Adaptive resistance to chemotherapy, a multi-FAK-torial linkage

Abstract

Oncogenes provide tumor cells with a growth and survival advantage. Directed therapies targeted to oncogenic mutations (such as BRAF V600E) are part of effective late-stage melanoma treatment. However, tumors with BRAF V600E mutations, in ~10% of colorectal cancer (CRC), are generally treatment-insensitive. Research has identified various “feedback” mechanisms that result in BRAF signal pathway reactivation in response to BRAF inhibition. Herein, we highlight key findings from Chen et. al. (this issue) showing that integrin-associated focal adhesion kinase (FAK) activation selectively occurs in BRAF V600E mutant CRC cells in response to pharmacological BRAF inhibition. FAK activation results in elevated β-catenin protein levels, β-catenin nuclear localization, and increased gene transcription. Small molecule inhibitors of β-catenin or FAK synergize with vemurafenib BRAF inhibitor to prevent BRAF V600E CRC cell proliferation in vitro and xenograft tumor growth in mice. This study complements findings linking FAK to β-catenin in intestinal tumorigenesis, resistance to radiotherapy, and cancer stem cell survival. Thus, FAK activation may occur as a frequent tumor cell “adaptive resistance” mechanism. Although FAK (PTK2) is not mutated in most cancers, targeting FAK activity in combinational approaches may limit tumor cell escape mechanisms and enhance durable responses to treatment.

Colorectal cancer (CRC) progression is a disease driven by the accumulation of genetic aberrations and oncogenic mutations. CRC tumors harboring BRAF V600E mutations are generally treatment-resistant and have an extremely poor prognosis (1). KRAS, a predictive biomarker in the management of CRC, functions upstream of BRAF and mutations in these oncogenes activate the mitogen-activated protein kinase (MAPK/ERK) signaling cascade. Unlike melanoma, various BRAF inhibitors are ineffective as monotherapy in BRAF-mutated CRC (2). This may be due in part to the rapid activation of epidermal growth factor receptor (EGFR) or phosphoinositide 3-kinase (PI3K) via proposed feedback activation loops or breakthrough signaling in response to pharmacological BRAF inhibition in CRC. Unfortunately, combinations of BRAF, EGFR, or PI3K inhibitors have limited efficacy in treating BRAF V600E CRC tumors (3). Thus, an alternative switch or BRAF “bypass” signaling pathway activation may continue to drive tumor growth upon BRAF inhibitor administration.

Focal adhesion kinase (FAK) activation in response to BRAF inhibition: connections to Wnt/β-catenin

In this issue, Chen et al. (4) show that increased FAK tyrosine (Y) Y397 phosphorylation (a marker of FAK activation) occurs within 4 to 8 hours after BRAF inhibitor (vemurafenib, dabrafenib, GDC-0879, PLX7904, or LY3009120) treatment of CRC cells containing BRAF V600E mutations (Colo205, Colo741, HT-29, LS411N, SW1417, and WiDr) but not in CRC cells with wildtype BRAF (DLD-1, RKO, and SW48). Sustained FAK activation after vemurafenib addition (measured up to 48 h) occurred in the presence of EGFR (erlotinib), MEK1 (trametinib) and ERK2 (SCH772984) inhibitors, suggesting that FAK activation was independent of BRAF-MEK-ERK-EGFR pathway reactivation. Notably, FAK Y397 phosphorylation was prevented by co-administration of FAK inhibitor (PF562271) and the combination of vemurafenib plus PF562271 potently blocked BRAF-mutant cell growth under conditions whereby ERK2 remained active. Together, these results suggest that FAK activation is distinct from a “BRAF pathway reactivation” process and that signaling linkages to targets other than ERK2 are functioning to promote BRAF V600E-mutant cell growth (Figure 1).

Figure 1.

Schematic of the BRAF/MEK/ERK pathway and an alternative bypass pathway in BRAF-mutant CRC. A) In the absence of BRAF inhibition, the BRAF/MEK/ERK pathway is active and supports cell growth. FAK is inactive, whereas GSK3β is activated and inhibits β-catenin. B) Upon BRAF inhibition, the BRAF/MEK/ERK pathway is inhibited. FAK is phosphorylated and inhibits GSK3β, allowing for bypass activation of β-catenin and continued cell proliferation.

FAK is a cytoplasmic tyrosine kinase that is best characterized as mediating integrin receptor signals from matrix proteins controlling cell movement and survival (5). FAK expression and Y397 phosphorylation are elevated in several advanced-stage solid cancers (6) and it is believed that FAK signaling is compartmentalized through interactions with a variety of effector proteins that localize FAK either to focal adhesions, endocytic vesicles, cell-cell junctions, or the cell nucleus (5). Although Chen et al. (4) did not investigate where Y397-phosphorylated FAK was localized in vemurafenib-treated BRAF V600E-mutant CRC cells, Jing Hu’s group previously showed that FAK promotes Wnt/β-catenin pathway activation and intestinal tumorigenesis by phosphorylating glycogen synthase kinase-3 beta (GSK3β) at Y216 (7).

GSK3β Y216 phosphorylation promotes the binding of ubiquitin E3 ligase β-TrCP (β-transducin repeats containing protein) and subsequent ubiquitination and degradation of GSK3β. Since GSK3β activity is a negative regulator of β-catenin, FAK phosphorylation of GSK3β at Y216 indirectly results in enhanced β-catenin stabilization by lowering GSK3β levels. Notably, pharmacological inhibition of FAK activity (7) or conditional FAK knockout (8) suppressed intestinal tumorigenesis in mice heterozygous for tumor suppressor adenomatous polyposis coli (APC) – a major regulator of the β-catenin “destruction complex” with GSK3β, axin, and casein kinase 1. Conditional FAK knockout has also revealed a role in β-catenin-driven hepatocarcinogenesis (9) and FAK can enhance β-catenin tyrosine phosphorylation and activation in endothelial and breast carcinoma cells (10,11). Together, these studies point to the importance of FAK in β-catenin regulation within multiple cell types.

In the current study by Chen et al. (4), treatment of HT-29 cells with BRAF inhibitors (dabrafenib, GDC-0879, or vemurafenib) resulted in a time-dependent increase in FAK Y397 phosphorylation, GSK3β Y216 phosphorylation, and β-catenin protein levels. Verurafenib increased nuclear β-catenin accumulation and Wnt target gene transcription in BRAF-mutant cells which was independent of microsatellite instability or APC mutation status. FAK inhibitor treatment or knockdown of FAK mRNA levels prevented vemurafenib-induced β-catenin accumulation. Co-targeting the FAK/β-catenin pathway with either PF562271 or ICG-001 (a Wnt/ β-catenin antagonist) in combination with vemurafenib, resulted in decreased cell proliferation in vitro and the synergistic prevention of HT-29 and BRAF V600E patient-derived xenograft tumor growth in mice. The results of Chen et al. (4) are notable for the identification of a novel FAK to Wnt/β-catenin parallel or “bypass” signaling mechanism “triggered” by BRAF inhibitor treatment of BRAF V600E-mutant CRC (Figure 1).

A multi-FAK-torial linkage to adaptive resistance

One of the open questions remaining is to determine the mechanism of how BRAF inhibitor addition triggers elevated FAK Y397 phosphorylation. Since PF562271 FAK inhibitor prevents vemurafenib-induced FAK Y397 phosphorylation, intrinsic FAK activity is needed. As increased FAK Y397 phosphorylation occurs in HT-29 tumors after vemurafenib treatment of mice, and after 4 to 24 h treatment of HT-29, Colo205, Colo741, LS411N, and WiDr cell treatment with vemurafenib, LY3009120, PLX7904, or GDC-0879 BRAF inhibitors in cell culture, increased FAK Y397 phosphorylation likely involves a specific but indirect tumor-cell intrinsic mechanism. As this effect was not observed in CRC cells without BRAF mutation, the cells are likely primed for a “switch” to an adaptive resistance mechanism. However, the slow time frame makes it difficult to establish cause-effect relationships of potential changes in autocrine-paracrine factor production or the release of FAK inhibitory “restraint” through alterations in protein-protein interactions.

FAK activation is regulated by both conformational changes and specific recruitment to active sites of intracellular signaling (6). Although experimental results in Chen et al (4) support the notion that FAK activation upon BRAF inhibitor treatment is independent of ERK/MAPK pathway reactivation, it is possible that constitutive ERK/MAPK pathway activation in BRAF V600E-expressing cells may also impact FAK in a negative regulatory manner. In particular, ERK/MAPK kinase-mediated phosphorylation of FAK at serine-910 has been shown to result in FAK Y397 de-phosphorylation by tyrosine phosphatase action (12). Thus, reduction of constitutive ERK/MAPK activity BRAF V600E-expressing cells by BRAF inhibitor addition could possibly lead to increased FAK Y397 phosphorylation associated with the decrease in ERK/MAPK-mediated FAK S910 phosphorylation.

Notably, increased FAK Y397 phosphorylation is also part of an intrinsic adaptive resistance mechanism of melanoma cells to RAF inhibition (13) and involved in acquired sorafenib (a combined tyrosine kinase and BRAF inhibitor) resistance of KRAS-mutant lung adenocarcinoma cells (14). Moreover, tumor stromal responsive changes to BRAF inhibitor administration revealed an important role for matrix-integrin-FAK signaling in creating a protective tumor microenvironment (15). Importantly, pharmacological BRAF and FAK inhibition exhibit synergy in both BRAF V600E melanoma and CRC pre-clinical models.

How FAK/β-catenin pathway activation could be pre-wired for activation is possibly associated with the generation and survival of cancer stem cells (CSCs). Both FAK and Wnt/β-catenin signaling have been connected to CSC plasticity and chemotherapy resistance (11). Chen et. al. (4) detected increased expression of Wnt pathway targets SOX9 and LGR5 in vermurafenib-treated patient-derived xenograft tumors in mice, suggestive of an increase in a CSC pool. However, it remains unknown whether LGR5⁺ CRC cells exhibit elevated FAK Y397 phosphorylation and nuclear β-catenin enrichment. Another potential connection between FAK/β-catenin and CSCs may also involve a Yes-associated protein 1 (YAP1) transcriptional complex. β-catenin forms a complex with YAP1 in the nucleus (16) and a FAK to YAP1 signaling axis supports stem cell-based tissue renewal in mice (17). Moreover, evidence is also accumulating that FAK and Wnt/β-catenin signaling are also supporting human ductal carcinoma in situ breast cancer (18) and KRAS mutated non-small cell lung cancer (19) resistance to radiotherapy. FAK inhibition resulted in susceptibility to ionizing radiation treatment associated with elevated DNA damage. Taken together, CSC-enriched populations exhibit resistance to radiotherapy and chemotherapy treatment – a similar population of cells that exhibit elevated FAK Y397 phosphorylation and Wnt/β-catenin activity.

Combinatorial clinical trial testing of FAK inhibitors

Although FAK inhibitor administration to mice have revealed anti-tumor activity as a monotherapy (6), early phase clinical trials with various FAK inhibitors have revealed an acceptable safety profile with 15 to 20% of unselected patients exhibiting disease stabilization (20,21). In additional to the studies described above, support for combinatorial approaches using FAK inhibitors come from pre-clinical studies showing that tumor-intrinsic FAK signaling supports pancreatic and squamous cell carcinoma tumor immune evasion via modulation of cytokine expression and effects on tumor-infiltrating T cells (22,23). Pharmacological FAK inhibition was associated with a depletion of T regulatory cells and an enhancement of CD8+ T cell-mediated anti-tumor response. Importantly, FAK inhibition rendered unresponsive KRAS and p53 tumor suppressor mutant pancreatic tumors responsive to T cell immunotherapy and PD-1 checkpoint inhibitor antagonists (23). Mouse survival was greatest using a combinatorial treatment of gemcitabine, FAK inhibitor, and anti-PD-1 antibody therapy. These studies are supporting Phase I/II clinical trials evaluating FAK inhibitor combination with pembrolizumab (PD-1 inhibitor) in pancreatic cancer, mesothelioma, and non-small cell lung carcinoma (Table 1).

Table 1.

Early phase clinical trials using FAK inhibitor as combination therapy.

	Clinical trial ID	FAK inhibitor	Malignancy	Phase	Other agent(s)	Results (if any)
Completed trials	NCT01938443	GSK2256098	Advanced solid tumors	Ib	Trametinib (MEK inhibitor)	Stable disease > 6 weeks = 67%a
	NCT01778803	VS-6063	Ovarian	I/Ib	Paclitaxel	Disease control rate = 64%b Objective response rate = 18%c
Ongoing trials	NCT02523014	GSK2256098	Meningioma	II	Vismodegib (hedgehog pathway inhibitor)
	NCT02546531	VS-6063	Advanced solid tumors	I	Pembrolizumab (PD-1 inhibitor), gemcitabine
	NCT02758587	VS-6063	Pancreatic, mesothelioma, non-small cell lung cancer (NSCLC)	I/IIA	Pembrolizumab (PD-1 inhibitor)
	NCT02943317	VS-6063	Ovarian	I/II	Avelumab (PD-L1 inhibitor)
	NCT03287271	VS-6063	Ovarian	I/II	Carboplatin, paclitaxel

^aStable disease includes those who do not meet criteria for progression (at least 20% increase of target lesions), partial response (at least 30% shrinkage of target lesions), or complete response (disappearance of all target lesions).

^bDisease control rate includes stable disease, partial responses, and complete responses.

^cObjective response rate includes partial responses and complete responses.

^dPD-L1 expressers only

Other combinatorial approaches have targeted key cell signaling molecules involved in metastasis and cell proliferation. Related to the findings from Chen et. al. (4), a Phase Ib trial is evaluating the MEK inhibitor trametinib combined with GSK2256098 (a FAK inhibitor) in patients with advanced solid tumors (NCT01938443), with particular interest in mesothelioma, where synergy has been demonstrated with this drug combination in vitro. Vismodegib, an inhibitor of smoothened receptor (SMO, part of the hedgehog signaling pathway), is administered with GSK2256098 in a Phase II study of patient with meningiomas harboring mutations in SMO or PTCH1 (protein patched homolog 1), a related signaling protein (NCT02523014). By co-targeting the FAK/β-catenin pathway in addition to the mutated hedgehog pathway, there is potential to prevent or combat adaptive resistance.

Advanced ovarian cancer presents a unique treatment challenge because, although tumors are initially sensitive to platinum-based chemotherapy, patients recur and ultimately develop resistance to chemotherapy, with few agents demonstrating anti-tumor activity in this setting. Three trials of combination therapy with FAK inhibitor specifically target recurrent ovarian cancer and aim to overcome mechanisms of drug resistance. In a Phase I/Ib trial of 22 patients with recurrent ovarian cancer, the small molecule FAK inhibitor VS-6063 was administered in combination with paclitaxel, resulting in a 64% disease control rate, including two complete and two partial responses (NCT01778803). The Resensitization of Carboplatin-resistant Ovarian Cancer with Kinase Inhibition of FAK (ROCK-IF) trial is a Phase I/II study of VS-6063 with paclitaxel as well as carboplatin in the same patient population (NCT03287271). Inhibition of the FAK/β-catenin pathway may restore sensitivity to carboplatin, leading to a more significant anti-tumor response.

Conclusion

The search for novel treatment approaches in drug-resistant cancers remains a focus of inquiry in the preclinical and clinical arenas. Activation of a FAK signaling linkage to β-catenin supports tumor cell survival in the setting of cytotoxic or targeted therapy. Targeting FAK in combination with therapies directed at other pathways may be a promising approach to turn off the adaptive “switch” promoting tumor cell proliferation.

Abbreviations

APC

adenomatous polyposis coli

BRAF

B-rapidly accelerated fibrosarcoma

CD8

cluster of differentiation 8

CSC

cancer stem cell

CRC

colorectal cancer

EGFR

epidermal growth factor receptor

ERK

extracellular signal-related kinase

FAK

focal adhesion kinase

GSK

glycogen synthase kinase

MAPK

mitogen-activated protein kinase

MEK

mitogen-activated protein kinase kinase

PD-1

programmed death-1

PD-L1

programmed death-ligand-1

PI3K

phosphoinositide 3-kinase

PTCH1

protein patched homolog 1

SMO

smoothened receptor

TrCP

transducin repeats containing protein

Y

tyrosine

YAP

Yes-associated protein