2.1. Melanoma is a major and costly cancer . . . . . . . . . . . . . . . . . . . . . . . . . . 1992.2. Uncertainties in epidemiology cause uncertainties in prevention. . . . . . . . . . . . . . 199
3. Human and animal model data on long wavelength ultraviolet causation of melanoma . . . . . 1993.1. Human epidemiological data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1993.2. Signature mutations in human melanomas do not suggest short wavelength ultraviolet . . 2003.3. Sunscreens do not protect humans against melanoma . . . . . . . . . . . . . . . . . . . 2003.4. Long wavelength ultraviolet melanoma causation data from animal models. . . . . . . . 200
4. Factors that potentiate the importance of wavelengths N320 nm in melanoma causation . . . . . 2014.1. Long wavelength ultraviolet penetrates skin to melanocytes better than short
wavelength ultraviolet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2014.2. Long wavelength ultraviolet is much more prevalent than short wavelength ultraviolet . . 2014.3. Melanin is a dominant short wavelength ultraviolet photosensitizer in melanocytes
Melanoma incidence and deaths continue to increase in mostpopulations (Longstreth, 1988; Woodhead et al., 1999),although major efforts in prevention and early detection incertain countries, such as Australia (Baade & Coory, 2005),have led to mortality being stabilized or even reduced. How-ever, it is still not precisely known which wavelengths insunlight cause human melanoma, or how these different wave-lengths act. This lack of knowledge profoundly hindersscientifically based prevention strategies and has led to contro-versies over how best to protect against melanoma, especial-ly regarding the role of sunscreens in melanoma prevention(Huncharek & Kupelnick, 2002; Christensen, 2003; Huncharek& Kupelnick, 2003; Marshall et al., 2003). This controversy haseven led to consumer lawsuits in the United States. Here, wereview the evidence supporting the importance of longwavelength ultraviolet (UVA) in causing melanoma, especiallywith regard to the role of melanin as an endogenous photo-sensitizer in the melanocyte; the evidence regarding sunscreenprevention of melanoma; and current measurement methodol-ogies of the UVA protection of sunscreens. Drawing upon ourrecent research (Wood et al., 2006), we describe a noveltechnique for measuring the UVA and melanocyte protectionfactors of sunscreens that may allow a resolution of some ofthe controversies in this field.
2. Background
2.1. Melanoma is a major and costly cancer
Melanoma rates have been increasing for many years in manycountries. SEER study analysis in the United States showsannual percent increases of 2.5% between 1992 and 2001 (Rieset al., 2004), with over 60,000 new cases of melanomadiagnosed in 2006, and with about 8000 deaths predicted bythe American Cancer Society. In White males, the highest riskgroup, the lifetime risk of diagnosis in the United States iscurrently 2.27%. Besides its increasing prevalence, the eco-nomic and emotional costs of melanoma are great, a currentestimate is of about US$3 billion annually in the United Statesalone (Grant et al., 2005). Furthermore, melanoma has adisproportionately high mortality in younger age groups, suchas 18–40 years old, with each death causing loss of almost19 years of life, among the highest for adult onset cancers(Ries et al., 2004). Even in countries such as Australia, where amajor public health campaign on sun exposure prevention andearlier detection has led to a decrease in mortality in youngerages (Baade & Coory, 2005), the incidence of melanoma hasnot decreased. Overall, it is clear that reducing melanoma
incidence, and hence mortality, is a necessary and worthwhileendeavor.
2.2. Uncertainties in epidemiologycause uncertainties in prevention
Despite long-time recognition of the association of melano-ma causation with exposure to sunlight, especially acuteintermittent exposures (Armstrong, 1988; Gandini et al.,2005; Whiteman et al., 2006), it is still unclear how sunlightcauses melanoma. Although many important genes involved inmelanoma have been characterized (Chudnovsky et al., 2005), itis still far from clear how melanoma is caused and which kindsof mutations (short wavelength ultraviolet [UVB] or oxidativesignature mutations) are involved in all the relevant genes.Because of these uncertainties in how UV causes melanoma,scientific and medical authorities are unable to offer detailedprevention advice, and this is especially the case when it comesto which wavelengths cause melanoma. This uncertainty pri-marily derives from a lack of appropriate action spectra, withthe only one available being in a highly-susceptible cross-bredXiphophorus fish model (Setlow et al., 1993). Just one result ofthis profound uncertainty has been paradoxical recommendationson whether sunscreen use is even protective, due to differentefficiencies of older sunscreen in blocking UVA and UVB(Huncharek & Kupelnick, 2002; Christensen, 2003; Huncharek& Kupelnick, 2003; Marshall et al., 2003) although poor UVAefficiency is still apparent even in ‘new’ UVA protectionsunscreens (Haywood et al., 2003). It is unclear as to how thiscontroversy will be resolved without definitive experimentalevidence, such as an action spectrum in a suitable model.
3. Human and animal model data onlong wavelength ultraviolet causation of melanoma
3.1. Human epidemiological data
A significant body of epidemiological evidence suggeststhat both UVA and UVB are involved in melanoma causation(Moan et al., 1999; Oliveria et al., 2001; Wang et al., 2001;Garland et al., 2003), whereas nonmelanoma skin cancers(NMSC) have primarily linked to UVB (de Gruijl, 1995).However, even for the much better understood situation inNMSC, there are recent indications that UVA may also beimportant (Agar et al., 2004). One of the major reasons forthis uncertainty is that the epidemiological evidence regardinglonger wavelengths than UVB is controversial, as sunlight is acomplex and changing mix of different UV wavelengths, so itis very difficult to accurately delineate the precise lifetimeexposures of individuals and entire populations to UVA and
UVB from available surrogates, such as latitude at diagnosisor exposure questionnaires (Eide & Weinstock, 2005). Even inthe clearer-cut case of tanning lamp exposure, there are stilluncertainties, deriving from the differing types of lamps usedat different times as tanning booth technologies have evolved(Miller et al., 1998). In short, although data of incidence as afunction of latitude does support the importance of wave-lengths N320 nm, (Moan et al., 1999), there is insufficientpower and too many confounders (variable UVB exposures,etc.) to base prevention strategies solely on these data.
3.2. Signature mutations in humanmelanomas do not suggest short wavelength ultraviolet
A powerful approach in assigning the mechanisms of skincancer causation has been to study the mutations in cancer-associated genes. This is possible because there are often‘signature’ mutations that allow determination of whether thesekey mutations were caused by direct DNA absorption of UVB,causing formation of pyrimidine dimers and 6-4 photoproducts,or whether they were caused by oxidative stress such as throughformation of 8-hydroxy-2′-deoxyguanine (Agar et al., 2004).Unlike the characteristic UVB fingerprint mutations of genesfrequently observed in human NMSC, there are as yet fewercharacteristic wavelength signatures in melanoma, althoughoxidative stress has most often been implicated. For example,BRAF mutations are common in melanoma in sun-exposedsites (Gorden et al., 2003; Pollock et al., 2003; Goldenberg-Cohen et al., 2005) but not in sun-protected sites, (Edwardset al., 2004; Helmke et al., 2004; Wong et al., 2005) and are alsovery common in sun-induced nevi. (Pollock et al., 2003; Kumaret al., 2004) BRAF is strongly linked to melanoma causation,including but not exclusively, causation of nevi as a preme-lanoma lesion (Michaloglou et al., 2005; Patton et al., 2005). Anoverwhelming proportion, about 90% of these sun-inducedBRAF mutations, are the single nucleotide substitution T1799A(V600E) (Brose et al., 2002; Davies et al., 2002), a lesion that isnot formed by UVB-induced pyrimidine dimers or 6–4 pho-toproducts but is instead suggestive of an oxidative stresslesion, although the precise mechanism is as yet uncertain(Edwards et al., 2004). As well as oxidative mutation signaturesin BRAF, other commonly mutated melanoma-related genesshow strong oxidative damage signatures. Papp et al. (1999)studied NRAS to show that in congenital melanocytic nevi, acondition with a high probability of transformation intomalignant melanoma, 10/18 lesions showed mutations atcodon 61 of NRAS with 9 of these 10 mutations being signa-tures of oxidative DNA damage (CAA to AAA). A similarlyhigh prevalence of CAA to AAA mutations in codon 61 ofNRAS has been observed in melanomas in familial melanomain which CDKN2 mutations predispose towards melanoma(Eskandarpour et al., 2003). Although some mutations inhuman cutaneous melanoma in TP53 (relatively rare) are C-Tmutations at dipyrimidine sites, indicative of direct UVB DNAdamage, the same mutations were observed at similar fre-quencies in non-ultraviolet (UV) exposed mucosal melanomas,arguing against a role for UVB causation of these mutations
(Ragnarsson-Olding et al., 2002). Thus it can be seen that,although not yet incontrovertible, the available evidencepoints away from pyrimidine dimer formation by UVB andtoward oxidative stress damage caused by longer wavelengths.
3.3. Sunscreens do not protect humans against melanoma
Few areas in the field are as controversial as to whethersunscreen use might contribute to melanoma causation. Thishypothesis arose as early sunscreens primarily screened UVBbut not UVA, as their activity was measured by their sunprotection factor (SPF) which is determined by their ability toprevent erythema by a process caused only by UVB (Anderset al., 1995). However, upon determination of the actionspectrum of melanoma causation in Xiphophorus (Setlow et al.,1993; Setlow, 1999), its convolution with solar irradiancespectra suggested that over 90% of melanomas could be causedby wavelengths longer than UVB such as UVA, with somecontributions by even visible light. Since older sunscreens onlyblocked UVB, they would not block the full range ofwavelengths causing melanoma, so their protection againstmelanoma would be minimal. Furthermore by allowing in-creased exposures by preventing UVB burning, they couldactually increase UVA exposures and melanoma prevalence(Garland et al., 1993; Setlow & Woodhead, 1994). Subse-quently, there has been much epidemiological investigation,recently, comprehensively meta-analyzed by Huncharek et al.(2002) and Dennis et al. (2003) to show there seems to be littleor no positive association of sunscreen use with melanomacausation, although most of these studies were based uponolder UVB only sunscreen, and although there remainscontroversy (Christensen, 2003; Huncharek & Kupelnick,2003; Marshall et al., 2003). However, upon careful consider-ation of these studies, it is best concluded that in fact the mostappropriate analysis of this data, representing over 5500 casesand 7800 matched controls (Dennis et al., 2003) is thatpredominantly UVB absorbing sunscreens appear unable toprevent melanoma in human populations. The corollary of thisis that UVB does not seem to be a significant cause ofmelanoma in human populations, although the impact ofconfounding factors can only be addressed by a placebocontrolled clinical trial of prevention (Armstrong, 2005).Although optimism that the use of modern sunscreens, withbetter UVA absorbance, has been voiced (Diffey, 2005) thereare real difficulties in measuring real-life UVA protection(Gasparro, 2000; Cole, 2001) and the UVA protection of evenmodern sunscreens of claimed UVA high protection has beenshown by some to be insufficient (Haywood et al., 2003).Furthermore, a recent study (Gallagher et al., 2005) indicatesthat “there is a significantly increased risk of melanomasubsequent to sunlamp, sunbed exposure.”
3.4. Long wavelength ultravioletmelanoma causation data from animal models
It is difficult to definitively assign human data to mechanismsmediated by particular wavelengths because environmental sun
Fig. 1. Transmission to melanocytes of different wavelengths of light through100 μm of epidermis and into the epidermal/dermal junction. The much greatertransmission of UVA and visible light is apparent. The assignment of UVB,UVA, and visible wavebands is also shown.
exposures are to a complex and varying mixture of wavelengths.The most definitive approach to study wavelength-dependentmechanisms has been to determine melanoma causation inanimal models, in which controlled exposures to definedwavelength light sources can be employed. The 2 majormodels in which both UVA and UVB have been shown tocause melanoma are the fish Xiphophorus model and themammalian Monodelphis domestica (Ley et al., 1987; Setlowet al., 1989, 1993; Kusewitt & Ley, 1996; Setlow, 1999; Leyet al., 2000; Ley, 2002; Noonan et al., 2003). Each of thesemodels exhibit both UVA and UVB causation of melanoma, orin the case of M. domestica the causation of the premelanomalesion, focal melanocytic hyperplasia (FMH). There are evensome reports that long-wavelength visible light can causemelanoma in Xiphophorus (Setlow, 1999). As discussedearlier, the only model in which the wavelength dependenceof melanoma causation has been studied is in Xiphophorus.However, there has been considerable controversy becausesusceptibility to UVA melanoma causation is not replicated inalbino models, such as the HGF/SF transgenic mouse, inwhich UVB predominates in inducing melanomas (Noonanet al., 2001; DeFabo et al., 2004). However, the best availablesupportive data summarized above is that melanin is thecentral photosensitizing chromophore in melanocytes that isresponsible for melanoma causation, with a photosensitizingactivity that ranges from UVB (Takeuchi et al., 2004) to visiblelight (Setlow, 1999). The corollary is that albino models withamelanotic melanocytes (Noonan et al., 2001, 2003) will lackthis chromophore, and so only would be sensitive to UVB(DeFabo et al., 2004) that can cause DNA damage directlythrough pyrimidine dimer formation. Furthermore, otherfactors, such as the age of exposure, can be important; evenappropriately pigmented models will only exhibit sensitivity tolonger than UVB wavelengths if UV irradiations are performedwhen the target cells (melanocytes) are expressing significantamounts of melanin. Thus exposures early in development(Robinson et al., 2000) when the melanocytes may not expressmelanin, may not be suitable models for UVA effects, even inotherwise suitable species.
4. Factors that potentiate the importanceof wavelengths N320 nm in melanoma causation
4.1. Long wavelength ultraviolet penetrates skinto melanocytes better than short wavelength ultraviolet
UVB (280–320 nm) is efficiently absorbed directly by DNAto cause damage but penetrates poorly into skin (Bruls et al.,1984). UVA (320–400 nm) is very poorly absorbed by DNA,and its photobiological effects are instead mediated by itsabsorption by other cellular chromophores to generate reactivespecies; however, UVA penetrates into skin much deeper andvisible light yet further, with Fig. 1 showing the penetrance oflight through 100 μm of epidermis (Bruls et al., 1984). This isthe approximate depth of the epidermal/dermal junction wherethe target cells, melanocytes, reside (Takayama et al., 1997;Otsuka et al., 1998; Noonan et al., 2001).
This transmission has been experimentally verified inhuman skin tumors, where deeper cells harbor proportion-ately more UVA-fingerprint mutations than UVB (Agar et al.,2004). Thus, penetration into skin by longer wavelengths isan important factor. This will be even more so in pigmentedmelanocytic nevi, especially if they are of sufficient depth, sothat only longer wavelengths of light can efficiently penetratethe entire depth of this lesion. Since nevi represent a sourceof genetically altered melanocytes for further transformation,this may yet further increase the importance of UVA andlonger wavelengths, as only surface cells can be acted uponby UVB, whereas many more cells throughout the depth ofthe lesion are UVA targets. We have obtained experimentalevidence of this phenomenon in the Xiphophorus model byshowing that in thicker skin of adult fish, the efficiency ofreactive melanin radical (RMR) formation at longer wave-lengths (404 and 434 nm) is increased to that of thinnerskinned juvenile fish (Wood et al., 2006).
4.2. Long wavelength ultraviolet is muchmore prevalent than short wavelength ultraviolet
The total flux of UVA and longer wavelengths at the earth'ssurface vastly exceeds that of UVB, with this differentialincreasing with increasing latitude, decreasing altitude, increas-ing daytime from solar noon, and temporal distance fromsummer solstice. Yet, most behavioral solar protection advice(avoid sun near to noontimes, avoid sunshine during summerdays, avoid sunshine near the equator) will selectively furtherincrease the ratio of UVA to UVB exposure. Furthermore,devices such as sun beds and tanning booths remain popularsources of intense UVexposure used by significant segments ofthe population (Diffey et al., 1990; Chen et al., 1998; Wanget al., 2001; Young, 2004) and there has been an ongoing trendtoward specifically enhancing UVA wavelength irradiance inthese devices (Miller et al., 1998). Thus, the exposure patternsof tanning booth users will even furthermore be weighted toUVA wavelengths.
4.3. Melanin is a dominant short wavelength ultravioletphotosensitizer in melanocytes causing oxidative damage
It has been shown in vitro that illumination of melanin byUVand blue light generates metastable and RMR that can reactwith biomolecules and also react with molecular oxygen andrelease reactive small molecule oxidants, such as superoxide,leading to hydrogen peroxide and hydroxyl radical withequivalent action spectra (Sarna et al., 1984; Sarna & Sealy,1984). We have recently developed techniques to generate anaction spectrum of UV-induced RMR formation in the skin ofthe Xiphophorus hybrid, which is the only model in which theaction spectrum for melanoma is known (Setlow et al., 1993),and show that the 2 are identical. This result indicates that, inthis model at least, melanin photosensitized free radicalproduction is the central event in melanoma causation. Andbecause the action spectrum for this extends far into the UVAand even visible (unlike pyrimidine dimers and 6-4 photo-products that are only caused by UVB in sunlight), this resultalso drives up the importance of UVA. Thus, the componentsthat Tyrrell (1995) shows contribute to a biological actionspectrum all drive the importance of UVA in this process.Although under ‘normal’ circumstances melanin probably actsas a protectant (Meredith & Sarna, 2006) we hypothesize athreshold dose of UV is required, perhaps to deplete or saturatemelanocyte antioxidant defenses, above which the net effect ofmelanin is to be a damaging photosensitizer. Such a hypothesisis in accord with the known epidemiologic (Armstrong, 1988;Moan et al., 1999; Woodhead et al., 1999; Armstrong &Kricker, 2001) and animal model data (Setlow et al., 1993;Noonan et al., 2001) showing that melanoma is most stronglyassociated with episodes of acute, sunburning, UV doses.
4.4. Long wavelength ultraviolet sunscreen effectivenessremains suboptimal compared to short wavelength ultraviolet
Although, or perhaps because, it is currently difficult todetermine UVA protection (Gasparro, 2000), the UVA screen-ing of many newer sunscreens that claim UVA protection issometimes insufficient (Haywood et al., 2003), leading tofurther weighting of UVA to UVB exposures. Although it isdesirable not to reignite speculation that older sunscreenslacking UVA protection may predispose to melanoma causation(Dobak & Liu, 1992; Garland et al., 1993; Setlow, 1999;Christensen, 2003; Huncharek & Kupelnick, 2003; Marshallet al., 2003), sound data shows that selective screening of UVBwavelengths might still be a concern. It is also important to notethat not only is the UVA filter avobenzone susceptible tophotodegradation, but it can photosensitize the degradation ofother sunscreen components (Sayre et al., 2005). Furthermore,although micronized and nanoparticulate inorganic sunscreenssuch as titanium dioxide show high UVA absorption andphotostability, there is ongoing evidence that they can exhibittoxicities of their own (Ahn et al., 2005; Gurr et al., 2005;Hussain et al., 2005; Sayes et al., 2006) that may, in the future,be of regulatory concern. Thus, the development of truly wide-spectrum sunscreens will need to continue.
5. Measures of long wavelengthultraviolet protection by sunscreens
5.1. Defining the requirements forlong wavelength ultraviolet protection measurement
The “gold standard” of sunscreen measurement is the SPF(Tyrrell, 1995; Nash et al., 2006). However, SPF measures anendpoint that is sensitive to UVB, and so does not measure theUVA protection of sunscreens. There remains an unmet need for aSPF-like measure of the ability of sunscreens to protect againstUVA, as exposure to UVA is known to cause a range of adverseoutcomes such as photoaging (Kligman, 1989), immunosuppres-sion (Ullrich, 2005), and melanoma. As a result, there has been aclear trend to increase the UVA protection ability of sunscreensover the last 15–20 years, and yet there is still no consensus onhow best tomeasure UVA protection, or how to communicate thisinformation to consumers. A range of in vitro and in vivomethodshave been proposed formeasuringUVAprotection, but none havemet with the widespread approval and recognition that SPFmeasurement has achieved, primarily because they have not metall of the most important requirements. These have beenexcellently reviewed recently (Nash et al., 2006). An idealmeasure of UVA protection would possess the followingattributes.
It would be determined by the measurement of a relevantbiological endpoint. In other words, it should be a surrogate forprotection against the risks that people use sunscreens to protectagainst and not measure some side-reaction or unimportantfactor.
The technique would be clinically applicable to humans.Although in vitro tests are useful adjuncts, the technique shouldbe able to measure parameters that can only be fully evaluatedin living skin, such as photostability, absorption and perme-ation, and water resistance.
The measurement should be accurate, reproducible, andpreferably objective and not dependent upon interpretation byan observer. In other words, the test should yield the sameresults at different testing sites and in skin-type matchedindividuals; removing operator interpretation of results wouldalso be desirable as this can be a source of inconsistency.Furthermore, it would ideally be a continuous, and notthreshold-dependent, type of measurement.
The technique should be able to measure UVA protectionusing broadband solar-simulated light. This is important, assunscreen filter photostability can be much lower with UVBwavelengths, so that a technique using irradiation with UVAalone will not provide an accurate assessment of how asunscreen will perform under real solar light (Nash et al., 2006).
Exposure levels should be low and at real-life levels. Somecurrent methods require very large UVA doses that may haveadverse effects, and that may also unrealistically stress thephotostability of the sunscreen filters.
The index derived from the measurements should becompatible with ‘dual-labeling’ with SPF and with an easy-to-communicate meaning. The SPF remains a gold standardfor consumer awareness and expectations and for regulatory
controls, and the information in the UVA protection indexshould be able to be listed together with the SPF, and theconsumer should have a ready idea of what this means.
The technique would not suffer from loss of resolution athigher protection factors. Because of the way it is performed,the resolution of SPF actually decreases as SPF increases (i.e., itis easy to differentiate SPF 2 from SPF 4, but more difficult todifferentiate SPF 25 from SPF 30).
It should be able to measure the screening effects of clothing.UV protective clothing is becoming more prevalent, and soconsumers should be able to use the measurement to choosebetween different garments, to choose between garments andsunscreens, or to assess likely effectiveness of multiple strategies.
Applicability to animal models. Since humans are subject tocomplex patterns of UV exposure, and photoeffects take manyyears to develop, rigorous determinations of much of themechanisms and effects of UV are best performed in animalmodels, and then correlated to humans. Thus, an ideal measurewould also be applicable to relevant animal models to correlateprotection with prevention of an outcome.
Cost and speed. The technique should be as inexpensive as iscompatible with providing the best data. Obviously, no industrywould appreciate a large and unnecessary cost being imposedby a regulatory agency, but analyzing the balance between costminimization and quality of consumer protection data shouldbe performed in light of the large size of the sunscreen andphotoprotection market and the prevailing costs of sunscreen (inthe United States, approximately US$10 per bottle). Finally, itwould be convenient if the measures were conducted rapidly,without the need for subsequent recall of volunteers to assessendpoints.
5.2. How do current methods compare?
A number of in vitro spectrophotometric techniques have beenproposed and developed, with the “critical wavelength” approachdeveloped by Diffey et al. (2000) receiving support and attention(Lim et al., 2001; Kullavanijaya & Lim, 2005; Nash et al., 2006),although there have been a range of related techniques proposedand accepted such as the Boots star system in theUnitedKingdom(Diffey et al., 2000). However, although they are undoubtedlyuseful and highly cost-effective, such in vitro techniques are notable to determine several important parameters such as skinphotostability, permeation or water resistance, nor is their directrelevance to clinical/biological endpoints easy to determine. Also,although the concept is readily grasped by a photobiologist, it isunlikely a typical consumer will be able to interpret the meaningand importance of the critical wavelength, or how to use this inproduct comparisons.
Fig. 2. Diagrammatic representation of potential
The in vivo techniques such as immediate pigment darkening(IPD) or persistent pigment darkening (PPD) also termedprotection factor UVA (PFA) suffer from a range of issues:neither reports upon a highly relevant endpoint; both are subjec-tive and prone to interpretation; both use only UVA; and PPDrequires high UVA doses. This lack of an optimal measurementmethod is partially responsible for the regulatory endorsements ofUVAprotection that have been somewhat limited when comparedto the widespread acceptance of SPF measurements. A UVAprotection measure that addresses all the relevant issues raisedwould enable the progress that is needed in this area.
5.3. Development and application ofelectron paramagnetic resonance measurements
It has long been suspected that melanin acts as a UVAphotosensitizer in melanocytes, and that the oxidative damagethis process causes contributes to the causation of melanoma(visually represented in Fig. 2). It has been shown that RMR canalso be observed in UV irradiated pigmented rabbit skin (Collinset al., 1995), suggesting these processes might occur in mela-nocytes in situ in the skin. We have recently developed electronparamagnetic resonance (EPR) techniques to enable accuratemeasurement of RMR in situ in skin, and used these to generatean action spectrum of UV-induced reactive melanin radicalformation in the skin of the Xiphophorus hybrid that is the onlymodel in which the action spectrum for melanoma is known(Setlow et al., 1993). We were able to show that the 2 actionspectra are identical from 303 to 434 nm, a range spanning bothUVB and UVA, showing that measurement of RMR can providea powerful surrogate for processes causing melanoma inpigmented species (Fig. 2). The literature of EPR of melaninsis extensive, and the interested reader is directed to a recentreview of the field for further details (Meredith & Sarna, 2006).
We had also noted that the extent of light-induced radical(LIR) formation at any particular wavelength was proportionalto the square root of irradiance, so that measuring LIR canprovide an accurate measurement of the amount of UV reachingthe melanocyte, the target cell for melanoma (Marrot et al.,1999). The availability of a new generation of low-frequencyEPR spectrometers make these measurements possible in skinin situ using the surface coils developed for other biomedicaluses (Liu et al., 1993; He et al., 2001, 2002; Salikhov et al.,2003; Liu et al., 2004). Thus RMR measurement may be apotentially valuable measurement of a relevant endpoint.
To determine how sensitive RMR measurement is to differentwavelengths, we convoluted the action spectrum of RMRformation (from our Xiphophorus studies) and for comparisonthe CIE action spectrum for erythema, with typical solar emission
spectra to determine the relative sensitivities of RMR anderythema to UVA and UVB fluxes typical of sunlight, or fullspectrum solar simulated light. These data are shown in Fig. 3 andwere normalized for the integrated area under the curve. It can beseen that erythema is overwhelmingly caused by UVB, with over80% of erythema caused by these wavelength and less than 20%by UVA. Thus, a sunscreen that blocked 99.9% of all UVAwavelengths, but not UVB, would have virtually no detectableSPF, as total removal of UVAwould only result in an SPF of 1.25.In contrast, it can be seen that over 95% of measured RMRwouldbe caused by UVA, with less that 5% by UVB, thus a sunscreenthat blocked 99.9% of UVB, but not UVA, would offer virtuallyno protection against RMR formation, and so would have nomelanocyte protection factor (MPF). This waveband selectivityleads to the possibility of using full-spectrum solar simulated lightand obtaining discrete measurements of UVB and UVA protec-tion by simply measuring the appropriate endpoint, SPF for UVBand MPF for UVA. Use of full spectrum solar simulated lightwould overcome many of the disadvantages of measurementtechniques that can only use filtered wavelength ranges. Further-more, the relative ratio of SPF to MPF gives a direct measure ofthe relative performance of UVA and UVB protection; an SPFmuch higher than the MPF would indicate low UVA protection,while equal SPF and MPF would indicate similar effectiveness inblocking UVA and UVB. Finally, the technique fulfills all therequirements detailed in Section 5.1, apart from cost as typicalEPR instrumentation ranges from US$100,000 to US$250,000.However, since the costs of 1 solar simulator and the scanningspectroradiometer required for quality control of its output are aboutUS$100,000, the instrumentation costs are not disproportionate.
5.4. Practical measurement of melanocyte protection factor
As previously detailed melanin also contains stable melaninradicals (SMR) formed upon polymerization, so that accuratemeasurement of LIR can be achieved by measuring the levels of
Fig. 3. The overall contributions of different wavelengths upon causation ofmelanoma and nonmelanoma skin cancer obtained by convoluting typical solaremission spectra (in this case Albuquerque, NM) with the relevant actionspectra: Setlow for melanoma (Setlow et al., 1993), SCUP for NMSC (de Gruijl,1995) normalized for area under curve=1.
SMR (simplymeasured before light irradiation of the sample) and(LIR+SMR) is measured during its irradiation with theappropriate light source by a simple shutter arrangement. Thisis shown diagrammatically in Fig. 4. This ratio is used to derivethe protection provided to the melanocytes in which this index ismeasured, termed the MPF. Since LIR is proportional to thesquare root of light intensity reaching themelanin-containing cell,we can define the ratio of light reaching the melanocyte withoutsunscreen to
MPF ¼ UV reaching melanocyte without sunscreenUV reaching melanocyte with sunscreen
¼SMRþscreen
LIRþUVþScreen
� �2
SMRControlLIRþUV Control
� �2 ð1Þ
How the MPF is calculated from EPR measurements.If a monochromatic protection factor (mPF) is required,
specific filtered wavebands can be used instead of the fullspectrum light used for MPF.
5.5. Wavelength dependence of sunscreen protection
It is also possible to perform similar experiments usingisolated wavebands across the UVB and UVA ranges in amanner similar to that previously used to determine the actionspectrum of RMR formation (Wood et al., 2006). By comparingthe effectiveness of the sunscreen to prevent RMR formation,
the wavelength-specific protection factor (the fold by whichlight of that wavelength is screened from the target cells) can bereadily calculated. Thus, the wavelength dependence of sun-screen effectiveness can also be measured in a way that does notrequire an assumption about the action spectra of a biologicalevent, or the wavelength distribution of the UV light the userwill be exposed to.
6. Conclusions
Continued advances in melanoma prevention will depend notonly upon new understanding of how melanoma is caused butalso in the better application of what is already known about thisdisease. UVA is likely a major causative factor in humanmelanoma, and although sunscreens are widely used for sunprotection, yet it is has been difficult to assess their UVAprotection abilities to a level that is as good as the measurement ofthe UVB protection that is determined by the gold standard ofSPF. Maximizing the UVA protection afforded by sunscreens byproviding new tools to optimize its measurement should helpsignificantly decrease the incidence of UVA induced phenomenasuch as photoaging, immunosuppresssion, and melanoma.
Acknowledgments
The author thanks Marianne Berwick, Ronald Ley, andRichard Setlow for their continued advice and friendship. Electronparamagnetic resonance instrumentation at UNMHSC is fundedin part by the NIH Center of Biomedical Research Excellence(NCRR P20-RR-15636). UV exposure facilities supported byNIEHS P30 ES-012072. This work funded by the AmericanCancer Society (ACSIRG-192) and NIH/NCI (CAA113687).
References
Agar, N. S., Halliday, G. M., Barnetson, R. S., Ananthaswamy, H. N., Wheeler,M., & Jones, A. M. (2004). The basal layer in human squamous tumorsharbors more UVA than UVB fingerprint mutations: a role for UVA inhuman skin carcinogenesis. Proc Natl Acad Sci U S A 101, 4954−4959.
Ahn, M. H., Kang, C. M., Park, C. S., Park, S. J., Rhim, T., Yoon, P. O., et al.(2005). Titanium dioxide particle-induced goblet cell hyperplasia: associ-ation with mast cells and IL-13. Respir Res 6, 34.
Anders, A., Altheide, H. J., Knalmann, M., & Tronnier, H. (1995). Actionspectrum for erythema in humans investigated with dye lasers. PhotochemPhotobiol 61, 200−205.
Armstrong, B. K. (1988). Epidemiology of malignant melanoma: intermittent ortotal accumulated exposure to the sun? J Dermatol Surg Oncol 14, 835−849.
Armstrong, B. K. (2005). Personal communication.Armstrong, B. K., & Kricker, A. (2001). The epidemiology of UV induced skin
cancer. J Photochem Photobiol B Biol 63, 8−18.Baade, P., & Coory, M. (2005). Trends in melanoma mortality in Australia:
1950–2002 and their implications for melanoma control. Aust N Z J PublicHealth 29, 383−386.
Brose, M. S., Volpe, P., Feldman, M., Kumar, M., Rishi, I., Gerrero, R., et al.(2002). BRAF and RAS mutations in human lung cancer and melanoma.Cancer Res 62, 6997−7000.
Bruls, W. A., Slaper, H., van der Leun, J. C., & Berrens, L. (1984). Transmissionof human epidermis and stratum corneum as a function of thickness in theultraviolet and visible wavelengths. Photochem Photobiol 40, 485−494.
Chen, Y. T., Dubrow, R., Zheng, T., Barnhill, R. L., Fine, J., & Berwick, M.(1998). Sunlamp use and the risk of cutaneous malignant melanoma: a
population-based case-control study in Connecticut, USA. Int J Epidemiol27, 758−765.
Christensen, D. (2003). Data still cloudy on association between sunscreen useand melanoma risk. J Natl Cancer Inst 95, 932−933.
Chudnovsky, Y., Khavari, P. A., & Adams, A. E. (2005). Melanoma genetics andthe development of rational therapeutics. J Clin Invest 115, 813−824.
Cole, C. (2001). Sunscreen protection in the ultraviolet A region: how tomeasurethe effectiveness. Photodermatol Photoimmunol Photomed 17, 2−10.
Collins, B., Poehler, T. O., & Bryden, W. A. (1995). EPR persistencemeasurements of UV-induced melanin free radicals in whole skin. Photo-chem Photobiol 62, 557−560.
Davies, H., Bignell, G. R., Cox, C., Stephens, P., Edkins, S., Clegg, S., et al. (2002).Mutations of the BRAF gene in human cancer. Nature 417, 949−954.
de Gruijl, F. R. (1995). Action spectrum for photocarcinogenesis. Recent ResultsCancer Res 139, 21−30.
DeFabo, E., Noonan, F. P., Fears, T., & Merlino, G. (2004). Ultraviolet B but notultraviolet A radiation induces melanoma. Cancer Res 64, 6372−6376.
Dennis, L., Beane Freeman, L., & VanBeek, M. (2003). Sunscreen use and therisk for melanoma: a quantitative review. Ann Intern Med 139, 966−978.
Diffey, B. L. (2005). Sunscreens and melanoma: the future looks bright. Br JDermatol 153.
Diffey, B. L., Farr, P. M., Ferguson, J., Gibbs, N. K., deGruijl, F. R., Hawk, J. L.,et al. (1990). Tanning with ultraviolet A sunbeds [see comments.] [erratumappears in BMJ 1990 Oct 20;301(6757):890.]. BMJ 301, 773−774.
Diffey, B. L., Tanner, P. R., Matts, P. J., & Nash, J. F. (2000). In vitro assessmentof the broad-spectrum ultraviolet protection of sunscreen products. J AmAcad Dermatol 43, 1024−1035.
Dobak, J., & Liu, F. T. (1992). Sunscreens, UVA, and cutaneous malignancy:adding fuel to the fire. Int J Dermatol 31, 544−548.
Edwards, R., Ward, M., Wu, H., Medina, C., Brose, M., P. V., et al. (2004).Absence of BRAF mutations in UV protected mucosal melanomas. J MedGenet 41, 270−272.
Eide, M. J., & Weinstock, M. A. (2005). Association of UV index, latitude, andmelanoma incidence in non-White populations—US surveillance, epidemiol-ogy, and end results (SEER) program, 1992 to 2001. Arch Dermatol 141,477−481.
Eskandarpour, M., Hashemi, J., Kanter, L., Ringborg, U., Platz, A., & Hansson,J. (2003). Frequency of UV-inducible NRAS mutations in melanomasof patients with germline CDKN2A mutations. J Natl Cancer Inst 95,790−798.
Gallagher, R. P., Spinelli, J. J., & Lee, T. K. (2005). Tanning beds, sunlamps, andrisk of cutaneous malignant melanoma. Cancer Epidemiol Biomark Prev14, 562−566.
Gandini, S., Sera, F., Cattaruzza, M. S., Pasquini, P., Picconi, O., Boyle, P., et al.(2005). Meta-analysis of risk factors for cutaneous melanoma: II. Sunexposure. Eur J Cancer 41, 45−60.
Garland, C. F., Garland, F. C., & Gorham, E. D. (1993). Rising trends inmelanoma. An hypothesis concerning sunscreen effectiveness. AnnEpidemiol 3, 103−110.
Garland, C., Garland, F., & Gorham, E. (2003). Epidemiologic evidence fordifferent roles of ultraviolet A and B radiation in melanoma mortality rates.Ann Epidemiol 13, 395−404.
Gasparro, F. P. (2000). Sunscreens; skin photobiology; and skin cancer: the needfor UVA protection and evaluation of efficacy. Environ Health Perspect 108(S1), 71−78.
Goldenberg-Cohen, N., Cohen, Y., Rosenbaum, E., Herscovici, Z., Chowers, I.,Weinberger, D., et al. (2005). T1799A BRAF mutations in conjunctivalmelanocytic lesions. Invest Ophthalmol Visual Sci 46, 3027−3030.
Gorden, A., Osman, I., Gai, W., He, D., Huang, W., Davidson, A., et al. (2003).Analysis of BRAF and N-RAS mutations in metastatic melanoma tissues.Cancer Res 63, 3955−3957.
Grant, W., Garland, C. F., & Holick, M. (2005). Comparisons of estimatedeconomic burdens due to insufficient solar ultravioloet irradiance andvitaminD and excess solar UV irradiance for the United States. PhotochemPhotobiol 81, 1276−1286.
Gurr, J. R., Wang, A. S., Chen, C. H., & Jan, K. Y. (2005). Ultrafine titaniumdioxide particles in the absence of photoactivation can induce oxidativedamage to human bronchial epithelial cells. Toxicology 213, 66−73.
Haywood, R., Wardman, P., Sanders, R., & Linge, C. (2003). Sunscreensinadequately protect against ultraviolet-A-induced free radicals in skin:implications for skin aging and melanoma? J Invest Dermatol 121, 862−868.
He, G., Samouilov, A., Kuppusamy, P., & Zweier, J. L. (2001). In vivo EPRimaging of the distribution and metabolism of nitroxide radicals in humanskin. J Magn Reson 148, 155−164.
He, G., Evalappan, S. P., Hirata, H., Deng, Y., Petryakov, S., Kuppusamy, P.,et al. (2002). Mapping of the B1 field distribution of a surface coilresonator using EPR imaging. Magn Reson Med 48, 1057−1062.
Helmke, B. M., Mollenhauer, J., Herold-Mende, C., Benner, A., Thome, M.,Gassler, N., et al. (2004). BRAF mutations distinguish anorectal fromcutaneous melanoma at the molecular level. Gastroenterology 127,1815−1820.
Huncharek, M., & Kupelnick, B. (2002). Use of topical sunscreens and the riskof malignant melanoma: a meta-analysis of 9067 patients from 11 case-control studies. Am J Public Health 92, 1173−1177.
Huncharek, M., & Kupelnick, B. (2003). Sunscreen use and malignantmelanoma risk: the jury is still out-response. Am J Public Health 93, 12.
Hussain, S. M., Hess, K. L., Gearhart, J. M., Geiss, K. T., & Schlager, J. J.(2005). In vitro toxicity of nanoparticles in BRL 3A rat liver cells. Toxicol inVitro 19, 975−983.
Kligman, L. H. (1989). The ultraviolet-irradiated hairless mouse: a model forphotoaging. J Am Acad Dermatol 21, 623−631.
Kullavanijaya, P., & Lim, H. W. (2005). Photoprotection. J Am Acad Dermatol52, 937−958 (quiz 959-962).
Kumar, R., Angelini, S., Snellman, E., & Hemminki, K. (2004). BRAFmutations are common somatic events in melanocytic nevi. J InvestDermatol 122, 342−348.
Kusewitt, D. F., & Ley, R. D. (1996). Animal models of melanoma. Cancer Surv26, 35−70.
Ley, R. D. (2002). Animal models of ultraviolet radiation (UVR)-inducedcutaneous melanoma. Front Biosci 7, d1531−d1534.
Ley, R. D., Applegate, L. A., Stuart, T. D., & Fry, R. J. (1987). UV radiation-induced skin tumors in Monodelphis domestica. Photodermatology 4,144−147.
Ley, R. D., Reeve, V. E., & Kusewitt, D. F. (2000). Photobiology of Mono-delphis domestica. Dev Comp Immunol 24, 503−516.
Lim, H. W., Naylor, M., Honigsmann, H., Gilchrest, B. A., Cooper, K., Morison,W., et al. (2001). American Academy of Dermatology ConsensusConference on UVA protection of sunscreens: summary and recommenda-tions. Washington, DC, Feb. 4, 2000. J Am Acad Dermatol 44, 505−508.
Liu, K. J., Gast, P., Moussavi, M., Norby, S. W., Vahidi, N., Walczak, T., et al.(1993). Lithiumphthalocyanine: a probe for electron paramagnetic resonanceoximetry in viable biological systems. Proc Natl Acad Sci U S A 90,5438−5442.
Liu, S., Shi, H., Liu, W., Furuichi, T., Timmins, G. S., & Liu, K. J. (2004).Interstitial pO2 in ischemic penumbra and core are differentially affectedfollowing transient focal cerebral ischemia in rats. J Cereb Blood FlowMetab 24, 343−349.
Longstreth, J. (1988). Cutaneous malignant melanoma and ultraviolet radiation:a review. Cancer Metastasis Rev 7, 321−333.
Marrot, L., Belaidi, J. P., Meunier, J. R., Perez, P., & Agapakis-Causse, C.(1999). The human melanocyte as a particular target for UVA radiation andan endpoint for photoprotection assessment. Photochem Photobiol 69,686−693.
Marshall, S., Poole, C., & Waller, A. (2003). Sunscreen use and malignantmelanoma risk: the jury is still out. Am J Public Health 93, 11−12.
Meredith, P., & Sarna, T. (2006). The physical and chemical properties ofeumelanin. Pigment Cell Res 19(6), 572−594.
Michaloglou, C., Vredeveld, L. C., Soengas, M. S., Denoyelle, C., Kuilman, T.,van der Horst, C. M., et al. (2005). BRAFE600-associated senescence-likecell cycle arrest of human naevi. Nature 436, 720−724.
Miller, S. A., Hamilton, S. L., Wester, U. G., & Cyr, W. H. (1998). An analysisof UVA emissions from sunlamps and the potential importance formelanoma. Photochem Photobiol 68, 63−70.
Moan, J., Dahlback, A., & Setlow, R. B. (1999). Epidemiological support for anhypothesis for melanoma induction indicating a role for UVA radiation.Photochem Photobiol 70, 243−247.
Nash, J. F., Tanner, P. R., & Matts, P. J. (2006). Ultraviolet A radiation: testingand labeling for sunscreen products. Dermatol Clin 24, 63−74.
Noonan, F. P., Recio, J. A., Takayama, H., Duray, P., Anver, M. R., Rush, W. L.,et al. (2001). Neonatal sunburn and melanoma in mice. Nature 413,271−272.
Noonan, F., Dudek, J., Merlino, G., & De Fabo, E. (2003). Animal models ofmelanoma: an HGF/SF transgenic mouse model may facilitate experimentalaccess to UV initiating events. Pigment Cell Res 16, 16−25.
Oliveria, S., Dusza, S., & Berwick, M. (2001). Issues in the epidemiology ofmelanoma. Expert Rev Anticancer Ther 1, 453−459.
Otsuka, T., Takayama, H., Sharp, R., Celli, G., LaRochelle, W., Bottaro, D.,et al. (1998). c-Met autocrine activation induces development of malig-nant melanoma and acquisition of the metastatic phenotype. Cancer Res58, 5157−5167.
Papp, T., Pemsel, H., Zimmermann, R., Bastrop, R., Weiss, D. G., &Schiffmann, D. (1999). Mutational analysis of the N-ras, p53, p16INK4a,CDK4, and MC1R genes in human congenital melanocytic naevi. J MedGenet 36, 610−614.
Patton, E. E., Widlund, H. R., Kutok, J. L., Kopani, K. R., Amatruda, J. F.,Murphey, R. D., et al. (2005). BRAF mutations are sufficient to promotenevi formation and cooperate with p53 in the genesis of melanoma. CurrBiol 15, 249−254.
Pollock, P. M., Harper, U. L., Hansen, K. S., Yudt, L. M., Stark, M., Robbins,C. M., et al. (2003). High frequency of BRAF mutations in nevi. Nat Genet33, 19−20.
Ragnarsson-Olding, B. K., Karsberg, S., Platz, A., & Ringborg, U. K. (2002).Mutations in the TP53 gene in human malignant melanomas derived fromsun-exposed skin and unexposed mucosal membranes. Melanoma Res 12,453−463.
Ries, L. A. G., Eisner, M. P., Kosary, C. L., Hankey, B. F., Miller, B. A., Clegg,L., et al. (2004). SEER Cancer Statistics Review, 1975-2001. http://seer.cancer.gov/csr/1975_2001
Robinson, E. S., Hill, R. H., Jr., Kripke, M. L., & Setlow, R. B. (2000). Themonodelphis melanoma model: initial report on large ultraviolet Aexposures of suckling young. Photochem Photobiol 71, 743−746.
Salikhov, I., Hirata, H., Walczak, T., & Swartz, H. M. (2003). An improvedexternal loop resonator for in vivo L-band EPR spectroscopy. J Magn Reson164, 54−59.
Sarna, T., & Sealy, R. C. (1984). Free-Radicals from Eumelanins: quantumYields and Wavelength Dependence. Arch Biochem Biophys 232, 574−578.
Sarna, T., Menon, I. A., & Sealy, R. C. (1984). Photoinduced Oxygen-Consumption in Melanin Systems .2. Action Spectra and Quantum Yieldsfor Pheomelanins. Photochem Photobiol 39, 805−809.
Sayes, C. M., Wahi, R., Kurian, P. A., Liu, Y., West, J. L., Ausman, K. D., et al.(2006). Correlating nanoscale titania structure with toxicity: a cytotoxicityand inflammatory response study with human dermal fibroblasts and humanlung epithelial cells. Toxicol Sci 92, 174−185.
Sayre, R. M., Dowdy, J. C., Gerwig, A. J., Shields, W. J., & Lloyd, R. V. (2005).Unexpected photolysis of the sunscreen octinoxate in the presence of thesunscreen avobenzone. Photochem Photobiol 81, 452−456.
Setlow, R. B. (1999). Spectral regions contributing to melanoma: a personalview. J Investig Dermatol Symp Proc 4, 46−49.
Setlow, R. B., &Woodhead, A. D. (1994). Temporal Changes in the Incidence ofMalignant-Melanoma: explanation from Action Spectra. Mutat Res 307,365−374.
Setlow, R. B., Woodhead, A. D., & Grist, E. (1989). Animal model forultraviolet radiation-induced melanoma: platyfish-swordtail hybrid. ProcNatl Acad Sci U S A 86, 8922−8926.
Setlow, R. B., Grist, E., Thompson, K., & Woodhead, A. D. (1993). Wavelengthseffective in induction of malignant melanoma. Proc Natl Acad Sci U S A 90,6666−6670.
Takayama, H., LaRochelle, W., Sharp, R., Otsuka, T., Kriebel, P., Anver, M.,et al. (1997). Diverse tumorigenesis associated with aberrant developmentin mice overexpressing hepatocyte growth factor scatter factor. Proc NatlAcad Sci U S A 94, 701−706.
Takeuchi, S., Zhang, W., Wakamatsu, K., Ito, S., Hearing, V., Kraemer, K., et al.(2004). Melanin acts as a potent UVB photosensitizer to cause an atypicalmode of cell death in murine skin. Proc Natl Acad Sci U S A, 15076−15081.
Tyrrell, R. M. (1995). Biological dosimetry and action spectra. J PhotochemPhotobiol B Biol 31, 35−41.
Ullrich, S. E. (2005). Mechanisms underlying UV-induced immune suppression.Mutat Res 571, 185−205.
Wang, S., Setlow, R., Berwick, M., Polsky, D., Marghoob, A., Kopf, A., et al.(2001). Ultraviolet A and melanoma: a review. J Am Acad Dermatol 44,837−846.
Whiteman, D. C., Stickley, M., Watt, P., Hughes, M. C., Davis, M. B., & Green,A. C. (2006). Anatomic site, sun exposure, and risk of cutaneous melanoma.J Clin Oncol 24, 3172−3177.
Wong, C., Fan, Y., Chan, T., Chan, A., Ho, L., Ma, T., et al. (2005). BRAF andNRAS mutations are uncommon in melanomas arising in diverse internalorgans. J Clin Pathol 58, 640−644.
Wood, S. R., Berwick, M., Ley, R. D., Walter, R. B., Setlow, R. B., & Timmins,G. S. (2006). UV causation of melanoma in Xiphophorus is dominated bymelanin photosensitized oxidant production. Proc Natl Acad Sci U S A 103,4111−4115.
Woodhead, A. D., Setlow, R. B., & Tanaka, M. (1999). Environmental factors innonmelanoma and melanoma skin cancer. J Epidemiol 9, S102−S114.
Young, A. R. (2004). Tanning devices—fast track to skin cancer? Pigment CellRes 17, 2−9.
