MET Enhances Amivantamab Binding to EGFR and Antibody-Dependent Cellular Toxicity.

1
Introduction
Amivantamab is a bispecific antibody against epidermal growth factor receptor (EGFR) and MET that has been approved by the Food and Drug Administration (FDA) for nonsmall cell lung cancer (NSCLC) with EGFR exon 20 insertion mutations, as well as NSCLC with common EGFR mutations, such as exon 19 deletion and L858R. Amivantamab in combination with cytotoxic anticancer drugs prolongs progression‐free survival compared with cytotoxic anticancer drugs alone in untreated advanced NSCLC with EGFR exon 20 insertion mutations [1]. Furthermore, in NSCLC with common EGFR mutations, the combination of amivantamab and lazertinib, a third‐generation EGFR‐tyrosine kinase inhibitor (TKI), significantly extends overall survival compared with the third‐generation EGFR‐TKI osimertinib [2, 3]. Therefore, amivantamab has become a key drug for the treatment of EGFR‐mutated NSCLC. Moreover, several clinical trials are ongoing for EGFR‐TKI‐resistant common EGFR mutation‐positive NSCLC and NSCLC with high MET expression [4, 5], and the indications for amivantamab may be expanded.
Amivantamab has several mechanisms of action [6]. Amivantamab binds EGFR and MET, thereby inhibiting the binding of their respective ligands (epidermal growth factor [EGF], transforming growth factor [TGF]‐α, and amphiregulin for EGFR and hepatocyte growth factor [HGF] for MET) to the receptor. Amivantamab‐bound EGFR and MET are internalized into the cytoplasm and degraded, thereby inhibiting EGFR and MET signaling. In addition, monocytes and macrophages can induce trogocytosis and ingest cell membranes via interaction with the Fc portion of amivantamab that binds to cancer cells via EGFR or MET. Similarly, amivantamab can bind immune cells, including Natural Killer (NK) cells, and induce antibody‐dependent cellular cytotoxicity (ADCC) in cancer cells. A previous report showed that the ADCC activity of amivantamab is mainly induced by binding the anti‐EGFR arm [7]. However, how the binding of the anti‐MET arm to MET enhances the antitumor effects of amivantamab has not been fully elucidated.
In this study, we investigated the role of MET in the binding of amivantamab to EGFR and induction of ADCC. We examined the binding of amivantamab to the EGFR‐ and/or MET‐extracellular domains (ECDs) using high‐speed atomic force microscopy (HS‐AFM) to observe the interactions of single or multiple proteins in real time under native solution conditions [8]. This was the first successful visualization and analysis of the trimers formed by amivantamab, EGFR‐ECD, and MET‐ECD. Furthermore, we clarified the involvement of MET in the amivantamab‐induced ADCC in NSCLC cells with common EGFR mutations using MET‐knockout cells.

2
Materials and Methods
2.1
High‐Speed Atomic Force Microscopy (HS‐AFM)
HS‐AFM was performed, as previously described [9], using USC‐F1.2‐k0.15 probes (NanoWorld). Amivantamab, amivantamab/EGFR‐ECD, amivantamab/MET‐ECD, and amivantamab/EGFR‐ECD/MET‐ECD complexes were visualized on bare mica surfaces in 1 mM Phosphate‐buffered saline (PBS). Amivantamab was purchased from Janssen Pharmaceuticals (Tokyo, Japan). His‐tagged recombinant human EGFR and recombinant human MET were purchased from Bio‐Techne R&D Systems (#11302‐ER) and Sino Biological (#10692‐H08H), respectively. For amivantamab observation, 5 μg/mL of amivantamab was incubated on a mica surface for 10 min, followed by rinsing with 10 mM PBS. For amivantamab–receptor complexes, 5 μg/mL of EGFR or MET was incubated on amivantamab‐adsorbed mica for 10 min and then rinsed with 10 mM PBS. The HS‐AFM images were mean‐filtered with a 0.7‐ or 1‐pixel radius in Fiji [10], to reduce noise.

2.2
Prediction of Amivantamab and Amivantamab–Receptor Structures
The amivantamab structure was predicted using AlphaFold 3 and the amino acid sequence from GenomeNet (Entry ID: D11894; Figure S1). The amivantamab/MET‐ECD complex was modeled based on the interaction of the MET Sema domain with the heavy chain II/light chain II in the Fab fragment of amivantamab (Protein Data Bank [PDB] ID: 6WVZ) [11]. The PDB structure of the Sema domain bound to heavy chain II/light chain II and the AlphaFold 3‐predicted structures of amivantamab and MET‐ECD were aligned using PyMOL (PyMOL Molecular Graphics System, version 2.4.1; Schrödinger LLC). The interaction of EGFR‐ECD with the heavy chain I/light chain I of amivantamab was predicted using AlphaFold 3, based on the EGFR–cetuximab interaction (PDB ID: 1YY9) [12]. Pairwise sequence alignment (PSA) using EMBOSS Needle demonstrated 92.1% and 91.8% sequence similarities between the cetuximab heavy chain and heavy chains I and II of amivantamab, respectively, and 87.9% and 88.3% between the cetuximab light chain and light chains I and II of amivantamab, respectively. The AlphaFold 3‐predicted structure of the heavy chain I/light chain I complex with EGFR‐ECD was aligned with the structure of amivantamab to construct the amivantamab/EGFR‐ECD complex.

2.3
Flexible Fitting for HS‐AFM Imaging of Amivantamab
Flexible fitting was applied to the AlphaFold 3‐predicted static molecular structure of amivantamab to infer atomistic‐precision models of structural dynamics from resolution‐limited HS‐AFM topographic imaging data. We employed normal mode flexible fitting AFM (NMFF‐AFM) [13] using the BioAFMviewer interface [14]. Raw HS‐AFM topographic images were preprocessed using automatic tilt correction of the AFM stage and a Gaussian filter with a standard deviation of 1 nm. AFM images of the atomistic amivantamab structure were simulated using a probe sphere radius (R) of 3 nm and a cone half angle (ϑ) of 20°. Flexible fitting was performed with the following parameters: elastic network cutoff distance = 8 Å, number of normal modes = 12, root mean square deviation (RMSD) of structural deformation = 0.5 Å, number of iterations = 100–200, and image similarity scoring = image correlation coefficient.

2.4
Gel Filtration Analysis of the Binding of Amivantamab to EGFR and/or MET
The binding of amivantamab (148 kDa), EGFR‐ECD (100 kDa), and MET‐ECD (120 kDa) was examined using gel filtration. The three substances (1000 nM each) were added to PBS containing 0.01% Tween 20. After 30 min, the mixture was filtered through Superdex 200, and 300 μL fractions were collected. The EGFR and MET levels in the fractions were detected using western blotting. Because EGFR was His‐tagged, an anti‐His‐tag antibody (#12698; Cell Signaling Technology) was used as the primary antibody. The human HGFR/c‐MET antibody (#MAB3582; R&D Systems), which recognizes the ECD, was used as the anti‐MET antibody. Detection was performed using the respective secondary antibodies, as previously reported [15]. The band intensities were examined using densitometry.

2.5
Cellular and Biochemical Assays
Human lung cancer cell lines (Table S1) were used to evaluate the effects of these compounds. Cell surface protein expression was analyzed using flow cytometry, cell viability was determined using a cell viability assay, and protein expression was confirmed using western blotting. ADCC was measured using a Lactate Dehydrogenase (LDH) release assay with peripheral blood mononuclear cells (MNCs) as effector cells. Detailed protocols for these assays are provided in the Supporting Information and Methods (Data S1).

3
Results
3.1
HS‐AFM Observations of Amivantamab Bound to EGFR and/or MET
First, the structures of amivantamab and amivantamab bound to EGFR and/or MET were predicted using AlphaFold 3 (Figure 1A–D). Based on previous studies [12] and AlphaFold 3 predictions, domain III of EGFR interacted with the Fab heavy chain I/light chain I of amivantamab (Figure 1B), and the Sema domain of MET bound to the Fab heavy chain II/light chain II of amivantamab (Figure 1C). The HS‐AFM observations of amivantamab revealed three domains, which were assumed to be one Fc domain and two Fab domains (Figure 1E; Movie S1). Time‐lapse images showed variations in the topographic maps of amivantamab, indicating the intrinsic flexibility of the protein. Normal mode analysis of the predicted structure suggested that domain rearrangements underlie this flexibility (Movie S2). The topographic maps of amivantamab were validated using flexible fitting (Figure 1F; Figure S2), which demonstrated significant relative changes between the Fab and Fc domains and between both Fab domains.
EGFR‐ and MET‐ECDs were also observed using HS‐AFM (Figure S3). The EGFR‐ECD structure matched the closed (tethered) conformation of the EGFR‐ECD, composed of domains I–IV (Figure S3A,C). The MET‐ECD had a globular Sema domain connected to a thin PSI‐IPT domain (Figure S3B,D), as previously reported [9].
Next, amivantamab and the EGFR‐ECD were incubated on mica surfaces at room temperature for 30 min for HS‐AFM observation of the amivantamab/EGFR complexes. Amivantamab reversibly bound to the predicted EGFR domain III through its Fab domain (Figure 2A,B; Movie S3), with the bound state persisting for 85% of the observation period (Figure 2C, Figure S4). The reversible binding of amivantamab to the EGFR‐ECD was consistent across multiple examples of the amivantamab/EGFR complex (Figure 2C, subwindows; Figure S4). Analysis of the time intervals between the binding and unbinding events showed that the bound state persisted longer than the unbound state.
Complexes of amivantamab with the MET‐ECD were also observed using HS‐AFM. Amivantamab reversibly bound to the Sema domain of MET through its Fab domain (Figure 3A,B; Movie S4). Amivantamab remained bound to MET for 87% of the observation period (Figure 3C, Figure S5), with longer intervals in the bound state that resembled the behavior of the amivantamab/EGFR complex. However, a comparison of the dynamics between the amivantamab/EGFR and amivantamab/MET complexes indicated that unbinding events occurred less frequently in the amivantamab/MET complex. Additional examples of the amivantamab‐MET complex exhibited stability with no dissociation (Figure 3C, subwindows; Figure S5), suggesting that amivantamab interacted more strongly with MET than with EGFR.
Furthermore, amivantamab was mixed with the EGFR‐ and MET‐ECDs and observed using HS‐AFM. Many molecules were tightly bound in the form of heterooligomers. Extended observation revealed that amivantamab bound to molecules presumed to be the ECDs of EGFR and MET, forming trimers (Figure 4A; Movie S5). Initially, the trimer (Figure 4A) was highly compact, which obscured the distinction between the receptors and amivantamab. At 9.7 s, both receptors dissociated from amivantamab, confirming that the initial complex was a trimer. These receptors reversibly bound to amivantamab. After 550 s, amivantamab interacted with both receptors and formed structures similar to those predicted for the amivantamab/EGFR/MET complex (Figure 4B).
Analysis of the binding and unbinding events (Figure 4C) revealed that amivantamab was bound to only one receptor for 22% of the observation period and was bound to both receptors for 78% of the time (Figure S6). Only a negligible portion of the observation period involved the complete dissociation of both receptors. During the observation period (< 550 s), when the complex was unstable, the time interval for the one‐receptor‐bound state was longer than that for the two‐receptor‐bound state. After 550 s, the complex contained both EGFR and MET and remained stable. Additional examples of the amivantamab/EGFR/MET complex were stable throughout the observation period (Figure 4C, subwindows; Figure S6). These results indicated that the amivantamab/EGFR/MET complex predominantly remained associated.

3.2
Enhancement of Amivantamab‐EGFR Binding by MET
To further assess the effect of MET on the binding of amivantamab to EGFR, we combined 1000 nM amivantamab (148 kDa), EGFR‐ECD (100 kDa), and MET‐ECD (120 kDa) and performed gel filtration. MET and EGFR in the fractions were detected using western blotting. In the fractions obtained by mixing amivantamab and MET‐ECD, MET was strongly detected in fractions 5–8 (presumably amivantamab‐bound MET) and weakly detected in fractions 10–12 (presumably free MET) (Figure 5A). Densitometric analysis of the bands showed that the complex‐to‐free monomer ratio of the MET‐ECD was 1:0.66 (Figure 5B). In contrast, in the fractions obtained by mixing amivantamab with the EGFR‐ and MET‐ECDs, MET was detected more strongly in fractions 4–7 (presumably MET in a ternary complex with amivantamab and EGFR) and more weakly in fractions 10–12 (presumably free MET) (Figure 5A). Densitometric analysis of the bands showed that the complex‐to‐free monomer ratio of the MET‐ECD was 1:0.27 (Figure 5B), indicating a decrease in free MET‐ECD.
In parallel, EGFR was detected using an anti‐His‐Tag antibody because the EGFR‐ECD was His‐tagged. In the fractions obtained by mixing amivantamab and EGFR, EGFR was strongly detected in fractions 7–11 (presumably amivantamab‐bound EGFR) and weakly detected in fractions 14 and 15 (presumably free EGFR) (Figure 5C). Densitometric analysis of the bands showed that the complex‐to‐free monomer ratio was 1:0.68 (Figure 5D). In contrast, in the fractions obtained by mixing amivantamab with EGFR and MET, EGFR was detected strongly in fractions 4–7 (presumably EGFR in a ternary complex with amivantamab and MET) and weakly in fractions 14 and 15 (presumably free EGFR) (Figure 5C). Densitometric analysis of the bands showed that the complex‐to‐free monomer ratio was 1:0.24 (Figure 5D), indicating a decrease in the free EGFR‐ECD.
These results indicated that the addition of MET to amivantamab and EGFR reduced free EGFR monomers, suggesting that MET enhanced and stabilized the binding of amivantamab to EGFR. Thus, amivantamab bound more strongly to EGFR and MET when they coexisted.

3.3
Amivantamab‐Induced ADCC in Human NSCLC Cell Lines With Common EGFR Mutations
Flow cytometry and western blot analyses using anti‐EGFR and anti‐MET antibodies showed that NSCLC cell lines with common EGFR mutations, such as exon 19 deletion and L858R, expressed varying levels of EGFR and MET (Figure 6A,B). Further western blot analysis using an anti‐EGFR del19‐specific antibody (recognizing del E746‐A750) detected bands in PC‐9, HCC827, H1650, and H2279 cells with EGFR exon 19 deletions (Figure 6B). H2935 cells had a unique EGFR exon 19 deletion mutation (E746–T751del, S752I); therefore, no bands were detected using the anti‐EGFR del19‐specific antibody, which was not unexpected. In contrast, EGFR in H1975 and HCC4011 cells was detected with the anti‐EGFR‐L858R‐specific antibody.
These cell lines showed different sensitivities to the EGFR‐TKI, osimertinib (Figure 6C). All cell lines showed low sensitivity to amivantamab alone, with < 25% cell viability suppression (Figure 6D). Next, the amivantamab‐induced ADCC was examined. In preliminary experiments with PC‐9 cells and human MNCs as target and effector cells, respectively, we observed ADCC activity in an effector: target [E:T] ratio‐dependent manner (Figure S7). At an E:T ratio of 20, ADCC was induced in an amivantamab concentration‐dependent manner by MNCs from all three donors (Figure S8). Amivantamab induced different levels of ADCC in each EGFR‐mutated NSCLC cell line (Figure 6E, Figure S9a). The ADCC sensitivity was positively correlated with EGFR expression but not with MET expression in EGFR‐mutated NSCLC cells (Figure 6F, Figure S9b).
To investigate whether MET expression in target cells affected the ADCC activity of amivantamab, we established MET knockout (MET‐KO) cells using CRISPR–Cas9 methods with PC‐9 and HCC827 cells. PC‐9 and HCC827 MET‐KO cells expressed EGFR at the same level as their parent lines; however, MET expression was lost (Figure 7A–C). Relatively high concentrations of amivantamab (≥ 10 ng/mL) induced approximately the same ADCC activity in both parent and MET‐KO lines. However, low concentrations of amivantamab (approximately 1 ng/mL) reduced the ADCC activity in the MET‐KO line (Figure 7D,E). These results indicated that MET enhanced the ADCC activity in the presence of low amivantamab concentrations.

4
Discussion
Amivantamab was selected from the EGFR‐MET bispecific antibodies consisting of eight anti‐EGFR antibodies and five anti‐MET antibodies [11]. It binds with high affinity (ED50 < 1 μg/mL) to the human NSCLC cell line A549, which expresses wild‐type EGFR, and does not induce MET phosphorylation. The anti‐EGFR arm of amivantamab binds to EGFR domain III at the same binding site as zalutumumab. The binding site for EGFR partially overlaps with that for cetuximab and inhibits the binding of EGFR to EGF. In contrast, the anti‐MET arm binds to the MET Sema domain in all complementary‐determining regions (CDRs), except for CDR‐H1. This overlaps with the binding site of MET on the HGF‐β chain and inhibits the binding of MET to HGF. In this study, we directly visualized the morphology of amivantamab alone, its heterodimers with the EGFR‐ or MET‐ECD, and its heterotrimers with the EGFR‐ and MET‐ECD, which were consistent with previous findings [11]. Our HS‐AFM results further suggested that amivantamab interacted more strongly with MET than with EGFR. These results are consistent with those of a previous study showing that amivantamab has approximately the same binding rate constant (K
on) for EGFR and MET (KonEGFR: 0.028 nM/min and KonMET: 0.040 nM/min) but a lower dissociation rate constant (K
off) for MET than for EGFR (KoffEGFR: 0.039 nM/min and KoffMET: 0.0028 nM/min) [16]. Importantly, the trimeric complex of amivantamab with EGFR‐ECD and MET‐ECD was substantially more stable than either dimeric complex: complete dissociation of both receptors was rarely observed in the trimer, whereas full receptor dissociation occurred during 13%–15% of the observation period for the dimeric complexes. Gel filtration analysis supported these findings, showing a marked reduction in free EGFR and MET monomers under trimeric conditions compared with dimeric mixtures. Taken together, these dynamic and equilibrium data demonstrate that amivantamab forms a highly stable trimeric complex, with MET playing a key role in stabilizing the cooperative interactions, which likely contributes to maintaining amivantamab's efficacy even at the low antibody concentrations expected in the tumor microenvironment.
Amivantamab is an antibody produced by an engineered Chinese hamster ovary (CHO) cell line defective in protein fucosylation and thus has low Fc fucosylation [17, 18]. Consequently, its binding affinity to FcγRIIIa and ADCC activity are enhanced [7, 19]. The ADCC activity of amivantamab is mainly induced by its EGFR‐binding arm [17, 20]. In line with these findings, we demonstrated that EGFR expression was correlated with ADCC sensitivity in NSCLC cells with common EGFR mutations. Interestingly, MET enhanced ADCC in the presence of low amivantamab concentrations. We speculate that in the presence of low amivantamab concentrations, MET may enhance the binding of amivantamab to EGFR, thereby efficiently inducing ADCC mediated by the anti‐EGFR arm. In contrast, in the presence of high amivantamab concentrations, sufficient amivantamab may bind to EGFR; therefore, the maximum ADCC activity may be induced regardless of the presence or absence of MET expression. Since only limited concentrations of antibodies are assumed to reach tumor cells in vivo, owing to the influence of the tumor microenvironment, the MET‐binding property of amivantamab may lead to superior clinical efficacy compared to anti‐EGFR bivalent antibodies.
Amivantamab has been approved for use in combination with lazertinib, a third‐generation EGFR‐TKI, in patients with untreated NSCLC with common EGFR mutations [2]. In addition, amivantamab in combination with cytotoxic anticancer drugs is effective in patients with NSCLC with common EGFR mutations who develop resistance to EGFR‐TKIs [21]. In this study, we used various NSCLC cell lines with common EGFR mutations, and their sensitivities to osimertinib varied among cell lines. Interestingly, amivantamab induced ADCC even in cell lines with a low sensitivity to osimertinib. These results suggest that ADCC is the predominant mechanism of action of amivantamab in patients with NSCLC with common EGFR mutations. In the future, strategies to enhance ADCC may be important to augment the antitumor effects of amivantamab.
To date, no bivalent anti‐EGFR antibodies have been approved for the treatment of NSCLC with common EGFR mutations; however, the bispecific antibody amivantamab has been approved because of its clinical efficacy. This study clarified the role of the MET‐binding arm of amivantamab in binding to EGFR and MNC‐mediated ADCC and demonstrated that MET promotes the binding of amivantamab to EGFR and enhances MNC‐mediated ADCC at low concentrations in NSCLC with common EGFR mutations. These findings may inspire future studies to enhance and optimize the therapeutic effects of amivantamab.

Author Contributions

Shigeki Sato: conceptualization, methodology, writing – original draft, writing – review and editing. Neval Yilmaz: investigation, methodology, writing – original draft, writing – review and editing. Sachiko Arai: investigation. Katsuya Sakai: conceptualization, investigation, methodology, writing – original draft, writing – review and editing. Hiroki Sato: methodology. Yuya Murase: investigation. Tsukasa Ueda: investigation. Hayato Koba: investigation. Shigeki Nanjo: methodology. Yuichi Tambo: methodology. Hiroshi Kotani: investigation, methodology. Koji Fukuda: methodology. Hiroaki Taniguchi: investigation. Romain Amyot: formal analysis, investigation, validation. Holger Flechsig: formal analysis, investigation, validation. Hideko Isozaki: investigation. Kunio Matsumoto: conceptualization, funding acquisition. Seiji Yano: conceptualization, funding acquisition, writing – original draft, writing – review and editing.

Ethics Statement
All experimental procedures of this study were reviewed and approved by the Ethics Committee of Kanazawa University Hospital (114857). In this study, informed consent has been obtained from them or their guardians, and written informed consent has been signed.

Conflicts of Interest
Seiji Yano received research grants from Janssen Pharma, Chugai, Eli‐Lilly, Boehringer‐Ingelheim, MSD, BlissBio, and Eisai, and speaking honoraria from Janssen Pharma, Chugai, AstraZeneca, Eli‐Lilly, Takeda, Boehringer‐Ingelheim, Pfizer, Amgen, Ono, Taiho, MSD, BMS, Daiichi Sankyo, and Nihon Kayaku. Seiji Yano is also an editorial board member of Cancer Science. All other authors declare no competing interests.

Supporting information

Figure S1: Amino acid sequence of amivantamab. The amino acid sequence of amivantamab is available from GenomeNet (Entry ID: D11894).

Figure S2: Flexible fitting of the amivantamab structure. (A) HS‐AFM images of amivantamab (Figure 1E) used for the validation of its predicted structure and measured dynamic topographies using flexible fitting. (B) Superposition of the predicted dynamic molecular conformations (Figure 1F) with the corresponding topographic maps. (C) Simulated AFM images representing the molecular conformations in (B).

Figure S3: Topographies of EGFR‐ECD and MET‐ECD. HS‐AFM images of the EGFR‐ECD (A) and MET‐ECD (B) on mica. HS‐AFM images were captured in 1 mM PBS at scan rates of 150 (A) and 200 (B) ms/frame. HS‐AFM images in (B), obtained using ‐an SiN probe without an ElectronBeam‐Deposited (EBD) tip, exhibit larger morphological features than other HS‐AFM images. Different orientations of the simulated AFM images of the EGFR‐ECD (C) and MET‐ECD (D). Simulated AFM images were generated using the BioAFMviewer software [14].

Figure S4: Additional examples of the amivantamab/EGFR complex. HS‐AFM images of two different amivantamab/EGFR‐ECD complexes. HS‐AFM images were acquired on mica in 1 mM PBS at scan rates of 105 (A) and 150 (B) ms/frame. (C) Binding and unbinding events throughout the duration of HS‐AFM imaging for three complexes from Figure 2A (black), (A) (red), and (B) (blue). EGFR was bound to amivantamab for 94% and 91% of the total observation period in (A) and (B), respectively.

Figure S5: Additional examples of amivantamab/MET complex. HS‐AFM images of two different amivantamab/MET‐ECD complexes. HS‐AFM images were acquired on mica in 1 mM PBS at scan rates of 250 (A) and 141 (B) ms/frame. (C) Binding and unbinding events throughout the total period of HS‐AFM imaging for three complexes from Figure 3A (black), (A) (red), and (B) (blue). MET was bound to amivantamab throughout the entire observation period in both (A) and (B).

Figure S6: Additional examples of amivantamab/EGFR/MET complex. HS‐AFM images of two different amivantamab/EGFR‐ECD/MET‐ECD complexes. HS‐AFM images were acquired on mica in 1 mM PBS at scan rates of 80–300 (A) and 300 (B) ms/frame. (C) Binding and unbinding events throughout the total period of HS‐AFM imaging for three complexes from Figure 4A (black), (A) (red), and (B) (blue). EGFR and MET were bound to amivantamab throughout the total observation period in both (A) and (B).

Figure S7: Effect of the E/T ratio on MNC mediation of amivantamab‐induced ADCC. PC‐9 cells were incubated with control IgG or amivantamab (1 μg/mL) and MNCs at the indicated E/T ratio. The ADCC was determined as described in the Materials and Methods. Bars indicate the standard deviation of triplicate cultures. †< 0.1, **p < 0.001, ***p < 0.0001.

Figure S8: Amivantamab‐induced ADCC in PC‐9 cells mediated by MNCs from different donors. PC‐9 cells were incubated with amivantamab and MNCs from three different donors at the indicated E:T ratio. The ADCC was determined as described in the Materials and Methods. Bars indicate the standard deviation of triplicate cultures.

Figure S9: Amivantamab‐induced ADCC in NSCLC cells with common EGFR mutations. (A) NSCLC cells were incubated with amivantamab (1 μg/mL) and MNCs at an E:T ratio of 20. The ADCC was determined as described in the Materials and Methods. Bars indicate the standard deviation of triplicate cultures. (B) Correlation between the MFIs of the MET expression in NSCLC cells and amivantamab‐induced ADCC.

Movie S1: HS‐AFM images acquired at intervals of 125 ms for amivantamab on mica surface at a scan size of 50 nm × 50 nm.

Movie S2: Normal mode analysis of amivantamab structure. Flexible motions of Fc and Fab domains corresponding to the six slowest normal modes are visualized. For computation the elastic network model with a cutoff distance of 8 Å was employed (the same as used for flexible fitting).

Movie S3: HS‐AFM images acquired at intervals of 150 ms for amivantamab/EGFR complex on mica surface at a scan size of 60 nm × 60 nm.

Movie S4: HS‐AFM images acquired at intervals of 255 ms for amivantamab/MET complex on mica surface at a scan size of 80 nm × 80 nm.

Movie S5: Combined sequential stacks of HS‐AFM images acquired at intervals of 168–200 ms for amivantamab/EGFR/MET complex on mica surface at a scan size of 80 nm × 80 nm.

Table S1: The type of EGFR mutations in human NSCLC cell lines used in this study.

Data S1: Materials and Methods.