Targeting RAL-GEF signaling and cytokines in KRAS driven lung cancer
October 9, 2015, by David Barbie
Given the challenges associated with targeting KRAS directly, much recent attention has been directed at inhibiting one or more of its multiple downstream signaling pathways. Indeed, small molecules that target kinases along the PI3K-AKT and RAF-MEK-ERK axes have been tested in clinical trials in various combinations. Unfortunately, despite being well-validated targets, the therapeutic window of combining these agents is limited, and to date most trials have shown lack of efficacy.
My laboratory has been focused on targeting the non-canonical IκB kinase TBK1, which resides downstream of the RAL-GEF target RALB, and has been shown across multiple in vitro studies to promote KRAS-dependent cell survival. While TBK1 signaling is complex, with roles proposed in NF-κB activation, AKT signaling, autophagy, and more recently mitophagy, we have focused in particular on the secreted cytokines induced by this kinase, which are involved in innate immunity. In KRAS-dependent cancers we discovered a key role for TBK1 and its homologue IKKε in regulating CCL5 and IL-6 production, which foster cell transformation, survival, and angiogenesis (Zhu et al., Cancer Discov 4, 452, 2014; PMID: 24444711). Fortuitously, we also discovered that a compound already in the clinic as a JAK1/2 inhibitor, CYT387, is also a potent TBK1/IKKε inhibitor, and disrupts both upstream CCL5/IL-6 production and downstream JAK/STAT pathway engagement. Treatment of murine KrasG12D induced lung cancers with single agent CYT387 was more effective than docetaxel chemotherapy and at least as effective as PI3K/MEK combinations based on historical data. Single agent treatment of murine KrasG12D; p53 mutant (KP) lung cancer was relatively ineffective, but given the aggressive nature of this mouse model, we suspected that combination therapy would be necessary to achieve response.
Consistent with work from other groups, we observed that TBK1 inhibition not only resulted in paradoxical induction of phosphorylated-TBK (pTBK1), but also increased pERK levels (observations that still remain unexplained mechanistically). By combining CYT387 treatment with MEK inhibition, we could prevent feedback pERK activation, which led to further suppression of IL-6 levels and cooperative impairment of cell viability in vitro. We therefore tested combination CYT387 with AZD6244 (MEK inhibitor selumetinib) treatment in KP mice, and remarkably observed regression of established tumors with durable responses for up to 8 weeks.
Helping to accelerate translation of these findings into the clinic, the rights to CYT387 were purchased by Gilead Sciences in 2013 (now renamed GS-0387 or momelotinib), whereupon it was moved into phase 3 studies versus the JAK inhibitor ruxolitinib for treatment of myelofibrosis. Around the same time, the MEK inhibitor trametinib received FDA approval for treatment of BRAF mutant melanoma. Together, these developments fostered the initiation of a phase 1B study of combination momelotinib and trametinib therapy in advanced/refractory KRAS mutant lung cancer, which is in dose escalation.
Currently, our laboratory is interested in identifying markers of response or resistance to this drug combination. For example we are performing preclinical studies in KrasG12D; Lkb1 mutant (KL) mouse models and correlating this with tumor suppressor genotypes in the clinical trial. We are also conducting in vitro studies of acquired resistance, to identify pathways and drug targets that may be incorporated for triple combination therapy. Finally, since it may be challenging to identify discrete genomic markers that correlate with response to therapies that target immune signaling pathways, we have adapted a microfluidic 3-dimensional co-culture system to grow primary tumor spheroids and measure dynamic cytokine production. Since current attempts to quantify tumor-associated cytokines rely on static determination of mRNA levels in frozen tumor samples or remote measurement in plasma, we hope that this technology will enable a more direct functional predictor of response.
About the Author
David Barbie obtained his M.D. from Harvard Medical School in 2002, and was an HHMI Medical Fellow in Ed Harlow’s lab at Massachusetts General Hospital. Following his medical oncology training and a post-doctoral fellowship in William Hahn’s lab, he started his own laboratory and is currently an Assistant Professor and Thoracic Oncologist at the Dana-Farber Cancer Institute.