Caspase-3 promotes genetic instability and carcinogenesis

Abstract

Apoptosis is typically considered an anti-oncogenic process since caspase activation can promote the elimination of genetically unstable or damaged cells. We report that a central effector of apoptosis, caspase-3, facilitates rather than suppresses chemical- and radiation-induced genetic instability and carcinogenesis. We found that a significant fraction of mammalian cells treated with ionizing radiation can survive despite caspase-3 activation. Moreover, this sublethal activation of caspase-3 promoted persistent DNA damage and oncogenic transformation. In addition, chemically induced skin carcinogenesis was significantly reduced in mice genetically deficient in caspase-3. Furthermore, attenuation of EndoG activity significantly reduced radiation-induced DNA damage and oncogenic transformation, identifying EndoG as a downstream effector of caspase-3 in this pathway. Our findings suggest that rather than acting as a broad inhibitor of carcinogenesis, caspase-3 activation may contribute to genome instability and play a pivotal role in tumor formation following damage.

Figure 1. Non-lethal activation of caspases 3 in MCF10A cells exposed to ionizing radiation

A). Diagram of the caspase 3 reporter gene. (Ub)₉, a nine-ubiquitin polyubiqutin domain that serves as the proteasome recognition signal that causes the rapid degradation of the reporter protein, which consists of EGFP and firefly luciferase linked by a flexible linker. B). Irradiated (0.5 Gy 600MeV ⁵⁶Fe ions) MCF10A cells with low (top panels) and high (lower panels) caspase 3 reporter activities after separation by a FACS sorter based on Luc-GFP expression levels 7 days after cellular exposure to radiation. C). Western blot analysis of caspase 3 cleavage and activation in cells with high and low Casp3Luc-EGFP reporter activities 14 days after irradiation. Cytochrome c was probed using cytosolic extracts while other proteins were probed with whole cell lysates. D). Flow cytometry profiles of MCF10A-Casp3EGFP reporter activities in cells exposed to different doses of x-rays. Cells were analyzed 4 days after irradiation. Cells were gated into 8 different groups for colony forming assays according to their fluorescence intensity levels (M1-M8). E). The mean (geometric) GFP fluorescence intensities of gated MCF10A cells. F). Distribution of MCF10A cells in each gate after different irradiation doses. For each radiation dose, M1+M2+…+M8=100%. G). Colony forming abilities of cells with different Casp3EGFP reporter activities. Cell from varying levels of reporter activities (M1-M8) were flow-sorted into individual wells of 96-well plates at 1 cell/well. Two weeks later, the numbers of MCF10A colonies on each 96-well plate were counted and plotted. Error bars in 1E, 1F, and 1G represent standard deviation. All values are derived from the average of three independent experiments. See also Figure S1.

Figure 2. A key role for activated caspase 3 in radiation induced foci formation

A). Typical micrographs of γH2AX foci in control (0 Gy) and 0.5 Gy ⁵⁶Fe ions irradiated MCF10A cells. B). The average number of γH2AX foci in caspase 3 reporter-transduced MCF10A cells exposed 0.5 Gy of ⁵⁶Fe ions 14 days after ⁵⁶Fe ions irradiation. The left two bars shows the numbers for non-irradiated vs 0.5 Gy irradiated cells while the right two bars show the numbers for irradiated MCF10A cells with high and low caspase 3 reporter activities based on GFP expression levels. C). The effect of caspase 3 expression attenuation on γH2AX foci formation. MCF10A cells transduced with an shRNA gene against the CASP3 gene (shCASP3) or control (shControl) were evaluated for γH2AX foci formation at 10 days after 0.5 Gy or sham irradiation. D). The effect of caspase 3 inhibition on radiation induced γH2AX foci formation in MCF10A cells. Cells transduced with a control lentiviral vector or those transduced with a dominant negative caspase 3 gene (CASP3DN) were evaluated for γH2AX foci formation with sham or 0.5 Gy of irradiation at 14 days after irradiation. E). Typical micrographs of γH2AX foci in control (0 Gy) and 3 Gy x-Ray irradiated MEF cells at 4 days after irradiation. F). The numbers of γH2AX foci in irradiated caspase 3 deficiency MEFs were significantly lower than those of caspase 3 proficiency MEFs at 4 days after irradiation. Error bars in 2B, 2C, 2D and 2F represent standard error of the mean (SEM). All values are derived from the average of three independent experiments. See also Figure S1–S3.

Figure 3. A key role for caspase 3 in radiation-induced DNA damage as determined by the comet assay

A). Typical examples of control and irradiated cells when run through electrophoresis during the comet assay. B). The fraction of DNA in the comet tail in caspase 3 reporter-transduced MCF10A cells exposed 0.5 Gy of ⁵⁶Fe ions 14 days after ⁵⁶Fe ions irradiation. The left two bars shows fractions for non-irradiated vs 0.5 Gy irradiated cells while the right two bars show the fractions for irradiated MCF10A cells with high and low caspase 3 reporter activities based on GFP expression levels. C). The effect of caspase 3 expression attenuation on the fraction of DNA in the comet tails. MCF10A cells transduced with an shRNA gene against the CASP3 gene (shCASP3) or control (shControl) were evaluated for the fraction of DNA in the comet tails at 10 days after 0.5 Gy or sham irradiation. D). The effect of caspase 3 inhibition on radiation induced DNA strand breaks as quantified by the amount of cellular DNA in the comet tail in MCF10A cells. Cells transduced with a control lentiviral vector or those transduced with a dominant negative caspase 3 gene (CASP3DN) were evaluated for the amount of DNA in the comet tails with sham or 0.5 Gy of irradiation at 14 days after irradiation. E). The effect of caspase 3 attenuation on the fraction of DNA in the comet tails in IMR90 cells. IMR90 cells transduced transduced with an shRNA gene against the CASP3 gene (shCASP3) or control (shControl) were evaluated for the fraction of DNA in the comet tails at 10 days after 0.5 Gy or sham irradiation. Error bars in 3B, 3C, 3D and 3E represent standard deviation. All values are derived from the average of three different samples within one experimental group. For each group, at least 50 cells from randomly chosen fields were scored by use of the Image J software (NIH) for DNA distribution. Two sided Student’s t-test was used to calculate the p-values. See also Fig. S3&S4.

Figure 4. A key role for caspase 3 activities in radiation induced chromosome aberrations

A). The effect of caspase 3 inhibition on radiation induced chromosome aberrations in MCF10A cells. Cells transduced with a control lentiviral vector or a dominant negative caspase 3 gene (casp3DN) were evaluated for chromosome aberrations with sham or 0.5 Gy of 600MeV ⁵⁶Fe ions irradiation at 14 days after irradiation. B). Radiation-induced chromosome aberrations in wild type or CASP3 gene knockout (CASP3KO) C57BL/6 mice. Results were obtained from bone marrow cells harvested from irradiated mice 3 days after 3 Gy x-ray irradiation. C). Whole chromosome painting results for chromosomes 1 (FITC) & 2 (rhodamine). A cellular chromosome spread showing a translocation of part of chromosome 2 (arrow in red) in irradiated wild type C57BL/6 mice. D). Quantitative summary of radiation-induced chromosome 1 &2 translocations in wild type or CASP3 gene knockout (CASP3KO) C57BL/6 mice. Results were obtained from bone marrow cells harvested from irradiated mice 3 days after. No translocations were identified in cells from either wild type or CASP3KO mice without irradiation. E). Radiation induced chromosomal aberrations in WT, Casp3KO, Casp7KO, and Casp3KOCasp7KO (DKO) mouse embryonic fibroblast cells 5 days post irradiation. Error bars in 4A, 4B, 4D, and 4E represent standard deviations. All values are derived from the average of three separate experiments. In each experiment, 150 cells for each cell type were counted without knowledge of the cell types to enumerate chromosome aberrations. Two-tailed Student’s t-test was used to calculate the p-values. See also Figure S4 and Tables S2&S3.

Figure 5. A facilitative role for caspase 3 in radiation-induced oncogenic transformation of MCF10A cells

A). Soft agar colony formation in MCF10A cells with high and low caspase 3 reporter activities. Irradiated (0.5 Gy of 600MeV ⁵⁶Fe ions) cells were sorted for high and low caspase 3 reporter expression by use of an FACS sorter. They were then cultured for 14 days and seeded in soft agar plates. B). The effects of small hairpin caspase 3 (shCASP3) gene expression on soft agar colony formation in irradiated MCF10A cells. C). The effects of dominant-negative caspase 3 (CASP3DN) expression on soft agar colony formation ability of irradiated MCF10A cells. Error bars in 5A, 5B, and 5C represent standard error of the mean (SEM). All values are derived from the average of three independent experiments. Student’s t-test was used to calculate the p-values in A–C. D). A xenograft tumor in nude mice after 5 weeks inoculation of irradiated MCF10A cells (indicated by arrow in red). E). H&E staining of tissues derived from a xenograft tumor formed from irradiated MCF10A cells in nude mice. The right panel shows an amplified image of part of the left panel. F). Tumorigenic abilities of different MCF10A cells. Only the wild type MCF10A cells irradiated with 0.5 Gy can form tumors in nude mice. For each group of tumor cells, 10 mice were used. G). Tumor growth kinetics from each of the 10 mice injected with the wild type MCF10A cells irradiated with 0.5 Gy of ⁵⁶Fe ions. Six out of ten mice showed tumor growth. See also Figure S5.

Figure 6. A facilitative role for caspase 3 in two-stage chemically induced skin carcinogenesis in wild type and Casp3 (−/−) C57BL/6 mice

A). Activated caspase 3 in DMBA+TPA treated pre-malignant mouse skin. Scale bars represent 50 µm. B). Quantitative measurement of caspase 3 activation in DMBA+TPA treated C57BL/6 mouse skin. Error bars show standard deviation (n=3). C). Photographs representative of skin tumor formation in wild type and Casp3 (−/−) C57BL/6 mice at 20 weeks after the initiation of two-stage chemical treatment. D). Tumor incidence in DMBA+TPA treated wild type (WT) (n=23) and Casp3 (−/−) (n=26) mice. p<0.001, logrank test. E). Average number of tumors per mouse in DMBA+TPA treated wild type and Casp3 (−/−) mice at 20 weeks after initiation of chemical treatment. The difference between the two groups is statistically significant (p <0.001, two-sided ANOVA test). F). Average tumor volume per mouse during the course of two-stage chemical induction. For each mouse, tumor burden represent the aggregate of all tumors in the mouse. The difference between the two groups is statistically significant (p <0.001, two-sided ANOVA test). Error bars in 6E&F represent standard error of the mean (SEM). See also Fig. S5&S6.

Figure 7. A significant role for endonuclease G as a downstream factor of caspase 3 in mediating radiation induced DNA damage and transformation

A). Immunofluorescence co-staining of endonuclease G (green) and mitochondria (red) in control (top panels) and irradiated (lower panels) MCF10A cells. Insets: DAPI staining of the same slides. B). Effect of caspase 3 inhibition (through CASP3DN or shCasp3 expression) on radiation-induced endoG migration from cytoplasmic to nuclear locations. C). Western blot analysis of endoG in the mitochondria, cytoplasmic and nuclear fractions of MCF10A cells with or without Casp3DN or shCasp3 expression. Mito marker, TBP (TATA binding protein) and β-actin were used as mitochondria, nuclear and cytoplasmic loading controls, respectively. D). Immunofluorescence co-staining of endoG (green) and γH2AX foci (red) in irradiated MCF10A cells. Insets: DAPI staining of the same slides. E). Fraction of cells with γH2AX foci among cells with endoG staining in the cytoplasm only or those in both the cytoplasm and nucleus. F). Effect of attenuating endoG expression on radiation induced γH2AX foci in MCF10A cells. G). Effects of attenuating endoG expression on radiation induced comet tail DNA amount (an estimate of DNA strand breaks) in MCF10A cells. H). Effects of attenuating endoG expression on the fraction of cells with radiation-induced chromosome aberrations in MCF10A cells. I). Effects of attenuating endoG expression on the soft agar colony forming abilities of irradiated MCF10A cells. In F–I, shEndoG represents shRNA2 from supplementary Fig.S7F. Error bars in 7B, 7E–I represent standard error of the mean (SEM). The values in 7B, 7E, 7G and 7H are derived from the average of three samples within of each cell group. The values in 7F and 7I are derived from the average of three separate experiments. For γH2AX foci, at least 200 cells were counted for each measurement. For comet assay, a minimum of randomly chosen 50 cells was measured automatically by Image J software (NIH) for DNA distribution. For chromosome aberration analysis, a minimum of 150 cells was counted without knowledge of cell identities. Student’s t-test was used to calculate the p-values. See also Figure S7.