EGFR signaling and pharmacology in oncology revealed with innovative BRET-based biosensors – Communications Biology

Ligands

The following compounds were used in this study: human recombinant Epidermal Growth Factor (EGF) (Cedarlane; ON, Canada; cat#CL10504), human recombinant Epiregulin (Cedarlane; ON, Canada; cat# CLCYT609), Gefitinib (Iressa®; Tocris Bioscience, United Kingdom; cat# 3000). and Osimertinib (Tagrisso®; MedChemexpress, NJ, USA; cat# HY-15772A).

Plasmids

Human pcDNA3.1(+)-EGFR-WT was kindly provided by Dr Michel Bouvier’s lab (IRIC, Montreal University, QC, Canada). Human pcDNA3.1(+)-EGFR-WT was used to generate EGFR mutants by mutagenesis: (i) EGFR-T790M: threonine in position 790 mutated in methionine; (ii) EGFR-C797S: cysteine in position 797 mutated in serine; (iii) EGFR-T790M/C797S: threonine in position 790 mutated in methionine and cysteine in position 797 mutated in serine; (iv) EGFR-G598V: glycine in position 598 mutated in valine; (v) EGFR-vIII, truncated between amino-acids 30 and 297; and (vi) EGFR-vIV, truncated between amino-acids 982 and 1054. Human SH2-domains of effector proteins were synthesized, and the generated synthetic DNA fragments were subcloned downstream the RlucII portion into a pCDNA3.1(+) vector after HindIII/XbaI digestion (Supplementary Fig. 1a, b). All plasmids were generated by Topgenetech (Montreal, QC, Canada). Their amino-acid sequences are presented in Supplementary Fig. 1. Both rGFP-CAAX and rGFP-FYVE constructs were previously described¹².

Cell culture

Human embryonic kidney 293 HEK293-SL cells were a gift from Dr Stéphane Laporte’s lab (McGill University, Montreal, QC, Canada)¹² and human epidermoid carcinoma A-431 cells (cat# CRL-1555), human adenocarcinoma HeLa cells (cat# CRM-CCL-2) and human adenocarcinoma MDA-MB-231 cells (cat# CRM-HTB-26) were obtained from the American Type Culture Collection (ATCC, VA, USA). All cell lines were cultured in a complete medium containing Dulbecco’s Modified Eagle Medium (DMEM; without sodium pyruvate, with 4.5 g/L glucose, without L-glutamine; Wisent, QC, Canada; cat# 319-030-CL) supplemented with 1% penicillin-streptomycin (PS; 100 U/mL penicillin and 100 μg/mL streptomycin; Wisent, QC, Canada; cat# 450-201-EL) and 10% fetal bovine serum (FBS; Wisent, QC, Canada; cat# 090150). Cells were divided twice a week and incubated at 37 °C in a humidified atmosphere with 5% CO₂.

Transfection

Forty-eight hours before the BRET experiments, HEK293-SL cells were co-transfected with the receptor and one of each RlucII-SH2 / rGFP-localization biosensors. For each transfection condition, the total amount of transfected DNA was kept constant at 1 µg per mL of cell culture to be transfected; whenever necessary, salmon sperm DNA (Invitrogen, CA, USA; cat# 15632011) was used as a DNA carrier to supplement plasmid DNA (i.e., biosensor and receptor). Plasmids quantities per mL of cell culture were assembled in 150 mM NaCl (pH 7.0) as follows: 250 ng of EGFR-WT or EGFR mutants, 10 ng of RlucII-SH2 construct, and 250 ng of rGFP-CAAX (PM) or rGFP-FYVE (EEs) plasmids. PEI (polyethylenimine 25 kDa linear; PolyScience, IL, USA; cat# 23966), previously diluted in 150 mM NaCl, was used as the transfection agent, and the PEI: DNA ratio (µg:µg) was fixed at 3:1. The PEI-containing solution was added to the DNA solution and the DNA/PEI mixture immediately vortexed for 5 s. The DNA/PEI mixture was incubated for at least 20 min at room temperature to allow for the formation of DNA/PEI complexes. During the incubation, HEK293-SL cells were detached, using Trypsin-0,05% EDTA (Wisent, QC, Canada; cat# 325-542-EL), counted and re-suspended in complete DMEM culture medium containing 2% of FBS to a concentration of 3.5 × 10⁵ cells/mL. At the end of the incubation period, the DNA/PEI mixture was added to the cells. Cells were finally distributed in 96-well plates (White Opaque 96-well /Microplates; Greiner, NC, USA; cat# 655083) at a density of 35,000 cells per well. Cells were maintained in culture for the next 48 h before BRET measurements.

A-431 cells were transfected exactly as described for HEK293-SL above, with the exception that no receptor-coding plasmid was transfected since A-431 cells express high levels of endogenous EGFR. HeLa and MDA-MB-231 were transfected using FuGENE® HD Transfection Reagent (Promega, Madison, WI, USA) and OPTI-MEM®I reduced serum medium (Invitrogen; cat# 31985062) as the DNA and FuGENE® HD dilution agent. Cells were distributed in 96-well plates at a density of 45,000 cells per well and maintained in culture for the next 48 h before BRET measurements.

Bioluminescence resonance energy transfer (BRET) measurements

At 48 h post-transfection, BRET experiments were performed according to the following protocol. Culture medium was aspirated and replaced with 100 µl of pH 7.0 Hank’s Balanced Salt Solution buffer (HBSS; without red phenol; with sodium bicarbonate, with calcium and magnesium, with HEPES; Wisent, QC, Canada; cat# 319-067CL) per well. Plates were incubated for at least 60 min at room temperature to allow equilibration of the transfected cells in the HBSS buffer.

Testing of agonist ligands

Increasing concentrations of test compound (EGF or Epiregulin) were added to each well using the HP D300e digital dispenser (Tecan, Switzerland). Compounds were assayed at 22 concentrations for each biosensor. Right after compound injection, 10 µl of 10 µM of luciferase substrate, e-Coelenterazine Prolume Purple (Methoxy e-CTZ; Nanolight Technologies, AZ, USA; cat# 369), were added to each well to a final concentration of 1 µM. Cells were then incubated with the test compounds at room temperature for 10 min (for measurement of responses at the plasma membrane (PM)) or 60 min (for measurement of responses at early endosomes (EE)) under agitation. BRET readings were collected with a 20 s integration time on a SPARK 10 M plate reader (Tecan, Switzerland) with an energy donor filter (400 nm ± 70 nm) and energy acceptor filter (515 ± 20 nm). BRET values were obtained by calculating the ratio of the light emitted by the energy acceptor (rGFP) over the light emitted by the energy donor (RlucII).

Testing of inhibitors

To study the inhibitory effects of Gefitinib at both the PM and EE, increasing Gefitinib concentrations were added to cells (as described above). BRET was recorded 30 min later as a measure of any potential agonist (including inverse agonist) activity of the compounds. Immediately following this initial BRET recording, EGF (EC₈₀) was directly added to each well (the same well treated with the inhibitor) using the HP D300e digital dispenser. Cells were then incubated at room temperature for an additional 10 min (for measurement of responses at the PM) or 60 min (for measurement of responses at EE), after which time a second BRET measurement was recorded to assess the inhibitory activity of the compounds. All BRET readings were collected, as previously mentioned, on a SPARK 10 M plate reader.

Real-time kinetics of agonist ligands

Prior to ligand addition, cells were incubated with 10 µl of 20 µM e-Coelenterazine Prolume Purple (2 µM final concentration) for 5 min at room temperature to stabilize the luciferase signal. Previously determined EC₈₀ of EGF or maximal concentration of Epiregulin was then injected using the HP D300e digital dispenser. BRET values were collected each 30 s for 60 min with a 20 s integration time on the SPARK 10 M plate reader. The same filters as described above were used.

Real-time kinetics of inhibition

After a 5-minute incubation with a final concentration of 2 µM Prolume Purple, cells were stimulated with an EC₈₀ of EGF, and BRET was recorded every 30 s for 30 min (for measurement of responses at the PM) or 60 min (for measurement of responses at EE). After 30 or 60 min of agonist stimulation, 10 µM of Gefitinib were injected, and BRET measurements were recorded every 30 s for another 30 min. In the case of A-431 cells, the same protocol was followed except for the total time of the PM kinetics assay: after 15 min of agonist stimulation, Gefitinib was added for 15 additional minutes.

BRET Calculations and analysis

BRET values were obtained by calculating the ratio of the light emitted by the energy acceptor (rGFP; 515 ± 20 nm)) over the light emitted by the energy donor (RlucII; 400 nm ± 70 nm). BRET ratios were then standardized using the equation below and represented as universal BRET (uBRET) values:

Constants A and B, obtained on the SPARK 10 M plate reader, correspond to the following values:

A = pre-established BRET ratio obtained from transfection of a background control (vector coding for RlucII alone).

B = pre-established BRET ratio obtained from transfection of positive control (vector coding for a GFP10-RlucII fusion protein).

The standardized BRET ratio is referred to as uBRET. uBRET data are expressed as the mean of at least three independent experiments ± SEM.

In order to compare basal (constitutive) activities between EGFR-WT and EGFR mutants (Figs. 6c and 7b), a control condition with no receptor and RlucII-SH2 / rGFP-CAAX (for PM) or rGFP-FYVE (for EE) transfected cells was added on each assay plate; using the mean of these ‘biosensor only’ controls, raw uBRET data were normalized and thus allowing to differentiate changes in basal activities due to EGFR mutations. Using the normalized values calculated for non-stimulated conditions, we generated histograms for constitutive activities comparison (vs.EGFR-WT).

Statistics and reproducibility

All experiments were performed in at least three biological replicates (see figure legends), and data are expressed as the mean of at least three independent experiments ± standard error of the mean (SEM). The statistical significance of the differences was tested using an ordinary one-way ANOVA with Bonferroni post hoc adjustment for multiple comparisons, using GraphPad Prism 8.0. A value of p < 0.05 was considered significant. All generated concentration-response curves were fitted and analyzed using the four-parameter logistic non-linear regression model in GraphPad Prism (v8.0, GraphPad Software Inc, CA, USA). Kinetics were fitted using the non-linear one-phase association model in Prism 8.

BRET imaging

BRET imaging was performed using a BRET microscope composed of an inverted microscope (Eclipse Ti-U, Nikon, NY, USA), an optical filter unit (Lambda 10-2, Sutter Instrument, CA, USA), and an EMCCD camera (HNü512, Nüvü cameras, QC, Canada) as described in Kobayashi et al.⁶³.

HEK293-SL cells were seeded 72 h prior measurement on poly-d-lysine-coated 35 mm glass bottom dishes (MatTek, MA, USA; cat# P35GC-1.5-14-C) at a density of 2–6 × 10⁵ cells per dish, and transfected at 48 h before BRET imaging with 250 ng EGFR-WT, 30 ng of RlucII-SH2(Grb2) construct, 250 ng of rGFP-CAAX (PM) or rGFP-FYVE (EEs) plasmids, completed to 1 µg final with ssDNA.

Just before the imaging experiment, cells were washed with pH 7.4 Modified HBSS (137.9 mM NaCl, 5.33 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 0.44 mM KH₂PO₄, 0.33 mM Na₂HPO₄, 10 mM HEPES). EGF (10 nM), Gefitinib (1 µM), and the luciferase substrate Coelenterazine-400a (10 µM; Nanolight Technologies, AZ, USA; cat# 340) were diluted in HBSS.

After the addition of the luciferase substrate, photon counting frames were continuously obtained and integrated for 10 s without a filter (total luminescence frames), then for 10 s with a long-pass filter (480LP, acceptor frames). The video represents the time-lapse recording with exposure of 50 s/channel/frame and a video rate of 6FPS (1 s → 300 s). Each frame was recorded with an EM gain of 3000 and 100-ms exposure time. Acceptor and total luminescence images were generated by repeating 25 integration cycles (total exposure time 250 s/channel) and integrating all frames with the same filter settings. Image analysis was performed using MATLAB 2019b (The MathWorks, Inc.). Images were treated with photometric correction⁶⁴ and iterative Poisson image denoising^{<a data-track="click" data-track-action="reference anchor" data-track-label="link" data-test="citation-ref" aria-label="Reference 65" title="Azzari, L., Vol. 23. (ed. A. Foi) 1086 – 1090 (IEEE Signal Processing Letters; 2016).” href=”#ref-CR65″ id=”ref-link-section-d132185815e1636″>65} filters. BRET levels used in BRET imaging correspond to the ratio of acceptor (rGFP) photon counts to donor photon counts calculated for each pixel^63,66. BRET levels are expressed as a color-coded heat map, with the lowest being black and purple and the highest red and white.