2.1.

Animals

In vivo studies were carried out on six outbred albino rats 1 year of age and the body mass of $\sim 400\mathrm{g}$. During all experimental studies, animals were kept in accordance with the European Convention for the Protection of animals used for experimental and other scientific purposes.³⁰ Prior to all treatments, the rats were anesthetized with Zoletil 50 (Virbac, France) in a dose of $0.05\mathrm{mg}/\mathrm{kg}$. On the sites under study, hair was thoroughly removed using a cream for depilation.

2.2.

Particles

Commercially available lanthanide-doped UCPs [${\mathrm{Y}}_2{\mathrm{O}}_3:\mathrm{Yb}$, Er] PTIR660-UF (Phosphor Technology, United Kingdom) with an average size of $\sim 1.6\mu \mathrm{m}$ and in-house synthesized ${\mathrm{CuInS}}_2/\mathrm{ZnS}$ QDs with an average size $\sim 20\text{-}\mathrm{nm}$ coated with polyethylene glycol (PEG)-based amphiphilic polymer (see Appendix) were used in the experiments.

The UCPs have a luminescent band near the 650-nm wavelength, upon 980-nm laser excitation [Fig. 1(a)]. The UCPs [${\mathrm{Y}}_2{\mathrm{O}}_3:\mathrm{Yb}$, Er] are a promising new generation of agents for bioimaging. Upconversion as a phenomenon utilizes a sequential absorption of multiple photons via long lifetime (metastable) and real ladder-like energy levels of trivalent lanthanide ions embedded in an appropriate inorganic host lattice to produce higher energy anti-Stokes luminescence.^31–33 It thereby converts two or more low-energy excitation photons of NIR light into shorter wavelength emission in NIR and visible ranges. High photostability, large blue shift of the upconversion luminescence relative to the wavelength of the exciting radiation, and absorption in the range of maximal transparency of tissues (600 to 1000 nm) make it possible to track single particles and perform their long-term detection.³⁴

Fig. 1

Luminescent spectra of (a) UCP (excitation wavelength 980 nm) and (b) QDs (excitation wavelength 405 nm).

The synthesized QDs have an irregular shape with the crystal size less than 10 nm and a hydrodynamic diameter around 20 nm.³⁵ The luminescent peak of QDs is at the 680-nm wavelength [Fig. 1(b)] with the quantum yield of $\sim 45\%$ measured at 488-nm excitation with Rhodamine 6G as a reference dye.

Aqueous suspensions of the particles were prepared immediately before their use. Concentrations of UCPs and QDs were $0.625\mu \mathrm{g}/\mu \mathrm{L}$ and $0.6\times {10}^{-6}\mathrm{mol}/\mathrm{L}$, respectively. To prevent aggregation of the particles, UCP and QD suspensions were sonicated in an ultrasonic bath before their topical application to the skin surface.

2.3.

Experimental Methods

2.3.1.

Particle delivery protocol

All six rats were subjected to the procedure of FLMA. For carrying out the FLMA procedure, a commercially available Palomar Lux 2940 pulse erbium laser (Palomar Medical Technologies Inc., USA) operating at the 2940-nm wavelength was used. Laser pulse energy was 3 J with a duration of 20 ms; the temporal profile of the pulse was modulated by three subpulses with duration of each about 5 ms. The laser beam was split into 64 microbeams using a microlens array. Thus, 64 vertical microchannels were created over the $8\times 8{\mathrm{mm}}^2$ skin area.

The animals were divided into two groups. In the first one (two animals), the efficiency of administration of UCP and QD suspensions was studied by using ultrasound or massage.

Five sites were marked as shown in Fig. 2(a): the site 0 corresponded to intact skin without any treatment, the sites 1 and 3 were treated by the UCP suspension, and the sites 2 and 4 by the QD suspension. The volume of both suspensions was equal to $23\mu \mathrm{L}$. The UCPs and the QDs were delivered into microchannels by using sonophoresis (sites 3 and 4) or mechanical massage (sites 1 and 2) for 5 min each.

Fig. 2

Scheme of the studied sites of rat skin in vivo for (a) the first group and (b) the second group of experimental animals.

A commercially available device Dynatron 125 (Dynatronics, USA), equipped with a 2.2-cm diameter US transducer was applied for sonophoresis in the continuous-wave mode with the frequency of 1 MHz and the power density of $1.5{\mathrm{W}/\mathrm{cm}}^2$. Remains of the suspensions were removed from the skin surface using distilled water.

In the second group (four animals), six sites were marked as shown in Fig. 2(b): the site 0 corresponded to intact skin without any treatment, the sites 1 and 4 were treated by UCP suspension, and the sites 2 and 5 by the QD suspension. The volume of the applied suspensions was the same as for the first group. The site 3 remained untreated by any particle suspension. This site (control) was used for the monitoring of the changes of skin state after FLMA. The UCPs and the QDs were delivered into the microchannels by using sonophoresis (the sites 1, 2, 4, and 5) for 5 min each.

2.3.2.

Particle imaging

Imaging of the particles was performed using different optical modalities, such as optical coherence tomography, luminescent spectroscopy, and confocal microscopy, as well as conventional optical microscopy (histological analysis).

The microchannels were visualized just after FLMA and after filling them by the particle suspensions in the first and the second groups. The imaging was performed with a commercially available spectral-domain optical coherence tomography (OCT) system OCP930SR (Thorlabs, USA), operating at the central wavelength of $930\pm 5\mathrm{nm}$ with $100\pm 5\mathrm{nm}$ full-width-at-half-maximum spectrum and having the following characteristics: numerical aperture of 0.22, optical power of 2 mW; the scanning length along the skin surface of 6 mm with the axial and the lateral resolutions on the air of 6.2 and $9.6\mu \mathrm{m}$, respectively.

To study the influence of ultrasound and massage on intensity of luminescent signal from particles inside the skin, we carried out the spectral measurement immediately after FLMA and particle delivery in the first group of animals.

For monitoring UCP and QD localization in the rat skin in vivo (the second group), luminescent spectroscopy was applied every day during the entire period of in vivo studies (nine days). Figure 3 schematically shows the experimental system for detection of UCP and QD luminescent spectra by a spectrometer QE6500 (Ocean Optics, USA). The excitation of UCPs was performed at 980-nm wavelength and that of QDs at 405-nm using a semiconductor laser module DMH980-200 (Laser Systems LAS, Russia, 20 mW) and a violet laser pointer Pen Style (HangZhou NaKu Technology Co. Ltd, China, 300 mW), respectively. The fiber-optic probe was fixed over the skin surface at the distance of 25 mm. Luminescent spectra were measured with a 100-ms acquisition time.

Fig. 3

Scheme of the experimental setup used for spatial localization of upconversion luminescent particles and QDs inside rat skin in vivo.

The withdrawal of the animals from the experiment and sampling of tissues for the luminescent and morphological study was performed on the first day (two rats from the first and the second groups) and on the last day of the experiment (two rats from the second group).

Unstained sections were studies with luminescent microscopy. Since luminescence of UCPs and QDs is excited by different wavelengths, we used different equipment for the observation. Luminescent images of skin histological sections with QDs were obtained at room temperature ($\sim 20°\mathrm{C}$) using a standard confocal luminescent microscope (Leica TCS SP8, Germany) with $10\times$ and $100\times$ magnification under 405-nm laser excitation. Images of UCPs were taken with an optical microscope OGME-PZ (Sapphire, Russia) equipped with a CCD camera (Videoscan 415/P-USB, Videoscan, Russia) under 980-nm laser excitation.

Then the sections were stained with hematoxylin–eosin according to the standard procedure.³⁶ Morphometric studies of histological preparations were done using a digital image analysis system $\mu \mathrm{Vizo}$-103 medical microvisor (LOMO, Russia).

3.1.

Optical Coherence Tomography Imaging

Figure 4 shows the series of OCT images of skin sites from the first group of rats immediately after FLMA alone (a), FLMA followed by QD delivery with massage (b), FLMA followed by UCP delivery with massage (c), FLMA followed by QD delivery with sonophoresis (d), and FLMA followed by UCP delivery with sonophoresis (e). In the images, two layers—epidermis (E) and dermis (D)—are clearly distinguishable. Microchannels (MC) look like cones. Figure 4(a) allows for evaluation of the separation between the centers of the laser-perforated channels as $\sim 1\mathrm{mm}$ with the diameter of their openings at skin surface as 150 to $250\mu \mathrm{m}$ and the depth as 450 to $500\mu \mathrm{m}$.

Fig. 4

The OCT images of FLMA-treated skin in vivo: (a) immediately after FLMA, (b) after QD delivery using a massage; (c) after UCP delivery using a massage; (d) after QD delivery using a sonophoresis, and (e) after UCP delivery using a sonophoresis. E, epidermis; D, dermis; and MC, microchannel.

As follows from OCT images (c) and (e), after both massage and sonophoresis, the largest part of the microchannel is filled up by the UCP suspension. It is well seen that QDs cannot be visualized with high-contrast on the background of the skin scattering [see images (b) and (d)]. UCPs provide high contrast of OCT visualization of the microchannels because of their much larger size than QDs ($1.6\mu \mathrm{m}$ versus 20 nm). Comparison of data presented in Figs. 4(c) and 4(e) shows that in contrast to the massage, the ultrasound treatment allows for deeper filling of the ablated channels in the skin. It is in a good agreement with the results of Ref. 12, where the increase of skin permeability for both in vitro and in vivo studies using ultrasound irradiation was shown.

3.2.

Luminescent Analysis

The spatial localization of the particles in the microchannels immediately after FLMA and the particles delivery can be observed on luminescent images presented in Fig. 5. Luminescence of UCPs was detected by optical microscope under 980-nm laser excitation [see Figs. 5(a) and 5(b)] and luminescent of QDs was detected by confocal microscope under 405-nm laser excitation [see Figs. 5(c) and 5(d)].

Fig. 5

Luminescent images of histological section of the perforated skin obtained using: (a, b) an optical microscope equipped with CCD camera under 980-nm laser excitation of the UCPs and (c, d) a confocal microscope under 405-nm laser excitation of the QDs. The fluorescing UCPs are marked with the red circle. The images were obtained from the depth of (c) $0\mu \mathrm{m}$ and (d) $40\mu \mathrm{m}$ immediately after FLMA and particle delivery.

It is well seen that near the skin surface, the particles were localized along the channel walls; the main volume of the suspensions, apparently, was removed from the channels during the washing-off. However, as we can see in Fig. 4, the suspension remained at the depth of 400 to $500\mu \mathrm{m}$. After healing of the tissue around the channels and as a result of their closing, the particles are accumulated within a localized space in tissue. QDs and UCPs were not detected in the histological sections with a confocal luminescent microscope after nine days of particle delivery.

Figure 6 shows the luminescent spectra of the UCPs [see Fig. 6(a)] and QDs delivered into skin by massage and sonophoresis [see Fig. 6(b)]. It can be seen that sonophoresis results in the increased luminescent signal from the particles in comparison with the massage, apparently, due to the increase of quantity of the particles penetrated into the channels.

Fig. 6

Luminescent spectra of the (a) UCPs (excitation wavelength 980 nm) and (b) QDs (excitation wavelength 405 nm) delivered into skin by massage and by sonophoresis. Asterisk represents luminescent bands of endogenous porphyrins.

Measured kinetics of luminescent intensity (Figs. 7 and 8) and histological analysis (Figs. 9 and 10) confirm that the depot of the particles can be created inside the microchannels.

Fig. 7

Luminescent spectra of the (a) UCPs (excitation wavelength 980 nm) and (b) QD (excitation wavelength 405 nm) embedded into rat skin by sonophoresis: with topically applied Omnipaque™ immediately after FLMA and particle delivery (1); without Omnipaque™ immediately after FLMA and particle delivery (2); with Omnipaque™ after eight days of particle delivery (3); and without Omnipaque™ after eight days of particle delivery (4). Asterisk represents luminescent bands of endogenous porphyrins.

Fig. 8

The time dependence of the luminescent maximum of the (a) UCPs (${\lambda }_{\max }=658\mathrm{nm}$) and (b) QDs (${\lambda }_{\max }=680\mathrm{nm}$) embedded into the rat skin by sonophoresis and with (black squares) and without (red circles) optical clearing by topical application of Omnipaque™.

Fig. 9

Histological sections of the skin: (a) intact skin, (b) skin after laser ablation without embedded particles, (c) with the UCPs delivered by massage, and (d) with the UCPs delivered by sonophoresis (the mean size of the skin damaged area $[260\pm 70]\times [110\pm 30]{\mu \mathrm{m}}^2$). The samples were stained with hematoxylin and eosin. The contoured zones indicate coagulation necrosis (b, c, d).

Fig. 10

Histological sections of the perforated skin with the UCPs delivered by sonophoresis after nine days. The samples were stained with hematoxylin and eosin. The contoured zones indicate coagulation necrosis.

Figure 7 shows the luminescent intensity produced by UCPs (a) and QDs (b) embedded into the rat skin in vivo.

Under excitation by 980-nm laser, skin autofluorescence in the red spectral region is negligible. By contrast, 405-nm radiation excites fluorescence from endogenous skin fluorophores: three bands of porphyrins in the 630 to 710-nm region and NADH near 480 nm.³⁷

The luminescent peak of QDs shifted in the skin from 680 to 670 nm overlaps with the fluorescent bands of porphyrins [see Fig. 7(b)]. The QD and UCP luminescence was observed for four days [see Fig. 7(b) (3, 4)] and eight days [see Fig. 7(a) (3, 4)], respectively, with a gradual decrease in the intensity with time.

Healing and regeneration of skin epidermis (about 4 to 5 days) can prevent penetration of 405-nm light to the QD depot in the skin. At the same time, 980-nm light penetrates significantly deeper and can excite luminescence of UCPs for more prolonged time during skin healing and regeneration and afterward.

The decrease of luminescent signal can also be connected with the diffusion of the particles from the observation zone to surrounding tissues. In Ref. 38, it was shown that contrast of OCT-images of human skin with embedded particles also decreased during a few days after their delivery. After the full regeneration of the skin integrity, quantity of the particles was so little that they did not register in the OCT images. Difference between time interval (days) during which luminescence from UCPs and QDs is detectable can be related to particle size and their diffusion from the initial delivery volume: larger UCPs remained localized inside the skin channels for a long time, whereas much smaller QDs diffuse from the channels and distributed in the surrounding tissues.

It is well seen that the maximal luminescent signal is observed after the skin optical clearing. Figure 8 shows kinetics of the maximal intensity of the luminescent bands of the UCPs (${\lambda }_{\max }=658\mathrm{nm}$) (a) and QDs (${\lambda }_{\max }=680\mathrm{nm}$) (b) embedded into the rat skin by sonophoresis and with and without optical clearing by topical application of Omnipaque™.

Figures 7 and 8 show that the topical application of the immersion agent Omnipaque™ allows for enhancement of the luminescent signal from the particles located inside skin dermis due to reduction of light scattering of tissue layers, where excitation and fluorescent signals propagate. This is a so-called optical clearing technique application that increases probing depth and contrast of optical imaging methods.²⁹ It was demonstrated, in particular, for fluorescent spectroscopy and microscopy.^20,27,29 Omnipaque™ has refractive index (1.439) close to refractive index of main skin scatterers—collagen fibers (1.43).²⁷ Thus, a partial replacement of the interstitial fluid by Omnipaque™ causes the matching of the refractive indices of the tissue components, its better homogeneity and reduction of scattering.

In Fig. 8(a), we can see that skin treatment by Omnipaque™ does not affect the intensity of the luminescent signal after four days post-FLMA and particle delivery. We suppose that after full regeneration of skin barrier function, Omnipaque™ does not penetrate effectively into skin dermis, and therefore, the 20-min agent application is not enough for signal enhancement.

3.3.

Histological Analysis

Microphotographs of histological sections of the studied skin areas from the first group measured immediately after UCPs delivery are shown in Fig. 9. The control sample [intact skin without any treatment; see Fig. 9(a)] represents skin with normal multilayer epithelium (epidermis). The skin appendages are visible in the dermis. QDs could not be detected in the histological sections with conventional optical microscopy due to their small size (20 nm).

For the samples taken from the FLMA-zone [ablation only; see Fig. 9(b)], crater-like defects crossing epidermis and dermis, regularly occurring on average every $600\mu \mathrm{m}$ with the development of coagulation necrosis in tissue around them, are observed. In Fig. 9(b), one of the craters covers an area of $\text{width}\times \text{depth}$ equal to $153\times 170\mu {\mathrm{m}}^2$ as the maximal depth of necrosis edge is $363\mu \mathrm{m}$. Coagulation necrosis of the thickness about $10\mu \mathrm{m}$ is formed around the perforated channel. In addition, full-blooded vessels around the zone of influence are observed.

Figure 9(c) shows UCPs on crater walls delivered by massage; this crater is an internal part of the microchannel. Outside, the channel is surrounded by a necrotic zone with a radius approximately equal to the crater depth. The size of the crater is $155\times 58\mu {\mathrm{m}}^2$.

It is important to note that sonophoresis increases the channel size. The form of the channel changes from conical to spherical. We anticipate that the necrotic tissue is detached and pushed out from the channel by sonophoresis. This hypothesis is confirmed by a thinner zone of coagulation necrosis observed in Fig. 9(d) than in Fig. 9(c).

Figure 10 presents a fragment of skin with a microchannel filled by UCPs of the rat from the second group. A depth of localization is $323\mu \mathrm{m}$. The size of the defect is $21\times 63\mu {\mathrm{m}}^2$.

Analysis of this histological section confirms our hypothesis that particles with the microscale dimension localized inside the channels remain there after the channel healing and do not penetrate in surrounding tissue. Thus, a depot of particles can be created for prolonged drug release.

QDs, on the contrary, can diffuse through the surrounding tissue and leave the channel zone. Therefore, we cannot see accumulation of the particles in the histological sections and luminescent signal to the ninth day post-treatment. It was shown earlier that the use of FLMA allowed for delivery of microparticles into dermis and creation of a particle depot.^11,38,39 For example, biocompatible ${\mathrm{CaCO}}_3$ microcontainers ($4.0\pm 0.8\mu \mathrm{m}$) containing ${\mathrm{Fe}}_3{\mathrm{O}}_4$ nanoparticles were delivered to the depth about $300\mu \mathrm{m}$ in dermis.¹¹ On the seventh day, the dissolving of the microcontainers and release of the content into dermis was observed. In Refs. 38 and 39, it was shown that particles with diameter about 100 nm and more can remain in skin in vivo during 30 days.