The pGEX-4T-1-Axl^−Ig1 recombinant plasmid was purchased from Beijing SinoGenoMax Research Center Co., Ltd. The Axl^−ECD-Fc was made in our laboratory (Beijing Key Laboratory of Therapeutic Gene Engineering Antibody, Beijing, China). The biotin-labeled Axl^−ECD-Fc was purchased from Beijing Jiaxuan Zhirui Biotechnology Co., Ltd. RPMI-1640 medium (cat. no. A1049101), DMEM (cat. no. C11995500BT), 0.25% trypsin-EDTA (cat. no. 25200-072) and FBS (cat. no. 1997802C) were purchased from Gibco (Thermo Fisher Scientific, Inc.). Freund's complete adjuvant (cat. no. F5881), Freund's incomplete adjuvant (cat. no. F5506), polyethylene glycol (PEG1450; cat. no. P7306) and hypoxanthine-aminopterin-thymidine (HAT) media supplement (cat. no. H0262) were purchased from Sigma-Aldrich (Merck KGaA). Escherichia coli BL21 (DE3) competent cells (cat. no. C1400), Glutathione S-transferase (GST) beads (cat. no. P2020), RIPA buffer (cat. no. R0010) and Giemsa solution (cat. no. G1010) were purchased from Beijing Solarbio Science and Technology Co., Ltd. HiTrap Protein G HP (cat. no. 17-0405-01) was purchased from GE Healthcare (Cytiva). R428 inhibitor (cat. no. CC2997) was purchased from ChemCatch (http://www.chemcatch.cn/index.php). The 3,3′,5,5,-tetramethylbenzidine (TMB) solution (cat. no. 00-4201-56) was purchased from Invitrogen (Thermo Fisher Scientific, Inc.).

Recombinant human Gas6 protein (cat. no. 885-GSB) and goat anti-Axl polyclonal antibody (cat. no. AF154) were purchased from R&D Systems, Inc. Anti-phosphorylated (p)-Axl (cat. no. 5724), anti-mouse IgG-HRP (cat. no. 7076), anti-rabbit IgG-HRP (cat. no. 7074) and anti-GAPDH-HRP (cat. no. 51332) antibodies were purchased from Cell Signaling Technology, Inc. Streptavidin-HRP (cat. no. 554066) was purchased from BD Biosciences. Anti-mouse IgG-FITC (cat. no. ab6785), anti-goat IgG-FITC (cat. no. ab6737) and anti-goat IgG-HRP (cat. no. ab6741) were purchased from Abcam. Goat anti-human IgG (GAH-IgG; cat. no. 01-10-06) was purchased from KPL, Inc., while mouse anti-GST IgG (cat. no. 66001-2-Ig) was purchased from ProteinTech Group, Inc.

A total of 4 female BALB/c mice were purchased from JOINN Laboratories. The animal experiments were approved by the Animal Ethics Committee of Suzhou Vocational Health College (Jiangsu, China; approval no. SWAE201905) and performed in accordance with the Guidelines of the Care and Use of Laboratory Animals of the Ministry of Health, China to minimize animal suffering and distress. The mice were monitored daily by an animal keeper and no mice died during the experiment. The mice were anesthetised with 2% isoflurane and blood was extracted from the tails to verify the immune titer. A total of 4 mice were euthanized by cervical dislocation and death was verified by respiratory and cardiac arrest. The duration of the experiment was 8 weeks and 3 days.

The human breast cancer cell line, MDA-MB-231 (cat. no. HTB-26) and the human NSCLC cell line H1299 (cat. no. CR-5803) were purchased from the American Type Culture Collection and were authenticated by Beijing ZhongYuan Ltd., in 2014. The SP2/0 mouse myeloma cell line (cat. no. CL-0217) was purchased from Procell Life Science&Technology Co., Ltd. The cells were cultured in DMEM containing 10% FBS and 100 U/ml penicillin-streptomycin at 37°C in a humidified incubator with 5% CO₂.

The protein sequence information for Axl (UniProtKB P30530) was determined using www.ebi.ac.uk. The crystal complex information for Gas6-Axl (no. 2C5D) was obtained from the Protein Data Bank (https://www.ncbi.nlm.nih.gov/Structure/pdb). Interaction analysis of Gas6-Axl was performed using Insight II 2000 software (Insight Software, Inc.).

Theoretical analyses were performed as described in our previous study (22). Briefly, based on the crystal complex structure (23), the Gas6-Axl theoretical complex model was constructed and optimized using the steepest decent and conjugate gradient method (discovery mode) under the consistent valence force field. With the optimized complex structure, the computer graphics technique and the distance geometry method (standard mode) were used to analyze the Gas6-Axl interaction mode. The superimposition method (standard mode) was used to identify the orientation of the main chain carbon atoms. Furthermore, the interaction binding free energy calculation method (discovery mode) was used to calculate the binding energy of Gas6-Axl.

The DNA sequence encoding the Axl functional domain (Axl^−Ig1) was confirmed by BLAST analysis (https://blast.ncbi.nlm.nih.gov/Blast.cgi, GenBank: BC032229.1) and cloned into the pGEX-4T-1 vector between BamHI and XhoI restiction sites, which allowed for the addition of a GST tag to the N-terminus of the Axl^−Ig1 domain (GST-Axl^−Ig1 fusion protein). The recombinant plasmid pGEX-4T-1-Axl^−Ig1 was transformed into Escherichia coli BL21 (DE3) competent cells. After optimization of the expression conditions, the GST-Axl^−Ig1 fusion protein expression was induced by the addition of 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 20°C for 6 h, followed by purification using GST beads according to the manufacturer's instructions. Then, the GST-Axl^−Ig1 fusion protein was analyzed using 12% SDS-PAGE and further validated using western blot analysis and the goat anti-Axl polyclonal antibody, AF154 and mouse anti-GST IgG, as separate primary antibodies, followed by anti-goat IgG-HRP and anti-mouse IgG-HRP as separate secondary antibodies.

Animal immunization was performed as described previously (24). Briefly, four BALB/c female mice (8-week-old, 18–20 g) were subcutaneously immunized with 100 µg GST-Axl^−Ig1 antigen in Freund's complete adjuvant on the first week, and in Freund's incomplete adjuvant on the fourth and sixth week. When the immune titer was >1:10,000, the mice were intravenously immunized with 100 µg GST-Axl^−Ig1 antigen without adjuvant for the final boost. Then, 3 days later, the mice were euthanized by cervical dislocation and the spleen cells from the immunized mice were fused with SP2/0 mouse myeloma cells at a ratio of 7:1 with PEG1450. The fused cells were cultured in RPMI-1640 medium with 20% FBS and HAT, then screened using ELISA and the hybridoma cell supernatant.

The hybridoma cell clones that bound to GST-Axl^−Ig1, but not GST were further screened using a BD FACSCalibur™ flow cytometer (cat. no. 342975; BD Biosciences) and FlowJo software (version 7.6; BD Biosciences) using H1299 cells (NSCLC cell line) with a high expression of Axl. The H1299 cell line was incubated with the hybridoma cell supernatant containing anti-Axl mAbs and detected using anti-mouse IgG-FITC (1:1,000). Goat anti-Axl polyclonal antibody AF154 (1:500) served as a positive control and was detected using anti-goat IgG-FITC (1:1,000). ELISAs were always performed before flow cytometry screening. After three rounds of ELISA, flow cytometry and limited dilution subclonal screening, hybridoma cell secreting anti-Axl mAbs were obtained. The hybridoma cells were cultured at 37°C with 5% CO₂, and 3×10⁶ hybridoma cells (6×10⁶ cells/ml) were seeded into the abdominal cavity of each mouse to boost the production of the antibody protein. Anti-Axl mAbs were purified from the mouse ascites using Protein G affinity chromatography and an ÄKTAprime^® plus system (cat. no. 11001313; Cytiva) with 20 mM sodium phosphate (pH 7.0) as binding buffer, and 0.1 M glycine-HCl (pH 2.7) as elution buffer according to the manufacturer's instructions. Anti-Axl mAbs were analyzed via reduced and non-reduced 12% SDS-PAGE, filtered under sterile conditions and stored at 4°C in PBS. Protein concentrations were determined using a BCA kit (Applygen Technologies, Inc.).

The anti-Axl mAbs were tested for binding with Axl fusion proteins, Axl^−ECD-Fc and GST-Axl^−Ig1. The ELISA plates were coated with 2 µg/ml Axl^−ECD-Fc and GST-Axl^−Ig1 in 0.1 M sodium carbonate buffer (pH 9.6) overnight at 4°C. After blocking with 10% FBS in PBS buffer at 37°C for 1 h, the plates were incubated with serial dilutions of mouse anti-Axl mAb (5.56, 1.85, 0.62, 0.21, 0.069, 0.023, 0.008, 0.0027 and 0.0009 µg/ml) at 37°C for 1 h. Subsequently, anti-mouse IgG-HRP (1:2,500) was added as the secondary antibody and incubated at 37°C for 30 min. The binding signals were visualized using the TMB substrate and the absorbance was measured at 450 nm using a SPECTRA MAX 190 ELISA reader (Molecular Devices, LLC). The ELISA plates were separately washed with TBS-Tween-20 (TBST; 50 mM Tris-HCl, 150 mM NaCl, 0.2% Tween-20, pH 7.4) three times for 5 min after coating, blocking and incubating with the primary and secondary antibodies.

The anti-Axl mAbs were tested for binding to the Axl fusion proteins. Axl^−ECD-Fc and GST-Axl^−Ig1 proteins were resolved on 12% SDS-PAGE and transferred onto a PVDF membrane, which was blocked with 5% skimmed milk at room temperature for 1 h. The membranes were incubated with mouse anti-Axl mAbs (10 µg/ml) overnight at 4°C, then with anti-mouse IgG-HRP (1:2,500) at room temperature for 1 h. The anti-Axl polyclonal antibody AF154 (1:500) was used as the positive control to recognize the Axl^−ECD and Axl^−Ig1 domains and was detected using the anti-goat IgG-HRP (1:2,500). Finally, the membranes were developed using an ECL detection system (SuperSignal Westpico Trial kit; Thermo Fisher Scientific, Inc.). The membrane was separately washed three times with TBST for 10 min after transferring, blocking and incubating with the primary and secondary antibodies.

The anti-Axl mAbs were tested for binding to Axl molecules on the surface of the cells. After trypsin digestion, the MDA-MB-231 cell lysates from 5×10⁵ cells were obtained using ice-cold RIPA buffer and were resolved using 12% SDS-PAGE. The subsequent steps were the same as aforementioned. The anti-Axl antibody AF154 was also used as the positive control.

After trypsin digestion, 3×10⁵ MDA-MB-231 cells were collected and incubated with mouse anti-Axl mAb (10 µg/ml) as the primary antibody at 4°C for 30 min. The cells were then incubated with anti-mouse IgG-FITC (1:1,000) as the secondary antibody at 4°C for 30 min. Finally, the cells were detected using a BD FACSCalibur™ flow cytometer (cat. no. 342975; BD Biosciences). The cells were separately washed three times with washing buffer (PBS containing 2% FBS) after digestion and incubation with the primary and secondary antibodies.

ELISA was performed as aforementioned. Briefly, the ELISA plates were coated with 0.5 µg/ml Gas6, and incubated with 2 µg/ml biotin-labelled Axl^−ECD-Fc, then with serial dilutions of anti-Axl mAb (500, 125, 32, 8, 2 and 0.5 µg/ml). Streptavidin-HRP (1:1,000) was added as the secondary antibody and binding signals were measured at 450 nm. The ELISA plates were washed three times separately with TBST for 5 min after coating, blocking and incubating with the primary and secondary antibodies.

Western blot analysis was performed as aforementioned. The MDA-MD-231 cells were seeded into 6-well plates at 3×10⁵ cells/well and cultured overnight. After incubation, 100 µg/ml (0.67 µmol/l) anti-Axl mAbs were added to the cells for 4 h, followed by the addition of 200 ng/ml Gas6 for 15 min. Neither Gas6 nor anti-Axl mAbs were added to the non-treated group (NT). Subsequently, the cell lysates were obtained using ice-cold RIPA buffer, and the proteins were resolved on 12% SDS-PAGE after quantification using a BCA kit (Applygen Technologies, Inc.). The proteins were transferred onto a PVDF membrane, followed by blocking with 5% skimmed milk. Goat anti-Axl antibody AF154 (1:500) and anti-goat IgG-HRP (1:2,500) were used to detect the expression level of Axl, and rabbit anti-p-Axl antibody (1:500) and anti-rabbit IgG-HRP (1:2,500) were used for the detection of p-Axl (Tyr702). Finally, the signal was detected using ECL detection reagents and semi-quantified using ImageJ software (version 18.0; National Institutes of Health). The membrane was washed three times with TBST for 10 min after transferring, blocking and incubating with the primary and secondary antibodies.

To assess cell migration, 8-µm pore size BD-Falcon 24 Fluoroblock Transwell inserts were used. After serum-starved culture overnight, 1×10⁵ MDA-MB-231 cells were seeded into the upper chamber containing serum-free DMEM. For Gas6-dependent migration, 200 ng/ml Gas6 was added to the lower chamber containing 10% FBS+ DMEM in the presence or absence of anti-Axl mAbs (100 µg/ml or 0.67 µmol/l). R428 (0.67 µmol/l) was selected as the positive control. Neither Gas6 nor anti-Axl mAbs/R428 were added to the NT group. After 4 h of migration at 37°C, the non-migrating cells on the top surface of the filter were removed using cotton swabs. The migrating cells on the bottom surface of the filter were fixed with 4% paraformaldehyde for 30 min at room temperature and stained with Giemsa solution for 10 min at room temperature. Images of the migrating cells were captured under a light microscope, then quantified using ImageJ software (version 1.52p; National Institutes of Health), based on five random visual fields in the different groups.

After trypsin digestion, 8×10⁵ MDA-MB-231 cells were seeded into 6-well plates in 2 ml DMEM per well. At 90% confluence, wounds were created by scratching the cell monolayer with a sterile p200 µl pipette tip. Then, the cells were washed with PBS and cultured in serum-free DMEM containing anti-Axl mAbs (100 µg/ml or 0.67 µmol/l) or R428 (0.67 µmol/l) as the positive control. In total, five images from random visual fields for each group were captured at 0 and 24 h post-wounding. The wounded area was measured using ImageJ software (version 1.52p; National Institutes of Health), and recorded as S₀ and S₂₄ respectively. The cell migration rate (%) was calculated with the following formula: (S₀-S₂₄)/S₀ × 100%.

All assays related to bioactivities were repeated three times. The data were presented as the mean ± SD. Statistical analysis was performed using GraphPad Prism software v.8.0 (GraphPad Software, Inc.). For the Axl, p-Axl level detection and cell migration assays, one-way ANOVA was performed and Dunnett's multiple comparisons test was used to compare all other groups with the Gas6 group or NT group. P<0.05 was considered to indicate a statistically significant difference.

The natural ligand, Gas6, binds to Axl and actives downstream signaling pathways. Thus, the current study first determined the functional domain of Axl that interacts with Gas6, before preparing an antigen containing the specific domain. Schematic representation of the Gas6-Axl interaction is showed in Fig. 1A. The protein sequence information of Axl (UniProtKB P30530) was obtained by searching www.ebi.ac.uk. As shown in Fig. 1A, the Axl protein consisted of 894 aa, with a signal peptide (1–25 aa) and a mature chain (26–894 aa). The mature chain of Axl consisted of an extracellular domain (26–451 aa), a transmembrane region (452–472 aa) and an intracellular domain (473–894 aa). There were two Ig-like domains in the extracellular domain, Axl Ig-like domain 1(Axl^−Ig1, 27–128 aa) and Axl Ig-like domain 2 (Axl^−Ig2, 139–222 aa).

Based on the crystal complex information (PDB no. 2C5D) (23) and the theoretical complex structure of Gas6-Axl (Fig. 1C), Axl^−Ig1 (27–128 aa) was determined as the major functional domain interacting with Gas6 using interaction mode analysis and the standard mode. As presented in Fig. 1B, 23 aa residues from Gas6 laminin G-like domains 1 (Gas6^−LG1) and 22 aa residues from Axl^−Ig1 interacted at the interface, and resulted in the formation of 13 hydrogen bonds and 120 non-banded contacts between Gas6 and Axl, which was a relatively stable interaction.

It was identified that R³⁰⁸, R³¹⁰, K³¹², R³¹³ and L³¹⁴ of Gas6^−LG1 served important roles in interacting with ≥3 aa residues of Axl^−Ig1 via hydrogen bonds and non-bonded contacts (Fig. 1B), and were considered as the key aa. With respect to Axl^−Ig1, E⁵⁶, E⁵⁹, D⁷³, S⁷⁴, T⁷⁵, Q⁷⁶, T⁷⁷ and Q⁷⁸ interacted with ≥1 of the key aa from Gas6^−LG1 (Fig. 1B), and were considered as the key aa. Thus, the antibody against this Axl^−Ig1 domain may block the Gas6-Axl binding and attenuate the downstream signals and bioactivities.

Furthermore, amino acids involved in Gas6-Axl interactions were analyzed and compared between a previous EMBO study (23) and the present study. As presented in Table I, according to Sasaki et al (23), 12 aa residues of Gas6^−LG1 and 9 aa residues of Axl^−Ig1 were involved in Gas6-Axl interactions, including 6 hydrogen bonds (no data available for the other contacts). Furthermore, 10 aa residues of Gas6 and 8 aa residues of Axl (italicized in Table I) were shared in the study by Sasaki et al and the present study. Consistently, R³⁰⁸, R³¹⁰ and K³¹² of Gas6 and E⁵⁶, E⁵⁹ and T⁷⁷ of the Axl^−Ig1 domain were considered as the key residues in Gas6-Axl binding by both studies (underlined in Table I). Furthermore, the key role of E⁵⁶, E⁵⁹ and T⁷⁷ of the Axl^−Ig1 domain in binding to Gas6 was confirmed by the study by Sasaki et al (23) and our previous study using mutants assays (22).

GST-Axl^−Ig1 protein expression was induced by 1 mM IPTG at 20°C for 6 h. As expected, an obvious expression band at 34 kDa was visible after induction following SDS-PAGE analysis, and most of the GST-Axl^−Ig1 protein was soluble in the ultrasonic supernatant (Fig. 2A). GST-Axl^−Ig1 protein with high purity was obtained after affinity purification using GST beads. Further western blot analysis revealed that the GST-Axl^−Ig1 protein could specifically bind with the anti-Axl antibody AF154 and anti-GST IgG (Fig. 2B), indicating that the GST-Axl^−Ig1 protein was correctly expressed and could be used as an antigen for immunizing animals.

In total, seven hybridoma cell clones that bound to GST-Axl^−Ig1, but not GST, were selected from six plates of 96-well cells from two rounds of ELISA screening (Fig. 3A) and were further screened using flow cytometry by binding to the H1299 cells with high Axl expression. After three rounds of ELISA and flow cytometry screening in turn, two clones were obtained with 39.42 and 38.80% binding ratio to the H1299 cells, separately (Fig. 3C). Finally, the hybridoma cell clones that specifically bound to the Axl^−Ig1 functional domain were obtained, named as Am1 and Am2. The two mAbs had the same isotype, IgG3 heavy chain and κ light chain.

After protein G purification, the specific antibody protein bands were observed at ~50 kDa and 25 kDa in reduced SDS-PAGE analysis, and at ~150 kDa in non-reduced SDS-PAGE analysis, and as expected with a high purity (Fig. 3B).

As demonstrated by ELISA and western blot analysis (Fig. 4A and B), Am1 or Am2 could bind with the Axl fusion proteins, Axl^−ECD-Fc and GST-Axl^−Ig1, which indicated that Am1 or Am2 could recognize the Axl^−ECD domain and Axl^−Ig1 functional domain. In the present study, goat anti-human IgG (GAH-IgG) was used as the control for recognizing the human IgG-Fc domain and anti-GST IgG was used as the control for recognizing the GST tag.

Using competitive ELISA, anti-Axl mAbs were detected for their ability to inhibit Gas6 binding with Axl^−ECD-Fc. Biotin-labeled Axl^−ECD-Fc and Am1 or Am2 were incubated with the ELISA plates coated with Gas6, and the captured biotin-labeled Axl^−ECD-Fc was detected using streptavidin-HRP. As shown in Fig. 4C, the decreasing signal indicated that Am1 and Am2 blocked Axl^−ECD-Fc binding to Gas6. Thus, the inhibitory effect of Am1 and Am2 on the Gas6-Axl interaction made it possible to neutralize the activity of Axl. Given that 125 µg/ml Am1 and Am2 had a notable inhibitory effect on the Gas6-Axl interaction, 100 µg/ml Am1 and Am2 was used in the subsequent assays.

Axl was notably overexpressed in highly invasive cells (1–3,6–8). The anti-Axl mAbs were detected for their binding to the natural Axl proteins on the surface of highly invasive MDA-MB-231 cell using western blot analysis and flow cytometry. As presented in Fig. 5A, Am1 or Am2 could bind to Axl proteins in the cell lysate or on the cell surface.

Gas6 binds to Axl and induces its phosphorylation, which leads to activation of the downstream signaling pathways and the proliferation and migration of tumor cells (2). Axl and p-Axl proteins were detected using western blot analysis to determine the effect of anti-Axl mAb binding with Axl in the MDA-MB-231 cells. For the NT group with no Gas6 stimulation, Axl protein expression was observed at a high level in the MDA-MB-231 cells, but p-Axl was almost undetectable (Fig. 5B). Furthermore, Gas6 stimulation for 15 min had no effect on Axl expression but increased p-Axl expression up to 165-fold compared with that in the NT group. It was also identified that Am1 or Am2 significantly decreased Axl expression from 0.930 to 0.483 or 0.480 folds compared with that in the NT group respectively (P<0.001), and p-Axl expression from 165.35 to 9.99 or 21.91 folds compared with that in the NT group, respectively (P<0.0001). The ratio of p-Axl/Axl is also shown in Fig. 5B. Therefore, it was suggested that Am1 and Am2 could reduce the expression level of Axl and significantly inhibit Axl phosphorylation, which may be the main mechanism preventing tumorigenesis.

Anti-Axl mAbs were evaluated for their ability to inhibit cell migration using a Transwell assay. The MDA-MD-231 cells, plated in the upper chamber of Transwell inserts, were stimulated to migrate into the lower chamber containing Gas6 and Am1 or Am2. Compared with that in the Gas6 group, the migration of the MDA-MD-231 cells was significantly attenuated by Am1 or Am2 (P<0.001) to a similar extent as that observed with the Axl inhibitor R428 (P=0.865 and P=0.735, respectively) (Fig. 5C). In addition, cell migration in the Gas6+Am1 or Gas6+Am2 groups was weaker compared with that in the NT group without Gas6 stimulation, indicating that Am1 and Am2 almost completely blocked the stimulation of Gas6.

In the wound healing assay, Am1 or Am2 were added to scratched MDA-MD-231 cells to verify their inhibitory effect on cell migration during the wound healing progress. Compared with that in the NT group, the wound healing speed at 24 h post-wounding was significantly impeded by Am1 (P<0.01) or Am2 (P<0.05), although the inhibitory effect of Am1 or Am2 on cell migration was a little weaker to that observed in the group treated with the Axl inhibitor R428 (Fig. 5D). Collectively, both the Transwell and wound healing assays indicated that Am1 and Am2 could effectively inhibit MDA-MD-231 cell migration.