Synthesis and in vitro antitumour activity of crassalactone D, its stereoisomers and novel cinnamic ester derivatives.

Naturally occurring styryl lactone, crassalactone D (1), unnatural 4-epi-crassalactone D (2), and the corresponding 7-epimers (3 and 4) have been synthesized starting from d-glucose. The key step of the synthesis is a new one-pot sequence that commenced with a Z-selective Wittig olefination of suitably functionalized sugar lactols with a stabilized ylide, (methoxycarbonylmethylene)-triphenylphosphorane, in dry methanol, to afford 1 or 3, in the mixtures with the corresponding 4-epimers (2 or 4, respectively). A number of 6-O-cinnamoyl derivatives of styryl lactones 1-4 have been prepared, bearing electron donating or electron withdrawing functionalities in the C-4 position of cinnamic acid residue. The synthesized products were evaluated for their in vitro antiproliferative activity against selected human tumour cell lines, whereupon very potent cytotoxicities have been recorded in many cases. SAR analysis indicated some important structural features responsible for biological activity, such as stereochemistry at the C-4 and C-7 positions, as well as the nature of a substituent at the C-4 position in the aromatic ring of cinnamoate moiety. Flow cytometry and Western blot analysis data gave insight in the mechanism underlying antiproliferative effects of the synthesized compounds.


Introduction
Crassalactone D (1, Fig. 1) is a naturally occurring styryl-lactone that shows notable cytotoxic effects against certain mammalian tumour cell lines. It was isolated from the tropical plant Polyalthia crassa by Tuchinda and co-workers in 2006 [1]. The relative configuration of 1 was established by X-ray diffraction analysis and its absolute stereochemistry was determined by NMR studies of its (R)-and (S)-MTPA esters. Two asymmetric total syntheses of crassalactone D (1) reported in 2009 [2,3] have been followed by a formal synthesis published in 2012 [4]. In the same year, we published a preliminary account describing a new chiral pool approach to the total synthesis of the natural product 1 and several new epimers [5]. The effects of the synthesized compounds on the proliferation of some human malignant cell lines have also been preliminary reported [5]. Herein, we wish to disclose full details of the synthesis and biological evaluation of styryl lactones 1-4, as well as some new 6-O-cinnamoyl derivatives (1a-d, 2a-d, 3a-d and 4a-d) along with their effects on the proliferation of certain malignant cell lines. Rationale for the preparation of cinnamic ester derivatives of crassalactone D and epimers arises from the fact that cinnamoates represent a promising class of anti-cancer agents [6]. Such concept of drug design and development is known as 'molecular hybridization'. It is based on the combination of pharmacophoric moieties of different bioactive substances to produce a new hybrid compound with improved affinity and efficacy. Additionally, this strategy can result in compounds presenting modified selectivity profile, different and/or dual modes of action and reduced undesired side effects [7]. Another reason for studying the cinnamic ester derivatives of type 1b-d, 2b-d, 3b-d, 4b-d is derived from our earlier findings that the introduction of either electron withdrawing or electron donating groups in the C-4 position of the cinnamic acid aromatic ring may increase antitumour activity of the resulting (+)-crassalactone B and C analogues [8,9]. Apoptotic signalling induced by the synthesized compounds was also studied.

Chemistry
Synthesis of crassalactone D (1) and the corresponding 7-epimer 3 is shown in Scheme 1. The starting compounds 5 and 7 were prepared from commercially available diacetone-D-glucose in three steps as we reported earlier [10]. Hydrolytic removal of the isopropylidene protective groups in both 5 and 7 afforded the corresponding lactols 6 and 8 in almost quantitative yields. Both products, and particularly stereoisomer 6, are rather hygroscopic. They were therefore used in the next synthetic step immediately after their brief isolation. Accordingly, lactol 6 was submitted to the reaction with a stabilized ylide, (methoxycarbonylmethylene)triphenylphosphorane (MCMP) in dry methanol, to afford (+)-crassalactone D (1) along with the corresponding 4-epimer 2 in 1:2 respective ratio and in 51% combined yield. This one-pot sequence is comprised of an initial Zselective Wittig olefination [11,12], followed by successive γ-lactonization, β-elimination and the final 5-endo-trig spirocyclization [5,10]. In order to increase the yield and/or epimer ratio we carried out the olefination step with (ethoxycarbonylmethylene)triphenylphosphorane (ECMP) [13,14]. However, the isomers 1 and 2 were isolated in the similar total yield (48%) and in the same isomer ratio (1/2 = 1:2).

<Insert Scheme 1>
In an attempt to generalize this methodology by using a similar substrate, but with the opposite stereochemistry at the C-5 position, we treated lactol 8 under the same Wittig olefination conditions. Gratifyingly, reaction of 8 with MCMP gave unnatural spiro-lactones 3 and 4 in 67% combined yield. Unfortunately, a drop in selectivity was observed, whereupon an equimolar mixture of 3 and 4 was obtained. However, when the Wittig olefination of 8 was performed with ECMP, products 3 and 4 were obtained in respective isomeric ratio of 1:4, in slightly lower total yield (46%). The products 1 and 2, as well as 3 and 4, were separated with difficulties. In order to obtain pure products several flash column chromatographic separations were required, followed by preparative TLC. This significantly decreased the yields of all products, and in particular the yields of minor stereoisomers. Fortunately, it was found that the less stable isomer 2 could be converted to the more stable 1, after treatment of their mixture (2:1 in favour of 2) with a solution of trifluoroacetic acid in chloroform, to give a 2:1 mixture of 1 and 2 (by 1 H NMR analysis). The similar results were obtained when the reaction was carried out in the presence of Lewis acids (for more details, see the Supplementary data). Greater stability of naturally occurring 1 compared to 2, could be a result of stabilization by anomeric effect and the related stereoelectronic effects [15,16].
In contrast, when pure 4 was treated under similar reaction conditions, a 1:1 mixture of 3 and 4 was obtained indicating a similarity in stability of both stereoisomers.
Both 1 H and 13 C NMR data and physical properties of compounds 1 and 2 were in agreement with those reported previously [2,3]. Stereochemistry of lactones 3 and 4 were confirmed by single crystal X-ray diffraction analysis (for the crystal structures of 3 and 4, see the Supplementary data).
Two independent procedures were used for the conversion of spiro-lactones 1-4 to the cinnamic esters 1a-d and 2a-d and the results are presented in Table 1.
<Insert Table 1> In the first experiments, a mixture of stereoisomeric lactones 1 and 2 (in 1:2 respective ratio) was treated with cinnamoyl chloride (entry 1) or with 4-nitrocinnamoyl chloride (entry 2) in the presence of DMAP in dry dichloromethane, to afford mixtures of 1a and 2a or 1b and 2b in 72 and 85% combined yields, respectively.
Next, the esterification step was carried out by using the Steglich esterification protocol [17]. We were pleased to find that treatment of mixture of 1 and 2 (in 1:2 respective ratio) with 4methoxycinnamic acid, in the presence of DCC and DMAP in anhydrous dichloromethane, gave the corresponding esters 1c and 2c in 32 and 60% respective yields (entry 3). Finally when the mixture of 1 and 2 was treated with 4-fluorocinnamic acid, under the same reaction conditions as those described above, the expected products 1d and 2d were obtained in 36 and 50% yields, respectively (entry 4).
The results related to the synthesis of 6-O-cinnamoyl derivatives 3a-d and 4a-d are presented in Table 2.
<Insert Table 2> Treatment of 3 with cinnamoyl chloride or with 4-nitrocinnamoyl chloride in the presence of DMAP in dry dichloromethane, gave the corresponding 6-O-cinnamoyl derivatives 3a and 3b in 68 and 74% respective yields (entries 1 and 2). The same synthetic protocol was then used for the conversion of 4 in 4a and 4b in 62 and 89% yields, respectively (entries 5 and 6). Next, the Steglich esterification protocol was used for the synthesis of the corresponding 6-O-(4-methoxycinnamoyl) or 6-O-(4-fluorocinnamoyl) derivatives (Table 2,  It should be noted that all C-6 signals of (7S)-isomers (3, 4, 3a-d, 4a-d) appeared downfield compared to the corresponding signals of (7R)-isomers (1, 2, 1a-d, 2a-d), presumably due to the steric compressions caused by the OH and Ph groups.
Structure and stereochemistry of esters 1b-d, 2a and 4a were confirmed by single crystal X-ray diffraction analysis (for the crystal structures see the Supplementary data). <Insert Table 3> Crassalactone D (1) and stereoisomers 2 and 4 exhibited submicromolar activity against K562 cell line (Table 3). Natural lactone 1 was 2-fold more active than commercial antitumour agent DOX, 5fold more active with respect to 2 and 4 and 10-fold more potent than 3. Compound 1 (IC 50 0.34 μM) was 6-, 14-and 104-fold more potent compared to 2, 4 and 3, respectively in the culture of  Finally, we considered influence of absolute stereochemistry at C-4 on antitumour activity. The data in Table 3 indicate that (4R)-stereoisomers showed higher potency toward majority of investigated cell lines compared to (4S)-isomers with one exception. Namely, 4-nitrocinnamoyl-crasslactone D (1b) exhibited higher cytotoxicity against five of eight cell lines under investigation compared to (4R)-analogue (2b).

In vitro antitumour activities and SAR
Our previous studies [8][9][10] indicated that the styryl lactones having the (7S)-stereochemistry represent more potent cytotoxic agents with respect to the corresponding (7R)-epimers. The results obtained from comparison of the IC 50 values of (7R)-and (7S)-crassalactone D derivatives, which showed that (7S)-stereoisomers were are more active then (7R)-analogues toward the majority of investigated cell lines, agrees well with our previous findings [8][9][10].

Detection of apoptosis and apoptotic pathways
The ability of cancer cells to evade this programmed cell death is one of the major characteristics that enables their uncontrolled growth [19]. The efficiency of chemotherapy in killing such cells depends on the successful induction of apoptosis, since defects in apoptosis signalling represents one of the causes of drug resistance [20].
Recent report disclosed by Choo et al [21] showed that some of natural styryl-lactones isolated from The observed apoptosis-inducing effect of investigated compounds (Table 4A) Table 4B as the percentage of specific apoptosis and necrosis [22].
<Insert Table 4> Although natural product 1 just slightly increases percentage of K562 cells in sub G1 phase with respect to control, it induced apoptosis in 22 and intrinsic (mitochondrial) pathways [24]. The conversion of procaspase 3 to caspase 3 results in the generation of the active 'executioner' caspase that subsequently catalyzes the hydrolysis of many protein substrates [24] and other downstream targets including PARP which primary function is to detect and repair DNA damage. Cleavage of PARP by caspases is considered to be a hallmark of apoptosis [25].
<Insert Figure 2> Western blot analysis revealed that five (1, 4a-d) of 20 evaluated compounds decreased expression of both anti-apoptotic Bcl-2 protein and pro-apoptotic Bax protein in K562 cells compared with control (Fig. 2). Compounds 1, 4a-d increased expression of activated caspase 3 or 85 kDa fragment of cleaved PARP or both which allowed us to hypothesize that these compounds induce apoptosis in caspase-dependent way .  Lactones 1a, 2a, 2d, 3, 3a, 3b, 3d, 4 decreased expression of anti-apoptotic Bcl-2 protein and increased of pro-apoptotic Bax protein while 1b-d, 2, 2b, 2c and 3c increased expression of both Bcl-2 and Bax. Majority of these compounds induced over-expression of activated caspase 3 subunit or 85 kDa fragment of cleaved PARP or both (with the exception of 1b and 3a which decreased both). Hence, these results have suggested that the triggered caspase-dependent apoptotic cell death was influenced with the Bcl-2 protein family.

Conclusions
A novel synthesis of crassalactone D (1) and its epimers 2-4 has been achieved starting from the commercially available chiral template, diacetone-D-glucose. Synthesis of natural 1 has been achieved employing a new one-pot process that started with a Z-selective Wittig reaction of lactol 6 with MCMP or ECMP as a key step, followed by acid-promoted equilibration of dominant isomer

X-ray Crystal Structure Analysis
Diffraction experiments were performed on an Oxford Diffraction Gemini S diffractometer. Crystal structures were solved by SHELXT [26] or SIR92 [27] and refined with SHELXL [28]. See the

MTT assay
The colorimetric MTT assay was carried out following reported procedure [18].

Cell cycle analysis
K562 cells were treated with tested compounds for 72 h at their IC 50 concentrations. After treatment, K562 cells were washed in cold PBS and incubated for 30 min in 70% ethanol on ice, centrifuged and incubated with 500 µL Rnase A (100 units/mL) and 500 µL propidium iodide (400 µL/mL) for 30 min at 37 °C. Cell cycle was analyzed by FACS Calibur E440 (Becton Dickinson) flow cytometer and the Cell Quest software. Results were presented as percentage of cell cycle phases.

Detection of apoptosis
Apoptosis of K562 cells was evaluated with an Annexin V-FITC detection kit. Treated cells from each sample were collected (800 rpm/5 min, Megafuge 1.0R, Heraeus, Thermo Fisher Scientific) and pellet was re-suspended in 1mL of phosphate buffer (PBS, pH 7.2). K562 cells were washed twice with cold PBS and then re-suspended in binding buffer to reach the concentration of 1×10 6 cells/mL. The cell suspension (100 µL) was transferred to 5mL culture tubes and mixed with Annexin V (5 µL) and propidium iodide (5 µL). The cells were gently vortexed and incubated for 15 min at 25 °C. After incubation, 400 mL of binding buffer was added to each tube and suspension was analyzed after 1 h on FACS Calibur E440 (Becton Dickinson) flow cytometer. Results were presented as percent of Annexin V positive gated cells. Percentage of specific apoptosis was calculated according to Bender et al. [22].

Western blot
For the Western blot, 50 µg of proteins per sample were separated by electrophoresis and electrotransferred to a PVDF membrane Hybond-P and then blotted with primary antibodies against Bcl-2, Bax, caspase 3 and PARP. β-Actin was used as internal control. Proteins were detected by an enhanced chemiluminescence (ECL Plus) kit that includes peroxidase-labelled donkey anti-rabbit and sheep anti-mouse secondary antibodies. Blots were developed with an ECL Plus detection system and recorded on the Amersham Hyperfilm. Images of protein expression were analyzed in ImageJ computer program (NIH Image, http://imagej.nih.gov) after minor levels adjustments.
Expression of proteins was measured by densitometry and compared with control sample.