MACROD2 Haploinsufficiency Impairs Catalytic Activity of PARP1 and Promotes Chromosome Instability and Growth of Intestinal Tumors

Published OnlineFirst June 7, 2018; DOI: 10.1158/2159-8290.CD-17-0909

Research Article

Anuratha Sakthianandeswaren1,2, Marie J. Parsons1,3, Dmitri Mouradov1,2, Ruth N. MacKinnon4,5, Bruno Catimel1,2, Sheng Liu1,2, Michelle Palmieri1,2, Christopher Love6, Robert N. Jorissen1,2, Shan Li1, Lachlan Whitehead1,2, Tracy L. Putoczki2,7, Adele Preaudet7, Cary Tsui8, Cameron J. Nowell9, Robyn L. Ward10, Nicholas J. Hawkins11, Jayesh Desai1,2,12, Peter Gibbs1,2,12, Matthias Ernst13,14, Ian Street1,2,15, Michael Buchert13,14, and Oliver M. Sieber1,2,3,16

abstract ADP-ribosylation is an important posttranslational protein modification that regulates diverse biological processes, controlled by dedicated transferases and hydrolases. Here, we show that frequent deletions (∼30%) of the MACROD2 mono-ADP-ribosylhydrolase locus in human colorectal cancer cause impaired PARP1 transferase activity in a gene dosage–dependent manner. MACROD2 haploinsufficiency alters DNA repair and sensitivity to DNA damage and results in chromosome instability. Heterozygous and homozygous depletion of Macrod2 enhances intestinal tumorigenesis in ApcMin/+ mice and the growth of human colorectal cancer xenografts. MACROD2 deletion in sporadic colorectal cancer is associated with the extent of chromosome instability, independent of clinical parameters and other known genetic drivers. We conclude that MACROD2 acts as a haploinsufficient tumor suppressor, with loss of function promoting chromosome instability, thereby driving cancer evolution.

SIGNIFICANCE: Chromosome instability (CIN) is a hallmark of cancer. We identify MACROD2 deletion as a cause of CIN in human colorectal cancer. MACROD2 loss causes repression of PARP1 activity, impairing DNA repair. MACROD2 haploinsufficiency promotes CIN and intestinal tumor growth. Our results reveal MACROD2 as a major caretaker tumor suppressor gene. Cancer Discov; 8(8); 1–18. ©2018 AACR.

See related commentary by Jin and Burkard, p. 921.

1Systems Biology and Personalised Medicine Division, The Walter and Australia. 13Olivia Newton-John Cancer Research Institute, Olivia Newton- Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia. John Cancer & Wellness Centre, Heidelberg, Victoria, Australia. 14School 2Department of Medical Biology, The University of Melbourne, Parkville, of Cancer Medicine, LaTrobe University, Heidelberg, Victoria, Australia. Victoria, Australia. 3Department of Surgery, The University of Melbourne, 15Cancer Therapeutics Cooperative Research Centre, Parkville, Victoria, Parkville, Victoria, Australia. 4Victorian Cancer Cytogenetics Service, St Australia. 16Department of Biochemistry & Molecular Biology, Monash Vincent’s Hospital Melbourne, Fitzroy, Victoria, Australia. 5Department University, Clayton, Victoria, Australia. of Medicine, The University of Melbourne (St Vincent’s Hospital), Fitzroy, Note: Supplementary data for this article are available at Cancer Discovery 6 Victoria, Australia. Department of Pathology, Peter MacCallum Cancer Online (http://cancerdiscovery.aacrjournals.org/). Centre, Parkville, Victoria, Australia. 7Inflammation Division, The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia. A. Sakthianandeswaren, M.J. Parsons, and D. Mouradov contributed equally 8Histology Facility, The Walter and Eliza Hall Institute of Medical Research, to this article. Parkville, Victoria, Australia. 9Drug Discovery Biology, The Monash Insti- Corresponding Author: Oliver M. Sieber, The Walter and Eliza Hall Institute tute of Pharmaceutical Sciences, Monash University, Parkville, Victoria, of Medical Research, 1G Royal Parade, Parkville, VIC 3052, Australia. Australia. 10Office of the Deputy Vice-Chancellor (Research), The Univer- Phone: 613-9345-2885. E-mail: [email protected] 11 sity of Queensland, Brisbane, Queensland, Australia. Faculty of Medicine, doi: 10.1158/2159-8290.CD-17-0909 The University of Queensland, Brisbane, Queensland, Australia. 12Depart- ment of Medical Oncology, Royal Melbourne Hospital, Parkville, Victoria, ©2018 American Association for Cancer Research.

OF1 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

INTRODUCTION establishes and amplifies the DNA-damage signal, recruiting repair factors and activating effector proteins involved in the ADP-ribosylation is a widespread posttranslational pro- DNA-damage response, including master regulators such tein modification at DNA lesions, which is governed by the as ATM, ATR, and DNA-dependent protein kinase (8). PAR activities of specific transferases and hydrolases. This modifi- synthesis is reversed by degradation by poly(ADP-ribose) cation regulates various biological processes, including DNA- glycohydrolase; however, removal of the terminal autoinhibi- damage response, chromatin reorganization, transcriptional tory mono-ADP-ribose from PARP1 to cause reactivation regulation, apoptosis, and mitosis (1–3). Genome-wide DNA requires MACROD2 recruitment and enzymatic activity (7). copy-number analyses across human cancers have revealed MACROD2 phosphorylation by ATM acts as a negative feed- common focal deletions of the MACROD2 mono-ADP- back loop, triggering MACROD2 nuclear export upon DNA ribosylhydrolase locus on chromosome 20p12.1 in multiple damage, thus temporally restricting its recruitment to DNA malignancies, including gastric and colorectal cancers (4, 5). lesions (9). However, the locus is considered a tissue-specific fragile site MACROD2 has further been implicated as a regulator of (6), and whether MACROD2 aberrations contribute to car- glycogen synthase kinase 3-beta (GSK3β), indicating a role in cinogenesis is unknown. the modulation of WNT signaling (10). MACROD2 reverses Recent studies have identified MACROD2 as a regulator PARP10-catalyzed mono-ADP-ribose-mediated inhibition of of PARP1, a principal sensor of DNA single-strand breaks GSK3β, activating GSK3β to phosphorylate β-catenin in the (SSB) and double-strand breaks (DSB; ref. 7). Following context of a protein complex with adenomatous polyposis coli binding to sites of DNA nicks or breaks, PARP1 polymer- (APC), axin, and other components. Phosphorylation targets izes PAR chains onto histones and other proteins, includ- β-catenin for ubiquitination and proteasomal degradation, ing itself. This auto- and substrate-PARylation by PARP1 preventing its translocation to the nucleus and interaction

AUGUST 2018 CANCER DISCOVERY | OF2

RESEARCH ARTICLE Sakthianandeswaren et al. with members of the T-cell factor/lymphoid enhancer factor microdeletions also exhibited aberrant MACROD2 tran- (TCF/LEF) transcription factor family to activate expression scripts lacking one or more exons, indicating pathogenicity of WNT target genes (11). in these cases. In five cell lines with multiple heterozygous These observations raise the possibility that somatic exonic or intronic MACROD2 aberrations, no wild-type tran- genomic aberrations in MACROD2 contribute to cancer devel- script was detected consistent with deletions affecting both opment through impairing the DNA-damage response or alleles. Aberrant transcripts were predicted to result in pre- promoting aberrant WNT signaling. Here, we present genetic, mature MACROD2 protein truncation in 28% (7 of 26) of biochemical, and functional data that reveal MACROD2 as cases, or in-frame exonic deletions involving the catalytic a caretaker tumor suppressor gene essential for the mainte- macrodomain of the protein in 72% (19 of 26) of cases (Sup- nance of cancer genome integrity. plementary Table S4). MACROD2 Somatic Mutation Burden in Human RESULTS Colorectal Cancer Focal Deletions at Chromosome 20p12.1 Target We next examined the contribution of somatic mutations MACROD2 in Human Colorectal Cancer to the burden of MACROD2 aberrations in colorectal cancer. To comprehensively characterize the human cancer types We sequenced all exons of MACROD2 in 102 sporadic colorec- in which the MACROD2 locus is subject to focal deletions tal cancers and 53 colorectal cancer cell lines. Detected vari- at chromosome 20p12.1, we analyzed The Cancer Genome ants were verified to not correspond to known SNPs, and for Atlas (TCGA)–derived DNA copy-number data from 10,575 primary tumors were confirmed to be somatically acquired by tumors representing 32 malignancies using Genomic Iden- sequencing of matched normal tissue. Results were combined tification of Significant Targets in Cancer (GISTIC; ref. with TCGA-derived mutation data for colorectal cancer (n = 12). Colorectal adenocarcinoma (COAD/READ) exhibited 536; Supplementary Table S5). the strongest evidence for recurrent focal DNA copy-num- Somatic MACROD2 mutations were found to be of low ber loss at MACROD2 (q = 5.89E−78), but targeting was prevalence, with 14 missense mutations detected in a total also observed in stomach adenocarcinoma (STAD), cervical of 691 cases. MACROD2 missense mutations were localized squamous cell carcinoma and endocervical adenocarcinoma throughout the protein with no apparent clustering (Fig. (CESC), esophageal carcinoma (ESCA), uterine corpus endo- 1D). Fifty percent (7 of 14) of the mutations were assigned metrial carcinoma (UCEC), uterine carcinosarcoma (UCS), as pathogenic based on the consensus from two in silico lung adenocarcinoma (LUAD), liver hepatocellular carci- prediction algorithms (PolyPhen and SIFT; refs. 13, 14; Sup- noma (LIHC), and thyroid carcinoma (THCA; Fig. 1A; Sup- plementary Table S5). Of the seven missense variants that plementary Table S1). mapped to the available crystal structure of the MACROD2 To fine-map the spectrum of somaticMACROD2 deletions macrodomain (PDB: 4IQY; ref. 7), three variants (Ala91Ser, in human colorectal cancer, we estimated absolute DNA copy- Gly100Ser, Thr187Ala) were predicted to interfere with bind- number states at chromosome 20p12.1 for 616 TCGA tumors ing of the cocrystalized ADP-ribose moiety to the catalytic and an in-house cohort of 651 tumors using single-nucleotide pocket (Fig. 1E). polymorphism (SNP) array data and OncoSNP (top, Fig. 1B; Considering the combined somatic copy-number altera- Supplementary Fig. S1; Supplementary Table S2). We detected tion (SCNA), transcript sequencing, and mutation data for frequent heterozygous and homozygous loss of variable size our colorectal cancer cell line, 61.3% (19 of 31) of cases with on 20p12.1, involving the MACROD2 locus in 27.9% (172/616) MACROD2 aberrations showed evidence of “two hits,” con- of TCGA and 32.3% (210/651) of in-house colorectal cancers. sistent with a tumor suppressor role. Nevertheless, 38.7% Deletion frequencies were similar across tumor stages (P = (12 of 31) of cell lines with MACROD2 aberrations exhibited 0.084). The majority of loss events (71.5%; 273 of 382) were only “one hit,” suggesting that inactivation of a single copy intragenic microdeletions in MACROD2 excluding any sur- of MACROD2 may be sufficient to impart a clonal growth rounding genes, with most mapping to a region from exons 4 advantage during tumor development. to 5. A similar deletion spectrum was also observed for a panel of 53 human colorectal cancer cell lines (bottom, Fig. 1B; MACROD2 Deficiency Enhances Growth of Supplementary Tables S3 and S4), with validation of detected Intestinal Tumors in a Haploinsufficient Manner homozygous exonic MACROD2 deletions by the absence of The majority (∼70%) of human sporadic colorectal can- reads in whole-exome sequencing (WES) data (Fig. 1C). cers are initiated by biallelic mutations in the APC tumor Although 50.5% (138 of 273) of MACROD2 microdeletions suppressor gene (15, 16). To test whether MACROD2 defi- involved exonic regions, 49.5% (135 of 273) were limited ciency could promote intestinal tumorigenesis, we crossed to introns. To test whether intronic deletions influenced Macrod2 knockout mice obtained from the Knockout Mouse MACROD2 gene expression, we sequenced MACROD2 mRNA Project Repository (The Jackson Laboratory; Fig. 2A), transcripts in our colorectal cancer cell lines, thereby avoid- which develop normally into adulthood (see Supplemen- ing contamination from nontumor cells that would other- tary Data), with the ApcMin/+ mouse model (17). ApcMin/+ wise be present in patient samples. Cell lines wild-type for mice harbor a truncating germline mutation in Apc and MACROD2 (n = 20) or with exonic microdeletions (n = 19) intestinal tumors arise spontaneously from loss of het- were verified to express full-length or deletion transcripts, erozygosity of the wild-type Apc allele (17), a mechanism respectively (Fig. 1C; Supplementary Fig. S2; Supplementary found in human sporadic colorectal cancer and familial Table S4). Moreover, 66.7% (4 of 6) of cell lines with intronic adenomatous polyposis (18, 19).

OF3 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

AB D 159 240 247 388 425 GISTIC −log (q-value) 10 20p12.1 Macro domain Glutamine-rich region 0 20 40 60

COAD/READ 23.8% STAD 10.4% * CESC 12.5% * ESCA * 17.4% UCEC *7.2% UCS *19.6% LUAD * 26.6% E LIHC * 7.7% THCA * 0.4% UVM * THYM TGCT SKCM A91S R17C

SARC Colorectal cancers PRAD PCPG PAAD A149S G100S OV MESO LUSC LGG LAML KIRP T187A KIRC D33Y KICH HNSC Cell lines GBM DLBC T58I CHOL BRCA BLCA MACROD2 ACC

C SNP array Whole-exome sequencing RNA sequencing Exon 4 Exon 6

GGAGGAGGTGATGTCATCCA Exon 4 Exon 6 ∆ INTRONIC GGAGGAGGTGATGTCATCCA CCTGCAAAATATGTCATCCA

Exon 4 Exon 6

GGAGGAGGTGATGTCATCCA

Exon 4 Exon 6 ∆ EXONIC HET

LIM1215 COLO320 SW48 HRA19 GGAGGAGGTGATGTCATCCA

Exon 3 Exon 5

SW480 GTCAATGCCGTGGATGGCTG Exon 3 Exon 8

∆ EXONIC HOM GTCAATGCCGGCTTTCCCAA GGCATTTATGGCTTTCCCAA LIM2405

1 2 3 45 678910 11 12 13 14 15 16 17 1 2 3 45 678910 1112 13 14 15 1617 MACROD2

Figure 1. Focal MACROD2 deletions, somatic mutations, and transcript-level consequences in human colorectal cancer. A, GISTIC analysis across 32 TCGA-analyzed tumor types identifying focal SCNAs targeting MACROD2 in colorectal cancer cancers (COAD/READ) and other malignancies of the stomach, esophagus, cervix, uterus, lung, liver, and thyroid (STAD, ESCA, CESC, UCEC, UCS, LUAD, LIHC, and THCA); *, q < 0.05 for GISTIC q statistic. The tumor types corresponding to the TCGA study abbreviations are listed in Supplementary Table S1. B, SNP array segmentation map for 651 in house primary colorectal cancers (top) and 53 colorectal cancer cell lines (bottom) showing the area surrounding MACROD2 on chromosome 20. Cases are sorted by region of loss at the MACROD2 locus; heterozygous (light blue bars) and homozygous (dark blue bars) deletions are labeled. Position and exons of the MACROD2 gene are indicated (red bars). Surrounding genes are indicated with gray arrows. C, SNP array, WES, and mRNA sequencing data for representative human colorectal cancer cell lines relating both exonic and intronic MACROD2 microdeletions to aberrant transcripts lacking one or more exons are shown. Red and green bars in SNP array plots are indicative of homozygous and heterozygous deletions, respectively. D, Somatic MACROD2 mutations detected in colorectal cancer tumors/cell lines from this study and in the TCGA-analyzed colorectal cancer tumors (blue triangles). E, Structural analysis of somatic mutations in the MACROD2 macrodomain (PDB = 4IQY). Ribbon diagram is shown (with alpha helices and beta sheets). Mutations are shown in orange and labeled.

AUGUST 2018 CANCER DISCOVERY | OF4

RESEARCH ARTICLE Sakthianandeswaren et al.

A Wild-type Exon 2 Exon 3

Macrod2−/− lacZ-p(A) loxP hUbCpro neor-p(A) loxP

B Apc Min/+ Macrod2+/+ ApcMin/+ Macrod2−/+ ApcMin/+ Macrod2−/−

C * * *** ** * ** 150 30 400 25 2 300 20 100 /mouse 2 15 200 10 50 umor number/mouse 100 area >3 mm Tumor T 5 Tumor area mm Tumor

0 0 0 − / + − / − / + − / − / + − / + / + / + / Min/ + Min/ + Min/ + Min/ + Min/ + Min/ + Min/ + Min/ + Min/ + Apc Apc Apc Apc Apc Apc Apc Apc Apc Macrod2 Macrod2 Macrod2 Macrod2 Macrod2 Macrod2 Macrod2 Macrod2 Macrod2

DEHCT116 iso LIM2405 MACROD2 −/− 100

) 300 800 3 MACROD2 +/+ MACROD2 * P = 0.019 MACROD2 −/+ GFP 50 MACROD2 −/− 600 * P = 0.009 200 rcent survival

lume (mm 400

Pe *** Apc Min/+Macrod2+/+ 100 ** Apc Min/+Macrod2−/+ 200 Apc Min/+Macrod2−/− 0 vo Tumor 0 0 50 100 150 200 0510 15 0510 15 Time (days) Time (days)

OF5 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

We observed a significant increase in the multiplicity of MACROD2 Deficiency Does Not Affect WNT intestinal adenomas in F2 ApcMin/+/Macrod2−/− and ApcMin/+/ Signaling in APC or CTNNB1 Mutated Intestinal Macrod2−/+ mice as compared with ApcMin/+/Macrod2+/+ mice Tumor Cells (35 animals per group), with an average number of 81.8, 81.3, and 65.6 polyps, respectively (P < 0.012 and P < 0.007, respec- MACROD2 is a positive regulator of GSK3β, which in turn tively; Fig. 2B and C). Concomitantly, total adenoma burden controls cellular β-catenin levels, nuclear relocalization, and (indicated by the overall area of tumors) was significantly engagement of downstream WNT signaling (10). However, the increased in both the ApcMin/+/Macrod2−/− and the ApcMin/+/ majority of human colorectal cancers carry inactivating muta- Macrod2−/+ mice (P < 0.033 and P < 0.044, respectively), with tions in APC, a central member of the β-catenin destruction evidence for a significant gene dosage–dependent enhance- complex, or constitutively activating mutations in CTNNB1 ment when considering large (>3 mm) adenomas (2.5-fold (encoding β-catenin; ref. 20). Whether the regulatory role of and 2-fold increase, P < 0.001 and P < 0.019, respectively; GSK3β is maintained in this context is uncertain. To assess the Fig. 2C). Accordingly, the median number of neoplastic cells impact of MACROD2 deficiency in human colorectal cancer per adenoma as assessed by image analysis of hematoxylin cells with WNT pathway mutations, we evaluated the expres- and eosin (H&E)–stained intestinal sections increased from sion of WNT target genes (CD44, AXIN2, and DVL1; TCF/LEF 1,212.5 in ApcMin/+/Macrod2+/+ mice to 1,808 and 1,865.5 in reporter assay) and β-catenin relocalization in representative ApcMin/+/Macrod2−/+ and ApcMin/+/Macrod2−/− mice, respec- CTNNB1–mutated (HCT116-MACROD2−/−, HCT116-MACROD2−/+ tively (P < 0.001 for both comparisons; Supplementary Fig. and HCT116-MACROD2+/+, LIM1215-MACROD2−/+) and APC- S3A–S3B). Overall survival of aging ApcMin/+/Macrod2−/− and mutated cell lines (LIM2405-MACROD2−/−, COLO320-MACROD2−/+, ApcMin/+/Macrod2−/+ mice as compared with ApcMin/+/Macrod2+/+ and LOVO-MACROD2+/+). For LIM1215, LIM2405, COLO320, mice was significantly decreased, with a median survival time and LOVO cell lines, MACROD2 was either reconstituted of 109.5, 126, and 142 days, respectively (P < 0.006 and P < or suppressed depending on their MACROD2 deletion status 0.013, respectively; Fig. 2D). No tumors were found in the (Supplementary Fig. S5A–S5B). HEK293T cells were used as intestines of animals from each of the three control groups APC and CTNNB1 wild-type control cells. (Apc+/+/Macrod2+/+, Apc+/+/Macrod2−/+, and Apc+/+/Macrod2−/−; Consistent with previous reports, MACROD2 suppression 15 animals per group). in HEK293T cells resulted in significant induction of WNT To determine whether MACROD2 deficiency could also target gene expression, increased TCF/LEF reporter activity, enhance growth of human colorectal cancer cells in vivo, and β-catenin nuclear relocalization in response to WNT3A we generated xenografts of MACROD2 wild-type, heterozy- stimulation (Supplementary Fig. S7A–S7F). However, in gous, and homozygous knockout HCT116 cells (β-catenin colorectal cancer cells with APC or CTNNB1 mutations, mutated) using CRISPR/Cas9 technology (Supplementary MACROD2 suppression or reconstitution did not affect the Figs. S4A–S4B and S5A–S5B). Isogenic HCT116-MACROD2−/− expression of WNT target genes (Fig. 3A), TCF/LEF reporter and HCT116-MACROD2−/+ cells exhibited significantly activity (Fig. 3B), or β-catenin subcellular localization (Fig. 3C increased establishment rates in CBA athymic nude mice and D). In addition, no association between MACROD2 dele- as compared with HCT116-MACROD2+/+ cells (P < 0.002 for tion status and WNT target gene expression was observed when both comparisons, Fig. 2E), and HCT116-MACROD2−/− cells integrating SNP array and RNA sequencing (RNA-seq) data further displayed increased tumor growth over a 20-day period from 208 TCGA-analyzed COAD/READ samples (Fig. 3E). (P = 0.007). Increased tumor growth was similarly observed Consistent with the findings in human colorectal cancer for shRNA-mediated MACROD2 knockdown in HCT116- cells, evaluation of intestinal polyps from ApcMin/+/Macrod2−/−, MACROD2+/+ cells (Supplementary Fig. S6A–S6B). Con- ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2+/+ mice showed versely, reconstitution of GFP-tagged wild-type MA CROD2 similar expression of WNT target genes by RNA-seq analysis in homozygous deleted LIM2405 cells (APC mutated) and qRT-PCR (Fig. 3F and G), and similar patterns of subcel- resulted in a significant reduction in tumor growth as com- lular β-catenin localization (Fig. 3H and I). Taken together, pared with GFP-control transfected cells (P < 0.001; Fig. 2E; these data indicate that MACROD2 deficiency has no major Supplementary Fig. S5A–S5B). Collectively, these results impact on WNT signaling—within the level detectable by the show that MACROD2 deficiency promotes the growth of assays used—in the context of APC- or CTNNB1–mutated intestinal tumors in a haploinsufficient manner. intestinal tumor cells.

Figure 2. MACROD2 deficiency enhances intestinal tumorigenesis inApc Min/+ mice and tumor xenograft growth of human colorectal cancer cell lines. A, Schematic diagram showing the strategy to disrupt the mouse Macrod2 gene via homologous recombination. Macrod2 knockout mice (Macrod2tm1.1(KOMP)Vlcg) were obtained from the Knockout Mouse Program Repository at The Jackson Laboratory. B, Representative sections of proximal small intestines from 130-day-old ApcMin/+/Macrod2−/−, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2+/+ mice stained with β-catenin to highlight tumors (black arrowheads; scale bar, 3 mm). C, Quantification of the number, area, and size distribution of tumors in the small intestines and colon ofApc Min/+/ Macrod2−/−, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2+/+ mice at 130 days. Data are means ± SEM of 35 mice per cohort for tumor number and 15 mice per cohort for tumor area/size measurements. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t test). D, Survival plot for aging ApcMin/+/Macrod2−/−, ApcMin/+/ Macrod2−/+ and ApcMin/+/Macrod2+/+ mice. n ≥ 10 mice per genotype. P values shown are for the Wald test. E, Human tumor xenograft establishment rate and growth in CBA athymic nude mice injected subcutaneously with isogenic HCT116-MACROD2−/−, HCT116-MACROD2−/+ and HCT116-MACROD2+/+ cells, or LIM2405-MACROD2−/− cells reconstituted with wild-type MACROD2 or GFP control. Data are means ± SEM of 12 or 8 tumors per condition representative of duplicate experiments, respectively. **, P < 0.01; ***, P < 0.001 (Compare Groups of Growth Curves test).

AUGUST 2018 CANCER DISCOVERY | OF6

RESEARCH ARTICLE Sakthianandeswaren et al.

A CTNNB1 mutant B E TCGA primary CRCs HCT116 iso LIM1215 MACROD2 −/+ HCT116 MACROD2 +/+ MACROD2 ATOH1 10 12 AXIN2 ns ns BIRC5 ns 10 BMP4 ns ns ns 8 CCND1 ns CD44 CLDN1 1 EFNB1 1 ETS1 4 FGF18 FOSL1 CTNNB1 mutant pression GAST HNF1A 0 0 0 ID2 CD44 AXIN2 DVL1 CD44 AXIN2 DVL1 siNEG siMACROD2 JAG1 JUN

gene ex APC mutant L1CAM +/+ +/+ LEF1 ay LOVO MACROD2 LIM2405 MACROD2 −/− LOVO MACROD2 5.0 MET ns 100 ns ns MMP7 100 MSL1 MYC

ns TCF/LEF reporter assay MYCBP ns 10 NRCAM WNT pathw WNT3A stimulated/unstimulated * WNT3A stimulated/unstimulated 2.5 PLAUR 10 ns PRARD TCF4

1 APC mutant VEGFA

/+ /+ /− 1 0 0.0 + − − CD44 AXIN2 DVL1 CD44 AXIN2 DVL1 siNEG siMACROD2 / / / MACROD2 + + MACROD2 − + MACROD2 − − Macrod2 Macrod2Macrod2 GFP MACROD2 shNEG shMACROD2

C D CTNNB1 mutant F Apc Min/+ Macrod2 LIM2405 HCT116 iso LIM1215 MACROD2 −/+ Macrod2 2 1.5 Atoh1 Axin2 ns ns Birc5 Bmp4 1 Bmyc Ccnd1 − WNT3A 1 Cd44

GFP Cldn1 0.5 Efnb1 Ets1 Fgf18 Fosl1

+ WNT3A 0 0 Hnf1a +/+ −/+ −/− MACROD2 MACROD2 MACROD2 GFP MACROD2 Id2 Jag1 APC mutant Jun L1cam +/+ −/− LOVO MACROD2 LIM2405 MACROD2 Lef1 1.5 2 Met ns Mmp7 ns Msl1 − WNT3A Mycbp Nuclear β -catenin intensity 1 Nrcam Plaur WNT3A stimulated/unstimulated 1 Ppard Tcf4

MACROD2-GFP 0.5 Vegfa + WNT3A

0 0 Macrod2 +/+ Macrod2 −/+ Macrod2 −/− shMACROD2 shNEG GFP MACROD2 GH I Apc Min/+ Macrod2+/+ Apc Min/+ Macrod2−/+ Apc Min/+ Macrod2−/− 100 ns ns ns 1.5 ns pression 10 ns 1 1 gene ex

ay 0.1 0.5 β -catenin

0 0 Cd44 Axin2 Dvl1 Min/+ Min/+ Min/+

% Nuclear β -catenin cells Apc Apc Apc / / +/+ −/+ −/− WNT pathw +/+ − + − + Macrod2 Macrod2 Macrod2 WNT3A stimulated/unstimulated Macrod2 Macrod2 Macrod2

Figure 3. MACROD2 deficiency does not affect WNT signaling inAPC- or CTNNB1-mutated intestinal tumor cells. A, Induction of WNT target gene expression in CTNNB1-mutated (HCT116-MACROD2−/−, HCT116-MACROD2−/+ and HCT116-MACROD2+/+, LIM1215-MACROD2−/+) and APC-mutated colorectal cancer cell lines (LIM2405-MACROD2−/−, COLO320-MACROD2−/+, and LOVO-MACROD2+/+) in response to stimulation with WNT3A for 24 hours. For APC-mutated cell lines, MACROD2 was either reconstituted or stably suppressed using shRNA. Data are means ± SEM for 6 replicates representative of duplicate experiments. ns, not significant; *,P < 0.05 (Student t test). B, TCF/LEF reporter activity for stable HCT116-MACROD2+/+ and LOVO-MACROD2+/+ cells transfected with siRNA-MACROD2 or siRNA-control in response to WNT3A stimulation for 48 hours. Twenty thousand events were acquired by flow cytometry. Data are means± SEM for 6 replicates representative of duplicate experiments. ns, not significant (Studentt test). C, Representative immunofluorescence images of β-catenin relocalization in LIM2405-MACROD2−/− cells reconstituted with MACROD2-GFP or GFP- control in response to WNT3A stimulation for 16 hours are shown. β-catenin, red; DAPI; blue; scale bar, 20 μm. D, Quantification of nuclearβ -catenin relocalization in MACROD2 reconstituted or stably suppressed colorectal cancer cells after WNT3A stimulation for 16 hours. Data are means ± SEM of >100 cells representative of duplicate experiments. ns, not significant (Studentt test). E–F, RNA-seq heat maps of WNT target gene expression for (E) TCGA-analyzed colorectal cancer (CRC; COAD/READ, n = 208) samples according to MACROD2 deletion status and (F) intestinal tumors from 130-day-old ApcMin/+/Macrod2−/−, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2+/+ mice. Expression levels were similar between genotypes for both cohorts. BH-adjusted P > 0.05, limma/voom. G, Relative mRNA abundance for β-catenin target genes in intestinal tumors from 130-day-old ApcMin/+/Macrod2−/−, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2+/+ mice by qRT-PCR. Data are means ± SEM of 6 replicates representative of duplicate experiments. ns, not significant (Studentt test). H, Representative sections of intestinal tumors from 130-day-old ApcMin/+/Macrod2−/−, ApcMin/+/Macrod2−/+, and ApcMin/+/ Macrod2+/+ mice stained with β-catenin (scale bar, 1 mm) and (I) quantification ofβ -catenin cellular localization. Data are means ± SEM of 10 tumors/ mice for 6 mice per cohort. ns, not significant (Studentt test).

OF7 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

MACROD2 Haploinsufficiency Causes Repression MACROD2 Haploinsufficiency Alters DNA Repair of PARP1 Transferase Activity and Sensitivity to DNA Damage The macrodomain of MACROD2 has previously been Impairment of PARP1 activity caused by loss of MAC- shown to be recruited to sites of laser-induced DNA damage ROD2 regulatory function is predicted to result in altered in a PARP1-dependent manner, and to regulate PARP1 activ- DNA repair and sensitivity of cells to genotoxic stress– ity through the removal of terminal autoinhibitory mono- induced DNA damage. Accordingly, comet assays on isogenic ADP-ribose (7, 21). HCT116-MACROD2−/− and HCT116-MACROD2−/+ cells iden- Given that MACROD2 microdeletions in human colorectal tified a significant, gene dosage–dependent increase for both cancer often produce in-frame exonic deletions involving DSBs (neutral comet assay) and SSBs (alkaline comet assay) as the macrodomain, we tested whether these resulted in a loss compared with HCT116-MACROD2+/+ when challenged with of protein recruitment to sites of DNA damage induced by either IR or doxorubicin (P < 0.010 for all comparisons; Fig. near-infrared (NIR) laser microirradiation. Corresponding 4E–F; Supplementary Fig. S10A–S10B). Equivalent results deletion transcripts were cloned from colorectal cancer cell were obtained when analyzing mouse embryonic fibroblasts lines and expressed in HeLa cells and shown to produce (MEF) generated from Macrod2−/−, Macrod2−/+, and Macrod2+/+ truncated proteins (MACROD2Δex1-8, MACROD2Δex4, mice (Supplementary Fig. S11A–S11B). Conversely, recon- MACROD2Δex5, and MACROD2Δex4-5; Supplementary stitution of wild-type MACROD2 in colorectal cancer cell Fig. S8). As anticipated, GFP-tagged MACROD2 deletion lines with homozygous MACROD2 deletions (LIM2405 and proteins lacking an intact macrodomain were not recruited SW480) or heterozygous MACROD2 deletions (COLO320 to DNA lesions, whereas full-length protein (MACROD2- and LIM1215) resulted in a significant reduction of DSBs and GFP) and a control truncated protein with an intact macro- SSBs (P < 0.001; Fig. 4G and H; Supplementary Fig. S12A and domain (MACROD2Δex11-17) exhibited rapid accumulation S10B). Reconstitution with MACROD2Δex4 protein lack- (Fig. 4A). ing the catalytic macrodomain did not result in rescue of To determine whether MACROD2 loss results in an MA CROD2-deleted cell lines (Supplementary Fig. S12C). increase of PARP1 autoinhibitory mono-ADP-ribosylation Pharmacologic inhibition of PARP1 (olaparib) increased following induction of DNA damage, we treated our isogenic DSBs and SSBs in HCT116-MACROD2+/+ and, to a lesser HCT116 MACROD2+/+, MACROD2−/+, and MACROD−/− cells extent, HCT116-MACROD2+/- cells, but not in HCT116- with the DNA topoisomerase II inhibitor doxorubicin and MACROD2−/− cells (Fig. 4I). Conversely, rescue of MACROD2- probed immunoprecipitated PARP1 with anti-mono(ADP- deleted cell lines with wild-type MACROD2 was blocked with ribose) antibody. As anticipated, PARP1 mono-ADP- PARP1 inhibition (olaparib), and treatment did not increase ribosylation levels (relative to total PARP1) were elevated DSBs and SSBs in MACROD2-deleted GFP-control cells (Fig. with MACROD2 loss posttreatment with doxorubicin in a 4J). Similar comet assay results for rescue and nonadditivity gene dosage–dependent manner (Fig. 4B). Conversely, res- were observed for an alternative PARP1 inhibitor (veliparib; cue of homozygous or heterozygous MACROD2-deleted cell Supplementary Fig. S13). lines (LIM2405-MACROD2−/− and LIM1215-MACROD2−/+) by The impact of MACROD2 haploinsufficiency on DNA expression of wild-type MACROD2 resulted in a decrease in repair was further demonstrated through differential for- mono-ADP-ribosylated PARP1 (Fig. 4B). mation of DNA damage–induced foci of phosphorylated To quantify the impact of homozygous and heterozygous H2AX (γ-H2AX) and phosphorylated ATM (pATM), mark- MACROD2 deficiency on PARP1 transferase activity in human ers of DSBs. Human colorectal cancer or MEF cells with colorectal cancer cells, we measured global protein PARylation homozygous or heterozygous MACROD2 deletions developed levels (Fig. 4C) and incorporation of biotinylated PAR onto his- significantly moreγ -H2AX foci as compared with MACROD2 tone proteins (Fig. 4D) following DNA damage with doxorubicin. wild-type cells when challenged with IR or doxorubicin, Isogenic HCT116-MACROD2−/− and HCT116-MACROD2−/+ whereas the inverse was observed upon MACROD2 recon- cells showed significant attenuation of both overall protein stitution in MACROD2−/− or MACROD2−/+ cells (Fig. 5A–B; PARylation and PAR incorporation onto histones as compared Supplementary Fig. S14A and S14B). Similar results were with HCT116-MACROD2+/+ cells (P < 0.010). Conversely, res- obtained for foci of pATM (Supplementary Fig. S15A and cue of homozygous or heterozygous MACROD2-deleted cell S15B). Reconstitution with MACROD2Δex4 protein again lines (LIM2405-MACROD2−/− and LIM1215-MACROD2−/+) by did not rescue γ-H2AX foci formation in MACROD2-deleted expression of wild-type MACROD2 resulted in a significant cell lines (Supplementary Fig. S16A and S16B). Moreover, increase in global protein and histone PARylation (P < 0.005 for a significant dose-dependent increase inγ -H2AX staining all comparisons). Similar results for PAR incorporation onto with MACROD2 loss was observed for intestinal polyps from histones were observed when cells were exposed to γ-irradiation ApcMin/+/Macrod2−/+ and ApcMin/+/Macrod2−/− mice as compared (Supplementary Fig. S9). Protein PARylation was reflective of with ApcMin/+/Macrod2+/+ mice (Fig. 5C and D). PARP1 activity, as the addition of PARP1 inhibitor (olaparib) Impairment of DNA repair was reflected by increased abrogated PAR incorporation in MACROD2 wild-type cells and tumor cell sensitivity to DNA damage. In clonogenic assays, MACROD2 rescued cells (Fig. 4C). isogenic HCT116-MACROD2−/− cells showed significantly Together, these findings indicate thatMACROD2 haplo- reduced viability as compared with HCT116-MACROD2+/+ insufficiency, due to whole-gene or catalytic macrodomain cells when treated with IR or doxorubicin (P < 0.010 for deletions, causes impaired PARP1 transferase activity in the all comparisons; Supplementary Fig. S17A–A17B). HCT116- context of human colorectal cancer cells. MACROD2−/+ cells showed reduced viability for IR (P < 0.05)

AUGUST 2018 CANCER DISCOVERY | OF8

RESEARCH ARTICLE Sakthianandeswaren et al.

A BDDox MACROD2-GFP LIM2405 LIM1215 2 *** / / * 1 59 240 247 388 425 MACROD2− − MACROD2− + * * MACROD2 +/+ 1 MACROD2 ×11-17del-GFP HCT116 iso MACROD2 −/+

591 240 242 HCT116 iso + / − / + − / + / − / + − / 0 −/− GFP MACROD2 GFP GFP GFP MACROD2 MACROD2 MACROD2 MACROD2 Dox − −−+++ −−++ −−++M.W µ g protein) 3

MACROD2 ×1-8del-GFP mADPr − / −110 240192 247 388 425 PARP1 −110 2 ** MACROD2 ×4del-GFP 1 LIM2405 1 91101 240 247 388 425 C LIM2405 LIM1215 HCT116 iso MACROD2−/− MACROD2−/+ 0 MACROD2 GFP MACROD2 ×5del-GFP +/+−/+ −/− GFP MACROD2MGFP ACROD2 Olaparib −− +− −+−−+ −− +− −+ −−+−− + 1 102 140 240 247 388 425 M.W 1.5 MACROD2 Dox −+ +− ++−++ −+ +− ++ −++−++ − / + −260 1 * MACROD2 ×4-5del-GFP pADPr activity (mUnits/ PARP1 191180 240 247 388 425 0.5

−110 LIM1215 PARP1 0 20 200 Macro domain −110 MACROD2 Vinculin 2 4 Time (s) Glutamine-rich region −110 Time (h) E F HCT116 iso I HCT116 iso MACROD2 +/+ HCT116 iso 500 **** 80 **** +/+ MACROD2 +/+ MACROD2 −/+ MACROD2 −/− **** MACROD2 *** ns + olaparib **** **** **** −/+ IR 40 MACROD2 250 **** *** MACROD2 −/+ il moment 20 + olaparib Ta Dox MACROD2 −/− 0 0 MACROD2 −/− Control IR DoxDox + olaparib

GHMACROD2 −/− J LIM2405 SW480 LIM2405 MACROD2 −/− MACROD2 GFP 300 **** ns **** LIM2405 100 **** IR SW480 200 **** ****

il moment **** il moment 50 LIM2405 100 Ta Ta SW480 Dox 0 0 MACROD2 −/+ Control IR Dox Control IR DoxDox LIM1215 COLO320 SW480 MACROD2 −/− MACROD2 GFP 300 LIM1215 60 **** **** ns **** IR 200 **** COLO320 40 **** ****

100 il moment LIM1215 il moment 20 Ta Ta Dox COLO320 0 0 Control IR Dox Control IR DoxDox GFP MACROD2 GFP + olaparib MACROD2 + olaparib

Figure 4. MACROD2 macrodomain deletion causes loss of protein recruitment to sites of DNA damage, increased PARP1 autoinhibitory mono-ADP- ribosylation and repression of PARP1 transferase activity, and increases cell sensitivity to genotoxic stress-induced DNA damage. A, Recruitment of EGFP-tagged MACROD2 macrodomain deletion mutants (MACROD2Δex11-17, MACROD2Δex1-8, MACROD2Δex4, MACROD2Δex5, and MACROD2Δex4-5) and wild-type MACROD2 control to sites of laser-induced DNA damage. Sites of laser microirradiation are indicated by the arrowheads. Images are representative for >5 nuclei representative of duplicate experiments. Scale bar, 10 μm. B, PARP1 immunoprecipitation and western blots for PARP1 mono-ADP-ribosylation levels using anti-mono(ADP ribose) antibody (mADPr), C, Western blots for global protein poly-ADP-ribosylation (PARylation) levels detected using anti-poly(ADP ribose) antibody (pADPr), and D, incorporation of biotinylated PAR onto histone proteins for isogenic HCT116-MAC- ROD2−/−, HCT116-MACROD2−/+ and HCT116-MACROD2+/+ cells, or LIM2405-MACROD2−/− and LIM1215-MACROD2−/+ cells reconstituted with wild-type MACROD2 or GFP control challenged with doxorubicin (Dox; 0.5 μmol/L) and/or the PARP1 inhibitor olaparib (1 μmol/L). For histone PARylation, data is presented as the mean ± SEM of 6 repeats representative of duplicate experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t test). E, Representa- tive images of γ-irradiation- (IR, 10 Gy) and doxorubicin- (0.5 μmol/L) induced DNA damage in MACROD2+/+, MACROD2−/+, and MACROD2−/− isogenic HCT116 cells, as measured by the alkaline comet assay (detects DNA-SSBs). F, Levels of DNA damage quantified by the tail moment in the comet assay. G, Representative images of IR- and Dox-induced DNA damage in LIM2405-MACROD2−/−, SW480-MACROD2−/−, LIM1215-MACROD2−/+, and COLO320- MACROD2−/+ cells reconstituted with wild-type MACROD2 or GFP control, as measured by the alkaline comet assay. H, Levels of DNA damage quantified by the tail moment in the comet assay.I–J, DNA damage in PARP1 inhibitor–treated (olaparib) and IR- or Dox-treated isogenic HCT116 cells (I) and LIM2405-MACROD2−/− and SW480-MACROD2−/− cells reconstituted with wild-type MACROD2 or GFP control (J), as measured by the alkaline comet assay. All comet assay data are presented as the mean ± SEM for >100 comets representative of duplicate experiments. ns, not significant; *,P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 (Student t test).

OF9 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

AC LIM2405 LIM1215 p-γH2AX −/− −/+ HCT116 iso MACROD2 MACROD2 + / MACROD2 MACROD2 MACROD2 +/+ MACROD2 −/+ MACROD2 −/− GFP GFP GFP GFP Macrod2 Min/ + Apc − / + Macrod2 Min/ + Ap c − / − 1h No IR 2h No IR Macrod2 + Min /

B LIM2405 LIM1215 Ap c HCT116 iso MACROD2 −/ − MACROD2 −/+ 15 8 6 D 80 ** **** **** 6 * **** IR 10 **** 4 **** **** 4 60 ******** *** 5 2 **** 2 40 0 0 0 6 20 8 *** **** 20 p- γ H2AX foci/cell **** 15 6 4 * 10 ** 4 *** **** H2AX–positive cells % of p- γ H2AX–positive Dox 2 ** 0 ** 5 2 *** ** Min/+ Min/+ Min/+ 0 0 0 Apc Apc Apc / / / 0124 0 1 24 0124 Macrod2 + + Macrod2 − + Macrod2 − − Time (h)

−/− −/+ EFLIM2405 MACROD2 −/− HCT116 iso LIM2405 MACROD2 LIM1215 MACROD2 2.0 25 20 30 **** **** **** 20 1.5 15 **** 20 *** **** IR 15 **** 1.0 **** 10 **** 10 **** 10 5 **** 0.5 5

+ cells 0 0 0 0 SW480 MACROD2 −/− 2.5 20 20 20 ****

%GFP **** **** **** **** 2.0 15 15 15 **** BRCA1 foci/cell **** 1.5 **** 10 10 10 **** 1.0 ****

**** Dox 5 5 5 0.5 0 0 0 0 0 0.5 12 0124 0124 RFP DR-GFP − pCBAScel MACROD2 DR-GFP − pCBAScel Time (h) RFP DR-GFP + pCBAScel MACROD2 DR-GFP + pCBAScel MACROD2 +/+ MACROD2 −/+ MACROD2 −/− GFP MACROD2

Figure 5. MACROD2 deficiency increases cell sensitivity to genotoxic stress-induced DNA damage. A, Representative images of γ-H2AX and BRCA1 foci in isogenic HCT116-MACROD2−/−, HCT116-MACROD2−/+ and HCT116-MACROD2+/+ cells, or LIM2405-MACROD2−/− and LIM1215-MACROD2−/+ cells reconstituted with wild-type MACROD2 or GFP control. Cells were subject to IR (2 Gy), collected at the times indicated and immunostained with anti-γ-H2AX antibody (yellow) or anti-BRCA1 antibody (pink) and DAPI (blue). Scale bar, 20 μm. B and F, Quantification of the number ofγ -H2AX foci (B) and BRCA1 foci (F) for IR or treatment with doxorubicin (Dox; 0.5 μmol/L) in >80 cells for the indicated colorectal cancer cell lines. Data are means ± SEM for a representative duplicate experiment. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 (Student t test). C, Representative images of γ-H2AX immunohistochemistry in intestinal tumors from 130-day-old ApcMin/+/Macrod2+/+, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2−/− mice. Scale bar, 200 μm. D, Quantification ofγ -H2AX–positive tumor cells for each genotype. Data are means ± SEM of 10 tumors/mice from 6 mice per cohort. ****, P < 0.0001 (Student t test). E, DR-GFP reporter assays in LIM2405-MACROD2−/− and SW480-MACROD2−/− cells reconstituted with wild-type MACROD2 or GFP control, transfected with/without the pBACSceI vector, and treated with/without olaparib (1 μmol/L). Scatter plots quantifying the percentage of GFP- positive cells are shown. Data are means ± SEM of 6 replicates representative of triplicate experiment. ****, P < 0.0001 (Student t test).

AUGUST 2018 CANCER DISCOVERY | OF10

RESEARCH ARTICLE Sakthianandeswaren et al.

AB + / + + /

+ *** 120 0/43 *** (%) clone 5 100 2 MACROD MACROD2 + / − + / − 80 4/104 (2.4%) clone 81 Cell count MACROD2 HCT116 iso

2 MACROD 60 − / Chromosome count − − / − 8/83 40 (9.6%) clone 76 2 MACROD

2 MACROD PI (FL2) /+ /+ /− + − − Passage 6 Passage 50 Diploid Aneuploid

MACROD2 MACROD2 MACROD2 C 12345678910111213141516171819202122X Y mar HCT116 ATCC

+/+

−/+

−/− Olaparib +/+

Figure 6. MACROD2 deficiency promotes chromosome instability. A, Metaphase analyses for isogenic HCT116-MACROD2−/−, HCT116-MACROD2−/+, and HCT116-MACROD2+/+ cells showing representative spreads at passages 6 and 50 and violin plots summarizing chromosome counts (>50 metaphases per genotype). ***, P < 0.001 (Levene test). B, Flow cytometry analysis for DNA content (red, euploid; yellow, aneuploid) of single cell clones from isogenic HCT116 cells at passage 50. Representative plots shown. C, Representative karyotypes from G-banding analysis for isogenic HCT116 MACROD2 knockout cells with/without treatment with olaparib (10 μmol/L). A total of 30 metaphases were karyotyped per genotype. but not doxorubicin treatment. Accordingly, suppression or treatment with IR or doxorubicin. HCT116-MACROD2−/− deletion of wild-type MACROD2 resulted in increased apop- and HCT116-MACROD−/+ cells developed significantly fewer tosis posttreatment as measured by the luminescent assay for BRCA1 foci as compared with HCT116-MACROD2+/+ cells, caspase 3/7 activity or by flow cytometric analysis of Annexin whereas the inverse was observed upon MACROD2 reconsti- V staining, whereas rescue of MACROD2-deleted cell lines tution in LIM2405-MACROD2−/− or LIM1215-MACROD2−/+ resulted in reduced apoptosis (P < 0.05 for all comparisons; cells (Fig. 5F; Supplementary Fig. S19). Comparable results Supplementary Fig. S18A–S18B). were observed in isogenic HCT116 MACROD2 knockout cells A well-established consequence of loss of PARP1 function for RAD51 foci (Supplementary Fig. S20). is impaired homologous recombination (HR)–mediated DSB Together, these data indicate that impairment of PARP1 repair (22). To test for the impact of MACROD2 deficiency activity caused by MACROD2 haploinsufficiency results in on HR-mediated DSB repair, we performed DR-GFP reporter altered DNA repair and increased sensitivity to DNA damage. assays on MACROD2−/− LIM2405 and SW480 cells with and without MACROD2 reconstitution. In this system, transient MACROD2 Haploinsufficiency Promotes expression of an I-SceI endonuclease generates a DSB at the Development of CIN integrated GFP-deletion gene sequences in which error-free Increased susceptibility to DNA damage is a major path- repair leads to a full-length GFP product detectable by flow way to chromosome instability (CIN) and aneuploidy in cytometry (23). As anticipated, reintroduction of RFP-tagged human cancer (24, 25). To examine whether MACROD2 hap- full-length MACROD2 into LIM2405-DR and SW480-DR loinsufficiency compromised maintenance of genome integ- cells resulted in significant enhancement of HR-mediated rity, we examined karyotype stability of MEFs from Macrod2 repair as compared with RFP control cells (P < 0.001 for knockout mice during long-term cell culture. Karyotype vari- both comparisons; Fig. 5E). To further evaluate HR-mediated ability was strikingly increased by passage 60 in Macrod2−/− DSB repair, we performed assessment of BRCA1 foci post- and Macrod2−/+ MEFs as compared with Macrod2+/+ MEFs

OF11 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

(P < 0.001 for both comparisons), with 82% (86 of 105), 47% associated with MACROD2 loss, with ∼10% of HCT116- (59 of 126), and 15% (17 of 116) of metaphases exhibiting MACROD2−/− and HCT116-MACROD2−/+ cells harboring >55 chromosomes, respectively (Supplementary Fig. S21A). more than two centrosomes, whereas HCT116-MACROD2+/+ Gross aneuploidy in Macrod2−/− and Macrod2−/+ MEFs was fur- cells showed no abnormal centrosome numbers (Fig. 7E and ther evident by flow cytometry analysis of single-cell clones. F). The chromosome missegregation phenotype associated Ninety-one percent (21 of 23) of Macrod2−/− MEF clones and with MACROD2 haploinsufficiency was phenocopied by treat- 53% (8 of 15) of Macrod2−/+ MEF clones were aneuploid, as ment of HCT116-MACROD2+/+ cells with PARP1 inhibitors compared with 16% (6 of 38) of Macrod2+/+ MEF clones (P < (olaparib and veliparib; Fig. 7B; Supplementary Fig. S22). 0.005 for both comparisons; Supplementary Fig. S21B). To examine the relationship between MACROD2 loss To confirm thatMACROD2 haploinsufficiency could and the extent of aneuploidy in human primary tumors, also promote the development of aneuploidy in human we calculated the estimated chromosome segment number colorectal cancer cells, we repeated the long-term culture (eCSN; Supplementary Methods) from SNP array data for experiments using our isogenic HCT116 MACROD2 knock- the 616 TCGA and 651 in-house colorectal cancers. eCSN out cells. Consistent with our results in MEFs, karyotype increased by ∼1.3-fold in MACROD2−/+ tumors and ∼1.8-fold variability was increased by passage 50 in HCT116-MAC- in MA CROD2−/− tumors as compared with the respective ROD2−/− and HCT116-MACROD2−/+ cells as compared with MACROD2+/+ tumors (P < 0.001 for both comparisons in HCT116-MACROD2+/+ cells (P < 0.011 for both compari- both cohorts; Fig. 7G). Concomitantly, the proportion of ane- sons), with 11.5% (8 of 61), 4% (3 of 74), and 2% (1 of 84) of uploid tumors increased from 72.7% in MACROD2+/+ to 83.0% metaphases exhibiting >50 chromosomes, respectively (Fig. and 92.3% in MACROD2−/+ and MACROD2−/− cases, respec- 6A). Aneuploidy in HCT116-MACROD2−/− and HCT116- tively (Fig. 7H). In multivariate analysis, adjusting for age at MACROD2−/+ cells was again confirmed by flow cytometry, diagnosis, gender, tumor location, stage, differentiation, and with 9.6% (8 of 83) of HCT116-MACROD2−/− clones and MSI status, eCSN remained independently associated with 2.4% (4 of 104) of HCT116-MACROD2−/+ clones showing MACROD2 deletion status in both cohorts (Supplementary aneuploidy, as compared with 0% (0 of 43) of HCT116- Table S7). Considering other proposed drivers of CIN in colo- MACROD2+/+ clones (P = 0.035 and P = 0.196, respectively, rectal cancer, including APC (26, 27), TP53 (28), FBXW7 (29), Z-test; Fig. 6B). and BCL9L (30), MACROD2 status remained an independent Further, G-banding analysis for isogenic HCT116 predictor of eCSN in both cohorts (Supplementary Table MA CROD2 knockout cells demonstrated increased structural S8; BCL9L mutation status only available for TCGA cohort). chromosome abnormalities in HCT116-MACROD2−/− and Our results identify MACROD2 as a critical caretaker gene in HCT116-MACROD2−/+ cells as compared with HCT116- human colorectal cancer, haploinsufficiency of which leads to MA CROD2+/+ cells, with 20, 12, and 5 aberrations in 30 CIN and aneuploidy. metaphases, respectively (Fig. 6C; Supplementary Table In summary, we propose a model in which partial or S6). Structural abnormalities included abundant intersti- complete loss of MACROD2 mono-ADP-ribosylhydrolase tial deletions and translocations (derivative chromosomes). function, because of whole-gene or catalytic domain dele- Cytogenetic analysis further confirmed our initial observations, causes impaired PARP1 transferase activity in human tion of increased numerical chromosome abnormalities (20, colorectal cancer by abrogating removal of PARP1 terminal 12, and 7 aberrations, respectively), but clarified that many autoinhibitory mono-ADP-ribose (Fig. 7I). Repression of of these represented marker chromosomes (unidentifiable PARP1 activity results in altered DNA repair and sensi- chromosomes). Notably, short-term treatment of HCT116- tivity to DNA damage. DNA repair deficiency, potentially MACROD2+/+ cells with the PARP1 inhibitor olaparib pro- enhanced by centrosome amplification, appears to culminate duced a similar spectrum of chromosome aberrations, with in CIN, thereby promoting intratumor heterogeneity and 19 structural and 22 numerical abnormalities, respectively cancer progression. (Fig. 6C; Supplementary Table S6). Our karyotype and flow cytometry analyses of long-term cell cultures suggest that MACROD2 haploinsufficiency pro- DISCUSSION motes CIN. To measure the impact of MACROD2 loss on CIN is a principal driver of cancer evolution (31), has mitotic chromosome segregation, we quantified the level of prognostic relevance in multiple cancer types (32–34), and mitotic defects—including anaphase bridges, lagging chromo- is associated with tumor multidrug resistance (35, 36). Our somes, and micronuclei—in our isogenic HCT116 MACROD2 studies identify MACROD2 as a major caretaker tumor sup- knockout cells by immunostaining for kinetochores and pressor gene in human colorectal cancer, with haploinsuf- microtubules (Fig. 7A). As anticipated, HCT116-MACROD2−/+ ficiency causing repression of PARP1 activity, altered DNA and HCT116-MACROD2−/− cells displayed a significant gene repair, and CIN. dosage–dependent increase in chromosome missegregation Our DNA copy-number analysis on 651 sporadic colorectal errors as compared with HCT116-MACROD2+/+ cells (P < cancers validates the original observation of recurrent focal 0.001 and P < 0.001, respectively; Fig. 7B). A gene dosage– MACROD2 deletions in TCGA-analyzed colorectal cancers, dependent increase in mitotic defects was further con- confirming a prevalence of∼ 30% (4). Integrating both data- firmed for tumors fromApc Min/+/Macrod2−/+ and ApcMin/+/ sets and expanding the analysis to 53 human colorectal Macrod2−/− mice as compared with ApcMin/+/Macrod2+/+ mice cancer cell lines, our report provides the first comprehensive (P < 0.001 and P < 0.001, respectively; Fig. 7C and D). fine-mapping data forMACROD2 deletions, establishing dele- We further noted evidence of centrosome amplification tion extent and status (heterozygous vs. homozygous), and

AUGUST 2018 CANCER DISCOVERY | OF12

RESEARCH ARTICLE Sakthianandeswaren et al.

A B HCT116 iso C Anaphase Lagging ns Anaphase Lagging Normal bridge chromosome Normal bridge chromosome Micronuclei **** 100 80 DNA

60 tumors

40 20 Macrod2 Percentage cells Percentage CENP-A /+

0 Mi n ib ib ib ib +/ + −/+ +/ + −/− +/ + −/ + −/ − Apc HCT116 iso β -tubulin MACROD2 2 MACROD MACROD2 MACROD2 2 MACROD MACROD2 MACROD2 mol/L olapar mol/L olapar mol/L olapar mol/L olapar **** **** µ µ µ µ D 100 80 0.5 0.5 0.1 0.5 Merge Micronuclei Lagging 60 Anaphase chromosomes 40 bridges Normal mitosis 20

Percentage cells Percentage 0 Apc Min/+ Apc Min/+ Apc Min/+

Macrod2 +/+ Macrod2 −/+ Macrod2 −/−

E GH I DAPI + DAPI + γ-tubulin β-tubulin Merge In-house In-house *** * *** ** −/+ or −/− 1 150 100 MACROD2 +/+ MACROD2 Centrosome DSB DSB 80 SSB SSB 100 60 2 DNA repair DNA repair Centrosomes 40 factors factors 50 pADPr pADPr 20 PARP1 PARP1 >2 0 Centrosomes 0 MACROD2 TCGA TCGA *** * PARP1 PARP1 MACROD2 *** ** deletion 100 F 150 100 MACROD2 DNA replication 80 * 80

Percent diploid / aneuploid Percent SSB conversion to DSB amplification PARP1 Centrosome 100 60 mADPr 60 40 PARP1 Estimated chromosome segment number 50 SSB repair 40 DSB repair defect of cells 20 Percentage *** 0

0 −/ − −/ + +/ + −/ − −/ + +/ + Chromosome 0 instability 12>2 Centrosome number per cell CROD2 PARP1 Active PARP1 Inactive MACROD2 +/+ MACROD2 −/+ MACROD2 −/− MACROD2 2 MACROD MACROD2 MACROD2 MACROD2 MA

Figure 7. MACROD2 deficiency promotes chromosome instability. A, Representative immunofluorescent images of a normal anaphase, anaphase bridges, lagging chromosomes, and micronuclei (scale bar, 10 μm), and B, quantification of segregation errors for isogenic HCT116-MACROD2−/−, HCT116-MACROD2−/+, and HCT116-MACROD2+/+ cells with/without olaparib treatment at the indicated concentrations. HCT116 cells are stained with DAPI (blue), CENP-A (green), and β-tubulin (red). Data are representative of triplicate experiments with 150 mitotic events scored. ns, not significant; ****, P < 0.0001 (Student t test). C, Representative images of a normal anaphase, anaphase bridges, and lagging chromosomes (scale bar, 10 μm), and D, quantification of segregation errors for tumors from 130-day-old ApcMin/+/Macrod2+/+, ApcMin/+/Macrod2−/+, and ApcMin/+/Macrod2−/−mice. Tumors are stained with H&E. Data are representative of triplicate experiments with 150 mitotic events scored. ****, P < 0.0001 (Student t test). E, Representative immunofluorescent images of centrosomes (scale bar, 10μ m), and F, quantification of centrosome numbers for isogenic HCT116-MACROD2−/−, HCT116- MACROD2−/+, and HCT116-MACROD2+/+ cells. HCT116 cells are stained with DAPI (blue), γ-tubulin (yellow), and β-tubulin (red). Data are representative of duplicate experiments with >100 cells scored. *, P < 0.05; ***, P < 0.001 (Student t test). G–H, eCSN (G) and aneuploidy (H) derived from SNP array data of TCGA (n = 616) and in-house (n = 651) colorectal cancers according to MACROD2 deletion status. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (Student t test and Pearson χ2 test, respectively). I, Proposed model in which partial or complete loss of MACROD2 mono-ADP-ribosylhydrolase function, because of whole-gene or catalytic domain deletions, causes impaired PARP1 transferase activity in human colorectal cancer by abrogating removal of PARP1 terminal autoinhibitory mono-ADP-ribose. Repression of PARP1 activity results in altered DNA repair and sensitivity to DNA damage. DNA repair deficiency, potentially enhanced by centrosome amplification, culminates in CIN.

OF13 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE demonstrating the impact of both exonic and intronic dele- tumor multiplicity and burden, and promoted the growth tions at the MACROD2 transcript level. MACROD2 was almost of human colorectal cancer xenografts. Previous evaluation exclusively subject to whole-gene deletions or microdeletions of PARP1-deficient mice for an alternative model of sporadic involving the catalytic macrodomain, with only rare somatic colon tumorigenesis (azoxymethane treatment) has found point mutations. About two thirds of tumors with MACROD2 increased colon tumorigenesis (44), demonstrating that when aberrations showed evidence of two hits. A similar pattern of combined with relevant drivers, PARP1 deficiency promotes targeting with predominant focal deletions has previously intestinal tumor development, supporting our observations been described for the PARK2 tumor suppressor in colorectal for MACROD2 deficiency. Notably, functions other than cancer (37). The high frequency of MACROD2 deletions in impairment of PARP1 may contribute to the protumorigenic colorectal cancer cells may be a consequence of the reported impact of MACROD2 loss. Previous studies have implicated susceptibility of this locus to breakage during replicative MACROD2 as a regulator of GSK3β and WNT signaling that stress (6). could provide such an additional protumorigenic drive (10). MACROD2 is a mono-ADP-ribosylhydrolase with a criti- However, we found no supporting evidence for this in APC or cal role in the regulation of the DNA-damage sensor PARP1 CTNNB1 mutated intestinal tumor cells. Studies to expand through controlling the removal of the terminal autoin- our knowledge of MACROD2 protein targets and related hibitory mono-ADP-ribose (7, 21). Our results indicate that pathways are clearly warranted. MACROD2 deficiency in colorectal cancer causes increased In summary, our findings reveal MACROD2 as a haplo- PARP1 mono-ADP-ribosylation and repression of PARP1 insufficient caretaker tumor suppressor gene, essential for transferase activity, leading to impaired DNA repair and sen- the maintenance of genome integrity. Although MACROD2 sitivity to genotoxic stress-induced DNA damage in a gene aberrations were particularly frequent in human colorectal dosage–dependent manner. The MACROD2 macrodomain is cancer, recurrent focal loss was also noted in cancers of required for PARP1-dependent recruitment to sites of DNA the stomach, esophagus, cervix, uterus, lung, liver, and damage (7), and accordingly MACROD2 proteins with in- thyroid, supporting a broader role of MACROD2 haploin- frame exonic deletions disrupting this region exhibited loss sufficiency in the development and evolution of human of accumulation. Our findings are supported by studies in malignancies. PARP1-deficient cells that similarly show altered DNA repair and sensitivity to DNA damage (38, 39). PARP1 inhibitor METHODS use in cells deficient in HR repair is a prime example of the Additional details are provided in the Supplementary Methods. therapeutic paradigm of synthetic lethality in cancer (38, 40, 41), and MACROD2 haploinsufficiency may constitute a Patients potential vulnerability in tumors that could be therapeuti- A total of 651 patients with stages 1–4 colorectal cancer were cally exploited. recruited from the Royal Melbourne Hospital (Parkville, VIC, Our data reveal MACROD2 as a caretaker gene in colo- Australia), Western Hospital Footscray (Footscray, VIC, Australia), rectal cancer, haploinsufficiency of which results in CIN. and St Vincent’s Hospital Sydney (Darlinghurst, NSW, Australia). MACROD2-deficient colorectal cancer cells displayed a gene Patients with familial polyposis syndromes, ulcerative colitis, or dosage–dependent increase in structural and numerical chro- Crohn disease–associated colorectal cancer were excluded. Writ- mosome abnormalities and were prone to chromosome mis- ten informed consent was obtained from all patients; the studies segregation errors. Similar to our findings for MACROD2 were conducted in accordance with recognized ethical guidelines deficiency, Parp1 knockout MEFs have been shown to accu- (National Statement on Ethical Conduct in Human Research, Com- mulate structural chromosome abnormalities (38, 39). monwealth of Australia) and approved by the Human Research Ethics Committee of the Walter and Eliza Hall Institute of Medical Cytogenetic analyses of Parp1 knockout MEFs have further +/+ Research (Parkville, VIC, Australia; WEHI HREC 12/19). Fresh- reported an increase in aneuploidy from ∼1% in Parp1 cells frozen tumor and matched normal tissues were retrieved from −/+ −/− to ∼5% and ∼20% in Parp1 and Parp1 cells, respectively hospital-associated tissue banks, including 60 stage 1, 208 stage 2, (42). Accordingly, PARP1 inhibitor treatment of MACROD2 297 stage 3, and 86 stage 4 cases. Colon tumors from caecum to wild-type colorectal cancer cells recapitulated the spectrum transverse colon were defined as proximal, and those from splenic of chromosome abnormalities and missegregation pheno- flexure to rectum as distal. Patients’ clinicopathologic characteris- type associated with MACROD2 loss. MACROD2 deletion tics were collected using a multisite database and are summarized in in primary tumors was associated with increased eCSN and Supplementary Table S2. aneuploidy, independent of clinical features and other known drivers of CIN. Interestingly, MACROD2-deficient colorectal Cell Lines cancer cells also showed a tendency to centrosome amplifica- A total of 53 colorectal cancer cell lines were studied: C10, C32, tion, a phenotype linked to PARP1 deficiency (43), which may C80, C99, C106, C135, CCK81, CoCM-1, COLO320, COLO678, contribute to the CIN phenotype. Together, these data are CX-1, Gp2D, HCA7, HCC2998, HCT15, HCT116, HDC54, HDC82, HDC87, HDC90, HRA19, HT115, HT55, KM12, LIM1215, LIM1899, consistent with a model in which MACROD2 loss promotes LIM2099, LIM2405, LIM2551, LOVO, LS180, LS411, LS513, NCI- cancer CIN at least in part mediated through the impairment H716, NCI-H747, RKO, RW2982, RW7213, SKCO-1, SNU-175, SNU- of PARP1. 283, SNU-C1, SNU-C2B, SW1222, SW1417, SW1463, SW48, SW480, Our in vivo studies support a role of MACROD2 as a tumor SW837, T84, VACO4S, VACO5, and VACO10. HEK293T cells were suppressor acting in a haploinsufficient manner. Homozy- used as APC and CTNNB1 wild-type controls for WNT stimula- gous and heterozygous MACROD2 deficiency enhanced intes- tion assays. Cells were cultured with DMEM and 10% FBS at 37°C Min/+ tinal tumorigenesis in Apc mice, as evident for both and 5% CO2. Cell lines were authenticated by short tandem repeat

AUGUST 2018 CANCER DISCOVERY | OF14

RESEARCH ARTICLE Sakthianandeswaren et al. analysis using the GenePrint 10 System (Promega) at the Australian group on two occasions. Tumor growth was measured using digital Genome Research Facility (AGRF; Parkville, VIC, Australia, February calipers on alternate days until days 20 to 24 after injection. Length, 2016) and confirmed to beMycoplasma free using the Lookout Myco- width, and height of the subcutaneous xenograft nodules were plasma PCR Detection Kit (Sigma-Aldrich, January 2017). Details of measured, and tumor volume calculated as π/6 × L × W × H. At the colorectal cancer cell lines are summarized in Supplementary Table experimental endpoint, all mice were euthanized, and tumors were S3. Treatments were performed with doxorubicin (topoisomerase removed and fixed in formalin. All animal procedures were approved II inhibitor, SelleckChem, S1208), olaparib (an effective catalytic and conducted in accordance with the Animal Ethics Committee of PARP1 inhibitor/effective PARP1 trapper, SelleckChem, S1060), veli- the Walter and Eliza Hall Institute of Medical Research (Parkville, parib (an effective catalytic PARP1 inhibitor/poor PARP1 trapper, VIC, Australia). SelleckChem, S1004), or γ-irradiation using a Cobalt-60 source as indicated. NIR Laser Microirradiation Assay NIR laser microirradiation was performed using a SpectraPhysics Microsatellite Instability Analysis MaiTai Ti:Saphhire laser with DeepSee attachment at 800 nm, with Tumor and matched normal DNA were PCR-amplified for the a 13-μs laser pulse delivered to the cell nucleus with a total energy Bethesda panel of microsatellite markers (BAT25, BAT26, D2S123, of 7.65 mJ. Microirradiated cells were imaged every 20 seconds for D5S346, and D17S250) using fluorescently labeled primers (45). 4 minutes using a 60× UPlanApo NA1.2 water immersion objective Reaction products were analyzed on a 3130xl Genetic Analyzer and an Olympus FV1000 confocal system attached to an Olympus (Applied Biosystems). Microsatellite instability-high was diagnosed IX-81 microscope (473 nm excitation, 510–550 nm detection). Dur- if instability was evident at two or more markers. ing the experiment, cells were kept at 37°C in a CO2-independent HEPES-based imaging medium (Invitrogen) supplemented with 10% SNP Array Analysis FBS (Invitrogen). SNP assays were performed using Human 610-Quad BeadChips (Illumina) at the AGRF. For patient specimens, matched primary can- Single-Cell Gel Electrophoresis (Comet) Assay cers and normal DNA samples were analyzed on the same BeadChip. Colorectal cancer cells (1 × 105 cells/well seeded) and MEF cells Raw SNP array data were processed using GenomeStudio software (1 × 105 cells/well seeded) were seeded into 24-well plates. Cells (Illumina), and call rate, genotype, log R ratio (LRR), and B allele were γ-irradiated with 10 Gy using a Cobalt-60 source or treated frequency data (BAF) were exported. Data have been deposited in the with 0.5 μmol/L doxorubicin [± olaparib (SelleckChem) or DMSO Gene Expression Omnibus (GSE115145). SNP array data for colo- vehicle control], and the degree of DNA damage analyzed at base- rectal cancer cell lines were previously deposited (GSE55832). For line and 2 hours after treatment under neutral (DSBs) or alkaline TCGA-analyzed tumor types, Genome-Wide Human SNP Array 6.0 (SSBs) conditions using the CometAssay Kit (Trevigen) as per (Affymetrix) CEL files were retrieved and processed using Affymetrix the manufacturer’s instructions. After electrophoresis, slides were Power Tools (Affymetrix) and PennCNV (46) to export LRR and BAF stained with SYBR Green Reagent (Bio-Rad), and comets were data. DNA copy-number aberrations (segments) in tumor and colo- imaged using a SPOT RT3 slide camera attached to a Nikon 90i rectal cancer cell line samples were identified using OncoSNP v2.18 microscope (Nikon). Comet Olive tail moments [(tail length) × (tail Suite as previously described (47). To account for noise, segmented fluorescence/(head + tail fluorescence))] were determined using regions were accepted when the difference between LRR means for MetaMorph Microscopy Automation and Image Analysis Software adjacent regions (including adjacent chromosomes, i.e., end of chro- (Molecular Devices). At least 100 randomly chosen comets were mosome 1 and start of chromosome 2) was greater than their LRR analyzed per sample. standard deviations. The eCSN is the total number of segments per sample. Aneuploidy was assigned if samples showed three or more nondiploid chromosomes, where individual chromosome copy num- Antibodies and Reagents bers were estimated based on the modal copy number for all SNPs Immunofluoresence and immunohistochemistry were performed within that chromosome. with the following antibodies: anti–β-catenin (BD Transduction Laboratories, #610153, 1:500 dilution), anti-γH2A.X (phospho S140 Tumor Burden Assessment [3F2]; Abcam, ab22551, 1:1,000 dilution), anti–phospho-histone On autopsy of knockout mice, the intestinal tract was opened H2A.X (Ser139; 20E3; Cell Signaling Technology, #9718, 1:200 longitudinally, fixed in methacarn, and stained with methylene blue, dilution), anti-BRCA1 (Sigma, SAB2702136, 1:1,000 dilution), and tumors were enumerated according to their size by two observers anti-RAD51 (Abcam, ab133534, 1:1,000 dilution), or anti-pATM (A. Sakthianandeswaren and M.J. Parsons) blinded to the genotype antibody (EP1890Y; Abcam, ab81292, 1:1,000 dilution). Western of the mice. Representative tumors and adjacent nontumor tissues blotting analyses were performed with the following antibodies: were harvested for DNA/RNA extraction. Swiss rolls of intestinal seg- anti-PARP1 antibody (Abcam, ab110915), anti-MACROD2 (in house; ments were prepared and fixed in 10% neutral buffered formalin for Supplementary Fig. S23), anti-(mono)ADP ribose (Merck Millipore, paraffin embedding. H&E-stained tissue sections were imaged, and MABE1076), anti-poly(ADP)ribose (Enzo, ALX-804-220-R100), anti- the number of neoplastic cells per adenoma was quantified by image vinculin (Sigma, V9264), anti-GFP (OriGene, TA150052), and anti- analysis using ImageJ software. PARP1 (Abcam, ab6079). Primer sequences used for assays detailed in the Supplementary Human Colorectal Cancer Cell Line Xenografts Methods are provided in Supplementary Tables S9, S10, and S11. Xenografts of isogenic HCT116-MACROD2 knockout cells, HCT116-shMACROD2 knockdown cells, LIM2405-MACROD2-GFP Metaphase Spreads reconstituted cells, and respective controls were generated by sub- Cells growing in exponential phase were incubated with 0.2 μg/mL cutaneous injection of 1 × 106 cells into the rear flanks of BALB/c colchicine (Sigma, C3915) overnight. Cells were harvested by mitotic athymic nude mice. Knockdown or reexpression of MACROD2 was shake off, incubated with 0.56% or 0.28% KCl at 37°C for MEFs and confirmed using qRT-PCR (see below): forward primer ′5 -TGACCT human colorectal cancer cells, respectively, fixed with 3:1 methanol/ TAGAAGAGAGACGCAAA-3′, reverse primer 5′-TCTTCACCTGG acetic acid (v/v), and dropped onto slides. Slides were air-dried GATGTTTCC-3′. Eight to 12 xenografts were analyzed for each and stained with Leishman’s stain (VWR International) for

OF15 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

4 minutes. Chromosome numbers were evaluated using an Axioimager Commons Data Portal (https://gdc-portal.nci.nih.gov/), filtering for Z2 (Zeiss) a under a 63× objective. At least 50 metaphases were ana- tumor samples and removing duplicates. GISTIC was run in “genegis- lyzed per sample. tic” mode, with default parameters. Differential gene expression analysis for RNA-seq data was performed with voom/limma using G-Banding Analysis the edgeR R package; results for WNT target genes (http://web.stan- ford.edu/group/nusselab/cgi-bin/wnt/target_genes) were adjusted Cells in exponential growth phase (± treatment with 10 μmol/L olaparib for 28 hours) were incubated with 0.2 μg/mL colchicine for multiple testing using the Benjamini–Hochberg procedure. All (Sigma, C3915) for 60 minutes. Metaphases were harvested according comparisons were two sided, and P values of <0.05 were considered to standard procedures (48). Thirty pseudodiploid metaphases with statistically significant. minimal overlaps were selected at 100× magnification, captured, and Disclosure of Potential Conflicts of Interest karyotyped at 1,000× magnification with an Ikaros karyotyping system (Metasystems). Karyotypes were written with reference to the pub- No potential conflicts of interest were disclosed. lished HCT116 karyotype (45, X, -Y, der(10)dup(10)(q24q26)t(10;16) (q26;q24), der(16)t(8;16)(q13;p13), der(18)t(17;18)(q21;p11.3); ref. 49) Authors’ Contributions as stemline according to International System for Human Cytogenetic Conception and design: A. Sakthianandeswaren, D. Mouradov, Nomenclature (2016; ref. 50). Changes from the stemline karyotype J. Desai, I. Street, M. Buchert, O.M. Sieber were classified as structural and/or numerical abnormalities. Mark- Development of methodology: A. Sakthianandeswaren, M.J. Parsons, ers and abnormal chromosomes were each counted as one structural R.N. MacKinnon, C. Tsui, C.J. Nowell, I. Street, O.M. Sieber abnormality. Gain or loss of a chromosome from the stemline karyo- Acquisition of data (provided animals, acquired and managed type was counted as one numerical abnormality. patients, provided facilities, etc.): M.J. Parsons, R.N. MacKinnon, B. Catimel, S. Liu, M. Palmieri, T.L. Putoczki, A. Preaudet, C. Tsui, DSB HR DNA Repair Assays R.L. Ward, N.J. Hawkins, J. Desai, P. Gibbs, O.M. Sieber DSBs in DR-GFP reporter cell lines (106 cells) were induced by Analysis and interpretation of data (e.g., statistical analysis, biosta- transfection with pCBASce plasmid according to the manufacturer’s tistics, computational analysis): A. Sakthianandeswaren, M.J. Parsons, instructions; transfection with empty vector control served as nega- D. Mouradov, R.N. MacKinnon, C. Love, R.N. Jorissen, L. Whitehead, tive control. HR repair activity was assessed by quantification of the C.J. Nowell, M. Ernst, M. Buchert, O.M. Sieber percentages of GFP-positive cells using a BD FACSCalibur instru- Writing, review, and/or revision of the manuscript: ment (BD Biosciences). Data were analyzed using CellQuest 3.2 A. Sakthianandeswaren, M.J. Parsons, D. Mouradov, C.J. Nowell, software (Becton Dickinson). For each experiment, 200,000 cells were R.L. Ward, J. Desai, P. Gibbs, M. Ernst, M. Buchert, O.M. Sieber scored per treatment group, and the frequency of recombination Administrative, technical, or material support (i.e., reporting or events was calculated from the number of GFP-positive cells divided organizing data, constructing databases): A. Sakthianandeswaren, by the number of cells analyzed following correction for transfection S. Liu, S. Li, C. Tsui, N.J. Hawkins, O.M. Sieber efficiency. Study supervision: O.M. Sieber

Chromosome Segregation and Centrosome Assays Acknowledgments Cells were grown overnight on glass cover-slips, fixed in 4% para- The authors thank the Victorian Cancer Biobank and BioGrid formaldehyde for 20 minutes, and permeabilized with 0.2% Triton Australia for provision of patient specimens and data, and Prof. X-100 for 15 minutes. Samples were blocked in 5% BSA/TBS and M. Schwab at the DKFZ for access to colorectal cancer cell lines. This labeled with rabbit CENP-A (CST, 2186S, 1:400), rabbit γ-tubulin work was supported by an NHMRC Project Grant (APP1079364 to (Abcam, ab11317, 1:500), and mouse β-tubulin (Sigma, T4026, 1:200) O.M. Sieber, M. Buchert, J. Desai, and I. Street), a Cancer Council overnight at 4°C. After incubation with Alexa Fluor–conjugated sec- Victoria Grant-in-Aid (APP1060964 to O.M. Sieber and R.L. Ward), ondary antibodies (Alexa 546 anti-mouse A11003, 1:1,000; Alexa 647 the Ludwig Institute for Cancer Research (O.M. Sieber), the Cancer anti-rabbit A21245, 1:1,000) and DAPI (Roche, 1:5,000), the slides Therapeutics Cooperative Research Centre (I. Street), an NHMRC were mounted with DPX (Sigma-Aldrich). Immunofluorescence analy- Senior Research Fellowship (APP1136119 to O.M. Sieber), a Cancer ses for segregation errors and centrosome number were performed Therapeutics CRC Top Up PhD Scholarship (M.J. Parsons), and the on a Leica SP8 confocal microscope equipped with a 63×/1.8 NA oil Victorian Government’s Operational Infrastructure Support Pro- immersion objective. At least 100 mitotic events were scored per gram (O.M. Sieber). genotype. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby Statistical Analysis marked advertisement in accordance with 18 U.S.C. Section 1734 Statistical analyses were performed using the statistical comput- solely to indicate this fact. ing software R (R Development Core Team, 2011). For univariate analyses, differences between groups were assessed using the Pearson Received August 11, 2017; revised April 16, 2018; accepted June 5, χ2 test for categorical variables, and the Student t test, Z-proportion 2018; published first June 7, 2018. test, or Levene test for continuous variables as indicated. Survival data for aging mice were analyzed using the Wald test. Multivariate analysis for the association between MACROD2 deletion status and eCSN, with adjustment for clinicopathologic and molecular fea- REFERENCES tures was performed using a quasi-Poisson generalized linear model. . 1 Schreiber V, Dantzer F, Ame JC, de Murcia G. Poly(ADP-ribose): novel The Compare Groups of Growth Curves (http://bioinf.wehi.edu.au/ functions for an old molecule. Nat Rev Mol Cell Biol 2006;7:517–28. software/compareCurves) permutation test was used to evaluate cell 2. Erener S, Mirsaidi A, Hesse M, Tiaden AN, Ellingsgaard H, Kostadi- line xenograft assays. Identification of focalMACROD2 deletions in nova R, et al. ARTD1 deletion causes increased hepatic lipid accumu- TCGA cohorts was performed using GISTIC v2.022 (12). Masked lation in mice fed a high-fat diet and impairs adipocyte function and Copy Number Segment data were retrieved from the Genomic Data differentiation. FASEB J 2012;26:2631–8.

AUGUST 2018 CANCER DISCOVERY | OF16

RESEARCH ARTICLE Sakthianandeswaren et al.

3. Hassa PO, Haenni SS, Elser M, Hottiger MO. Nuclear ADP-ribosyla- 25. Bakhoum SF, Kabeche L, Murnane JP, Zaki BI, Compton DA. DNA- tion reactions in mammalian cells: where are we today and where are damage response during mitosis induces whole-chromosome mis- we going? Microbiol Mol Biol Rev 2006;70:789–829. segregation. Cancer Discov 2014;4:1281–9. 4. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Dono- 26. Fodde R, Kuipers J, Rosenberg C, Smits R, Kielman M, Gaspar C, et al. van J, et al. The landscape of somatic copy-number alteration across Mutations in the APC tumour suppressor gene cause chromosomal human cancers. Nature 2010;463:899–905. instability. Nat Cell Biol 2001;3:433–8. 5. Dulak AM, Schumacher SE, van Lieshout J, Imamura Y, Fox C, Shim 27. Kaplan KB, Burds AA, Swedlow JR, Bekir SS, Sorger PK, Nathke IS. B, et al. Gastrointestinal adenocarcinomas of the esophagus, stom- A role for the adenomatous polyposis coli protein in chromosome ach, and colon exhibit distinct patterns of genome instability and segregation. Nat Cell Biol 2001;3:429–32. oncogenesis. Cancer Res 2012;72:4383–93. 28. Samassekou O, Bastien N, Lichtensztejn D, Yan J, Mai S, Drouin 6. Rajaram M, Zhang J, Wang T, Li J, Kuscu C, Qi H, et al. Two dis- R. Different TP53 mutations are associated with specific chromo- tinct categories of focal deletions in cancer genomes. PLoS One somal rearrangements, telomere length changes, and remodeling of 2013;8:e66264. the nuclear architecture of telomeres. Genes Chromosomes Cancer 7. Jankevicius G, Hassler M, Golia B, Rybin V, Zacharias M, Timinszky 2014;53:934–50. G, et al. A family of macrodomain proteins reverses cellular mono- 29. Mao JH, Perez-Losada J, Wu D, Delrosario R, Tsunematsu R, Nakay- ADP-ribosylation. Nat Struct Mol Biol 2013;20:508–14. ama KI, et al. Fbxw7/Cdc4 is a p53-dependent, haploinsufficient 8. Ciccia A, Elledge SJ. The DNA damage response: making it safe to tumour suppressor gene. Nature 2004;432:775–9. play with knives. Mol Cell 2010;40:179–204. 30. Lopez-Garcia C, Sansregret L, Domingo E, McGranahan N, Hobor S, 9. Golia B, Moeller GK, Jankevicius G, Schmidt A, Hegele A, Preisser Birkbak NJ, et al. BCL9L dysfunction impairs caspase-2 expression J, et al. ATM induces MacroD2 nuclear export upon DNA damage. permitting aneuploidy tolerance in colorectal cancer. Cancer Cell Nucleic Acids Res 2017;45:244–54. 2017;31:79–93. 10. Rosenthal F, Feijs KL, Frugier E, Bonalli M, Forst AH, Imhof R, et al. 31. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Macrodomain-containing proteins are new mono-ADP-ribosylhydro- Cell 2011;144:646–74. lases. Nat Struct Mol Biol 2013;20:502–7. 32. Carter SL, Eklund AC, Kohane IS, Harris LN, Szallasi Z. A signature 11. Niehrs C. The complex world of WNT receptor signalling. Nat Rev of chromosomal instability inferred from gene expression profiles Mol Cell Biol 2012;13:767–79. predicts clinical outcome in multiple human cancers. Nat Genet 12. Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, 2006;38:1043–8. Getz G. GISTIC2.0 facilitates sensitive and confident localization of 33. Mouradov D, Domingo E, Gibbs P, Jorissen RN, Li S, Soo PY, et al. the targets of focal somatic copy-number alteration in human can- Survival in stage II/III colorectal cancer is independently predicted by cers. Genome Biol 2011;12:R41. chromosomal and microsatellite instability, but not by specific driver 13. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, mutations. Am J Gastroenterol 2013;108:1785–93. Bork P, et al. A method and server for predicting damaging missense 34. Kronenwett U, Huwendiek S, Ostring C, Portwood N, Roblick UJ, mutations. Nat Methods 2010;7:248–9. Pawitan Y, et al. Improved grading of breast adenocarcinomas based 14. Kumar P, Henikoff S, Ng PC. Predicting the effects of coding non- on genomic instability. Cancer Res 2004;64:904–9. synonymous variants on protein function using the SIFT algorithm. 35. Lee AJ, Endesfelder D, Rowan AJ, Walther A, Birkbak NJ, Futreal PA, Nat Protoc 2009;4:1073–81. et al. Chromosomal instability confers intrinsic multidrug resistance. 15. Miyoshi Y, Nagase H, Ando H, Horii A, Ichii S, Nakatsuru S, et al. Cancer Res 2011;71:1858–70. Somatic mutations of the APC gene in colorectal tumors: mutation 36. Kuznetsova AY, Seget K, Moeller GK, de Pagter MS, de Roos JA, cluster region in the APC gene. Hum Mol Genet 1992;1:229–33. Durrbaum M, et al. Chromosomal instability, tolerance of mitotic 16. Powell SM, Zilz N, Beazer-Barclay Y, Bryan TM, Hamilton SR, Thibo- errors and multidrug resistance are promoted by tetraploidization in deau SN, et al. APC mutations occur early during colorectal tumori- human cells. Cell Cycle 2015;14:2810–20. genesis. Nature 1992;359:235–7. 37. Poulogiannis G, McIntyre RE, Dimitriadi M, Apps JR, Wilson CH, 17. Su LK, Kinzler KW, Vogelstein B, Preisinger AC, Moser AR, Luongo Ichimura K, et al. PARK2 deletions occur frequently in sporadic colo- C, et al. Multiple intestinal neoplasia caused by a mutation in the rectal cancer and accelerate adenoma development in Apc mutant murine homolog of the APC gene. Science 1992;256:668–70. mice. PNAS 2010;107:15145–50. 18. Christie M, Jorissen RN, Mouradov D, Sakthianandeswaren A, Li S, 38. Patel AG, Sarkaria JN, Kaufmann SH. Nonhomologous end joining Day F, et al. Different APC genotypes in proximal and distal spo- drives poly(ADP-ribose) polymerase (PARP) inhibitor lethality in radic colorectal cancers suggest distinct WNT/beta-catenin signalling homologous recombination-deficient cells. PNAS 2011;108:3406–11. thresholds for tumourigenesis. Oncogene 2013;32:4675–82. 39. Shall S, de Murcia G. Poly(ADP-ribose) polymerase-1: what have we 19. Crabtree M, Sieber OM, Lipton L, Hodgson SV, Lamlum H, Thomas learned from the deficient mouse model? Mutat Res 2000;460:1–15. HJ, et al. Refining the relation between ‘first hits’ and ‘second hits’ 40. Dedes KJ, Wilkerson PM, Wetterskog D, Weigelt B, Ashworth A, Reis- at the APC locus: the ‘loose fit’ model and evidence for differences Filho JS. Synthetic lethality of PARP inhibition in cancers lacking in somatic mutation spectra among patients. Oncogene 2003;22: BRCA1 and BRCA2 mutations. Cell Cycle 2011;10:1192–9. 4257–65. 41. Helleday T. The underlying mechanism for the PARP and BRCA 20. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, synthetic lethality: clearing up the misunderstandings. Mol Oncol et al. Activation of beta-catenin-Tcf signaling in colon cancer by 2011;5:387–93. mutations in beta-catenin or APC. Science 1997;275:1787–90. 42. d’Adda di Fagagna F, Hande MP, Tong WM, Lansdorp PM, Wang ZQ, 21. Timinszky G, Till S, Hassa PO, Hothorn M, Kustatscher G, Nijmeijer B, Jackson SP. Functions of poly(ADP-ribose) polymerase in control- et al. A macrodomain-containing histone rearranges chromatin upon ling telomere length and chromosomal stability. Nat Genet 1999;23: sensing PARP1 activation. Nat Struct Mol Biol 2009;16:923–9. 76–80. 22. Masutani M, Nakagama H, Sugimura T. Poly(ADP-ribosyl)ation in 43. Kanai M, Tong WM, Sugihara E, Wang ZQ, Fukasawa K, Miwa M. relation to cancer and autoimmune disease. Cell Mol Life Sci 2005; Involvement of poly(ADP-Ribose) polymerase 1 and poly(ADP- 62:769–83. Ribosyl)ation in regulation of centrosome function. Mol Cell Biol 23. Nakanishi K, Cavallo F, Brunet E, Jasin M. Homologous recombina- 2003;23:2451–62. tion assay for interstrand cross-link repair. Methods Mol Biol 2011; 44. Nozaki T, Fujihara H, Watanabe M, Tsutsumi M, Nakamoto K, 745:283–91. Kusuoka O, et al. Parp-1 deficiency implicated in colon and liver 24. Pino MS, Chung DC. The chromosomal instability pathway in colon tumorigenesis induced by azoxymethane. Cancer Sci 2003;94: cancer. Gastroenterology 2010;138:2059–72. 497–500.

OF17 | CANCER DISCOVERY AUGUST 2018 www.aacrjournals.org

MACROD2 Loss Impairs PARP1 Activity and Promotes CIN in Colorectal Cancer RESEARCH ARTICLE

45. Boland CR, Sato J, Saito K, Carethers JM, Marra G, Laghi L, et al. ous tumor samples from single nucleotide polymorphism genotyping Genetic instability and chromosomal aberrations in colorectal data. Genome Biol 2010;11:R92. cancer: a review of the current models. Cancer Detect Prev 1998;22: 48. Moorhead PS, Nowell PC, Mellman WJ, Battips DM, Hungerford 377–82. DA. Chromosome preparations of leukocytes cultured from human 46. Wang K, Li M, Hadley D, Liu R, Glessner J, Grant SF, et al. PennCNV: peripheral blood. Exp Cell Res 1960;20:613–6. an integrated hidden Markov model designed for high-resolution 49. Langer S, Geigl JB, Ehnle S, Gangnus R, Speicher MR. Live cell cata- copy number variation detection in whole-genome SNP genotyping pulting and recultivation does not change the karyotype of HCT116 data. Genome Res 2007;17:1665–74. tumor cells. Cancer Genet Cytogenet 2005;161:174–7. 47. Yau C, Mouradov D, Jorissen RN, Colella S, Mirza G, Steers G, et al. A 50. An international system for human cytogenomic nomenclature. statistical approach for detecting genomic aberrations in heterogene- Cytogenet Genome Res 2016;149:1–2.

AUGUST 2018 CANCER DISCOVERY | OF18

MACROD2 Haploinsufficiency Impairs Catalytic Activity of PARP1 and Promotes Chromosome Instability and Growth of Intestinal Tumors

Anuratha Sakthianandeswaren, Marie J. Parsons, Dmitri Mouradov, et al.

Cancer Discov Published OnlineFirst June 7, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-17-0909

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2018/06/07/2159-8290.CD-17-0909.DC1

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerdiscovery.aacrjournals.org/content/early/2018/07/20/2159-8290.CD-17-0909. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.