Inhibition of NOS1 promotes the interferon response of melanoma cells

Human melanoma A375 cell line and murine melanoma B16 cell line were maintained in our laboratory. Cells were cultured in RPMI-1640 medium (Gibco, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (BI, Salt Lake City, UT, USA), and maintained at \(37\;^{\circ }\hbox {C}\) in a humidified atmosphere of 5% CO2.

CRISPR/Cas9 Plasmid p2u6 was approved by Zhili Rong, Ph.D. [23] The gRNAs were designed using the online software (http://​crispr.​mit.​edu). The gRNAs were amplified by PCR using Q5 High-Fidelity DNA Polymerase (NEB). To construct plasmids, PCR products and Plasmid p2u6 linearized by BaeI digestion were used for Gibson assembly (NEBuilder HiFi DNA Assembly Master Mix). The sequence of gRNAs is listed in Table 1.

The genomic DNA was extracted from the cell using an SDS lysis buffer. The genomic DNA was subjected to PCR amplification to identify the mutations. The mutation was confirmed by sequencing the PCR product, and the primers used are listed in Table 2.

Total protein was extracted by a Protein Extraction Kit (Fude, Hangzhou, CN). The concentration in each sample was determined using the Bradford assay (Beyotime, Shanghai, CN). Equal amounts of proteins were separated on 8–15% SDS-PAGE. Then the samples were transferred to PVDF membranes (Millipore, USA) and blocking with 5% BSA for 1h. Before corresponding with HRP-conjugated secondary antibodies, the membranes were incubated with the primary antibodies, anti-NOS1 (Abcam, Cambridge, MA, USA) at a 1:1000 dilution and GAPDH (Proteintech, Wuhan, CN) at a 1:5000 dilution, respectively. Then, the immunoreactive bands were visualized by chemiluminescence (Bio-Rad, USA).

Cells were implanted into 96-well plates and cultured overnight. The next day, 10 \(\upmu \)l 5 mg/ml MTT (Sigma-Aldrich, St. Louis, USA) was added to each well. After incubated for 2–4 h at 37 \(^{\circ }\hbox {C}\), 100\(\upmu \)l of DMSO was added to each well. Swirl gently for 15min at RT. After purple precipitate was fully dissolved, record absorbance at 490nm by BioTek Instruments (Winooski, VT, USA).

Cells were implanted into 6-well plates. Fresh medium was added to plates every 3 days. On day 10, cells were stained with 0.005% crystal violet and photographed after fixed with 4% paraformaldehyde for 15min.

Cells were implanted into 96-well plates and cultured overnight. The next day, \(50 \; \upmu \)M EdU was added to each well for 2h at \(37 \; ^{\circ }\hbox {C}\). Subsequently, the cells were fixed with 4% paraformaldehyde for 30min. Then, incorporated EdU was detected using the Cell-Light EdU Apollo 567 in Vitro Kit (RiboBio, Guangzhou, CN).

Cells were pelleted and washed with ice-cold PBS buffer, then incubated with PI staining solution at \(4 \; ^{\circ }\hbox {C}\) for 30min. The cell-cycle distribution was determined by FACS Aria Flow cytometer (Beckton Dickson, San Jose, CA, USA). PI staining solution: 25mg PI + 25mg Sodium Citrate + 5 mg RNase A + 750 ml Triton X-100 in 250 ml ddH\({}_2\)O.

All animals in this study, C57BL/6 mice and BALB/c-nu mice (Female, 6weeks old), were all purchased from Guangdong Medical Laboratory Animal Center.

Tumor xenograft assay was performed as described below. Briefly, \(5\times 10{}^6\) A375 cells (NOS1 KO or WT) were injected subcutaneously into BALB/c-nu mice(\(\hbox {n}=5\)). Tumor size was measured by caliper every 3 days. And tumor volume was calculated according to the following equation: tumor volume (mm\({}^3\)) = (length (mm) \(\times \) width\({}^2\) (mm\({}^2\))) \(\times \; 0.5\). The xenograft experiment was terminated when tumors were no more than 1400 mm\({}^3\) in volume.

To construct lung metastasis models of melanoma, \(5\times 10{}^5\) B16 cells (NOS1 KO or WT) were intravenously injected into C57BL/6 mice (\(\hbox {n}=7\)). Mice were euthanized on day 11 post-injection, and lung tissue was isolated and photographed. Paraffin sections 5 (\(\upmu \hbox {m}\)) were stained with H&E, examined by microscopy and photographed.

Total RNA was isolated using an RNAiso Plus reagent (Takara, Shiga, Japan) and reverse-transcribed using a PrimeScript RT kit (Takara). Then, TB Green Premix Ex Taq II (Takara) and LightCycler 96 System (Roche Life Science) were performed on quantitative real-time PCR (qPCR) analyses. Gene expressions were calculated using the comparative \(2{}^{-\Delta \Delta CT}\) method. The primer sequences for qPCR are listed in Table 3.

Total RNA was extracted by TRIzol (Invitrogen, CA) and purified with NucleoSpin miRNA (Macherey-Nagel, Duren, DE). Target preparation for microarray processing was carried out according to the GeneChip WT PLUS Reagent Kit. A total of 500 ng RNA was used for a double-round of cDNA synthesis. The cDNAs were fragmented, biotinylated, and hybridized to the Affymetrix Clariom S Assay human (Affymetrix, Santa Clara, CA). Following hybridization, the microarrays were washed and stained with Streptavidin Phycoerythrin on the Affymetrix Fluidics Station 450. Microarrays were scanned by using Affymetrix GeneChip Command Console (AGCC, Thermo Fisher Scientific, Inc) which installed in GeneChip Scanner 3000 7G. The microarray data are deposited in the GEO database with the accession GSE166287.

The data were analyzed with Robust Multichip Analysis (RMA) algorithm using default analysis settings and global scaling as normalization method Values presented are log2 RMA signal intensity. Normalized data were further analyzed using a moderated t-test to screen out the DEGs (differentially expressed genes) with Fold Change \(>1.2\) or \(<-1.2\), \(\hbox {FDR}<0.05\). The volcano plot was constructed using the R studio web server iBio Tools v5.0. Kyoto Encyclopaedia of Genes and Genomes pathway enrichment analyses of DEGs were performed by the KOBAS online database(http://​kobas.​cbi.​pku.​edu.​cn/​kobas3). Cluego plug-in in Cytoscape software (3.8 version) was used for showing the ClueGO network diagram. GSEA software(https://​www.​gsea-msigdb.​org/​gsea/​index.​jsp) was used to perform the gene set enrichment analysis. A normalized enrichment score (NES) and p-value \(<0.05\) were used to determine statistical significance.

Total RNA was extracted using TRIzol (Invitrogen) and genomic DNA was removed using DNase I (TaKara). A total of \(1 \;\upmu \hbox {g}\) mRNA was isolated following the TruSeq RNA sample preparation kit (Illumina, San Diego, CA). The double-stranded cDNA was synthesized using a SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA). Then, the synthesized cDNA was degraded and selected. After quantified by TBS380, paired-end RNA-seq sequencing library was sequenced with the Illumina HiSeq xten/NovaSeq 6000 sequencer. The RNA-seq data are deposited in the SRA database with the accession PRJNA700578. The raw paired-end reads were analyzed by SeqPrep (https://​github.​com/​jstjohn/​SeqPrep) and Sickle (https://​github.​com/​najoshi/​sickle) with default parameters. Then clean reads were separately aligned using TopHat (http://​tophat.​cbcb.​umd.​edu/​, version2.​0.​0) software.

DEGs were screened by R statistical package software EdgeR (Empirical Analysis of Digital Gene Expression in R, http://​www.​bioconductor.​org/​packages/​2.​12/​bioc/​html/​edgeR.​html) with Fold Change \(>1.2\) or \(<-1.2\), \(\hbox {FDR}<0.05\). Kyoto Encyclopaedia of Genes and Genomes pathway enrichment analyses of DEGs were performed by the KOBAS online database(http://​kobas.​cbi.​pku.​edu.​cn/​kobas3).

Clinical data were all performed using the NCI BRB-Array tool [24] and Qlucore Omics Explorer 3.2 (https://​www.​genomeweb.​com/​resources/​new-product/​qlucore-omics-explorer-32).

The Student’s t-test was used to analyze the two-sample comparison. A p-value \(<0.05\) was considered to be statistically significant. The values are expressed as the mean values ±S.D. All the statistical analyses were performed using Prism software, version 8.0 (GraphPad, Inc, San Diego, CA, USA).

To generate NOS1-deleted melanoma cell lines, we constructed a CRISPR/Cas9 plasmid with a gRNAs targeting the exon six of NOS1 gene (Fig. 1A and Additional file 1: Fig. S1) and transfected into A375 cells. Six clones with NOS1 mutated alleles were selected with puromycin and four clones (A2, A4, A5, and A6) were identified mutated successfully by PCR assay (Fig. 1B). Western blot assay confirmed that NOS1 expression was completely abolished in the A5 clone (Fig. 1C), and Sanger sequencing analysis showed a 180bp deletion and one site mutation at NOS1 alleles (Fig. 1D). Thus, we generated a NOS1-knockout melanoma cell line NOS1-KO A375 (clone A5) for the following experiments.

First, we detected the effect of NOS1 deletion on cell growth in culture using the MTT assay and clone formation assay. The data showed that the proliferation rate and the clonogenic ability of NOS1-KO A375 cells decreased by about 50% compared to WT cells (Fig. 1E, Fig. 1F, \(\hbox {p}<0.001\), \(\hbox {p}<0.05\)). Harvesting the exponential growth cells and staining with EdU, we compared the number of EdU positive (EdU+) cells between NOS1-KO A375 cells with control cells. NOS1 deletion decreased the proliferation of A375 cells (Fig. 1G). Cell cycle analysis with PI staining showed that the G1, G2, and S-phase were 58.95%, 17.17%, and 23.87% in NOS1-KO cells and 50.84%, 27.51%, and 21.65% in WT A375 (Fig. 1H), indicating NOS1 knockout may reduce G1/S phase transition and prevent cell proliferation.

The xenograft mouse model experiment (NOS1-KO and WT A375 cells in BALB/c-nu mice, five mice for each genotype) showed that NOS1 deletion dramatically reduced the tumor sizes, which were measured by three diameters and tumor weights, (\(\hbox {p} < 0.05\), Fig. 1I). These results suggested that NOS1 is essential for melanoma cell growth, especially for xenograft tumors.

We next explored the effect of NOS1 deletion on the transcription profile of melanoma A375 cells. Using Affymetrix microarray, we identified 2203 differentially expressed genes (DEGs) in NOS1-deleted cells, including 1138 downregulated and 1065 upregulated genes (cutoffs of \(\hbox {p} < 0.05\) and \(\hbox {FC} \ge 1.2\) or \(\le 1.2\); Fig. 2A). KEGG analysis showed that the top ten pathways of downregulated DEGs are enriched in metabolic processes, while the top ten pathways of upregulated DEGs were involved in cancer signaling pathways cellular interaction, and focal adhesion (Fig. 2B).

The top one is metabolic pathways, which involved 164 genes. Network analysis using the CLUEGO of Cytoscape software showed that these metabolic genes were associated with twelve metabolic pathways, including purine metabolism, amino acid metabolism, carbon metabolism, and redox balance or biotransformation (Fig. 2C). Thus, transcription data and cell experiments in vitro indicated that NOS1 expression plays a role in cell cycle regulation of melanoma cells.

The transaminase GOT and GPT, and a key enzyme for the regeneration of arginine (ASS1) and CAD complex (carbamoyl-phosphate synthetase 2, aspartate transcarbamylase, and dihydroorotase) are downregulated in NOS1-deleted cells, implicating that NOS1 participates in ammonia metabolism and arginine regeneration. It has been reported that increased intracellular polyamine concentrations are essential for cell proliferation and tumorigenesis. Ornithine decarboxylase (ODC) is the rate-limiting enzyme for polyamine synthesis, and spermidine/spermine N-acetyltransferase (SAT1) a critical enzyme for polyamine exclusion, both of which are key regulators for intracellular polyamine concentration. We observed that the ODC expression decreased and the SAT1 level increased in NOS deleted cells, suggesting NOS1 plays a critical in regulating intracellular polyamine concentration (Fig. 2D).

Gene set enrichment analysis (GSEA) found that the gene expression of NOS1 deletion mainly enriched in 37 processes-related KEGG pathways (Additional file 2: Table S1). The up-regulation pathways (21 terms) mainly exhibited enrichment of immune disease and regulation, whereas the down regulation pathways (16 terms) showed enrichment of metabolism and cellular progresses. We noticed that the genes in JAK-STAT signaling pathway, a key signaling of innate immune regulation, significantly enriched in upregulated genes of NOS1 deletion (Fig. 2E). We tested the expression of components in JAK-STAT signal in NOS1-deleted cells using qPCR. Nine of 12 tested genes were upregulated in NOS1-deleted cells (Fig. 2F).

Dysfunction of JAK-STAT signal in cancer cells is crucial for cancer immune escape [21]. Therefore, we explored the role of NOS1 in the expression of genes in response to IFN\(\upalpha \) (ISGs). A375 cells were stimulated with IFN\(\upalpha \) (1000 IU/ml) for 12 hours with or without NOS1 inhibition (\(100\upmu \)M NPLA) and RNA-sequencing (RNA-Seq) was performed for gene expression profiling. The differential expression genes between cells with or without IFN\(\upalpha \) stimulation were defined as ISGs. 290 ISGs were identified in A375 (CON-ISGs), while 460 ISGs identified in the NPLA-treatment group (NPLA -ISGs). Workflow of genomic comparison of NOS1-inhibitor versus NOS1-WT A375 cells was present in Fig. 3A.

KEGG analysis of ISGs indicated that NOS1 deletion induced upregulation of cytokine-receptor interactions and downstream signals of JAK-STAT signaling, TOLL-like receptor signaling, and NODs receptor signaling, but not chemokines signaling pathways (Fig. 3B). These data suggested the NOS1 may act mainly on pathways of the immune reaction regulation but not the chemotaxis of immune cells.

KEGG analysis showed that NOS1-ISGs enriched in immunoreaction processes, cytokine-receptor interactions, JAK-STAT signaling, TOLL-like receptor signaling, and NODs receptor signaling, antigen processing and presentation, complement and coagulation cascades (Fig. 3C, D). NOS1 inhibition increased expression of TLR3, MYD88, STAT1/2, OAS1, AIM2, NOD2 and GBP3/5, key factors in response to IFN\(\upalpha \) stimulation. Thus, these data suggested that NOS1 is involved in innate immune escape of melanoma.

To validate the role of NOS1 in immune escape, we established lung metastasis of melanoma cells with or without NOS1 expression (NOS1-KO B16 and WT B16) in immune competent C57BL/6 mouse. MTT assay indicated consistently that the cell proliferation of NOS1-KO B16 was slower than WT B16 (\(\hbox {p} < 0.001\), Additional file 3: Fig. S2). \(5\times 10{}^5\) cells harvested at logarithm stage in culture were injected into C57BL/6 mice through tail vein for generating lung metastasis model (\(\hbox {n}=7\)). Ten days after injection, mice were sacrificed humanely and the number and size of metastatic nodules were evaluated. As shown in Fig. 4, the size of tumor nodes were significantly smaller in NOS1-KO groups (\(\hbox {p}<0.001\), Fig. 4A).

The histopathology with hematoxylin and eosin (H&E) staining showed that nodules of NOS1-KO melanoma have less number of melanoma cells and display poor cell morphology compared with the WT group. Ki67 staining also showed that proliferating cells in NOS1-KO nudes were significantly less than control groups (\(\hbox {p}<0.001\), Fig. 4B). The immune cell infiltration in lung nodules were detected by staining with the pan-T cell marker CD3 (\(\hbox {p}<0.05\), Fig. 4B). The percentages of CD3 positive cells in NOS1-KO nodules were significantly higher than the control group (Fig. 4B). The results confirmed that NOS1 knockout inhibited melanoma growth of lung metastasis and increased the immune cell infiltration. However, the immune inhibitory role modification by NOS1 expression might also contribute to the involve in its growth promotion in of melanoma.

To better understand the role of NOS1 expression on immune modulation and clinical significance in melanoma, we investigated a cohort of transcription data of melanoma under high dose of rIL2 treatment down load from GEO (GSE32611). We analyzed the correlation of NOS1 expression with the key components (TLR3, MYD88, STAT1/2, OAS1, NOD2 and GBP3/5) in immune response pathways, JAK-STAT, NOD-LIKE, TOLL-LIKE pathway. Before rIL2 treatment, only GBP3 was negatively correlated with NOS1 expression (Fig. 5A). However, expression of eight genes encoding key immune factors were all negatively correlated with NOS1 after immune therapy (TLR3, STAT1, and GBP3/5 with \(R{}^2> 0.25\); Fig. 5A). Consistent with our transcription data in vitro, NOS1 expression is correlated negatively with the function of immunomodulation pathways.

We further explored the clinical significance of NOS1 expression and its immune modulation from TCGA database of melanoma patients. Of 471 melanoma clinical samples, NOS1 expressed low level in most samples and found no different among four clinical stages (Additional file 4: Fig.S3) and could not predicate the survival prognosis of patients (data not shown). With stringent cut off (p-adjust \(< 0.05\)), 16 ISGs were identified in A375 (con-ISGs), while 42 ISGs identified in the NPLA-treatment group (NPLA -ISGs). Venn diagram classed 30 genes uniquely in NOS1 inhibited group, which defined as NOS1 modulated immune response gene (NOS1-ISGs) (Fig. 5B). Immune gene signatures have been shown to be associated with patient outcomes such as disease-free and overall survival, response to chemotherapy. We downloaded 471 transcriptional profiles associated with their clinical parameters. Through hierarchical clustering analysis (Qlucore Omics Explorer 3.2), 14 NOS1-ISGs self-organize 471 patients into major two groups, high or low NOS1-ISGs groups. Examined the prognostic performance of the two groups by Kaplan–Meier analysis showed that the overall survival time of the high-NOS1-ISGs group was significantly longer than that of the low -NOS1-ISGs group (\(\hbox {p}<0.001\), Fig. 5C). We further analyzed the tumor remission by using these 14 NOS1-ISGs in 19 patients under chemotherapy (GSE10282). As showed in Fig. 5D that the tumor remission in high-NOS1-ISGs group was more significant than low-NOS1-ISGs group.