Identification of plasma metabolite changes following phosphatidylinositol-3-kinase (PI3K) inhibition with GDC-0941, a potent and selective pan-Class I inhibitor of PI3K: preclinical discovery followed by clinical qualification in a Phase I clinical trial.


Year:

Session type:

Joo Ern Ang1,2, Rupinder Pandher1, Yasmin Asad1, Alan Henley1, Melanie Valenti1, Gary Box1, Alexis de haven Brandon1, Richard Baird1, Lori Friedman3, Mika Derynck3, Suzanne Eccles1, Stan Kaye1,2, Paul Workman1, Johann de Bono1,2, Florence Raynaud1,2
1Cancer Research UK Cancer Therapeutics Unit, The Institute of Cancer Research, Sutton, Surrey, UK, 2Drug Development Unit, The Royal Marsden NHS Foundation Trust, Sutton, Surrey, UK, 3Genentech Inc, South San Francisco, California, USA

Background

Phosphatidylinositol-3-kinase (PI3K) plays a key role in cellular metabolism. A mass spectrometry-based metabolomic platform was used to identify novel pharmacodynamic plasma metabolite biomarkers of systemic modulation of PI3K signalling.

Method

The preclinical screen included comparisons of tumor-bearing PTEN knockout (+/-) mice with their respective non-tumor-bearing wildtype littermates and athymic mice with or without PTEN-/- human tumor xenografts. The effect of GDC-0941-treatment versus vehicle was evaluated in xenograft-bearing mice and candidate biomarkers then measured in plasma samples of patients with advanced solid tumours treated in a Phase I clinical trial of GDC-0941.

Results

Thirty plasma metabolites including branched chain and aromatic amino acids, acylcarnitines and phospholipids were down-regulated in the PTEN null models and up-regulated following pharmacologic inhibition using GDC-0941. In patients treated with GDC-0941 (n=41, median age 54 years [range 37-72], 18 males and 23 females, median body mass index 26.4 km/m2 (SD 4.2]), time- and dose-dependent changes in 26 candidate metabolites were observed post-dose on day 1; the magnitudes of treatment-associated changes were greater than baseline variability. All these changes were consistent with the tumor-bearing mouse models apart from the long chain acylcarnitines where a significant decrease was observed in-keeping with changes in the non-tumor-bearing mice. All observed changes resolved after a 1-week drug washout and were recapitulated on day 15 after a week of continuous oral dosing.

Conclusion

This study provides the first link between systemic modulation of the PI3K pathway and changes in plasma metabolites which may have utility as minimally invasive pharmacodynamic biomarkers. These findings provide additional support for the association of insulin resistance with branched-chain amino acids and related metabolites following PI3K inhibition.