Two-photon tomography is frequently applied in in vivo dermatological research [116,117,118,119], mainly uses excitation in the range of 760–820 nm (“skin optical transparent window-I”), and provides high-quality-resolved images of skin structures from the surface down to ≈150–200 µm with a subcellular lateral (<0.6 µm) and axial (<2 µm) resolution. The standard 2PT operates with two channels—autofluorescence (AF) and second-harmonic generation (SHG)—which provide information on the distribution of AF-generating fluorophores representing the main morphological features of the skin (cells, melanin, and elastin) [116], and SHG shows the distribution of non-centrosymmetric collagen type I in the dermis [116] and recently discovered crystallized urea dendriform structures in the SC of glabrous skin [77]. An extended version of the 2PT, additionally combined with spectral imaging of the AF [120], FLIM [116], CARS [121], and/or RCM [122] channels, considerably expand the possibilities for molecular imaging and practical applications.

The spectral imaging of AF offers the advantage of seeing the distribution of emitted AF photons as a function of wavelength, which is important information for the choice of the appropriate AF transmission filters [120,123]. The 2PT-FLIM technique is a time-resolved technique that enables the detection of fluorescence lifetime decay curves, determining fast (>0.2 ns) and slow (<2.4 ns) lifetime components, which are described mainly by free and protein-bound NAD(P)H, respectively, with increased free NAD(P)H indicating reduced metabolic activity [124,125]. Fluorescence lifetimes are dependent on the chemical composition of the target fluorophores and their interaction with the surroundings [126], and vary for different skin constituents [119]. Thus, based on the individual lifetime characteristics, it was possible to image melanin [127], collagen type III [128], mast cells [129], and macrophages [130] in skin in vivo and ex vivo, to determine the metabolic changes during keratinocyte proliferation in vitro [131] and to differentiate and diagnose skin cancer ex vivo [132] using 2PT-FLIM. The 2PT-CARS technique is able to visualize the distribution of lipids and water in the skin in vivo and ex vivo [133]. The combination of 2PT with the reflectance regime provides a fast overview provided by RCLSM [122,134] with the possibility of detailed multimodal 2PT imaging of the area of interest [135]. The operating principle of 2PT-(AF, FLIM, RCM, SHG, and CARS) is described elsewhere [76,121,122].

The 2PT-AF (intensity and spectral channels) technique is a valuable non-invasive method to study skin penetration. If the applied substance exhibits intense two-photon-excited AF whose intensity exceeds skin AF and/or whose AF emission spectrum differs from that of the skin constituents, this can be visualized in the skin [120,136]. For instance, under two-photon excitation at 780 nm, the fluorescence emission maxima for the hydrophobic rhodamine B hexyl ester and hydrophilic sulforhodamine B fluorescent dyes at 578 and 586 nm, respectively, overlap minimally with skin AF. This allows their visualization and quantification in the skin, as well as an estimation of the effect of oleic acid as a penetration enhancer [137], and the visualization of the intracellular penetration pathway [138]. Skin penetration was determined for topically applied highly fluorescent 6-carboxyfluorescein in a core droplet of the tailorable nano-emulsions [114], fluorescently labelled dextran in combination with topical enhancers [139], gold nanoparticle dispersion [140], zinc oxide nanoparticles [141], nanobeads [142], FITC dextran, Texas Red [120], FITC-labeled PLGA nanoparticles [136], and sunscreen labeled with fluorescein dye [143]. The distribution and deposition in the SC, furrows, and hair follicles could be clearly visualized. In addition, the accumulation of the highly fluorescent chemotherapeutic drug doxorubicine exclusively within the cytoplasm in the perinuclear area has been demonstrated in vitro in chemo-sensitive living cells and their chemo-resistant variants using 2PT-AF [144], suggesting the possibility of visualizing doxorubicine in the skin of chemotherapy patients in vivo (inside–outside penetration). The major limitation of 2PT-AF (intensity and spectral channels) is the presence of skin intrinsic AF intensity and broad AF emission spectrum, which make this method not sensitive enough in most practical cases.

The 2PT-FLIM technique is a pseudo-chemical analysis, which significantly improves and extends the capabilities of 2PT-AF for skin penetration studies by providing additional information on the fluorescence lifetime parameters of the target substances. The applied substances, which have their own two-photon-excited AF, usually characterized by the combination of fast and slow AF lifetimes (that are normally used as parameters for bi-exponential fitting [76]) that are different from the AF lifetime values of the endogenous skin constituents, can potentially be evaluated in the skin [147]. [148], uncoated silver nanoparticles [149], nanobeads [142], nanodiamonds [102], minocycline and tazarotene—used for the treatment of acne vulgaris [150,151], anti-inflammatory compound GSK2894512—a drug used for the treatment of atopic dermatitis and psoriasis [152], ethinyl estradiol—used in hormonal therapy [153], bovine serum albumin and hyaluronic acid [154], sodium fluorescein [145,147], Nile red [112,145], 5-carboxyfluorescein-labelled liposomes [147], coated and uncoated zinc oxide nanoparticles [149,155,156], gold nanoparticles [157], and dendritic core multishell nanotransporters [112]. It was recently shown that the penetration of non-fluorescent propylene glycol can also be determined indirectly in skin based on the pH-dependent increase in the AF lifetime of SC components [145]. The 2PT-FLIM technique is also applicable for the determination of inside–outside penetration, which was recently demonstrated in vivo for carbon black tattoo ink particles diffusing from the dermis into the epidermis in old tattoos [158]. The results demonstrate the possibility of identifying the penetration profiles of multicomponent substances based on the differences in their fluorescence lifetime parameters [151]. The 2PT-FLIM technique can also be successfully used for the visualization of drugs, such as proretinal and retinal nanoparticles, delivered with microneedles [159]. To distinguish between the endogenous and exogenous skin FLIM data, a phasor approach, where the time signal is transformed into a pair of phasor co-ordinates representing the sine and cosine components of a Fourier transform, is highly advantageous [160]. Using a phasor plot, it is possible to visualize many fluorophores simultaneously according to their temporal characteristics [76].

SHG does not require electronic excitation of the molecules and is used to study the penetration of non-centrosymmetric SHG-active substances in the skin, such as zinc oxide nanoparticles, where the intercellular and follicular penetration pathways could be recognized [146]. No zinc oxide nanoparticles were found in viable epidermis [146], suggesting that there is no cellular toxicity induced and it can be safely used in sunscreens [161]. Chemical enhancers such as ethanol, oleic acid, and oleic acid–ethanol, however, facilitate the transdermal delivery of zinc oxide nanoparticles, which is due to the increase in lipid fluidity and/or the extraction of lipids from the SC [162]. This method is very sensitive for screening the SHG-active substances in the SHG-free epidermis. One limitation is glabrous skin whose SC may contain SHG-active crystallized urea dendriform structures [77], which cannot be separated from the applied formulation.

With CARS, the vibrational signatures of the molecules can be determined. Three laser beams (ps or fs pulses) are required to excite CARS: a “pump” (frequency ω_p), a “Stokes” (frequency ω_S), and a “probe” beam (frequency ω_pr) [163]. However, to simplify the setup, the “probe” and “pump” beams are often provided by one laser and have the same frequencies [123]. The interaction with the sample leads to the generation of a coherent optical signal at the anti-Stokes frequency ω_CARS = ω_pr + ω_p − ω_S = 2ω_p − ω_S, which is strongly enhanced when the energy difference between ω_p and ω_S matches the energy of molecular vibration Ω = ω_p − ω_S. The excitation of the CARS signal does not require electronic excitation of the molecules. The working principle of the 2PT-CARS is described in detail elsewhere [75,121,123].

The 2PT-CARS technique allows the multimodal imaging (imaging speed can be <1 s/image [164]) of human skin ex vivo and in vivo with a subcellular resolution [72,121,133,165,166], but is not commonly used in skin penetration studies. It has been shown that the use of 2PT-CARS can visualize the penetration of airborne carbonaceous particulate materials (2693 and 2840 cm⁻¹) [167], retinol (1594 cm⁻¹) [168], and elongated silica microparticles combined with tailorable nanoemulsions packed with glycerol (2845 cm⁻¹) [114] in skin ex vivo. Here, the 2PT-CARS is excited at 1040 nm for the “Stokes” and 803 nm for the “pump” beams (“skin optical transparent window-II and -I”, respectively). The 2PT-CARS technique can also be successfully used for the visualization of drugs, such as betamethasone dipropionate (1750 cm⁻¹), delivered with microneedles [169]. In vivo studies examined the penetration of omega-3-oil (2845 cm⁻¹) [133] and mineral oil (2845 cm⁻¹) [170] into the epidermis, but there was no effective separation between the applied oil and SC lipids, which appears to be a major limitation of this method in penetration studies. Although CARS bands are normally not overlapped with the fluorescence background [123], the strong overlap of lipid- and protein-related Raman bands is also a critical factor in evaluating their separate contributions in the HWN spectral region [171]. Successful quantitative determination is possible for substances whose Raman spectra do not overlap with the skin spectrum, such as deuterated glycerol [172].

In skin analysis, 3PT mainly uses longer excitation wavelengths (≈1200–2200 nm with an optimum at ≈1300 nm, ≈1700 nm [71,173], or 2200 nm [174]) corresponding to “skin optical transparent window-II, -III, and IV”, which are known for the deeper penetration of light into the skin compared to 2PT, where the typical excitation wavelengths are in the spectral region of 710–920 nm [123]—“skin optical transparent window-I”. The higher imaging depth of 3PT compared to 2PT is due to the lower absorption (mainly by water) and the lower scattering of both excitation and emission light. It has been reported that the possible photodamage with three-photon excitation is lower than with two-photon and even one-photon excitation, which is due to the lower absorption by water and lower heating [175]. The imaging speed can be <1 s/image [164]. The endogenous three-photon-excited AF of skin is weak (3PT-AF images are background-free); therefore, the high-contrast exogenous three-photon-excited fluorescent dyes, such as green and red fluorescent proteins [71], iridium (III) complexes [176], moxifloxacin [177], or nanoparticulate materials [178] are required to provide bioimaging. Third-harmonic generation (THG) does not require electronic excitation of the molecules and is sensitive to local differences in the third-order nonlinear susceptibility, refractive index, and dispersion, and is particularly generated by water–lipid or water–protein scaffold interfaces, as well as lipid bodies, fat cells, nerve fibers, membranes, intracellular vesicles [179], and blood capillaries [164]—it does not require exogenous dyes, i.e., is non-invasive.

Studies on skin penetration with 3PT are lacking in the literature. However, 3PT has a strong potential to visualize particulate substances in a similar way as shown for tumor-associated microparticles and aggregated intracellular vesicles in skin in vivo [179], gold nanorods in the skin and brain in vivo and ex vivo [180], the deposition of intravenously injected gold–silver nanocages in mouse liver ex vivo [181], or lipid droplets in mouse liver ex vivo [182]. The latter may mimic oil-in-water/water-in-oil pharmaceutical formulations. The combination with FLIM should greatly broaden the applicability of 3PT in future skin penetration research.

The major advances of 2PT over one-photon CLSM imaging in in vivo skin penetration research are its non-invasiveness, high-quality morphological imaging with a subcellular resolution due to signal generation in a small volume, and increased imaging depth to ≈150–200 µm (less in the presence of hair follicles). Since two-photon absorption occurs in a subfemtoliter volume for a very short time (fs), the skin undergoes negligible photobleaching [149] and phototoxicity [160]. The main drawbacks are the high cost of the 2PT device, the not-very-fast image acquisition (usually 3–12 s/image), and the small skin-screening areas (max. ≈ 300 µm × 300 µm) that make it difficult to find target areas quickly. The 3PT technique has the prospect of being used in skin penetration research.

Substances studied with 2PT-AF in skin should exhibit intense two-photon-excited AF whose intensity exceeds that of skin and/or whose AF emission spectrum differs from that of skin (for 2PT-AF and 2PT-FLIM). This can potentially be realized by using new materials excited in “skin optical transparent window-II” and emitted two-photon-excited AF in “skin optical transparent window-I”, where the intrinsic 2PT-AF intensity of the skin is minimal, such as monomeric and dimeric di-styryl-BODIPY dyes [183]. For the SHG analysis, target substances should be SHG-active. For the 2PT-CARS analysis, the most important requirement for the target substances is the presence of a Raman band that does not overlap with that of skin, preferably in the ≈1700–2820 cm⁻¹ spectral region. The overlap of the Raman bands of the substance and skin reduces the detection sensitivity. Potential limitations of using 3PT in skin penetration research include studying only those substances that exhibit three-photon-excited fluorescence and/or THG-active substances.