AMIGO2 attenuates innate cisplatin sensitivity by suppression of GSDME‐conferred pyroptosis in non‐small cell lung cancer

Abstract Non‐small cell lung cancer (NSCLC) accounts for approximately 85% of lung cancer. Cisplatin is commonly used in the treatment of many malignant tumours including NSCLC. The innate drug sensitivity greatly affects the clinical efficacy of cisplatin‐based chemotherapy. As a plasma membrane adhesion molecule, amphoterin‐induced gene and ORF‐2 (AMIGO2) initially identified as a neurite outgrowth factor has been recently found to play a crucial role in cancer occurrence and progression. However, it is still unclear whether AMIGO2 is involved in innate cisplatin sensitivity. In the present study, we provided the in vitro and in vivo evidences indicating that the alteration of AMIGO2 expression triggered changes of innate cisplatin sensitivity as well as cisplatin‐induced pyroptosis in NSCLC. Further results revealed that AMIGO2 might inhibit cisplatin‐induced activation of (caspase‐8 and caspase‐9)/caspase‐3 via stimulating PDK1/Akt (T308) signalling axis, resulting in suppression of GSDME cleavage and the subsequent cell pyroptosis, thereby decreasing the sensitivity of NSCLC cells to cisplatin treatment. The results provided a new insight that AMIGO2 regulated the innate cisplatin sensitivity of NSCLC through GSDME‐mediated pyroptosis.


| INTRODUC TI ON
Lung cancer is the leading cause of cancer-related mortality worldwide, accounting for nearly one fifth of cancer mortality. 1 Approximately 85% of the newly diagnosed lung cancer cases belong to non-small cell lung cancer (NSCLC), and response poorly to most of the therapeutic options. 2 The cytotoxic chemotherapy agents (e.g. cisplatin) are the first-line drugs for NSCLC treatment. 3 Although cisplatin-based chemotherapy improves the overall survival of NSCLC patients in clinic, many patients are insensitive to the chemical agents and display poor therapeutic efficacy. 4 Despite intense efforts to explore the regulation mechanisms for the response of tumour cells to chemotherapy, effective targets remain deficient for improving the innate cisplatin sensitivity in NSCLC.
Pyroptosis is a severer type of programmed cell death, which is mediated by gasdermin family members. [5][6][7] Cells undergoing pyroptosis display distinct characteristics such as ruptured cell membrane with large bubbles blowing from the membrane, increased propidium iodide (PI) uptake and elevated lactic dehydrogenase (LDH) activity. 8 Recent studies have found caspase-3 activated by chemotherapy agents can cleave one member of gasdermin family, gasdermin E (GSDME), to generate the N-terminal fragment (GSDME-NT), which perforates the plasma membrane and results in pyroptosis. 9 Moreover, it has been proved that GSDME is associated with clinical cisplatin therapeutic effects, with shorter survival time and higher mortality rate observed in NSCLC patients with lower GSDME expression after treatment with cisplatin. 10 These results implied that GSDME-mediated pyroptosis may play an essential role in the innate cisplatin sensitivity in NSCLC.
AMIGO2 is a leucine-rich repeat (LRR)-containing cell adhesion molecule initially identified as a neurite outgrowth factor. 11,12 Further investigations suggest that AMIGO2 is implicated in neuronal survival and regulation, 13 vascular development and angiogenesis, 14,15 as well as ischemia protection and immune regulation. 16,17 Recently, AMIGO2 has been found to play a potential etiologic role in carcinogenesis and tumour progression. SiRNA-mediated knockdown of AMIGO2 expression in gastric cancer cells resulted in abrogated tumorigenicity. 18 Forced expression of AMIGO2 promoted tumour cell adhesion to liver endothelial cells and enhanced liver metastasis. 19 AMIGO2 upregulated in melanomas served as a target gene of bromine domain and extraterminal domain (BET) for melanoma cell survival and metastasis. 20 Although accumulating evidences suggest that AMIGO2 is a potential target for cancer therapy, the roles and the underlying mechanisms of AMIGO2 in regulation of the chemosensitivity have not been determined. In this study, we proved that AMIGO2 attenuated the response of NSCLC to cisplatin treatment and revealed inhibition of the caspase cascade and GSDME-mediated pyroptosis served as a novel molecular mechanism for AMIGO2-triggered drug insensitivity. Biology, Jinan University). Cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and incubated at 37°C in a humidified atmosphere of 5% CO 2 .

| Reagents and antibodies
Cisplatin (DDP) was obtained from Yunnan Phytopharmaceutical Co.

| Generation of stable AMIGO2 knockdown and overexpression cell lines
The recombinant lentiviral plasmid harbouring AMIGO2 shRNA

( G C A A G G A A A G T A C AT T TG T T T C TC G AG A A A C A A ATG TA C
TTTCCTTGCTTTTTT) obtained from Hanyang biological company, and the recombinant plasmid pCDH510B/AMIGO2 containing full-length AMIGO2 coding sequence constructed previously by our lab were cotransfected with packaging auxiliary vectors (psPAX2 and pMD2. G) into 293T cells, respectively, to package lentiviruses, which were applied to infect cells to obtain cell lines stably knocking down or overexpressing AMIGO2 by puromycin screening.

| Flow cytometric analysis
Cells were seeded in 6-well plates at 3 × 10 5 cells/well and cultured overnight. Cells were treated with indicated concentrations of cisplatin for 48 h before subjection to analysis of the programmed cell death using the Annexin V-FITC/PI kit (Life) according to the manufacturer's instructions. Briefly, cells were double stained with Annexin V-FITC and PI in 1 × binding buffer for 15 min in dark at room temperature, and subjected to flow cytometric analysis using Attune NxT Acoustic Focusing Cytometer (Thermo Fisher Scientific).
Data were further analysed with Flowjo v10 software.

| Hoechst 33342/PI staining
Cells were seeded in 12-well plates and cultured overnight. After treatment with indicated concentrations of cisplatin, cells were stained with Hoechst 33342 (5 μg/mL) and PI (2 μg/mL) for 10 min at room temperature, and subjected to observation using Zeiss Axio Observer D1 fluorescence microscope with 20× objective lens (Carl Zeiss Micro-Imaging GmbH). Five fields per well were randomly selected for counting the stained cells by ImageJ software. Briefly, Aliquots of supernatants were transferred into a blank 96well plate at 120 μL/well, and 60 μL of LDH detection solution was added to each well. The plates were then incubated for 30 min in dark at room temperature with gentle shaking. The absorbance was measured at 490 nm on a spectrophotometric microplate reader.

| LDH release assay
The percentage of LDH release was calculated using the equation where LDH sample , LDH background and LDH maximum are the OD 490 measured for the drug treated, untreated and lysis solution-treated supernatants, respectively.
For determination of the LDH activity in tumour tissue using Lactate dehydrogenase assay kit (Nanjing Jiancheng), tumour tissues were cut into pieces and subjected to homogenization in 0.9% normal saline with tissue crusher. The supernatants were collected for measurement of the LDH activity as described above immediately.

| siRNA transfection
For silencing of target genes with small interfering RNA, the siRNA (5′-GUACA AAA GAU CUC GCUGA-3′) duplexes targeting AMIGO2 and the siRNA (5′-GGUCC UAU UUG AUG AUGAA-3′) duplexes targeting GSDME were synthesized by GenePharma, and transfected into cells using Lipofectamine®3000 transfection reagent (Invitrogen) according to the manufacturer's protocols. After grown in a serum-free medium for 6 h, cells were cultured in RPMI-1640 medium containing 10% FBS. The knockdown efficiency was determined by real-time quantitative PCR and western blot analysis.

| Real-time quantitative PCR analysis
Total RNA was extracted using TRIzol™ according to the manufac-

| Western blot analysis
Cells or tumour tissues were washed with cold PBS twice and subjected to preparation of protein samples using RIPA lysis buffer (Beyotime). After quantification with BCA protein assay kit (Thermo Fisher Scientific), protein extracts were loaded in 12% SDS-PAGE gel and transferred onto polyvinyl difluoride (PVDF) membranes (Millipore). The membrane was blocked with 5% nonfat dry milk for 1 h at room temperature, and incubated with primary antibodies at 4°C overnight. After washing three times with TBST (0.1% Tween-20), the membrane was incubated with anti-rabbit or antimouse IgG antibodies for 1 h at room temperature. The immunoblots on the membrane was revealed by a BeyoECL Plus kit (Beyotime), and developed on an x-ray film. Densitometry analysis for western blotting was performed using Quantity One version 4.6.2 software (Bio-Rad). Tumour volume was obtained using a sliding calliper and calculated as length × width 2 × 0.52. Tumours were extracted from the sacrificed mice, weighted, imaged and subjected to western blotting and immunohistochemical analysis. The animal experimental procedures complied with the guidelines for animal care of Jinan University (Guangzhou, China), and were approved by the Animal Ethics Committee of this institution.

| Immunohistochemical analysis
The tumours were diced, fixed in 4% paraformaldehyde, dehydrated and embedded in paraffin. After deparaffinized, tissue sections were incubated with anti-Ki67 and anti-CD34 monoclonal antibodies, followed by secondary antibodies, and indicated by HRP-streptavidin.
Images were acquired using an Inverted Fluorescence Microscope (Zeiss), and subjected to analysis of integrated optical density (IOD) and Area by Image-Pro® Plus v 6.0.

| Statistical analysis
Statistical analysis was performed with GraphPad Prism software.
In all experiments, comparisons between two groups were based on two-sided Student's t-test and one-way anova was used to test for differences among more groups. P values less than 0.05 were considered statistically significant.

| AMIGO2 reduced innate cisplatin sensitivity in NSCLC cells
In order to explore whether AMIGO2 is involved in the innate cisplatin sensitivity of NSCLC cells, stable clones of AMIGO2 knockdown cells A549/shAMIGO2, as well as AMIGO2 overexpressing cells H460/ AMIGO2 were constructed and characterized by RT-qPCR and immunoblotting ( Figure S1). The influences of alteration of AMIGO2 expression on innate cisplatin sensitivity were determined by the MTT method. The results indicated that knockdown of AMIGO2 increased the sensitivity to cisplatin treatment, with lower IC 50 value of A549/shAMIGO2 cells than that of the corresponding A549/shNC cells infected with the control lentivirus. Meanwhile, AMIGO2 siRNA targeting treatment on A549 cells also led to enhanced sensitivity to the cisplatin ( Table 1). In contrast, overexpression of AMIGO2 decreased the sensitivity evidenced by the higher IC 50 value detected in H460/AMIGO2 compared with H460/pCDH control cells ( Table 1). The results confirmed that AMIGO2 was implicated in the modulation of innate cisplatin sensitivity in NSCLC cells.

| AMIGO2 suppressed pyroptosis induced by cisplatin
Pyroptosis is a type of lytic programmed cell death. Accumulating evidences have suggested that the programmed cell death is associated  Figure 1C,D). LDH release assay applied for detection of the plasma membrane permeability change further indicated that A549/ siAMIGO2 cells exposed to cisplatin released more LDH than A549/siNC control cells ( Figure 1E). Similar results were achieved using AMIGO2 knockdown stable cell line A549/shAMIGO2 ( Figure S2). Consistent with these observations, we also found that overexpression of AMIGO2 significantly reduced both the percentage of PI-positive cells with pyroptotic morphology and the LDH release rate in cisplatin-treated H460 cells ( Figure 1F-H). The results implied that AMIGO2 had the potential of alleviating the pyroptosis triggered by cisplatin.

| Effects of AMIGO2 on the cisplatin-activated caspase cascade
It has been demonstrated that activation of the caspase cascade can elicit pyroptosis via the proteolytic cleavage of the gasdemin protein TA B L E 1 AMIGO2 modulated innate cisplatin sensitivity in NSCLC cells.  Figure 2, cisplatin significantly activated both extrinsic and intrinsic apoptotic initiators, caspase-8 and caspase-9, as well as the downstream executioner caspase-3. In the cisplatin-treated cells, we also found that GSDME, a member of the gasdemin protein family recently recognized as a substrate of caspase-3, was cleaved to generate the GSDME-NT fragment, which has been proved to be an effector of cell pyroptosis. Importantly, upon cisplatin treatment, AMIGO2 knockdown enhanced activations of the caspase cascade as well as GSDME cleavage in A549 cells (Figure 2A), while enforced expression of AMIGO2 attenuated the caspase and GSDME fragmentation in H460 cells ( Figure 2B). The results suggested that AMIGO2 might regulate cisplatin-induced pyroptosis through the caspase cascade.

| AMIGO2 reduced innate cisplatin sensitivity via suppression of GSDME-mediated pyroptosis
Previous studies have shown that patients with lower GSDME expression levels in tumour tissues have a shorter survival time and higher mortality after cisplatin-based treatment, implying that the expression levels of GSDME in tumour cells may affect the sensitivity to cisplatin-based therapy. 10 To further determine whether GSDME was involved in mediating the effects of AMIGO2 on cisplatin sensitivity, MTT assay was first applied to detect cell viability in A549 cells simultaneously transfected with AMIGO2 siRNA and GSDME siRNA. The results showed that knockdown of GSDME attenuated the suppression effect of AMIGO2 silence on cell survival ( Figure 3A). Further investigation indicated that knockdown of GSDME alleviated the promotion effects triggered by AMIGO2 knockdown on cisplatin-induced pyroptosis as evidenced by less GSDME cleavage, decrease of PI-positive cells as well as reduced release of LDH were observed in siAMIGO2 and siGSDME-transfected A549 cells compared with the control cells transfected with siNC ( Figure 3B-D). The results suggested that GSDME was involved in AMIGO2 regulating the pyroptosis induced by cisplatin. AMIGO2 might impair the response of NSCLC cells to cisplatin through inhibition of GSDME-mediated pyroptosis.

| AMIGO2 inhibited GSDME-mediated pyroptosis through (caspase-8 and caspase-9)/ caspase-3 cascade
It has recently been recognized that specific cleavage of GSDME by active caspase-3 triggers pyroptosis. Combined with the above results indicating that AMIGO2 suppressed the (caspase-8 and caspase-9)/caspase-3 cascade in A549 cells exposed to cisplatin, we speculated that AMIGO2 might inhibit GSDME-mediated py-  Figure 4C,D). The results implied that AMIGO2 suppressed GSDME-conferred pyroptosis via a blockade of (caspase-8 and caspase-9)/caspase-3 cascade. sults showed that either PDK1 or Akt phosphorylation (T308 and S473) was decreased in AMIGO2 deficient cells, whereas overexpression of AMIGO2 enhanced both PDK1 and Akt activation ( Figure 5A). It has been known that Akt is phosphorylated at T308 by PDK1 and at S473 by PDK2 family members. We found that the depletion of Akt (T308) activation with BX795 (5 μM), a specific inhibitor for PDK1, promoted

| Knockdown of AMIGO2 increased cisplatin sensitivity in vivo
The in vitro study suggested that AMIGO2 possessed the potential of reducing innate cisplatin sensitivity through modulation of GSDME- Similar results were also observed when cisplatin was administrated ( Figure 6G-I). Taken together, these findings suggested AMIGO2 depletion enhanced the cytotoxic efficacy of cisplatin through GSDME-mediated pyroptosis in vivo, which were conformed to the molecular mechanisms as revealed in vitro.

| DISCUSS ION
Non-small cell lung cancer (NSCLC) is the most common type of lung cancers with the highest morbidity and mortality worldwide.
F I G U R E 3 AMIGO2 reduced innate cisplatin sensitivity via suppression of GSDME-mediated pyroptosis. (A) A549 cells simultaneously transfected with AMIGO2 siRNA and GSDME siRNA were treated with 5 μg/mL cisplatin for 48 h prior to detection of cell viability by MTT method. (B) A549 cells simultaneously transfected with AMIGO2 siRNA and GSDME siRNA were treated with 5 μg/ mL cisplatin for 24 h, and subjected to western blot analysis with the indicated antibodies. (C) A549 cells simultaneously transfected with AMIGO2 siRNA and GSDME siRNA were treated with 5 μg/ mL cisplatin for 24 h, and dual stained with 2 μg/mL PI and 5 μg/mL Hoechst 33342 for 10 min in dark. After capturing the images by fluorescence microscopy (20× objective lens), PI-positive cells were calculated in five random fields and statistically analysed. (D) The percentage of LDH release in the culture supernatants from cells treated with cisplatin for 24 h was measured using LDH cytotoxicity assay kit. *p < 0.05; **p < 0.01.
Cisplatin is an effective first-line platinum agent widely used in NSCLC therapy, which has improved the overall survival at 1 year from 5% in untreated patients to 15% in patients administrated with the cisplatin-based therapy. 21  they also found that significant absolute AMIGO2 expression levels were observed in some lung cancer patients (20%). Of a variety of cancer lines including lung, colon, breast, ovary, cervical, prostate, skin, brain, kidney, liver, osteosarcoma and epidermis, the highest expression observed was in lung A549 and NCI-H1299 cells, implying that AMIGO2 might play vital roles in a subset of lung cancer. 18 In this study, AMIGO2 was found to be an essential mediator involved in the response of NSCLC to cisplatin, and depletion of AMIGO2 markedly enhanced innate sensitivity against cisplatin via GSDMEmediated pyroptosis in NSCLC cells. However, further clinical investigations are needed to elucidate the correlation between AMIGO2 expression and innate cisplatin sensitivity in NSCLC patients.
Pyroptosis is a form of programmed cell death originally thought to be executed by the gasdermin D (GSDMD) subjected to cleavage by the active caspase-1, -4, -5 or -11 to release its gasdemin-NT fragment with pore-forming activity. [23][24][25] Recent study found that another gasdermin, GSDME, competitively suppressed the association of caspase-3 with its apoptotic substrates by binding to caspase-3, and thus switch apoptosis to pyroptosis. 9 It has been proved that cisplatin could activate caspase-3 and induce pyroptosis in both cancer and normal cells with abundant GSDME expression. 9,26 Combined with the literature that indicated that GSDME promoted cisplatin sensitivity to inhibit lung carcinoma by mediating pyroptosis to trigger antitumor immunocyte infiltration, 10 we speculated that caspase-3-related cascade and its substrate GSDME might be implicated in the mechanisms of AMIGO2 affecting innate sensitivity against cisplatin. We first demonstrated that AMIGO2 suppressed the caspase and GSDME fragmentation induced by cisplatin. Knockdown of GSDME or blockade of caspase activations with the specific inhibitors further revealed that AMIGO2 impaired the cisplatin sensibility through suppression of the cisplatin-induced F I G U R E 4 AMIGO2 inhibited GSDMEmediated pyroptosis through (caspase-8 and caspase-9)/caspase-3 cascade. A549/shAMIGO2 cells pretreated with Av-DEVD-CHO (100 μM), Z-LEHD-FMK (50 μM) or Z-IETD-FMK (50 μM) for 4 h were exposed to 5 μg/mL cisplatin for 24 h. (A) MTT method was performed to assess cell viability. (B) Cells were lysed to obtain protein samples for western blot analysis with the indicated antibodies. (C) Cells were stained with PI and Hoechst 33342 prior to capturing the fluorescent images (20× objective lens). Five fields were randomly selected to calculate PI-positive cells for statistical analysis.
The Akt signalling pathway is associated with multiple physiological and pathological processes, including cell survival, angiogenesis, as well as cancer and neurological disorders. 27,28 The activation of Akt is triggered by phosphorylation at two sites, T308 and S473. PDK1 served as a primary activator of Akt is autophosphorylated at the site of S241 via binding to phosphatidylinositol-3,4,5-triphosphate (PIP3), resulting in the recruitment of PDK1 to the membrane and the subsequent phosphorylation of Akt at one site T308. 29 Another site at S473 was phosphorylated by PDK2 family members such as mTORC2 (mammalian target of rapamycin complex 2) or DNA-PK (DNA-dependent protein kinase). 30,31 Recent report indicated that AMIGO2 served as a scaffold protein at the plasma membrane to promote PDK1 membrane localization and PDK1/Akt activation in endothelial cells (ECs) and regulated the angiogenic function. 14 Consistent with these suggestions, we found that knockdown of AMIGO2 significantly inhibited the phosphorylation of PDK1 (S241), Akt (T308) and Akt (S473), while enhanced the expression of AMIGO2 stimulated the activations of the above signal molecules in NSCLC cells. Given AMIGO2 can promote the phosphorylation of Akt at both T308 and S473, whose initial phosphorylation was triggered by PDK1 and PDK2 family members respectively, it is reasonable to speculate that in addition to modulation of PDK1 activity for Akt phosphorylation at T308, AMIGO2 might have the potential of regulating the kinase activity of PDK2 family members for Akt phosphorylation at S473. Further investigations are required to clarify the possible roles of AMIGO2 in regulation of the PDK2 family members.
It has been known that the activated Akt not only directly inhibits the protease activity of caspase-9 by phosphorylating its serine residue at position 196 (S196), 32 but also impairs the kinase activity of the caspase-8 activator, ASK1 (apoptosis signal-regulated kinase 1) via phosphorylating its serine residue at position 83 (S83), and thus indirectly abrogates the protease activity of caspase-8. 33,34 Activation of Akt by AMIGO2 might be responsible for inhibition of caspase-3 by repressing both caspase-9 and caspase-8, resulting in suppression of GSDME cleavage and the subsequent cell pyroptosis, thereby reducing the cisplatin sensitivity of NSCLC cells.
Intriguingly, we found that inhibition of Akt phosphorylation at S473 with MK-2206 had little effect on the protease activity of caspase-9 and caspase-8, as wells as the GSDME fragmentation and pyroptosis, whereas abrogation of Akt phosphorylation at T308 with BX795 significantly enhanced the cleavage of caspase-9, caspase-8 and GSDME, and promoted pyroptosis in cisplatin-treated NSCLC cells overexpressing AMIGO2. The results suggested that phosphorylation of Akt at T308 rather than at S473 was essentially indispensable for the inhibitory effects of AMIGO2 on the caspase cascade and GSDME-mediated pyroptosis. The LDH release in the culture supernatants from cells was assessed using LDH cytotoxicity assay kit. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.
In summary, we have first provided evidences to reveal the role and the underlying mechanism of AMIGO2 in regulation of the innate cisplatin sensitivity in NSCLC cells. Our results demonstrated that AMIGO2 might inhibit cisplatin-induced activation of (caspase-8 and caspase-9)/caspase-3 via stimulating PDK1/Akt (T308) signalling axis, leading to attenuation of GSDME-conferred pyroptosis, thereby reducing the innate sensitivity of NSCLC cells against cisplatin (Graphical Abstract). Therefore, determination of the expression levels of AMIGO2 in NSCLC patients might give the hint for predicting the efficacy of cisplatin treatment. More importantly, AMIGO2 might be employed as an adjunctive therapeutic target for improving the response to cisplatin-based chemotherapy in a subset of NSCLC with abundant GSDME expression. Funding acquisition (equal); project administration (lead). Xiao-ping Wu: Conceptualization (lead); funding acquisition (lead); supervision (lead); writing -review and editing (lead).

ACK N O WLE D G E M ENTS
The work was supported by the National Natural Science Foundation

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors declare that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.