Next-Gen Sequencing Exposes Frequent MED12 Mutations and Actionable Therapeutic Targets in Phyllodes Tumors Andi K

Published OnlineFirst January 15, 2015; DOI: 10.1158/1541-7786.MCR-14-0578

Molecular Cancer Research

Next-Gen Sequencing Exposes Frequent MED12 Mutations and Actionable Therapeutic Targets in Phyllodes Tumors Andi K. Cani1, Daniel H. Hovelson2, Andrew S. McDaniel1, Seth Sadis3, Michaela J. Haller1, Venkata Yadati1, Anmol M. Amin1, Jarred Bratley1, Santhoshi Bandla3, Paul D. Williams3, Kate Rhodes4, Chia-Jen Liu1, Michael J. Quist1,5, Daniel R. Rhodes3, Catherine S. Grasso1,5, Celina G. Kleer1, and Scott A. Tomlins1,6,7

Abstract

Phyllodes tumors are rare fibroepithelial tumors with variable gene amplifications in IGF1R and EGFR. Taken together, this clinical behavior accounting for a small subset of all breast study defines the genomic landscape underlying phyllodes tumor neoplasms, yet little is known about the genetic alterations that development, suggests potential molecular correlates to histolog- drive tumor initiation and/or progression. Here, targeted next- ic grade, expands the spectrum of human tumors with frequent generation sequencing (NGS) was used to identify somatic altera- recurrent MED12 mutations, and identifies IGF1R and EGFR as tions in formalin-fixed paraffin-embedded (FFPE) patient speci- potential therapeutic targets in malignant cases. mens from malignant, borderline, and benign cases. NGS revealed mutations in mediator complex subunit 12 (MED12) affect- Implications: Integrated genomic sequencing and mutational ing the G44 hotspot residue in the majority (67%) of cases profiling provides insight into the molecular origin of phyllodes spanning all three histologic grades. In addition, loss-of-function tumors and indicates potential druggable targets in malignant mutations in p53 (TP53) as well as deleterious mutations in the disease. tumor suppressors retinoblastoma (RB1) and neurofibromin 1 (NF1) were identified exclusively in malignant tumors. High-level copy- Visual Overview: http://mcr.aacrjournals.org/content/early/2015/ number alterations (CNA) were nearly exclusively confined to 04/02/1541-7786.MCR-14-0578/F1.large.jpg. malignant tumors, including potentially clinically actionable Mol Cancer Res; 13(4); 613–9. 2015 AACR.

Introduction including cellular atypia, mitotic activity, stromal overgrowth, stromal cellularity, and tumor margins (1). However, this histo- Phyllodes tumors of the breast are relatively rare fibroepithelial pathologic classification often fails to predict which phyllodes tumors that account for approximately 1% of all breast neo- tumors will recur or metastasize after treatment and does not plasms. Like benign breast fibroadenomas, they are characterized accurately inform on treatment options. Although local recur- by proliferation of both stromal and epithelial components, but, rence after resection is most prevalent in histologically malignant in contrast, they have considerable malignant potential. Phyl- cases (approximately 30%, depending on width of excised mar- lodes tumors are classified as benign (65%), borderline gins), borderline, and benign tumors can also recur locally in (25%), and malignant (10%) based on histologic features, about 15% and 10% of cases, respectively, demonstrating the limitations of current prognostic approaches (2). Likewise, although approximately 10% of all phyllodes tumors progress 1Department of Pathology, Michigan Center for Translational Pathol- ogy, Ann Arbor, Michigan. 2Department of Computational Medicine to distant metastases, only approximately 20% of histologically and Bioinformatics University of Michigan Medical School, Ann Arbor, malignant cases do so (3, 4), leaving a substantial number of Michigan. 3Life Sciences Solutions, ThermoFisher Scientific, Ann borderline and even histologically benign cases that have meta- 4 fi Arbor, Michigan. Life Sciences Solutions, ThermoFisher Scienti c, static potential. Conversely, although most histologically benign Carlsbad, California. 5Department of Pathology, Oregon Health and Sciences University, Portland, Oregon. 6Department of Urology, Uni- cases will behave as such, there are a proportion of phyllodes versity of Michigan Medical School, Ann Arbor, Michigan. 7Compre- tumors classified as malignant and borderline that will behave in a hensive Cancer Center, University of Michigan Medical School, Ann benign manner. Current treatment guidelines for phyllodes Arbor, Michigan. tumors require wide surgical resection margins, but efficacious Note: Supplementary data for this article are available at Molecular Cancer treatment options for the 10% of all phyllodes tumors that Research Online (http://mcr.aacrjournals.org/). progress to metastatic disease are lacking and survival rates are Corresponding Author: Scott A. Tomlins, University of Michigan Medical School, dismal (3). 1524 BSRB, 109 Zina Pitcher Place, Ann Arbor, MI 48109. Phone: 734-764-1549; The key genetic alterations driving phyllodes tumor deve- Fax: 734-647-7950; E-mail: [email protected] lopment and molecular correlates with histologic grade and doi: 10.1158/1541-7786.MCR-14-0578 malignant behavior are poorly characterized. Comparative 2015 American Association for Cancer Research. genomic hybridization (CGH) and array CGH (aCGH) studies

www.aacrjournals.org 613

Cani et al.

have shown multiple recurrent, broad somatic chromosomal somatic variant settings. Variants were annotated using Annovar copy-number alterations (CNA) in phyllodes tumors, including (13). Called variants were filtered to remove synonymous or gains of chromosome 1q and losses in 13q, 6q, 9p; however, noncoding variants, those with flow corrected read depths their prognostic utility is unclear (5–9). Several genes have (FDP) less than 20, flow corrected variant allele containing been implicated in phyllodes tumor development by virtue of reads (FAO) less than 6, variant allele frequencies (FAO/FDP) being localized to areas of CNA, including EGFR,whichwas less than 0.10, extreme skewing of forward/reverse flow cor- recently shown by FISH to be amplified in 2% to 16% of cases rected reads calling the variant (FSAF/FSAR <0.2 or >5, or FSAF (10, 11). In addition, gene expression and IHC studies have or FSAR <3), or indels within homopolymer runs >4. Variants implicated various signaling pathways, including insulin-like occurring exclusively in reads containing other variants (single- growth factor (IGF) and Wnt/b-catenin, as being activated in nucleotide variants or indels) or those occurring in the last phyllodes tumors (1). To more comprehensively assess mapped base of a read were excluded. Variants with allele somatic molecular alterations in phyllodes tumors and identify frequencies >0.5% in ESP6500 or 1,000 genomes or those potential opportunities for personalized medicine, we per- reported in ESP6500 or 1,000 genomes with observed variant formed next-generation sequencing (NGS) of 15 formalin-fixed allele frequencies between 0.40 and 0.60 or >0.9 were consid- paraffin-embedded (FFPE) phyllodes tumors representing the ered germ line variants. High-confidence somatic variants pass- histologic grade spectrum. ing the above criteria were then visualized in IGV. We have previously confirmed that these filtering criteria identify variants that pass PCR validation with >95% accuracy (12). To prioritize Materials and Methods potential driving alterations, we used Oncomine software Case selection tools (powertools.oncomine.com) to annotate called variants, We identified a cohort of 15 archived, routine clinical FFPE which uses pan-cancer NGS data to identify genes as oncogenes phyllodes tumor specimens from the University of Michigan, or tumor suppressors, based on overrepresentation of hotspot Department of Pathology Tissue Archive. Clinicopathologic infor- or deleterious mutations, respectively. Variants in oncogenes mation for each case was obtained from the clinical archive. are then considered gain-of-function if at a hotspot and variants Hematoxylin and eosin (H&E)–stained slides for all cases were in tumor suppressors are considered loss-of-function if delete- reviewed by a board-certified Anatomic Pathologist (S.A. Tomlins) rious or at a hotspot (S.A. Tomlins; unpublished data). to ensure sufficient tumor content and confirm histologic grade. Copy-number analysis Targeted next-generation sequencing To identify CNAs, normalized, GC-content corrected read Targeted NGS of tumor tissue was performed with IRB approv- counts per amplicon for each sample were divided by those from al. For each specimen, 4 to 10 10-mm FFPE sections were cut a pool of normal male genomic DNA samples (FFPE and frozen from a single representative block per case, using macrodissection tissue, individual, and pooled samples), yielding a copy-number with a scalpel as needed to enrich for at least 50% tumor content ratio for each amplicon. Gene-level copy-number estimates were (as defined by areas of stromal overgrowth). DNA was isolated determined by taking the coverage-weighted mean of the per- using the Qiagen Allprep FFPE DNA/RNA Kit (Qiagen), according probe ratios, with expected error determined by the probe-to- to the manufacturer's instructions except for adding a 2 minute probe variance (12); a detailed article describing this technique is room temperature incubation and extending centrifugation time in submission (C.S. Grasso; submitted for publication). Genes to 5 minutes during the xylene deparaffinization (step 1) and with a log2 copy-number estimate of <1or>0.6 were considered ethanol washing of xylene (step 2). DNA was quantified using the to have high-level loss or gain, respectively. Qubit 2.0 fluorometer (Life Technologies). Targeted, multiplexed PCR-based NGS was performed on each Sanger sequencing to validate called somatic variants component using a custom panel comprised of 2,462 amplicons Bidirectional Sanger sequencing was performed over the targeting 130 genes and Ion Torrent–based sequencing. Genes observed MED12 mutation hotspot (G44) on all tumor samples. included in this panel were selected based on pan-cancer NGS and Genomic DNA (10 ng) was used as template in PCR amplifica- copy-number profiling data analysis that prioritized somatic, tions with Invitrogen Platinum PCR Supermix Hi-Fi (Life Tech- recurrently altered oncogenes, tumors suppressors, genes present nologies) with the suggested initial denaturation and cycling in high-level copy gains/losses, and known/investigational ther- conditions. Primer sequences were as previously reported apeutic targets. Barcoded libraries were generated from 20 ng of (14, 15) with the addition of universal M13 adaptors (M13 DNA per sample using the Ion Ampliseq library Kit 2.0 (Life forward: TGTAAAACGACGGCCAGT and M13 reverse: CAG- Technologies) according to the manufacturer's instructions with GAAACAGCTATGACC). PCR products were subjected to bidirec- barcode incorporation. Templates were prepared using the Ion tional Sanger sequencing for both primer pairs by the University PGM Template OT2 200 Kit (Life Technologies) on the Ion One of Michigan DNA Sequencing Core after treatment with ExoSAP- Touch 2 according to the manufacturer's instructions. Sequencing IT (GE Healthcare) and sequences were analyzed using SeqMan of multiplexed templates was performed using the Ion Torrent Pro software (DNASTAR). Personal Genome Machine (PGM) on Ion 318 chips using the Ion PGM Sequencing 200 Kit v2 (Life Technologies) according to the qPCR to validate copy-number variations manufacturer's instructions. EGFR, IGF1R, and CDKN2A copy-number changes were sub- Data analysis was performed essentially as described previ- jected to validation through quantitative real-time PCR (qPCR) ously (12) in Torrent Suite 4.0.2, with alignment by TMAP using for 12 samples with sufficient DNA. PH13, 14 and 30 had default parameters, and variant calling using the Torrent Variant insufficient DNA (no CNAs in these genes were identified by Caller plugin (version 4.0-r76860) using default low-stringency NGS) for qPCR and PH5 had sufficient DNA only for assessing

614 Mol Cancer Res; 13(4) April 2015 Molecular Cancer Research

Next-Generation Sequencing of Phyllodes Tumors

EGFR and CDKN2A (no IGF1R CNAs were identified by NGS). PH samples) and PH 22 (copy-number neutral by NGS) as the Primers and probes (50 FAM; ZEN/Iowa Black FQ dual quench- calibrator sample. ers) were designed using PrimerQuest (http://www.idtdna. com/Primerquest/Home/Index, hg 19 genome assembly) and Statistical analysis obtained from IDT. Assay specificity was confirmed using Comparisons of the number of mutations or CNAs per BLAST and BLAT and primers/probes in areas of SNPs were sample by tumor grade were performed using the Kruskal– excluded. Primer/probe sequences are available upon request. Wallis test with post-hoc pairwise comparison of subgroups qPCR reactions (15 mL) were performed in triplicate using using MedCalc 13.1.2.0. Comparison of the frequency of TaqMan Genotyping Master Mix (Applied Biosystems), 5 ng MED12 mutations by tumor grade was performed by the Fisher genomic DNA per reaction and a final concentration of exact test using R 3.1.0. 0.9 mmol/L each primer and 0.25 mmol/L probe in 384-well plates on the QuantStudio 12K Flex (Applied Biosystems). Results Automatic baseline and Ct thresholds were set using Quant- Studio 12K Flex Real-Time PCR System Software. Log2 copy We performed targeted NGS on a cohort of 15 FFPE phyllodes number of EGFR, CDKN2A,andIGF1R were determined by the tumors comprised of 5 cases each of benign, borderline, and DDCt method using the average Ct of DNMT3A, FBXW7,and malignant histologic grade; representative photomicrographs and MYO18A as the reference (copy-number neutral by NGS in all clinical characteristics of all patients are presented in Fig. 1A.

A PH-19 PH-08 PH-05 PH-06

Figure 1. Histology and clinicopathologic information for FFPE phyllodes tumors assessed by targeted NGS and integrative molecular heatmap of driving molecular alterations in phyllodes tumors. A, H&E-stained PH-19 PH-08 PH-05 PH-06 sections of representative benign (PH-19), borderline (PH-08), and malignant (PH-05 and PH-06) phyllodes tumors sequenced are shown. Top, 4 original magnification; bottom 20 magnification. Clinicopathologic information, including histologic grade, specimen type, patient age, procedure type, and tumor size, for all cases is given (Ex./Lump, excisional biopsy or lumpectomy; Mast, Case Grade Type Age Procedure Size (cm) mastectomy; Core bx, Core biopsy). PH-18 Benign Primary 47 Ex./Lump. 1.1 B B, targeted NGS of 15 FFPE phyllodes PH-19 Benign Primary 32 Ex./Lump. 1.5 PH-17 PH-13 PH-03 PH-05 PH-16 PH-14 PH-20 Benign Primary 42 Ex./Lump. 3.0 PH-30 PH-18 PH-19 PH-22 PH-20 PH-11 PH-04 PH-08 PH-06 tumors was performed to identify PH-22 Benign Primary 24 Ex./Lump. 15.0 Grade potentially driving/actionable PH-30 Benign Primary 33 Ex./Lump. 3.2 Type molecular alterations. All high- PH-04 Borderline Primary 37 Ex./Lump. 4.4 Age confidence, gain- or loss-of-function PH-08 Borderline Recurrence 21 Ex./Lump. 3.4 PH-11 Borderline Primary 30 Ex./Lump. 4.0 Procedure somatic mutations in oncogenes and PH-13 Borderline Primary 35 Ex./Lump. 3.0 tumor suppressors, in addition to PH-17 Borderline Primary 13 Ex./Lump. 9.1 MED12 high-level CNA are indicated for each PH-03 Malignant Primary 30 Ex./Lump. 8.1 TP53 fi PH-05 Malignant Recurrence 67 Mast. 7.0 case. Speci c alteration types are PH-06 Malignant Primary 26 Ex./Lump. 1.4 TERT indicated according to the legend PH-14 Malignant Lung Met 60 Core Bx. N/A IGF1R (Nonsyn SNV, nonsynonymous SNV; PH-16 Malignant Primary 39 Mast. 10.0 MCL1 Fs and Fp indel, frame-shifting and EGFR frame-preserving indels, CCNE1 respectively). Slashed boxes indicate MYC two alterations. Clinicopathologic information is shown above the ZNF217 heatmap according to the legend and NF1 as in A. Grade Type Alteration CDKN2A Benign Primary Copy gain Borderline Recurrence Copy loss PDGFRA Malignant Metastasis Nonsyn. SNV CCND1 Fs. indel Patient age Procedure Fp. indel CD44 ≤30 y Core biopsy Splice site 31–49 y Exc. / Lump. Nonsense RB1 ≥50 y Mastectomy BCL9

www.aacrjournals.org Mol Cancer Res; 13(4) April 2015 615

Cani et al.

We isolated an average of 0.65 mg DNA per case from 4 to 10 Plugin in Ion Reporter for assessing gain- or loss-of-function. 10-mm sections using macrodissection to enrich tumor content This analysis identified five loss-of-function alterations, including as needed. NGS was performed using a multiplexed PCR-based three mutations in TP53 (F270L in PH-14, Q192X in PH-16, custom Ion Torrent Ampliseq panel comprised of 2,462 ampli- and C242Y in PH3), and one mutation each in RB1 (E533X in cons targeting 130 genes and Ion Torrent–based sequencing on PH16) and NF1 (p.1152_1153del in PH-05), as shown in the the PGM. Targeted genes were selected on the basis of pan- integrative heatmap of driving alterations (Fig. 1B and Supple- cancer NGS and copy-number profiling data analysis to prior- mentary Tables S2 and S3). Intriguingly, these loss-of-function itize somatic, recurrently altered oncogenes, tumors suppres- alterations occurred exclusively in malignant tumors. sors and genes present in high-level CNAs. Detailed character- Copy-number analysis of NGS data demonstrated recurrent ization of this panel will be reported separately (S.A. Tomlins; low-level CNAs, including gain of chromosome 1q and loss of unpublished data). chromosome 13q, consistent with previous reports (5–9). These NGS of multiplexed templates on the Ion Torrent PGM gen- were more prevalent in malignant tumors (5/5), but were also erated an average of 1,011,571 mapped reads yielding 409 present in two borderline and one benign case (Fig. 3A and targeted base coverage across the 15 samples (Supplementary Supplementary Fig. S2). High-level CNAs were nearly exclusively Table S1). We identified a total of 26 high-confidence, likely present in malignant specimens 14 of 16 alterations, as shown somatic nonsynonymous or splice site altering point mutations in Fig. 3A and B. Of note, PH-03 showed high-level EGFR (copy- and short insertion/deletions (indels) across the 15 samples number ratio > 6) and IGF1R amplifications, whereas PH-16 (median 2; range, 0–4) as shown in Supplementary Tables S2 also showed a high-level IGF1R amplification (copy-number ratio and S3. The number of high-confidence somatic nonsynony- > 32, Fig. 3B). TERT amplifications were also observed in three mous mutations were not significantly different between the malignant tumors, whereas PH-05 harbored a focal high-level histologic grades (Kruskal–Wallis test, P ¼ 0.09), as shown in CDKN2A (p16INK4A) loss. We confirmed EGFR, IGF1R, and Supplementary Fig. S1A. Copy-number analysis of NGS data CDKN2A CNAs by qPCR as shown in Fig. 3C. yielded a total of 16 high-level CNAs (median 0; range, 0–6). The number of high level CNAs differed significantly between histologic grades (Kruskal-Wallis test, P ¼ 0.002), with malignant tumors harboring significantly more high-level CNAs per Discussion sample (median 2; range, 2–6) than borderline (median 0; range, We performed targeted NGS of 15 FFPE phyllodes tumors 0–0) or benign (median 0; range, 0–2) tumors (Kruskal–Wallis representing all three histologic grades to identify somatic test, post-hoc analysis, both P < 0.05), as shown in Supplementary alterations associated with tumor development and potential Fig. S1B. Prioritized likely gain- or loss-of-function somatic targetable alterations. Mutations in MED12 were present in 10 mutations in oncogenes and tumor suppressors (see below) of 15 cases (67%) and affected the known exon 2 G44 residue and high-level CNAs for each case are shown in an integrative hotspot through multiple mechanisms. No significant difference heatmap (Fig. 1B). in MED12 mutation frequency was observed across histologic By NGS, we found that MED12, which encodes subunit 12 of grades, although our cohort size is limited. Our IRB approved the Mediator complex (the multiprotein assembly that serves as a protocol does not allow NGS of matched normal tissue; how- general coactivator of transcription by RNA polymerase II) was ever, the observed MED12 variant allele frequencies are consis- mutated in 10 of 15 samples (67%; one sample with biallelic tent with somatic events as seen in other tumors. Similar MED12 mutations) by automated variant calling and visual read inspec- somatic mutations are frequent (50%–70%) in benign uterine tion in IGV (as some called variants were filtered due to skewed leiomyomas (14–16), but less common in malignant uterine read support). All mutations were localized to the exon 2 hotspot leiomyosarcomas (7%–30%; refs. 15, 16, 18, 19). Recently, Lim region near residue G44 (Fig. 2A and Supplementary Tables S2 and colleagues identified similar MED12 mutations in 59% of and S3), which has recently been reported to be recurrently benign breast fibroadenomas through exome sequencing (17). mutated at high frequency in uterine leiomyomas (14–16) and Given the morphologic similarity of breast fibroadenomas and benign breast fibroadenomas (17), and more rarely in uterine benign phyllodes tumors, frequent MED12 mutations in both leiomyosarcomas (15, 16, 18, 19). Five of 11 total MED12 entities support a closely related molecular origin. In addition, mutations were point mutations at G44 (3 p.G44S, 1 p.G44C, although our findings will need to be replicated in larger and 1 p.G44R) whereas three mutations were frame-preserving cohorts, the similar frequency of MED12 mutations across the deletions adjacent to or including G44 (p.38_43, p.41_49, and histologic spectrum of phyllodes tumors (in addition to benign p.42_51). Two mutations were intronic point mutations just fibroadenomas) suggests that MED12 mutations in the breast upstream of exon 2, at a previously reported splice acceptor site are early events that may be unrelated to malignant behavior, in causing retention of an additional 6 bases in the transcript (c.IVS-8 contrast with uterine leiomyomas and leiomyosarcomas, which p.E33_D34insPQ; refs. 14, 17). PH-11, which harbored a c.IVS-8 show notable differences in MED12 mutation frequencies as mutation, also harbored an additional intronic mutation further just described. Our results also support the evolution of malig- upstream (c.IVS-15), consistent with biallelic intronic MED12 nant phyllodes tumors from less aggressive fibroadenomas or mutations in this sample. There was no significant difference in phyllodes tumors (possibly through loss of key tumor suppres- the presence of MED12 mutations between tumors of different sors). Although MED12 hotspot mutations have been identified histologic grade (benign 4/5, borderline 4/5, malignant 2/5, infrequently in extrauterine or extramammary tumors (18, 20), Fisher exact test, P ¼ 0.5). All MED12 mutations were confirmed functional studies support a role for MED12 mutations affect- by bidirectional Sanger sequencing (Fig. 2B). ing the G44 hotspot in dysregulation of estrogen signaling in To prioritize potential driving alterations from the remaining estrogen responsive cells (17), and the Mediator complex is non-MED12 point mutations/indels, we used the Oncomine known to interact with the estrogen receptor (21).

616 Mol Cancer Res; 13(4) April 2015 Molecular Cancer Research

Next-Generation Sequencing of Phyllodes Tumors

A MED12 Alteration p. G44S (3) Nonsyn. SNV p. G44C Fp. indel c.IVS-8 (2) p. G44R Splice site c.IVS-15

Exon 2 Intron 1 p. 38_43delInsE p. 41_49delinsA p. 42_51delinsI

B PH-18 c. 130G>A; p. G44S PH-03 c. 113_127del15; p. 38_43delInsE Figure 2. Identiﬁcation of recurrent MED12 mutations in phyllodes tumors. NGS and Sanger sequencing identiﬁed 11 MED12 mutations in 10 of 15 phyllodes tumors subjected to NGS. A, schematic representation of MED12 intron 1 and exon 2 junction with locations of all PH-19 c. 130G>A; p. G44S PH-04 c. 122_145del24; p. 41_49delinsA observed mutations shown. Mutation type is indicated in the legend and the frequency of observed mutations is indicated in parentheses. B, bidirectional Sanger sequencing was performed on all specimens. Traces of MED12 cases with mutations are shown PH-30 c. 124_151del27; p. 42_51delinsI (only one trace direction shown) with PH-22 c130G>C; p. G44R the indicated nucleotide and amino acid changes noted. The mutation(s) position is indicated by the arrow.

PH-13 c. 130G>T; p. G44C PH-17 c. IVS-8; p. E33_D34insPQ

PH-11c. IVS-8, cIVS-15 ; p. E33_D34insPQ? PH-05 c. 130G>A; p. G44S

Although surgical resection of phyllodes tumors may be malignant cases). EGFR has been shown to be highly amplified curative, local recurrence is not uncommon and distant metas- in phyllodes tumors by FISH in up to 16% of cases (10, 11), tasis is associated with poor survival. Furthermore, the histo- consistent with our findings. Dysregulation of the IGF pathway logic features do not accurately predict clinical behavior of has been implicated in phyllodes tumors by IHC (1); however, phyllodes tumors. Hence, targetable alterations, particularly in IGF1R amplification has not been reported. malignant phyllodes tumors, may be useful for personalized Direct comparison of additional CNAs identified in our medicine strategies. Through copy-number analysis of NGS study and previous studies is challenging due to platform data (and confirmed by qPCR), we identified potentially clin- differences. However, broad, low-level gains in genes on 1q ically actionable high-level, focal amplifications of EGFR and and losses on 13q were observed in malignant as well as IGF1R in 7% and 13% of cases, respectively (1/5 and 2/5 borderline and benign tumors, consistent with previous

www.aacrjournals.org Mol Cancer Res; 13(4) April 2015 617

Cani et al.

A TERT EGFR 5.0 CDKN2A IGF1R 2.5 Other Figure 3. 0.0 Copy-number analysis of phyllodes −2.5 copy number ratio ﬁ

2 tumor identi es potential therapeutic

Log −5.0 Chr 1 Chr X targets in malignant samples. Copy- number analysis was performed from AR IL6 KIT NF1 NF2 VHL RB1 ATM APC PNP WT1 KDR MET FLT3 MYC JAK2 TP53 ABL1 TET2 AKT1 BCL9 TSC1 BAP1 CD44 TSC2 SOX2 MCL1 CDK4 CDK6 TERT GAS6 PTEN TIAF1 CDH1 KRAS MSH2 EGFR MTOR MDM4 BIRC3 MDM2 BIRC2 IGF1R MYCN NGS data. For each sequenced GATA3 SF3B1 STK11 CD274 APEX1 FGFR1 FGFR3 PTCH1 FGFR2 ERBB2 BRCA1 MYCL1 BRCA2 PPARG CCNE1 CCND1 PIK3R1 FBXW7 PIK3CA ATP11B BCL2L1 ZNF217 NKX2−1 ACVRL1 MYO18A CDKN2A NOTCH1 DNMT3A PDGFRA RPS6KB1 CSNK2A1 DCUN1D1

SMARCB1 phyllodes tumor, GC content PDCD1LG2 corrected, normalized read counts per amplicon were divided by those from B a composite normal sample, yielding a 3.0 PH-03 IGF1R copy-number ratio for each amplicon. 1.5 EGFR Gene-level copy-number estimates 0.0 were determined by taking the

−1.5 weighted mean of the per-probe copy number ratio 2 copy-number ratios. A, summary of −3.0 Chr 1 2 3 4 5 7 8 9 10 11 12 13 14 15 16 17 19 20 22 X Log gene-level copy-number ratios (log2) 3.0 ﬁ PH-05 for all pro led samples. Selected 1.5 genes of interest with high-level CNAs

0.0 are colored according to the legend. B, copy-number proﬁles for three −1.5 copy number ratio 2 CDKN2A malignant phyllodes tumors with −3.0 Log high-level CNAs. Log2 copy-number

5.0 ratios per amplicon are plotted (with PH-16 TERT 2.5 IGF1R each individual amplicon represented by a single dot, and each individual 0.0 gene indicated by different colors), −2.5 copy number ratio with gene-level copy-number 2 −5.0

Log estimates (black bars) determined by taking the weighted mean of the per- C probe copy-number ratios. Selected 6 high-level CNAs are indicated. C, IGF1R qPCR conﬁrmation of high-level CNAs EGFR in EGFR, IGF1R,andCDKN2A. qPCR on CDKN2A 3 genomic DNA from indicated samples was performed using DNMT3A, FBXW7,andMYO18A as the reference 1 genes. Normalized mean IGF1R (blue),

copy number ratio 0 EGFR CDKN2A 2 (red), and (green) log2 -1 copy-number ratios [using PH22 (no Log CNAs by NGS) as the calibrator] from -3 Tumor grade triplicate qPCR SD are plotted. Benign Borderline Malignant PH-20 PH-16 PH-19 PH-11 PH-17 PH-04 PH-08 PH-03 PH-05 PH-06 PH-18

reports. On the other hand, we did not observe gains in MDM2 Taken together, our results demonstrate frequent MED12 muta- or MDM4, which have been reported in previous aCGH studies, tions in phyllodes tumors, supporting a shared origin with benign and were targeted herein; we hypothesize this may be due to the breast fibroadenomas. In addition, our results suggest potential high TP53 alteration rate in our phyllodes tumors with high therapeutic targets in malignant tumors, including EGFR and numbers of CNAs. Finally, our panel did not target genes in IGF1R. Finally, as both driving somatic mutations/indels other some previously reported regions of CNA (such as 6q), pre- than MED12 and high-level, focal CNAs occurred exclusively in cluding comparisons of these alterations. malignant tumors in our cohort, such alterations may be useful Besides MED12 hotspot mutations, other potential driving for classification or prognostication in borderline tumors if con- somatic point mutations/indels, which included loss-of-function firmed in other cohorts. alterations in TP53, RB1, and NF1, occurred exclusively in malignant tumors. In addition, high-level, focal CNAs (such as those in fl EGFR and IGF1R) were only observed in malignant cases. Togeth- Disclosure of Potential Con icts of Interest er, these findings support molecular correlates to histologic grade. S.A. Tomlins has a separate sponsored research agreement with Com- pendia Bioscience/Life Technologies/ThermoFisher Scientificthatprovides Whether such molecular alterations may be useful in cases with access to the sequencing panel used herein. No other aspect of the study challenging histology or show prognostic potential can be inves- described herein was supported by Compendia Bioscience/Life Technolo- tigated in additional cohorts. gies/ThermoFisher Scientific. S. Sadis, S. Bandla, P.D. Williams, K. Rhodes,

618 Mol Cancer Res; 13(4) April 2015 Molecular Cancer Research

Next-Generation Sequencing of Phyllodes Tumors

andD.R.RhodesareemployeesofThermoFisherScientiﬁc. The other Administrative, technical, or material support (i.e., reporting or authors have no competing interests to declare. organizing data, constructing databases): D.H.Hovelson,A.M.Amin, P.D. Williams, C.-J. Liu Authors' Contributions Study supervision: A.K. Cani, P.D. Williams, K. Rhodes, S.A. Tomlins Conception and design: A.K.Cani,A.S.McDaniel,K.Rhodes,D.R.Rhodes, S.A. Tomlins Acknowledgments Development of methodology: A.K.Cani,S.Sadis,S.Bandla,K.Rhodes, The authors thank Javed Siddiqui and Mandy Davis for technical D.R. Rhodes assistance. Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.K. Cani, A.S. McDaniel, V. Yadati, A.M. Amin, J. Bratley, K. Rhodes Grant Support Analysis and interpretation of data (e.g., statistical analysis, biostatistics, S.A. Tomlins is supported by the A. Alfred Taubman Medical Research computational analysis): A.K. Cani, D.H. Hovelson, M.J. Quist, C.S. Grasso, Institute. C.G. Kleer, S.A. Tomlins Writing, review, and/or revision of the manuscript: A.K. Cani, D.H. Hovelson, A.S. McDaniel, S. Sadis, M.J. Haller, A.M. Amin, C.S. Grasso, Received October 27, 2014; revised December 9, 2014; accepted December C.G.Kleer,S.A.Tomlins 15, 2014; published OnlineFirst March 30, 2015.

References 1. Yang X, Kandil D, Cosar EF, Khan A. Fibroepithelial tumors of the breast: 11. Tse GM, Lui PC, Vong JS, Lau KM, Putti TC, Karim R, et al. Increased pathologic and immunohistochemical features and molecular mechan- epidermal growth factor receptor (EGFR) expression in malignant mam- isms. Arch Pathol Lab Med 2014;138:25–36. mary phyllodes tumors. Breast Cancer Res Treat 2009;114:441–8. 2. Tan PH, Thike AA, Tan WJ, Thu MM, Busmanis I, Li H, et al. Predicting 12. McDaniel AS, Zhai Y, Cho KR, Dhanasekaran SM, Montgomery JS, Pala- clinical behaviour of breast phyllodes tumours: a nomogram based on pattu G, et al. HRAS mutations are frequent in inverted urothelial neo- histological criteria and surgical margins. J Clin Pathol 2012;65:69–76. plasms. Hum Pathol 2014;45:1957–65. 3. Parker SJ, Harries SA. Phyllodes tumors. Postgrad Med J 2001;77:428–35. 13. Chang X, Wang K. wANNOVAR: annotating genetic variants for personal 4. Moffat CJ, Pinder SE, Dixon AR, Elston CW, Blamey RW, Ellis IO. Phyllodes genomes via the web. J Med Genet 2012;49:433–6. tumours of the breast: a clinicopathological review of thirty-two cases. 14. Makinen N, Mehine M, Tolvanen J, Kaasinen E, Li Y, Lehtonen HJ, et al. Histopathology 1995;27:205–18. MED12, the mediator complex subunit 12 gene, is mutated at high 5. Jones AM, Mitter R, Springall R, Graham T, Winter E, Gillett C, et al. A frequency in uterine leiomyomas. Science 2011;334:252–5. comprehensive genetic profile of phyllodes tumours of the breast detects 15. de Graaff MA, Cleton-Jansen AM, Szuhai K, Bovee JV. Mediator complex important mutations, intra-tumoral genetic heterogeneity and new genetic subunit 12 exon 2 mutation analysis in different subtypes of smooth changes on recurrence. J Pathol 2008;214:533–44. muscle tumors confirms genetic heterogeneity. Hum Pathol 2013;44: 6. Kuijper A, Snijders AM, Berns EM, Kuenen-Boumeester V, van der Wall E, 1597–604. Albertson DG, et al. Genomic profiling by array comparative genomic 16. Schwetye KE, Pfeifer JD, Duncavage EJ. MED12 exon 2 mutations in uterine hybridization reveals novel DNA copy number changes in breast phyllodes and extrauterine smooth muscle tumors. Hum Pathol 2014;45:65–70. tumours. Cell Oncol 2009;31:31–9. 17. Lim WK, Ong CK, Tan J, Thike AA, Ng CC, Rajasegaran V, et al. Exome 7. Lae M, Vincent-Salomon A, Savignoni A, Huon I, Freneaux P, Sigal-Zafrani sequencing identifies highly recurrent MED12 somatic mutations in breast B, et al. Phyllodes tumors of the breast segregate in two groups according to fibroadenoma. Nat Genet 2014;46:877–80. genetic criteria. Mod Pathol 2007;20:435–44. 18. Kampjarvi K, Makinen N, Kilpivaara O, Arola J, Heinonen HR, Bohm J, et al. 8. Lu YJ, Birdsall S, Osin P, Gusterson B, Shipley J. Phyllodes tumors of the Somatic MED12 mutations in uterine leiomyosarcoma and colorectal breast analyzed by comparative genomic hybridization and association of cancer. Br J Cancer 2012;107:1761–5. increased 1q copy number with stromal overgrowth and recurrence. Genes 19. Ravegnini G, Marino-Enriquez A, Slater J, Eilers G, Wang Y, Zhu M, et al. Chromosomes Cancer 1997;20:275–81. MED12 mutations in leiomyosarcoma and extrauterine leiomyoma. Mod 9. Lv S, Niu Y, Wei L, Liu Q, Wang X, Chen Y. Chromosomal aberrations and Pathol 2013;26:743–9. genetic relations in benign, borderline and malignant phyllodes tumors of 20. Barbieri CE, Baca SC, Lawrence MS, Demichelis F, Blattner M, Theurillat JP, the breast: a comparative genomic hybridization study. Breast Cancer Res et al. Exome sequencing identifies recurrent SPOP, FOXA1 and MED12 Treat 2008;112:411–8. mutations in prostate cancer. Nat Genet 2012;44:685–9. 10. Kersting C, Kuijper A, Schmidt H, Packeisen J, Liedtke C, Tidow N, et al. 21. Kang YK, Guermah M, Yuan CX, Roeder RG. The TRAP/Mediator coacti- Amplifications of the epidermal growth factor receptor gene (egfr) are vator complex interacts directly with estrogen receptors alpha and beta common in phyllodes tumors of the breast and are associated with tumor through the TRAP220 subunit and directly enhances estrogen receptor progression. Lab Invest 2006;86:54–61. function in vitro. Proc Natl Acad Sci U S A 2002;99:2642–7.

www.aacrjournals.org Mol Cancer Res; 13(4) April 2015 619

Next-Gen Sequencing Exposes Frequent MED12 Mutations and Actionable Therapeutic Targets in Phyllodes Tumors

Andi K. Cani, Daniel H. Hovelson, Andrew S. McDaniel, et al.

Mol Cancer Res 2015;13:613-619. Published OnlineFirst January 15, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-14-0578

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2015/01/14/1541-7786.MCR-14-0578.DC1

Visual A diagrammatic summary of the major findings and biological implications: Overview http://mcr.aacrjournals.org/content/13/4/613/F1.large.jpg

Cited articles This article cites 21 articles, 5 of which you can access for free at: http://mcr.aacrjournals.org/content/13/4/613.full#ref-list-1

Citing articles This article has been cited by 14 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/13/4/613.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/13/4/613. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.