AKT is directly phosphorylated at AKTTHR308 by PDK1 and this has been widely used as a measure of PI3K activity in in vitro tissue culture and in A66 vivo tumour xenograft experiments. However, the stability of phosphorylation of the AKTTHR308 site is generally thought not to be sufficiently robust for use in clinical studies. Other direct protein substrate targets of PDK1, such as SGKs, are yet to be explored as biomarkers of PI3K pathway inhibition.  AKTSER473 and some downstream proteins as preclinical and clinical biomarkers of PI3K activity. The downstream markers of activity include PRAS40THR246, a substrate of AKT, and RPS6SER240/244 and 4EBP1THR37/46. However, none of these biomarkers are perfect as they are not entirely specific for PI3K activation/inhibition because their phosphorylation can be influenced directly or indirectly by mTOR kinase activity.
In addition, they may be influenced by inputs from other pathways, for example RPS6 can be phosphorylated AZ 3146 by p90S6K, a protein kinase regulated by other signalling cascades including MEK/ERK. Nevertheless they can play a useful research role. A number of studies have used unbiased screening strategies with the aim of identifying better and more specific biomarkers of PI3K inhibition for use in the development of PI3K inhibitors. Andersen and colleagues have employed immunoaffinity precipitation followed by mass spectrometry of protein extracts from cells that were treated with inhibitors of PDK1, AKT or PI3K/mTOR.
The aim of this study was to find specific biomarkers of PI3K pathway inhibition, it successfully led to the identification and quantification of 375 nonredundant phosphopeptides that were relevant to PI3K pathway signalling, and which contained AKT and PDK1 recognition motifs. Of these, seventy one phosphopeptides were drug modulated and 11 were reduced by all three inhibitors examined. An example was phosphorylation of the ribosomal protein RPS6 that was the most strongly inhibited by all 3 inhibitors and phosphorylation of PRAS40THR246 which was the most affected following AKT and PI3K/mTOR inhibition. PRASTHR246 was validated in lung and breast cancer cell lines and predicted sensitivity to an AKT inhibitor. Importantly, the phospho PRASTHR246 epitope was more stable than the phospho AKTSER473 epitope commonly used for identifying tumours with AKT pathway activation, suggesting that this biomarker might be more suitable for clinical evaluation of PI3K pathway inhibition.
In particular it may be ideal for use in immunohistochemistry, which is often applied in clinical studies. The value of using ELISA based methodology to measure quantitatively the phosphorylation of pathway proteins that are both proximal and distal to PI3K has been demonstrated with several inhibitors including GDC 0941, with potency declining at more distal points. Interestingly, although inhibition of substrate phosphorylation was valuable as a measure of PI3K target inhibition, the degree of inhibition measured by immune assay did not predict sensitivity in this study. An alternative non biased approach for pharmacodynamic biomarker discovery is to use microarray expression profiling to identify gene signatures specifically associated with PI3K inhibition.