Synthesis and Biological Evaluation of Rigid Polycyclic Derivatives of the Diels-Alder Adduct Tricyclo[6.2.1.02,7]undeca-4,9-dien-3,6-dione

Abstract

Part of our research program concentrates on the discovery of new bioactive compounds prepared either by total synthesis or molecular transformation of compounds with bioactivity profiles. In this work we have focused our interest on chemical transformations of the Diels-Alder adduct tricyclo[6.2.1.0^2,7]undeca-4,9-dien-3,6-dione in order to obtain cage-like compounds and derivatives, followed by an evaluation of their biological activity.

Introduction

Since the discovery of the Diels-Alder reaction [1], it has become one of the most employed strategies in organic synthesis [2]. This all began in 1928, when the cycloaddition between cyclopentadiene (1) and p-benzoquinone (2), forming mostly the endo adduct tricyclo[6.2.1.0^2,7]-undeca-4,9-dien-3,6-dione (3), as shown in Scheme 1, was described.

Scheme 1. Preparation of tricyclo [6.2.1.0^2,7]undeca-4,9-dien-3,6-dione.

Scheme 1. Preparation of tricyclo [6.2.1.0^2,7]undeca-4,9-dien-3,6-dione.

A unique aspect of adduct 3 is its high symmetry, which allows for facile selective reactions at one or both carbonyl groups by means of classical and non-classical reagents, including microbial transformations [3]. Another feature to be considered is its cage-like framework, which forces functional groups into close spatial proximity, facilitating subsequent reactions, as exemplified, for instance, by the synthesis of Cookson’s diketone (4) [4] (Scheme 2).

Scheme 2. Preparation of Cookson’s diketone (4).

Scheme 2. Preparation of Cookson’s diketone (4).

Although compound 3 has been extensively studied, its use in the medicinal chemistry field has not been exploited very much. On the other hand, there is considerable interest nowadays in studying the biological activity of cage-like rigid polycyclic molecules, which owing to their structural peculiarities, may interact in a specific way with biological receptors and/or trap smaller chemical species in their interior. It is known, for example, that the incorporation of a rigid carbogenic structure in biologically active compounds often increases their physiological activity. One of their advantages is a slowing down of metabolic degradation and consequent prolongation of the activity [5,6,7].

Figure 1. Examples of bioactive cage-like polycyclic compounds.

Figure 1. Examples of bioactive cage-like polycyclic compounds.

Considering the important antiviral activity of amantadine (5) and rimantadine (6), additional studies related to cage-like polycyclic compounds led to the discovery of other compounds with similar properties, such as the pentacycloundecanes 7 and 8 [6] (Figure 1).

Accordingly, as a part of a research program aimed at finding new synthetic compounds with interesting bioactivity profiles, we decided to investigate the preparation of some rigid polycyclic oxygenated compounds, taking advantage of the structural versatility of the Diels-Alder adduct 3. The central idea in this work was to make accessible new structurally complex molecules which are both easy to prepare and endowed with potential biological activity.

Results and Discussion

Chemistry

The reaction shown in Scheme 1 is widely known in literature [2,8,9,10,11], however, from an experimental point of view, a few drawbacks can be noted, such as the use of toxic solvents and the need to perform the reaction at very low temperatures. These features prompted us to search for a more efficient and higher yielding methodology to generate 3. As a result of these studies the cycloaddition reaction was performed in a 5:1 mixture of hexane-ethyl acetate to both force the kinetic control of the reaction and to solubilize the reagents. This procedure led to a 97% yield of the endo isomer after recrystallization from hexane, as proven by its ¹H-NMR spectrum.

Scheme 3 and Scheme 4 show some of the compounds chosen for chemical transformation and screening for biological activity that were prepared from adduct 3. Compounds 9 and 10 (Scheme 3) were easily obtained using described methods [3,12]. As one may observe, the keto-alcohol 10 possesses appropriate structural and chemical functionality for the insertion of new groups and building rigid cage-like polycyclic molecules, in conformity with our initial plan.

Scheme 3. Selective reductions of tricyclo[6.2.1.0^2,7]undeca-4,9-dien-3,6-dione (3).

Scheme 3. Selective reductions of tricyclo[6.2.1.0^2,7]undeca-4,9-dien-3,6-dione (3).

The structures of compounds 3, 9 and 10 were confirmed by comparison of their spectroscopic data with those reported in the literature [8, 12 and 3, respectively]. Next, and starting from compound 10, five rigid polycyclic compounds were obtained, possessing various and diverse chemical functionalities, as shown in Scheme 4.

Scheme 4. Syntheses of derivatives of keto-alcohol 10.

Scheme 4. Syntheses of derivatives of keto-alcohol 10.

Gallate 12 was prepared by submitting 10 to simple esterification in refluxing toluene with the acid chloride 11 in the presence of pyridine. The structure and stereochemistry of the resulting ester 12 were assigned based on NMR spectral analyses, including HSQC, HMBC, COSY and NOESY experiments. The ¹H-NMR spectrum showed a multiplet at 5.52 ppm, assigned to the hydrogen attached to C6, which is probably deshielded by the ester carbonyl. It also showed a multiplet (4H) at 1.95-2.45 ppm, assigned to the two methylene groups at C4 and C5. The two doublets at 1.26 and 1.37 ppm (J = 8.4 Hz) correspond to the geminal methylene protons at C11. The broad signal at 3.24 ppm was ascribed to the hydrogen attached to C1 and the signals of the protons on C2, C7 and C8 were superimposed at around 2.91-3.03 ppm. The olefinic hydrogens of C9 and C10 appear as double-doublets at 5.93 and 6.2 ppm, respectively, while the aromatic ring singlets and methoxyls are observed at 7.21 and 3.83 ppm. A comparison of the NMR data of substance 12 with that of a similar tricyclic ketoester described by Marchand et al., which contains an endo-6-OC(O)Ar group were Ar = 3,5-dinitrophenyl [3], also confirmed the structure proposed for compound 12. Compound 13 was obtained along with 12 (Scheme 4), probably due to the mild acid conditions of reaction mixture favoring protonation of double bond followed by ring closure via the hydroxyl group [13] (Scheme 5). With the purpose of verifying this proposal, we checked the stability of compound 10. We have observed that 10 is stable in pure ethyl acetate, however, it is rapidly converted into 13 when a small amount of p-toluenesulfonic acid is added to the solution.

Scheme 5. Mechanistic proposal for the formation of compound 13.

Scheme 5. Mechanistic proposal for the formation of compound 13.

The structure of compound 13 was confirmed on the basis of its spectral data. In contrast to 10, no olefinic proton signals were observed in the ¹H-NMR spectrum. As main signals a multiplet at 4.22 ppm, ascribed to the hydrogen attached to C9, and the triplet at 4.41 ppm assigned to the hydrogen at C6 were observed. The ¹³C-NMR data were in agreement with the proposal structure. The appearance of C10 as an additional secondary carbon (CH₂) at 35.1 ppm was decisive for confirmation of the 10 → 13 transformation. Moreover, the mass spectrum registered the expected molecular ion (M⁺) at m/z = 178.

The bromoether 14 was obtained in quantitative yield from alcohol 10 by bromination with NBS in dichloromethane [14]. Treatment of tetracycle 14 with NaH in THF under reflux induced abstraction of the acidic proton at C2 and subsequent intramolecular cyclopropanation leading to formation of compound 15. Its structure was confirmed by spectroscopic studies comprising 2D experiments (COSY, HSQC, HMBC and NOESY). The ¹H-NMR spectrum showed the hydrogen at C6 as a multiplet at 4.35 ppm, whereas the C9 hydrogen was a double doublet at 4.55 ppm. The hydrogen of the cyclopropane ring (C10-H) is in a current shielding region and can be seen at 1.69 ppm, coupled with the C9-H (J = 5.0 Hz). The triplet at 2.51 ppm was attributed to the C8-H that is coupled with the C9-H and C7-H (J = 2.8 Hz). The ¹³C-NMR spectrum showed signals in accordance with the chemical transformation performed, mainly at 21.8 (C1), 33.1 (C2) and 35.3 (C10) ppm (see Table 1). The mass spectrum showed the molecular ion peak (M⁺) at m/z = 176.

When we attempted to prepare the highly strained compound 17 by reduction of the carbonyl in 14 from the convex face, followed by displacement of the bromide, only the known bromohydrin 16 [15] was formed (Scheme 4). Even forcing conditions, such as treatment of 16 with NaH in THF under reflux, failed to yield 17.

Bromohydrin 16 (m.p. 136-138 ^oC; Lit. [15], m.p. 133-134^oC), was characterized by means of its spectroscopic data (¹H- and ¹³C-NMR, IR) and mass spectrum (cf. Experimental Part). Its mass spectrum showed molecular ion peak at m/z 258/260 and a (M -18)⁺ peak at m/z 240/242, due to loss of one water molecule from the hexane ring system.

Table 1 list the experimental ¹³C-NMR data of compounds 13, 14, 15 and 16, allowing a rapid comparison of the chemical shifts of all carbons of these molecules and supporting the main transformations carried out on the 6,9-epoxytricyclo[6.2.1.0^2,7]undecane ring system (Figure 2).

Figure 2.

Figure 2.

Table 1. ¹³C-NMR (CDCl₃) spectral data of the compounds 13, 14, 15 and 16 ^a

Positionb	13 Cδ	14 Cδ	16 Cδ	15 Cδ
1	39.3	42.2	42.9	21.8
2	51.2	49.8	39.5	33.1
3	214.0	211.8	69.7	209.4
4	37.4	35.0	26.7	38.3
5	26.5	25.9	26.1	30.4
6	74.4	75.3	76.1	74.1
7	43.5	46.9	44.2	42.0
8	49.1	48.6	49.1	46.3
9	79.7	88.4	88.7	82.2
10	35.1	55.1	56.5	35.3
11	37.2	34.9	35.4	30.3

^a Data acquired at 75 MHz.^b Confirmed by bidimensional NMR techniques. Signal multiplicity of carbons in all compounds was assigned by DEPT-135 experiments.

Table 1. ¹³C-NMR (CDCl₃) spectral data of the compounds 13, 14, 15 and 16 ^a

Positionb	13 Cδ	14 Cδ	16 Cδ	15 Cδ
1	39.3	42.2	42.9	21.8
2	51.2	49.8	39.5	33.1
3	214.0	211.8	69.7	209.4
4	37.4	35.0	26.7	38.3
5	26.5	25.9	26.1	30.4
6	74.4	75.3	76.1	74.1
7	43.5	46.9	44.2	42.0
8	49.1	48.6	49.1	46.3
9	79.7	88.4	88.7	82.2
10	35.1	55.1	56.5	35.3
11	37.2	34.9	35.4	30.3

^a Data acquired at 75 MHz.^b Confirmed by bidimensional NMR techniques. Signal multiplicity of carbons in all compounds was assigned by DEPT-135 experiments.

Biological Evaluation

1. Assay for cytotoxic activity

Cytotoxic analyses by the MTT method are used in the screening program of the U.S. National Cancer Institute (NCI), which tests more than 10,000 samples per year [16]. It is a fast, sensitive and cheap methodology, described for the first time by Mosman [17] in 1983 and subsequently modified in 1996 by Alley et al. [18]. This evaluation allows one to easily determine the cytotoxicity of a particular compound, but it does not provide any insight into the mechanism of action [19].

The cytotoxic activity of compounds 3, 9, 10, 12, 14, 15, 16 and 18 are listed in Table 2, with the respective IC₅₀ values. Among the eight samples analyzed, compound 3 showed moderate albeit unspecific cytotoxicity, with IC₅₀ values of 11.34, 13.62 and 13.52 μg/mL towards the HL-60, HCT-8 and MDA-MB 435 cell lines, respectively. Similarly, the novel gallate 12 was cytotoxic towards the HCT-8, SF-295 and MDA-MB 435 cell lines, with IC₅₀ values of 6.68, 7.58 and 13.51 μg/mL, respectively.

Figure 3. Methyl gallate.

Figure 3. Methyl gallate.

On the other hand, methyl gallate 18 (Figure 3) did not show any cytotoxic effects on the cell growth of the tested cell lines, suggesting that the cage structure present in 12 may play an important role in any observed increase of activity. Compound 9 was noticeably cytotoxic towards the MDA-MB 435 cell line with an IC₅₀ of 10.4 μg/mL. On the other hand, the other samples did not show important cytotoxic effects.

Table 2. Cytotoxicity of compounds 3, 9, 10, 12, 14, 15, 16 and 18, based on inhibition of tumor cell growth.

Compound	Cellsa IC50b [µg/mL, confidence interval]
	HL-60	HCT-8		SF-295		MDA-MB 435
3	11.34 (8.68 – 14,80)		13.62 (5.52 – 33.63)		> 25		13.52 (11.57 – 15.80)
9	> 25		> 25		> 25		10.47 (4.80 – 22.05)
10	> 25		> 25		> 25		> 25
12	> 25		6.68 (4.22 – 10.56)		7.58 (6.12 – 9.39)		13.51 (11.16 – 16.37)
14	> 25		> 25		> 25		> 25
15	> 25		> 25		> 25		> 25
16	> 25		> 25		> 25		> 25
18	> 25		> 25		> 25		> 25

^a Cells were plated in 96-well plates incubated with a 5% CO₂ atmosphere, at 37ºC, for 72h, in presence of concentrations of pure compounds (0.39-25 μg/mL). Each concentration was tested in triplicate and the analyses were performed in duplicate.^b The IC₅₀ corresponds to inhibitory concentration for 50% of cellular growth (95% Confidence Interval).

Table 2. Cytotoxicity of compounds 3, 9, 10, 12, 14, 15, 16 and 18, based on inhibition of tumor cell growth.

Compound	Cellsa IC50b [µg/mL, confidence interval]
	HL-60	HCT-8		SF-295		MDA-MB 435
3	11.34 (8.68 – 14,80)		13.62 (5.52 – 33.63)		> 25		13.52 (11.57 – 15.80)
9	> 25		> 25		> 25		10.47 (4.80 – 22.05)
10	> 25		> 25		> 25		> 25
12	> 25		6.68 (4.22 – 10.56)		7.58 (6.12 – 9.39)		13.51 (11.16 – 16.37)
14	> 25		> 25		> 25		> 25
15	> 25		> 25		> 25		> 25
16	> 25		> 25		> 25		> 25
18	> 25		> 25		> 25		> 25

^a Cells were plated in 96-well plates incubated with a 5% CO₂ atmosphere, at 37ºC, for 72h, in presence of concentrations of pure compounds (0.39-25 μg/mL). Each concentration was tested in triplicate and the analyses were performed in duplicate.^b The IC₅₀ corresponds to inhibitory concentration for 50% of cellular growth (95% Confidence Interval).

2. Assay of hemolytic activity in Swiss mouse (Mus musculus) erythrocytes

In order to verify whether the observed cytotoxic and antimitotic activities are related to membrane disruption, compounds were tested for their ability to induce lysis of mouse erythrocytes. The erythrocyte membrane is a dynamic structure that can dictate significant changes in its interaction with drugs [20,21]. The results revealed the absence of hemolytic activity for compounds 3, 12, 9, 16 and 18 at the highest tested concentration (200 μg/mL), suggesting that both cytotoxic and antimitotic activities were not related to the lytic properties or membrane instability induced by these compounds, which may be probably caused by a more specific pathway. The others compounds were not tested.

Conclusions

From the results presented it may be concluded that a polycyclic framework could be a useful scaffold for generating new lead compounds with biological activity and further synthetic studies are warranted, aimed at the preparation of hybrid compounds similar, for instance, to compound 12, which represents a kind of hybrid between a polycyclic and a trimethoxy residue found in some cytotoxic molecules such as podophyllotoxin and colchicine. Another topic of potential interest would be molecular modeling studies aimed at revealing possible structural modifications that could be made in the prepared compounds so they would be active in the receptor site(s) of proteins involved in the structure-cytotoxicity activity relationships, such as tubuline.

Experimental Section

General

¹H- and ¹³C-NMR spectra were obtained at 300 and 75 MHz respectively on a Bruker AVANCE DPX-300 spectrometer. Chemical shifts were referenced to TMS and coupling constants are given in Hertz. Infrared (IR) spectra were obtained on a Perkin Elmer equipment model 783. Mass spectra were recorded on a GC/Mass Spectrometer model QP5000 (Shimadzu). Solvents were purified before use and chemical reagents were commercially available and used without previous treatment. p‑Benzoquinone was prepared by oxidation of hydroquinone [22]. Column chromatography was performed using Merck 60 Å (70-230 mesh) silica gel, eluting with a hexane-ethyl acetate gradient. TLC analyses employed Merck (60F254/0.2 mm) sílica gel plates Melting points were recorded on Köffler equipment and are reported without correction.

Chemistry Synthesis of rel-(1S,2S,7R,8R)-tricyclo[6.2.1.0^2,7]undeca-4,9-dien-3,6-dione (3)

Cyclopentadiene (14.6 g, 221 mmol) was added to a solution of p-benzoquinone (22 g, 203 mmol) in hexane-ethyl acetate (5:1), cooled at 0^oC. The resulting mixture was stirred for 4 hours. Afterwards, the solvent was removed under vacuum and the product was recrystallized from hexane to give the title compound as yellow crystals: 97% yield; m.p.: 71-75^o C; ¹H-NMR (CDCl₃) δ (ppm): 1.46 (dt, J = 8.8/1.3 Hz, 1H), 1.55 (dt, J = 8.8/1.8 Hz, 1H), 3.20 (m, 2H), 3.50 (m, 2H), 6.09 (t, J = 1.8 Hz, 2H), 6.60 (s, 2H); ¹³C-NMR (CDCl₃) δ (ppm): 43.4 (CH), 46.3 (CH), 49.8 (CH₂), 67.7 (CH), 131.3 (CH), 136.6 (CH).

Synthesis of rel-(2S,3S,4S,5S,6S,10S,1R,9R)-2-bromine-3,6-epoxytricyclo[6.2.1.0^5,10]undecan-9-ol (16)

NaBH₄ (0.034g, 0.89 mmol) was added to an ice-bath-cooled solution of compound 14 [14] (0.153 g, 0.6 mmol) in MeOH (1.6 mL). The ice bath was then removed and the reaction mixture was stirred for 1 hour at room temperature, before being quenched by addition of a drop of acetic acid. The mixture was extracted with CH₂Cl₂ (3 x 2 mL) and washed with distilled water. The organic layer was dried over anhydrous MgSO₄, filtered and the solvent was removed under vacuum. The product was purified by flash chromatography on silica gel column using hexane-ethyl acetate (2:1) as eluent to afford a white amorphous solid: 72% yield; m.p.: 136-138°C; ¹H-NMR (CDCl₃) δ (ppm): 1.36-1.42 (m, 1H), 1.50 (d, J = 11 Hz, 1H), 1.59-1.67 (m, 3H), 1.97 (ddd, J = 17.8/6.5/3.2 Hz, 1H), 2.13 (d, J = 11 Hz, 1H), 2.30 (m, 1H), 2.36 (m, 1H), 2.61 (broad signal, -OH), 2.90 (m, 1H), 3.95 (dd, J = 17.8/7.8 Hz, 1H), 4.15 (m, 1H), 4.22 (d, J = 2.0 Hz, 1H), 4.54 (d, J = 5 Hz, 1H); ¹³C- NMR (CDCl₃) δ (ppm): 26.1 (CH₂), 26.7 (CH₂), 35.4 (CH₂), 39.5 (CH), 42.9 (CH), 44.2 (CH), 49.1 (CH), 56.5 (CH), 69.7 (CH), 76.1 (CH), 88.7 (CH); IR (KBr) ν (cm^-1): 3269, 2975-2869; MS m/z: 260 [(M + 2)⁺, ⁸¹Br], 258 (M⁺, ⁷⁹Br), 242 [(M – H₂O + 2)⁺, ⁸¹Br], 240 [(M – H₂O)⁺, ⁷⁹Br].

Synthesis of rel-(1S,3S,7S,2R,8R)-6-oxo-3-(3,4,5-trimethoxyphenylcarbonyloxy) tricyclo[6.2.1.0^2,7]-undec-9-ene (12) and rel- (1S,2S,6S,8S,7R,9R)-6,9-epoxytricyclo[6.2.1.0^2,7]undecan-3-one (13)

Under a nitrogen atmosphere, thionyl chloride (0.270 mL; 3.6 mmol) was added to a solution of 3,4,5-trimethoxybenzoic acid (0.385 g; 1.8 mmol) in CH₂Cl₂ (5.5 mL). The mixture was refluxed for 4 hours. After evaporation of the solvent under reduced pressure the acid chloride 11 was dissolved in anhydrous toluene (5.5 mL). To this solution it was added compound 10 (0.318 g, 1.8 mmol) and pyridine (5.0 mL) and the mixture was refluxed during 24 hours. The reaction was monitored by TLC and after completion it was diluted with dichloromethane and the resulting solution was washed with a saturated solution of CuSO₄. The organic layer was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The products were purified by silica gel column chromatography using hexane-ethyl acetate (1:1) as eluent. Compound 12: white solid: 15% yield; m.p.: 42-46 °C; ¹H-NMR (CDCl₃) δ (ppm): 1.26 (d, J = 8.4 Hz, 1H), 1.37 (d, J = 8.4 Hz, 1H), 1.92-2.05 (m, 2H), 2.16-2.29 (m, 2H), 2.91-3.03 (m, 3H), 3.24 (broad signal, 1H), 3.83 (s, 9H), 5.52 (m, 1H), 5.93 (dd, J = 5.3/2.7 Hz, 1H), 6.12 (dd, J = 5.3/2.9 Hz, 1H), 7,21 (s, 2H); ¹³C-NMR (CDCl₃) δ (ppm): 25.6 (CH₂), 35.6 (CH₂), 44.4 (CH), 45.0 (CH), 45.4 (CH), 49.6 (CH₂), 51.7 (CH), 56.1 (CH₃), 60.8 (CH₃), 70.0 (CH), 106.8 (CH), 125.0 (C), 135.4 (CH), 135.8 (CH), 142.2 (C), 152.9 (C), 165.1 (C), 211.5 (C=O). Compound 13: colorless oil: 9 % yield; ¹H-NMR (CDCl₃) δ (ppm): 1.31 (d, J = 10.4, 1H), 1.35 (d, J = 3.2 Hz, 1H), 1.44 (m, 1H), 1.50 (m, 1H), 1.82 (tdd, J = 13.0/13.0/5.4/1.3 Hz, 1H), 2.12 (m, 1H), 2.19 (m, 1H), 2.26 (d, J = 5.4 Hz, 1H), 2.37-2.58 (m, 3H), 2.89 (t, J = 4.0 Hz, 1H), 4.22 (m, 1H), 4.40 (t, J = 6.3 Hz, 1H); ¹³C-NMR (CDCl₃) δ (ppm): 26.5 (CH₂), 35.1 (CH₂), 37.2 (CH₂), 37.4 (CH₂), 39.3 (CH), 43.5 (CH), 49.1 (CH), 51.2 (CH), 74.4 (CH), 79.7 (CH), 214.0 (C=O); MS m/z: 178 [M⁺].

Synthesis of 6,9-epoxytetracyclo[6.2.1.0^2,7.0^2,10]undecan-3-one (15)

A solution of compound 14 (0.60 g, 2.3 mmol) in THF (1.6 mL) was added to a mixture of 60% NaH (0.11 g, 2.8 mmol) cooled in water-ice bath. The mixture was then stirred for 1 hour at room temperature and 6 hours under reflux before extraction with dichloromethane (3 x 2 mL). The organic layer was dried over MgSO₄, filtered and the solvent was removed on a rotavapor. The product was purified by silica gel column chromatography using hexane-ethyl acetate (2:1) as eluent: 76% yield. ¹H-NMR (CDCl₃) δ (ppm): 1.69 (ddd, J = 5.0/1.5/0.7 Hz, 1H), 1.73 (d, J = 11.3 Hz, 1H), 1.79 (d, J = 11.3 Hz, 1H), 2.03 (m, 1H), 2.30-2.44 (m, 5H), 2.51 (t, J = 2.8 Hz, 1H), 4.35 (m, 1H), 4.55 (dd, J = 5.0/2.8 Hz, 1H); ¹³C-NMR (CDCl₃) δ (ppm): 21.8, (CH), 30.3 (CH₂), 30.4 (CH₂), 33.1 (C), 35.3 (CH), 38.3 (CH₂), 41.8 (CH), 46.3 (CH), 74.1 (CH), 82.2 (CH), 209.4 (C=O); MS m/z: 176 [M⁺].

Biological Evaluation

1. Assay for cytotoxic activity

The cytotoxicity of the compounds was tested against HL-60 (human leukemia), HCT-8 (human colon), SF-295 (human central nervous system) and MDA-MB 435 (human breast cancer) cell lines obtained from the National Cancer Institute, Bethesda, MD, USA. Cells were grown in RPMI-1640 medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 μg/mL streptomycin and 100 μg /mL penicillin and incubated at 37 °C with a 5% CO₂ atmosphere. Cells were plated in 96-well plates (10⁵ cells/well for adherent cells or 0.5 x 10⁵ cells/well for suspended cells in 100 μL of medium). After 24 h, the compounds (0.39 to 25.0 μg/mL) dissolved in DMSO 1% were added to each well and incubated for 72 h. Control group received the same amount of DMSO. Doxorubicin (Doxolem^®, Zodiac Produtos Farmacêuticos S/A, Brazil) was used as positive control. Tumor cell growth was quantified by the ability of living cells to reduce the yellow dye 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) to a purple formazan product [17]. At the end of the incubation, the plates were centrifuged and the medium was replaced by fresh medium (200 μL) containing 0.5 mg/mL MTT. Three h later, the MTT formazan product was dissolved in DMSO (150 μL), and absorbance was measured using a multiplate reader (Spectra Count, Packard, Ontario, Canada). The drug effect was quantified as the percentage of the absorbance of reduced dye at 550 nm in relation to control wells.

2. Swiss mouse (Mus musculus) erythrocyte hemolysis assay

The test was performed in 96-well plates following the method described by Costa-Lotufo et al. [23]. Each well received 100 μL of 0.85% NaCl solution containing 10mM CaCl₂. The first well was the negative control that contained only the vehicle (DMSO 10%), and, in the second well, test substance (100 μL) half was added. The compounds were tested at concentrations ranging from 3.9 to 200 μg/mL. The serial dilution continued until the 11th well. The last well received 0.1% Triton X-100 in 0.85% saline (20 μL) to produce 100% hemolysis (positive control). Then, each well received a 2% suspension of mouse erythrocytes in 0.85% saline containing 10 mM CaCl₂ (100 μL). After incubation at room temperature for 30 min and centrifugation, the supernatant was removed and the hemoglobin liberated was measured spectroscopically as absorbance at 540 nM.

Statistical Analysis

The IC₅₀ values and their 95% confidence intervals (CI 95%) were obtained by nonlinear regression using the GRAPHPAD program (Intuitive Software for Science, San Diego, CA, U.S.A.). For the hemolytic assay, the differences were analyzed by ANOVA followed by Student Newman Keuls test compared to negative control at a significance level of 5%.