Antibody dependent cellular cytotoxicity-inducing anti-EGFR antibodies as effective therapeutic option for cutaneous melanoma resistant to BRAF inhibitors

Introduction About 50% of cutaneous melanoma (CM) patients present activating BRAF mutations that can be effectively targeted by BRAF inhibitors (BRAFi). However, 20% of CM patients exhibit intrinsic drug resistance to BRAFi, while most of the others develop adaptive resistance over time. The mechanisms involved in BRAFi resistance are disparate and globally seem to rewire the cellular signaling profile by up-regulating different receptor tyrosine kinases (RTKs), such as the epidermal growth factor receptor (EGFR). RTKs inhibitors have not clearly demonstrated anti-tumor activity in BRAFi resistant models. To overcome this issue, we wondered whether the shared up-regulated RTK phenotype associated with BRAFi resistance could be exploited by using immune weapons as the antibody-dependent cell cytotoxicity (ADCC)-mediated effect of anti-RTKs antibodies, and kill tumor cells independently from the mechanistic roots. Methods and results By using an in vitro model of BRAFi resistance, we detected increased membrane expression of EGFR, both at mRNA and protein level in 4 out of 9 BRAFi-resistant (VR) CM cultures as compared to their parental sensitive cells. Increased EGFR phosphorylation and AKT activation were observed in the VR CM cultures. EGFR signaling appeared dispensable for maintaining resistance, since small molecule-, antibody- and CRISPR-targeting of EGFR did not restore sensitivity of VR cells to BRAFi. Importantly, immune-targeting of EGFR by the anti-EGFR antibody cetuximab efficiently and specifically killed EGFR-expressing VR CM cells, both in vitro and in humanized mouse models in vivo, triggering ADCC by healthy donors’ and patients’ peripheral blood cells. Conclusion Our data demonstrate the efficacy of immune targeting of RTKs expressed by CM relapsing on BRAFi, providing the proof-of-concept supporting the assessment of anti-RTK antibodies in combination therapies in this setting. This strategy might be expected to concomitantly trigger the crosstalk of adaptive immune response leading to a complementing T cell immune rejection of tumors.


BRAF and NRAS status
The mutational status of BRAF V600 (exon 15) and NRAS Q61 (exon 3) was determined in CM cell lines with Sanger sequencing as previously described (1,2).Primers sets used are listed in Supplemental Table 3.

Colony forming assay
Cells were seeded in 6-well plates at a density of 30.000 cells/well.After 24h, cells were added with DMSO (negative control), 2.5 µM gefitinib, 5 µM gefitinib, or 20 µg/ml cetuximab, with or without 4 µM PLX-4032.Complete medium with drugs was substituted every 3-4 days, and colonies were evaluated after 2 weeks from plating.Colony imaging was performed by a Chemidoc XRS+ imaging system and QuantityOne software (Bio Rad) after a 20-minute staining with a 0.05% crystal violet, 1% formaldehyde, 1% methanol solution.

Immunohistochemistry analysis of EGFR and AXL expression on melanoma tissues
Immunohistochemistry was performed as described (3) on formalin-fixed, paraffin-embedded (FFPE) tissues from pre-or post-therapy lesions surgically removed from advanced melanoma patients treated with the BRAFV600E inhibitor Vemurafenib.Sections were stained with antibodies to AXL (AF154, R&D Systems) and to EGFR (4267, Cell Signaling).Sections were evaluated by a semi quantitative scoring system (4).Each antigen was assessed for expression on neoplastic cells, or on intra-tumor stromal cells, or on extra-tumor stromal cells.For each marker the expression (E) was ranked from 1 to 4 as follows: 1: immunoreactivity in up to 25% cells; 2: expression in 26-50% cells; 3 expression in 51-75% cells: 4 expression in >76-100% cells.Immunostaining intensity (I) was ranked as low (1; fainter than internal controls), normal (2; as faint as controls), or strong (3; more intense than controls).E and I were combined into a single score (S), calculated as E×I.

Flow cytometry analysis
Flow cytometry analyses were performed essentially as previously described (5).AXL was detected by the Alexa-fluor488-conjugated anti-AXL Monoclonal Mouse IgG1(R&D sistems); a Mouse IgG1/FITC, Clone DAK-GO1 isotypic antibody (DAKO) served as negative control.Antibodies were used following the manufacturer's instructions.Data acquisition was performed with a FACSCanto II flow cytometer (Becton Dickinson), and analyzed with Diva 5.0 (Becton Dickinson) and FlowJo software (Tree Star, Inc).

Dose-response curves
CM cells were seeded in flat bottom 96-well plates at a density of 2500 cells/well.After 24h, scalar doses of PLX-4032, or an equal volume of DMSO, used as negative control, were added into triplicate wells.Cell viability was evaluated 72h after the addition of PLX-4032 by a standard MTT assay (Life Technologies).Generation of sigmoidal dose-response curves using a four-parameter nonlinear regression model, and calculation of PLX-4032 IC 50 values were achieved by the Prism 6.0 software (GraphPad Software).

Immunohistochemical expression of xenograft tumors
Mice bearing Mel 767 parental and/or Mel 767 VR tumors were euthanized at endpoint via CO 2 asphyxiation.Tumors were harvested and fixed in 4% paraformaldehyde overnight.Fixed tissues were dehydrated by submerging in increasing concentrations of ethanol (70-100%), followed by clearing with xylene, and finally infiltrated using Tissue-Tek VIP 6 automated tissue processor prior to embedding.Paraffinized tissues were then embedded in paraffin wax for microtome sectioning.Tissue blocks were sectioned at 4 µm thickness and mounted onto Superfrost® Plus glass microscope slides (Menzel Gläser).Sectioned tumour samples were de-paraffinized followed by enzymatic antigen retrieval and endogenous peroxidase activity blocked by hydrogen peroxide.Sections were then incubated with primary antibodies, mouse-monoclonal anti-human EGFR (31G7, Abcam, 1:100) 36 o C for 1 hour, followed by incubation with the appropriate horse radish peroxidase (HRP)-conjugated secondary antibodies at room temperature for 1 hour.Staining was then visualised with 3,3'-diaminutesobenzidine (DAB) and haematoxylin counterstain via the Ventana anti-HQ HRP automated detection system.Slides were imaged using VS120 slide scanners (Olympus) at 40x magnification.

Analysis of TCGA Skin Cutaneous Melanoma (SKCM) cohort
Analysis of TCGA SKCM cohort (dataset TCGA Firehoose Legacy) was performed using https://www.cbioportal.org/and https://portal.gdc.cancer.gov/web portals.Only samples with BRAF V600 mutations were considered for the further analysis.Considering the ratio of the expression values of MITF/AXL genes (7), the sampled from the TCGA-SKCM cohort were classified as resistant (R, 32 Samples) or sensitive (S, 93 Samples) .The cut-off of low MITF/AXL ratio was set considering the first quartile of data distribution.The expression values of transcripts for RTK, RTK ligands, and transcription factors were obtained from normalized values downloaded from the web portals.Statistical differences between groups were evaluated by Student's t test, pvalues ≤0.05 were considered significant.

Pathway analysis
EGFR-related pathways analysis was conducted with Ingenuity Pathway Analysis software (winter release, December 2023) on genes differentially expressed between the 2 conditions of interest, namely, genes differentially expressed between R vs S in the TCGA SKCM, or between VR and P cells.

Statistical analysis
For RNA-Seq analysis DESeq2 was used for differential expression analysis.Wald-test was employed for hypothesis testing when comparing two groups.A BH adjusted pvalue=0.05was set.Analysis for significance was performed by parametric Student's t-test.At least three independent experiments were taken in consideration for each experiment.Values were expressed as mean ± SD.The p-values <0.05 were considered significant.