Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Chem Commun (Camb). Author manuscript; available in PMC 2017 April 14.
Published in final edited form as:
PMCID: PMC5271582

Discovery of a 19F MRI sensitive salinomycin derivative with high cytotoxicity towards cancer cells


Salinomycin is a promising anti-cancer agent which selectively targets cancer stem cells. To improve its potency and selectivity, an analogs library of salinomycin was generated by site-specific modification and CuAAc derivatization. Through a cytotoxicity analysis of the library, a fluorinated analog with high potency, selectivity, and 19F MRI sensitivity was discovered as a novel theranostic agent.

Graphical Abstract

An external file that holds a picture, illustration, etc.
Object name is nihms771407u1.jpg

The first 19F MRI sensitive theranostic agent was discovered through site-specific modification of salinomycin and click library derivatization.

Salinomycin, a polyether ionophore antibiotic isolated from Streptomyces albus, has been commercially used in veterinary medicine for decades. In 2009, salinomycin was identified as a selective breast cancer stem cells inhibitor with activity 100-fold higher than paclitaxel.1 Since then, its cytotoxicity against a variety of cancer stem cells and cancer cells has been discovered.2 Although the detailed mechanism of these effects is still unclear,3 salinomycin is already in clinical trials for a panel of cancers. Therefore, generation of highly potent, highly specific salinomycin analogs is of great interest. Further, it is desired if the analogs can also be used to probe the modes of action of salinomycin.

As a complex natural product with 18 chiral centres, 5 rings, and multiple reactive groups, the site-specific modification of salinomycin is very challenging. Currently, the modification strategies are limited to esterification or amidation of its carboxylic group, acylation of its hydroxyl groups, and hydrogenation of its double bond (Fig. 1).4 Recently, Strand’s group developed a strategy for site-specific acylation of the hydroxyl groups in salinomycin and found that C20-acylated analogs with the least bulky substituents had the highest potency against cancer cells.4f,5 The X-Ray crystal and molecular modelling of salinomycin show that the C20 hydroxyl group is not involved in ion chelation, and its acylation actually poses steric hindrance for ion chelation.6 We envisioned that the inversion of the C20 configuration could relieve the steric hindrance and therefore enhance ion chelation and potency. In this way, a highly valuable conjugation site is also available for targeted delivery of salinomycin without compromising its ion chelation and potency.7 To this end, inversion of the C20 hydroxyl group with an azido group is preferred because it could provide easy access to various salinomycin analogs and conjugates using the Cu-catalyzed azide-alkyne cycloaddition (CuAAC) reaction8 under mild conditions (Fig. 1). Herein, we report a convenient strategy for the C20-specific modification of salinomycin with an azido group and the consequent generation of a library of analogs through the CuAAC reaction, as well as a detailed cytotoxicity assay of the library.

Fig. 1
Methods for modification of salinomycin 1.

After esterification of salinomycin 1 with TMS(CH2)2OH using a reversible strategy for carboxylic group masking (Scheme 1),4f the selective inversion of the C20 configuration with an azido nucleophile was investigated. Because the allylic C20 hydroxyl group is more reactive than C9 and C28 ones,4f,6 selective tosylation of 2 provided C20 tosylate S1 in high yield (See supporting Information). However, many attempts at nucleophilic substitution of the tosylate group with an azido group resulted in the decomposition of S1. Fortunately, an azido group was successfully introduced into 2 at C20 through the Mitsunobu reaction with diphenylphosphoryl azide (DPPA) as a nucleophile. Azide 3 was obtained with a 73% yield and no C9- or C28-substituted side product was isolated. Unmasking the carboxylic group with tetrabutylammonium fluoride (TBAF) provided the key intermediate 4 on a multi-gram scale with a 68% yield. To generate a diverse library of salinomycin analogs using the CuAAC reaction, a panel of 26 alkynes were selected, including 9 phenyl-containing, 6 nitrogen-containing, 7 hydroxyl-containing, and 4 ether-containing alkynes. As expected, each alkyne was conveniently coupled with azide 4, respectively, in the presence of CuSO4 and sodium ascorbate under mild conditions to give triazols 5a-z with good yields.

Scheme 1
Synthesis of salinomycin-based click library.

To illustrate the structures of triazols 5a-z, single-crystal X-ray structures of 5f-Na+ and 5f-K+ were obtained (Fig. 2). It was found that 5f can adopt a suitable conformation to chelate ions, Na+ or K+, with its oxygen atoms. As expected, the Mitsunobu reaction took place exclusively at C20 with the inversion of the configuration, and, consequently, the CuAAC reaction constructed a trizol sidearm which stretched out of the ion chelation pocket. In this way, both the steric effect and chelation perturbation of ion chelation were successfully avoided.

Fig. 2
Single-crystal X-ray structures of 5f-Na+ and 5f-K+.

The cytotoxicity of these salinomycin analogs 5a-z together with salinomycin 1 and azide 4 was then evaluated in 4T1 murine breast cancer cells using a MTT assay (Table 1). Surprisingly, azide 4 showed a 27-fold loss of potency as compared to salinomycin. Out of the library, all 9 phenyl-containing triazols 5a-i displayed higher or comparable potency to salinomycin. These EC50 values compare very favourable to bulky substituents without a hetero atom, which display EC50 of 2.29–2.46 μM for 5b-d. With the exception of 3-pyridine substituted 5k, the EC50 of 5 nitrogen-containing triazols 5j-o were much higher than that of salinomycin. In these cases, the strong chelation ability of nitrogen may pose a perturbation on ion chelation.4f The same trend was observed in the 7 hydroxyl-containing triazols 5p-v. It is noteworthy that no significant cytotoxicity was observed for triazols 5t-v with ethylene glycol units which are known to chelate ions. Among the 4 ether-containing triazols 5w-z, perfluoro-tert-butyl ether 5w turned out to be the most potent one with an EC50 of 1.52 μM, which is over 2-fold more potent than salinomycin. Notably, perfluoro-tert-butyl is a bulky group ideal for 19F magnetic resonance imaging (19F MRI).9 Based on these observations, it is clear that bulky substituents on the C20 trizol is preferred for achieving high potency. In contrast, C20 hydroxyl acyl analogs with the least bulky substituent exhibited the highest potency.4f Therefore, inversion of the C20 configuration indeed improves the potency.

Table 1
EC50 of salinomycin 1, azide 4 and trizol 5a-z on murine 4T1 cell.

Based on the initial assay, triazols 5b, 5d, 5k, 5w and salinomycin were selected for further assay on human cells (Table 2). Besides human hepatic cells (L02), a panel of cancer cells were selected, including human glioblastoma cells (U87), cervical cancer cells (Hela), epithelial colorectal adenocarcinoma cells (Caco2), and breast cancer cells (MCF-7). Compared to salinomycin, these triazols displayed dramatically lower cytotoxicity against normal L02 cells and significantly higher cytotoxicity towards cancer cells. For example, 5d, 5k and 5w displayed over 2-fold higher cytotoxicity towards U87, and 5b exhibited 2.9-fold higher cytotoxicity towards MCF-7 than that of salinomycin. It is noteworthy that, compared to salinomycin, 5w showed 29.5-fold higher cytotoxicity towards Caco2 with an IC50 of 0.44 μM.

Table 2
IC50 of salinomycin 1, analogs 5b, 5d, 5k and 5w on a panel of cells.

To evaluate the clinical potential of triazols 5b, 5d, 5k, and 5w, the selectivity index (SI) was calculated with salinomycin as a comparison (Table 3). It is an indication of a drug with efficacy against cancer cells when SI > 1.0. For selected cancer cells, salinomycin showed a low SI except for in Hela cells. In contrast, these triazols had a significantly higher SI than salinomycin. For example, 5b displayed the highest SI of 8.85 for Hela and 5d displayed the highest SI of 5.89 for Caco2. Among these triazols, 5w is very promising because it had high potency and SI for Hela and Caco2 cells.

Table 3
Selectivity index (SI) of salinomycin 1, triazols 5b, 5d, 5k and 5w on a panel of cancer cells over normal cell L02.

With 9 symmetrical fluorine atoms, fluorinated trizol 5w is also a valuable 19F NMR/MRI probe for better understanding the mechanism of salinomycin’s selective cytotoxicity on CSCs. In recent years, 19F MRI has been widely used in tracking targets of interest10 and monitoring biological reactions.11 Compared to other imaging technologies, 19F MRI is able to provide high-contrast in vivo images without an endogenous background and ionizing radiation. Trizol 5w gives a strong singlet 19F NMR peak at −70.50 ppm from its nine symmetrical fluorines. Regardless of the size and chelation pattern of the ions, little chemical shift change was observed when 5w was chelated with a panel of ions (Fig. 3a), which can dramatically simplify the downstream 19F MRI study. 19F MRI relaxation experiments indicated that 5w has appropriate relaxation times for 19F MRI with a longitudinal relaxation time T1 of 843 ms and a transverse relaxation time T2 of 445 ms. Then, a 19F MRI phantom experiment of 5w was carried out on an array of its solutions in DMSO. With a short scan time of 128 seconds, 5w was imaged by 19F MRI at a concentration as low as 1.1 mM (or 10 mM in 19F concentration, Fig. 3b). Because of its high cytotoxicity, selectivity, and 19F MRI sensitivity, 5w is a promising 19F MRI traceable mechanism probe for CSCs and a drug candidate for imaging-guided cancer therapy.

Fig. 3
(a) 19F NMR of 5w and its ion complexes. (b) 19F MRI phantom images of 5w (Upper: coloured; Lower: black-white).

In summary, a fluorinated salinomycin analog with high potency and selectivity against cancer cells and high 19F MRI sensitivity was discovered. The Mitsunobu reaction enables the site-specific modification of salinomycin with high efficacy, while the CuAAC reaction provides a convenient way to generate a salinomycin analog library under mild conditions. Importantly, inversion of the C20 configuration of salinomycin can relieve steric hindrance, enhance ion chelation, improve potency, and provide a valuable conjugation site for targeted delivery. The results presented here could serve as a starting point for the discovery of clinically useful salinomycin-based anti-cancer agents, 19F NMR/MRI-guided mechanism study of salinomycin’s effects on CSCs, and 19F MRI-guided cancer therapy. Using the fluorinated salinomycin analog as a mechanism probe to study its modes of action on CSCs is currently in progress and will be published in due course.

Supplementary Material



We are thankful for financial support from the National Natural Science Foundation of China (21372181 and 21572168), the Key Laboratory of Synthetic Chemistry of Natural Substances (Shanghai Institute of Organic Chemistry), and the National Institutes of Health (R00CA153929).


Electronic Supplementary Information (ESI) available: Synthesis of library compounds, copies of 1H NMR, 13C NMR, 19F NMR, IR and HRMS of compounds, and single-crystal X-ray diffractograms of 5f-Na+ (CCDC 1448688) and 5f-K+ (CCDC 1449564). See DOI: 10.1039/x0xx00000x

Notes and references

1. Gupta PB, Onder TT, Jiang G, Tao K, Kuperwasser C, Weinberg RA, Lander ES. Cell. 2009;138:645. [PMC free article] [PubMed]
2. (a) Tang QL, Zhao ZQ, Li JC, Liang Y, Yin JQ, Zou CY, Xie XB, Zeng YX, Shen JN, Kang T, Wang J. Cancer Lett. 2011;311:113. [PubMed](b) Schenk M, Aykut B, Teske C, Geise NA, Weitz J, Welsch T. Cancer Lett. 2015;358:161. [PubMed]
3. (a) Huczyński A. Chem Biol Drug Des. 2012;79:235. [PubMed](b) Huczyński A. Bioorg Med Chem Lett. 2012;22:7002. [PubMed]
4. (a) Huczyński A, Janczak J, Stefańska J, Antoszczak M, Brzezinski B. Bioorg Med Chem Lett. 2012;22:4697. [PubMed](b) Huczyński A, Janczak J, Antoszczak M, Wietrzyk J, Maj E, Brzezinski B. Bioorg Med Chem Lett. 2012;22:7146. [PubMed](c) Antoszczak M, Maj E, Stefańska J, Wietrzyk J, Janczak J, Brzezinski B, Huczyński A. Bioorg Med Chem Lett. 2014;24:1724. [PubMed](d) Antoszczak M, Popiel K, Stefańska J, Wietrzyk J, Maj E, Janczak J, Michalska G, Brzezinski B, Huczyński A. Eur J Med Chem. 2014;76:435. [PubMed](e) Antoszczak M, Maj E, Napiorkówska A, Stefańska J, Augustynowicz-Kopec E, Wietrzyk J, Janczak J, Brzezinski B, Huczyński A. Molecules. 2014;19:19435. [PubMed](f) Borgström B, Huang X, Pošta M, Hegardt C, Oredsson S, Strand D. Chem Commun. 2013;49:9944. [PubMed](g) Sprott KML, Choi H, Fang F, Shan M, Lazarova TI, Li L, Siddiqui MA, Larouche-Gauthier R, Lemire A. 20140371285 A1. US. 2013
5. (a) Miyazaki Y, Shibuya M, Sugawara H, Kawaguchi O, Hirsoe C. J Antibiot. 1974;27:814. [PubMed](b) Miyazaki Y, Kinashi H, Otake N, Mitani M, Yamanishi T. Agric Biol Chem. 1976;40:1633.(c) Hammann P, Raether W, Vertesy L. J Antibiot. 1993;46:523. [PubMed]
6. (a) Paulus EF, Kurz M, Matter H, Vértesy L. J Am Chem Soc. 1998;120:8209.(b) Huang X, Borgström B, Månsson L, Persson L, Oredsson S, Hegardt C, Strand D. ACS Chem Biol. 2014;9:1587. [PubMed]
7. (a) Zhao P, Dong S, Bhattacharyya J, Chen M. Mol Pharm. 2014;11:2703. [PubMed](b) Huczyński A, Antoszczak M, Kleczewska N, Lewandowska M, Maj E, Stefańska J, Wietrzyk J, Janczak J, Celewicz L. Eur J Med Chem. 2015;93:33. [PubMed](c) Antoszczak M, Sobusiak M, Maj E, Wietrzyk J, Huczyński A. Bioorg Med Chem Lett. 2015;25:3511. [PubMed]
8. Meldal M, Tornøe CW. Chem Rev. 2008;108:2952. [PubMed]
9. (a) Jiang ZX, Liu X, Jeong EK, Yu YB. Angew Chem Int Ed. 2009;48:4755. [PMC free article] [PubMed](b) Jiang ZX, Yu YB. J Org Chem. 2010;75:2044. [PubMed]
10. (a) Ahrens ET, Flores R, Xu H, Morel PA. Nat Biotechnol. 2005;23:983. [PubMed](b) Vivian D, Cheng K, Khuranan S, Xu S, Kriel EH, Dawson PA, Raufman JP, Polli JE. Mol Pharmaceutics. 2014;11:1575. [PMC free article] [PubMed]
11. (a) Mizukami S, Takikawa R, Sugihara F, Hori Y, Tochio H, Walchli M, Shirakawa M, Kikuchi K. J Am Chem Soc. 2008;130:794. [PubMed](b) Bruemmer KJ, Merrikhihaghi S, Lollar CT, Morris SN, Bauer JH, Lippert AR. Chem Commun. 2014;50:12311. [PubMed]