Haploinsufficiency of an RB–E2F1–Condensin II Complex Leads To

Published OnlineFirst April 16, 2014; DOI: 10.1158/2159-8290.CD-14-0215

RESEARCH ARTICLE

Haploinsufﬁ ciency of an RB–E2F1–Condensin II Complex Leads to Aberrant Replication and Aneuploidy

Courtney H. Coschi 1 , 3, Charles A. Ishak 1 , 3, David Gallo 5 , 8, Aren Marshall 1 , 3, Srikanth Talluri1 , 3, Jianxin Wang 9 , Matthew J. Cecchini 1 , 3, Alison L. Martens 1 , 3, Vanessa Percy 1 , Ian Welch 4, Paul C. Boutros 6 , 7 , 9, Grant W. Brown 5 , 8, and Frederick A. Dick 1 , 2 , 3

ABSTRACT Genome instability is a characteristic of malignant cells; however, evidence for its contribution to tumorigenesis has been enigmatic. In this study, we demonstrate that the retinoblastoma protein, E2F1, and Condensin II localize to discrete genomic locations including major satellite repeats at pericentromeres. In the absence of this complex, aberrant replication ensues followed by defective chromosome segregation in mitosis. Surprisingly, loss of even one copy of the retinoblastoma gene reduced recruitment of Condensin II to pericentromeres and caused this phenotype. Using cancer genome data and gene-targeted mice, we demonstrate that mutation of one copy of RB1 is associated with chromosome copy-number variation in cancer. Our study connects DNA replication and chromosome structure defects with aneuploidy through a dosage-sensitive complex at pericentromeric repeats.

SIGNIFICANCE: Genome instability is inherent to most cancers and is the basis for selective killing of cancer cells by genotoxic therapeutics. In this report, we demonstrate that instability can be caused by loss of a single allele of the retinoblastoma gene that prevents proper replication and condensation of pericentromeric chromosomal regions, leading to elevated levels of aneuploidy in cancer. Cancer Discov; 4(7); 1–14. ©2014 AACR.

See related commentary by Hinds, p. xxx.

INTRODUCTION repair is undertaken before advancement into mitosis (2 ). Recent evidence suggests exceptions to this rule, as damage Fidelity of DNA replication and cell division are critical observed before mitosis is transmitted through M-phase ( 3, 4 ). processes in multicellular organisms. Unrepaired errors can These DNA lesions are often associated with replication stress be passed on to daughter cells and contribute to cancer ( 1 ). (4, 5), and their ability to evade checkpoints suggests that their In general, damaged DNA signals the cell cycle to arrest, and impact on genome instability and cancer may be signiﬁ cant ( 6 ).

Authors’ Afﬁ liations: 1 London Regional Cancer Program; 2 Children’s Corresponding Author: Frederick A. Dick, Cancer Research Labs, 790 Com- Health Research Institute; 3 Department of Biochemistry, and 4 Veterinary missioners Road East, London, ON N6A 3R7, Canada. Phone: 519-685- Services, Western University, London; Departments of 5 Biochemistry, 8620; Fax: 519-685-8616; E-mail: [email protected] 6 7 8 Medical Biophysics, and Pharmacology and Toxicology, Donnelly Cen- doi: 10.1158/2159-8290.CD-14-0215 tre, University of Toronto; and 9 Informatics and Biocomputing Program, Ontario Institute for Cancer Research, Toronto, Ontario, Canada ©2014 American Association for Cancer Research. Note: Supplementary data for this article are available at Cancer Discovery Online (http://cancerdiscovery.aacrjournals.org/).

OF1 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

Replication stress can lead to fork stalling and chromo- imbalances in nucleotide pools (23 ). Thus, an in-depth somal aberrations (7 ). It can create short gaps in sequence, understanding of pRB function in genome stability has yet or unresolved replication intermediates (3 , 7 , 8 ). Much of to emerge, in part because these phenotypes are not mutu- the evidence for replication stress phenotypes has been ally exclusive. demonstrated using repetitive elements in yeast that In humans, the RB1 gene is lost in hereditary retinoblas- require the Condensin complex for fi delity of replication toma, and survivors face an elevated risk of other cancers, and accurate segregation in mitosis (7 , 9 , 10 ). Mammals such as osteosarcomas, throughout their lives (26 ). RB1 loss contain two Condensin complexes, but only Condensin occurs by the “two hit” model proposed by Knudson ( 27 ). II is constitutively nuclear, suggesting it may play roles in Because loss of heterozygosity was the rate-limiting step both DNA replication and chromosome condensation (11 ). in the genesis of retinoblastoma, Knudson concluded that In addition to its role in mitotic chromosome condensa- heterozygosity likely did not contribute to tumorigenesis. tion, chromosome shape, and controlling recombination, This premise is recapitulated in Rb1 +/− mice that develop Condensin II has also been implicated in DNA replication pituitary tumors characterized by loss of the remaining and damage repair (12 ). In particular, Condensin II func- wild-type allele (28 ). Curiously, a number of experiments tions in the resolution of sister chromatids immediately have suggested that loss of one copy of pRB may contribute following replication in S-phase (13 ). Despite these roles to cancer progression in other contexts. Crosses between for Condensin II, we know little about how it is recruited Rb1- and Trp53-defi cient mice revealed that Rb1+ /−; Trp53 −/− to specifi c genomic locations, such as repetitive sequences, mice develop some tumors without losing the remaining to carry out these functions. Rb1 allele (29 ). In addition, loss of one copy of Rb1 in mouse The retinoblastoma protein (pRB) is generally thought of osteoblasts or embryonic stem (ES) cells confers genome

as a regulator of the G 1 to S-phase transition (14 ). However, instability (30, 31). Consequently, phenotypes of Rb1 het- evidence suggests that pRB also contributes to functions erozygosity have been previously reported, but a clear link

beyond G1 . For example, pRB has been implicated in an to cancer incidence or progression is lacking. In addition, a S-phase checkpoint to repair DNA breaks and to regulate mechanism that preserves genome stability in which the rela- initiation of DNA replication (15–18 ). In addition, pRB tive supply of pRB is critical has yet to be described. facilitates chromosome condensation in mitosis, particu- pRB and Condensin II contribute to the integrity of larly at pericentromeres, although there is no evidence that pericentromeric heterochromatin, which is prone to the it physically localizes to this region (19–22 ). pRB-defi cient effects of replication stress. In this study, we demonstrate cells have increased genome instability characterized by that Rb1 -mutant cells experience replication stress, par- spontaneous DNA breaks and chromosome missegrega- ticularly at pericentromeric regions. We show that a com- tion in mitosis (23 ). Surprisingly, these outcomes have not plex composed of pRB, E2F1, and Condensin II localizes been attributed to a specifi c mechanism of pRB action. For to pericentromeric repeats. Intriguingly, replication and example, altered regulation of E2F transcription has been chromosome structure defects were recapitulated in hetero- implicated in shifts in intermediary metabolism (24, 25), zygous Rb1 -mutant fi broblasts, indicating that this genomic misexpression of spindle assembly checkpoint genes, and instability phenotype is gene-dosage dependent. Similarly,

JULY 2014CANCER DISCOVERY | OF2

RESEARCH ARTICLE Coschi et al.

A –/– Rb1+/+ Rb1ΔL/ΔL Rb1 x x x z z z z x z x z x

DAPI y y y y y y γ H2AX

B C D Fra16d Rb1+/+ Rb1ΔL/ΔL Rb1–/– Chr 8 P < 0.002 P 0.75 < 0.00001 116 mb 117 mb 118 mb 119 mb 120 mb

0.60 0.75 Input 0.45 0.60 0.45 Rb1+/+ 0.30 0.30 Rb1ΔL/ΔL Proportion of cells 0.15 colocalized foci 0.15 –/–

Proportion of cells with Rb1 0.00 0.00 012345+ Number of γH2AX foci Wwox E F Rb1+/+ G Rb1+/+ Rb1ΔL/ΔL Rb1–/– 0.0015 0.008 * * 0.006 0.26 0.001 LINE Tandem rep. Major sat. ERV SINE 0.004 Rb1ΔL/ΔL Input (%) 0.0005 Input (%) –/– Rb1 0.002

0 0 –0.12 IgG γH2AX IgG γH2AX

Figure 1. Rb1 -mutant MEFs exhibit increased γH2AX foci that are enriched at pericentromeric DNA. A, immunofl uorescence microscopy of γH2AX staining is shown in red and cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Red arrows, colocalization; white arrows, lack of colocalization; scale bars, 10 μm. B, the quantity of foci per cell was determined for each genotype and compared using a χ 2 test. Rb1 +/+, n = 103; Rb1Δ L/ΔL, n = 114; Rb1− /−, n = 63. C, the proportion of cells with γH2AX foci colocalizing with DAPI-rich foci was determined by confocal 3D rendering and compared using a χ2 test. D, ChIP-seq analysis was performed for γH2AX, and tracks comparing abundance at Fra16d among different Rb1 genotypes are shown. A γ red box indicates this locus on a chromosome 8 ideogram. E, heatmap to show log2 ratios of the abundance of Rb1 -mutant H2AX precipitable reads per million mapped reads versus wild-type γH2AX precipitated reads. ERV, endogenous retrovirus. F and G, ChIP-qPCR for γH2AX to quantitate major satellite (major sat.) repeats. *, P < 0.05 using a t test; n = 3. we demonstrate that normal RB1 +/− fi broblasts from patients RESULTS with hereditary retinoblastoma exhibit the same aberrant replication characteristics followed by mitotic errors. Using Defective Rb1 Causes Deposition of gH2AX at genotype and copy-number variation data from the Cata- Pericentromeric Repeats logue of Somatic Mutations in Cancer (COSMIC) database, In normal human fi broblasts, loss of RB results in γH2AX we demonstrate that RB1 +/− lymphoma and sarcoma can- foci (32 ). We compared γH2AX foci levels in mouse embryonic cer cell lines exhibit as much genomic instability as RB1 −/− fi broblasts (MEF) from Rb1 + /+ and Rb1 −/− animals (Fig. 1A ). In cell lines. Finally, using gene-targeted mice bearing a single addition, we analyzed a targeted mutant (Rb1 ΔL) that antago- mutant allele that is defective for recruiting Condensin II nizes interactions between Condensin II and the LXCXE to chromatin (called Rb1 Δ L), we demonstrate that tumors binding cleft region on pRB, while leaving E2F interactions from Rb1 ΔL/+ ; Trp53− /− mice have increased chromosome copy- and many pRB-dependent aspects of cell-cycle control intact number variation compared with Trp53 − /− controls. This (19, 20, 33, 34). There was a signifi cant increase in γH2AX provides proof-of-principle that dosage sensitivity of a pRB– foci in Rb1 − /− MEFs compared with wild-type ( Fig. 1B ). DAPI E2F1–Condensin II complex compromises replication fi delity counterstaining produces punctate foci at pericentromeric and leads to aneuploidy. regions. Analysis by confocal microscopy revealed that γH2AX

OF3 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

accumulation was more frequent at DAPI-rich spots in Rb1 −/− maintenance complex (MCM) 3, and DNA Polδ at pericentro- MEFs than in controls ( Fig. 1C ). A similar, less pronounced meres in Rb1 ΔL/ΔL cells ( Fig. 2F–H ). Together these data indicate effect was observed in Rb1 ΔL/ΔL cells with respect to γH2AX that replication of this genomic region is defective. Investiga- focus abundance and localization ( Fig. 1B and C ). This indi- tion of γH2AX staining of mitotic cells suggests similarities cates that defi ciency in pRB–LXCXE–dependent functions between this replication phenotype and reports of under- alone is suffi cient to cause γH2AX foci. replication (Supplementary Fig. S2). Furthermore, telomere To further investigate the localization of γH2AX depo- analysis suggests no loss of integrity (Supplementary Fig. S3). sition, we performed chromatin immunoprecipitation In sum, Rb1ΔL/ΔL cells have abnormal replication rates and (ChIP)-sequence analysis in which γH2AX was precipitated molecular markers that are consistent with replication stress. from wild-type, Rb1 ΔL/ΔL, and Rb1 −/− chromatin. Sequence Although not entirely consistent with phenotypes generated by reads were mapped onto unique and repetitive genomic treating cells with hydroxyurea, reduced RPA levels at pericentro- regions to compare their relative abundance. Analy- meres and elevated levels of replication machinery and γH2AX sis on a megabase scale did not reveal a particular bias further suggest that replication in this region is impaired. The of γH2AX enrichment within damage-susceptible frag- simplest explanation for γH2AX deposition at major satellite ile regions (Fig. 1D ). Model-based Analysis of ChIP-Seq repeats is that they are the result of replication stress. (MACS) fails to detect reproducible enrichment in Rb1ΔL/ΔL and Rb1 − /− cells at this location (Supplementary Fig. S1). pRB, E2F1, and Condensin II Form a In accordance with elevated quantities of γH2AX foci in Complex at Pericentromeres Rb1ΔL/ΔL and Rb1 −/− MEFs, failure to detect local regions of A similar phenotype of spontaneous γH2AX foci has been enrichment compared with wild-type implies that γH2AX reported for a number of known pRB-interacting proteins with levels are uniformly increased across unique regions of the roles in replication. Increased γH2AX has been reported in cells genome in Rb1 Δ L /Δ L and Rb1 − /− cells. defi cient for the Condensin II subunit structural maintenance Mapping reads to nonunique regions of the genome of chromosome protein 2 ( SMC2; 38 ) and E2F1 ( 39 ). Notably, revealed that major satellite repeats, which comprise much of loss of other activator E2Fs did not cause γH2AX foci ( 39 ). In the DAPI-rich pericentromeric foci, were enriched relative to MEFs, shRNA depletion of CAP-D3 (a Condensin II subunit) other nonunique comparators ( Fig. 1E ). ChIP-qPCR analysis caused similar patterns of γH2AX foci as described above (Sup- also demonstrated that γH2AX is elevated at major satellites plementary Fig. S4). We also discovered similar γH2AX stain- in Rb1 ΔL/ΔL and Rb1 −/− genotypes compared with wild-type ing in E2f1 − /− cells (Supplementary Fig. S4). (Fig. 1F and G). Therefore, defi ciency for pRB increases Because of these similarities, we investigated whether pRB, γH2AX deposition throughout the genome, with particular E2F1, and CAP-D3 localize to pericentromeric heterochroma- enrichment at pericentromeric repeats, and loss of pRB– tin in interphase cells. ChIP-qPCR for major satellites in chro- LXCXE interactions alone contributes to this effect. matin precipitated with pRB, CAP-D3, and E2F1 antibodies revealed that they are all present at these repeats ( Fig. 3A Aberrant DNA Replication in Rb1 -Mutant MEFs and B). Furthermore, CAP-D3 recruitment was dependent The spontaneous formation of γH2AX foci in Rb1-mutant on E2F1 and pRB because its binding was diminished in cells, and their frequent association with major satellite repeats, E2f1− /− , Rb1− /− , and Rb1Δ L/ΔL MEFs ( Fig. 3B and C ). To bet- is inconsistent with DNA double-strand breaks (35 ). Because ter understand Condensin II recruitment to chromatin and Rb1− /− cells have multiple confounding defects that contribute its relationship to pRB, we performed ChIP-seq for CAP- to genome instability, we focused on the Rb1 ΔL/ΔL mutant as it D3 in wild-type and Rb1 ΔL/ΔL MEFs. This revealed 8,056

maintains normal G1 –S regulation during proliferation ( 33 ). peaks of local enrichment for CAP-D3 within unique genome We pulse-labeled Rb1 Δ L/ΔL cells with CIdUrd and IdUrd and sequences, and 6,979 of these were lost in the Rb1 ΔL/ΔL cells. stained DNA ﬁ bers to visualize replication fork progression Intriguingly, CAP-D3 was not recruited to the promoters of (Fig. 2A ). This revealed a net increase in fork rate in Rb1 Δ L/ΔL well-known pRB–E2F transcriptional targets such as Cyclin mutants compared with controls (Fig. 2B ). Accelerated forks, E1 ( Ccne1), Mcm3 , and p107 ( Rbl1; Fig. 3D ). However, it was although less common than slowed forks, are also associated recruited to the Lmnb2 replication origin in a pRB-dependent with replication stress and are prone to stalling (5 , 36 , 37 ). In manner (Fig. 3D ). This is signiﬁ cant because pRB and E2F1 addition, elevated replication fork asymmetry (Fig. 2C ), and are known to localize to this origin (16 , 18 ), even though it elevated levels of replication protein A (RPA) 32 phospho- does not possess a canonical E2F DNA sequence element. serine33, further suggest replication stress (Fig. 2D and Sup- We isolated the putative pRB–E2F1–Condensin II complex plementary Fig. S2). Thus, from a genome-wide perspective, to investigate its properties. We generated a biotinylated major Rb1Δ L/ΔL cells replicate their DNA abnormally and display satellite repeat probe, and mixed it with nuclear extract from markers of replication stress. wild-type and Rb1 − /− cells. Major satellite-associated com- Because γH2AX shows greater accumulation at major satel- plexes were collected on streptavidin beads, and proteins were lite repeats, we investigated this region further. We used ChIP- detected by immunoblotting. SMC2, CAP-D3, pRB, and E2F1 qPCR assays of RPA to investigate its levels at major satellites were isolated together from wild-type extracts, but none of in Rb1 Δ L /Δ L MEFs ( Fig. 2E ). This revealed lower levels of RPA at these components interacted with the major satellite sequence pericentromeres in Rb1 ΔL/ΔL cells, suggesting that the genera- when pRB was absent (Fig. 4A ). Major satellites were bound tion of ssDNA needed for replication or break repair is inhib- in a sequence-dependent manner as a control probe failed to ited. Furthermore, ChIP-pPCR assays reveal increased levels of precipitate them, and their interaction was sensitive to com- proliferating cell nuclear antigen (PCNA), minichromosome petition by exogenous major satellite sequences (Fig. 4B ). We

JULY 2014CANCER DISCOVERY | OF4

RESEARCH ARTICLE Coschi et al.

DNA

CIdUrd

IdUrd

B C D 7 1,000 P = 0.015 6 P = 0.0001 800 5 L L/ Δ 600 +/+ Δ 4 Rb1 Rb1

3 400 RPA pS33 - 1 1.9 2 Asymmetry (%) Fork rate (kb/min) 200 1 RPA32 -

0 0 Rb1+/+ Rb1ΔL/ΔL Rb1+/+ Rb1ΔL/ΔL Median: 1.31 1.51 Median: 32.3 43.1 E F G H Rb1+/+ Rb1ΔL/ΔL 0.04 0.020 0.002 0.10 * * * 0.0016 0.08 0.03 0.015

0.0012 0.06 0.02 0.010 0.0008 0.04 Input (%) 0.01 0.005 0.0004 * 0.02

0 0.00 0.00 0.000 IgG RPA IgG PCNA IgG MCM3 IgG DNA Polδ

Figure 2. Replication abnormalities in Rb1 -mutant cells. A, fl uorescence microscopy of a replicating fork is shown by CIdUrd labeling (P1, green) and IdUrd (P2, red), and light blue bars mark unlabeled DNA. B, fork rates were measured for Rb1 + /+ (n = 718) and Rb1Δ L/ΔL (n = 855). Median values were compared using the Mann–Whitney U test. C, diverging forks, as shown in A, were measured and the percentage asymmetry was determined, Rb1 + /+ (n = 141) and Rb1 ΔL/ΔL ( n = 174). Median values were compared as in B. D, extracts were blotted for RPA and the indicated phosphorylated RPA-S33. Rela- tive intensity of RPA pS33 is shown below each lane. E, ChIP-qPCR analysis of RPA at major satellite repeats in Rb1 + /+ and Rb1 ΔL/ΔL MEFs. *, P < 0.05 using a t test; n = 3. F–H, ChIP-qPCR analysis of PCNA, MCM3, and DNA Polδ occupancy at major satellites. also reconstituted this complex using nuclear extracts from Previous reports indicate that binding of E2F1 to pRB human C33A cells that express HA-E2F1/HA-DP1 (Fig. 4C ). through an alternate confi guration called the “specifi c” inter- Major satellite probes were mixed with extract that contained action changes the sequence specifi city of E2F1, reducing GST-RB Large Pocket (LP). Figure 4D demonstrates that affi nity for canonical E2F recognition sequences (40 ). We neither E2F1 nor CAP-D3 was recruited to the major satellite used MEFs from two new Rb1 -mutant mouse lines to inves- probe without GST-RBLP. This suggests that E2F1’s ability to tigate whether the “specifi c” interaction explains cooperative interact with these sequences is not autonomous, as it is in the DNA binding at major satellites by pRB and E2F1. Rb1 Δ G recognition of cell-cycle target genes, but requires a coopera- contains two amino acid substitutions (R461E and K542E) tive contribution from pRB. Furthermore, ChIP of E2F1 from that prevent binding to activator E2Fs to regulate transcrip- Rb1− /− MEFs revealed cooperativity in vivo as E2F1 binding was tion at cell-cycle promoters; however, it maintains the ability diminished in the absence of pRB (Fig. 4E ). to bind to E2F1 through the separate “specifi c” mechanism

OF5 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

A 0.0008 0.0008 Figure 3. pRB, E2F1, and Condensin II localize to peri- Rb1+/+ centromeric DNA. A, ChIP-qPCR for major satellite repeats 0.0006 0.0006 Rb1+/+ < Rb1–/– using antibodies against pRB and CAP-D3. *, P 0.05 using a t test; n = 3. B, chromatin was precipitated with E2F1 and 0.0004 0.0004 CAP-D3 antibodies and major satellite repeats were ampli- ﬁ ed. C, ChIP-qPCR using a CAP-D3 antibody to precipitate Input (%) Input (%) major satellite repeats in Rb1 + /+, Rb1 −/−, and Rb1 Δ L/ΔL MEFs. 0.0002 0.0002 D, ChIP-seq was performed for CAP-D3 using chromatin * + + Δ Δ from Rb1 / and Rb1 L/ L MEFs and tracks are shown for 0 0 selected genomic regions. E2F binding sites present in the IgG pRB IgG CAP-D3 promoters of Ccne1, Mcm3 , and Rbl1 are indicated by red B 0.04 0.0006 boxes, and regions of signiﬁ cant enrichment (MACS peaks) E2f1+/+ E2f1+/+ are indicated in green. E2f1–/– E2f1–/– 0.03 * 0.0004 0.02 * Input (%) Input (%) 0.0002 0.01

0 0 C IgG E2F1 IgG CAP-D3 0.0009 0.0008 Rb1+/+ Rb1+/+ Δ Δ Rb1–/– 0.0006 Rb1 L/ L 0.0006 0.0004 * Input (%) 0.0003 Input (%) 0.0002 * 0 0 IgG CAP-D3 IgG CAP-D3 D Ccne1 promoter Mcm3 promoter

Rb1+/+

Rb1ΔL/ΔL MACS Peaks

Ccne1 Mcm3 1 kb 1 kb Rbl1 promoter Lmnb2 origin Rb1+/+

Rb1ΔL/ΔL MACS Peaks

Rbl1 Timm13 Lmnb2 1 kb 1 kb

(41, 42). Conversely, Rb1 Δ S is unable to bind E2F1 through the γH2AX focus location, we suggest that these proteins form “speciﬁ c” interaction because of an F832A substitution, but it a complex to facilitate DNA replication and chromosome maintains its ability to bind activator E2Fs in a manner that condensation, and its absence preferentially affects major regulates canonical E2F transcription ( 42, 43 ). ChIP for CAP- satellite repeats. D3 in Rb1 ΔG/ΔG MEFs demonstrated no change in its loading at pericentromeric heterochromatin, whereas in Rb1 Δ S/ΔS cells, g H2AX Distribution, Replication Stress Markers, CAP-D3 was absent ( Fig. 4F ). and Recruitment of Condensin II Are Sensitive to This indicates that pRB, E2F1, and Condensin II form Rb1 Gene Dosage a stable complex with major satellite repeats, and other Major satellites account for 3% of the mouse genome, unique loci in the genome such as the Lmnb2 origin (Supple- suggesting that gene dosage of Rb1 may be important mentary Fig. S5). On the basis of phenotypic similarities of for supplying enough of this complex to ensure stability

JULY 2014CANCER DISCOVERY | OF6

RESEARCH ARTICLE Coschi et al.

A B C Major satellite Input 10% Input pulldown Probe: GST- +/+ +/+ –/– –/– Maj. sat. ++–– NE RBLP Rb1 Rb1 Rb1 Rb1 Control ––+– CAP-D3 Competitor: SMC2 pRB Maj. Sat. –+–– GST-RBLP CAP-D3 CAP-D3 HA-E2F1 HA-DP1 pRB pRB E2F1 E2F1 E2F1

D E F Major satellite Rb1+/+ pulldown Rb1+/+ Rb1–/– 0.025 0.4 Rb1ΔG/ΔG NE: + + Rb1ΔS/ΔS GST-RBLP: 0.02 + – 0.3 CAP-D3 0.015 * 0.2

GST-RBLP Input (%) Input (%) 0.01

0.1 HA-E2F1 0.005 * HA-DP1 0 0.0 IgG E2F1 IgG CAP-D3

Figure 4. pRB, E2F1, and Condensin II form a complex on pericentromeric DNA. A, nuclear extracts (NE) were mixed with streptavidin beads and a biotinylated major satellite (maj. sat.) repeat containing probe. Precipitated proteins were analyzed by SDS-PAGE and Western blotting. B, wild-type nuclear extracts were mixed with the indicated probes and/or competitor DNA. C, RB1 -defi cient C33A cells were transfected with HA-E2F1 and HA-DP1 expression vectors and nuclear extracts were prepared. Western blots of relevant proteins are shown. D, nuclear extracts were mixed with streptavidin beads and biotinylated major satellite probes, either with or without GST-RBLP. Associated proteins were precipitated and analyzed by Western blotting. E and F, ChIP-qPCR for E2F1 was performed, and major satellite repeat DNA was amplifi ed by real-time PCR. *, P < 0.05 using a t test; n = 3. F, ChIP-qPCR using a CAP-D3 antibody was performed to precipitate major satellite repeats from Rb1 + /+, Rb1 ΔG/ΔG, and Rb1 Δ S/ΔS MEFs. of these repeats. Fluorescence microscopy revealed a sig- the centromere. To extend our investigation of Rb1 gene nifi cant increase in γH2AX foci in both Rb1+ /− and Rb1 ΔL/+ dosage sensitivity, we performed ChIP-qPCR for CAP-D3 MEFs compared with the wild-type control (Fig. 5A and B). at pericentromeric heterochromatin in Rb1 ΔL/ + cells and Moreover, γH2AX foci were found at pericentromeric het- determined that it was reduced at major satellites in Rb1 ΔL/ + erochromatin in Rb1 +/− and Rb1 ΔL/+ MEFs more frequently MEFs relative to wild-type control ( Fig. 5G ). This suggests than in Rb1 + /+ controls, and were almost as abundant as that loss of LXCXE interactions in even one copy of Rb1 their homozygous mutant counterparts ( Fig. 5C ). Western reduced the supply of this complex at these repeats. In addi- blotting revealed elevated γH2AX levels in Rb1Δ L /Δ L , Rb1 +/−, tion, heterozygous Rb1 -mutant MEFs displayed anaphase and Rb1− /− MEFs relative to Rb1 +/+ (Fig. 5D). Similarly, bridges, tangled centromeres, and aneuploidy as reported phosphorylated RPA-S33 displays increased abundance in previously for homozygous Rb1 mutants (Supplementary all Rb1-mutant genotypes (Fig. 5E ). As with Rb1 −/− and Fig. S7 and S8). Rb1ΔL/ΔL , ChIP-seq analysis of the distribution of γH2AX in These data demonstrate that single Rb1 null and Rb1 ΔL Rb1 ΔL/+ MEFs also revealed enrichment at major satellites alleles compromise the ability to prevent γH2AX foci and (Fig. 5F ). Using ChIP-seq data from all three Rb1-mutant aneuploidy. γH2AX and phosphorylated RPA-S33 levels are genotypes, we compared the reproducibility of enrichment similar between heterozygous and homozygous genotypes, among repeat elements (Supplementary Fig. S6). This and these translate into a similar frequency of mitotic errors. revealed that major and minor satellites have a greater On the basis of these direct comparisons of homozygous proportion of reads than wild-type. This is noteworthy mutant, heterozygous, and wild-type genotypes, we describe because minor satellites are adjacent to major satellites at Rb1 as haploinsuffi cient for this function.

OF7 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

A B C DAPI Rb1+/+ γH2AX P < 0.002 Rb1ΔL/+ P Rb1ΔL/ΔL < 0.002 Rb1+/– Rb1+/+ 0.7 Rb1–/– 0.7

0.6 0.6

0.5 0.5 Rb1+/+ Rb1ΔL/+

0.4 0.4 ΔL/ΔL Rb1ΔL/+ Rb1 Rb1+/– 0.3 0.3 Rb1–/– colocalized foci colocalized

Proportion of cells 0.2 0.2 Proportion of cells withProportion of cells 0.1 0.1 Rb1+/– 0.0 0.0 0 1 2 3 4 5+ Number of γH2AX foci D E Δ L/+ Δ L/ L –/– +/+ +/– Δ L / Δ L/+ +/+ +/– –/– Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 γH2AX - RPA pS33 - 1 0.9 1.4 1.6 1.6 1 1.8 1.8 1.7 2.2 RPA32 - Histones -

F G 4

3 0.26 Rb1+/+ Rb1ΔL/+ Major sat. ERV SINE LINE Tandem rep. 2 Rb1ΔL/+ * Rb1ΔL/ΔL Rb1–/– 1 Percentage of Percentage input

–0.12 0 IP: IgG CAP-D3

Figure 5. Rb1 heterozygous MEFs have elevated γH2AX and underloading of CAP-D3 at major satellites. A, immunoﬂ uorescence microscopy of γH2AX staining in Rb1+ /+, Rb1 ΔL/+, and Rb1 +/− MEFs is shown. Scale bars, 10 μm. B, the quantity of γH2AX foci per cell was determined for Rb1+ /+, Rb1 Δ L/+, and Rb1 + /− MEFs and compared with homozygous mutants from Fig. 1. The quantity of foci was compared using a χ 2 test. Rb1 ΔL/+, n = 114; Rb1+ /−, n = 63. C, quantiﬁ cation of DAPI foci colocalization with γH2AX foci by confocal microscopy. The proportion of colocalization was compared using a χ 2 test. D, γH2AX levels were detected by Western blotting and Coomassie staining of histones was used as a loading control. Relative intensity of γH2AX is shown below. E, RPA32 and RPA32 phospho serine 33 levels were detected by Western blotting. Relative intensity of RPA pS33 is shown below. F, heatmap γ depicting log2 ratios of H2AX precipitated sequence tags in the indicated mutants relative to wild-type control. G, ChIP-qPCR using a CAP-D3 antibody in Rb1+ /+ and Rb1 ΔL/+ MEFs was used to detect major satellite repeats. *, P < 0.05 using a t test; n = 3.

Human RB1 + /- Cells Exhibit g H2AX Foci, us to determine whether a similar phenotype is present in Phosphorylated RPA-S33, Mitotic Defects, normal fi broblasts from patients with hereditary retinoblas- and Aneuploidy toma (RB1 +/−). We obtained patient fi broblasts, confi rmed their heterozygous status by sequencing (Supplementary The kinetics of RB1 loss in retinoblastoma gave rise to the Fig. S9), and compared them with unrelated human fi bro- now famous “two-hit” hypothesis (27 ). However, our data on blast cells (IMR90, BJ, and WI38). This revealed that patient haploinsuffi ciency in Rb1 -mutant mouse fi broblasts motivated fi broblasts exhibited increased γH2AX foci ( Fig. 6A and B).

JULY 2014CANCER DISCOVERY | OF8

RESEARCH ARTICLE Coschi et al.

A DAPI B γ H2AX P < 0.00001 0.60 0.50 IMR90 Pooled control 0.40 GM_01408 0.30 GM_01123 GM_06418 0.20

Proportion of cells 0.10 GM_06418 0.00 054321 + Number of γH2AX foci

C D Metaphase plate 06:09:34 06:42:34 06:51:34

IMR90

IMR90 GM_01408 GM_01123 GM_06418 33’ RPA pS33 - 02:27:34 06:15:34 06:27:34 1 2.0 1.8 1.6 GM_01408 RPA32 - 228’ E Cell line Cell Anaphase bridges Prophases Average time from onset of divisions and condensation to onset of observed metaphases anaphase (min) IMR90 31 19 13 45 BJ 42 14 17 61 WI38 18 33 49 GM_01123 10 9# 7 211† GM_01408 31 19# 20 85† GM_06418 84#6 120† FGHI Whole chromosome Total segments Total abnormal Chromothriptic changes segments regions NS NS NS NS * * * * * 40 * 600 * 600 15 *

30 400 400 10 20 200 200 5 10 Quantity per cell line Quantity per cell line Quantity per cell line 0 Quantity per cell line 0 0 0 /– /– /– /– –/– –/– –/– –/– +/+ +/+ +/+ +/+ + + + + Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1 Rb1

Figure 6. Normal RB1 + /− cells have elevated γH2AX foci, mitotic errors, and aneuploidy. A, immunofl uorescence microscopy for γH2AX in control and RB1 + /− patient fi broblasts. Scale bars, 10 μm. B, quantitation of γH2AX foci in control and each patient fi broblast isolate. Foci were compared for patient fi broblasts against pooled control data (IMR90, BJ, and WI-38) using a χ 2 test. Pooled control, n = 246; GM_01408, n = 79; GM_01123, n = 93; GM_06418, n = 71. C, Western blotting was used to detect phosphorylated RPA pS33 and total RPA32. Relative intensity of RPA pS33 is displayed. D, cells were transduced with H2B-GFP, and video microscopy was performed to capture phase contrast and GFP images over 15 hours. The left-most image shows the onset of prophase, and the middle image is of the metaphase plate just before the onset of anaphase (elapsed time since the onset of prophase is shown in yellow). The right-most image shows cells in telophase. Scale bars, 50 μm. E, summary of video microscopy data showing the number of divisions observed along with the number that contained anaphase bridges. Similarly, the number of cells observed to complete prophase and metaphase is shown along with the average length of these two phases. Each patient fi broblast is compared with pooled control data. #, P < 0.05, χ2 test; †, P < 0.05, t test. F–I, quantitation of genomic abnormalities in mesenchymal-derived cancer cell lines that are wild-type ( n = 15), hemizygous (n = 10), or null ( n = 12) for RB1 . Means were compared using a t test. *, P < 0.05. NS, not signifi cant.

OF9 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

As with heterozygous mouse fi broblasts, they also contained tion of the wild-type and Rb1 Δ L alleles in tumor DNA from elevated levels of phosphorylated RPA-S33 compared with Rb1Δ L/+; Trp53− /− mice showed retention of the wild-type allele control ( Fig. 6C ). We also used video microscopy of H2BGFP- (Fig. 7B ). Thymic lymphomas from Rb1 ΔL/+; Trp53 −/− mice expressing cells to search for mitotic defects ( Fig. 6D ). All demonstrated a more aggressive histology compared with three patient fi broblast lines exhibited a signifi cant delay in Trp53− /− controls as evidenced by smaller cytoplasmic-to- progression to metaphase from the onset of chromosome nuclear area (Fig. 7C ). This is very similar to the histology condensation ( Fig. 6E ). Finally, patient cells also showed described in Rb1 ΔL/ΔL; Trp53 −/− thymic lymphomas ( 19 ), and a signifi cant increase in mitoses with anaphase bridges Rb1Δ L/+; Trp53 − /− mice exhibited a shift in tumor spectrum (Fig. 6E ). Taken together, these experiments suggest that from thymic lymphoma, commonly reported in Trp53 −/− normal human fi broblast cells heterozygous for RB1 are mice, toward sarcomas (Supplementary Fig. S10). Thus, characterized by replication and mitotic errors similar to our one defective copy of Rb1 can have similar effects on disease mouse data. progression as mutation of both alleles. Because thymic lym- Given the direct impact of haploinsuffi ciency for RB1 phomas were common among all genotypes, we investigated on γH2AX abundance and mitotic errors, we sought evi- their relative levels of chromosome gains and losses. Figure +/− dence for aneuploidy in RB1 cancers. Using copy-number 7D shows log2 ratio plots from a male versus female control variation data from COSMIC ( 44 ), we asked whether cancer- hybridization and representative Rb1 ΔL/+; Trp53 −/− tumors. derived cell lines heterozygous for RB1 exhibit increased The number of whole-autosome gains and losses in thymic levels of genomic abnormalities. Because retinoblastoma lymphomas from Rb1 ΔL/+; Trp53 −/− mice was greater than survivors are highly prone to second primary neoplasms that Trp53− /− controls, and statistically indistinguishable from are of mesenchymal origin (26 ), we compared copy-number Rb1Δ L/ΔL; Trp53 −/− tumors ( Fig. 7E ). data for RB1 +/+, RB1 + /− , and RB1− /− cell lines from this germ Earlier experiments demonstrated that one Rb1 ΔL allele layer. We sorted genomic abnormalities into four categories: can cause replication and chromosome structure–associated whole-chromosome gains and losses, total genomic seg- phenotypes. Taken together with this tumor study, our data ments, total abnormal genomic segments, and chromothrip- suggest that gene dosage-sensitive effects on replication and tic regions. RB1 + /− and RB1− /− cancers exhibited signifi cantly mitotic chromosomes have the potential to compromise pRB more whole-chromosome changes than cell lines retain- function in vivo leading to aneuploidy. ing both wild-type copies of RB1 (Fig. 6F). Importantly, RB1+ /− lines exhibited as many whole-chromosome changes as RB1 −/− cells. This trend was observed for the other meas- DISCUSSION ures of instability as RB1 + /− and RB1 −/− lines exhibited sig- The pRB–E2F1–Condensin II complex identifi ed in this nifi cantly more abnormalities than wild-type and were not study affects chromatin from S-phase to mitosis. Our fi nd- statistically different from each other (Fig. 6G–I ). To con- ings suggest how replication and structural defects at peri- fi rm these results, we obtained representative lines to verify centromeres may be connected to common chromosomal RB1 copy number and expression of pRB (Supplementary abnormalities in cancer. Reduced occupancy by this complex Fig. S9). at pericentromeric repeats may lead to unresolved replica- These data indicate that loss of one copy of RB1 may tion intermediates and compromise the integrity of centro- exhibit haploinsuffi ciency in humans, namely through the meres and kinetochores. Misshapen kinetochores can lead ability of pRB to prevent the accumulation of γH2AX foci, to merotelic microtubule attachments, lagging anaphase phosphorylated RPA-S33, and mitotic errors. Moreover, this chromosomes, and ultimately gains and losses of whole chro- is associated with increased chromosomal abnormalities in mosomes in daughter cells ( 45 ). Alternatively, centromere cell lines at a level comparable with that found in RB1 −/− fusions or recombination with other chromosomes ( 7, 8 ) cells. A number of possibilities exist that may connect rep- may lead to gains or losses of whole chromosome arms, and lication and chromatin structure defects at pericentromeric these are the most common segmental chromosome changes sequences with common chromosomal aberrations found in in human cancer (46 ). Finally, lagging chromosomes created cancer, and they are discussed below. by these mechanisms are more likely to become incorporated into micronuclei and undergo chromothripsis in subsequent Haploinsuffi ciency of Rb1 Contributes to cell cycles (47 ). In this way, defects in pericentromeric chro- Aneuploidy in a Mouse Model of Cancer mosomal regions may have the ability to cause a wide spec- We also sought evidence for haploinsuffi ciency in vivo trum of chromosomal abnormalities, similar to those found using mouse models. Because Rb1 − /− mice are inviable, com- in RB1 +/− and RB1− /− cancer cell lines. Intriguingly, augmenta- parisons between heterozygous, homozygous mutant, and tion of Cohesin function can suppress replication and chro- wild-type populations are not possible. However, Rb1 Δ L/ΔL mosomal abnormalities caused by loss of RB1 function. This mice are viable and can therefore be used to study whether suggests that these chromosomal defects may potentially be haploinsuffi ciency in Rb1 ΔL/+ may contribute to oncogenesis. suppressed therapeutically (48 ). We crossed Rb1Δ L/+ mice with Trp53 −/− mice as previously We were surprised to discover that pRB’s function with reported (19 ). Rb1 Δ L / +; Trp53 −/− mice exhibited a signifi - E2F1 and Condensin II was sensitive to gene dosage. Knud- cant decrease in survival compared with Trp53 − /− controls son’s hypothesis, that both copies of RB1 are lost during tum- (Fig. 7A ), which was not statistically different from previ- origenesis ( 27 ), has greatly shaped our genetic understanding ously reported Rb1 Δ L/ΔL; Trp53 − /− mice ( 19 ). PCR amplifi ca- of cancer. In humans, hereditary retinoblastoma survivors

JULY 2014CANCER DISCOVERY | OF10

RESEARCH ARTICLE Coschi et al.

A 120 Rb1+/+; Trp53–/– 100 Rb1ΔL/+; Trp53–/– Rb1ΔL/ΔL; Trp53–/– 80 60 * 40 *

20 Percentage survival Percentage 0 0 50 100150 200 250 300 Days B Thymic lymphoma Sarcoma TE + T Tu T Tu T Tu T Tu T Tu T Tu T Tu ΔL WT

C Rb1ΔL/ΔL; Rb1ΔL/+; Trp53–/– Trp53 –/– Trp53 –/–

D E * Chrom. #: 1 5 10 15 XY Female 20 NS control * vs. male 15 control

Rb1ΔL/+;

Trp53–/– losses and gains

ratio 5 vs. 2

control Log Number of whole-chromosome 0 –/– –/– –/– Control ; Trp53 ; Trp53 ; Trp53 Δ L +/+ Δ L/+ Δ L/ Rb1 Rb1 Rb1

Figure 7. Haploinsufﬁ ciency of Rb1 contributes to aneuploidy. A, Kaplan–Meier survival proportions are shown for Rb1 Δ L/+; Trp53 −/− mice (n = 52). Pre- viously reported data for Rb1 Δ L/ΔL; Trp53 − /− and Trp53 −/− mice is included as a comparison. Asterisks indicate populations that are signiﬁ cantly different from the Trp53− /− control (log-rank test, P < 0.05). B, Rb1 tumor genotypes (Tu) were determined by PCR; matched tail DNA (T) is shown as a comparison. C, representative images of H&E-stained tumors from Trp53 − /− control, Rb1Δ L/ΔL; Trp53 − /−, and Rb1 ΔL/+; Trp53 −/− compound mutants. Scale bars, 100 μm.

D, control, or tumor DNA, was used for array comparative genomic hybridization (CGH). Representative graphs show log 2 ratio values plotted against chromosome number. Individual chromosomes are shown in different colors. E, whole-chromosome changes (among autosomes) for Rb1 ΔL/+; Trp53 −/− tumors is plotted against their genotype. Previously reported control, Trp53− /−, and Rb1 ΔL/ΔL; Trp53 −/− data are shown for comparison. The control male versus control male hybridization is shown in blue; the male versus female hybridizations are shown in yellow. The means were compared between genotypes using a t test, *, P < 0.05. NS, not signiﬁ cant.

OF11 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

acquire early-onset second primary cancers more frequently Western Blotting and Analysis of pRB, E2F1, than the general population ( 26 ). Thus, it is interesting to and Condensin II speculate that the underlying genomic instability we observe Nuclear extracts from the indicated cells were prepared as described + − in RB1 / patient fi broblasts could contribute to this increased previously (42 ). For isolation of pRB–E2F1–Condensin II complexes, penetrance. The secondary tumors in retinoblastoma survi- a major satellite repeat probe was generated by PCR using a bioti- vors are often sarcomas (26 , 49 ), which is consistent with the nylated primer. Extracts were mixed with the probe and complexes genome instability characteristics of mesenchymal cancers were precipitated using streptavidin dynabeads. Western blotting highlighted in this study. identifi ed captured proteins. GST pull-downs were carried out by Our data bring together concepts of replication and mito- standard methods. Antibodies used in this study can be found in the Supplementary Methods. sis and demonstrate how abundance of a pRB–E2F1–Con- +/− densin II complex may link them. RB1 patient fi broblasts Tumor Incidence and Genomic Analyses are haploinsuffi cient in their ability to prevent the accumula- The Rb1 Δ L-mutant strain (Rb1 tm1Fad) has been previously described tion of γH2AX and mitotic errors, suggesting that RB1 hap- (33 ). The Trp53 − /− mice were obtained from The Jackson Labora- loinsuffi ciency in this functional context may contribute to tory. All animals were housed, handled, and analyzed as previously tumorigenesis in humans. described (19 ). Tissues were processed for staining according to standard methods. Array comparative genomic hybridization (CGH) data were processed as previously reported (19 ), and are available in METHODS GEO (GSE51876). A list of cell line data analyzed from COSMIC is Cell Lines and Culture Methods present in the Supplementary Methods. Primary MEFs were prepared and cultured according to standard Δ Δ Δ Δ Disclosure of Potential Confl icts of Interest methods. Rb1 G/ G and Rb1 S/ S introduce R461E/K542E and F832A coding substitutions into the murine Rb1 gene, respectively ( 41, 42 ). No potential confl icts of interest were disclosed. The generation of Rb1 Δ S/ΔS mice will be published elsewhere. Mitotic chromosome spreads were prepared and H2BGFP transduction of Authors’ Contributions MEFs was as before (19 ). For DNA combing, cells were pulse-labeled Conception and design: C.H. Coschi, F.A. Dick for 30 minutes with CIdUrd, followed by 30 minutes of IdUrd as Development of methodology: C.H. Coschi, C.A. Ishak, A. Marshall, described previously ( 50 ). S. Talluri, J. Wang, V. Percy, I. Welch, P.C. Boutros, F.A. Dick Primary patient fi broblasts were obtained from the Coriell Insti- Acquisition of data (provided animals, acquired and managed tute for Medical Research (Camden, NJ) and cultured as recom- patients, provided facilities, etc.): C.A. Ishak, D. Gallo, S. Talluri, mended. RB1 genotyping was performed by Impact Genetics Inc. We M.J. Cecchini, A.L. Martens, V. Percy, I. Welch created a FUtdTW lentiviral vector that expresses H2BGFP for analy- Analysis and interpretation of data (e.g., statistical analysis, biosta- sis of human fi broblasts. All cell lines and their culture conditions are tistics, computational analysis): C.H. Coschi, C.A. Ishak, D. Gallo, described in the Supplementary Methods. A. Marshall, S. Talluri, J. Wang, I. Welch, P.C. Boutros, G.W. Brown, F.A. Dick Staining and Microscopy Writing, review, and/or revision of the manuscript: C.H. Coschi, Stained cells were examined on an Olympus Fluoview FV1000 C.A. Ishak, P.C. Boutros, G.W. Brown, F.A. Dick confocal microscope system, and z stacks at intervals of 7 μm were Administrative, technical, or material support (i.e., report- collected using the Olympus Fluoview FV1000 Viewer. Collapsed and ing or organizing data, constructing databases): C.A. Ishak, 3D rendered images were used to determine whether colocalization A. Marshall of γH2AX foci coincided with pericentromeric DNA. Study supervision: P.C. Boutros, F.A. Dick Live-cell microscopy was carried out as described previously (19 ). FISH analysis of aneuploidy and telomeres was carried out using pre- Acknowledgments viously reported methods ( 33 ). CIdUrd and IdUrd staining of DNA The authors thank Nathalie Bérubé and Kristin Kernohan for fi bers was carried out as described previously ( 50 ) with modifi cations discussions and advice. described in the Supplementary Methods. Fork velocity was calcu- lated as IdUrd length/time. Percentage asymmetry is [(long IdUrd Grant Support track − short IdUrd track) − 1] × 100. C.H. Coschi was a member of the Cancer Research and Technology Transfer (CaRTT) training program and was a Canadian Institutes ChIP and Analysis of Health Research (CIHR) doctoral award recipient. D. Gallo was Cycling cells were fi xed in 1% formaldehyde, except ChIP experi- supported by an NSERC (PGS-D) award. A. Marshall holds a Natural ments to detect PCNA, MCM3, and DNA Polδ that were fi xed in Sciences and Engineering Research Council (NSERC) CGS-M fellow- 1% formaldehyde and ethylene glycol bis[succinimidylsuccinate] ship. C.A. Ishak, S. Talluri, and M.J. Cecchini are also members of (EGS). ChIP methods and the source of PCR primers are described CaRTT, and M.J. Cecchini was supported by a CIHR MD/PHD award. in the Supplementary Methods. For ChIP-seq experiments, γH2AX F.A. Dick is the Wolfe Senior Fellow in Tumor Suppressor Genes. This and CAP-D3 antibodies were used to precipitate chromatin, and study was conducted with support from the Ontario Institute for 150 ng of DNA was used for library preparation and sequenced Cancer Research to P.C. Boutros through funding from the Govern- on an Illumina Hi-Seq 2000 at The Centre for Applied Genom- ment of Ontario and the CIHR to G.W. Brown (MOP79368) and F.A. ics (Sick Kids, Toronto, ON, Canada). Unique sequence reads Dick (MOP64253 and MOP89765). were aligned to the mm9 genome assembly or repeat containing The costs of publication of this article were defrayed in part by indexes and analyzed as described in the Supplementary Methods. the payment of page charges. This article must therefore be hereby Raw sequence data are available in the Gene Expression Omnibus marked advertisement in accordance with 18 U.S.C. Section 1734 (GEO; GSE55041). solely to indicate this fact.

JULY 2014CANCER DISCOVERY | OF12

RESEARCH ARTICLE Coschi et al.

Received February 27, 2014; revised April 8, 2014; accepted April 23. Manning AL , Dyson NJ . RB: mitotic implications of a tumour sup- 12, 2014; published OnlineFirst April 16, 2014. pressor. Nat Rev Cancer 2012 ; 12 : 220 – 6 . 24. Nicolay BN , Gameiro PA , Tschop K , Korenjak M , Heilmann AM , Asara JM , et al. Loss of RBF1 changes glutamine catabolism. Genes Dev 2013 ; 27 : 182 – 96 . REFERENCES 25. Reynolds MR , Lane AN , Robertson B , Kemp S , Liu Y , Hill BG , et al. 1. Harper JW , Elledge SJ . The DNA damage response: ten years after. Control of glutamine metabolism by the tumor suppressor Rb. Onco- Mol Cell 2007 ; 28 : 739 – 45 . gene 2014 ; 33 : 556 – 66 . 2. Jackson SP , Bartek J . The DNA-damage response in human biology 26. Abramson DH , Ellsworth RM , Zimmerman LE . Nonocular cancer in and disease. Nature 2009 ; 461 : 1071 – 8 . retinoblastoma survivors. Trans Sect Ophthalmol Am Acad Ophthal- 3. Torres-Rosell J , De Piccoli G , Cordon-Preciado V , Farmer S , Jarmuz mol Otolaryngol 1976 ; 81 : 454 – 7 . A , Machin F , et al. Anaphase onset before complete DNA replication 27. Knudson AGJ . Mutation and cancer: statistical study of retinoblasto- with intact checkpoint responses. Science 2007 ; 315 : 1411 – 5 . mas. Proc Natl Acad Sci U S A 1971 ; 68 : 820 – 3 . 4. Lukas C , Savic V , Bekker-Jensen S , Doil C , Neumann B , Pedersen R S, 28. Jacks T , Fazeli A , Schmitt EM , Bronson RT , Goodell MA , Weinbe r g R A. et al. 53BP1 nuclear bodies form around DNA lesions generated by Effects of an Rb mutation in the mouse . Nature 1992 ; 359 : 295 – 300 . mitotic transmission of chromosomes under replication stress. Nat 29. Williams BO , Remington L , Albert DM , Mukai S , Bronson RT , Jacks Cell Biol 2011 ; 13 : 243 – 53 . T . Cooperative tumorigenic effects of germline mutations in Rb and 5. Aguilera A , Gomez-Gonzalez B . Genome instability: a mechanistic p53. Nat Genet 1994 ; 7 : 480 – 4 . view of its causes and consequences. Nat Rev Genet 2008 ; 9 : 204 – 17 . 30. Zheng L , Flesken-Nikitin A , Chen PL , Lee WH . Deﬁ ciency of retino- 6. Burrell RA , McClelland SE , Endesfelder D , Groth P , Weller MC , blastoma gene in mouse embryonic stem cells leads to genetic insta- Shaikh N , et al. Replication stress links structural and numerical bility. Cancer Res 2002 ; 62 : 2498 – 502 . cancer chromosomal instability. Nature 2013 ; 494 : 492 – 6 . 31. Gonzalez-Vasconcellos I , Anastasov N , Sanli-Bonazzi B , Klymenko 7. Dulev S , de Renty C , Mehta R , Minkov I , Schwob E , Strunnikov A . O , Atkinson MJ , Rosemann M . Rb1 haploinsufﬁ ciency promotes Essential global role of CDC14 in DNA synthesis revealed by chro- telomere attrition and radiation-induced genomic instability . Cancer mosome underreplication unrecognized by checkpoints in cdc14 Res 2013 ; 73 : 4247 – 55 . mutants. Proc Natl Acad Sci U S A 2009 ; 106 : 14466 – 71 . 32. Pickering MT , Kowalik TF . Rb inactivation leads to E2F1-medi- 8. Chan KL , Palmai-Pallag T , Ying S , Hickson ID . Replication stress ated DNA double-strand break accumulation. Oncogene 2006 ; 25 : induces sister-chromatid bridging at fragile site loci in mitosis. Nat 746 – 55 . Cell Biol 2009 ; 11 : 753 – 60 . 33. Isaac CE , Francis SM , Martens AL , Julian LM , Seifried LA , Erdmann 9. Clemente-Blanco A , Mayan-Santos M , Schneider DA , Machin F , N , et al. The retinoblastoma protein regulates pericentric heterochro- Jarmuz A , Tschochner H , et al. Cdc14 inhibits transcription by RNA matin. Mol Cell Biol 2006 ; 26 : 3659 – 71 . polymerase I during anaphase. Nature 2009 ; 458 : 219 – 22 . 34. Talluri S , Isaac CE , Ahmad M , Henley SA , Francis SM , Martens A L,

10. Ide S , Miyazaki T , Maki H , Kobayashi T . Abundance of ribosomal RNA et al. A G 1 checkpoint mediated by the retinoblastoma protein that is gene copies maintains genome integrity. Science 2010 ; 327 : 693 – 6 . dispensable in terminal differentiation but essential for senescence . 11. Losada A , Hirano T . Dynamic molecular linkers of the genome: the Mol Cell Biol 2010 ; 30 : 948 – 60 . fi rst decade of SMC proteins. Genes Dev 2005 ; 19 : 1269 – 87 . 35. Kim JA , Kruhlak M , Dotiwala F , Nussenzweig A , Haber JE . Hetero- 12. Wang BD , Eyre D , Basrai M , Lichten M , Strunnikov A . Condensin chromatin is refractory to gamma-H2AX modifi cation in yeast and binding at distinct and specifi c chromosomal sites in the Saccharomy- mammals. J Cell Biol 2007 ; 178 : 209 – 18 . ces cerevisiae genome. Mol Cell Biol 2005 ; 25 : 7216 – 25 . 36. Dominguez-Sanchez MS , Barroso S , Gomez-Gonzalez B , Luna R , 13. Ono T , Yamashita D , Hirano T . Condensin II initiates sister chroma- Aguilera A . Genome instability and transcription elongation impair- tid resolution during S phase. J Cell Biol 2013 ; 200 : 429 – 41 . ment in human cells depleted of THO/TREX. PLoS Genet 2011 ; 14. Bertoli C , Skotheim JM , de Bruin RA . Control of cell cycle transcrip- 7: e1002386 . tion during G1 and S phases. Nat Rev Mol Cell Biol 2013 ; 14 : 518 – 28 . 37. Lossaint G , Larroque M , Ribeyre C , Bec N , Larroque C , Decaill et 15. Knudsen KE , Booth D , Naderi S , Sever-Chroneos Z , Fribourg AF , C , et al. FANCD2 binds MCM proteins and controls replisome Hunton IC , et al. RB-dependent S-phase response to DNA damage. function upon activation of s phase checkpoint signaling. Mol Cell Mol Cell Biol 2000 ; 20 : 7751 – 63 . 2013 ; 51 : 678 – 90 . 16. Avni D , Yang H , Martelli F , Hofmann F , ElShamy WM , Ganesan S , 38. Samoshkin A , Dulev S , Loukinov D , Rosenfeld JA , Strunnikov AV . et al. Active localization of the retinoblastoma protein in chromatin Condensin dysfunction in human cells induces nonrandom chromo- and its response to S phase DNA damage. Mol Cell 2003 ; 12 : 735 – 46 . somal breaks in anaphase, with distinct patterns for both unique and 17. Wells J , Yan PS , Cechvala M , Huang T , Farnham PJ . Identifi cation of repeated genomic regions. Chromosoma 2012 ; 121 : 191 – 9 . novel pRb binding sites using CpG microarrays suggests that E2F 39. Chong JL , Wenzel PL , Saenz-Robles MT , Nair V , Ferrey A , Hagan J P, recruits pRb to specifi c genomic sites during S phase. Oncogene et al. E2f1-3 switch from activators in progenitor cells to repressors in 2003 ; 22 : 1445 – 60 . differentiating cells. Nature 2009 ; 462 : 930 – 4 . 18. Mendoza-Maldonado R , Paolinelli R , Galbiati L , Giadrossi S , Giacca 40. Dick FA , Dyson N . pRB contains an E2F1-specifi c binding domain M . Interaction of the retinoblastoma protein with Orc1 and its recruit- that allows E2F1-induced apoptosis to be regulated separately from ment to human origins of DNA replication . PLoS ONE 2010 ; 5 : e13720 . other E2F activities. Mol Cell 2003 ; 12 : 639 – 49 . 19. Coschi CH , Martens AL , Ritchie K , Francis SM , Chakrabarti S , Berube 41. Cecchini MJ , Thwaites M , Talluri S , Macdonald JI , Passos DT , Chong NG , et al. Mitotic chromosome condensation mediated by the retino- JL , et al. A retinoblastoma allele that is mutated at its common E2F blastoma protein is tumor-suppressive . Genes Dev 2010 ; 24 : 1351 – 63 . interaction site inhibits cell proliferation in gene targeted mice . Mol 20. Longworth MS , Herr A , Ji JY , Dyson NJ . RBF1 promotes chromatin Cell Biol 2014 ; 34 : 2029 – 45 . condensation through a conserved interaction with the Condensin II 42. Cecchini MJ , Dick FA . The biochemical basis of CDK phosphoryla- protein dCAP-D3. Genes Dev 2008 ; 22 : 1011 – 24 . tion-independent regulation of E2F1 by the retinoblastoma protein. 21. Manning AL , Longworth MS , Dyson NJ . Loss of pRB causes cen- Biochem J 2011 ; 434 : 297 – 308 . tromere dysfunction and chromosomal instability. Genes Dev 43. Julian LM , Palander O , Seifried LA , Foster JE , Dick FA . Charact eriza- 2010 ; 24 : 1364 – 76 . tion of an E2F1-specifi c binding domain in pRB and its implications 22. van Harn T , Foijer F , van Vugt M , Banerjee R , Yang F , Oostra A , for apoptotic regulation. Oncogene 2008 ; 27 : 1572 – 9 . et al. Loss of Rb proteins causes genomic instability in the absence of 44. Forbes SA , Tang G , Bindal N , Bamford S , Dawson E , Cole C, et al. mitogenic signaling. Genes Dev 2010 ; 24 : 1377 – 88 . COSMIC (the Catalogue of Somatic Mutations in Cancer): a resource

OF13 | CANCER DISCOVERYJULY 2014 www.aacrjournals.org

RB Dosage in DNA Replication and Aneuploidy RESEARCH ARTICLE

to investigate acquired mutations in human cancer . Nucleic Acids 48. Manning AL , Yazinski SA , Nicolay B , Bryll A , Zou L , Dyson NJ. Sup- Res 2010 ; 38 : D652 – 7 . pression of genome instability in pRB-deﬁ cient cells by enhancement 45. Gordon DJ , Resio B , Pellman D . Causes and consequences of aneu- of chromosome cohesion. Mol Cell 2014 ; 53 : 993 – 1004 . ploidy in cancer. Nat Rev Genet 2012 ; 13 : 189 – 203 . 49. Friend SH , Horowitz JM , Gerber MR , Wang XF , Bogenmann E , Li F P, 46. Beroukhim R , Mermel CH , Porter D , Wei G , Raychaudhuri S , Dono - et al. Deletions of a DNA sequence in retinoblastomas and mesenchy- van J , et al. The landscape of somatic copy-number alteration across mal tumors: organization of the sequence and its encoded protein . human cancers. Nature 2010 ; 463 : 899 – 905 . Proc Natl Acad Sci U S A 1987 ; 85 : 2234 – 8 . 47. Crasta K , Ganem NJ , Dagher R , Lantermann AB , Ivanova EV , Pan 50. Yang J , O’Donnell L , Durocher D , Brown GW . RMI1 promotes DNA Y , et al. DNA breaks and chromosome pulverization from errors in replication fork progression and recovery from replication fork stress. mitosis. Nature 2012 ; 482 : 53 – 8 . Mol Cell Biol 2012 ; 32 : 3054 – 64 .

JULY 2014CANCER DISCOVERY | OF14

Haploinsufficiency of an RB−E2F1−Condensin II Complex Leads to Aberrant Replication and Aneuploidy

Courtney H. Coschi, Charles A. Ishak, David Gallo, et al.

Cancer Discovery Published OnlineFirst April 16, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/2159-8290.CD-14-0215

Supplementary Access the most recent supplemental material at: Material http://cancerdiscovery.aacrjournals.org/content/suppl/2014/04/16/2159-8290.CD-14-0215.DC1 http://cancerdiscovery.aacrjournals.org/content/suppl/2021/03/04/2159-8290.CD-14-0215.DC2

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://cancerdiscovery.aacrjournals.org/content/early/2014/06/10/2159-8290.CD-14-0215. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.