Metabolic regulation of EGFR effector and feedback signaling in pancreatic cancer cells requires K-Ras

Abstract

Nutrient stress driven by glutamine deficiency activates EGFR signaling in a subset of KRAS-mutant pancreatic ductal adenocarcinoma (PDAC) cells. EGFR signaling in the context of glutamine starvation is thought to be instigated by the transcriptional upregulation of EGFR ligands and functions as an adaptation mechanism to allow PDAC cells to maintain metabolic fitness. Having a clear view of the intricate signaling cascades potentiated by the metabolic induction of EGFR is important in understanding how these effector pathways influence cancer progression. In this study, we examined the complex signaling that occurs in PDAC cells when EGFR is activated by glutamine deprivation. We elucidate that the metabolic activation of EGFR is principally mediated by HB-EGF, and that other members of the ErbB receptor tyrosine kinase family are not activated by glutamine starvation. Additionally, we determine that glutamine depletion-driven EGFR signaling is associated with a specific receptor phosphorylation known to participate in a feedback loop, a process that is dependent on Erk. Lastly, we determine that K-Ras is required for glutamine depletion-induced Erk activation, as well as EGFR feedback phosphorylation, but is dispensable for Akt activation. These data provide important insights into the regulation of EGFR signaling in the context of metabolic stresses.

1. Introduction

The EGFR pathway is upregulated during KRAS-driven pancreatic tumorigenesis and inhibition of EGFR signaling limits tumor initiation, despite the presence of oncogenic KRAS [1–3]. Recently, we demonstrated that metabolic stress caused by glutamine deprivation potentiates EGFR signaling in a subset of pancreatic ductal adenocarcinoma (PDAC) cells that harbor oncogenic KRAS mutations [4]. As an adaptation to nutrient deficiencies, EGFR stimulates macropinocytosis, an amino acid supply pathway, through the upregulation of Pak. In this way, PDAC cells that are deprived of glutamine, which is the most depleted amino acid in PDAC tumors relative to normal pancreatic tissue, can maintain their proliferation and survive [5,6]. In addition to Pak, EGFR has numerous other effectors, such as Erk and Akt, that are critical regulators of almost every aspect of tumor initiation and maintenance. Therefore, having a clear picture of the complex signaling cascades potentiated by the metabolic induction of EGFR is critical to understanding how these effector pathways influence cancer progression.

EGFR is a member of the ErbB family of receptor tyrosine kinases (RTKs) that includes ErbB1–4. EGFR, also known as ErbB1, is activated by several different ligands, many of which are transcriptionally upregulated by glutamine depletion in PDAC cells [4]. These metabolically driven ligands have the capacity to activate EGFR through receptor autophosphorylation and drive downstream signaling. In the context of nutrient stress, it is unclear whether these ligands act in concert or if a single EGFR ligand is responsible for the observed signaling output. Additionally, whether other ErbB family members are affected by nutrient stress is unknown. EGFR activation is tightly regulated by feedback mechanisms, some of which involve downstream effector pathways that control EGFR signaling dynamics. The extent to which these feedback mechanisms are employed by PDAC cells in the setting of glutamine starvation is uncertain. During pancreatic carcinogenesis, EGFR and oncogenic KRAS cooperate to maximally activate Erk signaling [1], but how these signals are integrated from a metabolic perspective remains to be elucidated.

In this study, we examined the complex signaling that occurs in PDAC cells when EGFR is activated by nutrient deficiency. We find that HB-EGF, which we previously demonstrated to be upregulated by glutamine deprivation [4], is responsible for the metabolic activation of EGFR. Although exogenous HB-EGF had the capacity to activate multiple ErbB family members, we find that only EGFR is phosphorylated in glutamine-starved cells. Additionally, we determine that glutamine depletion-driven EGFR signaling is associated with the negative feedback phosphorylation of Thr 669 (T669), a process that is wholly dependent on Erk. Lastly, we determine that K-Ras is required for metabolic upregulation of Erk, as well as T669 EGFR phosphorylation, but is dispensable for Akt phosphorylation, which may occur through p38 MAPK signaling. These data provide important insights into the regulation of EGFR signaling in the context of nutrient stress.

2. Materials and Methods

2.1. Cell culture

AsPC-1 cells were maintained in 100 units/mL penicillin/streptomycin under 5% CO2 at 37°C and routinely tested for mycoplasma contamination using ABM’s PCR Mycoplasma Detection Kit. Cells were cultured in RPMI (Corning) supplemented with 10% fetal bovine serum (FBS) and 1 mM sodium pyruvate. Cells have been authenticated via short tandem repeat fingerprinting.

2.2. Glutamine deprivation

All glutamine deprivation experiments were performed for 24 hours, unless otherwise indicated. Cells were plated in complete culture media, which was exchanged with glutamine-free media 1–3 days after cell seeding. For glutamine deprivation experiments, glutamine-free RPMI (Corning) was used.

2.3. Reagents and chemicals

Human EGF (Cat. No. E9644) and human HB-EGF (Cat. No. SRP3052) were obtained from Sigma. Erlotinib (Cat. No. E-4007) was obtained from LC Laboratories. Gefitinib was obtained from Fisher Scientific (Cat. No. 5052093). SB202190 (Cat. No. 507979) and SCH772985 (Cat. No. S7101) were obtained from Selleck Chemical. Human epiregulin (EREG) (Cat. No. 1195-EP-025/CF) was obtained from R&D Systems. Bafilomycin (Cat. No. NC0910155) was obtained from Cayman Chemical. Treatment with erlotinib, gefitinib and lapatinib was done for 2 hours by directly adding to the media. Treatment with bafilomycin was done for 1 hour by directly adding to the media. Treatment with EGF, HB-EGF, or EREG was done by adding directly to the media. For functional antibody assessments, goat anti-HB-EGF (Cat. No. AF259NA) and goat IgG negative control (Cat. No. AB108C) were obtained from R&D Systems and incubated at the indicated concentrations for 24 hours.

2.4. Immunoblotting and antibodies

Cells were lysed in RIPA buffer (10mM Tris-HCl [pH 8.0], 150mM NaCl, 1% sodium deoxycholate, 0.1% SDS, 1% Triton X-100) with protease and phosphatase inhibitors (Roche). Protein concentrations were measured using the DC Protein Assay Kit (Bio-Rad). SDS–PAGE and protein transfer were performed using Mini Gel Tank and Mini Blot Module (Life Technologies). Immunoblotting was detected using near-infrared fluorescence and the Odyssey CLx imager (LI-COR). Quantitative analysis of immunoblots was performed using Image Studio Lite software (LI-COR). The following primary antibodies were used: p-EGFR (Y1068) (CST, 3777), pEGFR (T669) (Abcam, ab227017), EGFR (CST, 4267), p-Erk1/2 (T202/Y204) (CST, 4370), Erk1/2 (CST, 4695), p-Akt (S473) (CST, 4060), pan-Akt (CST, 4691), K-Ras (sc-30) and β-actin (Sigma, A1978). At least three independent experiments were performed in all cases.

2.5. siRNA transfection

Lipofectamine RNAiMAX Transfection Reagent (Invitrogen) was used with a final siRNA concentration of 25 nM according to the manufacturer’s protocol. 24 hours after the second transfection, the transfected cells were trypsinized and plated for experiments. The negative control siRNA and siRNAs targeting EGFR, ErbB2, ErbB3 and KRAS (Silencer Select) were purchased from Life Technologies.

3. Results

3.1. Metabolic regulation of ErbB effector signals occur via selective activation of EGFR by HB-EGF

We previously demonstrated that glutamine depletion enhances EGFR phosphorylation and upregulates downstream effector signaling pathways, such as Erk and Akt [4]. This effector signal potentiation is thought to occur via the transcriptional upregulation of EGFR ligands, including AREG, BTC, EREG, HB-EGF and TGFA [4]. Intriguingly, in vivo, the only EGFR ligand that was upregulated in tumor regions that displayed glutamine depletion was HB-EGF; therefore, we sought to explore whether HB-EGF was the critical player in modulating signal transduction in response to glutamine starvation. To examine this, we treated AsPC-1 cells with HB-EGF neutralizing antibodies and measured EGFR activation by assessing phosphorylation at Y1068 (p-EGFR^Y1068). As previously reported, we detected increased levels of p-EGFR^Y1068 upon glutamine withdrawal and this enhancement was selectively suppressed by an HB-EGF-specific antibody (Fig. 1A), suggesting that HB-EGF inactivation is sufficient to abrogate metabolic activation of EGFR. Because HB-EGF can also activate other ErbB receptor tyrosine kinase (RTK) family members, we sought to examine the effects of HB-EGF on the phosphorylation of ErbB2, ErbB3 and ErbB4. While we were unable to detect ErbB4 protein via western blot in these cells, HB-EGF treatment did lead to elevated levels of activated ErbB2 (p-ErbB2^Y1248) and ErbB3 (pErbB3^Y1289; Fig. 1B). Epiregulin (EREG), like HB-EGF, is upregulated at the protein level by glutamine starvation in AsPC-1 cells [4]; however, EREG treatment led to nominal induction of pEGFR^Y1068 and did not activate ErbB2 or ErbB3 (Fig. 1B). Importantly, although exogenous addition of HB-EGF had the capacity to enhance the activation of multiple ErbB receptors, only EGFR exhibited increased phosphorylation in the setting of glutamine deprivation (Fig. 1C). Interestingly, ErbB3 total protein levels were actually diminished by starvation of glutamine (Fig. 1C).

Fig. 1. Metabolic regulation of ErbB effector signals occur via selective activation of EGFR by HB-EGF.

(A) Immunoblots assessing levels of EGFR and p-EGFR^Y1068 under glutamine-containing or glutamine-free conditions. Cells were treated for 24 hours with the indicated concentration of anti-HB-EGF or IgG control. (B) Immunoblots assessing protein expression in cells treated with the indicated EGFR ligand at 50 ng/mL for 5 min. (C) Immunoblots assessing protein expression in cells under glutamine-containing or glutamine-free conditions.

3.2. Glutamine depletion-dependent EGFR activation triggers an Erk-dependent feedback loop

EGFR activation and signaling is tightly regulated in cancer cells and one such mechanism of regulation is lysosome-dependent degradation of the receptor itself [7]. To investigate whether activated EGFR is targeted for lysosome-mediated proteolysis as a means of controlling signal potentiation, we starved cells of glutamine and treated with bafilomycin A1 (Baf). Indeed, after 1 hour of Baf treatment, we detected accumulation of pEGFR^Y1068 (Fig. 2A). Lysosome-dependent degradation was specific to EGFR, as the effector pathway readouts pAkt^S473 and p-Erk1/2 were not affected by Baf (Fig. 2A). As a negative feedback mechanism to control EGFR signals, phosphorylation of the receptor at Thr 669 (p-EGFR^T669) is thought to modulate EGFR signaling by either triggering endocytosis and/or by preventing dimerization [8,9]. We found that glutamine deprivation led to the robust enhancement of p-EGFR^T669, which also accumulated upon Baf treatment (Fig. 2A). EGFR activation by glutamine depletion may be triggering p-EGFR^T669 since partial reduction of EGFR activation by low doses of the inhibitors erlotinib and gefitinib led to near total abrogation of p-EGFR^T669, while only marginally affecting p-EGFR^Y1068 (Fig. 2B). As part of the negative feedback loop, Erk is thought to play a pivotal role in Thr 669 phosphorylation [8]. This seems to be the case in the context of glutamine starvation since treatment with the Erk inhibitor SCH772984 led to abrogation of p-EGFR^T669 levels, while leaving p-EGFR^Y1068 intact (Fig. 2C).

Fig. 2. Glutamine depletion-dependent EGFR activation triggers an Erk-dependent feedback loop.

(A) Immunoblots assessing protein levels in cells under glutamine-containing or glutamine-free conditions. Bafilomycin (Baf) was used for 1 hour at 250 nM. DMSO was used as a vehicle-only control. p-EGFR/EGFR ratios are shown relative to control. (B) Immunoblots assessing protein levels in cells under glutamine-containing or glutamine-free conditions. Erlotinib (Erl) and gefitinib (Gef) were used at 5 μM for 2 hours. DMSO was used as a control. (C) Immunoblots assessing protein levels in glutamine-starved cells treated with SCH772984 (SCH) at the indicated concentration for 1 hour. DMSO was used as a control.

3.3. K-Ras mediates EGFR-dependent Erk, but not Akt, activation in response to metabolic stress

We previously deciphered that K-Ras is required for EGFR-dependent Pak activation in the setting of glutamine depletion[4]. To investigate whether K-Ras is also required for activation of Akt or Erk, we performed siRNA-mediated knockdown experiments and evaluated levels of pAkt^S473 and p-Erk1/2 upon glutamine withdrawal. KRAS knockdown led to the suppression of p-Erk1/2, but p-Akt^S473 levels were still enhanced by glutamine starvation (Fig. 3A). Interestingly, the increase in Akt activation in response to glutamine deprivation may involve p38 MAPK signaling, since treatment with SB202190, a potent p38 MAPK inhibitor targeting p38α/β[10], reduced levels of p-Akt^S473 (Fig. 3B). Consistent with K-Ras mediating glutamine depletion-induced Erk activation, KRAS knockdown led to complete suppression of p-EGFR^T669 induction (Fig. 3C), suggesting that K-Ras plays an important role in EGFR feedback regulation.

Fig. 3. K-Ras mediates EGFR-dependent Erk, but not Akt, activation in response to metabolic stress.

(A) Immunoblots assessing protein levels in cells under glutamine-containing or glutamine-free conditions. Cell were transfected with siRNAs targeting KRAS or a non-targeting control siRNA (siNC1). (B) Immunoblots assessing protein levels in cells under glutamine-containing or glutamine-free conditions. SB202190 (SB) was used at the indicated concentration for 4 hours. DMSO was used as a control. (C) Immunoblots assessing protein levels in in cells under glutamine-containing or glutamine-free conditions.

4. Discussion

Altogether, our data suggest a model in which glutamine deprivation drives the upregulation of HB-EGF, which selectively activates EGFR effector pathways, some of which require K-Ras (Fig. 4). We demonstrate that HB-EGF is the critical ligand that activates EGFR in response to metabolic stress, and that while exogenous HB-EGF has the capacity to activate additional ErbB family members, this activation does not occur in the context of glutamine depletion. We show that EGFR feedback mechanisms that depend on Erk signaling are instigated in PDAC cells upon glutamine withdrawal, suggesting that the metabolic activation of EGFR signal transduction is finely controlled. We previously outlined how K-Ras is required to enhance Pak activation in response to glutamine depletion [4]. Here, we extend our findings and delineate that K-Ras is required to enhance the activation of Erk, while it may be dispensable for Akt signaling. The observed Akt activation may involve p38 MAPK signaling, or other signal transduction mechanisms, such as PI3K or H/N-Ras activation. Our work underscores the extent of signal potentiation complexity in PDAC cells and sheds light on how tumors can integrate metabolic inputs to control different aspects of cancer signaling that regulate cellular fitness and growth.

Fig. 4. Model depicting EGFR signal activation in response to nutrient stress.

Activating phosphorylations are shown in green and feedback phosphorylations are shown in yellow.

Highlights.

• Metabolic activation of EGFR is selectively mediated by HB-EGF

• Glutamine depletion-driven EGFR signaling elicits feedback phosphorylation that is dependent on Erk

• K-Ras is required for glutamine depletion-induced Erk activation and feedback phosphorylation