CRIF1−CDK2 Interface Inhibitors - ACS Publications - American

Communication Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

pubs.acs.org/JACS

CRIF1−CDK2 Interface Inhibitors: An Unprecedented Strategy for Modulation of Cell Radiosensitivity Qian Ran,†,§ Yang Xiang,†,§ Preyesh Stephen,‡,§ Chun Wu,† Tang Li,‡ Sheng-Xiang Lin,*,‡ and Zhongjun Li*,† †

Department of Blood Transfusion, Irradiation Biology Laboratory, Xinqiao Hospital, Chongqing, 400037, China Axe Molecular Endocrinology and Nephrology, CHU Research Center and Laval University, Québec City, Québec G1V 4G2, Canada

Downloaded via UNIV OF WINNIPEG on January 18, 2019 at 00:05:09 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

integrity.3 Previously, we reported that CRIF1 was involved in the G1 phase arrest of leukemia cells in the bone marrow microenvironment through interaction with CDK2 and by acting as a CDK2 inhibitor.4,5 Accumulating evidence suggests that key negative regulators of CDKs may represent cancer therapeutic targets, such as WEE1, CHK1, and CHK2. Numerous drugs targeting these key regulators are undergoing clinical tests, and some of them are promising.6 Thus, monitoring the regulatory role of CRIF1 could reveal a novel therapeutic target and a potential sensitizing approach for cancer therapy. Unlike the previous efforts focusing on the ATP binding site of CDKs, our strategy involves an indirect targeting of CDK2 by making use of its natural cell cycle regulator, CRIF1. The CDK2 structure may be probed indirectly, by evaluating the relative abilities of a series of ligand structures to bind on CRIF1 and suppress the CRIF1 negative regulation to CDK2, through which we expect to overcome the nonspecificity barrier faced in previous attempts of CDK targeting. Despite knowledge regarding the negative regulation of CDK2 by CRIF1,5,7 there is no information elucidating the interface between CDK2 and CRIF1, by which a binding site can be detected and used in virtual screening. We used ClusPro,8 ZDOCK,9 and FireDock10 to investigate the interface in the CRIF1−CDK2 complex. Protein−protein docking studies revealed two possible binding sites (1: at a small α helix, 2: at a long α helix) on CRIF1. The binding sites on the small alpha helical region involve Glu88, Glu84, and Glu81. However, as observed in PDB ID 3J7Y, CRIF1 is a component in a large mitoribosome. To reflect this physiological environment, CDK2 was further docked on the large complex (containing the ribosomal part of CRIF1). It revealed that there is not enough space to accommodate CDK2 in the limited available space near this small α-helical region of CRIF1. To account for this physiological constraint, Arg200, Arg214, and Asp210 of CDK2 interacting with His120, Glu116, and Gln112 in the long alpha helical region of CRIF1 were considered to form the binding interface (Figure 1). A virtual screening using 40 678 compounds from the Life Chemicals database at 10 Å around His120 was carried out

ABSTRACT: Cyclin-dependent kinases (CDKs) are historic therapeutic targets implicated in tumorigenic events due to their critical involvement in the cell cycle phase. However, selectivity has proven to be a bottleneck, causing repeated failures. Previously, we reported CR6interacting factor 1 (CRIF1), acting as a cell cycle negative regulator through interaction with CDK2. In the current report, we identified the CRIF1−CDK2 interaction interface by in silico studies and shortlisted interface inhibitors through virtual screening on CRIF1 using 40 678 drug-like compounds. These compounds were tested by cell proliferation assay, and four of these molecules were found to selectively inhibit the proliferation of osteosarcoma (OS) cell lines, but do not affect normal bone mesenchymal stem cells (BMSC). A binding study reveals significant affinities of the inhibitors on CRIF1. More importantly, treatment of the OS cells with a combination of ionizing radiation (IR) and the bestperforming inhibitors remarkably increased IR inhibition potential from 19.9% to 59.6%. This occurred by selectively promoting G2/M arrest and apoptosis related to CDK2 overactivation in OS cells but not in BMSC and was supported by significant CDK2 phosphorylation modifications. Knocking down of CRIF1 by siRNA treatment showed similar effects to the interface inhibitors. Together we substantiate the identification of novel lead molecules, which may provide a new treatment to overcome selectivity issues and enhance the radiosensitivity of tumor cells, opening a conceptually novel strategy of CDK-targeting for different cancer types.

C

yclin-dependent kinases (CDKs) as important cell cycle regulators have been a historic therapeutic target in tumorigenic events. Although this approach has proven challenging, numerous CDK inhibitors have failed in clinical trials primarily from the lack of a therapeutic window. Furthermore, high homology among CDK family members has hindered selective targeting of CDKs. CR6-interacting factor 1 (CRIF1) is a de novo component in the large subunit of the mitoribosome that performs multiple functions in physiological processes in the human body. CRIF1 is proposed to play important roles in embryonic development,1 cellular stress,2 and mitochondrial membrane © XXXX American Chemical Society

Received: September 20, 2018

A

DOI: 10.1021/jacs.8b10207 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

Figure 1. CRIF1−CDK2 interaction model showing CDK2 (colored in yellow), mitoribosomal region (colored in cyan), and CRIF1 (colored in magenta). Interface amino acids are shown as sticks and colored in pink (for CRIF1) and gray (for CDK2) and indicated as a zoomed-in view in the inset figure. An alternative interface proposed but not considered is shown in the orange circle. Figure 2. Inhibition potential of the best ranked molecules. (a1−a4) Chemical structure of each inhibitor molecule and their docked pose on CRIF1 (colored in magenta, surface view). Docked molecule (cyan, stick) and the amino acids involved in hydrogen bonds (green) and the hydrophobic interactions (blue) are shown. A detailed view of the interaction is shown in Figure S1. The IC50 plots of inhibitors in cell proliferation of U2OS and Saos2 are shown in b1 and b2, respectively.

using the program Gold version 5.4.1. These compounds included 28 345 anticancer compounds, 9391 compounds based on fragment-based virtual screening, and 2942 natural product compounds. All compounds passed the Lipinsky’s rule of five, Veber criteria, and dissimilarity evaluation. The best 15 compounds were chosen based on scoring algorithms, Gold, and ChemPLP (Table S1). The binding pattern in the best scored ligand−protein complexes potentially contained multiple interactions dominated by hydrophobic amino acids (Figure 2a). The 15 best compounds were selected for experimental screening. Initially a tetrazolium salt (WST-8) assay was carried out to study the effect of ligands on the proliferation of U2OS and Saos2 cells, cell lines widely employed in solid tumor studies. Human normal bone mesenchymal stem cells (BMSC) were used as a control in the cell studies described here. As shown in Figure 2b1,b2, a dose-dependent inhibition of proliferation was observed with the best IC50s measured in U2OS as 12.7 ± 4.0, 23.2 ± 6.7, 26.4 ± 7.7, and 39.2 ± 7.6 μM for F1142-3225, F0922-0913, F0922-0915, and F1602-0609, respectively. A similar trend, listed in Table S2, was observed in Saos2. Incompetent cell cycle checkpoints, resulting in erratic responses to cellular damage, provide selective growth advantages for human cancers. These checkpoints are potential therapeutic targets, and targeting them may even increase the sensitivity of tumor cells to standard chemotherapy and radiotherapy.11 The 15 best scored small molecules were tested for their effect on cell cycle checkpoints using U2OS and Saos2 cell lines. F1602-0609, F0922-0913, F0922-0915, and F1142-3225 at a concentration of 2 μM induced a significant modification in the cell cycle distribution of U2OS (Figure 3a) and Saos2 cells (Figure S2). The effect of F0922-0913 was the most significant, advancing the cells from G1 phase to S phase by +13%. An increased cell population in S phase was observed in

both cell types with a slightly higher value observed in U2OS cells compared with Saos2. While the distribution of control U2OS cells was 47.7% in the G1 phase, the distributions of inhibitor-treated cells were 36.8%, 34.7%, 38.0%, and 39.8% for F1602-0609, F0922-0913, F0922-0915, and F1142-3225 respectively. The reduction in cells in the G1 phase observed in response to the inhibitors could be explained as redistribution of cells into the S phase, while the G2 phase remained stable. Apparently, there was no significant druginduced effect on cell cycle phases in BMSC cell lines (Figure S3). In addition, we also found that these four inhibitors significantly reduced CDK2 inhibitory phosphorylation (tyrosine 15 (Y15)) levels, while increasing threonine 160 (T160) phosphorylation levels (Figure S4), indicating enhanced CDK2 activity facilitated through CRIF1 inhibition.12 These suggest that the increase in the S phase is caused by an increased activity of CDK2. The above results indicate that the inhibitor molecules discussed here can interfere with the CRIF1−CDK2 interaction, leading to an excess or significant CDK2 activation, promoting the transition of cells from G1 to S phase. However, it should be noted that CDK2 is only one of multiple components affecting the G1 to S transition, and a multitargeting strategy of cancer cells could produce a significantly better therapeutic outcome. Radiation therapy is a general procedure in cancer treatment and works by stimulating DNA damage. Relying on the hypothesis that a radiation-induced cell cycle checkpoint B

DOI: 10.1021/jacs.8b10207 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

G2/M phase after radiation is beneficial to reach higher cellular radiosensitivity.17−20 Following combined treatment with inhibitor and radiation, T160 phosphorylation of CDK2 and the S807 and S811 phosphorylation of retinoblastoma protein (Rb), which is a specific phosphorylation substrate of CDK2,12 were significantly elevated in comparison with the control group (Figure S6). Additionally, the S795 phosphorylation of Rb, which is a specific phosphorylation substrate of CDK4, had no effect. This indicates that the inhibitor treatment in osteosarcoma (OS) cells after IR can selectively and significantly increase CDK2 activity, resulting in G2/M arrest. Intrinsic fluorescence (Figure 4a) titrations have quantitatively demonstrated the significant binding of these molecules

Figure 3. Cell cycle and radiosensitization. Effects of inhibitorinduced cell cycle modulation in U2OS cells before and after IR are shown (a). Effect of drug molecules in radiosensitization estimated by a cell proliferation assay is shown for U2OS (b) and Saos2 (c) cells.

activation provides repair time for sublethal DNA damage,13 it is probable that the abolition of arrests in G1 or G2 may increase the sensitivity of tumor cells to DNA-damaging agents.14,15 Molecules improving radiosensitization effects can find applications in a wide variety of cancer treatments to address failures of 30−50% of radiation therapy at normal tissue tolerance dose of ionizing radiation (IR) in many cancer types.16 U2OS, Saos2, and BMSC cell lines were treated with 6 Gy of IR, inhibitors, or both to investigate the effects of inhibitors on the viability of IR-treated cells. At concentrations up to 2 μM, all inhibitors except F1142-3225 had no effect on the viability of IR− (IR untreated) U2OS and Saos2 cells. In contrast, all inhibitors showed significant radiosensitizing effects in U2OS and Saos2 cells at a concentration of 2 μM. F1142-3225 showed the strongest sensitization, with an increased proliferation inhibition rate from 24.9% to 54.7% in U2OS (Figure 3b) and from 19.7% to 59.6% in Saos2 (Figure 3c). However, BMSC did not show any effect following either treatment with inhibitors alone or with the inhibitors in combination with IR (Figure S5). In addition, the combination of IR and inhibitor treatments significantly induced G2/M arrest in U2OS and Saos2 cells, which were accompanied by a decreased population of cells in G1 and S phases (see IR− and IR+ in Figure 3a and Figure S2). As expected, the combination treatment did not result in a significantly increased population of BMSC in the G2/M phase. Many reports in the literature support that the increase in the distribution ratio of cells in

Figure 4. Binding of drug molecules on CRIF1 and their effect on IRinduced apoptosis in U2OS cells. Fluorescence titration on CRIF1 for representative drug molecules F1142-3225 and F0922-0915 is shown in a. Percentage of apoptotic cells before IR and after IR in U2OS by the inhibitors (b) and siRNA treatment (c) as well as in BMSC by the inhibitors (d) and siRNA treatment (e) is shown.

on CRIF1. Two representative candidates (F0922-0915 and F1142-3225) showed KD values of 5.09 ± 1.71 and 5.13 ± 0.23, respectively. The inhibitor binding for a representative candidate was further validated in isothermal titration calorimetry (for F1142-3225 see Figure S7) and found to have a similar KD of 6.4 μM. Additionally, the effects of CRIF1 disruption by siRNAmediated knockdown in U2OS, Saos2, and BMSC (Figure S8) for radiosensitization (Figure S9) and CDK2 regulatory phosphorylation (Figure S10) were studied, aiming to demonstrate the inhibitors are acting by this disruption of the CRIF1−CDK2 interaction. Cancer cells with CRIF1 knockdown showed increased sensitivity to IR, which could be associated with changes in CDK2 regulatory phosphorylation. The radiosensitization and CDK2 regulatory phosphorylation studies in CRIF1-silenced cells reflect a similar C

DOI: 10.1021/jacs.8b10207 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

■

behavior by CRIF1-targeted compounds, supporting the role of these molecules in disrupting the CRIF1−CDK2 interactions. The G2/M arrest following IR is attributed to the activation of the G2/M checkpoint that is induced by DNA damage. This arrest is only released following complete repair of DNA, whereas the nonrepairable cells undergo apoptosis,21 as observed in Figure 4b−e. The cell apoptosis induced by such combinatorial treatment of inhibitors and IR was studied in U2OS and BMSC cell lines. As shown in Figure 4b,c, a significantly higher level of apoptosis was seen in response to the combined application of the inhibitor molecules and IR or to CRIF1 knockdown and IR in U2OS cells, while the BMSC were not affected (Figure 4d,e). Those differences could be due to the fact that tumor cells are frequently dependent on individual cyclins and CDKs, and hence inhibitors may selectively target cancer cells.22 Similarly, the level of γH2AX in the radiation−inhibitor combination treatment group was much higher than that of the control group (Figure S11), indicating that the increase in apoptosis caused by the inhibitor may be related to impaired DNA damage repair caused by the G1 checkpoint inactivation. Evidently, the combinatorial treatment of inhibitor and IR produced a synergistic effect. These interface inhibitors ultimately appear to be associated with the CRIF1−CDK2 complex producing a significant impact on G1/S checkpoint activation and favoring elevated cell count in G2/M phase, when applied after IR. In theory, inhibitors may block the interaction between CRIF1 and CDK2, thus promoting cell cycle progression at the G1/S transition. Radiation would normally induce events of DNA damage resulting in cell cycle arrests at both G1 and G2 checkpoints followed by a damage repair response. However, as the inhibitors prevent the binding of CRIF1 on CDK2, CDK2 is overactivated and cells cannot arrest in the G1 phase, resulting in cells with DNA damage directly entering the G2 phase. Toxicity is a major concern with CDK-targeting agent, as nonselectivity was the major cause of failure in most trials. Cytotoxicity using a WST-8 assay was performed in BMSC and Vero cells to determine the nontoxic concentrations of each inhibitor molecule. Although 80 μM F1142-3225 had some negative impact on cell viability (Figure S12), this concentration is 6 times higher than the IC50 obtained for F11423225. No toxicity was observed up to 160 μM for any other inhibitor molecule tested, which suggests a large therapeutic window for these drug molecules. Of importance is that the combination of inhibitors and IR selectively kills OS cells through the promotion of G2/M arrests and apoptosis. All inhibitors tested had similar effects on cell cycle arrest and proliferation reduction of OS cells, but not on normal cells. This indicates their ability to distinguish normal and cancer cells, providing lead compounds with potential as future anticancer drugs.

■

Communication

AUTHOR INFORMATION

Corresponding Authors

*[email protected] *[email protected] ORCID

Tang Li: 0000-0001-6622-9471 Sheng-Xiang Lin: 0000-0001-9149-375X Author Contributions §

Q. Ran, Y. Xiang, and P. Stephen contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by Mitacs Canada acceleration project (No. 148361) and National Natural Science Foundation of China (No. 81402634, No. 81472915).

■

REFERENCES

(1) Kwon, M. C.; Koo, B. K.; Moon, J. S.; Kim, Y. Y.; Park, K. C.; Kim, N. S.; Kwon, M. Y.; Kong, M. P.; Yoon, K. J.; Im, S. K.; Ghim, J.; Han, Y. M.; Jang, S. K.; Shong, M.; Kong, Y. Y. Crif1 is a novel transcriptional coactivator of STAT3. EMBO J. 2008, 27 (4), 642−53. (2) Chen, L.; Ran, Q.; Xiang, Y.; Xiang, L.; Chen, L.; Li, F.; Wu, J.; Wu, C.; Li, Z. Co-Activation of PKC-delta by CRIF1Modulates Oxidative Stress in Bone Marrow Multipotent Mesenchymal Stromal Cells after Irradiation by Phosphorylating NRF2 Ser40. Theranostics 2017, 7 (10), 2634−2648. (3) Kim, S. J.; Kwon, M. C.; Ryu, M. J.; Chung, H. K.; Tadi, S.; Kim, Y. K.; Kim, J. M.; Lee, S. H.; Park, J. H.; Kweon, G. R.; Ryu, S. W.; Jo, Y. S.; Lee, C. H.; Hatakeyama, H.; Goto, Y.; Yim, Y. H.; Chung, J.; Kong, Y. Y.; Shong, M. CRIF1 is essential for the synthesis and insertion of oxidative phosphorylation polypeptides in the mammalian mitochondrial membrane. Cell Metab. 2012, 16 (2), 274−83. (4) Zhang, X.; Ran, Q.; Li, Z.; Liu, Y.; Liang, X.; Chen, X. Cell cycle arrest of Jurkat cells by leukemic bone marrow stromal cells: possible mechanisms and involvement of CRIF1. Transplant. Proc. 2011, 43 (7), 2770−2773. (5) Ran, Q.; Hao, P.; Xiao, Y.; Xiang, L.; Ye, X.; Deng, X.; Zhao, J.; Li, Z. CRIF1 interacting with CDK2 regulates bone marrow microenvironment-induced G0/G1 arrest of leukemia cells. PLoS One 2014, 9 (2), No. e85328. (6) Pilie, P. G.; Tang, C.; Mills, G. B.; Yap, T. A., State-of-the-art strategies for targeting the DNA damage response in cancer. Nat. Rev. Clin. Oncol. 2018, DOI: 10.1038/s41571-018-0114-z. (7) Chung, H. K.; Yi, Y. W.; Jung, N. C.; Kim, D.; Suh, J. M.; Kim, H.; Park, K. C.; Song, J. H.; Kim, D. W.; Hwang, E. S.; Yoon, S. H.; Bae, Y. S.; Kim, J. M.; Bae, I.; Shong, M. CR6-interacting factor 1 interacts with Gadd45 family proteins and modulates the cell cycle. J. Biol. Chem. 2003, 278 (30), 28079−88. (8) Kozakov, D.; Hall, D. R.; Xia, B.; Porter, K. A.; Padhorny, D.; Yueh, C.; Beglov, D.; Vajda, S. The ClusPro web server for proteinprotein docking. Nat. Protoc. 2017, 12 (2), 255−278. (9) Pierce, B. G.; Wiehe, K.; Hwang, H.; Kim, B. H.; Vreven, T.; Weng, Z. ZDOCK server: interactive docking prediction of proteinprotein complexes and symmetric multimers. Bioinformatics 2014, 30 (12), 1771−3. (10) Mashiach, E.; Schneidman-Duhovny, D.; Andrusier, N.; Nussinov, R.; Wolfson, H. J. FireDock: a web server for fast interaction refinement in molecular docking. Nucleic Acids Res. 2008, 36, W229−32. (11) Hartwell, L. H.; Kastan, M. B. Cell cycle control and cancer. Science 1994, 266 (5192), 1821−8. (12) Mueller, S.; Huard, J.; Waldow, K.; Huang, X.; D’Alessandro, L. A.; Bohl, S.; Borner, K.; Grimm, D.; Klamt, S.; Klingmuller, U.; Schilling, M. T160-phosphorylated CDK2 defines threshold for HGF

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.8b10207. Experimental procedures, supporting tables, and supporting figures (PDF) D

DOI: 10.1021/jacs.8b10207 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society dependent proliferation in primary hepatocytes. Mol. Syst. Biol. 2015, 11 (3), 795. (13) Branzei, D.; Foiani, M. Regulation of DNA repair throughout the cell cycle. Nat. Rev. Mol. Cell Biol. 2008, 9 (4), 297−308. (14) Kim, J. S.; Pirnia, F.; Choi, Y. H.; Nguyen, P. M.; Knepper, B.; Tsokos, M.; Schulte, T. W.; Birrer, M. J.; Blagosklonny, M. V.; Schaefer, O.; Mushinski, J. F.; Trepel, J. B. Lovastatin induces apoptosis in a primitive neuroectodermal tumor cell line in association with RB down-regulation and loss of the G1 checkpoint. Oncogene 2000, 19 (52), 6082−90. (15) Reinhardt, H. C.; Aslanian, A. S.; Lees, J. A.; Yaffe, M. B. p53deficient cells rely on ATM- and ATR-mediated checkpoint signaling through the p38MAPK/MK2 pathway for survival after DNA damage. Cancer Cell 2007, 11 (2), 175−89. (16) Le Chevalier, T.; Arriagada, R.; Quoix, E.; Ruffie, P.; Martin, M.; Tarayre, M.; Lacombe-Terrier, M. J.; Douillard, J. Y.; Laplanche, A. Radiotherapy alone versus combined chemotherapy and radiotherapy in nonresectable non-small-cell lung cancer: first analysis of a randomized trial in 353 patients. J. Natl. Cancer Inst. 1991, 83 (6), 417−23. (17) Ho, W. S.; Sizdahkhani, S.; Hao, S.; Song, H.; Seldomridge, A.; Tandle, A.; Maric, D.; Kramp, T.; Lu, R.; Heiss, J. D.; Camphausen, K.; Gilbert, M. R.; Zhuang, Z.; Park, D. M. LB-100, a novel Protein Phosphatase 2A (PP2A) inhibitor, sensitizes malignant meningioma cells to the therapeutic effects of radiation. Cancer Lett. 2018, 415, 217−226. (18) Hao, S.; Song, H.; Zhang, W.; Seldomridge, A.; Jung, J.; Giles, A. J.; Hutchinson, M. K.; Cao, X.; Colwell, N.; Lita, A.; Larion, M.; Maric, D.; Abu-Asab, M.; Quezado, M.; Kramp, T.; Camphausen, K.; Zhuang, Z.; Gilbert, M. R.; Park, D. M., Protein phosphatase 2A inhibition enhances radiation sensitivity and reduces tumor growth in chordoma. Neuro Oncol. 2018, 20, 799. (19) Jin, Y.; Lyu, Y.; Tang, X.; Zhang, Y.; Chen, J.; Zheng, D.; Liang, Y. Lupeol enhances radiosensitivity of human hepatocellular carcinoma cell line SMMC-7721 in vitro and in vivo. Int. J. Radiat. Biol. 2015, 91 (2), 202−8. (20) Wang, Y.; Mei, H.; Shao, Q.; Wang, J.; Lin, Z. Association of ribosomal protein S6 kinase 1 with cellular radiosensitivity of nonsmall lung cancer. Int. J. Radiat. Biol. 2017, 93 (6), 581−589. (21) Martin, L. M.; Marples, B.; Coffey, M.; Lawler, M.; Lynch, T. H.; Hollywood, D.; Marignol, L. DNA mismatch repair and the DNA damage response to ionizing radiation: making sense of apparently conflicting data. Cancer Treat. Rev. 2010, 36 (7), 518−27. (22) Otto, T.; Sicinski, P. Cell cycle proteins as promising targets in cancer therapy. Nat. Rev. Cancer 2017, 17 (2), 93−115.

E

DOI: 10.1021/jacs.8b10207 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX