Molecular imaging provides the ability to dynamically and non-invasively monitor specific biological pathways in a single living subject. Recent advances in the field have enabled imaging of protein-protein interactions using split reporter constructs (14
), and others(32
) have used molecular imaging to monitor processes such as cell proliferation or protein degradation. Here we describe a strategy for imaging apoptosis using a luciferase complementation assay with enhanced signal to noise over our previous efforts (35
). Increased signal to noise was achieved by reducing interaction between NLuc and CLuc fragments by steric hinderance (to decrease background) and by increasing NLuc and CLuc fragment complementation via the high affinity interaction of pepA and pepB (to increase signal). Although some self-complementation is observed by NLuc and CLuc proteins when coexpressed in culture (14
), addition of pepA and pepB to the split luciferase fragments increased bioluminescence activity observed by ~20-fold (data not shown). Additionally, expression of ANLuc and BCLuc components in parallel as a hybrid protein (ANLucBCLuc) hindered self-complementation between the NLuc and CLuc fragments, reducing bioluminescence activity observed to ~38% of coexpressed NLuc and CLuc fragments (data not shown). Bioluminescence observed by expression of ANLucBCLuc was reduced to ~2% of that observed by coexpression of pepANLuc and pepBCLuc fragments (data not shown). This approach both lowered background and increased luciferase signal, greatly enhancing the signal-to-noise of the ANLucBCLuc reporter construct. The highest bioluminescence activity exhibited in early studies was approximately 6–8 fold over pretreatment values (35
). In contrast, activation of the ANLucBCLuc reporter has consistently resulted in an induction of bioluminescence activity of 6–75 fold over pretreatment values under various conditions.
Time and dose dependence studies revealed that bioluminescence activity and cleavage of ANLucBCLuc occurs in conditions wherein caspases-3 is active. In dose dependence studies, increased bioluminescence activity was observed at concentrations lower than detectable active caspases-3 by western blot analysis (). This suggests that the reporter may be a more sensitive measure of caspase-3 activation than traditional western blot analysis. Although not investigated here, this reporter may also be activated by active caspase-7, which cleaves proteins containing the DEVD cleavage motif (37
The quantitative nature of this reporter technology was demonstrated by comparing bioluminescence activity and active caspase-3 staining of tumors from animals treated with vehicle, 70mg/kg, or 140mg/kg TMZ as shown in and . Bioluminescence activity in tumors of animals receiving 140mg/kg TMZ increased ~12-fold over that observed in tumors of animals treated with vehicle, which correlated well with the ~12-fold increase in tumor cells staining positive for active caspase-3. Additionally tumors from animals treated with 70mg/kg TMZ had a ~5-fold increase in bioluminescence activity compared to control, which also paralleled the ~5-fold increase in cells staining positive for active caspase-3 from these tumors. This suggests that ANLucBCLuc can be used in a quantitative manner to determine whether increasing doses of drug are adding therapeutic benefit by enhancing apoptosis.
The utility of this technology in an experimental therapeutics setting was demonstrated in various contexts. First, the ability of this reporter to non-invasively report on the efficacy of combination therapies is demonstrated in and . Treatment of orthotopic GBM tumors to single modality therapy (TMZ or RT) as well as combination therapy (TMZ + RT) in both short term (36 hours) and long term (two week) studies revealed enhanced apoptotic cell death when animals were treated with combinatorial therapy. Second, the utility of the reporter in scheduling optimization was demonstrated. For example, data presented in show that application of a single dose of RT to animals pretreated with TMZ 18 hours prior was sufficient to induce a significant enhancement in apoptosis, while RT alone did not result in a significant induction of apoptosis.
An apoptosis reporter such as ANLucBCLuc that can be utilized to monitor apoptosis dynamically and non-invasively provides the opportunity to rationally design fractionated therapeutic protocols. For example, data presented in demonstrate that a single dose of TMZ results in peak apoptosis three hours after administration and subsides significantly after 12 hours. Re-treatment of animals at this time would be rational in that systemic toxicity would be minimized while still providing prolonged periods of increased apoptosis. Indeed, experiments described in were designed based on this concept and thus animals were treated daily with lower doses of TMZ.
Results obtained from the combination therapy experiments conducted in this study are supported by various published studies using TMZ and RT treatment. Treatment of tumors with RT alone did not induce apoptosis as detected by our reporter ( and ); however, these tumors had decreased growth compared to tumors from control mice (). This data supports studies demonstrating that tumor response to RT results in only modest levels of apoptosis in most tumor cell lines (38
). Instead, tumors treated with DNA damaging therapy such as irradiation undergo cell death via necrosis or “mitotic catastrophe”(40
). Administration of the DNA alkylating agent TMZ in conjunction with radiation increases dsDNA damage to levels that activate apoptotic machinery and forms the basis of radio-sensitizing activity of TMZ (24
). Thus, tumors of animals receiving both TMZ + RT had dsDNA damage to an extent that enabled activation of the apoptotic cascade, resulting in increased bioluminescence activity compared to animals treated with individual therapies. Additionally, the results of our TMZ and RT combination therapy experiments are consistent with data from a preclinical setting where TMZ + RT treatment was found to be synergistic in tumors with low O6
-methylguanine-DNA methyl-transferase (MGMT) levels. D54 cells express low levels of MGMT, which may provide an additional mechanism for the radio-sensitization effect of TMZ (24
Although bioluminescence imaging of apoptosis using the technology described here is limited to testing of potential chemotherapeutic drugs using animals in a research setting, we believe that information garnered using this type of technology with regards to schedule and efficacy of combination therapies has potential to impact the clinical setting. In fact, TMZ + RT combination treatment of brain tumors is showing significant promise in patients.
Although not investigated here, we believe that the ANLucBCLuc reporter would be of significant benefit in studies of other pathologies wherein dysregulation of apoptosis plays an important role and have begun development of transgenic animals that initiate expression of the reporter in a tissue specific manner. Additionally, we are developing reporters for other cytosolic proteases that play a role in various disease processes. Further, replacing the reporter platform with other reporter molecules (e.g. B-galactosidase) that, unlike luciferase, do not require ATP as a cofactor and would remain active in non-viable cells could be developed. Recently, firefly luciferase fragments with novel split sites enabling more effective silencing of luciferase have been reported (42
). We are constructing caspase reporters using these improved split sites to further reduce the background signal of our reporter system. In summary, we have shown this reporter system can be utilized to sensitively, dynamically, and quantitatively image apoptosis. This will enable development or enhancement of therapeutic protocols by providing a platform to investigate dosing, scheduling, or the efficacy of combination therapies in living subjects.