Cell-Cycle Gene Alterations in 4,864 Tumors Analyzed by Next

Published OnlineFirst May 11, 2016; DOI: 10.1158/1535-7163.MCT-16-0071

Cancer Biology and Signal Transduction Molecular Cancer Therapeutics Cell-Cycle Gene Alterations in 4,864 Tumors Analyzed by Next-Generation Sequencing: Implications for Targeted Therapeutics Teresa Helsten1, Shumei Kato2, Maria Schwaederle1, Brett N. Tomson3, Timon P.H. Buys3, Sheryl K. Elkin3, Jennifer L. Carter3, and Razelle Kurzrock1

Abstract

Alterations in the cyclin-dependent kinase (CDK)-retinoblas- existing and mutually exclusive genetic aberrations showed that toma (RB) machinery disrupt cell-cycle regulation and are being CCND1, CCND2, and CCND3 aberrations were all positively targeted in drug development. To understand the cancer types associated with CDK6 aberrations [OR and P values, multivariate impacted by this pathway, we analyzed frequency of abnormal- analysis: CCND1 and CDK6 (OR ¼ 3.5; P < 0.0001), CCND2 ities in key cell-cycle genes across 4,864 tumors using next- and CDK6 (OR ¼ 4.3; P ¼ 0.003), CCND3 and CDK6 (OR ¼ 3.6; generation sequencing (182 or 236 genes; Clinical Laboratory P ¼ 0.007)]. In contrast, RB1 alterations were negatively associ- Improvement Amendments laboratory). Aberrations in the cell- ated with multiple gene anomalies in the cell-cycle pathway, cycle pathway were identiﬁed in 39% of cancers, making this including CCND1 (OR ¼ 0.25; P ¼ 0.003), CKD4 (OR ¼ 0.10; pathway one of the most commonly altered in cancer. The P ¼ 0.001), and CDKN2A/B (OR ¼ 0.21; P < 0.0001). In con- frequency of aberrations was as follows: CDKN2A/B (20.1% of clusion, aberrations in the cell-cycle pathway were very common all patients), RB1 (7.6%), CCND1 (6.1%), CCNE1 (3.6%), CDK4 in diverse cancers (39% of 4,864 neoplasms). The frequencies (3.2%), CCND3 (1.8%), CCND2 (1.7%), and CDK6 (1.7%). and types of alterations differed between and within tumor Rates and types of aberrant cell-cycle pathway genes differed types and will be informative for drug development strategies. between cancer types and within histologies. Analysis of co- Mol Cancer Ther; 15(7); 1682–90. 2016 AACR.

Introduction binding sites), and subsequent phosphate groups are added by the cyclin E and CDK2 complex, to further hyperphosphorylate The cyclin D–cyclin-dependent kinase (CDK)–retinoblastoma the RB protein before entering into the next phase of cell cycle (Fig. (RB) pathway is a key gatekeeper for the G phase of the cell cycle. 1 1; ref. 7). This pathway can be altered through multiple mechan- Aberrations in the cell-cycle pathway have been implicated in isms, including loss of RB1, increased signaling through CDK4/6 human cancer that leads to tumor proliferation, chromosomal ampliﬁcation, aberrations in CCND1/CCNE1, or inactivation of instability, and attenuation of genomic integrity (1–3). pathway inhibitors including CDKN2A/B (5, 6). The RB tumor suppressor protein is crucial in regulating the G 1 Alterations in the cell-cycle pathway have been described in phase. The G phase is also controlled by the cyclin D and CDK4/6 1 multiple cancers (8–12) and are associated with poorer clinical complex, leading to phosphorylation of the RB protein and outcome in some tumor types (13–15), including acute lympho- subsequent E2F-mediated transcription of target genes that are blastic leukemia (9), ovarian (10), colon cancer (12), and medul- required for G cell-cycle progression (4–6). Recent literature 1 loblastoma (11). suggests that the cyclin D1 and CDK4/6 complex adds just one To overcome cyclin pathway abnormalities in cancers, there are phosphate group to the RB (out of 14 different phosphate- several inhibitors in clinical trials, with varying selectivity for speciﬁc members of the CDK family (2). Among them, palbociclib (PD0332991), a potent and selective CDK4/6 inhibitor, 1Center for Personalized Cancer Therapy, UC San Diego Moores Can- demonstrated prolonged progression-free survival (PFS) in cer Center, La Jolla, California. 2Department of Investigational Cancer patients with hormone receptor–positive, Her2-negative meta- Therapeutics, The University of Texas MD Anderson Cancer Center, Houston, Texas. 3N-of-One, Inc., Lexington, Massachusetts. static breast cancer, in combination with the hormone modulator letrozole, as compared with letrozole alone [PALOMA-1 trial, Note: Supplementary data for this article are available at Molecular Cancer median PFS of 20.2 vs. 10.2 months, respectively (P ¼ 0.0004); Therapeutics Online (http://mct.aacrjournals.org/). ref. 16]. These observations led to accelerated FDA approval in T. Helsten and S. Kato contributed equally to this article. February 2015. Additional CDK4/6 inhibitors are in clinical trials, Corresponding Author: Shumei Kato, Department of Investigational Cancer including LEE011 (17), abemaciclib (LY-2835219; ref. 4), and the Therapeutics, The University of Texas MD Anderson Cancer Center, 1400 CDK1/2/5/9 inhibitor dinaciclib (MK-7965, formerly SCH- Holcombe Blvd, Houston, TX 77030. Phone: 713-795-9246; Fax: 713-745- 727965; refs. 18, 19). 3885; E-mail: [email protected] With novel promising CDK inhibitors in clinical trials (4, 16– doi: 10.1158/1535-7163.MCT-16-0071 19), the comprehensive analysis of cell-cycle gene aberrations 2016 American Association for Cancer Research. among diverse cancer types is of interest. Next-generation

1682 Mol Cancer Ther; 15(7) July 2016

Cell-Cycle Gene Aberrations in Cancer

sequencing (NGS) technology makes rapid and accurate identi- Statistical analysis fication of these aberrations feasible. Herein, we report the Descriptive statistics were used to summarize the baseline molecular characteristics of CCND1, CCND2, CCND3, CDK4/ patient characteristics. The Fisher exact test was used to assess 6, CCNE1, CDKN2A/B, and RB1 in 4,864 patients with diverse the association between categorical variables. All tests were two cancers interrogated by NGS. sided. Statistical analyses were carried out using SPSS version 22.0. Results Materials and Methods Overview of frequency of alterations in the key cell-cycle Patients pathway genes (CCND1, CCND2, CCND3, CDK4/6, CDKN2A/B, CCND1/2/3, CDK4/6, CCNE1, CDKN2A/B We investigated the , CCNE1, and RB1) RB1 and aberration status of patients with diverse malignancies Among the 4,864 patients, the most common cancers were that were referred for NGS from December 2011 to November breast (10.4%) and lung adenocarcinoma (8.4%; Table 1). N ¼ 2013 ( 4,864; Table 1 and Supplementary Tables S1 Thirty-nine percent of patients had at least one cell-cycle and S2). pathway anomaly. Among 1,918 patients with aberrant cell- cycle pathway genes, CDKN2A/B alterations were the most Tissue samples and mutational analysis common [51.0% (978/1,918)], followed by aberrations in RB1 We collected sequencing data from 4,864 cancers whose for- [19.2% (369/1,918)], CCND1 [15.4% (295/1,918)], CCNE1 malin-fixed, paraffin-embedded (FFPE) tumor samples were [9.2% (176/1,918)], and CDK4 [8.1% (156/1,918)]. CCND3 submitted to a Clinical Laboratory Improvement Amend- [4.5% (86/1,918)], CCND2 [4.4% (84/1,918)], and CDK6 ments–certified laboratory for genomic profiling (Foundation [4.4% (84/1,918)] were less common. Fourteen percent Medicine). Samples were required to have a surface area 25 (274/1,918) of patients had more than one abnormality in 2 3 mm , volume 1mm, nucleated cellularity 80%, and tumor these cell-cycle pathway genes (Table 1; Fig. 2). content 20% (20). The methods used in this assay have been – fl validated and previously reported (20 22). Brie y, 50 to 200 ng Overview of cancer diagnosis and abnormalities in cell-cycle fi of genomic DNA was extracted and puri ed from the submitted pathway genes FFPE tumor samples. This whole-genome DNA was subjected to Aberrations in the cell-cycle pathway were most frequently shotgun library construction and hybridization-based capture detected among glioblastoma [73.8% (62/84)], followed by those before paired end sequencing on the Illumina HiSeq2000 plat- with squamous cell carcinoma (SCC) of the esophagus [71.4% form. Hybridization selection is performed using individually (15/21)] and transitional cell carcinoma of the bladder [67.8% synthesized baits targeting the exons of 182 or 236 cancer-related (61/90)]. Papillary thyroid carcinoma was less frequently asso- genes and the introns of 14 or 19 genes frequently rearranged in ciated with cell-cycle aberrations [9.5% (2/21); Fig. 3; Table 1]. cancer. Sequence data were processed using a customized analysis Overall, the frequencies of cell-cycle pathway aberrations among pipeline (20). Sequencing was performed with an average cancer diagnoses from our current report were similar to those > sequencing depth of coverage greater than 250 , with 100 at found in the cBioPortal data (Fig. 4). >99% of exons. This method of sequencing allows for detection of copy number alterations, gene rearrangements, and somatic Distribution of specific gene alterations mutations with 99% specificity and >99% sensitivity for base CCND1. CCND1 amplification was seen in 6.1% of patients (295/ substitutions at 5 mutant allele frequency and >95% sensitivity 4,864). CCND1 amplification was most commonly identified in for copy number alterations. A threshold of 8 copies for gene patients with SCC of esophagus [42.9% (9/21); Supplementary Fig. amplification with 6 copies considered equivocal (except for S1A; Supplementary Tables S1 and S2]. Overall, the results from the ERRB2, which is considered equivocally amplified with 5 cop- cBioPortal data were similar, with 7.8% (466/6,009) of cancer cases ies) was used. Synonymous mutations were not included in the having CCND1 aberrations (Supplementary Fig. S2A). analysis. Also of note, promoter/enhancer regions for the CDKN2A/B RB1 and loci were not interrogated, and only coding CCND2. CCND2 aberrations were documented in only 1.7% (84/ regions were evaluated in the current report. The submitting 4,864) of all cancer types. The most common alteration was fi physicians provided speci cation of tumor types. The database amplification [96.4% (81/84)]; mutations were infrequent fi was deidenti ed with only diagnosis available. NGS data were [3.6% (3/84)]. CCND2 aberrations were most commonly seen collected and interpreted by N-of-One, Inc. For this study, the in patients with adrenal carcinoma [11.5% (3/26); Supplemen- dataset of 4,864 sequenced tumors was queried for alterations in tary Fig. S1B; Supplementary Tables S1 and S2]. These data are CCND1, CCND2, CCND3, CDK4/6, CCNE 1, CDKN2A/B , and consistent with the data derived from the cBioPortal, where RB1 genes. This study and data analysis herein was performed in CCND2 alterations across all cancers were also infrequent, being accordance with UCSD IRB guidelines. aberrant in 2.5% (152/6,009) of cases, including a few cases of deletion [0.3% (20/6,009); Supplementary Fig. S2B]. cBio Cancer Genomics Portal data Publicly available datasets containing genomic information CCND3. CCND3 aberrations were seen in 1.8% (86/4,864) of from a diverse array of cancer types were investigated for all cancer types; amplification [94.2% (81/86)], mutations [3.5% cell-cycle pathway genes (CCND1, CCND2, CCND3, CDK4/6, (3/86)], and fusions [2.3% (2/86)] were observed. CCND3 anom- CCNE 1, CDKN2A/B, and RB1) using cBio Cancer Genomics alies were most commonly found in esophageal adenocarcinoma Portal data (cBioPortal; http://cbioportal.org, March 2016) to [11.6% (8/69); Supplementary Fig. S1C; Supplementary Tables compare with the current study (see Supplementary Methods for S1 and S2]. According to cBioPortal, CCND3 was aberrant in 2.5% additional information; refs. 23, 24). (149/6,009) of cancer cases, with 7.5% (14/186) of esophageal

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1683

Helsten et al.

Table 1. Summary of cell-cycle gene aberrations by tumor type: Primary cancer diagnosis with n 20a Tumor type Any alteration CCND1 CCND2 CCND3 CDK4 CDK6 CDKN2A/B CCNE1 RB1 Breast carcinoma (n ¼ 507) 203 (40.0%) 87 (17.2%) 13 (2.6%) 15 (3.0%) 12 (2.4%) 10 (2.0%) 40 (7.9%) 28 (5.5%) 38 (7.5%) Lung adenocarcinoma (n ¼ 409) 171 (41.8%) 14 (3.4%) 5 (1.2%) 7 (1.7%) 19 (4.6%) 4 (1.0%) 98 (24.0%) 20 (4.9%) 27 (6.6%) Sarcoma (n ¼ 348) 186 (53.4%) 9 (2.6%) 6 (1.7%) 13 (3.7%) 40 (11.5%) 0 (0%) 70 (20.1%) 14 (4.0%) 61 (17.5%) Colorectal adenocarcinoma (n ¼ 298) 30 (10.1%) 2 (0.7%) 8 (2.7%) 4 (1.3%) 1 (0.3%) 2 (0.7%) 6 (2.0%) 3 (1.0%) 8 (2.7%) Carcinoma of unknown primary (n ¼ 269) 112 (41.6%) 10 (3.7%) 3 (1.1%) 3 (1.1%) 6 (2.2%) 5 (1.9%) 70 (26.0%) 7 (2.6%) 23 (8.6%) Ovarian serous carcinoma (n ¼ 169) 45 (26.6%) 2 (1.2%) 5 (3.0%) 6 (3.6%) 2 (1.2%) 1 (0.6%) 7 (4.1%) 20 (11.8%) 6 (3.6%) Pancreatic ductal adenocarcinoma (n ¼ 159) 65 (40.9%) 2 (1.3%) 4 (2.5%) 2 (1.3%) 2 (1.3%) 3 (1.9%) 49 (30.8%) 5 (3.1%) 4 (2.5%) Melanoma (n ¼ 135) 62 (45.9%) 5 (3.7%) 1 (0.7%) 2 (1.5%) 4 (3.0%) 1 (0.7%) 49 (36.3%) 0 (0%) 6 (4.4%) Gastric adenocarcinoma (n ¼ 134) 47 (35.1%) 10 (7.5%) 1 (0.7%) 4 (3.0%) 3 (2.2%) 8 (6.0%) 22 (16.4%) 6 (4.5%) 2 (1.5%) NSCLC (n ¼ 124) 47 (37.9%) 7 (5.6%) 0 (0%) 0 (0%) 2 (1.6%) 3 (2.4%) 27 (21.8%) 2 (1.6%) 13 (10.5%) Cholangiocarcinoma (n ¼ 114) 36 (31.6%) 3 (2.6%) 1 (0.9%) 2 (1.8%) 2 (1.8%) 4 (3.5%) 27 (23.7%) 2 (1.8%) 2 (1.8%) HNSCC (n ¼ 107) 59 (55.1%) 26 (24.3%) 1 (0.9%) 0 (0%) 1 (0.9%) 5 (4.7%) 47 (43.9%) 0 (0%) 5 (4.7%) Lung SCC (n ¼ 92) 54 (58.7%) 9 (9.8%) 4 (4.3%) 1 (1.1%) 1 (1.1%) 4 (4.3%) 40 (43.5%) 2 (2.2%) 5 (5.4%) Bladder urothelial (transitional cell) carcinoma (n ¼ 90) 61 (67.8%) 18 (20.0%) 0 (0%) 3 (3.3%) 5 (5.6%) 2 (2.2%) 32 (35.6%) 1 (1.1%) 14 (15.6%) Glioblastoma (n ¼ 84) 62 (73.8%) 0 (0%) 1 (1.2%) 0 (0%) 10 (11.9%) 4 (4.8%) 45 (53.6%) 0 (0%) 10 (11.9%) Endometrial adenocarcinoma (n ¼ 79) 18 (22.8%) 2 (2.5%) 0 (0%) 1 (1.3%) 1 (1.3%) 1 (1.3%) 3 (3.8%) 7 (8.9%) 7 (8.9%) Esophageal adenocarcinoma (n ¼ 69) 39 (56.5%) 9 (13.0%) 1 (1.4%) 8 (11.6%) 2 (2.9%) 8 (11.6%) 17 (24.6%) 10 (14.5%) 1 (1.4%) Prostate adenocarcinoma (n ¼ 63) 14 (22.2%) 6 (9.5%) 0 (0%) 1 (1.6%) 2 (3.2%) 0 (0%) 4 (6.3%) 1 (1.6%) 2 (3.2%) Renal cell carcinoma (n ¼ 58) 12 (20.7%) 2 (3.4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 10 (17.2%) 0 (0%) 2 (3.4%) Ovarian epithelial carcinoma (n ¼ 53) 20 (37.7%) 3 (5.7%) 3 (5.7%) 1 (1.9%) 1 (1.9%) 1 (1.9%) 8 (15.1%) 8 (15.1%) 1 (1.9%) Head and neck adenoid cystic carcinoma (n ¼ 48) 6 (12.5%) 0 (0%) 0 (0%) 0 (0%) 1 (2.1%) 0 (0%) 4 (8.3%) 0 (0%) 1 (2.1%) Gallbladder adenocarcinoma (n ¼ 45) 16 (35.6%) 1 (2.2%) 1 (2.2%) 1 (2.2%) 0 (0%) 0 (0%) 10 (22.2%) 5 (11.1%) 2 (4.4%) SCLC (n ¼ 43) 27 (62.8%) 1 (2.3%) 0 (0%) 1 (2.3%) 1 (2.3%) 0 (0%) 2 (4.7%) 1 (2.3%) 26 (60.5%) Hepatocellular carcinoma (n ¼ 43) 13 (30.2%) 2 (4.7%) 0 (0%) 0 (0%) 1 (2.3%) 1 (2.3%) 6 (14.0%) 0 (0%) 5 (11.6%) Neuroendocrine carcinoma (n ¼ 40) 18 (45.0%) 2 (5.0%) 3 (7.5%) 0 (0%) 1 (2.5%) 0 (0%) 3 (7.5%) 3 (7.5%) 7 (17.5%) Mesothelioma (n ¼ 36) 16 (44.4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 16 (44.4%) 1 (2.8%) 0 (0%) Cutaneous SCC (n ¼ 35) 23 (65.7%) 1 (2.9%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 19 (54.3%) 0 (0%) 3 (8.6%) SCC (n ¼ 35) 19 (54.3%) 5 (14.3%) 2 (5.7%) 0 (0%) 0 (0%) 0 (0%) 15 (42.9%) 1 (2.9%) 1 (2.9%) Salivary gland adenocarcinoma (n ¼ 31) 11 (35.5%) 2 (6.5%) 0 (0%) 1 (3.2%) 2 (6.5%) 0 (0%) 6 (19.4%) 0 (0%) 0 (0%) Large cell lung carcinoma (n ¼ 30) 18 (60.0%) 2 (6.7%) 1 (3.3%) 0 (0%) 2 (6.7%) 0 (0%) 9 (30.0%) 1 (3.3%) 5 (16.7%) Gastroesophageal junction carcinoma (n ¼ 29) 18 (62.1%) 4 (13.8%) 0 (0%) 0 (0%) 1 (3.4%) 6 (20.7%) 7 (24.1%) 4 (13.8%) 1 (3.4%) GIST (n ¼ 29) 14 (48.3%) 0 (0%) 0 (0%) 0 (0%) 1 (3.4%) 0 (0%) 11 (37.9%) 0 (0%) 2 (6.9%) Neuroblastoma (n ¼ 28) 10 (35.7%) 1 (3.6%) 0 (0%) 0 (0%) 5 (17.9%) 1 (3.6%) 2 (7.1%) 0 (0%) 1 (3.6%) Adrenal carcinoma (n ¼ 26) 9 (34.6%) 0 (0%) 3 (11.5%) 0 (0%) 4 (15.4%) 0 (0%) 4 (15.4%) 0 (0%) 1 (3.8%) Cervical adenocarcinoma (n ¼ 25) 5 (20.0%) 0 (0%) 0 (0%) 1 (4.0%) 0 (0%) 0 (0%) 2 (8.0%) 2 (8.0%) 0 (0%) Head and neck carcinoma (n ¼ 24) 14 (58.3%) 5 (20.8%) 0 (0%) 1 (4.2%) 2 (8.3%) 0 (0%) 8 (33.3%) 1 (4.2%) 2 (8.3%) Appendix adenocarcinoma (n ¼ 23) 8 (34.8%) 3 (13.0%) 1 (4.3%) 1 (4.3%) 0 (0%) 1 (4.3%) 1 (4.3%) 1 (4.3%) 1 (4.3%) Anal SCC (n ¼ 23) 7 (30.4%) 4 (17.4%) 0 (0%) 1 (4.3%) 0 (0%) 0 (0%) 2 (8.7%) 0 (0%) 1 (4.3%) Astrocytoma (n ¼ 23) 6 (26.1%) 0 (0%) 0 (0%) 0 (0%) 1 (4.3%) 0 (0%) 4 (17.4%) 0 (0%) 1 (4.3%) Clear cell renal cell carcinoma (n ¼ 23) 5 (21.7%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 5 (21.7%) 0 (0%) 0 (0%) Anaplastic astrocytoma (n ¼ 22) 11 (50.0%) 0 (0%) 1 (4.5%) 0 (0%) 1 (4.5%) 1 (4.5%) 8 (36.4%) 0 (0%) 0 (0%) Pancreatic neuroendocrine tumor (n ¼ 22) 9 (40.9%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 6 (27.3%) 1 (4.5%) 3 (13.6%) Salivary gland carcinoma (n ¼ 22) 9 (40.9%) 0 (0%) 0 (0%) 0 (0%) 2 (9.1%) 1 (4.5%) 5 (22.7%) 0 (0%) 2 (9.1%) Cervical SCC (n ¼ 22) 5 (22.7%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (9.1%) 0 (0%) 3 (13.6%) Esophageal SCC (n ¼ 21) 15 (71.4%) 9 (42.9%) 1 (4.8%) 0 (0%) 0 (0%) 1 (4.8%) 8 (38.1%) 0 (0%) 2 (9.5%) Renal pelvis urothelial carcinoma (n ¼ 21) 12 (57.1%) 3 (14.3%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 8 (38.1%) 0 (0%) 2 (9.5%) Papillary thyroid carcinoma (n ¼ 21) 2 (9.5%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (9.5%) 0 (0%) 0 (0%) Abbreviations: GIST, gastrointestinal stromal tumor; HNSCC, head and neck squamous cell carcinoma; NSCLC, non-small cell lung cancer; SCC, squamous cell carcinoma; SCLC, small cell lung cancer. aPathologic diagnosis per submitting physician.

adenocarcinoma harboring CCND3 aberration (Supplementary CDK6. CDK6 amplification was seen in 1.7% (84/4,864) of all Fig. S2C). malignancies, with the most common tumor type affected being gastroesophageal junction carcinoma [20.7% of patients (6/29)], CDK4. CDK4 amplification was seen in 3.2% (156/4,864) of all followed by esophageal adenocarcinoma [11.6% (8/69); Supple- cancer types. CDK4 amplification was most commonly found in mentary Fig. S1E; Supplementary Tables S1 and S2]. This was neuroblastoma [17.9% (5/28)], followed by glioblastoma consistent with cBioPortal data, where CDK6 is aberrant in 2.7% [11.9% (10/84); Supplementary Fig. S1D; Supplementary Tables (165/6,009) of cancer cases, including CDK6 amplification in S1 and S2]. According to the cBioPortal, CDK4 aberrations were 14.5% (27/186) of esophageal adenocarcinoma (Supplement- also seen in 3.2% (191/6,009) of cancer cases (Supplementary ary Fig. S2E). Fig. S2D). As neuroblastoma data in cBioPortal only reflect mutations and not amplification, we are unable to directly com- CDKN2A/B. CDKN2A/B aberrations were seen in 20.1% (978/ pare cBioPortal alterations with our incidence of CDK4 amplifi- 4,864) of patients. These aberrations were most commonly cation in neuroblastoma. However, CDK4 amplification was seen CDKN2A/B loss [63.4% (620/978)], followed by CDKN2A altera- in 16.7% (47/281) of glioblastoma cases, which is similar to the tions [36.6% (358/978)]. Among the 358 samples analyzed for incidence in our cases (Supplementary Fig. S2D). CDKN2A alteration, 49.4% (177/358) harbored at least one

1684 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Cell-Cycle Gene Aberrations in Cancer

Figure 1. Function of cell-cycle proteins and inhibitors. The RB tumor suppressor protein plays a pivotal role in the negative control of the cell cycle. It is responsible for amajorG1 checkpoint, blocking S-phase entry and cell growth. Phosphorylation leads to functional inactivation of RB. Loss of RB cell cycle–suppressive functions can be mediated through multiple mechanisms: loss of RB1, increased signaling through CDK4 and CDK6 ampliﬁcation, overexpression or aberration of cyclin D/E, and loss of the inhibitory function of gene products, such as CDKN2A/B, the latter leading to CDK4/6 activity. Cyclin E1 and

CDK2 complex further hyperphosphorylates RB and facilitate G1–S progression. CDK inhibitors targeting CDK4/6 such as palbociclib, LY-2835219, and LEE011 as well as CDK1/2/5/9 inhibitors such as MK-7965 are under clinical investigation. frameshift or nonsense alteration, 30.2% (108/258) harbored at frequent in those with ovarian epithelial carcinoma [15.1% (8/ least one structural alteration [including loss: 20.1% (72/358), 53)] and adenocarcinoma of esophagus [14.5% (10/69); Supple- splice site alterations: 5.9% (21/358), partial loss: 2.2% (8/358), mentary Fig. S1G; Supplementary Tables S1 and S2]. According to indel: 1.1% (4/358), and rearrangement: 0.8% (3/358)], and cBioPortal, CCNE1 aberrations were also seen in 5.0% (303/ 25.1% (90/358) harbored at least one missense alteration 6,009) of cancer cases (Supplementary Fig. S2G). [including inactivating alterations: 15.9% (57/358), variant of unknown significance: 8.7% (31/258), and no effect: 0.6% RB1. Aberrations in the RB1 tumor suppressor gene were docu- (2/358); Supplementary Fig. S1F; Supplementary Tables S1 and mented in 7.6% (369/4,864) of all cancer types. Among 369 cases S2]. As opposed to CDKN2A/B aberrations, CDKN2C alterations with RB1 alteration, 53.9% (199/369) harbored at least one were rare, seen in only 0.25% (12/4,864) of cases (Supplement- nonsense or frameshift alteration, 45.3% (167/369) harbored at ary Table S2). According to the cBioPortal, CDKN2A/B was aberrant least one structural alteration [including loss: 29.0% (107/369), in 18.0% (1,084/6,009) of cases, with 44.8 % (13/29) of cutaneous partial loss: 1.9% (7/369), rearrangement: 0.5% (2/369), partial SCC harboring CDKN2A/B aberrations; these data are similar to the duplication: 0.5% (2/369), amplification: 0.3% (1/369), and observations in the current report (Supplementary Fig. S2F). splice site alterations: 13.0% (48/369)], and 2.7% (10/369) harbored at least one missense alteration [including inactivating CCNE1. CCNE1 was amplified in 3.6% (176/4,864) of pati- alterations: 1.6% (6/369) and variants of unknown significance: ents with all cancer types. CCNE1 amplifications were most 1.1% (4/369)]. RB1 aberrations were commonly identified in

Figure 2. Frequencies of alterations in cell-cycle genes. Among 4,864 patients with diverse malignancies, 1,918 (39%) were found to have alterations in cell-cycle pathway (CCND1/2/3, CDK4/6, CDKN2A/B, CCNE1,andRb1). The most common aberration was in CDKN2A/B (978/1,918; 51%), followed by RB1 (369/1,918; 19%), CCND1 (295/1,918; 15%), CCNE1 (176/1,918; 9%), and CDK4 (156/1,918; 8%). CDK6 (84/1,918; 4%), CCND2 (84/1,918; 4%), and CCND3 (86/1,918; 4%) alterations were relatively uncommon. Fourteen percent (274/1,918) of patients had more than one aberration in the cell- cycle pathway.

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1685

Helsten et al.

All 39.4% Glioblastoma 73.8% Esophageal squamous cell carcinoma 71.4% Bladder urothelial (transional cell) carcinoma 67.8% Cutaneous squamous cell carcinoma 65.7% Small cell lung carcinoma 62.8% Gastroesophageal juncon carcinoma 62.1% Large cell lung carcinoma 60.0% Lung squamous cell carcinoma 58.7% Head and neck carcinoma 58.3% Renal pelvis urothelial carcinoma 57.1% Esophageal adenocarcinoma 56.5% Head and neck squamous cell carcinoma (HNSCC) 56.2% Squamous cell carcinoma 54.3% Sarcoma 53.4% Anaplasc astrocytoma 50.0% GIST (Gastrointesnal stromal tumor) 48.3% Melanoma 45.9% Neuroendocrine carcinoma 45.0% Mesothelioma 44.4% Lung adenocarcinoma 41.8% Carcinoma unknown primary 41.6% Salivary gland carcinoma 40.9% Pancreac neuroendocrine tumor 40.9% Pancreac ductal adenocarcinoma 40.9% Breast carcinoma 40.0% Non-small cell lung carcinoma (NSCLC) 37.9% Ovarian epithelial carcinoma 37.7% Neuroblastoma 35.7% Gallbladder adenocarcinoma 35.6% Salivary gland adenocarcinoma 35.5% Gastric adenocarcinoma 35.1% Appendix adenocarcinoma 34.8% Adrenal carcinoma 34.6% Cholangiocarcinoma 31.6% Anal squamous cell carcinoma 30.4% Hepatocellular carcinoma 30.2% Ovarian serous carcinoma 26.6% Astrocytoma 26.1% Endometrial adenocarcinoma 22.8% Cervical squamous cell carcinoma 22.7% Prostate adenocarcinoma 22.2% Clear cell renal cell carcinoma 21.7% Renal cell carcinoma 20.7% Cervical adenocarcinoma 20.0% Head and neck adenoid cysc carcinoma 12.5% Colorectal adenocarcinoma 10.1% Papillary thyroid carcinoma 9.5% 0% 10% 20% 30% 40% 50% 60% 70% 80%

Figure 3. Frequencies of cumulative cell-cycle gene aberrations (CCND1/2/3, CDK4/6, CDKN2A/B, CCNE1,andRB1) by cancer type. This ﬁgure includes primary cancer diagnosis with n 20. Please see Supplementary Table 1 for complete list of cancer diagnosis with n < 20. Cell-cycle gene aberrations were most commonly seen in patients with glioblastoma (73.8%) and least common in patients with papillary thyroid carcinoma (9.5%). GIST, gastrointestinal stromal tumor; HNSCC, head and neck squamous cell carcinoma.

patients with small cell lung cancer [SCLC; 60.5% (26/43); monly associated with alterations in RB1 [23.5% (4/17)] as Supplementary Fig. S1H; Supplementary Tables S1 and S2]. The opposed to 7.5% (38/507) in other breast cancers. CDKN2A/B data in our current report are similar to the incidence reported in aberrations were also more commonly seen in patients with cBioPortal, where 7.6% (456/6,009) of cancer cases have RB1 metaplastic breast cancer [23.5% (4/17)] when compared with aberrations. Consistent with our ﬁndings, SCLC is the diagnosis those with other breast cancers [7.9% (40/507); Table 1 and most commonly associated with RB1 aberration in the cBioPortal Supplementary Table S1]. [74.5% (82/110); Supplementary Fig. S2H]. Lung cancers. Aberrations in cell-cycle elements were relatively Distribution of speciﬁc gene alterations by cancer type common in patients with non–small cell lung cancer [NSCLC; Breast cancers. Forty percent of patients (203/507) with breast adenocarcinoma, 41.8% (171/409); SCC, 58.7% (54/92); large cancer had alterations in cell-cycle pathway genes; these altera- cell carcinoma, 60.0% (18/30)] and SCLC [62.8% (27/43)]. tions were strikingly common in the small subset of patients with Alterations in RB1 were more commonly associated with SCLC metaplastic breast carcinoma [70.6% (12/17); Table 1 and Sup- [60.5% (26/43)] as opposed to NSCLC (5.4%–16.7%; Table 1). plementary Table S1]. Metaplastic breast cancer was more com- On the other hand, CDKN2A/B aberrations were more commonly

1686 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Cell-Cycle Gene Aberrations in Cancer

All (current report) All (cBioPortal) Glioblastoma Glioblastoma (TCGA, Cell 2013) Bladder urothelial (transional cell) carcinoma Bladder urothelial carcinoma (TCGA, Nature 2014) Cutaneous squamous cell carcinoma Cutaneous squamous cell carcinoma (DFCI, Clin Cancer Res 2015) Small cell lung carcinoma Small Cell Lung Cancer (U Cologne, Nature 2015) * Lung squamous cell carcinoma Lung squamous cell carcinoma (TCGA, Nature 2012) Esophageal adenocarcinoma Esophageal carcinoma (TCGA, Provisional) Head and neck squamous cell carcinoma (HNSCC) Head and neck squamous cell carcinoma (TCGA, Nature 2015) Sarcoma Sarcoma (TCGA, Provisional) Melanoma Skin cutaneous melanoma (TCGA, Provisional) Mesothelioma Mesothelioma (TCGA, Provisional) Lung adenocarcinoma Lung adenocarcinoma (TCGA, Nature 2014) Pancreac neuroendocrine tumor Pancreac neuroendocrine tumors (Johns Hopkins University, Science 2011) * Pancreac ductal adenocarcinoma Pancreac adenocarcinoma (TCGA, Provisional) Breast carcinoma Breast invasive carcinoma (TCGA, Cell 2015) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

Figure 4. Cumulative cell-cycle gene aberrations (CCND1/2/3, CDK4/6, CDKN2A/B, CCNE1,andRB1) between current report and cBioPortal. In the current report, 39.4% (1,918/4,864) of patient samples were found to have aberrations in cell-cycle pathway genes, while it was found in 38.7% (2,326/6,009) from cBioPortal (http://cbioportal.org). , cBioPortal data available for mutation only. Included cancer diagnoses with cell-cycle gene aberrations in 40% of patients from current report and when cBioPortal data were available for corresponding diagnosis. observed in NSCLC (24.0%–43.5%) when compared with SCLC cycle pathway genes [71.4% (15/21)]. Cervical SCC was the least [4.7% (2/43); Table 1]. frequently associated with molecular anomalies in cell-cycle pathway genes [22.7% (5/22); Table 1 and Supplementary Table S1]. Central nervous system tumors. Amongst central nervous system (CNS) tumors, glioblastoma was most frequently associated with Adenocarcinoma. Among adenocarcinomas, esophageal adeno- aberrations in cell-cycle pathway genes [73.8% of patients (62/ carcinoma was most commonly associated with alterations 84)], followed by anaplastic astrocytoma [50% (11/22)], glioma in the cell-cycle pathway [56.5% (39/69)], whereas such altera- [38.5% (5/13)], oligodendroglioma [35.3% (6/17)], astrocytoma tions were least common in colorectal cancer [10.1% (30/ [26.1% (6/23)], and meningioma [22.2% (4/18)]. Among 298); Table 1 and Supplementary Table S1]. patients with CNS tumors, CCND1 aberrations were only discerned in meningioma [5.6% (1/18)]. Alterations in RB1 were more commonly detected in glioblastoma [11.9% (10/84)] and Coexisting genomic aberrations among cell-cycle gene oligodendroglioma [11.8% (2/17)] and were less regularly aberrations observed in astrocytoma [4.3% (1/23)]; they were not detected We investigated whether certain cell-cycle gene aberrations in individuals with anaplastic astrocytoma, glioma, and menin- were correlated with one another using the Fisher exact test and gioma (Table 1 and Supplementary Table S1). multivariate logistic regression analysis (Fig. 5; Supplementary Table S3). CCND1, CCND2, and CCND3 aberrations were all SCC. Among SCC histology in various primary cancer types, esoph- positively associated with CDK6 aberrations [OR and P values ageal SCC was most commonly associated with aberrations in cell- after multivariate analysis: CCND1 and CDK6 (OR ¼ 3.5;

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1687

Helsten et al.

Figure 5. Co-altered cell-cycle genes. Among multiple cell cycle–associated gene aberrations, positive and negative relationships for co-alterations were seen. For example, there was a positive correlation between CDK6 and CCND3 aberrations, with OR of 4.5. In contrast, there was a negative correlation (OR ¼ 0.15) between RB1 and CDK4 anomalies. ORs are from univariate analysis (Fisher exact test). Solid black boxes indicate the association was statistically signiﬁcant after multivariate analysis (Supplementary Table S3).

P < 0.0001); CCND2 and CDK6 (OR ¼ 4.3; P ¼ 0.003); CCND3 In our dataset, among cancers originating from breast, aber- and CDK6 (OR ¼ 3.6; P ¼ 0.007); Fig. 5; Supplementary Table rations in cell-cycle pathway were more commonly seen in S3E]. On the other hand, RB1 mutations/loss were negatively patients with metaplastic breast cancer when compared with associated with multiple aberrations in the cell-cycle pathway, breast cancer patients in general [70.6% (12/17) vs. 40.0% including CCND1 (OR ¼ 0.25; P ¼ 0.003), CKD4 (OR ¼ 0.10; P ¼ (203/507); Table 1 and Supplementary Table S1]. Metaplastic 0.001), and CDKN2A/B (OR ¼ 0.21; P < 0.0001; Fig. 5; Supple- breast cancer is a rare subtype that is typically negative for mentary Table S3H). either estrogen/progesterone receptors or HER2 (27) and thus is usually treated as triple-negative breast cancer. However, women with metaplastic breast cancer experience detrimental Discussion clinical outcomes when compared with patients with triple- This study represents a comprehensive overview of aberrations negative breast cancer, (28) and thus, there is no standard in relevant cell-cycle pathway genes in a large number of cancer therapy for metaplastic breast cancer. However, recent data tissues interrogated by NGS (N ¼ 4,864). Overall, 39.4% of suggest susceptibility to a combination of liposomal doxoru- cancers (1,918/4,864) had alterations in the key cell-cycle path- bicin, bevacizumab, and temsirolimus, perhaps because these way genes, making the CDK/CCND/RB1 axis one of the most tumors also commonly harbor PI3K/AKT/mTOR pathway frequently disturbed in cancer (Fig. 2). Our findings were also aberrations (27, 29, 30). Our results suggest that, for patients comparable with the data from cBioPortal (23, 24), where 38.7 % with metaplastic breast cancer, an aberrant cell-cycle pathway (2,326/6,009) of cases harbored aberrations in these cell-cycle may be a potential therapeutic target to be considered, with pathway genes (Fig. 4 and Supplementary Fig. S2). the caveat being that about one third of the anomalies were in The frequency of cell-cycle pathway aberrations varies by dis- the CCNE1 and RB1 gene, which would confer resistance to ease, with glioblastoma having the highest rate of abnormalities CDK4/6 inhibitors. (73.8%), whereas colorectal adenocarcinoma has a much lower Among patients with lung cancer, frequencies of aberrations rate (10.1%), consistent with previous reports on smaller num- in cell-cycle pathway genes were similar between NSCLC bers of patients (Fig. 3; refs. 4, 8, 13). [adenocarcinoma, 41.8% (171/409); SCC, 58.7% (54/92); Certain cancer diagnoses were significantly associated with large cell carcinoma, 60.0% (18/30)] and SCLC [62.8% (27/ different cell-cycle pathway aberrations. For example, esophageal 43); Table 1]. However, RB1 aberrations were a hallmark of SCC was highly associated with CCND1 aberrations (OR ¼ 14.5; SCLC [60.5% (26/43)] as opposed to NSCLC (5.4%–16.7%). In P < 0.0001); gastroesophageal junction carcinoma, with CDK6 contrast, CDKN2A/B aberrations were more common among aberration (OR ¼ 19.7; P < 0.0001); and SCLC, with RB1 aber- NSCLC (24.0%–43.5%) when compared with SCLC (4.7%; ration (OR ¼ 21; P < 0.0001; Supplementary Table S3). Table 1), consistent with previous reports (31–33). It is inter- The selective CDK4/6 inhibitor palbociclib has been FDA esting to note that both SCC of lung and SCLC are associated approved in hormone receptor–positive, Her2-negative metastat- with tobacco exposure, yet the aberration pattern in cell-cycle ic breast cancer. To date, a correlation between cell-cycle gene genesisdifferent. aberrations and response of breast cancer to palbociclib has not We have also evaluated whether the same histology among been shown; however, the question remains of interest. Upregu- different cancer types shares similar patterns of aberrations lation of CDK4/6 may occur via CDK4 or CDK6 amplification, (Table 1 and Supplementary Table S1). When focusing on inactivation of CDKN2A/B,orCCND1 amplification, which sub- cancer with SCC or adenocarcinoma histologies in various sequently augments phosphorylation and inactivation of RB1 cancer types, aberrations in cell-cycle elements occurred fre- (4, 5). Meanwhile, resistance to CDK4/6 inhibitors may be quently, but rates still differed (Table 1). These results suggest mediated by RB1 or CCNE1 aberrations, which subsequently that histologic classification may be further substratified (34). leads to activation of E2F (25, 26). Of interest in this regard, head and neck and other tumors

1688 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Cell-Cycle Gene Aberrations in Cancer

infected with human papilloma virus (HPV) do not generally these limitations, this study provides a very large and compre- exhibit cell-cycle pathway abnormalities, whereas HPV-nega- hensive analysis of cell-cycle gene alterations in a wide range of tive tumors do harbor these alterations, which are often asso- human cancers. ciated with TP53 mutations (35). In conclusion, we have interrogated 4,864 patients with Understanding the mutual exclusivity among some genomic canceranddemonstratedthataberrations in cell-cycle pathway aberrations has been therapeutically important. For example, genes are extremely common, being discerned in 39% of tissues ALK rearrangements, EGFR or KRAS mutations are mutually across cancer types (Figs. 2 and 3). However, the frequencies exclusive in patients with NSCLC (36). We have reported here and the types of aberrant cell-cycle pathway differ between and a comprehensive analysis of coexisting genetic aberrations and within cancer types. Of interest in this regard, we and others mutually exclusiveness amongst cell-cycle gene aberrations. have previously shown that patients with metastatic cancer We observed that RB1 alterations were significantly less fre- mostly have distinct molecular portfolios, highlighting the quently associated with CCND1 (OR ¼ 0.25; P ¼ 0.003), CDK4 heterogeneity of cancer and that matching patients with geno- (OR ¼ 0.10; P ¼ 0.001), and CDKN2A/B (OR ¼ 0.21, P < mically targeted therapy can have salutary effects (41–47). 0.0001) aberrations (Fig. 5 and Supplementary Table S3H). Aberrations in the cell-cycle pathway may also have prognostic Aberrations in CDKN2A/B were also less associated with CDK4 value. For instance, it has been reported that cell-cycle pathway aberrations (OR ¼ 0.20; P < 0.0001; Fig. 5; Supplementary gene alterations are independently associated with poor clinical Table S3F). These data are consistent with previous smaller outcomes, including overall survival (13–15). Further study studies reporting that CDKN2A, CDK4,andRB1 were mutually willberequiredtoelucidatetheimpactofeachofthediverse exclusive in glioblastomas and CDKN2A, CDK6,andRB1 cell-cycle pathway genes and their correlation to response with were mutually exclusive in lung adenocarcinomas (37). In the various inhibitors being deployed in the clinic, including contrast, we have observed that CDK4/6 aberrations were but not limited to palbociclib (PD0332991; ref. 16), LEE011 positively associated with CCND 1/2/3 abnormalities (Fig. (17), abemaciclib (LY-2835219; ref. 4), and the CDK1/2/5/9 5; Supplementary Table S3A–S3E). The latter observations inhibitor, dinaciclib (MK-7965, formerly SCH-727965; refs. 18, support the notion that CDK4/6 and CCND 1/2/3 cooperate 19). On the basis of the frequent finding of aberrations in closely for cell-cycle progression. However, RB1 or CDKN2A/B the cell-cycle pathway across diverse malignancies, prosecution aberrations alone may be sufficient to perturb cell-cycle of this pathway in multiple tumor types merits further progression. exploration. There are several limitations to these data. First, the dataset was not clinically annotated; hence, correlation with pheno- fl typic characteristics and outcome was not feasible. Second, Disclosure of Potential Con icts of Interest S.K. Elkin and J.L. Carter have ownership interest (including patents) in methylation status (which could change gene expression) and fi N-of-One, Inc. R. Kurzrock is employed at and has ownership interest correlation between gene ampli cation and protein expression (including patents) in Curematch, Inc. and Novena, Inc.; reports receiving were not evaluated in this study. As we have not evaluated commercial research grants from Foundation Medicine, Genentech, Guar- normal tissues, the impact of germline mutations is not dant, Merck Serono, Pfizer, and Sequenom; and is a consultant/advisory addressed. Similarly, as multiple samples were not analyzed board member for Actuate Therapeutics and Sequenom. No potential con- fl fromindividualcases,wewereunabletoevaluateintratumoral icts of interest were disclosed by the other authors. heterogeneity. Third, the number of cases with each malignancy relied on the number of specimens submitted by physicians for Authors' Contributions fi molecular pro ling; this introduces the possibility of sample Conception and design: T. Helsten, S. Kato, B.N. Tomson, S.K. Elkin, J.L. Carter, size bias. In this regard, it is important to note that the dataset R. Kurzrock only included small numbers of hematologic malignancies Development of methodology: T. Helsten, S.K. Elkin (Supplementary Table S1). This is important because cell-cycle Acquisition of data (provided animals, acquired and managed patients, genes may be affected in them as well. For instance, CCND1 provided facilities, etc.): B.N. Tomson, S.K. Elkin Analysis and interpretation of data (e.g., statistical analysis, biostatistics, (also known as BCL1) is rearranged in about 60 % of mantle computational analysis): T. Helsten, S. Kato, M. Schwaederle, B.N. Tomson, cell lymphomas (38), and translocations are also seen in about T.P.H. Buys, S.K. Elkin, R. Kurzrock 15% to 20% of multiple myeloma specimens (39). Interest- Writing, review, and/or revision of the manuscript: T. Helsten, S. Kato, ingly, a preliminary study showed activity for the CDK4/6 M. Schwaederle, B.N. Tomson, T.P.H. Buys, S.K. Elkin, J.L. Carter, R. Kurzrock inhibitor in mantle cell lymphoma (40). Further study of this Study supervision: T. Helsten family of genes is needed in hematologic malignancies. Fourth, CDKN2D, RBL1, RBL2, E2F, CCNE2, other genes including Grant Support CCNA1, CCNA2,andCDK2 that are known to influence cell- R. Kurzrock is funded in part by the Joan and Irwin Jacobs philanthropic cycle pathway were not evaluated in the current report, and fund. further investigation is required as to their impact. A final limitation of the data was that diagnosis was determined on Received February 11, 2016; revised April 12, 2016; accepted April 29, 2016; the basis of the attending physician designation. Yet despite published OnlineFirst May 11, 2016.

References 1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2. Malumbres M, Barbacid M. Cell cycle, CDKs and cancer: a changing 2011;144:646–74. paradigm. Nat Rev Cancer 2009;9:153–66.

www.aacrjournals.org Mol Cancer Ther; 15(7) July 2016 1689

Helsten et al.

3. Rocca A, Farolfi A, Bravaccini S, Schirone A, Amadori D. Palbociclib (PD 25. Wiedemeyer WR, Dunn IF, Quayle SN, Zhang J, Chheda MG, Dunn GP, 0332991): targeting the cell cycle machinery in breast cancer. Expert Opin et al. Pattern of retinoblastoma pathway inactivation dictates response to Pharmacother 2014;15:407–20. CDK4/6 inhibition in GBM. Proc Natl Acad Sci U S A 2010;107:11501–6. 4. Asghar U, Witkiewicz AK, Turner NC, Knudsen ES. The history and future of 26. Taylor-Harding B, Aspuria PJ, Agadjanian H, Cheon DJ, Mizuno T, Green- targeting cyclin-dependent kinases in cancer therapy. Nat Rev Drug Discov berg D, et al. Cyclin E1 and RTK/RAS signaling drive CDK inhibitor 2015;14:130–46. resistance via activation of E2F and ETS. Oncotarget 2015;6:696–714. 5. Shapiro GI. Cyclin-dependent kinase pathways as targets for cancer treat- 27. Moulder S, Helgason T, Janku F, Wheler J, Moroney J, Booser D, et al. ment. J Clin Oncol 2006;24:1770–83. Inhibition of the phosphoinositide 3-kinase pathway for the treatment of 6. Sheppard KE, McArthur GA. The cell-cycle regulator CDK4: an emerging patients with metastatic metaplastic breast cancer. Ann Oncol 2015;26: therapeutic target in melanoma. Clin Cancer Res 2013;19:5320–8. 1346–52. 7. Narasimha AM, Kaulich M, Shapiro GS, Choi YJ, Sicinski P, Dowdy SF. Cyclin 28. Rayson D, Adjei AA, Suman VJ, Wold LE, Ingle JN. Metaplastic breast D activates the Rb tumor suppressor by mono-phosphorylation. Elife 2014;3. cancer: prognosis and response to systemic therapy. Ann Oncol 1999; 8. The Cancer Genome Atlas Research Network. Comprehensive genomic 10:413–9. characterization defines human glioblastoma genes and core pathways. 29. Moroney J, Fu S, Moulder S, Falchook G, Helgason T, Levenback C, et al. Nature 2008;455:1061–8. Phase I study of the antiangiogenic antibody bevacizumab and the mTOR/ 9. Iacobucci I, Ferrari A, Lonetti A, Papayannidis C, Paoloni F, Trino S, et al. hypoxia-inducible factor inhibitor temsirolimus combined with liposomal CDKN2A/B alterations impair prognosis in adult BCR-ABL1-positive acute doxorubicin: tolerance and biological activity. Clin Cancer Res 2012;18: lymphoblastic leukemia patients. Clin Cancer Res 2011;17:7413–23. 5796–805. 10. Kusume T, Tsuda H, Kawabata M, Inoue T, Umesaki N, Suzuki T, et al. The 30. Moulder S, Moroney J, Helgason T, Wheler J, Booser D, Albarracin C, et al. p16-cyclin D1/CDK4-pRb pathway and clinical outcome in epithelial Responses to liposomal Doxorubicin, bevacizumab, and temsirolimus in ovarian cancer. Clin Cancer Res 1999;5:4152–7. metaplastic carcinoma of the breast: biologic rationale and implications for 11. Mendrzyk F, Radlwimmer B, Joos S, Kokocinski F, Benner A, Stange DE, stem-cell research in breast cancer. J Clin Oncol 2011;29:e572–5. et al. Genomic and protein expression profiling identifies CDK6 as novel 31. The Cancer Genome Atlas Research Network. Comprehensive molecular independent prognostic marker in medulloblastoma. J Clin Oncol 2005; profiling of lung adenocarcinoma. Nature 2014;511:543–50. 23:8853–62. 32. The Cancer Genome Atlas Research Network. Comprehensive genomic 12. Xing X, Cai W, Shi H, Wang Y, Li M, Jiao J, et al. The prognostic value characterization of squamous cell lung cancers. Nature 2012;489:519–25. of CDKN2A hypermethylation in colorectal cancer: a meta-analysis. 33. Peifer M, Fernandez-Cuesta L, Sos ML, George J, Seidel D, Kasper LH, et al. Br J Cancer 2013;108:2542–8. Integrative genome analyses identify key somatic driver mutations of 13. Kato S, Schwaederle M, Daniels GA, Piccioni D, Kesari S, Bazhenova L, et al. small-cell lung cancer. Nat Genet 2012;44:1104–10. Cyclin-dependent kinase pathway aberrations in diverse malignancies: 34. Schwaederle M, Elkin SK, Tomson BN, Carter JL, Kurzrock R. Squamous- clinical and molecular characteristics. Cell Cycle 2015;14:1252–9. ness: next-generation sequencing reveals shared molecular features across 14. Schwaederle M, Daniels GA, Piccioni DE, Fanta PT, Schwab RB, Shima- squamous tumor types. Cell Cycle 2015;14:2355–61. bukuro KA, et al. Cyclin alterations in diverse cancers: outcome and 35. Seiwert TY. Ties that bind: p16 as a prognostic biomarker and the need co-amplification network. Oncotarget 2015;6:3033–42. for high-accuracy human papillomavirus testing. J Clin Oncol 2014; 15. Schwaederle M, Daniels GA, Piccioni DE, Kesari S, Fanta PT, Schwab RB, 32:3914–6. et al. Next generation sequencing demonstrates association between tumor 36. Gainor JF, Varghese AM, Ou SH, Kabraji S, Awad MM, Katayama R, et al. suppressor gene aberrations and poor outcome in patients with cancer. ALK rearrangements are mutually exclusive with mutations in EGFR or Cell Cycle 2015;14:1730–7. KRAS: an analysis of 1,683 patients with non-small cell lung cancer. Clin 16. Finn RS, Crown JP, Lang I, Boer K, Bondarenko IM, Kulyk SO, et al. The Cancer Res 2013;19:4273–81. cyclin-dependent kinase 4/6 inhibitor palbociclib in combination with 37. Negrini S, Gorgoulis VG, Halazonetis TD. Genomic instability–an evolving letrozole versus letrozole alone as first-line treatment of oestrogen receptor- hallmark of cancer. Nat Rev Mol Cell Biol 2010;11:220–8. positive, HER2-negative, advanced breast cancer (PALOMA-1/TRIO-18): a 38. Rimokh R, Berger F, Delsol G, Digonnet I, Rouault JP, Tigaud JD, et al. randomised phase 2 study. Lancet Oncol 2015;16:25–35. Detection of the chromosomal translocation t(11;14) by polymerase chain 17. Infante JR, Shapiro G, Witteveen P, Gerecitano JF, Ribrag V, Chugh R, et al. A reaction in mantle cell lymphomas. Blood 1994;83:1871–5. phase I study of the single-agent CDK4/6 inhibitor LEE011 in pts with 39. Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic events and advanced solid tumors and lymphomas. J Clin Oncol 32:5s, 2014(suppl; host interactions. Nat Rev Cancer 2002;2:175–87. abstr 2528^). 40. Leonard JP, LaCasce AS, Smith MR, Noy A, Chirieac LR, Rodig SJ, et al. 18. Mita MM, Joy AA, Mita A, Sankhala K, Jou YM, Zhang D, et al. Randomized Selective CDK4/6 inhibition with tumor responses by PD0332991 in phase II trial of the cyclin-dependent kinase inhibitor dinaciclib (MK- patients with mantle cell lymphoma. Blood 2012;119:4597–607. 7965) versus capecitabine in patients with advanced breast cancer. Clin 41. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E, Breast Cancer 2014;14:169–76. et al. Intratumor heterogeneity and branched evolution revealed by multi- 19. Stephenson JJ, Nemunaitis J, Joy AA, Martin JC, Jou YM, Zhang D, et al. region sequencing. N Engl J Med 2012;366:883–92. Randomized phase 2 study of the cyclin-dependent kinase inhibitor 42. Kurzrock R, Giles FJ. Precision oncology for patients with advanced cancer: dinaciclib (MK-7965) versus erlotinib in patients with non-small cell lung the challenges of malignant snowflakes. Cell Cycle 2015;14:2219–21. cancer. Lung Cancer 2014;83:219–23. 43. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, Brannigan 20. Frampton GM, Fichtenholtz A, Otto GA, Wang K, Downing SR, He J, et al. BW, et al. Activating mutations in the epidermal growth factor receptor Development and validation of a clinical cancer genomic profiling test based underlying responsiveness of non-small-cell lung cancer to gefitinib. on massively parallel DNA sequencing. Nat Biotechnol 2013;31:1023–31. N Engl J Med 2004;350:2129–39. 21. Thomas RK, Nickerson E, Simons JF, Janne PA, Tengs T, Yuza Y, et al. 44. Shaw AT, Kim DW, Mehra R, Tan DS, Felip E, Chow LQ, et al. Ceritinib Sensitive mutation detection in heterogeneous cancer specimens by mas- in ALK-rearranged non-small-cell lung cancer. N Engl J Med 2014;370: sively parallel picoliter reactor sequencing. Nat Med 2006;12:852–5. 1189–97. 22. Wagle N, Berger MF, Davis MJ, Blumenstiel B, Defelice M, Pochanard P, 45. Tsimberidou AM, Iskander NG, Hong DS, Wheler JJ, Falchook GS, Fu S, et al. High-throughput detection of actionable genomic alterations in et al. Personalized medicine in a phase I clinical trials program: the MD clinical tumor samples by targeted, massively parallel sequencing. Cancer Anderson Cancer Center initiative. Clin Cancer Res 2012;18:6373–83. Discov 2012;2:82–93. 46. Wheler J, Lee JJ, Kurzrock R. Unique molecular landscapes in cancer: 23. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio implications for individualized, curated drug combinations. Cancer Res cancer genomics portal: an open platform for exploring multidimensional 2014;74:7181–4. cancer genomics data. Cancer Discov 2012;2:401–4. 47. Wheler JJ, Parker BA, Lee JJ, Atkins JT, Janku F, Tsimberidou AM, et al. 24. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. Unique molecular signatures as a hallmark of patients with metastatic Integrative analysis of complex cancer genomics and clinical profiles using breast cancer: implications for current treatment paradigms. Oncotarget the cBioPortal. Sci Signal 2013;6:pl1. 2014;5:2349–54.

1690 Mol Cancer Ther; 15(7) July 2016 Molecular Cancer Therapeutics

Cell-Cycle Gene Alterations in 4,864 Tumors Analyzed by Next-Generation Sequencing: Implications for Targeted Therapeutics

Teresa Helsten, Shumei Kato, Maria Schwaederle, et al.

Mol Cancer Ther 2016;15:1682-1690. Published OnlineFirst May 11, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-16-0071

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2016/05/11/1535-7163.MCT-16-0071.DC1

Cited articles This article cites 45 articles, 17 of which you can access for free at: http://mct.aacrjournals.org/content/15/7/1682.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/15/7/1682.full#related-urls

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

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

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/15/7/1682. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.