Introduction

Cancer incidence has increased globally in the twenty-first century [40]. Among them, cutaneous melanoma (CM) has become a public health problem due to its ability to metastasize and its significant contribution to the global statistics data on cancer deaths [28, 38]. This neoplasm’s pathology originates from the malignant transformation of epidermal melanocytes mainly due to excessive or unprotected sunbathing, which leads to DNA mutations by ultraviolet (UV) radiation [7, 31].

Even with new therapeutic possibilities, several side effects and possible pharmacological resistance still limit the effectiveness of cancer treatments [18]. When CM progresses to stages III and IV, these problems are more pronounced [30]. In this context, phenolic compounds, such as rosmarinic acid (RA), have been indicated promising in an adjuvant therapeutic perspective associated with pharmacological use [10]. This phenolic acid is easily and naturally found in plants from Boraginaceae and Lamiaceae families, such as rosemary (Rosmarinus officinalis L.) [34]. Some research has already shown the anticancer effect of RA in colon carcinoma [20], prostate cancer [19], and breast cancer [29]. Although works are showing the effects of RA in the field of cancer and of CM, the effects of RA in humans are still poorly understood.

Recently, the purinergic system, a ubiquitous and sophisticated cell–cell communication extracellular signaling pathway [6], has been shown to play an essential role in cancer pathophysiology, such as in lung cancer [49], glioblastoma [1] and melanoma [27]. The molecules involved in this cell signaling pathway mainly include the nucleotides adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP) and the nucleoside adenosine (Ado), whose levels are regulated by ectonucleotidases NTPDase (CD39), 5′-nucleotidase (CD73) and adenosine deaminase (ADA). The signaling molecule’s action on P1 and P2 receptors is implicated in several cellular outcomes [5, 6].

Although other signaling pathways have shown a solid association with the anticancer action of RA, the potential for purinergic signaling modulation has yet to be explored in the experimental context. In a previous study, we proposed hypothetically that RA may be a modulator of the purinergic system [10]. Continuing to research RA properties, in this research we investigated the effects of RA on melanoma cell viability, purinergic signaling modulation, and cytokine levels. Thus, this is the first study that focused on experimentally investigating the purinergic signaling as a target of RA in melanoma context.

Materials and methods

Chemicals, reagents and equipment

All chemicals and reagents used were of analytical grade, purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO, USA) and Merck (Darmstadt, Germany). RA (99% purity) (Fig. 1) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Cell culture plates and flasks used for culture procedures were obtained from Gibco™ Thermo Fisher Scientific (Grand Island, NY, USA) and Invitrogen Life Technologies (Carlsbad, CA, USA). Molecular biology reagents were purchased from Invitrogen and Applied Biosystems (Waltham, Massachusetts, USA). Flow cytometry was analyzed in an Accuri™ C6 Plus cytometer (BD Biosciences) and by the FlowJo V10 software.

Fig. 1
figure 1

Source: Pubchem

Molecular structure of rosmarinic acid (RA).

Full size image

Cell culture and treatment with RA

The human metastatic melanoma cell line SK–MEL–28 was purchased from the Cell Bank of Rio de Janeiro (CBRJ), Brazil. Cells were grown in flasks with Dulbecco’s modified Eagle’s medium (DMEM) with glucose (4,500 mg/L), L–glutamine, containing antibiotics/antifungal (1% penicillin/streptomycin) and supplemented with 10% fetal bovine serum. The cells were grown under adequate conditions in a humidified and controlled atmosphere of 5% carbon dioxide (CO2) at 37 °C. RA was dissolved in the appropriate culture medium to obtain different concentrations, and the cells were treated for 24 h at concentrations of 400 µM and 800 µM according to previous studies that found optimal RA concentrations to be used [9]. The negative control group cells received only culture medium.

Apoptosis assay

The detection of the apoptosis-inducing capacity of RA was assessed by an Annexin V–FITC and propidium iodide (PI) apoptosis detection kit (BD Biosciences), used according to the manufacturer’s instructions. Briefly, after exposure to RA, cells were harvested and washed twice with cold PBS. Then, cells were resuspended at 1 × 106 cells/mL in 100 μL binding buffer, mixed with 5 μL Annexin V–FITC and 5 μL PI (20 μg/mL), and then incubated for 15 min at room temperature in the dark. Finally, the reaction was mixed with 400 μL binding buffer and read on the BD Accuri™ C6 Plus at every 20,000 events.

Gene expression of CD39 and CD73

We used the real-time quantitative polymerase chain reaction (RT–qPCR) analysis to assess the gene expression of CD39 and CD73. First, we obtained the RNA from cell culture samples with TRIzol™ reagent (Invitrogen™) and quantified it spectrophotometrically (Thermo Scientific™ Varioskan™ LUX). Reverse transcription into cDNA was performed with a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems™) by the addition to each 10 μL sample of 10 μL of a mix containing 1 μL of MultiScribe™ Reverse Transcriptase. The steps of the reaction were 25 °C for 10 min, 37 °C for 120 min, 85 °C for 5 min, and a final hold step of 4 °C for 30 min, performed using a thermal cycler. The RT–qPCR reaction was performed using 17 μL of a mix containing the SYBR Green PCR Master Mix (Applied Biosystems™) and 3 μL of the cDNA sample. The parameters used were a pre-activation step of 10 min at 95 °C, followed by a cycling step of 40 cycles at 95 °C for 15 s, 60 °C for 60 s, and finally, a melting step with a melting curve of 60 °C to 95 °C with an increase of 1 °C every 5 s. The relative expression of each gene was represented as the fold expression about the control and calculated using the comparative ΔΔCT value. GAPDH was used as the housekeeping gene to normalize gene expression. The forward and reverse sequences of oligos (5′-3′) used for each gene were as follows: GAPDH (F): CTCCTCACAGTTGCCATGTA; GAPDH (R): GTTGAGCACAGGGTACTTTATTG; CD39 (F): GCCCTGGTCTTCAGTGTATTAG; CD39 (R): CTGGCATAACCTACCTACTCTTTC; CD73 (F): GTGCCTTTGATGAGTCAGGTAG; CD73 (R): TTCCTTTCTCTCGTGTCCTTTG.

Assessment of CD39, CD73 and A2A protein expression

SK–MEL–28 cells were cultured under treatment conditions for 24 h, followed by trypsinization using trypsin, and counted in a hemocytometer. Around 1 × 106 cells were centrifuged for 5 min at 400 g and washed twice with PBS with 10% FBS. The cell pellets were suspended and incubated for 30 min with purified mouse anti–human CD39 (clone A1 (RUO), catalog Nº. 567157), anti–human CD73 (clone AD2 (RUO), catalog Nº. 550257), and anti-human A2A (catalog Nº. MA5–31611) antibodies conjugated with fluorescent phycoerythrin (PE) (BD Pharmingen TM) (1:10). Then, the same number of cells was incubated without antibodies (non–stained). All samples were washed with PBS, and 10,000 events were immediately acquired by flow cytometry (BD ACCURI C6) and analyzed by FlowJo V10 software. Each target (CD39, CD73, and A2A) was analyzed separately and incubated with respective antibodies. The results were expressed as a percentage (%) of CD39 or CD73 or A2A positive cells relative to the control.

Enzymatic activity of ectonucleotidases

The hydrolysis of nucleotides such as ATP, ADP, AMP, and the nucleoside Ado was employed to evaluate the alterations in purinergic system enzyme activities. Briefly, after protein adjustments of cells, 20 μL of samples were added to the reaction mixture of each enzyme and pre-incubated at 37 °C for 10 min. The enzymatic reaction was initiated by adding the specific substrates for each enzyme: ATP and ADP for ATPase/ADPase and AMP for AMPase. After incubation at 37 °C for 70 min, the reactions were stopped by the addition of trichloroacetic acid (TCA, 10%), and the released inorganic phosphate due to ATP, ADP, and AMP hydrolysis was determined by using malachite green as the colorimetric reagent. A standard curve was prepared with KH2PO4. Controls were performed to correct for non-enzymatic hydrolysis. The absorbance was measured at 630 nm, and enzyme-specific activities were reported as nmol/Pi/min/mg of protein [24, 39].

For the Ado breakdown, assessed by ADA activity, was performed based on the measurement of ammonia produced when this enzyme is activated, following a previously published method (Giusti and Galanti, 1984). In brief, 50 μL of cell suspension reacted with 21 mM of adenosine (pH 6.5) at 37 °C for 60 min. After incubation, the reaction was stopped by adding 167.8 mM sodium nitroprusside, 106.2 mM phenol, and a sodium hypochlorite solution. Lastly, absorbance was read at 620 nm, and values were expressed as units/liter (U/L).

Assessment of cytokine levels by flow cytometry

The levels of cytokines released from melanoma cells, interleukin–4 (IL–4), interleukin–6 (IL–6), interleukin–10 (IL–10), interferon–γ (IFN–γ), and tumor necrosis factor–α (TNF–α), were verified by BD™ CBA Human Th1/Th2 Cytokine Kit II (catalog Nº. 551,809) following the manufacturer’s recommendations. The method principle of the kit is based on the formation of sandwich complexes (capture bead/cytokine/fluorescent detection reagent) of cytokines with conjugated specific antibody beads of known size and fluorescence, whose intensity can be measured in flow cytometry equipment and is proportional to the cytokine level. Briefly, 50 µL of supernatant samples were mixed with 50 µL of Human Th1/Th2–II PE Detection Reagent in a microtube, followed by 3 h of incubation at room temperature and protected from light. After time elapsed, the samples were washed with 1 mL of washing buffer and centrifuged at 200 g for 5 min. Then, the supernatant from each test tube was discarded, and the sandwich bead pellet was resuspended with 300 µL of buffer for analysis by flow cytometry.

Protein determination

We used the Bradford [4] method to assess the protein concentration in samples using bovine serum albumin as the standard. When necessary, the protein of samples was adjusted according to each protocol.

Statistical analysis

All measurements were statistically performed by analysis of variance followed by the appropriate post hoc test using GraphPad Prism 9 software. The differences between the untreated (CT) and treated cells were evaluated by unpaired Student’s t–test or through the variance analysis one–way ANOVA followed by Tukey’s post hoc test. All data are expressed as mean ± standard deviation. The differences in the probability of rejecting the null hypothesis at < 5% (P < 0.05) were considered statistically significant. Statistical significance was defined for p–values of *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001.

Results

RA induces apoptosis in melanoma cells

The apoptosis-induced in melanoma cells after 24 h of RA treatment is shown in Fig. 2. It is notably that RA significantly induced apoptosis in melanoma cells at concentrations of 400 µM (P < 0.0001) and 800 µM (P < 0.0001) in comparison to control (Fig. 2A–B). These results showed that RA exhibits strong antineoplastic effects on melanoma cells through induction of apoptosis.

Fig. 2
figure 2

Apoptosis-induced melanoma cells after treatment with RA. RA also significantly induced apoptosis of melanoma cells at concentrations of 400 µM and 800 µM (A–B). Control (CT). All experiments were performed independently at least three times and in three replicates. Data are presented as mean ± SD. Statistical analysis: one–way ANOVA followed by a post hoc Tukey’s multiple comparisons test. Values with P < 0.05 were considered statistically significant. *(P < 0.05); **(P < 0.01); ***(P < 0.001); ****(P < 0.0001). All data indicate differences from the control group

Full size image

RA modulates expression of purinergic signaling components in melanoma cells

We searched for a possible modulatory effect of RA on the gene and protein expression of CD39 and CD73 in melanoma cells (Fig. 3A–F). After 24 h, the treatment with RA downregulated the gene expression of CD39 at concentration of 800 µM (P = 0.0183) when compared to the control group (Fig. 3A). However, this effect was no observed on the protein expression of CD39 (Fig. 3B–C). Regarding the CD73, we also found a reduction in gene expression at concentration of 800 µM (P = 0.0003), but in contrast to CD39, the protein expression of CD73 also was strongly reduced (P < 0.0001) in this treatment concentration (Fig. 3D–F).

Fig. 3
figure 3

Gene and protein expression of the CD39 and CD73. We employed RT–qPCR to assess the gene expression of CD39 and CD73 in melanoma cells. To verify the protein expression, we used flow cytometry. At concentration of 800 µM, the treatment with RA reduced the gene expression of CD39 (A) and of CD73 (D). Similarly, the same RA treatment concentration was able to decreased of CD73 protein expression (E–F). There was no statistical significance for CD39 protein expression (B–C). Control (CT). All experiments were performed independently at least three times and in three replicates. Data are presented as mean ± SD. Statistical analysis: one–way ANOVA followed by a post hoc Tukey’s multiple comparisons test. Values with P < 0.05 were considered statistically significant. *(P < 0.05); **(P < 0.01); ***(P < 0.001); ****(P < 0.0001). All data indicate differences from the control group

Full size image

In addition, we investigated the effect of RA on the protein expression of the A2A receptor (Fig. 4A–B). Interestingly, we found that RA treatment was able to reduce A2A expression in melanoma cells at concentrations of 400 µM (P < 0.0001) and 800 µM (P < 0.0001) when compared to control.

Fig. 4
figure 4

Protein expression of the receptor A2A. After 24 h of treatment with RA, the expression of the receptor A2A significantly decreased at both concentrations of 400 µM and 800 µM in melanoma cells (A–B). Control (CT). All experiments were performed independently at least three times and in three replicates. Data are presented as mean ± SD. Statistical analysis: one–way ANOVA followed by a post hoc Tukey’s multiple comparisons test. Values with P < 0.05 were considered statistically significant. *(P < 0.05); **(P < 0.01); ***(P < 0.001); ****(P < 0.0001). All data indicate differences from the control group

Full size image

RA modulates activity of ectonucleotidases in melanoma cells

Figure 5 (A–D) shows the enzymatic activity of ectonucleotidases in the hydrolysis of nucleotides ATP, ADP and AMP, and the Ado nucleoside after 24 h of treatment with RA. Curiously, at concentration of 800 µM, the treatment significantly decreased the ADP hydrolysis (P < 0.0001) (Fig. 5B). Likewise, the treatment with RA at concentration of 800 µM greatly reduced the AMP hydrolysis (P < 0.0001) (Fig. 5C). No statistical significance was observed in ATP hydrolysis and Ado breakdown (Fig. 5A and D).

Fig. 5
figure 5

Enzymatic activity of ectonucleotidases. We assessed the ATPase/ADPase/AMPase activity on ATP, ADP, and AMP hydrolysis, respectively. In addition, we also assessed the breakdown of adenosine (Ado) by the ADA activity. Treatment of 800 µM for 24 h o RA significantly decreased the hydrolysis of ADP (B) and AMP (C) in melanoma cells. There was no statistical significance for ATP hydrolysis (A) and Ado breakdown (D) in both tested concentrations of RA. All experiments were performed independently at least three times and in three replicates. Data are presented as mean ± SD. Statistical analysis: one–way ANOVA followed by a post hoc Tukey’s multiple comparisons test. Values with P < 0.05 were considered statistically significant. *(P < 0.05); **(P < 0.01); ***(P < 0.001); ****(P < 0.0001). All data indicate differences from the control group

Full size image

RA modulates the cytokine levels in melanoma cells

We also investigated the levels of cytokines after 24 h of RA treatment (Fig. 6A–E). We found that RA is a phenolic compound that modulates the levels of cytokines in melanoma cells, such as IL–4, that had increased levels at concentrations of 400 µM (P = 0.0002) and 800 µM (P = 0.0005) in comparison to the control group (Fig. 6A). The levels of IL–6 were also increased at concentrations of 400 µM (P = 0.0045) and 800 µM (P = 0.0010) (Fig. 6B). We also found increased levels for IL–10 at concentrations of 400 µM (P = 0.0011) and 800 (P = 0.0016) of RA (Fig. 6C). The levels of IFN–γ were increased at concentration of 800 µM (P = 0.0152) (Fig. 6D). In addition, we detected elevated levels of TNF–α after treatment with RA in melanoma cells in 400 µM (P = 0.0007) and 800 µM (P = 0.0004) (Fig. 6E).

Fig. 6
figure 6

Cytokine levels. RA treatment increased the levels of cytokines IL–4 (A), IL–6 (B), IL–10 (C), and TNF–α (E) in both concentrations of 400 µM and 800 µM. In the case of INF–ɣ, treatment with RA increased the levels of this cytokine at concentrations of 800 µM (D). Control (CT). All experiments were performed independently at least three times and in three replicates. Data are presented as mean ± SD. Statistical analysis: one–way ANOVA followed by a post hoc Tukey’s multiple comparisons test. Values with P < 0.05 were considered statistically significant. *(P < 0.05); **(P < 0.01); ***(P < 0.001); ****(P < 0.0001). All data indicate differences from the control group

Full size image

Discussion

Several studies have shown the multiple biological properties of RA, such as anti-inflammatory [25], anti–oxidant [35], and neuroprotection [16]. In the field of oxidative stress, Venza et al. [44] showed that the molecule ROS has a significant role in melanoma, being that high levels of ROS are involved in the pathophysiology of this cancer type. Interestingly, in a recent article, we proved that RA can reduce both intracellular and extracellular ROS levels in melanoma cells besides modulates genes related to cell death [9].

An important cell signaling pathway called purinergic signaling, has been linked to pathophysiology of CM [27, 49]. Unfortunately, there is a poor prognosis in the advanced stages of CM due to insufficient effectiveness of available treatments [45]. Even if the evidence about RA's antineoplastic potential is highlighted, research on purinergic signaling as a target of RA in the CM context needs to be improved. In a previous study, we hypothesized that RA may be a modulator of the purinergic system in melanoma context [10]. In this ground–breaking study, we proved that RA is able to downregulates the expression of CD73 and A2A, and modulates the ectonucleotidases activity. Taken all together, these purinergic signaling modulation cause increasing in the levels of several cytokines that culminates in apoptosis of melanoma cells.

In previous experimental work, we reported that RA has an antineoplastic effect on melanoma cells by reducing viability [9]. Here we found that RA induces apoptosis at concentrations of 400 µM and 800 µM (Fig. 2A–B). Similar, a study involving CT26 colorectal cancer cells showed that RA reduces cell viability after 24 h of treatment [17]. Reports from the literature also showed that RA can induce apoptosis in breast [29] and prostate [19] cancer cells, but this effect had not been well explored in SK–MEL–28 lineage until then. Thus, our results reinforce the antineoplastic effect of RA and add new findings in the melanoma context.

Sequentially, we continue searching for a possible modulatory effect on the components of the purinergic system. We found that RA significantly reduced the expression of the CD73 at concentration of 800 µM (Fig. 3D–F). In an animal model, tumor size reduction was related to CD73 inhibition [13]. In humans, overexpression of CD73 has been described in many cancer cell types, such as breast cancer, colorectal cancer, ovarian cancer, gastric cancer, and gallbladder cancer [14]. The expression and activity of CD73 seem to be associated with tumor invasion and metastasis [8]. It was also shown that patients with advanced melanoma have high rates of CD73 expression [32]. Thus, our results bring to light the possibility of RA pharmacotherapeutic applications having the CD73 as a target against melanoma. In addition, although associated with a hydrolytic activity, CD73 also has a non–enzymatic action, promoting cell migration in the extracellular matrix (ECM) through the activation of focal adhesion kinase (FAK) in melanoma cells [41]. Thus, together with our previous study in which we found that RA inhibits cell migration [9], now we have support to the hypothesis that this inhibition of adhesion of melanoma cells by may occurs by the effect of RA on CD73 expression.

Alterations in the enzymatic activity of ectonucleotidases were highlighted in previous studies involving patients with lung cancer [50], uterine cervical neoplasia [26], and melanoma [27]. Considering the studies above and the results presented in this work, which indicated that RA is a phenolic compound that modulates the expression of components of the purinergic system, we advanced our study to understand its action on the purinergic enzymatic cascade in detail. For this, assays to verify the enzymatic activity of ectonucleotidases were performed. After 24 h of treatment, we found that RA interfered in the enzymatic activity of ectonucleotidases, reducing significantly the hydrolysis of ADP (Fig. 5B), but did not have the same effect on ATP hydrolysis (Fig. 5A). Although CD39 is an enzyme capable of hydrolyzing ATP into ADP and ADP into AMP [12], it is not the only enzyme with this function. Alkaline phosphatases (APs) can also specifically hydrolyze ADP into AMP [2], which may justify the reduction in ADP hydrolysis that we found, not being strictly related to a previous hydrolysis of ATP by CD39. The hydrolysis of AMP was also decreased at a concentration of 800 µM (Fig. 5C). These outcomes are explained by reduction in CD73 expression after treatment with RA (Fig. 3E–F), which promotes changes in the purinergic signaling cascade and reduces Ado formation.

Regarding Ado, the CD73/ADA enzyme axis tightly regulates the levels of this nucleoside at the end of the purinergic cascade. This is due to CD73 promoting the conversion of AMP into Ado, while ADA converts Ado into inosine by an irreversible deamination reaction [3]. Thus, the ADA is crucial in controlling Ado levels [8]. The presence of increased levels of Ado in the tumor microenvironment (TME) is related to tumor immunosuppression [11]. Reduced ADA activity and increased P1 receptor expression (activated by Ado) in patients with advanced–stage lung cancer have been attributed to pro-tumor effects by increasing IL–6 and TNF–α levels and decreasing IL–17 and INF–γ [49]. However, in this study, we did not find statistical significance for ADA activity after treatment with RA (Fig. 5D), and thus, in opposition to the studies mentioned above, our results indicated that purinergic cascade was more prone to a decrease in Ado formation in melanoma cells. Hence, our results relative to CD73 downregulation indicate that RA’s antineoplastic effect may be linked to two mechanisms: reducing Ado production and inhibiting cell adhesion.

As shown, CD73 is responsible for converting AMP into Ado and thus exerts essential control over the immune response against tumors [8]. In cancer cells, the activation of the A2A receptor by Ado also participates in the immune response, promoting a pro–tumoral condition via the PIK3/AKT signaling pathway [42], with an increase in angiogenesis, tumor growth, and reduction of the immune response, such as IL–2, IL–6, and IFN–γ production [33, 47]. Recently, it has been shown that simultaneous inhibition of CD73 and A2AR exhibits a synergistic effect against cancer [46]. Thus, the blockade of A2AR may be an exciting cancer immunotherapy [48]. In melanoma, blocking A2A is linked to restoring the antitumor immune response [36]. Kim et al. [21] suggested hypothetically that RA plays a negative allosteric role in regulating the A2A receptor. Intriguingly, found that RA significantly reduces the expression of A2A (Fig. 4A–B).

We must highlight that purinergic signaling closely connects with the immune system, promoting many physiological mechanisms. Interestingly, in this field, evidence has shown that melanocytes exhibit macrophage phagocytic-like functions and may actively participate in immune responses, particularly as antigen-processing or antigen–presenting cells [22]. Moreover, this cell type can produce cytokines, such as IL–4, IL–6, and IFN–γ, in inflammatory processes [15]. Considering that CM is derived from melanocytes and our results regarding the CD73 and A2A receptors, we searched for cytokine levels in melanoma cells after treatment with RA. In this in vitro study, we found that RA treatment increased the levels of IL–4, IL–6, IL–10, IFN–γ, TNF–α (Fig. 6A–E) at the same concentrations as RA significantly decreased A2A receptor expression.

Two cytokines involved in the immune response that play critical roles in the inflammatory process are IL–6 and IL–10. While IL–6 is an important pro-inflammatory mediator that induces acute phase responses and the production of other cytokines, IL–10 is related to the suppression of inflammation [37, 43]. Our study also showed increased cytokine levels, emphasizing IL–6 and IL–10. Although simultaneous increases in both IL–6 and IL–10 levels can be explained by the assumption that cytokine release is an inflammation regulatory balance, data from the literature supports that IL–6 may have growth-inhibitory effects against melanoma [23].

Conclusion

In an unprecedented-like manner, we proved that RA downregulated expression of CD73 and A2A in melanoma cells. Reduction in CD73 expression promoted a modulation in hydrolysis of ADP and AMP. All this purinergic signaling modulations by RA, increased the levels of cytokines, reversing the immunosuppression, and led melanoma cells to apoptosis. Thus, our in vitro results confirm RA's potential to be used to induce melanoma cell apoptosis, having CD73 and A2A as target when reversion of immune suppression is desired. We suggest further studies in animal models and clinical trials focusing on the modulation of purinergic signaling by RA in melanoma.