MCF7 (cat. no. SCSP-531), 293T (cat. no. SCSP-502), MDA-MB-231 (cat. no. SCSP-5043) and HeLa (cat. no. SCSP-504) cells were purchased from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences. MCF7 cells were cultured in MEM (Procell Life Science & Technology, Co., Ltd.) supplemented with 0.01 mg/ml insulin and 10% fetal bovine serum (FBS). The 293T, MDA-MB-231 and HeLa cells were maintained in DMEM (Gibco; Thermo Fisher Scientific, Inc.) containing 10% FBS. All cells were incubated at 37˚C in a humidified atmosphere with 5% CO₂. Cell line authentication was performed by short tandem repeat profiling at the Forensic Science Center, Jining Medical University. IL-6 was purchased from PeproTech, Inc. R428 and tocilizumab were obtained from MedChemExpress. R428 was used as previously described (25), and tocilizumab was used at 25 µg/ml.

AXL plasmid (cat. no. 105932, based on pcDNA4/TO/myc/HIS) was obtained from Addgene, Inc. The STAT3 reporter plasmid (SIE vector) and pRL-TK vector were purchased from Promega Corporation. Lentivirus for shRNA targeting AXL (shAXL: 5'-GGGTGACAATGTGGGAGATTG-3') and control shRNA (Ctrl: 5'-TTCTCCGACGTGTCACGT-3') were purchased from Shanghai GenePharma Co., Ltd. as pre-packaged, ready-to-use viral particles (3rd-generation, packaged in 293T cells; titer 10⁹ TU/ml). Transfection of plasmids (1 µg for one well of 12-well plate) was performed using Lipo8000 (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. Briefly, Lipo8000 and plasmids were incubated in DMEM without FBS for 5 min and added the cell medium for 4 h at 37˚C. A total of 44-48 h later, cells were harvested for analysis. For lentiviral transfection, viruses (5 µl/ml) and polybrene (2 µg/ml) were added to the cell medium at MOI=5, incubated at 37˚C/5% CO₂ for 16 h before medium change. After 3 days cells were treated with puromycin (2 µg/ml) for 1 week to select transfected cells. Then, cells were cultured at puromycin (0.5 µg/ml) for maintenance. These stable cell lines were used for experiments.

Nuclear and cytoplasmic proteins were extracted using a nuclear and cytoplasmic protein extraction kit (cat. no. P0027; Beyotime Institute of Biotechnology) according to the manufacturer's instructions. These experiments were repeated three times.

Western blot analysis was performed as previously reported (25,26). Total cellular proteins were extracted using RIPA buffer supplemented with protease and phosphatase inhibitors sourced from Beyotime Institute of Biotechnology. Subsequently, nuclear and cytoplasmic proteins were isolated using a dedicated nuclear-cytoplasmic protein extraction kit (Beyotime Institute of Biotechnology). The protein concentration was precisely measured using the BCA Protein Assay kit (Beyotime Institute of Biotechnology). Equal quantities (10 µg) of these proteins were resolved through 10% SDS-PAGE and then transferred to PVDF membranes. The membranes were first blocked for 1 h at room temperature in a 5% non-fat milk solution. They were then incubated overnight at 4˚C with an array of primary antibodies: AXL (1:1,000; cat. no. 8661), phosphorylated (p)-AKT (1:1,000; cat. no. 4060), STAT3 (1:1,000; cat. no. 9139), Histone H3 (1:1,000; cat. no. 9715), p-STAT3 (1:1,000; cat. no. 9145), AKT (1:1,000; cat. no. 4691; Cell Signaling Technology, Inc.), ERK1 + ERK2 (1:10,000; cat. no. ab184699), p-ERK1 + ERK2 (1:10,000; cat. no. ab76299) (all from Abcam), GAPDH antibody (1:1,000; cat. no. 60004-1-Ig; Proteintech Group, Inc.) and p-AXL (1:200; cat. no. AF2228; R&D Systems, Inc.). Post-incubation with the primary antibodies, the membranes were further incubated with peroxidase-conjugated goat anti-rabbit (cat. no. ZB-2301) or goat anti-mouse (cat. no. ZB-2305) secondary antibodies, both from OriGene Technologies Inc., at a dilution of 1:5,000 for 1 h at room temperature. Immunoreactive bands were detected using ECL reagent from Beyotime Institute of Biotechnology. GAPDH and histone H3 were employed as loading controls for cytoplasmic and nuclear proteins, respectively. All the experiments were repeated at least three times. ImageJ software (version 1.53m; National Institutes of Health) was used for the densitometric analysis.

Cells underwent co-transfection with an SIE vector serving as a STAT3 reporter, a pRL-TK vector acting as an internal control, and either an empty vector (EV) or an AXL plasmid using Lipo8000 (Beyotime Institute of Biotechnology) according to the manufacturer's protocol. For each well of a 12-well plate, a total of 1 µg plasmid DNA was transfected, comprising 0.5 µg AXL plasmid (or empty vector), 0.45 µg STAT3 reporter plasmid (SIE vector), and 0.05 µg pRL-TK internal control plasmid. The transfection complexes were incubated with cells for 4 h at 37˚C in a humidified incubator with 5% CO₂, after which the medium was replaced with fresh complete medium carefully. At 48 h post-transfection, luciferase activity was determined by employing the Dual Luciferase Reporter Gene Assay kit from Beyotime Institute of Biotechnology (cat. no. RG029M), strictly adhering to the manufacturer's guidelines. In the inhibitor experiments, cells were transfected with 0.45 µg STAT3 reporter plasmid (SIE vector), and 0.05 µg pRL-TK internal control plasmid. 24 h later, the cells were incubated with R428 at a concentration of 1, 3 µM or DMSO for another 24 h followed by luciferase activity determination. This experiment was repeated three times with triplicate wells for each condition, and data was normalized by comparison with Renilla luciferase activity.

Total RNA was extracted using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.). First-strand cDNA was synthesized using PrimeScript RT Master Mix (Takara Biotechnology, Co., Ltd.) according to the manufacturer's instructions. qPCR was performed using SYBR-Green PCR Master Mix (Vazyme Biotech Co., Ltd.) on a QuantStudio 5 Real-Time PCR system (Thermo Fisher Scientific, Inc.). Each qPCR reaction was performed in a final volume of 20 µl, containing 10 µl SYBR Green Master mix (2X), 0.5 µl forward and reverse primers (10 µM), 8 µl DEPC treated water and 1 µl cDNA. The following thermocycling conditions were used for the qPCR: 95˚C for 30 sec, 40 cycles of 95˚C for 10 sec and 60˚C for 30 sec. The primer sequences for IL-6 and 18S were as follows: IL-6 forward, 5'-CACAGACAGCCACTCACCTC-3' (sense) and reverse, 5'-TTTTCTGCCAGTGCCTCTTT-3' (antisense); 18S rRNA, 5'-GAGGATGAGGTGGAACGTGT-3' (sense) and reverse, 5'-GGACCTGGCTGTATTTTCCA-3' (antisense). 18S was used as an internal control. Relative gene expression was calculated using the 2^-ΔΔCq method (27). This experiment was repeated three times with triplicate wells for each condition.

The IL-6 concentration in the cell culture supernatants was measured using an IL-6 ELISA kit [cat. no. EK106; Hangzhou Multi Sciences (Lianke) Biotech Co., Ltd.] according to the manufacturer's protocol. This experiment was repeated three times with triplicate wells for each condition.

Gene expression datasets GSE102484 (n=683) (28) and GSE9893 (n=155) (29) were downloaded from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/). GSEA was performed to compare IL6_JAK_STAT3 signaling between the AXL-high and AXL-low groups as reported in a previous study by the authors (21). Protein expression data of AXL and p-STAT3 (Y705) from breast invasive carcinoma were obtained from The Cancer Proteome Atlas (TCPA, n=901) (https://www.tcpaportal.org/tcpa/download.html); the samples whose AXL expression marked with NA (n=156) were excluded. The association between the expression of AXL and p-STAT3 was analyzed with SPSS (version 13.0; SPSS, Inc.) using Pearson's correlation analysis (n=745). The association analysis between AXL mRNA and IL-6 mRNA in BC was performed by gepia2 using Pearson's correlation analysis as the method to calculate the correlation coefficient (http://gepia2.cancer-pku.cn/#correlation) (30).

Data are presented as the mean ± standard deviation (SD). Differences between two groups were analyzed using a two-tailed unpaired Student's t-test. Comparisons among multiple groups were performed by one-way analysis of variance (ANOVA) followed by Dunnett's post hoc test using GraphPad Prism 8 (GraphPad Software, Inc.; Dotmatics). P<0.05 was considered to indicate a statistically significant difference.

To investigate the effects of AXL on STAT3 activation, AXL was overexpressed in MCF7 and 293T cells. Western blot analysis revealed that AXL overexpression significantly increased p-STAT3 (Y705) levels without altering total STAT3 (t-STAT3) expression (Fig. 1A and B). Moreover, AXL overexpression enhanced p-STAT3 (Y705) levels in both cell lines in a concentration-dependent manner (0, 1 and 2 µg) (Fig. 1C and D).

Subcellular fractionation analysis revealed that AXL overexpression increased nuclear STAT3 levels in MCF7 cells, indicating an enhanced STAT3 nuclear translocation (Fig. 1E). A port of AXL was also detected in the nucleus, consistent with a previous study (31). Dual luciferase reporter assay confirmed that AXL overexpression significantly augmented STAT3 transcriptional activity in MCF7 and 293T cells (Fig. 1F and G). Additionally, AXL overexpression upregulated IL-6 mRNA expression and secretion, as measured using RT-qPCR and ELISA, respectively (Fig. 1H and I). Furthermore, GEPIA2 analysis further validated an association between AXL and IL-6 mRNA in clinical BC samples (P=4.1x10^-7) (Fig. 1J). Treatment with the IL-6 receptor-neutralizing antibody, tocilizumab, partly abrogated the AXL-induced activation of STAT3 (Fig. 1K). These results suggest that AXL activates STAT3, at least in part by promoting IL-6 secretion.

To further confirm the role of AXL in STAT3 activation, AXL was knocked down in MDA-MB-231 and HeLa cells, which express high endogenous AXL. The results of western blot analysis demonstrated that the shRNA-mediated AXL knockdown significantly reduced STAT3 activation (Fig. 2A and B). Pharmacological inhibition with the selective AXL inhibitor, R428, also decreased AXL and STAT3 activation in the MDA-MB-231 cells in a concentration-dependent manner (Fig. 2C and D). Furthermore, R428 treatment dose reduced STAT3 transcriptional activity in a concentration-dependent manner (Fig. 2E).

GSEA analysis of the GSE102484 and GSE9893 datasets revealed that IL6_JAK_STAT3 signaling was significantly enriched in AXL-high BC tumors compared with AXL-low tumors [normalized enrichment score (NES)=1.57, P=0.027 for GSE102484; NES=1.58, P=0.008 for GSE9893] (Fig. 3A and B). TCPA dataset analysis also revealed a positive correlation between AXL protein expression and p-STAT3 (Y705) levels (Fig. 3C). These clinical data support a close association between AXL and STAT3 activation in BC.