Search for New Type of Anticancer Drugs with High Tumor Specificity and Less Keratinocyte Toxicity

Problems of Current Anticancer Drugs

It is well known that administration of anticancer agents induces skin toxicity (1-7). This prompted us to re-evaluate the cytotoxicity and tumor specificity of anticancer drugs. For this purpose, we established an in vitro assay system, using four human oral squamous cell carcinoma (OSCC) cell lines (Ca9-22, HSC-2, HSC-3 and HSC-4), three human mesenchymal normal oral cells [gingival fibroblasts (HGFs), pulp cells (HPCs), periodontal ligament fibroblasts (HPLFs)] and two human epithelial normal oral cells [buccal mucosal human oral keratinocytes (HOKs) and primary human gingival epithelial cells (HGEPs)] (Figure 1). Cells were incubated for 48 h with increasing concentrations of test agents, and the relative viable cell number was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide method. Tumor-selectivity index (TS) was determined by dividing the mean of the concentration that reduced the viable cell number by 50% (CC₅₀) of each agent against normal cells by its mean CC₅₀ against tumor cells (Figure 2A).

We first investigated the tumor specificity of anticancer drugs using OSCC cell lines and normal mesenchymal cells (MCs) (referred to as System 1). Many anticancer drugs, such as docetaxel, 5-fluorouracil (5-FU), methotrexate, mitomicin C, etoposide, daunorubicin, doxorubicin, SN-38 (active metabolite of irinotecan), camptothecin and gefitinib, showed excellent specificity (TS=10-1000) (Figure 2B) (8). This validated the present method for evaluating anticancer activity.

We next evaluated the tumor specificity of anticancer drugs using OSCC cell lines and normal epithelial cells (ECs.) (referred to as System 2). In this case, anticancer agents showed cytotoxicity to both of these cell types to comparable extents, producing a much lower TS value (9). It was unexpected that doxorubicin would induce apoptosis [loss of cell surface microvilli, chromatin condensation, nuclear fragmentation (Figure 3A) and caspase-3 activation (Figure 3B)] in HOKs (9). It is therefore imperative to explore new anticancer drugs with less keratinocyte toxicity.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Cells used for calculation of CC₅₀ and tumor specificity (TS) value of previously investigated compounds.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

A: Exemplary calculation of tumor specificity (TS) value using the concentration that reduced the viable cell number by 50% (CC₅₀) values for tumor and normal cells. B: Comparison of TS values for antitumor agents and various natural and synthetic compounds. Data are cited from reference (8).

In Search of Antitumor Agents with Less Toxicity to Keratinocytes

We have synthesized a total of 133 compounds, which are classified into nine groups, by introducing various functional groups into distinct backbone structures (Figure 4). The most potent compound in each group was: (2E)-3-(3,4-dihydroxyphenyl)-N-[2-(4-hydroxyphenyl)ethyl]-2-propena-mide [1] (among 12 phenylpropanoid amides) (10), (2E,4E)-N-[2-(3,4-dihydroxyphenyl)ethyl]-5-(3,4-methylenedioxyphenyl)-2,4-pentadienamide [2] (among 12 piperic acid amides) (11), (E)-3-[2-(4-hydroxyphenyl)ethenyl]-6-methoxy-4H-1-benzopyran-4-one [3] (among 15 3-styrylchromones) (12, 13), (E)-3-[2-(4-chlorophenyl)ethenyl]-7-methoxy-2H-1-benzopyran [4] (among 16 3-styryl-2H-chromenes) (14), (Z)-N-[2-(3,4-dihydroxyphenyl)ethyl]-9-octadecenamide [5] (among 18 oleoylamides) (15), (3E)-2,3-dihydro-3-[(3,4-dihydroxyphenyl)methylene]-7-methoxy-4H-1-benzopyran-4-one [6] (among 17 3-benzylidenechromanones) (16), (2E)-1-(2,4-dimethoxyphenyl)-3-(4-methoxyphenyl)-2-propen-1-one [7] (among 15 chalcones) (17), (2E,4E)-5-(3,4-methylenedioxyphenyl)-2,4-pentadienoic acid (4-methoxyphenyl)methyl ester [8] (among 11 piperic acid esters) (18) and (2Z)-2-[(4-hydroxyphenyl)methylene]-3(2H)-benzofuranone [9] (among 17 aurones) (19).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Keratinocyte toxicity induced by doxorubicin (DXR). HSC-2 cells were incubated for 0 or 24 h with 10 μM doxorubicin (A), or with different concentrations of doxorubicin (B), and then processed for observation under transmission electron microscopy (A) or subjected to western blot analysis (B). Cited from (9) with permission. PARP: Poly (ADP-ribose) polymerase; GAPDH: glyceraldehyde 3-phosphate dehydrogenase.

Among these nine compounds, [3] and [4] had the highest TS values (69.0 and 59.9, respectively), as assessed with System I [OSCC vs. MCs]. Their TS values (59.9-69.0) were comparable with that of doxorubicin (63.7±95.0) and 5-FU (13.1±21.1), but greatly exceeded that of resveratrol (TS=2.4), a stilbene with anticancer activity (20) (Figure 4).

Compounds [3] and [4] also had the highest TS values when assessed with System II (OSCC vs. ECs: 204.5 and >85.1, respectively) (Figure 4). Their TS values were much higher than those of doxorubicin (1.7±1.9) and 5-FU (1.4±1.1) (Figure 4).

Treatment of HSC-2 cells with [4] induced mitochondrial vacuolization and inhibition of autophagy (as evidenced by loss of microtubule-associated protein 1A/1B-light chain 3 (LC3)-II at an early stage, followed by the induction of apoptosis (as evidenced by cleavage of poly (ADP-ribose) polymerase and caspase-3). Compound [4] increased the intracellular levels of diethanolamine and cytidine diphosphate-choline, whereas it reduced the level of choline, suggesting down-regulation of the glycerophospholipid pathway (13).

Estimation of TS by Chemical Descriptors

Six descriptors that correspond greatly with cytotoxicity against normal cells (N) and tumor cells (T), and with tumor specificity (T–N) of nine groups of compounds are listed in Table I. Generally, these descriptors did not overlap with each other. Tumor specificity of 3-styrylchromones was well corrected with molecular size (12). T–N can be estimated by molecular diameter (largest value in the distance matrix defined by the elements Dij), vsurf_DD23 (interaction with hydrophobic probe assumed surrounding the molecule) and R3 OH (4’-hydroxy substitution in the phenyl group of styryl moiety) as: T–N=0.607(±0.169)diameter – 0.121 (±0.035)vsurf_DD23 + 1.11 (±0.235)R3OH – 7.17 (±2.26), with n=15, R²=0.764, Q²=0.570. s=0.308 (Figure 5A).

Tumor selectivity of 3-styryl-2H-chromenes correlated well with six descriptors (std_dim3, BCUT_SLOGP_1, vsurf_D4, vsurf_R, vsurf_D5 and E_oop) which reflect structure connectivity and conformation, hydrophobicity, surface rugosity and out-of-plane potential energy (Table I). The T–N value of [4] can be estimated using two descriptors (vsurf_R and E_oop) as: T–N=32.1(±4.39)vsurf_R + 121(±17)E_oop – 46.1(±6.2), with n=16, R²=0.870, Q²=0.821, s=0.145 (Figure 5B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Nine groups of compounds synthesized by chemical modification of back-bone structures present in naturally occurring compounds. OSCC: Oral squamous cell carcinoma; TS: tumor specificity; MC: mesenchymal cells; EC: epithelial cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Multiple regression models for the estimation of the difference in cytotoxicity between tumor and normal cells (T–N) of 3-styrylchromones (A) and 3-styryl-2H-chromenes (B). Data are cited from references (12) and (14).

Future Directions

In our work, we have demonstrated that [3] and [4] are two new compounds that showed the highest TS and potency-selectivity expression values among the compounds tested, and they exhibited much less keratinocyte toxicity compared to doxorubicin and 5-FU (Figure 4). Chemical modification of these lead compounds by introduction of appropriate functional groups is a crucial step towards manufacturing new types of anticancer drugs with reduced keratinocyte toxicity.