Oxidative Stress in Malignant Melanoma Enhances Tumor Necrosis Factor-α Secretion of Tumor-Associated Macrophages That Promote Cancer Cell Invasion

Abstract

Aims: Malignant melanoma is well known for abundant reactive oxygen species (ROS) that exist in the primary tumor environment. Within this microenvironment, tumor-associated macrophages (TAMs) play substantial roles in multiple steps of tumor development in terms of tumor growth, invasion, and metastasis. We therefore aimed to determine whether this high-level ROS in primary melanoma is capable to promote tumor invasiveness by influencing TAM properties. Moreover, we wanted to further investigate probable underlying mechanisms. Results: We characterized malignant melanoma TAMs as a heterogeneous phenotype, which possesses both M1 and M2 markers. We also revealed a role for high-level intracellular ROS in enhancing proinvasion signature of TAMs by strongly increasing their tumor necrosis factor α secretion, which is possibly attributed to ROS-enhanced peroxisome proliferator-activated receptor γ (PPARγ) translocation mediated by MAPK/ERK kinase 1. Innovation: This is the first study demonstrating that high levels of ROS in the primary melanoma environment can influence TAM behaviors. Furthermore, we are also the first to indentify that nucleus-to-cytoplasm translocation of PPARγ is significantly upregulated by ROS and responsible for the proinvasiveness capacity of melanoma TAMs. Conclusion: Taken together, our data describe how a high level of ROS plays a critical role in enhancing the proinvasion characteristic of TAMs in malignant melanoma. Antioxid. Redox Signal. 19, 1337–1355.

Introduction

H uman malignant melanoma is notorious worldwide for its rapidly rising rate of incidence and comparatively poor prognosis. To date, there is still no effective treatment for advanced disease (16, 46). Compared to other types of tumors, melanoma is relatively unique due to the extraordinarily high levels of oxidative stress in the primary tumor environment (18, 27). Because of structurally aberrant melanosomes of melanoma cells, exogenous attacks, and the infiltration of immune cells, large amounts of reactive oxygen species (ROS) are generated within tumor sites. These events exert selective pressure on the tumor cells and orchestrate cellular signaling networks, aggravating tumor cell proliferation, drug resistance, and malignancy (18, 63). Additionally, oxidative stress is also inevitably imposed on other melanoma tissues (48, 62) and stromal cells such as epithelial cells and T cells (50).

Innovation

Our study is the first to demonstrate that reactive oxygen species (ROS), which is robustly generated in primary melanoma, has the ability to influence behaviors of tumorassociated macrophages (TAMs). In the melanoma microenvironment, tumor necrosis factor (TNF)-α secretion of TAMs is greatly enhanced by ROS, thereby increasing tumor invasion and even metastasis. Furthermore, this enhancementof TNF-α secretion is proved to be contributed by ROS-increased peroxisome proliferator-activated receptor γ translocation through the MAPK/ERK kinase 1-mediated pathway. These findings accomplish a large step toward defining the role of ROS in the modulation of immune cells in malignant melanoma, and also suggest several approaches for modifying antimelanoma therapies.

The progression and pathogenesis of melanoma strongly depends on the interactions between cancer cells and the various stromal cells in the host (31, 70). Among them, tumor-associated macrophages (TAMs), which are majorly derived from circulating monocytes and are acknowledged as the most abundant leukocytes in melanoma lesions (constituting up to 30% of tumor cells) (24, 56). Although some reports describe their antitumor traits as M1 macrophages (classically activated macrophages) (34), they have been increasingly characterized as M2 macrophages (alternatively activated macrophages) within many kinds of tumors, closely associated with tumorigenesis, tumor development, and tumor metastasis (43, 51). Accumulating evidence has indicated that increased figure of TAMs infiltrating melanoma and high level of macrophage markers expression in melanoma tissue are strongly related with poor prognosis (57, 26). One crucial property of TAMs is their ability to enhance the aggressiveness of cancer cells (41). It has been reported that the secretion of diverse factors such as the epidermal growth factor (19), urokinase plasminogen activator (uPA) (22), and matrix metalloproteinases (MMPs) (33) by TAMs can accelerate tumor cell intravasation and therefore facilitate tumor metastasis. Recently, the tumor necrosis factor (TNF)-α, a wide-known inflammatory cytokine, has been indicated to be responsible for the proinvasiveness characteristic of TAMs. TAM-secreted TNF-α was found to facilitate the invasiveness of MCF-7 and SK-BR-3 breast cancer cells by increasing MMPs in the coculture medium (21). Furthermore, the invasive ability of T47D ductal carcinoma cells was enhanced by TNF-α produced by TAMs (69). In addition, TAM-generated TNF-α can activate the nuclear factor κB (NF-κB)-Snail pathway in different types of tumor cells, promoting their invasiveness (65).

It has been demonstrated that ROS play a substantial role in upregulating inflammatory cytokine secretion of macrophages in several pathological conditions (15, 42). Although the level of oxidative stress has been reported to decrease in TAMs in advanced colon carcinoma and thymoma (13, 14, 67), melanoma TAMs could be one of the potential regulating targets of abundant ROS within the melanoma microenvironment.

Peroxisome proliferator-activated receptor γ (PPARγ) is a multifunction nuclear receptor that has been proved to have anti-inflammation effects in various macrophage models, inhibiting the expression of many proinflammation genes, including the genes of IL-1, IL-6, TNF-α, IL-12 and inducible nitric oxide synthase (iNOS) (2, 3, 45). Until now, several PPARγ regulating processes, which have influence on its activity or expression have been proposed, including ligand-induced activation, proteasomal degradation of activated PPARγ, MAPK-mediated phosphorylation, and newly found subcellular compartmentalization possibly due to MAPK/ERK kinase 1 (MEK-1) binding. Moreover, PPARγ has been also reported to be regulated by oxidative stress (40), providing us another potential target in interrogating the molecular mechanisms in melanoma TAMs.

Considering ROS has been employed as one of the therapeutic targets in increasing number of clinical studies, there is a necessity to thoroughly characterize the role of ROS in the influence of melanoma TAM behaviors. Herein, we are the first to underscore a critical role of oxidative stress in the regulation of TAM-promoted invasion in malignant melanoma. Furthermore, we have also elucidated the relevant molecular mechanisms involving ROS-enhanced TNF-α secretion resulted from PPARγ translocation.

Results

High level of ROS exists in malignant melanoma TAMs

We established subcutaneous melanoma models using B16F1 (a low metastatic variant) and B16F10 (a high metastatic variant) cells in C57BL/6 mice. First, the presence and location of TAMs were examined by H&E staining and immunofluorescence. As shown in Supplementary Figure S1A (Supplementary Data are available online at www.liebertpub.com/ars), apparent inflammatory cell infiltration could be seen at the edge of both tumors where is called “invasive fronts” (indicated by the white arrow). Consistently, F4/80 positive cells were also observed at a considerable level in the peripheral part of both F1 and F10 melanomas, while few were observed in the intratumoral mass of both tumor types (Fig 1A). The existence of abundant ROS in malignant melanoma is widely reported. Therefore, we hypothesized that there may be some disparity in the ROS levels between subcutaneous melanomas established from different variants. To address this, we tested ROS abundance using in situ dihydroethidium (DHE) staining of the tumor tissue. As expected, although both tumor tissues had distinctly more ROS compared to the normal skin control, the ROS level in the B16F10 tumor was significantly higher than its B16F1 counterpart (Fig. 1B). Consistent with this, the glutathione disulfide (GSSG)/glutathione (GSH) ratio of the tumor tissue lysate also showed similar results (Fig. 1C). We then assessed the intracellular ROS level of TAMs isolated from these tumors with both 2′,7′-dichlorofluorescin-diacetate (DCFH-DA) and DHE probes. As shown in Figure 1D, after transient stimulation (2 h) with a tumor cell conditioned medium postisolation, TAMs from the B16F10 melanoma line exhibited much higher intracellular ROS levels than their B16F1 counterpart. To further study melanoma TAMs in vitro, we introduced the tumor-conditioned medium-educated macrophages (TCMEMs), an in vitro model commonly used to mimic the phenotypes and functions of TAMs (44, 52). Interestingly, regardless of the specific tumor microenvironment, similar results were seen regarding the intracellular ROS levels in RAW 264.7, ANA-1, and peritoneal TCMEMs using the DCHF-DA probe (Fig. 1E, Fig. 1B, C). To validate this result, we repeated this experiment in RAW264.7 macrophages with another ROS detection kit based on the Fenton Reaction, exhibiting a similar ROS increasing pattern (Fig. 1F).

FIG. 1.

High levels of reactive oxygen species (ROS) exist in tumor-associated macrophages (TAMs) of malignant melanomas. (A) To verify the macrophage infiltration in malignant melanoma with 3 weeks growth postsubcutaneous inoculation, F4/80-labeled macrophages were observed via immunofluorescence in both peripheral (outer) and intratumoral (inner) areas. (B) After 3 weeks growth, B16F1 and B16F10 subcutaneous tumors were quickly removed from sacrificed mice, and then immediately frozen at −80°C. The ROS levels were detected with in situ dihydroethidium (DHE) staining on frozen sections. (C) The glutathione disulfide (GSSG)/glutathione (GSH) ratios of tumoral tissue from both B16F1 and B16F10 melanomas with 3 weeks growth and normal skin tissue were tested using GSH and GSSG Detection Kit. (D) Three weeks after inoculation, the TAMs were isolated from subcutaneous melanomas using density gradient centrifugation (percoll) followed by purification via sorting with the F4/80 antibody. Then, intracellular ROS levels of these cells were tested after transient stimulation (90 min) with a relevant tumor-conditioned medium (TCM). (E) For time sequential detection of the intracellular ROS level during the education process of TCM-educated macrophages (TCMEMs), RAW 264.7 macrophages were incubated with TCMs for either 0 h, 6 h, 12 h, 24 h, 36 h, or 48 h. The incubation was followed by flow cytometry analysis with 2′,7′-dichlorofluorescin-diacetate (DCFH-DA) staining, and the measurement of histogram peak shifting was shown as geometric mean (GM). (F) RAW 264.7 macrophages were stimulated with TCMs for either 0 h, 6 h, 12 h, 24 h, 36 h, or 48 h. Then, the intracellular peroxide levels were examined using Hydrogen Peroxide Assay Kit, the relative H₂O₂ concentration (count one ROOH molecule as one H₂O₂ molecule) of cell lysate (1 million cells in 100 μl lysate buffer) was shown in the figure. *p<0.05.

TAMs of malignant melanoma exhibit enhanced antioxidant system, heterogeneous phenotype, and a proaggressiveness signature

To further explore the redox status and functional phenotype of malignant melanoma TAMs, we performed cDNA microarrays to interrogate the gene expression pattern of B16F10 TAMs by comparing them with normal peritoneal macrophages. A total of 35556 genes were analyzed, among them 621 genes altered by more than fourfold, with 381 upregulated and 240 downregulated (Supplementary Table. S1). To gain deeper insight into the redox condition of melanoma TAMs, we utilized clustering analysis. As seen in Figure 2A, among 33 obviously altered genes (fold-change <−1.5 or > 1.5, p-value <0.05), which are closely related to antioxidant systems, only four of them were downregulated, while others exhibited upregulation, further confirming the results from Figure 1. We continued this study with interrogating the generating mechanism of macrophage intracellular ROS. The NADPH oxidase 2 (NOX2) complex is widely known for massive ROS production during respiratory burst when macrophages confront pathogens (6). Surprisingly, our gene profiling data revealed that all the members in NOX2 underwent apparent downregulation (Table. 1). However, the nuclear cofactor PPARγ coactivator 1-α (PGC1-α) and its downstreaming product transcription factor A, mitochondrial (TFAM), which are inseparately linked to mitochondria biosynthesis (58) were hugely elevated, so as to NADH dehydrogenase (ubiquinone) flavoprotein 2 (NDUFV2) on complex I and the Rieske Fe-S Protein (complex III), which are two main ROS-producing points on the mitochondrial respiratory chain complex (29, 38), indicating mitochondrial ROS generation is one of the main sources of TAMs intracellular oxidative stress. Next, we set out to examine the plasticity of melanoma TAMs. Thirty most renowned and acknowledged M1 and M2 activation markers (15 each, p value < 0.05) (7, 32) were analyzed according to our gene profiling data, clustering figure Figure 2B reveals that 9 m1 genes and 5 m2 genes were increased, while others display diminution of their expressions, suggesting these TAMs had an intermediate phenotype that does not fit the characteristic of either M1 macrophages or M2 macrophages. This conclusion was verified by the measurements of CD206 expression, the secreted TNF-α concentration, the iNOS activity, and the arginase-1 (Arg-1) activity in RAW264.7 TCMEMs (Supplementary Fig. S2). On top of these analyses, pathway analysis was also introduced based on strongly upregulated genes (fold-change >4). Twenty-one pathways were found statistically significant under a p value 0.05 (Supplementary Table. S2). Among them, seven pathways are linked to the metabolism process, showing a much greater energy expenditure of TAMs than peritoneal control. Other signaling pathways such as the toll-like receptor signaling pathway, cell adhesion molecules, focal adhesion, and leukocyte transendothelial migration depict a signature of inflammatory activation and a pro cell migration activity, indicating roles of these cells in tumor invasion and metastasis. Consistent with this finding, several protease genes that contribute tumor invasiveness such as mmp2, mmp9, mmp12, spp1 (osteopontin), ctsl (cathepsin L), and plau (urokinase) were also significantly elevated (Table 1). In general, the data of microarray study indicated that malignant melanoma TAMs, which are under dramatic oxidative stress exhibit a phenotype not fitting the orthodox M1/M2 paradigm and possibly possess a proagressiveness signature.

FIG. 2.

Microarray analysis of TAMs isolated from B16F10 melanoma. Three weeks after subcutaneous plantation, TAMs of B16F10 melanoma were purified and subjected to the assessment of gene expression profile. (A) Heat map of 33 genes (p-value < 0.05, fold-change > 1.5 or < −1.5) closely linked with antioxidant system is shown. (B) Clustering figure based on 30 genes (p-value < 0.05) of most renowned and acknowledged M1 and M2 macrophage markers is exhibited. (C) Ten most significant genes were verified with real-time polymerase chain reaction. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Table 1.

Microarray Assay of Tumor-Associated Macrophages Isolated from B16F10 Melanoma

Gene category	Gene symbol	Protein symbol/name	p-value	Fold-change
NOX2 complex	cybb	NOX2	2.29E-08	−4.86698
	ncf2	P67phox	0.00016044	−1.42517
	ncf1	P47phox	4.84E-06	−1.56401
	ncf4	P40phox	0.00104345	−1.61279
	cyba	P22phox	6.93E-06	−2.54903
Mitochondria biogenesis & respiratory chain complex	ppargc1a	PGC-1α	1.85E-08	13.7079
	tfam	TFAM	0.00049442	1.59863
	ndufv2	NDUFV2	5.02E-05	2.90273
	uqcrfs1	Rieske Fe-S protein	0.00048207	1.36161
Proteases	mmp8	MMP8	5.14E-06	7.06739
	mmp9	MMP9	2.68E-06	2.34575
	mmp12	MMP12	4.86E-06	1.92327
	spp1	osteopontin	2.17E-08	13.3246
	ctsl	cathepsin L	2.86E-05	1.6189
	plau	urokinase	8.80E-07	4.0491

TAMs of malignant melanoma significantly enhance tumor cell invasiveness by TNF-α secretion

Our histology and microarray data have already suggested a vital role of TAMs in promoting tumor cell invasion and metastasis in malignant melanoma. To address this conjecture, we sorted TAMs from both B16F1 and B16F10 subcutaneous melanomas and subjected them to an in vitro assay that was widely utilized previously. As shown in Figure 3A, TAMs from both B16F1 and B16F10 tumors possess more capacity to enhance the tumor cell invasion compared with peritoneal macrophage controls. Interestingly, TAMs isolated from B16F10 tumors gave the greatest rise to tumor cell invasion, no matter cocultured with B16F1 tumor cells or B16F10 cells. Tumor migration ability is also vital in the metastasis process, we measured the promigration capacity of TAMs from two types of melanomas using the wound-healing assay, yielding quite similar enhancing patterns with an invasion test (Supplementary Fig. S3A). We also analyzed the proinvasion capacity of RAW264.7 and ANA-1 TCMEMs with the invasion assay. Likewise, macrophages that were incubated with the B16F10 tumor-conditioned medium (TCM) seemed to have the greatest ability to enhance tumor cell invasion (Fig. 3B and Supplementary Fig. S3B). Next, we set out to elucidate the molecular mechanisms underlying the proinvasion signature of melanoma TAMs. Recently, the inflammatory cytokine TNF-α has been demonstrated to induce tumor invasiveness in several tumor types. Therefore, we hypothesized that it could play the same role in malignant melanoma cells. Our microarray analysis already found more than threefold increase of tnf gene trancription in melanoma TAMs compared to peritoneal controls; this was confirmed with the measurement of TAM-secreted TNF-α by detecting its concentration in the postisolation culture medium with the enzyme-linked immunosorbent assay (ELISA) test. Resembling F1CM and F10CM RAW 264.7 TCMEMs, B16F10 TAMs also produced much more TNF-α than their B16F1 counterpart (Supplementary Fig. S2A, Fig. 3C). To verify whether TNF-α is capable to enhance tumor aggressiveness, we performed an invasion assay with only B16F1 or B16F10 melanoma cells in the presence of exogenous recombinant TNF-α. The results showed that TNF-α can dramatically give rise to their invasiveness in a concentration-dependent manner (Supplementary Fig. S3C). Consistently, the neutralization of TNF-α in a cell invasion assay with an anti-TNF-α antibody significantly decreased TCMEM-promoted tumor invasion (Fig. 3D). To further substantiate the role of TNF-α in the proinvasion trait of melanoma TAMs, we brought in TNF-α knockout mice. Notably, as seen in Figure 3E, the absence of TNF-α nearly completely depleted peritoneal macrophage TCMEM-induced tumor invasiveness compared with wild-type (WT) control. We also observed the development of in vivo metastasis in knock-out (KO) mice. Although there were negligible differences of lung metastasis and inguinal lymphatic gland metastasis on day 21 after foot-pad plantation between KO mice and WT counterpart (Supplementary Fig. S3D, E), much more severe metastasis could be seen in popliteal lymph nodes of TNF-α KO mice (Fig. 3F), indicating a role of TNF-α in promoting lymphatic metastasis of malignant melanoma. Taken together, these data confirm that TNF-α secretion, to a great extent, was responsible for proinvasion and prometastasis signatures of malignant melanoma TAMs.

FIG. 3.

Malignant melanoma TAMs significantly enhance tumor cell invasiveness by tumor necrosis factor (TNF)-α secretion. (A) TAMs isolated from malignant melanoma cells grown for 3 weeks were subjected to in vitro invasion assay with the Transwell (8 μm) apparatus. The results are shown as both microscopic pictures (qualitative data) and optical density (OD.) values of resolubilized crystal violet measured at a wavelength of 600 nm (quantitative data). The data were first grouped into “F1 Invasion” and “F10 Invasion” by the type of tumor cells in the Transwell insert, and then subcategorized by various kinds of macrophages in the well beneath the insert. “Negative Control” means there were no macrophages. “F1 TAM” or “F10 TAM” represents the TAMs isolated from B16F1 or B16F10 subcutaneous melanoma. (B) RAW264.7 macrophages stimulated with B16F1 or B16F10 TCM for 48 h were subjected to in vitro invasion assay. “No Stimulation” represents unstimulated macrophages, while “F1CM” or “F10CM” stands for the macrophages incubated with B16F1 TCM or B16F10 TCM. (C) Around 10000 peritoneal macrophages or the same number of melonama TAMs were cultured in 150 μl of a regular medium for 12 h after purification. Then, the TNF-α concentration in the culture medium was examined by enzyme-linked immunosorbent assay (ELISA) assay. (D) To address the significance of TNF-α in TCMEM-induced tumor invasion, RAW 264.7 TCMEMs were subjected to an in vitro invasion test with or without the addition of 1:50 TNF-α antibody in the coculture medium. “F1CM+TNF-α Ab” or “F10CM+TNF-α Ab” means when macrophages educated with B16F1 TCM or B16F10 were in invasion assay, the TNF-α antibody was added in the coculture system. (E) Peritoneal macrophages of TNF-α^−/− mice were incubated with B16F1 or B16F10 TCM for 48 h, and then proinvasion capacity of these TCMEMs was examined with invasion assay, compared with wild-type peritoneal macrophage counterparts. (F) Three weeks after foot-pad subcutaneous inoculation in TNF-α^−/− or wild-type mice with 1million B16F10 melanoma cells, the popliteal lymph node of the left leg was taken out for photographing and subjected to H&E staining. The metastatic area on histology image is emphasized by black arrows. *p<0.05, **p<0.01.

High level of oxidative stress strengthens the invasion-promoting phenotype of malignant melanoma TAMs

We hypothesized that the large amount of ROS found in malignant melanoma may critically affect the proinvasion signature of TAMs. Therefore, antioxidants N-acetylcysteine (NAC), Vitamin E (Ve), or pro-oxidant H₂O₂ were introduced in the 48-h education process of TCMEMs to elevate or reduce intracellular oxidative stress. As observed in Figure 4A and Supplementary Figure S4, the intracellular ROS level of TCMEMs was strongly affected by these antioxidants and oxidants. Next, these TCMEMs were subjected to in vitro invasion assay. The result reveals that the proinvasion capacity of TCMEMs whose education process was accompanied with a low ROS level was apparently attenuated. However, the addition of H₂O₂ in the TCMEM education process could considerably enhance the proinvasion signature of these cells (Fig. 4B). Moreover, the result from the TNF-α ELISA assay seems compatible with the data of invasion assay (Fig. 4C). Likewise, the diminution of intracellular ROS during the education process obviously decreased the promigration capacity of TCMEMs in wound-healing assay (Fig. 4D). Hence, the high level of intracellular ROS in melanoma TAMs is likely to facilitate TAM-enhanced tumor invasiveness by increasing TNF-α secretion.

FIG. 4.

The high level of intracellular ROS in TAMs enhances the protumor cell invasion ability of TAMs in malignant melanomas. (A) To analyze the role of intracellular ROS in the proinvasion property of TCMEMs, RAW264.7 macrophages were incubated with B16F1 or B16F10 TCM with or without N-acetylcysteine (NAC) (10 mM), vitamin E (Ve) (10 mM), or H₂O₂ (100 μM) for 48 h. Then, the intracellular ROS levels of these macrophages were tested by flow cytometry with DCFH-DA staining. (B) RAW 264.7 macrophages were stimulated with melanoma TCM for 48 h with or without NAC (10 mM), Ve (10 mM) or H₂O₂ (100 μM). Then, they were subjected to invasion assay to test their proinvasiveness capacity. The results are exhibited both qualitatively and quantitatively. (C) To investigate the role of intracellular ROS in TNF-α production by TCMEMs, RAW 264.7 macrophages were incubated with melanoma TCM for 48 h with or without NAC (10 mM), Ve (10 mM), or H₂O₂ (100 μM). Then, concentrations of TNF-α in the culture medium were detected by the ELISA test. (D) RAW 264.7 macrophages were stimulated with melanoma TCM for 48 h with or without the presence of 10 mM Ve. Then, the promigration capacity of these cells was tested by wound-healing assay with Transwell apparatus (0.4 μM). Images were taken at the same location of the wound field at both 0 h and 24 h. *p<0.05, **p<0.01, ***p<0.001. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

PPARγ functioned against proinvasiveness signature of malignant melanoma TAMs by downregulating TNF-α production

It has been demonstrated that PPARγ plays a considerable role in regulating the secretion of various inflammatory cytokines in different kinds of macrophages. Hence, we presumed that it also has a similar function in melanoma TAMs. To address this hypothesis, we knocked down PPARγ in RAW264.7 macrophages using small interference RNAs (siRNAs) and performed tumor cell invasion assay in RAW264.7 TCMEMs. Efficient knockdown of the PPARγ protein was shown in Supplementary Figure S5. As we expected, a decrease of PPARγ in TCMEMs led to a significant increase in proinvasion capacity of these cells (Fig. 5A). Consistently, the TNF-α level in the culture medium of PPARγ knockdown TCMEMs was also significantly higher than scrambled control (Fig. 5B). The functions of PPARγ are strongly regulated by its binding ligands. Therefore, we introduced both potent selective agonist and antagonist of PPARγ (GW1929 and GW9662, respectively), in our study to further elucidate the role of this nuclear receptor. As shown in Figure 5C and D, incubation with GW1929 distinctly reduced TNF-α secretion by TCMEMs, resulting in a decreased ability to enhance tumor invasiveness. In contrast, the presence of GW9662 strongly gave rise to the TNF-α production level as well as proinvasion capacity of TCMEMs. To further explore the significance of TNF-α secretion decrease in the downregulation of TAM-enhanced tumor invasiveness by PPARγ, we stimulated TNF-α knockout peritoneal macrophages with melanoma TCMs with the presence of GW9662 or H₂O₂. Then, we subjected these TCMEMs to invasion assay. Remarkably, the enhancement of TAM-induced tumor invasion by GW9662 or H₂O₂ was totally abrogated by the loss of TNF-α (Fig. 4E). Hence, these data prove that PPARγ can downregulate proinvasion capacity of TAMs by attenuating their TNF-α secretion.

FIG. 5.

Peroxisome proliferator-activated receptor γ (PPARγ) functioned against proinvasiveness signature of malignant melanoma TAMs by downregulating TNF-α production. (A) Normal TCMEMs and TCMEMs with small interference RNA (siRNA)-induced PPARγ silencing were subjected to invasion assay. Both qualitative (microscopic images) and quantitative (OD. values) results are shown herein. (B) The concentration of TNF-α in the culture medium of normal TCMEMs or TCMEMs with PPARγ depletion was examined by the ELISA test. (C) RAW264.7 macrophages were stimulated with TCMs for 48 h with or without a PPARγ agonist GW1929 (10 μM) or a PPARγ antagonist GW9662 (10 μM). The cells were then subjected to invasion assay and quantitative results are exhibited. (D) RAW264.7 macrophages were incubated with melanoma TCM for 48 h with or without GW1929 (10 μM) or GW9662 (10 μM). Then, TNF-α levels in the culture medium were detected with ELISA assay. (E) Peritoneal macrophages isolated from both wild-type and TNF-α^−/− mice were incubated with melanoma TCMs for 48 h with or without H₂O₂ (100 μM) and GW9662 (10 μM). Then, these cells were subjected to invasion assay for the assessment of their proinvasion capacity. Quantitative results as OD. values are shown herein. *p<0.05, **p<0.01, ***p<0.001.

PPARγ translocates from the nucleus to the cytoplasm in TAMs of malignant melanoma

As TNF-α secretion was considerably high in melanoma TAMs, we hypothesized its expression level was downregulated in these cells. However, our data showed that there was no apparent decrease of PPARγ expression in RAW264.7 TCMEMs or even TCMEMs under higher oxidative stress (Supplementary Fig. S6A). Subcellular localization alteration has been implicated to be equally important for the regulation of PPARγ functions. Thus, we turned to examine the subcellular distribution of PPARγ in melanoma TAMs with immunofluorescent staining. Compared with peritoneal macrophages, there was an obvious decrease of PPARγ in the nucleus, while the PPARγ level dramatically rises in the cytoplasm (Fig. 6A). These data indicated a translocation of PPARγ during the activation of macrophages within the tumor microenvironment. We also confirmed this phenomenon in TCMEMs utilizing both Western blot assays and immunofluorescent staining. As seen in Figure 6B and Supplementary Figure S6B, nucleus-to-cytoplasm translocation occurred in both stimulated RAW264.7 and ANA-1 macrophages. Moreover, when examined every 24 h in RAW264.7 cells, a gradual subcellular distribution change of PPARγ was observed (Fig. 6C). To investigate whether the abundant intracellular ROS was linked with PPARγ translocation, we analyzed PPARγ subcellular distribution in TCMEMs under various oxidative states. As shown in Fig. 6D, the presence of antioxidant NAC or Ve could effectively block this translocation. On the contrary, the addition of H₂O₂ was capable to enhance the nucleus-to-cytoplasm shuttle of PPARγ. Taken together, these data indicate a vital role of intracellular oxidative stress in downregulating PPARγ function by promoting its translocation from the nucleus to the cytoplasm in melanoma TAMs.

FIG. 6.

PPARγ translocates from the nucleus to the cytoplasm in TAMs of malignant melanoma. (A) The subcellular distribution of PPARγ in peritoneal macrophages or both kinds of TAMs was examined by immunofluorescence (IF). (B) Both nuclear and cytoplasmic expression of PPARγ in RAW264.7 TCMEMs were examined by Western blot. The quantitative data were generated using ImageJ software based on three independent experiments. (C) RAW 264.7 macrophages were incubated with TCMs for 24 h or 48 h. Then, the subcellular localization of PPARγ in these cells was analyzed by IF. (D) To investigate the role of intracellular ROS in PPARγ translocation in RAW264.7 TCMEMs, RAW 264.7 macrophages were stimulated with TCM for 48 h with or without NAC (10 mM), Ve (10 mM), or H₂O₂ (100 μM). Next, both nuclear and cytoplasmic expression of PPARγ were examined by Western blot. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Activated MEK-1 is responsible for PPARγ translocation in TAMs of malignant melanoma

To elucidate the mechanism underlying PPARγ translocation in malignant melanoma TAMs, we specifically analyzed MEK-1 that belongs to the MAP kinase family and has been shown to facilitate PPARγ nucleus-to-cytoplasm localization shift after its phosphorylation by directly binding to PPARγ (9). We began with the measurement of MEK-1 subcellular distribution in RAW264.7 TCMEMs. Similar to PPARγ, MEK-1 also showed an apparent translocation from the nucleus to the cytoplasm (Fig. 7A). Next, we examined the localization alteration over time of both PPARγ and MEK-1 to interrogate the linkage between these two proteins. The result displays two very similar changing patterns of the subcellular distribution of these proteins (Fig. S7A). We further examined their distributions in TCMEMs under different oxidative stresses. Again, two identical altering patterns were observed (Fig. 7B), indicating a close correlation between PPARγ and MEK-1 translocations in melanoma TAMs. Then, we continued this study with the MEK-1-specific inhibitor PD98059 to test whether the loss of MEK-1 activation would affect PPARγ translocation. This inhibitor exhibited great effectiveness to suppress MEK-1 phosphorylation without disturbing PPARγ expression compared with the control level (Supplementary Fig. S7B, C). As we expected, the nucleus-to-cytoplasm shifting of PPARγ in RAW264.7 TCMEMs was significantly blocked by the addition of PD98059 during TCM stimulation. Nevertheless, proinvasion capacity of TCMEMs whose 48-h education process was covered with PD98059 was just partially dampened compared with normal TCMEM counterparts (Fig. 7D), suggesting in melanoma TAMs there are other regulating mechanisms that suppress anti-inflammation function of PPARγ besides subcellular translocation. The phosphorylation of PPARγ on several sites like S112 by several MAP kinases (extracellular regulated protein kinases [ERK], p38, c-Jun N-terminal kinases …) can also lead to dysfunction of this nuclear receptor (10). Therefore, we examined the phosphorylation levels of PPARγ (S112) as well as ERK1/2(T202/T204), a downstreaming target of MEK-1, in the RAW264.7 TCMEM nucleus. As shown in Supplementary Figure S7D, the ratio of the phosphorylated protein level to the total protein level significantly increased compared with the control ratio, implicating that the phosphorylation of PPARγ could play a role in inhibiting PPARγ function in melanoma TAMs.

FIG. 7.

Activated MEK-1 is responsible for PPARγ translocation in TAMs of malignant melanoma. (A) Both nuclear and cytoplasmic expression of MEK-1 in RAW264.7 TCMEMs were examined by Western blot. (B) To investigate the subcellular distribution of both PPARγ and MEK-1 in TCMEMs under various oxidative states, RAW264.7 macrophages were incubated with TCMs with or without NAC (10 mM), Ve (10 mM), or H₂O₂ (100 μM) for 48 h. Thereafter, the cells were lysed and nucleus proteins and cytosolic proteins were extracted for Western blot. Again, relative integrated optical density (IDO) values shown in the figure were generated by ImageJ. (C) RAW264.7 macrophages were incubated with melanoma TCM for 48 h with or without 50 μM PD98059, and then PPARγ expression in the nucleus and cytoplasm was analyzed using Western blot. (D) RAW264.7 macrophages were stimulated with melanoma TCM for 48 h with or without 50 μM PD98059, 100 μM H₂O₂, or both 50 μM PD98059 and 100 μM H₂O₂. Next, the proinvasion capacity of these TCMEMs was assessed by invasion assay and quantitative results are shown in this figure. *p<0.05.

Discussion

Malignant melanoma is more deadly than many other types of cancers due to its high risk for metastasis, which results in a poor survival rate (the 5-year survival rate is under 10%) (5). Moreover, without proper treatment, melanoma cells can spread very rapidly to other organs, resulting in formidable difficulty in melanoma therapies (17, 37). Hence, thoroughly elucidating the mechanisms underlying the complicated metastatic process of malignant melanoma seems quite urgent. In the tumor microenvironment, residing stromal cells bidirectionally interact with tumor cells, generating a permissive/conducive environment that facilitates tumor cells to invade into blood vessels and lymphatic vessels. In this study, we focused on the interplay between TAMs and melanoma cells. Although there have been quite a number of studies depicting various aspects of melanoma TAMs based on plasticity, phenotype, and functions (24), few of them have been aware of the possible role of prevailing oxidative stress in the modulation of these cells, which could also drastically affect the malignancy of melanoma cells. Therefore, our study filled the gap between TAMs and the high level of ROS in the field of melanoma research.

ROS was once recognized as a strong weapon of the immune system to kill tumor cells (23). However, when malignant melanoma cells successfully escape the apoptosis triggered by abundant ROS (30), persistent ROS existence tends to favor melanoma survival, proliferation, and metastasis via activating or enhancing several related pathways (54, 25). We presumed the intracellular oxidative stress of melanoma stromal cells, such as macrophages, could also be affected. Hence, we not only verified the oxidative stress of B16F1 and B16F10 melanoma tissues, but also measured the ROS level of TAMs and TCMEMs, generating positive results (Fig. 1D, E). Consistently, our cDNA microarray analysis also confirmed a great enhancement of genes relevant to the antioxidant system in B16F10 TAMs, which can be explained as the response of macrophage cells intending to quench high oxidative stress (Fig. 2A). Moreover, the result of DHE staining (Fig. 1D) showed a great increase of intracellular superoxide anion in TAMs. Since superoxide anion has a high reactivity and considerably poor membrane permeability (55), this result provides the information that at least part of the intracellular ROS was endogenously generated. This conjecture was further verified by the elevation of intracellular oxidative stress of TCMEMs in the absence of ROS-excreting melanoma cells. The data from microarray and real-time polymerase chain reaction (PCR) verification rule out the possibility of deliberate cellular ROS production by the NOX2 complex or even other NOX enzymes (Table. 1, data not shown). Nevertheless, there is a dramatic increase of the gene level of PGC1-α, which has been reported as a master regulator of inductive mitochondria biogenesis, as well as an important mitochondrial transcriptional factor TFAM, implicating that ROS production from increased number of mitochondria may contribute to intracellular oxidative stress. This assumption is further validated by the gene elevation of NDUFV2 (Fig. 2C), the ROS-generating position of complex one, as well as the considerable increase of metabolic signaling pathways (Supplementary Table. S2). In addition, the mitochondrial ROS generation is also likely to be enhanced by various stresses prevailing in the melanoma microenvironment, such as hypoxia and pro-oxidants. However, this hypothesis still needs our future investigation.

The M1/M2 activation model of macrophages has been widely recognized during the last decade. However, the increasing number of recent studies suggested that this model probably does not entirely describe the complexity and heterogeneity of TAMs. In our study, we found that several genes of both M1 and M2 markers were upregulated in melanoma TAMs. Furthermore, the activities of Arg-1 and iNOS, the expression of CD206 and the secretion of TNF-α were also elevated together in melanoma TCMEMs. These data and the observations of other groups (61, 18a) demonstrated that melanoma TAMs possess a heterogeneous phenotype that does not belong to either M1 or M2 polarized macrophages. Hence, functional studies of TAMs seem to have a larger significance than simply interrogating their plasticity. Our analysis of signaling pathways based on microarray data proposes a proaggressiveness signature of macrophages (Supplementary Table. S2), which is verified by the raise of gene expressions of MMP2, MMP9, MMP12, osteopontin, cathepsin L, and urokinase (Table. 1). This proaggressiveness property was confirmed by both the in vitro invasion test and wound-healing assay.

The experiments conducted by us proposed a vital role of one particular inflammatory cytokine, TNF-α, in TAM-induced tumor invasiveness (Fig. 3). TNF-α was used to be discussed as a potent role tumor killer substantially produced by M1 macrophages to help the immune system inhibit tumor growth and progression (8, 53). Nevertheless, reports in recent years have shown that TNF-α also plays a critical role in tumorigenesis (49), tumor progression (4), and especially tumor invasiveness (66). Existing evidence proposed a relatively upstream position of TNF-α initiating the formation of a permissive/conducive microenvironment for tumor invasion. Kim and his colleagues found TNF-α induce expression of uPA and β-catenin activation in human breast epithelial cells (28). Hagemann et al. proved that coculturing TAMs with tumor cells can enhance the expression of MMPs, especially MMP2 and MMP9, in a TNF-α-dependent manner (21). Wu et al. concluded that TNF-α induces protein stabilization of Snail and b-catenin by inhibiting GSK-3b-mediated phosphorylation through NF-κB and Akt signaling pathways, which contribute to epithelial-mesenchymal transition (EMT) induction and invasion in tumor cells (65). Our experiments underscore the importance of the upstream position of TNF-α by exhibiting nearly complete loss of proinvasion capacity of TCMEMs derived from peritoneal macrophages of TNF-α^−/− mice. Moreover, we found the lack of TNF-α can abolish melanoma metastasis to the popliteal lymph node, even though tumor metastasis is a much more complex process involving multiple steps.

The anti-inflammatory effect of the nuclear receptor PPAR-γ has been observed in several human and murine monocyte/macrophage models (45). Our data further corroborated the universality of this effect of PPARγ in TAMs, showing a dramatic increase of TNF-α production after PPARγ silencing or deactivation (Fig. 5B, D). The inflammation-suppressing activity of PPARγ was proved to strongly depend on the regulations of its activation, such as modulation by natural binding ligands. Recently, there has been an increased awareness of the importance of PPARγ modulation by oxidative stress. It has been reviewed that ROS could elicit several cross talks between PPARγ and NRF2, FOXO as well as Wnt/β-catenin signaling pathways, resulting in expression and activation changes of PPARγ itself (40). Our study shed some light on ROS-mediated PPARγ regulation by demonstrating oxidative stress enhanced PPARγ nucleus export in melanoma TAMs. Subcellular compartmentalization of PPARγ has been assumed as a major mechanism modulating protein functions in various kinds of cells. However, the real effects of the alteration of PPARγ subcellular distribution in macrophages still need further investigation. Several distinct intranuclear mechanisms have been reported to contribute to the suppression of inflammatory genes by PPARγ, including transrepressing NF-κB (47), inhibiting the clearance of NCoR corepressor complexes (39) and upregulating IL-1 receptor antagonist (36). Although one study implicated that cytoplasm PPAR-γ may desensitize activated macrophages by attenuating cytosol-to-membrane translocation of protein kinase Cα (PKCα) via direct binding (59), the preponderance of evidence has revealed that nuclear localization is crucial for PPAR-γ interacting with other nuclear factors and transcriptional elements, thus exerting its anti-inflammatory function in macrophages. Hence, we suggest that the nucleus-to-cytoplasm translocation downregulates the inflammatory repression conducted by PPAR-γ in macrophages. This conjecture is strongly supported by our data as howsoever the intracelluar oxidative stress of TCMEMs changed, TNF-α secretion was always negatively correlated with the nucleus expression level of PPAR-γ (Fig. 4A, C; Fig. 7B). One model proposed by Burgermeister et al. to elucidate the mechanisms underlying nucleus export of PPAR-γ focuses on its direct binding with activated MEK-1 (11). Our observations not only corroborated this concept, but also granted the ROS-enhanced MEK-1/PPAR-γ translocation with a pathological significance for the first time.

The almost complete abolishment of PPAR-γ translocation by the MEK-1 inhibitor results in partly repressed proinvasion capacity of TCMEMs, indicating the existence of other molecular events that downregulate the PPAR-γ activity. PPAR-γ phosphorylation has been reported as another process that affects its functions (12). Among various phosphorylation sites, the phosphorylation of S112 (PPARγ2) was widely believed as a potent way to deactivate this nuclear receptor. Our data indicated this S112 phosphorylation may be also included in PPARγ modulation mediated by intracellular ROS in melanoma TAMs. Interestingly, a recent study conducted by Knethen et al. showed that CK-II mediated S16 and S21 phosphorylation of PPARγ could be critical in the shuttling of PPARγ from the nucleus to the cytosol, linking PPARγ phosphorylating status with its subcellular distribution (60). Hence, the phosphorylation of different sites on PPARγ could have various regulating effects. The studies to find and characterize new ROS-regulated phosphorylation sites and the linkage between different regulating mechanisms deserve further attention.

As noted earlier, we demonstrated a remarkable regulation mediated by intracellular ROS in melanoma TAMs. The intracellular oxidative stress promote nucleus-to-cytoplasm translocation of PPAR-γ mediated by MEK-1, resulting in the increase of TNF-α secretion, thus enhance the invasiveness and metastasis of melanoma cells. These molecular mechanisms mediated by ROS are well illustrated in Fig. 8. During the last decade, increasing attention of therapists has been attracted in the development of new therapies against melanoma using ROS as a therapeutic target. Due to the dual role of ROS, both pro-oxidant- and antioxidant-based anticancer approaches have been employed. Although some reports implicated ROS scavenging showed some effects in preventing melanoma growth, much more efforts have been spent on chemotherapy and radiotherapy that elevate ROS generation or inhibit the antioxidant system to directly kill melanoma cells (20, 18). However, nearly all of these studies neglect the critical role of fine-tuned complex interactions between stromal cells and tumor cells. Our study pointed out that inflammatory cytokine TNF-α secreted by TAMs, which is enhanced by ROS can be critical in melanoma lymphatic metastasis (Fig. 3F). Accordingly, how to specifically reduce the oxidative stress in TAMs or how to boost the melanoma ROS level without evoking inflammatory responses in TAMs should be considered in the future modification of melanoma therapies.

FIG. 8.

Proposed mechanistic explanation for ROS enhancing the proinvasion property of TAMs in malignant melanoma. Straight and curved rigid arrows in this figure represent the enhancing effects, which have been confirmed by our study or other reports. Straight dash arrows means the effects need our further investigation. Polyline arrows stand for the translocation and phosphorylation of proteins.

Materials and Methods

Reagents and antibodies

RPMI 1640 and the Dulbecco's modified Eagle's medium (DMEM) were purchased from Gibco. All of the chemicals were purchased from Sigma, unless otherwise stated. Anti-mouse PE-CD206 (M1) was purchased from BD Pharmingen. Anti-mouse PE-F4/80 was purchased from eBiosience. Anti-mouse PPARγ (81B8), anti-mouse MEK-1 (30C8), anti-mouse p-MEK-1 (S298), anti-mouse ERK (137F5), anti-mouse p-ERK1/2 (T202/T204), and mentioned loading controls were purchased from Cell Signalling. p-PPARg (S112) was provided by Abcam.

Mice and treatment

Male C57BL/6 mice were purchased from the Model Animal Research Center of the Nanjing University, Nanjing, China and bred in our animal facilities under specific pathogen-free conditions. After a 5-day quarantine, the mice were randomized into groups. Melanoma tumors were induced by subcutaneous injection of 10⁵ B16F1 or B16F10 murine melanoma cells. The control groups were treated with the solvent only.

Cell culture and culture supernatant treatment

B16F1 cells, B16F10 cells (Nanjing Keygen Biotech. Co., Ltd.), RAW264.7 cells (China Center for Type Culture Collection), and ANA-1 cells (Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences), were cultured in the DMEM and RPMI 1640 medium, respectively, supplemented with 10% fetal bovine serum (Gibco), 2 mM l-glutamine, 100 units/ml penicillin, and 100 units/ml streptomycin. The cells were kept in a humidified atmosphere of 5% CO₂ at 37°C. The culture supernatant of the B16F1 cells and the B16F10 cells was collected after 24 h. Then, the culture supernatants of the RAW264.7 and the ANA-1 cells were replaced by the culture supernatant of the B16F1 cells or the B16F10 cells.

Histology and immunofluorescence assays

Formalin-fixed paraffin-embedded samples were sectioned and stained with H&E for tumor histology analysis. The sections were blocked for 1 h with 3% bovine serum albumin (BSA) and 0.05% Tween 20 (diluted in 50 mM Tris-HCl pH 7.6 containing 150 mM NaCl). The sections were then probed with antibodies against the anti-mouse F4/80 antigen (eBioscience) followed by anti-rat IgG fluorescein isothiocyanate.

The frozen sections of TAMs were stained with DHE purchased from Beyotime. Then, they were observed at the appropriate fluorescence wavelength using a confocal microscope.

The cells were spread on a coverslip treated with polylysine and were fixed with 4% paraformaldehyde for 10 min. The cells were perforated using 0.5% Triton X-100 for 15 min. After being blocked in phosphate-buffered saline (PBS) containing 0.1% Tween-20 with 5% BSA at room temperature for 30 min, the cells were incubated with a primary antibody overnight at 4°C. The cells were then incubated with a secondary antibody for 1 h at room temperature. The cells were then observed at the appropriate fluorescence wavelength using a confocal microscope.

Flow cytometry analysis

The DCFH-DA was purchased from Beyotime (Haimen). The saponin (0.05%) (Sigma-Aldrich) was diluted in PBS and used as a membrane permeator for inner membrane staining. The cells were analyzed on a fluorescence-activated cell sorting (FACS) Calibur cytometer using Cellquest software (Becton Dickinson). The statistics presented are based on 10,000 events gated on the population of interest.

Cytokine assays by specific ELISA

The TNF-α concentration in 150 μl supernatants of 10,000 TAMs or 2 ml supernatants of 1 million TCMEMs was assessed using a standard sandwich ELISA according to the instruction manual. The TNF-α concentrations were expressed in nanograms per milliliter, as calculated from the calibration curves from serial dilutions of murine recombinant standards (eBioscience) in each assay. The sensitivity of the TNF-α assay was 20 pg/ml.

Western blotting

The cell protein was extracted by whole cell lysis or a nuclear protein extract kit purchased from Beytotime, which contained protease and phosphatase inhibitors. The cell debris was removed by centrifugation at 4°C, and the supernatants were collected and stored at −70°C until use. The protein amounts were determined using the bicinchoninic acid protein assay (Pierce), and a dot blot analysis was performed as described previously (11).

Cell invasion assay

The extensively utilized invasion assay in this study is well illustrated in Figure 9 and described in figure legend.

FIG. 9.

Examine the proinvasion capacity of macrophages via invasion assay. In a 24-well plate, 100,000 RAW264.7/ANA-1/peritoneal macrophages were educated with melanoma TCM for 48 h in the presence or absence of NAC/Ve/H₂O₂ to change the intracellular ROS level, with or without GW1929/GW9662 to alter the PPARγ activity or with or without PD98059 to inhibit the MEK-1 activity. Ten thousand B16F1 or B16F10 tumor cells were transferred into a Transwell chamber (8.0-μM pore size) (Costar) coated with Matrigel. Then, tumor cells and RAW264.7/ANA-1/peritoneal macrophages TCMEMs or 100,000 TAMs isolated from subcutaneous melanoma were cocultured in the same medium with 10% fetal bovine serum in the presence or absence of the TNF-α antibody. After 14 h, cells that remained on the top side of the membrane were removed carefully with a cotton tip and the invading cells adhered on the other side were stained with a 0.1% crystal violet solution. After a phosphate-buffered saline rinse for three times, microscopic images were taken as qualitative results. Or the crystal violet on these cells was resolubilized into 10% acetic acid for absorption spectroscopic analysis at 600 nm to generate quantitative data.

Wound-healing assay

B16F1 or B16F10 cells were cultured in 96-well plates at 40,000/well as confluent monolayers. The monolayers were incubated in the absence of serum for 16 h and wounded in a line across the well with a 10-μl standard pipette tip. The wounded monolayers were then washed twice with serum-free media to remove cell debris and cocultured with 10,000 macrophages in Transwell chambers (0.4-μm pore size) (Costar) in the medium with 10% fetal calf serum (FCS). The microscopic images of the same location of a cell-free wound were recorded at 0 h and 24 h.

Microarray analysis

Total RNA of B16F10 TAMs was prepared from three independent experiments. Microarray experiments were performed by Gene Tech (Shanghai, China) with 20 μg of total RNA, which was labeled with fluorescent nucleotides and hybridized to murine cDNA slides. Hybridized slides were interrogated via an Agilent scanner, data were subjected to background subtraction and normalization, and analyzed with statistical analysis of microarrays software package.

Quantitative real-time PCR verification

The RT reaction was performed on 150 ng of total RNA with avian myeloblastosis virus reverse transcriptase (Promega). Quantitative real-time PCR was performed using QuantiTect SYBR Green I (Qiagen). PCRs were performed in a total volume of 20 μl (1×QuantiTect SYBR Green Master Mix) in the ABI Prism 7000 system (Applied Biosystems). The absolute number of copies of the gene of interest in the experimental cDNA samples was calculated from the linear regression of a standard curve. The expression of the measured genes in each sample was normalized for actin expression.

The details of primers are listed as follows:

arg1:+5′ GAATCTGCATGGGCAACCT 3′

− 5′ CAGGGTCTACGTCTCGCAAG 3′

gsta4:+5′ GATTGCCGTGGCTCCATTTA 3′

− 5′ CAACGAGAAAAGCCTCTCCGT 3′

il1r2:+5′ AAGGATGTGGGTGAAGGGTAAC 3′

− 5′ CTCACAGTGGGATGCGTTTC 3′

nos2:+5′ TGGAGCGAGTTGTGGATTGTC 3′

− 5′ GCCTCTTGTCTTTGACCCAGTAG 3′

mmp9:+5′ CGAAGCGGACATTGTCATCC 3′

− 5′ GTCGTCGAAATGGGCATCTC 3′

ndufv2:+5′ ACAACAGGACCACTATCTCGGTC 3′

− 5′ GGCTCTATGAATGGGCAACAC 3′

cybb:+5′ TGAATGCCAGAGTCGGGATTT 3′

− 5′ CGAGTCACGGCCACATACA 3′

ppargc1a:+5′ CAACAATGAGCCTGCGAACA 3′

− 5′ CATCAAATGAGGGCAATCCG 3′

prdx1:+5′ TGTCCCACGGAGATCATTGC 3′

− 5′ CACAGAAGCGCCAATCACTT 3′

tnf:+5′ CTGTGAAGGGAATGGGTGTTC 3′

− 5′ CAGGTCACTGTCCCAGCATC 3′

siRNA transfection

The RAW264.7 cells (10⁵ cells in 300 μl) were mixed with 100 μl of the electroporation buffer in a cuvette, which contained siRNA oligonucleotides (Stealth siRNA duplex oligoribonucleotides, Invitrogen). The cuvette was immediately transferred to the electroporator and the device was turned on. After 1–2 min, the cuvette was placed on ice, and then cells were plated in six-well plates. The treatments were performed 24 h after the siRNA transfection. The sequences of the PPARγ siRNA1 are 5′-ACAGGCCUCAUGAAGAAUUTT-3′ and 5′-AAUUCUUCAUGAGGCCUGUUG-3′. The sequences of the PPARγ siRNA2 are 5′-UGGAAGACCACUCCCACUCTT-3′ and 5′-GAGUGGGAGUGGUCUUCCATT-3′. For a control, we used a scrambled siRNA, whose sequences are 5′-UUCUCCGAACGUGUCACGUTT-3′ and 5′-ACGUGACACGUUCGGAGAATT-3′.

Primary TAM flow sorting

A mouse subcutaneous melanoma model was established for 3 weeks. The tumors were minced in the DMEM with 0.1% collagenase I (1 h, 37°C), passed through a 19G needle, and filtered using a 40-μm cell strainer (BD Falcon). The cells were centrifuged through density gradient centrifugation and washed with PBS. The cells were then incubated with 10% FCS and an anti-F4/80 antibody for 30 min. The tumor cells were sorted as F4/80⁻cells. The TAMs were sorted as F4/80⁺ cells.

Measurement of the intracellular GSSG/GSH

The intracellular GSH content in the melanoma tissue was determined using a GSH and GSSG assay kit purchased from Beyotime (Haimen) according to the manufacturer's instructions.

Statistical analysis

The data are expressed as the mean±standard deviation. The statistical analysis was performed by the Student's t-test when only two value sets were compared. A one-way ANOVA followed by a Dunnett's test were used when the data involved three or more groups. p<0.05, p<0.01, or p<0.001 were considered statistically significant and indicated by *, **, or ***, respectively.

Author Contribution

The project and researchers were funded and supervised by Prof. Pingping Shen. Xuzhu Lin was chiefly responsible for experiment designing and manuscript drafting. Wei Zheng and Jing Lin generated the majority of the data in this manuscript. The rest of the work mentioned in this study was contributed by other authors.

Footnotes

Acknowledgments

We thank Adebusola Alagbala Ajibade (Baylor College of Medicine, U.S.A.) and Jun Cui (the Nanjing University, China) for critical comments and grammatical correction of the manuscript. We also thank Dr. Ning Su (professor, School of Basic Medical Sciences, Southeast University, China) for excellent technical assistance.

This work was supported by the National Natural Science Foundation of China (project No. 91013015, 81273527, 31070764), and the Science Fund for Creative Research Groups (No. 81121062).

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

This reference has been deleted.

Aoki

, Maeno

, Aoyagi

et al. Pioglitazone, a peroxisome proliferator-activated receptor gamma ligand, suppresses bleomycin-induced acute lung injury and fibrosis. Respiration, 77:311–319. 2008.

Asada

, Sasaki

, Suda

, Chida

, Nakamura

. Antiinflammatory roles of peroxisome proliferator-activated receptor {gamma} in human alveolar macrophages. Am J Respir Crit Care Med, 169:195–200. 2004.

Balkwill

. TNF-alpha in promotion and progression of cancer. Cancer Metastasis Rev, 25:409–416. 2006.

Barth

, Wanek

, Morton

. Prognostic factors in 1,521 melanoma patients with distant metastases. J Am Coll Surg, 181:193–201. 1995.

Bedard

, Krause

. The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology. Physiol Rev, 87:245–313. 2007.

Biswas

, Mantovani

. Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nat Immunol, 11:889–896. 2010.

Bonavida

, Ng

, Jazirehi

, Schiller

, Mizutani

. Selectivity of TRAIL-mediated apoptosis of cancer cells and synergy with drugs: the trail to non-toxic cancer therapeutics (review) Int J Oncol, 15:793–802. 1999.

Burgermeister

, Chuderland

, Hanoch

, Meyer

, Liscovitch

, Seger

. Interaction with MEK causes nuclear export and downregulation of peroxisome proliferator-activated receptor {gamma} Mol Cell Biol, 27:803–817. 2007.

10.

Burgermeister

, Seger

. PPARgamma and MEK interactions in cancer. PPAR Res, 2008:309–469. 2008.

11.

Burgermeister

, Seger

. MAPK kinases as nucleo-cytoplasmic shuttles for PPARgamma. Cell Cycle (Georgetown, Tex), 6:1539–1548. 2007.

12.

Burns

, Vanden Heuvel

. Modulation of PPAR activity via phosphorylation. Biochim Biophys Acta, 1771:952–960. 2007.

13.

Corzo

, Condamine

, Lu

et al. HIF-1α regulates function and differentiation of myeloid-derived suppressor cells in the tumor microenvironment. J Exp Med, 207:2439–2453. 2010.

14.

Corzo

, Cotter

, Cheng

et al. Mechanism regulating reactive oxygen species in tumor-induced myeloid-derived suppressor cells. J Immunol, 182:5693–5701. 2009.

15.

Cruz

, Rinna

, Forman

, Ventura

ALM

, Persechini

, Ojcius

. ATP activates a reactive oxygen species-dependent oxidative stress response and secretion of proinflammatory cytokines in macrophages. J Biol Chem, 282:2871–2879. 2007.

16.

Cummins

, Cummins

, Pantle

, Silverman

, Leonard

, Chanmugam

. Cutaneous malignant melanoma. Mayo Clinic, 81:500–507. 2006.

17.

Fidler

. Melanoma metastasis. Cancer Control, 2:398–404. 1995.

18.

Fruehauf

, Trapp

. Reactive oxygen species: an Achilles heel of melanoma? Expert Rev Anticancer Ther, 8:1751–1757. 2008.

19.

Gal

, Tapmeier

, Muschel

. Plasticity of tumor associated macrophages in a metastatic melanoma model in the mouse [abstract] Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31–Apr 4, Chicago, IL. Philadelphia (PA): AACR. Cancer Res, 2012; 72,8 SupplAbstract nr 402DOI:1538-7445.AM2012-402.

20.

Goswami

, Sahai

, Wyckoff

et al. Macrophages promote the invasion of breast carcinoma cells via a colony-stimulating factor-1/epidermal growth factor paracrine loop. Cancer Res, 65:5278–5283. 2005.

21.

Gupta

, Hevia

, Patchva

, Park

, Koh

, Aggarwal

. Upsides and downsides of reactive oxygen species for cancer: the roles of reactive oxygen species in tumorigenesis, prevention, and therapy. Antioxid Redox Signal, 16:1295–1322. 2012.

22.

Hagemann

, Robinson

, Schulz

, Trumper

, Balkwill

, Binder

. Enhanced invasiveness of breast cancer cell lines upon co-cultivation with macrophages is due to TNF-alpha dependent up-regulation of matrix metalloproteases. Carcinogenesis, 25:1543–1549. 2004.

23.

Hildenbrand

, Dilger

, Hörlin

, Stutte

. Urokinase and macrophages in tumour angiogenesis. Br J Cancer, 72:818–823. 1995.

24.

, Brindle

. Immune cell-induced synthesis of NO and reactive oxygen species in lymphoma cells causes their death by apoptosis. FEBS Lett, 579:2833–2841. 2005.

25.

Hussein

. Tumour-associated macrophages and melanoma tumourigenesis: integrating the complexity. Int J Exp Pathol, 87:163–176. 2006.

26.

Ibañez

, Notcovich

, Policastro

, Durán

. Reactive oxygen species in the biology of melanoma. Breakthroughs in Melanoma Research. Rijeka, Crotia: InTech Europe, 1–30. 2011.

27.

Jensen

, Schmidt

, Møller

, Høyer

, Maniecki

, Sjoegren

, Christensen

, Steiniche

. Macrophage markers in serum and tumor have prognostic impact in American Joint Committee on Cancer stage I/II melanoma. J Clin Oncol, 27:3330–3337. 2009.

28.

Joosse

, De Vries

, Van Eijck

, Eggermont

AMM

, Nijsten

, Coebergh

JWW

. Reactive oxygen species and melanoma: an explanation for gender differences in survival? Pigment Cell Melanoma Res, 23:352–364. 2010.

29.

Kim

, Kim

, Na

, Surh

. TNF-α induces expression of urokinase-type plasminogen activator and β-catenin activation through generation of ROS in human breast epithelial cells. Biochem Pharmacol, 80:2092–2100. 2010.

30.

Kushnareva

, Murphy

, Andreyev

. Complex I-mediated reactive oxygen species generation: modulation by cytochrome c and NAD (P)+ oxidation-reduction state. Biochem J, 368:545–553. 2002.

31.

Leiter

, Schmid

, Kaskel

, Peter

, Krahn

. Antiapoptotic bcl-2 and bcl-xL in advanced malignant melanoma. Arch Dermatol Res, 292:225–232. 2000.

32.

, Satyamoorthy

, Herlyn

. Dynamics of cell interactions and communications during melanoma development. Crit Rev Oral Biol Med, 13:62–70. 2002.

33.

Mantovani

, Sica

, Sozzani

, Allavena

, Vecchi

, Locati

. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol, 25:677–686. 2004.

34.

Marconi

, Bianchini

, Mannini

, Mugnai

, Ruggieri

, Calorini

. Tumoral and macrophage uPAR and MMP-9 contribute to the invasiveness of B16 murine melanoma cells. Clin Exp Metastasis, 25:225–231. 2008.

35.

Martin-Manso

, Galli

, Ridnour

, Tsokos

, Wink

, Roberts

. Thrombospondin 1 promotes tumor macrophage recruitment and enhances tumor cell cytotoxicity of differentiated U937 cells. Cancer Res, 68:7090–7099. 2008.

36.

This reference has been deleted.

37.

Meier

, Chicheportiche

, Juge-Aubry

, Dreyer

, Dayer

. Regulation of the interleukin-1 receptor antagonist in THP-1 cells by ligands of the peroxisome proliferator-activated receptor gamma. Cytokine, 18:320–328. 2002.

38.

Melnikova

, Bolshakov

, Walker

, Ananthaswamy

. Genomic alterations in spontaneous and carcinogen-induced murine melanoma cell lines. Oncogene, 23:2347–2356. 2004.

39.

Murphy

. How mitochondria produce reactive oxygen species. Biochemical J, 417:1–13. 2009.

40.

Pascual

, Fong

, Ogawa

, Gamliel

, Li

, Perissi

, Rose

, Willson

, Rosenfeld

, Glass

. A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-γ Nature, 437:759–763. 2005.

41.

Polvani

, Tarocchi

, Galli

. PPARγ and oxidative stress: con (β) catenating NRF2 and FOXO. PPAR Res, 2012; 2012–641087.

42.

Pollard

. Macrophages define the invasive microenvironment in breast cancer. J Leukoc Biol, 84:623–630. 2008.

43.

Pruett

, Cheng

, Fan

, Tan

, Sebastian

. Oxidative stress and sodium methyldithiocarbamate-induced modulation of the macrophage response to lipopolysaccharide in vivo. Toxicol Sci, 109:237–246. 2009.

44.

Qian

, Pollard

. Macrophage diversity enhances tumor progression and metastasis. Cell, 141:39–51. 2010.

45.

Qian

, Zhang

, Liu

. Tumor-secreted PGE2 inhibits CCL5 production in activated macrophages through cAMP/PKA signaling pathway. J Biol Chem, 286:2111–2120. 2010.

46.

Rigamonti

, Chinetti-Gbaguidi

, Staels

. Regulation of macrophage functions by PPAR-α, PPAR-γ, and LXRs in mice and men. Arterioscler Thromb Vasc Biol, 28:1050–1059. 2008.

47.

Rigel

, Russak

, Friedman

. The evolution of melanoma diagnosis: 25 years beyond the ABCDs. CA Cancer J Clin, 60:301–316. 2010.

48.

Ricote

, Glass

. PPARs and molecular mechanisms of transrepression. Biochim et Biophys Acta, 1771:926–935. 2007.

49.

Sander

, Hamm

, Elsner

, Thiele

. Oxidative stress in malignant melanoma and non-melanoma skin cancer. Br J Dermatol, 148:913–922. 2003.

50.

Sethi

, Sung

, Aggarwal

. TNF: a master switch for inflammation to cancer. Front Biosci, 13:5094–5107. 2008.

51.

Söderberg

, Barral

, S derstr m

, Sander

, Rosén

. Redox-signaling transmitted in trans to neighboring cells by melanoma-derived TNF-containing exosomes. Free Radic Biol Med, 43:90–99. 2007.

52.

Solinas

, Germano

, Mantovani

, Allavena

. Tumor-associated macrophages (TAM) as major players of the cancer-related inflammation. J Leukoc Biol, 86:1065–1073. 2009.

53.

Solinas

, Schiarea

, Liguori

et al. Tumor-conditioned macrophages secrete migration-stimulating factor: a new marker for M2-polarization, influencing tumor cell motility. J Immunol, 185:642–652. 2010.

54.

Song

, Kang

, Lee

, Jeung

, Yang

. Upregulation of tumor necrosis factor-[alpha] expression by trans10-cis12 conjugated linoleic acid enhances phagocytosis of RAW macrophages via a peroxisome proliferator-activated receptor [gamma]-dependent pathway. Cytokine, 37:227–235. 2007.

55.

Storz

. Reactive oxygen species in tumor progression. Front Biosci, 10:1881–1896. 2005.

56.

Toledano

, Fourquet

, D'Autréaux

. Cellular signaling by reactive oxygen species: biochemical basis and physiological scope. Signal Transduct Pathways Mech Dis. Berlin, Heldelberg: Springer, 2009; 313–336.

57.

Van Kempen

, van Muijen

, Ruiter

. Stromal responses in human primary melanoma of the skin. Front Biosci, 10:2922–2931. 2005.

58.

Varney

, Johansson

, Singh

. Tumour-associated macrophage infiltration, neovascularization and aggressiveness in malignant melanoma: role of monocyte chemotactic protein-1 and vascular endothelial growth factor-A. Melanoma Res, 15:417–425. 2005.

59.

Ventura-Clapier

, Garnier

, Veksler

. Transcriptional control of mitochondrial biogenesis: the central role of PGC-1α Cardiovasc Res, 79:208–217. 2008.

60.

Von Knethen

, Soller

, Tzieply

et al. PPARγ1 attenuates cytosol to membrane translocation of PKCα to desensitize monocytes/macrophages. J Cell Biol, 176:681–694. 2007.

61.

von Knethen

, Tzieply

, Jennewein

, Brüne

. Casein-kinase-II-dependent phosphorylation of PPARgamma provokes CRM1-mediated shuttling of PPARgamma from the nucleus to the cytosol. J Cell Sci, 123:192–201. 2010.

62.

Wang

, Ge

, Xiao

, Lopez-Coral

, Azuma

, Somasundaram

, Zhang

, Wei

, Xu

, Rauscher

. Melanoma-derived conditioned media efficiently induce the differentiation of monocytes to macrophages that display a highly invasive gene signature. Pigment Cell Melanoma Res, 25:493–505. 2012.

63.

Wheeler

, Smutney

, Samulski

. Secretion of extracellular superoxide dismutase from muscle transduced with recombinant adenovirus inhibits the growth of B16 melanomas in mice. Mol Cancer Res, 1:871–881. 2003.

64.

Wittgen

HGM

, van Kempen

LCLT

. Reactive oxygen species in melanoma and its therapeutic implications. Melanoma Res, 17:400–409. 2007.

65.

This reference has been deleted.

66.

, Deng

, Rychahou

, Qiu

, Evers

, Zhou

BPH

. Stabilization of snail by NF-kappa B is required for inflammation-induced cell migration and invasion. Cancer Cell, 15:416–428. 2009.

67.

, Zhou

. TNF-alpha/NF-kappa B/Snail pathway in cancer cell migration and invasion. Br J Cancer, 102:639–644. 2010.

68.

Youn

, Nagaraj

, Collazo

, Gabrilovich

. Subsets of myeloid-derived suppressor cells in tumor-bearing mice. J Immunol, 181:5791–5802. 2008.

69.

This reference has been deleted.

70.

Zhu

, Mulcahy

, Mohammed

et al. IL-17 expression by breast-cancer-associated macrophages: IL-17 promotes invasiveness of breast cancer cell lines. Breast Cancer Res, 10:R95. 2008.

71.

Zucker

, Cao

. Measurement of matrix metalloproteinases in serum of patients with melanoma: snarled in technical pitfalls. Clin Cancer Res, 11:5069–5070. 2005.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.67 MB