Triple-negative breast cancer (TNBC) is a highly aggressive breast cancer subtype characterized by lack of estrogen, progesterone and normal level of the HER2 receptor. Current treatment options for TNBC are restricted to chemotherapy, thus novel therapeutic strategies is a pressing clinical need. Receptor tyrosine kinases (RTKs) including EGFR and c-Met are often highly expressed in TNBC and their co-expression is associated with poor patient prognosis/survival. However, targeting these RTKs with single agents invariably lead to drug resistance contributed partly by network-encoded mechanisms including feedback interference and/or activation of compensatory pathways. We recently found that the non-receptor tyrosine kinase PYK2 is a common downstream effector of EGFR and c-Met in TNBC cell lines. Interestingly, we established multiple interplayed positive feedback loops between EGFR, PYK2, STAT3 and c-Met, which suggest synergism in co-targeting these network nodes. However, which target combinations may yield synergistic efficacy and why remain unclear.
Using an integrated systems approach combining predictive modelling and wet-lab experimentation, we developed a quantitative model of the EGFR-PYK2-c-Met interaction network in TNBC. Our model predicts STAT3 and ERK activation respond in ultrasensitive, switch-like manners to EGFR and/or PYK2 gradient inhibition; and co-inhibition of EGFR/PYK2 yield synergistic effects. Additionally, model simulations allowed us to rank different drug combinations targeting EGFR, c-Met, PYK2 and STAT3. These predictions were subsequently validated in TNBC cell lines. Next, using public patient data we generated patient-specific models and explored their prognostic value by examining whether they could identify TNBC patients subsets that would benefit from EGFR/PYK2 (and others) combination therapy. Simulations showed that patients with highest PYK2 expression belong to the basal-like (BL) subtype, and that a subgroup of BL patients displaying concomitant moderate/high PYK2 and moderate/high EGFR expression exhibit exclusively strong synergistic effect, consistent with our experimental data.
Our systems modelling provides a powerful approach which enables rational designs of drug combinations and predictive prioritization of combinatorial therapies in an individualized manner. This work is therefore a concrete step towards personalized treatment for triple-negative breast cancer.