UVA-Irradiation Induces Melanoma Invasion via the ...

ORIGINAL ARTICLE

UVA-Irradiation Induces Melanoma Invasion via the Enhanced Warburg Effect York Kamenisch1,2,4, Tarza S.A. Baban2,4, Winfried Schuller2, Anna-Katharina von Thaler3, Tobias Sinnberg2, Gisela Metzler2, Ju¨rgen Bauer2, Birgit Schittek2, Claus Garbe2, Martin Rocken2 and Mark Berneburg1 Melanoma is a malignant tumor in which UVA (320e400 nm) radiation is considered to be an important risk factor. But the role of UVA in melanoma progression toward an invasive phenotype is still not adequately investigated. For most proliferating tumor cells the preference of aerobic glycolysis has been described as the Warburg effect. Here we investigate the effect of UVA irradiation on changes in the Warburg effect and tumor progression toward invasive potential. On UVA irradiation, melanoma cell lines from initial tumors show an induction of the Warburg effect with increased glucose consumption and lactate production, which is at least partially mediated by reactive oxygen species. Associated with UVA treatment and enhanced lactic acid production, tumor-relevant proteases and in situ invasion is upregulated. Simultaneously, UVA increases intracellular concentrations of progression marker transketolase and activated protein kinase Akt, both involved in metabolic changes that increase with proliferation. Using invasion assays we show that lactic acid, resulting from the UVA enhanced and partially reactive oxygen species-mediated Warburg effect, increases the invasive potential of all melanoma cell lines investigated. Therefore, we demonstrate in melanoma cells that production of lactic acid, induced by UVA irradiation, increases invasiveness of melanoma cells via expression of tumorrelevant proteases. Journal of Investigative Dermatology (2016) 00, -e-; doi:10.1016/j.jid.2016.02.815

INTRODUCTION Melanoma is a malignant skin tumor characterized by high morbidity and mortality (Howe et al., 2001; Jemal et al., 2010; Jhappan et al., 2003; Leiter et al., 2014). Melanomas develop from initial tumor cells via radial and vertical growth, ultimately leading to metastasis, and epidemiological evidence indicates that UV radiation is involved in the generation of melanoma (Bald et al., 2014; Fears et al., 2002; Tucker and Goldstein, 2003). Recently it was shown that UVB irradiation enhances perivascular invasion of melanoma cells (Bald et al., 2014). Although this UVB irradiation correlates with sunburn reactions in the skin, solar UV radiation with physiological relevance consists, to the largest extent, of UVA radiation (320e400 nm) (Baczynska et al., 2013; Parisi and Wong, 2000; Turnbull and Parisi, 2003). The effect of UVA irradiation is partially

1

Department of Dermatology, University Hospital Regensburg, Regensburg, Germany; 2Department of Dermatology, Eberhard Karls University, Tu¨bingen, Germany; and 3Department of Neurology, Eberhard Karls University, Tu¨bingen, Germany

4

These authors contributed equally to this work.

Correspondence: Mark Berneburg, Department of Dermatology, University Hospital Regensburg, 93042 Regensburg, Germany. E-mail: Mark. [email protected] Abbreviations: ANOVA, analysis of variance; 2-DG, 2-deoxy-D-glucose; IM, initial melanoma; MM, metastasizing melanoma; MMP, matrix metalloproteinase; ROS, reactive oxygen species; SD, standard deviation; TKTL-1, transketolase 1; TIMP1, tissue inhibitor of metalloproteinase; uPA, urokinasetype plasminogen activator Received 30 June 2014; revised 20 January 2016; accepted 14 February 2016; accepted manuscript published online 13 May 2016; corrected proof published online XXX XXXX

mediated by reactive oxygen species (ROS), finally leading to intracellular oxidative damage (Beissert and Loser, 2008; Kappes et al., 2006). This is of particular importance because—as opposed to UVB—UVA reaches basal layers of the epidermis containing melanocytes as well as the dermis at pathophysiologically relevant doses. Despite its physiological relevance, the contribution of UVA to solar-induced melanoma is still discussed controversially as studies with a Xiphophorus hybrid fish model could not detect UVA-induced melanomas (Mitchell et al., 2010). In addition to this, epidemiological data report melanoma on usually sun-protected body sites (Levell et al., 2009), but there are also epidemiological data that show a clear correlation between sunburn and melanoma in sun-sensitive patients (Newton-Bishop et al., 2011). Furthermore, a single neonatal UVA irradiation, mimicking sunburn reactions in childhood, in a transgenic mouse model on an albino inbred background did not significantly induce melanomas (De Fabo et al., 2004; Noonan et al., 2001; Zaidi et al., 2011). But interestingly in pigmented transgenic mice, a single neonatal dose of UVA irradiation (Noonan et al., 2012) induces oxidative DNA damage in melanocytes and is sufficient to induce melanotic tumors. Although these studies strongly support the role of a single neonatal high dose of UVA during melanoma pathogenesis, they do not model other modes of UV exposure, such as the repeated exposure of humans to low doses of solar UV irradiation in everyday life. In addition, it is still not clear which role exposure to UVA irradiation plays during early melanoma progression. Two important features during progression of initial melanoma are (i) invasion, needing specific proteases, and (ii) proliferation, needing a specific metabolic setting.

ª 2016 The Authors. Published by Elsevier, Inc. on behalf of the Society for Investigative Dermatology.

www.jidonline.org

1

Y Kamenisch et al.

Ultraviolet A and Melanoma

Proliferating cells, including tumor cells, turn to glycolysis for energy production with subsequent lactate fermentation even in the presence of oxygen, a metabolic characteristic described by Warburg (1927), thus called the Warburg effect. Increased glycolysis is associated with the activation of Akt (Elstrom et al., 2004), a protein kinase also involved in antiapoptotic signaling (Majewski et al., 2004). Interestingly, Akt phosphorylation at serine 473 is present in the majority of immunohistologically investigated melanomas (Dhawan et al., 2002). Another tumor-relevant metabolic change is the activation of the pentose phosphate pathway that provides tumor cells with components for nucleotide synthesis. A key enzyme of this pathway is transketolase like 1 (TKTL1), which is used as a tumor marker for highly proliferative cancers (Diaz-Moralli et al., 2011; Langbein et al., 2006). An important step in melanoma progression is invasion of adjacent tissue, facilitated by proteases such as matrix metalloproteinases (MMPs) and urokinase-type plasminogen activator (uPA). Particularly, it was shown that MMP9 and uPA are activated during invasion of melanoma cells (Bianchini et al., 2006; Tang et al., 2013). To investigate the role of UVA irradiation for melanoma progression and invasion via the Warburg effect, we exposed cell lines of initial melanoma with chronic sublethal doses of UVA. Here we show a UVA-dependent increase in glucose uptake and lactate production, and a pH decrease. This lactate production subsequently led to increased expression of MMPs and uPA, resulting in increased invasiveness of UVA-treated melanoma cells. We provide a functional link between the UVA-induced Warburg effect with enhanced lactic acid production and enhanced expression of MMP and uPA, which finally promotes enhanced invasion. RESULTS The UVA-induced Warburg effect mediated by ROS

The Warburg effect is characterized by a preference for glycolysis (enhanced glucose consumption) and subsequent lactic acid fermentation. Melanoma cells from initial melanomas (IM) and metastasizing melanomas (MM) (Supplementary Table S1 and Supplementary Figure S1 online) were irradiated with UVA (6 J/cm2) three times daily for 4 consecutive days. Glucose consumption and lactate production of IM and MM cells were measured in the absence or presence of UVA exposure. On UVA irradiation all melanoma cell lines showed increased (Student’s t-test; P < 0.01) glucose consumption (Figure 1a and c) and lactate production (Figure 1b and d) (mean value and standard deviation [SD] of at least four independent experiments and Student’s ttest shown; asterisks represent P < 0.01 and P < 0.05). This effect was also observed for different doses of UVA irradiation in IM and MM cells, partially showing a significant dosedependent upregulation of the Warburg effect (analysis of variance [ANOVA], post-test for linear trend) (Supplementary Figure S2a and b online). Despite different genetic backgrounds, every IM and MM investigated tended, when compared with unirradiated control, to enhanced UVA-induced glucose consumption (Figure 1e) and lactate production (Figure 1f) (mean value and SD of at least three independent experiments shown). As a consequence of the increased release of lactic acid in the medium, the pH in the 2

Journal of Investigative Dermatology (2016), Volume

-

medium decreased compared with unirradiated control (Figure 1g and h) (mean value and SD of at least three independent experiments shown and Student’s t-test shown; asterisks represent P < 0.01). Because exposure to UVA irradiation generates ROS (Meewes et al., 2001), we tested whether the UVA-induced Warburg effect in melanoma cells is ROS mediated by coincubation with ROS quencher Trolox. The addition of 20 mM Trolox significantly decreases the UVA-induced glucose consumption and lactate production (Figure 1i and j) (mean value and SD of at least three independent experiments and two-way ANOVA, Bonferroni multiple comparisons post-test shown; asterisk represents P < 0.05). The inhibitor of glycolysis 2-deoxy-D-glucose (2DG) reversely blocks activity of glucose kinase, a key enzyme of the glycolytic pathway. In consequence, we found that 5.5 mM of 2DG significantly diminished (Student’s t-test; P < 0.05) UVA-induced glucose consumption and increased lactate production (Figure 1k and l) (glucose consumption and lactate production presented as the mean value with SD of triplicates as representative of at least four independent experiments). To recapitulate the situation in human skin in vivo more closely, skin reconstructs (Backvall et al., 2002) with different melanoma cells (IM and MM) treated either with or without repetitive UVA doses showed more lactate production on UVA irradiation compared with unirradiated control (Figure 1meo) (mean value and SD of at least three independent experiments and Student’s t-test shown; asterisks represent P < 0.01 and P < 0.05). UVA irradiation enhances transketolase activity

Transketolase isoform TKTL1 is a key enzyme of the pentose phosphate pathway and part of the metabolic network supporting growth of tumor cells (Resendis-Antonio et al., 2010). To investigate the effect of UVA irradiation on TKTL1 expression, melanoma cells (IM and MM) were UVAirradiated (as described above) and expression of TKTL1 was detected by immunocytochemical staining. All melanoma cell lines (IM and MM) show enhanced TKTL1 expression on UVA irradiation (Supplementary Figure S2d and e). Simultaneously, the same UVA treatment increased total transketolase activity in melanoma cells (IM and MM) in a colorimetric assay (Smith et al., 2006). Increased transketolase activity on UVA treatment (Figure 2a) was visible in every investigated IM and MM, independent from their different genetic background (mean value and SD of three independent experiments). This UVA-enhanced transketolase activity in IM melanoma cells was dependent on UVA-induced ROS, as the addition of Trolox attenuated UVA-induced transketolase activity (Figure 2b and c) (mean value and SD of at least three independent experiments and two-way ANOVA, Bonferroni multiple comparisons post-test shown; asterisk represents P < 0.05). Furthermore, UVA-enhanced transketolase activity showed partially a dose-dependent upregulation of transketolase activity (Supplementary Figure S2c). Skin reconstructs with different melanoma cells (IM and MM) treated either with or without repetitive UVA doses also showed elevated transketolase activity on UVA irradiation

Y Kamenisch et al.


Figure 1. UVA induces the Warburg effect. Melanoma cell lines from initial (IM) and metastasizing melanoma (MM) or skin reconstructs with IM and MM were treated repetitively with UVA with 6 J/cm2 three times daily for 4 consecutive days. UVA increases glucose consumption and lactate production in IM (a, b) and in MM (c, d). (e, f) UVA increases glucose consumption and lactate production in IM and MM. (g, h) UVA decreases the pH value of the medium of IM and MM. (i, j) Treatment of melanoma cells with reactive oxygen species (ROS) quencher Trolox during UVA irradiation attenuates UVA-induced glucose consumption and lactate production. (k, l) UVA-induced glucose consumption and lactate production could be decreased in the presence of 5.5 mM glycolytic inhibitor 2-deoxy-D-glucose (2DG). (meo) UVA increases lactate production in skin reconstructs with IM and MM melanoma cells.

www.jidonline.org

3

Y Kamenisch et al.


Figure 2. Repetitive UVA irradiation enhances transketolase activity. (a) UVA increases transketolase activity in IM and MM cells, and the addition of ROS quencher Trolox attenuates this effect in (b) IM and (c) MM. (def) UVA irradiation increases transketolase activity in skin reconstructs with IM and MM melanoma cells. (g) UVA irradiation enhances activation of proteinkinase Akt. Melanoma cell lines (IM and MM) were UVA-irradiated with subsequent detection of phosphorylated (Ser 473) Akt and ß actin and the level of phosphorylated (Ser 473) Akt relative to ß actin was quantified densitometrically. IM, initial melanoma; MM, metastasizing melanoma; ROS, reactive oxygen species.

compared with unirradiated control (Figure 2def) (mean value and SD of at least three independent experiments and Student’s t-test shown; asterisk represents P < 0.05). Hyperactivation of the proteinkinase Akt by phosphorylation at the serine residue 473 is associated with increased glucose metabolism and significantly correlates with TKTL1 expression in tumors (Volker et al., 2008). Melanoma cells (IM and MM) were exposed to UVA, and after treatment, the levels of phosphorylated AKT were determined by western blot. Repetitive UVA irradiation significantly (Student’s t-test; P < 0.05) enhances the level of phosphorylated Akt in IM cells and it was also increased in MM cells (Figure 2g) (mean value and SD of at least three independent experiments and Student’s t-test shown; asterisk represents P < 0.05). Persistence of the UVA-induced Warburg effect after UVA irradiation

To test whether the UVA-induced Warburg effect persists beyond cessation of UVA treatment, IM cells were 4


-

UVA-irradiated with 6 J/cm2 for 4 days and afterward seeded for growth on a new plate for another 5 days without irradiation. Subsequently, glucose consumption, lactate production, and transketolase activity (Figure 3a) were measured. When compared with unirradiated controls, UVA-mediated induction of glucose consumption and lactate production continued (Student’s t-test; P < 0.01) 5 days after cessation of the last UVA irradiation (Figure 3b and c) (mean value and SD of at least three independent experiments and two-way ANOVA, Bonferroni multiple comparisons post-test shown; asterisks represent P < 0.01). Again, this UVA-induced longterm persistence of elevated glucose consumption and lactate production was ROS mediated, similar to the immediate UVA-induced Warburg effect, as the addition of Trolox during UVA treatment again decreased this long-term persistence (Figure 3b and c). Similar to the UVA-induced Warburg effect, UVA-induced transketolase activity also persisted, as transketolase activity 5 days after irradiation was still significantly higher when compared with unirradiated controls

Y Kamenisch et al.


Figure 3. UVA enhanced glucose consumption and lactate production persists 5 days after UVA treatment and this long-term effect is partially ROS mediated. (a) The melanoma cell lines (IM) were treated with and without 6 J/cm2 UVA and with and without the ROS quencher Trolox and with subsequent recovery without irradiation for 5 days. After recovery of 5 days’ glucose consumption, lactate production and transketolase activity were measured. (b) Higher glucose consumption, (c) higher lactate production, and (d) higher transketolase activity, induced by UVA, are significantly (two-way ANOVA, Bonferroni multiple comparisons post-test; asterisk represents P < 0.01) elevated 5 days after UVA treatment compared with unirradiated control. Persistence of UVA-induced elevated glucose consumption, lactate production, and transketolase treatment is ameliorated through the addition of ROS quencher Trolox. ANOVA, analysis of variance; IM, initial melanoma; ROS, reactive oxygen species.

(Figure 3d) (mean value and SD of at least three independent experiments and two-way ANOVA, Bonferroni multiple comparisons post-test shown; asterisks represent P < 0.01). This UVA-induced persistent increase in transketolase activity could in part be attenuated by Trolox (Figure 3d). MMP and uPA expression is upregulated after exposure to lactate

Expression of MMPs and uPA is important for melanoma invasion. Lactate concentrations of 11 mM are described in tumors in vivo (Serganova et al., 2011), and lactate levels of approximately 16 mM were measured on UVA irradiation of melanoma cells in vitro (Figure 1b and d). To determine whether lactic acid alone, a major derivate of the UVA-induced Warburg effect, is able to induce expression of MMPs and uPA in melanoma cells, IM cells were treated with 20 mM lactic acid for 12 hours. After lactic acid treatment, we analyzed transcription of MMP1, MMP2, MMP3, MMP9, MMP13, MMP15, tissue inhibitor of metalloproteinase (TIMP1), and uPA. Lactic acid increased the transcription of all investigated proteases (MMPs and uPA) relevant for tumor progression (Figure 4a and Supplementary Figure S3a and b online). In addition, we found a slight induction of TIMP1 in IM melanoma cells visible, but compared with the induction of transcription levels of most MMPs and uPA, these effects of lactate on TIMP1 transcription were negligible (Figure 4a) (transcriptional level of MMP, uPA, TIMP1 relative to housekeeping gene beta actin shown in 2DDCt in logarithmic scale; data presented as the mean value with SD of at least four independent experiments). To show that either UVA irradiation or UVA irradiation-associated lactic acid is capable of inducing secretion of active tumor, relevant MMP2 and MMP9 zymography was performed. UVA irradiation induced secretion of active MMP2 and MMP9 in IM and MM cells and UVA-induced activity of MMP2 and MMP9 was slightly reduced in the presence of ROS quencher, indicating a partially ROS-mediated mechanism of

UVA-induced MMP2 and MMP9 induction (Figure 4b and c, Supplementary Figure S3c and d) (IM and MM treated with and without UVA or 20 mM lactic acid; Student’s t-test; asterisk represents P < 0.05; data shown as the mean value with SD of at least three independent experiments). Furthermore, treatment of lactic acid alone was also capable of facilitating secretion of MMP2 and MMP9 when compared with control (Figure 4b and c, Supplementary Figure S3c and d). Additionally, UVA irradiation enhanced secretion of uPA in IM and MM cells (Supplementary Figure S4c and d online). The UVA-enhanced Warburg effect increases invasion via production of lactic acid

To test whether production of lactic acid, which can be induced by UVA-mediated ROS, is functionally capable of enhancing invasion of IM and MM cells, we performed in vitro invasion assays. In these assays, IM and MM cells were grown on an artificial collagen matrix where they invade this matrix according to their invasive potential. Melanoma cells were treated with and without UVA irradiation (6 J/cm2 for 4 days), ROS quencher Trolox, glycolytic inhibitor 2DG, or lactic acid (Figure 4d and e) (in vitro invasion assay of melanoma cells [IM and MM] treated with and without UVA, 20 mM lactic acid, ROS quencher Trolox, and 2DG; one-way ANOVA with Newman Keuls multiple comparison post-test; asterisk represents P < 0.05). Data are presented as the mean value with SD of at least three independent experiments. In the invasion assay, UVA irradiation significantly increased invasion of melanoma cells, and this effect was partially ROS mediated as Trolox significantly attenuated UVA-induced invasion (Figure 4d and e). Similarly, the inhibitor of glycolysis (2DG) significantly attenuated UVA-induced invasion in IM and MM cells (Figure 4d and e). Treatment of melanoma cells with lactic acid significantly increased invasion to an extent comparable to UVA irradiation (Figure 4d and e). These data show that the addition of lactic acid can mimic the UVA-induced increase of the www.jidonline.org

5

Y Kamenisch et al.

Ultraviolet A and Melanoma Figure 4. Increased UVA-associated lactic acid enhances invasion. (a) Lactic acid enhances transcription of MMP and uPA in melanoma cell lines (IM). (b, c) UVA irradiation or lactic acid alone increases MMP2 and MMP9 secretion. (d, e) UVA irradiation or lactic acid alone increased invasion, whereas the addition of ROS quencher Trolox or inhibitor of glycolysis 2DG previous to irradiation inhibited UVA-induced invasive potential. (f) In human skin reconstructs with epidermal melanoma cells (IM and MM), UVA induces more initial invasion of melanoma cells in the dermis than in control, detected with melanomaspecific S100 staining and counted per field of vision. Scale bar represents 70 mm. (g) Proposed model for UVAenhanced invasion via production of ROS, lactic acid, and expression of MMP and uPA. 2DG, 2-deoxy-Dglucose; IM, initial melanoma; MM, metastasizing melanoma; MMP, matrix metalloproteinase; ROS, reactive oxygen species; uPA, urokinase-type plasminogen activator.

6


-

Y Kamenisch et al.


invasive potential in vitro. To investigate to which extent the UVA- or lactic acid-induced invasion is facilitated either by MMP proteases or by uPA proteases we tested the invasive potential of UVA-treated cells with and without inhibitors of MMP and uPA. The presence of MMP inhibitor (GM6001) as well as the presence of uPA inhibitor (amiloride) partially abrogated UVA-induced invasive potential (Supplementary Figure S4a and b), indicating that MMP and uPA are functionally needed for UVA-induced in vitro invasion. Furthermore, UVA irradiation of skin reconstructs with melanoma cells (IM and MM) enhanced expression of MMP9 and uPA (Supplementary Figure S4e and f). In addition to this, UVA irradiation increased initial dermal invasion of melanoma cells (IM and MM) in skin reconstructs (Figure 4f) (mean value with SD of at least three independent experiments shown; Student’s t-test; asterisk represents P < 0.05), which was verified by a histopathologist. DISCUSSION In many patients, initial small melanomas are not discovered and patients continue solar exposure and many patients even increase their sun exposure after diagnosis of initial melanoma (Idorn et al., 2014). The importance of single neonatal high doses of UVB or UVA radiation for initiation of melanoma was shown previously in mice (Noonan et al., 2003, 2012). But despite its physiological relevance, it is still not clear how UVA irradiation influences progression of human initial melanoma cells. The dose of UVA irradiation we used is within the physiological range to which an individual can readily be exposed to during a summer day in Europe (http:// www.soda-is.com/eng/map/maps_for_free.html#uv). Here we demonstrate that UVA induces the tumor-associated Warburg effect. The literature indicates that tumor cells benefit from this Warburg effect as part of the glycolysisderived pyruvate can be used for further anabolic pathways such as amino acid synthesis or fatty acid synthesis (Vander Heiden et al., 2009). Although UVA irradiation decreased survival of the tumor cells in a dose-dependent way (Supplementary Figure S1a and b), the remaining viable cells have a higher rate of glucose consumption and lactate production per cell than unirradiated controls (Figure 1aef, Supplementary Figure S2a and b). These metabolic changes seem to be partially mediated by ROS because the addition of ROS quenchers was able to prevent a UVA-induced increase in the Warburg effect (Figure 1i and j). Two immediate effects of UVA irradiation, activation of epidermal growth factor receptor and production of ROS, are well known (El-Abaseri et al., 2006; Lopez-Camarillo et al., 2012). Activation of epidermal growth factor receptor leads to activation of pyruvate kinase M2, which then induces the expression of glycolytic enzymes, leading to an increased Warburg effect (Yang et al., 2012). Recently, it was shown that ROS via induction of hypoxia inducible factor can lead to metabolic reprogramming toward the Warburg effect (Haigis et al., 2012). Instead of ATP generation, the pentose phosphate pathway is the main source of riboses needed for nucleoside synthesis (Boros et al., 1997) and nicotinamide adenine dinucleotide phosphate, and transketolase as the key enzyme of this pathway is used as a prognostic marker for cancer therapy (Langbein et al., 2006). So both pentose phosphate

pathway and glycolysis enhanced by the Warburg effect are helpful for proliferating cells. Interestingly, the UVA-induced Warburg effect persists for 5 days after cessation of the last UVA irradiation (Figure 3b and c), indicating a long-lasting change in the metabolism of tumor cells that goes beyond the time typically needed to repair UVA-induced DNA damage. Although lactic acid is detrimental for cellular viability, high levels of lactic acid are angiogenic (Dhup et al., 2012) and immunosuppressive via inhibition of tumor necrosis factor secretion in monocytes (Dietl et al., 2010). In addition, macrophages treated with lactic acid inhibited activation of CD8þ T cells (Ohashi et al., 2013). Here we show that high levels of lactic acid induce transcription and secretion of MMPs and uPA (Figure 4aec, Supplementary Figures S3a and b and S4c), which are important for tumor invasion. The effect of lactic acid on MMP2 and MMP9 and uPA secretion is stronger than the effect of UVA in melanoma cells (Figure 4b and c, Supplementary Figure S4c), which indicates that other mechanisms than UVA may be responsible for MMP2 and MMP9 and uPA secretion in melanoma cells. Additionally, it is known that UVA induces MMPs in dermal cells. Furthermore, it was previously shown in other tumors that lactic acid is capable of increasing tumor cell migration by activation of MMP2 (Baumann et al., 2009). In our hands, initial melanoma cells show a fast transcriptional upregulation of MMP2, MMP3, MMP9, MMP13, MMP15 as well as uPA, compared with untreated control, 12 hours after the addition of lactic acid in physiological concentrations. In contrast to this, no significant upregulation of the inhibitors of MMPs, TIMP1, can be found. It has been shown that induction of MMP 2, MMP3, MMP9, MMP13, and uPA proteases is important during melanoma progression (Bianchini et al., 2006; Corte et al., 2005; Girouard et al., 2012; Ji et al., 2013; Orimoto et al., 2008; Rotte et al., 2012; Tang et al., 2013; Tas et al., 2005, 2008; Zigrino et al., 2009). So far it has not been reported that MMP15 transcription is induced in melanoma. Lactic acid can also be produced in hypoxic conditions (Pasteur effect), which is often observed in melanoma (Scott et al., 2011) and many other tumors. Interestingly, lactic acid-induced expression of MMPs and invasion in distant areas could be one efficient strategy of melanoma cells to escape from unfavorable areas with hypoxia, high lactic acid concentrations, and senescence inducing conditions (Mo et al., 2013). Finally, the finding that in matrigel-based assays, UVA irradiation-induced lactic acid increases invasion of melanoma cells provides a functional link between melanoma invasion and the UVA-induced Warburg effect (Figure 4d and e, Supplementary Figure S3e). Furthermore, we could show that uPA and MMP activity are functionally important for UVA-induced invasion (Supplementary Figure S4a and b). UVA also induced the Warburg effect and invasion in melanoma cells of metastasizing melanomas, although to a smaller extent than in melanoma cells of initial melanoma. This is in line with the notion that UVA irradiation is of particular relevance at an early stage of metastasis (IM cells) rather than for cells that derive from a tumor that already did metastasize. Furthermore, UVA irradiation may be more relevant to IM that are in the upper layers of the skin and are usually in the penetration range of physiological UVA www.jidonline.org

7

Y Kamenisch et al.


irradiation but not distant lymph nodes or other organs, which are targets for metastasizing melanoma cells. With regard to the role of oxidative stress, it is important to note that in our system, ROS quenchers cannot completely abrogate UVA-mediated effects, indicating that either cumulative ROS production during UVA treatment exceeds the ROS detoxification capacity of Trolox or other mechanisms such as activation of mitogen-activated protein kinase, NF-kB, ß-selectin, or chemokine signaling may also be relevant. Taken together, at least partially, UVA irradiation enhances the Warburg effect in an ROS-dependent fashion with increased lactate production, which induces MMP and uPA transcription, finally leading to invasion of tumor cells, which is also visible in skin reconstructs. This indicates that UVA irradiation is not only important in initiation of melanoma, but also during progression. These findings support therapeutic strategies of using MMP inhibitors together with antioxidants and inhibitors of glycolysis. Furthermore, our data support the therapeutic recommendation of continuous sun protection.

Cell culture Human melanoma cells of different progression levels (vertical growth phase, radial growth phase, and MM) (Supplementary Table S1) were cultured in DMEM as described previously (Berneburg et al., 2005; Kamenisch et al., 2010; Koch et al., 2001) and in Supplementary Materials and Methods online. Patient consent for experiments was not required for experiments with established cell lines.

Skin reconstructs Preparation of skin reconstructs and S100 staining was performed as described previously (Meier et al., 2000). Briefly, layers of collagen, fibroblasts, keratinocytes, and different melanoma cells were cultivated and treated with and without UVA irradiation as described in Supplementary Materials and Methods.

UVA irradiation Irradiation was carried out as described (Berneburg et al., 2005; Kamenisch et al., 2010; Koch et al., 2001). Briefly, cells were irradiated with a sellasol (SELLAS, Ennepetal, Germany) UVA irradiation device as described in Supplementary Materials and Methods.

Statistical analysis Unless mentioned otherwise, data are shown as the mean with SD of at least three independent experiments, and statistical significance was tested with paired and unpaired Student’s t-test, two-way ANOVA with Bonferroni multiple comparisons after test analysis, one-way ANOVA with post-test for a linear trend and one-way ANOVA with Newman Keuls multiple comparison post-test analysis, and a P value < 0.05 was considered significant.

Chemical treatment Cells were treated as described in Supplementary Materials and Methods.

Lactate and glucose measurement The same number of melanoma cells were treated with and without 4 days of repetitive UVA irradiation and aliquots of the supernatant of the treated cells were used for lactate and glucose measurement Journal of Investigative Dermatology (2016), Volume

Protein measurement Protein measurements were performed using the BCA Protein Assay Reagent A and B (Pierce, Rockford, IL).

Transketolase measurement Transketolase activity was measured colorimetrically (Smith et al., 2006) as described in Supplementary Materials and Methods.

Quantitative analysis of MMP and uPA transcription via realtime quantitative reverse transcription-PCR After treatment with lactic acid, RNA was isolated out of the cells according to protocol (Nucleospin RNA protocol) and cDNA was made by reverse transcriptase PCR according to protocol (BIO-RAD, Germany). Real-time PCR was performed on a Roche LightCycler using SYBR Green Supermix (BIO-RAD, Germany). Primers are listed in Supplementary Table S2 online. Subsequent to real-time PCR, products were verified on gel by gel electrophoresis with marker (PEQ gold Ultra Low Range DNA Ladder I).

Zymography For zymography the supernatant of IM and MM cells treated with UVA, Trolox, or lactic acid was used and zymography was performed as described in Supplementary Materials and Methods.

MATERIALS AND METHODS

8

and normalized to the number of cells after treatment as described in Supplementary Materials and Methods.

-

Assay of uPA For quantitative detection of human uPA, the uPA assay according to the protocol of uPA (URK) Human ELISA Kit (abcam, Germany) was performed as described in Supplementary Materials and Methods.

In vitro cell invasion assays Cell invasion assays were performed according to protocols of Millipore Cell invasion Assay protocol (Millipore, Germany). Briefly, 250,000 melanoma cells were incubated per insert for 24 hours with and without treatment, and the fraction of invaded cells was determined fluorometrically with a fluorescence plate reader using a 480/520 nm filter set. CONFLICT OF INTEREST The authors state no conflict of interest except the conflict of interest of MR Rocken. He has received grants (or other support as listed) from AB Science (Study), Abbott Laboratories (Study), Abbott Pharmaceuticals (Study), Almirall Hermal (Grant/Honorarium; Study and consulting), AstraZeneca (Study), Bayer (Study) Biogen Idec (Grant/Honorarium; Study, consulting and speaker), Bristol-Myers Squibb (Study), Bundesministerium fu¨r Bildung und Forschung (Research project), Celgene (Study), Deutsche Dermatologische Gesellschaft (Officer), Deutsche Forschungsgemeinschaft (Research project), Deutsche Krebshilfe (Research project), European Union (Research project), European Academy of Dermatology and Venereology (Officer), Galderma (Grant/ Honorarium; Study and consulting), GSK (Study), Hokusai (Study), JanssenCilag (Grant/Honorarium; Study and consulting), Johnson & Johnson (Grant/ Honorarium; Study and consulting), Lilly (Study), Merck (Study), MSD Sharp & Dohme (Study), Novartis (Study), Pfizer (Study), Philogen (Study), Regeneron (Honorarium; Consulting), Roche (Study), sterna biologicals (Honorarium; Speaker), Wilhelm Sander-Stiftung (Research project).

ACKNOWLEDGMENTS The authors thank Nurgu¨l Du¨zenli, Evelyn Maczey, and Eva Mu¨ller-Hermelink for their excellent technical assistance. This work was partially funded by Deutsche Krebshilfe 106223 (Die Rolle UV induzierter mitochondrialer DNS Mutationen und deren Reparatur in der Pathogenese des malignen Melanoms: Untersuchungen zum Vorliegen und zur funktionellen Relevanz in vitro und in vivo). This work was partially funded by Ro DFG 764/14-1 REDOX and Wilhelm Sander Stiftung 2012.056.2 (Differenzierung). This work was partially funded by Kurdistan Regional Government Ministry of Higher Education and Scientific Research; Deutsche Krebshilfe 110210 and the Melanoma Research Network; and the Deutsche Forschungsgemeinschaft (DFG) Klinische Forschergruppe KFO 262, Project 11.

Y Kamenisch et al.

Ultraviolet A and Melanoma SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at www. jidonline.org, and at http://dx.doi.org/10.1016/j.jid.2016.02.815.

REFERENCES Backvall H, Wassberg C, Berne B, Ponten F. Similar UV responses are seen in a skin organ culture as in human skin in vivo. Exp Dermatol 2002;11: 349e56. Baczynska KA, Pearson AJ, O’Hagan JB, Heydenreich J. Effect of altitude on solar UVR and spectral and spatial variations of UV irradiances measured inWagrain, Austria in winter. Radiat Prot Dosimetry 2013;154: 497e504. Bald T, Quast T, Landsberg J, Rogava M, Glodde N, Lopez-Ramos D, et al. Ultraviolet-radiation-induced inflammation promotes angiotropism and metastasis in melanoma. Nature 2014;507:109e13. Baumann F, Leukel P, Doerfelt A, Beier CP, Dettmer K, Oefner PJ, et al. Lactate promotes glioma migration by TGF-beta2-dependent regulation of matrix metalloproteinase-2. Neuro Oncol 2009;11:368e80. Beissert S, Loser K. Molecular and cellular mechanisms of photocarcinogenesis. Photochem Photobiol 2008;84:29e34. Berneburg M, Gremmel T, Kurten V, Schroeder P, Hertel I, von Mikecz A, et al. Creatine supplementation normalizes mutagenesis of mitochondrial DNA as well as functional consequences. J Invest Dermatol 2005;125: 213e20. Bianchini F, D’Alessio S, Fibbi G, Del Rosso M, Calorini L. Cytokinedependent invasiveness in B16 murine melanoma cells: role of uPA system and MMP-9. Oncol Rep 2006;15:709e14. Boros LG, Puigjaner J, Cascante M, Lee WN, Brandes JL, Bassilian S, et al. Oxythiamine and dehydroepiandrosterone inhibit the nonoxidative synthesis of ribose and tumor cell proliferation. Cancer Res 1997;57: 4242e8. Corte MD, Gonzalez LO, Corte MG, Quintela I, Pidal I, Bongera M, et al. Collagenase-3 (MMP-13) expression in cutaneous malignant melanoma. Int J Biol Markers 2005;20:242e8.

Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin 2010;60:277e300. Jhappan C, Noonan FP, Merlino G. Ultraviolet radiation and cutaneous malignant melanoma. Oncogene 2003;22:3099e112. Ji MK, Shi Y, Xu JW, Lin X, Lin JY. Recombinant snake venom metalloproteinase inhibitor BJ46A inhibits invasion and metastasis of B16F10 and MHCC97H cells through reductions of matrix metalloproteinases 2 and 9 activities. Anticancer Drugs 2013;24:461e72. Kamenisch Y, Fousteri M, Knoch J, von Thaler AK, Fehrenbacher B, Kato H, et al. Proteins of nucleotide and base excision repair pathways interact in mitochondria to protect from loss of subcutaneous fat, a hallmark of aging. J Exp Med 2010;207:379e90. Kappes UP, Luo D, Potter M, Schulmeister K, Runger TM. Short- and longwave UV light (UVB and UVA) induce similar mutations in human skin cells. J Invest Dermatol 2006;126:667e75. Koch H, Wittern KP, Bergemann J. In human keratinocytes the Common Deletion reflects donor variabilities rather than chronologic aging and can be induced by ultraviolet A irradiation. J Invest Dermatol 2001;117: 892e7. Langbein S, Zerilli M, Zur Hausen A, Staiger W, Rensch-Boschert K, Lukan N, et al. Expression of transketolase TKTL1 predicts colon and urothelial cancer patient survival: Warburg effect reinterpreted. Br J Cancer 2006;94: 578e85. Leiter U, Eigentler T, Garbe C. Epidemiology of skin cancer. Adv Exp Med Biol 2014;810:120e40. Levell NJ, Beattie CC, Shuster S, Greenberg DC. Melanoma epidemic: a midsummer night’s dream? Br J Dermatol 2009;161:630e4. Lopez-Camarillo C, Ocampo EA, Casamichana ML, Perez-Plasencia C, Alvarez-Sanchez E, Marchat LA. Protein kinases and transcription factors activation in response to uv-radiation of skin: implications for carcinogenesis. Int J Mol Sci 2012;13:142e72. Majewski N, Nogueira V, Robey RB, Hay N. Akt inhibits apoptosis downstream of BID cleavage via a glucose-dependent mechanism involving mitochondrial hexokinases. Mol Cell Biol 2004;24:730e40.

De Fabo EC, Noonan FP, Fears T, Merlino G. Ultraviolet B but not ultraviolet A radiation initiates melanoma. Cancer Res 2004;64:6372e6.

Meewes C, Brenneisen P, Wenk J, Kuhr L, Ma W, Alikoski J, et al. Adaptive antioxidant response protects dermal fibroblasts from UVA-induced phototoxicity. Free Radic Biol Med 2001;30:238e47.

Dhawan P, Singh AB, Ellis DL, Richmond A. Constitutive activation of Akt/ protein kinase B in melanoma leads to up-regulation of nuclear factorkappaB and tumor progression. Cancer Res 2002;62:7335e42.

Meier F, Nesbit M, Hsu MY, Martin B, Van Belle P, Elder DE, et al. Human melanoma progression in skin reconstructs : biological significance of bFGF. Am J Pathol 2000;156:193e200.

Dhup S, Dadhich RK, Porporato PE, Sonveaux P. Multiple biological activities of lactic acid in cancer: influences on tumor growth, angiogenesis and metastasis. Curr Pharma Des 2012;18:1319e30.

Mitchell DL, Fernandez AA, Nairn RS, Garcia R, Paniker L, Trono D, et al. Ultraviolet A does not induce melanomas in a Xiphophorus hybrid fish model. Proc Natl Acad Sci U S A 2010;107:9329e34.

Diaz-Moralli S, Tarrado-Castellarnau M, Alenda C, Castells A, Cascante M. Transketolase-like 1 expression is modulated during colorectal cancer progression and metastasis formation. PLoS One 2011;6:e25323.

Mo J, Sun B, Zhao X, Gu Q, Dong X, Liu Z, et al. Hypoxia-induced senescence contributes to the regulation of microenvironment in melanomas. Pathol Res Pract 2013;209:640e7.

Dietl K, Renner K, Dettmer K, Timischl B, Eberhart K, Dorn C, et al. Lactic acid and acidification inhibit TNF secretion and glycolysis of human monocytes. J Immunol 2010;184:1200e9.

Newton-Bishop JA, Chang YM, Elliott F, Chan M, Leake S, Karpavicius B, et al. Relationship between sun exposure and melanoma risk for tumours in different body sites in a large case-control study in a temperate climate. Eur J Cancer 2011;47:732e41.

El-Abaseri TB, Putta S, Hansen LA. Ultraviolet irradiation induces keratinocyte proliferation and epidermal hyperplasia through the activation of the epidermal growth factor receptor. Carcinogenesis 2006;27:225e31. Elstrom RL, Bauer DE, Buzzai M, Karnauskas R, Harris MH, Plas DR, et al. Akt stimulates aerobic glycolysis in cancer cells. Cancer Res 2004;64:3892e9. Fears TR, Bird CC, Guerry Dt, Sagebiel RW, Gail MH, Elder DE, et al. Average midrange ultraviolet radiation flux and time outdoors predict melanoma risk. Cancer Res 2002;62:3992e6. Girouard SD, Laga AC, Mihm MC, Scolyer RA, Thompson JF, Zhan Q, et al. SOX2 contributes to melanoma cell invasion. Lab Invest 2012;92:362e70. Haigis MC, Deng CX, Finley LW, Kim HS, Gius D. SIRT3 is a mitochondrial tumor suppressor: a scientific tale that connects aberrant cellular ROS, the Warburg effect, and carcinogenesis. Cancer Res 2012;72:2468e72. Howe HL, Wingo PA, Thun MJ, Ries LA, Rosenberg HM, Feigal EG, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 2001;93: 824e42. Idorn LW, Datta P, Heydenreich J, Philipsen PA, Wulf HC. A 3-year follow-up of sun behavior in patients with cutaneous malignant melanoma. JAMA Dermatol 2014;150:163e8.

Noonan FP, Dudek J, Merlino G, De Fabo EC. Animal models of melanoma: an HGF/SF transgenic mouse model may facilitate experimental access to UV initiating events. Pigment Cell Res 2003;16:16e25. Noonan FP, Recio JA, Takayama H, Duray P, Anver MR, Rush WL, et al. Neonatal sunburn and melanoma in mice. Nature 2001;413:271e2. Noonan FP, Zaidi MR, Wolnicka-Glubisz A, Anver MR, Bahn J, Wielgus A, et al. Melanoma induction by ultraviolet A but not ultraviolet B radiation requires melanin pigment. Nat Commun 2012;3:884. Ohashi T, Akazawa T, Aoki M, Kuze B, Mizuta K, Ito Y, et al. Dichloroacetate improves immune dysfunction caused by tumor-secreted lactic acid and increases antitumor immunoreactivity. Int J Cancer 2013;133:1107e18. Orimoto AM, Neto CF, Pimentel ER, Sanches JA, Sotto MN, Akaishi E, et al. High numbers of human skin cancers express MMP2 and several integrin genes. J Cutan Pathol 2008;35:285e91. Parisi AV, Wong JC. An estimation of biological hazards due to solar radiation. J Photochem Photobiol B 2000;54:126e30. Resendis-Antonio O, Checa A, Encarnacion S. Modeling core metabolism in cancer cells: surveying the topology underlying the Warburg effect. PLoS One 2010;5:e12383.

www.jidonline.org

9

Y Kamenisch et al.

Ultraviolet A and Melanoma Rotte A, Martinka M, Li G. MMP2 expression is a prognostic marker for primary melanoma patients. Cell Oncol (Dordr) 2012;35:207e16.

Tucker MA, Goldstein AM. Melanoma etiology: where are we? Oncogene 2003;22:3042e52.

Scott DA, Richardson AD, Filipp FV, Knutzen CA, Chiang GG, Ronai ZA, et al. Comparative metabolic flux profiling of melanoma cell lines: beyond the Warburg effect. J Biol Chem 2011;286:42626e34.

Turnbull DJ, Parisi AV. Spectral UV in public shade settings. J Photochem Photobiol B 2003;69:13e9.

Serganova I, Rizwan A, Ni X, Thakur SB, Vider J, Russell J, et al. Metabolic imaging: a link between lactate dehydrogenase A, lactate, and tumor phenotype. Clin Cancer Res 2011;17:6250e61. Smith ME, Kaulmann U, Ward JM, Hailes HC. A colorimetric assay for screening transketolase activity. Bioorg Med Chem 2006;14:7062e5. Tang ZY, Liu Y, Liu LX, Ding XY, Zhang H, Fang LQ. RNAi-mediated MMP-9 silencing inhibits mouse melanoma cell invasion and migration in vitro and in vivo. Cell Biol Int 2013;37:849e54. Tas F, Duranyildiz D, Oguz H, Camlica H, Yasasever V, Topuz E. Circulating levels of vascular endothelial growth factor (VEGF), matrix metalloproteinase-3 (MMP-3), and BCL-2 in malignant melanoma. Med Oncol 2008;25:431e6. Tas F, Duranyildiz D, Oguz H, Disci R, Kurul S, Yasasever V, et al. Serum matrix metalloproteinase-3 and tissue inhibitor of metalloproteinase-1 in patients with malignant melanoma. Med Oncol 2005;22:39e44.

10


-

Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 2009;324:1029e33. Volker HU, Hagemann C, Coy J, Wittig R, Sommer S, Stojic J, et al. Expression of transketolase-like 1 and activation of Akt in grade IV glioblastomas compared with grades II and III astrocytic gliomas. Am J Clin Pathol 2008;130:50e7. Warburg O, Wind F, Negerlein E. The metabolism of tumors in the body; 1927. Yang W, Zheng Y, Xia Y, Ji H, Chen X, Guo F, et al. ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect. Nat Cell Biol 2012;14:1295e304. Zaidi MR, Davis S, Noonan FP, Graff-Cherry C, Hawley TS, Walker RL, et al. Interferon-gamma links ultraviolet radiation to melanomagenesis in mice. Nature 2011;469:548e53. Zigrino P, Kuhn I, Bauerle T, Zamek J, Fox JW, Neumann S, et al. Stromal expression of MMP-13 is required for melanoma invasion and metastasis. J Invest Dermatol 2009;129:2686e93.