Synthesis and evaluation of antimigratory and antiproliferative activities of lipid-linked [13]-macro-dilactones

Abstract

The biological activities of a family of novel, lipid-linked 13-membered-ring macro-dilactones are reported. These [13]-macro-dilactones were synthesized by diacylation of functionalized diols, followed by ring-closing metathesis under conditions we had previously reported. Antimigratory, cytostatic and cytotoxic activities of the compounds against cancer cells were evaluated. Compound 13 was the most potent in the series, while compound 10 had the broadest concentration range of subtoxic antiproliferative activity. These compounds share common structural components, namely the [13]-macro-dilactone templated by an octyl α-glucoside 4,6-diol.

A general objective of bioorganic chemistry and chemical biology is to generate synthetic small molecules that mimic the robust biological activities of complex natural products. Macrolactones are an enticing class of molecules in this regard because they possess a variety of molecular architectures and activities. Consideration of migrastatin (1)¹ and iso-migrastatin (iso-1)² (Figure 1) illustrates some key points. These macrolactones are products of a polyketide synthase pathway.³ The macrocycles thus contain an even number of atoms; 1 has a 14-membered ring and iso-1 has a 12-membered ring. As their names imply, these and related molecules have been shown to inhibit the migration of a variety of human and mouse cancer cells^1a,4a–g,4h and tumor metastasis in mouse xenograft models.^4d,5 Cell migration is a fundamental biological phenomenon involved in a range of normal and disease states, including embryonic development, wound healing, blood vessel formation (angiogenesis), immune system function and tumor metastasis. Cell migration has become recognized as a process that could be targeted for drug development, particularly for the treatment of cancer at the level of angiogenesis (which is involved in the early development of malignant solid tumors and depends on the growth and migration of vascular endothelial cells) and cancer cell invasion and metastasis.⁶

Figure 1.

Migrastatin (1), iso-migrastatin (iso-1), “macroketone” 2, quinic acid-based macrolide 3 and [13]-macro-dilactone 4.

The antimigratory activity of iso-1 exceeds that of 1 by several orders of magnitude,^4g and 1 is thought to be a shunt metabolite of iso-1.^3b A number of synthetic and semisynthetic analogs of migrastatin have also been shown to possess considerable antimigratory activity.4b–h Recently, the synthetic macroketone 2, which potently inhibits cell migration,^4b has been shown to bind and inhibit fascin, a crosslinker of actin filaments involved in cell migration, suggesting that fascin could be an antimetastatic drug target.⁵ In addition, the natural product-like molecule 3, which contains a macrocycle fused to a quinic acid scaffold, is also a potent inhibitor of cell migration.^4e The ability of 2 and 3 to recapitulate the biological activity of the natural product itself suggests that other macrocycles of intermediate complexity may similarly inhibit the migration of cells.

We recently reported on the synthesis of some macro-dilactones that complement natural [12]- and [14]-macrocycles because they contain 13 atoms in their ring (e.g., 4 in Figure 1).⁷ A structural analysis of the new [13]-macro-dilactones demonstrated that they possessed helical chirality that arose from the size of the ring, the planar nature of their alkene and ester functional groups, and the stereochemistry of individual backbone stereocenters. As a consequence of the helical chirality, epoxidation of the double bond embedded in the macrocyclic backbone was highly selective. The unique molecular topology exhibited by the [13]-macro-dilactones and the fact that their ring size was similar to the migrastatin and iso-migrastatin isomers prompted us to consider their potential ability to inhibit cell migration. We were additionally inspired by the remarkable activity of the natural macrolactones despite their relatively sparse functionality and stereochemistry. Reported here are the synthesis and antimigratory and antiproliferative activities of a family of [13]-macro-dilactones (Figure 2). A unifying feature among the active [13]-macro-dilactones was the presence of an octyl (C₈) glycoside. It is worth noting that two natural products containing a [13]-macrolactone, spongidepsin⁸ and amphidinolide P,⁹ both have modest antiproliferative activity.

Figure 2.

[13]-Macro-dilactones and related molecules synthesized and evaluated in this study.

[13]-Macro-dilactones and related molecules evaluated in this investigation are shown in Figure 2. The molecules were designed to contain up to three components that might impart biological activity. The primary feature was the [13]-macro-dilactone itself, which is common among 4–13. The nature of the 1,3-diol used to template the macro-dilactone was also varied. Macrocycles 4 and 7, for example, utilized acyclic 1,3-diols from 1,3-butane diol and an acylated threoninol,¹⁰ respectively. The 1R,2S and 1S,2S isomers of 2-hydroxycyclohexyl methanol¹¹ were used to template 5 and 6. The remainder of the macrocycles were templated via the C4 and C6 hydroxyls of a protected glucoside, as in 8–13. The attachment of a lipophilic alkyl chain to the molecule represented the final structural feature.

The [13]-macro-dilactones in Figure 2 were synthesized by a strategy we have previously described.⁷ Scheme 1 depicts the synthesis of [13]-macro-dilactone 10. It is illustrative of the approach taken for all the [13]-macro-dilactones. Preparation of 10 commenced with the protection of octyl glucoside 21¹² as its 4,6-O-benzylidene. The α-anomer of the benzylidene, 19a, was isolated in 20% yield, as was the β-anomer, from an initial ~1:1 mixture of C1-anomers. The C2 and C3 hydroxyl groups were then efficiently converted to their corresponding benzyl ethers to deliver 20 (92% yield). The 4,6-O-benzylidene of 20 was then removed, followed by acylation of the exposed hydroxyl groups with 4-pentenoic acid to give 14 in 60% yield over two steps. Ring-closing metathesis of the alkene moieties in 14 provided 10 (50% yield).

Scheme 1.

Synthesis of octyl glucoside-fused [13]-macro-dilactone 10

With the [13]-macro-dilactones in hand, the next objective was to assay their antimigratory activity. In addition, intermediates 14–20 from the synthesis of the macrocycles were also tested. We employed a scratch-wound assay, where a small wound – mechanically scratched in a cell monolayer – triggers cell migration and closure of the wound. The progress of wound closure was followed over time,¹³ essentially as previously described.¹⁴ The activity of these compounds was evaluated in BT-20 human breast carcinoma cells, T47D human breast carcinoma cells, MDA-MB-231 human breast carcinoma cells, MDA-MB-435 human melanoma cells, 4T1 mouse breast carcinoma cells and Madin-Darby canine kidney (MDCK) epithelial cells. Compounds 10, 11, 12, 13, 19a and 19b displayed weak antimigratory activity in BT-20, MDA-MB-435 and MDCK cells but not in T47D cells, MDA-MB-231 or 4T1 cells. Of these, 11, 12 and 13 were the most bioactive, and they appeared to have the greatest activity in BT-20 cells of all the cell lines tested. The activity of these compounds was thus further examined in BT-20 cells. The rate of wound closure over a range of compound concentrations was determined from digital microscope images.^14,15 Potential cytotoxicity was determined at the end of each experiment. The concentration-response profiles revealed a very narrow range of subtoxic antimigratory activity between the minimum inhibitory concentration (MIC) and the minimum lethal concentration (MLC), as shown in Table 1. This precluded determination of meaningful half maximal inhibitory concentration (IC₅₀) values for the compounds’ antimigratory activity. We concluded that the effects on wound closure observed at subtoxic concentrations were likely due to incipient toxicity.

Table 1.

Antimigratory activity of compounds 11, 12 and 13 in BT-20 breast cancer cells

Compound	MICa (µM)	MLCb (µM)
11	75	125
12	50	125
13	50	75

^aMinimum inhibitory concentration defined as the lowest concentration at which there was a statistically significant decrease by Student’s t-test in the mean rate of migration in the wound closure assay compared to parallel controls from three independent experiments for each compound.

^bMinimum lethal concentration defined as the lowest concentration at which there was evidence of cytotoxicity based on the trypan blue dye exclusion assay conducted at the end of each wound closure experiment.

We next evaluated the effects of the compounds on the viability and growth of BT-20 cells.¹⁶ We first tested these compounds at 100 µM and found that compounds 9, 10, 11, 12, 13, 15, 17, 19a and 19b displayed either subtoxic antiproliferative (cytostatic) or cytotoxic activity at this concentration (Figure 3). Compounds that were considered cytostatic at 100 µM (9, 10, 11, 12 and 19b) reduced the rate of cell proliferation over 48 h compared to controls treated with dimethyl sulfoxide (DMSO) alone but did not reduce cell numbers below the initial values. Compounds that were considered cytotoxic at this concentration (13, 15, 17 and 19a) not only inhibited cell growth but also caused a reduction in cell numbers from the initial values due to cell death.

Figure 3.

Antiproliferative activity at 100 µM in BT-20 breast cancer cells. Cells were plated at uniform density onto 96-well tissue culture plates and allowed to attach and grow for 24 h. At that point, the mean cell number (“initial”) was determined in a tetrazolium salt-based assay. Compounds were added to experimental cultures, which were grown for another 48 h, along with parallel controls that had been treated with 1% DMSO alone. Values represent means with error bars signifying standard errors of the mean for three independent experiments for each treatment (with four replicate wells for each experiment for 12 total), normalized to the mean initial cell number. Black bars represent compounds that had no statistically significant effect on cell growth. Blue bars represent compounds that appeared cytostatic at 100 µM, defined as mean cell numbers that were significantly lower by Student’s t-tests than the mean control cell number at 48 h but not significantly reduced from the mean initial cell number. Red bars represent compounds that appeared cytotoxic at 100 µM, defined as cell numbers that were significantly lower by Student’s t-tests than the mean initial cell number.

Compounds 10, 11, 12, 13, 15, 17, 19a and 19b exhibited either pronounced cytostatic or cytotoxic activity in the initial assay at 100 µM and were tested over a range of concentration to establish concentration-response profiles for each compound. Compound 9 only very weakly inhibited cell proliferation, and, when tested over a range of concentrations, the concentration-response profile was virtually flat.¹⁷ Compounds 10, 11, 12, 13, 15, 17, 19a and 19b, on the other hand, displayed cytostatic activity that in most cases appeared separable from the cytotoxicity observed at higher concentrations (Table 2). The criterion for this conclusion was based on the MLC/MIC ratio (“therapeutic index,” often also defined as half-maximal lethal concentration divided by the IC₅₀). By this measure, therefore, these compounds have subtoxic cytostatic activity, particularly in the case of 10, with a therapeutic index of 30.

Table 2.

Activities of antiproliferative compounds in BT-20 breast cancer cells

Compound	MICa (µM)	IC50b (µM)	95% Confidence intervalc (µM)	MLCd (µM)	Therapeutic index (MLC/MIC)
10	10	85.1	82.8 – 87.4	300	30
11	50	91.4	89.2 – 93.6	300	6
12	50	107.0	102.6 – 111.5	>300	>6
13	10	38.1	37.3 – 38.8	50	5
15	25	52.9	51.2 – 53.8	75	3
17	25	61.1	60.1 – 62.1	100	4
19a	10	47.1	43.8 – 50.1	75	7.5
19b	50	72.1	69.4 – 74.8	300	6

^aMinimum inhibitory concentration, defined as the lowest concentration at which there was a statistically significant reduction by Student’s t-test in mean cell number from the mean control cell number at 48 h from three independent experiments (with four replicates for each treatment per experiment for 12 replicates total) in a tetrazolium-salt-based assay.

^bHalf-maximal concentration for inhibition of cell growth, calculated from mean cell numbers at 48 h for a range of concentrations of each compound.

^c95% confidence interval for the IC₅₀.

^dMinimum lethal concentration, defined as the lowest concentration at which there was a statistically significant reduction by Student’s t-test in mean cell number at 48 h from the mean initial cell number.

The compounds possessing antiproliferative activity, 10, 11, 12, 13, 15, 17, 19a and 19b, share common structural components that had been included in their design. To our surprise, the [13]-macro-dilactone unit was not essential for activity. All of the bioactive compounds contain a glucosyl unit and an octyl (C₈) chain. Furthermore, the majority of these molecules contain an α-linkage between the glucose unit and the alkyl chain. However, the most promising compounds, as measured by either therapeutic index (compound 10) or IC₅₀ (compound 13), contained the [13]-macro-dilactone. In the case of 10 in particular, the cytostatic and cytotoxic activities were clearly separable, with a wide concentration range for its subtoxic antiproliferative activity. Based on this and its micromolar IC₅₀, it is likely that 10 does not act by merely having a nonspecific effect on cells, such as disrupting the integrity of cellular membranes. Instead, this compound may target some factor(s) involved in cell cycle progression. In summary, we have demonstrated the application of our published synthesis to access a novel class of antiproliferative agents.