Anticancer activity of cationic porphyrins in melanoma tumour-bearing mice and mechanistic in vitro studies

Background Porphyrin TMPyP4 (P4) and its C14H28-alkyl derivative (C14) are G-quadruplex binders and singlet oxygen (1O2) generators. In contrast, TMPyP2 (P2) produces 1O2 but it is not a G-quadruplex binder. As their photosensitizing activity is currently undefined, we report in this study their efficacy against a melanoma skin tumour and describe an in vitro mechanistic study which gives insights into their anticancer activity. Methods Uptake and antiproliferative activity of photoactivated P2, P4 and C14 have been investigated in murine melanoma B78-H1 cells by FACS, clonogenic and migration assays. Apoptosis was investigated by PARP-1 cleavage and annexin-propidium iodide assays. Biodistribution and in vivo anticancer activity were tested in melanoma tumour-bearing mice. Porphyrin binding and photocleavage of G-rich mRNA regions were investigated by electrophoresis and RT-PCR. Porphyrin effect on ERK pathway was explored by Western blots. Results Thanks to its higher lipophylicity C14 was taken up by murine melanoma B78-H1 cells up to 30-fold more efficiently than P4. When photoactivated (7.2 J/cm2) in B78-H1 melanoma cells, P4 and C14, but not control P2, caused a strong inhibition of metabolic activity, clonogenic growth and cell migration. Biodistribution studies on melanoma tumour-bearing mice showed that P4 and C14 localize in the tumour. Upon irradiation (660 nm, 193 J/cm2), P4 and C14 retarded tumour growth and increased the median survival time of the treated mice by ~50% (P <0.01 by ANOVA), whereas porphyrin P2 did not. The light-dependent mechanism mediated by P4 and C14 is likely due to the binding to and photocleavage of G-rich quadruplex-forming sequences within the 5′-untranslated regions of the mitogenic ras genes. This causes a decrease of RAS protein and inhibition of downstream ERK pathway, which stimulates proliferation. Annexin V/propidium iodide and PARP-1 cleavage assays showed that the porphyrins arrested tumour growth by apoptosis and necrosis. C14 also showed an intrinsic light-independent anticancer activity, as recently reported for G4-RNA binders. Conclusions Porphyrins P4 and C14 impair the clonogenic growth and migration of B78-H1 melanoma cells and inhibit melanoma tumour growth in vivo. Evidence is provided that C14 acts through light-dependent (mRNA photocleavage) and light-independent (translation inhibition) mechanisms.


Background
Photodynamic therapy (PDT) is a rapidly expanding therapeutic modality for the treatment of a number of diseases including cancer [1][2][3]. PDT employs a photosensitizer which, upon irradiation, produces singlet oxygen ( 1 O 2 ) that damages cells [4,5]. One major objective of PDT is the search of new photosensitizers with high water solubility, low dark cytotoxycity, high capacity to penetrate the plasma membrane and generate 1 O 2 , and an ability to interact with specific cellular targets [6][7][8].
In previous studies from our laboratory we synthesized expanded porphyrins, composed of a macrocycle of five pyrroles, that exhibit a photodynamic activity in cancer cells at micromolar concentrations, either as free molecules or complexed to Zn or Lu [9,10]. In addition, we examined squaraines [11] and pheophorbide a [12][13][14][15], a chlorophyll derivative with a tetrapyrrolic macrocycle which is active at a nanomolar concentration range, comparable to that of verteporfin and temoporfin: two well-known photosensitizers used in clinic [16,17]. However, like many other photosensitizers, pheophorbide a is not very soluble in water, and this reduces PDT efficacy. A group of potential photosensitizers that are completely soluble in water are the cationic porphyrins TMPyP2 (called P2), with four 2-methypyridyl groups, and TMPyP4 (called P4), with four 4-methylpyridyl groups ( Figure 1A). Porphyrin P4 has been extensively studied for its capacity to bind to an unusual nucleic acid conformation called G-quadruplex, formed by DNA and RNA sequences composed of blocks of guanines [18][19][20][21]. DNA and RNA Gquadruplexes (G4-DNA and G4-RNA) are stabilized by the stacking upon each other of at least two G-tetrads (a planar arrangement of four guanines each forming two Hoogsteen hydrogen bonds with the neighboring bases) and by alkali metal ions (Na + or K + ) that coordinate to O6 of guanines and lie in the central cavity of the structure [22]. Over the last decade G4-DNA and G4-RNA have attracted interest of several researchers providing growing evidence that these quadruplexes are involved in the regulation of gene expression. Compared to double-stranded DNA, single-stranded RNA is unconstrained and thus prone to fold into complex secondary/tertiary structures containing double-stranded and G-quadruplex elements. A bioinformatic study by Huppert et al. [23] revealed that about 3000 5′-UTR of the human transcriptome contains one or more G-quadruplex motifs. This brought the authors to the conclusion that these tertiary structures should regulate translation. They also demonstrated that an RNA quadruplex-forming sequence within the 5′-UTR of the NRAS transcript inhibited translation in vitro [24]. Ever since, there has been a growing literature on the possible functions of RNA G-quadruplexes [25][26][27]. As G4-RNA can indeed inhibit translation, the use of small molecules to inhibit the function of mRNA looks quite attractive [28].
In our study we focus on bifunctional G4 RNA-interacting agents, i.e. molecules capable not only to bind to RNA G-quadruplexes with a high affinity but also to generate singlet oxygen ( 1 O 2 ) upon irradiation. In most human cancers the RAS-MEK-ERK pathway that controls proliferation is hyperactive [29,30]. The activator of this pathway is protein RAS which is expressed by the ras genes (KRAS, NRAS and HRAS). These genes are characterized by the presence in the 5′ end of mRNA of a guanine rich untranslated region (5′-UTR) containing quadruplexforming motifs which can serve as targets for small molecules with photosensitizing properties. Of particular interest is KRAS, as it contains a 192-nt 5′-UTR with 45% guanines that can fold into a cluster of G4-RNA structures and is in principle capable of providing multiple binding sites for cationic porphyrins [18][19][20][21]. The 5′-UTR of NRAS also harbors a G-rich motif forming a very stable G-quadruplex structure.
Against this background we have hypothesized a new strategy to down-regulate in cancer cells the mitogenic RAS/MEK/ERK pathway [31]. We reasoned that the porphyrins delivered to the cells should bind to mRNA, in particular at G-rich quadruplex-forming sequences in the 5′-UTRs of the ras genes. Upon irradiation, the porphyrins mediate a photoprocess leading to the degradation of mRNA, inhibition of ERK pathway and cell proliferation. This therapeutic approach has been tested in murine amelanotic melanoma B78-H1 cells both in vitro (cell cultures) and in vivo (B78-H1 melanoma cells transplanted in C57/BL6 mice). B78-H1 melanoma cells derive from the B16 clone and have an activated RAS/MEK/ERK pathway [32]. As bifunctional photosensitizing agents we tested the cationic porphyrins tetra-meso (N-methyl-4-pyridyl) porphine (called P4) and C14-alkyl derivative tri-meso(N-methyl-4-pyridyl), meso(N-tetradecyl-4-pyridyl) porphine (called C14). As a control, we used the positional isomer tetra-meso(N-methyl-2-pyridyl) porphine (called P2), which does not bind to G4-RNA. In the following we demonstrate that the cationic porphyrins C14 and P4 strongly inhibit the growth of melanoma cells both in vitro and in vivo. This anticancer effect correlates with the capacity of these porphyrins to bind to the G-rich region of mitogenic ras genes.

Results and discussion
Singlet oxygen generation by porphyrins C14, P4 and P2 The structures of porphyrins C14, P4 and P2 are shown in Figure 1A. Due to their cationic charges they are very soluble in aqueous solutions and obey the Lambert-Beer law over a wide concentration range (not shown). Their capacity to generate 1 O 2 was examined by the 9,10-dimethylantracene (DMA) photobleaching assay [33]. DMA is a fluorescent dye [λ ex = 375 nm, λ em = 436 nm] that reacts selectively with 1 O 2 to form a non-fluorescent endoperoxide derivative. An equimolar solution of DMA and porphyrin (10 μM each) was irradiated with a lamp (60 mW) for different periods of time up to 1000 s and the residual fluorescence was recorded between 380 and 550 nm. The residual fluorescence was plotted as a function of time and the experimental curve was best-fitted to a first order decay equation y = exp (-k⋅t) where k (s -1 ) is the rate constant and t the time(s) ( Figure 1B). We obtained k = 0.0058 ± 0.0004 s -1 for P4 and C14 and k = 0.0021 ± 0.0006 s -1 for P2. Considering that the quantum yield of singlet oxygen generation (ϕΔ) of P4 was reported to be 0.51 [34], we found that the ϕΔ of C14 and P2 were respectively 0.51 and 0.18 ( Figure 1C). It is worth noting that porphyrin P2, being non planar due to steric clashes between the 2-methyl-pyridyl groups and the 3-pyrrol hydrogens, shows a lower ϕΔ.

Cellular uptake and phototoxicity of the cationic porphyrins
The uptake of the cationic porphyrins by melanoma B78-H1 cells has been investigated by FACS, taking advantage of the fact that the porphyrins emit red fluorescence when they are excited at 488 nm. Figure 2A shows a typical FACS analysis of B78-H1 cells treated with 10 μM porphyrins for 2, 4, 8 and 24 h. As observed with human Panc-1 cells [31], C14 is taken up by melanoma cells more efficiently than P2 or P4. After an incubation of 24 h, the fluorescence of the cells treated with C14 is 25 and 50-fold higher than that observed with P4 and P2, respectively. The data show that the addition to P4 of the lipophilic chain significantly enhances the cellular uptake.
The photoactivity of the porphyrins in B78-H1 cells (24 h after delivery) has been evaluated by resazurin assays ( Figure 2B). Without irradiation the porphyrins are not cytotoxic (not shown), but when they are irradiated with a halogen lamp at a fluence of 7.2 J/cm 2 , C14 and P4 decrease the metabolic activity of the cells, at concentrations < 1 μM, while P2 shows some bioactivity only at concentrations > 20 μM. A dose-response assay is shown in Figure 2B. From these plots we estimated IC 50 values of about 10 and 200 nM for C14 and P4 and 30 μM for P2. These values correlate with the different uptake of the three porphyrins.
The impact of the photoactivated porphyrins on the clonogenic growth of B78-H1 melanoma cells was also examined ( Figure 3A). The cells treated with the porphyrins were seeded in plates and after an incubation of 7 days the colonies were stained with methylene blue. The data show that the untreated cells formed colonies uniformly distributed in the plate (control). In contrast, the cells treated with P4 and light (7.2 J/cm 2 ) showed~30% reduction in the number of colonies, while C14/light completely arrested the clonogenic growth. In contrast, porphyrin P2/light did not have any inhibitory effect on colony formation.
As a next step, the inhibition of proliferation and migration of B78-H1 cells caused by the cationic porphyrins was assessed by the scratch-wound healing assay. This assay is based on the assumption that a denuded area created in a plate (80% confluent) will quickly heal thanks to proliferating and migrating cells. But as soon as proliferation and migration are inhibited by the effector molecules, the cells will lose their healing capacity and the area will remain denuded. Figure 3B reports a typical experiment showing that B78-H1 cells untreated or treated with P2 and light are able to heal the wound in 24 h. Instead, when the cells are treated with an IC 50 dose of P4 or C14 and light, they are not able to grow and migrate with a sufficient rate to cover the denuded area.
Having established that photoactivated P4 and C14 arrest proliferation, we asked if this effect was due to apoptosis. An early event occurring in apoptosis is the translocation of phosphatidylserine from the inner to the outer leaflet of the plasma membrane, thus exposing it to the external cell environment. Annexin V, a phosphatidylserine recognizing protein labeled with FITC, can be used to detect this event by FACS. Early apoptosis and late apoptosis/necrosis can be distinguished by using annexin V and propidium iodide (PI) together ( Figure 4A). The percentage of apoptotic/necrotic cells is reported in Table 1. It can be seen that C14 (10 nM) strongly induces apoptosis (43%) and necrosis (38%) in the melanoma B78-H1 cells, whereas P4 (250 nM) appears less efficient in triggering apoptosis (26%) and necrosis (11%). The treatment with P4 produces, however, a larger amount of cell debris than with C14 (18% against 9%). From this experiment we concluded that C14 and P4 promoted in melanoma cells both apoptosis and necrosis and not only necrosis as previously found with fibrosarcoma cells [35]. To confirm the presence of apoptosis in B78-H1 cells treated with the porphyrins, we measured by immunoblotting the cleavage of poly-(ADP ribose)-polymerase (PARP-1) by caspases, as this is considered one of the best markers for apoptosis [36]. In keeping with annexin/PI, Figure 4B shows that P4 and C14 induced the cleavage of PARP-1, while P2 did not. The extent of cleavage is proportional to the percentage of apoptotic cells found by FACS. Indeed, the cells treated with C14 showed 43% apoptosis and~50% PARP-1 cleavage, while those treated with P4 had 26% apoptosis and~10% PARP-1 cleavage.
Given the strong impact on clonogenic growth and mobility caused by P4 and C14 in B78-H1 melanoma cells, we asked whether these molecules are also able to inhibit the growth of a B78-H1 melanoma tumour subcutaneously transplanted in mouse. To address this question, we first had to know how the porphyrins distribute in the mouse body, in order to determine the time point at which the porphyrins show a high accumulation in the tumour after delivery.

Biodistribution of the cationic porhyrins
The time-dependent distribution of P2, P4 and C14 after injection in the peritoneum (i.p.) of female C57/BL6 mice bearing a subcutaneous B78-H1 melanoma tumour of about 6-8 mm was examined as previously described (B) Scratch-wound assay of B78-H1 cells plated at density of 6 × 10 5 cells/well in a 6-well plate. After attachment, the cells were treated with 7.5 nM C14, 250 nM P2 and P4. A denudated area was created across the diameter of the dish with a yellow tip. Cells were washed with PBS and further incubated in complete medium. After light irradiation (7.2 J/cm 2 ), pictures were taken by an epiluminescent microscope (at 10-fold magnification) to evaluate the migration of the cells. The experiment has been perfomed in duplicate. [37]. Porphyrins P2 and P4 were injected (and not treated with light) at a concentration of 30 mg/Kg, while C14, due to its higher cytoxicity, was used at a concentration of 9 mg/Kg. For each molecule three animals were sacrificed at each time point: 1, 3, 6, 9 and 24 h post-injection. The amount of porphyrin recovered from the various organs was measured as described in Materials and Methods and the percentage of injected dose per gram of organ (% ID/g) was determined and reported in Figure 5. At 1, 3 and 6 h post-injection, the porphyrins were detected in all the organs, except in the brain. It is noteworthy that the amount of injected dose in the tumour is higher than in normal tissues, except liver and kidneys. Surprisingly, P4 showed a higher tumour accumulation than C14 (at 9 h after delivery % ID/g P4~2.5 versus % ID/g C14~0.9). It is possible that the alkyl group reduces the capacity of C14 to diffuse from the peritoneus. Indeed, the presence of C14 in the intestine is greater than that of P4, at 3, 6 and 9 h post injection. The amount of P2 in the organs is relatively low, and this correlates with the low cellular uptake of this molecule.
The presence of C14 and P4 in liver, duodenum and blood (not shown) up to 3 h post-injection suggests that the porphyrins are recycled via the enterohepatic pathway, but at higher time points, the urinary output  prevails over the bile-gut recycling. Furthermore, the fact that the porphyrines are practically not found in the spleen, suggests that their elimination through the reticolo-endothelial system is limited, contrarily to what has been observed with more hydrophobic photosensitizers [38].
Porphyrins P4 and C14 reduce tumour growth and increase the mouse median survival time To evaluate in vivo the anticancer property of P2, P4 and C14, C57/BL6 mice with a subcutaneous B78-H1 melanoma tumour of about 6-8 mm were randomized into groups of 8 animals each (the animals were prepared as reported in Materials and Methods). In Figure 6A we report the results obtained with P2 and P4. Group 1 was untreated, groups 2 and 3 were i.p. injected with P2 (30 mg/ Kg) and P4 (30 mg/Kg), respectively, groups 4 and 5 were injected with P2 (30 mg/Kg), P4 (30 Kg/mg) then irradiated at the Q-1 band (620-690 nm) with a diode laser at 660 ± 5 nm (fluence of 193 J/cm 2 ), i.e. in the optical therapeutic window where light is harmless and shows its maximum depth of tissue penetration, as previously described [37,39] (Additional file 1: Figure S1). Irradiation was carried out~7 h after delivery, when the molecules showed a significant accumulation in the tumour. We performed three porphyrin-light treatments,  Figure 6C). The treatment with P2 and P4 without irradiation did not delay tumour growth and did not increase the m.s.t. compared to control ( Figure 6A,C). In contrast, when the tumour was irradiated with the laser, a significant delay in tumour growth was obtained with P4, but not with P2, in keeping with the scarce uptake and photosensitizing property of P2. As a further control, we measured the rate of tumour growth when the mice were irradiated in the absence of porphyrin, but no effect was observed (not shown). The impact of P4 on tumour growth was significant after the second treatment (day 7) compared to both the untreated group (P < 0.001 by ANOVA) and the group treated with P4 without irradiation (P < 0.01). The Kaplan-Maier survival curves showed that treatment with P4 and light increased the m.s.t. by5 0% compared to the untreated group (from 40 to 60 days) or the group treated with P2 and light (from 42 to 60 days) (P < 0.01). In Figure 6B we show the behaviour of C14. Porphyrin C14, at the concentration of 3 mg/Kg, inhibited tumour growth and increased the m.s.t., as does P4 at 30 mg/Kg ( Figure 6B,C). However, contrarily to P4, porphyrin C14 affected somewhat tumour growth, even in the absence of irradiation. After the second PDTtreatment (day 7), tumour growth in the groups treated with C14/dark and C14/light was lower than in the untreated group (P < 0.05 and P < 0.01, respectively). The dual effect of C14 appears more evident when the survival curves are examined. Compared to the untreated group, C14 with and without irradiation increased the median survival time from 40 to 60 and 52 days, respectively (P < 0.01 for both groups). The likely mechanism by which C14-dark reduces tumour growth will be discussed in the next sections.
In summary, photoactivated C14 and P4 showed a remarkable capacity to delay tumour growth in a melanoma mouse model and to increase the survival of the treated mice. Since the cationic porphyrins have a high affinity for RNA G-rich sequences folded into G-quadruplex structures [31], we interrogated whether the antitumour activity of the porphyrins correlates with their binding to RNA.
The cationic porphyrins bind to G-rich quadruplex-forming sequences of mRNA Previous studies have shown that mitochondria are major intracellular targets for hydrophobic photosensitizers, while cationic P4 seems to locate in the lysosomes [35]. So far however, research on PDT did not focus on nucleic acids as targets for photosensitizers. Considering that P4 and C14 accumulate in the cytoplasm and are cationic in nature, they should interact with negatively charged mRNA. There is a vast literature on porphyrin P4's capacity to bind to DNA, in particular at G-rich quadruplexforming sequences occurring in the promoters of the genes and at the ends of the chromosomes [40][41][42][43][44][45][46][47]. It seems therefore reasonable to assume that if a quadruplex structure is extruded by genomic G-rich sequences, these structures should form more easily by unconstrained single-stranded mRNA. We thus hypothesized that cationic porphyrins could bind to G4-RNA structures formed by mRNA. Considering that in all stages of melanoma the key signaling cascade stimulating proliferation is the RAS/ MEK/ERK pathway [48], we focused on the ras genes that encode for protein RAS: the pathway activator [49]. The ras genes show a high homology and their transcripts are composed of six exons of which the exon at the 5′-end is untranslated (5′-UTR) [50]. The 5′-UTR of KRAS has a guanine content of 44% and five quadruplex-forming sequences [31] (Figure 7A). Also NRAS [24] and HRAS include in their 5′-UTR G-rich sequences that can fold into G4-RNA structures (Additional file 2: Figure S2). Quadruplex formation by mRNA sequences can easily be detected by circular dichroism (CD). Typical CD spectra at various temperatures for the KRAS sequence that we called utr-2 are shown in Figure 7B. The spectra clearly show that utr-2 forms a parallel quadruplex with a T M of 54°C in 100 mM KCl [51]. In Table 2 we report the melting data for some quadruplex-forming sequences of the ras genes. These folded RNA structures are strong targets for cationic porphyrins. Indeed, upon binding to the G4-RNA, the porphyrin's Soret band undergoes strong hypochromic and bathochromic effects ( Figure 7C). Analyzing the data with a simple binding equation (Sigma plot 11.0), we found that the interaction between C14 and utr-2 G4-RNA is characterized by a K D of~5 × 10 -7 M.

The porphyrins mediate the photocleavage of mRNA
Next, we asked if mRNA is indeed degraded by the photoactivated porphyrins. To address this question we used three 32 P-radiolabelled RNA fragments adopting one of the possible conformations present in mRNA: G4-RNA, duplex and single-stranded RNA. The three radiolabelled RNA substrates were incubated with  Figure 8A. In the absence of irradiation the porphyrins, even at r = 6, did not promote any degradation (lane 2). Irradiation, in the absence of porphyrin, did not affect RNA either (lane 14). In contrast, when the samples were irradiated in the presence of the porphyrins, a photochemical process leading to RNA degradation took place. The highest cleavage is observed with the G4-RNA substrate (KRAS utr-2 at r = 3 and 6) (lanes 4 and 5), in keeping with the fact that P4 and C14 show a higher affinity for the quadruplexes than the duplex or single-stranded substrates [31].
Finally, we checked if the level of cellular mRNA is actually reduced by the photoactivated porphyrins. By quantitative RT-PCR, we found that 1 h after irradiation P2 did not appreciably decrease the level of ras transcript, while 15 or 30 nM C14 reduced KRAS mRNA by~50% compared to the control (cells tretead with porphyrin but non irradiated). Porphyrin P4 also reduced the mRNA but only at the highest concentration (500 nM) ( Figure 8B). These data are in keeping with those obtained in Panc-1 cells and support our hypothesis that the cationic porphyrins target mRNA [31].

C14 exhibits also an intrinsic light-independent anticancer activity
Several studies have hypothesized that the formation of secondary or tertiary structures within the 5′-UTR of mRNAs may have important functions in the regulation of translation [25]. Since Balasubramanian and coworkers have demonstrated that an RNA-forming sequence located in NRAS mRNA inhibited translation in vitro, there has been a growing interest in G4-RNA in the 5′-UTR of mRNA [24]. To interfere with the translation process, small molecules binding to G4-RNA in 5′-UTR of mRNA have been used [28]. By stabilizing G4-RNA, these ligands should impair the assembly and/or scanning of the 43S ribosomal complex along mRNA. For example, quinolone dicarboxamide derivatives and bisquinolinum compounds showed translational inhibition [52,53]. The finding that C14 without irradiation slows down tumour growth and increases the median survival time by 35% (from 40 to 52 days) may be rationalized in terms of translation inhibition. Due to its high cellular uptake, C14 accumulates in the cytoplasm more than P4. Its binding to G4-RNA in the 5′-UTR of ras mRNA (KRAS and NRAS) may inhibit the translation process. Indeed, when we measured by immunoblotting the level of total protein RAS in B78-H1 cells treated with 1 or 5 μM P2, P4 and C14 (in the absence of light), we found that the protein was significantly reduced by 5 μM C14, but not by P2 and P4 ( Figure 9A). The drop of RAS protein caused by C14 in the dark is likely to be the cause of the observed decline of tumour growth and increase of survival. Further investigation is necessary for more insight into this point.

Photoactivated P4 and C14 inhibit the ERK pathway and NF-kB
Our data showed that photoactivated P4 and C14 inhibit the metabolic activity and the clonogenic growth of melanoma B78-H1 cells by apoptosis and necrosis. We therefore asked how this process is activated. Evidence that the RAS/MEK/ERK signaling pathway plays a central role in melanoma has been reported [54,55]. Protein RAS initiates sequential phosphorylations leading to P-ERK which, in turn, activates downstream proteins enhancing proliferation and survival. About 90% of melanoma tumours carries a high level of P-ERK [55]. Indeed, we found that P-ERK in B78-H1 cells is aberrantly high, though the ras genes are not mutated (Additional file 3: Figure S3). This suggests that the ERK signaling pathway can also be activated by perturbations of components upstream of RAS [56]. Previous studies have demonstrated that P-ERK induces the expression of various cytokines and chemokines that are activators of NF-κB [57,58]. In earlier work we found that in melanoma B78-H1 cells NF-κΒ controls apoptosis via a loop involving anti-apoptotic Snail and pro-apoptotic RKIP [59]. When B78-H1 cells are treated with porphyrin P4 or C14 and then irradiated, the resulting down-regulation RAS expression is accompanied by a significant decrease of the P-ERK and NF-κB ( Figure 9B). A decrease of NF-κB in B78-H1 cells brings about a repression of anti-apoptotic Snail and an activation of pro-apoptotic RKIP [59]. Apoptosis is thus activated, as documented by PARP-1 cleavage and annexin V/PI assays. Considering the role of NF-κB in the tumourigenesis of melanoma, combined treatments with cationic porphyrin/ light and NF-κB inhibitors should be effective against melanoma [60].

Conclusions
In summary, we have documented that tetracationic, mesosubstituted porphyrins P4 and C14 are effective photosensitizers for the photodynamic treatment of melanoma. A previous study reported that P4 and C14 internalize in human fibrosarcoma cells by endocytosis, causing cell death by necrosis [35]. In melanoma B78-H1 cells, photoactivated P4 and C14 behave in a more complex way, as they cause cell death by apoptosis and necrosis. Our data suggest that an intracellular target of the porphyrins is mRNA, in particular the G-rich quadruplexforming sequences located in the 5′-UTR of the mitogenic ras genes. Although we cannot rule out that these molecules can also bind to other gene transcripts, the cationic porphyrins are particularly active in cancer cells showing a RAS/MEK/ERK dependence, such as B78-H1 (Additional file 3: Figure S3) and Panc-1 [31,61] cells. Evidence that the porphyrins tightly bind to the G-rich quadruplexforming sequences in the KRAS transcripts is provided. Due to their capacity to generate singlet oxygen upon irradiation, P4 and C14 photocleave KRAS mRNA, thus reducing the level of protein RAS and the activation of downstream RAS/MEK/ERK and NF-κB pathways. This photodynamic action against the ras transcripts is likely to be responsible of the inhibitory effects on cell growth seen in vitro and on tumour growth seen in vivo. In priciple, this strategy has the advantage of minimizing undesired secondary effects, as the photoprocess mediated by the porphyrins leading to the degradation of mRNA takes place only in the irradiated tumour. The data reported are promising, however, to fully understand the mechanism of tumour suppression in vivo, further studies are necessary. As for the two active porphyrins, we found that C14 is a stronger anticancer drug than P4 as it: (i) internalizes in the melanoma cells more than P4; (ii) inhibits the clonogenic growth more than P4; (iii) acts through a lightdependent mechanism (photocleavage of mRNA) as well as a light-independent mechanism (blockade of translation through the stabilization of G4-RNA at the 5′-UTR of the ras genes). Indeed, C14 at 3 mg/Kg produces in vivo the same effect as P4 does at 30 mg/Kg. By improving the biodistribution of C14, for instance by incorporating the molecule in nanoparticles to exploit their enhanced permeability and retention (EPR) effect, this cationic porphyrin may have a great potential in the cure of melanoma skin tumours.

UV spectroscopy and circular dichroism
UV-vis spectra have been obtained with a Jasco V-530 UV/VIS spectrophotometer. CD spectra have been collected with a JASCO J-600 spectropolarimeter equipped with a thermostatted cell holder. RNA samples in 50 mM Tris-HCl, pH 7.4, 100 mM KCl were 10 μM. The spectra were recorded in 0.5 cm quartz cuvette at increasing temperature. Ordinate is expressed in mdeg.

Cell culture, metabolic activity assay and PDT treatment
In this study we used murine amelanotic melanoma B78-H1 cells. The cells were maintained in exponential growth in Dulbecco's modified eagle's medium containing 100 U/ml penicillin, 100 mg/ml streptomycin, 20 mM L-glutamine and 10% fetal bovine serum (Euroclone, Milano, Italy). The cells were seeded in a 96-well plate at a density of 5 × 10 3 cells/well. The following day they have been treated with the porphyrins in the dark and after 24 h, irradiated with metal halogen lamp with an irradiation of 8 mW/cm 2 for 15 min (fluence 7.2 J/ cm 2 ). 24 h after irradiation the metabolic activity was determined by the resazurin assay following the manufacturer's instructions (Sigma-Aldrich, Milan, Italy). The fluorescence was measured with a spectrofluorometer Spectra Max GeminiXS (Molecular Devices, Sunnyvale, CA). The data are presented as the percentage of metabolic activity compared to untreated cells. The data are the average of at least three independent experiments.
The protocol for PDT was the same for each type of experiment. B78-H1 cells have been plated, the day after the cells have been treated with porhyrins. Then, after 24 h the cells have been irradiated with the halogen lamp (fluence 7.2 J/cm 2 ). As P2 is an isomer of P4 and does not produce any effect on the cells, this molecule was used at the same dose of P4, as a negative control.

Clonogenic assay
B78-H1 cells have been treated with porphyrin/light and seeded in 60 mm Petri dish, at a density of 5 × 10 3 cells. After one week of growth, the colonies were fixed and stained with 2.5% methylene blue in 50% ethanol for 10 min. The images were obtained by Gel DOC 2000 Bio-Rad (Milan, Italy). The stained colonies (>50 cells) have been counted by an Image Scanner equipped with Image Quant TL software (Amersham Biosciences). The experiment was performed in duplicate.

Annexin V-propidium iodide assay
Apoptosis was assessed by annexin V, a protein that binds to phosphatidylserine (PS) residues, which are exposed on the cell surface of apoptotic cells. B78-H1 cells were seeded in a 6-well plate at density of 5 × 10 5 cells/well. After one day, the cells were treated with P2, P4 or C14 for 24 h and irradiated for 15 min with a halogen lamp (7.2 J/cm 2 ). After light activation, cells were washed with PBS, trypsinized, and pelleted. Pellets were suspended in 100 μL Hepes buffer added with 2 μL of annexin V and 2 μL of propidium iodide, PI (annexin-V FLUOS Staining kit, Roche, Penzberg, Germany) and incubated for 10 min at 25°C in the dark. Cells were immediately analyzed by FACS (Becton-Dickinson, San Jose, United States). A minimum of 10 4 cells per sample were acquired in list mode and analyzed using Cell Quest software. The cell population was analyzed by FSC light and SSC light. The signal was detected by FL1 (annexin V-FITC) and FL-2 (PI). The dual parameter dot plots combining annexin V-FITC and PI fluorescence show the vial cell population in the lower left quadrant, the early apoptotic cells in the lower right quadrant, and the late apoptotic or necrotic cells in the upper right quadrant.

The scratch-wound assay
The cells were seeded in a 6-well plate at a density of 6 × 10 5 cells/well and grown for 24 h to 80% confluence.
A denuded area was created across the diameter of the dish by a yellow tip. After treatment with the porphyrin and light, the cells were washed with PBS and further incubated in a complete medium. Pictures were taken with an epiluminescent microscope Leica DMI6000B (Leica Microsystem, Heidelberg, Germany) at a 10-fold magnification to evaluate cell growth and migration [62].

Organ extraction and porphyrin biodistribution in mice with a B78-H1 tumour
The amount of porphyrin in the various organs was determined using the method of Villanueva and Jori [37]. The molecules were solubilized in saline solution and administered by intraperitoneal injection (i.p.) at the mice (P2 and P4, 30 mg/kg; C14, 9 mg/kg). The animals were sacrificed by cervical dislocation at different time points after administration (1, 3, 6, 9 and 24 h) (three mice for each time point). The organs (brain, kidney, duodenum, spleen, lung, liver, tumour and blood) were collected and homogenized in MeOH:DMSO (4:1 v/v). The homogenates were centrifuged and the amount of P2, P4 or C14 present in the supernatant was measured by fluorescence (Ex. 420 nm, Em. 450 to 750 nm). We did not observe any interference from other compounds present in the extracts. The amount of porphyrin in the various organs was determined by means of a calibration curve, which was obtained by plotting the fluorescence intensity against porphyrin concentration, using standard calibration porphyrin solutions. The curve was linear in the range from 0 to 2 μg/100 mg tissue with a correlation coefficient r 2 = 0.9994. Blood samples were taken from the left ventricle and centrifuged at 3000 g for 10 min to separate the plasma and stored at −80°C. Serum samples were diluted with defined volumes of 2% SDS so that the absorbance of cationic porphyrins at 423 nm was lower than 0.1 and analysed at the spectrofluorimeter. The porphyrin amounts were determined by interpolation on a standard curve plotted with known amounts of cationic porphyrins in 2% SDS, and reported in terms of mg/mg tissue or mg/ml of serum. The assay was highly reproducible with errors <8%.

Antitumour activity of the cationic porphyrins
Female 6-week old C57/BL6 mice were obtained from Harlan-Nossan (Italy) and were maintained in a conventional animal house for 2 weeks. All procedures with animals were carried out in accordance with the National Health Institute Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care Committee at the University of Trieste (approval number given by the Italian Ministry of Health: 6/2011-B). In particular every effort was made to avoid unnecessary pain to the animals. The mice, weighing 20 g, were implanted into the upper flank by subcutaneously injection with 2 × 10 6 B78-H1 amelanotic melanoma cells harvested from a cell culture. After 2 weeks the tumours reached a size of 6-8 mm (along the largest diameter) and the mice were randomized into groups of 8 mice each and injected intraperitoneally with the porphyrins (P2 and P4, 30 mg/Kg; C14, 3 mg/Kg). About 7 h after injection, when the porphyrins were present in the tumour in significant amounts, the mice were anesthetized with Zoletil® + Xylazine (15 mg/kg + 5 mg/kg; i.p.), shaved in the tumour area and irradiated with a laser BWN-660-60E [B&WTEK, Inc, Newark, DE, USA] at 660 ± 5 nm, fluence of 193 J/cm 2 . Three porphyrin/light treatments have been carried out at days 1, 7 and 14. The mice were examined every 2 days for changes in weight, appearance of side effects or signs of sickness. The size of the tumour was measured every 2-3 days by a caliper. The mass (mg) of the tumour was calculated assuming a tumour density of 1 and a tumour volume given by π/6 · a 2 · b where a and b are the shorter and larger axes (cm), respectively. The median survival time is defined as the animal's life spanning from the inoculation of tumour cells till death.