Inhibition of interferon-gamma-stimulated melanoma progression by targeting neuronal nitric oxide synthase (nNOS)

Abstract

Interferon-gamma (IFN-γ) is shown to stimulate melanoma development and progression. However, the underlying mechanism has not been completely defined. Our study aimed to determine the role of neuronal nitric oxide synthase (nNOS)-mediated signaling in IFN-γ-stimulated melanoma progression and the anti-melanoma effects of novel nNOS inhibitors. Our study shows that IFN-γ markedly induced the expression levels of nNOS in melanoma cells associated with increased intracellular nitric oxide (NO) levels. Co-treatment with novel nNOS inhibitors effectively alleviated IFN-γ-activated STAT1/3. Further, reverse phase protein array (RPPA) analysis demonstrated that IFN-γ induced the expression of HIF1α, c-Myc, and programmed death-ligand 1 (PD-L1), in contrast to IFN-α. Blocking the nNOS-mediated signaling pathway using nNOS-selective inhibitors was shown to effectively diminish IFN-γ-induced PD-L1 expression in melanoma cells. Using a human melanoma xenograft mouse model, the in vivo studies revealed that IFN-γ increased tumor growth compared to control, which was inhibited by the co-administration of nNOS inhibitor MAC-3-190. Another nNOS inhibitor, HH044, was shown to effectively inhibit in vivo tumor growth and was associated with reduced PD-L1 expression levels in melanoma xenografts. Our study demonstrates the important role of nNOS-mediated NO signaling in IFN-γ-stimulated melanoma progression. Targeting nNOS using highly selective small molecular inhibitors is a unique and effective strategy to improve melanoma treatment.

Figure 1

IFN-γ increased melanoma invasion potential as detected by matrigel invasion assay. The represented image (a) was from A375 metastatic melanoma cells treated with IFN-γ (100 units/mL) for 16 h. The invasive cells that traversed the Matrigel and spread to the lower surface of the polyethylene filter membrane were stained with hematoxylin and eosin. The numbers in 10 vision fields were counted using a light microscope. (b) The invasive cell numbers after IFN-γ treatment increased compared to control. *p < 0.05 compared to control.

Figure 2

(a,b) Effects of IFN-α and IFN-γ treatment on nNOS expression levels in human metastatic melanoma A375 cells. Cells were treated with IFN-α or IFN-γ for various timepoints and whole cell lysates were collected. Samples were subjected to Western blot analysis for nNOS. A control of the protein loading was performed by detecting actin. Full length blots are presented in Supplemental Figure S2a,b, respectively. (c) Intracellular nitric oxide levels of A375 cells detected with a microplate reader using a DAF-FM fluorescence probe after IFN-α or IFN-γ (100 units/mL) treatment for 24 h, respectively. *p < 0.05 compared to control. (d) nNOS inhibitor reduced intracellular nitric oxide levels in A375 cells. Cells treated with 250 units/mL of IFN-γ with or without nNOS inhibitor, MAC-3-190, for 4 h followed by flow cytometry analysis.

Figure 3

(a) Effects of IFN-γ on STAT3 and phospho-STAT3 expression levels in melanoma. Primary melanoma wm115 cells were treated with IFN-γ or IFN-α (250 units/mL) for various timepoints. Whole cell lysates were collected for Western blot analysis to detect STAT3 and p-STAT3 levels, respectively. Specific nNOS inhibitor MAC-3-190 (3 µM) inhibited the activation of STAT3 (b) and STAT1 (c) expressions induced by IFN-γ. A375 cells were treated with IFN-γ (250 units/mL) with or without MAC-3-190 for 48 h. Shown is the mean ± SD, n = 3 for each experiment, *p < 0.05 compared to control, and #p < 0.05 compared to IFN-γ treatment. Full length blots are presented in Supplementary Figure S3a-c, respectively. (d) Chemical structures of novel nNOS inhibitors MAC-3-190 and HH044.

Figure 4

(a) Heat map of reverse phase protein array (RPPA) showing distinct effects of IFN-α and IFN-γ on protein expression levels in human melanoma cells. Three melanoma cell lines (wm115, Sk-Mel-28 and A375) were treated with 250 units/mL of interferons for 48 h. Whole cell lysates were collected and subjected to RPPA assay. The top 10 upregulated proteins by IFN-γ were selected from 302 proteins and phosphorylation of key signaling molecules. Red, above median; green, below median. All the data points were normalized for protein loading and transformed to linear values. The heatmaps included in supplemental data were generated in Cluster 3.0 (http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm) as a hierarchical cluster using Pearson Correlation and a center metric. (b) PD-L1, c-Myc and HIF1α were significantly induced by IFN-γ. Average changes of three cell lines detected by RPPA are shown in the figure. *p < 0.05 in comparison to that of control. (c) Distinct effects of IFN-α and IFN-γ on the expression levels of PD-L1. Human melanoma A375 cells were incubated with 250 units/mL of IFN-α or IFN-γ for 48 h, followed by detection of PD-L1 levels using flow cytometry. (d,e) The induction of PD-L1 by IFN-γ was diminished by the co-treatment of nNOS inhibitors. A375 melanoma cells were exposed to IFN-γ (250 units/mL) with or without 3 µM of nNOS inhibitor MAC-3-190 or HH044 for 48 h. The relative expression of PD-L1 on the cell surface was determined by flow cytometry. Representative histograms out of two independent experimental replicates are shown. *p < 0.05 compared to control; #p < 0.05 compared to IFN-γ alone. (f) Expression of PD-L1 in metastatic melanoma A375 cells detected by immunofluorescence staining. A375 cells were plated on coverslips and allowed to adhere overnight to 75% confluence then treated with IFN-α or IFN-γ (250 units/mL) with or without MAC-3-190 (3 μM) of 72 h. Cells were then fixed and permeabilized with 4% formaldehyde and methanol. Samples were blocked in blocking buffer containing 5% horse serum for 1 h. The slides were then allowed to incubate in a 1:50 PD-L1 antibody dilution overnight at 4 ºC and DAPI reagent for 1 h. Representative images are shown stained with PD-L1 antibody (green) and DAPI (blue fluorescence) (100× magnification). Representative images for two experimental replicates are shown.

Figure 5

Anti-melanoma activity of novel nNOS inhibitors. Metastatic melanoma A375 cells were injected to nude mice subcutaneously on the flank. The growth of tumor was measured daily and tumor volumes were determined using digital calipers (Fisher Sci) by using the formula tumor volume (mm³) = [Length × (Width²)]/2. Data are represented as mean ± SD. (a) nNOS inhibitor HH044 (10 mg/kg, i.p. daily) markedly inhibited the tumor growth of human melanoma in vivo compared to control (Control, n = 5; HH044, n = 4). (b) HH044 significantly decreased the final mass of xenograft tumors with no significant change in lung and body weight. *p < 0.05 compared to control; ns, p > 0.05, compared to control (Control, n = 5; HH044, n = 4). (c) PD-L1 expression of HH044 treated tumors was significantly decreased as detected by flow cytometry. *p < 0.05 compared to control (Control, n = 5; HH044, n = 4). Single cell suspensions of harvested tumor xenografts were stained with Alexa Fluor 488 conjugated PD-L1 antibody. The relative expression levels of PD-L1 were determined by the average fluorescence density as detected by flow cytometry. (d) nNOS inhibitor MAC-3-190 (5 mg/kg, i.p. daily) diminished the tumor growth stimulated by IFN-γ (1000 units, i.p. daily). *p < 0.05 compared to control; #p < 0.05 compared to IFN-γ treatment (Control, n = 7; IFN-γ, n = 11; IFN-γ + MAC-3-190, n = 5). (e,f) Expression levels of PD-L1 induced by IFN-γ treatment were inhibited by nNOS inhibitor MAC-3-190 in vivo. Metastatic melanoma A375 cells were injected to nude mice subcutaneously on the flank. IFN-γ (1000 units/day) was injected intraperitoneally once daily and nNOS inhibitor MAC-3-190 was administered i.p. daily at a dosing of 5 mg/kg for 21 days. The expression of PD-L1 in xenograft tumor samples were detected by immunohistochemistry staining in T-cell non-infiltrated area. By the end of study, xenograft tumors from different treatment groups were collected and specimens were fixed in a 10% formalin solution and embedded in paraffin wax for automatic processing using the Ventana Benchmark Ultra machine. Images of PD-L1 staining (brown) were captured in CD8-negative areas at 20× and 100× magnification, respectively. PD-L1 positive cells were quantified using ImageJ (https://imagej.nih.gov/ij/index.html) and represented as percentage of PD-L1 positive staining in the graph. Representative sections of each condition are shown. *p < 0.05 compared to control group; #p < 0.05, compared to IFN-γ group (n = 4 of each treatment).

Figure 6

nNOS plays a central role in interferon-γ-mediated melanoma progression. Studies have shown that UV radiation, especially at sunburn doses, causes immunological and inflammatory effects, damages the skin and stimulates the production of IFN-γ. IFN-γ is shown to promote inflammation, melanomagenesis and disease progression both in a transgenic mouse model and melanoma patients. Our study shows that IFN-γ triggers the activation of nNOS-NO signaling cascades associated with the activation of nuclear transcription factor STAT3. Abnormally high levels of NO fuel melanoma proliferation and facilitate the escape of cancer cells from immune surveillance by inducing the expression of PD-L1, which negatively regulates T-cell responses to tumor cells. nNOS inhibitors not only effectively reduce the production of NO, but also inhibit IFN-γ-stimulated PD-L1 expression and the activation of STAT1/3 signaling. Both in vitro and in vivo studies demonstrate that targeting nNOS-NO using small molecular inhibitors is a promising strategy for melanoma therapy.