Highly Efficient Near-IR Photothermal Microneedles with Flame-Made Plasmonic Nanoaggregates for Reduced Intradermal Nanoparticle Deposition

Abstract

Near-infrared (NIR) photothermal therapy by microneedles (MNs) exhibits high potential against skin diseases. However, high costs, photobleaching of organic agents, low long-term stability, and potential nanotoxicity limit the clinical translation of photothermal MNs. Here, photothermal MNs are developed by utilizing Au nanoaggregates made by flame aerosol technology and incorporated in water-insoluble polymer matrix to reduce intradermal nanoparticle (NP) deposition. The individual Au interparticle distance and plasmonic coupling within the nanoaggregates are controlled by the addition of a spacer during their synthesis rendering the Au nanoaggregates highly efficient NIR photothermal agents. In situ aerosol deposition of Au nanoaggregates on MN molds results in the fabrication of photothermal MNs with thin plasmonic layers. The photothermal performance of these MN arrays is compared to ones made by three methods utilizing NP dispersions, and it is found that similar temperatures are reached with 28-fold lower Au mass due to reduced light scattering losses of the thin layers. Finally, all developed photothermal MN arrays here cause clinically relevant hyperthermia at benign laser intensities while reducing intradermal NP deposition 127-fold compared to conventional MNs made with water-soluble polymers. Such rational design of photothermal MNs requiring low laser intensities and minimal NP intradermal accumulation sets the basis for their safe clinical translation.

1. Introduction

Photothermal therapy (PTT) has attracted attention as an effective, minimally invasive treatment for skin diseases.^[1–3] In PTT, nanoparticles (NPs) can be used to convert electromagnetic radiation to heat. Increasing the local temperature in the skin may improve the therapeutic outcome of various diseases, such as skin tumors and cancer, or bacterial, parasitic, and viral skin infections.^[1] Target temperatures for PTT often range from 38 to 55 °C with a heating above ≈50 °C causing protein denaturation and tissue coagulation which may interfere with successful therapy.^[4] The therapeutic effect of hyperthermia is to i) induce cellular apoptosis and necrosis in cells, ii) increase the blood flow, and iii) elicit immune responses.^[1,4] Furthermore, combining hyperthermia with pharmaceutical treatments may reduce the required drug dose, which is specifically relevant for the prevention of further development of drug resistance.^[5]

For successful application of PTT in dermatological conditions, it is required that the heat is evenly distributed into potentially deep layers of the skin.^[6] Microneedle (MN) arrays have been extensively studied in recent years for PTT-based skin therapy^[7,8] and the photothermal agents that have been delivered via MNs include silica-coated lanthanum hexaboride,^[9–11] gold nanorods^[6,12,13] and nanocages,^[14] Prussian blue NPs,^[15] graphene oxide NPs,^[16,17] black phosphorus quantum dots,^[18] indocyanine green,^[19–21] CuO₂ NPs,^[22] and Nb₂C nanosheets.^[23] In all those examples, the photothermal agent is active in the near-infrared (NIR) wavelength region which avoids nonspecific thermal damage to tissue.^[24] Importantly, MN-assisted delivery of PTT can enhance treatment outcomes in cancers and reduce bacterial infections while distributing the hyperthermia evenly and deep into the skin.^[8]

However, the broad employment of photothermal MN arrays is prohibited due to high production costs of some photothermal agents (e.g., gold nanorods), their poor long-term stability, and/or degradation by photobleaching (e.g., indocyanine green).^[5,25] The lack of high photothermal efficiency of some proposed PTT-based MNs systems also requires high laser intensities up to 5 W cm⁻².^[9,13,26] Such high laser intensities may limit clinical translation of MN arrays for PTT since laser intensities above 0.3–1.0 W cm⁻² at 780–1050 nm are the maximum permissible exposure limit to skin according to the American National Standard for Safe Use of Lasers.^[27] Furthermore, almost all reports to this date have delivered the photothermal agent inside the skin with limited studies on potential toxicity of such intradermal deposition of NPs. Gold nanorods,^[28] black phosphorus nanosheets,^[29] and graphene NPs^[30] may be toxic to dermal cells in vitro and in vivo in zebrafish^[31] and mouse models.^[32,33] An extensive understanding of the long-term biosafety of intradermal NPs deposition and their subsequent accumulation in the human body is not yet established.^[33,34] To improve clinical translation of PTT in dermatology, the intradermal deposition of NPs should be avoided in medical devices.

To address these challenges, we introduce here four fabrication methods of PTT-based MN arrays made from water-insoluble polymers to facilitate removal of the photothermal NPs from the skin after treatment. We i) utilized cost-effective flame-made Au nanoaggregates as the photothermal agent allowing for tunable plasmonic coupling and production in industrial scale,^[35] ii) directly deposited such flame-made Au nanoaggregates as thin films in MN mold cavities resulting in highly efficient photothermal MN arrays, iii) compare various fabrication methods using Au NPs powders to elucidate how the fabrication protocol affects final MN morphology and photothermal performance, and iv) study the effect of polymer (water-soluble vs -insoluble) and MN fabrication on intradermal NPs accumulation ex vivo with a porcine skin model.

2. Results and Discussion

2.1. Fabrication of Photothermal MN Arrays

Plasmonic Au/SiO₂ NPs were produced via flame synthesis and utilized for the fabrication of photothermal MNs, as shown in Figure 1. In Figure 1a, the flame-based synthesis of Au/SiO₂ is illustrated in which a liquid solution containing Au and Si organometallic precursors is dispersed and combusted,^[36] resulting in the formation of NPs that can be 1) directly deposited, by thermophoresis, onto a water-cooled substrate (e.g., MN molds), or 2) collected as nanopowder by vacuum-aided filtration. Figure 1b shows the fabrication methods for photothermal MN arrays, termed as deposition, coating, incorporation, and sandwich. For the deposition method, the NPs are directly deposited above the flame on the surface of an MN mold. To remove the deposited NPs outside of the mold cavities, they are wiped away with a wet tissue. Alternatively, a stainless-steel mask can be employed to cover the noncavity surface of the MN mold, although in that case the variability of the alignment of the mask onto the MN mold may affect the reproducibility of the final photothermal performance (Figure S1, Supporting Information). Therefore, we chose to remove the NPs from the noncavity surface of the MN mold by wiping with a wet tissue for the remainder of the study. Subsequently, the water-insoluble polymer polymethylmethacrylate (PMMA) is casted into the mold, dried, and detached to obtain MN arrays with thin photothermal coatings. For the remaining three methods, NPs are collected as a powder and dispersed either in water (coating and sandwich method) or in 10 wt% PMMA (incorporation method). In the coating method, NPs are concentrated as thick coatings on the MNs by centrifuging NPs dispersions into the MN molds, evaporating the solvent, and casting and drying PMMA. The same procedure is followed for the sandwich method, however, before addition of the NPs a thin polymer layer made with 10 wt% PMMA is casted into the MN mold, which results in the NPs being sandwiched within two polymer layers. For the incorporation method, NPs dispersed in 10 wt% PMMA are concentrated in the mold MN cavities, followed by the addition of a pure PMMA support layer.

Figure 1. Schematic illustration of the fabrication of water-insoluble photothermal MN arrays loaded with Au/SiO₂ nanoaggregates.

a) Illustration of aerosol-based synthesis of Au/SiO₂ nanoaggregates by combusting an Au/SiO₂-containing, liquid precursor in a methane/oxygen supported flame. The Au nanoaggregates can be 1) directly deposited as coatings onto MN molds, or 2) collected as a powder. b) Illustrations of the different fabrication methods for Au/SiO₂–MN arrays, referred to depending on their production as, deposition, coating, incorporation, and sandwich.

2.2. Characterization of Plasmonic Au/SiO₂ Nanoaggregates

We chose Au as the metal agent for fabrication of the photothermal MNs because of its high photothermal light-to-heat conversion efficiency, biological stability, and low cytotoxicity.^[37] However, to render spherical Au NPs suitable for biomedical application, their localized surface plasmon resonance frequency needs to be shifted into the NIR region, which we obtained by using SiO₂ as a dielectric spacer.^[36,38] It has been previously shown that tuning the dielectric SiO₂ spacer content in flame-made plasmonic nanoaggregates controls the interparticle distance of individual nanospheres which directly affects their optical extinction by plasmonic coupling.^[36,38–40]

First, we studied the properties of Au nanocoatings as a function of SiO₂ content to fine-tune the photothermal effect in the NIR region. The UV/vis spectra in Figure 2a obtained from deposited Au/SiO₂ nanoaggregates as films on glass coverslips (t_d = 80 s) for various SiO₂ contents (0–50 wt%) show an increased extinction in the NIR range for low SiO₂ contents. This increased extinction in the NIR spectrum is further corroborated by the color of the NP films as shown in the top-view images in the inset of Figure 2a. Figure 2b–e shows b,d) top-view and c,e) side-view scanning electron microscope (SEM) images of the NP coatings produced with 4 or 50 wt% SiO₂, both exhibiting a porous structure. The side-view images indicate higher densities of films with lower SiO₂ content.

Figure 2.

a) UV/vis spectra normalized to 530 nm and with shaded NIR region of Au/SiO₂ nanoaggregate coatings deposited on glass coverslips for 80 s with varying SiO₂. Inset shows digital top-view images of deposited cover slips. SEM images of top-view and side-view images of Au/SiO₂ nanoaggregate coatings deposited on glass coverslips for 80 s with b,c) 4 and d,e) 50 wt% SiO₂ content, respectively. f,g) HAADF images and EDX/EELS elemental mapping images with false coloration for element identification (red: oxygen, green: silicon, blue: Au,) of single Au/SiO₂ nanoaggregate for f) 4 and g) 50 wt% SiO₂ content.

Second, to investigate whether the tuning of the plasmonic coupling can be attributed to varying dielectric spacer content in the Au/SiO₂ nanocoatings produced here, we analyzed the nanoscale morphology of Au/SiO₂ NPs via scanning transmission electron microscopy (STEM) images combined with energy-dispersive X-ray spectroscopy (EDX) and energy-loss spectroscopy (EELS). Figure 2f,g shows the high-angle annular dark field (HAADF) STEM images and corresponding EDX maps generated by fusing together the EELS and EDX datasets.^[41,42] The elemental mapping of NPs was retrieved from nanocoatings produced with low (4 wt%) and high (50 wt%) SiO₂ content. In both cases, the Au NPs are embedded in Si- and O-rich material which encapsulates the nanoaggregate and separates primary Au cores indicating that the SiO₂ acts as a spacer material. Furthermore, by controlling the nominal Si concentration stoichiometrically in the liquid precursor solution, the interparticle distance between the Au nanospheres is tuned with low SiO₂ contents bringing the primary Au nanospheres closer together than high SiO₂ contents. This change in interparticle distance as a function of SiO₂ content was further verified by measuring the distribution of distance between primary Au NPs in transmission electron microscopy (TEM) images (Figures S3 and S4, Supporting Information). Increasing the SiO₂ content from 2 to 50 wt% results in an increase of interparticle distance by almost 10 nm (>80% change), in qualitative agreement with the literature.^[38] However, the interparticle distance should only be used for qualitative comparison since the 2D nature of TEM images does not reflect a true interparticle distance in 3D nanoaggregates.

Finally, to identify the optimal SiO₂ content for maximum NIR photothermal efficiency of the Au nanoaggregates, we irradiated the as-deposited films of various thicknesses and SiO₂ contents with an NIR laser at 808 nm (1 W cm⁻²) and measured their photothermal response with a thermal camera. Increasing the NP deposition duration t_d yields thicker nanofilms with higher extinction and Au mass present which in turn yield higher photothermal responses (Figure S2, Supporting Information). Furthermore, Au/SiO₂ nanoaggregates with 2–6 wt% SiO₂ content at t_Dep = 80 s exhibit the highest photothermal response overall. However, from the UV/vis spectra in Figure 2a, the sample with 2–6 wt% SiO₂ content does not exhibit the highest extinction values at λ = 808 nm, indicating that the extinction of the samples may partially originate from light scattering rather than absorption, since the latter is primarily responsible for the plasmonic photothermal effect.^[43] The low NIR extinction values for the 50 wt% SiO₂ content films results from the largest Au interparticle distance in these nanoaggregates and, thus, lowest plasmonic NIR photothermal response (Figure S2, Supporting Information). The SiO₂ content (wt%), therefore, dictates the Au interparticle distance and plasmonic coupling enabling the photothermal response of these films in the NIR. It should be noted that the primary Au NPs have a similar crystal size for 1–6 wt% SiO₂ contents while the 0 and 0.5 wt% SiO₂ have slightly larger crystal sizes that may be responsible for their photothermal response (Figure S2e, Supporting Information). Furthermore, 1–6 wt% SiO₂ films have a minor decrease of around 8 nm in average d_TEM (Figure S5, Supporting Information) indicating that the spectral differences may be primarily attributed to the spatial orientation of the individual Au NPs within the plasmonic nanoaggregates. However, the extinction of spectra from the films (4 wt% SiO₂, Figure S6, Supporting Information) shows a red-shifted peak and higher extinction in the NIR region compared to dispersed NPs in solution indicating that the controlled plasmonic coupling may be partially attributed to their aerosol deposition as films. Based on these findings, we selected the 4 wt% SiO₂ content Au/SiO₂ nanoaggregates as the photothermal agent for the fabrication of MN arrays.

2.3. Morphology of MN Arrays

To study the morphology of photothermal Au-MN arrays produced with the various fabrication methods, microscopy images were obtained as summarized in Figure 3. Figure 3a shows dark-field stereomicroscopic overview images of the MN arrays (deposition: t_Dep = 90 s, coating, incorporation, sandwich: NPs content = 10 mg g⁻¹) with uniformly distributed needles. MN arrays made by the deposition method were produced containing needle bases (B) of 300 µm (opposed to 200 µm used for the other methods) to increase the NPs content in the cavities.^[44] In all cases, the MNs appear darker than the transparent PMMA support layer, confirming that the plasmonic photo-thermal NPs are present only in the MNs. Representative side-view microscopy images of single needles in the bright and dark-field are presented in Figure 3b,c, respectively, showing the formation of sharp needles for all methods. The presence of a thin Au coating in the deposition method cannot be verified qualitatively through the microscopy images probably due to its low thickness. However, micrographs of the cross-sections of MN molds after deposition of Au NPs show purple coloration typical to Au NPs in the needle cavities, indicating the NP presence there (Figure S7a,b, Supporting Information). After detachment of MN arrays from such deposited molds, the purple coloration cannot be distinguished in the cross-sections of the molds but instead on the produced MN arrays demonstrating the removal of the NPs from the mold along with the MN array (Figure S7c–e, Supporting Information). A darkening in purple color on MN arrays produced with increasing deposition duration suggests the presence of thicker NP coatings (Figure S7e, Supporting Information). For needles of the coating, incorporation, and sandwich method, the presence of the Au NPs can be determined in the bright (Figure 3b) and dark field (Figure 3c) images. The NPs in the coating method are dispersed in an aqueous solution and concentrated in the MN mold cavities resulting in the formation of a superficial NPs coating on the photothermal MNs. Interestingly, Dong et al. followed a similar fabrication method for incorporation of Au nanocages (NCs) in photothermal MN arrays resulting in the concentration of Au NCs only in the needle tips, which may be associated with other centrifugal forces and NPs concentration used.^[14] In the incorporation method, the NPs are first dispersed in the polymer before concentrated in the MN cavities and the images indicate that the NPs are distributed in the polymeric MNs, which is in coherence with similar reports preparing photothermal MN arrays by dispersing NPs in the polymer solution.^[9,45] Finally, the MNs of the sandwich method have a dense NPs middle section with a transparent polymeric needle tip. Therefore, by rational design and selection of the fabrication protocol, the incorporation of plasmonic NPs into MN matrix may be spatially controlled in the MNs.

Figure 3. Images of Au/SiO₂ (4 wt% SiO₂ content) MN arrays produced with four different fabrication methods (deposition, coating, incorporation, sandwich).

a) Stereomicroscopic dark-field overview of the whole MN array for each fabrication method. b) Bright-field and c) dark-field microscopy for single needles of Au/SiO₂ MN arrays. d) SEM image of MNs arrays tilted at 45° with contrast and brightness adjusted for each fabrication method to visualize the full array.

To study the distribution of Au/SiO₂ NPs in the MNs further, we performed SEM and EDX spectroscopy as shown in Figure 3d (see also Figure S8, Supporting Information). SEM images at low and high magnification with adjusted contrast and brightness to visualize NPs (white) and polymeric needles (gray) are shown in Figure 3d (see also Figure S8a, Supporting Information). For the deposition method, distribution of fine Au spots can be distinguished around the MNs, whereas needles produced with the coating method show a thick, patchy Au coating in coherence with the bright-field microscopy images presented above. The NPs in the deposition method are directly deposited from the hot flame onto the mold thus allowing the film formation at nanoaggregate level. The thick coating of the NPs in case of the coating method is probably caused by the centrifugation of highly concentrated (10 mg mL⁻¹) aqueous NPs dispersion into MN molds. The incorporation method results in more evenly distributed Au NPs in the needle compared to the coating method, however, areas of higher relative Au content can be distinguished, too. This may be explained by insufficient stabilization of the NPs in the polymeric solution causing agglomeration of the NPs during centrifugation. Finally, the sandwich method results in polymeric MNs with the lowest Au content of the various methods, indicating a complete outer polymer layer that fully encapsulates the Au NPs in the needle core.

To qualitatively compare the superficial Au content for MNs produced with different methods, side-view SEM images with identical contrast and brightness were obtained and the relative abundance of Au (red), C (dark blue), and O (bright blue) is measured by EDX and overlayed on SEM images (Figure S8b,c, Supporting Information). The elemental mapping from EDX measurements further underpins the above description on the spatial distribution of Au NPs. MN arrays produced via the deposition method results in MNs with finely sputtered Au aggregates while the coating and incorporation methods lead to the presence of thick Au coating on the MN surface. MNs fabricated following the sandwich method show almost no Au presence on the MN surface. Such encapsulation of plasmonic NPs into MN matrix may be beneficial for reducing any unwanted interaction between the nanomaterial and the biological tissue. We continued to evaluate the effect the different incorporation of photothermal NPs in MNs may have on their final photothermal response.

2.4. Photothermal Effect of Au-MN Arrays

The photothermal effect of Au-MN arrays in air was studied by recording their temperature increase by a thermal camera under continuous laser irradiation (λ = 808 nm, 1 W cm⁻²) over time. Figure 4a shows IR thermal images of MNs arrays made by all fabrication methods during laser irradiation for at least 60 s (deposition: t_Dep = 90 s, coating, incorporation, sandwich: NPs content = 10 mg g⁻¹). The highest temperature is reached at the MNs of the array, indicating the presence of plasmonic NPs there. Figure 4b–e shows the temperature increase over time for MN arrays produced with all methods for increasing deposition time t_Dep = 0–120 s (deposition method) or increasing NP content 0–20 mg g⁻¹ (other three methods). MN arrays of all production methods reach similar ΔT_max up to ≈30 °C after lasering for 2.5 min. The MN arrays of the coating and sandwich method reach ΔT_max of ≈30 °C at NPs concentrations of 10 mg_NPs g⁻¹_PMMA while those made by the incorporation method require concentrations of 20 mg_NPs g⁻¹_PMMA. However, such high NP concentrations in the incorporation method resulted in needles with broken tips (data not shown), therefore, 10 mg_NPs g⁻¹_PMMA was selected as the optimal NP concentration for the three fabrication methods. Furthermore, to reach similar ΔT_max for MN arrays made by the deposition, we needed to utilize higher MN bases of 300 µm (opposed to 200 µm used for the other methods) to increase the NPs content in the cavities and thus the photothermal effect. For direct comparison of the photothermal performance of MNs of all fabrication methods, we prepared MN arrays utilizing bases of 300 µm (Figure S9, Supporting Information). MN arrays made by the sandwich method outperform those made by the other three methods in terms of ΔT_max and reproducibility. The MN arrays made with the deposition method reach reproducible values up to t_Dep = 90 s, but for t_Dep = 120 s the coefficient of variation is very high (57%), therefore, we chose t_Dep = 90 s and B = 300 µm for the deposition method. For the other three methods, we chose B = 200 µm for further studies since clinically relevant temperatures can be achieved while the amount of NPs added to the mold during the fabrication is reduced. Finally, MN arrays of all methods exhibit a high stability upon repeated heating (Figure S10, Supporting Information). Such heat increases in air at laser intensities of 1 W cm⁻² are similar^[12,14] or better^[6] to previous reports utilizing Au nanomaterials for production of photothermal MNs arrays for improved antitumor treatments in mice.

Figure 4.

Photothermal heating of Au/SiO₂ (4 wt% SiO₂ content) MN arrays. a) Thermal images of MN arrays irradiated with laser light at 808 nm at 1 W cm⁻² produced with the deposition method for t_Dep = 90 s and B = 300 µm, or the coating, incorporation, or sandwich method utilizing 10 mg g⁻¹ NPs and B = 200 µm. b–e) Average temperature increase around the MN center point over time under laser irradiation for deposition, coating, incorporation, or sandwich method, respectively. Data shown as mean ± SD, n = 3.

We studied the relationship between the content of plasmonic Au nanoaggregates to the photothermal performance of the MN arrays by plotting the ΔT_max as a function of Au content per MN array (analyzed via inductively coupled plasma-mass spectrometry (ICP-MS)) as shown in Figure5a (and Figure S11, Supporting Information). For all methods, we can control the Au content in the MN arrays by changing the deposition time (deposition method) or the NPs concentration added to the MN molds (for the coating, incorporation, and sandwich methods) that in turn affects the final photothermal performance (Figure S11a,b, Supporting Information). For the coating (circles), incorporation (triangles), and sandwich (inverse triangles) methods, the ΔT_max increases for increasing Au content and plateaus for >82 µg Au content, after which a further increase of the Au content does not enhance the photothermal effect. However, the ΔT_max for MN arrays produced with the deposition method (squares) reaches similar high levels as with arrays made by the other methods with at least 16-fold and up to 28-fold lower Au content (Figure 5a and Figure S11c, Supporting Information).

Figure 5.

Maximum photothermal heating as a function of gold content of Au/SiO₂ (4 wt% SiO₂ content) MN arrays. a) The ΔT_max after 2.5 min irradiation at 808 nm at 1 W cm⁻² as a function of Au content per MN array as quantified via ICP-MS fitted using a linear (deposition method) or asymptotic regressions (coating, incorporation, sandwich method). b) ΔT_max per Au mass for all production methods. Data shown as mean ± SD, n = 3.

To further understand the temperature increase for the MN arrays made by all methods, we plotted the ΔT_max per Au mass (°C µg⁻¹ of Au) for each fabrication method as shown in Figure 5b. Producing photothermal MN arrays via the deposition method leads to at least nine times higher photothermal efficiency compared to the other methods. This improved photothermal efficiency for thin Au coatings produced with the deposition method may be due to various effects. First, different properties of nanoaggregates sprayed on the MN mold compared to those collected as powder such as increasing primary NP or agglomerate sizes^[46–48] could explain differences in their photothermal efficiency.^[49–53] However, upon measuring the interparticle distance and primary size of STEM images of NPs collected from MN molds or as powders (Figure S12, Supporting Information), we found no detectable difference in primary particle distance or size. Second, the plasmonic coupling of the Au/SiO₂ (4 wt% SiO₂ content) nanoaggregates may be hampered in MN arrays made by the coated, incorporation, and sandwich methods due to the centrifugation step involved in the fabrication. During the centrifugation, the NPs form large agglomerates/flocs where primary Au NPs of subsets of nanoaggregates may overlap in areas of relatively low SiO₂ spacer content. A decrease of NPs distance may increase the scattering of light due to i) reduced plasmonic coupling and ii) NPs effectively behaving like larger particles.^[54] Such scattering losses in turn decrease photothermal efficiency which largely depends on light absorption.^[55] This is further corroborated by the consistently lower ΔT_max per Au mass for increasing NP concentration in all fabrication methods (Figure 5b). In summary, MN arrays made by the deposition method outperform those from the other methods in terms of photothermal efficiency, however, an increased MN base length of minimum 300 µm is required to obtain clinically relevant heating and for t_Dep > 90 s their reproducibility is reduced.

2.5. In-Skin Hyperthermia

We further explored the potential application of the photothermal MN arrays for hyperthermia treatments by performing ex vivo porcine skin experiments under NIR laser radiation as shown in Figure 6. Figure 6a shows microscopy images of skin cross-sections after MN array application for arrays made by all methods (a–d) and we observe successful insertion resulting in formation of cavities of around 50–150 µm depth. To further analyze the insertion capability of photothermal MNs arrays, we also performed the parafilm insertion method^[56] and the compression test^[57] (Figure S13, Supporting Information). All tested MN arrays pierce the most upper two parafilm layers fully, while for the coating, incorporation, and sandwich methods, needles broke off and remained in the parafilm after removal of the arrays. In the compression test, the MN arrays with a base diameter of 300 µm had the lowest height reduction likely associated with increased mechanical strength due to decreasing aspect ratio.^[58] MN arrays with base diameter of 200 µm were compressed to 37–52% of their initial height with the MN arrays produced via the incorporation method showing highest resistance against compression. This is in coherence with previous reports showing that nanofillers in polymeric MNs may improve strength against compression.^[14,59]

Figure 6.

Skin insertion and hyperthermia of MN array application. Microscopy images of cross-section of porcine skin after Au/SiO₂ (4 wt% SiO₂ content) MN array application made following a) deposition (t_Dep = 90 s), b) coating, c) incorporation, and d) sandwich method (NPs content =10 mg g⁻¹). e) Illustration of experimental set-up to measure hyperthermia by photothermal MN arrays in skin under laser light irradiation at 808 nm at 0.5 W cm⁻². f) Dermal temperature increase over time after irradiation of MN arrays for 10 min. Data shown as mean ± SD, n = 3.

After confirming successful penetration of the MNs into skin, we explored the hyperthermia in the skin under NIR radiation. Figure 6e illustrates the experimental set-up to study hyperthermia after MN application in porcine skin.^[13] MN arrays were placed on pieces of full-thickness (≈1.2 mm) neonatal porcine skin which were secured on a holder and the temperature change at the dermis was recorded with a thermal camera. The laser was applied from the top of the MN array at a laser intensity of 0.5 W cm⁻². The ΔT_max was plotted as a function of time in Figure 6f for 10 min laser irradiation of MN arrays for all methods (deposition: t_Dep = 90 s, coating, incorporation, sandwich: NPs content = 10 mg g⁻¹). The MN arrays heat up the skin by 13–18 °C after 10 min of NIR radiation depending on the fabrication method. Of the four methods, the coating and sandwich methods enable the highest temperature increase in the skin, which is in coherence with their photothermal performance in air in comparison to the other methods for either of the tested MN dimensions (base 200 and 300 µm). Therefore, the MNs arrays here could be employed for MN-assisted intradermal photothermal treatment, for which a temperature increase of 5–20 °C is desired.^[1] Importantly, all MN arrays produced here outperform those of previously reported MN patches utilizing Au nanorods, since similar heat increase in the skin is achieved at four times lower laser intensity.^[13]

2.6. Intradermal Au Deposition

To study whether we can effectively reduce any unwanted intradermal NP deposition from photothermal MNs, we performed ex vivo porcine skin experiments with MN arrays made by all fabrication methods reported here. The majority of publications to date develop photothermal MNs that deliver the photothermal agents into the skin, which might be a concern in terms of nanotoxicity and related side effects.^[8] Here, we describe the fabrication of photothermal MN arrays using a water-insoluble polymer (PMMA) to reduce the intradermal NP deposition. To test whether the use of PMMA will result in reduced intradermal NP deposition compared to MN arrays fabricated with the water-soluble polymer poly(vinyl alcohol) (PVA), we produced MN arrays using both polymers (PMMA and PVA) for all fabrication methods (deposition: t_Dep = 90 s, coating, incorporation, sandwich: NPs content = 10 mg g⁻¹). MNs arrays were applied to skin samples for 10 min and the skin was digested with aqua regia and analyzed for Au content via ICP-MS (Figure 7).^[60] Additionally, MN arrays produced with PMMA were also exposed to laser irradiation (1 W cm⁻²) for 10 min to investigate the effect of heat on the intradermal NPs deposition. Figure 7a shows top-view digital images of skin samples after insertion of photothermal MN arrays made of water-soluble PVA or water-insoluble PMMA. After application of water-soluble MN arrays, the dark colored array of intradermally deposited NPs is clearly visible even though the skin surface was wiped with a tissue, resulting in a “tattoo-like” skin coloration. This higher intradermal NP deposition from water-soluble MNs was also quantified via ICP-MS (Figure 7b) with up to 127-fold reduction for MNs produced with water-insoluble instead of soluble polymer. Thus, the use of PMMA instead of PVA allows for effective removal of the NPs along with the MN array after treatment.

Figure 7. Intradermal deposition of Au nanoparticles for nondissolvable and dissolvable Au/SiO₂ (4 wt% SiO₂ content) MN arrays.

a) Top-view of skin samples after MN array application. b) Total gold content in skin per MN area after application of MN arrays quantified via ICP-MS. Nondissolvable (with and without laser irradiation at 808 nm at 1 W cm⁻²) and dissolvable MN arrays were applied to the skin for 10 min. Data shown as mean ± SD, n = 3. Diamonds indicate data below 0.02 µg mm⁻² and not visible due to the scale in the plot.

Using water-insoluble PMMA, the deposition method overall leads to the lowest accumulation of Au in the skin with an average 93-fold reduction compared to the other methods. The significant reduction of intradermal Au accumulation is most likely caused by the overall much lower Au content per MN array of the deposition method compared to the other methods and suggests a successful removal of the photothermal NPs with the MN array. It should be noted that laser radiation of the MNs arrays does not seem to increase the release of NPs from the MNs. Now, upon plotting the intradermal Au relative to the total amount of Au in MNs arrays (Figure S14, Supporting Information), we observe that the sandwich method releases the lowest fraction of NPs with 13 times less compared to the deposition method. This may be attributed to the particle-free “protective layer” around the needles, effectively encapsulating the NPs in the core of the needles and thus limiting the tissue– NP interface. To summarize, the deposition method allows the lowest release of NPs into the skin tissue while the sandwich method presents a novel protocol to reduce the contact interface of the tissue and NPs which may be easily implemented in standard laboratories. Overall, reducing the tissue–NP interface of photothermal MN arrays may reduce concern about side effects associated with the use of metal NPs thus opening new possibility of NPs-loaded MN therapies in dermatology.

3. Conclusion

In this work, we presented four different fabrications methods to produce water-insoluble, photothermal MN arrays using flame-made Au/SiO₂ nanoaggregates. The NIR photothermal response was controlled by tuning the interparticle distance in the nanoaggregates during their aerosol synthesis. Furthermore, we successfully controlled the incorporation of NPs in the MN arrays via the different fabrication methods which allowed for tuning of the final photothermal response. We employed direct aerosol deposition of Au nanoaggregates on MN molds as a novel fabrication of highly efficient photothermal MNs with ninefold improvement of the photothermal performance per Au mass. Highly efficient photothermal MN arrays made by the deposition method may be specifically attractive for clinical translation because the nanomaterial content and thus the fabrication costs may be significantly reduced. However, this method requires access to the flame aerosol reactor and optimization of the deposition process to avoid inefficient Au precursor consumption. Therefore, the three alternative methods utilizing photothermal NP powders may be also an attractive solution as they can be easily implemented in any standard laboratory. All photothermal MN arrays were able to heat up deep skin layers to therapeutic target temperatures at clinically relevant laser intensity of 0.5 W cm⁻². Finally, we showed for the first time how MN fabrication protocols can be exploited to reduce deposition of photothermal agents in the skin by using water-insoluble polymers and producing an outer protective polymer layer. This work advances the knowledge and understanding of the fabrication of NP-based MN arrays and lays the foundation for successful clinical translation of such photothermal MN arrays.

4. Experimental Section

Synthesis and Characterization of Au/SiO₂ Nanoaggregates and Nanostructured Films

The synthesis of Au/SiO₂ nanoaggregates was performed with flame spray pyrolysis which is described in detail elsewhere.^[61] In brief, a precursor solution was prepared by dissolving gold acetate (99.9%, Alfa Aesar) at 0.1 m under reflux at 110 °C for 1.5 h in 1:1 acetonitrile (99.8%, Sigma-Aldrich) and 2-ethylhexanoic acid (≥99 %, Sigma-Aldrich). Stoichiometric amounts of hexamethyldisiloxane (≥98%, Sigma-Aldrich) were added to achieve the desired SiO₂ wt% in the final Au/SiO₂ product. Using a syringe pump, the precursor solution was fed through a capillary at 10 mL min⁻¹, dispersed by 3 L min⁻¹ of oxygen (>99.5 %, Strandmöllen AB) at a pressure drop of 1.8 bar and ignited using a pilot flame of premixed oxygen/methane (>99.5 %, AGA Gas AB) at flow rates of 3.2 and 1.5 L min⁻¹. The NPs were collected as a powder on a glass fiber filter (Type GF6, Hahnemühle) with the aid of a vacuum pump (Mink MM 1144 BV Busch). For the production of MN arrays via the deposition method, the NPs were deposited on MN molds (B300, H600, Micropoint Technology) attached to a water-cooled (4 °C) holder at 35 cm distance for 20–180 s. For production of nanostructured films on flat surfaces, the NPs combusted at precursor and oxygen flow rate of 5 mL min⁻¹ and 5 L min⁻¹, respectively, were deposited on cover slips (20 × 20 mm, VWR) attached to a water-cooled (16 °C) holder at a height of 25 cm above the burner for 30–80 s.

UV/vis spectral scans at 300–1100 nm were performed for NPs dispersed in water at 0.1 mg mL⁻¹ and for nanostructured films deposited on cover slips. The X-ray diffraction of the nanostructured films was collected using CuKα radiation (Rigaku MiniFlex 600). Copper grids were dipped into NPs dispersions at 0.5 mg mL⁻¹ in EtOH for TEM (Talos 120C G2, 120 kV LaB6 source). HAADF on STEM images and EDX/EELS data fusion was performed as described by Thersleff et al.^[41,42] Elliptical regions of interest around primary particles were drawn on STEM images using ImageJ, and exported for size distribution analysis. Interparticle distances were calculated by estimating the mean distance between a circle and the two closest neighbors based on the circle center points and diameters, TEM images and elliptical regions of interest can be found in Figure S3 in the Supporting Information. Heat response of glass cover slips was recorded using a thermal imaging camera (Testo 871) under laser irradiation at 808 nm from a fiber-coupled diode laser with top-hat diffuser (Laser Century). The laser power after the diffuser was measured using a thermal optical power meter (Thorlabs S425C) and the diffuser height was adjusted to achieve a beam intensity of 1 W cm⁻².

Fabrication and Examination of Photothermal Au/SiO₂ MNs Arrays

Au/SiO₂ MN arrays were produced via mold and casting method using PMMA (MW 120k, Sigma-Aldrich) which was dissolved in ethyl lactate (≥98 %, Sigma-Aldrich) for 1.5 h at 150 °C. For the deposition method, Au/SiO₂ (4 wt% SiO₂ content) were deposited in MN molds and NPs on the noncavity surface were wiped off using an EtOH-soaked tissue. Alternatively, the NPs were deposited only in the cavities of the MN mold by employing a micromachined stainless steel mask (Ascilion AB) with openings matching the cavities of the MN mold. 30 wt% PMMA solution was casted into the mold, centrifuged for 30 min at 3500 rpm, and left to dry overnight. For the sandwich method, a first layer of 10 wt% PMMA was cast into the MNs by centrifuging for 30 min, removing the surplus with a tissue and drying the layer for 1 h at 50 °C and 1 h at 100 °C. For the coated and sandwich method, Au/SiO₂ (4 wt% SiO₂ content) powders were dispersed in water at varying concentration via bath sonication for 10 min. 50 µL of the NPs dispersion was added to the MN molds (B200, H600, Micropoint Technology), centrifuged as above, surplus solution was removed, and the solvent was evaporated from the cavities at 80 °C for 1 h. Then, 50 mg of 10 wt% PMMA was added, centrifuged, and surplus polymer was removed. 30 wt% PMMA was added as the support layer, centrifuged for 5 min, degassed for 5 min, centrifuged again for 30 min, and left to dry overnight. For the incorporation method NPs powder was dispersed in 10 wt% PMMA using a bath sonicator for 10 min. 50 µL was added to the molds, centrifuged for 30 min, and the surplus was removed. 30 wt% PMMA was added as the support layer and processed as detailed above. Finally, MN arrays were imaged using light microscopy and dark field microscopy (Leica MZ10). After sputter-coating the MN arrays with a carbon nanolayer (Quorum Q150T, QuorumTech), SEM images were obtained using the Phenom Pharos SEM (Thermo Fisher Scientific) with integrated EDX detector (15 kV, 176 µA, 192 ns dwell time). Heat response of MN arrays was examined as stated above. In brief, the temperature was recorded using a thermal imaging camera (Testo 871, accuracy ± 2 °C, sensitivity < 0.08 °C) under laser irradiation at 808 nm at 1 W cm⁻². The center point of the MN array in the thermal images was selected and the average temperature for 36 pixels around the center was plotted. High-resolution thermal images were obtained using an IR thermal camera (Ti480 Pro, Fluke) of MNs after irradiation at 1 W cm⁻².

Parafilm Insertion and Mechanical Testing of Au/SiO₂ MN Arrays

MN arrays were inserted at 32 N for 30 s into eight layers of Parafilm M using a Texture Analyser (TA.XT.plusC, Stable Micro Systems) with pretest, test, and post-test speed of 1, 0.5, and 1 mm s⁻¹, respectively.^[1] The layers of parafilm were separated, imaged via dark field microscopy, and the number of holes was quantified. Mechanical compression of MN arrays was performed using the Texture Analyser with similar conditions as stated above, with the difference of pre- and post-test speed at 0.5 mm s⁻¹. MN arrays were imaged before and after compression and reduction of needle height was calculated.

Skin Insertion of MN Arrays

Full-thickness porcine skin was obtained from stillborn piglets from a local breeder, which was except from ethical requirements according to the Swedish Board of Agriculture and stored at −20 °C. MN arrays were inserted into thawed porcine skin with thumb pressure for 2 min. Skin samples were embedded in optimal cutting temperature medium on dry ice and frozen for at least 24 h at −80 °C. The samples were cut into sections (10 µm) via cryostat (ThermoFisher Scientific).

Examination of In-Skin Hyperthermia

Porcine skin was thawed in phosphate-buffered saline (0.1 m, pH 7.4) at room temperature and the hair was shortened using scissors. The skin was mounted onto Styrofoam holder using needles with the dermal side facing the thermal camera to monitor the dermal temperature increase over time. MNs arrays were applied on the epidermis and lasered at 808 nm as stated above at a beam intensity of 0.5 W cm⁻². Coefficient of variation was calculated by dividing the standard deviation with the mean.

ICP-MS Analysis of Au Content in MN Arrays and in Ex Vivo Porcine Skin

For examination of deposition of NPs from MN arrays into skin, 8 mm biopsies of porcine skin were obtained and MN arrays were applied for 10 min (with and without lasering at 1 W cm⁻²). Skin biopsies or MN arrays containing NPs were dissolved in 1 mL aqua regia for at least 48 h. The samples were diluted in 50% aqua regia to a maximum concentration of 100 µg L⁻¹ Au ions. 0.5 mL of diluted samples were diluted with 4.5 mL dH₂O and 20 µL of internal Bi standard at 1.25 mg_Bi L⁻¹ was added prior to analysis by ICP-MS as reported previously (iCAP Q; Thermo Scientific, Waltham, MA, USA).^[60] Standard solution of 0, 0.1, 0.5, 1, 5, 10, 50, and 100 µg_Au L⁻¹ were prepared in 2% aqua regia with 20 µL Bi standard (5 µg_Bi L⁻¹). ¹⁹⁷Au and ²⁰⁹Bi isotopes were detected in triplicate for each sample using flatapole collision/reaction cell with helium as the collision gas (5.1 mL min⁻¹). Argon gas was used for cooling, as the auxiliary and nebulizer gas at 14, 0.79, and 1.0 L min⁻¹, respectively. Recovery of the internal standard varied from 79% to 114% and the Au concentrations were corrected accordingly. The limits of detection and quantification were calculated as 0.007 and 0.01 µg_Au L⁻¹, which was derived from three or tenfold standard deviation of the lowest Au standard concentration, respectively, added to the limit of blank.^[62] Diamonds in the final graphs indicate measured data that are not visible due to the scale in the plot.

Supplementary Material

Supporting Information is available from the Wiley Online Library or from the author.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.