Mitochondria as oncotarget: a comparison between the tetracycline analogs doxycycline and COL-3

Abstract

Tetracyclines have anticancer properties in addition to their well-known antibacterial properties. It has been proposed that tetracyclines slow metastasis and angiogenesis through inhibition of matrix metalloproteinases. However, we believe that the anticancer effect of tetracyclines is due to their inhibition of mitochondrial protein synthesis, resulting in a decrease of the mitochondrial energy generating capacity. Several groups have developed analogs that are void of antibacterial action. An example is COL-3, which is currently tested for its anticancer effects in clinical trials. We have undertaken a comparative study of the tetracycline analogs COL-3 and doxycycline, which has an antibacterial function, to further investigate the role of the mitochondrial energy generating capacity in the anticancer mechanism and, thereby, evaluate the usefulness of mitochondria as an oncotarget. Our experiments with cultures of the human A549, COLO357 and HT29 cancer cells and fibroblasts indicated that COL-3 is significantly more cytotoxic than doxycycline. Mitochondrial translation assays demonstrated that COL-3 has retained its inhibitory effect on mitochondrial protein synthesis. Both drugs caused a severe decrease in the levels of mitochondrially encoded cytochrome-c oxidase subunits and cytochrome-c oxidase activity. In addition, COL-3 produced a marked drop in the level of nuclear-encoded succinate dehydrogenase subunit A and citrate synthase activity, indicating that COL-3 has multiple inhibitory effects. Contrary to COL-3, the anticancer action of doxycycline appears to be based specifically on inhibition of mitochondrial protein synthesis, which is thought to affect rapidly proliferating cancer cells more than healthy tissue. Doxycycline is likely to cause less side effects that COL-3.

Keywords: COL-3; cancer; doxycycline; mitochondria; tetracycline.

Figure 1. Chemical structures of tetracycline, and its analogs doxycycline and COL-3

Figure 2. COL-3 is more cytotoxic than DC

Viability dose-response curves of cultured fibroblasts, and the A549, COLO357 and HT29 cancer cell lines treated with doxycycline or COL-3 for 5 days. Experiments were performed in quadruplicate. Results are shown as mean percentage of surviving cells relative to vehicle-treated cells ± standard deviation.

Figure 3. DC decreases the proliferation rate of A549 cells, whereas COL-3 kills A549 cells after 5 d of treatment

(A) Time course proliferation curves of A549 cells treated with vehicle, 19.1 µM doxycycline, or 8.1 µM COL-3 over a 5-day time period. Experiments were performed in quadruplicate. Results are shown as mean number of cells ± standard deviation. Asterisks denote statistically significant differences compared to vehicle-treated cells (p < 0.05). (B) Micrographs of A549 cell cultures treated with vehicle, 19.1 µM doxycycline or 8.1 µM COL-3 for 5 d. Arrows indicate dying cells. Scale bar: 100 μm.

Figure 4. DC and COL-3 decrease mitochondrial translation

(A) De novo mitochondrial protein synthesis of A549 cells treated with vehicle, 19.1 µM doxycycline, or 8.1 µM COL-3. Cells were pre-treated with the drugs for 1 hour, followed by a 1-hour pulse labeling with [³⁵S]-methionine in the presence of the drugs and emetine to block cytosolic protein synthesis. Denatured, 30-μg protein samples were subjected to 12% acrylamide gel electrophoresis. The gel was stained with Coomassie to verify even loading, followed by autoradiography to reveal the labeled mitochondrial translation products, indicated on the left. The migration of protein standards is indicated in the center. (B) Mean MTCO2 + MTCO3 labeling signals of the drug-treated cells relative to the mean value of vehicle-treated cells of four independent experiments. Error bars indicate standard deviation. Asterisks denote statistically significant differences (p < 0.05).

Figure 5. DC decreases the levels of mitochondrially synthesized proteins, whereas COL-3 decreases the levels of mitochondrially as well as cytosolically synthesized proteins

(A) Western blot image of denatured, 10-μg protein samples from A549 cells treated with vehicle, 19.1 µM doxycycline, or 8.1 µM COL-3 over a 5-day time period and subjected to 4–20% gel electrophoresis. The blot was developed with antibodies raised against SDHA, MTCO1, MTCO2, TOMM20 and β-actin as loading control. The migration of a protein standard marker (M) is indicated on the right. (B−E) Mean MTCO1 (B), MTCO2 (C), SDHA (D) and TOMM20 (E) signals in the treated cells over a 5-day time period, relative to the signal of cells at t = 0 in four independent experiments. Error bars indicate standard deviation. Asterisks denote statistically significant differences from vehicle-treated cells (p < 0.05).

Figure 6. DC decreases the activity of cytochrome-c oxidase, whereas COL-3 decreases the activity of cytochrome-c oxidase as well as citrate synthase

Cytochrome-c oxidase (A) and citrate synthase (B) activity in A549 cells treated with vehicle, 19.1 µM doxycycline, or 8.1 µM COL-3 over a 5-day time period. Assays were performed in quadruplicate. Results are shown as mean value ± standard deviation. Asterisks denote statistically significant differences from vehicle-treated cells (p < 0.05).

Figure 7. Absence of active cleaved caspase 3 indicates that DC and COL-3 do not induce apoptosis

(A) Western blot image of denatured, 10-μg protein samples from A549 cells treated with vehicle, 19.1 µM doxycycline, or 8.1 µM COL-3 over a 5-day time period and subjected to 12% gel electrophoresis. A549 cells treated with 1 µM staurosporine (Stau.) for 3 h served as an apoptosis positive control sample. The blot was sequentially developed with antibodies raised against caspase 3 and β-actin as loading control. The migration of a protein standard marker is indicated on the right. (B) Mean procaspase 3 signals in the treated cells, relative to the signal of cells at t = 0 in three independent experiments. Error bars indicate standard deviation.