Transcriptional profiling of apoptosis

Abstract

In a recent issue of Cell Cycle, Galluzi et al.1 describe comparative gene expression profiles of the cellular response to three different stressors, with the striking conclusion that cisplatin (CDDP)-induced apoptosis does not rely on direct effects either on mitochondrial integrity or nuclear transcriptional reprogramming due to DNA damage. While their approach gives important insight into the wide range of possibilities a cell can use to cope with stress at the transcriptional level, the paper also highlights the challenges to define cellular death pathways at the systems level. Classifying cell death modalities dates back to the times when the most powerful approach was thorough morphological observation by light and electron microscopy in the ’70s, allowing us to limit the catalog of cell death to merely three types.2 In the following decades, with the introduction of biochemical approaches and of the series of novel stressor substances tested, the picture became always fuzzier and more complex. The latest effort to categorize the regulated death pathways listed 13 modalities, defined by the combination of at least 30 unique biochemical processes and their sensitivity to numerous pharmacological and genetic modifiers.2 Recognizing the complexity of the systems involved, modeling based on the fairly well-characterized principal protein components of cell death pathways (considering apoptotic, regulated necrotic and autophagy pathways) and posttranslational biochemical modifications gained importance.3 Finally, further realizing that the cell death pathways depicted by classical biochemical studies might just scratch the surface, in recent years a series of studies applied large-scale, unbiased gene expression and proteomic approaches to identify novel players in cell fate determination. These studies were particularly boosted by the need to identify novel targets in the pharmacological treatment of drug-resistant cancer.4 While many cellular stress pathways impinge on gene expression (e.g., unfolded protein response and endoplasmic reticulum stress), modulation of transcription is envisaged to be the most relevant in defining the cellular response to genotoxic stress, which directly targets the integrity of DNA. Indeed, a recent meta-analysis of a large set of gene expression profiles underlying the cellular response to ionizing radiation-induced double-strand breaks confirmed the central role of the transcriptional targets of p53, mediating DNA repair and survival, senescence or programmed cell death.5 In contrast, the study by Galluzzi et al.,1 using CDDP, which rather causes intra- and interstrand links as the primary mechanism of DNA damage,6 showed that the enriched transcriptionally modified pathways belong to classes not directly associated with cell death induction, particularly when compared with C2-ceramide and CdCl2, classic inducers of mitochondrial apoptosis. The finding was corroborated by the limited modulation of CDDP-induced reduction of clonogenicity in genetically modified yeast clones, lacking several components of the apoptotic pathway. The study accompanies a previous effort by the same group,7 where, using genome-wide shRNA profiling, they identified a set of genes that are able to inhibit or enhance the toxicity caused by CDDP (CDDP response modifiers, CRMs). While it would have been expected that the primary target of the CDDP-induced transcriptional stress response will include this gene set, strikingly, only about 10% of the CRM genes were found significantly up or downregulated at the transcriptional level in the present study. This led to the cautious conclusion of the authors that it is likely that regulation at the translational and posttranslational levels is the essential factor in triggering the actual cell death execution machinery following CDDP treatment. Accordingly, the transcriptional regulation accompanying CDDP toxicity is either negligible or responsible for only secondary adaptive stress responses. Indeed, recent studies suggested that DNA damage can cause massive changes in global translational profiles,8 and the direct interaction of CDDP with a wide range of other cellular components could give rise to extensive posttranslational modification.9 Notably, the work illustrates the power of combining bioinformatic and experimental approaches to identify new transcriptional targets in the DNA damage response network, but also indicates the formidable challenges when aiming to define and classify cell death pathways based on unbiased genome-wide systems analyses. It will certainly take a lot of effort to get there, but now it also appears achievable if the combined approaches of high-content imaging, transcriptomic and proteomic techniques will be adopted by a larger community of cell stress-focused research groups.