Rapid Access to Ironomycin Derivatives by Click Chemistry.

Salinomycin, a natural carboxylic polyether ionophore, shows a very interesting spectrum of biological activities, including selective toxicity toward cancer stem cells (CSCs). Recently, we have developed a C20-propargylamine derivative of salinomycin (ironomycin) that exhibits more potent activity in vivo and greater selectivity against breast CSCs compared to the parent natural product. Since ironomycin contains a terminal alkyne motif, it stands out as being an ideal candidate for further functionalization. Using copper-catalyzed azide-alkyne cycloaddition (CuAAC), we synthesized a series of 1,2,3-triazole analogs of ironomycin in good overall yields. The in vitro screening of these derivatives against a well-established model of breast CSCs (HMLER CD24low/CD44high) and its corresponding epithelial counterpart (HMLER CD24high/CD44low) revealed four new products characterized by higher potency and improved selectivity toward CSCs compared to the reference compound ironomycin. The present study highlights the therapeutic potential of a new class of semisynthetic salinomycin derivatives for targeting selectively the CSC niche and highlights ironomycin as a promising starting material for the development of new anticancer drug candidates.

Introduction

Tumors are composed of heterogeneous cell populations that comprise cancer cells and stromal cells, which includes cancer-associated fibroblasts (CAFs), endothelial cells, and immune cells.[1] The paradigm of cancer stem cells (CSCs) defines the existence of rare populations of cancer cells capable of self-renewal, tumor initiation, and maintenance. CSCs are also typically refractory to conventional treatments, are prone to metastatic dissemination, and can cause cancer relapse.[2−8] Thus, the development of new lead structures that can selectively target CSCs is of great interest and represents a potential improvement in the arsenal of modern anticancer treatment strategies.

Salinomycin (Sal, Scheme ), a natural polyether ionophore isolated from Streptomyces albus, has been identified in a landmark study by Gupta et al. as a small molecule that preferentially targets breast CSCs.[9]Sal reduced the proportion of CSCs by >100-fold relative to that of paclitaxel, a widely used drug in breast cancer therapy.[9] Importantly, Sal has also been shown in clinical pilot studies to have therapeutic potential in cancer patients by inducing partial regression of therapy-resistant and heavily pretreated tumors.[10] Mechanistically, many pathways and molecular complexes that confer resistance and survival of CSCs have been identified to be significantly modulated after Sal treatment. This includes Wnt/β-catenin, PI3K/Akt/mTOR, K-Ras, as well as Hedgehog signaling.[11−15] With respect to the very promising biological activity of Sal, a number of its semisynthetic analog structures have been developed,[16−30] some of which exhibit relevant activity against cancer cells with stemlike characteristics.[25,30−34]

Scheme 1

Synthetic Access to the 1,2,3-Triazole Analogs of Ironomycin

Reagents and conditions: (a) activated MnO2, CH2Cl2, RT; (b) propargylamine, AcOH, MeOH, RT, then CeCl3·7H2O, NaBH3CN, MeOH, RT; (c) RN3, CuI, DIPEA, ACN, RT; (d) TMSN3, CuI, MeOH/DMF, RT.

Recently, we have demonstrated that the introduction of the propargylamine moiety to the C20 position of Sal (ironomycin, Scheme ) improves significantly the ability of the parent structure to selectively target breast CSCs, both in vitro and in vivo.[33] Compared to chemically unmodified ionophore, ironomycin showed ∼10-times higher potency against mesenchymal HMLER CD24low/CD44high cells over those deprived of CSC properties, with an IC50 value of about 100 nM.[33] The mechanism of action (MoA) of ironomycin and Sal underlying their selectivity toward CSCs involves accumulation of these compounds in lysosomes and iron retention in this organelle, which eventually leads to the production of reactive oxygen species (ROS) and cell death consistent with ferroptosis.[33] Since cells in the mesenchymal state are addicted to iron,[35−37] this cell state exhibits a heightened vulnerability to ironomycin and Sal.

As ironomycin contains a terminal alkyne moiety, it is an ideal substrate for further chemical transformations to generate products with enhanced anti-CSC activity, which could be of great value for the development of new anticancer lead drugs. In the present work, we describe the rapid access to potent 1,2,3-triazole derivatives of ironomycin obtained by means of click chemistry (Scheme ). Next, we evaluated the products against a well-established model of breast CSCs, together with the corresponding cancer cells deprived of stemlike characteristics (Table ). Based on the results, we identified four analogs (2–4 and 8) that have the ability to preferentially kill CSCs with remarkably low IC50 values that compete favorably with ironomycin. All products were also less toxic against normal breast cell line MCF10A compared to ironomycin (Table ).

Table 1

Antiproliferative Activity (IC50 in μM) with Standard Deviation and Values of the Selectivity Index (SI) of 1,2,3-Triazole Derivatives of Ironomycin, Measured at 72 h in HMLER CD24low/CD44high, HMLER CD24high/CD44low, MCF7, and MCF10A Cellsa

	HMLER CD24low/CD44high	HMLER CD24high/CD44low	SI (HMLER)	MCF7	MCF10A
ironomycin	0.13 ± 0.01	1.75 ± 0.06	13.5	0.41 ± 0.07	0.12 ± 0.02
1	3.34 ± 0.08	10.31 ± 1.00	3.1	2.91 ± 0.54	6.06 ± 0.40
2	0.03 ± 0.004	0.98 ± 0.30	32.7	0.10 ± 0.05	2.51 ± 0.71
3	0.03 ± 0.002	0.45 ± 0.05	15.0	0.10 ± 0.05	1.26 ± 0.67
4	0.04 ± 0.01	0.74 ± 0.17	18.5	0.43 ± 0.02	2.35 ± 1.16
5	0.46 ± 0.03	2.42 ± 0.20	5.3	3.02 ± 0.52	12.09 ± 2.00
6	10.85 ± 0.90	25.26 ± 2.80	2.3	>12.5	>12.5
7	8.83 ± 1.10	23.71 ± 0.90	2.7	2.48 ± 0.41	>12.5
8	0.09 ± 0.02	1.13 ± 0.30	12.6	0.49 ± 0.10	0.53 ± 0.23
9	9.29 ± 1.40	27.95 ± 2.90	3.0	12.67 ± 0.18	11.98 ± 1.08

Selectivity index (SI) was defined as IC50(HMLER CD24high/CD44low)/IC50(HMLER CD24low/CD44high). Each IC50 value was determined in biological triplicate (three independent biological experiments), and each triplicate was determined in at least technical duplicate.

Results and Discussion

Analog Design and Synthesis

Our key starting material, ironomycin, was synthesized according to a two-step protocol we previously published,[33] with slight modifications. Very briefly, in the first step, the allylic C20-hydroxyl was chemoselectively and quantitatively oxidized to the corresponding C20-ketone (oxoSal, Scheme ), using activated MnO2. Next, oxoSal was transformed to obtain the desired product ironomycin via chemoselective and diastereoselective reductive amination, using Luche reaction conditions (Scheme ). The NMR data of both products oxoSal and ironomycin were in agreement with the literature.[33]

To improve the potency and selectivity of Sal, Liu and co-workers recently synthesized a series of 1,2,3-triazole analogs at the C1/C20 positions by means of click chemistry.[20] They found that the introduction of the triazole motif at the C20 position can improve the inhibition of breast cancer cell proliferation.[20] On the other hand, Jiang and co-workers generated a library of Sal 1,2,3-triazoles with the opposite stereochemistry at the C20 position and further identified a 19F MRI sensitive product as a promising agent against cancer cells.[23]

Thus, having access to ironomycin, we synthesized a series of novel 1,2,3-triazole derivatives 1–9 (Scheme ). Regiospecific copper-catalyzed azide–alkyne cycloaddition (CuAAC) of ironomycin with selected commercially available organic azides was successfully completed using the Meldal–Sharpless protocol,[38,39] which led to the diversely 1,4-disubstituted 1,2,3-triazoles 1–8 (Scheme ). To establish a structure–activity relationship (SAR), various organic azides were chosen that give rise to products with various types of substituents, such as shorter or longer aliphatic chains containing heteroatoms (oxygen, sulfur), more sterically hindered tert-butyl and adamantane motifs, or aromatic rings. Furthermore, to expand the structural diversity at the C20 position, we also obtained the 1H-1,2,3-triazole analog 9 (Scheme ), following the Cu(I)-catalyzed reaction of ironomycin with TMSN3 in MeOH/DMF.[40] Despite the reactivity of polyfunctionalized molecules, we found that the reactions proceeded cleanly and resulted in the production of nine derivatives with satisfactory yields and in good quantities for further biological evaluation.

The spectral data of compounds 1–9 are described in the Supporting Information. In the 1H NMR spectra of the target derivatives recorded in CD2Cl2, the acetylene proton peak observed for ironomycin as a narrow triplet at δ 2.30 ppm (t, J = 2.0 Hz, 1H) was no longer observed, while a singlet characteristic of a triazole proton was found in the δ 7.79–8.93 ppm range. With respect to product 9, the NMR data suggested the zwitterionic nature of the molecule. In particular, in the 13C NMR spectrum, the signal of the C1 carboxyl group was shifted downfield (184.0 ppm) which is characteristic of a zwitterionic structure, while the singlet observed in the 1H NMR spectrum at 4.99 ppm (s, 2H) corresponds to the ammonium group. Finally, HRMS analysis also supported the formation of the expected products, with [M + H]+ as the main peaks.

To study the ionic binding properties of the synthesized derivatives, we investigated whether 2 has the ability to directly interact with iron(II). 1H NMR experiments showed that addition of 0.5 equiv of iron(II) chloride to a CD3OD solution of 2 resulted in line broadening and shifting of specific proton signals, including triazole, thiomethylene, and the C18–C19 olefinic protons (Supporting Information, Figure S28), indicating interactions between 2 and iron(II) in solution. A similar trend was observed for ironomycin previously.[33] Importantly, while 2 was stable under these conditions, chemically unmodified Sal has been found to degrade in the presence of iron(III) chloride over time.[33]

Anti-CSC Activity Studies

Next, we screened ironomycin together with the library of 1,2,3-triazole derivatives against mesenchymal HMLER CD24low/CD44high cells, a well-established model of human breast CSCs,[41,42] and compared it with their epithelial counterparts HMLER CD24high/CD44low (Table ).

Interestingly, we found that the antiproliferative activity of the evaluated products strongly depends on the nature (type) of the substituent attached to the triazole ring. Those containing polar motifs (1 and 5–7, Scheme ), including a labile thioester, were essentially less active and less selective than the reference ironomycin; a similar observation was made for the N-unsubstituted 1,2,3-triazole analog 9 (Table ). In this group, compound 6 with ethyl ether units and the hydroxyl-containing product 7 exhibited low cytotoxicity, suggesting that the metal chelation ability is perturbed by these chemical modifications. Conversely, derivatives containing sterically hindered apolar moieties, such as adamantane or aromatic rings (2–4 and 8, Scheme ), exhibited up to a 4-fold increase in potency, with an improved selectivity for the mesenchymal state of cells comparable to ironomycin (Table ). The analogs with adamantane or phenyl substituents showed consistently lower active concentrations in cell viability assays than those with tert-butyl or aliphatic motifs. Remarkably, analog 2 was characterized by an IC50 value of 30 nM toward HMLER CD24low/CD44high cells and was the most potent of the series. This compound also showed higher selectivity than ironomycin (SI = 32.7 versus 13.5). Compared to the structurally similar analog 8, compound 2 exhibited improved biological activity; this suggests that the sulfur atom of the thioether group may contribute to iron coordination and/or bring flexibility. It is worth noting that the activity of products 3 and 4 was also improved with lower IC50 values compared to ironomycin, yet with comparable SI.

Interestingly, our new family of compounds also showed toxicity against MCF7 breast cancer cells, generally with low IC50 values, with 2 and 3 showing IC50 values of 100 nM, a 4-fold improvement over reference ironomycin (Table ). Moreover, we generally observed lower toxicity of these analogs against MCF10A normal breast cells (Table ), showing a potential therapeutic window to selectively target cancer cells.

Conclusions

In summary, the marked antiproliferative effect of ironomycin on the mesenchymal state of cancer cells prompted us to synthesize a series of its 1,2,3-triazole derivatives to further explore potential improvements. The versatile and practical approach described in this study provides a convenient and rapid access to potent semisynthetic small molecules that represent valuable tools for cancer cell biology studies.

All derivatives were assessed for their antiproliferative activity and selectivity for the mesenchymal state of cancer cells using a well-established model of CSCs. Notably, four products (2–4 and 8) showed more potent activity against HMLER CD24low/CD44high cells compared to the landmark compound ironomycin. Specifically, derivative 2 exhibits one of the highest potencies described thus far against this model cell line, laying down the foundation for the development of new anticancer drugs.

Experimental Section

General Information

Detailed descriptions of the general procedures, equipment, and measurement parameters can be found in the Supporting Information. While C20-oxosalinomycin (oxoSal) was synthesized following exclusively the protocol published by us previously,[33] the procedure for the preparation of ironomycin was slightly modified compared to the original one.[33]

Briefly, under argon atmosphere, C20-oxosalinomycin (oxoSal; 1.0 g, 1.3 mmol, 1.0 equiv) was dissolved in anhydrous MeOH (50 mL), followed by the addition of activated MS 4 Å, propargylamine (830 μL, 13.0 mmol, 10.0 equiv), and glacial acetic acid (500 μL). The yellow solution was stirred for 3 h at RT, prior to the addition of CeCl3·7H2O (500 mg, 1.3 mmol, 1.0 equiv) and the dropwise addition of a solution of NaBH3CN (117 mg, 1.8 mmol, 1.4 equiv) in anhydrous MeOH (20 mL) over 20 h at RT, using a syringe pump. After addition of all of the reducing agent, the reaction mixture was stirred for an additional 3 h. Then, the yellow cloudy reaction mixture was filtered to remove the insoluble components. The filtrate was evaporated to dryness, and CH2Cl2 (40 mL) was added to the yellow oily residue. After mixing, the organic solvent containing ironomycin was filtered off. The filtrate was then extracted twice with an aqueous solution of H2SO4 (15 mM, 2 × 30 mL) and once with distilled water (30 mL). The organic layer was separated, dried over MgSO4, evaporated under reduced pressure, and purified chromatographically using first the CombiFlash system (cyclohexane/EtOAc 0% → 100%) and then the HPLC equipped with a C18-reverse phase column (gradient: ACN/H2O/FA 1/1/0.1 to ACN/FA 1/0.1) to give the pure product of the reaction, ironomycin as a white amorphous solid (374 mg, 37% yield).

General Procedure for the Preparation of 1,4-Disubstituted 1,2,3-Triazoles of Ironomycin 1–8

The respective organic azide (1.1 equiv) and DIPEA (3.0 equiv) were introduced to a solution of ironomycin (1.0 equiv) in anhydrous ACN under argon atmosphere, followed by the addition of catalytic CuI (0.1 equiv) in one portion. The reaction mixture was stirred at RT, typically for 24 h. After the consumption of the propargyl precursor (TLC control), the organic solvent was evaporated in vacuo. The oily residue was then dissolved in a small portion of CH2Cl2 and extracted a few times with 10% aq EDTA solution. The organic phases were separated and concentrated under reduced pressure. Purification on silica gel using the CombiFlash system followed by HPLC equipped with a C18-reverse phase column (gradient: ACN/H2O/FA 1/1/0.1 to ACN/FA 1/0.1) gave the pure products of the click reaction, 1–8 (17–43% yield) as amorphous solids. The NMR and HRMS spectra of compounds 1–8 are included in the Supporting Information (Figures S1–S24).

Compound 1

Yield: 32 mg, 43%. Isolated as a white amorphous solid, > 95% pure by NMR, and a single spot by TLC. Rf: 0.68 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (400 MHz, CD2Cl2) δ 8.29 (s, 1H), 6.30 (d, J = 10.3 Hz, 1H), 6.17 (s, 2H), 6.14 (dd, J = 10.2, 4.9 Hz, 1H), 4.64–4.54 (m, 2H), 4.04 (dd, J = 10.1, 1.6 Hz, 1H), 3.96 (dd, J = 10.9, 5.4 Hz, 1H), 3.80–3.71 (m, 2H), 3.65 (dd, J = 10.0, 1.8 Hz, 1H), 3.62 (d, J = 10.2 Hz, 1H), 3.52 (d, J = 5.0 Hz, 1H), 2.86 (dt, J = 10.8, 3.8 Hz, 1H), 2.62–2.50 (m, 2H), 2.14 (ddd, J = 11.6, 8.4, 2.9 Hz, 1H), 2.04–0.62 (m, 65H) ppm; 13C NMR (101 MHz, CD2Cl2) δ 214.7, 179.6, 177.5, 143.6, 129.3, 128.2, 126.8, 106.5, 98.8, 89.0, 77.5, 76.1, 75.7, 73.6, 71.6, 71.2, 70.3, 68.4, 55.7, 54.7, 50.4, 48.7, 42.8, 40.3, 39.8, 39.0, 38.3, 36.3, 32.7, 31.4, 31.0, 29.4, 28.4, 27.0 (3C), 25.2, 23.3, 22.3, 20.4, 17.6, 16.8, 15.7, 14.5, 13.2, 12.8, 12.5, 11.3, 7.2, 6.5 ppm, one signal overlapped; HRMS (ESI+) m/z [M + H]+ Calcd for C51H85N4O12 945.6159, Found 945.6147.

Compound 2

Yield: 26 mg, 28%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.67 in CH2Cl2/acetone 33%. UV-active and strain green with PMA; 1H NMR (500 MHz, CD2Cl2) δ 8.23 (s, 1H), 7.41–7.34 (m, 2H), 7.34–7.27 (m, 3H), 6.34 (d, J = 10.2 Hz, 1H), 6.15 (dd, J = 10.1, 5.1 Hz, 1H), 5.62–5.55 (m, 2H), 4.66 (d, J = 13.4 Hz, 1H), 4.58 (d, J = 13.4 Hz, 1H), 4.03 (d, J = 10.0 Hz, 1H), 3.92 (dd, J = 10.5, 5.6 Hz, 1H), 3.79–3.70 (m, 2H), 3.66 (d, J = 9.7 Hz, 1H), 3.59 (d, J = 10.1 Hz, 1H), 3.54 (d, J = 5.0 Hz, 1H), 2.84 (dt, J = 10.8, 3.3 Hz, 1H), 2.62–2.51 (m, 2H), 2.15 (ddd, J = 11.5, 7.3, 4.0 Hz, 1H), 1.99–0.64 (m, 56H) ppm; 13C NMR (126 MHz, CD2Cl2) δ 215.3, 180.0, 143.2, 132.8, 132.6 (2C), 130.2, 129.7 (2C), 128.8, 128.0, 125.6, 106.3, 98.7, 88.9, 77.5, 76.2, 75.7, 73.4, 71.7, 71.2, 68.4, 55.7, 54.6, 50.5, 48.9, 43.1, 40.5, 39.6, 38.2, 36.3, 32.7, 31.6, 31.0, 29.4, 28.5, 27.0, 25.0, 23.3, 22.3, 20.5, 17.6, 16.9, 15.7, 14.5, 13.2, 12.63, 12.61, 11.3, 7.2, 6.6 ppm, one signal overlapped; HRMS (ESI+) m/z [M + H]+ Calcd for C52H81N4O10S 953.5668, Found 953.5655.

Compound 3

Yield: 17 mg, 33%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.49 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (500 MHz, CD2Cl2) δ 8.27 (s, 1H), 6.38 (d, J = 10.0 Hz, 1H), 6.25 (dd, J = 9.5, 5.5 Hz, 1H), 4.81 (d, J = 13.1 Hz, 1H), 4.61 (d, J = 13.1 Hz, 1H), 4.07 (d, J = 10.0 Hz, 1H), 3.97 (dd, J = 10.0, 4.9 Hz, 1H), 3.79–3.65 (m, 4H), 3.55 (d, J = 10.0 Hz, 1H), 2.90–2.81 (m, 1H), 2.58–2.48 (m, 2H), 2.28–2.11 (m, 11H), 2.01 (dd, J = 7.9, 5.9 Hz, 2H), 1.91–0.63 (m, 59H) ppm; 13C NMR (126 MHz, CD2Cl2) δ 214.8, 180.3, 140.6, 131.0, 127.9, 122.7, 106.3, 98.7, 88.9, 77.5, 76.5, 75.7, 73.4, 71.6, 71.2, 68.0, 59.6, 55.5, 55.0, 50.6, 49.0, 43.8, 43.2 (3C), 40.8, 39.5, 38.1, 36.3 (3C), 36.2, 32.7, 31.7, 31.0, 30.0 (3C), 29.5, 28.4, 27.0, 24.8, 23.1, 22.2, 20.6, 17.5, 16.8, 15.7, 14.5, 13.2, 12.8, 12.5, 11.3, 7.3, 6.6 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C55H89N4O10 965.6573, Found 965.6563.

Compound 4

Yield: 27 mg, 39%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.49 in 100% EtOAc. UV-active and strain green with PMA; 1H NMR (500 MHz, CD2Cl2) δ 8.93 (s, 1H), 8.07 (s, 4H), 6.40 (dd, J = 19.5, 8.8 Hz, 2H), 5.07 (d, J = 12.9 Hz, 1H), 4.86 (d, J = 12.9 Hz, 1H), 4.06 (dd, J = 10.4, 5.8 Hz, 1H), 3.96 (d, J = 9.9 Hz, 1H), 3.81–3.70 (m, 4H), 3.55 (d, J = 10.1 Hz, 1H), 2.96–2.85 (m, 1H), 2.61 (s, 3H), 2.46 (dd, J = 21.1, 8.7 Hz, 2H), 2.20 (ddd, J = 10.8, 8.3, 2.1 Hz, 1H), 2.05–0.60 (m, 53H), 0.38 (d, J = 5.7 Hz, 3H) ppm; 13C NMR (126 MHz, CD2Cl2) δ 216.0, 196.9, 180.5, 143.1, 140.8, 136.8, 131.9, 130.1 (2C), 128.3, 124.9, 120.2 (2C), 106.9, 99.1, 89.2, 77.5, 76.9, 75.9, 73.4, 71.8, 71.1, 68.5, 55.5, 55.4, 50.6, 49.0, 43.9, 41.0, 39.3, 37.9, 36.1, 32.6, 31.5, 31.0, 29.4, 28.5, 27.0, 26.8, 25.0, 23.1, 22.3, 20.6, 17.5, 17.1, 15.6, 14.5, 13.1, 12.6, 12.2, 11.3, 7.2, 6.5 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C53H81N4O11 949.5896, Found 949.5883.

Compound 5

Yield: 16 mg, 27%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.37 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (400 MHz, CD2Cl2) δ 8.22 (s, 1H), 6.36 (d, J = 10.2 Hz, 1H), 6.18 (dd, J = 10.1, 5.5 Hz, 1H), 4.76 (d, J = 13.2 Hz, 1H), 4.65 (d, J = 13.2 Hz, 1H), 4.30 (dt, J = 7.2, 2.2 Hz, 2H), 4.04 (dd, J = 10.2, 1.6 Hz, 1H), 3.95 (dd, J = 10.8, 5.4 Hz, 1H), 3.80–3.71 (m, 2H), 3.67 (dd, J = 10.0, 1.6 Hz, 1H), 3.64 (d, J = 5.5 Hz, 1H), 3.57 (d, J = 10.1 Hz, 1H), 2.86 (ddd, J = 14.6, 11.4, 5.4 Hz, 3H), 2.59–2.50 (m, 2H), 2.30 (s, 3H), 2.17 (ddd, J = 11.7, 8.4, 3.0 Hz, 1H), 2.07–1.15 (m, 34H), 0.96–0.81 (m, 13H), 0.81–0.62 (m, 13H) ppm; 13C NMR (101 MHz, CD2Cl2) δ 215.0, 195.7, 180.1, 141.9, 130.9, 127.9, 126.0, 106.3, 98.7, 88.9, 77.5, 76.4, 75.7, 73.4, 71.6, 71.2, 68.3, 55.6, 54.4, 50.5, 49.9, 49.0, 43.1, 40.7, 39.5, 38.1, 36.2, 32.7, 31.7, 31.0, 30.8, 29.49, 29.46, 28.6, 28.5, 27.0 (2C), 24.9, 23.3, 22.3, 20.5, 17.6, 16.9, 15.7, 14.5, 13.2, 12.6 (2C), 11.3, 7.2, 6.6 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C51H85N4O11S 961.5930, Found 961.5920.

Compound 6

Yield: 29 mg, 38%. Isolated as a cream amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.20 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (500 MHz, CD2Cl2) δ 8.17 (s, 1H), 6.31 (d, J = 10.2 Hz, 1H), 6.12 (dd, J = 10.1, 4.7 Hz, 1H), 4.59 (q, J = 13.4 Hz, 2H), 4.52–4.43 (m, 2H), 4.07 (d, J = 9.9 Hz, 1H), 3.95 (dd, J = 10.6, 5.5 Hz, 1H), 3.86 (t, J = 5.5 Hz, 2H), 3.76 (dt, J = 13.3, 4.8 Hz, 2H), 3.66 (d, J = 9.6 Hz, 1H), 3.64–3.51 (m, 9H), 3.48 (dd, J = 5.6, 3.5 Hz, 2H), 3.32 (s, 3H), 2.85 (dt, J = 10.9, 3.3 Hz, 1H), 2.62–2.50 (m, 2H), 2.20–2.11 (m, 1H), 2.04 (dd, J = 21.3, 10.3 Hz, 1H), 1.99–1.13 (m, 28H), 0.98–0.81 (m, 13H), 0.81–0.62 (m, 13H) ppm; 13C NMR (126 MHz, CD2Cl2) δ 214.7, 179.8, 142.4, 129.5, 127.9, 126.3, 106.2, 98.6, 88.9, 77.5, 76.1, 75.7, 73.5, 72.3, 71.6, 71.2, 71.0, 70.79, 70.76, 69.8, 68.3, 59.0, 55.7, 54.3, 50.40, 50.36, 48.9, 42.8, 40.4, 39.8, 38.3, 36.3, 32.7, 31.5, 31.0, 29.4, 28.5, 27.0, 25.2, 23.4, 22.3, 20.4, 17.6, 16.7, 15.8, 14.4, 13.2, 12.7, 12.5, 11.3, 7.2, 6.6 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C52H89N4O13 977.6421, Found 977.6407.

Compound 7

Yield: 27 mg, 36%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.16 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (400 MHz, CD2Cl2) δ 8.58 (s, 1H), 6.44 (d, J = 10.0 Hz, 1H), 6.31 (dd, J = 10.0, 6.2 Hz, 1H), 4.95 (d, J = 13.2 Hz, 1H), 4.61 (d, J = 13.2 Hz, 1H), 4.55–4.37 (m, 2H), 4.07–3.94 (m, 2H), 3.83–3.62 (m, 4H), 3.54 (d, J = 10.1 Hz, 1H), 3.48 (t, J = 6.1 Hz, 2H), 2.81 (dt, J = 11.0, 3.8 Hz, 1H), 2.60–2.45 (m, 2H), 2.20–2.10 (m, 1H), 2.09–0.56 (m, 59H) ppm; 13C NMR (101 MHz, CD2Cl2) δ 215.0, 180.5, 140.5, 132.2, 128.0, 127.7, 107.0, 99.3, 89.3, 77.5, 77.1, 76.0, 73.3, 71.9, 71.1, 67.9, 58.1, 55.3, 55.2, 50.5, 49.6, 46.6, 43.6, 40.9, 39.3, 38.1, 35.9, 32.7, 32.4, 31.4, 31.0, 29.5, 28.5, 27.0, 24.8, 22.9, 22.2, 20.8, 17.5, 16.9, 15.6, 14.5, 13.1, 12.6, 12.2, 11.4, 7.3, 6.5 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C48H81N4O11 889.5896, Found 889.5885.

Compound 8

Yield: 8.0 mg, 17%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.47 in CH2Cl2/acetone 33%. Strain green with PMA; 1H NMR (500 MHz, CD2Cl2) δ 8.18 (s, 1H), 7.38–7.30 (m, 5H), 6.36 (d, J = 10.2 Hz, 1H), 6.20 (dd, J = 10.1, 5.3 Hz, 1H), 5.47 (s, 2H), 4.70 (d, J = 13.4 Hz, 1H), 4.61 (d, J = 13.4 Hz, 1H), 4.02 (dd, J = 10.2, 1.6 Hz, 1H), 3.90 (dd, J = 10.9, 5.6 Hz, 1H), 3.73 (dt, J = 5.7, 5.2 Hz, 2H), 3.68 (dd, J = 10.0, 1.6 Hz, 1H), 3.63 (d, J = 5.4 Hz, 1H), 3.58 (d, J = 10.1 Hz, 1H), 2.85 (dt, J = 10.9, 3.7 Hz, 1H), 2.59–2.52 (m, 2H), 2.14 (ddd, J = 11.9, 7.8, 3.9 Hz, 1H), 1.94–0.64 (m, 56H) ppm; 13C NMR (126 MHz, CD2Cl2) δ 215.3, 180.1, 142.4, 135.7, 130.7, 129.3 (2C), 128.8, 128.5 (2C), 127.7, 126.1, 106.3, 98.8, 88.9, 77.5, 76.3, 75.8, 73.4, 71.6, 71.1, 68.5, 55.6, 54.7, 50.5, 48.9, 43.2, 40.6, 39.6, 38.2, 36.3, 32.7, 31.5, 31.0, 29.4, 28.5, 26.9, 25.0, 23.2, 22.3, 20.5, 17.6, 17.0, 15.7, 14.4, 13.2, 12.6, 12.5, 11.4, 7.2, 6.6 ppm, one signal overlapped; HRMS (ESI+) m/z [M + H]+ Calcd for C52H81N4O10 921.5947, Found 921.5938.

Synthesis of the 1H-1,2,3-Triazole Analog of Ironomycin (9)

Trimethylsilyl azide (28 μL, 0.21 mmol, 1.5 equiv) was added to an anhydrous DMF/MeOH solution (2.0 mL, 9:1) of catalytic CuI (1.3 mg, 0.007 mmol, 0.05 equiv) and ironomycin (102 mg, 0.14 mmol, 1.0 equiv) under argon in a microwave vial. The reaction mixture was stirred at 100 °C for 6 h. After the consumption of the propargyl precursor (TLC control, mobile phase: cyclohexane/EtOAc 50%), the organic solvents were evaporated in vacuo. The oily residue was then dissolved in a small portion of CH2Cl2 and extracted a few times with 10% aq EDTA solution. The organic phases were separated, dried over MgSO4, and concentrated under reduced pressure. Purification by HPLC equipped with a C18-reverse phase column (gradient: ACN/H2O/FA 1/1/0.1 to ACN/FA 1/0.1) gave the pure product of the click reaction, 9 (18% yield) as a white amorphous solid. The NMR and HRMS spectra of compound 9 are included in the Supporting Information (Figures S25–S27).

Compound 9

Yield: 20 mg, 18%. Isolated as a white amorphous solid, >95% pure by NMR, and a single spot by TLC. Rf: 0.12 in cyclohexane/EtOAc 50%. Strain green with PMA; 1H NMR (400 MHz, CD2Cl2) δ 7.79 (s, 1H), 6.55 (d, J = 9.9 Hz, 1H), 6.09 (dd, J = 9.8, 6.5 Hz, 1H), 4.99 (s, 2H), 4.10–4.04 (m, 2H), 3.84 (q, J = 6.9 Hz, 1H), 3.72 (dd, J = 10.6, 3.4 Hz, 1H), 3.67 (dd, J = 10.0, 1.6 Hz, 1H), 3.57 (d, J = 10.3 Hz, 1H), 3.51 (d, J = 6.5 Hz, 1H), 2.80 (dt, J = 11.1, 4.1 Hz, 1H), 2.65–2.55 (m, 2H), 2.34–2.16 (m, 1H), 2.06–0.59 (m, 58H) ppm; 13C NMR (101 MHz, CD2Cl2) δ 217.4, 184.0, 135.2, 134.9, 128.8, 126.8, 106.0, 98.9, 89.6, 77.5, 77.0, 76.1, 73.6, 71.4, 71.1, 68.5, 55.7, 52.5, 50.8, 50.6, 41.7, 39.8, 39.0, 37.7, 35.9, 32.7, 31.2, 30.6, 29.3, 28.7, 27.4, 26.2, 23.1, 22.8, 20.6, 17.6, 17.1, 15.5, 14.4, 13.2, 12.4, 12.0, 11.3, 7.1, 6.5 ppm; HRMS (ESI+) m/z [M + H]+ Calcd for C45H75N4O10 831.5478, Found 831.5468.

Cell Culture

HMLER cells naturally repressing E-cadherin, obtained from human mammary epithelial cells infected with a retrovirus carrying hTERT, SV40, and the oncogenic allele H-rasV12, were cultured in DMEM/F12 (Gibco, 31331-028) supplemented with 10% FBS, 10 μg mL–1 insulin (Sigma-Aldrich, I0516), 0.5 μg mL–1 hydrocortisone (Sigma-Aldrich, H0888), and 0.5 μg mL–1 puromycin (Life Technologies, A11138-02) and were a generous gift from Alain Puisieux (INSERM). All cells were incubated at 37 °C with 5% CO2. HMLER CD44low/high cells stained with CD24-APC and CD44-PE antibodies were sorted by FACS using an Aria IIu (BD Biosciences) to obtain isolated CD24low/CD44high and CD24high/CD44low cell populations. HMLER CD24low/CD44high cells were supplemented with 10 ng mL–1 human epidermal growth factor (EGF, Miltenyi Biotec, 130-093-750, 100 ng mL–1), while HMLER CD24high/CD44low cells were grown without EGF. MCF7 cells (ATCC, HTB-22) were cultured in DMEM GlutaMAX (ThermoFisher Scientific, 61965059) supplemented with 10% FBS (Gibco, 10270-106) and a penicillin–streptomycin mixture (BioWhittaker/Lonza, DE17-602E). MCF10A cells (ATCC, CRL-10317) were cultured in DMEM/F12 supplemented with 10% Horse Serum (Invitrogen, 16050-122), 10 μg mL–1 insulin, 10 ng mL–1 EGF, 0.5 μg mL–1 hydrocortisone, 100 ng mL–1 cholera toxin (Sigma-Aldrich, C8052), and 1 × PenStrep (Invitrogen, 15070-063).

Cell Viability Assay (IC50)

The cell viability assay was carried out by plating 1000 cells per well in 96-well plates. Cells were treated for 72 h in a range between 12 nM and 50 μM using serial dilutions, following the manufacturer’s protocol. Very briefly, CellTiter-Blue reagent (G8081, Promega) was added after 72 h of treatment, and cells were incubated for 3 h before recording fluorescence intensities (λex 560/20 nm; λem 590/10 nm) using a PerkinElmer Wallac 1420 Victor2Microplate Reader.

The IC50 cell viability curves for synthesized compounds against HMLER CD24low/CD44high, its isogenic cell line HMLER CD24high/CD44low, MCF7, as well as MCF10A cells are included in the Supporting Information (Figures S29–S31).