Ligand‐independent EphA2 contributes to chemoresistance in small‐cell lung cancer by enhancing PRMT1‐mediated SOX2 methylation

Abstract

Chemoresistance is the crux of clinical treatment failure of small‐cell lung cancer (SCLC). Cancer stem cells play a critical role in therapeutic resistance of malignant tumors. Studies have shown that the role of erythropoietin‐producing hepatocellular A2 (EphA2) in tumors is complex. This study aimed to test the hypothesis that ligand‐independent activation of EphA2 modulates chemoresistance by enhancing stemness in SCLC. We verified that EphA2 was activated in chemoresistance sublines in a ligand‐independent manner rather than a ligand‐dependent manner. Ligand‐independent EphA2 enhanced the expression of stemness‐associated biomarkers (CD44, Myc, and SOX2), accelerated epithelial–mesenchymal transition (EMT) and reinforced self‐renewal to drive the chemoresistance of SCLC, while the P817H mutant EphA2 neutralized intrinsic function. Co‐immunoprecipitation (co‐IP) and GST‐pull down experiments were conducted to verify that EphA2 directly interacted with PRMT1. Moreover, EphA2 increased the expression and activity of PRMT1. Whereafter, PRMT1 interacted with and methylated SOX2 to induce stemness and chemoresistance in SCLC. Pharmacological inhibition of EphA2 showed a synergistic anti‐tumor effect with chemotherapy in preclinical models, including patient‐derived xenograft (PDX) models. These findings highlight, for the first time, that the EphA2/PRMT1/SOX2 pathway induces chemoresistance in SCLC by promoting stemness. EphA2 is a potential therapeutic target in SCLC treatment.

This study showed that ligand‐independent EphA2 induces chemoresistance in SCLC. Ligand‐independent EphA2 directly interacted with PRMT1, which augmented stemness to promote chemoresistance in SCLC while methylating SOX2 at R43. Finally, inhibition of EphA2 confirmed the anti‐tumor effect both in vitro and in vivo experiments.

Abbreviations

CI

combination indexes

co‐IP

co‐immunoprecipitation

EMT

epithelial–mesenchymal transition

EphA2

Erythropoietin‐producing hepatocellular A2

PDX

patient‐derived xenograft

RTKs

receptor tyrosine kinases

SCLC

small‐cell lung cancer

SFE

sphere formation efficiency

1. INTRODUCTION

Small‐cell lung cancer accounts for approximately 14% of all types of lung cancers and is characterized by extreme malignancy, with an overall 5‐year survival rate of only 7%. ¹ , ² SCLC patients are prone to developing multidrug resistance, which is the main reason for clinical treatment failure. ³ , ⁴ There had been a small breakthrough in the first‐line treatment strategy for SCLC over the past decades until the emergence of atezolizumab, which in combination with carboplatin and etoposide represents the current first‐line treatment for SCLC. ⁵ , ⁶ Other combinations used as a second‐line treatment include olaparib plus temozolomide. ⁷ , ⁸ Unfortunately, the efficacy of this treatment is modest and limited to a small subset of patients. ⁹ , ¹⁰ Moreover, more and more molecules have been recognized as novel potentially therapeutic targets against SCLC. ASCL1, NeuroD1, YAP1, and POU2F3 have been defined as molecular subtypes of SCLC. ¹¹ In addition, SOX2 is frequently amplified in SCLC. ¹² Whereas loss of Trp53 and Rb1 has been described to provide a preclinical mouse model to identify novel therapeutic targets against SCLC. ¹³ Nevertheless, new drugs that have been approved for the clinical treatment of SCLC are still limited. Hence, it is vital to investigate the molecular mechanisms of chemoresistance and to identify potential treatment strategies for SCLC.

Small molecules are one of the two main approaches for targeted cancer therapy. ¹⁴ The main targets for small molecules are kinases, especially tyrosine kinase. ¹⁵ Erythropoietin‐producing hepatocellular A2 belongs to the Eph subfamily, which is the largest RTK family. ¹⁶ EphA2 is highly expressed in multiple cancers, including lung cancer, esophageal carcinoma, colorectal cancer, and breast cancer. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ EphA2 signaling in cancer is complex and in a context‐dependent dual manner: (1) by canonical ephrin‐A ligand‐dependent signaling, EphA2 is autophosphorylated at Y588, which generally inhibits oncogenic signaling; or (2) by noncanonical ligand‐independent signaling, whereby EphA2 is phosphorylated at S897, driving downstream protumorigenic signaling upon crosstalk with Akt and other RTKs and signaling molecules. ²¹ , ²² , ²³ EphA2, which drives BRAF inhibitor resistance in melanoma, is phosphorylated at S897 by Akt. ²⁴ , ²⁵ , ²⁶ EphA2 suppresses anti‐tumor immunity by regulating PTGS2 in pancreatic ductal adenocarcinoma. ²⁷ Particularly, EphA2 plays an important role in stemness maintenance in glioblastoma. ²⁸ However, the effect of EphA2 on the chemoresistance of SCLC and the underlying molecular mechanism is still unclear.

Cancer stem cells are capable of self‐renewal and cell plasticity, which means that cancer cells can reversibly transit between epithelial and mesenchymal states. ²⁹ Cancer stem cells are the main reason for cancer initiation, tumor recurrence, therapy resistance, and metastasis. ³⁰ , ³¹ Drug resistance in SCLC could be attributed to the existence of a resistant cancer stem cell subpopulation. ³² Stemness properties are indicated to contribute to the intratumoral heterogeneity of cancer cells and govern therapeutic resistance. ³³ Studies have shown some biomarkers to be stemness‐associated biomarkers in SCLC, such as ALDH1, CD44, CD99, CD133, MYC, and SOX2. ³² , ³⁴ , ³⁵

Our previous study suggested that drug‐resistant subline H69AR possessed significantly higher mRNA expression of EPHA2 than drug‐sensitive subline H69 in SCLC, ³⁶ implying that EphA2 might induce chemoresistance in SCLC by reinforcing the cancer stemness. We performed CCK‐8 assays, western blot, and sphere formation assays in SCLC cell lines to investigate the role of EphA2 in the regulation of chemoresistance in SCLC. We used mouse xenograft models and PDX to assess the function of EphA2 in vivo. Finally, we identified the molecular mechanisms by which EphA2 regulates chemoresistance in SCLC.

2. MATERIALS AND METHODS

2.1. Cell lines and cell cultures

Human SCLC cell lines (NCI‐H69, NCI‐H69AR, and NCI‐H446) were obtained from the American Type Culture Collection (Manassas, VA, USA). The chemoresistant subline H446CDDP was induced by culturing H446 cells in progressively increasing concentrations of cisplatin (up to 0.5 μg/ml) over 12 months. H446CDDP was cultured in RPMI 1640 medium (Gibco, Waltham, WA, USA) supplemented with 10% FBS (Gibco, Waltham, WA, USA) and cisplatin (0.5 μg/ml). H69 and H446 cells were cultured in RPMI 1640 medium supplemented with 10% FBS, while H69AR was cultured in RPMI 1640 medium supplemented with 20% FBS. The chemoresistant sublines H69AR and H446CDDP were transferred to a drug‐free medium for at least 2 weeks before any experiment. All cell lines were maintained in a humidified incubator at 37°C with the presence of 5% CO₂. All cell lines were not contaminated with mycoplasma.

2.2. Vectors

The retroviral vectors and FLAG‐tagged plasmids of EPHA2 ^WT , EPHA2 ^A785S , EPHA2 ^P817H , EPHA2 ^Y930D , the lentiviral interference vectors of EPHA2, negative controls, and siRNAs for AKT were purchased from GenePharma (Suzhou, China). The shRNAs for PRMT1 and negative counterparts were obtained from Genechem (Shanghai, China). shRNA and siRNA sequences are provided in Table S1.

2.3. RNA isolation and RT‐qPCR

Samples were homogenized in RNAiso Plus buffer (TaKaRa, Kyoto, Japan), followed by RNA purification according to the manufacturer's instructions. Concentration and purity for each RNA extract were checked by comparing the OD_260/280 ratio on a NanoDrop 2000 instrument (Thermo Fisher Scientific, Waltham, WA, USA). First‐strand cDNA was synthesized from 1000 ng of total RNA using the FastKing RT Kit (Tiangen, Beijing, China). RT‐qPCR was performed in 10‐μl reactions on a Bio‐Rad CFX Connect device using the Talent qPCR PreMix (SYBR Green) kit (Tiangen, Beijing, China). Cycling conditions, as recommended by the manufacturer, were programmed. Sequences of RT‐qPCR primers are listed in Table S2.

2.4. Western blot

Protein extracts were purified from SCLC cells using a protein extraction kit (CWbio, Jiangsu, China). The BCA Protein Assay Kit (CWbio, Jiangsu, China) was used for protein quantification. Cell lysates supplemented with SDS‐PAGE loading buffer (Beyotime Biotechnology, Shanghai, China) were boiled at 100°C for 10 min. Protein samples were then electrophoresed in SDS‐PAGE gels and transferred onto PVDF membranes (Merck Millipore, Billerica, MA, USA). The membranes were blocked with 5% BSA and incubated with primary antibodies at 4°C overnight, washed three times with Tris‐buffered saline Tween (TBST) buffer, and incubated with secondary antibodies to exclude non‐specific binding. Last, signals were illuminated using chemiluminescence regents and digitally recorded under a Bio‐Rad imaging system (Bio‐Rad, Hercules, CA, USA).

2.5. Sphere formation assay

Single cells were suspended and seeded in 6‐well plates in densities of 2000, 1000, or 500 cells per well in serum‐free medium ³⁷ , ³⁸ and cultured for 14 days. Spheres were collected and the self‐renewal capacity of cells was estimated by SFE. The SFE was calculated based on the following formula: SFE = (number of spheres counted/number of seeded cells) × 100. ³⁹

2.6. CCK‐8 assay

Cell counting kit‐8 assay was performed according to the manufacturer's protocol. In total, 5000–20,000 cells were seeded into each well in a 96‐well plate supported by complete medium. Cells were then treated with gradient concentrations of drugs for 24 h. Optical density was assessed at 450 nm on an Absorbance Microplate Reader (Biotek, Winooski, VT, USA).

2.7. Co‐immunoprecipitation assay

Cells were collected and resuspended in freshly precooled IP lysis buffer (Thermo Fisher Scientific, Waltham, WA, USA) containing 1× protease and phosphatase inhibitor cocktail (CWbio, Jiangsu, China). The protein extraction was prepared as suggested by the manufacturer. The concentration of cell lysate was detected using the BCA Protein Assay Kit (CWbio, Jiangsu, China). Cell lysates were incubated with primary antibodies or IgG overnight at 4°C with gentle agitation. The immunoprecipitated complexes were captured by protein A/G magnetic beads (Thermo Fisher Scientific, Waltham, WA, USA) for 2 h at room temperature.

2.8. Immunofluorescence staining and confocal microscopy

Cells were seeded onto sterile matrix‐coated glass coverslips, washed with precooled PBS, and fixed in 4% paraformaldehyde for 10 min at room temperature. A permeabilization step was performed by incubating the coverslips in 0.1% Triton X‐100 in PBS buffer for 10 min at room temperature. Cells were blocked with 1% BSA prepared in PBS for 30 min at room temperature, incubated with primary antibodies (diluted in 1% BSA/PBST) at 4°C overnight, washed with precooled PBS, and then incubated with Alexa Fluor secondary antibodies (Thermo Fisher Scientific, Waltham, WA, USA) diluted in 1% BSA/PBST for 1 h at room temperature. Nuclei were counterstained with DAPI (Telenbiotech, Guangzhou, China). Images were acquired using the Leica TCS SP8 X confocal microscope (Leica, Wetzlar, Germany).

2.9. GST pull‐down assay

The GST‐tagged plasmids of EPHA2 ^WT , EPHA2 ^P817H , and His‐tagged plasmid of PRMT1 were transformed into E. coli BL21(DE3). The purified GST‐EphA2^WT or GST‐EphA2^P817H fusion proteins were immobilized in Glutathione Sepharose (GE Healthcare, Chicago, IL, USA) and then incubated with the purified His‐PRMT1 protein at 4°C overnight. The eluted samples were detected by western blot analysis using the GST antibody and His antibody.

2.10. Tumor xenograft experiments

Male BALB/c nude mice aged 4–6 weeks old were purchased from the Experimental Animal Center of Southern Medical University (Guangzhou, China). All mice used in this research were housed and manipulated in the Central Laboratory of Zhujiang Hospital, Southern Medical University complying with the institutional guidelines for animal use and care. Mice were accommodated for adaption for 1 week before being injected subcutaneously with cells. For in vivo chemosensitivity assays, the mice were injected intraperitoneally with PBS containing chemotherapeutics or PBS containing the vehicle (polyethylene glycol 300). At day 7, the mice received an intraperitoneal injection of etoposide (7 mg/kg) once every 2 days and cisplatin (3 mg/kg) once every 8 days. For in vivo sensitivity assay of ALW‐II‐41‐27 (MedChemExpress, Monmouth Junction, NJ, USA), the mice were randomized into four groups. Mice received empty vehicle (polyethylene glycol 300), chemotherapy as described above, ALW‐II‐41‐27 (20 mg/kg/day), or a combination of ALW‐II‐41‐27 and chemotherapy. The size of xenografts was monitored every 3 days. Mice were euthanized 21 days after injection and tumor xenografts were collected. Euthanasia was initiated by CO₂ suffocation followed by cervical dislocation. Xenograft tumors were collected following euthanasia.

2.11. Patient‐derived xenografts (PDX)

SCLC PDX models have been described elsewhere. ³ , ⁴⁰ Fresh tissues were obtained from SCLC patients who had undergone surgical treatment. SCLC tissue samples were cut into 2–4 mm fragments, and implanted subcutaneously into B‐NSG mice (BIOCYTOGEN, Beijing, China). The animals were monitored daily. When the tumor volume exceeded 1000–1500 mm³, animals were euthanized, and xenograft fragments were transplanted in new B‐NSG mice for passaging.

2.12. Human tissue samples

In total, 150 SCLC tissue specimens and 20 paired adjacent non‐tumor lung tissue specimens were collected from clinically diagnosed SCLC patients in Zhujiang Hospital of Southern Medical University (Guangzhou, China). None of the patients had received chemotherapy prior to obtaining the specimens. Institutional Review Board approval was obtained at the appropriate ethics committee of Southern Medical University. This study was performed in accordance with the Declaration of Helsinki.

2.13. Reagents and antibodies

The reagents and antibodies used in this study are listed in Table S3.

2.14. Data analysis

Ordinary one‐way ANOVA (multiple comparisons were included when necessary) was used to analyze the possible differences between groups. Mann–Whitney test was performed to analyze the expression differences between SCLC tissue samples and normal lung tissue samples. Kaplan–Meier analysis was performed in SPSS (V25.0, IBM, US). GraphPad Prism 8 (GraphPad Software, San Diego, CA, US) was used for statistical analyses and graph presentation.

3. RESULTS

3.1. Ligand‐independent phosphorylation of EphA2 is related to the chemoresistance of SCLC

Previous reports indicated that the expression of EPHA2 was intimately correlated with chemoresistance in SCLC. ³⁶ To further investigate the association between EPHA2 and SCLC, the expression levels of EPHA2 in SCLC and paired normal lung tissue were analyzed. RT‐qPCR assays exhibited markedly higher EPHA2 expression in SCLC tissues than paired normal lung tissues (Figure 1A). Furthermore, to evaluate the impact of EPHA2 on the clinicopathologic properties of SCLC, survival analysis was performed among 150 clinically confirmed SCLC patients. Kaplan–Meier analysis indicated that high expression levels of EPHA2 might be related to poor overall survival in SCLC (Figure 1B). However, multivariate analysis suggested that EPHA2 expression might not be an independent prognostic factor (Table S4). Considering that EphA2 signaling in cancers is complex, even diametrically opposite, ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ the results suggested that EphA2 signaling in SCLC might also be complex.

FIGURE 1.

Ligand‐independent phosphorylation of EphA2 is related to chemoresistance of SCLC. (A) mRNA expression of EPHA2 was compared in normal lung tissue (n = 20) and SCLC (n = 150) FFPE samples by RT‐qPCR. Data are presented in the form of mean ± SD based on independent biological replicates, ****p ≤ 0.0001. (B) Kaplan–Meier analysis of the overall survival of 150 patients with SCLC divided into high (n = 75) and low groups (n = 75), separated by the median level, based on EPHA2 expression levels. *p ≤ 0.05. (C) Western blot analysis shows that the expression of EphA2 (total and pS897) was enhanced in drug‐resistant sublines H69AR and H446CDDP in comparison with drug‐sensitive sublines H69 and H446, while EphA2‐pY588 showed the opposite. (D) Western blots analysis shows that the expression of ephrin‐A1 was decreased in H69AR and H446CDDP compared with H69 and H446. The expression of Akt‐pS473 showed the opposite. The expression of Akt showed no significant difference between H69 and H69AR, and between H446 and H446CDDP. (E) Western blots analysis shows that EphA2‐pS897 expression was decreased in H69AR and H446CDDP after ephrin‐A1 treatment, while EphA2‐pY588 showed the opposite. EphA2 (total) expression remained unaltered after ephrin‐A1 treatment. (F) Western blot analysis showed that Akt‐pS473 was decreased in H69AR and H446CDDP after MK2206 treatment. Akt (total) expression remained unaltered. (G) Western blot analysis shows that EphA2‐pS897 expression was decreased in H69AR and H446CDDP after MK2206 treatment. EphA2 (total) expression remained unaltered. Western blot analysis shows inhibition of Akt expression by Akt siRNA suppressed phosphorylation of EphA2‐S897 in H69AR (H) and H446CDDP (I), while EphA2 (total) expression remained unaltered. (J) Expression levels of CD44, Myc, and SOX2 were higher in drug‐resistant sublines (H69AR and H446CDDP) compared with drug‐sensitive sublines (H69 and H446). (K) Western blots show that the expression levels of N‐cadherin and vimentin were higher in drug‐resistant sublines compared with drug‐sensitive sublines, while E‐cadherin expression was the opposite. These experiments were repeated at least three times and representative images are shown.

To clarify whether EphA2 functions in a ligand‐dependent manner or a ligand‐independent manner in SCLC, the expression levels of EphA2‐pY588, EphA2‐pS897, Akt, and ephrin‐A1 were measured in SCLC cell lines. Western blots analysis showed that the expression levels of EphA2 and EphA2‐pS897 were higher in chemoresistant sublines compared with chemosensitive parent lines, whereas the expression levels of EphA2‐pY588 were the opposite (Figures 1C and S1A,B). Moreover, the activity of Akt was increased in drug‐resistant sublines compared with drug‐sensitive sublines, whereas the expression of ephrin‐A1 was the opposite (Figure 1D). These results indicated that EphA2 might be involved in the regulation of chemoresistance in SCLC. In addition, EphA2 might function in a ligand‐dependent manner in chemosensitive parent lines with ephrin‐A1, whereas EphA2 might function in a ligand‐independent manner in chemoresistant sublines containing low ephrin‐A1 levels. To further investigate the role of ephrin‐A1 on EphA2 in chemoresistant sublines, H69AR and H446CDDP cells were treated with ephrin‐A1 (Figure 1E). These results suggested that restoring ephrin‐A1 caused the phosphorylation of Y588 in EphA2 and dephosphorylation of S897 in EphA2 in chemoresistant sublines. To further verify whether EphA2‐S897 was activated by Akt in chemoresistant sublines, Akt inhibitor and siAKT were used. Western blot analysis showed that EphA2‐pS897 was decreased in H69AR and H446CDDP after treatment with the Akt inhibitor MK2206 or after interference with Akt expression (Figure 1F–I). These findings suggested that phosphorylation of EphA2‐S897 was ligand independent and could be induced by Akt in SCLC.

To reveal the connection between the cancer stemness and SCLC, the expression of stem cell biomarkers (CD44, Myc, and SOX2) was examined in SCLC cell lines. Western blot results revealed that the expression of stemness‐associated biomarkers was higher in drug‐resistant sublines compared with drug‐sensitive sublines (Figure 1J). In addition, drug‐resistant sublines displayed EMT (Figure 1K). These results indicated that the cancer stemness was also involved in the regulation of the chemoresistance in SCLC.

3.2. Wild‐type EphA2 enhances cancer stemness and chemoresistance of SCLC, while P817H mutation neutralizes intrinsic function

To further explore the effect of EphA2 on chemoresistance in SCLC, the knockdown of EPHA2 was established in drug‐resistant SCLC cell sublines H69AR (Figure 2A) and H446CDDP (Figure S2A). Cell counting kit‐8 (CCK‐8) assays showed that chemoresistance was reversed by the downregulation of EPHA2 (Figures 2B and S2B). At the same time, CD44, Myc, and SOX2 also decreased after EPHA2 was knocked down in H69AR (Figure 2C) and H446CDDP (Figure S2C). In addition, the phenotypic profile appeared to be remodeled from the mesenchymal to the epithelial phenotype (Figures 2D and S2D). Sphere formation assays showed that the self‐renewal capacity of cells was prominently attenuated in shEPHA2 groups compared with the control group (Figures 2E and S2E). These results indicated that chemoresistance and stemness were suppressed when the expression of EphA2 was interfered in chemoresistant sublines H69AR and H446CDDP.

FIGURE 2.

Wild‐type EphA2 enhances the cancer stemness and chemoresistance of SCLC, whereas the P817H mutation neutralizes intrinsic function. (A) Western blots show that the expression levels of EphA2 (total and pS897) were significantly reduced in H69AR cells, which were stably transfected with shEPHA2. (B) IC₅₀ values of epirubicin (ADM), cisplatin (CDDP), and etoposide (VP‐16) were significantly reduced in response to shEPHA2 transfection compared with the control group in H69AR cells. The number of independent biological replicates = 3, ***p ≤ 0.001, ****p ≤ 0.0001. (C) Western blots show that the expression levels of CD44, Myc, and SOX2 were decreased in H69AR cells with silenced EPHA2. (D) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in H69AR cells with silenced EPHA2, while E‐cadherin expression was increased. (E) SFE of shEPHA2 groups in H69AR cells. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (F) Western blots show the changes in expression levels of EphA2 (total and pS897) in H69 cells, which were stably transfected with wild‐type EPHA2 (WT), A785S, P817H, or Y930D mutant EPHA2. (G) IC₅₀ values of ADM, CDDP, and VP‐16 in H69 cells with upregulated wild‐type or mutant EphA2, respectively. The number of independent biological replicates = 3, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. (H) Western blots show that the expression levels of CD44, Myc, and SOX2 were increased in H69 cells overexpressing wild‐type, A785S, or Y930S mutant EphA2. (I) Western blots showed the expression levels of N‐cadherin and vimentin were increased in H69 cells overexpressing wild‐type, A785S, or Y930S mutant EphA2, while E‐cadherin expression was reduced. (J) SFE of wild‐type and mutant EphA2 groups in H69 cells. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (K) H69 cells stably overexpressed wild‐type EphA2, P817H mutant EphA2, or the control. (L) H69AR cells were stably transfected with shEPHA2 or shNC. Each group of cells was injected into mice, followed by chemotherapy (CDDP + VP‐16) or empty vehicles were injected intraperitoneally as indicated (n = 5 mice for each group). (M) Tumor weights of EphA2^WT groups and EphA2^P817H groups. Data are displayed as the mean ± SD, n = 5, *p ≤ 0.05, **p ≤ 0.01. (N) Tumor weights of shEPHA2 groups. Data are displayed as the mean ± SD, n = 5, ***p ≤ 0.001. Each group of cells was injected into mice, followed by chemotherapy (CDDP + VP‐16) or empty vehicles were injected intraperitoneally as indicated (n = 5 mice for each group). These experiments were repeated at least three times and representative images are shown.

Subsequently, we wondered whether stemness and chemoresistance of SCLC would be enhanced by EphA2 overexpression in chemosensitive counterparts, H69 and H446. Hence, H69 and H446 cell lines were transfected with wild‐type EPHA2 (Figures 2F and S2F). In agreement with our assumption, EphA2^WT prominently increased the chemoresistance of SCLC (Figures 2G and S2G). In addition, EphA2^WT increased the expression of stem cell biomarkers, facilitated EMT, and aggregated self‐renewal in H69 (Figures 2H–J) and H446 (Figure S2H–J) cells. These findings suggested that EphA2^WT magnified stemness to drive chemoresistance in SCLC.

EPHA2 mutations occurred in 3%–8% of clinical SCLC specimens, including P817H and Y930D mutations, ⁴¹ , ⁴² whereas the A785S mutation was identified in the SCLC cell line DMS454. However, the role of these mutations in SCLC remains unknown. To investigate the role of A785S, P817H, and Y930D mutations in SCLC, H69 and H446 cell lines were transfected with the mutant EPHA2 overexpression lentiviral vector (Figures 2F and S2F). Interestingly, EphA2^A785S‐pS897 and EphA2^Y930D‐pS897 were increased after transfection, whereas the phosphorylation level of EphA2‐S897 in the EphA2^P817H group was similar to that of the control group. Moreover, EphA2^A785S and EphA2^Y930D promoted stemness to induce chemoresistance in SCLC in a ligand‐independent manner in vitro, whereas EphA2^P817H neutralized the intrinsic function of EphA2^WT (Figures 2G–L and S2G–L).

Furthermore, to determine whether EphA2 promoted chemoresistance in vivo, H69 and H69AR cells with altered EphA2 expression were transplanted subcutaneously into nude mice. Tumor growth was obviously faster in the EphA2^WT groups than that in the EphA2^P817H groups and the corresponding control groups after chemotherapy treatment (Figures 2K,M and S2K). In contrast, the EPHA2 knockdown significantly slowed down tumor growth (Figures 2L,N and S2L). These results suggested that EphA2 promoted the chemoresistance of SCLC cells in vivo.

3.3. EphA2 directly interacts with PRMT1

To investigate the pathway that transduced the signals of EphA2, we performed co‐IP and liquid chromatography‐tandem mass spectrometry (Figure S3A, Table S5). In total, 142 genes were distinguished in the NC group, whereas 192 genes were identified in the EphA2 overexpression group, among which 75 genes were unique to the EphA2 overexpression group (Figure S3B). Among those genes, HDAC2, PRMT1, ROCK1, and YWHAG were involved in the regulation of the cancer stemness. ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ RT‐qPCR assays showed that PRMT1 might be regulated by EphA2 in H69AR cells, but other genes were not affected by EphA2 (Figure S3C). Western blot analysis indicated that PRMT1 was correlated with multidrug resistance, including cisplatin resistance, in SCLC cell lines (Figure 3A). Furthermore, PRMT1 and PRMT5 jointly regulated cell apoptosis in non–SCLC. ³⁰ The expression of PRMT5 was measured in SCLC cell lines, too. However, PRMT5 might not be involved in the regulation of platinum resistance (Figure 3A). Moreover, the downregulation of EphA2 reduced the expression of PRMT1 and the specific catalysate of PRMT1, H4R3me2a, ⁴⁹ but did not affect the expression of PRMT5 (Figures 3B and S3D). Importantly, ligand‐independent EphA2 activation increased the expression and activity of PRMT1 in H69 and H446 (Figures 3C and S3E). This suggested that PRMT1 might be involved in the regulation of chemoresistance of SCLC, and that EphA2 regulated the expression and methyltransferase activity of PRMT1.

FIGURE 3.

EphA2 directly interacts with PRMT1. (A) Expression of PRMT1 was higher in chemoresistant sublines H69AR and H446CDDP compared with chemosensitive sublines H69 and H446. A significant difference in PRMT5 expression was detected between H69 and H69AR, no significant difference was detected between H446 and H446CDDP. (B) Expression of PRMT1 and H4R3me2a was decreased in shEPHA2 groups compared with the control groups. (C) Expression of PRMT1 and H4R3me2a was higher in EphA2^WT, EphA2^A785S, and EphA2^Y930D groups in comparison with EphA2^P817H and control groups in H69. (D,E) Co‐IP was performed in H69 cells using anti‐FLAG and anti‐PRMT1 antibodies, which were transfected with plasmids of FLAG‐tagged wild‐type and mutant EPHA2 or negative control. The co‐IP products were then analyzed by western blot. (F–I) Co‐IP was performed in H69 and H69AR cells using anti‐EphA2 and anti‐PRMT1 antibodies. The co‐IP products were then analyzed by western blot. (J) GST‐pull down shows that wild‐type EphA2 directly interacted with PRMT1. (K) Immunofluorescence staining shows the subcellular localization of EphA2 (total and pS897) in H69AR and H446 cells overexpressing wild‐type or P817H mutant EphA2. Scale bars = 10 μm. (L) Immunofluorescence staining shows that wild‐type EphA2 colocalized with PRMT1 in H69AR and H446 cells overexpressing wild‐type EphA2. Scale bar = 10 μm. These experiments were repeated at least three times and representative images are shown.

To further confirm whether EphA2 interacted with PRMT1, a co‐IP assay was conducted in SCLC cell lines. The results indicated that PRMT1 preferentially interacted with EphA2^WT rather than EphA2^P817H (Figure 3D,E). Moreover, the co‐IP assay also revealed that, whereas EphA2^WT‐pS897 remained at a high level, EphA2^P817H‐pS897 was significantly restricted. Endogenous EphA2 expressed a similar affinity to interact with PRMT1 (Figure 3F–I). In addition, the GST pull‐down assay also confirmed the direct binding between EphA2 and PRMT1, as well as the attraction of PRMT1 toward EphA2^WT (Figure 3J). These results, for the first time, illustrate that EphA2 directly interacts with PRMT1 at the protein level.

PRMT1 is mainly located in the nucleus. Many kinases transport to the nucleus after phosphorylation. ⁵⁰ , ⁵¹ Therefore, we wondered whether the nuclear translocations of EphA2^WT or EphA2^P817H were different. An immunofluorescence assay showed that the EphA2^WT and EphA2^WT‐pS897 proteins were primarily present in the nucleus and colocalized with PRMT1, whereas the EphA2^P817H protein was mainly restricted to the cytoplasm and was minimally adjacent to the subcellular distribution of PRMT1 (Figure 3K,L). These findings indicated that the interaction between EphA2 and PRMT1 depended on the ligand‐independent phosphorylation of EphA2.

3.4. PRMT1 augments stemness and chemoresistance in SCLC

The role of PRMT1 in SCLC remains unknown. To investigate the effect of PRMT1 on chemoresistance in SCLC cell lines, PRMT1 was overexpressed in H69 and H446 (Figures 4A and S4A). In concert with the overexpression of PRMT1, the expression levels of stem cell biomarkers were also increased (Figures 4B and S4B). Intriguingly, in addition, SCLC cell lines were found to expedite EMT and a robust self‐renewal capacity (Figures 4C,D and S4C,D). Similarly, PRMT1 overexpression promoted chemoresistance in SCLC cell lines (Figures 4E and S4E). These findings suggested that PRMT1 could augment the cancer stemness to enhance chemoresistance in SCLC.

FIGURE 4.

PRMT1 augments stemness and chemoresistance in SCLC. (A) Western blots show the changes in expression levels of PRMT1 in H69 cells, which were stably transfected with His‐tagged PRMT1. (B) Western blots show that the expression levels of CD44, Myc, and SOX2 were increased in H69 cells overexpressing PRMT1. (C) Western blots show that the expression levels of N‐cadherin and vimentin were increased in H69 cells overexpressing PRMT1, while E‐cadherin expression was reduced. (D) SFE of PRMT1 groups in H69 cells. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (E) IC₅₀ values of ADM, CDDP, and VP‐16 in H69 cells with upregulated PRMT1. The number of independent biological replicates = 3, ***p ≤ 0.001. (F) Western blots show that the expression levels of PRMT1 were reduced in H69 cells with co‐transfected wild‐type or P817H mutant EPHA2 and shPRMT1. (G) Western blots show that the expression levels of CD44, Myc, and SOX2 were reduced in H69 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1. (H) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in H69 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1, whereas E‐cadherin expression was increased. (I) SFE of shPRMT1 groups in H69 cells with wild‐type or mutant EPHA2. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (J) IC₅₀ values of ADM, CDDP, and VP‐16 in H69 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1. The number of independent biological replicates = 3, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. These experiments were repeated at least three times and representative images are shown.

To further evaluate whether the knockdown of PRMT1 essentially compromised the viability of EphA2‐expressing cells, we co‐transfected H69 and H446 cells with two types of lentiviral vectors: one vector containing EPHA2 ^WT , EPHA2 ^P817H , or an empty control (Mock), and the other constructed with shPRMT1 or shNC (Figures 4F and S4F). H69 and H446 cells that were transfected with EPHA2 ^WT and EPHA2 ^P817H vectors showed greater stemness inhibition or sensitivity to shPRMT1‐mediated toxicity of chemotherapy than the shNC‐transfected cells (Figures 4G–J and S4G–J). These results revealed for the first time that ligand‐independent EphA2 phosphorylation built up stemness to strengthen chemoresistance through PRMT1 in SCLC.

3.5. PRMT1 combines and methylates SOX2

How PRMT1 regulated the chemoresistance of SCLC was still elusive. PRMT1 methylated SOX2 at R90, R98, R113, and R115 positions in embryonic stem cells. ²⁹ However, whether PRMT1 interacts with SOX2 was yet to be explored.

To further explore the relationship between PRMT1 and SOX2 in SCLC, the interaction between them was examined by co‐IP assay. The results indicated that endogenic PRMT1 interacted with SOX2 (Figure 5A–D). Moreover, immunofluorescence staining revealed a strong colocalization of PRMT1 and SOX2 in the nucleus (Figure 5E). These results revealed the interaction between PRMT1 and SOX2. Moreover, our results also demonstrated that SOX2 underwent methylation in H69AR (Figure 5F). To identify whether SOX2 was methylated by PRMT1, a small molecule inhibitor of PRMT1, AMI‐1, was used to inhibit the activity of PRMT1 (Figure 5G). The expression levels of both total SOX2 and methylated SOX2 were fundamentally decreased after H69AR was incubated with AMI‐1 for 48 h (Figure 5H). To identify the precise methylation sites of SOX2, we purified SOX2 protein using co‐IP after H69AR was treated with empty vehicle or AMI‐1. The prominent band for SOX2 from the SDS‐PAGE gel was excised and subjected to mass spectrometry, which led to the identification of R43 as a novel methylation site on SOX2 (Figure 5I). These results, for the first time, clarified that PRMT1 not only interacted with SOX2, but also methylated SOX2 at R43, and that the EphA2/PRMT1/SOX2 pathway depended on the activity of PRMT1.

FIGURE 5.

PRMT1 combines and methylates SOX2. (A–D) Co‐IP was performed in H69 and H69AR cells using anti‐PRMT1 and anti‐SOX2 antibodies. The co‐IP products were then analyzed by western blot. (E) Immunofluorescence staining shows that PRMT1 colocalized with SOX2 in H69AR, H446, and H446CDDP cells. Scale bars = 10 μm. (F) Co‐IP was performed in H69AR cells using anti‐asymmetric di‐methyl arginine motif (aDMA) and IgM antibodies. The co‐IP products were then analyzed by western blot using an anti‐SOX2 antibody. (G) Western blots show that the expression level of H4R3me2a was reduced in response to treatment with AMI‐1 for 48 h in a dose‐dependent manner in H69AR cells. PRMT1 expression remained unaltered. (H) Co‐IP was performed in H69AR cells using anti‐aDMA antibody and IgM. H69AR cells were treated with empty vehicle or AMI‐1 for 48 h. The co‐IP product was analyzed by western blotting using an anti‐SOX2 antibody. (I) Mass spectrometry analysis of arginine methylation sites of SOX2 protein. These experiments were repeated at least three times and representative images are shown.

3.6. Inhibiting ligand‐independent EphA2 phosphorylation suppresses the growth of SCLC in vitro and in vivo

To further investigate whether chemoresistance of SCLC would be affected only when the catalytic activities of EphA2 and PRMT1 were inhibited, a small molecule inhibitor of EphA2, ALW‐II‐41‐27, and small molecule inhibitor of PRMT1, AMI‐1, were used. First, H69AR and H446CDDP cells were treated with ALW‐II‐41‐27. EphA2‐pS897 was significantly reduced after ALW‐II‐41‐27 treatment (Figures 6A and S5A). CCK‐8 assay indicated that IC₅₀ values of epirubicin, cisplatin, and etoposide were significantly reduced in response to ALW‐II‐41‐27 treatment in chemosensitive parent lines and chemoresistant sublines (Figures 6C and S5C). Combination indexes suggested that the combination of ALW‐II‐41‐27 and the chemotherapeutic drug produced synergistic anti‐tumor effects (Figure S5E). In addition, dasatinib, another small molecule inhibitor of EphA2, similarly exhibited synergistic anti‐tumor effects after combination with a chemotherapeutic drug in chemosensitive sublines and chemoresistant sublines (Figures 6B,D and S5B,D,F).

FIGURE 6.

Inhibiting ligand‐independent EphA2 suppresses the growth of SCLC in vitro and in vivo. (A) Western blots show that the expression level of EphA2‐pS897 was reduced in response to treatment with ALW‐II‐41‐27 (ALW) for 2 h in a dose‐dependent manner in H69AR cells. The total EphA2 expression remained unaltered. (B) Western blots show that the expression level of EphA2‐pS897 was reduced in response to treatment with dasatinib (Dasa) for 24 h in a dose‐dependent manner in H69AR cells. The total EphA2 expression remained unaltered. (C) IC₅₀ values of ADM, CDDP, and VP‐16 were reduced in H69 and H69AR cells. H69 and H69AR cells were treated with empty vehicle, or 0.75 or 1.5 μmol/L of ALW‐II‐41‐27. The number of independent biological replicates = 3, ***p ≤ 0.001, ****p ≤ 0.0001. (D) IC₅₀ values for ADM, CDDP, and VP‐16 were reduced in H69 and H69AR cells. H69 and H69AR cells were treated with empty vehicle, 2.5 or 5 μmol/L of dasatinib. The number of independent biological replicates = 3, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. (E) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in response to treatment with ALW‐II‐41‐27, in H69AR cells, while E‐cadherin expression was increased. (F) Western blots show the changes in expression levels of CD44, Myc, SOX2, and PRMT1, in response to ALW‐II‐41‐27 treatment, in H69AR cells. (G) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in response to the treatment with AMI‐1, in H69AR cells, while E‐cadherin expression was increased. (H) Western blots show the changes in expression levels of CD44, Myc, and SOX2, in response to treatment with AMI‐1, in H69AR cells. (I) SFE of H69AR cells, which were treated with ALW‐II‐41‐27. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (J) Immunofluorescence shows the changes in subcellular localization of EphA2 and EphA2‐pS897 in response to treatment with ALW‐II‐41‐27, in H446 cells overexpressing wild‐type or P817H EphA2. Scale bars = 10 μm. (K) Effect of chemotherapy (CDDP + VP‐16), ALW, or the combination of ALW‐II‐41‐27 and chemotherapy on tumor growth using SCLC PDX models. (L) Effect of chemotherapy (CDDP + VP‐16), ALW‐II‐41‐27, or the combination of ALW‐II‐41‐27 and chemotherapy on tumor growth in vivo. Nude mice were engrafted with H69AR cells subcutaneously. These experiments were repeated at least three times and representative images are shown.

Moreover, ALW‐II‐41‐27 substantially reversed EMT, reduced the expression of stem cell biomarkers, and inhibited the self‐renewal capability in H69AR and H446CDDP (Figures 6E,F,I and S5G–I). In addition, ALW‐II‐41‐27 also reduced the expression of PRMT1 and SOX2, (Figures 6F and S5H). Immunofluorescence staining revealed that the nuclear translocation of total EphA2 and EphA2‐pS897 was blocked by ALW‐II‐41‐27 in the EphA2^WT group, but not in the EphA2^P817H group (Figure 6J). These findings suggested that the EphA2/PRMT1/SOX2 pathway relied on the phosphorylation of EphA2‐S897. In contrast, ALW‐II‐41‐27 treatment effectively suppressed the cancer stemness of SCLC, enhanced the sensitivity to chemotherapeutic drugs, and blocked the EphA2/PRMT1/SOX2 pathway in vitro.

In addition, a small molecule inhibitor of PRMT1, AMI‐1, substantially reversed EMT and reduced the expression of stem cell biomarkers, including SOX2, in H69AR and H446CDDP (Figures 6G,H and S5J–L). These results indicated that AMI‐1 inhibited stemness in SCLC and the EphA2/PRMT1/SOX2 pathway depending on the activity of PRMT1.

To further determine whether ALW‐II‐41‐27 modulated chemoresistance of SCLC in vivo, PDX models and subcutaneous xenograft models generated by H69AR and H446CDDP were established. Tumor growth was effectively restrained in mice that were given chemotherapeutic drugs ALW‐II‐41‐27 or the combined treatment, in vivo. Additionally, the combined usage of ALW‐II‐41‐27 and chemotherapeutic drugs exerted an intensified inhibitory regulation on tumor growth (Figures 6K,L and S5M–S). Collectively, these results revealed a potentially synergistic effect of ALW‐II‐41‐27 with chemotherapy in SCLC models, both in vitro and in vivo.

4. DISCUSSION

In this study, we found that ephrin‐A1 was overexpressed in chemosensitive SCLC cell lines, whereas Akt‐pS473 has a low expression level. High expression of ephrin‐A1 caused phosphorylation on Y588 in EphA2 and dephosphorylation on S897 in EphA2 in SCLC cell lines. In addition, SCLC cell lines developed chemoresistance, expression of ephrin‐A1 was markedly reduced, and Akt‐pS473 levels increased. The repression of EphA2‐pS897 by ephrinA1 was alleviated and ligand‐independent phosphorylation of EphA2‐pS897 was activated by Akt. We further demonstrated that EphA2 enhanced stemness to induce chemoresistance in a ligand‐independent manner, whereas P817H mutation neutralized intrinsic function. Furthermore, EphA2 was transported into the nucleus and directly interacted with PRMT1 to increase the expression and activity of PRMT1. Then, PRMT combined and methylated SOX2 to fortify stemness and chemoresistance. The EphA2/PRMT1/SOX2 pathway depended on the activity of EphA2 and PRMT1. Both ALW‐II‐41‐27 and AMI‐1 blocked the pathway in vitro. Last, the combination of ALW‐II‐41‐27 and chemotherapeutic drugs showed a synergistic anti‐tumor effect in vitro and in vivo.

Upregulation of EphA2 is found in many types of cancers. ⁵² , ⁵³ Ligand‐independent phosphorylation of EphA2‐S897 is mediated by kinases including Akt, RSK, and PKA, and other RTKs. ⁵⁴ , ⁵⁵ , ⁵⁶ Ligand‐independent EphA2 signaling contributes to therapy resistance including chemoresistance, immunotherapy resistance, and targeted therapy resistance in tumors. ¹⁸ , ²⁷ , ⁴⁸ , ⁵⁷ Ligand‐independent activated EphA2 induces chemoresistance in gastric cancer by interacting with YAP. ⁵⁸ The RSK1/2‐EphA2‐GPRC5A axis contributes to chemoresistance in high‐grade serous ovarian cancer. ⁴⁸ The genetic and pharmacological inhibition of EphA2 suppresses the proliferation of SCLC in vitro. ⁵⁹ Epigenome‐wide DNA methylation analysis revealed that EPHA2 was associated with the response to Aurora kinase inhibitors and a PLK‐1 inhibitor in SCLC. ⁶⁰ However, the role of EphA2 in the chemoresistance of SCLC was unclear. Moreover, this study demonstrated that EphA2 enhanced stemness to induce chemoresistance in SCLC and that the P817H mutation was a loss‐of‐function mutation. These studies suggested that EphA2 regulated the chemoresistance of SCLC through multiple approaches (e.g., proliferation and stemness) and that EphA2 might be a hub regulator for therapy resistance in SCLC, and even in other malignant tumors.

Furthermore, ALW‐II‐41‐27 is an ATP‐competitive EphA2 inhibitor, which can inhibit the activity of EphA2‐Y588, but theoretically cannot inhibit the activity of EphA2‐S897. ⁶¹ A series of evidence manifested that ALW‐II‐41‐27 was a multikinase inhibitor that inhibited the activity of Akt and other kinases, in addition to EphA2. ²⁴ , ⁶² In tandem with other studies, ²⁵ , ⁵⁶ our work indicated that Akt activated the phosphorylation of EphA2 at S897. Hence, we concluded that ALW‐II‐41‐27 inhibited the phosphorylation of EphA2‐S897 by inhibiting the activity of Akt. In fact, ALW‐II‐41‐27 suppresses the phosphorylation of EphA2‐S897 and Akt in renal cell carcinoma and melanoma, ²⁴ , ⁶² which supported our inference. Hence, the expression and activity of Akt and EphA2 must be considered and detected before applying ALW‐II‐41‐27. Alternatively, it may be counterproductive. Although the biological effects produced by ALW‐II‐41‐27 were comprehensive, resulting in acting on all targets, our results certified that EphA2‐S897 was a vital indirect target for ALW‐II‐41‐27, and that EphA2 was a potential therapeutic target in SCLC.

PRMT1 regulates the chemoresistance of multiple malignant tumors including ovarian cancer, esophageal cancer, and pancreatic cancer. ⁵³ , ⁶³ , ⁶⁴ , ⁶⁵ However, the role of PRMT1 in SCLC was unclear. In this study, we demonstrated that EphA2‐mediated PRMT1 activation promoted the expression and molecular functions of SOX2. PRMT1 methylated arginine residues of Axin, which increased the stability of Axin. ⁶⁶ Similarly, PRMT1‐mediated methylation might intensify the expression of SOX2 by stabilizing SOX2; SOX2 plays a crucial role in maintaining the stemness and drug resistance of tumors including SCLC. ⁵³ , ⁶⁷ , ⁶⁸ , ⁶⁹ Overall, PRMT1 plays an important role in signal transduction and regulation of chemoresistance in malignant tumors including SCLC. Furthermore, PRMT1 contributes to the major type I protein‐arginine methyltransferase activity in mammalian cells and tissues. ⁷⁰ , ⁷¹ Hence, the focus and difficulty in the development of new PRMT1‐targeted drugs in the future will be how to improve the selectivity of drugs.

In conclusion, this study demonstrates for the first time that the EphA2/PRMT1/SOX2 signaling axis is a novel pathway regulating the chemoresistance of SCLC and that EphA2 is a potential therapeutic target in SCLC. Our study validates that the combination of ALW‐II‐41‐27 and chemotherapeutic drugs produces a synergistic anti‐tumor effect in SCLC. This study provides a potential treatment strategy for patients with SCLC, especially those with chemoresistance.

AUTHOR CONTRIBUTIONS

Conceptualization: Shumei Liang, Juan Peng, and Linlang Guo; methodology: Shumei Liang, Qiuping Wang, and Yang Wen; funding acquisition: Juan Peng, Linlang Guo, and Qiongyao Wang; resources and material support: Yu Wang and Man Li; data curation and formal analysis: Shumei Liang, Qiuping Wang, Yang Wen, Yu Wang, and Man Li; technical support: Yu Wang, Man Li, and Qiongyao Wang; writing (original draft): Shumei Liang; writing (review and editing): Juan Peng and Linlang Guo. All authors read and approved the final paper.

CONFLICT OF INTEREST

The authors declare no potential conflicts of interest. None of the authors of this manuscript are current Editors or Editorial Board Members of Cancer Science.

ETHICAL APPROVAL

The current research was approved and supervised by the Ethics Committee of Zhujiang Hospital of Southern Medical University (Approval: 2017‐BLK‐001). All mouse studies were conducted according to the Institutional Animal Care and Use Committee (IACUC)‐approved animal protocols in accordance with the Southern Medical University's institutional guidelines (Approval: LAEC‐2020‐063). All patients provided informed consent before the study began, and clinical tissue samples were anonymized and obtained under the approval of the Zhujiang Hospital of Southern Medical University Institutional Review Board. This study was performed in accordance with the Declaration of Helsinki. Registry and the registration no. of the study/trial: N/A.

Supporting information

FIGURE S1 H69AR and H446CDDP are multidrug resistance cell lines. (A) IC₅₀ values of cisplatin (CDDP), etoposide (VP‐16) and irinotecan (CTP‐11) were higher in H69AR cells compared with H69 cells. The number of independent biological replicates = 3, *p ≤ 0.05, ***p ≤ 0.001. (B) IC₅₀ values of cisplatin (CDDP), carboplatin (CBDCA), etoposide (VP‐16) and irinotecan (CTP‐11) were higher in H446CDDP cells compared with H446 cells. The number of independent biological replicates = 3, *p ≤ 0.05, **p ≤ 0.01****p ≤ 0.0001.

FIGURE S2 Wild‐type EphA2 enhances the the cancer stemness and chemoresistance of small‐cell lung cancer (SCLC), while P817H mutation neutralizes intrinsic function. (A) Western blots showed that the expression levels of EphA2 (total and pS897) were significantly reduced in H446CDDP cells, which were stably transfected with shEPHA2. (B) IC₅₀ values of epirubicin (ADM), cisplatin (CDDP) and etoposide (VP‐16) were significantly reduced in response to shEPHA2 transfection compared with the control group in H446CDDP cells. The number of independent biological replicates = 3, ****p ≤ 0.0001. (C) Western blots showed that the expression levels of CD44, Myc, and SOX2 were decreased in H446CDDP cells with silenced EPHA2. (D) Western blots showed the expression levels of N‐cadherin and vimentin were reduced in H446CDDP cells with silenced EPHA2, while E‐cadherin expression was increased. (E) Sphere formation efficiency (SFE) of shEPHA2 groups in H446CDDP cells. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (F) Western blots showed the changes in expression levels of EphA2 (total and pS897) in H446 cells with upregulated wild‐type or mutant EphA2, respectively. (G) IC₅₀ values of ADM, CDDP, and VP‐16 in H446 cells with upregulated wild‐type or mutant EphA2, respectively. The number of independent biological replicates = 3, ***p ≤ 0.001, ****p ≤ 0.0001. (H) Western blots showed that the expression levels of CD44, Myc, and SOX2 were increased in H446 cells with upregulated wild‐type, A785S, or Y930D mutant EphA2, respectively. (I) Western blots showed that the expression levels of N‐cadherin and vimentin were increased in H446 cells with upregulated wild‐type, A785S, or Y930D mutant EphA2, while E‐cadherin expression was reduced. (J) SFE of wild‐type and mutant EphA2 groups in H446 cells. Data are presented as the mean ± SD based on three independent biological replicates, *p ≤ 0.05, ***p ≤ 0.001. (K) The growth curve of tumor volumes of the wild‐type or P817H mutant EphA2 groups. Data are displayed as the mean ± SD, n = 5, ****p ≤ 0.0001. (L) The growth curve of tumor volumes of the shEPHA2 groups. Data are presented in the form of mean ± SD, n = 5, ****p ≤ 0.0001. These experiments were repeated at least three times and representative images are shown.

FIGURE S3 PRMT1 co‐expresses with EphA2. (A) FLAG‐tag negative control and EPHA2 overexpression plasmids were transiently transfected into H69, followed by co‐immunoprecipitation (co‐IP) of anti‐FLAG antibody. The co‐IP products were then analyzed using western blot and liquid chromatography tandem mass spectrometry. (B) Venn diagram showed that 75 genes were unique to the EphA2 overexpression group compared with the control group. (C) RT‐qPCR assays showed that PRMT1 mRNA expression was reduced in H69AR cells with silenced EPHA2, but that other gene (HDAC1, ROCK1, and YWHAG) mRNA expression remained unaltered. (D) Western blots showed that the expression levels of PRMT1 and H4R3me2a were decreased in H446CDDP cells with silenced EPHA2. (E) Western blots showed that the expression of PRMT1 and H4R3me2a was higher in EphA2^WT, EphA2^A785S, and EphA2^Y930D groups in comparison with the EphA2^P817H and control groups in H446. These experiments were repeated at least three times and representative images are shown.

FIGURE S4 PRMT1 augments stemness and chemoresistance in small‐cell lung cancer (SCLC). (A) Western blots show the changes in expression levels of PRMT1 in H446 cells, which were stably transfected with His‐tagged PRMT1. (B) Western blots show that the expression levels of CD44, Myc, and SOX2 were increased in H446 cells overexpressing PRMT1. (C) Western blots show that the expression levels of N‐cadherin and vimentin were increased in H446 cells overexpressing PRMT1, while E‐cadherin expression was reduced. (D) Sphere formation efficiency (SFE) of PRMT1 groups in H446 cells. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (E) IC₅₀ values of ADM, CDDP, and VP‐16 in H446 cells with upregulated PRMT1. The number of independent biological replicates = 3, ***p ≤ 0.001. (F) Western blots show that the expression levels of PRMT1 were reduced in H446 cells with co‐transfected wild‐type or P817H mutant EPHA2 and shPRMT1. (G) Western blots show that the expression levels of CD44, Myc, and SOX2 were reduced in H446 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1. (H) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in H446 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1, while E‐cadherin expression was increased. (I) SFE of shPRMT1 groups in H446 cells with wild‐type or mutant EPHA2. Data are presented in the form of mean ± SD based on three independent biological replicates, ****p ≤ 0.0001. (J) IC₅₀ values of ADM, CDDP, and VP‐16 in H446 cells with co‐transfected wild‐type or mutant EPHA2 and shPRMT1. The number of independent biological replicates = 3, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. These experiments were repeated at least three times and representative images are shown.

FIGURE S5 Inhibiting ligand‐independent EphA2 suppresses the growth of small‐cell lung cancer (SCLC) in vitro and in vivo. (A) Western blots show that the expression levels of EphA2‐pS897 were reduced in response to treatment with ALW‐II‐41‐27 (ALW) for 2 h in a dose‐dependent manner in H446CDDP cells. The total EphA2 expression remained unaltered. (B) Western blots show that the expression level of EphA2‐pS897 was reduced in response to treatment with dasatinib (Dasa) for 24 h in a dose‐dependent manner in H446CDDP cells. The total EphA2 expression remained unaltered. (C) IC₅₀ values of ADM, CDDP, and VP‐16 were reduced in H446 and H446CDDP cells. H446 and H446CDDP cells were treated with empty vehicle, or 0.75 or 1.5 μmol/L of ALW‐II‐41‐27. The number of independent biological replicates = 3, ***p ≤ 0.001, ****p ≤ 0.0001. (D) IC₅₀ values of ADM, CDDP, and VP‐16 were reduced in H446 and H446CDDP cells. H446 and H446CDDP cells were treated with empty vehicle, or 2.5 or 5 μmol/L of dasatinib. The number of independent biological replicates = 3, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. (E) Combination index (CI) of ALW‐II‐41‐27 and chemotherapeutic drugs. CI ≤ 0.3, strong synergy; CI = 0.3–0.9, synergy; CI = 0.9–1.1, additive; CI ≥1.1, antagonism. (F) CI of dasatinib and chemotherapeutic drugs. (G) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in response to treatment with ALW‐II‐41‐27, in H446CDDP cells, while E‐cadherin expression was increased. (H) Western blots show the changes in expression levels of CD44, Myc, SOX2, and PRMT1 in response to ALW‐II‐41‐27 treatment in H446CDDP cells. (I) SFE of H446CDDP cells, which were treated with ALW‐II‐41‐27. Data are presented in the form of mean ± SD based on three independent biological replicates, **p ≤ 0.01, ***p ≤ 0.001. (J) Western blots show that the expression level of H4R3me2a was reduced in response to treatment with AMI‐1 for 48 h in a dose‐dependent manner in H446CDDP cells. PRMT1 expression remained unaltered. (K) Western blots show the changes in expression levels of CD44, Myc and SOX2 in response to treatment with AMI‐1, in H446CDDP cells. (L) Western blots show that the expression levels of N‐cadherin and vimentin were reduced in response to treatment with AMI‐1, in H446CDDP cells, while E‐cadherin expression was increased. (M) Growth curve of tumor volumes in each group of SCLC patient‐derived xenograft (PDX) models. Data are presented in the form of mean ± SD, n = 3. ****p ≤ 0.0001. (N) Tumor weights of all groups in PDX models. Data are displayed as the mean ± SD, n = 3. *p ≤ 0.05, ****p ≤ 0.0001. (O) Growth curve of tumor volumes in each group of H69AR models. Data are presented in the form of the mean ± SD, n = 4. ****p ≤ 0.0001. (P) Tumor weights of H69AR groups. Data are displayed as the mean ± SD, n = 4. *p ≤ 0.05, ***p ≤ 0.001, ****p ≤ 0.0001. (Q) Growth curve of tumor volumes in each group of H446CDDP models. Data are presented in the form of mean ± SD, n = 4. ****p ≤ 0.0001. (R) Tumor weights of H446CDDP groups. Data are presented in the form of mean ± SD, n = 4. *p ≤ 0.05, ****p ≤ 0.0001. (S) Effect of ALW‐II‐41‐27, chemotherapy (CDDP + VP‐16), or a combination of ALW‐II‐41‐27 and chemotherapy on tumor growth in vivo. Nude mice were engrafted with H446CDDP cells subcutaneously. These experiments were repeated at least three times and representative images were shown.

TABLE S1 Sequences of shRNA and siRNA

TABLE S2 Sequences of RT‐qPCR primers

TABLE S3 Reagents and antibodies

TABLE S4 Univariate analyses of clinicopathologic parameters in association with the overall survival rate in small‐cell lung cancer (SCLC) patients

TABLE S5

Liang S, Wang Q, Wen Y, et al. Ligand‐independent EphA2 contributes to chemoresistance in small‐cell lung cancer by enhancing PRMT1‐mediated SOX2 methylation. Cancer Sci. 2023;114:921‐936. doi: 10.1111/cas.15653