Essential roles of Caspase-3 in facilitating Myc-induced genetic instability and carcinogenesis

The mechanism for Myc-induced genetic instability is not well understood. Here we show that sublethal activation of Caspase-3 plays an essential, facilitative role in Myc-induced genomic instability and oncogenic transformation. Overexpression of Myc resulted in increased numbers of chromosome aberrations and γH2AX foci in non-transformed MCF10A human mammary epithelial cells. However, such increases were almost completely eliminated in isogenic cells with CASP3 gene ablation. Furthermore, we show that endonuclease G, an apoptotic nuclease downstream of Caspase-3, is directly responsible for Myc-induced genetic instability. Genetic ablation of either CASP3 or ENDOG prevented Myc-induced oncogenic transformation of MCF10A cells. Taken together, we believe that Caspase-3 plays a critical, unexpected role in mediating Myc-induced genetic instability and transformation in mammalian cells. DOI: http://dx.doi.org/10.7554/eLife.26371.001


Introduction
One of the hallmarks of cancer is increased genomic instability (Hanahan and Weinberg, 2011). In addition to DNA replication errors and /or mutations induced by exposure to DNA damaging agents, over-expression of oncogenes have been shown to induce genomic instability (Bartkova et al., 2005;Gorgoulis et al., 2005). One such oncogene is Myc. As one of the most widely studied oncogenes in cancer biology, it is mutated or over-expressed in multiple types of cancer (Pelengaris et al., 2002). Myc is involved in driving cellular proliferation and promoting stem cell self-renewal under normal circumstances. When myc is overexpressed in a cell, it can cause increased genomic instability and promote carcinogenesis (Karlsson et al., 2003;Ray et al., 2006). Despite numerous studies, the mechanisms involved in myc-induced genomic instability and transformation remain controversial. There are conflicting reports on the mechanism of myc-induced genomic instability and transformation. Several studies suggest that myc-induced genomic instability and carcinogenesis is a result of an overabundance of reactive oxygen species (ROS) (Vafa et al., 2002;Felsher and Bishop, 1999). However, it has also been reported that myc overexpression can cause DNA damage and transformation in the absence of ROS (Ray et al., 2006).
Over-expression of myc has been shown to induce apoptosis (Evan et al., 1992;Harrington et al., 1994). Until recently, apoptosis has been widely recognized as an anti-carcinogenic process based on the assumption that it is utilized by the host to eliminate damaged cells, including those suffering DNA damage (Hanahan and Weinberg, 2011;Reed, 1999). However, there has been increasing evidence that apoptosis may in fact been involved in promoting carcinogenesis (Tang et al., 2012;Ichim et al., 2015;Liu et al., 2015). Mammalian cells exposed to external stress can survive activation of the apoptotic cascade and incur increased genetic instability and oncogenic transformation. Studies show that mammalian cells that survive apoptosis experience increased mitochondria membrane permeability (MOMP) (Ichim et al., 2015) and sublethal caspase 3 activation (Liu et al., 2015), which lead to activation of downstream endonucleases such as CAD and endoG, which in turn cause increased genetic instability and oncogenic transformation.
In the current study, we carried out experiments to examine the potential roles of the cellular apoptotic machinery in Myc-induced mutagenesis and carcinogenesis. We show sublethal activation of caspase 3 and endonuclease G plays an essential role in Myc-induced genetic instability and oncogenic transformation in human cells.

Results and discussion
Previously, we and others have shown that mammalian cells exposed to external stress such radiation and chemicals could survive the activation of apoptotic caspases. Among the cells that survive caspase activation, elevated DNA damage, such as DNA double strand breaks were observed. In this study, we set out to examine the hypothesis that sublethal activation of apoptotic caspases are involved in Myc-induced genetic instability. Our hypothesis is based on the well-established evidence that Myc over-expression in mammalian cells promotes caspase activation and cell death (Karlsson et al., 2003;Ray et al., 2006).
To investigate the effects of Myc expression we used a recombinant lentivirus to transduce the c-Myc gene under the control of a constitutively active CMV promoter into MCF10A cells, an immortalized but not transformed human mammary epithelial cell line Tait et al., 1990). We reasoned that if some of the Myc-expressing cells can survive caspase activation, they may possess higher levels genomic instability, similar to those cells that were exposed to ionizing radiation (Liu et al., 2015). We quantified Myc's ability to activate Casp3 by immunofluorescence staining of cleaved Casp3 ( Figure 1A,B). Roughly 6% of Myc-expressing MCF10A cells were observed with having relatively normal nuclei and cleaved caspase 3, as compared to control MCF10A cells where only~1% of the cells were observed with cleaved caspase 3. We also examined the relationship between Casp3 activation and cellular survival by use of a reporter system described in a prior publication (Liu et al., 2015). Our data indicate that irrespective of Casp3 activation status in the presence or absence of Myc expression, 40% or more of the individually sorted MCF10A cells eLife digest Healthy cells can become cancerous if their DNA is damaged and not repaired properly, leading to changes in the DNA known as mutations. The cells tend to accumulate more and more mutations -a phenomenon known as genomic instability -as they transition into cancer cells. A protein called Myc is known to promote genomic instability and contributes to many types of cancers. However, high levels of Myc proteins in a cell can also activate proteins that trigger a process called apoptosis, which makes the cell commit suicide. This role does not appear to fit with the cancer-promoting properties of Myc because apoptosis is generally thought to protect against cancer by helping to remove damaged cells from the body.
Two of the proteins that Myc activates are known as caspase-3 and endonuclease G. In healthy cells, caspase-3 triggers a series of events that lead to apoptosis, while endonuclease G cuts up DNA in preparation for cell death. However, it is not known how these proteins affect the cancerpromoting properties of Myc.
Here, Cartwright, Liu et al. used a gene editing technique called CRISPR-Cas9 to examine how these apoptosis proteins affect the ability of Myc to promote cancer. Increasing the production of Myc in healthy human mammary cells transformed these cells into breast cancer cells that were capable of forming tumors when injected into mice. However, in cells that could not make caspase-3 or endonuclease G, increasing the production of Myc did not lead to the cells becoming cancerous.
Further experiments show that when Myc levels are high, the activation of caspase-3 and endonuclease G is not sufficient to kill the cells. As a result, endonuclease G causes damage to the DNA and promotes genomic instability. Future studies may focus on understanding exactly how and when the apoptosis proteins can contribute to cancer growth. With this knowledge, it may be possible to prevent or treat Myc-driven cancers by changing how these apoptosis proteins behave.     In order to rule out the possibility that our CASP3KO cells suffered off-target effects during the generation process, we re-expressed Casp3 in CASP3KO MCF10A cells (Figure 1-figure supplement 6A) and examined for Myc-induced gH2AX foci. Our results indicate that Casp3 re-expression restored Myc-induced DNA damage foci (Figure 1-figure supplement 6B). In a parallel experiment, we expressed a dominant-negative CAPS3 gene (dnCASP3) in Casp3KO cells. DnCasp3 differs from wild-type Casp3 in only a single amino acid that eliminates its cleavage activities (Stennicke and Salvesen, 1997). In contrast to wild-type Casp3 re-expression, dnCASP3 re-expression did not restore the ability of Myc to induce gH2AX foci (Figure 1-figure supplement 6C).

DAPI
In order to make sure that our observations so far are not restricted to MCF10A cells, we generated CASP3 gene knockout cells from hTERT immortalized BJ1 human fibroblast cells (Figure 1figure supplement 7A, Table 3) to assess the effects of Casp3 on myc-induced genomic instability. Similar to MCF10A cells, we observed that Myc overexpression resulted in statistically significant increases in gH2AX foci ( Since Myc-induced genomic instability is intimately associated with its ability to transform mammalian cells, we investigated Myc-induced tumorigenicity in MCF10A cells. We initially evaluated  Table 3. Mutations at target sequences in various CRISPR knockout MCF10A and BJ-1 hTERT cells. Myc-induced tumorigenicity of MCF10A cells by use of the soft agar colony forming assay, a wellestablish assay that evaluates the anchorage independence ability of putative tumor cells. Our results indicate that Myc overexpression in control MCF10A resulted in a significant increase in the number of observed soft agar colonies (Figure 2A,B). However, such increases were completely absent in Casp3 knockout cells (Figure 2A,B). The causative role for Casp3 in this process was further demonstrated in Casp3KO cells with Casp3 re-expression, the ability of Myc to induce soft agar colony formation was restored ( Figure 2C). In control MCF10A cells, expression of an exogenous Casp3 caused no increase in Myc-induced soft agar colony formation ( Figure 2C).

MCF10A MCF10A+Myc
Casp3 The following source data is available for figure 2: Source data 1. Data for Figure 2B. In a further experiment, we examined the ability of Myc-transduced control or CASP3KO cells to form tumors in nude mice. Despite injection of 4 Â 10 6 cells each, only the group of nude mice that were injected with Myc-transduced control MCF10A cell were able to form tumors after 8 weeks of observation ( Figure 2D), thereby confirming that both Myc over-expression and an intact CAPSP3 gene are required for oncogenic transformation. A more detailed examination showed that in Myc over-expressing MCF10A cells, five out of eight injected sites formed tumors, albeit with varying growth kinetics ( Figure 2E). The in vivo data here are striking in that Casp3 deficiency completely blocked the ability of Myc to transform MCF10A cells. In vitro, although Casp3 deficiency significantly reduced Myc-induced soft agar colony formation, it was not completely blocked. The discrepancy between the in vitro and in vivo assays perhaps reflected the different properties that the two assays are measuring.
We next sought to determine the downstream effectors of Casp3 in mediating Myc-induced genomic instability and oncogenic transformation. In a previous study, endonuclease G (endo G), an apoptotic endonuclease that normally resides within the mitochondria and migrates to the nucleus after activation of apoptotic cascade (Parrish et al., 2001;Li et al., 2001), is responsible for much of the Casp3-induced genomic instability after stress exposure. In order to evaluate if endoG is involved in Myc-induced genomic stability, we determined the cellular location of endoG in control and Myc expressing MCF10A cells by use of immunofluorescence staining. Our results show that there was a significant increase in the fraction of cells with endoG nuclear migration in Myc-expressing MCF10A vs control cells (11% vs 2%, Figure 3A,B). However, the increase was completely eliminated in CASP3 knockout cells ( Figure 3B).
In order to determine if endoG plays a causative role in Myc-induced genomic damage and transformation, we created MCF10A cells with ENDOG gene knockout by use of CRISPR/Cas9 ( Figure 3C). We then evaluated the abilities of Myc to induce gH2AX foci and oncogenic transformation. Our results show that ENDOG deletion was able to eliminate both Myc-induced gH2AX foci ( Figure 3D and Figure 3-figure supplement 1) and oncogenic transformation as evaluated by use of the soft agar colony-forming assay ( Figure 3E).
To further determine the relationships among Myc expression, CASP3 status, ENDOG status, and apoptosis, we carried out flow cytometry analysis of Annexin V and PI (propidium iondine) staining, which allowed for the identification of different stages of cellular apoptosis. Our results (Figure 3figure supplement 2) indicate that CASP3KO caused small reductions in both Annexin V+ and Annexin V+/PI+ MCF10A cells when compared with control cells. However, ENDOG knockout caused small increases in fractions of Annexin V+/PI-and Annexin V-/PI+ cells when compared with control cells. Myc expression, on the other hand, caused increases in fractions of Annexin V+/PI+ cells I all three cell populations. Overall, while the relative changes could be sizable, the absolute changes caused by the knockouts or Myc expression in terms of PI+ or Annexin V+ cells were small. Increased ROS has been previously implicated in Myc-induced carcinogenesis. Our data so far has suggested strongly that the Casp3 activation and endoG release from the mitochondria played decisive roles in Myc-induced carcinogenesis. In order to determine if ROS levels in MCF10A cells correspond with oncogenic transformation, we did DCFDA-based flow cytometry analysis ROS levels in various MCF10A-derived cells (Figure 3-figure supplement 3). Our results indicate that CASP3 or ENDOG knockout caused small increase and decrease in MCF10A cells (Figure 3-figure supplement 3, top panels), respectively. Forced MYC expression increased ROS in wild-type MCF10A cells but not in CASP3KO or ENDOGKO MCF10A cells (Figure 3-figure supplement 3, mid-panels). Further, Myc was able to induce significantly more ROS in wild-type MCF10A cells than CASP3KO cells, but about equal levels of ROS in wild type vs ENDOGKO cells (Figure 3-figure supplement 3, lower panels). Those data suggest that although ROS production appeared to track with carcinogenesis in wild-type and CASP3KO cells, it did not in ENDOGKO cells.
In order to determine whether Casp3 activation and endoG release is the result of partial damage of many mitochondria vs severe damage to a small number of mitochondria, a quantitative PCR (qPCR) analysis of cytoplasmic mitochondrial DNA (mtDNA) levels was done (Figure 3-figure supplement 4) Figure 3 continued on next page this, we speculate that Casp3 and other upstream factors that promote mitochondrial leakage form a positive loop to promote mitochondrial leakage with or without Myc expression.
To further examine if endoG leakage from the mitochondria and migration is sufficient to cause genomic instability and oncogenic transformation, we engineered a modified endoG protein where the native mitochondrial localization signal is switched to a nuclear localization signal (NLS-EndoG, Figure 4A) and transduced it into MCF10ACASP3KO cells with or without Myc expression ( Figure 4B). Immunofluorescence staining confirmed the nuclear localization of the engineered endoG ( Figure 3C). We then determined the incidence of gH2AX foci in transduced cells. Our results indicate the NLS-endoG expression restored the depleted gH2AX foci induction by Myc in CASO3KO cells ( Figure 4D). In fact, NLS-endoG alone was sufficient to induce gH2X foci in CASP3KO cells to levels induced by Myc ( Figure 4D).
We next examined the influence of NLS-EndoG on oncogenic transformation by use of the soft agar assay. Our results show that NLS-EndoG restored the ability of Myc to induce oncogenic transformation in Casp3 knockout MCF10A cells ( Figure 4E). However, NLS-EndoG expression alone was not sufficient to induce oncogenic transformation despite its ability to induce DNA damage foci to levels similar to Myc expression ( Figure 4E). Those results suggest while endoG nuclear migration is a necessary condition for Myc-induced oncogenic transformation, it is not sufficient by itself. Additional activities of Myc are clearly required in the transformation process.
A further tumor formation experiment was conducted in nude mice to examine the role of NLS-EndoG in oncogenic transformation. NLS-EndoG restored the ability of Myc to induce oncogenic transformation in CASP3KO MCF10A cells ( Figure 4F), with five out of eight inoculations formed tumors. On the other hand, NLS-EndoG alone was not able to make MCF10A cells with CASP3 gene knockout to become tumorigenic, consistent with results from soft agar colony formation assay ( Figure 4E).
Despite being one of the first oncogenes identified and having numerous studies dedicated to discovering its roles in cancer biology, there are still many gaps in our knowledge about Myc. The present study provides significant new insights into the roles of Myc in carcinogenesis. In particular, it resolves two apparently paradoxical observations regarding Myc: its ability to stimulate apoptosis and to induce genomic instability and oncogenic transformation. In many instances, such as in the case of p53, apoptosis induction is anti-carcinogenic due to its ability to remove damaged cells from the body. However, in the case of Myc, the apoptotic machinery, is exploited by Myc as a vehicle to cause genomic instability and induce oncogenic transformation. Most strikingly, our data suggest that Casp3 and Endo G, two well-established apoptosis effectors, are activated and required for Myc-induced oncogenic transformation. This finding suggest that Myc-induced activation of apoptosis, instead of being a result of Myc induced cellular stress, is actually part and parcel of Myc's capacity to induce mammalian transformation.
One important piece of information that remain missing is how Myc induces mitochondrial leakage and Casp3 activation. While we do not have any experimental evidence at present, it is possible that deregulation in mitochondrial biogenesis, which is known to be stimulated by Myc, may be responsible for it, as suggested previously (Dang et al., 2005;Zhang et al., 2007;Ahuja et al., 2010). This possibility should be investigated in future studies since it may holds the key to Myc's powerful oncogenic abilities.
On the surface, our discovery appears to be contrary to the established paradigm that apoptosis is a key barrier for carcinogenesis (Hanahan and Weinberg, 2011). However, it is consistent with an increasing body of literature that suggest a pro-oncogenic role for apoptosis and some apoptotic factors (Tang et al., 2012;Ichim et al., 2015;Liu et al., 2015). The key conceptual advance in the studies is the realization of that cells exposed to stress can survive caspase activation (Tang et al., 2012;Ichim et al., 2015;Liu et al., 2015;Ding et al., 2016;Ichim and Tait, 2016). The observation of such survival in development (Ding et al., 2016), chemical exposure (Tang et al., 2012;  Ichim et al., 2015), and radiation exposure (Liu et al., 2015) indicate that it is a wide spread phenomenon. Our observation of cells surviving Myc-induced caspase activation is a significant extension of those earlier observations and may provide important insights into the relationship between apoptosis and oncogene-induced transformation beyond that of Myc.
Taken together, our study provides crucial evidence that Casp3 and endo G, two key factors in the canonical apoptosis pathway, play essential, facilitative roles in Myc-induced oncogenic transformation.

Materials and methods
Cell lines and tissue culture medium was composed of Dulbacco's modified Eagle's medium (DMEM)/F12 (Sigma) supplemented with 5% donor horse serum (Sigma), 20 ng/ml of epidermal growth factor (EGF; R&D Systems), 0.5 mg/ml hydrocortisone (Sigma), 100 ng/ml cholera toxin (Sigma), 10 mg/ml insulin (Invitrogen), and 100 units/ml penicillin and 100 mg/ml streptomycin. hTERT immortalized, non-transformed human fibroblast cell line, hTERT BJ-1, was a kind gift from Dr. Takamitsu Kato of Colorado State University. hTERT BJ-1 growth medium was composed of DMEM supplemented with 10% fetal bovine serum (Sigma) and 100 units/ml penicillin and 100 mg/ml streptomycin. The identities of both MCF10A and hTERT-BJ1 were authenticated through STR profiling methods. Throughout the course of the experiments, the cells were also checked periodically for the absence of mycoplasma contamination.
gH2AX foci assay For gH2AX foci assays, the cells were synchronized in G1 using the well-established double-thymidine block protocol (Bostock et al., 1971). Briefly, cells were plated on glass-bottom 35 mm petri dishes (MatTek, Ashland, MA) and cultured with growth medium for overnight. They were incubated with 2 mM thymidine for 18 hr, washed 2x with PBS, and incubated for 10-12 hr in growth media. They were then incubated for 15-18 hr with 2 mM thymidine. After synchronization, cells were fixed with 4% PFA and permeabilized and blocked in PBS containing 0.1% Triton X-100, 5% goat serum, and 1% BSA. Cells were incubated with a primary antibody against gH2AX (Upstate Biotechnology, Lake Placid, NY), wash with PBS and incubated with a secondary antibody conjugated with Alexa Fluor 488 (Invitrogen). Cells were mounted with mounting medium (Vector Laboratories) containing DAPI. Fluorescent images of gH2AX were acquired with a Zeiss fluorescence microscope with a 63x oil objective (Axio Observer Z1). For each experimental group we observed 150 cells in triplicate.

Chromosome aberration analysis
We carried out chromosome aberration analysis in cultured cells following an established protocol (Savage, 1976). We analyzed for various chromosome/chromatid aberrations that include breaks/ gaps, dicentrics, centric/acentric rings, and translocations. Each data points represent data from 50 cells in triplicate.

CRISPR/Cas9-mediated gene knockout and lentivirus production
We made various cells deficient in various genes by use of the CRISPR/Cas9 technology. Singleguided RNA (sgRNA) sequences targeting the genes were generated with the use of a free online CRISPR design tool (crispr.mit.edu). The sgRNA sequences used were listed in Table 2. Annealed double stranded sgRNA oligos were ligated into the lentiCRISPR vector (Cong et al., 2013) (deposited by Dr. Feng Zhang to Addgene, Cambridge, MA) at BsmBl site, which co-express cas9 and sgRNA in the same vector. The constructed CRISPR lentivirus vectors were then packaged according to a standard protocol. To produce lentiviral vectors, lentiviral plasmids with the target genes were transduced into 293 T cells together with second-generation packaging plasmids (psPAX2, pMD2.G) following previously published procedures: http://tronolab.epfl.ch/lentivectors.

Immunofluorescence staining
Cells were cultured on glass-bottom 35 mm petri dishes. Cells were fixed with 4% paraformaldehyde (PFA) in PBS for 15 min, permeabilized and blocked with PBS containing 5% goat serum, 0.1% Triton X-100, and 1% bovine serum albumin (BSA) for 45 min. Fixed cells were incubated with primary antibodies for cleaved Caspase-3, gH2AX, or EndoG overnight at 4C, followed by incubation with appropriate Alexa Fluor 488-conjugated secondary antibodies (Invitrogen, Carlsbad, CA) for 1 hr and mounted with mounting medium (Vector Laboratories, CA) containing DAPI. Fluorescent images were acquired with a Zeiss fluorescence microscope with a 63x oil objective (Axio Observer Z1).

Soft-agar assay
The soft-agar assay was carried out following an established procedure (Cifone and Fidler, 1980). About 1 Â 10 4 MCF10A cells in growth medium were plated into six-well plates with 1.5 ml 0.3% (m/v) low-melting agar (BD, Sparks, MD, which was overlaid onto 1.5 ml 0.6% (w/v) bottom agar layer. Soft-agar colonies were maintained at 37˚C and fed twice a week by drop-wise addition of growth medium for colony formation. After 21 days in culture, the colonies were counted after staining with 0.005% crystal violet.

Flow cytometry-based analysis of ROS
In order to measure reactive oxygen species, the cells were labeled with DCFDA (20 mM) according to the manufacturer's instruction that comes with the ROS kit (Abcam, Cambridge, MA). The cells were then analyzed by use of flow cytometry.

Tumor formation in nude mice
Animal experiments conducted in this study were approved by the Duke University Institutional Animal Use and Care Committee (protocol# A195-14-08). To confirm the tumorigenicity of myc overexpressing MCF10A cells, about 4 Â 10 6 cells were injected subcutaneously into the flanks of 6-8 week-old, female athymic nude mice (Jackson Laboratories) in 50 ml of 1:1 Matrigel (Corning) and PBS. After inoculation, the growth of tumors was evaluated once a week for 8 weeks.

Statistical analysis
Student's t-test was used to evaluate the significance of differences between different experimental groups. In most cases, a p-value of less than 0.05 was deemed as significant while a p-value of more than 0.05 was deemed not significant.