Madecassic Acid Derivatives as Potential Anticancer Agents: Synthesis and Cytotoxic Evaluation

Abstract

A series of novel madecassic acid (1) derivatives was synthesized, and their cytotoxicity was evaluated against the NCI-60 panel of cancer cell lines. Several analogues exhibited broad-spectrum cytotoxic activities over all nine tumor types represented in the panel, with more potent antiproliferative activities observed against selected cancer cell lines, including multidrug-resistant phenotypes. Among them, compound 29 showed GI₅₀ (50% growth inhibition) values ranging from 0.3 to 0.9 μM against 26 different tumor cell lines and selectivity for one colon (COLO 205) and two melanoma (SK-MEL-5 and UACC-257) cell lines at the TGI (total growth inhibition) level. The mode of action of 29 was predicted by CellMiner bioinformatic analysis and confirmed by biochemical and cell-based experiments to involve inhibition of the DNA replication process, particularly the initiation of replication, and disruption of mitochondrial membrane potential. The present findings suggest this novel madecassic acid derivative may have potential as an anticancer therapeutic lead for both solid and hematological tumors.

Graphical Abstract

Natural products and their derivatives remain a rich source of successful drug leads.^1–5 A historical assessment of all FDA-approved new molecular entities (NMEs) reveals that natural products and their analogues represent over one-third of all NMEs.⁶ In the past decade there has been enormous interest in terpenoid natural products and their semisynthetic derivatives due to their remarkable structural diversity and wide range of pharmacological activities.^7–10 Among all the terpenoids, pentacyclic triterpenoids are the most potent compounds endowed with anti-inflammatory and antitumor activities.^11–18 There are hundreds of publications and patents describing the synthesis and biological properties of pentacyclic triterpenoids, which is indicative of wide interest in these compounds by academic and pharmaceutical industry research groups.^19–21 Madecassic acid (1) is a naturally occurring pentacyclic triterpenoid found in the traditional medicinal plant Centella asiatica (L.) Urban.²² This metabolite is known to possess important pharmacological activities, such as wound healing,²³ antioxidant,²⁴ anti-inflammatory,²⁵ and antidiabetic²⁶ effects. Furthermore, a recent study²⁷ reported evidence for an apoptotic effect of 1 in the colon cancer cell line CT26.

Compared to some other triterpenoid scaffolds, only a limited number of madecassic acid derivatives are known, and few of these have been investigated with respect to their cytotoxic activity.^28,29 In order to search for novel antitumor compounds and provide insight into the influence of substituent modification at the C-2, C-3, and C-23 positions of 1 on cytotoxic activity, a library of madecassic acid derivatives comprising different ester substituents in the A-ring and various functionalities in ring B was synthesized and evaluated for in vitro cytotoxic activities. Previous studies have reported that the introduction of furoyl and cinnamic acid groups to the pentacyclic skeleton of triterpenes can significantly improve the cytotoxicity of these compounds and increase their selectivity toward tumor cells.^29–32 Taking this into account, a panel of 2-furoylate and cinnamate madecassic acid derivatives was prepared in an attempt to improve the compound cytotoxicity and selectivity profiles. The impact of oxidative modification of the A-ring of 1 on its cytotoxic activity was also investigated. For this purpose, a series of novel diosphenol madecassic acid analogues was designed and synthesized. Accumulating evidence shows that incorporation of five-membered heterocyclic motifs into pentacyclic triterpenoid scaffolds can improve their cytotoxic activities and pharmacological properties significantly.^19,33–37

Although cyclic enol ethers are found in a number of bioactive compounds,³⁸ they have not been exploited biologically as oxygen-containing heterocycles in triterpene chemistry. This fact motivated us to synthesize a series of novel five-membered cyclic enol ethers from 1.

Biological evaluation of the new madecassic acid analogues using the U.S. National Cancer Institute (NCI) 60 human cancer cell line screen revealed potent antiproliferative activity against several cancer cell lines, including multidrug-resistant phenotypes. Based on its potency and cell line selectivity, analogue 29 was identified as the most promising lead compound. The molecular mechanisms underlying the cytotoxic activity of this agent were predicted using the publicly accessible web tool CellMiner (https://discover.nci.nih.gov/cellminer/)³⁹ and subsequently validated by a series of cell-based and biochemical assays.

RESULTS AND DISCUSSION

A series of 29 madecassic acid derivatives was synthesized, and their structures were fully confirmed by comprehensive spectroscopic analyses. With the exception of compounds 2–4,^29,40 11,^29,40 12,^29,41 and 19–23,^29,40,42 which are known, the remaining compounds are new. As shown in Scheme 1, treatment of 1 with anhydrous potassium carbonate (K₂CO₃) and methyl iodide in dimethylformamide (DMF) afforded the methyl ester derivative 2 in 82% yield. Acetylation of 2 with acetic anhydride in the presence of 4-dimethylaminopyridine (DMAP) in tetrahydrofuran (THF) at room temperature gave the triacetylated derivative 3 in good yield (68%). The 2α-acetylation of 2 was achieved by protection of the C-3 and C-23 hydroxy groups as an acetonide 4, which allowed the DMAP-catalyzed acetylation of the C-2 hydroxy group to give intermediate 5 in high yield (76%). Deprotection of acetonide 5 with 1 M HCl in THF at room temperature gave the desired monoacetylated 6 in 87% yield. Characteristic changes in chemical shifts of key NMR signals in the starting material 2 compared to the products 3 and 6 were utilized for structure characterization and product identification. The 1H NMR spectrum of 2α,3β,23-triacetate 3 revealed signals due to acetate methyl groups at δ_H 2.06, 2.03, and 1.98 ppm, and these showed strong HMBC correlations with oxymethine carbons at δ_C 65.4 (C-23), 75.0 (C-3), and 70.0 (C-2), respectively. The ¹H NMR spectrum of 6 showed a deshielded shift of the H-2 signal (δ_H 5.08, dt, J = 10.7, 4.5 Hz), and correlations in the HMBC spectrum from both H-2 and an acetate methyl (δ_H 2.08, 3H, s) to the carbonyl carbon at δ_C 171.9 established an acetyl group attached to C-2.

Scheme 1.

Synthesis of Madecassic Acid Derivatives 2–10

The 2α-(2-furoyl) and 2α-cinnamic ester derivatives 7 and 9 were synthesized by reaction of 4 with 2-furoyl chloride and trans-cinnamoyl chloride, respectively, in dry benzene and DMAP at 40 °C under a nitrogen atmosphere (Scheme 1). Deprotection of 7 and 9 with 1 M aqueous HCl in THF gave the desired 2-substitued 2-furoyl 8 and cinnamoyl-3β,6β,23-trihydroxy 10 products in moderate to good yields. Characteristic signals in each 1H NMR spectrum allowed ready identification of these derivatives. The introduction of a furan ring in derivative 8 was confirmed by the presence of three aromatic peaks at δ_H 7.57 (1H, m, H-5′), 7.19 (1H, br d, J = 3.2 Hz, H-3′), and 6.51 (1H, m, H-4′). The ¹H NMR spectrum of the 2α-cinnamic ester derivative 10 showed two multiplets, one between δ_H 7.53–7.51 (2H, m) and the other between δ_H 7.40–7.38 (3H, m) for the aromatic protons. The olefinic CH protons were observed as doublets (J = 16.0 Hz) at δ_H 7.70 (1H, O=C–CH=CH–) and 6.43 (1H, O=C–CH=CH–).

As depicted in Scheme 2, treatment of 3 with thionyl chloride and pyridine afforded the anhydro derivative 11, which was then deacetylated with potassium hydroxide (KOH) in MeOH to give 12. Reaction of 12 with dry acetone and a catalytic amount of HCl in the presence of activated molecular sieves gave rise to the acetonide derivative 13, which was oxidized with pyridinium dichromate (PDC) in CH₂Cl₂ to give the 2-oxo derivative 14. Deprotection of the acetonide group in 14 with concentrated HCl (37%) and THF under reflux provided the nor-diosphenol 15. The ¹H NMR spectrum of 15 showed characteristic signals for four tertiary methyl groups at δ_H 2.00 (3H, s, H-24), 1.28 (3H, s, H-25), 1.03 (3H, s, H-26), and 0.93 (3H, s, H-27) and two secondary methyl signals at δ_H0.92 (3H, m, H-30) and 0.83 (3H, d, J = 6.4 Hz, H-29). The methyl ester appeared as a sharp singlet at δ_H 3.63 (3H, s). The spectrum for 15 did not contain the methylene doublet signals for H-23 that were observed in 2. Instead, a low-field methyl signal at δ_H 2.00, assignable to CH₃–C=C, was observed for this compound. Furthermore, the ¹H NMR spectrum of 15 displayed an additional CH signal at δ_H 6.30 (s, H-1) attributed to the olefinic proton of the enol. The presence of only 30 ¹³C NMR signals confirmed the loss of the secondary carbon resonating at δ_C 68.4 assigned to C-23 in 2. The 2-hydroxy-Δ^1,4-dien-3-one subunit was established from key ¹H–¹³C correlations from the less shielded olefinic proton at δ_H 6.30 (H-1) to C-3 (δ_C 181.7), oxygenated C-2 (δ_C 144.4), C-9 (δ_C 45.1), and the C-25 methyl carbon (δ_C 21.7). HMBC correlations from the vinyl methyl protons (δ_H 2.00) to C-3 (δ_C 181.7), C-5 (δ_C 167.5), and C-4 (δ_C 126.0) confirmed the connectivity of this methyl group to C-4.

Scheme 2.

Synthesis of Madecassic Acid Derivatives 11–18

For the preparation of 2,6-dioxo derivative 17, acetonide 4 was oxidized with PDC in CH₂Cl₂ to give 16, which was deprotected under acidic conditions to give 17 in 58% yield. The hydroxy ketone 17 was then subjected to a base-catalyzed enolization with 10% methanolic KOH under reflux for 3 h to furnish diosphenol 18 (Scheme 2). The NMR spectra of this compound closely resembled those of 15, with changes focused on signals associated with the B-ring. The most obvious differences were the absence of the H-6 methylene signals in the 1H NMR spectrum and the presence of an additional carbonyl signal at δ_C 205.9 ppm (assigned to C-6) in the ¹³C NMR spectrum of 18.

Oxidation of 3 with Jones reagent in acetone afforded ketone 19, which was in turn treated with KOH in methanol to give the corresponding triol derivative 20 in 65% yield. Cyclodehydration of 20 under acidic conditions led to the formation of cyclic enol ether 21 (Scheme 3). The ¹H NMR spectrum of this compound revealed the absence of a singlet at δ_H 2.48 (assigned to H-5 in 20) and a deshielded shift of the H-23 methylene protons (δ_H 4.21 and 4.10, 1H, d, J = 9.0 Hz, each). The absence of the carbonyl peak at δ_C 212.2 in the ¹³C NMR spectrum and the presence of an additional quaternary carbon peak at δ_C 114.1 assigned to C-5 confirmed the placement of a double bond at C-5–C-6 and the presence of a five-membered cyclic vinyl ether ring. As shown in Scheme 3, compound 3 was also converted to α,β-unsaturated ketone 22 in 51% yield by oxidation with a mixture of potassium permanganate and iron(III) sulfate.⁴³ The trihydroxy compound 23 was obtained in 65% yield by deacetylation of 22 and subsequently transformed into cyclic enol ether derivative 24 using the same procedure as described previously for the synthesis of 21.

Scheme 3.

Synthesis of Madecassic Acid Derivatives 19–24

Cyclodehydration to the five-membered enol ether 25 was carried out by the acid-catalyzed rearrangement of 17 in 66% yield. In order to obtain the diosphenol 26, compound 25 was subjected to a base-rearrangement using 10% methanolic KOH, according to the same procedure described previously for the synthesis of 18 (Scheme 4). The IR spectrum of 26 was consistent with a diosphenol functionality with absorptions for a conjugated ketone (1674 cm⁻¹) and an enol OH (3364 cm⁻¹). A singlet at δ_H 6.11 in the ¹H NMR spectrum assigned to the C-1 vinylic proton and ¹³C NMR signals for an olefinic carbon at δ_C 126.5 (assigned to C-1) and a carbonyl carbon at δ_C 198.6 supported a diosphenol group. We also investigated if cleavage of ring A between C-2 and C-3 of 25 would lead to a compound with more potent inhibitory effects on the growth of cancer cells. Reaction of 25 with hydroxylamine hydro-chloride in EtOH in the presence of sodium acetate (NaOAc) afforded the corresponding hydroxyimino alcohol 27 (Scheme 4). The presence of a hydroxy group at C-3 was indicated by the presence of a broad IR absorption band in the 3248 to 3583 cm⁻¹ region and by a methine singlet at δ_H 4.20 corresponding to H-3 in the ¹H NMR spectrum. The ¹³C NMR spectrum contained a signal at δ_C 158.0, which was more shielded by 52 ppm compared with the corresponding carbon in 25 as a result of forming the C=N– bond. The α-hydroxyoxime 27 was then subjected to Beckmann rearrangement in pyridine under reflux with p-toluenesulfonyl chloride (p-TsCl) to provide the 2,3-seco-aldehydonitrile 28 in 64% yield. Successful preparation of 28 was confirmed by an IR band at 2243 cm⁻¹ that corresponded to nitrile vibrations, a ¹H NMR signal for the formyl proton at δ_H 9.82, and characteristic ¹³C NMR signals for nitrile (δ_C 118.4) and formyl group (δ_C 202.2) carbons. Esterification of 25 with either 2-furoyl chloride or trans-cinnamoyl chloride was then accomplished selectively at the 3-position to give compounds 29 and 30, respectively, in moderate to good yields (Scheme 4).

Scheme 4.

Synthesis of Madecassic Acid Derivatives 25–30

The synthesized compounds were submitted to the NCI for evaluation of their cytotoxic activity against 60 human tumor cell lines. Nineteen of these compounds (1, 2, 3, 6, 8, 10, 12, 15, 17, 18, 21, 22, 24, 25, 26, 27, 28, 29, and 30) were selected for initial screening based on the diversity they added to the NCI small-molecule collection.^44–46 Results for each compound in the single-dose assay are reported in Tables S1 and S2 (Supporting Information). Thirteen of the tested compounds (2, 3, 6, 8, 10, 12, 15, 17, 25, 26, 27, 28, and 29) satisfied predetermined inhibition thresholds set by the NCI and were thus selected for complete dose–response studies with five different test concentrations (0.01, 0.1, 1, 10, and 100 μM). Dose–response curves (% growth vs sample concentration) of the test compounds against each cell line in the NCI screen (Figures S39–S51) and the mean bar graphs obtained for each compound (Figures S52–S64) can be found in the Supporting Information. A comparative summary of the single-dose mean growth inhibition (%) for all selected compounds, and for those that passed the initial one-dose screening test, the mean (GI₅₀, μM) and range (μM) values, and the most sensitive cell line is provided in Table 1.

Table 1.

Antiproliferative Activity of Select Madecassic Acid Derivatives

compound	NSC numbera	one-dose mean growth (%)b	five-dose mean GI50 (μM)c	five-dose GI50 range (μM)d	most sensitive cancer cell line
madecassic acid (1)	783350	99.2	NTe	NTe	HOP-92, NSCLCf
2	787805	−54.1	12.9	18.8	PC-3, prostate
3	787221	8.0	1.7	7.8	K-562, leukemia
6	787213	−79.2	11.0	14.2	UO-31, renal
8	787214	−83.0	2.3	2.3	RXF 393, renal
10	785396	−4.8	2.4	3.8	A498, renal
12	787216	−55.4	17.0	9.6	SK-MEL-5, melanoma
15	783353	31.7	12.0	18.5	CCRF-CEM, leukemia
17	787807	−22.4	10.0	20.4	RPMI-8226, leukemia
18	783354	67.8	NTe	NTe	CCRF-CEM, leukemia
21	785612	85.8	NTe	NTe	SR, leukemia
22	783351	45.8	NTe	NTe	PC-3, prostate
24	783352	51.5	NTe	NTe	MDA-MB-435, melanoma
25	787806	−62.9	18.2	9.9	HOP-92, NSCLCf
26	787219	−58.9	8.9	18.7	MOLT-4, leukemia
27	787804	−76.9	6.6	16.9	786–0, renal
28	787220	−18.1	4.9	12.7	MOLT-4, leukemia
29	787217	27.2	2.3	99.7	COLO 205, colon
30	787218	78.2	NTe	NTe	RPMI-8226, leukemia

^aNational Service Center number assigned by the Developmental Therapeutics Program, NCI to compounds tested in the NCI-60 assay.

^bGrowth percent at 10 μM vs negative control.

^cAverage GI₅₀ value of each compound across the 60 cell lines.

^dDifference between the GI₅₀ values for the most resistant cell line and the most sensitive cell line.

^eNot tested.

^fNon-small-cell lung cancer.

SAR analysis was performed to understand the influence of the various functionalities in the semisynthetic analogues on their cytotoxic activity. As shown in Table 1, compounds 3, 8, 10, 26, 27, 28, and 29 had mean GI₅₀ (concentration of compound that inhibits cell growth by 50%) values of <10 μM in the NCI-60 cell line panel. Structure–activity correlations revealed that the introduction of acetyl groups to the C-2 OH, C-3 OH, and C-23 OH positions in 3 enhanced activity, resulting in a compound 7.5-fold more active than its precursor 2. Likewise, compounds 8 and 10, which contain furoyl and cinnamoyl groups at the C-2 position, respectively, exhibited strikingly more potent growth inhibitory activities than the parent compound 2 or the C-2 monoacetylated derivative 6. In the diosphenol series of madecassic acid derivatives, compound 26, which also bears a cyclic enol ether moiety, was 2-fold more potent than its parent 3α-hydroxy-2-ketone 25 and far more active than the 2,3-dihydroxy derivative 21, which was inactive in the one-dose 60-cell assay. Thus, it can be inferred that the presence of a diosphenol functionality in 26 was crucial for its antiproliferative activity. In the cyclic enol ether series, conversion of the ketone group at C-2 in 25 into an oxime in 27 resulted in an enhancement of activity. Notably, the cytotoxic activity of oxime 27 was further increased by cleavage of the A-ring with resultant formation of the 2,3-seco aldehydonitrile derivative 28. Despite its relatively modest cytotoxicity (mean GI₅₀ = 4.9 μM), this promising and versatile scaffold provides a new skeleton with potential for further optimization. Finally, introduction of different substituents at the C-3 position of 25 showed that only introduction of a 2-furoyl group at C-3 improved the antiproliferative activity (8-fold), resulting in the potent analogue 29 (mean GI₅₀ = 2.3 μM).

Compounds 3, 8, 10, and 29 displayed the most potent cytotoxic activity with significant inhibition for most of the 60 cell lines, and their mean GI₅₀, TGI (concentration of compound that totally inhibits cell growth), and LC₅₀ (concentration of compound that kills 50% of cells) values across each cell line are shown in Table S3 (Supporting Information). Compound 29 exhibited broad-spectrum antiproliferative activity with GI₅₀ values of <5.0 μM for 69% and <1.0 μM for 45% of the tested cell lines (Table S3, Supporting Information). Strong growth inhibition (GI₅₀ ≤ 1.0 μM) was observed against all leukemia cell lines and many cell lines derived from solid tumors. Among the tumor subpanels, selectivity greater than 100-fold was observed between NCIH460 and NCI-H226 in the NSCL (non-small-cell lung) cancer panel and between MCF7, T-47D, MDA-MB-468, and HS 578T in the breast cancer panel. A difference in potency of more than 50-fold was also observed between NCI-H23 and NCI-H226 in NSCL, OVCAR-4, and OVCAR-5 in ovarian and CAKI-1 and RXF 393 in renal cancer. Large differential inhibitory responses between cell lines suggests a specific target or cellular pathway is modulated by the test compound rather than a general mechanism such as acute toxicity or cell lysis. With respect to the total growth inhibition effect of 29, the COLO 205 (TGI 3.0 μM, colon), SK-MEL-5 (TGI 2.6 μM, melanoma), and UACC-257 (TGI 3.0 μM, melanoma) cell lines were the most sensitive. At the LC₅₀ level of cytotoxicity, most cell lines were not sufficiently impacted at the high test concentration of 100 μM, with the exception of the UACC-257 melanoma cell line (LC₅₀ 50.1 μM). Interestingly, the low GI₅₀ values observed for 29 against cell lines that naturally express high levels of multidrug transporter MDR1 (P-glycoprotein), such as the colon adenocarcinoma HCT-15 (GI₅₀ 0.5 μM) and the renal adenocarcinoma UO-31 (GI₅₀ 1.7 μM), suggest that this compound is not a substrate of the transporter and thus may be effective in overcoming MDR1-mediated drug resistance in vitro. Due to its potent cytotoxic activities and its distinctive pattern of selectivity in the NCI-60 panel, 29 was selected for further experiments to explore its mechanism of action.

Differential sensitivity of cell lines in the NCI screen offers an opportunity to relate variations in gene expression in these cells to the molecular pharmacology of the test compounds. We performed bioinformatics studies using the CellMiner web application to identify correlations between the 60-cell line gene expression patterns and drug sensitivity profiles using a Pearson correlation coefficient (r). This pharmacogenomics approach is based on the premise that genes with patterns of expression across the cell lines that correlate with compound sensitivity are candidate clinical biomarkers of drug efficacy, potentially even direct effectors of drug action, or targets for novel drug development. By analyzing the NCI-60 cell lines for a correlation between their transcriptome and their sensitivity to the cytotoxic effects of 29 using CellMiner, we found 196 genes that were moderately to highly correlated (Pearson correlation coefficient ≥0.50) with its in vitro antiproliferative activity (Table S4, Supporting Information). Gene ontology enrichment analysis of the candidate genes based on the biological processes and cellular components in which they participate revealed two main functional categories that were significantly enriched: DNA replication and mitochondrial electron transport (Figure 1 and Table S4, Supporting Information). The enrichment of genes involved in DNA replication in the first functional group suggests a possible impact of 29 on this process. Interestingly, ORC1, subunit 1 of the highly conserved origin recognition complex (ORC) that is essential for the initiation of DNA replication,⁴⁷ MCM6, a DNA licensing factor that is also involved in DNA replication initiation,⁴⁸ and CDC25A, a member of the CDC25 family of phosphatases that is required for progression from G₁ to the S phase of the cell cycle⁴⁹ were among the highly correlated genes that were enriched in this category. This suggested that compound 29 may affect the initiation of DNA replication. Significant enrichment of four genes (NDUFA9, NDUFB10, NDUFS7, NDUFS3) involved in mitochondrial electron transport chain complex I assembly and function^50–53 suggested this complex is a second potential target of 29.

Figure 1.

Network visualization of gene ontology (GO) enrichment analysis based on gene expression profiles that correlated with the cytotoxicity profile of 29 across the NCI-60 panel cell lines (correlation coefficient ≥0.50). Each node represents an enriched GO term (FDR < 0.05), and an edge represents existing genes shared between connecting enriched GO terms. Overlap of genes between nodes is indicated by edge thickness. Edges are not shown where the overlap coefficient is less than 0.5. GO terms that were very general (contain more than 300 genes) were excluded from the map.

The most prevalent functional group in our gene ontology (GO) enrichment analysis was composed of genes involved in the initiation of DNA replication, which suggested that 29 modulates this initiation process. To test this hypothesis, COLO 205 cells were analyzed for cell cycle progression after being treated with 10 and 20 μM 29 for 48 h (Figure 2). Flow cytometric analysis of DNA content in cells treated with vehicle alone (DMSO) showed that 63.8% of the cell population was in the G₀/G₁ phase. There were also distinct populations of cells in the S (16.3%) and G₂/M (18.6%) phases. However, cell cycle progression was dramatically blocked at the G₀/G₁ phase, and a significant reduction in the number of cells in the S and G₂/M phases was observed when cells were treated with compound 29. In the presence of 10 μM 29, fewer cells exited out of the G₀/G₁ phase (88.4% vs63.8%), and consequently fewer cells progressed to the S (3.8% vs 16.3%) and G₂/M (6.6% vs 18.6%) phases. A progressive decrease in the number of cells exiting the G₀/G₁ phase (89.1%) was observed with a higher dose of 29 (20 μM), accompanied by a concomitant decrease in the number of the cells in the S (3.3%) and G₂/M (6.2%) phases. As expected, the addition of a microtubule polymerization inhibitor (nocodazole) caused a clear shift in the cell population from the G₀/G₁ to the G₂/M phase. This result is in complete agreement with our GO analysis, suggesting that 29 exerts its growth inhibitory activity, at least in part, by blocking the initiation of DNA replication through cell-cycle arrest at the G₀/G₁ phase.

Figure 2.

Compound 29 arrests the G₁ to the S phase cell cycle transition in COLO 205 cells. After 48 h of incubation, cells were harvested and stained with propidium iodide (PI) followed by Cellometer imaging and analysis. (A) Representative histograms of cell cycle analysis. The x-axis indicates the PI fluorescence intensity that correlates with the nuclear DNA content. The results shown are of DMSO, compound 29 (10 and 20 μM, respectively), and nocodazole (2 μM for 24 h) treated samples. (B) Bar graph depicting the cell population (%) per cell cycle phase. Values represent the means ± SD of three independent experiments. Differences between treated and control groups were considered statistically significant at a p value < 0.01 (**).

The electrical potential across the inner mitochondrial membrane (Δψ_m) is the central parameter that controls mitochondrial respiration, ATP synthesis, and Ca²⁺ accumulation, as well as the generation of reactive oxygen species.^54–56 We investigated the effect of 29 on the mitochondrial membrane potential in COLO 205 cells to assess its impact on the mitochondrial respiratory chain complex I, which was the second major functional category identified in the gene ontology enrichment analysis (Figure 3). After COLO 205 cells were exposed to different concentrations (10 and 20 μM) of 29 for 48 h, mitochondrial dysfunction was assessed using the fluorescent probe JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide). The JC-1 monomer emits a green fluorescent signal corresponding to a low membrane potential. At high membrane potential, JC-1 aggregates and shows a red fluorescence signal. As shown in Figure 3, in the control (DMSO), JC-1 emits a red fluorescence (JC-aggregates) due to an intact mitochondrial transmembrane potential. However, in the presence of 29, cells showed a progressive loss of red JC-aggregate fluorescence and the appearance of cytoplasmic diffusion of green monomer fluorescence (JC-1 monomers) in a dose-dependent manner. These results indicate that 29 induces a concentration-dependent decrease of mitochondrial membrane potential and thus causes mitochondrial dysfunction.

Figure 3.

Mitochondrial dysfunction induced by compound 29. Photomicrographs of transmembrane potential-dependent JC-1 staining of mitochondria in control and treated COLO 205 cancer cells. In untreated cells, high mitochondrial polarization is indicated by a red fluorescence due to JC1-aggregate formation by the concentrated dye. In cells treated with compound 29 for 48 h, due to a collapse of mitochondrial potential, the JC-1 dye remained in the cytoplasm in its monomeric form, which fluoresces as a green color.

In conclusion, the madecassic acid derivatives 2–30 were synthesized and screened for cytotoxic activity against the NCI-60 cancer cell line panel. All the tested semisynthetic derivatives showed better antiproliferative activities than madecassic acid (1) itself. Compound 29, a cyclic enol ether derivative bearing a 2-furoyl moiety at C-3, exhibited sub-μM potencies against 26 different tumor cell lines and revealed particular selectivity for one colon (COLO 205) and two melanoma (SK-MEL-5 and UACC-257) cell lines at the TGI level. Using the bioinformatic tool CellMiner, tumor cell functions related to growth inhibition by 29 were predicted to include initiation of DNA replication and mitochondrial electron transport. Follow-up analysis confirmed that treatment with 29 results in cell cycle arrest at the G₁/S transition and disruption of the mitochondrial membrane potential in tumor cells. These experimental results support the mechanisms of cytotoxic activity that were predicted from the CellMiner analysis. In light of the present results, madecassic acid derivative 29 represents a potential lead for the development of new anticancer agents and merits further investigation.

EXPERIMENTAL SECTION

General Experimental Procedures.

Madecassic acid (1) was purchased from Santa Cruz Biotechnology Inc., in over 95% purity. Other reagents and solvents were purchased from Sigma-Aldrich Co., Merck Co., and VWR Portugal and used without further purification. Solvents were dried over standard drying agents according to the usual procedures. Thin-layer chromatographic (TLC) analysis and preparative TLC were carried out on Kieselgel 60HF254 and Kieselgel 60HF254/Kieselgel 60G from Merck Co., respectively. Column chromatographic separations were performed using Kieselgel 60 (230–400 mesh) from Merck Co. Melting points were determined by using open capillary tubes on a Büchi B-540 melting point apparatus and are uncorrected. ¹H, ¹³C, DEPT-135, HSQC, and HMBC NMR experiments were performed in CDCl₃ or C₆D₆ and recorded on Bruker Avance 400 and Bruker Avance III spectrometers operating at 400 and 600 MHz for ¹H and 100 and 150 MHz for ¹³C, respectively. The Bruker Avance III NMR spectrometer was equipped with a 3 mm cryogenically cooled probe. Spectra were calibrated to residual solvent signals at δ_H 7.26 and δ_C 77.16 (CDCl₃) and δ_H 7.16 and δ_C 128.06 (C₆D₆). IR spectra were recorded on a PerkinElmer Spectrum 2000 FT-IR spectrometer using NaCl circular cell windows. HRESIMS were performed with an Agilent 6530B Accurate Mass QTOF mass spectrometer.

Synthesis and Structural Characterization of Compounds 2–30.

The synthesis and spectroscopic characterization of 29, including its intermediates 2, 4, 16, 17, and 25, is described below. The synthetic procedures and spectroscopic data of the remaining compounds are given in the Supporting Information.

Methyl 2α,3β,6β,23-tetrahydroxyurs-12-en-28-oate (2).

Compound 2 was prepared according to the literature²⁹ from madecassic acid (1) (400 mg, 0.79 mmol). The crude product was purified by flash column chromatography with an isocratic elution of petroleum ether (40–65 °C)/EtOAc 1:25 (v/v) to afford 2 as a white solid (338 mg, 82%): mp 177.3–179.1 °C; IR ν_max (NaCl) 3358, 3020, 2922, 2852, 1742, 1634, 1463, 1242, 1168 cm⁻¹; ¹H NMR (CDCl₃, 600 MHz) δ_H 5.30 (1H, br t, J = 3.5 Hz, H-12), 4.41 (1H, br s, H-6), 3.84 (1H, dt, J = 10.3, 4.4 Hz, H-2), 3.75 (1H, d, J = 10.4 Hz, H-23a), 3.60 (3H, s, COOCH₃), 3.54 (1H, d, J = 10.4 Hz, H-23b), 3.37 (1H, d, J = 9.5 Hz, H-3), 2.26 (1H, d, J = 11.3 Hz, H-18), 2.08 and 2.01 (each 1H, m, H-11), 2.00 (1H, m, H-16a), 1.99 (1H, m, H-1a), 1.82 (1H, m, H-15a), 1.78 (1H, m, H-7a), 1.68 (1H, m, H-16b), 1.67 (1H, m, H-22a), 1.66 (1H, m, H-9), 1.59 (1H, m, H-22b), 1.50 (1H, m, H-21), 1.47 (1H, m, H-7b), 1.40 (3H, s, H-25), 1.33 (1H, m, H-19),1.28 (1H, m, H-21b), 1.21 (3H, s, H-24), 1.19 (1H, m, H-5), 1.06 (1H, m, H-15b), 1.05 (3H, s, H-27), 1.03 (3H, s, H-26), 1.01 (1H, m, H-20), 0.96 (1H, m, H-1b), 0.95 (3H, d, J = 6.5 Hz, H-30), 0.87 (3H, d, J = 6.5 Hz, H-29); ¹³C NMR (CDCl₃, 150 MHz) δ_C 178.2 (COOCH₃, C-28), 137.6 (C, C-13), 125.7 (CH, C-12), 79.0 (CH, C-3), 69.0 (CH, C-2), 68.4 (CH₂, C-23), 68.0 (CH, C-6), 53.0 (CH, C-18), 51.7 (COOCH₃), 49.0 (CH, C-5), 48.9 (CH, C-9), 48.7 (C, C-17), 48.2 (CH₂, C-1), 43.4 (C, C-4), 42.7 (C, C-14), 40.9 (CH₂, C-7), 39.0 (CH, C-19), 38.7 (C, C-10 and CH, C-20), 37.7 (C, C-8),36.8 (CH₂, C-22), 30.8 (CH₂, C-21), 28.1 (CH₂, C-15), 24.3 (CH₂, C-16), 23.9 (CH₃, C-27), 23.4 (CH₂, C-11), 21.3 (CH₃, C-30), 18.9 (CH₃, C-25), 18.5 (CH₃, C-26), 17.2 (CH₃, C-29), 14.6 (CH₃, C-24); HRQTOFMS m/z 541.3508 [M + Na]⁺ (calcd for C₃₁H₅₀O₆Na, 541.3505), Δ = 0.55 ppm.

Methyl 2α,6β-dihydroxy-3β,23-isopropylidenedioxyurs-12-en-28-oate (4).

Compound 4 was prepared according to the literature⁵⁷ from 2 (1000 mg, 1.93 mmol). The crude product was purified by flash column chromatography with a gradient elution of petroleum ether (40–65 °C)/EtOAc from 8:1 to 1:1 (v/v) to afford 4 as a white solid (930 mg, 86%): mp 161.5–162.2 °C; IR ν_max (NaCl) 3483, 3047, 2924, 2869, 1735, 1653, 1453, 1364, 1230, 1200, 1113 cm⁻¹; ¹H NMR (C₆D₆, 600 MHz) δ_H 5.36 (1H, br t, J = 3.5 Hz, H-12), 3.89 (1H, dt, J = 10.4, 4.4 Hz, H-2), 3.72 (1H, br s, H-6), 3.66 (1H, d, J =10.3 Hz, H-23a), 3.42 (3H, s, COOCH₃), 3.31 (1H, d, J = 10.3 Hz, H-23b), 3.24 (1H, d, J = 9.4 Hz, H-3), 2.46 (1H, d, J = 11.3 Hz, H-18), 2.04 (1H, m, H-1a), 1.66 (3H, s, H-24), 1.62 (1H, m, H-9), 1.54 (3H, s, CH₃ of acetonide), 1.44 (3H, s, H-25), 1.28 (3H, s, CH₃ of acetonide), 1.12 (3H, s, H-26), 1.05 (3H, s, H-27), 1.02 (1H, m, H-1b), 0.97 (3H, d, J = 6.4 Hz, H-29), 0.92 (3H, m, H-30), 0.78 (1H, br s, H-5); ¹³C NMR (C₆D₆, 150 MHz) δ_C 177.3 (COOCH₃, C-28), 137.6 (C, C-13), 126.3 (CH, C-12), 99.5 [C(OCH₃)_2, acetonide],83.0 (CH, C-3), 72.9 (CH₂, C-23), 67.3 (CH, C-6), 65.6 (CH, C-2),53.3 (CH, C-18), 52.8 (CH, C-5), 51.3 (COOCH₃), 49.2 (CH₂, C-1), 48.5 (CH, C-9), 48.3, 42.7, 40.9, 39.6, 39.3, 39.2, 37.8, 37.3, 37.1,31.0, 30.2 (CH₃, acetonide), 28.5, 24.7, 24.1 (CH₃, C-27), 23.4, 21.4 (CH₃, C-30), 19.5 (CH₃, C-25), 19.4 (CH₃, acetonide), 18.4 (CH₃, C-26), 17.4 (CH₃, C-29), 15.3 (CH₃, C-24); HRQTOFMS m/z 581.3824 [M + Na]⁺ (calcd for C₃₄H₅₄O₆Na, 581.3818), Δ = 1.03 ppm.

Methyl 2,6-dioxo-3β,23-isopropylidenedioxyurs-12-en-28-oate (16).

A solution of pyridinium dichromate (300 mg, 0.8 mmol) and acetic anhydride (0.23 mL, 2.41 mmol) in dry CH₂Cl₂ (35 mL) was stirred for 30 min at room temperature. To the mixture was slowly added 4 (450 mg, 0.81 mmol) in dry CH₂Cl₂ (10 mL). After 8 h at reflux in anhydrous conditions, the reaction was completed (monitored by TLC). The reaction mixture was evaporated under reduced pressure, and the resulting crude diluted with diethyl ether (100 mL) and filtered through a Celite pad. The filtrate was washed with H₂O (3 × 100 mL) and brine (100 mL). The organic phase was dried over anhydrous MgSO₄. Filtration and evaporation of the solvent at reduced pressure gave a crude solid, which was subjected to flash column chromatography with a gradient elution of petroleum ether (40–65 °C)/EtOAc from 7:1 to 5.5:1 (v/v) to afford 16 as a white solid (224 mg, 50%): mp 192.9–193.3 °C; IR ν_max (NaCl) 3055, 2925, 2872, 1733, 1717, 1714, 1636, 1456, 1366, 1220, 1198, 1148 cm⁻¹; ¹H NMR (C₆D₆, 600 MHz) δ_H 5.24 (1H, br t, J = 3.5 Hz, H-12), 4.15 (1H, d, J = 10.4 Hz, H-23a), 4.01 (1H, s, H-3), 3.38 (1H, m, H-23b), 3.37 (3H, s, COOCH₃), 2.44 (1H, d, J = 11.5 Hz, H-18),2.22 (1H, br s, H-5), 2.20 (1H, d each, J = 13.8 Hz, H-7a), 2.06 (1H, d, J = 13.0 Hz, H-1a), 1.97 (1H, m, H-9), 1.93 (1H, d, J = 13.9 Hz, H-7b), 1.69 (3H, s, H-24), 1.57 (3H, s, CH₃, acetonide), 1.44 (1H, d, J = 12.8 Hz, H-1b), 1.24 (3H, s, CH₃, acetonide), 1.06 (3H, s, H-27),0.97 (3H, d, J = 6.4 Hz, H-29), 0.92 (3H, m, H-30), 0.83 (3H, s, H-25), 0.73 (3H, s, H-26); ¹³C NMR (C₆D₆, 150 MHz) δ_C 207.6 (C-6=O), 200.4 (C-2=O), 176.9 (COOCH₃, C-28), 138.2 (C, C-13), 125.2 (CH, C-12), 99.7 [C(OCH₃)_2, acetonide], 80.7 (CH, C-3),70.5 (CH₂, C-23), 61.1 (CH, C-5), 54.3 (CH₂, C-1), 53.1 (CH, C-18), 51.3 (COOCH₃), 49.7 (CH₂, C-7), 48.1, 47.8 (CH, C-9), 47.5,45.5, 42.6, 40.7 (C, C-4), 39.24, 39.17, 36.8, 30.8, 30.0 (CH₃, acetonide), 28.2, 24.32 (CH₃, C-27), 24.28, 23.5, 21.3 (CH₃, C-30),19.1 (CH₃, C-25), 18.7 (CH₃, acetonide), 17.6 (CH₃, C-26), 17.3 (CH₃, C-29), 14.0 (CH₃, C-24); HRQTOFMS m/z 577.3504 [M + Na]⁺ (calcd for C₃₄H₅₀O₆Na, 577.3505), Δ = −0.17 ppm.

Methyl 2,6-dioxo-3β,23-dihydroxyurs-12-en-28-oate (17).

Compound 17 was prepared according to the literature⁵⁷ from 16 (180 mg, 0.32 mmol). The crude product was purified by flash column chromatography with a gradient elution of petroleum ether (40–65 °C)/EtOAc from 3:1 to 2:1 (v/v) to afford 17 as a white solid (95 mg, 58%): mp 128.7–130.0 °C; IR ν_max (NaCl) 3452, 3051, 2920, 2853, 1739, 1717, 1714, 1646, 1456, 1377, 1244, 1190 cm^–1; ¹H NMR (CDCl₃, 600 MHz) δ_H 5.32 (1H, br t, J = 3.5 Hz, H-12), 4.31 (1H, s, H-3), 3.67 (1H, d, J = 10.4 Hz, H-23a), 3.59 (3H, s, COOCH₃), 3.56 (1H, d, J = 10.4 Hz, H-23b), 3.27 (1H, s, H-5), 2.68 (1H, d, J = 12.6 Hz, H-7a), 2.58 (1H, d, J = 12.8 Hz, H-1a), 2.42 (1H, m, H-9), 2.36 (1H, d, J = 12.5 Hz, H-1b), 2.30 (1H, d, J = 11.2 Hz, H-18), 1.98 (1H, d, J = 12.9 Hz, H-7b), 1.26 (3H, s, H-27), 0.96 (9H, m, H-24, H-25, and H-30), 0.89 (3H, d, J = 6.4 Hz, H-29), 0.80 (3H, s, H-26); ¹³C NMR (CDCl₃, 150 MHz) δ_C 210.6 (C-6=O), 210.4 (C-2=O), 177.8 (COOCH₃, C-28), 138.4 (C, C-13), 124.5 (CH, C-12), 76.6 (CH, C-3), 65.4 (CH₂, C-23), 56.2 (CH, C-5), 53.3 (CH₂, C-1), 52.8 (CH, C-18), 51.7 (COOCH₃), 50.2 (CH₂, C-7), 48.1, 47.9 (CH, C-9), 47.8, 47.3, 46.1, 42.6, 39.0, 38.9, 36.5, 30.7, 28.1, 24.3 (CH₃, C-27), 24.1, 23.7, 21.3 (CH₃, C-30), 18.6 (CH₃, C-25), 17.23 (CH₃, C-26), 17.18 (CH₃, C-27), 13.2 (CH₃, C-24); HRQTOFMS m/z 537.3190 [M + Na]⁺ (calcd for C₃₁H₄₆O₆Na, 537.3192), Δ = −0.37 ppm.

Methyl 2-oxo-3β-hydroxy-6,23-epoxyursa-5,12-dien-28-oate (25).

To a solution of 17 (100 mg, 0.19 mmol) in MeOH (15 mL) was added 0.15 mL of concentrated HCl (37%). After 4 h at room temperature the reaction was completed (monitored by TLC). The solvent was evaporated under reduced pressure, and the residue was diluted with H₂O (50 mL) and extracted with diethyl ether (3 × 50 mL). The combined organic layers were washed with H₂O (3 × 50 mL) and brine (50 mL). The organic phase was dried over anhydrous MgSO₄. Filtration and evaporation of the solvent at reduced pressure gave a crude solid, which was subjected to preparative TLC [petroleum ether (40–65 °C)/EtOAc (4:1, v/v)] to afford 25 as a white solid (62 mg, 66%): mp 94.3–96.2 °C; IR ν_max (NaCl) 3481, 3022, 2923, 2854, 1739, 1722, 1716, 1456, 1378, 1238, 1191 cm⁻¹; ¹H NMR (CDCl₃, 600 MHz) δ_H 5.29 (1H, m, H-12), 4.46 (1H, d, J =9.2 Hz, H-23a), 4.24 (1H, s, H-3), 4.11 (1H, d, J = 9.2 Hz, H-23b),3.61 (3H, s, COOCH₃), 2.62 (1H, d, J = 12.5 Hz, H-1a), 2.36 (1H, d, J = 17.9 Hz, H-7a), 2.28 (1H, d, J = 11.1 Hz, H-18), 2.19 (1H, m, H-11a), 2.18 (1H, d, J = 12.6 Hz, H-1b), 1.97 (1H, m, H-9), 1.96 (1H, m, H-11b), 1.69 (1H, m, H-7b), 1.13 (3H, s, H-24), 1.12 (3H, s, H-27), 1.10 (3H, s, H-25), 0.95 (1H, d, J = 6.1 Hz, H-30), 0.89 (3H, d, J = 6.3 Hz, H-29), 0.85 (3H, s, H-26); ¹³C NMR (CDCl₃, 150 MHz) δ_C 210.3 (C-2=O), 178.0 (COOCH₃, C-28), 152.8 (C, C-6), 138.0 (C, C-13), 124.4 (CH, C-12), 111.4 (C, C-5), 82.6 (CH₂, C-23), 82.2 (CH, C-3), 55.8 (CH₂, C-1), 54.3, 52.5 (CH, C-18), 51.7 (COOCH₃), 48.1, 44.6 (CH, C-9), 41.7, 41.6, 39.8, 39.1, 39.0,36.6, 30.7, 29.8 (CH₂, C-7), 27.9, 24.7 (CH₃, C-27), 24.1, 23.8 (CH₂, C-11), 22.7 (CH₃, C-25), 21.5 (CH₃, C-24), 21.3 (CH₃, C-30), 17.9 (CH₃, C-26), 17.2 (CH₃, C-29); HRQTOFMS m/z 497.3263 [M +H]⁺ (calcd for C₃₁H₄₅O₅, 497.3267), Δ = −0.80 ppm.

Methyl 2-oxo-3β-(2-furoyloxy)-6,23-epoxyursa-5,12-dien-28-oate (29):

To a stirred solution of 25 (150 mg, 0.30 mmol) in dry benzene (8 mL) were added 2-furoyl chloride (0.12 mL, 1.2 mmol, 4 equiv) and DMAP (146.6 mg, 1.2 mmol, 4 equiv). After 3 h 30 min at 40 °C under a nitrogen atmosphere, the reaction was completed (monitored by TLC). The solvent was evaporated under reduced pressure, and the residue was diluted with H₂O (60 mL) and extracted with diethyl ether (3 × 60 mL). The combined organic layers were washed with H₂O (3 × 60 mL) and brine (60 mL). The organic phase was dried over anhydrous MgSO₄. Filtration and evaporation of the solvent at reduced pressure gave a crude solid, which was subjected to preparative TLC [petroleum ether (40–65 °C)/EtOAc (6:1, v/v)] to afford 29 as a white solid (105 mg, 59%): mp 135.7–137.2 °C; IR ν_max (NaCl) 3111, 3063, 2924, 2854, 1739, 1723, 1717, 1682, 1457, 1378, 1232, 1180, 1045 cm⁻¹; ¹H NMR (CDCl₃, 600 MHz) δ_H 7.61 (1H, m, H-5′), 7.26 (1H, d, J = 3.7 Hz, H-3′), 6.54 (1H, m, H-4′), 5.39 (1H, s, H-3), 5.31 (1H, br t, J = 3.6 Hz, H-12), 4.38 and 4.12 (each 1H, d, J = 9.3 Hz, H-23), 3.62 (3H, s, COOCH₃), 2.61 (1H, d, J = 12.6 Hz, H-1a), 2.38 (1H, d, J = 18.0 Hz, H-7a), 2.33 (1H, d, J = 12.6 Hz, H-1b), 2.30 (1H, d, J = 11.4 Hz, H-18), 2.02 (1H, m, H-9), 1.73 (1H, d, J = 18.1 Hz, H-7b), 1.38 (3H, s, H-24), 1.16 (3H, s, H-25), 1.14 (3H, s, H-27), 0.96 (3H, d, J = 6.4 Hz, H-30), 0.90 (3H, d, J = 6.4 Hz, H-29), 0.87 (3H, s, H-26); ¹³C NMR (CDCl₃, 150 MHz) δ_C 202.8 (C-2=O), 178.0 (COOCH₃, C-28), 157.8 (OCO, C-6′), 153.3 (C, C-6), 146.9 (CH, C-5′), 144.1 (C, C-2′), 138.0 (C, C-13), 124.5 (CH, C-12), 119.0 (CH, C-3′), 112.1 (CH, C-4′), 111.5 (C, C-5), 83.3 (CH, C-3), 82.1 (CH₂, C-23), 56.8 (CH₂, C-1), 52.5 (C, C-4 and CH, C-18), 51.7 (COOCH₃), 48.1,44.5 (CH, C-9), 41.7, 41.6, 39.4, 39.1, 39.0, 36.6, 30.7, 29.9 (CH₂, C-7), 27.9, 24.7 (CH₃, C-27), 24.1, 23.8, 22.6 (CH₃, C-24), 22.5 (CH₃, C-25), 21.3 (CH₃, C-30), 18.0 (CH₃, C-26), 17.2 (CH₃, C-29); HRQTOFMS m/z 591.3317 [M + H]⁺ (calcd for C₃₆H₄₇O₇, 591.3322), Δ = −0.85 ppm.

NCI-60 Cytotoxicity Drug Screen.

The NCI-60 cell line panel is organized into nine subpanels with diverse histology representing leukemia, melanoma, non-small-cell lung, colon, kidney, ovarian, breast, prostate, and central nervous system cancers. Details of the NCI-60 cell line screening protocols and reporting procedures have been described previously.^44,46,58,59 Briefly, test compounds were assayed at a single dose concentration (10 μM) in the full NCI-60 cancer cell line panel. Upon initial indication of activity in the single-dose experiment, compounds were subsequently tested at five doses starting at 100 μM and decreasing by logarithmic dilution to a final concentration of 0.01 μM. Cell viability after 48 h of incubation was visualized using sulforhodamine B as reported previously.^44,46 Through the use of a time zero cell control, the total cell growth can be determined for each cell line, thus allowing calculations of GI₅₀, TGI, and LC₅₀.

CellMiner and Gene Ontology Enrichment Analysis.

Analysis of the GI₅₀ data from the NCI-60 cell line screening for compound 29 was performed using the publicly accessible web tool CellMiner (http://discover.nci.nih.gov/cellminer/). GO enrichment analysis with BiNGO (ver. 3.0.3), a Cytoscape (ver. 3.3.0) plug-in,⁶⁰ was used for functional profiling of genes with expression profiles that were moderately to highly correlated (correlation coefficient ≥0.5) with the test compounds’ cytotoxicity profile across the NCI-60 panel of cancer cell lines. Significantly enriched biological processes and cellular components were evaluated using a hypergeometric test corrected for multiple hypothesis testing (p < 0.05) using a Benjamini–Hochberg false discovery rate (FDR) correction and then visualized using the Enrichment Map (ver. 2.0.1) plug-in⁶¹ developed for Cytoscape.

Cell Cycle Analysis.

COLO 205 cells were cultured at 37 °C with 5% CO₂ in RPMI medium (Invitrogen, Carlsbad, CA, USA) containing 100 U/mL penicillin, 100 μg/mL streptomycin (Invitrogen), 10% fetal bovine serum (Sigma-Aldrich, St Louis, MO, USA), and either compound 29 (10 and 20 μM) or DMSO (negative control). Nocodazole was purchased from Sigma-Aldrich and used as a positive control (2 μM for 24 h). At the end of the culture period, cells were trypsinized and then centrifuged at 1000 rpm for 5 min, and the pellet was resuspended in 200 μL of phosphate-buffered saline. Cells were fixed by adding 500 μL of 200 proof ethanol (Sigma-Aldrich) and incubated for 15 min on ice. Cells were then centrifuged at 2000 rpm for 8 min, and the pellet was resuspended in 200 μL of propidium iodide/RNase staining solution (Cellometer PI cell cycle kit, Nexcelom Bioscience, Lawrence, MA, USA). The COLO 205 cells were then incubated for 40 min at 37 °C before performing imaging cytometry analysis. Population histograms were generated with FCS Express 4 software.

Mitochondrial Membrane Potential Analysis.

To investigate the effect of 29 on mitochondrial membrane potential, COLO 205 cells (5 × 10⁵ cells, cultured as described above) were grown on sterile coverslips placed in a six-well plate and incubated overnight. The medium was aspirated, and cells were treated with 29 (10 and 20 μM) or DMSO for 48 h. After treatment, 1 mL of fresh medium containing 4 μg/mL of JC-1 (Life Technologies, Eugene, OR, USA) was added to each well, and plates were incubated at 37 °C for 30 min. Cells were washed two times with fresh medium, and coverslips were placed on microscope slides and imaged immediately by a Nikon Eclipse TE2000-S inverted microscope using a 40× objective at room temperature.