kit (Applied Biosystems). qPCR was per- (Submission no. SUB5567117). formed with TaqMan Universal Master Mix and Probes (Life RNA-sequencing reads were mapped to the human reference Technologies), and analyzed using the DDCt method. Data are genome (GRCh38) using STAR aligner and genes were presented as mean SD. The difference between two groups was quantified using featureCounts for all genes annotated in evaluated using the two-tailed Student t test. P values less than Gencode v28 (19, 20). Quality of sequencing data was assessed 0.05 were considered statistically significant. using Picard (http://broadinstitute.github.io/picard) and MultiQC (17). Genes with counts per million less than one in Western blot analysis two-thirds of samples or more were considered too lowly Cell lysates were prepared in Laemmli Buffer (Bio-Rad). Thirty expressed and excluded. Differential gene expression (DGE) anal- micrograms of total protein was resolved on a 7%–12% ysis comparing parental versus O1R and parental versus O3R (n ¼ SDS-polyacrylamide gel and probed with corresponding primary 3 for each) was then performed using linear modeling in limma antibodies. with TMM normalization and voom transformation (21) DGE results were plotted using GLIMMA (22). Sashimi plots were Whole-exome sequencing prepared using integrated genome viewer (23). Samples were prepared in accordance with the manufacturer's instructions for the KAPA Hyper Prep and NimbleGen SeqCap EZ TaqMan CNV confirmation MedExome kits. Briefly, 100 ng of DNA was sheared using a Confirmation of exome copy number calls for CUL2 and CRBN Covaris M220 sonicator (adaptive focused acoustics). DNA frag- was performed using TaqMan copy number assays. Three prede- ments were end-repaired, adenylated, ligated with KAPA dual signed assays each for CUL2 and CRBN (Supplementary Table S1) indexing adapters, and amplified by 11 cycles of PCR with the were ordered along with the reference assays for TERT and RNaseP KAPA hyper prep kit. Samples were pooled into a precapture genes. Experiments were performed according to the manufac- library and targeted capture was performed using the NimbleGen turer's instructions using 20 ng DNA per reaction, 2 TaqMan SeqCap EZ MedExome capture probe set. Captured libraries were genotyping master mix, and both a target (CUL2 or CRBN) and then enriched by 14 cycles of PCR and final libraries were reference assay (TERT or RNaseP). Each assay combination was evaluated using Qubit (Thermo Fisher Scientific) and Agilent run in quadruplicate for parental, O1R, O3R, and a no template Tape Station and were sequenced on an Illumina NextSeq control. qPCR was run on the ABI 7500 and analyzed using the sequencer using 2 75 bp read length. The raw sequencing data CopyCaller Software employing the DDCt method for normal- is deposited at SRA (submission no. SUB5567117). ization of copy calls relative to the reference assay and parental Exome sequencing reads were processed through the Sentieon samples. TNscope Pipeline (Sentieon Inc.) for alignment and somatic variant calling (parental sample designated at "normal", O1R Sanger sequencing and O3R designated as "tumor"). Quality of sequencing data was Sequencing primers were designed using Primer3 software assessed using Picard (http://broadinstitute.github.io/picard) (version 1.1.4, http://www.sourceforge.net) and were purchased

www.aacrjournals.org Mol Cancer Ther; 18(7) July 2019 1303

Zhang et al.

from Integrated DNA Technologies (Supplementary Table S2). 98C; 30 seconds at 55C; and 60 seconds at 72C). Amplicons Genomic DNA was ampliﬁed by singleplex PCR using the FailSafe were bidirectionally sequenced using Big Dye Terminator version PCR System (Epicentre). Thermal cycling was performed with 40 1.1 technology on an ABI 3130xl System (Applied Biosystems). cycles [30 seconds at 98C; 30 seconds at 62C(0.5C each Sequence analysis was performed using the Sequencher software cycle); 60 seconds at 72C], followed by 25 cycles (30 seconds at version 5.4.6 (Gene Codes Corporation).

Figure 1. Cancer cells develop drug resistance after chronic exposure to BET-PROTACs. A, Sensitivity to ARV-771 or ARV-825 of parental OVCAR8 (black) and resistant

derivatives O1R (blue) and O3R (red) was assessed by CTG assays. Data represent mean SD. N ¼ 3. The table lists the IC50 value fold change of resistant cells over parental cells for each compound. B, Cells were treated with indicated concentration of ARV-771 or ARV-825 for 48 hours and harvested for Western blot analysis using cleaved PARP antibody. b-Actin was used as loading control. C, Caspase 3/7 assay was performed in parental and resistant cells treated with a series dilution of ARV-771 or ARV-825 for 24 hours. Data represent mean SD. N ¼ 3.

1304 Mol Cancer Ther; 18(7) July 2019 Molecular Cancer Therapeutics

Defects in E3 Ligase Complex Cause Resistance to BET-PROTACs

Results prostate cancer (refs. 6, 24, 25; Supplementary Fig. S2). While no Chronic exposure to BET-PROTACs leads to drug resistance in stable resistant clones were obtained from SKM-1, MV4:11, or cancer cell lines LNCaP cells incubated with either of the VHL- or CRBN-based Drug resistance often arises following chronic exposure to BET-PROTACs, several resistant clones emerged from the cancer therapeutic agents. To investigate whether PROTACs are OVCAR8 cells. These resistant clones exhibited greater than subject to similar drug resistance issues, we exposed two AML 40 IC50 increase over the parental cell line for the VHL-based (SKM-1 and MV4:11) and two solid tumor (LNCaP and OVCAR8) BET-PROTAC ARV-771 (6) or the CRBN-based BET-PROTAC cell lines to increasing concentrations of VHL- or CRBN-based ARV-825 (ref. 5; Fig. 1A). Withdrawing BET-PROTAC from culture BET-PROTACs over a 4 month period. The establishment of the media for 2 months did not diminish resistance, indicating that resistant cell lines is illustrated in Supplementary Fig. S1A. resistance to BET-PROTACs in these cells may result from stable Although all the four cell lines underwent apoptosis after BET- genetic changes (Supplementary Fig. S1B). In addition, the resis- PROTAC treatment, SKM1, MV4:11, and LNCaP cells were much tant clones maintained similar sensitivities as the parental cell more sensitive to these compounds compared with OVCAR8, lines to the bromodomain inhibitor ABBV-075 (16), suggesting likely due to the strong dependency on BET proteins in AML and that the BET bromodomains in these cells retain the ability of

Figure 2. BET-PROTACs fail to induce BET protein degradation in the resistant cells. A, Cells were treated with indicated concentration of ARV-771 or ARV-825 for 16 hours and harvested for Western blot analysis using antibodies against BET family proteins and c-Myc. b-Actin was used as loading control. B, Cells were treated with 0.05 mmol/L of ARV-771 or ARV-825 for indicated time and harvested for Western blot analysis using antibodies against BRD4. b-Actin was used as loading control.

www.aacrjournals.org Mol Cancer Ther; 18(7) July 2019 1305

Zhang et al.

binding BET inhibitors (Supplementary Fig. S1C). It is notewor- BET-PROTACs fail to induce BET protein degradation in the thy that the O1R cells, which are resistant to the VHL-based resistant cells BET-PROTAC ARV-771, remained sensitive to the CRBN-based It has been shown that BET-PROTACs inhibit cell growth and BET-PROTAC ARV-825. Conversely, the O3R cells, which are cause apoptosis by inducing degradation of BET family proteins. resistant to the CRBN-based BET-PROTAC ARV-825, remained As expected, treatment of ARV-771 for 16 hours caused degrada- sensitive to the VHL-based BET-PROTAC ARV-771 (Fig. 1A). tion of BRD2, BRD3, and BRD4, and inhibited the well- Consistent with what was observed in the proliferation assays, established BET target gene Myc in the parental and the O3R cells analysis of PARP cleavage and caspase activation further con- in a dose-dependent manner (Fig. 2A). In contrast, ARV-771 failed firmed that the O1R cells were resistant to ARV-771- but not to induce significant BET protein degradation or Myc inhibition in ARV-825–induced apoptosis, and the O3R cells were resistant to the O1R cells even at the very high concentration of 3 mmol/L. ARV-825- but not ARV-771–induced apoptosis (Fig. 1B and C). Similarly, treatment of ARV-825 for 16 hours caused BET protein Taken together, these results demonstrate that the O1R cells are degradation and Myc inhibition in the parental and O1R cells at as specifically resistant to the VHL-based BET-PROTAC (ARV-771), low as 30 nmol/L (Fig. 2A). However, ARV-825 did not degrade and the O3R cells are specifically resistant to the CRBN-based BRD2 or BRD4 protein efficiently at up to 3 mmol/L and only BET-PROTAC (ARV-825). partially degraded BRD3 protein in O3R cells at higher

Figure 3. Exome sequencing identified mutations and deletion in CUL2 gene. A, A model of CUL2 ligase. Adapted from Chen and colleagues (35). B, CUL2 somatic mutations identified through exome sequencing of parental and O1R samples. C, Sashimi plot of RNA-seq read coverage of CUL2 exons 11–14. Exon 12 skipping present in O1R replicates (red) but not in parental replicates (black). Solid peaks depict reads per kilobase per million; reads mapped within individual exons. Reads that bridge different exons are shown as arcs. Intron deletion and T435Hfs deletion indicated by blue and red arrows, respectively. D, CUL2 deletion detected in O1R by copy number calling on exome-sequencing data (relative to parental). E, CUL2 deletion confirmation by TaqMan CNV assay.

1306 Mol Cancer Ther; 18(7) July 2019 Molecular Cancer Therapeutics

Defects in E3 Ligase Complex Cause Resistance to BET-PROTACs

Figure 4. Resistance toward VHL-based BET PROTAC ARV-771 derives from depletion of CUL2. A, qPCR analysis of CUL2 mRNA level in parental, O1R, and O3R cells. Data represent mean SD. N ¼ 3. B, Western blot analysis of CUL2 protein level in parental, O1R, and O3R cells. b-Actin was used as loading control. Quantiﬁcation of CUL2 protein level is shown below. C, Western blot analysis of CUL2 in O1R cells expressing empty vector or CUL2 cDNA. b-Actin was used as loading control. D, Dose–response curves for ARV-771 in O1R cells expressing empty vector or CUL2 cDNA. Data represent mean SD. N ¼ 3. E, O1R cells expressing empty vector or CUL2 cDNA were treated with indicated concentration of ARV-771 for 16 hours and harvested for Western blot analysis. b-Actin was used as loading control.

concentration (Fig. 2A). This is consistent with previous reports in proteins, especially BRD4, detected (Supplementary Fig. S1D). other cancer types that BRD3 is more readily degraded by BET- Interestingly, the O1R cells were found to possess multiple PROTACs compared with BRD2 and BRD4 (6, 24, 26). Time genomic alterations that impact the gene encoding CUL2, a course studies further established that extending BET-PROTAC critical component of the VHL–CRL complex (Fig. 3A). The treatment to 48 hours still could not trigger BRD4 degradation in alterations include a frame-shift mutation predicted to cause the resistant cells (Fig. 2B). These results collectively demonstrate nonsense-mediated mRNA decay by introducing a premature that the O1R and O3R cells are resistant to BET protein degrada- stop codon at position 442 in the cullin homology domain; an tion induced by VHL- or CRBN-based BET-PROTACs, respectively. intronic mutation upstream of exon 12 with concomitant The absence of cross resistance to both VHL- and CRBN-based skipping of exon 12; and a large-scale deletion of 42 Mb BET-PROTAC suggests that the proteasome degradation machin- encompassing the CUL2 gene (Fig. 3B–D). To investigate ery downstream of protein ubiquitination remains functional in the copy number of this region, we used SNP minor allele these resistant cells. frequencies across chromosome 10 in the parental, O1R, and O3R cells (Supplementary Fig. S3A). The pattern suggested the Resistance to VHL-based BET-PROTAC is caused by CUL2 loss presence of other chromosomal aberrations preexisting in the due to multiple genomic alterations at the CUL2 locus parental, making the estimation of exact copy number chal- To unravel the mechanism of resistance to VHL-based lenging. Follow-up experiments confirmed the frame shift and BET-PROTAC in the O1R cells, we examined the genomic and the intronic mutations by Sanger sequencing and the copy transcriptional differences such as acquired mutations, gene number loss by TaqMan CNV analysis (Fig. 3E; Supplementary expression changes, and copy-number variation (CNV) Fig. S3C and S3D). Consistent with the abnormalities at the between the parental and O1R cells using whole-exome genomic level, RNA-seq revealed CUL2 to be one of the most sequencing and RNA-seq. No genomic or mRNA expression significantly downregulated genes in the O1R cells compared alterations of BET family proteins or VHL were found in the with parental cells (Supplementary Fig. S4A). qPCR and O1R cells, and the protein level of VHL is comparable in the Western blot analysis further established the significant reduc- parental and the O1R cells (Supplementary Fig. S1D). Although tion of CUL2 mRNA and CUL2 protein in the O1R cells the protein levels of BRD4 and BRD3 are slightly decreased in compared with the parental cells (Fig. 4A and B). In contrast, O1R and O3R cells, there is still significant amount of these two the O3R cells (resistant to the CRBN-based BET-PROTAC) were

www.aacrjournals.org Mol Cancer Ther; 18(7) July 2019 1307

Zhang et al.

Figure 5. Resistance to CRBN-based BET PROTAC is caused by the loss of Cereblon (CRBN) gene due to chromosomal deletion. A, CNV plot of chromosome 3 from O3R cells. B, CRBN deletion conﬁrmation by TaqMan CNV assay. C, qPCR analysis of CRBN mRNA level in parental, O1R, and O3R cells. Data represent mean SD. N ¼ 3. D, Western blot analysis of CRBN level in parental, O1R, and O3R cells. b-Actin was used as loading control. Quantiﬁcation of CRBN protein level is shown below. E, Western blot analysis of CRBN in O3R cells expressing empty vector or CRBN cDNA. b-Actin was used as loading control. F, Dose–response curves for ARV-825 in O3R cells expressing empty vector or CRBN cDNA. Data represent mean SD. N ¼ 3. G, O3R cells expressing empty vector or CRBN cDNA were treated with indicated concentration of ARV-825 for 16 hours and harvested for Western blot analysis. b-Actin was used as loading control.

devoid of these genomic abnormalities in the CUL2 gene, and The hypoxia inducible factors HIF-1A and HIF-2A are physio- expression of CUL2 was comparable with the parental cells logic substrates of the VHL–CRL complex. Under normoxic con- (Fig. 4A and B). ditions, the VHL–CRL complex mediates HIF-1 ubiquitination

1308 Mol Cancer Ther; 18(7) July 2019 Molecular Cancer Therapeutics

Defects in E3 Ligase Complex Cause Resistance to BET-PROTACs

and degradation, consequently preventing HIF-1–dependent Although the protein levels of BRD3 and BRD4 are slightly transcription under normoxia (27). Gene set enrichment analyses decreased in O1R and O3R cells, there is still significant amount revealed significant upregulation and enrichment of the HIF1A of these two proteins, especially BRD4, detected. Considering and HIF2A pathways under normoxic conditions in the O1R cells that BRD4 is the major player in c-Myc regulation among the compared with the parental cells (Supplementary Fig. S4C). qPCR three BRD proteins, this subtle change in BRD4 protein level is analysis also demonstrated the upregulation of HIF-1 target genes not likely to contribute to the acquired drug resistance (28). SOD2, VEGF, and GLUT1 in the O1R cells compared with the Rather, resistance to both classes of BET-PROTACs was primar- parental cells or the O3R cells (Supplementary Fig. S4D). These ilyattributedtogenomicalterationsthatimpactcorecompo- results collectively support that CUL2 loss in the O1R cells nents of the corresponding E3 ligase complexes. In cells that compromises the function of the VHL–CRL complex. were resistant to ARV-771 or ARV-825, the proteasome degra- To determine whether CUL2 loss directly contributes to the dation machinery also remained intact. The preference of resistance to VHL-based BET-PROTAC, we created O1R-derived targeting the E3 ligases components over the proteasome cell lines that express exogenous CUL2 (Fig. 4C). As shown machinery for resistance development is intriguing. We suspect in Fig. 4D and E, overexpression of CUL2 in the O1R cells that the particular vulnerability of E3 ligase components for resensitized the cells to ARV-771 in the cell proliferation assay resistance development may relate to the redundancy and/or and restored ARV-771–induced BET protein degradation, nonessentiality of these components for cell survival, while demonstrating that CUL2 loss is responsible for acquired resis- compromising proteasome function in cells could be lethal or tance in O1R cells. significantly compromise cell fitness.Themultiplegenomic alterations at the CUL2 locus suggest the existence of strong Resistance to CRBN-based BET PROTAC is caused by the loss of selection pressure for the cells to abolish CUL2 function in the CRBN gene due to chromosomal deletion VHL–CRL complex for resistance development. Interestingly, in Whole-exome sequencing and RNA-seq were also carried out cells that are resistant to the CRBN-based BET-PROTAC to determine the genomic and transcriptional differences such ARV-825, resistance was primarily attributed to CRBN deletion as acquired mutations, gene expression changes, and CNV and no genomic/transcription alterations were observed for between the O3R cells and the parental or the O1R cells. While CUL4, the CUL2 equivalent of the CRBN–CRL complex. CRBN no genomic or expression alterations of BET family proteins appears to be a common point of resistance development for were found in the O3R cells, a 12 Mb deletion on chromosome CRBN-targeting agents. It has been shown that deletion of 3 that encompasses the CRBN gene was found in O3R but not CRBN was the primary cause of resistance to iMiDs in myeloma in the O1R cells (Fig. 5A). On the basis of SNP minor allele cells (29). The expression level of CRBN has also been iden- frequencies, copy number was estimated as three copies for tified as a prognostic or predictive biomarker for patients with chromosome 3p in the parental and the O1R cells and one copy multiple myeloma receiving iMiDs (30–32). In addition to in the region containing CRBN of the O3R cells (Supplemen- multiple myeloma, both genetic alterations and differential tary Fig. S3B). The estimated loss of two copies was consistent expression of CUL2 and CRBN have been observed in many with TaqMan CNV analysis, which demonstrated a greater than other cancer types based on The Cancer Genome Atlas dataset 50% reduction in genomic DNA level relative to parental (refs. 33, 34; Supplementary Fig. S5). Interestingly, cell lines (Fig.5B).ConsistentwiththeCRBN gene deletion, RNA-seq with lower CUL2 (Toledo and U266B1) or CRBN (TOV112D differential expression analysis identified CRBN as one of the and HCT116) expression are less sensitive to ARV-771 and most significantly downregulated genes in the O3R cells com- ARV-825, respectively (Supplementary Fig. S6A and S6B). pared with parental cells (Supplementary Fig. S4B). qPCR and Importantly, siRNA-mediated knockdown of CUL2 or CRBN Western blot analysis confirmed significant downregulation of conferred resistance to ARV-771andARV-825inLNCaPcells, the CRBN mRNA and CRBN protein in the O3R cells (Fig. 5C suggesting that there is an association between the expression and D). level of CUL2 or CRBN and sensitivity to VHL- or CRBN-based To determine whether the reduction of CRBN directly contri- PROTACs (Supplementary Fig. S6C and S6D). However, the butes to the resistance to CRBN-based BET-PROTAC in the O3R expression level of these two proteins may not be the only cells, we created O3R-derived cell lines that express high levels of determining factor for PROTAC sensitivity. The fact that CRBN (Fig. 5E). As shown in Fig. 5F and G, increasing CRBN only the OVCAR8 cell line readily developed resistance to these expression in the O3R cells resensitized the cells to ARV-825 in compounds may be attributed to the genetic background of proliferation assays and restored ARV-825–induced BET protein the parental cells. Consistently, OVCAR8 cells harbor an ATM degradation. mutation (p.V613L) and exhibit a mutational signature associated with double-stranded break repair defects with elevated numbers of larger indels (>3 bp; Supplementary Fig. S7). These Discussion underlying genetic attributes may play a role in the differential PROTACs have emerged as a promising novel modality for ability of the OVCAR8 cells to gain resistance to the the development of next-generation therapeutics. In this study, BET-PROTACs compared with other cell lines tested in this we reported that resistance to both VHL- and CRBN-based study. BET-PROTACs can occur in cancer cells following chronic In summary, we report here, for the first time, the mechanisms treatment. However, unlike what is often observed for other of acquired resistance to PROTACs in cancer cell lines. These targeted therapeutics, such as kinase inhibitors, cells that were results highlight the critical involvement of E3 ligase complexes in resistant to ARV-771 and ARV-825 did not contain secondary resistance development and lay the foundation for future mutations that affect compound binding to the target. investigation.

www.aacrjournals.org Mol Cancer Ther; 18(7) July 2019 1309

Zhang et al.

Disclosure of Potential Conflicts of Interest Acknowledgments Y. Shen has ownership interest (including stock, patents, etc.) in AbbVie. No We would like to thank Erin Murphy, Areej Ammar, Elina Dilmukhametova, potential conflicts of interest were disclosed by the other authors. and Ken Idler from the AbbVie Genomic Technologies team for help in preparing the samples for sequencing analysis. Thank you to Arne Grundstad Authors' Contributions from the AbbVie Computational Genomics team for help in processing the exome sequencing data. A special thank you to Relja Popovic from the AbbVie Conception and design: L. Zhang, B. Riley-Gillis, Y. Shen Pharmacogenomics team for key discussions guiding the initiation and design Development of methodology: L. Zhang, B. Riley-Gillis of the project. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Zhang, B. Riley-Gillis Analysis and interpretation of data (e.g., statistical analysis, biostatistics, The costs of publication of this article were defrayed in part by the payment of advertisement computational analysis): L. Zhang, B. Riley-Gillis, P. Vijay page charges. This article must therefore be hereby marked in Writing, review, and/or revision of the manuscript: L. Zhang, B. Riley-Gillis, accordance with 18 U.S.C. Section 1734 solely to indicate this fact. P. Vijay, Y. Shen Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Zhang Received October 2, 2018; revised January 28, 2019; accepted May 2, 2019; fi Study supervision: L. Zhang, Y. Shen published rst May 7, 2019.

References 1. Bosu DR, Kipreos ET. Cullin-RING ubiquitin ligases: global regulation and (ABBV-075/Mivebresib), a potent and orally available bromodomain activation cycles. Cell Div 2008;3:7. and extraterminal domain (BET) family bromodomain inhibitor. 2. Raina K, Crews CM. Chemical inducers of targeted protein degradation. J Med Chem 2017;60:8369–84. J Biol Chem 2010;285:11057–60. 17. Ewels P, Magnusson M, Lundin S, Kaller M. MultiQC: summarize analysis 3. Fischer ES, Bohm K, Lydeard JR, Yang H, Stadler MB, Cavadini S, et al. results for multiple tools and samples in a single report. Bioinformatics Structure of the DDB1-CRBN E3 ubiquitin ligase in complex with thalid- 2016;32:3047–8. omide. Nature 2014;512:49–53. 18. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, et al. 4. Cardote TAF, Gadd MS, Ciulli A. Crystal structure of the Cul2-Rbx1-EloBC- VarScan 2: somatic mutation and copy number alteration discovery in VHL ubiquitin ligase complex. Structure 2017;25:901–11. cancer by exome sequencing. Genome Res 2012;22:568–76. 5. Lu J, Qian Y, Altieri M, Dong H, Wang J, Raina K, et al. Hijacking the E3 19. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. ubiquitin ligase cereblon to efficiently target BRD4. Chem Biol 2015;22: STAR: ultrafast universal RNA-seq aligner. Bioinformatics 2013;29: 755–63. 15–21. 6. Raina K, Lu J, Qian Y, Altieri M, Gordon D, Rossi AM, et al. PROTAC- 20. Liao Y, Smyth GK, Shi W. featureCounts: an efficient general purpose induced BET protein degradation as a therapy for castration-resistant program for assigning sequence reads to genomic features. Bioinformatics prostate cancer. Proc Natl Acad Sci U S A 2016;113:7124–9. 2014;30:923–30. 7. Zhou B, Hu J, Xu F, Chen Z, Bai L, Fernandez-Salas E, et al. Discovery of a 21. Law CW, Chen Y, Shi W, Smyth GK. voom: Precision weights unlock small-molecule degrader of bromodomain and extra-terminal (BET) pro- linear model analysis tools for RNA-seq read counts. Genome Biol teins with picomolar cellular potencies and capable of achieving tumor 2014;15:R29. regression. J Med Chem 2018;61:462–81. 22. Su S, Law CW, Ah-Cann C, Asselin-Labat ML, Blewitt ME, Ritchie ME. 8. Sun Y, Zhao X, Ding N, Gao H, Wu Y, Yang Y, et al. PROTAC-induced BTK Glimma: interactive graphics for gene expression analysis. Bioinformatics degradation as a novel therapy for mutated BTK C481S induced ibrutinib- 2017;33:2050–2. resistant B-cell malignancies. Cell Res 2018;28:779–81. 23. Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, 9. Zhang C, Han XR, Yang X, Jiang B, Liu J, Xiong Y, et al. Proteolysis targeting Getz G, et al. Integrative genomics viewer. Nat Biotechnol 2011;29: chimeras (PROTACs) of anaplastic lymphoma kinase (ALK). Eur J Med 24–6. Chem 2018;151:304–14. 24. Asangani IA, Dommeti VL, Wang X, Malik R, Cieslik M, Yang R, et al. 10. Sun B, Fiskus W, Qian Y, Rajapakshe K, Raina K, Coleman KG, et al. Therapeutic targeting of BET bromodomain proteins in castration-resistant BET protein proteolysis targeting chimera (PROTAC) exerts potent prostate cancer. Nature 2014;510:278–82. lethal activity against mantle cell lymphoma cells. Leukemia 2018; 25. Zuber J, Shi J, Wang E, Rappaport AR, Herrmann H, Sison EA, et al. RNAi 32:343–52. screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. 11. Saenz DT, Fiskus W, Qian Y, Manshouri T, Rajapakshe K, Raina K, Nature 2011;478:524–8. et al. Novel BET protein proteolysis-targeting chimera exerts superior 26. Bai L, Zhou B, Yang CY, Ji J, McEachern D, Przybranowski S, et al. Targeted lethal activity than bromodomain inhibitor (BETi) against post-mye- degradation of BET proteins in triple-negative breast cancer. Cancer Res loproliferative neoplasm secondary (s) AML cells. Leukemia 2017;31: 2017;77:2476–87. 1951–61. 27. Kapitsinou PP, Haase VH. The VHL tumor suppressor and HIF: insights 12. Churcher I. Protac-induced protein degradation in drug discovery: from genetic studies in mice. Cell Death Differ 2008;15:650–9. breaking the rules or just making new ones? J Med Chem 2018;61: 28. Delmore JE, Issa GC, Lemieux ME, Rahl PB, Shi J, Jacobs HM, et al. BET 444–52. bromodomain inhibition as a therapeutic strategy to target c-Myc. Cell 13. Bondeson DP, Mares A, Smith IE, Ko E, Campos S, Miah AH, et al. Catalytic 2011;146:904–17. in vivo protein knockdown by small-molecule PROTACs. Nat Chem Biol 29. Zhu YX, Braggio E, Shi CX, Bruins LA, Schmidt JE, Van Wier S, et al. 2015;11:611–7. Cereblon expression is required for the antimyeloma activity of lenalido- 14. Winter GE, Buckley DL, Paulk J, Roberts JM, Souza A, Dhe-Paganon S, et al. mide and pomalidomide. Blood 2011;118:4771–9. DRUG DEVELOPMENT. Phthalimide conjugation as a strategy for in vivo 30. Broyl A, Kuiper R, van Duin M, van der Holt B, el Jarari L, Bertsch U, et al. target protein degradation. Science 2015;348:1376–81. High cereblon expression is associated with better survival in patients with 15. Bondeson DP, Smith BE, Burslem GM, Buhimschi AD, Hines J, newly diagnosed multiple myeloma treated with thalidomide mainte- Jaime-Figueroa S, et al. Lessons in PROTAC design from selective nance. Blood 2013;121:624–7. degradation with a promiscuous warhead. Cell Chem Biol 2018;25: 31. Bila J, Sretenovic A, Jelicic J, Tosic N, Glumac I, Fekete MD, et al. Prognostic 78–87. significance of cereblon expression in patients with multiple myeloma. 16. McDaniel KF, Wang L, Soltwedel T, Fidanze SD, Hasvold LA, Liu D, Clin Lymph Myeloma Leuk 2016;16:610–5. et al. Discovery of N-(4-(2,4-Difluorophenoxy)-3-(6-methyl-7-oxo- 32. Heintel D, Rocci A, Ludwig H, Bolomsky A, Caltagirone S, Schreder M, et al. 6,7-dihydro-1H-pyrrolo[2,3-c]pyridin-4-yl)phenyl)ethanesulfonamide High expression of cereblon (CRBN) is associated with improved clinical

1310 Mol Cancer Ther; 18(7) July 2019 Molecular Cancer Therapeutics

Defects in E3 Ligase Complex Cause Resistance to BET-PROTACs

response in patients with multiple myeloma treated with lenalidomide and 34. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio dexamethasone. Br J Haematol 2013;161:695–700. cancer genomics portal: an open platform for exploring multidimensional 33. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. cancer genomics data. Cancer Discov 2012;2:401–4. Integrative analysis of complex cancer genomics and clinical proﬁles using 35. Chen Z, Sui J, Zhang F, Zhang C. Cullin family proteins and tumorigenesis: the cBioPortal. Sci Signal 2013;6:pl1. genetic association and molecular mechanisms. J Cancer 2015;6:233–42.

www.aacrjournals.org Mol Cancer Ther; 18(7) July 2019 1311

Acquired Resistance to BET-PROTACs (Proteolysis-Targeting Chimeras) Caused by Genomic Alterations in Core Components of E3 Ligase Complexes

Lu Zhang, Bridget Riley-Gillis, Priyanka Vijay, et al.

Mol Cancer Ther 2019;18:1302-1311. Published OnlineFirst May 7, 2019.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-18-1129

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2019/05/07/1535-7163.MCT-18-1129.DC1

Cited articles This article cites 35 articles, 9 of which you can access for free at: http://mct.aacrjournals.org/content/18/7/1302.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/18/7/1302.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/18/7/1302. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.