IQGAP1 scaffold-kinase interaction blockade selectively targets RAS-MAP kinase-driven tumors

Abstract

Upregulation of the ERK1 and ERK2 (ERK1/2) MAP kinase (MAPK) cascade occurs in >30% of cancers, often through mutational activation of receptor tyrosine kinases or other upstream genes, including KRAS and BRAF. Efforts to target endogenous MAPKs are challenged by the fact that these kinases are required for viability in mammals. Additionally, the effectiveness of new inhibitors of mutant BRAF has been diminished by acquired tumor resistance through selection for BRAF-independent mechanisms of ERK1/2 induction. Furthermore, recently identified ERK1/2-inducing mutations in MEK1 and MEK2 (MEK1/2) MAPK genes in melanoma confer resistance to emerging therapeutic MEK inhibitors, underscoring the challenges facing direct kinase inhibition in cancer. MAPK scaffolds, such as IQ motif-containing GTPase activating protein 1 (IQGAP1), assemble pathway kinases to affect signal transmission, and disrupting scaffold function therefore offers an orthogonal approach to MAPK cascade inhibition. Consistent with this, we found a requirement for IQGAP1 in RAS-driven tumorigenesis in mouse and human tissue. In addition, the ERK1/2-binding IQGAP1 WW domain peptide disrupted IQGAP1-ERK1/2 interactions, inhibited RAS- and RAF-driven tumorigenesis, bypassed acquired resistance to the BRAF inhibitor vemurafenib (PLX-4032) and acted as a systemically deliverable therapeutic to significantly increase the lifespan of tumor-bearing mice. Scaffold-kinase interaction blockade acts by a mechanism distinct from direct kinase inhibition and may be a strategy to target overactive oncogenic kinase cascades in cancer.

Figure 1

Diminished tumorigenesis in Iqgap1 knockout mice. (a) Representative clinical appearance of tumors in mice wild type (left), heterozygous (middle) or null (right) for Iqgap1 over time as indicated (two mice in each group are shown at each time point). Scale bar, 5 mm. (b) Quantification of the formation of the tumors in a over time. Color bars denote tumor volume, as indicated in the legend; abortive papillomas <1 mm3 are not shown (n = 7 mice per genotype). TPA, 12-O-tetradecanoylphorbol-13-acetate.

Figure 2

The ERK-binding WW domain of IQGAP1 inhibits RAS-driven invasion in human tissue. (a) Confocal stack images showing the expression of keratin (red), basement membrane protein type VII collagen (Col VII; green) and the nuclei of keratinocytes and fibroblasts (blue) of RAS-driven organotypic human epidermal neoplasia treated with shRNAs targeting scrambled control (shControl) or IQGAP1 (shIQGAP1.a, shIQGAP1.b and shIQGAP1.c). The images are shown on the x-y plane (top) or the z-x plane (middle); note the blockade of invasive destruction of the basement membrane by tissue depletion of IQGAP1. Scale bars (top and middle), 25 μm. On the bottom is immunohistochemistry staining for IQGAP1. The black dashed lines denote the basement membrane separation of the epidermis and underlying dermis. Scale bar (bottom), 25 μm. (b) Immunohistochemistry staining for pERK1/2 (top) and total ERK1/2 (bottom) of the tissue in a treated with siRNAs targeting a nonfunctional control (siControl), ERK1/2 (siERK1/2) or IQGAP1 (siIQGAP1.a) sequence. Black arrowheads denote residual pERK staining in siIQGAP1.a-treated tissue. The black dashed lines denote the basement membrane. Scale bar, 25 μm. (c) Cross-species conservation of the IQGAP1 WW domain; divergent amino acids are listed in red. Pink shading indicates highly conserved residues across the WW domains from multiple proteins. Red asterisks denote residues mutated in the mutant WW domain construct. aa, amino acids. (d) Confocal stacks colored as in a after treatment with lentivirus corresponding to empty vector control, the IQGAP1 WW domain (WW) or the WW domain mutant with mutation of the two tyrosine residues noted in c (mutant WW). Scale bars, 25 μm. (e) Immunoblots for the Myc tag, pERK1/2 and total ERK1/2 in the cells in d. Cell lysates were also probed with antibodies to actin to verify equal loading.

Figure 3

Exogenous WW peptide delivery inhibits neoplastic invasion and diminishes tumorigenesis in vivo. (a) Expression of keratin (red), basement membrane protein type VII collagen (green) and the nuclei of keratinocytes and fibroblasts (blue) of RAS-driven organotypic human epidermal neoplasia treated with 10 μM Myc-tagged scrambled control (SCR peptide) or 10 μM Myc-tagged IQGAP1 WW (WW peptide) daily. Images are shown in the x-y plane (first row) or the z-x plane (second row). Also shown is immunohistochemistry staining for the Myc tag (third row), pERK1/2 (fourth row) and total ERK1/2 (fifth row). Black arrowheads denote residual pERK staining in WW peptide tissue. All scale bars, 25 μm. (b) Survival of 35-day-old Ptf1a^+/Cre; Kras^+/G12D; Trp53^flox/flox mice with established tumor burden and palpable pancreas mass randomized to receive WW peptide (WW), scrambled peptide placebo (SCR) or gemcitabine by intraperitoneal injection. (c) Median survival of the mice in b. Statistical significance was calculated by log-rank test. NS, not significant. Data are shown as the mean ± s.d. (d) Clinical images of tumors from the mice in b. Scale bars, 1 cm. Black arrowheads denote tumor nodules. s, spleen; d, duodenum. The WW peptide–treated tumor is from day 99, and the SCR-treated control tumor is from day 51. (e) Immunohistochemistry staining for the Myc tag of tissue from the mice in b along with an untreated control mouse.

Figure 4

Exogenous WW peptide bypasses PLX-4032 resistance. (a) Viability of PLX-4032–sensitive BRAF^V600E melanoma parental cell lines treated with DMSO, 5 μM PLX-4032, 10 μM scrambled peptide or 10 μM WW peptide, as indicated. (b) Dose response of the cells in a to increasing doses of PLX-4032 (0–50 μM). (c) Viability of PLX-4032–resistant subclones derived from the parental lines in a treated with DMSO, 5 μM PLX-4032, 10 μM scrambled peptide or 10 μM WW peptide, as indicated. (d) Dose response of the cells in c to increasing doses of PLX-4032 (0–50 μM). Data (a–d) are shown as the mean ± s.d. (e) Immunoprecipitation (IP) of IQGAP1 protein and associated ERK1/2 from human primary melanocytes (MC) or BRAF^V600E SK-Mel-28 melanoma cells after daily treatment with 10 μM scramble control (SCR) or WW peptide for 6 d. (f) Quantification of the relative amounts of ERK1/2 as a function of the immunoprecipitated IQGAP1 in e. (g) Model of WW-mediated SKIB action in RAS-MAPK–hyperactive cancer. In untreated cancer cells (left), Iqgap1 scaffolds RAF, MEK and ERK MAPKs and facilitates neoplasia-enabling signal transduction. In the presence of WW SKIB (right), ERK MAPK is sequestered away from the IQGAP1 scaffold, the amounts of pERK1/2 are diminished and neoplastic response is attenuated.