Visualizing Cell Death In Vivo
By: Carol A. Rouzer, VICB Communications
Published: February 25, 2014
A new probe used with positron emission tomography detects cancer cells that have responded to chemotherapy by undergoing programmed cell death.
Apoptosis or programmed cell death (Figure 1) is a process that cells may undergo in response to senescence or damage. Unneeded cells also undergo apoptosis during the process of tissue remodeling during development. Thus, apoptosis is critical to the processes of embryogenesis, tissue homeostasis, and inflammation/immunity. The ability to avoid undergoing apoptosis is one mechanism by which tumor cells, which are often highly damaged, continue to grow and divide. Indeed, many effective antitumor agents overcome this resistance and induce cancer cells to undergo apoptosis. In these cases, the ability to detect apoptotic cells in vivo could provide a means of assessing the response to therapy. This led VICB member Charles Manning and his laboratory to search for imaging agents directed towards the caspase enzymes, which are major players during the apoptotic response. Now, they report the discovery of a peptide-based positron emission tomography (PET) imaging probe that distinguishes cancer cells that have entered apoptosis in response to chemotherapy in vivo [M. R. Hight, et al. (2014) Clin. Cancer Res., published online February 26, doi:10.1158/1078-0432].
Figure 1. Changes in cells undergoing apoptosis. (a) The two HeLa cells on the left are healthy, while those on the right have undergone apoptosis due to exposure to the anticancer drug daunorubicin. Note the cell rounding and membrane blebbing. (b) These HeLa cells were treated with actinomycin D. Some remain healthy, while others have undergone apoptosis. Hoechst staining (blue) shows that the nuclei of the apoptotic cells are condensed, and the DNA is fragmented. (c) The HeLa cells on the left are healthy. Hoechst staining (blue) reveals normal nuclei, and a fluorescent stain (green) reveals abundant thread-like mitochondria. The cells on the right have undergone apoptosis due to treatment with actinomycin D. Note how the mitochondria have fragmented in these cells, even before the nuclei have become condensed. Image reproduced by permission from Macmillan Publishers Ltd, from R. C. Taylor, et al. (2008) Nat. Rev. Mol. Cell Biol., 9, 231, copyright 2008.
The researchers began their work with the inhibitor Val-Ala-Asp(OMe)-fluoromethylketone (VAD-FMK, Figure 2), which inhibits many caspase enzymes. The availability of two co-crystal structures of caspase 3, each with a different bound inhibitor, enabled the researchers to construct a molecular model to predict the binding pose of VAD-FMK in the enzyme’s active site. They then used the model to explore the effects of modifications of the peptide’s structure on the binding interaction.
Figure 2. Structures of the parent peptide VAD-FMK, and the modified peptides evaluated in this study.
Key to development of a radiotracer imaging agent is the introduction of a radionuclide, such as 18F. However, due to the short half-life of PET radioisotopes, initial chemical development and in vitro validation of prospective compounds is performed using non-radioactive, but otherwise chemically identical, analogues. Concentrating on the peptide’s N-terminus, the Manning group evaluated binding energies of Z-VAD-FMK [benzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethylketone], IZ-VAD-FMK [4-iodobenzyloxycarbonyl-Val-Ala-Asp(OMe)-fluoromethylketone], and FB-VAD-FMK [4-fluorobenzylcarbonyl—Val-Ala-Asp(OMe)-fluoromethylketone] (Figure 2). The predicted binding energy scores of all four peptides were in the nanomolar to micromolar range, though FB-VAD-FMK was notable for its better ability to form hydrogen bonds with residues in the caspase 3 active site.
Focusing on the FB-VAD-FMK peptide, the investigators performed direct inhibition studies using caspases-3, -6, -7, and -8. The peptide exhibited a potency similar to that of the parent peptide, VAD-FMK, with highest activity against caspase-8. Its potency was also similar to that of VAD-FMK in intact DiFi colorectal cancer cells treated with cetuximab to induce caspase-3 and -7 activity. Further studies using recombinant caspase-3 indicated that FB-VAD-FMK was the most potent inhibitor of the three modified peptides. Furthermore, among the three peptides, FB-VAD-FMK was the least lipophilic, an important finding because prior studies had suggested that IZ-VAD-FMK was too lipophilic for in vivo applications.
Together, these results supported the hypothesis that radiolabeled FB-VAD-FMK could be used as a PET imaging agent to detect apoptotic cells in vivo. To test this hypothesis, the investigators synthesized radiolabeled [18F]FB-VAD-FMK and used PET imaging to examine its biodistribution in mice. They detected the compound in many tissues, though accumulation in brain, lung, and bone was minimal. After 60 minutes, most of the radioactivity was present in the kidneys and liver, consistent with renal and hepatobiliary excretion. Most of the radiolabel was present as the parent compound, indicating minimal metabolism.
The investigators treated mice bearing both SW620 and DLD-1 human colorectal cancer xenografts with an aurora kinase inhibitor (AZD-1152) known to induce apoptosis in responsive cancer cells. They then used [18F]FB-VAD-FMK and PET imaging to search for apoptosis in the tumors. The results indicated uptake of the radiolabel only in the SW620 xenografts from mice that had been treated with AZD-1152. Immunohistochemistry of the excised tumors confirmed the presence of active caspase enzymes only in these tumors, indicating that the SW620 but not the DLD-1 xenografts had responded to the treatment.
Further studies focused on the COLO-205 and LIM-2405 colorectal cancer cell lines. Both of these lines carry a V600E mutation in the gene for BRAF, a kinase involved in the Ras signaling pathway. They established xenografts of both tumor cell lines in mice, and then treated them with a BRAF inhibitor (PLX-4720) or an inhibitor of PI3K/mTOR (BEZ-235), a downstream effector of RAS that does not depend on BRAF. Alternatively, they treated the mice with both inhibitors or no inhibitors. Following treatment PET imaging with [18F]FB-VAD-FMK revealed uptake of the radiotracer only in COLO-205 xenografts from mice that had been treated with both drugs. Consistently, only these tumors exhibited a significant reduction in volume, and histochemistry confirmed caspase induction only in these tumors (Figure 3).
Figure 3. (Left) PET images using the [18F]FB-VAD-FMK radiotracer of a COLO-205 xenograft in mice treated with drug vehicle or with the combination of BEZ-235 and PLX-4720. The arrows indicate the location of the tumor. (Center) Immunohistochemistry of the imaged tumors showing the presence of activated caspase (brown) in the BEZ-235/PLX-4720-treated tumors. (Right) Percent change in tumor volume in COLO-205 xenografts for mice treated with vehicle, BEZ-235 alone, PLX-4720 alone, or the combination of the two drugs. The asterisk indicates a significant decrease in tumor volume. Figure kindly provided by Matthew Hight and Charles Manning. Copyright 2014.
Together the results confirm that [18F]FB-VAD-FMK can detect apoptotic cells in vivo. More importantly, the probe was able to distinguish between colorectal cancer xenografts that had responded to therapy and those that did not. These highly promising results indicate that further exploration of the clinical utility of [18F]FB-VAD-FMK is certainly warranted.