Terminal α2,6-sialylation of epidermal growth factor receptor modulates antibody therapy response of colorectal cancer cells

Abstract

Background

The epidermal growth factor receptor (EGFR) is a key protein involved in cancer development. Monoclonal antibodies targeting EGFR are approved for the treatment of metastatic colorectal cancer (CRC). Despite the beneficial clinical effects observed in subgroups of patients, the acquisition of resistance to treatment remains a major concern. Protein N-glycosylation of cellular receptors is known to regulate physiological processes leading to activation of downstream signaling pathways. In the present study, the role of EGFR-specific terminal ⍺2,6-sialylation was analyzed in modulation of the malignant phenotype of CRC cells and their resistance to monoclonal antibody Cetuximab-based therapy.

Methods

Glycoengineered CRC cell models with specific sialyltransferase ST6GAL1 expression levels were applied to evaluate EGFR activation, cell surface glycosylation and therapeutic response to Cetuximab.

Results

Glycoproteomic analysis revealed EGFR as a major target of ST6Gal1-mediated ⍺2,6-sialylation in a glycosite-specific manner. Mechanistically, CRC cells with increased ST6Gal1 expression and displaying terminal ⍺2,6-sialylation showed a marked resistance to Cetuximab-induced cytotoxicity. Moreover, we found that this resistance was accompanied by downregulation of EGFR expression and its activation.

Conclusions

Our data indicate that EGFR ⍺2,6-sialylation is a key factor in modulating the susceptibility of CRC cells to antibody targeted therapy, thereby disclosing a potential novel biomarker and providing key molecular information for tailor made anti-cancer strategies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-021-00606-z.

Introduction

Colorectal cancer (CRC) is a major health problem, being the third most common cancer and the second leading cause of cancer-related death worldwide [1]. Having a considerably poor prognosis, almost 50 % of CRC patients develop metastatic disease, which is therapeutically targeted using monoclonal antibodies (mAbs) directed against angiogenic or mitogenic cell surface receptors [2]. Cetuximab and Panitumumab are two mAbs currently applied in the clinic for targeting the extracellular domain of the epidermal growth factor receptor (EGFR). These therapeutic mAbs inhibit EGFR activation by competing with EGF in binding to subdomain III of the receptor’s extracellular region [3–5], thereby blocking tumor growth [2]. However, this EGFR targeted therapy is approved only for metastatic CRC patients carrying tumors without activating mutations in the gene coding for the downstream protein KRAS, which represents 30–40 % of CRC cases [6, 7]. Moreover, only a subset of selected patients (10–20 %) without other known activating mutations in EGFR-downstream molecules has been found to respond to mAb treatment, stressing the need for uncovering additional mechanisms that underlie resistance to anti-EGFR mAb therapy [8].

The complex post-translational process of protein glycosylation is known to be crucial in the regulation of many physiological processes involved in several pathological conditions [9–11]. A specific cancer-associated glycan pattern has been shown to be associated with CRC, including the altered expression of terminal sialylated glycan structures [12]. This aberrant sialylation profile is mainly due to differential expression of specific sialyltransferases and sialidases that participate in distinct glycosylation steps [13–15]. The β-galactoside ⍺2,6-sialyltransferase 1 (ST6Gal1) is an enzyme responsible for the addition of a neuraminic (sialic) acid to a galactose residue in the terminal position of N-glycans in the Golgi complex [16, 17]. ST6Gal1 has been shown to be differentially expressed during CRC progression [18, 19] and its increased activity, together with alterations in its products, ⍺2,6-linked sialic acid (⍺2,6NeuAc) terminal structures, is able to interfere and regulate proliferative, invasive, metastatic and angiogenic processes, as well as epithelial to mesenchymal transition [10, 20–23]. Sialylation has also been described as a regulator of cell surface retention of several glycoproteins through different mechanisms, including disruption of protein-protein and galectin lattice interactions, which can ultimately affect drug responses of malignant cells [24, 25].

Glycosylation of specific proteins is crucial in the recognition of targeted therapies and for patients´ clinical responses. EGFR is a glycoprotein that harbors in its extracellular region eleven predicted N-glycosylation sites [26, 27] and previous in vitro studies have shown that EGFR dimerization, activation and internalization can be modulated by variations in its N-glycan composition [28–30]. Furthermore, the cytotoxic effect of Gefitinib, an EGFR tyrosine kinase inhibitor, was found to be impaired by the overexpression of ST6Gal1, suggesting that EGFR ⍺2,6-sialylation contributes to Gefitinib resistance in CRC cells [31]. While the relevance of α2,6-sialylation in many types of cancer including CRC has been shown, the mechanism through which it impairs the response to mAb treatment directed to this target remains elusive.

In this study, we established a map of EGFR-specific ⍺2,6-sialylation in order to understand the molecular mechanism underlying CRC resistance to mAb treatment. We developed glycoengineered CRC cell lines and unraveled a site-specific signature of cancer-associated glycosylation structures of the EGFR protein. We found that ST6Gal1 overexpressing CRC cells show resistance to Cetuximab, revealing a key role of terminal ⍺2,6-sialylation induced by ST6Gal1 in mAb treatment resistance of CRC cells. Furthermore, we show that ST6Gal1 overexpression downregulates EGFR expression at the cell surface, suggesting that this may be a mechanism underlying mAb resistance and determining the lack of therapeutic response. These findings suggest a key role of ⍺2,6-sialylation in tumor cell biology and targeted therapy, which may ultimately lead to the development of novel therapeutic strategies with relevant clinical applications.

Materials and methods

Cell lines and treatment

Five human CRC cell lines (Caco-2, COLO-205, HT-29, RKO and SW48) were purchased from the American Type Culture Collection (ATCC). Cells were cultured at 37 °C in a 5 % CO₂ atmosphere. Caco-2 and HT-29 cells were cultured in DMEM, High Glucose, Pyruvate medium (Gibco). COLO-205, RKO and SW48 cells were cultured in RPMI-1640 GlutaMAX™, HEPES medium (Gibco). The culture media was supplemented with 10 % heat-inactivated fetal bovine serum (FBS). The cell lines were genotyped by short tandem repeat (STR) profiling and were routinely tested for mycoplasma contamination.

For functional assays with the anti-EGFR mAb Cetuximab (Erbitux®) (Merck KGaA, Darmstadt, Germany), cells were seeded in 6-well culture plates at a density of 5.0 × 10⁵ cells/well. Cells were allowed to adhere for 24 h in complete growth medium and then treated with 10 µg/ml Cetuximab for 72 h, in the absence of antibiotic selection and FBS, in order to remove potential protective factors present in the serum. Untreated cells were used in the absence of FBS as controls for serum-deprivation effects. No serum-deprivation effects were found between untreated cells cultured with and without FBS.

Stable transfection of ST6GAL1 in SW48 cells

A full-length human ST6Gal1 cDNA expression construct was designed and prepared using forward primer 5’-AAAGCTAGCCACCATGATTCACACCAACCTGAAGAAAAAGTTCAG-3’ containing a NheI site and reverse primer 5’-AAAAAGCTTAGCAGTGAATGGTCCGGAAG-3´ containing a HindIII site. This cDNA was cloned into a pcDNA3.1/Hygro(+) plasmid (Invitrogen). SW48 cells were seeded at a density of 5 × 10⁵ cells/well and transfected with the plasmid pcDNA3.1/Hygro(+) containing the full-length ST6GAL1 construct or with the corresponding empty vector (Mock cells) using Lipofectamine™ 2000 (Invitrogen), according to the manufacturer’s instructions. The transfected cell pool was selected for 72 h in the presence of 0.1 mg/ml hygromycin B gold (InvivoGen, San Diego, CA, USA) and then seeded at single-cell density under antibiotic selection. Single-cell isogenic clones were expanded, validated for ST6Gal1 overexpression (OE) and maintained under antibiotic selection. The genetic status of SW48 Mock and ST6Gal1 OE cells was confirmed by Sanger sequencing for the genes KRAS and NRAS (exon 2 (codons 12, 13), exon 3 (codons 59, 61) and exon 4 (codons 117, 146)), BRAF (exon 15 (codon V600)) and EGFR (exon 18,19,20 and 21). No mutations were found.

CRISPR/Cas9-mediated knock out of ST6GAL1 in Caco-2 cells

CRISPR/Cas9-mediated knockout (KO) of the ST6GAL1 gene in Caco-2 cells was performed as previously described [32]. Briefly, Caco-2 cells seeded at a density of 1.5 × 10⁵ cells/well were simultaneously transfected with a plasmid containing a guide RNA (gRNA) with the sequence TGTATCCTCAAGCAGCACCC, which targets exon 1 of the ST6GAL1 gene, and a plasmid encoding the gene for Cas9 endonuclease and a GFP reporter. A FACS Aria II system (BD Biosciences, San Jose, CA, USA) was used for fluorescence-activated single cell sorting of transfected Caco-2 cells to obtain different isogenic clones. Indels at the ST6GAL1 gene of isolated clones were evaluated by IDAA PCR. Three clones were selected for indel validation by Sanger sequencing and by the bioinformatic web tool TIDE [33] and used for the following experiments.

RT-qPCR assay

RNA was extracted using TRI Reagent (Sigma-Aldrich, St. Louis, MO, USA) and reverse transcribed into cDNA using SuperScript® Reverse Transcriptase (Invitrogen), following the manufacturer’s instructions. TaqMan™ Universal PCR Master Mix II, no UNG, and ST6GAL1 TaqMan™ probe (s00949382) (Applied Biosystems®, Foster City, CA, USA) were used to carry out the PCR reaction. 18 S Ribosomal 5 expression levels (TaqMan™ 18S5 probe Hs99999901) were used as endogenous control for normalization of target gene abundance. ST6GAL1 mRNA expression was quantified using the ΔΔCt method. RT-qPCR was performed in triplicate for each biological sample.

Western and lectin blot assays

Total cell lysates were extracted using Lysis Buffer 17 (R&D Systems, Minneapolis, MN, USA) supplemented with a PhosSTOP phosphatase inhibitor cocktail (Sigma-Aldrich). Total protein lysates were separated by polyacrylamide gel electrophoresis (SDS-PAGE) (ST6Gal1 – 25–50 µg; SNA – 5 µg; EGFR/pEGFR – 20 µg) and blotted onto nitrocellulose membranes (GE Healthcare, Chicago, IL, USA). For protein detection, membranes were blocked for 1 h at room temperature (RT), followed by overnight (ON) incubation with primary antibody at 4 °C and 1 h incubation with horseradish peroxidase-conjugated secondary antibody at RT. For lectin detection, membranes were blocked with 2 % polyvinylpyrrolidone ON at 4 °C. Next, membranes were incubated with biotinylated SNA (D1/3000, Vector Laboratories, Burlingame, CA, USA) for 1 h at RT, followed by incubation with peroxidase-conjugated streptavidin (D1/100,000, GE Healthcare). Specific signals were detected using the enhanced chemiluminescence (ECL) detection method. The following primary antibodies working dilutions were used: anti-ST6Gal1 (D1/500, R&D Systems), anti-EGFR (clone D38B1, D1/1000, Cell Signaling Technology, Danvers, MA), anti-phosphorylated EGFR (Y1086) (D1/2000, Thermo Fisher Scientific, Waltham, MA, USA). Anti-β-actin (clone I-19, D1/4000, Santa Cruz Biotechnology, Dallas, TX, USA) and anti-⍺-Tubulin (clone DM1A, D1/10 000, Sigma-Aldrich) were used as loading controls. As secondary antibodies horseradish peroxidase-conjugated polyclonal anti-Rabbit (D1/25,000) and polyclonal anti-Mouse (D1/5000, Jackson Immunoresearch, Cambridgeshire, United Kingdom) were used.

For EGR stimulation, cells were seeded and treated with Cetuximab as previously described. After 48 h, cells were stimulated with 10 ng/ml EGF (Upstate Cell Signaling Solutions, Temecula, CA, USA) for 10 min. Protein samples were separated by SDS-PAGE and probed with anti-EGFR and anti-phosphorylated EGFR (Y1086) antibodies, as described above.

Immunofluorescence assay

Fixed cells were incubated with anti-ST6Gal1 antibody (D1/200) ON at 4 °C, followed by incubation with Alexa-Fluor™ 488-conjugated secondary antibody (D1/500, Invitrogen) for 45 min at RT. For ⍺2,6NeuAc detection, cells were incubated with SNA (D1/500) for 1 h at RT and labeled with FITC-conjugated streptavidin (D1/1000) for 45 min at RT. Nuclear counter staining was performed using DAPI (Sigma-Aldrich). Slides were mounted in VectaShield (Vector Laboratories). Images were acquired at 400x and 630x (inserts) magnification, with a Zeiss Axiocam MR apparatus, using AxioVision v4.8 software (Carl Zeiss, Oberkochen, Germany).

EGFR immunoprecipitation and PNGase F treatment

Protein G Sepharose Fast Flow beads (GE Healthcare Life Sciences) were conjugated with an anti-EGFR antibody (Cell Signaling Technology) for 2 h at 4 °C. In parallel, 1000 µg protein from whole cell lysates was pre-cleared for 2 h at 4 °C. Protein extracts were then incubated with antibody-conjugated beads ON at 4 °C. All conjugation steps were performed with gentle rotation. Antibody-protein bead complexes were washed and, subsequently, subjected to elution and denaturation in Laemmli buffer (BioRad) for separation by SDS-PAGE. Deglycosylation of EGFR was achieved by PNGase F treatment. Immunoprecipitated EGFR was incubated with denaturating buffer for 10 min at 100 °C followed by ON incubation with 2 units PNGase F at 37 °C (New England BioLabs® Inc., Ipswich, MA, USA), prior to separation by SDS-PAGE.

Cell surface lectin labeling assay

Cell surface glycosylation was assessed using lectin staining and flow cytometry. In total 10⁵ cells were stained with each lectin for 30 min on ice. Then, cells were incubated with FITC-conjugated streptavidin for 30 min on ice. Prior to acquisition, 1 µg/ml propidium iodide was added to the cells. Signals were detected using a BD FACSCantoII system (BD Bioscience) and data were analyzed using FlowJo software (BD Bioscience). The following lectins (Vector Laboratories) were used: SNA FITC-conjugated (D/2000), Maackia amuresins lectin I (MAL-I) (D1/500), Aleuria aurantia lectin (AAL) (D1/100), Lotus tetragonolobus lectin (LTL) (D1/200) and Ulex europaeus agglutinin I (UEA-I) (D1/200).

LC-MS/MS glycoproteomic EGFR analysis

EGFR was immunoprecipitated from whole cell lysates as described above, and separated by SDS-PAGE using NuPAGE 4–12 % Bis-Tris Protein Gels (Invitrogen) for sample clean-up. Protein gels were stained for 3 h using a Colloidal Blue staining kit (Invitrogen) and rinsed with distilled water ON. A ~ 170 kDa band was excised for further protein identification and glycoproteomic analysis. Briefly, excised bands were reduced with 10 mM DTT, alkylated with 50 mM iodoacetamide and in-gel digested with trypsin using a Proteineer DP digestion robot (Bruker, Billerica, MA, USA). Tryptic peptides were analyzed via on-line C18-nano-HPLC-MS using an Easy nLC 1000 gradient HPLC system (Thermo Fisher Scientific) and an Orbitrap Fusion™ Lumos™ Tribrid™ (Thermo Fisher Scientific) mass spectrometer. Fractions were injected onto a homemade precolumn (100 μm × 15 mm; Reprosil-Pur C18-AQ 3 μm, Dr. Maisch, Ammerbuch, Germany) and eluted using a homemade analytical nano-HPLC column (15 cm × 50 μm; Reprosil-Pur C18-AQ 3 μm). The gradient was run from 0 to 50 % solvent B (100/0.1 water/formic acid FA v/v) in 20 min. The nano-HPLC column was drawn to a tip of ∼5 μm and acted as the electrospray needle of the MS source. MS/MS spectra were acquired in a data-dependent mode (top-10) with a normalized collision energy of 32 % and recording of the MS/MS spectrum in the Orbitrap. To increase the fragmentation of N-glycosylated peptides, the HexNAc oxonium ion at m/z = 204.087 was set as Product Ion Trigger, and three additional MS/MS scans of the corresponding precursor ion were performed with HCD normalized collision energies of 32, 37 and 41 %, respectively, and a collision-induced dissociation (CID) energy of 35 V.

In a post-analysis process, raw data were first converted to peak lists using Proteome Discoverer version 2.2 (Thermo Electron), and next submitted to the Uniprot Homo sapiens database (67,911 entries), using Mascot v. 2.2.04 for protein identification. Mascot searches were performed with a tolerance of 5 and 20 ppm for precursor and fragment ions, respectively, and trypsin was selected as enzyme. Up to two missed cleavages were allowed. Methionine oxidation and protein N-terminal acetylation were set as a variable modification and carbamidomethyl on cysteines was set as a fixed modification. The data were also analyzed using Byonic™ version 2.13.2 (Protein Metrics, Cupertino, CA, USA) with default settings [34]. In addition to the settings for common modifications as described above, the “N-glycan 309 mammalian no sodium” database was used for the search and assignment of N-glycosylated peptides. Glycopeptide assignment was performed by manual spectra interpretation using Xcalibur™ software (Thermo Fisher Scientific).

Metabolic activity assay

Metabolic activity was measured using a resazurin reduction assay. Cells were seeded in triplicate in a 96-well plate at a density of 5 × 10³ cells/well and treated with increasing concentrations of Cetuximab (0.01–100 µg/ml). Cells treated with hydrogen peroxide (H₂O₂) and 10 µg/ml mouse IgG1 were used as positive and negative controls, respectively. After 72 h, 0.02 mg/ml resazurin (Sigma-Aldrich) was added to the cells and incubated for 3 h at 37 °C. Fluorescence intensity of resorufin (ex. 530 nm; em. 590 nm), product of the resazurin reduction inside metabolically active mitochondria, was measured using a Synergy™ Mx microplate reader (Agilent, Santa Clara, CA, USA).

Apoptosis assay

Cells were seeded as described above and treated for 72 h with 10 µg/ml Cetuximab in culture medium supplemented with 1 % FBS. After treatment, the cells were detached, washed and stained using an Annexin V-FITC Apoptosis Detection Kit (Invitrogen) following the manufacturer’s instructions.

EGFR cell surface expression assay

Cetuximab-treated cells were stained with anti-EGFR monoclonal antibodies Cetuximab (D1/500) and Matuzumab (D1/2000) (Merck KGaA) for 30 min on ice. Next, the cells were incubated with anti-human Alexa-Fluor™ 488-conjugated secondary antibody for 30 min on ice. Prior to image acquisition, 1 µg/ml propidium iodide was added to the cells.

Statistical analysis

Quantitative data are presented as mean ± SD of three independent replicates. Flow cytometry results were analyzed using multiple comparison one-way ANOVA. Statistical analysis of the metabolic activity assays was performed using a two-way ANOVA. Mock and WT were used as control groups for the SW48 ST6Gal1 OE model and the Caco-2 ST6Gal1 KO model, respectively. All statistical analyses were performed using Prism software (GraphPad Inc.). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Results

Expression of ST6Gal1 and EGFR in colorectal cancer cells

The ⍺2,6NeuAc structure has been described as a tumor-associated glycan, and an up-regulation of ST6Gal1 can be commonly found in multiple types of carcinomas, including CRC [15]. In order to determine the expression of ST6Gal1, a panel of five CRC cell lines (Caco-2, COLO-205, HT-29, RKO and SW48) was evaluated. These cell lines were previously described as negative for RAS mutations [35]. Quantitative reverse transcription PCR (RT-qPCR) analysis revealed different levels of ST6GAL1 mRNA expression among the CRC cell lines, with Caco-2 having substantially higher levels of the transcript (Fig. 1a). The expression of ST6Gal1 at the protein level, as well as the detection of the enzyme product ⍺2,6NeuAc by reactivity with Sambucus nigra agglutinin (SNA), was further characterized. Similar to the ST6GAL1 transcript levels, different protein expression levels of the sialyltransferase ST6Gal1 and its product ⍺2,6NeuAc were found among the CRC cell lines tested (Fig. 1b). Intermediate mRNA levels were observed in COLO-205, HT-29 and RKO cells, which did not fully correlate with the respective protein levels, probably reflecting cell line-dependent variations in ST6Gal1 expression. Caco-2 exhibited the highest levels of ST6GAL1 mRNA and protein expression in contrast to SW48 cells, in which the endogenous expression of ST6Gal1 was undetectable. Nevertheless, the levels of ⍺2,6NeuAc followed the same pattern as found for the ST6Gal1 protein expression levels. Based on these findings, and their contrasting ST6Gal1 and ⍺2,6NeuAc expression levels, the Caco-2 and SW48 cell lines were selected for further characterization. Immunofluorescent labeling confirmed the complete absence of ST6Gal1 protein expression in SW48 cells (Fig. 1c, upper panel). In Caco-2 cells, a punctuated signal in the cytoplasmic compartment (Golgi-like) was observed, consistent with the known residency of ST6Gal1 in the Golgi apparatus. Evaluation of reactivity with the SNA reinforced the lack of ⍺2,6NeuAc epitopes in SW48 cells and revealed an intense signal at the cell membrane of Caco-2 cells (Fig. 1c, lower panel).

Fig. 1.

⍺2,6-sialylation and EGFR expression in colorectal cancer (CRC) cell lines. a RT-qPCR analysis of ST6GAL1 transcript levels in Caco-2, COLO-205, HT-29, RKO and SW48 CRC cell lines. b Western blot analysis of ST6Gal1 protein levels and detection of ⍺2,6-linked sialic acid (⍺2,6NeuAc) terminal structures by CRC cell reactivity with Sambucus nigra agglutinin (SNA). c Immunofluorescence labeling for ST6Gal1 (upper panels) and ⍺2,6NeuAc detection by reactivity with SNA (lower panels) in Caco-2 and SW48 CRC cells. Nuclei (blue) stained with DAPI. Scale bar and insert scale bar represent 40 and 20 μm, respectively. d Western blot analysis of total EGFR expression and its phosphorylated form (pEGFR Y1086). e Deglycosylation shift in EGFR following PNGase F digestion of immunoprecipitated EGFR from Caco-2 and SW48 whole cell lysates

The EGFR protein is N-glycosylated and has previously been described as a target of ST6Gal1-mediated ⍺2,6-sialylation [36]. To assess the expression of EGFR and its activation status, a detailed analysis was performed in whole cell lysates from Caco-2 and SW48 cells. EGFR expression was detected in both cell lines, with SW48 displaying higher total receptor levels compared to Caco-2. Additionally, endogenous activation of EGFR (phosphorylated EGFR site Y1086) was only observed in SW48 cells, being undetectable in Caco-2 cells (Fig. 1d). To confirm the presence of N-linked glycans in EGFR, immunoprecipitated EGFR protein from Caco-2 and SW48 whole cell lysates was treated with PNGase F, an enzyme capable of removing all N-glycans from the protein backbone. Upon enzymatic N-deglycosylation, a shift in the EGFR molecular weight (MW) was observed in both cell lines, with EGFR migrating from its expected MW (~ 170 kDa) to a lower MW (Fig. 1e), indicating that EGFR is N-glycosylated in both cell lines.

Impact of altered α2,6-sialylation on EGFR activation

Given their contrasting expression levels of ST6Gal1 and its glycan product, the EGFR-positive SW48 and Caco-2 cell lines were used for a detailed assessment of the functional impact of ⍺2,6-sialylation in CRC cell phenotypes and for the specific EGFR protein glycosylation profile. A ST6Gal1 overexpressing (OE) model was established by stably transfecting SW48 wild-type (WT) cells with an expression vector containing the full-length ST6GAL1 construct or an empty vector (Mock). Four single-cell isogenic clones were selected and validated for ST6Gal1 expression. RT-qPCR showed an upregulation of the ST6GAL1 mRNA level in all SW48 ST6Gal1 OE cell clones (SW48 ST6Gal1 OE 1–4) compared to SW48 WT and Mock cells, both of which displayed undetectable ST6GAL1 transcript levels (Fig. 2a). Overexpression of the ST6Gal1 protein was demonstrated by Western blot analysis, with all SW48 ST6Gal1 OE cell clones showing a band corresponding to the expected ST6Gal1 MW (~ 56 kDa) (Fig. 2b). Furthermore, an increase in ⍺2,6NeuAc glycans in high MW proteins was observed as evaluated by reactivity with SNA, confirming overexpression of a functional ST6Gal1 protein in the stably transfected cells. Immunofluorescent labeling revealed an intense Golgi-like signal in all SW48 ST6Gal1 OE cell clones (Fig. 2c, upper panel), along with the presence of ⍺2,6NeuAc epitopes at the cell membrane (Fig. 2c, lower panel). No immunofluorescent signal was detected in SW48 WT and Mock cells. To evaluate the impact of ST6Gal1 overexpression on the basal activation of EGFR, whole cell lysates from SW48 ST6Gal1 OE cell clones were immunoblotted against total and phosphorylated EGFR. We found that ST6Gal1 overexpression did not alter total EGFR expression and activation compared with cells lacking endogenous ST6Gal1 expression (Fig. 2d).

Fig. 2.

Modulation of ST6Gal1 expression in CRC cell lines. a Characterization of SW48 cells stably transfected with empty (Mock) or ST6GAL1 (OE 1–4) plasmids by RT-qPCR analysis of ST6GAL1 mRNA levels. b Western blot analysis of ST6Gal1 protein levels and detection of its product ⍺2,6NeuAc by reactivity with SNA. c Immunofluorescence labeling of ST6Gal1 (upper panels) and detection of ⍺2,6NeuAc by reactivity with SNA (lower panels). Nuclei (blue) stained with DAPI. Scale bar represents 40 μm. d Western blot analysis of total and phosphorylated EGFR (pEGFR Y1086) expression. e Characterization of Caco-2 WT and ST6Gal1 Knock-Out (KO 17, KO 33 and KO 42) cell clones by RT-qPCR analysis of ST6GAL1 mRNA levels. f Western blot analysis of ST6Gal1 protein levels and detection of its product ⍺2,6NeuAc by reactivity with SNA. g Immunofluorescence labeling of ST6Gal1 (upper panels) and detection of ⍺2,6NeuAc by reactivity with SNA (lower panels). Nuclei (blue) stained with DAPI. Scale bar and insert scale bar represent 40 and 20 μm, respectively. h Western blot analysis of total and phosphorylated EGFR (pEGFR Y1086) expression

To further address the functional role of EGFR ⍺2,6-sialylation in CRC cell malignancy, the ST6GAL1 gene was knocked-out (KO) in Caco-2 cells using the CRISPR/Cas9 technology, after which three isogenic clones were selected (KO 17, KO 33 and KO 42). Nucleotide insertions and deletions (indels) at the ST6GAL1 locus were detected through Indel Detection by Amplicon Analysis (IDAA) PCR (ESM_1 Fig. S1a) and further validated by Sanger sequencing and by the bioinformatic web tool Tracking of Indels by DEcomposition (TIDE). The indel incidence within each clonal cell population is depicted in Online Resource 1b (KO 17: -2/+1; KO 33: +1; KO 42: -7/+1). A decrease in ST6GAL1 transcript levels in all Caco-2 ST6Gal1 KO cell clones was observed (Fig. 2e), indicating a mRNA clearing mechanism possibly by degradation. Western blot analysis revealed a complete abrogation of ST6Gal1 expression and a downregulation of ⍺2,6NeuAc structures in Caco-2 ST6Gal1 KO cell clones (Fig. 2f). Additionally, the Golgi-like punctuated signal observed in Caco-2 WT cells was completely abolished in Caco-2 ST6Gal1 KO cell clones (Fig. 2g, upper panel). Reactivity with SNA revealed the presence of ⍺2,6NeuAc at the surface of Caco-2 WT cells and a complete abrogation of these structures in Caco-2 ST6Gal1 KO cell clones (Fig. 2g, lower panel). Basal activation of EGFR in Caco-2 ST6Gal1 KO cell clones was evaluated by Western blot analysis, which revealed absence of alterations in total EGFR expression levels between Caco-2 WT and ST6Gal1 KO cell clones (Fig. 2h). In addition, we found that abrogation of ST6Gal1 did not induce any alteration in the EGFR activation status in Caco-2 cells.

ST6Gal1 shifts cell surface glycosylation pattern by altering N-glycan terminal motifs

Alterations in the sialic acid and fucose composition of terminal N-glycans have been associated with CRC [12, 37]. To assess the impact of ST6Gal1 expression in the overall cell surface glycosylation pattern, glyco-profiling of SW48 ST6Gal1 OE and Caco-2 ST6Gal1 KO cell clones was performed using lectin flow cytometry. Sialylation can occur either in ⍺2,6- or ⍺2,3-linkage, with different sialyltransferases competing for the same underlying substrates. Labeling of SW48 ST6Gal1 OE cell clones with SNA confirmed an upregulation of ⍺2,6NeuAc in SW48 ST6Gal1 OE cell clones compared to SW48 WT and Mock cells (Fig. 3a). In contrast, abrogation of ST6Gal1 abolished cell surface ⍺2,6-sialylation in Caco-2 cells. Detection of ⍺2,3-linked sialic acid (⍺2,3NeuAc) and Galβ1-4GlcNAc motifs was assessed by the reactivity of cells with Maackia amuresins lectin I (MAL-I). A decrease in MAL-I median fluorescence intensity (MFI) was observed in all SW48 ST6Gal1 OE cell clones compared with SW48 WT and Mock cells (Fig. 3b). Capping of N-glycans with sialic acid in ⍺2,6-linkage may inhibit MAL-I binding to ⍺2,3NeuAc and Galβ1-4GlcNAc motifs, resulting in a decrease in MAL-I reactivity. No alteration was observed in MAL-I signal in Caco-2 ST6Gal1 KO cell clones, suggesting the existence of alternative compensatory mechanisms for N-glycan capping besides ⍺2,3-sialylation.

Fig. 3.

Lectin glyco-profiling of SW48 ST6Gal1 overexpression and Caco-2 ST6Gal1 knock-out models. Cell surface glycosylation of SW48 WT, Mock and ST6Gal1 overexpression cell clones (OE 1–4) and Caco-2 WT and ST6Gal1 knock-out cell clones (KO 17, KO 33 and KO 42) was analyzed by flow cytometry. Lectin labeling of a ⍺2,6NeuAc by Sambucus nigra agglutinin (SNA), b ⍺2,3-linked sialic acid (⍺2,3NeuAc) by Maackia amurensis lectin I (MAL-I), c ⍺1,2-, ⍺1,3/4- and ⍺1,6-linked fucose by Aleuria aurantia lectin (AAL), d ⍺1,2-linked fucose by Lotus tetragonolobus lectin (LTL) and e Ulex europaeus agglutinin I (UEA-I). Data are presented as one representative flow cytometry histogram and median fluorescent intensity (MFI) normalized for binding of secondary antibody or unstained control. Comparisons between multiple groups were made using one-way ANOVA, with Mock and WT used as the control group for the SW48 ST6Gal1 overexpression model and Caco-2 ST6Gal1 knock-out model, respectively; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001

The presence of fucosylated structures was assessed using lectins sensible for fucose-residues linkage. Aleuria aurantia lectin (AAL) indiscriminately recognizes ⍺-linked fucose structures, including terminal ⍺1,2-, ⍺1,3/4- and core ⍺1,6-fucosylation. No major change in the overall fucosylation pattern of SW48 ST6Gal1 OE cell clones as well as of Caco-2 ST6Gal1 KO cell clones was found, compared with the respective control cells (Fig. 3c). Terminal fucosylation was assessed by reactivity of the cells with two lectins which recognize ⍺1,2-linked fucose residues, Lotus tetragonolobus lectin (LTL) and Ulex europaeus agglutinin I (UEA-I). SW48 ST6Gal1 OE cell clones showed a marked decrease in ⍺1,2-linked fucose structures as demonstrated by reductions in the levels of LTL (Fig. 3d) and UEA-I (Fig. 3e) labeling. In addition, Caco-2 ST6Gal1 KO cell clones showed an increase in the MFI of both lectins. Since terminal sialylation is the main competitor of terminal fucosylation, it is no surprise that upregulation of the expression of ST6Gal1 leads to increased levels of SNA labeling with concomitant decreases in ⍺1,2-fucosylation. On the contrary, reducing the expression of ST6Gal1 will lead to decreased levels of terminal sialylation and increased levels of ⍺1,2-fucosylation. Taken together, these findings support a regulatory mechanism involving sialyl- and fucosyltransferases capable of changing cell surface glycosylation patterns.

Overexpression of α2,6-sialylation triggers glycosite‐specific loss of multi‐fucosylated species on EGFR

The extracellular region of EGFR is a well-established molecular target of N-glycosylation. The removal of all N-linked glycans from EGFR results in a decrease of over 20 kDa in the receptor’s total MW in CRC cells (Fig. 1e). EGFR harbors 11 putative N-glycosylation sites distributed across its extracellular region: N128 in subdomain I (ligand-binding), N175, N196 and N271 in subdomain II (dimerization arm), N352, N361, N413 and N444 in subdomain III (ligand-binding and Cetuximab-binding), and N528, N568 and N603 occurring within subdomain IV [26, 27]. To confirm the addition of ⍺2,6NeuAc to EGFR from SW48 ST6Gal1 OE cells, the EGFR was immunoprecipitated from whole cell lysates after which ⍺2,6NeuAc was assessed by reactivity with SNA. A clear band around 170 kDa, matching the expected MW of the fully glycosylated EGFR, was solely detected in SW48 ST6Gal1 OE samples, confirming that EGFR is a carrier of ⍺2,6NeuAc in ST6Gal1 OE cells (Fig. 4a). A MS-based glycoproteomic analysis was subsequently applied to fully assess the glycosite-specific alterations induced by ST6Gal1 overexpression across the EGFR extracellular domain. Briefly, EGFR immunoprecipitated from SW48 WT, Mock and ST6Gal1 OE 1 and 3 cells was subjected to SDS-PAGE (Fig. 4b). The EGFR band was excised and processed for MS-based protein identification and glycopeptide analysis by liquid chromatography tandem mass spectrometry (LC-MS/MS). EGFR was identified as the major glycosylated protein in all samples. The protein coverage ranged from 62 to 71 % and from 70 to 76 % in the first and second biological replicates, respectively.

Fig. 4.

EGFR site-specific glycan composition in SW48 cells. a Detection of 2,6NeuAc structures in immunoprecipitated EGFR from SW48 WT, Mock and ST6Gal1 OE (OE 1–4) cell clones by reactivity with SNA. b Colloidal Blue gel staining of immunoprecipitated EGFR from SW48 WT, Mock and ST6Gal1 OE cell clones. c Upper panels: extracted ion chromatograms of the N444 glycopeptide modified with a mono-sialylated biantennary N-glycan (peak 1, peptide + GlcNAc(4)Man(3)Gal(1)GalNAc(1)Fuc(1)NeuAc(1), ion at m/z 1414.321, [M + 3 H]³⁺), and the N528 glycopeptide modified with a di-sialylated biantennary N-glycan (peak 1, peptide + GlcNAc(4)Man(3)Gal(2)Fuc(1)NeuAc(2), ion at m/z 1201.799, [M + 3 H]³) in WT, Mock and ST6Gal1 OE EGFR. Bottom panels: isotopic distribution of the N444 and N528 glycopeptides modified with the referred mono- and di-sialylated biantennary N-glycans in ST6Gal1 OE 1 EGFR. d Schematic representation of EGFR glycosylation following site-specific assignment and structural glycan characterization in SW48 WT, Mock and ST6Gal1 OE cell clones

Out of the eleven predicted EGFR N-glycosylation sites, five were assigned and had their glycosylation status characterized in all samples (ESM_2 Table S1; ESM_2 Table S2). Glycosites N175, N444 and N528, besides carrying high-mannose chains, were found to be modified with bi-, tri- and tetra-antennary N-glycan species with terminal galactosylation and fucosylation in WT and Mock samples. In ST6Gal1 OE samples, on the other hand, these three glycosites were found to be decorated by terminal sialylation. Two illustrative examples of this site-specific glycosylation shift are given in Fig. 4c. The glycopeptide 432-QHGQFSLAVVSLNITSLGLR-452 modified with a mono-sialylated bi-antennary N-glycan (GlcNAc(4)Man(3)Gal(1)GalNAc(1)Fuc(1)NeuAc(1), m/z 1414.321, [M + 3 H]³⁺), and the glycopeptide 522-DCVSCRNVSR-531 modified with a di-sialylated bi-antennary N-glycan (GlcNAc(4)Man(3)Gal(2)Fuc(1)NeuAc(2), m/z 1201.799, [M + 3 H]³⁺) were only detected in ST6Gal1 OE samples, and not in WT and Mock samples. The higher energy collision dissociation (HCD) MS/MS spectra of the referred ST6Gal1 OE glycopeptides, where distinct NeuAc-containing oxonium ions could be clearly detected, and the collision-induced dissociation (CID) MS/MS spectra of the referred ST6Gal1 OE glycopeptides, further support the presence of terminal sialylated motifs (ESM_1 Fig. S2a; Fig. 2b). A schematic representation of the receptor’s overall glycosylation profile, for both WT, Mock and ST6Gal1 OE EGFR is depicted in Fig. 4d. Glycosites N352 and N361 were found to carry oligomannosidic N-glycan chains both in WT and Mock samples, and in ST6Gal1 OE samples. Indeed, overexpression of ST6Gal1 is not expected to affect the levels of high-mannose N-glycans. Glycosite 413, carrying terminally fucosylated N-glycan species, was solely mapped and characterized in Mock EGFR samples, since the required protein coverage was not obtained in ST6Gal1 OE samples. For similar reasons, glycosite 603 was only mapped and characterized in ST6Gal1 OE EGFR samples, where it was found to carry sialylated N-glycan chains. One could postulate, however, that the same glycosylation shift observed in N175, N444 and N528 is likely to have occurred in both these glycosites.

Immunoprecipitation of EGFR from Caco-2 WT and ST6Gal1 KO whole cell lysates, followed by SNA reactivity, confirmed the abrogation of ⍺2,6NeuAc in EGFR from all Caco-2 ST6Gal1 KO cell clones (ESM_1 Fig. S3a). Caco-2 cells express significantly lower levels of EGFR compared to SW48 cells (Fig. 1d), explaining the low yield of immunoprecipitated EGFR in this cell line. SDS-PAGE separation and staining of immunoprecipitated EGFR revealed a faint band between 98 and 198 kDa, which was excised for further glycoproteomic analysis (ESM_1 Fig. S3b). Due to limited protein coverage, only glycosylation sites N444, N361 and N352 were identified and characterized in both Caco-2 WT and ST6Gal1 KO EGFR samples (ESM_1 Fig. S3c; ESM_2 Table S3). As observed for SW48 EGFR, glycosites N352 and N361 were found to be solely modified by oligomannosidic N-glycan structures, which remained unaffected following abrogation of ST6Gal1 expression. The glycosite N444 differed in its glycan composition between Caco-2 WT and ST6Gal1 KO EGFR samples, with KO EGFR harboring terminally fucosylated structures, instead of sialylated structures, the latter only being found in Caco-2 WT EGFR samples.

Overexpression of ST6Gal1 protects colorectal cancer cells from Cetuximab cytotoxicity

N-glycosylation is an important determinant in the regulation of EGFR-mediated processes, affecting the activation of EGFR downstream signaling pathways and shaping the resistance of cells to apoptosis [36, 38]. The functional impact of altered ST6Gal1 expression in CRC cell resistance to Cetuximab, a therapeutic anti-EGFR mAb, was further explored using phenotypical assays. Two SW48 ST6Gal1 OE cell clones (OE 1 and OE 3) and three Caco-2 ST6Gal1 KO cell clones (KO 17, KO 33 and KO 42) and their respective controls were cultured in increasing concentrations of Cetuximab. A subsequent metabolic activity assay revealed a higher capacity of SW48 ST6Gal1 OE cell clones to resist Cetuximab treatment and to maintain their metabolic activity, compared to Mock cells (Fig. 5a). On the other hand, Caco-2 WT and ST6Gal1 KO cells showed no alteration in their metabolic activity when treated with Cetuximab (Fig. 5b). In the absence of Cetuximab treatment, no difference was found in the proliferation (ESM_1 Fig. S4a) and apoptosis (ESM_1 Fig. S4b) of SW48 ST6Gal1 OE cell clones compared with Mock cells, as well as in Cetuximab binding to cell surface EGFR (ESM_1 Fig. S4c).

Fig. 5.

Impact of ST6Gal1 overexpression in CRC cell response to Cetuximab. a Metabolic activity following treatment with increasing concentrations of the anti-EGFR mAb Cetuximab, measured by resazurin reduction assay in SW48 WT, Mock and ST6Gal1 overexpression cell clones (OE 1 and OE 3) and b Caco-2 WT and ST6Gal1 KO cell clones (KO 17, KO 33 and KO 42). Hydrogen peroxidase (H₂O₂) was used as positive control and mouse IgG1 as Cetuximab isotope control. c Quantification of Cetuximab-induced cell death by Annexin V/Propidium Iodate apoptosis assay in SW48 WT, Mock and ST6Gal1 OE cell clones treated with 10 µg/ml Cetuximab. d Expression of cell surface EGFR following treatment of SW48 WT, Mock and ST6Gal1 OE cell clones with 10 µg/ml Cetuximab by binding to the anti-EGFR mAbs Cetuximab and e Matuzumab. f Representative Western blots of three independent analyses of total and phosphorylated EGFR (pEGFR Y1086) levels in SW48 WT, Mock and ST6Gal1 OE cell clones treated with 10 µg/ml Cetuximab and stimulated with 10 ng/ml EGF (in serum starvation conditions). Comparisons between multiple groups were made using two-way ANOVA (metabolic assay) and one-way ANOVA, with Mock and WT used as the control groups for the SW48 ST6Gal1 overexpressing model and the Caco-2 ST6Gal1 knock-out model, respectively; **p < 0.001; ***p < 0.001; ****p < 0.0001

Given the resistance of Caco-2 cells to Cetuximab and their unaltered EGFR activation status upon ST6Gal1 abrogation (Fig. 2h), subsequent functional assays were performed solely with SW48 cells. An Annexin V/Propidium Iodide apoptosis assay revealed a significant decrease in Cetuximab-induced cell death in both SW48 ST6Gal1 OE cell clones treated with 10 µg/ml Cetuximab (Fig. 5c). EGFR expression at the cell membrane of Cetuximab-treated SW48 cells was measured using different anti-EGFR mAbs. We found a decrease in EGFR levels at the cell surface of SW48 ST6Gal1 OE cell clones treated with Cetuximab, as demonstrated by a significant decrease in EGFR detection by Cetuximab (Fig. 5d) and Matuzumab (Fig. 5e). The cells were further stimulated with EGF following treatment with Cetuximab (in serum starvation conditions). Subsequent Western blot analysis revealed no major alterations in total EGFR expression in all tested conditions (Fig. 5f). EGF was capable of inducing EGFR activation in all SW48 cells. However, EGF-induced activation was severely diminished when cells were previously treated with Cetuximab. Additionally, we found that Cetuximab treatment decreased endogenous activation of EGFR in SW48 cells. Interestingly, a decrease in endogenous activation of EGFR was observed in SW48 ST6Gal1 OE cell clones compared to SW48 WT and Mock cells, even in the absence of Cetuximab treatment, highlighting the effect of ⍺2,6-sialylation on the activation of EGFR.

Discussion

CRC is a solid tumor with a poor clinical outcome worldwide. Metastatic CRC can be therapeutically targeted using mAbs directed against the EGFR (e.g. Cetuximab), a critical receptor tyrosine kinase (RTK) in CRC. Despite evident clinical benefits, several elusive drug resistance mechanisms still hamper the ultimate therapeutic efficacy of these mAbs. Thus, unveiling the molecular mechanisms that underlie resistance to therapeutic antibodies, such as anti-EGFR mAbs, is warranted. Alterations in RTK glycosylation have been shown to disrupt proliferative and anti-apoptotic cellular programs, thereby promoting cancer progression and drug resistance [39]. In fact, EGFR is known to be heavily N-glycosylated and several studies have shown the importance of N-glycosylation on the activity of the receptor, affecting the successful implementation of EGFR-targeted therapies [36, 38, 40].

In this study, we addressed the role of ST6Gal1-mediated ⍺2,6-sialylation in the response of CRC cells to therapy with the anti-EGFR mAb Cetuximab. Using CRC cell lines without RAS mutations, we mimicked the genetic mutational status of CRC patients eligible for Cetuximab treatment. By modulating ST6Gal1 expression, we identified cancer-associated glycosylation structures in specific sites of the EGFR, which may lead to the development of novel therapeutic strategies that take these features into consideration. In fact, EGFR glycosite N444, located at the Cetuximab binding domain, was found to exhibit upregulation of sialylated structures. In addition, we found that cellular hyper-sialylation mediated by ST6Gal1 overexpression protected CRC cells from the cytotoxic effects induced by Cetuximab.

Indeed, ST6Gal1 is frequently upregulated in many types of cancer, including CRC, where its overexpression correlates with a poor patient prognosis [18]. There is growing evidence indicating that ST6Gal1 may act as a malignant cell phenotype regulator [41], due its protective role against numerous assaults, including chemotherapy [42], radiotherapy [43], serum deprivation [44] and hypoxia [45]. Moreover, ST6Gal1 has been found to promote drug resistance by regulating stem cell transcriptional factors, granting malignant cells a cancer stem cell-like phenotype [46]. Interestingly, ST6Gal1 expression in normal colon epithelium is restricted to the base of the crypts, where the colon stem cell niche resides [47]. Several glycoproteins relevant for malignant transformation have been described as targets for ST6Gal1-mediated ⍺2,6-sialylation. Besides EGFR, ST6Gal1 is capable of targeting β1-integrin, promoting the motility and invasion of CRC cells [48]. Sialylation of tumor necrosis factor receptor 1 (TNFR1) and Fas death receptor precludes receptor internalization and inhibits apoptotic downstream signaling pathways [49]. Recently, ST6Gal1-mediated sialylation was also found to be present in exomeres [50], a type of extracellular vesicle recently identified [51]. ST6Gal1 from exomeres was shown to sialylate target proteins at recipient cells, including β1-integrin. Interestingly, exomeres containing a ligand for EGFR were found to be capable of inducing receptor activation in recipient cells and to promote tumor growth, supporting the importance of both ST6Gal1 and EGFR signaling in different aspects of cancer progression [50].

Several studies have reported the impact of altered ST6Gal1 expression on EGFR function and activation. We found that ST6Gal1 overexpression leads to a decrease in cell surface EGFR expression following treatment with Cetuximab. Moreover, lower levels of EGFR activation were found in CRC cells with ST6Gal1 overexpression, even in the absence of the therapeutic antibody. Previous studies have reported that EGFR ⍺2,6-sialylation enhances EGFR activation in ovarian and pancreatic cancer cells treated with the RTK inhibitor Gefitinib, protecting cells from its cytotoxic effects [44]. This RTK inhibitor has also been found to promote downregulation of EGFR activation in CRC cells overexpressing ST6Gal1 [31], which corroborates our findings in CRC cells targeted with Cetuximab. Nevertheless, the role of ST6Gal1 in EGFR activation status, particularly upregulation of the sialyltransferase, has been associated with drug resistance phenotypes. These effects of ST6Gal1 on EGFR activation hint at a complex interplay between sialylated receptors involved in drug resistance signaling pathways. Additionally, our findings show that EGFR location at the cell membrane is important in a therapeutic setting. This is in line with previous studies showing that depletion of lipid rafts sensitizes cancer cells to therapy with EGFR RTK inhibitors [52]. Moreover, interaction between EGFR and the galectin lattice, present at the extracellular domain of the plasma membrane, has been shown to be weakened when hypersialylation is induced in cancer cells, leading to increased EGFR internalization [53].

Overall, our study shows the impact of altered ST6Gal1 expression in site-specific glycosylation of the oncogenic protein EGFR and its implications for targeted cancer treatment. Our results highlight the key role of glycosylation in tumor cell biology and suggest a novel mechanism of resistance in EGFR-specific monoclonal antibody therapy in CRC.

Abbreviations

⍺2,3NeuAc

⍺2,3-linked sialic acid

⍺2,6NeuAc

⍺2,6-linked sialic acid

AAL

Aleuria aurantia lectin

BrdU

bromodeoxyuridine

CRC

colorectal cancer

CRISPR/Cas9

Clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9

ECL

enhanced chemiluminescence

EGFR

epidermal growth factor receptor

FBS

fetal bovine serum

gRNA

guide RNA

IDAA

Indel Detection by Amplicon Analysis

Indel

insertions and deletions

KO

Knock-Out

LTL

Lotus tetragonolobus lectin

mAb

monoclonal antibody

MAL-I

Maackia amurensis lectin I

MFI

median fluorescence intensity

OE

Overexpression

ON

overnight

RT

room temperature

RTK

receptor tyrosine kinase

SNA

Sambucus nigra agglutinin

ST6Gal1

β-Galactoside ⍺2,6-sialyltransferase 1

TIDE

Tracking of Indels by DEcomposition

TNFR1

tumor necrosis receptor 1

UEA-I

Ulex europaeus agglutinin I

WT

wild-type

Author contributions

J.G. and C.A.R. designed and supervised the study. J.G.R., H.O.D., C.G., M.B., A.M., A.H.d.R. and J.G performed experiments; J.G.R., H.O.D., C.G., P.J.H., J.L., A.A., P.A.V., M.W., J.G. and C.A.R performed the formal analyses; J.G.R., H.O.D., J.G. and C.A.R wrote the original manuscript. All authors reviewed and edited the manuscript. All authors have read and agreed to the final version of the manuscript.

Funding

This research was funded by Merck KGaA, Darmstadt, Germany, and by FEDER funds through the Operational Programme for Competitiveness Factors COMPETE 2020 (POCI-01-0145-FEDER-016585; POCI-01-0145-FEDER-007274) and national funds through the Foundation for Science and Technology (FCT), under the projects: PTDC/BBB-EBI/0567/2014 to C.A.R and UID/BIM/04293/2013; PTDC/MED-QUI/29,780/2017 to C.G., and the project NORTE-01-0145-FEDER-000029, supported by Norte Portugal Regional Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF). J.G.R. was supported by a FCT PhD grant (SFRH/BD/136,736/2018); H.O.D. was supported by a FCT PhD grant (PD/BD/128,407/2017) through the FCT PhD Programmes and by Programa Operacional Potencial Humano (POPH), specifically by the BiotechHealth Programme (Doctoral Programme on Cellular and Molecular Biotechnology Applied to Health Sciences), with the reference PD/0016/2012 funded by FCT. M.B. was supported by the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement n.º 748,880.

Data availability

The mass spectrometry proteomic data supporting the conclusions of this article are available in the ProteomeXchange Consortium via the PRIDE partner repository, under the project name “EGFR N-Glycosylation Profile from Colorectal Cancer Cells”, with the dataset identifier PXD017914.

Code availability

Not applicable.

Declarations

Conflict of interest

The authors declare that they have no conflict of interest. 

Ethics approval

Not applicable. 

Consent to participate

Not applicable. 

Consent for publication

Not applicable.