2020 ASCO Annual Meeting

Test measuring signaling activity in live patient tumor cells identifies PI3KCA WT patients who may benefit from PIK3 inhibitors
Lance G Laing, Salmaan Khan, Ian A MacNeil, Catherine Kuzmicki, Benjamin Rich, Abhay Shukla, Kelly Brass, Brian Sullivan; Celcuity Inc. Minneapolis, MN

Background: Biological factors other than PIK3CA status, such as aberrant GPCR-linked signaling, may be important to measure when identifying patients eligible for PI3K inhibitors.  A new assay, the CELx PI3K test, using an impedance biosensor was developed to measure ex vivo live tumor cell response to specific S1P agonists and PI3K antagonists to diagnose breast tumors with PI3K-involved hyperactive signaling. This study set out to: 1) compare CELx PI3K test results and xenograft results using cell lines with PIK3CA mutations; and 2) assess whether PI3K-involved hyperactive S1P signaling is found in PIK3CA WT breast cancer patient tumors.

Methods: A panel of 17 fresh HER2-/PIK3CA WT tumor cells from breast cancer patients and three PIK3CA mutated breast tumor cell lines were obtained. Live cell response to an S1P agonist, PI3K-α antagonist (alpelisib), PI3K-γ antagonist (IPI-549), and a pan-PI3K inhibitor (taselisib) were measured using an xCELLigence RTCA impedance biosensor. From these responses, PI3K-involved signaling was quantified and characterized as normal or abnormal using a previously determined cutpoint.  For the xenograft study, 16 NSG mice were injected with HCC1954 PIK3CA mutated breast cancer cells and randomly assigned to either the control or taselisib group (10 mg/kg).

Results: Four of the 17 PIK3CA WT tumor cells had abnormal levels of combined PI3K-α and PI3K-γ signaling.  Only one of the three PIK3CA mutated breast tumor cell lines (BT20) had abnormal levels of PI3K-α and pan-PI3K involved signaling.  The HCC1954 cell line had normal PI3K-α and abnormal pan-PI3K signaling. CAL-51 reported normal PI3K-α and pan-PI3K signaling.  The normal levels of PI3K-α signaling found in the HCC1954 and CAL-51 cell lines correlated with previously reported xenograft studies that found alpelisib had no anti-tumor effect.  The xenograft study reported here using HCC1954 cells found taselisib induces a significant anti-tumor effect (T/C ratio = 0.21; p = 0.009; t-test).

Conclusions:  A sub-set of PIK3CA WT patient breast cancer tumors had abnormal PI3K-involved signaling comparable to levels found in PI3KCA mutated cell lines.  Abnormal pan-PI3K signaling and normal PI3K-α signaling in the HCC1954 cell line correlated with xenograft results.  This study thus suggests that measurement of PI3K-involvement in hyperactive S1P signaling in live patient breast cancer cells may provide a means to identify breast cancer patients who may or may not benefit from treatment with PI3K inhibitors.