Combined inhibition of Aurora A and p21-activated kinase 1 as a new treatment strategy in breast cancer
- 932 Downloads
The serine-threonine kinases Aurora A (AURKA) and p21-activated kinase 1 (PAK1) are frequently overexpressed in breast tumors, with overexpression promoting aggressive breast cancer phenotypes and poor clinical outcomes. Besides the well-defined roles of these proteins in control of cell division, proliferation, and invasion, both kinases support MAPK kinase pathway activation and can contribute to endocrine resistance by phosphorylating estrogen receptor alpha (ERα). PAK1 directly phosphorylates AURKA and its functional partners, suggesting potential value of inhibiting both kinases activity in tumors overexpressing PAK1 and/or AURKA. Here, for the first time, we evaluated the effect of combining the AURKA inhibitor alisertib and the PAK inhibitor FRAX1036 in preclinical models of breast cancer.
Combination of alisertib and FRAX1036 was evaluated in a panel of 13 human breast tumor cell lines and BT474 xenograft model, with assessment of the cell cycle by FACS, and signaling changes by immunohistochemistry and Western blot. Additionally, we performed in silico analysis to identify markers of response to alisertib and FRAX1036.
Pharmacological inhibition of AURKA and PAK1 synergistically decreased survival of multiple tumor cell lines, showing particular effectiveness in luminal and HER2-enriched models, and inhibited growth and ERα-driven signaling in a BT474 xenograft model. In silico analysis suggested cell lines with dependence on AURKA are most likely to be sensitive to PAK1 inhibition.
Dual targeting of AURKA and PAK1 may be a promising therapeutic strategy for treatment of breast cancer, with a particular effectiveness in luminal and HER2-enriched tumor subtypes.
KeywordsBreast cancer Targeted therapy Aurora A AURKA p21-Activated kinase 1 PAK1
The serine-threonine kinases Aurora A (AURKA) and p21-activated kinase 1 (PAK1) are frequently overexpressed in breast tumors and associated with aggressive tumor phenotypes and poor clinical outcomes [1, 2, 3, 4]. AURKA controls centrosome maturation, timing of mitotic entry, assembly of the bipolar spindle, and chromosome alignment in metaphase . AURKA overexpression occurs in over 90% of breast carcinomas [3, 5]. Increased AURKA activity overrides the mitotic spindle assembly checkpoint, inducing resistance to anti-mitotic agents , while inhibition of AURKA increases the activity of microtubule inhibitors [7, 8]. In interphase, overexpressed AURKA stabilizes C-MYC  and stimulates the PI3K/AKT/mTOR pathway, promoting chemotherapeutic resistance .
Increased PAK1 activity is also common in breast cancer, typically due to amplification of the PAK1 gene (30% of breast carcinomas) . Like AURKA, PAK1 stimulates multiple pro-oncogenic pathways, including AKT, C-MYC, and β-catenin [11, 12], promoting proliferation, motility, and invasion [11, 13]. PAK1-dependent upregulation of cyclin D1 is important for G1/S transition . Although AURKA and PAK1 function within overlapping but distinct signaling pathways, PAK1 is capable of AURKA activation: both directly, by phosphorylating serine S342 and threonine T288 in the activation loop , and indirectly, by phosphorylation of the AURKA-activating protein partners LIMK1 and ARPC1b [15, 16, 17].
Of relevance to breast cancer, both AURKA and PAK1 phosphorylate estrogen receptor alpha (ERα) (on serines S118 (PAK1) and S305 (PAK1 and AURKA)) supporting ligand-independent transcription of ERα-dependent genes promoting proliferation, invasion, and endocrine resistance [4, 18]. The AURKA inhibitor alisertib synergized with tamoxifen in preclinical studies  and showed activity in patients with hormone receptor-positive (HR +) breast cancer . PAK1 inhibition has been reported to abrogate tamoxifen resistance .
Based on these activities of AURKA and PAK1, we hypothesized that combined inhibition of both could have synergistic anti-tumor effects in breast cancer [13, 21]. In this study, we explored the consequences of combination treatment with the alisertib and FRAX1036, a highly selective inhibitor of PAK1 and two paralogous group 1 PAK kinases, PAK2 and PAK3 .
Materials and methods
See “Supplementary Materials” for additional details on cell lines, cell culture, antibodies for Western blot and IHC, drug formulations for xenograft experiments, and statistical analysis.
Tumor cell lines, media, and reagents
Human breast cancer cell lines from the American Type Culture Collection were cultured in standard conditions. We confirmed negative mycoplasma testing and STR profile for each cell line. Alisertib was purchased from MedChem Express (Monmouth Junction, NJ). FRAX1036 was synthesized by AK and WW .
Cell viability assay
Cells were grown on 96-well plates for 24 h before treatment with drug(s) or vehicle. Cell viability was measured by CellTiterGlo assay (Promega, Madison, WI) after 72 h of treatment. Each drug concentration was evaluated in triplicate, with ≥ 3 biological repetitions. We determined synergy by Chou-Talalay method .
Protein lysates were prepared with RIPA lysis buffer (Thermo Fisher, Waltham, MA) containing protease/phosphatase inhibitor (Roche Diagnostic, Indianapolis, IN). Each blot was repeated with ≥ 3 preparations of lysates. Signal intensity was quantified by NIH ImageJ Software, or Odyssey imager software (Li-Cor Bioscience, Lincoln, NE), normalized to vinculin or GAPDH, and compared by two-tailed t test and one-way ANOVA.
All animal experiments were approved by the FCCC Institutional Animal Care and Use Committee. NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice from the FCCC breeding colony were maintained under pathogen-free conditions. Estrogen pellets were implanted subcutaneously into 6- to 8-week-old mice as described ; simultaneously, mice were injected in mammary fat pads with 107 BT474 cells (N = 45 mice). Treatment consisted of alisertib (15 mg/kg twice a day), FRAX1036 (20 mg/kg daily) or combination of drugs; control group received vehicle solution twice a day; all agents were administered by oral gavage.
To assess short-term signaling, after tumor volume reached 600 mm3, mice were treated for 3 days with vehicle, alisertib, FRAX1036, or combination of drugs, then euthanized and tumors were frozen for Western blots. To assess long-term responses, once tumors reached 150 mm3, mice were treated for 21 days, then euthanized, and tumors collected for analysis.
IHC was performed according to standard protocols. Results were quantitated with Aperio ePathology (Leica Biosystems, Buffalo Grove, IL) and analyzed by Mann–Whitney and Kruskal–Wallis tests.
Cell cycle analysis by fluorescence-activated cell sorting (FACS)
Non-synchronized growing cells were fixed with ethanol at 24 and 72 h after treatment with drug(s) or vehicle, then mixed with propidium iodide solution (BD Pharmingen, San Diego, CA) before FACS (BD Biosciences, San Diego, CA); data were analyzed by one-way ANOVA.
In silico analysis of expression and zGARP scores for the genes of interest and correlation with FRAX1036 and alisertib activity in vitro
Methods for deriving z-score normalized Gene Activity Ranking Profile (zGARP) score have been described in detail [25, 26]. zGARP scores for AURKA, CCND1, MYC, PAK1-3, and TFF1 were extracted from . For PAK1-3, we selected the smallest of the zGARP scores for each cell line. RNAseq fragments per kilobase million (FPKM) values were extracted from [25, 27, 28, 29]. For each gene, ranks were calculated across cell lines indicated in Results in each dataset. Ranks for gene/cell line pairs were averaged across the sets of RNAseq data. Pearson correlation coefficients and p values were calculated using GraphPad Prism for the drug IC50 versus zGARP score.
Alisertib and FRAX1036 synergize predominantly in luminal and HER2-enriched breast cancer cell lines
Considering the maximum tolerated doses of alisertib and FRAX1036 in vivo [30, 31] and clinically relevant doses of alisertib in humans [32, 33], we selected a fixed molar ratio of FRAX1036 to alisertib of 1:1.5 for assessment in cell lines (Fig. 1, S1). Synergy between alisertib and FRAX1036 was detected in four of five luminal cell lines, particularly at lower drug concentrations (Fig. 1, S2); activity of alisertib and FRAX1036 combination exceeded efficacy of fulvestrant in these cell lines (Fig. S3). Alisertib and FRAX1036 also synergized in 3 of 4 HR-/HER2 + tumor cell lines, but only in 1 of 4 TNBC cell lines (Fig. 1, S2).
Alisertib and FRAX1036 change cell cycle compartmentalization and decrease activity of ERα and MYC in tumor cell lines
Notably, the drug combination significantly inhibited phosphorylation of ERα(S118) and ERα(S305) in both cell lines (Fig. 2e, f). Alisertib and FRAX1036 also inhibited phosphorylation of ERα(S118) in both cell lines, although to a lesser degree than the combination. Expression of C-MYC was reduced more by the alisertib/FRAX1036 combination than by single agents in both lines (Fig. 2g, h).
Activity of combined versus monoagent alisertib and FRAX1036 in BT474 tumor xenografts
By regression analysis, reduction in the BT474 tumor growth rate compared to vehicle was significant in mice treated with alisertib or alisertib/FRAX1036 (p < 0.001), but not in FRAX1036-treated mice (Fig. 3a, b). Tumor control with combination therapy was better than with monotherapy (p < 0.001 combination versus FRAX1036, p = 0.003 combination versus alisertib, p value for synergy p = 0.014). Although FRAX1036 produced initial responses, they were lost after 10 days (Fig. 3b). Considering the difference of FRAX1036 activity in vivo and in vitro, tumor microenvironment likely plays a strong role in resistance mechanisms, based on emerging understanding of PAK function . After 21 days, tumor volume averaged 930 mm3 in vehicle-treated mice, 826 mm3 in FRAX1036-treated mice, 188 mm3 in alisertib-treated mice, and 55 mm3 in mice treated with the combination (Fig. 3c). Final tumor volumes differed significantly between the alisertib or the combination versus vehicle (p < 0.05); further, tumor volume with the combination treatment was smaller comparing to monagent alisertib (p = 0.004). Importantly, only the alisertib/FRAX1036 combination reduced tumor volume compared to the initial volume (~ 150 mm3) (Fig. 3c), with histopathological analysis indicating one case of near complete response (residual tumor volume of 16 mm3) and one case of complete response in treated mice. All therapies were well tolerated, with weight of drug- and vehicle-treated mice not significantly differing (Fig. 3d).
Immunohistopathological (IHC) assessment of xenografts
To better characterize treatment-induced cell cycle arrest, we evaluated cyclin D1, and the mitotic cyclin B1. Alisertib significantly reduced cyclin B1 expression (Fig. 4a, e), consistent with the requirement of AURKA for G2/M transition . FRAX1036 significantly decreased cyclin D1 expression (Fig. 4a, f), reflecting the essential role of PAK1 in induction of this gene . Combination therapy reduced expression of both cyclins to a much greater extent than with either single agent, suggesting quiescent or moribund cells (Fig. 4a, e, f).
C-MYC  and trefoil factor 1 (TFF1)  are canonical downstream effectors of ERα. After 3 weeks of treatment, all tumors treated with combination therapy had very low to undetectable expression of C-MYC, which was significantly different from the control or single agents (Fig. 4a, g). In contrast, FRAX1036 numerically increased C-MYC levels versus all other treatment groups, suggesting a rebound effect and potential escape mechanism. Combination therapy significantly decreased TFF1 expression, with a more modest reduction seen in single agent alisertib-treated tumors (Fig. 4a, h). Expression of the apoptotic marker cleaved caspase-3 (CC3) was increased in tumors treated with the combination of alisertib and FRAX1036 compared to control vehicle-treated cells (Fig. 4a, i). However, the number of CC3 positive cells was small, potentially indicating alternative mechanisms of cell death are also involved, such as necrosis, mitotic catastrophe, or senescence. In sum, these results indicated functional activity of combined alisertib/FRAX1036 in xenografts, reflected in decreased tumor volume, reduced cellularity, suppressed Ki67, altered cell cycle checkpoints, and depressed ERα signaling.
Alisertib and FRAX1036 inhibit PAK1 and ERα signaling following transient treatment of BT474 tumors in vivo
Differential response to alisertib and FRAX1036 correlates with AURKA and MYC zGARP scores
A database of gene essentiality in tumor cell lines has been determined by shRNA knockdown and characterized by z-score normalized Gene Activity Ranking Profile (zGARP) score [25, 26]. The zGARP score reflects changes in gene expression and cellular proliferation after treatment of tumor cells with shRNAs . Response to a targeted agent may correlate with gene essentiality even if it does not correlate with gene expression . We correlated zGARP scores for AURKA, PAK1-3, CCND1, C-MYC, and TFF1 with response to alisertib and FRAX1036 in our cell line experiments (Fig. 6b, c). In ERα + lines, the strongest predictor of response to alisertib was the strength of dependence on C-MYC, a relationship not observed in ERα- lines (Fig. 6b, c). ERα + cell lines highly sensitive to shRNA C-MYC knock down required higher concentrations of alisertib for growth inhibition, compared to less dependent cell lines. Weaker, but suggestive relationships with alisertib response in ERα + lines were found for dependence on CCND1 and the alisertib target, AURKA (Fig. 6b). Similar analysis performed for FRAX1036 (Fig. 6b, c) revealed correlation with dependence on PAK2 and PAK3, both of which are FRAX1036 targets along with PAK1, as well as weaker correlation with dependence on TFF1 (Fig. 6b, c). Intriguingly, the strongest interrelationship found was positive correlation of sensitivity to FRAX1036 with dependence on AURKA in ERα + cell models, suggesting that cells with strong requirements of AURKA might be more sensitive to PAK inhibition (Fig. 6b, c). Because zGARP scores were developed to predict individual drug sensitivity , we did not analyze correlation of zGARP scores with the efficacy of two drugs in combination, which is a limitation of our analysis.
Our results indicate that combined inhibition of AURKA and PAK1 is of potential value for the treatment of breast cancer, with greatest efficacy seen in luminal HR + and HER2 + subtypes in vitro. This could be explained by the interaction of AURKA and PAK1 with ERα (phosphorylation leading to ligand-independent activation), and with HER2 [4, 18]. AURKA promotes epithelial-mesenchymal transition and stem cell properties of ER + breast tumors in a mechanism involving overexpression of HER2 , while PAK1 is an essential mediator of HER2 signaling in mammary tumors dependent on this protein . Correspondingly, our analysis of the METABRIC dataset showed significantly worse overall survival in patients with co-alterations of AURKA and PAK1, 2, or 3, with the greatest differences noted in patients with luminal A (HR +/HER2-) and B (HR +/HER2 +) tumors (Fig. S7).
The potency of the combination in luminal cell lines is likely due at least in part to the decreased phosphorylation of ERα at both the S305 and S118 residues, seen both in vitro and in xenograft experiments. Greater disruption of cell cycle control with the combination is also likely to contribute. In the BT474 xenograft model, the combination effectively inhibited signaling proteins linked to G1 and G2/M cell cycle control and ERα-activation, including cyclin B1, TFF1, C-MYC, and cyclin D1. This was consistent with the FACS analysis showing the combination arrested BT474 cells in both G1 and G2/M phases. One limitation of the present work is that we did not use cell sorting to separate mouse stromal cells from human breast cancer cells in these experiments; this may have led to somewhat diminished apparent effect of the drugs on phosphorylation of ERα.
We have expected a synergistic effect of alisertib and FRAX1036 on cell cycle and suppression of tumor growth because of more effective suppression of AURKA in the settings of PAK1 inhibition . However, alisertib treatment also decreased phospho-PAK1/2/3, possibly via inhibition of phospho-AKT that can activate PAK1 [10, 38]. Notably, in silico analysis showed strong positive correlation of sensitivity to FRAX1036 with dependence on AURKA in ERα + tumors, providing a rationale to combine AURKA and PAK1-inhibitors.
The combination effectively inhibited expression of the transcription factor and proto-oncogene C-MYC, a protein frequently overexpressed in breast tumors, and implicated in poor clinical outcomes [36, 40]. Despite intense investigations, no effective strategies exist to target C-MYC. C-MYC upregulates the expression of AURKA , while AURKA activity protects C-MYC from degradation . AURKA signals through C-MYC to induce telomerase, supporting tumor immortalization . In kinase-independent functions, AURKA interacts with heterogeneous nuclear ribonucleoprotein K to activate C-MYC promoter, enhancing breast cancer stem cell phenotypes . In parallel, C-MYC is as a downstream target of PAK1: PAK1 inhibition decreases C-MYC expression and signaling [12, 45]. Significant downregulation of C-MYC after combined treatment with AURKA and PAK1 inhibitors observed in our study is an exciting and clinically important finding. Our analysis of correlations with zGARP scores identified dependence on C-MYC as the strongest predictor of response to alisertib in ERα + lines. Luminal cell lines sensitive to C-MYC knock down required higher concentrations of alisertib for growth inhibition. While cell lines highly dependent on C-MYC have more compensatory mechanisms to escape alisertib-induced C-MYC downregulation, co-treatment with PAK1 inhibitors may abrogate these mechanisms, allowing response to lower doses of alisertib.
Together, our results provide evidence that dual inhibition of AURKA and PAK1 is of value in breast cancer. Enhanced anti-tumor activity of this combination is based on multiple mechanisms, including enhanced inhibition of phosphorylation of AURKA, PAK1, and ERα, as well as decreased expression of cell cycle proteins and C-MYC. Although resistance developed in vivo to single agent FRAX1036, addition of FRAX1036 to alisertib conferred significant advantages and lead to cases of complete or near complete tumor response, consistent with the concept that combination targeted therapy is beneficial because of synergistic anti-tumor effect and prevention of the selection of drug-resistant subclones during therapy .
One limitation of our study is that we examined the effects of the combination in a single in vivo model—further studies in PDXs and breast tumor cell organoids will be useful to confirm and extend our findings. In our study, as proof of concept, we used a prototype PAK1/2/3 inhibitor FRAX1036. Newer, more potent and selective PAK1 inhibitors now in development [47, 48] should be evaluated in combination with AURKA inhibitors in further studies. Alisertib was shown to be active in preclinical studies and early clinical trials in combination with microtubule inhibitors [7, 8, 49] or fulvestrant . Given the findings of our study, evaluation of the combination of AURKA and PAK1 inhibitors together with other targeted or chemotherapeutic agents, such as tamoxifen, aromatase inhibitors, HER2-inhibitors, or taxanes, would be of interest. As genomic characterization of breast cancers becomes more advanced, understanding of the landscape of oncogenic drivers may help inform the optimal use of these valuable therapeutics.
We thank Troy Schedin and Veronica Wessells (University of Colorado Denver) for assistance with IHC, and Brian Egleston (Fox Chase Cancer Center) and Patrick Blatchford (University of Colorado Denver) for help with statistical analysis.
The study and authors were funded by American Cancer Society Seed Award (to ES), Robert F. and Patricia Young Connor Endowed Chair in Young Women’s Breast Cancer Research Award (to VFB), National Institute of Health R01 CA142928 (to JC), National Institute of Health R01 DK108195 (to EAG), a subsidy of the Russian Government to support the Program of Competitive Growth of Kazan Federal University (to IGS), and National Cancer Institute Core Grant P30 CA06927 (to Fox Chase Cancer Center, FCCC).
Compliance with ethical standards
Conflicts of interest
All authors declare that they have no conflict of interest.
This study has results of animal experiments. All animal experiments have been performed in compliance with the institutional regulations (including the Institutional Animal Care and Use Committee, and Institutional Biosafety Committee regulations), and with the current US laws. All applicable international, national, and institutional guidelines for the care and use of animals were followed.
Research involving human and animal participants
This study does not contain any research with human participants performed by any of the authors.
- 4.Zheng XQ, Guo JP, Yang H, Kanai M, He LL, Li YY, Koomen JM, Minton S, Gao M, Ren XB et al (2013) Aurora-A is a determinant of tamoxifen sensitivity through phosphorylation of ERalpha in breast cancer. Oncogene 42:4985Google Scholar
- 6.Zhou N, Singh K, Mir MC, Parker Y, Lindner D, Dreicer R, Ecsedy JA, Zhang Z, Teh BT, Almasan A et al (2013) The investigational Aurora kinase A inhibitor MLN8237 induces defects in cell viability and cell-cycle progression in malignant bladder cancer cells in vitro and in vivo. Clin Cancer Res 19(7):1717–1728PubMedPubMedCentralGoogle Scholar
- 7.Jensen JS, Omarsdottir S, Thorsteinsdottir JB, Ogmundsdottir HM, Olafsdottir ES (2012) Synergistic cytotoxic effect of the microtubule inhibitor marchantin A from Marchantia polymorpha and the Aurora kinase inhibitor MLN8237 on breast cancer cells in vitro. Planta Med 78(5):448–454CrossRefPubMedGoogle Scholar
- 8.Kozyreva VK, Kiseleva AA, Ice RJ, Jones BC, Loskutov YV, Matalkah F, Smolkin MB, Marinak K, Livengood RH, Salkeni MA et al (2016) Combination of eribulin and aurora A inhibitor MLN8237 prevents metastatic colonization and induces cytotoxic autophagy in breast cancer. Mol Cancer Ther 15(8):1809–1822CrossRefPubMedPubMedCentralGoogle Scholar
- 19.Melichar BAA, Havel L, Lockhart A et al (2013) Phase I/II study of investigational Aurora A kinase inhibitor MLN8237 (alisertib): updated phase II results in patients with small cell lung cancer, non-small cell lung cancer, breast cancer, head and neck squamous cell carcinoma, and gastroesophageal cancer. J Clin Oncol 31:605Google Scholar
- 21.Tentler JJ, Ionkina AA, Tan AC, Newton TP, Pitts TM, Glogowska MJ, Kabos P, Sartorius CA, Sullivan KD, Espinosa JM et al (2015) p53 Family members regulate phenotypic response to Aurora Kinase A inhibition in triple-negative breast cancer. Mol Cancer Ther 14(5):1117–1129CrossRefPubMedPubMedCentralGoogle Scholar
- 24.DeRose YS, Gligorich KM, Wang G, Georgelas A, Bowman P, Courdy SJ, Welm AL, Welm BE (2013) Patient-derived models of human breast cancer: protocols for in vitro and in vivo applications in tumor biology and translational medicine. Curr Protoc Pharmacol 60:14–23Google Scholar
- 31.Manfredi MG, Ecsedy JA, Chakravarty A, Silverman L, Zhang M, Hoar KM, Stroud SG, Chen W, Shinde V, Huck JJ et al (2011) Characterization of Alisertib (MLN8237), an investigational small-molecule inhibitor of aurora A kinase using novel in vivo pharmacodynamic assays. Clin Cancer Res 17(24):7614–7624CrossRefPubMedGoogle Scholar
- 32.Yang JJ, Li Y, Chakravarty A, Lu C, Xia CQ, Chen S, Pusalkar S, Zhang M, Ecsedy J, Manfredi MG et al (2014) Preclinical drug metabolism and pharmacokinetics, and prediction of human pharmacokinetics and efficacious dose of the investigational Aurora A kinase inhibitor alisertib (MLN8237). Drug Metab Lett 7(2):96–104CrossRefPubMedGoogle Scholar
- 33.Kelly KR, Shea TC, Goy A, Berdeja JG, Reeder CB, McDonagh KT, Zhou X, Danaee H, Liu H, Ecsedy JA et al (2014) Phase I study of MLN8237–investigational Aurora A kinase inhibitor–in relapsed/refractory multiple myeloma, non-Hodgkin lymphoma and chronic lymphocytic leukemia. Invest New Drugs 32(3):489–499CrossRefPubMedGoogle Scholar
- 35.Wang F, Li H, Yan XG, Zhou ZW, Yi ZG, He ZX, Pan ST, Yang YX, Wang ZZ, Zhang X et al (2015) Alisertib induces cell cycle arrest and autophagy and suppresses epithelial-to-mesenchymal transition involving PI3 K/Akt/mTOR and sirtuin 1-mediated signaling pathways in human pancreatic cancer cells. Drug Des Devel Ther 9:575–601PubMedPubMedCentralGoogle Scholar
- 36.Carroll JS, Swarbrick A, Musgrove EA, Sutherland RL (2002) Mechanisms of growth arrest by c-myc antisense oligonucleotides in MCF-7 breast cancer cells: implications for the antiproliferative effects of antiestrogens. Can Res 62(11):3126–3131Google Scholar
- 39.D’Assoro AB, Liu T, Quatraro C, Amato A, Opyrchal M, Leontovich A, Ikeda Y, Ohmine S, Lingle W, Suman V et al (2014) The mitotic kinase Aurora–a promotes distant metastases by inducing epithelial-to-mesenchymal transition in ERalpha(+) breast cancer cells. Oncogene 33(5):599–610CrossRefPubMedGoogle Scholar
- 47.Ndubaku CO, Crawford JJ, Drobnick J, Aliagas I, Campbell D, Dong P, Dornan LM, Duron S, Epler J, Gazzard L et al (2015) Design of Selective PAK1 Inhibitor G-5555: improving properties by employing an unorthodox low-pK a polar moiety. ACS Med Chem Lett 6(12):1241–1246CrossRefPubMedPubMedCentralGoogle Scholar
- 49.Falchook G, Coleman RL, Roszak A, Behbakht K, Matulonis U, Ray-Coquard I, Sawrycki P, Duska LR, Tew W, Ghamande S et al (2019) Alisertib in combination with weekly paclitaxel in patients with advanced breast cancer or recurrent ovarian cancer: a randomized clinical trial. JAMA Oncol 5(1):e183773CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.