The TRPA1 Channel Amplifies the Oxidative Stress Signal in Melanoma

Macrophages (MΦs) and reactive oxygen species (ROS) are implicated in carcinogenesis. The oxidative stress sensor, transient receptor potential ankyrin 1 (TRPA1), activated by ROS, appears to contribute to lung and breast cancer progression. Although TRPA1 expression has been reported in melanoma cell lines, and oxidative stress has been associated with melanocytic transformation, their role in melanoma remains poorly known. Here, we localized MΦs, the final end-product of oxidative stress, 4-hydroxynonenal (4-HNE), and TRPA1 in tissue samples of human common dermal melanocytic nevi, dysplastic nevi, and thin (pT1) and thick (pT4) cutaneous melanomas. The number (amount) of intratumoral and peritumoral M2 MΦs and 4-HNE staining progressively increased with tumor severity, while TRPA1 expression was similar in all samples. Hydrogen peroxide (H2O2) evoked a TRPA1-dependent calcium response in two distinct melanoma cell lines (SK-MEL-28 and WM266-4). Furthermore, H2O2 induced a TRPA1-dependent H2O2 release that was prevented by the TRPA1 antagonist, A967079, or Trpa1 gene silencing (siRNA). ROS release from infiltrating M2 MΦs may target TRPA1-expressing melanoma cells to amplify the oxidative stress signal that affects tumor cell survival and proliferation.

Mitochondrial macromolecules derived from 4-HNE are involved in the initiation and progression of cancer [29]. Using monoclonal antibodies specific for 4-HNE protein adducts and biochemical assays, differences in 4-HNE levels between normal and tumoral tissues have been shown in kidney [30], brain [31], colorectal [32], and thyroid [33] cancer. Malignant melanocyte transformation has been associated with oxidative stress, as antioxidant defenses are compromised in pathological conditions [34]. Furthermore, elevated levels of 4-HNE have been reported in a series of five dysplastic nevi and 35 cases of cutaneous malignant melanomas, as compared to five cases of benign nevi [35].
Here, in a series of common dermal melanocytic nevi (dermal nevi), dysplastic nevi, and thin (pT1) and thick (pT4) malignant melanomas, we simultaneously explored the presence of M1 and M2 TAM subtypes, the level of the oxidative stress end-product 4-HNE, and the expression of the TRPA1 channel. In addition, we studied the expression of TRPA1 in two different melanoma cell lines and its ability to sense and amplify the oxidative stress signal. Results indicate that increased peritumoral and intratumoral M2 TAM are positively associated with 4-HNE levels, and that TRPA1 stimulated by ROS amplifies the oxidative stress signal in cultured melanoma cells.

Tissue Collection
The study cohort included formalin-fixed paraffin-embedded (FFPE) tissues from ten commonly acquired (dermal) melanocytic nevi, nine dysplastic nevi and 28 invasive primary cutaneous melanomas. FFPE samples were retrospectively retrieved from the archive of the Division of Pathology, Department of Health Sciences, University of Florence, Italy. Consecutive cases of nevi, pT1, and pT4 primary cutaneous melanoma, histopathologically diagnosed between 2002 and 2018 at the Section of Anatomic Pathology, Department of Health Sciences, University of Florence, with accessible medical records and sufficient FFPE tumor specimens, were included in the study. In addition, FFPE samples were divided according to TNM classification into two groups according to Breslow thickness (mm): pT1 ≤ 1 mm and pT4 > 4 mm; the specimens showing lentiginous proliferation of melanocytes at the dermoepidermal junction, extending at least three rete ridges beyond lateral margins of the dermal component, were considered dysplastic nevi.

Immunohistochemistry
Immunohistochemistry was performed on representative FFPE whole tumor sections, 3-µm thick, of dermal and dysplastic nevi and invasive primary cutaneous melanoma. For CD68 and CD163 staining, sample processing was performed with automated immunostainer ( . For TRPA1 staining, sample processing was performed manually as follows: antigen retrieval was routinely performed by immersing the slides in a thermostat bath containing 10 mM Citrate Buffer (pH 6.0) for 20 min at 99 • C followed by cooling for 20 min at room temperature. Endogenous peroxidase activity was blocked with a blocking solution (Bloxall, ready to use, Vector, Burlingame, CA, USA) for 10 min. Sections were incubated with anti-TRPA1 antibody (#ACC-037, rabbit polyclonal, 1:100, Alomone, Israel) for 1 h at room temperature. The signal was developed with Vectastain Universal ABC-AP KIT (#AK5200, Vector, Burlingame, CA, USA). Nuclei were counterstained with Mayer's hematoxylin.

Image Analysis
For each dermal and dysplastic nevi and invasive primary cutaneous melanoma, a representative histopathological whole slide, stained with hematoxylin, eosin, anti-CD68, anti-CD163, and anti-TRPA1, was digitally scanned at 400× magnification with Aperio AT2 (Leica Biosystems, Milan, Italy) into whole slide digital images (WSI). Each SVS format file was imported into Halo Link ® (Indica Labs, Albuquerque, NM, USA) image management system. Detection of immune-stained positive cells was performed using Halo ® Multiplex immunohistochemistry (IHC) analysis software version v3.1.1076.308, based on cytonuclear features such as stain, size, and roundness. For CD68+ve and CD163+ve cell counting exclusively in the lesion region, we combined the multiplex IHC module with DenseNet AI (plugin) classifier (Indica Labs, Albuquerque, NM, USA). This method allowed us to count immune-stained positive cells in intratumoral and peritumoral areas, excluding any CD68+ve and CD163+ve cells that were outside the lesion, automatically. TRPA1 analysis was performed using the Halo ® Area Quantification analysis software version 2.1.7 (Indica Labs, Albuquerque, NM, USA), based on the staining intensity and distribution of TRPA1 staining.  (2 mM), penicillin (100 U/mL), and streptomycin (100 mg/mL) under a humidified atmosphere of 5% CO 2 at 37 • C. All cells were used without further authentication.

qRT-PCR
Total RNA was extracted from SK-MEL-28 and WM 266-4 cells using the RNeasy Mini kit (Qiagen SpA, Hilden, Germany), according to the manufacturer's protocol. RNA concentration and purity were assessed spectrophotometrically by measuring the absorbance at 260 nm and 280 nm. RNA was reverse transcribed with the Qiagen QuantiTect Reverse Transcription Kit (Qiagen SpA, Hilden, Germany), following the manufacturer's protocol. For mRNA relative quantification, rt-PCR was performed on Rotor Gene ® Q (Qiagen SpA, Hilden, Germany). The relative abundance of mRNA transcripts was calculated using the delta CT method and normalized to GAPDH levels. The sets of primers were as follows: GAPDH-Seq1 (GAPDH-sequence1): 5 -ACATCGCTCAGACACCATG -3 , GAPDH-Seq2

siRNA Transfection
The TriFECTa ® RNAi kit (Integrated DNA Technologies, IDT, Coralville, IA, USA) was used to silence the Trpa1 gene in SK-MEL-28 and WM 266-4 cell lines. Transfections of siRNA into cells were performed according to each manufacturer's instructions. Briefly, one day before the transfection, SK-MEL-28 and WM 266-4 were plated to 6-well plates (1 × 10 6 cells/well) in 2.5 mL of growth medium without antibiotics until 70-80% of confluence. Cells were transfected as follows: siRNA and lipofectamine RNAiMAX (#13778-075, Life Technologies, Waltham, MA, USA) were diluted in Opti-MEM I Reduced Serum Medium (#31985-047, Gibco ® -Life Technologies, Carlsbad, CA, USA) separately before being mixed by pipetting. The final medium amount per well was 2.5 mL. The siRNA-RNAiMAX mix was added to 6-well cell culture plates. The final concentrations of lipofectamine RNAiMAX and siRNA were 3 µL/mL and 10 nM, respectively. Cells were then incubated for 24 h under a humidified atmosphere of 5% CO 2 at 37 • C. Knock-down efficiency was detected at 24 h by qRT-PCR comparative quantitation analysis.

Statistical Analysis
Results are expressed as mean ± standard error of the mean (SEM). For multiple comparisons, a one-way analysis of variance (ANOVA) followed by the post-hoc Bonferroni's test or Dunnett's test was used. Statistical analyses were performed on raw data using Graph Pad Prism 8 (GraphPad Software Inc., San Diego, CA, USA). EC 50 values were determined from non-linear regression models using Graph Pad Prism 8 (GraphPad Software Inc.). A Spearman's rank-order correlation was run to assess the relationship between 4-HNE and CD68+ve or CD163+ve cells, and data was visualized with simple scatterplots (IBM Corp. SPSS Statistics, version 26). p values less than 0.05 (p < 0.05) were considered significant. Statistical tests used and the sample size for each analysis are listed in the figure legends.

TAM Quantification in Dermal and Dysplastic Nevi, and Melanoma Tissues
We first investigated, by immunohistochemistry, the density of CD163+ve and CD68+ve TAMs, separately, in both the intratumoral and peritumoral compartment. We observed that CD163+ve TAMs were higher in thicker pT4 melanomas than those found in dermal nevi (Figure 1a-d). A significant increase of CD163+ve TAMs was also observed between deeply invasive (pT4) and thin (pT1) melanomas (Figure 1a-d). Although a tendency to higher numbers of CD68+ve TAMs was observed from dermal nevi to pT4 melanomas, no significant difference in the four categories of samples analyzed was found (Figure 1a-d).

4-HNE Levels in Dermal and Dysplastic Nevi, and Melanoma Tissues and the Correlation with TAMs
The presence of the oxidative stress byproduct, 4-HNE, was explored in dermal and dysplastic nevi and melanoma tissues. Immunofluorescence analysis revealed that 4-HNE expression was significantly higher in pT1 melanomas compared to dermal nevi and in pT4 melanomas compared to dermal and dysplastic nevi (Figure 2a). In addition, thicker (pT4) melanomas showed an increased 4-HNE staining compared to thinner (pT1) melanomas ( Figure 2a).
By running a Spearman's correlation, we assessed the relationship between 4-HNE and CD68+ve or CD163+ve cells in the four different types of samples. A positive, statistically significant correlation between 4-HNE and CD163+ve cells was found in samples from dermal and dysplastic nevi and pT1 and pT4 melanomas.   By running a Spearman's correlation, we assessed the relationship between 4-HNE and CD68+ve or CD163+ve cells in the four different types of samples. A positive, statistically significant correlation between 4-HNE and CD163+ve cells was found in samples from dermal and dysplastic nevi and pT1 and pT4 melanomas. In particular, there was a moderate positive correlation between 4-HNE and intratumoral CD163+ve cells (rs = 0.405, n = 47, p = 0.005) and a stronger positive correlation between 4-HNE and peritumoral CD163+ve cells (rs = 0.650, n = 45, p < 0.0001). No correlation was found between 4-HNE and either intratumoral CD68+ve cells (rs = −0.022, n = 45, p = 0.887) or peritumoral CD68+ve cells (rs = 0.073, n = 44, p = 0.636) (Figure 2b).

Expression and Function of TRPA1 in Dermal and Dysplastic Nevi, Melanoma Tissues, and Melanoma Cell Lines
The expression of TRPA1, a major chemosensory receptor for ROS [50], was evaluated in tissue samples of dermal and dysplastic nevi and pT1 and pT4 malignant melanomas. In skin adjacent to the lesion, TRPA1 was found in cutaneous nerves, which are used as staining positive control (Figure 3a). Normal melanocytes generally showed faint to moderate TRPA1 staining, although the immunolabelling was occasionally difficult to evaluate, due to the presence of a clear perinuclear halo, resulting from marked cytoplasmic retraction. In dermal and dysplastic nevi, as well as in melanoma cells, TRPA1 immunoreactivity was comparable and mostly confined to the cell cytoplasm and peripheral (plasma or cell membrane) membrane without nuclear pattern (Figure 3a). However, the quantification of TRPA1+ve marked area revealed no significant differences in TRPA1 expression between the analyzed samples (Figure 3a).  Finally, the expression and functional activity of TRPA1 was evaluated in two different melanoma cell lines, SK-MEL-28 (malignant melanoma) and WM266-4 (derived from a metastatic site of a malignant melanoma) cells. TRPA1 mRNA (Figure 3b) and protein ( Figure 3c) were found in both cell lines. To verify that the TRPA1 mRNA was translated into a functional protein, we evaluated the calcium response to two channel agonists, allyl isothiocyanate (AITC) and hydrogen peroxide (H 2 O 2 ). Both compounds induced a concentration-dependent calcium response that was abolished by the selective TRPA1 channel antagonist, A967079 (Figure 3d,e). TRPA1 activation by electrophilic and reactive endogenous and exogenous agents results in a calcium-dependent NOX1 stimulation that amplifies the oxidative stress signal [25,38,51]. Here, we asked whether TRPA1 could increase ROS production in the two different melanoma cell lines. Exposure of SK-MEL-28 and WM266-4 cells to either AITC or H 2 O 2 increased H 2 O 2 content in the medium, a response that was prevented by the TRPA1 selective antagonist, A967079, and by Trpa1 gene silencing (siRNA) (Figure 3f). Thus, a functional TRPA1 channel expressed by melanoma cells is targeted by ROS to amplify the oxidative stress signal.

Discussion
While proinflammatory M1 MΦs are mainly involved in the initial phase of cancer development by creating a mutagenic microenvironment [52], in more advanced stages of cancer, TAMs often differentiate into anti-inflammatory M2 MΦs, which enhance tumor growth by creating an immunosuppressive TME by producing proangiogenic molecules and proteolytic enzymes to promote angiogenesis and tumor invasion [10]. A higher number of CD68+ve TAMs has been reported with increasing depth of melanoma invasion and ulceration [13][14][15] and in malignant compared to benign melanocytic lesions [18]. Both CD68+ve and CD163+ve TAMs were more abundant in metastatic compared to non-metastatic melanomas [16,17], and in malignant compared to benign melanocytic lesions [18], a finding which has been associated with tumor progression and metastasis aas well as poor clinical outcome. A massive MΦ infiltration has been linked to angiogenesis in cutaneous melanoma, as the numbers of MΦs and neovascularization and vascular endothelial growth factor-A increased significantly with augmented tumor depth [13,14]. A dense CD163+ve MΦ infiltration in melanoma stromal tissue and CD68+ve MΦ infiltration at the invasive front were associated with poor overall survival [16]. Moreover, high MΦ counts correlated with markers of aggressive disease, such as Breslow thickness, ulceration, mitotic rate [15], and tumor recurrence [18].
Here, we confirmed that higher numbers of CD163+ve TAMs are associated with a more invasive melanoma phenotype, as indicated by Breslow thickness. The higher density of CD163+ve TAMs in thicker pT4 melanomas, in comparison with thinner pT1 melanomas and dermal nevi, might suggest an association between immunosuppression and cancer progression. However, different from a previous study [18], we only found a tendency to an increased density of CD68+ve TAMs from dermal and dysplastic nevi to pT1 and pT4 melanomas. Failure to reach statistical significance may, in part, depend on the limited sample size and intrinsic variability of study materials. MΦs are among the major sources of oxidative stress [19], which may exert antioncogenic roles [21,53]. However, opposing evidence proposes that ROS exert selective pressure on the tumor cells and orchestrate cellular signaling networks that may aggravate tumor cell proliferation, malignancy, and drug resistance [54][55][56]. Compared to other types of tumors, melanoma is relatively unique, due to its extraordinarily high levels of oxidative stress in TME [55]. A previous comparative analysis of 4-HNE expression revealed a low level in common nevi that significantly increased in dysplastic nevi and malignant melanomas, while melanoma metastases lost the 4-HNE content [35].
We found progressively increasing levels of 4-HNE in association with severity of disease, with more intense 4-HNE staining in deeply invasive pT4 than in thin pT1 melanomas. Although these results suggest an intimate relationship between 4-HNE and the tumori-genesis process, they cannot determine whether the biomarker exerts a proactive role or remains a functionless end-product of tumor-associated oxidative stress. In malignant and advanced tumors, TAMs are biased toward the M2 phenotype, which favors tumor malignancy [11,57,58], and the increased oxidative stress is intimately linked to tumor progression. Thus, the most parsimonious explanation is that increased 4-HNE levels result from different components. Initially, the oxidative stress burden is generated mainly by invading peritumoral and intratumoral MΦs, with a possible higher contribution of CD163+ve cells. In a second phase, TRPA1, in melanoma cells targeted by ROS from MΦs, amplifies the oxidative stress in TME.
Recently, altered expression of a subset of TRP channels has been described throughout various cancer types, where their presence may confer selective growth and survival advantages to tumor cells, thus promoting cancer development and progression [59]. Some TRPs may regulate cellular functions of human melanoma. TRPM2, TRPM8, TRPV1, and TRPV2 have been identified in melanoma cell lines, where their expression and function confer susceptibility to apoptosis and necrosis [47,60,61]. Functional TRPV1, TRPM8, and TRPA1 were also expressed in malignant human uveal melanoma tissues and cell lines [49].
Limitations of the present study are: (1) the TRPA1 antibody was not extensively used in human tumor samples and (2) in vitro data were obtained in melanoma cell lines and not in primary cultures of melanoma cells. Nevertheless, given the fundamental role of ROS in a wide range of cellular responses, including tumor cell proliferation, migration, and survival [54,55], and a prominent role of TRPA1 as a sensor and amplifier of oxidative stress [25,37,38,51], we explored the role of TRPA1 in regulating oxidative stress production in melanoma cells. The expression of TRPA1 in two different melanoma cell lines was associated with their ability to evoke a calcium response by channel agonists AITC and H 2 O 2 . Importantly, H 2 O 2 elicited a TRPA1-dependent H 2 O 2 release, as it was attenuated by pharmacological antagonism or gene silencing of the channel. These findings suggest that, in melanoma, as reported under circumstances of inflammation or tissue injury [38,51], TRPA1 behaves as an oxidative stress sensor and amplifier.
TRPA1 activation is associated with a non-canonical oxidative-stress defense, as well as a canonical ROS-neutralizing mechanism, thus upregulating anti-apoptotic pathways that favor cancer progression [36]. However, there is also evidence that highlights the prooncogenic role of oxidative stress in melanoma [62]. Solar ultraviolet radiation, a major risk factor for melanoma [63], elicits melanin formation via TRPA1 in melanocytes [46]. TRPA1, which exhibits the highest sensitivity to oxidants due to the presence of hyperreactive cysteines in cytoplasmic domains [36,37], previously [48] and in this study, has been found functionally expressed in melanoma cells. Our finding that H 2 O 2 elicits a TRPA1-dependent calcium response in melanoma cells extends to melanoma the hypothesis [36] that TRPA1 promotes antiapoptotic pro-oncogenic programs. However, our additional result that channel activation by H 2 O 2 in melanoma cell lines amplifies the oxidative stress signal underlines the need of further studies to fully understand the complex and sometimes opposing relationship between MΦs, oxidative stress, and TRPA1 in the initiation and progression of some tumors, including melanoma ( Figure 4). and TRPV2 have been identified in melanoma cell lines, where their expression and function confer susceptibility to apoptosis and necrosis [47,60,61]. Functional TRPV1, TRPM8, and TRPA1 were also expressed in malignant human uveal melanoma tissues and cell lines [49].
Limitations of the present study are: 1) the TRPA1 antibody was not extensively used in human tumor samples and 2) in vitro data were obtained in melanoma cell lines and not in primary cultures of melanoma cells. Nevertheless, given the fundamental role of ROS in a wide range of cellular responses, including tumor cell proliferation, migration, and survival [54,55], and a prominent role of TRPA1 as a sensor and amplifier of oxidative stress [25,37,38,51], we explored the role of TRPA1 in regulating oxidative stress production in melanoma cells. The expression of TRPA1 in two different melanoma cell lines was associated with their ability to evoke a calcium response by channel agonists AITC and H2O2. Importantly, H2O2 elicited a TRPA1-dependent H2O2 release, as it was attenuated by pharmacological antagonism or gene silencing of the channel. These findings suggest that, in melanoma, as reported under circumstances of inflammation or tissue injury [38,51], TRPA1 behaves as an oxidative stress sensor and amplifier. TRPA1 activation is associated with a non-canonical oxidative-stress defense, as well as a canonical ROS-neutralizing mechanism, thus upregulating anti-apoptotic pathways that favor cancer progression [36]. However, there is also evidence that highlights the prooncogenic role of oxidative stress in melanoma [62]. Solar ultraviolet radiation, a major risk factor for melanoma [63], elicits melanin formation via TRPA1 in melanocytes [46]. TRPA1, which exhibits the highest sensitivity to oxidants due to the presence of hyperreactive cysteines in cytoplasmic domains [36,37], previously [48] and in this study, has been found functionally expressed in melanoma cells. Our finding that H2O2 elicits a TRPA1dependent calcium response in melanoma cells extends to melanoma the hypothesis [36] that TRPA1 promotes antiapoptotic pro-oncogenic programs. However, our additional result that channel activation by H2O2 in melanoma cell lines amplifies the oxidative stress signal underlines the need of further studies to fully understand the complex and sometimes opposing relationship between MΦs, oxidative stress, and TRPA1 in the initiation and progression of some tumors, including melanoma ( Figure 4).  Data Availability Statement: Data presented in this study are available on request from the corresponding author.