Natural product derivative Gossypolone inhibits Musashi family of RNA-binding proteins
The Musashi (MSI) family of RNA-binding proteins is best known for the role in post-transcriptional regulation of target mRNAs. Elevated MSI1 levels in a variety of human cancer are associated with up-regulation of Notch/Wnt signaling. MSI1 binds to and negatively regulates translation of Numb and APC (adenomatous polyposis coli), negative regulators of Notch and Wnt signaling respectively.
Previously, we have shown that the natural product (−)-gossypol as the first known small molecule inhibitor of MSI1 that down-regulates Notch/Wnt signaling and inhibits tumor xenograft growth in vivo. Using a fluorescence polarization (FP) competition assay, we identified gossypolone (Gn) with a > 20-fold increase in Ki value compared to (−)-gossypol. We validated Gn binding to MSI1 using surface plasmon resonance, nuclear magnetic resonance, and cellular thermal shift assay, and tested the effects of Gn on colon cancer cells and colon cancer DLD-1 xenografts in nude mice.
In colon cancer cells, Gn reduced Notch/Wnt signaling and induced apoptosis. Compared to (−)-gossypol, the same concentration of Gn is less active in all the cell assays tested. To increase Gn bioavailability, we used PEGylated liposomes in our in vivo studies. Gn-lip via tail vein injection inhibited the growth of human colon cancer DLD-1 xenografts in nude mice, as compared to the untreated control (P < 0.01, n = 10).
Our data suggest that PEGylation improved the bioavailability of Gn as well as achieved tumor-targeted delivery and controlled release of Gn, which enhanced its overall biocompatibility and drug efficacy in vivo. This provides proof of concept for the development of Gn-lip as a molecular therapy for colon cancer with MSI1/MSI2 overexpression.
KeywordsGossypolone Musashi RNA-binding protein Colon cancer Liposomes
Adenomatous polyposis coli
Cellular thermal shift assay
Drug loading content
Drug loading efficiency
Dynamic Light Scattering
(3, 4-Dimethoxyphenyl) methanimine gossypol
Nuclear magnetic resonance
RNA Recognition Motifs
Surface plasmon resonance
Transmission electron microscopy
The expression of the RNA-binding protein Musashi-1 (MSI1) is elevated in a variety of human cancers, including glioblastoma, breast, colon and lung cancers [1, 2, 3, 4, 5, 6, 7, 8, 9, 10], with higher levels corresponding to poor prognosis [3, 4, 5, 10, 11, 12]. Msi1 was first identified in Drosophila where it plays a role in neural development and asymmetric cell division in the adult sensory organ . Subsequently, Msi1 homologs were identified in other species, with higher levels in stem and undifferentiated cells [1, 2, 14, 15, 16, 17]. Musashi-1 typically plays a role in post-transcriptional regulation of target mRNAs [18, 19, 20, 21, 22]. Up-regulation of MSI1 in cancers appears to associate with elevated Notch/Wnt signaling, as MSI1 targets Numb [22, 23] and APC (adenomatous polyposis coli)  are negative regulators of Notch and Wnt signaling, respectively [24, 25]. CDKN1A (P21), a negative regulator of cell cycle progression, is also a direct MSI1 target . In all three cases, MSI1 blocks target mRNA translation. Knocking down MSI1 using siRNA , miRNA  and a small molecule inhibitor  led to decreased xenograft tumor growth. Taken together, these results point to MSI1 as a potential therapeutic target.
Our previous study identified (−)-gossypol as a small molecule inhibitor of MSI1 that reduced cancer cell proliferation and xenograft growth . More recent screening in our lab using an FP assay revealed more potent and/or specific inhibitors of MSI1. One inhibitor with a Ki of 12 ± 2 nM against full length MSI1 was gossypolone (Gn), it had a higher affinity than (−)-gossypol (Ki = 476 ± 273 nM) . Gn also showed similar affinity towards Musashi-2 (MSI2) in FP assay (Ki = 7.0 ± 0.3 nM against full length MSI2). MS12 is another Musashi family member that plays both redundant and independent roles as MSI1 in neural stem cells [28, 29]. In cancer, MSI2 expression is elevated in hematologic malignancies [30, 31, 32, 33, 34, 35, 36], colorectal adenocarcinomas , lung , pancreatic cancers [39, 40, 41], and glioblastoma . MSI1 and MSI2 share sequence and structure similarity, especially their N-terminal RNA recognition motifs (RRMs). The residues that recognize r(GUAGU) are highly conserved between MSI1 and MSI2 . Thus, Gn can potentially be used as a MSI1/2 dual inhibitor.
Gn is a major metabolite of gossypol , and is oxidized in the liver by P450 enzyme . Gn shares similar biological activities as gossypol [46, 47, 48, 49, 50, 51, 52], including as an inhibitor of Bcl-2 family with a Ki of 0.28 μM toward Bcl-xL . However, in colon cancer cell assays, the same concentration of Gn was less potent than (−)-gossypol .
To address this problem, we introduce a new liposome-based Gn nanocarrier. Liposomes have long been used as nanocarriers for targeted cancer therapy and have demonstrated biocompatibility and controlled drug release in previous studies [53, 54, 55, 56]. Particularly, compared with unmodified liposomes, some PEGylated liposomes were reported to be less entrapped by reticuloendothelial cells and lead to enhanced drug delivery to solid tumors in vivo [57, 58, 59]. In the present study, PEGylated liposomes were used to improve the bioavailability of Gn as well as to achieve tumor-targeted delivery and controlled release of Gn, which enhances its overall biocompatibility and drug efficacy in vivo.
Cell culture and reagents
Gossypolone (Gn) was prepared as previously described . (3, 4-Dimethoxyphenyl)methanimine gossypol (MP-Gr) was synthesized from gossypol . The Gn and MP-Gr powder were dissolved in DMSO at 20 mM as stock solutions. L-α-phosphatidylcholine (EPC) and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000] (PEG-DSPE) were purchased from Avati Polar Lipids, Inc. (Alabama, USA) . DiR (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotricarbocyanine iodide) was purchased from Invitrogen (Carlsbad, CA).
Cell growth, MTT, colony formation, western blot analysis, Caspase-3 activation assay, RT-PCR and quantitative real-time PCR were carried out according to our previous publications [27, 62, 63, 64, 65, 66]. Protein expression and purification, FP competition assay, SPR, NMR, and Wnt luciferase reporter assay were carried out as previously described . The primer sequences, the primary and the secondary antibodies used were from Lan et al. . Live cell imaging was carried out using EVOS FL Auto Cell Imaging System (Invitrogen, Thermo Fisher Scientific) and images were cropped and processed using ImageJ (NIH).
For all cell based studies, the DMSO concentration was 0.1% except where indicated below (for CETSA).
The AutoDock4.2.6 program  was used for docking calculations. The three-dimensional structure of Musashi1’s RBD1 in complex with RNA (PDB: 2RS2) was used to dock the gossypolone compound at the MSI1 RBD1 - RNA interface. A grid box of size 40*44*56 Å with 0.375 Å spacing centered around residue F23 was used for docking. A total of 200 docking runs were carried out using the Lamarckian genetic algorithm. The docked conformation with lowest energy was selected as the final predicted binding mode.
Cellular thermal shift assay (CETSA)
CETSA was carried out according to Molina et al. . For Gn dose CETSA, the HCT-116 β/W cell lysates with different concentrations of Gn were incubated for 30 min and heated individually at 52 °C for 3 min (StepOnePlus™ Real-Time PCR System, Applied Biosystems/Life Technologies) followed by cooling for 3 min at 25 °C. The soluble fractions were analyzed by western blot. The concentration of DMSO in each sample is 3.3%. Musashi-1 antibody used for CETSA was anti-MSI1 (01–1041, Millipore, Billerica, MA). Western band intensities were measured using Image Studio Ver 4.0 (LI-COR Bioscience, Lincoln, NE), and normalized to α-Tubulin.
Preparation and characterization of gossypolone-encapsulated liposomes (Gn-lip)
Gn-lip was formed using a mixture of Gn, EPC, PEG-DSPE, and cholesterol in chloroform, at a molar ratio of 30/85/6/9. The solution was dried under vacuum to form a thin film of Gn/carrier mixture, which was then dissolved in DPBS to produce Gn-encapsulated liposomes. Blank liposomes were prepared similarly without the addition of Gn. To prepare the samples for TEM image, both Gn-lip and blank liposomes were diluted in DI water, respectively. The suspensions were applied to a grid and negatively stained by 4% uranyl acetate. Images of liposomes were acquired using FEI Tecnai G2 Polara 200 kV TEM (FEI Company, OR, USA). The size distribution and zeta potential of liposomes in DI water were measured at 25̊ C using a Malvern instrument (Nano-ZS90, Malvern, UK). The size stability of Gn-lip for 3 months was investigated at 4 °C. The drug loading efficiency (DLE%) and drug loading content (DLC%) of Gn were determined using filtration method. Gn-lip solution was filtered using an ultra centrifugal filter unit (MWCO 3000 Da, Amicon®, Merck KGaA, Germany). The concentration of free drug in the filtrate was determined using a UV-vis spectrophotometer. The DLE% and DLC% of Gn were calculated as follows: DLE% = (weight of loaded Gn ÷ total weight of input Gn) × 100%; DLC% = (weight of loaded Gn ÷ total weight of Gn-lip) × 100%.
The viabilities of HCT-116 and DLD-1 cells in the presence of free Gn or Gn-lip were determined using MTT-based assay, as described previously.
Biodistribution of DiR-loaded liposomes in tumor-bearing mice
In vivo drug efficacy of Gn in DLD-1 tumor-bearing nude mice
The in vivo experiments were carried out with 5 to 6-week-old female athymic NCr-nu/nu nude mice purchased from the Harlan laboratory (Indianapolis, IN). After alcohol preparation of the skin, mice were inoculated subcutaneously with 200 μL DLD-1 cell suspension (1 × 106 cells) in plain DMEM on both flanks using a sterile 23-gauge needle. When tumors reached 40 mm3 on average, the mice were randomized into 2 groups. Group 1 (10 mice, 20 tumors) was given vehicle as the control; group 2 (5 mice, 10 tumors) was given 10 mg/kg Gn-lip. Gn-lip was administrated intravenously 2 times weekly for 3.5 weeks. Tumor size and body weight of each mouse were measured twice a week, and tumor volumes were determined as a × b2/2, in which a and b represent the longest and shortest diameter of the tumors, respectively. All animal experiments were carried out according to the protocol approved by the Institutional Committee for the Use and Care of Animals of University of Kansas.
Using Prism 5.0 software (GraphPad Prism), one-way ANOVA and t-Test were used to analyze the in vitro data, two-way ANOVA was used to analyze the in vivo data. A threshold of P < 0.05 was defined as statistically significant.
Gossypolone disrupts the Musashi-numb RNA interaction
Gossypolone directly binds to the RBD1 of MSI1 protein
To confirm the direct binding of Gn to MSI1, we carried out additional assays. First, we tested Gn in a SPR-based binding assay. In SPR, GB1-tagged MSI1-RBD1 was immobilized on the sensor chip and the level of response increases with increasing amount of material bound to the surface. As shown in Fig. 1b, at 5 μM, the response was 50, while at 10 μM, the response was 200. The SPR assay showed that Gn binds to MSI1-RBD1 in a dose-dependent manner.
The binding of Gn to MSI1-RBD1 was also confirmed using NMR (Fig. 1c). Results of NMR titrations of 15N MSI1-RBD1 with Gn showed that the RNA-binding residues (K93, F23, and W29) were primarily affected by Gn (Fig. 1c). The backbone amide peaks of K93, F23, and W29; including the side chain peak of W29, showed changes in peak positions as well as decreased peak intensities with increasing amounts of Gn, whereas most non-RNA binding residues remained unaffected. These NMR results suggested that these residues are involved in the binding of Gn to the RNA-binding pocket of MSI1-RBD1. Using computational docking, we then built a model of Gn bound to the RNA-binding pocket of MSI1-RBD1 (Fig. 1d); this model is in agreement with the NMR observation that these particular residues (K93, F23 and W29) are responsible for the interaction of MSI1-RBD1 with Gn.
Gn targets MSI1 in cells
To test drug-target engagement in cells, we used the CETSA to determine the thermal stability of target protein MSI1. When a protein is stabilized with addition of a ligand, the bound proteins can stay in solution whereas unbound proteins denature and precipitate with increasing temperatures . The advantage of CETSA is that one can evaluate the compounds in a cellular context, thus allowing us to identify the compounds with poor bioavailability that otherwise have high affinity in biochemical assays. Fig. 1e showed the concentration-dependent target engagement of Gn with MSI1, that is, more MSI1 protein is stabilized at higher Gn concentration.
Gn inhibits cell proliferation, induces apoptosis and autophagy in colon cancer cell lines
Additional file 2: Video 1. (AVI 2064 kb)
Additional file 3: Video 2. (AVI 2645 kb)
Gn down-regulates Notch/Wnt signaling through MSI1
Compare with (−)-gossypol, Gn was less effective in downregulating Notch/Wnt signaling through MSI1 in cells. For example, Wnt target gene AXIN2 mRNA levels were 50% (10 μM Gn treatment) versus 20% (10 μM (−)-gossypol treatment) compared to DMSO control (set as 1) in HCT-116 cells, and 80% versus 40% in DLD-1 cells (Fig. 4b, ). In our biophysical assays, Gn showed a better affinity to MSI1 (Fig. 1a, ). We thus sought to introduce a carrier for delivering Gn in vivo.
Characterizations of Gn-lip
The viabilities of cells in the presence of free Gn or Gn-lip were similar, for both HCT-116 and DLD-1 cells (Fig. 5c). The IC50 value of free Gn and Gn-lip were 12 μM and 13 μM for DLD-1 cell, respectively, and were 8 μM and 11 μM for HCT-116 cell, respectively. Although the IC50 values of Gn-lip was a little higher than that of free Gn, cytotoxicity of Gn was not significantly compromised by encapsulation. The increased IC50 values of Gn-lip might be due to the sustained release of Gn from the liposomes.
In vivo drug efficacy of Gn in DLD-1 tumor-bearing nude mice
In this study, we sought to identify additional small molecule inhibitors of Musashi-1(MSI1). Such inhibitors can downregulate Notch/Wnt signaling and block cell cycle progression through MSI1. To this end, we used FP-based screening and identified Gn as a potential MSI1 inhibitor. We further confirmed Gn binding to MSI1 using SPR and using NMR, identified the amino acids in the RNA-Recognition Motif 1 that were involved in the binding. Using colon cancer cell lines, we showed that Gn inhibited cell growth, induced autophagy, inhibited Notch/Wnt signaling in these cells and led to apoptotic cell death. However, compared to 10 μM (−)-gossypol , the same concentration of Gn was less active in cell assays. This result could be due to poor water-solubility of Gn. Therefore, we used a liposomal carrier to deliver Gn in animals. The liposome could efficiently increase the apparent solubility of Gn in water. We showed that Gn loaded liposomes induced apoptosis, inhibited tumor growth and Notch/Wnt signaling in DLD-1 xenograft model. Our study identified a new target for Gn and provided a new delivery method for this poorly bioavailable compound.
MSI1 is an RNA-binding protein that promotes cell proliferation and survival through Notch/Wnt signaling (Fig. 7d). Inhibiting MSI1 is a promising therapeutic strategy for preventing cancer cell proliferation and progression. Here we identified Gn as a more potent inhibitor of MSI1 compared to (−)-gossypol  in our binding assays. Our NMR studies showed that Gn bound to the same residues as the cognate RNA. The NMR peaks of the RNA-binding residues (K93, F23 and W29) showed the most changes upon titration of Gn (Fig. 1c). This suggests that these residues are involved in the binding to Gn, consistent with our docked model. The majority of non-RNA binding residues were unaffected when titrated with Gn, suggesting that the interaction between Gn and MSI1-RBD1 tends to be more localized near the RNA binding pocket as compared to the binding event between (−)-gossypol and MSI1-RBD1 . This might explain the improved potency of Gn compared to (−)-gossypol. Additional X-ray crystallography studies will be helpful in determining the high resolution structure of the complex between MSI1-RBD1 and Gn.
Based on the sequence identity of RBDs/RRMs of MSI1 and MSI2, we tested the binding of Gn towards MSI2 and showed that Gn can disrupt the binding of MSI2 to Numb RNA. Like MSI1, MSI2 is also a member of the Musashi family of RNA-binding protein and shares similar roles as MSI1 in stem cells [28, 29]. In colorectal cancer initiation and maintenance, a recent study demonstrated the functional redundancy between MSI1 and MSI2 . Thus using a MSI1/MSI2 dual inhibitor such as Gn in patients with MSI overexpression suggests an improved therapeutic outcome.
The introduction of PEGylated liposomes improved the dispersion of Gn in aqueous environment, thus making it possible to produce injectable drug solutions, which is essential for a better bioavailability of Gn. As shown in the results of MTT assay (Fig. 5c), the encapsulation of Gn using liposomes did not compromise the cytotoxicity of Gn in vitro. Finally, the Gn-lip exhibited a significant tumor inhibition efficacy and a low systemic toxicity in the mice. Our work is important in that our study provides a proof-of-concept to develop the Gn-lip as a novel molecular therapy for colon cancer with MSI overexpression.
The limitation of our current inhibitors is that they are not specific to MSI1/MSI2, because Gn and the previously reported (−)-gossypol  are both Bcl-2 family inhibitors as well [46, 47, 48, 49, 50, 51, 52, 62, 63, 76, 79]. With Gn treatment, we saw apoptosis/autophagy induction via Bcl-2 and cell proliferation inhibition via Wnt and/or Notch signaling pathways. Our goal is to develop potent and specific MSI1/MSI2 inhibitors, and ultimately move these new inhibitors into clinical applications in the treatment of cancers with MSI overexpression. Towards this goal, our future efforts will focus on utilizing computer modeling and medicinal chemistry for identifying new chemical scaffolds that selectively inhibit MSI1/2.
Gn was identified as a MSI1/2 duo inhibitor in this study. It disrupted binding of MSI1 to its target mRNAs by binding to the RBD1 of MSI1. Gn inhibited colon cancer cell growth, induced autophagy, down-regulated Notch/Wnt signaling and led to apoptotic cell death. The introduction of tumor-targeted liposomes significantly improved the bioavailability of Gn, meanwhile maintaining its drug efficacy. Gn-lip has promising antitumor effects and biocompatibility in vivo, warranting further study to determine its suitability for cancer treatment.
We are grateful to OpenEye Scientific Software (Santa Fe, NM) for providing an academic license for the use of OMEGA, ROCS, MolProp, and QuacPac. We thank the NCI/DTP Open Chemical Repository (http://dtp.cancer.gov) libraries.
This study was supported in part by National Institutes of Health grants R01 CA178831 and CA191785 (to LX, JA, KN); AI074856 (to R.N.D.); NIH COBRE-PSF at KU (P30 GM110761) Pilot Project, Department of Defense Breast Cancer Research Program Breakthrough Level II grant BC151845, and Kansas Bioscience Authority Rising Star Award (to LX). No involvement of the funding bodies in any part of the study.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
LL, HL, and LX contributed to conception and design of the study. LL, HL, AS, CA, JY, SL, RM and PG acquired and analyzed the data. LL, HL, AS, RM, KL, RG, JK, RND, JA, KN and LX interpreted data. LL and HL drafted the manuscript; LL, HL, AS, JK, RND, JA, KN and LX revised it critically for important intellectual content. All authors gave final approval of the version to be published. LL and HL contributed equally to this work.
Ethics approval and consent to participate
None of the cell lines used in this study require ethics approval for their use. All animal experiments were carried out according to the protocol approved by the Institutional Committee for the Use and Care of Animals of University of Kansas.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 9.Yokota N, Mainprize TG, Taylor MD, Kohata T, Loreto M, Ueda S, Dura W, Grajkowska W, Kuo JS, Rutka JT. Identification of differentially expressed and developmentally regulated genes in medulloblastoma using suppression subtraction hybridization. Oncogene. 2004;23(19):3444–53.CrossRefPubMedGoogle Scholar
- 27.Lan L, Appelman C, Smith AR, Yu J, Larsen S, Marquez RT, Liu H, Wu X, Gao P, Roy A, et al. Natural product (−)-gossypol inhibits colon cancer cell growth by targeting RNA-binding protein Musashi-1. Mol Oncol. 2015;Google Scholar
- 28.Sakakibara S, Nakamura Y, Yoshida T, Shibata S, Koike M, Takano H, Ueda S, Uchiyama Y, Noda T, Okano H. RNA-binding protein Musashi family: roles for CNS stem cells and a subpopulation of ependymal cells revealed by targeted disruption and antisense ablation. Proc Natl Acad Sci U S A. 2002;99(23):15194–9.CrossRefPubMedPubMedCentralGoogle Scholar
- 37.Shan W, Ning L, Maryam Y, Angela N-D, Fan L, Kimberly P, Shilpa R, Gerard M, Yarden K, Brian DG, et al. Transformation of the intestinal epithelium by the MSI2 RNA-binding protein. Nat Commun. 2015;6Google Scholar
- 38.Kudinov AE, Deneka A, Nikonova AS, Beck TN, Ahn YH, Liu X, Martinez CF, Schultz FA, Reynolds S, Yang DH, et al. Musashi-2 (MSI2) supports TGF-beta signaling and inhibits claudins to promote non-small cell lung cancer (NSCLC) metastasis. Proc Natl Acad Sci U S A. 2016;113(25):6955–60.CrossRefPubMedPubMedCentralGoogle Scholar
- 39.Sheng W, Dong M, Chen C, Wang Z, Li Y, Wang K, Li Y, Zhou J. Cooperation of Musashi-2, numb, MDM2, and P53 in drug resistance and malignant biology of pancreatic cancer. FASEB J. 2017;Google Scholar
- 49.Shaomeng Wang DY, Liang Xu: Small molecule antagonists of Bcl-2 family proteins (US patent 8163805 B2). In., vol. US8163805 B2.Google Scholar
- 69.Liu H, Li K, Xu L, Wu DC. Bilayered near-infrared fluorescent nanoparticles based on low molecular weight PEI for tumor-targeted in vivo imaging. J Nanopart Res. 2014:16(12).Google Scholar
- 76.Paoluzzi L, Gonen M, Gardner JR, Mastrella J, Yang D, Holmlund J, Sorensen M, Leopold L, Manova K, Marcucci G, et al. Targeting Bcl-2 family members with the BH3 mimetic AT-101 markedly enhances the therapeutic effects of chemotherapeutic agents in in vitro and in vivo models of B-cell lymphoma. Blood. 2008;111(11):5350–8.CrossRefPubMedGoogle Scholar
- 77.Li K, Liu H, Gao W, Chen M, Zeng Y, Liu J, Xu L, Wu D. Mulberry-like dual-drug complicated nanocarriers assembled with apogossypolone amphiphilic starch micelles and doxorubicin hyaluronic acid nanoparticles for tumor combination and targeted therapy. Biomaterials. 2015;39:131–44.CrossRefPubMedGoogle 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. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.