Melanin-Binding Colorants: Updating Molecular Modeling, Staining and Labeling Mechanisms, and Biomedical Perspectives

Abstract

Melanin and melanoma tumors are two fields of increasing interest in biomedical research. Melanins are ubiquitous biopigments with adaptive value and multiple functions, and occur in the malignant melanoma. Although several chemical structures have been proposed for eumelanin, molecular modeling and orbitals indicate that a planar or spiral benzoquinone-porphycene polymer would be the model that better explains the broad-band light and ultrasound absorption, electric conductivity, and graphite-like organization shown by X-ray crystallography and electron microscopy. Lysosomes and melanosomes are selectively labeled by vital probes, and melanin also binds to metal cations, colorants, and drugs, with important consequences in pharmacology, pathology, and melanoma therapy. In addition to traditional and recent oncologic treatments, photodynamic, photothermal, and ultrasound protocols represent novel modalities for melanoma therapy. Since eumelanin is practically the ideal photothermal and ultrasound sensitizer, the vibrational decay from photo-excited electrons after NIR irradiation, or the electrochemical production of ROS and radicals after ultrasound absorption, induce an efficient heating or oxidative response, resulting in the damage and death of tumor cells. This allows repetitive treatments due to the remaining melanin contained in tumoral melanophages. Given that evolution and prognosis of the advanced melanoma is still a concern, new biophysical procedures based on melanin properties can now be developed and applied.

1. Introduction

The study of interactions between small molecular ligands (colorants, drugs) and macromolecular substrates is a biomedical field of steadily growing interest and involves the chemical structure and properties of both components. The term colorant has a wider connotation than dye [1], but here both will be applied interchangeably. A colorant is defined as a substance that is used to impart color to another material [2]. Commonly, detailed physico-chemical features of small dye molecules are better known than those of biopolymers, and often the knowledge of dye-substrate binding mechanisms allows a deeper understanding of macromolecular structure.

The early history (discovery and development) of dyes and drugs is closely interwoven, establishing the foundations of modern chemotherapy [3]. Numerous dyes and their derivatives have found medical applications as diagnostic and therapeutic agents, and new dyes have been developed from the research on antiparasite diseases [4,5]. Thus, the azo dye chrysoidin led to the antibacterial p-sulfamido derivative (prontosil), some bisazo trypanocide dyes (Trypan red and blue) led to the discovery of arsenicals and aryl-urea derivatives (suramin) [6,7,8], etc. Other examples are antimalarial quinacrine and chloroquine, antibacterial acridines and triarylmethanes, antiseptic xanthenes, tumor-markering and psychotropic phenothiazines, antitumoral and antiviral photosensitizing (photodynamic) agents, antitrypanosomal aromatic diamidine fluorochromes, etc. [3,7,9,10,11,12,13].

On the other hand, the application of physico-chemical parameters to colorants (e.g., quantitative structure activity relationships, QSAR) has allowed explaining cell staining and labeling mechanisms, as well as to correct misleading chemical formulations, and to design new and improved probes and drugs [14,15,16,17,18]. As macromolecular substrate, melanin has been scarcely used to study staining processes, but at present a great interest is devoted to analyze the selective binding of some dyes and drugs to this remarkable cell component, which is probably the most ancient group of natural pigments. All Prokaryota and Eukaryota, the latter including Protista, Fungi, Plantae, and Animalia, have melanins [19,20,21].

Melanins are indole- and catechol-type biopolymers, and in mammals, indole-melanins correspond to eumelanin (brown-black) and pheomelanin (yellow-red). They occur mainly in the skin and derivatives, retinal pigmented epithelium, and central nervous system (neuromelanin, in the substantia nigra, locus coeruleus, and stria vascularis of the cochlea). In plants and fungi, melanins generally correspond to the catechol-type (allomelanins) [22]. Early and recent reviews deal with the main features of melanin chemistry, properties, and applications [19,23,24,25,26,27,28,29,30,31,32,33].

Melanocytes are dendritic cells derived from neuro-ectodermal melanoblasts that have migrated to the skin during vertebrate embryogenesis. They are observed in hair follicles, and within and near the basal layer of the epidermis interspersed between keratinocytes, with whom they form adherent junctions with 5–8 neighboring cells using E-cadherin molecules [34]. The biosynthesis of melanin occurs within melanosomes, which are round or oval organelles (~1 µm in diameter) with a clear lysosomal lineage [35,36], and the pigment is then transferred to keratinocytes through dendritic processes. Within melanomas, in addition to melanocytes, melanin-containing macrophages (melanophages) show few to an abundance of variably sized pigmented lysosomal granules (compound melanosomes) often forming coarse and vacuolated structures [37]. Sepia melanin and melanoma melanosomes are also phagocytized by cultured keratinocytes [38].

The malignant melanoma is one of the most aggressive human tumors, showing high resistance to standard treatments. Melanoma also occurs in experimental, companion, and larger domestic animals [37,39,40,41,42,43]. Giemsa- or Papanicolaou-stained cell smears from melanoma aspirates and tumor imprints can be used for cytopathology diagnosis [37,44].

Although the molecular structure of eumelanin is still poorly known, an overwhelming evidence indicates that it corresponds to an indole polymer with a high degree of conjugated double bonds [19,32]. A model based in a benzoquinone-porphycene (BQPo) macrocycle has been suggested [45,46,47]. It could fulfill most characteristics of eumelanin [33], namely a graphite-like structure with broad-band light absorption, semi- and photo-conductivity, strong binding affinity for dyes and drugs, and electron microscopic and X-ray crystallographic features.

On the other hand, eumelanin has a striking affinity for binding to a great number of ligands, examples being colorants, drugs, heavy metals, toxic compounds, and other xenobiotics. In addition to provide a mechanism for elimination of dangerous agents in the skin, ligands can also improve some melanin properties that can be useful in melanoma treatments such as photodynamic (PDT), photothermal (PTT), and sonodynamic (sono-electrochemical, SET) therapy [33]. When the precise melanin type is not referred, the simple name of melanin will be used along this review.

On account of the difficulty in successfully treating the malignant melanoma in advanced and metastatic stage [48], further studies on new therapeutic modalities based on chemical and physical approaches are still necessary. Due to the close relation between chemistry, properties, and implications of melanin in human diseases, the aim of this review is to summarize proposed molecular structures and binding mechanisms of colorants and drugs, as well as their consequences regarding melanoma therapy. New melanoma treatments need to be based on a deeper understanding of the molecular structure of melanin and its interaction with colorants and drugs. Therefore, in agreement with recently suggested rules and criteria for review articles [49], these themes based on molecular mechanics and orbitals will be here illustrated, also attempting to offer seminal perspectives for innovative therapeutic conceptions and developments.

2. Molecular Structure of Eumelanin

The detailed molecular structure of eumelanin is poorly known, although an overwhelming amount of evidence indicates that it corresponds to a highly conjugated aromatic polymer [19,32]. Previous and recent reviews on eumelanin have been published [32,50,51,52,53], but at present no agreement has been reached regarding its precise chemical structure. In spite of methodological limitations, several polymeric linear models from the precursors 3,4-dihydroxy-phenylalanine (DOPA), DHI, and IQ [32] have been proposed (Figure 1). Although the precursor dihydroxy-indole-2-carboxylic acid (DHICA) can also play a role in melanin biosynthesis, it has not taken into account in this work because of the limitation of using the C2 site for polymerization, and the paler color when compared with the DHI polymer. Dopamine (DA) is also a melanin precursor, and although synthetic poly-dopamine (PDA) melanin does not contain the 2-carboxylic group, its properties correspond to a true eumelanin [19].

Possible models of eumelanin polymers are a linear flexible poly 4-7 IQ [19,54], a planar rigid poly 1-7,3-4 IQ [55]), a planar tetramer of a benzoquinone-porphyrin (BQP, cyclic 2-7 IQ [56,57]), and a benzoquinone-porphycene (poly-BQPo [33,45,47]). In contrast with flexible chain models for eumelanin, which have been recently reviewed [32], the poly-BQPo model based in the 2,2′-IQ dimer and the resulting 7-7 tetramer has been first described [45,46] and recently revisited [47]. In this model (Figure 1), the building block is either a benzoquinone (BQ) or benzocatechol (BC) derivative of the porphycene ring (Po), which is a structural isomer of the porphyrin ring. A possible synthesis mechanism based on radical precursors has been described [33]. Taking into account the reactive C2, C3, C4, and C7 on IQ or DHI units, it can be expected that a polymer with all these sites forming covalent bonds with other units would be the most favorable and stable structure. This rationale is just satisfied for the BQPo model, in which these carbons are involved in polymerization (see Figure 1).

Chemical structures of melanin monomers are presented in Figure 2, showing formal, mesomeric, and π*-conjugated structure of the precursor IQ, as well as the formation of hydrated IQ and ether DHI dimer.

In the earliest formulation, Olivieri and Nicolaus [45] proposed the present BQPo model as a “porphyrin-like graphitic structure” for DHI-melanin. A planar, three-layer stacked model appeared in their work as Figure 4, which illustrated a polymer with ether bridges, causing a non-shown curvature. Therefore, the correct planar stacked model without ether bridges is here represented as Figure 3.

To substantiate chemical structures and processes, molecular modeling was performed following previous descriptions [32,33,47]. Inspection of molecular orbitals (MOs) of eumelanin models shows the conjugation changes induced by photo-excitation: ground and excited singlet states of molecules (S₀ and S₁, respectively) result in different MOs. The highest-occupied (HOMO), and lowest-unoccupied (LUMO) MOs represent the energy levels of the ground and excited molecule, respectively [18]. In HOMO-d and LUMO+d, d is from 0 to the maximum energy level, and the HOMO-LUMO separation corresponds to the forbidden Fermi’s energy gap (Eg) between the valence band (VB) and conduction band (CB) of semiconductors. The orbital phases are denoted by colors or signs of π- orbital lobes. Fused lobes with the same color or sign are in-phase, and those with isolated (unfused) lobes with different colors or signs are out-of-phase. In MO images, positive and negative signs have nothing to do with charge. In the case of the planar tetra-BQPo model (Figure 4), the LUMO+0 pattern shows a clear longitudinal orientation, whereas HOMO-0 has rather transversal components.

Quinone compounds containing the vicinal carbonyls (=HC−CO−CO−CH=) from benzoquinones (eumelanin, allomelanin) show a longitudinal LUMOs pattern. In these graphitic materials (melanins, carbon pigments), the massive dissipation of the electronic energy from excited MOs as vibrational energy produces a great amount of heat, namely, a photothermal response [32]. The increase in absorption and dark color of DHI-melanin by further oxidation is explained by the conversion of catechols to quinones [52]. It seems logical to assume that, in pigments with increased number of linear benzoquinones, black color and broad-band absorption spectra are closely related to the high π*-conjugation and extended longitudinal LUMO components, with a reduction in the Eg and an increase in semi-conductivity.

MOs of some linear eumelanin models have been described in previous reviews [32,33]. Regarding the BQPo model (Figure 5A), the formation of ether bridges [45] leads to a chain curvature and potential spiral organization of the poly-BQPoe model [33]. In this case, the excited state LUMO+1 of dodeca-BQPoe (Figure 5B) shows an extended longitudinal π*–electron conjugation, as well as fused MOs, whereas the ground HOMO-0 has a transversal conjugation (not shown). The energy levels of HOMO and LUMO states (Figure 5C) appear again as compact “semiconductor-like” VB and CB, respectively.

From the morphological point of view, transmission electron microscopy (TEM) studies of natural and synthetic eumelanin reveal stacked planar electron-dense sheets, which is characteristic of a graphitic organization. Graphite-like soft curved or marked wavy layers, as well as concentric onion-like structures, are also seen in eumelanins observed by TEM (Figure 6) in all cases with a ~3.4 Å-interlayer separation, which is also confirmed by X-ray diffraction analysis [26,59,60,61,62].

On the other hand, it is known that carbon black nano- and micro-particles have onion-like structures formed by concentrically arranged graphene sheets. Examples are closed-cage fullerenes spheres [63,64], onion-like graphitic spheres [65], nautilus-like fullerene spirals [66,67], and graphene nano-scrolls structured as Archimedean spirals [68,69,70,71,72]. The occurrence of spiral fullerenes and graphenes support the possibility that an oligo- or poly-BQPoe chain could also grow as spiral structures [33]. To this respect, the similarity between both BQPoe and graphene spirals and scrolling processes is astonishing (compare Figure 5A (this work) and Figure 1 from [70]).

It is most exciting to consider that the circular or spiral organization of sp2 carbon sheets [64] might also occur in eumelanin protomolecules (e.g., BQPoe model; see Figure 5). A supramolecular organization of spiral melanin units could be formed by aggregation of smaller and then larger globular structures [73,74]. Spherical onion-like proto-particles (20–50 Å) can be arranged by about four turns (~42 Å) of the Archimedean-type spiral BQPoe model [33], and then they could aggregate into larger globular bodies following co-axial and/or side-to-side parallel linkages. It has been proposed that the supramolecular structure of natural and synthetic eumelanins could derive from a fractal organization of globular units involving different hydration levels, with low pHs and high ionic strength promoting aggregation to larger structures [27,51], in agreement with the hypothesis that small (proto) particles are capable of aggregating to build the final eumelanin structure [75].

The proposed stacked planar BQPo and spiral BQPoe models agree with most structural views of melanins [46,60,76,77]. In the local-order-global-disorder model, ordered nanostructured proto-particles grow first, then aggregate to form disordered spherical particles, that self-arrange again to form large globular structures [78]. Eumelanin was claimed to be composed of onion-like concentric circles [26], which agrees with the idea that the eumelanin proto-molecules could assume fullerene-like closed forms [45,46], or spiral globular structures [33].

3. Spectroscopy and Fluorescence of Melanins

As in the case of graphite and graphene materials, melanins show a broad-band photonic absorption with exponential decay from the ultraviolet (UV) to visible and near-infrared (NIR) region [28,32,79]. It is accepted that this spectrum fits better to graphitic materials and inorganic semiconductors with VB and CB (and small Eg, ~0.5–1.5 eV) [31,80] than to small organic chromophores with structured absorption peaks typically related to π-bonding and π*-antibonding localized orbitals. The conjugated structure of eumelanins also allows easy redox changes and reversible equilibrium between quinone and catechol groups. Another unusual and intriguing feature of eumelanin is its striking binding capacity to a large number of inorganic cations, colorants, and drugs, a topic which will be discussed in the next Section.

Although highly compacted eumelanin is a very dark or even black material, the microscopic color of melanosomes in unstained skin sections is pale brown (Figure 7A,C), which can slightly modify the color of the pigment in subsequently stained sections. Eumelanin is capable of dissipating >99.9% of absorbed UV-visible radiation through a non-radiative decay mode [81]. Since the radiative decay of eumelanin is nearly zero, it is expected that its fluorescence emission should be negligible. In fact, melanosomes from unstained paraffin skin sections show no fluorescence at all under UV, blue, or green excitation (Figure 7B). Easy visualization of melanosomes in melanoma cell cultures can be achieved using phase contrast optics (Figure 7C).

The reported yellow autofluorescence of natural and synthetic eumelanin, neuromelanin, and pheomelanin (λexc: ~470 nm, λem: ~540 nm) seems to be caused by degradation products induced by H₂O₂ oxidation and/or UV/violet irradiation [83,84,85]. In keeping with this, fluorescent quantification of melanin can be made after oxidative degradation [86]. The fluorescence of opio-melanins is also due to low- and middle-molecular weight fractions formed during oxidative photo-bleaching [87]. In ophthalmoscopy, where the lifetime of autofluorescence from the retinal pigment epithelium (RPE) is recorded, the most relevant emission is not due to melanin but to lipofuscin, NAD(P)H, FAD, collagen, elastin, and carotenoids [88]. The lipofuscin A2E is a relevant fluorophore of the RPE [89]. Under 785 nm near-infrared (NIR) excitation, eumelanin shows a very weak broad-band emission between 820 and 920 nm, which appears superimposed with the Raman scatter at 880 and 890 nm [90], possibly contributing to the NIR autofluorescence in ophthalmoscopy. Interestingly, a non-invasive recording of Raman spectra has been proposed for the quantification of eu- and pheomelanin [91].

On the other hand, formaldehyde-induced fluorescence (FIF) in melanocytes was identified as due to the reaction of the aldehyde with DOPA, yielding an isoquinoline derivative [92,93,94]. A green FIF was also observed in pheomelanin-containing cells that arises from 5-S-cysteinyl-DOPA rather than DOPA, allowing for the detection of melanoma metastasis by serum and urine analysis [95].

In contrast, fluorescence quenching by melanin can reduce the background autofluorescence of cells subjected to incorporation of fluorescence-labeled mRNA probes [96]. Likewise, on account of the broad-band absorption, fluorescence quenching by graphene oxide has been proposed for removing the signal of free single-stranded fluorescent DNA nanoprobes (attached to the graphene surface), whereas those hybridized as DNA duplexes remain far away of the surface and retain the fluorescence signal [97]. Interestingly, in spite of its broad-band absorption, which could quench emission by an inner filter or screen effect, melanin allows the emission of bound fluorophores in microscopic preparations, appearing highly fluorescent after appropriate labeling, a feature that will be described later.

4. Staining Mechanisms of Melanin

Eumelanin is the melanin type mainly stained or labeled in melanosomes, but studies on the staining of pheomelanin or allomelanin seem rather overlooked. Although eumelanin is a dark pigment, light microscopic visualization depends on its concentration in the melanosome. It must be noted that cultured melanomas often show weakly or non-pigmented cells. The well-known B16 melanoma line can differ significantly in pigmentation, as well as the expression of antigens, growth rate in mice, metastatic potential, etc. [39]. In contrast, the canine OL melanoma cells show conspicuous melanosomes (see Figure 7C). Microscopic analysis of cultured cells should allow the observation of melanosomes [98] using appropriate optics, but unfortunately detailed morphological studies are often not performed, or the absence of melanosomes is subtlety dismissed.

Several immunohistochemical methods regarding melanoma diagnosis and prognosis are available [99], but the classical silver staining by the Fontana-Masson (FM) method for melanin remains useful [100]. In this argentaffin reaction, which does not require any developer, the silver diammine cation (Ag[NH₃]₂)⁺ from an ammoniacal silver nitrate solution is reduced by melanin to a dark colloidal metallic Ag⁰, but chromaffin cells and lipofuscin pigments are also revealed by this method. In addition to FM, Warthin-Starry (WS) silver stain is also used to reveal melanin [101,102]. In the WS stain, hydroquinone is added to the silver solution, increasing the sensitivity and specificity of detection, possibly by pre-reduction in melanin or by reducing the silver bound to melanin [102].

Most staining reactions are based on the affinity of eumelanin for cationic compounds and dyes (basophilia). Eumelanin is negatively charged at pH values above 4 [103], and then some staining reactivity can be related to its polyanionic nature. Melanin reacts with metal cations, and can be stained by metal-dye complexes (e.g., aluminum-hematein), and cationic colorants. Reactivity of catechols and quinones seems to be also responsible for strong adhesive binding to particles and surfaces [27,104,105,106].

Taking into account most of the reports on binding of inorganic cations to melanin, the following alkali, alkali earth, and transition metals can be mentioned: Na, Mg, Al, K, Ca, Ti, V, Cr, Mn, Fe, Co, Cu, Zn, Sr, Mo, Ag, Cd, Sb, Hg, Pb, As, etc. [19,30,77,78,107,108,109,110]. Although native sepiomelanin appears as a Ca and Mg salt, exchange reactions with the above-mentioned cations are possible. The amount and binding affinity of melanin for some metals can be very large, binding to Ca²⁺, Mg²⁺, Sr²⁺, and Cu²⁺ being 4, 5, 14, and 34 times stronger than EDTA, respectively. For Ca, Mg, Fe, Cu o Zn, the saturation levels of binding are ~3–4 indole units per ion [79]. Binding for metal ions is about 6 × 10²⁰ sites per gram of dried melanin [107].

Diagnostic and microscopic applications of metal binding to melanin are abundant. Melanoma detection by magnetic resonance imaging (MRI) could be improved by using melanin uptake of ⁶⁷Ga and ¹¹¹In [111]. Doping of eumelanin and synthetic melanin with titania (TiO₂) nanoparticles, and Fe³⁺ or Gd³⁺ metal ions with catechol or quinone units (Figure 8), produces nanoplatforms for multimodal imaging and therapeutic applications, enhancing light absorption, photocatalysis, and photothermal effect [30,78,108,109,110].

In the case of TiO₂, the O atoms of IQ ligands forms a charge-transfer complex, which also promotes polymerization of the monomer [109] (Figure 8A). The iron uptake reaction gives a blue-green color [100], which is typical of ferric-catechol complexes [7,78]. Fe³⁺ hexa-coordinated by O atoms from catechol, forms a mononuclear complex with typical square-planar geometry and octahedral arrangement [78] (Figure 8B). Likewise, borate esters with tetrahedral geometry of anionic boron may be formed using the O5 and O6 catechol atoms of DHI. Strong melanin binding of trivalent lanthanides (La, Pr, Nd, Sm, Gd, Dy, Ho, Er, Tm) with the O ligands of IQ has been also reported [107,108]. Using molecular docking of cations on the melanin BQPo model, binding of Ca²⁺ and lanthanides (Eu³⁺, Dy³⁺) to O5 and O6 have been found (Figure 8C).

Interestingly, the luminescence of Eu³⁺ under UV excitation [18] and its binding to melanin (Figure 9) can be microscopically visualized due to its red (620 nm) emission when complexed with IQ, which act as a ligand UV antenna for Eu³⁺ photoexcitation. Likewise, Pr, Nd, Sm, Dy, Ho, and Er are NIR-emitting lanthanides that when bound to IQ units of melanin could induce selective tumor-damaging luminescent heating.

After using the common nuclear staining of paraffin tissue sections by the cationic blue aluminum-hematein complex (H), the pale brown color of melanosomes is shifted to blue-violet, which depends on the binding of Al ions of the complex to melanin anionic sites. When followed by eosin Y staining (H&E), melanosomes are visualized in dark violet-brown or even black color (Figure 10). In addition to H, other colored or fluorescent dye-metal lakes are expected to stain melanosomes.

Numerous cationic dyes can stain eumelanin [7,33,100]. Examples are methylene blue, toluidine blue, thionine, Nile blue, and acridine orange. The precise binding mechanism of cationic dyes to eumelanin has been rather overlooked. In contrast to classical views [100], melanosomes from both normal skin and melanoma cells give metachromatic staining and fluorescence when stained with thionine, toluidine blue, and acridine orange. A clear example is toluidine blue metachromatic staining of keratinocyte melanosomes in the skin of Eubalaena australis (Figure 11).

The polyanionic character of melanin could explain its metachromatic staining reaction with appropriate colorants. In the case of double-stranded nucleic acids, intercalation between base pairs is favored when compared with external electrostatic binding to phosphate sites [18,114]. The same preference seems to occur for melanin when the graphitic BQPo model is analyzed for dye docking using molecular modeling: cationic acridine and thiazine colorants prefer face-to-face hydrophobic binding (lower free energy) instead of edge-to-edge ionic interactions (higher free energy).

It is well known that planar aromatic compounds can intercalate into host lattices forming inclusion complexes [115]. In the case of tissue components such as polysaccharides, lignin, nucleic acids duplexes, etc., bathochromic changes in the absorption spectra are induced by monomerization of colorants, which remain trapped between aliphatic chains or aromatic rings of the biopolymer [116,117].

Since Lerman’s formulation of the intercalative binding mode into DNA [118], numerous colorants have been found to bind to nucleic acid duplexes by intercalation [114,119], using strong face-to-face hydrophobic and π–π electron interactions between planar dyes and base pairs. Therefore, on account of the aromatic character of eumelanin, intercalation of planar cationic dyes should be the preferential binding mode, as occurs with nucleic acid duplexes stained by acridine, thiazine, azine, oxazine, and xanthene dyes. A gallery of representative cationic colorants and fluorochromes involved in melanin binding is presented in Figure 12.

Stacked dye aggregates commonly found in classical tissue histochemistry give metachromatic reactions, which are most apparent for toluidine blue [120,121], and acridine orange [18,122]. They have the advantage that ortho- and metachromatic peaks (and colors) are well defined and spectrally separated.

In this respect, the copper-phthalocyanine (CuPc) macrocycle is an intriguing model dye regarding binding to eumelanin. A tetrasulfonated CuPcS4 dye deposited on multiwall carbon nanotubes has been used as electronic conductor in photovoltaics [123]. The cationic derivative Alcian blue pyridine variant (ABPV) was also applied in TEM studies of PDA-melanin [62]. No increase in electron contrast over that of a graphitic structure was apparent, but a new composite was formed by layer-by-layer film deposition of PDA-melanin and ABPV, which displayed an electrical conductivity five orders of magnitude higher than that of pure melanin films [62].

Interestingly, clear 610 nm metachromatic spectra (peak/shoulder: 610/670 nm) appear after addition of successive monomeric ABPV-PDA layers [62]. This is an intriguing finding, because dye intercalation involves binding of isolated monomers, canonically leading to an orthochromatic reaction (red or bathochromic spectral shift), and not to metachromasia (hypsochromic shift). Spectroscopic analysis confirms that in the presence of Sepia melanin (SM), the absorption peak of the AB-SM complex remains metachromatic (blue), as it would correspond to a highly aggregated state and not to the intercalated monomers (Figure 13). This unexpected metachromatic reaction can be explained on account of the massive occurrence of overlapping LUMOs in the AB/SM intercalated complex, allowing the π*–electron coupling between all CuPc chromophores [33]. Thus, long-range CuPc “aggregates” are established from CuPc intercalated monomers.

Positive charges (a) on the ring of phthalocyanine colorants (e.g., cuprolinic blue and its analogous zinc complex [124,125,126]), or (b) on the side chains (e.g., AB [127,128], and Alcian blue pyridine variant [129]) could be advantageous but not essential for intercalative binding and staining of eumelanin. Although there are no structural details of the dye binding to this substrate, it is logical to assume that, as it occurs in the face-to-face CuPc binding to the graphitic surface of carbon nanotubes [123], the same mechanism could also be possible for eumelanin.

5. Lysosome and Melanosome Labeling

Selective labeling of lysosomes and endocytic cell compartments (endocytosis vesicles, endosomes, phagosomes, trans-Golgi network, lysosome remnants such as age pigments or lipofuscin granules) is a well-known process. The main vital probes are acridine orange, acridine yellow, alizarin, auramine O, 3,4-benzpyrene, Bismarck brown, brilliant cresyl blue, coupled azo dyes, eosin Y, erythrosin B, euchrysine 3R, Evans blue, FITC-dextran, fluorescein, hematein and hematoxylin, hydroethidine, inorganic nanoparticles, Lucifer yellow CH, methylene blue, neocyanine, neutral red, Nile blue, phloxine B, phosphine, primulin, proflavine, purpurin, quinacrine, thionine, toluidine blue, trypan blue, etc. [5,18,130,131,132,133,134,135,136,137,138,139]. Numerous lysosome probes are also efficient photosensitizers for photodynamic therapy (PDT) [140,141].

On the other hand, eosinophil, basophil, and azurophil leucocyte granules, sperm cell acrosomes, secretion (zymogen) granules, and particularly melanosomes, are acidic compartments that also have a clear lysosome lineage [18,35,36,142,143,144]. The intramelanosomal pH range is as low as 3–5, and the organelle contains proton-translocating ATPase, hydrolases, cathepsines B and L, and acid phosphatase [35]. In addition to colorants and fluorescent probes, an interesting feature is the high number of xenobiotics (nanoparticles, colloids, latex microparticles, metals, drugs) that can be up taken by lysosomes and melanosomes [35,130].

Selective lysosome labeling with vital colorants and fluorochromes show metachromatic color, although toluidine blue (TB) and methylene blue (MB) give violet and blue colors, respectively; both dyes are just metachromatic, but the shift is only appreciated for TB, because the shift for MB takes place entirely within the blue spectral region, and it is not appreciated by the eye. In the case of CuPc, the colorant shows a spectral metachromasia in spite of the monomeric binding in PDA-ABPV films [62]. After acridine orange (AO) labeling, bright orange emission is observed in lysosomes from live cells, which is due to the low lysosomal pH and the aggregation of AO⁺, reaching local metachromatic concentrations (Figure 14A), whereas a weaker AO⁺ accumulation produces a green orthochromatic emission in nucleoli, chromatin and cytoplasm [18]. Selective lysosome labeling by other photosensitizing dyes such the porphyrin ZnTPP, and the porphycene TPPo are shown in Figure 14B,C, respectively.

Lysosomal photosensitizers such as AlPc2, AlPc4, CF3, Pc13, PoD, Py₃MeO-TBPo, SnEt2, TPP, TPPS2a, TPPS4, TMPyP, ZnPc, ZnTPP, etc. [140,141,145,146,147,148,149] should also be suitable to label melanosomes, and then for dye-improved photothermal and ultrasonic melanoma therapy. It must be noted that some well-known lysosome probes can also label the Golgi apparatus [150]. Lysosome-labeling PDT dyes that are water insoluble can be delivered within DPPC liposomes to enter live cells, increasing selective uptake into tumor cells [151]. Regarding porphycenes, as the double N-protonation of the Po core at low pH is possible, some amount of TPPo could remain trapped in lysosomes as a dication [141,152].

The selective AO labeling of cultured canine OL melanoma cells (Figure 15), illustrates the strong metachromatic labeling of melanosomes (and lysosomes). It is worth to note that AO uptake in both organelles depends on the low pH, but dye aggregation occurs due to its high concentration in lysosomes, and to massive dye binding to eumelanin in melanosomes.

The possible molecular structure and electron orbitals of this colorant bound to eumelanin bis-BQPoe units is shown in Figure 16A,B, as modeled intercalative binding of two AO⁺ into two sheets of the units. Molecular docking indicates that both aromatic acridine rings approach and remain bound face-to-face on one sheet of BQPoe. In the intercalated complex, the fused MOs from LUMO+5 are clearly visible (Figure 16C).

Lysosomes and melanosomes share numerous properties, and when they occur simultaneously, differential identification should involve unshared properties: (a) the brown color of melanosomes characterizes this organelle; (b) silver staining or immunocytochemistry (for tyrosinase or melanosomal proteins) to detect only melanosomes; (c) autofluorescence, which occurs in lysosomes [18] but not in melanosomes. In comparison with melanosomes of skin and skin melanoma, neuromelanin [77], and melanin from RPE are very special cases, because they have a complex organization, with melanosomes, lipids, lipofuscin, and melanosomes encased in melano-lipofuscin granules [25].

Macrocyclic colorants are also suitable to fit tightly between eumelanin layers of the poly-BQPo model, allowing to develop fused LUMOs [33,47]. The possible intercalative binding of TPPo into two BQPoe units is shown in Figure 17A,B. When MOs are analyzed, the fused LUMO pattern is clearly observed (Figure 17C). Although different in shape, fused LUMOs of intercalated TPPo-BQPoe eumelanin also occur at many other energy levels.

Figure 18 emphasizes the hydrophobic binding mode of two copper phthalocyanine (CuPc) chromophores from the AB dye with three aromatic BQPoe units of this eumelanin model. Obviously, stacked arrays of fused LUMOs permits interactions between successive dye chromophores through MOs coupling with the host eumelanin layers, this feature allowing the metachromatic reaction of intercalated dye monomers.

In the intercalative binding mode, the planar ligands slip between aromatic units of eumelanin layers, and remain trapped as an inclusion complex or “graphitic sandwich”, allowing both linear and stacked fused LUMOs [33]. Therefore, massive π*–π* electron interactions are found in dye-intercalated BQPo or BQPoe model sheets. A schematic diagram of coupled LUMOs is shown in Figure 19. These coupling patterns just occur in aggregated colorants commonly found in tissue histochemistry, leading to a strong metachromatic shift.

6. Binding of Colorants and Drugs to Melanin

Natural and synthetic eumelanins are capable of binding small molecules with high dielectric constants (dimethyl sulfoxide, formamide, methanol), inducing reversible conductivity changes of as much as ten orders of magnitude [154]. Hydrated DOPA-melanin complexed with diethylamine subjected to 10 kHz-ultrasound (US) shows a resistivity reduction in ~6 orders of magnitude [155]. The high affinity of eumelanin for organic compounds allows the establishment of charge-transfer complexes between the pigment (electron acceptor) and the included compound (electron donor), leading to a strongly increased conductivity [80].

High binding affinity of eumelanin for drugs and colorants was found as an unexpected adverse effect of treatments with some neuroleptic and antimalarial drugs. Chronic administration of antidepressant phenothiazines [156,157], and high-dose chloroquine therapy [158] produced chorioretinopathy, suggesting an association between toxic effects of drugs and affinity for melanin. Ototoxicity and disturbances of the skin and hair pigmentation were also reported. Drug binding affinity has been calculated in silico by simple free-energy methods [159], as well as by solvation energy relationships [160], in vitro-in vivo correlations [161], and microscale thermophoresis [162].

Using biochemical analysis, typical colorants, drugs and chemicals have been found to bind strongly to melanin [74,163,164,165,166,167,168,169,170,171,172,173,174]. Among these melanin-binding ligands, colorants are acridine orange, methyl orange, Biebrich scarlet, Bismarck brown Y, Congo red, Buffalo black NBR, methylene blue, thionine, toluidine blue, fluorescein, neutral red, neutral violet, safranine O, Janus green B, Nile blue A, bromophenol blue, malachite green, fast green FCF, alizarin red S, primulin, tetramethyl-p-phenylenediamine, etc. [159,163]. Most of these colorants are also selective probes for lysosome labeling.

Likewise, numerous drugs from different chemical groups bind strongly to melanin (for references, see [74,163,164,165,166,167,168,169,170,171,172,173,174]). Examples are: aflatoxin B₁, p-aminobenzoic acid, aminoglycoside and tetracycline antibiotics, amitriptyline, aniline, benzamides (e.g., ¹⁸F-labeled picolinamide or nicotinamide derivatives), and antipsychotic and anti-emetic benzamides, benzidine, carcinogenic polycyclic hydrocarbons (benzo(a)pyrene, dimethyl-benz(a)anthracene), chloroquine, chlorpromazine, dexamethasone, diclofenac, fluoro-quinolones, herbicides (paraquat), indole acetic acid, iodoquine, MPP⁺ (1-methyl-4-phenylpyridinium), methotrexate, N-nitrosamines, 2-naphthoic acid, papaverine, timolol), quinine, quinidine, phenothiazines, thiouracil, thioureas, trimethyl-psoralen, etc. Benzamides are also used as melanotropic carriers for cytostatic drugs [171].

Therapeutic uses of melanin-binding agents (named melanin or melanoma “seekers”) also have limitations. It has been estimated that 7% of primary skin melanomas, 13% of regional lymph-node metastases, and 31% of distant metastases lack melanin [175]. The occurrence of amelanotic melanomas is a disadvantage, because melanoma seekers do not accumulate in those tumors. However, amelanotic melanomas may contain minor amounts of melanin [176], which could be still sufficient to work as a selective therapeutic target.

Melanin-binding ligands can be related to neurotoxic effects, Parkinsonism, and melanoma induction [170]. Interestingly, some of these ligands are known DNA-intercalators, and analogies between DNA- and melanin-binding modes have been pointed out [165]. Among the wide collection of melanin-binding ligands, numerous aromatic compounds are obvious candidates for an intercalative binding mode, according to the models already mentioned (see Figure 16, Figure 17 and Figure 18). Thiazine ligands (almost planar promethazine and chlorpromazine, with ring torsion angle of ~20°), and planar methylene blue bind strongly to melanin, likely by intercalation.

Interestingly, the ability to label melanosomes can be profited by using radionuclide-containing colorants or agents in therapeutic protocols for advanced melanomas [177]. Binding of radionuclide-labeled dyes and drugs to eumelanin has been applied for melanoma diagnosis and therapy [166,174,178,179,180]. It is worth to note that in classical studies on melanin-bearing tissues (melanomas, uveal tract, hair follicles), whole body autoradiography has shown labeling after administration of radioactive chloroquine and chlorpromazine [166].

Radioactive ligands such as ³⁵S-methylthionine bromide, ¹²⁵I- and ²¹¹At-methylene blue have been used to irradiate pigmented melanoma in animal models [179,181,182]. Mono- and di-radioiodinated methylene blue [183], and ¹²⁵I-radiolabeled acridine and acridone derivatives [184] have also been used. ¹²⁵Iodine and ²¹¹Astatine are Auger-electron and alfa-particle emitters, respectively.

Several compounds with a thioureylene group (e.g., thiouracil and 6-alkyl derivatives, thioureas, methimazole) are incorporated into melanin during its synthesis, and thiouracil derivatives such as ¹²⁵I-5-iodo-2-thiouracil, ³⁵S-2-thiouracil, and ¹⁰B-5-dihydroxyboryl-2-thiouracil have been applied as selective tumor seekers for radio-scanning diagnosis or treatment of malignant melanoma and metastasis [169,180], in the latter case using slow neutron irradiation. Targeted radiotherapy of pigmented melanoma using ¹³¹I-labeled picolinamide agents has been carried out [185]. An interesting review has been published on imaging of melanoma metastases with the positron emitter ¹⁵F attached to melanotropic carriers [186].

Likewise, targeting of melanocytes in vitro and in vivo has been achieved using octapeptide derivatives of α-melanocyte-stimulating hormone (α-MSH) for binding to the specific receptor MC1R, containing either a chelator for ¹¹¹In, ^67/68Ga or ⁹⁰Y (allowing accumulation of radioactivity in experimental melanomas) [187], or a photosensitizer (methylene blue or pyropheophorbide derivative) for PDT [188]. Finally, due to its high ability to retain heavy metals, selective melanin seekers could be simple radionuclide cations for direct binding to O or N atoms of the pigment.

7. Melanoma and Ultrasound Therapy

On the other hand, using ultrasound (US) microscopy at frequencies near 1 GHz, the comparison of iris tissue from albino and pigmented rabbits showed that melanin is a particularly strong acoustic attenuator at those frequencies. In addition, other high US-absorbing structures are cell nuclei, rod and cone outer segments, and red blood cells [189]. Acoustic waves can traverse longer distances than photons in condensed matter, and as such, penetration of US into deep tissues can be orders of magnitude higher than light. Accordingly, targeting and therapy of melanomas subjected to sono-sensitizers (SSs) and US treatments show good responses and promising perspectives. Taking into account the binding of numerous ligands to melanin and the strong photonic and US absorption of this pigment, it seems logical to expect an improved photothermal and ultrasono-activity.

An emerging approach for the therapy of tumors and atherosclerosis has been called sonodynamic therapy (SDT), which involves the use of a SS followed by exposure to low intensity US [48,190,191]. As SDT generates damaging ROS and radicals, a more appropriate denomination for this modality should be “sono-electrochemical therapy” (SET), to avoid the misleading use of the term “dynamic” [33]. Employed US parameters for SET are ~1–2 MHz frequencies and ~0.5–10 W/cm² energy densities [192].

Berberine, curcumin, hypericin, and protoporphyrin IX are natural SSs with increasing importance [190,191]. Regression of experimental B16 melanoma has been achieved using several SSs such as TiO₂ nanoparticles, chloro-aluminum phthalocyanine disulfonate, rose Bengal, and nickel ferrite/carbon nano-composites (NiFeO/C), revealing that SET may be more effective than PDT in treating advanced melanotic melanomas [48]. TiO₂ nanoparticles subjected to US (1 MHz, 1 W/cm², 2 min) produced significant regression of the C32 melanoma [193]. It is tempting to assume that generic biomembranes and lysosome uptake of nanoparticles [138], and piezoelectric activity of TiO₂ are involved in the therapeutic response. As xanthene colorants accumulate into lysosomes, it would be expected that rose Bengal could also label melanosomes. It is noteworthy realizing that the strong US-absorbing eumelanin from melanomas is just a highly suitable endogenous SS for SET itself.

Melanocytes are killed in vitro by US (phonons) in proportion to their melanosome content [194]. Melanin-binding drugs and colorants induce additional toxicity in melanocytes subjected to US, and cell damage increases sharply with temperature, from 7 to 37 °C. The absorption coefficient of melanin exposed to 1-MHz US increases and reaches a value approaching maximal absorption at 60 °C. Absorbed US converts the mechanical phonon energy into cytotoxic products through phonon–electron interactions within the melanosome [194]. In early experiments, cultured melanotic tumor cells treated with chlorpromazine were preferentially killed by 10 kHz US irradiation [155], with radical production and DNA damage. Likewise, 1 MHz-US induced melanin degradation and death of melanin-containing cells, which were potentiated by previous treatment with melanin-binding drugs such as chlorpromazine and kanamycin [194,195].

Therefore, the concept of melanin as an inert biopolymer is in conflict with the experimental evidence that it is an unusually efficient energy conversion medium with possible biological implications, and capable of cytoprotective as well as cytotoxic action. Cytoprotection is observed with the conversion of photons or excited-state energy of free radicals into phonons. Cytotoxic effects are achieved by driving this process in reverse. In keeping with these facts, electron–phonon interactions are known [196,197]. A photon–electron–phonon coupling model for energy interactions in melanin and ligand-bound melanin is shown in Figure 20.

Therefore, melanin-, melanin/ligand-based SET, and melanin/lanthanide-based US-hyperthermia could represent new, elegant, and non- or minimally invasive therapies for melanoma.

8. Biomedical Perspectives

At present there are important advances in imaging and therapeutic protocols for melanomas and neurologic diseases using properties of melanin and melanin-like nanoparticles [32,78,200,201,202]. A better knowledge of melanin structure and binding capacity toward colorants and drugs will be surely reflected in understanding mechanistic aspects of drug-induced pathological processes, leading to a progress of new biophysical approaches for photodynamic, photothermal, and ultrasonic therapy of melanomas. In this review, we have described molecular modeling advances (mainly planar BQPo and spiral BQPoe models) to explain the chemical organization of eumelanin, as well as the possible intercalation of colorants into graphitic eumelanin layers. These processes result in linear and fused LUMOs in melanin-dye complexes, allowing photon–electron–phonon interactions, and leading to metachromatic staining and labeling of melanosomes. Features based on US-absorbing melanin-ligand complexes induce biological and antitumoral effects, in agreement with previous proposals [33,47].

If corroborated by further studies, these novel conceptions may open interesting perspectives and efficient strategies regarding melanoma therapy, as well as innovative developments with potential biomedical and biotechnological applications. Regarding melanin affinity and melanoma seekers, it is conceivable that the great quantity of catechol groups makes eumelanin a suitable substrate to form borate esters [18]. When the isotope ¹⁰B is used and then irradiated with slow neutrons, the products of the nuclear fission (⁷Li and ⁴He nuclei) due to neutron capture could damage specifically melanoma cells because of their very short braking distance in condensed matter. In keeping with this, a borate ester of DOPA [203] has been proposed as a possible precursor of melanin for ¹⁰B neutron-capture therapy of melanoma.

Regarding dye–melanin interaction, suitable molecular fitting should be a relevant factor: most colorants that label lysosomes also bind to melanin, and could also intercalate between its aromatic layers. In particular, possible intercalative binding of the porphycene dyes TPPo and PdTPPo with melanin would be based on the excellent geometrical matching of these ligands with the BQPo model for eumelanin (see Figure 17), because the host unit and the ligand dyes have the same porphycene ring.

Hypericin induces a necrotic mode of cell death in pigmented melanoma cells and melanocytes (possibly due to the release of toxic melanin precursors from melanosomes), as well as apoptosis in non-pigmented melanoma cells and keratinocytes [204]. Taking into account the aromatic structure of hypericin, it is tempting to speculate that its lysosomal localization [205] could allow uptake and intercalative binding to melanin also in melanosomes. Colorants from acridine, thiazine, xanthene, azine, oxazine, porphyrin, phthalocyanine, and other groups are also well suited to intercalate into graphitic melanin layers. In all these cases of melanin–colorant interactions, an enhanced biological response of melanoma cells and tumors would be expected to occur. Therefore, better understanding of the molecular structure of melanin, as well as binding possibilities and melanin staining and labeling mechanisms, will surely improve the rationale and efficiency of PDT, PTT, and SET protocols for melanoma therapy.