Fluorescence of black skin cancer
Berlin, 6 june 2007
Matthias Scholz and Dieter Leupold
Due to the increasing incidence of skin cancer, the interest in using the autofluorescence of human skin tissue as a non-invasive tool for early detection of malignant degeneration is enforced. Focus is especially on malignant melanotic melanoma1, the most dangerous type of skin cancer with a high probability of early induction of metastases. - Melanin synthesis is a main phenotype in the pathophysiology in neoplastic transformation of melanocytes2, and the melanin function in tissue in response to sunlight has been characterized as that of a “two-edged sword”3: protective like a parasol or deleterious by induction of (or at least contributing to) malignant degeneration4,5. It therefore appears to be well-founded to expect indications of malignant transformations in the melanin fluorescence from tissue. But melanin has an extremely low quantum yield, so that in the conventionally excited skin autofluorescence melanin emission is masked by the contributions from several other fluorophores. This drawback can be overcome by a stepwise two-photon fluorescence excitation, unique for melanin. Realized for the first time with nanosecond laser pulses, a specific fluorescence spectrum of malignant melanoma has been found. Its use as a basis for early detection of black skin cancer has just been started, with promising first results.
Despite long-standing research activities fluorescence diagnostics have not played a role in the differential diagnosis of melanocytic pigmented lesions. The main reason is that the autofluorescence of skin tissue is usually the sum of contributions from several endogeneous and possibly also exogeneous fluorophores which depends on a manifold of internal and external factors, e.g. the excitation conditions. Since recently however, there has been a special attempt to follow the melanocyte degeneration via melanin fluorescence.
The main obstacle to identify melanin fluorescence in the complex autofluorescence spectrum of skin tissue is its fluorescence quantum yield in the order of 10-4.
A first step towards a relative enhancement of melanin fluorescence compared to the fluorescence of all other skin fluorophores was realized by the laser-based process of two-photon excitation of fluorescence. It is already standard technique to excite fluorophores in tissue, instead by one photon with an energy corresponding to the near UV or blue spectral range, rather by two photons with half of that energy in order to realize a larger penetration depth due to the so-called spectral window of skin tissue. With respect to melanin this offers a further, special adventage: When using fs laser pulses in the spectral range around 800 nm, simultaneous two-photon absorption allows population of fluorescent levels red-shifted from 400 nm. Examples include flavines, elastin, lipo-pigments, porphyrins and to a certain extent also NADH. All these fluorophores have negligibly small one-photon absorption cross-sections in the 800 nm range. Fundamentally different from this is melanin which has a sufficient
(one-photon) absorption there, so that with 800 nm fs pulses a stepwise two-photon absorption can be realized (assuming that the excited state absorption is also sufficient). It results in a new, green-yellow emission. As has been shown for several melanin solutions, the spectral band shape of this emission is somewhat different and red-shifted as compared to the 400 nm (one-photon) excited fluorescence6,7. A schematic representation of the two excitation mechanisms is shown in fig. 1.
Fig.1
The consequences of the differing mechanisms for population of the two-photon excited fluorescent levels are obvious from the analytical expressions for their normalized population densities:
nsim(Imax, t0) = 0.38 σ(2)tI2max for simultaneous two-photon absorption (1)
nstep(Imax, t0) = 0.14 σ1σ2t2I2max for stepwise two-photon absorption (2)
Here, Imax is the peak intensity of the laser pulse at time t0 , t is the pulse duration
(FWHM), σ(2) is the two-photon absorption cross section, σ1 and σ2 are the one-photon absorption cross sections of the two consecutive absorption steps (for further details cp. ref.7). These population densities are proportional to the respective fluorescence intensity. Assuming for example t=100fs and average values for the cross sections, from (1, 2) an at least two orders of magnitude higher population density nstep as compared to nsim results.
This means, with two-photon excitation the low fluorescence quantum yield of melanin can be partly compensated and may result in detectable melanin contributions to the autofluorescence of human skin tissue.
This has been demonstrated experimentally: Excitation of pigmented skin tissue with two photons from an 800nm fs laser results in a certain change of the spectral shape of the autofluorescence as compared to corresponding excitation with one photon at 400 nm. This change can be attributed to melanin fluorescence6.
Equations (1), (2) indicate a possibility for a further relative enhancement of stepwise excited melanin fluorescence as compared to simultaneous two-photon excitation of the other fluorophores: nstep (melanin) is proportional to the square of the pulse duration, t2, whereas nsim (usual fluorophores) is proportional to t. This means, within the validity range of the above equations, prolongation of the pulse duration results in a strong discrimination of the two processes in favour of melanin. If, for example, t changes from 10-13s to 10-11s, a decrease of Imax by 2 orders of magnitude gives no change in nstep , but a corresponding decrease by 2 orders of magnitude in nsim.
To make use of this theoretical result in practice suggests the observation of this type of melanin fluorescence not only with fs-pulse excitation, but with ps- or, better still, ns-pulse excitation. Such experiments had previously failed until recently, when we adopted a sensitive detection system with a multichannel multiplier. Now the two-photon excited melanin fluorescence can be measured at least up to the longest pulse durations we have in use (8 ns).
This ns-pulse excitation allows us to optically extract the melanin fluorescence from the autofluorescence of skin tissue by proper adjustment of pulse duration and intensity. As an example, in fig. 2 (a) the fluorescence from a malignant melanotic melanoma (formalin-fixed, paraffin embedded, for further characterization cp. figure legend), two-photon-excited with 810nm/2.5ns pulses, is shown. This fluorescence is characterized by a dominating band peaking at 600 nm. The same fluorescence behaviour has been observed with several other histopathologically characterized melanotic melanomas. Therefore it can be stated, that the occurrence of this 600 nm-band is a fingerprint of malignant melanoma tissue. In melanocytic pigmented lesions of common naevus-type, under the same excitation conditions, a fluorescence band in the blue spectral range, peaking at around 460 nm,is dominating. This fact is exemplified by fig. 2(b), taken from a melanocytic compound naevus of Clark type (also formalin-fixed, paraffin embedded). The same spectrum as shown in fig. 2(b) was obtained in the vast majority of all investigated areas (50µm in diameter each) of several other histopathological naevi preparations (other compound naevi, junctional naevi, Spitz naevus). In few areas the fluorescence spectrum looks like an intermediate between the band shapes of fig 2(a) and 2(b). In these cases, in the histopathological findings of the respective naevi, there are hints to melanocytic degenerations.
Fig. 2(a)
Fig. 2(b)
Therefore the intensity ratio of the two fluorescence bands may serve as an indicator of degeneration from benign, common melanocytic lesions via naevi with melanocytic dysplasia8 to malignant melanoma. Very probably, the both fluorescence bands shown in fig. 2(a) and 2(b) belong to different types of melanin (eumelanin, pheomelanin), which is currently under intense investigation.
Preliminary results with freshly excised skin tissue samples (measured prior to the formalin treatment) indicate a similar fluorescence spectroscopic behaviour as the corresponding histopathological samples. In particular there is a more than 100nm difference between the maxima of the “benign” and “malign” fluorescence bands. This holds true even for carcinoma in situ. Therefore, by the described non-invasive method early detection of black skin cancer will be possible.
In summary, the possibility to separate the fluorescence of melanin from that of other species in mixtures of fluorophores by stepwise two-photon excitation with ns-pulses allows a very sensitive attempt for the investigation of melanin-based problems, e. g. in dermatology, ophthalmology, pharmacy and cosmetics. One special example is early detection of black skin cancer.
References
1. Gray-Schopfer, V., Wellbrock, C. & Marais, R. Melanoma biology and new
targeted therapy. Nature 445, 851-857 (2007).
2. Jimbow, K. Salopek T.G., Dixon, W.T., Searles, G. E. & Yamada, K. C. The
epidermal melanin unit in the pathophysiology of malignant melanoma. Am J Dermatopathol.13, 179-188 (1991).
3. Hill, H. Z. Melanin - the two-edged sword? Photochem. Photobiol. (Spec. Iss.) 63, 41S-42S (1996).
4. Wood, S. R. et al. UV causation of melanoma in Xiphorus is dominated by melanin photosensitized oxidant production. PNAS 103, 4111 – 4115 (2006).
5. Lin, J. Y. & Fisher D. E. Melanocyte biology and skin pigmentation. Nature 445, 843-850 (2007).
6. Teuchner, K. et al. Femtosecond Two-photon excited Fluorescence of
Melanin. Photochem Photobiol 70, 146-151 (1999).
7. Teuchner, K. et al. Fluorescence Studies of Melanin by Stepwise Two-
Photon Femtosecond Laser Excitation. J Fluorescence 10, 275-281 (2000).
8. Pavel, S. et al. Disturbed melanin synthesis and chronic oxidative stress in
dysplastic naevi. Europ. J. Cancer 40, 1423-1430 (2004).
Figure legends
Fig.1 Mechanisms of two-photon excited fluorescence
A, simultaneous two-photon absorption via a virtual intermediate level with a two-photon absorption cross section σ(2)
B, stepwise absorption of two photons via a real intermediate level (excited electronic state) with one-photon absorption cross sections σ1 and σ2
Fig. 2 Two-photon excited fluorescence of pigmented melanocytic lesions of human skin tissue (histopathological samples of standard preparation, formalin-fixed, paraffin embedded). Excitation pulse wavelength 810 nm, pulse duration 2.5 ns
Fig. 2(a) Representative example of fluorescence from a nodular malignant melanoma of Clark level V, maximum tumor thickness 7.7 mm, measured in the center of the tumor, 3.0 mm below the skin surface.
Fig. 2(b) Representative example of fluorescence from a compound naevus of Clark type
Acknowledgements We thank R. Eichhorn, K. Hoffmann and M. Stücker for making available the skin tissue samples and G. Stankovic for assistance. This work was supported in part by a grant from the Bundesministerium für Bildung und Forschung der Bundesrepublik Deutschland
<zurück>