Assays were performed in duplicate
Molecular Editing of Kinase-Targeting Resorcylic Acid Lactones (RAL): Fluoroenone RAL
Article first published online: 5 MAR 2010
Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Volume 5, Issue 5, pages 670–673, May 3, 2010
How to Cite
Jogireddy, R., Barluenga, S. and Winssinger, N. (2010), Molecular Editing of Kinase-Targeting Resorcylic Acid Lactones (RAL): Fluoroenone RAL. ChemMedChem, 5: 670–673. doi: 10.1002/cmdc.201000047
- Issue published online: 27 APR 2010
- Article first published online: 5 MAR 2010
- Manuscript Received: 3 FEB 2010
- Agence National de la Recherche
- irreversible inhibitors;
- resorcylic acid lactones;
Kinases have become one of the most intensely pursued target class for therapeutic intervention, particularly for applications in oncology.1 Most small-molecule inhibitors are heterocycles reminiscent of the adenosine scaffold, which target the nucleotide binding site of the kinase. Resorcylic acid lactones (RAL) bearing a suitably positioned cis-enone, such as hypothemycin,2, 3 LL-Z1640-2,4, 5 L-783277,6 radicicol A7, 8 (Figure 1), have been shown to irreversibly inhibit select kinases and represent a unique pharmacophore that has been shown to be effective in vivo.9 Through a bioinformatic analysis of the kinome, Santi and co-workers identified 46 kinases that could potentially be targeted by this family of compounds.2 Our efforts to expand the diversity of this natural pharmacophore has led to the identification of several tolerated modifications, such as substitution of the benzylic carbon by an oxygen which offers simpler synthetic access.7, 10, 11
Based on the fact that these compounds react with the cysteine residue in the kinase nucleotide binding site through a Michael addition, modulation of the electronic properties of the enone may further enhance the inhibitory properties of this pharmacophore. Previously, we have shown that substitution at the β-position of the enone with a methyl group abolishes activity.11 These results were concurrently confirmed by researchers from Eisai.12
To further pursue modulation of the Michael acceptor, we reasoned that a fluorine substituent at the α-position of the enone may accentuate the Michael acceptor properties while constituting a fairly neutral steric permutation. This “molecular editing”13 may provide further benefits from a synthetic perspective, as well as favorably impede the propensity of these cis-enones to isomerize to the thermodynamically more favorable and biologically less active trans-enone. As shown in Figure 2, the synthesis could capitalize on our previously established strategy,7, 10, 14 namely, disconnection of the ester and benzylic position. However, the difference of reactivity between a fluoroalkene and an alkene could be harnessed to introduce the diol moiety by dihydroxylation chemistry.
The synthesis of fluoroenones 1 and 2 began with the straightforward conversion of ester 5 into fluoroenal 7 via aldehyde 6 relying on a Horner–Wadsworth–Emmons reaction with triethyl 2-fluoro-2-phosphonoacetate to install the desired E-fluoroenoal (Scheme 1). Reaction of aldehyde 7 with alkyl lithium 8 afforded the desired alcohol as a diasteromeric mixture (diastereoisomeric ratio (dr), 1:1), which was reduced to the cis-alkene using Lindlar's catalyst (H2, Pd/CaCO3) to afford compound 9. As the hydroxy group will ultimately be oxidized to the ketone, the lack of selectivity is inconsequential. Benzoyl protection of alcohol 9 thus afforded 10 wherein the terminal OPMB group was converted to an iodide 11 (DDQ; I2, PPh3, imidazole). Attempts to alkylate a phenol using 11 led to rapid elimination of the iodide to give a conjugated triene (product not shown). To circumvent this elimination, which was thought to be facilitated by the presence of the alkene, compound 10 was dihydroxylated (dr, 2.5:1 in favor of the desired isomer 12) and converted to acetonide 13. Conversion of the terminal protected PMB ether into an iodide afforded the key intermediate 14.
Fluoroenones 1 and 2 were synthesized from key intermediate 14 via the route shown in Scheme 2. Following previously established chemistry, fragment 14 was coupled to the aromatic moiety 15 bearing a benzylic selenide by alkylation followed by oxidation/elimination of the selenide to afford 17. Conversely, alkylation of phenol 16 with fragment 14 yielded the ether analogue 18 in excellent yield. Global deprotection of the silyl groups (TMSE and TBDPS) followed by Mitsunobu macrocyclization, selective benzoate deprotection and oxidation of the resulting alcohol afforded products 19 and 20 from 17 and 18, respectively, in good overall yield. Global deprotection of 19 and 20 using aqueous hydrofluoric acid and resin-bound sulfonic acid, respectively, afforded the desired fluoroenone resorcylic acids 1 (from 19) and 2 (from 20). We and others had previously noted the sensitivity of the cis-enone system of LL-Z1640-2 and related analogues to acid conditions that could lead to isomerization.10, 15, 16 As anticipated based on the higher thermodynamic stability of E-fluoroalkene, analogues 1 and 2 proved to be more resistant to epimerization, and no trace of isomerization was observed even with prolonged reaction times. Interestingly, the different diastereoisomers originating from the nonselective dihydroxylation proved to have significant difference in the kinetics of deprotection, which offered a convenient means to separate them.
To gage the indiscriminate reactivity of these enone resorcylic acids with thiols, fluoroenone 2 and the parent analogue lacking the fluoride were treated with one equivalent of DTT (2 mM) as a prototypical thiol nucleophile in phosphate-buffered saline (pH 7.4) containing 1 % DMSO. Interestingly, a clear difference in reactivity was observed with the fluoroenone 2 reacting slower. With fluoroenones 1 and 2 in hand, we tested their efficacy against VEGF-R2, which was frequently found to be the most inhibited kinase in our profile of cis-enone resorcylic acid library,7, 11 and KIT, which represents a less inhibited kinase. As shown in Table 1, fluoroenones 1 and 2 are less active than the natural product LL-Z1640-2, however, they maintain an inhibition level that is interesting: 6.8 nM and 60.5 nM for 1 and 2, respectively, against VEGF-R2 compared with 2.63 nM for LL-Z1640-2, and their selectivity for VEGF-R2 relative to KIT remains unaffected. The products originating from the minor diastereoisomers of the dihydroxylation were significantly less active.
|Compound||VEGF-R2 [nM]||KIT [nM]|
The cellular activity of fluoroenones 1 and 2 against VEGF-R2 was tested in HUE cells, a spontaneously immortalized HUVEC clone known to overexpress VEGF-R2. The results paralleled those of the enzyme inhibition assays. As shown in Figure 3, the natural product LL-Z1640-2 is a very potent inhibitor with an IC50 value of 6.5 nM. Direct comparison with fluoroenone 2 revealed an IC50 value of 17 nm, while the analogue containing the synthetically more accessible ether macrocycle remained very potent with an IC50 value of 70 nM.
Thus the fluoroenone modification is well tolerated but does not enhance the cellular efficacy of VEGF-R2 inhibition. During the course of our investigation, a similar modification was reported by researchers from Esai,12 however, their medicinal chemistry efforts focused on the inhibition of inflammation-related pathways (MEKs and their down-stream regulatory effect on cytokines).12, 17 A selected analogue (E6201) indeed shows a >60- and >200-fold selectivity17 for MEK1 relatively to VGF-R2 and PDGFRβ, which is a significantly different pattern relatively to the parent natural product (LL-Z1640-2) for which VEGF-R2 and PDGFRβ are most inhibited.11 It was further shown that the addition of a methyl substituent at the γ-position relative to the enone (C3, numbering shown in Figure 2) enhanced the metabolic stability with a tolerable loss of activity (tenfold).12 It should be noted, however, that several cis-enone have already been shown to be effective in vivo. As the inhibition is irreversible, the relative rate of metabolic instability and irreversible inhibition coupled to target turn over should be considered.
Detailed procedures for the synthesis of compounds 1 and 2 are described in the Supporting Information.
Fluoroenone 1: Rf=0.19 (CH2Cl2/MeOH, 9:1); 1H NMR (CD3OD, 400 MHz): δ=6.99 (d, J=15.3 Hz, 1 H), 6.40 (d, J=2.4 Hz, 1 H), 6.28 (d, J=2.4 Hz, 1 H), 6.09 (ddd, J=15.1, 8.9, 6.4 Hz, 1 H), 5.95 (ddd, J=21.9, 11.6, 2.7 Hz, 1 H), 5.28–5.20 (m, 1 H), 4.90 (br s, 1 H), 4.09–4.04 (m, 1 H), 3.50–3.39 (m, 1 H), 2.62–2.53 (m, 1 H), 2.31–2.17 (m, 2 H), 1.50 ppm (d, J=6.2 Hz, 3 H), n.b., 4OH signals not visible; 13C NMR (CD3OD, 100 MHz, 25 °C): δ=196.4 (d, 2JC,F=36.8 Hz, CO), 172.8 (CO), 166.9 (C), 164.1 (C), 155.2 (2×d, 1JC,F=254.6 Hz, C), 145.0 (C), 133.9 (CH), 132.3 (CH), 123.9 (d, 2JC,F=18.4 Hz, CH), 109.4 (CH), 103.5 (C), 102.9 (CH), 81.7 (d, 3JC,F=3.8 Hz, CH), 74.8 (CH), 74.2 (CH), 36.6 (CH2), 33.4 (d, 3JC,F=6.1 Hz, CH2), 20.9 ppm (CH3); HRMS (MALDI-TOF): m/z [M+Na]+ calcd for C18H19FO7Na: 389.1013; found: 389.1024.
Fluoroenone 2: Rf=0.18 (CH2Cl2/MeOH, 9:1); 1H NMR (CD3OD, 400 MHz): δ=6.02 (d, J=1.9 Hz, 1 H), 5.98 (d, J=1.9 Hz, 1 H), 5.88 (ddd, J=21.8, 11.8, 2.7 Hz, 1 H), 5.44–5.29 (m, 1 H), 5.03 (d, J=4.3 Hz, 1 H), 4.24–4.18 (m, 2 H), 3.97–3.88 (m, 1 H), 3.11–3.00 (m, 1 H), 2.55–2.44 (m, 1 H), 1.94–1.86 (m, 1 H), 1.74–1.61 (m, 1 H), 1.38 ppm (d, J=5.9 Hz, 3 H), n.b., 4OH signals not visible; 13C NMR (CD3OD, 100 MHz, 25 °C): δ=196.8 (d, 2Jc,F=35.9 Hz, CO), 172.6 (CO), 167.1 (C), 165.4 (C), 163.6 (C), 155.6 (2xd, 1JC,F=255.3 Hz, C), 119.5 (d, 2JC,F=17.6 Hz, CH), 96.6 (CH), 96.4 (C), 92.8 (CH), 80.8 (d, 3JC,F=2.9 Hz, CH), 73.1 (CH), 69.5 (CH), 65.8 (CH2), 33.1 (d, 3JC,F=5.9 Hz, CH2), 31.3 (CH2), 20.8 ppm (CH3); HRMS (MALDI-TOF): m/z [M+Na]+ calcd for C17H19FO8Na: 393.0962, found: 393.0942.
Enzymatic inhibition of VEGF-R2 and KIT: A radiometric protein kinase assay (33PanQinase® Activity Assay) was used for measuring the kinase activity of VEGF-R2 and KIT kinases (Proqinase, Freiburg, Germany). All kinase assays were performed in 96-well FlashPlates™ from Perkin–Elmer (Boston, USA) using 50 μL of assay buffer (60 mM HEPES-NaOH, pH 7.5, 3 mM MgCl2, 3 mM MnCl2, 3 μM Na3VO4, 1.2 mM DTT, 50 μg mL−1 PEG2000, 1 μM [γ-33P]ATP), 20 ng of kinase and a generic substrate (polyGluTyr) with 1 % DMSO. The test compound concentration ranged from 10 μM to 0.1 nM (semilog dilution). The assays were performed by premixing the ATP solution with the test compound and addition of this solution to the kinase/substrate solution. After 60 min at 30 °C, the reaction was stopped with 50 μL of 2 % H3PO4, plates were aspirated and treated with 200 μL of 0.9 % NaCl (2×) and the level of 33P incorporation was determined with a microplate scintillation counter (MicroBeta, Wallac). Assays were run in duplicate.
Cellular VEGF-R2 inhibition assay: HUE cells, a spontaneously immortalized HUVEC clone known to overexpress VEGF-R2, were plated in ECGM (PromoCell, Germany), supplemented with 10 % fetal calf serum (FCS), with 25.000 cells per well in 48-well cell culture dishes. After 6 h, the medium was exchanged against ECBM (PromoCell, Germany) without FCS, and the cells were starved overnight. Prediluted test samples in DMSO (from 0.1 mM to 30 nM in half logarithmic steps) were added to the cell culture medium (1:100) resulting in a final DMSO concentration of 1 %. After incubation for 90 min at 37 °C, cells were stimulated for 7 min at room temperature with 100 ng mL−1 hVEGF-A165. As a positive control, one row in each plate was treated with the known kinase inhibitor staurosporine. The mean value of those wells was used as the background, which was substracted from all other data points. One row was treated with DMSO alone and represents the maximal gain in activation with cells that received 1 % DMSO and had been stimulated with hVEGF-A165. The mean of those wells was set to 100 %. Quantification of VEGF-R2 phosphorylation was assessed in 96-well plates via sandwich ELISA using a respective VEGF-R2-specific capture antibody and an antiphosphotyrosine detection antibody. Raw data were converted into percent receptor autophosphorylation relative to stimulated controls, which were set to 100 %. IC50 value determination was done with GraphPad Prism 5.01 software with constrain of bottom to 0 and top to 100 using a nonlinear regression curve fit with variable hill slope. The equation is a four-parameter logistic equation. All compounds were tested in triplicate.
This work was funded by grants from the Agence National de la Recherche (ANR) and Conectus.
Please note: Minor changes have been made to this publication in ChemMedChem EarlyView. The Editor.
Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors.
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.