Synthesis and PI3 Kinase Inhibition Activity of a Wortmannin-Leucine Derivative

Wortmannin is a potent covalent inhibitor of PI3K that shows substantial in vivo toxicity and thus is unsuitable for systemic therapeutic applications. One possible approach to minimize systemic toxicity is to generate a latent wortmannin pro-drug that will be selectively activated in target tissues. To test this approach, a wortmannin derivative with a leucine linker attached to C20 has been synthesized and tested for inhibition of PI3K activity in prostate cancer cells. Analysis of PI3K pathway inhibition by Wormannin-Leu (Wn-L) and intact Wortmannin (Wn) showed that attachment of Leu at C-20 decreased potency of PI3K pathway inhibition 10-fold compared to intact wortmannin, yet exceeded the potency of a competitive PI3K inhibitor LY294002.


Introduction
A significant proportion of advanced prostate cancers are characterized by increased PI3K signaling, however no survival benefits from PI3K inhibitors tested in clinical trials have been reported [1][2][3][4][5][6][7].One potential strategy to improve therapeutic efficacy is to generate prostate-selective PI3K inhibitor prodrugs by attaching a peptide that prevents entry of the PI3K inhibitor into cells.To test this concept, a prodrug of quercetin analogue LY294002 has been generated by attaching a peptide Mu-LEHSSKLQL, in which the internal sequence HSSKLQ is a substrate for the prostate-specific antigen (PSA) protease.This produced a latent PI3K inhibitor prodrug Mu-LEHSSKLQL-LY294002 which activation is dependent on PSA cleavage.Once activated, Mu-LEHSSKLQL-LY294002 can specifically inhibit PI3K in PSA-secreting prostate cancer cells and induce apoptosis with a potency comparable to that of the original LY294002 compound [8].However, for effective inhibition of PI3K in vivo, generating prostate-selective PI3K inhibitors with lower IC50 would be beneficial.To generate a more potent prostate-selective PI3K inhibitor, we tested the feasibility of replacing LY294002 (IC50 ~1.4 mcM) with Wortmannin (1) (IC50 4.7 nM) (Scheme 1).Because PSA is predicted to cleave Mu-LEHSSKLQL peptide after glutamine (Q), it leaves leucine (Leu) attached to the base inhibitor that potentially can diminish inhibitor potency.Since earlier publications demonstrated that various moieties can be attached to wortmannin via C-20 [9,10] we tested the effect of Leu attached at this position on the compound potency and we report the results of those studies here.

Chemistry
The synthesis of compound 10 (Scheme 2) was accomplished by the reaction of Boc-protected leucine (2) with propargylamine (3) under standard peptide coupling conditions (diisopropylcarbodiimide (DIC), hydroxybenzotriazole (HOBt)) to give the propargyl amide (4) in 90% yield.Azido amine compound (5) was prepared as described previously [11].The click reaction between 4 and 5 produced compound 6 as a 2.3:1 mixture of regioisomers, which were purified by SCX2 resin in 64% yield.We presume based on literature precedent [12], that the trans isomer (6) shown was the major isomer due to steric interactions in the cycloaddition.For simplicity this scheme just depicts this isomer.Compound 6 was methylated at the terminal amino group via reductive amination using formalin and sodium triacetoxyborohydride (STAB-H) to form compound 7 in low yield (14%) after extensive purification.The reaction was not selective and a mixture of mono and dialkylated products, along with unreacted starting material (6) were present.As this was a proof of concept project, we decided not to optimize any one synthetic step unless the inhibition activity at the end warranted it later.We did not add the primary amine (6) directly to wortmannin at C-20 since other primary amines were known to add there reversibly [10].Compound 7 was coupled with wortmannin (8) and the product (9) was purified by silica gel chromatography in 38% yield.The Boc protecting group was removed (phosphoric acid, toluene) to form compound (10).

Biological Activity
To assess the biological activity of new PI3K inhibitor compounds, the experiments were conducted in C4-2 prostate cancer cells that represent castration-resistant metastatic prostate cancer.C4-2 cells are characterized by constitutive activation of PI3K pathway due to the loss of expression of PI3 phosphatase PTEN.A protein kinase PKB/AKT is among best characterized downstream effectors of PI3K pathway.Activation of PI3K leads to accumulation of PI3P in the plasma membrane.PI3P binds to PH domains of AKT and PDK kinases and recruits AKT and PDK kinases to plasma membrane.This in turn leads to phosphorylation of AKT at T308 by PDK1, and to phosphorylation at S473 by rapamycin-insensitive mTOR:Rictor:Gβ complex.Thus, phosphorylation of AKT at S473 and T308 faithfully reflect activation of PI3K pathway in most cell lines and routinely used to monitor the PI3K activity [13].
Analyses of phosphorylation of S473AKT in C4-2 cells have been used in this study to assess the PI3K inhibition by Wn-L. Figure 1 shows representative western blots that illustrate inhibition of S473AKT phosphorylation by Wn (1) and by Wn-L (10) which contains the leucine linker.Comparison of PI3K pathway inhibition by Wormannin-Leu (Wn-L) (10) and intact Wortmannin (Wn) (1) showed that attachment of Leu at C-20 decreased potency of PI3K pathway inhibition by 10 fold since comparable inhibition of Akt phosphorylation at S473 was observed in C42 cells treated with 100 nM of Wn-L and 10 nM of Wn (Figure 1).Future studies of the effects of Leu attached at various positions of Wn are needed to determine which Wn-L derivative compound will retain the inhibitory potency of intact wortmannin.
Quantitation of Western blots by image J software showed that respective IC50 s for Wn and Wn-L were 100-and 10-fold lower comparing to the well characterized PI3K inhibitor ZSTK474.Thus, 50% inhibition of S473AKT phosphorylation was observed at lower concentrations of Wn and Wn-L than 50% inhibition by ZSTK474.Comparison of Wn-L and Wn, showed that inhibition of S473AKT phosphorylation by compound Wn-L at 300 nM was comparable to inhibition by Wn at 30 nM; whereas inhibition by Wn-L at 100 nM was comparable to inhibition by Wn at 10 nM.Thus, Wn-L is a 10-fold less potent PI3K inhibitor comparing to Wn.

General Information
Unless otherwise noted, solvents and reagents were used without purification.1,2-dichloroethylene (DCE) was dried over 4 Å molecular sieves for 48 h prior to use.Volatile solvents were removed under reduced pressure using a Buchi rotary evaporator.Infrared (IR) spectra were obtained using a Nicolet iS550 FT IR spectrophotometer using a diamond crystal attenuated total reflection (ATR) accessory and reported as wave numbers.Melting points were determined using differential scanning calorimetry (DSC) on a TA Instruments Differential Scanning Calorimeter Model Q100.Thin layer chromatography (TLC) was performed on glass-backed precoated silica gel plates (0.25 mm thick with 60 F254) and were visualized using one or both of the following manners: UV light (254 nm) and staining with I 2 impregnated silica.Flash flash chromatography was performed using the Biotage Isolera One using pre-loaded Silicycle 25g high performance (14-40 µM) columns. 1 H nuclear magnetic resonance (NMR) spectra were obtained at 400 MHz as indicated as solutions in CDCl 3 with 0.05% v/v tetramethylsilane (TMS) unless indicated otherwise. 13C-NMR were obtained at 100 MHz as shown in the indicated deuterated solvent.Chemical shifts are reported in parts per million (ppm, δ), and referenced to TMS, and coupling constants are reported in Hertz (Hz).Spectral splitting patterns are designated as s, singlet; d, doublet; t, triplet; q, quartet; quint, quintuplet; sex, sextet; sept, septuplet; m, multiplet; comp, overlapping multiplets of magnetically nonequivalent protons; br, broad; and app, apparent.Copies of NMR spectra and mass spectra for all new compounds reported here are included in the Supplementary Materials.