Advertisement
Cancer Biology and Translational Studies

Small Molecule–Mediated Activation of RAS Elicits Biphasic Modulation of Phospho-ERK Levels that Are Regulated through Negative Feedback on SOS1

Jennifer E. Howes, Denis T. Akan, Michael C. Burns, Olivia W. Rossanese, Alex G. Waterson and Stephen W. Fesik
DOI: 10.1158/1535-7163.MCT-17-0666 Published May 2018

Abstract

Oncogenic mutation of RAS results in aberrant cellular signaling and is responsible for more than 30% of all human tumors. Therefore, pharmacologic modulation of RAS has attracted great interest as a therapeutic strategy. Our laboratory has recently discovered small molecules that activate Son of Sevenless (SOS)–catalyzed nucleotide exchange on RAS and inhibit downstream signaling. Here, we describe how pharmacologically targeting SOS1 induced biphasic modulation of RAS-GTP and ERK phosphorylation levels, which we observed in a variety of cell lines expressing different RAS-mutant isoforms. We show that compound treatment caused an increase in phosphorylation at ERK consensus motifs on SOS1 that was not observed with the expression of a non-phosphorylatable S1178A SOS1 mutant or after pretreatment with an ERK inhibitor. Phosphorylation at S1178 on SOS1 is known to inhibit the association between SOS1 and GRB2 and disrupt SOS1 membrane localization. Consistent with this, we show that wild-type SOS1 and GRB2 dissociated in a time-dependent fashion in response to compound treatment, and conversely, this interaction was enhanced with the expression of an S1178A SOS1 mutant. Furthermore, in cells expressing either S1178A SOS1 or a constitutively membrane-bound CAAX box tagged SOS1 mutant, we observed elevated RAS-GTP levels over time in response to compound, as compared with the biphasic changes in RAS-GTP exhibited in cells expressing wild-type SOS1. These results suggest that small molecule targeting of SOS1 can elicit a biphasic modulation of RAS-GTP and phospho-ERK levels through negative feedback on SOS1 that regulates the interaction between SOS1 and GRB2. Mol Cancer Ther; 17(5); 1051–60. ©2018 AACR.

Introduction

RAS proteins are small GTPases that serve as signaling transducers by coupling extracellular receptor activation to intracellular effector pathway signaling. Canonically, RAS proteins cycle from a GDP-bound “inactive” state to a GTP-bound “active” state. The transition between the inactive to the active GTP-bound state is catalytically regulated by guanine nucleotide exchange factors (GEF), such as Son of Sevenless 1 (SOS1; ref. 1). Membrane localization of SOS1 enhances its local protein concentration and enables the activation of RAS (2). Although RAS is tethered to the membrane through a C-terminal CAAX box motif (1), SOS1 is not constitutively membrane bound and relies on posttranslational modifications and protein–protein interactions to facilitate membrane recruitment (2). Classically, autophosphorylation of ligand-activated growth factor receptors results in the binding of adaptor proteins, such as growth factor receptor–bound protein-2 (GRB2), that recruit SOS1 to the membrane and facilitate its interaction with RAS (3, 4), stimulating downstream effector pathway signaling events.

Mutationally activated RAS protein is implicated in the oncogenic progression of three of the four most prevalent cancers in the United States (colon, lung, and pancreas; ref. 5). Hyperactivation of RAS signaling confers capabilities, known as the Hallmarks of Cancer (6) that cause the cell to become malignant (7). Inactivation of oncogenic RAS leads to dramatic tumor regressions in multiple mouse models, making RAS an important therapeutic target (8–10). Early work in our laboratory led to the discovery of small molecules that bind directly to KRAS and inhibit SOS-mediated nucleotide exchange (11). Analogously, a group from Genentech identified small molecules that bind directly to RAS and also inhibit SOS-mediated nucleotide exchange (12). Using a different strategy, the Shokat group discovered small molecules that irreversibly bind to and inhibit the oncogenic G12C mutant form of KRAS. Efforts to indirectly target RAS include the use of farnesyl- and geranylgeranyl-transferase inhibitors to interfere with RAS membrane association (13), RAF, MEK, and ERK inhibitors to block RAS-mediated downstream signaling events (14) and synthetic lethal strategies, such as the use of proteasome inhibitors in the context of mutant RAS-driven tumors (15, 16).

Our laboratory has recently discovered compounds that indirectly target RAS through the activation of SOS-mediated nucleotide exchange (17). We identified aminopiperidine indole compounds (henceforth referred to as RAS activator compounds) that bind to the RAS:SOS:RAS ternary complex in a pocket on the CDC25 domain of SOS1, adjacent to the switch II region of RAS. A group from AstraZeneca also identified fragments that bind at this site (18). Although compound-mediated activation of RAS to a GTP-bound state seems like a counterintuitive strategy, we observed that compounds that activated SOS1-mediated nucleotide exchange on RAS actually caused a striking inhibition of downstream signaling in the MAPK signaling pathway. Specifically, we observed biphasic changes in ERK phosphorylation characterized by an increase in phospho-ERK1/2T202/Y204 at lower compound concentrations and a decrease at higher concentrations (17). To follow up on this exciting discovery, we sought to understand the biological mechanism of RAS activator compound action on intracellular ERK signaling.

Materials and Methods

Cell culture and compounds

HeLa (ATCC CCL-2), HEK-293 (ATCC CRL-1573), NCI-H358 (ATCC CRL-5807), NCI-H1792 (ATCC CRL-5895), HCT 116 (ATCC CCL-247), and NCI-H1299 (ATCC CRL-5803) cells were obtained from the ATCC and cultured in DMEM or RPMI supplemented with 10% (v/v) FBS where appropriate. Cell lines were authenticated by STR profiling using PowerPlex 16HS technology and were tested negative for mycoplasma using the eMYCO Plus Kit PCR test in October 2017 (Genetica DNA Laboratories). After thawing from liquid nitrogen, cells were passaged at least twice before use in experiments and passaged for a maximum of 25 times.

Compound 1 was synthesized as described previously (17). HSP90 inhibitor 17-(allylamino)-17-demethoxygeldanamycin (17-AAG; ref. 19) and ERK1/2 inhibitor SCH772984 (20) were purchased from Selleckchem. EGF was purchased from R&D Systems.

Active RAS-GTP pull-down and Western blotting

Cells were seeded to reach 70% confluency after 24-hour incubation and subsequently treated as indicated. The vehicle control (VC) for compound 1 was DMSO, and the vehicle control for EGF was PBS. Lysates were resolved by SDS-PAGE and transferred onto Immobilon-FL PVDF membranes (Millipore). Membranes were probed with primary antibodies as indicated, incubated with labeled secondary antibody, and scanned using an Odyssey imager (LiCor). Multiple independent blots were run using the same lysate from each experiment. Membranes were cut to allow probing for multiple molecular weight proteins. Where appropriate, pull-down blots were probed with two different species of antibodies, using multiple colors to detect exactly the same molecular weight on the same membrane, for example, pSPP (rabbit = green) and HA (mouse = red). Levels of RAS-GTP were determined using an Active RAS Pull-Down and Detection Kit according to the manufacturer's instructions (Thermo Fisher Scientific #16117). Primary antibodies used were anti-HA (BioLegend #901501); anti-SOS1 (Santa Cruz Biotechnology #SC-256); anti-SPRY4 (Abcam #ab103114); anti-V5 (Invitrogen #46-0705); and antibodies targeting pan-RAS (K-, N-, and HRAS isoforms; #3339), total ERK1/2 (#9102), phospho-ERK1/2T202/Y204 (#9106), total AKT (#2920), phospho-AKTS473 (#4060), phospho-AKTT308 (#4056) GRB2 (#3972), phospho-RSKS380 (#11989), phospho-RSKT359 (#8753), phospho-RSKT573 (#9346), total RSK (#9355), DUSP4 (#5149), DUSP16 (#5523), SPRY2 (#14954), phospho-SPP (#14390), phospho-Y (#8954), and GAPDH (#2118) were obtained from Cell Signaling Technology. Molecular weight markers are indicated in kDa on each blot.

Immunoprecipitation

Cells were seeded and incubated for 24 hours to reach 70% confluency. Subsequently, cells were transfected with an HA-SOS1WT plasmid (a kind gift from Dr. Bar-Sagi, NYU School of Medicine, New York, Addgene #32920; ref. 21), or with a V5-SPRY2 plasmid (donor vector was a kind gift from Dominic Esposito, NIH National Cancer Institute, Frederick, Addgene #70614) using Lipofectamine 2000 (Life Technologies) and incubated overnight. Mock control cells were transfected with lipid only. Cells were treated with compound or vehicle control alone for the times and concentrations indicated. Cells were lysed on ice using lysis buffer [150 mmol/L NaCl, 10 mmol/L Tris-HCl pH 7.4, 1% (v/v) Triton X-100, 10% (v/v) glycerol, 1 mmol/L EDTA, and supplemented with protease and phosphatase inhibitors (Roche)]. Protein concentration of each sample was assessed using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific #23225). Lysates of equal protein concentration were incubated at 4°C for 1 hour with anti-HA magnetic beads for SOS1 immunoprecipitation (Cell Signaling Technology #11846), or anti-V5 magnetic beads for SPRY2 immunoprecipitation (MBL #M167-11). Samples were washed using PBS with 0.1% TWEEN 20 (v/v), eluted in SDS loading buffer, and incubated at 95°C for 10 minutes with 10 mmol/L DTT. For the SOS1-GRB2 coimmunoprecipitation, samples were eluted from the beads in SDS loading buffer and incubated at 50°C for 10 minutes, prior to boiling the eluent at 95°C with 10 mmol/L DTT. Input VC-treated samples were run on the immunoprecipitation Western blots to compare protein levels between the input and immunoprecipitation samples.

Site-directed mutagenesis and cloning

The HA-SOS1S1178A construct was made using the QuikChange II Site Directed Mutagenesis Kit (Agilent #200523) by following the manufacturer's instructions. Mutations were introduced to the HA-SOS1WT plasmid. The primers used for mutagenesis were as follows:

Forward: 5′-ATTATGTCTAAGCACTTGGACGCCCCCCCAGC-3′; reverse: 5′-GCTGGGGGGGCGTCCAAGTGCTTAGACATAAT-3′. Primers were PAGE purified and synthesized by Eurofins Genomics.

The HA-SOS1CAAX construct was made by cloning the hypervariable region of KRAS 4B onto the C-terminus of the HA-SOS1WT construct (protein sequence: KMSKDGKKKKKKSKTKCVIM). The HA-SOS1WT plasmid was double digested using BsmI and SbfI (New England Biolabs). The CAAX box was ligated in the HA-SOS1WT plasmid using two preannealed oligomers. The oligomers used to insert KRAS 4B CAAX box onto HA-SOS1WT were as follows:

Forward 5′-CATTCTTCCAAGATGAGCAAAGATGGTAAAAAGAAGAAAAAGAAGTCAAAGACAAAGTGTGTAATTATGTAACCTGCA-3′, reverse: 5′-GGTTACATAATTACACACTTTGTCTTTGACTTCTTTTTCTTCTTTTTACCATCTTTGCTCATCTTGGAAGAATGGG-3′. Oligomers were synthesized by Eurofins Genomics.

The V5-SPRY2 construct was made using Gateway Cloning (Invitrogen). The SPRY2 coding region from the donor vector was recombined into the pcDNA-DEST40 vector for expression in mammalian cells. All constructs were validated by sequencing prior to use.

Results

Treatment with a RAS activator compound causes biphasic changes in RAS-GTP and phospho-ERK levels

To gain a mechanistic understanding of RAS activator–mediated signaling, we investigated how RAS activator compound 1 (compound 4 from ref. 17; Fig. 1A) modulated changes in RAS-GTP levels and downstream phospho-ERK signaling events. To determine that the effects on ERK phosphorylation elicited by compound 1 treatment were not due to off-target kinase inhibition, we tested compound 1 in a kinase activity panel and did not identify any substantial inhibition against 342 protein kinases (Supplementary Table S1). Next, we determined the effects of different concentrations of 1 on RAS-GTP and phospho-ERK1/2T202/Y204 levels in cancer cells. Consistent with our previous observations (17), RAS-GTP levels increased in response to treatment with 50 to 100 μmol/L concentrations of 1 treatment (Fig. 1B). Furthermore, biphasic modulation of ERK phosphorylation was observed, characterized by an increase in phospho-ERK at lower concentrations and a decrease at higher concentrations of 1. Notably, when assessing the effects of 1 by dose response, we observed a disconnect between RAS-GTP levels and downstream signaling patterns in ERK phosphorylation, where an increase in RAS-GTP levels did not directly correspond to an increase in signaling downstream as anticipated (Fig. 1B). To gain a better insight, we assessed the effects of 1 on RAS-GTP levels and downstream ERK phosphorylation over a time course of 180 minutes using an activating concentration of 1 (Fig. 1C). A rapid increase in RAS-GTP levels was observed after 5 minutes of treatment with 1 that peaked after 10 to 20 minutes and remained detectable until 90 minutes, then RAS-GTP levels were reduced back to baseline after 120 minutes of treatment (Fig. 1C). Subsequent to heightened RAS-GTP levels, a peak in phospho-ERK protein levels occurred after 20 to 30 minutes of treatment with 1 (Fig. 1C). Moreover, phospho-ERK levels decreased in conjunction with RAS-GTP levels after 90 minutes of treatment with 1. Indicative of ERK activation (22), we also observed a similar biphasic pattern in phosphorylation on RSK at S380 and T359 (Fig. 1C). Phosphorylation at both of these sites on RSK was also induced by EGF stimulation. Phosphorylation on RSK at T573 was not induced by either 1 treatment or EGF stimulation (Fig. 1C). Furthermore, as active ERK is known to induce the expression of DUSP and SPRY proteins that can negatively regulate RAS–MAPK signaling (23), we assessed the effects of compound treatment on total protein levels of DUSP4, a nuclear phosphatase known to inactivate ERK1/2 in response to growth factor stimulation (24), DUSP16, a cytoplasmic phosphatase that regulates JNK signaling and can sequester activated ERK from its substrates (25), and SPRY2 (26) and SPRY4 (27), both antagonists of receptor tyrosine kinase–mediated MAPK signaling. We did not observe any increase in total protein expression of DUSP4, DUSP16, SPRY2, or SPRY4 in response to 1 treatment for up to 180 minutes in HeLa cells (Supplementary Fig. S1).

Figure 1.
Figure 1.

Treatment with 1 causes modulation of RAS-GTP and phospho-ERK levels. A, Chemical structure of 1 (compound 4 from ref. 17). B, HeLa cells were treated with various concentrations of 1 for 30 minutes. C, HeLa cells were treated with 50 μmol/L 1 over a time course of up to 180 minutes. EGF (50 ng/mL for 5 minutes) was used as a positive control. D, HeLa cells were stimulated with EGF (50 ng/mL) over a time course of up to 60 minutes. Two biological repeats of each experiment were independently conducted.

Importantly, we show that this biphasic modulation of both RAS-GTP levels and ERK phosphorylation in response to 1 treatment is also observed in cells that express mutant RAS isoforms (Supplementary Fig. S2). We treated NCI-H358 (KRAS heterozygous G12C), NCI-H1792 (KRAS heterozygous G12C), HCT 116 (KRAS heterozygous G13D), and NCI-H1299 cells (NRAS heterozygous Q61K) with 1 over a time course of up to 180 minutes, and we observed various degrees of biphasic RAS-GTP and phospho ERK modulation. Interestingly, the degree of RAS-GTP fluctuation in RAS-mutant cell lines was not as striking as in HeLa cells [KRAS homozygous wild-type (WT)]. However, EGF treatment also did not markedly induce RAS-GTP levels in cells expressing mutant RAS isoforms. We hypothesized this is due to constitutive activation of RAS in RAS-mutant cell lines (28). Therefore, we proceeded to explore the mechanism of compound action using cell lines expressing WT KRAS.

The biphasic modulation of RAS-GTP and ERK phosphorylation that we observed in response to 1 treatment is similar to a phenotype reported with growth factor stimulation (29). Consistent with the literature, after 1 minute of stimulation with EGF, we observed an increase in RAS-GTP, leading to heightened levels of phospho-ERK that returned to a level just above baseline after 10 to 15 minutes (Fig. 1D). Interestingly, an HSP90 chaperone inhibitor (17-AAG) can also induce similar fluctuations in phospho-ERK signaling at high treatment concentrations through the induction of a stress response in cells (30, 31). However, to our knowledge, the effects of 17-AAG on RAS-GTP levels have not been assessed. Therefore, we treated HeLa cells with a high concentration of 17-AAG over a time course of up to 180 minutes, to induce a stress response (Supplementary Fig. S3). We show that there is a similarity between 1 treatment and 17-AAG treatment on phospho-ERK levels in HeLa cells. However, 17-AAG did not induce any changes in RAS-GTP levels, unlike that observed with 1 treatment. These data demonstrate that 1 treatment at higher concentrations was not simply causing biphasic phospho-ERK modulation due to a stress response because we also observed modulation of RAS-GTP; the anticipated phenotype of a SOS agonist. Furthermore, lower concentrations of 1 treatment also caused changes in ERK phosphorylation (Supplementary Fig. S4).

Treatment with 1 induces phosphorylation of SOS1 at S1178 within an ERK consensus motif

Due to the biphasic changes in RAS-GTP and phospho-ERK levels observed in response to 1 treatment, we hypothesized that a negative feedback loop may be induced to dampen the effects of increased signaling. In the context of growth factor signaling, activated ERK phosphorylates the C-terminus of SOS1 on serine residues at the ERK consensus phosphorylation motif serine-proline-proline (SPP; ref. 32). Phosphorylation of SOS1 by activated ERK inhibits the interaction between SOS1 and GRB2 (32) and uncouples downstream activation of RAS (29). Due to the phenotypic similarities between treatment with EGF and 1 (Fig. 1C and D), we hypothesized that the same negative feedback mechanism may be elicited in response to compound-mediated induction of RAS-GTP and phospho-ERK levels. To test whether treatment with 1 induced phosphorylation on SOS1 at ERK consensus motifs, we used an antibody to detect phosphorylated serine residues at SPP motifs on immunoprecipitated SOS1 (Fig. 2A). We overexpressed an HA-tagged SOS1 construct and treated cells with an activating concentration of 1 over a time course of up to 90 minutes. Subsequently, HA-SOS1 was isolated by immunoprecipitation and resolved by Western blot to assess phosphorylation of serine residues at any SPP motif on HA-SOS1 (Fig. 2A). In response to treatment with 1, we observed an increase over time in the levels of phospho-SPP in cells expressing HA-SOS1WT (Fig. 2A). Higher levels of phospho-SPP modifications were observed after 90 minutes of treatment with 1, concurrent with heightened levels of phospho-ERK (Fig. 2A). Similarly, increases in SPP phosphorylation levels on HA-SOS1WT and downstream ERK phosphorylation levels were observed in response to EGF stimulation (Fig. 2B). A small shift in apparent molecular weight of SOS1 was also detected after HA-SOS1WT–expressing cells were stimulated with either compound 1 or EGF, suggestive of increased SOS1 phosphorylation over time (Fig. 2A, B, and C).

Figure 2.
Figure 2.

SOS1 is phosphorylated at ERK consensus sites in response to stimulation with 1 and EGF. Levels of phospho-Ser Pro Pro peptide in HA-SOS1 immunoprecipitation (IP) eluent from HeLa cells overexpressing either an HA-tagged WT SOS1 construct (HA-SOS1WT) or an HA-tagged S1178A SOS1 construct (HA-SOS1S1178A). Cells were treated with 1 (A; 50 μmol/L) over a time course of up to 90 minutes, EGF (B; 50 ng/mL) over a time course of up to 20 minutes, and pretreatment with 1 μmol/L ERK inhibitor SCH772984 for 60 minutes followed by 1 treatment (50 μmol/L) over a time course of up to 90 minutes (C). Two biological repeats of each experiment were independently conducted.

Phosphorylation of SOS1 on S1178 (within an ERK consensus motif) is known to disrupt the interaction between SOS1 and GRB2 and has been specifically implicated in the control of RAS activation in the context of growth factor stimulation (32, 33). Therefore, we tested whether S1178 on SOS1 was also phosphorylated in response to treatment with 1. Consistent with the literature (32), after stimulation with EGF in cells overexpressing a nonphosphorylatable mutant form of SOS1, HA-SOS1S1178A, SPP phosphorylation on SOS1 was not observed to the same degree relative to HA-SOS1WT phospho-SPP levels (Fig. 2B). Although HA-SOS1S1178A cannot be phosphorylated at position 1178, residual SPP phosphorylation on HA-SOS1S1178A was still detected after EGF stimulation (Fig. 2B), likely because the C-terminus of SOS1 contains multiple other SPP motifs (32). Interestingly, similarly to EGF stimulation, markedly less SPP phosphorylation was observed in response to 1 treatment in cells expressing the HA-SOS1S1178A mutant, as compared to HA-SOS1WT expressing cells. Under these circumstances, SOS1 phosphorylation at SPP motifs did not increase over time, independent of phospho-ERK levels that behave in a similar manner to cells expressing HA-SOS1WT (Fig. 2A). These data show that HA-SOS1WT is phosphorylated in response to treatment with 1 and led us to hypothesize that active phospho-ERK is the kinase responsible for this phosphorylation event, as S1178 is at an ERK consensus SPP motif, and SPP SOS1 phosphorylation increases in parallel with elevated phospho-ERK levels (Fig. 2A). To test this, we pretreated cells expressing HA-SOS1WT with the ERK1/2 inhibitor SCH772984 for 60 minutes and then assessed the levels of SPP phosphorylation on SOS1 in response to 1 treatment over a time course of up to 90 minutes (Fig. 2C). Following pretreatment with SCH772984 alone, we observed an anticipated decrease in downstream phosphorylation of RSKT359 and, as previously reported, on ERK1/2 itself (20). Importantly, after combined ERK inhibitor pretreatment and treatment with 1, we observed less phosphorylation of SPP motifs on SOS1, as compared with the effects of 1 treatment alone. Similarly, pretreatment with an ERK inhibitor caused less SPP phosphorylation on SOS1 in response to EGF stimulation, as compared to the effects of EGF stimulation alone (Fig. 2C). These data show that ERK is the kinase responsible for phosphorylation of SPP motifs on SOS1 at S1178 in the context of 1 treatment.

Treatment with 1 causes dissociation between SOS1 and GRB2 that is mediated through SOS1 phosphorylation at S1178

After discovering that treatment with 1 caused phosphorylation on S1178 of SOS1 (Fig. 2A), we hypothesized that the interaction between SOS1 and GRB2 may be disrupted in response to treatment with 1. To test this, we assessed the effects of treatment with 1 over time on the interaction between SOS1 and GRB2 using coimmunoprecipitation. We used 293 cells due to their high level of GRB2 expression. FBS stimulation was used as a negative control for the association between SOS1 and GRB2 (34), and consistent with the literature, FBS stimulation inhibited the interaction between HA-SOS1WT and GRB2 below the baseline levels of the vehicle control (Fig. 3A). The HA-SOS1WT–GRB2 interaction decreased over time in response to treatment with 1. In parallel, ERK phosphorylation gradually increased over time (Fig. 3A). This is consistent with the hypothesis that treatment with 1 increases phospho-ERK levels that result in the negative regulation of SOS1 activity through phosphorylation at S1178.

Figure 3.
Figure 3.

The GRB2-SOS1 association is reduced over time after treatment with 1. Two hundred and ninety-three cells were transiently transfected with HA-SOS1WT (A), HA-SOS1S1178A (B), HA-SOS1CAAX (C), and treated with 25 μmol/L 1 over 90 minutes. D, 293 cells were transiently transfected with HA-SOS1WT and pretreated with 1 μmol/L ERK inhibitor SCH772984 or a vehicle control for 1 hour where indicated, followed by treatment with 25 μmol/L 1 for either 60 or 90 minutes. Where indicated, cells were treated with FBS (20% v/v) for 15 minutes. Two biological repeats of each experiment were independently conducted.

Next, we used a SOS1 S1178A nonphosphorylatable mutant (32) to prevent inhibition of the SOS1–GRB2 association through ERK-mediated phosphorylation on SOS1. After overexpression of HA-SOS1S1178A, some enhancement of the baseline interaction between SOS1 and GRB2 was observed, compared with cells overexpressing HA-SOS1WT (Fig. 3B). This is consistent with the literature showing that phosphorylation at S1178 of SOS1 negatively regulates the association between SOS1 and GRB2 (32). After overexpression of HA-SOS1S1178A, we hypothesized that there would be no disruption of the SOS1–GRB2 interaction in response to 1 treatment. After a 60- to 90-minute treatment with 1, we did not see marked inhibition of the HA-SOS1S1178A–GRB2 association below the vehicle control baseline, as was observed in HA-SOS1WT cells (Fig. 3A and B). Similarly, the GRB2–SOS1 association was maintained at baseline after stimulation with FBS in cells overexpressing HA-SOS1S1178A (Fig. 3B). Furthermore, we used a SOS1 CAAX box construct (2) to constitutively localize SOS1 to the cell membrane. The baseline SOS1–GRB2 association was lower in cells expressing HA-SOS1CAAX as compared with cells expressing HA-SOS1WT, and this may have been due to the position of the CAAX tag on SOS1 interfering with C-terminal GRB2 binding domain of SOS1 (Fig. 3C). Similarly to cells expressing HA-SOS1S1178A, we did not observe marked disruption of the GRB2-HA-SOS1CAAX association after stimulation with FBS or 1 treatment (Fig. 3C). Overall, these data show that 1 disrupts the GRB2–SOS1 interaction, and this may be due to phosphorylation of S1178 on SOS1 by active ERK.

As there was still a partial disruption of the HA-SOS1S1178A–GRB2 association in response to 1 treatment (Fig. 3B), we hypothesized that active ERK-mediated phosphorylation on SOS1 may not be the sole mechanism mediating the SOS1–GRB2 association in the presence of 1. To test this, we pretreated cells expressing HA-SOS1WT with ERK1/2 inhibitor SCH772984 for 60 minutes and then assessed the HA-SOS1WT–GRB2 association in the presence of 1 or FBS. Importantly, ERK inhibitor pretreatment prevented compound 1–induced dissociation between SOS1 and GRB2 that we observed with 1 treatment or FBS treatment alone (Fig. 3D). However, there was still partial disruption of the HA-SOS1WT–GRB2 association in the presence of the ERK inhibitor, indicating that other negative feedback mechanisms may regulate this association. One possibility is that SPRY2 is phosphorylated at Y55 in response to ERK pathway activation (26). Phospho-SPRY2Y55 is known to negatively regulate RAS-GTP levels through interaction with GRB2, resulting in sequestration of GRB2 and SOS1 from the membrane (26). Therefore, we assessed phospho-SPRY2 levels in response to 1 treatment over a time course of up to 90 minutes. To achieve this, we overexpressed V5-SPRY2 alone (Supplementary Fig. S5) or in combination with HA-SOS1S1178A (Supplementary Fig. S6) in 293 cells, immunoprecipitated SPRY2 using the V5 tag and assessed tyrosine phosphorylation on SPRY2. We did not observe tyrosine phosphorylation of SPRY2 in 293 cells, as has been previously reported in HeLa cells (26). These data suggest that SPRY2 phosphorylation on tyrosine residues does not mediate cellular response to 1 treatment in 293 cells.

We also assessed the effects of compound treatment on DUSP16 and DUSP4 protein levels and saw small increases in the levels of both proteins at later time points of 1 treatment in 293 cells, as compared with the vehicle control (Fig. 3D). Overall, although there may be multiple cell line–specific mechanisms modulating molecular responses to 1 treatment, our data suggest that treatment with 1 causes dissociation of SOS1 and GRB2 through ERK-mediated phosphorylation of SOS1, and this is the primary mechanism modulating cellular responses to compound.

Changes in RAS-GTP levels in response to treatment with 1 are mediated through negative feedback on SOS1 at S1178

SOS1 is localized to the membrane through interactions with GRB2 (4). As RAS is constitutively membrane bound, membrane localization of SOS1 facilitates activation of RAS to its GTP-bound state (2, 4, 33). Because overexpression of HA-SOS1S1178A reduced compound-mediated dissociation of SOS1 and GRB2 (Fig. 3B), we anticipated that in cells expressing HA-SOS1S1178A or constitutively membrane bound HA-SOS1CAAX, the biphasic changes in RAS-GTP levels in response to 1 over time would be diminished, as compared with the response of cells expressing HA-SOS1WT. To test this, we overexpressed HA-SOS1WT, HA-SOS1S1178A, and HA-SOS1CAAX in HeLa cells and assessed RAS-GTP levels in response to an activating concentration of 1, in a time-dependent fashion. After HA-SOS1WT overexpression, we observed a similar biphasic pattern of RAS-GTP stimulation and phospho-ERK activation as in HeLa cells without SOS1 overexpression (Figs. 4A and 1C). However, after overexpression of either HA-SOS1S1178A or HA-SOS1CAAX, RAS-GTP levels remained elevated at later time points, independent of phospho-ERK levels (Fig. 4B and C). Consistent with the literature (35), these data suggest that membrane localization of SOS1 either through CAAX box expression or constitutive GRB2 association is the key to maintaining RAS-GTP levels, and based on these results, we suggest that phosphorylation of SOS1 at S1178 negatively regulates SOS1-mediated RAS activation, in the context of treatment with 1.

Figure 4.
Figure 4.

RAS-GTP levels are maintained above baseline in response to treatment with 1 in cells overexpressing S1178A and CAAX SOS1 mutants. HeLa cells were transiently transfected with HA-SOS1WT (A), HA-SOS1S1178A (B), HA-SOS1CAAX (C), and treated with 1 (50 μmol/L) over a time course of up to 150 minutes. EGF (50 ng/mL for 5 minutes) was used as a positive control. Two biological repeats of each experiment have been independently conducted.

We propose a negative feedback mechanism through which treatment with 1 leads to changes in RAS-GTP and phospho-ERK signaling over time (Fig. 5A and B). The responses to compound are split into two main phases, the early and late responses. Initially, 1 binds to SOS1 and stimulates the exchange of GDP for GTP on RAS, characterized by a rapid induction in RAS-GTP levels. Increased levels of RAS-GTP result in enhanced downstream signaling, as evidenced by an increase in phospho-ERK levels. Subsequently, a negative feedback loop causes SOS1 to be phosphorylated at the C-terminus by the active form of ERK. Phosphorylation of SOS1 at S1178 disrupts the interaction between GRB2 and SOS1, hence delocalizing SOS1 from the vicinity of RAS. This results in a gradual decrease over time in RAS-GTP and consequently in phospho-ERK levels at later stages of response to compound.

Figure 5.
Figure 5.

Negative feedback model for modulation of RAS-GTP and phospho-ERK signaling by RAS activator compound 1. A, Early signaling responses to compound (up to 30 minutes posttreatment). Activator compound 1 (hexagon) binds to SOS1, stimulates exchange of RAS-GDP to RAS-GTP, and subsequently activates downstream signaling. Activated ERK phosphorylates SOS1 and can also increase DUSP protein expression levels in a cell line–specific manner. B, Late signaling responses to activator compounds (past 60 minutes treatment). Phosphorylation of SOS1 disrupts the interaction between GRB2, hence preventing association with RAS. RAS-GTP levels and downstream signaling are inhibited. DUSPs can also inhibit ERK activation.

Discussion

Our laboratory has discovered compounds that bind to SOS1 and activate RAS, resulting in biphasic signaling in the MAPK pathway (17). The experiments outlined here help define the biological mechanism of action of a RAS activator compound. Overall, our data show that a decrease in ERK phosphorylation is achieved via induction of a negative feedback loop to override compound-mediated stimulation of RAS-GTP. Furthermore, membrane localization of SOS1 may be important in the context of 1 treatment; substantiating the mechanistic hypothesis that compound-mediated phosphorylation of SOS1 by ERK can negatively regulate its localization in proximity to RAS through dissociation with GRB2.

We propose that modulation of the association between SOS1 and GRB2 is the main mechanism mediating cell response to compound; however, our data also suggest that additional feedback mechanisms may contribute to the biphasic signaling phenotype that we observe. There are multiple alternative molecular mechanisms that regulate SOS1 membrane localization, such as PIP2 binding via the pleckstrin homology domain of SOS1, and allosteric binding of RAS to the CDC25 domain of SOS1 (2, 36). We show that total DUSP4 and DUSP16 protein levels were increased in response to compound treatment in 293 cells, but not in HeLa cells, and so even though we observe similar biphasic signaling phenotypes in response to RAS activator compound treatment across different cell lines, multiple feedback mechanisms may occur to achieve the same overall phenotype, and these could include changes in DUSP and SPRY total- and phospho-protein levels. Nonetheless, our experiments show that the SOS1–GRB2 association is important for mediating response to 1 in cells, and consistent with our findings, targeted inhibition of the SOS1–GRB2 interaction using a reactive peptide also causes inhibition of phospho-ERK levels (37).

Here, we have shown that treatment with compound 1 can lead to activation of RAS–GTP and of downstream ERK signaling. Although seemingly counterintuitive, we speculate that rapid hyperactivation of an oncogenic pathway may be a valid therapeutic strategy. In support of this, early experiments using transforming retroviruses showed that overexpression of oncogenes can result in growth arrest and cell death, for example, activation of RAS in rat embryonic fibroblasts leads to growth arrest (38), and overexpression of active RAS in glioblastoma cells results in autophagic cell death (39). Furthermore, a publication from the Varmus laboratory showed that mutual exclusivity of oncogenic KRAS and EGFR mutations is likely due to synthetic lethality and not functional redundancy (40). In cells coexpressing both EGFR and RAS mutations, concurrent increases in both phospho-ERK levels and PARP cleavage were observed. On the basis of these results, it was suggested that signaling agonists could be utilized therapeutically to provoke synthetic lethal events in the presence of another mutant oncogene (40). Thus, it may be possible that a RAS activator compound could provide a useful approach to therapeutically impair oncogenic signaling.

Disclosure of Potential Conflicts of Interest

RAS activator compounds have been licensed to Boehringer Ingelheim.

Authors' Contributions

Conception and design: J.E. Howes, D.T. Akan, M.C. Burns, O.W. Rossanese, A.G. Waterson, S.W. Fesik

Development of methodology: J.E. Howes, D.T. Akan, M.C. Burns

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.E. Howes, D.T. Akan, M.C. Burns

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.E. Howes, D.T. Akan, M.C. Burns

Writing, review, and/or revision of the manuscript: J.E. Howes, D.T. Akan, M.C. Burns, O.W. Rossanese, A.G. Waterson, S.W. Fesik

Study supervision: A.G. Waterson, S.W. Fesik

Acknowledgments

This work was supported by a Research Investigator Award from the Lustgarten Foundation awarded to S.W. Fesik in 2015.

We thank our colleagues in the Fesik group (Vanderbilt University) for helpful discussions, and we acknowledge P. Patel, J. Abbott, and A. Little (Vanderbilt University, Fesik Lab) for synthesizing RAS activator compound 1 and T. Sobolik (Vanderbilt University) for help with DNA preparation. We acknowledge B. Papke and C. Der (University of North Carolina, Chapel Hill) for discussions and experimental ideas.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Footnotes

  • Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).

  • Received July 12, 2017.
  • Revision received December 11, 2017.
  • Accepted January 11, 2018.
  • Published first February 13, 2018.

References

  1. 1.↵
    1. Vigil D,
    2. Cherfils J,
    3. Rossman KL,
    4. Der CJ
    . Ras superfamily GEFs and GAPs: validated and tractable targets for cancer therapy? Nat Rev Cancer 2010;10:842–57.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Gureasko J,
    2. Galush WJ,
    3. Boykevisch S,
    4. Sondermann H,
    5. Bar-Sagi D,
    6. Groves JT,
    7. et al.
    Membrane-dependent signal integration by the Ras activator Son of sevenless. Nat Struct Mol Biol 2008;15:452–61.
    OpenUrlCrossRefPubMed
  3. 3.↵
    1. Buday L,
    2. Downward J
    . Epidermal growth factor regulates p21ras through the formation of a complex of receptor, Grb2 adapter protein, and Sos nucleotide exchange factor. Cell 1993;73:611–20.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Langlois WJ,
    2. Medh J,
    3. Leitner JW,
    4. Sasaoka T,
    5. Olefsky JM,
    6. Draznin B
    . Insulin and epidermal growth factor influence guanine nucleotide-releasing factor by distinct mechanisms. Endocrinology 1994;135:2412–7.
    OpenUrlCrossRefPubMed
  5. 5.↵
    American Cancer Society. Cancer Facts and Figures. Atlanta, GA: American Cancer Society; 2017.
  6. 6.↵
    1. Hanahan D,
    2. Weinberg RA
    . The hallmarks of cancer. Cell 2000;100:57–70.
    OpenUrlCrossRefPubMed
  7. 7.↵
    1. Pylayeva-Gupta Y,
    2. Grabocka E,
    3. Bar-Sagi D
    . RAS oncogenes: weaving a tumorigenic web. Nat Rev Cancer 2011;11:761–74.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Chin L,
    2. Tam A,
    3. Pomerantz J,
    4. Wong M,
    5. Holash J,
    6. Bardeesy N,
    7. et al.
    Essential role for oncogenic Ras in tumour maintenance. Nature 1999;400:468–72.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Ying H,
    2. Kimmelman AC,
    3. Lyssiotis CA,
    4. Hua S,
    5. Chu GC,
    6. Fletcher-Sananikone E,
    7. et al.
    Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose metabolism. Cell 2012;149:656–70.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Fisher GH,
    2. Wellen SL,
    3. Klimstra D,
    4. Lenczowski JM,
    5. Tichelaar JW,
    6. Lizak MJ,
    7. et al.
    Induction and apoptotic regression of lung adenocarcinomas by regulation of a K-Ras transgene in the presence and absence of tumor suppressor genes. Genes Dev 2001;15:3249–62.
    OpenUrlAbstract/FREE Full Text
  11. 11.↵
    1. Sun Q,
    2. Burke JP,
    3. Phan J,
    4. Burns MC,
    5. Olejniczak ET,
    6. Waterson AG,
    7. et al.
    Discovery of small molecules that bind to K-Ras and inhibit SOS-mediated activation. Angew Chem Int Ed Engl 2012;51:6140–3.
    OpenUrlCrossRefPubMed
  12. 12.↵
    1. Maurer T,
    2. Garrenton LS,
    3. Oh A,
    4. Pitts K,
    5. Anderson DJ,
    6. Skelton NJ,
    7. et al.
    Small-molecule ligands bind to a distinct pocket in Ras and inhibit SOS-mediated nucleotide exchange activity. Proc Natl Acad Sci U S A 2012;109:5299–304.
    OpenUrlAbstract/FREE Full Text
  13. 13.↵
    1. Berndt N,
    2. Hamilton AD,
    3. Sebti SM
    . Targeting protein prenylation for cancer therapy. Nat Rev Cancer 2011;11:775–91.
    OpenUrlCrossRefPubMed
  14. 14.↵
    1. Roberts PJ,
    2. Der CJ
    . Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the treatment of cancer. Oncogene 2007;26:3291–310.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Downward J
    . RAS synthetic lethal screens revisited: still seeking the elusive prize? Clin Cancer Res 2015;21:1802–9.
    OpenUrlAbstract/FREE Full Text
  16. 16.↵
    1. Luo J,
    2. Emanuele MJ,
    3. Li D,
    4. Creighton CJ,
    5. Schlabach MR,
    6. Westbrook TF,
    7. et al.
    A genome-wide RNAi screen identifies multiple synthetic lethal interactions with the Ras oncogene. Cell 2009;137:835–48.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Burns MC,
    2. Sun Q,
    3. Daniels RN,
    4. Camper D,
    5. Kennedy JP,
    6. Phan J,
    7. et al.
    Approach for targeting Ras with small molecules that activate SOS-mediated nucleotide exchange. Proc Natl Acad Sci U S A 2014;111:3401–6.
    OpenUrlAbstract/FREE Full Text
  18. 18.↵
    1. Winter JJ,
    2. Anderson M,
    3. Blades K,
    4. Brassington C,
    5. Breeze AL,
    6. Chresta C,
    7. et al.
    Small molecule binding sites on the Ras:SOS complex can be exploited for inhibition of Ras activation. J Med Chem 2015;58:2265–74.
    OpenUrlCrossRefPubMed
  19. 19.↵
    1. Schulte TW,
    2. Neckers LM
    . The benzoquinone ansamycin 17-allylamino-17-demethoxygeldanamycin binds to HSP90 and shares important biologic activities with geldanamycin. Cancer Chemother Pharmacol 1998;42:273–9.
    OpenUrlCrossRefPubMed
  20. 20.↵
    1. Morris EJ,
    2. Jha S,
    3. Restaino CR,
    4. Dayananth P,
    5. Zhu H,
    6. Cooper A,
    7. et al.
    Discovery of a novel ERK inhibitor with activity in models of acquired resistance to BRAF and MEK inhibitors. Cancer Discov 2013;3:742–50.
    OpenUrlAbstract/FREE Full Text
  21. 21.↵
    1. Yang SS,
    2. Van Aelst L,
    3. Bar-Sagi D
    . Differential interactions of human Sos1 and Sos2 with Grb2. J Biol Chem 1995;270:18212–5.
    OpenUrlAbstract/FREE Full Text
  22. 22.↵
    1. Dalby KN,
    2. Morrice N,
    3. Caudwell FB,
    4. Avruch J,
    5. Cohen P
    . Identification of regulatory phosphorylation sites in mitogen-activated protein kinase (MAPK)-activated protein kinase-1a/p90rsk that are inducible by MAPK. J Biol Chem 1998;273:1496–505.
    OpenUrlAbstract/FREE Full Text
  23. 23.↵
    1. Lake D,
    2. Correa SA,
    3. Muller J
    . Negative feedback regulation of the ERK1/2 MAPK pathway. Cell Mol Life Sci 2016;73:4397–413.
    OpenUrl
  24. 24.↵
    1. Lawan A,
    2. Al-Harthi S,
    3. Cadalbert L,
    4. McCluskey AG,
    5. Shweash M,
    6. Grassia G,
    7. et al.
    Deletion of the dual specific phosphatase-4 (DUSP-4) gene reveals an essential non-redundant role for MAP kinase phosphatase-2 (MKP-2) in proliferation and cell survival. J Biol Chem 2011;286:12933–43.
    OpenUrlAbstract/FREE Full Text
  25. 25.↵
    1. Masuda K,
    2. Katagiri C,
    3. Nomura M,
    4. Sato M,
    5. Kakumoto K,
    6. Akagi T,
    7. et al.
    MKP-7, a JNK phosphatase, blocks ERK-dependent gene activation by anchoring phosphorylated ERK in the cytoplasm. Biochem Biophys Res Commun 2010;393:201–6.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Hanafusa H,
    2. Torii S,
    3. Yasunaga T,
    4. Nishida E
    . Sprouty1 and Sprouty2 provide a control mechanism for the Ras/MAPK signalling pathway. Nat Cell Biol 2002;4:850–8.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Leeksma OC,
    2. Van Achterberg TA,
    3. Tsumura Y,
    4. Toshima J,
    5. Eldering E,
    6. Kroes WG,
    7. et al.
    Human sprouty 4, a new ras antagonist on 5q31, interacts with the dual specificity kinase TESK1. Eur J Biochem 2002;269:2546–56.
    OpenUrlPubMed
  28. 28.↵
    1. Prior IA,
    2. Lewis PD,
    3. Mattos C
    . A comprehensive survey of Ras mutations in cancer. Cancer Res 2012;72:2457–67.
    OpenUrlAbstract/FREE Full Text
  29. 29.↵
    1. Nakayama K,
    2. Satoh T,
    3. Igari A,
    4. Kageyama R,
    5. Nishida E
    . FGF induces oscillations of Hes1 expression and Ras/ERK activation. Curr Biol 2008;18:R332–4.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Koga F,
    2. Xu W,
    3. Karpova TS,
    4. McNally JG,
    5. Baron R,
    6. Neckers L
    . Hsp90 inhibition transiently activates Src kinase and promotes Src-dependent Akt and Erk activation. Proc Natl Acad Sci U S A 2006;103:11318–22.
    OpenUrlAbstract/FREE Full Text
  31. 31.↵
    1. Zhang T,
    2. Li Y,
    3. Zhu Z,
    4. Gu M,
    5. Newman B,
    6. Sun D
    . MEK inhibition potentiates the activity of Hsp90 inhibitor 17-AAG against pancreatic cancer cells. Mol Pharm 2010;7:1576–84.
    OpenUrl
  32. 32.↵
    1. Corbalan-Garcia S,
    2. Yang SS,
    3. Degenhardt KR,
    4. Bar-Sagi D
    . Identification of the mitogen-activated protein kinase phosphorylation sites on human Sos1 that regulate interaction with Grb2. Mol Cell Biol 1996;16:5674–82.
    OpenUrlAbstract/FREE Full Text
  33. 33.↵
    1. Boriack-Sjodin PA,
    2. Margarit SM,
    3. Bar-Sagi D,
    4. Kuriyan J
    . The structural basis of the activation of Ras by Sos. Nature 1998;394:337–43.
    OpenUrlCrossRefPubMed
  34. 34.↵
    1. Waters SB,
    2. Yamauchi K,
    3. Pessin JE
    . Insulin-stimulated disassociation of the SOS-Grb2 complex. Mol Cell Biol 1995;15:2791–9.
    OpenUrlAbstract/FREE Full Text
  35. 35.↵
    1. Aronheim A,
    2. Engelberg D,
    3. Li N,
    4. al-Alawi N,
    5. Schlessinger J,
    6. Karin M
    . Membrane targeting of the nucleotide exchange factor Sos is sufficient for activating the Ras signaling pathway. Cell 1994;78:949–61.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Findlay GM,
    2. Pawson T
    . How is SOS activated? Let us count the ways. Nat Struct Mol Biol 2008;15:538–40.
    OpenUrlPubMed
  37. 37.↵
    1. Yu Y,
    2. Nie Y,
    3. Feng Q,
    4. Qu J,
    5. Wang R,
    6. Bian L,
    7. et al.
    Targeted covalent inhibition of Grb2-Sos1 interaction through proximity-induced conjugation in breast cancer cells. Mol Pharm 2017;14:1548–57.
    OpenUrl
  38. 38.↵
    1. Hirakawa T,
    2. Ruley HE
    . Rescue of cells from ras oncogene-induced growth arrest by a second, complementing, oncogene. Proc Natl Acad Sci U S A 1988;85:1519–23.
    OpenUrlAbstract/FREE Full Text
  39. 39.↵
    1. Overmeyer JH,
    2. Kaul A,
    3. Johnson EE,
    4. Maltese WA
    . Active ras triggers death in glioblastoma cells through hyperstimulation of macropinocytosis. Mol Cancer Res 2008;6:965–77.
    OpenUrlAbstract/FREE Full Text
  40. 40.↵
    1. Unni AM,
    2. Lockwood WW,
    3. Zejnullahu K,
    4. Lee-Lin SQ,
    5. Varmus H
    . Evidence that synthetic lethality underlies the mutual exclusivity of oncogenic KRAS and EGFR mutations in lung adenocarcinoma. Elife 2015;4:e06907.
    OpenUrlAbstract/FREE Full Text
View Abstract
Back to top
View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
Small Molecule–Mediated Activation of RAS Elicits Biphasic Modulation of Phospho-ERK Levels that Are Regulated through Negative Feedback on SOS1
Jennifer E. Howes, Denis T. Akan, Michael C. Burns, Olivia W. Rossanese, Alex G. Waterson and Stephen W. Fesik
Mol Cancer Ther May 1 2018 (17) (5) 1051-1060; DOI: 10.1158/1535-7163.MCT-17-0666
Digg logo Reddit logo Twitter logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement