Published OnlineFirst December 30, 2016; DOI: 10.1158/1541-7786.MCR-16-0319 Cell Cycle and Senescence Molecular Cancer Research Stromal Senescence By Prolonged CDK4/6 Inhibition Potentiates Tumor Growth Xiangnan Guan1, Kyle M. LaPak1,2, Rebecca C. Hennessey2, Christina Y.Yu2,3, Reena Shakya4, Jianying Zhang3, and Christin E. Burd1,2,4 Abstract Senescent cells within the tumor microenvironment (TME) blasts, melanoma cell lines exhibited genotype-dependent pro- adopt a proinflammatory, senescence-associated secretory phe- liferative responses. However, in vivo, PD-0332991–treated notype (SASP) that promotes cancer initiation, progression, fibroblasts enhanced the growth of all melanoma lines tested and therapeutic resistance. Here, exposure to palbociclib (PD- and promoted the recruitment of Gr-1–positive immune cells. 0332991), a CDK4/6 inhibitor, induces senescence and a These data indicate that prolonged CDK4/6 inhibitor treatment robust SASP in normal fibroblasts. Senescence caused by pro- causes normal fibroblasts to enter senescence and adopt a longed CDK4/6 inhibition is DNA damage–independent and robust SASP. Such senescent cells suppress the antitumor associated with Mdm2 downregulation, whereas the SASP immune response and promote melanoma growth in immu- elicited by these cells is largely reliant upon NF-kB activation. nocompetent, in vivo models. Based upon these observations, it was hypothesized that the exposure of nontransformed stromal cells to PD-0332991 Implications: The ability of prolonged CDK4/6 inhibitor treat- would promote tumor growth. Ongoing clinical trials of ment to induce cellular senescence and a robust SASP in primary CDK4/6 inhibitors in melanoma prompted a validation of this cells may hinder therapeutic efficacy and promote long-term, hypothesis using a suite of genetically defined melanoma cells gerontogenic consequences that should be considered in clin- (i.e., Ras mutant, Braf mutant, and Ras/Braf wild-type). When ical trials aiming to treat melanoma and other cancer types. cultured in the presence of CDK4/6i-induced senescent fibro- Mol Cancer Res; 15(3); 237–49. Ó2016 AACR. Introduction fore, senescent cells are not phenotypic equivalents and likely to contribute to distinct biological outcomes in vivo. Cellular senescence is a process in which cells with the potential Emerging data reveal the presence of senescent cells within the to divide permanently exit the cell cycle, but remain viable and tumor microenvironment. For example, using a genetically mod- metabolically active. Senescent cells are characterized by numer- ified mouse model wherein activation of the senescence biomark- ous cellular phenotypes, including insensitivity to mitogenic er, p16INK4a, induces firefly luciferase expression, the accumula- stimuli, flattened morphology, increased senescence-associated tion of senescent stromal cells can be visualized both in auto- ß-galactosidase activity (SA-ß-gal), shortened telomeres, elevated chthonous tumor transplants and spontaneous neoplasms (4). cyclin-dependent kinase inhibitor expression, changes in chro- In humans, p16INK4a-positive stromal cells accumulate around matin structure, pervasive DNA damage foci, and resistance to large-cell lung carcinomas and ductal carcinoma in situ lesions apoptosis and activation of the proinflammatory senescence- of the breast and pancreas (5, 6). Of note, stromal p16INK4a associated secretory phenotype (SASP; refs. 1, 2). Notably, not expression in breast cancer is more predictive of disease recurrence every senescent cell exhibits all of these characteristics. Instead, the than HER2, PR, or ER status (7), suggesting that senescent stromal triggering event (e.g., oncogene activation, telomere attrition, cells are indicative of poor prognosis. prolonged CDKi expression, or DNA damage) and originating Numerous in vitro coculture studies indicate that the SASP of cell type appear to dictate which phenotypes ensue (2, 3). There- senescent stromal cells influences cancer initiation, progres- sion, and therapeutic response; however, few studies extend 1Department of Molecular Genetics, The Ohio State University, Columbus, Ohio. these observations to in vivo models (8–11). Of the publica- 2Department of Cancer Biology and Genetics, The Ohio State University, Columbus, tions that do address how senescent stromal cells influence Ohio. 3Department of Biomedical Informatics, The Ohio State University, Columbus, 4 tumor growth in vivo, only one has been conducted in an Ohio. The Ohio State University Comprehensive Cancer Center - Arthur G. James immunoproficient mouse model (12). Consequently, mechan- Cancer Hospital and Richard J. Solove Research Institute, Columbus, Ohio. isms by which paracrine SASP signals emanating from the Note: Supplementary data for this article are available at Molecular Cancer senescent tumor stroma might alter malignant cell clearance Research Online (http://mcr.aacrjournals.org/). by the immune system are understudied, a particularly perti- Corresponding Author: Christin E. Burd, The Ohio State University, Biomedical nent point given increasing interest in combining established Research Tower, Rm 586, 460 W. 12th Avenue, Columbus, OH 43210. Phone: 614- treatment modalities (i.e., chemotherapy, radiotherapy, and 688-7569; Fax: 614-292-6356; E-mail: [email protected] molecularly targeted therapies) with immunotherapy (13). doi: 10.1158/1541-7786.MCR-16-0319 Senescence induced by pharmaceutical inhibitors of CDK4/6 Ó2016 American Association for Cancer Research. is of growing biological and clinical relevance. Developed to www.aacrjournals.org 237 Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2017 American Association for Cancer Research. Published OnlineFirst December 30, 2016; DOI: 10.1158/1541-7786.MCR-16-0319 Guan et al. combat frequent inactivation of the CDK/Cyclin-RB signaling sequence inserted between the BamHI and NotI restriction sites. axis in human cancer, these compounds induce a permanent GFP from pEGFP-N3 (Clontech) was inserted into the resulting cell-cycle arrest in many tumor-derived cell lines. CDK4/6 multiple cloning sequence. Transduced tumor cells were selected inhibitors (e.g., LY2835219, LEE011, G1T28, and P1446A- with 3 mg/mL puromycin. 05) are under clinical investigation for a variety of tumor types, Cell lines were tested for mycoplasma using Mycoplasma Plus including melanoma. Given early results suggesting that CDK4/ PCR Primers (Agilent Technologies) and identity-verified at mul- 6 inhibitors enhance the efficacy of other targeted melanoma tiple time points during the study. Identity verification was con- therapies (14–16), it is likely these drugs will be soon approved ducted by PCR for all alleles unique to the study cell lines (e.g., for use in patients with metastatic melanoma. One potential alterations in Stk11, Ink4a/Arf, NRas, and Braf). In addition, concern surrounding the use of these drugs stems from recent BrafV600E and NRasQ61R mutations were sequence verified using in vitro data showing that extended exposure to PD-0332991 PCR products generated from genomic DNA. can trigger cellular senescence in normal fibroblasts (17). Given the known tumor-promoting effects of the SASP (2) as well Senescence induction as the contribution of senescent cells to biological aging (1), it To generate senescent MEFs, fibroblasts cryopreserved two days is logical to examine the effects of these drugs on normal after isolation were thawed, grown in culture for 48 hours, and tissues. However, no study to date has extensively characterized then plated at a density of 4 Â 105 cells per 10-cm plate. Two days the phenotype of CDK4/6 inhibitor–induced senescence in later, cells were treated to induce senescence. For UV-induced normal fibroblasts or determined the effect of these stromal senescence, MEFs were irradiated with two doses of 3 mJ/cm2 UV cells on tumor growth. administered 48 hours apart using a Stratalinker 1800 (Strata- Here, we set out to determine how stromal senescence gene). MEFs were allowed to recover for 48 hours under normal induced by prolonged PD-0332991 treatment influences mel- growth conditions prior to any experimental assessments. For anoma cell proliferation both in vitro and in vivo.Tothisend,we mitomycin C–induced senescence, MEFs were exposed to 10 mg/ extensively compared the phenotypes of fibroblasts triggered to mL mitomycin C (Abcam) for 2.5 hours and then cultured in enter senescence via PD-0332991 treatment to those triggered to growth media for 4 days to establish senescence. For CDK4/6 enter senescence by other, melanoma-relevant signals (i.e., UV inhibitor–induced senescence, MEFs were treated with 4 mmol/L irradiation and DNA-damaging chemotherapy). Then, employ- PD-0332991 (Sigma, 827022-33-3) for 8 days, adding new drug ing a panel of syngeneic murine melanoma cell lines represent- and media on day 4. During in vitro assays, PD-0332991–treated ing major genetic subtypes of human melanoma [BRAF mutant cells were trypisinized, washed with PBS, and then plated in (40%–60% of melanomas), NRAS mutant (15%–30% of mel- normal growth media for at least 24 hours before the start of anomas), and NRAS/BRAF wild-type (<20% of melanomas)], any experiments. Prior to in vivo injections, PD-0332991–treated the ability of these senescent fibroblasts to influence cancer cell cells were trypsinized and thoroughly washed with PBS to remove proliferation was assessed both in vitro and in an immunocom- any residual drug. petent murine model. Our results reveal that CDK4/6 inhibitor– To test