Multiphoton microscopy for the investigation of dermal
penetration of nanoparticle-borne drugs
Frank Stracke1,C, Barbara Weiss2, Claus-M. Lehr2,
Karsten König1, Ulrich F. Schaefer2, Marc Schneider2
1) Fraunhofer Institute for Biomedical Technology, Ensheimer Str. 48, 66386 St. Ingbert,
2) Department of Biopharmaceutics & Pharmaceutical Technology, Saarland University, 66123
C) Corresponding author:
eMail: email@example.com, phone / fax: +49 (0)6894 980-166 / -152
Short title: “Multiphoton microscopy of particle penetration”
CLSM: Confocal laser scanning microscopy
FWHM: Full width half maximum
MPM: Multiphoton microscopy
NA: Numerical aperture
NADH: Nicotineamide adenine
dinucleotide, hydrogenated form
NIR: Near infrared
PMT: Photomultiplier tube
ROI : Region of interest
x, y : lateral dimensions
z : normal dimension / subsurface depth
This document is a preprint version of :
Multiphoton Microscopy for the Investigation of Dermal Penetration of Nanoparticle-Borne
Formal publication in:
Journal of Investigative Dermatology advance online publication 18 May 2006; doi:
Frank Stracke, Barbara Weiss, Claus-Michael Lehr, Karsten König, Ulrich F Schaefer and
Multiphoton microscopy of a dually fluorescence-labeled model system in excised
human skin is employed for high resolution three dimensional visualization in order to study
the release, accumulation and penetration properties of drugs released from nanoscale carrier
particles in dermal administration. Polymer particles were covalently labeled with fluorescein
while Texas Red as a drug-model was dissolved in the particle to be released to the
formulation matrix. Single nanoparticles on skin could easily be localized and imaged with
diffraction limited resolution. The temporal evolution of the fluorescent drug-model
concentration in various skin compartments over more than five hours was investigated by
multiphoton spectral imaging of the same area of the specimen. The three dimensional
penetration profile of the drug-model in correlation with skin morphology and particle
localization information are obtained by a multiple laser line excitation experiment.
Multiphoton microscopy combined with spectral imaging was found to allow non invasive
long term studies of particle-borne drug-model penetration into the skin with sub cellular
resolution. By dual color labeling a clear discrimination between particle-bound and released
drug-model was possible. The introduced technique was shown to be a powerful tool in
revealing the dermal penetration properties and pathways of drugs and nanoscale drug
vehicles on microscopic level.
The encapsulation of active substances is a common pharmaceutical strategy to
modify the transport and release properties of a drug. Especially to nanoparticulate systems
great potential is attributed in the field of drug delivery. This is partly due to the fact, that
sensitive drugs can be hidden from degradation in the particles (Daniels, 2006; Volodkin et
al., 2004). Further powerful properties of nanoscale drug carrier are the sustained release
(Daniels, 2006; El-Samaligy et al., 1986) of the active substances resulting in an extended
activity or enhanced uptake (Alvarez-Roman et al., 2004c; Lombardi Borgia et al., 2005) and
the possible reduction of adverse effects (Lamprecht et al., 2001). Functional coatings of the
particles may allow the targeted accumulation and release of drugs at their therapeutic sites
(Dinauer et al., 2005; Kotrotsiou et al., 2005; Wartlick et al., 2004).
Widely used nanoparticle formulations are based on poly(lactic acid) (PLA),
poly(glycolic acid) (PGA), and their co-polymers, poly(lactide-co-glycolide) (PLGA), which
are known for their good biocompatibility and degradability through natural pathways
(Brannon-Peppas, 1995). In oral and parenteral applications these solid biodegradable
polymeric nanoparticle have already shown their advantage over liposomes by their increased
stability (Hans and Lowman, 2002; Ravi Kumar et al., 2003; Soppimath et al., 2001).
Nanoscale polymeric drug vehicles have also been proposed for transdermal delivery
(Alvarez-Roman et al., 2004c; Kohli and Alpar, 2004; Lombardi Borgia et al., 2005; Luengo
et al., accepted). Penetration (Alvarez-Roman et al., 2004c; Luengo et al., accepted),
permeation (Luengo et al., accepted) and accumulation (Toll et al., 2004) of some particle-
borne drugs and drug models after topical application have been investigated by conventional
techniques and confocal microscopy of single stained particles.
Established methods for the investigation of drug penetration into the skin are mostly
destructive: a representative sample of a defined skin layer is isolated and extracted for
chemical analysis (tape stripping method, cryo-sectioning) (Brain et al., 2002; Wagner et al.,
2001). The result of such an experiment is an area averaged depth profile of the drug in the
skin to a certain time (Luengo et al., accepted). The depth profiles for different incubation
times have to be investigated with different samples, neglecting the individual characteristics
of biological specimens. To evaluate and optimize novel dermal drug delivery strategies using
nanoscale drug carriers, more versatile techniques are required. Such a technique must allow
the discrimination between free and carrier-bound drug, the tracing of the carrier
nanoparticles, the allocation of microscopic delivery pathways to specific dermal sites and
time studies on the same skin area. After application of nanoparticles as topical vehicles one
can imagine different routes of drug delivery. It could be assumed that the whole
nanoparticulate system is taken up without being destroyed (Kohli and Alpar, 2004) or that
the nano-carrier is decomposed close to the skin surface and thereafter the active substance
penetrates in dependence on the local environment (acidification or absorption of drug /
nanoparticle-complexes) as speculated in Luengo et al. for the enhanced long term uptake of
flufenamic acid (Luengo et al., accepted). Furthermore, a direct diffusion from the carrier into
the stratum corneum is reasonable as described by Bouwstra et al. (Meuwissen et al., 1998;
Van Kuijk-Meuwissen et al., 1998a). In any case, it is essential to distinguish between
particles, particle-bound drug and released drug. Herein we describe how multiphoton laser
scanning microscopy (MPM) and confocal laser scanning microscopy (CLSM) can be very
beneficial tools in order to meet all these demands in one experiment. In particular
multiphoton microscopy enables repeated non-invasive investigations of skin tissue down to
the dermis with virtually no out-of-focus effects of the scanning laser beam (Konig and
Riemann, 2003). Due to multiple labeling techniques in combination with multiphoton
spectral imaging or selective excitation of the labels a clear discrimination between particle
and free drug model is possible, as well as tracking of single particles. Due to the excitation of
endogenous fluorophores of the skin by multiphoton excitation and the correlation of the
resulting autofluorescence image to the drug fluorescence pattern the identification of
accumulative spots and penetration pathways is possible with sub-cellular resolution (Yu et
al., 2003; Yu et al., 2002). Multiphoton microscopy provides several considerable advantages
over conventional fluorescence and confocal microscopy (Konig, 2000; Xu et al., 1996).
Three of which are relevant to the present investigation: the concentration of all light-matter
interactions to the focal volume, the convenient separation of fluorescence from scattered
excitation light due to the large blue-shift of fluorescence and the capability to excite
compounds which else require ultraviolet excitation, in particular native fluorophores as
NADH and keratin (Huang et al., 2002; Konig and Riemann, 2003; Pena et al., 2005). The
confined interaction volume at the focal point is due to the In-dependence of n-photon
absorption processes on the illumination intensity I. Hence already two-photon absorptions
are confined to a sub-femtoliter focal volume, in which the illumination intensity is
sufficiently high. Since the excitation with near infrared (NIR) lasers matches the optical
window of biological matter (700 – 1100 nm) virtually no single-photon absorptions occur in
the illumination cones. As a consequence, no fluorescence is generated outside the focal
volume and therefore three dimensional spatial resolution is an inherent feature of
multiphoton laser scanning microscopy. Furthermore out of focus photo-damage is drastically
reduced and light penetration depth into tissue is significantly enhanced (Centonze and White,
1998; Konig and Riemann, 2003).
Since most pharmaceutical substances are basically non fluorescent, the usage of
appropriate fluorescent model compounds is reasonable. Such a model compound has to
match the molecular size, charge, membrane permeability, distribution, and diffusion
coefficients as good as possible. A fluorescent label fixed to the actual drug molecule changes
these properties and thus the penetration behavior considerably. Hence labeling only makes
sense if specific interactions of the drug to certain sites are investigated. In contrast the non-
superficial fluorescent labeling of the nanoscale carrier particles doesn’t change the particle’s
pharmacokinetics significantly. In addition, the fate of the nano-carrier itself and its role in the
changed uptake behavior may be investigated. In this work a two color labeling technique was
used to trace the migration of the nanoparticles and to observe the release and uptake of the
drug-model compound. To this end fluoresceinamine was covalently linked to the polymeric
particle material and Texas Red was physically resolved in the particle matrix. It is shown,
that individual sub-diffraction sized nanoparticles can be localized, traced and spectrally
analyzed. Due to the two-color staining a clear discrimination between free and particle-
bound dye was achieved. The method turned out to allow stable measurements on excised
human skin over hours with no significant drift of the specimen.
Multiphoton fluorescence imaging
The multiphoton optical sections were recorded from the skin surface down to the
bottom of the shown dermatoglyph at z = 42 µm over five hours under identical conditions
(Figure 1). It was found, that the sub-diffraction-limit sized particles can easily be detected
and localized, as long as their mean distance is well above this limit. They appear as lateral
diffraction limited spots with widths of about 0.5 µm, which is in reasonable agreement with
theoretical predictions for the minimum achievable full width half maximum FWHM of
302 nm1. A typical fluorescence profile of two particles in situ is displayed in Figure 2. The
minimal fluorescence spot size may be broadened by Brownian motion and distorted by flux
motions of the gel in the dermatoglyphs. In case of the bright spots the detector went into
saturation, which additionally caused a considerable broadening of the spot size. The mean
distance between the nanoparticles in the present formulation is on the order of 5 µm. Three
dimensional tracing of individual particles is easily possible under the outlined conditions and
allows detailed studies on the migration of nanoscale drug carriers in the skin.
The significant endogenous fluorescence of keratin under two-photon excitation
enables imaging of the outermost layer of the stratum corneum and hence the dermal
topography. The PLGA particles are obviously not able to penetrate the stratum corneum and
stay in the gel-filled dermatoglyphs over the entire observation time (Figure 3). This
corresponds with former findings of the authors in which no penetration of PLGA particles
loaded with flufenamic acid into the human skin could be observed (Luengo et al., accepted).
1 The theoretical FWHM was calculated by convolution of a sphere profile of d = 290 nm with
the squared intensity point spread function IPSF². The ISPF2 was derived according (Zipfel et
It is noteworthy, that even after more than five hours swelling, shrinking and stress relaxation
motions of the skin sample only lead to minute deformations within the field of view. No drift
of the specimen occurred.
No significant changes in the background fluorescence intensity of the ointment
matrix or in the stratum corneum as a consequence of the Texas Red release and accumulation
were observed in the multiphoton fluorescence mode. The reason for this finding is the
comparable low two-photon absorption cross-section of Texas Red at λ = 800 nm (Figure 4)
while fluorescein and keratin are efficiently excited.
Multiphoton spectral imaging
In order to investigate the distribution of the drug-model Texas Red as a function of
time, two-photon spectral imaging was applied. In this technique the luminescence signal
from the specimen is spectrally resolved and the spectrum is stored for each pixel or voxel,
respectively. From these data, fluorescence spectra for arbitrary regions of interest can be
calculated. The laser wavelength of 800 nm leads to an image which is dominated by the
fluorescein emission and the endogenous fluorescence of the stratum corneum. But due to the
different sensitivity spectrum of the META detector array the sensitivity loss in the red
spectral range is less drastic compared to the PMTs. Even faint contributions of the Texas Red
fluorescence to any pixel or region of interest (ROI) of the image can now be isolated from
the other emissions by spectral separation. In the present study a 145 × 145 µm2 area from the
middle of Figure 3 was investigated. The spectral images were recorded to similar times as
the optical sections. In Figure 5a a typical example of such a spectral image is displayed in
true color mode. The nanoparticles appear clearly green, indicating that their fluorescence
originates predominantly from the covalently bound fluorescein. The outermost layer of the
stratum corneum shows a heterogeneous distribution of colors. For an interpretation the
fluorescence spectrum of the respective region is taken (Figure 6a, ROI 2). The reddish color
of the deeper skin is mainly caused by residual scattered excitation light beyond 700 nm
(Figure 6a, ROI 3).
Three regions of interest are indicated in Figure 5a: ROI 1 contains the gel matrix
including the nanoparticles, ROI 2 contains the keratinous surface layer of the stratum
corneum and ROI 3 covers the deeper layers down to approximately 20 µm. Average
fluorescence spectra are calculated from the ROIs at different times to reveal the evolution of
the Texas Red concentration in these regions. Due to the small deformations of the specimen,
especially in vertical direction, the ROIs do not enclose exactly identical skin domain at the
different times. This may be a cause for small discrepancies in the spectral analysis. But since
the ROIs were always chosen to enclose only skin compartments according to the definitions
above, the spectral analysis will still reveal the accurate trends. To rule out errors due to
varying offsets and fluctuations in the absolute signal, the Texas Red fluorescence is
determined relative to an emission, which is expected to be constant in time. For ROI 1 the
fluorescein emission from the nanoparticles at 525 nm is used, for ROI 2 the keratin
autofluorescence around 510 nm and for ROI 3 the back scattered excitation light at 714 nm.
Linear baselines are applied to reduce the influence of the background (Figure 6a). Since in
particular the autofluorescence in ROI 2 and the Texas Red emission in ROI 3 show only dim
intensities, the determined intensity ratios for ROI 2 and ROI 3 have large uncertainties.
Anyway the rough trends are visible in those plots.
Whereas the Texas Red content in the nanoparticles showed up to be low and
basically constant already from the first measurement at t = 30 minutes (Figure 5c), the
concentration of Texas Red in the gel matrix drops significantly with time (Figure 5b, 7b).
The Texas Red concentration in the superficial layer of the stratum corneum is also declining,
but less rapid (Figure 6c). The observation of declining concentration in this compartment
from the earliest measurement at t = 30 minutes on means, that the vast fraction of Texas Red
was taken up from the gel to the stratum corneum surface almost immediately after
application. Furthermore the release of Texas Red from the nanoparticles to the matrix started
well before application to the skin. Probably the predominant fraction was released to the
solvent even before suspension of the particles to the hydrogel. Since the concentration of the
released dye is decreasing in the gel-filled dermatoglyphs and in the stratum corneum surface
it must have penetrated the skin (or metabolism of the dye occurred, which is unlikely). In
fact an increase of Texas Red concentration in the deeper stratum corneum and stratum
granulosum is evident (Figure 6d), indicating a slow penetration of the dye. The fits to the
plots are mono-exponential decays and growth, respectively. The fit curves have no
theoretical pharmacokinetic background, but allow a convenient comparison of decay and rise
times. Advanced interpretations of the release, uptake and penetration of the dye will be
possible only on the basis of an appropriate pharmacokinetic model. The decay times of the
mono-exponential fits are τROI 1 = 35±6 minutes, τROI 2 = 148±319 minutes and the rise time is
τROI 3 = 59±38 minutes.
After 320 minutes a multitracking experiment visualizes the distribution of Texas
Red, fluorescein and keratin fluorescence (Figure 7). In this technique the signals from the
different fluorophores are separated by their excitation spectra, not by their fluorescence
spectra as done in spectral imaging. As opposed to the spectral images, which are recorded
with one excitation wavelength, herein the successive use of three excitation sources may lead
to small mismatches between the channels. The vertical mismatch of the NIR two-photon
excitation image to the images excited by visible laser sources is due to chromatic aberrations
of the objective. The small average lateral shift from the fluorescein image, excited by the
488 nm argon ion laser line, to the Texas Red image, excited by the 543 nm helium neon laser
line, is due to the microscope optics and could be compensated by image processing.
Furthermore a multitracking experiment yields an overlay of successive scan images. If the
specimen shows dynamics, any motion is reflected by shifts from one scan to the next. One
must keep these mismatches in mind when performing colocalisation studies on different
dyes. In Figure 7d an average lateral shift of about 6µm is evident between the fluorescein
image and the Texas Red image. The shift is not uniform for all particles, but varies due to
flux motions within the hydrogel during the acquisition times of the three single images of the
multitracking experiment. The particle pattern in the two-photon excited image exhibits no
correlation with the visibly excited images, indicating a vertical mismatch of the focal planes
exceeding the normal size of the focal volumes (approximately 1 µm). This mismatch is also
evident in the missing congruency of the dermatoglyph borders in the two-photon excited
image and the 543 nm excited image (Figure 7). The keratin autofluorescence seems to be
located not at the surface of the Texas Red stained dermatoglyph, but some microns inside of
the skin. This is a consequence of the vertical mismatch between the NIR and visible focus.
Nevertheless detailed distribution analysis for each isolated dye can be performed as well as a
rough colocalisation study, since fluorescein and keratin are not excited by the 543 nm laser
line and Texas Red is virtually exclusively excited by this excitation source (Figure 4).
The 488 nm excited image (green) proves, that fluoresceinamine is strictly bound to
the particles and not released during the time of observation. Previous experiments showed
that released fluorescein is rapidly bound to the keratinous layer of the stratum corneum. In
the present study no such accumulation is found. The fluorescence spots of the particles are
mostly broadened due to saturation. In the 543 nm excited images (orange) the distribution of
Texas Red is visible. The predominant fraction of the dye is to be found within the skin, but
the particles are also slightly observable. Obviously there is still a certain amount of the dye
stored in the nanoparticles, though the release process has proceeded far. The released Texas
Red penetrated the skin and accumulated in the stratum corneum down to approximately
20 µm (Figure 7b). In Figure 7c, recorded at a depth of 32 µm, the penetration into deeper
skin compartments can be recognized, since the walls of the dermatoglyph are steep enough to
reveal the diffusion from the superficial layers inwards. The distribution of Texas Red in the
skin is not uniform, but structured. It is conspicuous, that the superficial accumulation is
strongest, where the skin adjoins to large gel-filled spaces. Beneath the superficial layer the
distribution of the drug-model is more uniform, which indicates faster diffusion of the
compound in these skin compartments. The fundamental discrepancy between the visible
excited images and the two-photon image proves that the skin visible structures from the two-
photon image are predominantly due to keratin autofluorescence and not due to the superficial
accumulation of Texas Red. xz-sections can be extracted from the multitracking z-stacks,
which nicely exhibit the concentration profile in normal direction (Figure 8). The highest
concentrations of Texas Red in the deeper layers are found under superficial areas with
pronounced keratin autofluorescence. This supports the hypothesis of a fast resorption of
Texas Red from the hydrogel to the keratinous compartments and a slower diffusion from
there to the skin tissue underneath. A vertical mismatch of several microns between the
543 nm excited Texas Red fluorescence and the multiphoton excited endogenous keratin
fluorescence due to the chromatic aberration of the optics is apparent in all xz-sections.
A major challenge in the design of nanoscale drug delivery entities is the
development of mechanisms which triggers the release of the drug when the nanoparticle
attains its therapeutic site. For dermal applications such mechanisms may consist of an
intrinsic recognition element for specific sites or compounds of the skin and a thereby
controlled release step. Another approach to this end is to modify the particle surface in a way
that they accumulate at the therapeutic site and to start the drug release by external stimuli
like light illumination or the delayed application of a kick-off agent. A pure diffusive release
of a drug from suspended particles starts at the moment of suspension and after a certain
storage time the drug is predominantly dissolved in the suspension matrix. Hence this way of
formulation compromises all advantages of nano particular drug delivery. Since so far no such
intelligent trigger techniques are established for topical application of nanoscale drug carriers,
the simple diffusive release mechanism is observed in the present work to demonstrate the
capabilities of laser scanning microscopy in this field. As a consequence it was found, that the
major fraction of Texas Red was already released to the gel matrix at the moment of the first
measurement. This observation demonstrates that diffusive release of drugs from the particle
cannot be employed for the practical use of nanoscale drug carriers and in general stresses the
need for the design of intelligent nano carriers with controllable release behavior. The three
dimensional microscopic and spectral resolution of the utilized techniques showed up to have
a more versatile potential for the evaluation of such smart nanoparticle formulations than
conventional penetration study methods.
In general, multiphoton and confocal microscopy of dually labeled nanoparticles in
human skin biopsies have been demonstrated to be very suitable techniques to investigate the
migration, accumulation, release and penetration of nanoparticle-borne drugs in dermal
application. By different fluorescence spectra of the covalently fixed and the physically
dissolved dye discrimination between particle-bound and released compounds is possible. By
performing spectral imaging the quantitative analysis of the release process is much less
interfered by background emissions and crosstalk errors than in a conventional two detector
channel study with a dichroic beamsplitter. Furthermore an emission can be attributed to a
certain compound with a high degree of accuracy by resolving the fluorescence spectrum.
It was shown, that tracing of even single particles of about 300 nm diameter in the
gel matrix inside the dermatoglyphs is not a difficult task. The observation depth of
multiphoton microscopy even in turbid media is on the order of several hundred microns
(Centonze and White, 1998) and is mostly limited by the working distance of the applied high
numerical aperture objectives. Hence single fluorescent nanoparticles should be observable
within the skin down to the dermis. This is of particular importance if particles are
investigated, which are able to penetrate the epidermal layers. No penetration of particles into
the skin was found in the present study. Due to the three dimensional sub cellular resolution
and the possibility of repeated non-invasive investigations of the same skin area detailed
information on the penetration pathway of particle-bound and free drug-models are
Since confocal microscopy with visible excitation does not provide such enhanced
observation depths, the multitracking technique is primarily adequate for investigations on the
upper skin layers. Problems may arise from possible lateral mismatches, the chromatic shift of
the focal planes in case of strongly differing excitation wavelengths and the time lag between
the single scans. The influences of these interferences have to be carefully regarded in the
interpretations of multitracking studies. Nevertheless it can be a powerful tool if the applied
fluorophores are basically excited exclusively by the chosen laser lines, as demonstrated in
the present study.
The specimen mounting fulfilled the objectives of keeping the skin sample in place
with an accuracy of few microns in three dimensions, to avoid desiccation and to minimize
swelling and shrinking effects. In the present study excised, frozen and thawed human skin
was used. The properties of a skin specimen treated accordingly will certainly differ from skin
in vivo and freshly excised biopsies, as e.g. the pH depth profile changes rapidly after excision
(Wagner et al., 2003) and the endogenous NADH fluorescence of the vital epidermal layers is
virtually vanished after frozen storage. In spite of these physiological changes permeation
measurements did not exhibit a changed passage behavior for the investigated compounds
after frozen storage of human skin, indicating that the non vital superficial layer of the stratum
corneum was the main penetration barrier (Wagner et al., 2004). A multiphoton microscopy
experiment on freshly excised human skin or in vivo could reveal even more detailed
information on the drug penetration into the vital skin layers, since the cellular
autofluorescence would allow the allocation of the drug pathways on single cell level.
By deliberate variation of the fluorescent drug-model, correlations of the microscopic
penetration behavior with various physicochemical properties of the drug-models could be
investigated. An intelligent release mechanism providing for a defined initial time of release
from the nanoscale drug carrier would allow to study the accurate evolution of drug
concentrations in the diverse skin compartments and to derive a pharmacokinetic model of the
drug uptake from nano particular formulations.
Concluding, we demonstrated the benefits of multi-color labeling of biodegradable
nanoparticles and the intriguing insights into the penetration behavior of particle-borne drugs
due to the combination with multiphoton microscopy and confocal laser scanning microscopy.
The usage of two fluorescent dyes of well separated absorption and emission spectra enabled
the investigation of the transport of a fluorescent drug model in situ. This might be extended
to multiple loading of nanoparticles with two or more drug models which differ in spectral
and physicochemical properties for direct comparison. Furthermore relevant pharmaceutical
compounds with native fluorescence may be investigated as well as intelligent release
mechanisms. The kinetics of the drug transport from the initial formulation to the
subcutaneous compartments can be studied in elementary steps, as the enrichment of particles
in certain dermal sites (Toll et al., 2004), the release of the drug from the particles, its uptake
into the stratum corneum, the diffusion into the deeper skin layers etc.. The determination of
the enrichment in dependence of time as well as the visualization of the ‘structured’ diffusion
process into the skin envisages the potential of this approach. In addition, the covalent label to
the nanoparticle itself enables the investigator to follow the fate of the nano-carrier, its uptake,
accumulation or decomposition. This might be very meaningful in particular for the
exploration of the penetration function of hair shafts. First evidence was observed that these
follicles might play an important role (Alvarez-Roman et al., 2004b; Toll et al., 2004; Van
Kuijk-Meuwissen et al., 1998b).
Methods & Material
Poly(L-lactide-co-glycolide) (PLGA) (Resomer RG 50:50 H) was kindly provided by
Boehringer Ingelheim (Boehringer Ingelheim GmbH & Co. KG, Ingelheim, Germany). 5-
Fluoresceinamine (FA) and 1-ethyl-3-(3-Dimethylaminopropyl)-carbodiimide hydrochloride
(DMAP) were obtained from Sigma (Sigma Chemical Co., St. Louis, MO, USA).
Polyvinylalcohol (PVA) (Mowiol 4-88) was purchased from Kuraray (Kuraray Specialities
GmbH, Frankfurt am Main, Germany). Texas Red® was provided by Atto-Tec (Atto-Tec
GmbH, Siegen, Germany). We grateful acknowledge the kind provision of Natrosol® 250 M
hydrogel (Aqualon, Hercules Inc., DE, USA) by J. Luengo. All other chemicals are of
Polymer labeling and preparation of dual color nanoparticles
FA bound PLGA (FA-PLGA) was prepared based upon the method described by
Horisawa et al. (Horisawa et al., 2002). Briefly, PLGA (3.07g) and FA (0.0583g) were
dissolved entirely in 30ml of acetonitrile with 0.0408g of DMAP and incubated at room
temperature for 24h under light protection and gentle stirring. The resultant FA-PLGA was
precipitated by the addition of purified water and separated by centrifugation. The polymer
was rinsed from excessive reagents (dissolution in acetone and precipitation with ethanol in
terms) and then lyophilized (Alpha 2-4 LSC, Martin Christ Gefriertrocknungsanlagen GmbH,
Texas Red nanoparticles were prepared from FA-PLGA employing a single emulsion
method (oil in water). The emulsion was formed between an organic FA-PLGA solution (2%
(w/v) in ethyl acetate) which additionally contained 30µl of a saturated Texas Red solution
and 5ml of a PVA solution (1% in demineralized water) under stirring on a magnetic stirrer
for 2h. Then, the emulsion was homogenized using an Ultra-Turrax® T 25 Mixer (Janke und
Kunkel GmbH & Co., Staufen, Germany) at 13,500rpm and the organic solvent was removed
by a rotary evaporator. The mean particle size was determined to d = 290±5 nm using
photocorrelation spectroscopy (Zetasizer®3000HSA, Malvern Instruments GmbH,
Herrenberg, Germany) and the homogeneity of the particles was verified using scanning
probe microscopy (BioScope, Veeco, Santa Barbara, USA) (Figure 9). To prepare the
nanoparticle ointment a Natrosol® gel (3% w/w) was mixed with an aqueous suspension of
the nanoparticles in a 1:1 ratio.
Excised human skin from Caucasian female patients, who had undergone abdominal
plastic surgery, was used. The procedure was approved by the Ethical Committee of the
Caritas-Traegergesellschaft, Trier, Germany (6th July 1998). Adequate health and no medical
history of dermatological disease were required. After excision the skin was cut into 10 × 10
cm2 pieces and the subcutaneous fatty tissue was removed from the skin specimen using a
scalpel. Afterwards the surface of each specimen was cleaned with water, wrapped in
aluminum foil and stored in polyethylene bags at –26°C until use. Previous investigations
have shown that no change in the penetration characteristics occurs during the storage time of
6 months (Bronaugh et al., 1986; Wagner et al., 2004).
Laser Scanning Microscopy:
The microscopy studies were performed on a versatile laser scanning microscope
(LSM 510 NLO META, Carl Zeiss Jena GmbH, Germany) for conventional confocal
microscopy with multiple excitation laser lines and multiphoton excitation microscopy.
Herein, the 488 nm laser line of the internal argon ion laser and the 543 nm line of the internal
helium neon laser were applied for confocal imaging. For multiphoton microscopy a
femtosecond pulsed titan:sapphire laser at λ = 800 nm with 90 fs pulse width and 80 MHz
repetition rate (Coherent Vitesse) was coupled into the microscope.
The META scanning and detection module offers different ways to detect the
fluorescence light. First, the complete emission is distributed between two sensitive
photomultiplier tubes (PMT) by means of neutral and dichroic mirrors. Second, the emission
signal is spectrally dispersed by a diffraction grating and guided on a 32 channel
photomultiplier array for spectral analysis. Each channel detects a spectral range of about
10nm in this arrangement. Three different imaging modes were applied in the present study:
(a) the multiphoton fluorescence mode for the cumulative detection of the emitted
fluorescence. Here the fluorescence intensity is recorded by one PMT. The laser line is
blocked by a 650 nm shortpass beamsplitter and a 685 nm shortpass filter. This detection
mode is the most sensitive one, but any spectral information is lost. (b) In the multiphoton
spectral imaging mode the emitted light is separated from the laser line by the 650 nm
shortpass beamsplitter and recorded spectrally resolved by means of the grating and the
detector array. Each pixel of a spectral image contains the data of the 32 detector channels, so
that emission spectra for each pixel or for defined areas of the image are accessible. The
displayed images are true color coded. The spectral data is converted into appropriate RGB
values. (c) The multitracking mode is actually an overlay of three fluorescence intensity
images, subsequently acquired at different excitation wavelengths. An overlay procedure with
two visible excitation laser lines was utilized on similar specimens by Alvarez-Roman et al.
(Alvarez-Roman et al., 2004a). The excitation wavelengths are chosen to be preferably
absorbed by exclusively one fluorophore (Figure 4). The 488nm argon ion laser line is
predominantly absorbed by fluorescein, the 543 nm helium neon laser line by Texas Red
exclusively. Two-photon excitation at 800 nm is strong for fluorescein and the endogenous
fluorophore keratin, but poor for Texas Red.
Two-photon excitation spectra (Figure 4) were acquired according to a modified
`excitation fingerprinting´ procedure introduced by Dickinson et al. (Dickinson et al., 2003).
In contrast to the excitation fingerprinting technique, where the excitation power is kept
constant for different wavelengths, herein the laser power calibration is adjusted for constant
photon flux at different wavelengths to yield correct two-photon excitation spectra (Schneider
et al., 2005). One-photon excitation and fluorescence spectra were acquired with a Hitachi FL
4500 fluorometer using 50 µM aqueous solutions of fluorescein isothiocyanate (FITC)
dextran and Texas Red. FITC dextran was used in order to investigate a compound preferably
akin to FA-PLGA while showing good water solubility. The spectral properties of the
fluorescein derivatives are almost unaffected by different substituents in the 5-position.
For microscopy, disks of 0.8 cm diameter were punched out of frozen skin and
placed surface-up on a microscopy slide inside a circular vertical spacer. After application of
the ointment to the skin surface a cover slide was fixed onto the spacer by double-faced
adhesive tape in a way that the skin surface was gently pressed against the cover slide. This
setup provides constant specimen thickness and prevents desiccation of the skin sample. The
observation was performed through the cover slide by means of a 40×/1.3 NA Oil Objective
(Plan Neofluar, Carl Zeiss Jena GmbH, Germany).
This work is part of the PhD thesis of Ms. Barbara Weiss. We grateful acknowledge
the kind provision of Natrosol® hydrogel by J. Luengo and the supply of excised human skin
samples by K.-H. Kostka (Department of Plastic and Hand Surgery, Caritaskrankenhaus,
Alvarez-Roman R, Naik A, Kalia Y, Guy RH and Fessi H (2004a). Skin penetration and
distribution of polymeric nanoparticles. Journal of Controlled Release 99:53-62.
Alvarez-Roman R, Naik A, Kalia YN, Fessi H and Guy RH (2004b). Visualization of skin
penetration using confocal laser scanning microscopy. European Journal of
Pharmaceutics and Biopharmaceutics 58:301-316.
Alvarez-Roman R, Naik A, Kalia YN, Guy RH and Fessi H (2004c). Enhancement of topical
delivery from biodegradable nanoparticles. Pharmaceutical Research 21:1818-
Brain KR, Walters KA and Watkinson AC: Methods for Studying Percutaneous Absorption,
Walters KA (ed): Dermatological and Transdermal Formulations 1st ed., pp 197
(Marcel Dekker Inc., New York 2002).
Brannon-Peppas L (1995). Recent advances on the use of biodegradable microparticles and
nanoparticles in controlled drug delivery. International Journal of Pharmaceutics
Bronaugh RL, Stewart RF and Simon M (1986). Methods for Invitro Percutaneous-
Absorption Studies. VII. Use of Excised Human-Skin. J Pharm Sci 75:1094-1097.
Centonze VE and White JG (1998). Multiphoton Excitation Provides Optical Sections from
Deeper within Scattering Specimens than Confocal Imaging. Biophys J 75:2015-
Daniels J (2006). How polymeric microspheres deliver the goods. Pharmaceutical
Technology Europe 18:30-32.
Dickinson ME, Waters CW, Fraser SE, Simbuerger E and Zimmermann B (2003).
Multiphoton excitation spectra in biological samples. Journal of Biomedical
Dinauer N, Von Briesen H, Balthasar S, Weber C, Kreuter J and Langer K (2005). Selective
targeting of antibody-conjugated nanoparticles to leukemic cells and primary T-
lymphocytes. Biomaterials 26:5898-5906.
El-Samaligy MS, Rohdewald P and Mahmoud HA (1986). Polyalkyl cyanoacrylate
nanocapsules. Journal of Pharmacy and Pharmacology 38:216-218.
Hans ML and Lowman AM (2002). Biodegradable nanoparticles for drug delivery and
targeting. Current Opinion in Solid State & Materials Science 6:319-327.
Horisawa E, Kubota K, Tuboi I, Sato K, Yamamoto H, Takeuchi H et al. (2002). Size-
dependency of DL-lactide/glycolide copolymer particulates for intra-articular
delivery system on phagocytosis in rat synovium. Pharmaceutical Research
Huang SH, Heikal AA and Webb WW (2002). Two-photon fluorescence spectroscopy and
microscopy of NAD(P)H and flavoprotein. Biophys J 82:2811-2825.
Kohli AK and Alpar HO (2004). Potential use of nanoparticles for transcutaneous vaccine
delivery: Effect of particle size and charge. International Journal of
Konig K (2000). Multiphoton microscopy in life sciences. J Microsc 200:83-104.
Konig K and Riemann I (2003). High-resolution multiphoton tomography of human skin with
subcellular spatial resolution and picosecond time resolution. Journal of
Biomedical Optics 8:432-439.
Kotrotsiou O, Kotti K, Dini E, Kammona O and Kiparissides C (2005). Nanostructured
materials for selective recognition and targeted drug delivery. Journal of Physics:
Conference Series 10:281-284.
Lamprecht A, Ubrich N, Yamamoto H, Schaefer UF, Takeuchi H, Maincent P et al. (2001).
Biodegradable Nanoparticles for Targeted Drug Delivery in Treatment of
Inflammatory Bowel Disease. The Journal of Pharmacology and Experimental
Lombardi Borgia S, Regehly M, Sivaramakrishnan R, Mehnert W, Korting HC, Danker K et
al. (2005). Lipid nanoparticles for skin penetration enhancement--correlation to
drug localization within the particle matrix as determined by fluorescence and
parelectric spectroscopy. Journal of Controlled Release 110:151-163.
Luengo J, Weiss B, Schneider M, Ehlers A, Stracke F, Koenig K et al. (accepted). Influence
of nanoencapsulation on human skin transport of flufenamic acid. Skin
Pharmacology and Physiology.
Meuwissen MEMJ, Junginger HE, Bouwstra JA, Janssen J and Cullander C (1998). A cross-
section device to improve visualization of fluorescent probe penetration into the
skin by confocal laser scanning microscopy. Pharmaceutical Research 15:352-
Pena AM, Strupler M, Boulesteix T, Godeau G and Schanne-Klein MC (2005). Spectroscopic
analysis of keratin endogenous signal for skin multiphoton microscopy: erratum
(vol 13, pg 6268, 2005). Opt Express 13:6667-6667.
Ravi Kumar MNV, Sameti M, Kneuer C, Lamprecht A and Lehr C-M: Polymeric
nanoparticles for drug and gene delivery, Nalwa HS (ed): Encyclopedia of
Nanoscience and Nanotechnology, pp 1-19 (American Scientific Publishers,
Schneider M, Barozzi S, Testa I, Faretta M and Diaspro A (2005). Two-photon activation and
excitation properties of PA-GFP in the 720-920 nm region. Biophys J 89:1346-
Soppimath KS, Aminabhavi TM, Kulkarni AR and Rudzinski WE (2001). Biodegradable
polymeric nanoparticles as drug delivery devices. Journal of Controlled Release
Toll R, Jacobi U, Richter H, Lademann J, Schaefer H and Blume-Peytavi U (2004).
Penetration profile of microspheres in follicular targeting of terminal hair follicles.
Journal of Investigative Dermatology 123:168-176.
Van Kuijk-Meuwissen MEMJ, Junginger HE and Bouwstra JA (1998a). Interactions between
liposomes and human skin in vitro, a confocal laser scanning microscopy study.
Biochimica et Biophysica Acta - Biomembranes 1371:31-39.
Van Kuijk-Meuwissen MEMJ, Mougin L, Junginger HE and Bouwstra JA (1998b).
Application of vesicles to rat skin in vivo: A confocal laser scanning microscopy
study. Journal of Controlled Release 56:189-196.
Volodkin DV, Sukhorukov GB and Larionova NI (2004). Protein encapsulation via porous
CaCO3 microparticles templating. Biomacromolecules 5:1962-1972.
Wagner H, Kostka K-H, Lehr C-M and Schaefer UF (2001). Interrelation of permeation and
penetration parameters obtained from in vitro experiments with human skin and
skin equivalents. Journal of Controlled Release 75:283-295.
Wagner H, Lehr C-M, Schaefer UF and Kostka K-H (2003). pH profiles in human skin:
Influence of two in vitro test systems for drug delivery testing. European Journal
of Pharmaceutics and Biopharmaceutics 55:57-65.
Wagner H, Schaefer UF, Kostka K-H and Adelhardt W (2004). Effects of various vehicles on
the penetration of flufenamic acid into human skin. European Journal of
Pharmaceutics and Biopharmaceutics 58:121-129.
Wartlick H, Michaelis K, Balthasar S, Kreuter J, Langer K and Strebhardt K (2004). Highly
specific HER2-mediated cellular uptake of antibody-modified nanoparticles in
tumour cells. Journal of Drug Targeting 12:461-471.
Xu C, Zipfel W, Shear JB, Williams RM and Webb WW (1996). Multiphoton fluorescence
excitation: New spectral windows for biological nonlinear microscopy. Proc Natl
Acad Sci U S A 93:10763-10768.
Yu B, Blankschtein D, Langer R, Kim KH and So PTC (2003). Visualization of oleic acid-
induced transdermal diffusion pathways using two-photon fluorescence
microscopy. Journal of Investigative Dermatology 120:448-455.
Yu B, Kim KH, So PTC, Blankschtein D and Langer R (2002). Topographic heterogeneity in
transdermal transport revealed by high-speed two-photon microscopy:
Determination of representative skin sample sizes. Journal of Investigative
Zipfel WR, Williams RM and Webb WW (2003). Nonlinear magic: Multiphoton microscopy
in the biosciences. Nature Biotechnology 21:1369-1377.
Figure 1: 325 × 325 µm2 multiphoton optical sections of human skin treated with a
hydrogel suspension of two color labeled nanoparticles. The sub-surface depths of the
displayed images are (a) –6µm, (b) –9µm, (c) –12µm, (d) –15µm, (e) –24mm, (f) –33µm. The
keratin autofluorescence clearly shows the surface of the dermatoglyphs. The excitation
power for the multiphoton optical sections was PEX = 5 mW, the pixel acquisition time tPx =
Figure 2: Fluorescence intensity profile of two multiphoton excited nanoparticles in
situ. The bimodal fit yields a full width at half maximum (FWHM) of 0.5µm. This is a
reasonable value for the diffraction limited resolution of nanoscale particles under the present
Figure 3: 325 × 325 µm2 multiphoton optical sections at a depth of –27µm 15, 50 and
315 minutes after application of the nanoparticular formulation.
Figure 4: One- (solid line) and two-photon (open circles plus B-spline fit) excitation
spectra are displayed in the left panels. The right panels show the related fluorescence spectra
of fluorescein isothiocyanate (FITC) dextran (a) and Texas Red (b). All spectra are
normalized, the one- and two-photon abscissa are aligned for equal transition energies. One
and two-photon excitation spectra may differ considerably because of the converse selection
rules of the related absorption processes. The excitation wavelengths at λ1P = 488 nm and
543 nm, as well as at λ2P = 800 nm are accentuated by dotted vertical lines.
Figure 5: Multiphoton spectral imaging: (a) True color representation of a spectral
image at z = –21 µm with plots of the ROIs used for spectral analysis. (b) Fluorescence
spectra series of ROI 1 showing the decline of Texas Red in the gel matrix (including
particles). The Texas Red content in the particles is low and shows no clear trend as depicted
in the series of the average spectra of 15 particles per image in (c). Since the Texas Red
content of the particles does not change significantly, the decline of Texas Red in ROI 1 must
occur in the gel matrix itself. The excitation power for the multiphoton spectral images was
PEX = 15 mW, the pixel acquisition time tPx = 3.2 µs.
Figure 6: Quantitative spectral analysis of the Texas Red concentration in different
skin compartments. Scheme (a) shows the baselines and spectral positions used for calculating
the emission ratios used to determine the relative contents of the drug-model Texas Red in the
ROIs. Panels (b, c, d) display the temporal evolution of the emission ratios as a measure of
Texas Red content in the ROIs. In ROI 2 the Texas Red emission at 612 nm is divided by the
average keratin autofluorescence intensity at 500 nm and 525 nm because of the interference
of the broad autofluorescence emission with noise effects around 500 nm.
Figure 7: Multitracking experiments: (a, b, c) 325 × 325 µm2 combined optical
sections in –6, -16 and –32 µm depth. (d) A detail image of particles and skin surface. Each
panel consists of a multiphoton excited image (predominantly keratin autofluorescence and
fluorescein, grey-scale, top left image), a 488 nm excited image (fluorescein, green-scale, top
right image) and a 543 nm excited image (Texas Red, orange-scale, bottom left image), as
well as an overlay of which (bottom right image). The excitation powers for the optical
sections in the multitracking study were PEX(800 nm) = 10 mW, PEX(488 nm) = 50 µW and
PEX(543 nm) = 36 µW, the pixel acquisition time was always tPx = 3.2 µs.
Figure 8: xz-multitracking sections composed from a stack of xy-optical sections
with 3 µm distance in between. This representation nicely reveals the penetration profiles of
Texas red into the skin and the correlation of the penetration behavior with skin morphology.
The color coding is in accordance with Figure 7.
Figure 9: (a) 2×2µm2 scanning force microscopy image of an air-dried aqueous
suspension of the PLGA nanoparticles on a glass substrate. The image was acquired in the
tapping mode, 0.2Hz scan speed. The topography from 180nm to 550nm altitude is encoded
in the grey scale. (b) Histogram of the particle diameter d obtained by photocorrelation
spectroscopy. The line is a Gaussian fit to the data.
Figure 1: 325 × 325 µm2 multiphoton optical sections of human skin treated with a hydrogel suspension of two
color labeled nanoparticles. The sub-surface depths of the displayed images are (a) –6µm, (b) –9µm, (c) –12µm, (d)
–15µm, (e) –24mm, (f) –33µm. The keratin autofluorescence clearly shows the surface of the dermatoglyphs. The
excitation power for the multiphoton optical sections was PEX = 5 mW, the pixel acquisition time tPx = 3.2 µs.
Figure 2: Fluorescence intensity profile of two multiphoton excited nanoparticles in situ. The bimodal fit yields a
full width at half maximum (FWHM) of 0.5µm. This is a reasonable value for the diffraction limited resolution of
nanoscale particles under the present conditions.
Figure 3: 325 × 325 µm2 multiphoton optical sections at a depth of –27µm 15, 50 and 315 minutes after application
of the nanoparticular formulation.
Figure 4: One- (solid line) and two-photon (open circles plus B-spline fit) excitation spectra are displayed in the left
panels. The right panels show the related fluorescence spectra of fluorescein isothiocyanate (FITC) dextran (a) and
Texas Red (b). All spectra are normalized, the one- and two-photon abscissa are aligned for equal transition
energies. One and two-photon excitation spectra may differ considerably because of the converse selection rules of
the related absorption processes. The excitation wavelengths at λ1P = 488 nm and 543 nm, as well as at λ2P =
800 nm are accentuated by dotted vertical lines.
Figure 5: Multiphoton spectral imaging: (a) True color representation of a spectral image at z = –21 µm with plots
of the ROIs used for spectral analysis. (b) Fluorescence spectra series of ROI 1 showing the decline of Texas Red in
the gel matrix (including particles). The Texas Red content in the particles is low and shows no clear trend as
depicted in the series of the average spectra of 15 particles per image in (c). Since the Texas Red content of the
particles does not change significantly, the decline of Texas Red in ROI 1 must occur in the gel matrix itself. The
excitation power for the multiphoton spectral images was PEX = 15 mW, the pixel acquisition time tPx = 3.2 µs.
Figure 6: Quantitative spectral analysis of the Texas Red concentration in different skin compartments. Scheme (a)
shows the baselines and spectral positions used for calculating the emission ratios used to determine the relative
contents of the drug-model Texas Red in the ROIs. Panels (b, c, d) display the temporal evolution of the emission
ratios as a measure of Texas Red content in the ROIs. In ROI 2 the Texas Red emission at 612 nm is divided by the
average keratin autofluorescence intensity at 500 nm and 525 nm because of the interference of the broad
autofluorescence emission with noise effects around 500 nm.
Figure 7: Multitracking experiments: (a, b, c) 325 × 325 µm2 combined optical sections in –6, -16 and –32 µm
depth. (d) A detail image of particles and skin surface. Each panel consists of a multiphoton excited image
(predominantly keratin autofluorescence and fluorescein, grey-scale, top left image), a 488 nm excited image
(fluorescein, green-scale, top right image) and a 543 nm excited image (Texas Red, orange-scale, bottom left
image), as well as an overlay of which (bottom right image). The excitation powers for the optical sections in the
multitracking study were PEX(800 nm) = 10 mW, PEX(488 nm) = 50 µW and PEX(543 nm) = 36 µW, the pixel
acquisition time was always tPx = 3.2 µs.
Figure 8: xz-multitracking sections composed from a stack of xy-optical sections with 3 µm distance in between.
This representation nicely reveals the penetration profiles of Texas red into the skin and the correlation of the
penetration behavior with skin morphology. The color coding is in accordance with Figure 7.
Figure 9: (a) 2×2µm2 scanning force microscopy image of an air-dried aqueous suspension of the PLGA Download full-text
nanoparticles on a glass substrate. The image was acquired in the tapping mode, 0.2Hz scan speed. The topography
from 180nm to 550nm altitude is encoded in the grey scale. (b) Histogram of the particle diameter d obtained by
photocorrelation spectroscopy. The line is a Gaussian fit to the data.