One of the hallmarks of cancer is that even within different tumour types, there is an enormous degree of heterogeneity. Ultimately, in simple terms this means that individual patients will respond to different therapies depending upon their underlying biology.   The challenge, therefore, is defining and categorising the subtypes and working out which are the passenger and driver oncogenes, since the latter will cause aberrant tumour growth and survival, while the former may result as a consequence of changing pathway activity.

This morning I was researching gliomas and came across this old paper (March 2006) that looks at molecular subtypes of gliomas i.e. glioblastomas and astrocytomas.  The article concluded:

“Recent evidence suggests that gliomas may arise from a cell type with neural stem cell-like properties. The current work demonstrates that prognostic subtypes of glioma resemble key stages in neurogenesis and implicates signaling pathways that play critical roles in regulation of forebrain neurogenesis in control of tumor aggressiveness. Longitudinal analysis of glioma cases reveals a frequent pattern of disease progression into the mesenchymal phenotype, a state associated with robust angiogenesis.

This work suggests that molecular classification of glioblastoma may predict response to targeted therapies and suggests that greater understanding of neurogenesis in the adult forebrain may yield novel therapeutic insights for glial malignancies.”

The reason I was curious about this particular paper was because following the AACR Special Conference on PI3K and mTOR that I attended last week, it made sense to look at the literature on mTOR, PI3K and AKT in more detail.

In the glioma research, it was interesting to see what predicted poor prognosis:

“A robust two-gene prognostic model utilizing PTEN and DLL3 expression suggests that Akt and Notch signaling are hallmarks of poor prognosis versus better prognosis gliomas, respectively.”

Now, while Akt and Notch signalling may be important, it doesn’t mean that they make idea targets for drug therapy.  PTEN loss of function is also a difficult target at present and it isn’t clear if it is a driver per se.  What was very clear at AACR last week was that for every action there is an equal and opposite reaction, meaning that targeting one part of a pathway may lead to switching of aberrant activity to another part of the pathway as it adapts to the changing environment.

Neal Rosen from MSKCC gave perhaps one of the best talks of the AACR meeting. He succinctly and simply put out a few constructs based on what we know so far. I will summarise some of the talks in a conference report (sign up on the top right column), but what was relevant to the paper on gliomas is that while at first sight it might make sense to target Akt, that strategy will have consequences.

According to Rosen, in general, inhibiting PI3K also stimulates HER3 expression and phosphorylation, as well as other receptor tyrosine kinases in many cell lines.  In other words, we may need a multi-targeting approach based on the original aberrant driver, the adaptive pathway and the ligand driving activity.

Double and triple combinations make sense from a scientific perspective, but they will also incur far higher costs and more complex clinical trial designs. Who knows whether other adaptive mechanisms will also evolve as a result of pursuing that strategy?  It brings vividly to mind Frank McCormick’s wac-a-mole approach that he described last year at AACR on the challenges of targeting the PI3K pathway in general, irrespective of upstream or downstream targets.

Progress is slowly being made, but we have a long way to go yet with the PI3K-mTOR pathway, although I’m hopeful of some positive progress soon. Certainly there will be some new data emerging on the biology at AACR in April and clinical data at ASCO in June.


ResearchBlogging.orgPhillips, H., Kharbanda, S., Chen, R., Forrest, W., Soriano, R., Wu, T., Misra, A., Nigro, J., Colman, H., & Soroceanu, L. (2006). Molecular subclasses of high-grade glioma predict prognosis, delineate a pattern of disease progression, and resemble stages in neurogenesis Cancer Cell, 9 (3), 157-173 DOI: 10.1016/j.ccr.2006.02.019