X-ray Crystallography-Guided Design, Antitumor Efficacy, and QSAR Analysis of Metabolically Stable Cyclopenta-Pyrimidinyl Dihydroquinoxalinone as a Potent Tubulin Polymerization Inhibitor

Abstract

Small molecules that interact with the colchicine binding site in tubulin have demonstrated therapeutic efficacy in treating cancers. We report the design, syntheses, and antitumor efficacies of new analogues of pyridopyrimidine and hydroquinoxalinone compounds with improved drug-like characteristics. Eight analogues, 5j, 5k, 5l, 5m, 5n, 5r, 5t, and 5u, showed significant improvement in metabolic stability and demonstrated strong antiproliferative potency in a panel of human cancer cell lines, including melanoma, lung cancer, and breast cancer. We report crystal structures of tubulin in complex with five representative compounds, 5j, 5k, 5l, 5m, and 5t, providing direct confirmation for their binding to the colchicine site in tubulin. A quantitative structure-activity relationship analysis of the synthesized analogues showed strong ability to predict potency. In vivo, 5m (4 mg/kg) and 5t (5 mg/kg) significantly inhibited tumor growth as well as melanoma spontaneous metastasis into the lung and liver against a highly paclitaxel-resistant A375/TxR xenograft model.

Figure 1.

Design and synthesis of novel dihydroquinoxalinone pyrimidine analogues.

Figure 2.

Binding patterns of dihydroquinoxalinone pyrimidine analogues with tubulin and their localization within the cell. (A) Tubulin polymerization assay of 5m (10 μM) and 5t (10 μM) using the tubulin protein from the Bovine brain origin. Identical concentrations of paclitaxel and colchicine were used in this assay as standard controls. (B) Morphology and α-tubulin distribution of A375/TxR melanoma cells in the interphase (top) and mitosis (bottom) after treatment with 2 nM of colchicine, paclitaxel, 5m, and 5t in vitro. Cells were stained for α-tubulin (red) and the nucleus (blue).

Figure 3.

X-ray structures of tubulin-RB3-SLD-TTL proteins in complex with 5j, 5k, 5l, 5m, and 5t. (A) Original complex with compound 2a at a 2.6 Å resolution that formed the basis of this study (PDB ID: 6N47). (B) Compound 5j at a 2.9 Å resolution. (C) Compound 5k at a 2.8 Å resolution. (D) Compound 5l at a 2.9 Å resolution. (E) Compound 5m at a 2.7 Å resolution. (F) Compound 5t at a 2.7 Å resolution. (G) Overlay of compounds 5m and 5t. (H) Colchicine-bound tubulin complex (PDB ID: 5XIW). In all panels, the tubulin α-monomer is shown in cyan and the β-monomer is shown in gold.

Figure 4.

Effects of compounds 5m and 5t on A375/TxR cell migration. Scratches were created by the wound maker. Representative images of wound healing were captured by IncuCyte after 12 or 24 h treatment with 5m (A) and 5t (B), and blue lines demonstrate the wound edges of the cell monolayer. Wound closure is shown as the percent of relative wound density at each time point.

Figure 5.

Score plot of PCA of QSAR of drug candidates on (A) melanoma cancer cell lines and (B) breast cancer cell lines.

Figure 6.

Dihydroquinoxalinone pyrimidine analogue 5m inhibited the tumor growth in the A375/TxR xenograft study. 10 mg/kg paclitaxel treatment was included as the reference control. 5m was intravenously (IV) administered with two different doses (2 and 4 mg/kg) for every 2 times per week. Similarly, paclitaxel was administered at a dose of 10 mg/kg (IV) with the same frequency as 5m. Tumor growth of the injected tumor (A) and the mice body weight (B) were measured during the therapy. After 21 days of treatment, the tumors were excised and the final tumor weight was measured (C) and the image with all tumors was captured (D). The significant differences between groups were determined by one-way ANOVA, followed by Dunnett’s multiple comparison test (*p < 0.05, ****p < 0.0001 vs control). Data = mean ± SD. Isotonic saline was administered by the IV route in the control group.

Figure 7.

Dihydroquinoxalinone pyrimidine analogue 5t inhibited the tumor growth of A375/TxR xenografts in NSG mice. The mice were treated with intravenous injections of 2.5 or 5 mg/kg 5t 2 times per week. The paclitaxel-treated group (10 mg/kg) was used as a positive control. (A) Tumor growth curve of inoculated A375/TxR xenografts. (B) Changes of mice body weight. (C) Final tumor weight at the study endpoint. (D) Representative tumor images in this study. Data are presented as the mean ± SD. *p < 0.05, ****p < 0.0001 vs control, as analyzed by one-way ANOVA, followed by Dunnett’s multiple comparison test.

Figure 8.

Necrosis in A375/TxR tumors caused by 5m or 5t treatment. Tumors were harvested, fixed, embedded, sectioned, and stained with H&E; slides from the 5m xenograft model (upper panel) and 5t xenograft model (lower panel) were scanned using a Pannoramic FLASH III system; and representative images were captured using CaseViewer. Necrotic cells are indicated by yellow arrows, and cells in the mitotic phase are indicated by black arrows.

Figure 9.

5m displayed anti-lung and liver spontaneous metastases effects in the A375/TxR subcutaneous xenograft model. After 21 days of treatment, the mice bearing A375/TxR tumors described in Figure 6 were sacrificed, and the lung and liver tissues were harvested, fixed, and stained with H&E or the anti-human mitochondria antibody. The number of lung (A) or liver (B) metastases was counted in all the mice by averaging the percentage of the metastatic area of three to four representative H&E images of each mouse (****p < 0.0001 vs control). The anti-human mitochondria IHC staining of lung (C) and liver (D) tissues indicated dense tissue distribution (red triangle) that suggested the presence of melanoma metastases in vehicle or paclitaxel-treated mice, and the metastases were reduced with 5m treatment. Images were obtained using a Keyence BZ-X700 microscope. Brown staining indicates the metastases.

Figure 10.

In vivo efficacy of 5t in lung or liver spontaneous metastasis. Scatter plots of the mean ± SEM show the quantification of metastases present in the lung (A) and liver (B) through counting the average of the percentage of the metastatic area in three to four representative H&E images of each mouse (*p =0.016, ****p < 0.0001 vs control). Immunohistochemical staining for anti-human specific mitochondria to detect metastases that indicates dense tissue distribution (red triangle) in lung (C) and liver (D) sections. Images were obtained using a Keyence BZ-X700 microscope. Brown staining indicates the metastases.

Scheme 1.

Design and Synthesis of Modified C- and D-Ring Pyridopyrimidine Analogues

Scheme 2.

Design and Synthesis of Modified A- and B-Ring Dihydroquinoxalinone Analogues

Scheme 3.

Synthesis of Ethylamine-Substituted B-Ring Dihydroquinoxalinone Analogues