Licochalcone A inhibits EGFR signalling and translationally suppresses survivin expression in human cancer cells

Dysfunction of epidermal growth factor receptor (EGFR) signalling plays a critical role in the oncogenesis of non–small‐cell lung cancer (NSCLC). Here, we reported the natural product, licochalcone A, exhibited a profound anti‐tumour efficacy through directly targeting EGFR signalling. Licochalcone A inhibited in vitro cell growth, colony formation and in vivo tumour growth of either wild‐type (WT) or activating mutation EGFR‐expressed NSCLC cells. Licochalcone A bound with L858R single‐site mutation, exon 19 deletion, L858R/T790M mutation and WT EGFR ex vivo, and impaired EGFR kinase activity both in vitro and in NSCLC cells. The in silico docking study further indicated that licochalcone A interacted with both WT and mutant EGFRs. Moreover, licochalcone A induced apoptosis and decreased survivin protein robustly in NSCLC cells. Mechanistically, we found that treatment with licochalcone A translationally suppressed survivin through inhibiting EGFR downstream kinases ERK1/2 and Akt. Depletion of the translation initiation complex by eIF4E knockdown effectively inhibited survivin expression. In contrast, knockdown of 4E‐BP1 showed the opposite effect and dramatically enhanced survivin protein level. Overall, our data indicate that targeting survivin might be an alternative strategy to sensitize EGFR‐targeted therapy.

novel approaches to counteract resistance is an urgent demand in clinic treatment.
Many chemotherapy drugs exhibited adverse side effects.
Identification of new anti-tumour chemicals from natural compounds might be a safer alternative for anti-tumour treatment.
Licochalcone A (Lico A), a flavonoid extracted from licorice root, exerts a wide range of pharmacological effects in the treatment of human diseases, including inflammation, infections and gastric ulcers. Recently, the in vitro and in vivo studies have demonstrated that licochalcone A inhibits multiple human solid tumours, 8 including lung, 9 gastric, 10 prostate, 11 liver 8 and ovarian cancer. 12 Licochalcone A promotes cell-cycle arrest, induces apoptosis, reduces angiogenesis and metastasis, etc. [13][14][15] However, the direct targets of licochalcone A in human cancer cells have not been elaborated, and the effect of licochalcone A on EGFR signalling has not been reported.
In the present study, we demonstrated that licochalcone A suppressed the activation of both wild-type and mutant EGFRs, and translationally suppressed survivin expression in NSCLC cells. The specific targeting of the EGFR-survivin axis might provide opportunities for NSCLC prevention and treatment.

| MTS assay
Cells were suspended at the concentration of 2 × 10 4 , and 100 μL of aliquot was seed into 96-well plates, followed by incubation with licochalcone A or osimertinib at various time-points. Cell viability was analysed with MTS assay (Promega).

| Lentiviral package and stable lines generation
The EGFR cDNA clones, including L858R EGFR, L858R/T790M EGFR, Del E746-A750 EGFR and wild-type EGFR, were obtained from Origene and used for virus package. For virus infection, polybrene (8 μg/mL) and virus supernatant were added to cell culture medium and maintained for 24 hours. The infected cells were selected with puromycin (1 μg/mL) for 7 days. The CRISPR-Cas9-based EGFR knockout was performed with the EGFR sgRNA (TGAGCTTGTTACTCGTGCCT) following the standard protocols.

| Anchorage-independent growth
The colony formation in soft agar was performed as described previously. 17 Cells were suspended (8000 cells/mL) in 1 mL of 0.3% agar with Eagle's basal medium containing 1% antibiotics, 10% FBS, and osimertinib or different concentrations of licochalcone A overlaid into six-well plates containing a 0.6% agar base. The cultures were maintained for 2 weeks in a 37°C, 5% CO 2 incubator.

| Flow cytometry
NSCLC cells were treated with osimertinib or licochalcone A and dissociated with trypsin. Flow cytometry analysis was conducted as described previously. 18 Briefly, cells were washed with PBS and centrifuged, followed by suspending at a final concentration of 1 × 10 6 cells/mL. The staining buffer which contains propidium iodide and Annexin V (5 µL) was added to the cell suspension and incubated for 15 minutes in the dark. The apoptotic cells were determined with a FACSort Flow Cytometer (BD, San Jose, CA, USA).
Briefly, the active EGFRs (100 ng) were incubated with 500 μmol/L angiotensin II, osimertinib or licochalcone A for 5 minutes at room temperature. The ATP mixture (0.25 μmol/L ATP and 25 mmol/L MgAc containing 10 μCi [γ-32P] ATP) was added into the reaction and incubated at 30°C for 15 minutes, and transferred onto P81 papers. The papers were washed subsequently with 0.75% phosphoric acid and acetone. The scintillation counter was used for radioactive incorporation analysis.

| Molecular modelling
Homology modelling: The three-dimensional structure of exon 19 deletion mutation (residues 696-984) EGFR was modelled based on the wild-type (WT) EGFR crystal structure using Modeller, 20

| Immunohistochemical (IHC) staining
IHC staining for xenograft tumour tissues was performed as described previously. 22 Briefly, tissue sections were deparaffinized and rehydrated, followed by immersing into boiling sodium citrate buffer (10 mmol/L, pH 6.0) for 10 minutes for antigen retrieval. The tissue section was washed with PBS twice and incubated with 3% H 2 O 2 in methanol for 10 minutes. Tissues were blocked with 50% goat serum albumin and incubated with the primary antibody overnight in a humidified chamber at 4°C, followed by hybridizing with the secondary antibody at room temperature for 1 hour. The target protein was visualized with DAB substrate.

| Statistical analysis
Statistical analysis was performed with SPSS 16.0 (SPSS, Inc). The quantitative data were expressed as means ± SD. Student's t test or one-way ANOVA was used for significant differences determination. A probability value less than 0.05 was used as the criterion for statistical significance.

| Licochalcone A effectively inhibits the growth of both osimertinib-sensitive and osimertinibresistant NSCLC cells
Previous studies have demonstrated that licochalcone A ( Figure  . The ligands were shown in the sticks, whereas proteins were depicted in cartoon representation with key residues indicated as grey sticks. Hydrogen bonds are shown as red dashed lines, and the cation-π interaction is shown as yellow dashed line. Besides, the relating residues or ligands were well labelled. C, Licochalcone A inhibits EGFR signalling in NSCLC cells. The NSCLC cells were treated with licochalcone A and osimertinib, and whole-cell extracts (WCE) were subjected to IB analysis. D, Ba/F3 cells stable cells were pre-treated with licochalcone A or osimertinib for 2 h and then treated with EGF for 15 min. WCE was collected and subjected to IB analysis osimertinib, markedly suppressed colony formation of A549 cells dose-dependently ( Figure 1G-J, Figure S1A-C). These results indicate that licochalcone A suppresses the growth of either WT or mutant EGFR expression NSCLC cells, but has no obvious cytotoxic effect on non-tumour lung cells.

| Licochalcone A binds and inhibits EGFRs activities
To better understand the anti-tumour mechanism of licochalcone A, we examined whether licochalcone A could affect EGFR signalling pathway. The in vitro EGFR kinase assays showed that either licochalcone A or osimertinib significantly suppressed the activity of the activating mutant EGFRs, including EGFR Del E746-A750, EGFR L858R/T790M and EGFR L858R (Figure 2A deletion might have changed the shape of the pocket, in which licochalcone was predicted to interact with Met793, Thr790 and Glu762 by hydrogen bonding ( Figure 3B). Our data indicate that licochalcone might be a good hit, especially for designing novel EGFR inhibitors with selectivity to different mutation types.
Immunoblotting analysis suggested that EGFR activity was decreased in response to licochalcone A treatment in NSCLC cells.
However, licochalcone A can not significantly suppress the activity of WT EGFR at the dose of 5 μmol/L, which is consistent with our in vitro kinase assay that a higher concentration of licochalcone A was required for blocking of WT EGFR activity ( Figure 3C). Akt and ERK kinases are two primary downstream targets of EGFR kinase. Our data indicated that both licochalcone A and osimertinib dramatically inhibited the phosphorylation of Akt and ERK1/2 in HCC827, H1975 and H3255 cells in a dose-dependent manner ( Figure 3C). However, only licochalcone A, but not osimertinib, inhibited WT EGFR, Akt and ERK1/2 phosphorylation in A549 cells. We further examined EGFR signalling in Ba/F3 stable cells carrying various EGFRs, including Del E746-A750, L858R or L858R/T790M mutants, and WT.
The immunoblotting data showed that licochalcone A or osimertinib exhibited a similar inhibitory effects on Del E746-A750, L858R and L858R/T790M mutant-expressing stable cell lines ( Figure 3D).

Consistently, licochalcone A exhibited a stronger inhibitory effect on
WT EGFR than that of osimertinib ( Figure 3D). Taken together, our data indicate that licochalcone inhibits the activation of both wildtype and mutant EGFRs.

| Licochalcone A induces apoptosis in NSCLC Cells
HCC827, H1975 and A549 cells were pre-treated with inhibitors of apoptosis and necroptosis, such as z-VAD-fmk, GSK'873 or necrostatin-1. The MTS data showed that only z-VAD-fmk rescued licochalcone A-induced cell death ( Figure 4A), which is confirmed by the trypan blue exclusion assay ( Figure 4B). These results indicate that licochalcone A promoted apoptosis in NSCLC cells. Treatment with licochalcone A or osimertinib promoted the protein level of cleaved caspase 3 and cleaved PARP in HCC827, H3255 and H1975 cells ( Figure 4C). However, the osimertinib-induced apoptosis in WT EGFR expression A549 cell was compromised when compared with that of licochalcone A-treated A549 cells ( Figure 4C). Also, lico-

| Survivin is translationally regulated by licochalcone A in NSCLC cells
To further confirm that EGFR signalling is required for survivin expression in NSCLC cells, we examined the survivin protein level with EGFR inhibitor treatment. We found that osimertinib inhibited the phosphorylation of EGFR in HC827, H1975 and H3255 cells F I G U R E 5 Licochalcone A translationally regulates survivin in NSCLC cells. A, HCC827 cells were treated with licochalcone A, and the mRNA level of survivin was analysed by quantitative RT-PCR. B, IB analysis of HCC827 cells treated with licochalcone A and MG132. C, HCC827 cells were treated with licochalcone A for 24 h and co-cultured with cycloheximide (CHX) for various time-points, and WCE was subjected to IB analysis (top). The line chart shows the half-life of survivin from IB analysis (bottom). D, IB analysis of survivin in HCC827-shGFP and HCC827-sh4E-BP1 stable cells. E, IB analysis of HCC827-shGFP and HCC827-sh4E-BP1 stable cells treated with licochalcone A or DMSO. F, IB analysis of survivin in HCC827-shGFP and HCC827-sheIF4E stable cells. G, HCC827-shGFP and HCC827-sheIF4E stable cells were treated with DMSO or licochalcone A and subjected to IB analysis. H and I, HCC827 cells were treated with ERK1/2 inhibitor PD98059, Akt inhibitor MK2206, licochalcone A, or a combination of PD98059 and MK2206 for 24 h, and WCE was subjected to IB analysis (H) or quantitative RT-PCR analysis (I). J, Co-immunoprecipitation (co-IP) and IB analysis of HCC827 cells treated with DMSO or licochalcone A robustly. Consistently, the protein level of survivin was decreased in these EGFR mutant cells, but not in A549 with wild-type EGFR ( Figure S4A). Furthermore, knockdown of EGFR with siRNA reduced survivin expression robustly in HCC827 and H1975 cells ( Figure   S4B). We further generated EGFR knockout stable cells in A549 cells using sgRNA. The data revealed that overexpression of EGFR activation mutant Del E746-A750 or L858R/T790M restored survivin protein level in A549 cells with EGFR knockout ( Figure S4C). The qRT-PCR data revealed that licochalcone A did not affect the mRNA level of survivin ( Figure 5A). Moreover, incubation with MG132 F I G U R E 6 Licochalcone A inhibits xenograft tumour growth. A-d, HCC827 (A), H1975 (B), H3255 (C) and A549 (D) xenograft tumours were treated with vehicle control, licochalcone A, or osimertinib, tumour volume (left) and tumour weight (right) were recorded. **P < .01, ***P < .001. ns, not statistically significant. E, IHC staining analysis of p-EGFR, Ki67 and survivin in HCC827 xenograft tumours. F, Quantitative analysis of p-EGFR, Ki67 and survivin from E. ***P < .001 failed to rescue licochalcone A-induced down-regulation of survivin ( Figure 5B). Also, cycloheximide chase assay showed that the protein degradation rate was similar in licochalcone A-and DMSO-treated HCC827 cells ( Figure 5C). These data indicate that the reduction of survivin expression was not caused by the transcription suppression or protein degradation. Strikingly, we found that knockdown of 4E-BP1 by shRNA, which resulted in the activation of cap-dependent translation increased the protein level of survivin ( Figure 5D). Consistently, the mRNA level of survivin was similar in all of these treated groups ( Figure 5I). Additionally, treatment with licochalcone A reduced the phosphorylation of 4E-BP1 in NSCLC cells ( Figure   S4D). Importantly, the co-immunoprecipitation (co-IP) data demonstrated that licochalcone A enhanced the interaction between eIF4E and 4E-BP1 ( Figure 5J), which further confirmed the inhibition of cap-dependent translation in licochalcone A-treated NSCLC cell.
Together, our data suggest that down-regulation of Akt and ERK signalling by licochalcone A decreased survivin expression through inhibition of cap-dependent translation in NSCLC cells.

| Licochalcone A suppresses xenograft tumour growth in vivo
We next examined the in vivo anti-tumour activity of licochalcone A using the xenograft mouse model. When tumour volume reached around 100 mm 3 , treatment with licochalcone A, osimertinib or vehicle control was initiated. Our data indicated that every 2 days dosing of licochalcone A delayed the tumour growth of HCC827 ( Figure 6A), H3255 ( Figure 6B) and H1975 ( Figure 6C) xenografts.
Osimertinib blocked tumour growth in the HCC827 and H3255 xenograft tumours, but failed in A549 ( Figure 6D) xenograft with EGFR WT. In contrast, licochalcone A reduced tumour size significantly. Moreover, the phosphorylated EGFR, ki67 and total protein level of survivin were examined by immunohistochemical analysis.
As shown in Figure 6E,F, licochalcone A suppressed EGFR kinase activity. Consistently, the protein levels of Ki67 and survivin were reduced with licochalcone A or osimertinib treatment. These results suggest that licochalcone A inhibits the in vivo tumour growth of both osimertinib-sensitive and osimertinib-resistant xenografts.

| D ISCUSS I ON
With the development of diagnosis and treatment, especially the successful clinical application of EGFR TKI, the overall survival of NSCLC patients has improved significantly during the past years. [24][25][26] Currently, osimertinib is the only irreversible third-generation EGFR-TKI approved for the treatment of EGFR-activating mutations and the EGFR T790M mutation in patients with EGFR oncogene addiction. 27 Survivin plays a critical role for cancer cell survival and metastasis in multiple human cancer cells. 33,34 Previous studies have shown that survivin is highly expressed in human cancers, including lung, 35 ovarian, 36 cervical 37 and colorectal cancer, 38 glioblastoma 39 and T cell lymphoma. 40 A recent study demonstrated that survivin is overexpressed on cancer stem cells and required for maintaining cancer stem cell properties. 41 Furthermore, survivin could be processed and presented by dendritic cells and activates the CTL response in vitro or in a murine melanoma model in vivo. 42,43 Survivin protein expression is involved in the progression of NSCLC 35 and decreases survivin by anti-tumour compound T21-inhibited NSCLC cell growth and T21-induced apoptosis. 44,45 Moreover, metformin promotes survivin degradation through AMPK/PKA/GSK-3β-axis, which reduces the cell viability of NSCLC cells. 46 This evidence indicates survivin is a fantastic target in cancer treatment. Indeed suppression of transactivation of survivin through direct binding to its promoter, the small molecule inhibitors YM-155 and terameprocol (EM-1421) decreased survivin protein and induced apoptosis in human cancer cells. [47][48][49] Additionally, targeting survivin enhanced tumour chemo-and radio-sensitivity. [50][51][52] Our data demonstrated that licochalcone A translationally regulates survivin expression, but exhibits no significant effect on mRNA level and protein stability. Importantly, licochalcone A-mediated survivin down-regulation is partly dependent on the suppression of EGFR downstream Akt and ERK1/2 signallings, which is consistent with the previous report that reduces survivin protein by brexpiprazole overcomes EGFR TKI resistance in lung and pancreatic cancer. 53 Overall, this study investigated the anti-tumour efficacy of licochalcone A in NSCLC cells. Through suppression of EGFR signalling and decrease in survivin expression, licochalcone A exhibited profound antitumour potential in either EGFR WT or mutant NSCLC cells. Our studies provided new insights into the role of licochalcone A in cancer treatment and suggested licochalcone A might be a therapeutic agent against this devastating disease.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated and analysed during this study are available from the corresponding author on reasonable request.