Background

Prostate cancer is the number one cancer diagnosed in men today. While it occurs to certain extent throughout the world (least commonly in Eastern/Southern Asia), it is viewed as the major public health threat in Western Europe and, especially, the United States [1]. In the US alone, it has been estimated [2] that 186,295 new cases of prostate cancer (mostly - among men over fifty) were diagnosed in 2008, accounting for 25% of all cancers diagnosed in men that year and 10% of the total cancer-related mortality. Appropriate diet (including dietary supplements) and exercise are currently the common themes for prostate cancer prevention while classical treatments are limited to surgery, radiation therapy, and hormone therapy. Chemotherapy of late-stage prostate cancer is still largely experimental; however, it may lead to increased survival in the future [3]. Specifically, small molecules as well as antibodies targeted at disrupting vital signaling pathways in cancerous cells have a potential to provide new basis for innovative treatment of proliferative disorders as prostate cancer in the years to come [4].

Results and discussion

The present study was a part of an ongoing effort [5] in our laboratories to find novel antiproliferative agents as potential treatments for prostate cancer. It was aimed at identifying new small heterocyclic molecules in Chemical Diversity Research Institute collection (parts of that can be accessed at http://www.chemdiv.com) that would be specifically inhibitory to DU-145 human prostate carcinoma cell line (a 'classical' cell line of androgen-independent prostate cancer [6]) while exhibiting no non-specific (general) cytotoxicity. High-throughput screening of a highly diverse set of over 5,000 compounds comprising over 200 chemical classes led to several confirmed hit classes.

Among these, one hit compound, 1 (Figure 1), that exhibited inhibition of DU-145 cell proliferation in dose-response manner, attracted our attention due to its drug-likeness [7], structural simplicity and the presence in its structure of two distinct types of peripheral appendages (thus allowing for informative SAR exploration).

Figure 1
figure 1

Structure of the hit compound identified in DU-145 cell proliferation HTS.

Full size image

Herein we report on the synthesis and potency characterization of twenty analogs of the hit compound 1 leading to the initial SAR generalizations in this chemical class and to significant improvement of the inhibitory potency as well as to confirmation that the most potent compounds synthesized in this work inhibit DU-145 proliferation in a cell-specific manner rather than via general cytotoxicity.

To confirm that the observed activity of 1 was indeed associated with the N-aryl-N-(3-aryl-1,2,4-oxadiazol-5-yl)amines chemotype and to obtain initial SAR clues for further potency optimization, we synthesized a library of 176 analogs 2 of the hit compound. According to the literature [8], N-aryl-N-(3-aryl-1,2,4-oxadiazol-5-yl)amines 2 can be prepared in moderate yields by reaction of 2 eq. of carbodiimides with amidoximes (3). The target products 2 were isolated and purified chromatography on silica gel using 5-25% EtOAc in hexanes as a mobile phase. The amidoximes 3 were conveniently synthesized in nearly quantitative yields from corresponding benzonitriles (R1CN) (Scheme 1).

Scheme 1
scheme 1

Preparation of N-aryl-N-(3-aryl-1,2,4-oxadiazol-5-yl)amines 2 [8]. Reagents and conditions: (a) hydroxylamine hydrochloride (1 eq.), EtOH, reflux, 5-10 h; (b) aq. Na2CO3, extraction with EtOAc; (c) 0.2 M solution of 3 in dry DMF, 2.1 eq. of R2-N = C = N-R2 carbodiimides, 120°C, 18-36 h.

Full size image

Compounds 2 were tested at 10 μM concentration for inhibition of DU-145 cell proliferation. 19 compounds with inhibition >50% were selected for dose-response experiments to determine IC50 values. The results are presented in Table 1.

Table 1 Structures, preparative yields (from 3), and DU-145 proliferation inhibition data of compounds 2a-s.
Full size table

There was a clear preference of electron-rich or unsubstituted phenyls in the 'western' aryl portion of the molecule with notorious presence of meta-methoxy substituent in 14 active compounds out of 19 compounds for which dose-response curves were obtained. As to the 'eastern' arylamino portion, the best activity was observed for para-methoxyanilino substituent (2k and 2s). Notably, presence of electron-withdrawing or ortho-substituents in the arylamino portion of 2 rendered such compounds less active or completely devoid of inhibitory potency. This observation is illustrated by several examples provided in Figure 2.

Figure 2
figure 2

Examples of inactive or poorly active (% inhibition < 20% at 10 μM) compounds 2.

Full size image

These SAR observations are useful in determining the 'activity' chemistry space for future optimization of the series. To test our generalizations, we decided to combined the para-methoxy substituent in the 'eastern' arylamino portion of the molecule with another beneficial feature - namely, meta-methoxy substituent in the 'western' aryl portion (vide supra). Indeed, the activity of the synthesized SAR-guided compound 2t (Figure 3) improved nearly three-fold compared to 2k, thus demonstrating the synergy of effects of these two substituents on the inhibitory potency of the series.

Figure 3
figure 3

The 'best' SAR-merging compound 2t.

Full size image

Having improved the antiproliferative potency of the series by two orders of magnitude, we then investigated the specificity of the cytotoxicity of the most potent compounds. The general, non-specific cytotoxicity of the most potent three representative compounds (2j, 2s and 2t with IC50 values of 0.55, 0.38 and 0.029 μM, respectively) was measured on HepG2 and DU-145 cell lines using a known cytotoxic compound, digitonin, as a positive control (Figure 4). None of these compounds exhibited cytotoxicity in excess of 20% at concentrations as high as 50 μM, while digitonin itself caused 100% cell death in nanomolar concentration range. Thus, the observed antiproliferative activity of the most potent compounds in the series is not due to non-specific cytotoxicity but must be attributed to a specific cellular target and is selective towards human prostate carcinoma cells.

Figure 4
figure 4

Cytotoxicity of the most active compounds prepared in this work and digitonin measured on HepG2 (A) and DU-145 (B) cells.

Full size image

Methods

Cell sources: the prostate cancer Du-145 cell line was purchased from the American Type Cell Collection (HTB-81). Du-145 was cultured in RPMI-1640 complemented with 10% fetal bovine serum and 2 mM L-glutamine. To our delight, 80% of the library screened exhibited >20% inhibition of DU-145 proliferation. The hepatocellular carcinoma HepG2 cell line was purchased from the American Type Cell Collection (HB-8065) and was cultured in DMEM complemented with 10% fetal bovine serum and 2 mM L-glutamine.

Assays description

1. Biological assay to determine inhibition of proliferation of the DU-145 cells

Du-145 cells were plated in 384-well plate at the density of 4000 cells per well. 4 mM solutions of compounds in DMSO were diluted 100 times with medium and added to cells to a final concentration of 20 μM (40 μL of cell culture plus 40 μL of compound solution, final concentration of DMSO - 0.5%). Taxol at the final concentration of 1 μM was used as a positive control. The cells were incubated with compounds for 3 days. Alamar Blue was then added to the cells to a final concentration of 50 μM. After incubation for 4-6 hours at 37°C, plate fluorescence was read using fluorescence plate reader Wallac 1420 (530 nm excitation filter, 590 nm emission filter). Proliferation inhibition was calculated using formula:

where

F negative : DMSO added to the cells (viable cells) and

F positive : taxol (1 uM) added to cells (cell count on 1st day of incubation)

2. Cytotoxicity assay

For cytotoxicity experiments, higher seeding densities (10,000 or 20,000 cells per well) were used to more accurately observe potentially diminishing number of cells. This is in contrast with the measurement of antiproliferative activity where initial seeding density (4,000 cells per well) was chosen to allow more room for cell proliferation. Thus, Du-145 or HepG2 cells were plated in 384-well plate at the density of 10,000 or 20,000 cells per well, respectively. Serial dilutions (200×) of the tested compounds in DMSO were prepared. After that compounds were diluted 100 times in medium and added to the cells (40 μL of cell culture plus 40 μL of compound solution, final concentration of DMSO - 0.5%). The cells were incubated with compounds overnight. Known toxic compound digitonin was used as control. The next day Alamar Blue was added to the cells to final concentration of 50 μM. After incubation for 4-6 hours at 37°C, plate fluorescence was read using fluorescence plate reader Wallac 1420 (PerkinElmer) (530 nm excitation filter, 590 nm emission filter). Compound cytotoxity was calculated using formula:

where

F negative - DMSO was added to cells (viable cells)

F positive - digitonin was added to cells (dead cells)

Conclusions

In conclusion, we have undertaken hit expansion and SAR exploration of a new antiproliferative chemical series, N-aryl-N-(3-aryl-1,2,4-oxadiazol-5-yl)amines. This study led to identification of the new 'best' peripheral moieties which, when combined in the same compound (2t) led to significant improvement of the DU-145 proliferation inhibitory potency (>200-fold compared to the initial hit compound 1). We also demonstrated that the observed antiproliferative activity of the compounds belonging to the studied chemotype was not due to non-specific cytotoxicity. The compound 2t represents a promising new lead for development of novel therapeutic agents for treatment of androgen-independent prostate cancer. The specific cellular mechanism of action of this compound remains to be investigated and will be presented in subsequent communication.