An inducible caspase-3 ‘tumour death switch’ model to identify and quantify circulating biomarkers of apoptotic cell death using iTRAQ isobaric tags and mass spectrometry


Session type:

Kathryn Simpson1,Cassandra Hodgkinson1,Michael Walker1,Martin Dawson1,Lee Lancashire1,Cong Zhou1,Marion MacFarlane2,Anthony Whetton3,Caroline Dive1
1Paterson Institute for Cancer Research, Manchester Academic Health Science Centre, Christie Hospital, University of Manchester, Manchester, United Kingdom,2MRC Toxicology Unit, University of Leicester, Leicester, United Kingdom,3School of Cancer and Enabling Sciences, Manchester Academic Health Science Centre, Wolfson Molecular Imaging Centre, University of Manchester, Manchester, United Kingdom


Many emerging mechanism-based cancer therapeutics promote tumour cell death, yet there is currently a lack of reliable biomarkers to monitor this. We have developed a caspase-3-based tumour ‘death switch’ model in which apoptosis is controlled and synchronously induced. Using this model in our iTRAQ-based clinical proteomics workflow we aim to identify novel biomarkers of tumour apoptosis with potential clinical application.


HT29 colorectal cancer cells were stably transfected with a doxycycline inducible, constitutively active recombinant caspase-3 (revC3). Xenografts were grown in SCID/bg mice and plasma samples were taken at time-points following doxycycline administration.  Immunohistochemistry was carried out on tumours for cleaved caspase-3, cleaved PARP and caspase-cleaved cytokeratin-18 (CK18-neo). Plasma CK18 was assayed using M30/M65 ELISAs (Peviva). Proteomics was carried out on media from ‘death switched’ cell lines in vitro and pooled plasma from ‘death switched’ xenograft-bearing mice using our iTRAQ-based MALDI-TOF/TOF workflow.


Tumour regression via apoptosis was confirmed in xenografts by immunohistochemistry of cleaved PARP, cleaved caspase-3 and CK18-neo. CK18-neo in tumours correlated with its appearance in the plasma. In vitro proteomic analysis identified 235 proteins released into the media following death switch activation. Of these, 164 were significantly up-or down regulated and 72 were altered at early time-points following death switch activation, indicating specific secreted markers of apoptosis. In the more complex system of mouse plasma, 110 human proteins were identified, with 83 candidates for tumour markers (FDR of ?5%, ?2 peptides/protein, ?95% confidence).


The death switch model causes rapid and synchronous apoptosis when grown as xenografts and has been used to validate plasma CK18-neo as a surrogate biomarker of apoptosis. This model was successfully used in our newly developed proteomics workflow for clinical sample analysis. We have been able to identify several candidate proteins that may prove to be plasma biomarkers of tumour apoptosis.