Up-regulation of microRNA miR-101-3p enhances sensitivity to cisplatin via regulation of small interfering RNA (siRNA) Anti-human AGT4D and autophagy in non-small-cell lung carcinoma (NSCLC)

ABSTRACT The emergence of drug resistance hinders the treatment of malignant tumors, and autophagy plays an important role in tumor chemotherapy resistance. However, its mechanism in non-small cell lung cancer (NSCLC) has not been well-researched. We aim to investigate the role of miR-101-3p in cisplatin-resistant via regulation of autophagy-related protein 4D (ATG4D) and autophagy. Cell viability, apoptosis, fluorescence intensity of GFP-LC3 and RFP-GFP-LC3 were determined using Cell Counting Kit-8 (CCK-8) assay, flow cytometry, and Laser scanning confocal microscope analysis, respectively. The levels of LC3II/LC3I, P62 and ATG4D were detected by Western blot. The results showed that the sensitivity to cisplatin in NSCLC cells was up-regulated by miR-101-3p mimics treatment, inducing promoting cell apoptosis and inhibiting autophagy. Further mechanistic study identified that ATG4D was a direct target of miR-101-3p. Moreover, ATG4D siRNA also could reverse miR-101-3p inhibitor-induced the up-regulation of ATG4D and the ration of LC3II/LC3I, the down-regulation of p62 expression. Our findings indicated that miR-101-3p could regulate sensitivity to cisplatin of NSNCC cells by regulating autophagy mediated by ATG4D. Therefore, miR-101-3p may act as a potential therapeutic target for the treatment of NSCLC.


Introduction
Lung cancer, as the most common malignant cancer, was associated with high mortality rate all over the world, approximately 80-85% of total lung cancer cases is non-small cell-lung cancer (NSCLC) [1]. Cisplatin is widely used as the firstline chemotherapy drug for treatment in hepatocellular carcinoma (HCC), triple-negative breast cancer, bladder cancer and ovarian, also including NSCLC [2][3][4][5][6]. Cisplatin binding could result in DNA damage and induce cell apoptosis. However, cisplatin chemotherapy is often limited by inherent and acquired drug resistance. Therefore, it is urgent for us to search effective targets to improve treatment in NSCLC patients.
MicroRNAs (miRNAs), a class of small endogenous conserved RNAs, modulate a different kind of gene expression by interacting with their mRNAs, leading to mRNA degradation or inhibition of protein translation [7]. Accumulating evidence indicated that miRNAs are participate in various biological processes in cancer cells, including cell proliferation, apoptosis, invasion and migration [8,9]. Moreover, up-regulation or down-regulation miRNAs could help regulate cancer cell drug resistance, including NSCLC [10,11]. MiR-126-3p mimics could increase melanoma's sensitivity to dabrafenib by regulating disintegrin and metalloproteinase 9 (ADAM9) [12]. Inhibition of miR-214 could reverse resistance to erlotinib in NSCLC via increasing LIM Homeobox 6 (LHX6) [13]. The level of miR-101-3p was reduced in many cancers including NSCLC, indicating that it may act as a tumor suppressor [14][15][16]. Up-regulation of miR-101-3p enhanced sensitivity to oxaliplatin in HCC cells by inhibiting autophagy and Beclin-1 14 . Transfection of miR-101-3p mimics also sensitized bladder urothelial carcinoma cells to cisplatin via inhibiting enhancer of zeste homolog 2 (EZH2) expression [17]. It has been demonstrated that miR-101-3p could regulate cell autophagy in various cells, and it may be a potential autophagy inhibitor [14,18]. However, the potential mechanisms of miR-101-3p regulating drug resistance in NSCLC have not been explored. Therefore, we need to elucidate the effect of miR-101-3p on cisplatin-resistant NSCLC.
Autophagy is a highly conserved catabolic process that involves cells self-digesting through a double-membrane organelle called autophagosome [19,20]. Autophagy is composed of a series of molecular events that lead to the formation of autophagosomes, the autophagosomes engulf the intracellular material and finally fuse with lysosomes to degrade its contents [21]. It has been reported that autophagy can promote tumor progression and resistance to treatment in cancers. Over-expression of miRNA-22 could inhibit cisplatin resistance by regulation of autophagy and apoptosis in osteosarcoma [22]. Transfection of miRNA-1301 could inhibit cisplatin-resistance in human ovarian cancer cells by regulation of epithelial-to-mesenchymal transition (EMT) and NF-κB signaling pathway [23]. This study investigated the effect of miR-101a-3p alterations in drug-resistant NSCLC cells and the relationship between cisplatin resistance, miR-101-3p and autophagy. However, whether autophagy has a cytoprotective effect and how it influences the acquired resistance to cisplatin through specific molecular mechanisms was unclear. Therefore, the present study revealed that miR-101-3p may be a potential target, which can be used to counteract the drug resistance and improve treatment in NSCLC patients.

Cell culture and cell viability
The human NSCLC cells and the normal lung epithelial cells needed for this study include (A549 PC-9, NCI-H1299, HCC827(ATCC) and MCR5 were purchased from National Collection of Authenticated Cell Cultures (Shanghai, China). All cells were maintained in PRMI 1640medium (GIBCO) contained with 10% FBS (GIBCO) in the incubator with a special environment (5% CO 2 , 37°C). When the cells were grown to a certain extent, we digested the cells, inoculated the cells in a 96-well plate at a density of 5000/well, cultured for 24 hours until the cells adhered to the wall of culture flask, and then added different concentrations of drugs to incubate for 48 hours. At last, we added 10-µl Cell Counting Kit-8 (CCK-8; Dojindo) solution that was from CCK-8 assay to detect cell viability for incubation 3 hours and then measured the OD value using a microtiter reader at 450 nm.

Transfection
MiR-101-3p mimics, inhibitors or negative controls were designed from RiboBio Co., Ltd Company (Guangzhou, China). AGT4D siRNA and Negative control was purchased from Santa Cruz. According to the instruction, lipofectamine 2000 (Invitrogen, USA) was used for transfection. The relative sequence is shown in Table 1.

Cell proliferation
The cells were cultivated into 96-well plates and examined by a Click-iTEdU Imaging kit (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. NSCLC cells were incubated with the IC50 concentration of Cisplatin for 24 h at 37°C, fixed by 4% paraformaldehyde at room temperature for 20 min, permeabilized by 0.5% Triton-X100 at room temperature for 15-20 min, then followed by 10 μM EdU for 2 h at room temperature for 20 min. Cell nuclei were stained by 5 μg/ml Hoechst 33,342 (Invitrogen; Thermo Fisher Scientific, Inc.) for 30 min at room temperature.

Flow cytometry analysis
The number of apoptotic cells was determined by an Annexin V-FITC/propidium iodide (PI) staining kit (BCD Biosciences, San Jose, CA, USA). In brief, following transfection at 48 h, the cells were cultivated into 6-well plates, and stained with 5 μL Annexin V-FITC and 10 μL PI for 15 min at 4°C in the dark. Finally, the cells were analyzed using a FACScan flow cytometer (Becton-Dickinson, Bedford, MA, USA).

Dual luciferase reporter assay
The wild type (WT) 3ʹ-UTR and mutated (Mut) 3ʹ-UTR of the ATG4D gene containing the predicted target sites for miR-101a-3p were constructed by GenePharma. The NSCLC cell were seeded into 24-well plates and co-transfected with 50 ng recombinant luciferase vectors, 5 ng pRL-TK vectors and 50 nM miR-101-3p mimics or inhibitor, and miR negative control (miR-NC) using Lipofectamine 2000. After transfection for 48 h, the relative luciferase activities were measured using the Dual Luciferase Reporter Assay System (Promega).

RNA extraction and qRT-PCR
We used TRIzol reagent to extract total RNA from cultured HCC cells; then, total RNA was reverse transcribed into cDNA by Moloney murine leukemia virus (M-MLV) reverse transcriptase (TAKARA, Dalian, China) according to the manufacturer s instructions. We used SYBR® Premix Ex Taq™ II (Takara) to analyze qRT-PCR. The 2 −ΔΔCt method was used to calculate the relative gene expression normalized by GAPDH and U6. The relative sequences of primers are listed in Table 2.

Laser scanning confocal microscope
Cells were transfected with pSELECT-GFP-LC3 (InvivoGen, San Diego, CA) using Lipofectamine reagent (Invitrogen) following the manufacturer's instructions. Briefly, cells stably transfected with monomeric red fluorescent protein (mRFP)green fluorescent protein (GFP)-LC3 adenovirus were subjected to different treatments. After 48 h, the cells were fixed with 4% paraformaldehyde (Sigma, USA). Laser confocal fluorescence microscopy was used to observe the green (GFP) and red (mRFP) fluorescence. Autophagosomes appeared as yellow dots and autolysosomes appear as red dots in merged images. The autophagic flux was determined by the increase in the percentage of red dots in the merged images.

Electron microscopy detection
Cells were harvested and fixed in 2.5% glutaraldehyde, and then post-fixed in 1% osmium tetroxide buffer, dehydration with ethanol, embedded using epoxy resin, stained with 3% uranyl acetate and lead citrate, and examined by transmission electron microscope. Autophagic vacuoles per cell were examined by counting 10 cells for each sample.

Nude mouse xenograft model
Twenty BALB/c nude mice (5 weeks,18-20 g old) were obtained from Charles River (Beijing, China). The mice were adaptively fed for 1 week, light/night cycle for 12 hours, indoor temperature control at 23 ± 2°C, humidity 55 ± 5%, and free drinking water. 2 × 10 7 A549 were injected subcutaneously into the left lower limb of nude mice, and the tumor reached around 50 − 100 mm 3 , the mice were randomly divided into 3 groups with 5 animals in each group (blank control group, cisplatin treatment group, miR-101-3p agomiR+ cisplatin group). In this study, we used miR-101-3p agomiR (RiboBio, Guangzhou, China) to modify miR-101-3p mimics in vivo studies. The cisplatin (2.0 mg/kg), miR-101-3p agomiR (10 nmol), was injected intraperitoneally every 3 days for 3 weeks. We measured the tumor volume every two days and drawn the tumor growth curve. On the 5th and last day, nude mice were photographed in vivo. The tumor volume was recorded every 2 days (Tumor volume (mm 3 ) = length × width [2]/2).

TUNEL analysis
The tissue sections were fixed in 4% paraformaldehyde for 24 h, infiltrated with wax and embedded with paraffin. Then, add 100 μL TUNEL reaction mixture to the specimen and add cover glass or parafilm to react at 37°C for 1 h in a dark humidified box; wash with PBS for three times, add 100 μL streptavidin-labeled HRP (diluted in 1:500 PBS) for 30 minutes; wash it again, add 100 μL diaminobenzidine (DAB) mixture for 10 minutes, rinse with deionized water for 5 times; counterstain with hematoxylin, rinsed with tap water after about 3 seconds; dehydrate with gradient alcohol, transparent xylene for 1 minute, seale with neutral glue. Observe and take pictures under an immunofluorescence microscope (Olympus).

Immunocytochemistry
The sections are routinely deparaffinized in water, dipped with xylene for 10 min; immersed with gradient ethanol from high concentration to low concentration for 5 min each time; then fixed with 4% paraformaldehyde for 15 min; Place the slices in Hydrogen Peroxide Block and incubate for 10 minutes, PBS immersion for 5 min, repeat 2 times, add Ultra V Block dropwise, incubated at room temperature for 5 minutes to block nonspecific background staining, wash with PBS for 5 minutes, and repeat twice. Drop the primary antibody working solution and incubate at 37°C for 2 hours; drop the Primary Antibody Enhancer and incubate at room temperature for 20 minutes, add the enzyme-labeled secondary antibody dropwise, and incubate at room temperature for 30 minutes in the dark. Finally, add 1-2 drops of DAB incubate for 10 minutes. Finally, hematoxylin was used as a counterstain, the slices were rinsed thoroughly with tap water, counterstained, dehydrated, transparent, mounted and observed under light microscopy (Olympus).

Statistical analysis
The experimental data were presented as the mean ± standard deviation (SD) and the data were analyzed using GraphPad Prism 8.0 (GraphPad Software, Inc.). The differences between two groups or more than two groups were compared using an unpaired Student's t-test or one-way ANOVA, respectively. The results were considered statistically significant at P < 0.05. All experiments were repeated three times.

MiR-101-3p mimics increased sensitivity to cisplatin and promoted cell apoptosis
To further reveal the effect of miR-101-3p in NSCLC cells, we transfected with miR-101-3p mimics or inhibitor under cisplatin treatment and used CCK-8 assay, EdU and flow cytometry to examine cell viability, proliferation and apoptosis, respectively, the results showed that overexpression of miR-101-3p by miR-101-3p mimics increased cell viability and proliferation, however, miR-101-3p inhibitor had an opposite effect. The above results indicate that upregulation or down-regulation of miR-101-3p could regulate cisplatin sensitivity in NSCLC cells (Figure 2 A-B). Almost immediately, we found that upregulation of miR-101-3p enhanced the number of apoptosis cells, while miR-101-3p inhibitor reduced it (Figure 2c). This is enough to demonstrate that miR-101-3p could regulate sensitivity to cisplatin by regulating cell viability and apoptosis.

MiR-101-3p could interact with ATG4D expression
To elucidate the molecular mechanisms of miR-101-3p mediated autophagy, StarBase 3 (http://starbase. sysu.edu.cn/index.php) and TargetScan (http://www. targetscan.org/vert_71/) were used to analyze the expression of miRNA in NSCLC tissues and normal tissue or predicted the potential target genes of miRNA, and then found that ATG4D was a target gene for miR-101-3p (Figure 3a). Furthermore, ATG4D, as an autophagy-related gene, could regulate the autophagic flow at the autophagosome stage before fusion with lysosomes [25]. The results of Luciferase reporter assay showed that transfection with miR-101-3p mimic could down-regulate ATG4D WT luciferase activity, while it has no significant difference in ATG4D Mut between the two groups, which further verified that miR-101-3p could regulate ATG4D expression (Figure 3b). Almost immediately, we transfected with a miR-101a-3p mimic and inhibitor into NSCLC cells and determined the level of ATG4D, showing that the expression level of ATG4D was significantly downregulated by miR-101a-3p mimic treatment, whereas inhibition of miR-101a-3p upregulated ATG4D expression in NSCLC cells (Figure 3c).

The sensitivity effect of miR-101-3p was related with autophagy and regulation of ATG4D
To explore whether miR-101-3p enhanced cisplatin sensitivity by regulating autophagy and ATG4D, the expression of autophagy-related protein ATG4D, p62 and LC3I/LC3II in different treatment group (Control, Cisplatin, miR-101-3p mimic+ Cisplatin, miR-101-3p inhibitor + Cisplatin) were detected by Western blot analysis. Transfection with miR-101a-3p mimic could inhibit cisplatin-induced the upregulation of ATG4D expression and the increase of LC3II/LC3I ratio, the down-regulation of p62 expression, but miR-101a-3p inhibitor had the opposite effect (Figure 4 A). The expression of autophagosomes (mRFP + GFP) and autolysosomes (mRFP) represented autophagy level. Confocal microscopy analysis showed that co-treatment with miR-101-3p mimic and cisplatin could reduce the expression of autophagy in NSCLC cells compared with cisplatin treatment (Figure 4b). TEM analysis determining the features of autophagy ultrastructure indicated that cisplatin treatment  A-B. CCK-8 and EdU assay analyzing cell viability and proliferation after being transfected with miR-101-3p mimics or inhibitor, followed by treatment with cisplatin. C. Flow cytometry determined cell apoptosis following miR-101-3p mimics or inhibitor transfection, followed by treatment with cisplatin. *P < 0.05, **P < 0.01, ***P < 0.001 vs. Control.
3-MA treatment enhanced cell viability, while miR-101-3p mimics combined with 3-MA, the effect of increasing the sensitivity to cisplatin disappeared; RAPA treatment decreased cell viability, and compared with RAPA group, the effect of sensitivity to cisplatin had no significant difference in RAPA +miR-101-3p inhibitor together ( Supplementary  Figure 1a-B).

Knockdown of ATG4D reversed the role of miR-101-3p inhibitor on the sensitivity to cisplatin
To further confirm miR-101-3p regulated cisplatin sensitivity by regulation of ATG4D and autophagy, we co-transfected with miR-101-3p inhibitor and ATG4D siRNA to NSCLC cells, and compared the changes of cell activity, autophagy-related protein expression, number of autophagosomes and autolysosomes, and autophagy ultrastructure by CCK-8 assay, Western blot, confocal microscopy and TEM analysis, respectively. The results found that the miR-101-3p inhibitor treatment could decrease sensitivity to cisplatin, while knockdown of ATG4D could reverse miR-101-3p inhibitorinduced resistance to cisplatin in NSCLC cells (Figure 5 A); ATG4D siRNA also could reverse miR-101-3p inhibitor-induced the up-regulation of ATG4D and the ration of LC3II/LC3I, the down-regulation of p62 expression (Figure 5b); the total number of autophagosomes and autolysosomes was reduced after co-transfection with miR-101-3p inhibitor and ATG4D siRNA ( Figure 5c); moreover, the number of autophagosome that labeled red arrow was also reduced following combined with miR-101-3p inhibitor and ATG4D siRNA (Figure 5d). These data indicated that ATG4D could reverse the role of miR-101-3p inhibitor on the sensitivity to cisplatin by autophagy.

Up-regulated miR-101-3p could inhibit tumor the effect of transplanted tumor growth in vivo
Next, we further verify the effect of miR-101-3p in vivo, in vivo imaging showed that compared with control, the tumor volume decreased after cisplatin treatment, and compared to the cisplatin group, the tumor volume reduced following transfection with miR-101-3p mimic, the difference was statistically significant (Figure 6a). MiR-101-3p up-regulation plus cisplatin could obviously inhibit tumor grown, which had a similar effect on in vitro experiments ( Figure  6b). Immunohistochemistry analysis showed that the cisplatin group had obvious brown positively Figure 5. Knockdown of ATG4D reversed the role of miR-101-3p inhibitor on the sensitivity to cisplatin. A. The sensitivity to cisplatin transfected by miR-101-3p inhibitor, ATG4D siRNA or miR-101-3p inhibitor plus ATG4D siRNA under cisplatin treatment was detected by CCK-8 assay. B. The expression of autophagy-related protein LC3II/LC3I, p62 and ATG4D was examined by Western Blot. *P < 0.05 vs. Control. C. The level of autophagy after co-transfection with miR-101-3p inhibitor and ATG4D siRNA were examined by confocal microscopy analysis. D. TEM analysis determining the features of autophagy ultrastructure. expressed cells, and the level of ATG4D was reduced in the cisplatin group, while it was decreased following combined with miR-101-3p agomiR (Figure 6c). The above data revealed that miR-101-3p was up-regulated to inhibit ATG4D expression in the transplanted tumor tissues.

Discussion
Cisplatin is one of the best and first chemotherapeutic drugs for the treatment of NSCLC [26]; however, cisplatin does not show its highest potential because of side effects and drug resistance. Therefore, an understanding of the mechanisms of cisplatin resistance will be important to improve the efficacy of cisplatin treatment. It is well known that miRNAs are associated with drug resistance in various cancer cells [27,28]. Treatment with miR-423 inhibitor could significantly enhance sensitivity to drugs in breast cancer by reducing cell viability and increasing the apoptotic cells. We used TCGA and GEO databases to verify some miRNAs that are associated with cisplatin sensitivity in NSCLC (number: GSE135918). We screened top 15 differential expression miRNAs, identified that only miR-101-3p was significantly downregulated in 5 paired primary LUAD tissues and had a significant difference in survival curve. It was also down-regulated in NSCLC cancer cells, which suggested that miR-101-3p may be more suitable for the research objective of this article.
Autophagy is a conservative metabolic process, which is induced quickly to ensure the survival of tumor cells facing a harsh tumor microenvironment, for instance, hypoxia, starvation or chemotherapy drugs. Autophagy is also associated with the tumor development and progression, such as cell viability, cell resistance, and cell metastasis [29][30][31]. It has been reported that regulation of autophagy may be as an effective treatment strategy to improve the efficacy of many antitumor drugs [32][33][34]. Recent studies also revealed that miRNAs could regulate drug treatment-induced autophagy in many different cancers [27,28]. For instance, transfection with miR-874 mimics could sensitize gastric cancer cells to chemotherapy through regulating ATG16L1 and inhibiting autophagy [35]. Transfection with miR-153-3p mimics also could increase sensitivity to gefitinib in NSCLS cells via inhibiting autophagy and ATG5 [36]. Based on gene database and StarBase software analysis, we identified that the suitable miRNA (miR-101-3p) which was down-regulated in LUAD tissue and NSCLC cells. Later, it was also found that miR-101-3p could decrease the regulation of cisplatin induced autophagy in NSCLC cells. The activity of autophagy was associated with the number of LC3-II or the rate of LC3-II /LC3-I [37,38]. P62 can select the degradation of autophagy by combining cargo with autophagy Immunohistochemistry analysis of the expression of ATG4D in different treated groups. *P < 0.05, **P < 0.01. mechanism, which was also one of the most famous autophagy receptors [39]. The present study has demonstrated that transfection with miR-101a-3p and treatment with cisplatin could inhibit LC3II/LC3I ratio, the expression of autophagosomes and autolysosomes, increase p62 expression in NSCLC cells, but miR-101a-3p inhibitor had the opposite.
Autophagy-related protein 4D (ATG4D) and mTOR have been identified to play an important effect in the formation of autophagosome [40]. Up-regulation of miR-101 could enhance cisplatin-induced apoptosis and sensitivity to cisplatin in HCC cells by inhibiting autophagy [41]. However, the understanding of the mechanisms of miR-101-3p on resistance to cisplatin in NSCLC cells was not further researched. This study identified that ATG4D was a target gene for miR-101-3p in NSCLC cells and inhibition of ATG4D reversed the role of miR-101-3p inhibitor on the sensitivity to cisplatin in NSCLC cells. It also demonstrated that transfected with ATG4D siRNA also could reverse miR-101-3p inhibitorinduced the up-regulation level of autophagy. However, we also proved that miR-101-3p mimic could enhance the sensitivity to cisplatin in vivo. Additionally, further identification of the precise mechanisms underlying of ATG4D is needed, which will be the direction and focus of our future research.

In conclusion
Taken together, our findings demonstrated that over-repression of miR-101-3p enhanced cisplatin sensitivity by inhibiting cell viability and increasing cell apoptosis, inducing autophagy in NSCLC cells, and the potential mechanism was related with regulating of ATG4D expression, suggesting that the miR-101a-3p-ATG4D axis may become a novel prognostic and therapeutic target for overcoming cisplatin resistance in NSCLC.

Availability of Data and Materials
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.