The Daily Sign-Out

Systems biology for studying cancer

This month's issue of Future Oncology has a well-written editorial from Dr. John McDonald from Georgia Tech regarding systems biology in cancer research.  This is a free download and I highly recommend it as a nice overview contrasting the traditional "reductionist" approach to cancer cell biology research versus the emerging application of a systems biology approach.

First, it is appreciated that the traditional approach has grudgingly yielded some valuable insight into cancer cell biology which has led to the development of "rationally designed" targeted therapies.  But while many of these therapies have had some initial impact for patients enrolled in clinical trials (and even some dramatic responses), I think it is fair to say that they also have not radically shifted overall survival or long-term outlook of the most common lethal cancers.

However, we are beginning to realize the potential for combining high-throughput molecular technologies with computational strategies and capabilities for combining multidimensional datasets to find inferred network relationships.  Dr. McDonald refers to this as the "top-down" approach.  He contrasts this with the traditional reductionist "bottom-up" approach which links human oncogenes or tumor suppressor genes with known interacting genes previously characterized in experimental models using cancer cell lines or animal models to identify signaling pathways implicated with cancer.  He stresses that these approaches are by no means mutally exclusive but indeed are complementary.

The systems approach starts with establishing a list of component parts.  Using new technologies, we can now speak of the genome, transcriptome, proteome, metabolome, etc.  The revolutionary step is using sophisticated computer models to detemine how these various components interact and which of these interactions are relevant for cancer initiation and progression and which represent potential druggable targets.

Again, this is a nice appetizer for additional reading on this exciting area!

Too many SOX can be a bad thing

Nature Genetics published a fascinating article online on October 4 by Bass et al. identifying SOX2 amplification in lung and esophageal squamous cell carcinomas (SqCC).  It is refreshing to read a good study of squamous cell carcinomas since there has been so much published recently on lung adenocarcinomas.

This is a multi-institutional study that first examined 40 esophageal SqCC and 47 lung SqCC using SNP arrays to determine copy number changes and GISTIC (Genomic Identification of Significant Targets in Cancer) to score the significance of recurrent chromosomal gains and losses in order to identify peak regions likely to harbor tumor-driving genes.

The most significant amplified peak in lung SqCC was at chromosome 3q26.33, found in 11/47 lung SqCC and 6/40 esophageal SqCC.  In lung SqCC, the peak in this segment contained four genes, including SOX2; in esophageal SqCC, the peak only included SOX2.  Next, the authors measured SOX2 mRNA levels by RT-PCR in 27 lung SqCC with matched SNP data cases.  Cases with SOX2 amplification were found to have higher mRNA expression but they also found several cases without 3q26.33 amplification that had high SOX2 mRNA expression.  The authors made similar findings with esophageal SqCC.  The authors then performed an arrayed RNA interference screen targeting SOX2, neighboring genes, additional candidate genes and controls in 4 SqCC cell lines (2 lung, 2 esophageal) for differential effects on proliferation.  Suppression of SOX2 had the largest differential anti-proliferative effects on 3q26.33 amplified cell lines amongst all genes tested.  Further, they showed that SOX2 knockdown reduced anchorage-independent growth.  These findings were confirmed with rescue of the effects of SOX2 suppression by restoring wild-type SOX2 in a mutant cell line.

Additional work shows that SOX2 amplification and mRNA levels are significantly higher in lung SqCC compared with lung adenoca (which differentially showed NKX2-1 amplification and increased mRNA levels cf. to lung SqCC).  SOX2 alone could not transform immortalized tracheobronchial cells but could do so in cooperation with the epithelial isoform of FGFR2.

As it turns out SOX2 is an very interesting transcription factor intimately involved in forgut development and is required for proper differentiation of esophageal squamous mucosa and in multiple respiratory cell types.

This is a very dense paper presenting the results of multiple experiments of which I have summarized only a part.  What does all this mean for the practicing pathologist or oncologist?  The identification of SOX2-dependent pathways in lung squamous cell carcinomas is an exciting finding that may represent a novel therapeutic target--which has been conspicuously lacking so far in lung SqCC.

Novel classification for gastric cancer using pathways

A multi-institutional study published online on Oct. 1 in PLoS Genetics examined 301 gastric carcinomas using gene expression profiles and a pathway prediction model to identify major oncogenic pathways involved in gastric cancer and relate expression of these pathways to patient survival.  They found proliferation/stem cell-related, Wnt/beta-catenin, and NF-kappaB pathways are deregulated in over 70% of the gastric cancers they studied.  Furthermore, the found patterns of pathway co-activation associated with survival.  This is a free access paper so please download and read.

Reading the "Materials and Methods" of this paper is worthwhile in itself because the strategy the authors used is well-conceived and explained.  More importantly, this study demonstrates the feasibility of using gene expression profiles to perform in silico experiments to make pathway predictions as a basis for in vitro experiments--another example of "systems thinking" in pathology and oncology.