Two-photon excitation fluorescence correlation spectroscopy (TPFCS) has been applied in connection to measurements of the point spread function (PSF) for quantitative analysis of sulphorhodamine B (SRB) in excised human skin. The PSF was measured using subresolution fluorescent beads embedded in the skin specimen. The PSF, measured as full width at half maximum (FWHM) was found to be 0.41 ± 0.05 µm in the lateral direction, and 1.2 ± 0.4 μm in the axial direction. The molecular diffusion of SRB inside the skin ranged between 0.5 and 15.0 × 10−8 cm2/s. The diffusion coefficient is not dependent on depths down to 40 µm. The fluorophores were found to accumulate on the upper layers of the skin. This work is the first TPFCS study in human skin. The results show that TPFCS can be used for quantitative analyses of fluorescent compounds in human skin.
© 2010 OSA
Two-photon laser scanning fluorescence microscopy (TPLSM) has become a powerful tool for performing imaging of optically dense biological tissue [1–4] such as human skin. The main advantage of using TPLSM rather than confocal fluorescence microscopy is the enhanced penetration depth, which is an effect of the non-linearity of the excitation process in combination with the lower absorption coefficient for the near infrared (NIR) light. Due to the non-linearity, a confocal pinhole is not necessary for TPLSM, since all emitted photons originate from the focal spot, giving inherent optical sectioning properties. In addition, a more effective detection setup with direct, not descanned, detectors can be used.
TPLSM has been successfully applied for microscopic investigations of intact human skin. For example, TPLSM has been used for imaging of normal skin [5–7], studying transdermal drug delivery [8,9] and for clinical studies of skin cancer [10,11]. The human skin is an organ with a highly complex structure, composed of several layers with unique optical properties . The outermost layer, i.e. the epidermis, can be further subdivided into different layers distinguished by variations in cell size, shape and structure [13,14]. The topmost of these layers is the stratum corneum (SC) which has a thickness between 10 and 20 μm and consists mostly of dead flake-like cells lacking nuclei. SC forms an effective barrier towards external influences such as xenobiotics, nanoscale particles, different sorts of radiation and protects us from dehydration. Below the SC is the stratum granulosum (SG) layer, which consists of regularly distributed flattened cells. In TPLSM, SG is distinguished by fluorescent cells with dark cell nuclei growing in a honeycomb pattern . Further down in the epidermis are the layers stratum spinosum and stratum basale, which are visually characterised by decreasing cell size in the lateral direction.
Even though TPLSM provides valuable three-dimensional information regarding the morphology and distribution of fluorophores in human skin, quantitative analysis is problematic. Because of the scattering and absorbing properties of biological tissue, the excitation light intensity has to be increased with increasing imaging depth. Thus, the relation between the recorded fluorescence signal and the amount of fluorophores in the sample is complex. Attempts have been made to use TPLSM for investigating skin penetration using different delivery systems , but only relative information can be extracted. One possible way to extract data is to combine TPLSM with fluorescence correlation spectroscopy (FCS) . In FCS, the laser beam is not scanned across the sample, but focussed at a specific spot . Diffusing fluorophores will randomly move in and out of the excitation volume, which can give rise to fluctuations in the fluorescence signal. Mathematic autocorrelation analysis of these fluctuations gives information about the diffusion time and the amount of fluorophores in the focal volume, which is related to the diffusion coefficient and the fluorophore concentration, respectively. Similar to for laser scanning microscopy, two-photon excitation can be used in FCS (TPFCS) to enhance the measurement depth in biological tissue.
As suggested earlier, TPFCS can be used for measuring the diffusion properties in tumours . In order to perform detailed quantification analysis using FCS, the actual size of the focal volume must be determined. Since the optical properties of human skin are complex, we present a method where the point spread function (PSF) of the system is determined in connection to the TPFCS measurements. Earlier reports of PSF measurements in human skin are limited to measurements on fluorescent beads placed on top of skin samples cut at 45 degree angles . This might not yield the true PSF as the beads are actually not located in the skin. In addition, the interface between the skin and the beads might cause additional optical distortion. Thus, in this study we measure the PSF on beads embedded in human skin, to give a more correct measurement of the focal volume. The results from these measurements were then incorporated in the TPFCS measurements for quantitative analysis of the diffusion of sulphorhodamine B (SRB) applied to excised human skin.
To our knowledge, no previous study of TPFCS in human skin has been reported. The work presented in this paper shows that it is a useful method for quantifying the concentration of fluorophores at different locations in the skin. This is of great importance, e.g. for topical drug delivery studies.
2. Materials and Methods
2.1. Experimental Setup
The TPFCS system, hereafter called the MRC600 microscope, is shown in Fig. 1 It consists of a modified inverted Zeiss Axiovert 135 TV microscope (Carl Zeiss...), a Bio-Rad MRC600 scan box (Bio-Rad CellScience, UK, now Zeiss CellScience) and two external GaAsP photomultiplier tubes (Hamamatsu, Japan) (PMT A and PMT B in Fig. 1). A 63x water immersion objective (Zeiss ‘C-Apochromat’ 63x/1.2 W Corr) was used for the TPFCS and imaging. A mode-locked femtosecond pulsed Tsunami Ti:Sapphire laser, pumped by a Millennia Xs 10W diode-pump laser (Spectra-Physics, California, USA) was used as excitation light source. This has a wavelength output range of 700 to 1000 nm and 83 MHz repetition rate. The excitation wavelength was set to 840 nm and the pulse characteristics are shown in Fig. 1. The pump power was set to 5.50 W and remained unchanged during the measurements. The laser power at the sample was controlled by a Pockels cell (Conoptics, Connecticut, USA), and measured through a 10x dry objective with similar transmission as the 63x objective. Laser powers in the rage 0.5 – 25 mW were used. Generally, the power was increased when measuring deeper into the tissue.
A 570 LP dichroic mirror separates the emission into a green and a red channel. In front of the PMT for the green channel is a 525/50 emission filter and in front of the PMT for the red channel is a 580/150 emission filter. The TPFCS measurements were made using ALV software and the fluorescence intensity was hardware autocorrelated by a ALV-5000 card (ALV-Laser Vertriebsgesellschaft m.b.H, Germany).
2.2. Sample preparation
Human skin specimens from Caucasian females, collected as leftovers from breast reduction surgery, were used in this work. The specimens were collected in direct connection to surgery, cut into 1 x 1 cm pieces and stored in −70 ° C for a maximum of six months. The samples were thawed in room temperature before use. Two skin samples from the same patient were prepared for each experiment, in order to measure the PSF and TPFCS under similar conditions. As much as possible of the subcutis was removed by a scalpel. In two of the three experiments, the samples were tape stripped 30 times by using a conventional surgical tape to remove the SC to facilitate the penetration of fluorescent beads into the sample.
The PSF measurements were performed by imaging orange subresolution fluorescent beads (PS-Speck Microscope Point Source Kit, Molecular Probes) with excitation/emission maxima at 540/560 nm and a diameter of 0.175 ± 0.005 μm. For the application of beads into skin the sonicated ready-to-use PS-Speck bead suspension was diluted 1:50 in PBS. For PSF measurements in solution 2 µl of undiluted bead suspension was mounted between a regular microscope glass slide and a Nr. 1.5 cover slip.
The fluorescent dye Sulforhodamine B (SRB) (Sigma-Aldrich, Steinheim, Germany), which has an emission maximum at 586 nm, was used as a model drug for the TPFCS experiment. SRB was diluted in PBS to a concentration of 3.4 µM before application to the skin samples. As a control, TPFCS was also performed on different concentrations (340, 34 and 3.4 nM) of SRB in solution,
The skin samples were mounted in flow-through diffusion chambers  for application of PBS solution containing either subresolution beads or SRB. The acceptor chamber was filled with phosphate-buffered saline (PBS) (Sigma-Aldrich, Steinheim, Germany), the diffusion cell was covered with parafilm and aluminium foil and was kept at a constant temperature of 30° C for 20 hours. Then the sample was thoroughly washed with PBS, dried with Kleenex and mounted between a microscope slide and a home-made image chamber. The imaging chamber consisted of a Nr 1.5 cover slip (Menzel-Gläser, Braunschweig, Germany), with a measured thickness of 0.18 mm, and double-coated adhesive tape allowing sufficient space for the sample.
The skin sample exposed to either the subresolution fluorescent beads or the SRB solution was placed on the MRC600 microscope stage. For the PSF measurements the skin surface, i.e. the zero level for the tissue depth, was found by imaging the skin autofluorescence using the green channel. This zero level was defined in different ways depending on the pre-treatment of the skin. For the unstripped samples, the SC containing flake-like cells with high fluorescence was set as the zero level. For the tape-stripped samples, where the SC had been removed, the SG with cells appearing in a honeycombed pattern was set as zero. Then the specimen was scanned using the red channel while focusing deeper into the skin with a step size of 1 µm. When one or more beads could be visualized their depth was noted and a z-stack of the bead/beads was collected. In this study, beads were imaged at depths ranging from 0 – 40 µm in the skin. The image size in the lateral plane of the MRC600 microscope is 768 x 512 pixels. For the smallest possible step size (0.18 µm), an xy-pixel size of 0.03µm (zoom factor 10), a Kalman average of 4 frames and a scan speed of 4 seconds per frame was used for collecting the z-stacks of the beads.
For the TPFCS measurements the zero level of the skin was found by imaging the SRB fluorescence intensity using the red channel. The level with the highest intensity was set as the surface. The depth was measured by using the focus motor to step 1 µm into the tissue. The single-plane images, taken to record the positions of the TPFCS measurements, were done using the red channel for visualizing SRB. A zoom factor of 1 (xy-pixel size 0.3 µm), a scan speed of 4 seconds per frame and a Kalman average of 4 frames were used.
2.4. TPFCS-data collection
The skin sample that had been exposed to SRB was mounted in the MRC600-microscope. The sample was first studied visually by scanning, using the red channel, as described above. Once the areas of interest (inside or outside cells and at different depths) had been identified, the laser beam was focussed at the desired spot. The fluorescence intensity was recorded for 10 seconds and this measurement was repeated 15 times for each position. Each run was autocorrelated and saved individually for analysing. For each FCS measurement, an image was recorded and the location of the measurement noted.
2.5. PSF Analysis
The analysis of the PSF of the MRC600 microscope was done by imaging subresolution beads in three dimensions, i.e. taking so-called z-stacks of the beads. The PSF of a TPLSM/TPFCS setup (with other words; the three-dimensional resolution, the detection volume or the focal volume) is the same as the excitation volume, as opposed to in confocal imaging/FCS where a pinhole is necessary for limiting the detection to the focal plane. Hereafter this volume will be referred to as either the PSF or the focal volume.
Two different softwares were used for the PSF analysis: ImageJ (National Institutes of Health, USA), for inspecting the z-stacks and making the axial reconstructions of the beads, and MatLab for the intensity analysis.
The intensity distribution from line profile through a cross section of the beads, both in the lateral and axial direction, was described by a one dimensional Gaussian probability density function:Fig. 2 .
The full width at half maximum (FWHM) is related to the Gaussian standard deviation, σ, by:
The width of the PSF, ω, corresponding to the width at which the intensity had has decreased to 1/e, is related to the FWHM by:
Knowing the 1/e widths in lateral, ωxy, and axial, ωz, directions, the two-photon focal volume (VTPE) can be estimated by :
The measured values of ωxy and ωz were compared to those calculated from the empirically derived equations .
The PSFs were measured from the recorded fluorescence of the beads in both the axial and lateral directions by fitting Eq. (1) to the intensity profile data. The intensity profiles were obtained by selecting the pixels located in a straight line through the intensity maximum. The optical properties of the skin sometimes caused non-uniform astigmatic distortion of the PSF in the axial direction. In these cases the intensity profile was obtained in a slightly tilted direction to ensure correct determination of ωz .
2.6. TPFCS Analysis
The autocorrelation function (G(τ)) of the fluorescence fluctuations, δF(t), is defined as :
F(t) is the time dependent fluorescence intensity, 〈F(t)〉 is the average of the fluorescence intensity and τ is the time lag. If the focal volume is assumed to be a three-dimensional Gaussian distribution, then the autocorrelation function in Eq. (7) can be calculated for free three-dimensional diffusion :
Here r 0 and z 0 are the 1/e2-widths of the measurement volume (i.e. the PSF), which are related to the FWHM by:
τD is the average diffusion time across the measurement volume and N is the average number of molecules in this volume. If Eq. (9) is fitted to the autocorrelation data, then the number of molecules, N, and the diffusion time, τD, can be obtained. Usually, the parameter z 0/ r 0 is combined to the so-called structure parameter S, which describes how elongated the PSF is in the axial direction compared to the lateral direction. S can either be fixed by a PSF measurement or allowed to vary during the fit. The diffusion coefficient, D, can then be calculated from r 0 and the diffusion time obtained by the fitting procedure according to
The fluorescence trace from each individual TPFCS run was inspected manually by using Matlab. The autocorrelated data from these runs were rejected if the fluctuating fluorescence signal deviated strongly from a horizontal mean line or if it contained a peak higher than ~40% of the mean value. These deviations indicate possible movements in the experimental equipment or in the skin specimen due to heating or other instabilities. An average of 6 out of 15 runs was saved for each measurement. Saving all of the runs, would have yielded deviations that strongly contribute to an incorrect correlation, leading to misinterpretation of the diffusion of the fluorophore. The autocorrelation function for three-dimensional diffusion Eq. (9) was fitted to the average value of the saved runs, as described above. The structure parameter S was either set to values obtained from the PSF measurements or was free to vary.
3.1. PSF measurements
In order to measure the PSF in human skin, excised skin samples were exposed to subresolution fluorescent beads and imaged using TPLSM by the MRC600 microscope. Figure 3 shows a TPLSM image obtained from skin exposed to orange beads for 20 hours. This image is taken with a second TPLSM setup (Bio-Rad Radiance 2100 MP), dedicated for imaging. This microscope was used to be able to visualize the beads together with the weak autofluorescence, since the green channel at the MRC600 microscope has a too narrow emission filter bandwidth to image the autofluorescence deeper than the very top layer in the skin. The green signal corresponds to the autofluorescence, from which the tissue morphology can be determined, while the red signal corresponds to the fluorescence emitted from the beads. As shown by the figure, the fluorescent beads are embedded in the skin. The honeycomb structure of the SG layer with cells containing non-fluorescent nuclei can be distinguished at a depth of 20 µm.
The PSF of beads located at different depths ranging from 0 µm down to 40 µm into the skin tissue was investigated by imaging using the MRC600 microscope. Figure 4 shows typical intensity profiles obtained in the lateral and axial directions of a fluorescent bead located at a depth of 20 μm in a sample of human skin. As illustrated by the figure, the images in the lateral direction showed a symmetric circular disc with the intensity maxima located at the centre, while the cross section of the image in the axial direction has an elongated shape, as expected. The mean value ( ± std) of the FWHM for the depths (0 – 40 µm) was 0.41 ± 0.05 µm in lateral direction and 1.2 ± 0.4 µm in the axial direction. These values, in both the lateral and axial directions, are higher than the calculated values of 0.27 µm and 0.69 μm, respectively, obtained by using Eqs. (5) and (6).
Interestingly, for the examined range of depths (0—40 µm), the widths of the FWHMs were not found to depend on the depth, see Fig. 5(a) and 5(b), and hence neither did the focal volume, see Fig. 5(c), calculated from Eq. (4). Linear regressions of the data in Fig. 5 gave R2 = 0.03 in the lateral direction, R2 = 0.01 in the axial direction and R2 = 0.01 for the volume. The size of the PSF volume ranged between 0.2 and 0.8 fl, with an overall mean value of 0.39 ± 0.17 fl. No difference in PSF was observed when comparing the tape-stripped samples with the samples not receiving any pretreatment. The corresponding PSF widths that were measured in solution were 0.41 ± 0.03 µm in the lateral direction, and 1.1 ± 0.4 µm in the axial direction.
3.2. TPFCS results
Two TPFCS measurements, one from inside a cell and one from the intercellular space, are presented in Fig. 6 . The tissue morphology is visualised by the SRB fluorescence image, see Fig. 6(a). The cell structure observed from the SRB fluorescence resembles the structures observed using the skin autofluorescence as seen in Fig. 3. Examples of fluorescence curves from saved runs and rejected runs (crossed) are shown in Fig. 6(b)–6(f). The average autocorrelation data points, resulting from the saved runs, together with the fitted curves are presented in the graphs in Fig. 6(d) (inside a cell) and Fig. 6(g) (intercellular space). The diffusion times and average numbers of molecules within the focal volume are shown in the respective graphs.
The curve fitting using the autocorrelation function for the three dimensional free diffusion was performed in three different ways; 1) by using an S parameter calculated from the PSF measurements on fluorescent beads in the skin at the corresponding tissue depth; 2) by using an S parameter calculated from the PSF measurements in solution; 3) by letting the S parameter vary in the fitting routine. The extracted values of the diffusion times for the three different fit types are shown in Table 1 for different tissue depths. As shown by the table, the two methods where the S parameter is pre-determined yielded similar results, since the PSF did not differ much between the skin and solution, On the other hand, when the S parameter is allowed to vary the diffusion time deviates significantly.
The diffusion coefficient was calculated from the diffusion time by using Eq. (12). Only the values of the diffusion times obtained from the fit where the S parameter came from the PSF measurements in skin were included in the calculations. Also the values of the width r0 were obtained from the same PSF data.
These values of D are shown as function of tissue depth in Fig. 7 . Analogous to the diffusion times presented in Table 1, the value of the diffusion coefficient was not found to be dependent on tissue depth since the PSF, i.e. the measurement volume did not vary with depth, see Fig. 5.
In order to visualize the uptake of fluorescent compounds into tissue, e.g. in drug delivery models, using TPLSM the fluorescence intensity profile is generally plotted [8,9,20]; however, this intensity profile cannot be directly converted to the concentration of fluorophores at the different tissue depths. On the other hand, when using TPFCS the detected concentration of molecules present in the focal volume should be independent of the intensity of the excitation light, assuming there are no bleaching or saturation effects. Figure 8 shows the measured number of molecules in the focal volume by performing TPFCS measurements on a skin sample exposed to SRB as a function of depth. This curve resembles an intensity profile curve, but since it contains quantitative information about the number of detected fluorophores it can be directly converted to the fluorophore concentration profile.
The concentration of SRB in the applied solution, 3.4 µM, corresponds to approximately 600 molecules in the focal volume (0.3 fl). With the TPFCS measurements, the detected numbers of molecules in the focal volume in the superficial layers of the skin was as high as 5000. This corresponds to an approximately 10-fold increase in concentration at the skin surface, implying that the fluorophores are accumulating in the SC, giving a higher concentration than in the solution. Deeper into the tissue, there is a decrease in the concentration and the amount of fluorophores in the skin tissue is close to that in the applied solution. The “control” measurement of the 3.4 nM SRB solution in water resulted in a value of 3 molecules in the focal volume, which corresponds rather well with a calculated value of 0.6 molecules.
As discussed earlier, the value of D was not found to depend on tissue depth, Fig. 7. However, as presented by Fig. 6, TPFCS allowed for detailed measurement of molecular diffusion based on cellular localisation. As shown by the figure, both differences in number of molecules and diffusion time could be discerned. These measurements were performed in the SG, where it was easy to clearly identify the cells and the intercellular space. The number of molecules in the focal volume was found to be lower inside the cell (N = 110 ± 10) compared to in between cells (N = 210 ± 10). On the other hand, the diffusion time measured in the cell was τD = 1.7 ± 0.2 ms, compared to τD = 2.9 ± 0.1 ms in the intercellular space. This implies faster diffusing molecules in the cytoplasm than in between the cells.
In this study, we combine TPFCS with PSF measurements for quantitative analysis of fluorophores absorbed into excised samples of human skin. TPLSM is powerful tool to visualize the distribution and uptake of fluorescent chemical compounds in skin; however, quantification is difficult due to the highly light scattering properties of human skin. TPFCS can be applied to measure the diffusion of low molecular concentrations in tissue . For successful FCS analysis, correct values of the size and shape of the focal volume has to be incorporated in the calculations. The size of the focal volume in TPLSM and TPFCS depends on the properties and the performance of the imaging system, e.g. the excitation wavelength, the numerical aperture of the objective, aberrations and misalignments. However, the optical features of the sample are also of great importance. Thus, in order to apply TPFCS for quantitative analysis of fluorophores in skin, the actual focal volume, i.e. the PSF, has to be determined.
The PSF was measured at different depths in skin samples exposed to subresolution beads in diffusion chambers for 20 hours. The obtained values of FWHM in both the lateral direction and the axial direction are nearly twice as large as the corresponding calculated values obtained from Eq. (5) and Eq. (6). Our results of PSF in human skin obtained in the lateral direction agree well with the results by others ; however, in the axial direction we found a slightly smaller value of the PSF width, even though their resolution should be better than ours since they use a slightly shorter wavelength and larger NA. This might be due to the fact that our measurements are performed on beads embedded in the skin, while the earlier report is based on measurements of beads actually located outside the tissue. As the skin is a highly optically turbid media, we expected the axial FWHM to be strongly influenced by the imaging depth. Surprisingly, our measurements show that the PSF measured in skin down to 40 μm is similar to the PSF obtained in solution. This implies that the distortions caused by the optical system are dominant. On the other hand, the beads in skin occasionally showed a slightly tilted appearance, probably due to astigmatism. This bent shape has earlier been described in optically turbid samples .
The insertion of fluorescent nanoparticles into skin is problematic due to the inherent barrier properties of SC. Others have measured the PSF by placing fluorescent subresolution beads onto skin samples of various thicknesses , but it is preferable to have the beads located in the skin sample. In an earlier study by our group, we used a syringe to insert the beads into the skin . This procedure allowed for the insertion of a large number of particles; however, as the injection caused undesirable tissue damage we now applied the beads using diffusion chambers allowing for passive diffusion of the beads into the skin. In order to facilitate the penetration of the particles into the skin, tape-stripping was performed to remove SC in some cases. We could not observe any difference in PSF comparing the samples with or without pre-treatment. This was surprising as we had expected the SC to play a major role for the optical properties of skin. Nevertheless, the pretreatment did have an effect of the penetration ability; more beads could be detected at depths down to 40 µm in the samples where SC was removed.
From the TPFCS measurements it could be concluded that the diffusion of SRB inside human skin was not found to vary with tissue depth. Instead, the local variations of the diffusion time based on cellular localization could be detected. The mean values of the diffusion coefficient for SRB in skin were found to be in the range 4 to 11.0 × 10−8 cm2/s. This can be compared to the diffusion coefficient of 0.5 × 10−5 cm2/s measured in solution. Our results suggest that the diffusion of SRB in skin is in average two orders of magnitude slower that the diffusion of SRB in solution. Due to the molecular interaction between the SRB molecules and the lipids and proteins present in the skin, these diffusion values seem reasonable. Furthermore, we found that the diffusion rate is different in the intercellular and intracellular space. Thus the suggested TPFCS method can be used to measure and quantify the molecular diffusion of fluorescent xenobiotics delivered to human skin in more detail.
The TPFCS measurements not only yield data on the molecular diffusion, but also the number of fluorescent molecules present in the focal volume. Since the focal volume is determined by the PSF measurement, this result can be directly converted to the fluorophore concentration inside the skin. In order for successful TPFCS analysis, the average number of molecules in the focal volume should be between 0.1 and 1000 for optimal conditions for the autocorrelation analysis . In this study, we choose a concentration corresponding to 600 molecules of the SRB solution applied to the skin samples. The detected number of molecules was found to be in the range 300 – 5000. This means that in the superficial layers of the skin, i.e. the SC, there is an accumulation of the fluorophore corresponding to a concentration almost 10 times the one in the applied solution.
Earlier studies using TPLSM for investigation of the uptake of compounds into skin have been restricted to qualitative or relative analysis of the results, and quantification has been problematic [8,9,20]. In the present study we show that TPFCS can be used for quantitative analysis of the distribution and diffusion of fluorophores applied to skin. This is of great importance as increased understanding of the processes involved in the exposure of skin to xenobiotics, not only enable design and investigations of new drug delivery systems, but also form the basis for improved understanding on how skin exposure may lead to contact allergy or skin cancer.
We acknowledge Christina Halldin at the Department of Dermatology, Sahlgrenska University Hospital, for collecting the skin samples. We also thank the Centre for Cellular Imaging at the Sahlgrenska Academy, University of Gothenburg for the use of imaging equipment and for support from the staff. Financial support was obtained from Göteborgs Science Centre for Molecular Skin Research, and the Swedish Research Council (VR 621-206-3700).
References and links
2. P. T. C. So, C. Y. Dong, B. R. Masters, and K. M. Berland, “Two-photon excitation fluorescence microscopy,” Annu. Rev. Biomed. Eng. 2(1), 399–429 (2000). [CrossRef]
6. K. Koenig and I. Riemann, “High-resolution multiphoton tomography of human skin with subcellular spatial resolution and picosecond time resolution,” J. Biomed. Opt. 8(3), 432–439 (2003). [CrossRef]
7. B. R. Masters, P. T. C. So, and E. Gratton, “Multiphoton excitation microscopy of in vivo human skin. Functional and morphological optical biopsy based on three-dimensional imaging, lifetime measurements and fluorescence spectroscopy,” Ann. N. Y. Acad. Sci. 838(1 ADVANCES IN O), 58–67 (1998). [CrossRef] [PubMed]
8. B. Yu, C. Y. Dong, P. T. C. So, D. Blankschtein, and R. Langer, “In vitro visualization and quantification of oleic acid induced changes in transdermal transport using two-photon fluorescence microscopy,” J. Invest. Dermatol. 117(1), 16–25 (2001). [CrossRef] [PubMed]
9. J. Bender, C. Simonsson, M. Smedh, S. Engström, and M. B. Ericson, “Lipid cubic phases in topical drug delivery: visualization of skin distribution using two-photon microscopy,” J. Control. Release 129(3), 163–169 (2008). [CrossRef] [PubMed]
10. J. Paoli, M. Smedh, A. M. Wennberg, and M. B. Ericson, “Multiphoton laser scanning microscopy on non-melanoma skin cancer: morphologic features for future non-invasive diagnostics,” J. Invest. Dermatol. 128(5), 1248–1255 (2008). [CrossRef]
11. E. Dimitrow, M. Ziemer, M. J. Koehler, J. Norgauer, K. König, P. Elsner, and M. Kaatz, “Sensitivity and specificity of multiphoton laser tomography for in vivo and ex vivo diagnosis of malignant melanoma,” J. Invest. Dermatol. 129(7), 1752–1758 (2009). [CrossRef] [PubMed]
14. J. Kanitakis, “Anatomy, histology and immunohistochemistry of normal human skin,” Eur. J. Dermatol. 12(4), 390–399, quiz 400–401 (2002). [PubMed]
15. G. Alexandrakis, E. B. Brown, R. T. Tong, T. D. McKee, R. B. Campbell, Y. Boucher, and R. K. Jain, “Two-photon fluorescence correlation microscopy reveals the two-phase nature of transport in tumors,” Nat. Med. 10(2), 203–207 (2004). [CrossRef] [PubMed]
16. E. H. P.Schwille, “Fluorescence Correlation Spectroscopy, An introduction to its Concepts and Applications.”
17. K. K. F. Fischer, S. Puschmann, R. Wepf, I. Riemann, V. Ulrich, and P. Fischer, “Characterization of multiphoton laser scanning device optical parameters for image restoration,” Proc. SPIE 5463, 140–145 (2004). [CrossRef]
18. W. J. Addicks, G. L. Flynn, and N. Weiner, “Validation of a flow-through diffusion cell for use in transdermal research,” Pharm. Res. 04(4), 337–341 (1987). [CrossRef]
19. P. Schwille, U. Haupts, S. Maiti, and W. W. Webb, “Molecular dynamics in living cells observed by fluorescence correlation spectroscopy with one- and two-photon excitation,” Biophys. J. 77(4), 2251–2265 (1999). [CrossRef] [PubMed]
20. K. Samuelsson, C. Simonsson, C. A. Jonsson, G. Westman, M. B. Ericson, and A. T. Karlberg, “Accumulation of FITC near stratum corneum-visualizing epidermal distribution of a strong sensitizer using two-photon microscopy,” Contact Dermat. 61(2), 91–100 (2009). [CrossRef]
21. D. E. M. Na Ji and E. Betzig, “Adaptive optics via pupil segmentation for high-resolution imaging in biological tissues,” Nat. Methods 7, ••• (2010).
22. S. Guldbrand, C. Simonsson, M. Smedh, and M. B. Ericson, “Point spread function measured in human skin using two-photon fluorescence microscopy,” Proc. SPIE 7367 (2009).