Highly sensitive near-infrared SERS nanoprobes for in vivo imaging using gold-assembled silica nanoparticles with controllable nanogaps

Background To take advantages, such as multiplex capacity, non-photobleaching property, and high sensitivity, of surface-enhanced Raman scattering (SERS)-based in vivo imaging, development of highly enhanced SERS nanoprobes in near-infrared (NIR) region is needed. A well-controlled morphology and biocompatibility are essential features of NIR SERS nanoprobes. Gold (Au)-assembled nanostructures with controllable nanogaps with highly enhanced SERS signals within multiple hotspots could be a breakthrough. Results Au-assembled silica (SiO2) nanoparticles (NPs) (SiO2@Au@Au NPs) as NIR SERS nanoprobes are synthesized using the seed-mediated growth method. SiO2@Au@Au NPs using six different sizes of Au NPs (SiO2@Au@Au50–SiO2@Au@Au500) were prepared by controlling the concentration of Au precursor in the growth step. The nanogaps between Au NPs on the SiO2 surface could be controlled from 4.16 to 0.98 nm by adjusting the concentration of Au precursor (hence increasing Au NP sizes), which resulted in the formation of effective SERS hotspots. SiO2@Au@Au500 NPs with a 0.98-nm gap showed a high SERS enhancement factor of approximately 3.8 × 106 under 785-nm photoexcitation. SiO2@Au@Au500 nanoprobes showed detectable in vivo SERS signals at a concentration of 16 μg/mL in animal tissue specimen at a depth of 7 mm. SiO2@Au@Au500 NPs with 14 different Raman label compounds exhibited distinct SERS signals upon subcutaneous injection into nude mice. Conclusions SiO2@Au@Au NPs showed high potential for in vivo applications as multiplex nanoprobes with high SERS sensitivity in the NIR region. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s12951-022-01327-7.


Background
In vivo imaging is a powerful tool for observing the localized effects of drugs as well as biological phenomena in living tissues or organs. However, conventional imaging methods, such as magnetic resonance imaging or molecular imaging based on fluorophores, usually lack multiplex capability [1][2][3]. Moreover, fluorescence imaging suffers from sensitivity issue owing to the autofluorescence of living animal tissues in the presence of visible light. To overcome these problems, in vivo imaging using near-infrared (NIR) light has attracted considerable attention owing to the good penetration ability of NIR radiation into tissues [4][5][6]. To observe effects of different drugs simultaneously and their multiple tumortargeting abilities using in vivo imaging techniques, multiplexing capacity is one of the virtues for in vivo imaging probes [7]. However, widely used NIR imaging probes, such as fluorescent dyes and upconversion luminescent nanoparticles (NPs), still exhibit issues of spectral overlap, hindering multiplex imaging [8][9][10].
Many researchers have attempted to build Au nanostructures to embed controllable nanogaps or to be effective as SERS nanoprobes, overcoming the limited SERS enhancement performance of Au NPs [36][37][38]. Wang et al. reported seed-mediated growth method for Au nanostars that exhibited strong absorbance in the NIR region and showed the possibility of imaging and treating cancer cells through photothermal therapy (PTT) [39]. Ding et al. reported sea urchin-like, flower-like, meatballlike, and polyhedral Au mesopores of various sizes and shapes [40]. Au NP-assembled nanostructures exhibited strong SERS signals, generating multiple hotspots within the nanogaps between small Au NPs [41]. However, Au NP-assembled nanostructures generated without seedmediated growth method (usually with a long tedious synthesis process) were heterogenous in shape and had uncontrolled nanogaps.
We recently prepared Au-assembled silica (SiO 2 ) NPs by precisely controlling the size of Au NPs on the surface of SiO 2 NPs [42]. Currently, there are no studies on the relationship between nanogaps within Au NPs and their SERS characteristics, which can be a critical feature for NIR SERS imaging. In this study, Au NP-assembled SiO 2 NPs (SiO 2 @Au@Au NPs) with small nanogaps were synthesized to develop NIR-active SERS nanoprobes. Various types of nanogaps as SERS hotspots were generated by controlling the degree of Au NP growth on the surface of SiO 2 NPs. Our SiO 2 @ Au@Au NPs showed single-particle level detection sensitivity under 785-nm NIR laser photoexcitation and were applied for in vitro imaging using HCT 116 cell line. To evaluate them as potential SERS nanoprobes, SiO 2 @Au@Au NPs labeled with 4-fluorobenzenethiol (4-FBT) were used to investigate the signal penetration depth in the porcine tissue and the detectable concentration limit upon subcutaneous injection. SiO 2 @Au@ Au NPs labeled with 14 types of Raman labeling compounds (RLCs) exhibited distinct Raman spectra and unique bands upon subcutaneous injection. The highly enhanced SERS signals and spectroscopic features of SiO 2 @Au@Au NPs indicate that our NIR nanoprobes have the potential for use in multiplex imaging with various RLCs in vivo.

Preparation of SiO 2 @Au@Au NPs
SiO 2 @Au was synthesized using a previously reported method [42]. Au NPs (3 nm) were prepared using the Turkevich method. Briefly, 1.5 mL of 0.2 M NaOH, 12 μL of THPC, and 1.5 mL of HAuCl 4 solution (50 mM) were added to 47.5 mL of DW. The mixture was vigorously stirred for 1 h and stored in a refrigerator for at least two days. In addition, 62 μL of APTS and 40 μL of NH 4 OH were added to 1 mL of SiO 2 NPs (50 mg/mL), and the mixture was stirred overnight at 700 rpm to produce aminated SiO 2 NPs (SiO 2 -NH 2 NPs), which were then washed several times with EtOH by centrifugation; subsequently, 10 mL of Au NPs and 200 μL of SiO 2 -NH 2 NPs (10 mg/mL) were mixed and stirred overnight. SiO 2 @ Au NPs were obtained after washing several times with EtOH by centrifugation, which were dispersed in 2 mL of DW containing 2 mg of PVP.
SiO 2 @Au@Au NPs were prepared according to a method described in our previous study [42] with slight modifications. Briefly, SiO 2 @Au@Au NPs were synthesized using the seed-mediated growth method with SiO 2 @Au NP as a seed and Au precursor. To grow Au into SiO 2 @Au seed, 200 μL of SiO 2 @Au NPs (1 mg/ mL) was dispersed in 9.8 mL DW containing 10 mg of PVP. This suspension was stirred after adding 20 μL of HAuCl 4 (10 mM) and treated with 40 μL of AA (10 mM) every 5 min until the desired concentration of Au 3+ was achieved (50, 100, 200, 300, 400, and 500 μM), which was then washed several times with EtOH by centrifugation to obtain SiO 2 @Au@Au 50 -SiO 2 @Au@Au 500 NPs.

Labeling SiO 2 @Au@Au with Raman compounds
An RLC solution (2 mM) was prepared and added to 1 mL of SiO 2 @Au@Au NPs (1 mg/mL). The mixture was vigorously shaken for 1 h at 25 ℃, and thus obtained RLC-conjugated SiO 2 @Au@Au NPs were washed several times with EtOH by centrifugation. Subsequently, Raman-labeled SiO 2 @Au@Au (SiO 2 @Au@Au-RLC) NPs were redispersed in 1 mL of EtOH.

SERS measurement of SiO 2 @Au@Au-RLC NPs
SiO 2 @Au@Au-RLC suspension (1 mg/mL) was injected into a capillary tube. SERS spectrum of each NP was measured thrice using microscopic Raman system. Measurement was carried out under 532-nm photoexcitation at 1 mW, 660-nm photoexcitation at 1.2 mW, and 785nm photoexcitation at 2.1 mW laser power using a × 10 objective lens with 5-s acquisition time.

Calculation of SERS enhancement factor (EF)
Further Raman spectroscopic studies and bioapplication experiments were carried out using SiO 2 @Au@Au 500 , where the SERS EF of SiO 2 @Au@Au 500 -4-FBT NPs under 785-nm photoexcitation was estimated using the following equation: EF = (I SERS × N normal )/(I normal × N SERS ), where I SERS and I normal indicate the intensity of the Raman band from SERS and normal Raman, respectively, and N normal and N SERS are the numbers of 4-FBT molecules in the pure and assembled forms, respectively, on the surface of SiO 2 @Au@Au 500 -4-FBT NPs. The Raman signal intensity was measured for both pure 4-FBT and singleparticle level using identical × 100 objective lens (0.90 NA, Olympus) under the following conditions: 0.3 mW laser power and 5-s acquisition time. The 4-FBT peak at 1075 cm −1 was used to estimate the EF. I SERS was defined as an average value of the peak intensities of 20 individual particles. The probing volume (18.84 μm 2 ) for the normal Raman measurement was approximated using a cylindrical form with a diameter of 2 μm and height of 6 μm. Assuming that 4-FBT molecules form a monolayer on the surface of NPs, N SERS was calculated based on the surface area of NPs (assuming that SiO 2 @Au@Au 500 -4-FBT has a spherical shape, r = 115 nm) and the molecular footprint of 4-FBT (0.383 nm 2 /molecule) [43].

Cytotoxicity of SiO 2 @Au@Au 500 -4-FBT in HCT 116 cells
HCT 116 cells (human colon cancer cell line) were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS and 1% penicillin/streptomycin at 37 °C in humidified air with 5% CO 2 . Cytotoxicity of NPs was estimated using the crystal violet assay. Cells were seeded in 96-well plates and incubated with different concentrations (0, 1.95, 3.90, 7.81, 15.63, 31.25, and 62.50 mg/mL) of SiO 2 @Au@Au 500 -4-FBT NPs at 37 °C for 24 h. After incubation, the culture medium was removed, and the cells were fixed with 4% paraformaldehyde for 1 h. Then, the cells were washed with DW and air dried. The cells in each well were treated with 100 μL of 0.5% crystal violet solution. After 10 min, the solution was removed and the plates were washed with DW and air dried. Subsequently, the cells were lysed with 1% SDS, and absorbance was measured using VICTOR X3 multilabel plate reader (PerkinElmer, Waltham, MA, USA) at 570 nm.

SERS imaging of HCT 116 cells
Cells were seeded in a 60-mm dish and incubated with 50 μg/mL SiO 2 @Au@Au 500 -4-FBT NPs at 37 °C for 24 h. After incubation, the culture medium was removed, and the cells were washed thrice with 1 × PBS. Cells were then fixed with 4% paraformaldehyde for 1 h, washed with PBS, and dried at room temperature. Then, the SERS mapping images were obtained by point-by-point mapping (step size: 1 μm) using a × 100 objective lens with a 785-nm excitation source, 0.3-mW laser power, and 1-s acquisition time.

Depth profile evaluation of SiO 2 @Au@Au SERS signal
To evaluate the depth profile of SiO 2 @Au@Au SERS signal, NPs were injected into the porcine tissue, and the Raman spectra were measured. First, 15 μL of SiO 2 @ Au@Au 500 -4-FBT (1 mg/mL) was dispersed in DW and injected into the porcine tissue with a 26-gauge syringe at different depths (1, 3, 5, 7, and 9 mm). SERS signals of NPs inside the tissue were measured immediately after injection using a × 10 objective lens with a 785-nm excitation source, 2.1-mW laser power, and 10-s acquisition time.

In vivo multiplexing SERS imaging
To conduct multiplexing SERS imaging in nude mice, 14 different types of RLCs

Results and discussion
Characterization and SERS properties of SiO 2 @Au@Au NPs SiO 2 @Au@Au NPs were prepared using the method described in our previous study, with modifications [42]. Briefly, SiO 2 @Au@Au was prepared by introducing Au NPs into SiO 2 to facilitate the growth of Au (Fig. 1). SiO 2 NPs were prepared according to the Stöber method (Additional file 1: Fig. S1). Subsequently, Au NPs were introduced into SiO 2 following treatment with APTS. SiO 2 @Au was used as a seed in the seed-mediated growth method (Additional file 1: Fig. S2). SiO 2 @Au seed (195.30 ± 13.16 nm) contained several Au NPs of very small size (3 nm) attached to the SiO 2 NP surface. It is imperative to control the size of Au NPs on the SiO 2 core to achieve a gap-enhanced SERS efficacy to create a strong local field between the Au NP gaps. In this regard, SiO 2 @Au@Au NPs were fabricated using SiO 2 @Au NP as a seed with varying concentrations of Au precursor (50, 100, 200, 300, 400, and 500 μM). After application of the growth method, the six prepared SiO 2 @Au@ Au NPs were 212.80 ± 7.35, 213.54 ± 7.14, 215.81 ± 8.30, 219.56 ± 9.36, 229.47 ± 9.85, and 229.48 ± 7.27 nm in size, corresponding to Au precursor concentrations of 50, 100, 200, 300, 400, and 500 μM, respectively (Fig. 2). With addition of higher concentration of Au precursor, the overall size of the NPs increased owing to the growth of Au NPs. The maximum concentration of Au precursor was 500 μM to prevent formation of merged structures (loss of particle morphology feature of Au NP) and seedings exclude Si NPs (Additional file 1: Fig. S3). The seed-mediated growth method allowed dense packing of Au NPs on the SiO 2 core surface, in contrast to the direct attachment of large-sized Au NPs on the SiO 2 core (Additional file 1: Fig. S4). Figure 3a shows the absorbance of each prepared SiO 2 @Au@Au NP. The absorbance increased at all wavelengths with increase in Au precursor concentration, particularly in the NIR region. In addition, the maximum absorption wavelength (λ max ) showed a red-shift with an increase in the concentration of Au precursor. This phenomenon is attributed to strong plasmonic coupling caused by the growth of Au NPs on the SiO 2 NP surfaces. As the absorbance changed, the color of the NPs dispersed in the solvent (EtOH) changed from light pink to dark blue (Fig. 3b). To investigate the SERS characteristics of SiO 2 @ Au@Au, SERS spectra of the six SiO 2 @Au@Au NPs after treatment with 4-FBT were measured using three different laser lines (532, 660, and 785 nm) ( Fig. 3c; Additional file 1: Fig. S4). Raman signals were not detectable for any of the six SiO 2 @Au@Au-4-FBT NPs at 532 nm (Additional file 1: Fig. S4a). This was due to the relatively weak plasmonic resonances of the six SiO 2 @Au@Au NPs irradiated with light at a wavelength of 532 nm. SERS spectra obtained using a 660-nm laser revealed distinct bands for SiO 2 @Au@ Au NPs treated with 200, 300, 400, and 500 μM Au precursor (Additional file 1: Fig. S4b). SERS signals measured using a 785-nm laser were stronger than those obtained using a 532-nm laser and 660-nm laser except the SiO 2 @Au@Au NPs treated with 50 μM Au precursor, for which no signal was detected ( Fig. 3c; Additional file 1: Fig. S5). A comparison between 4-FBT SERS signal of SiO 2 @Au@Au 500 at 1075 cm −1 peak showed that the Raman intensity with 785-nm photoexcitation was 7.7 times higher than that measured using 660-nm laser (Fig. 3d).
SERS spectra of SiO 2 @Au@Au NPs captured using 660-nm and 785-nm lasers showed enhanced Raman signals with an increase in the number of Au NPs on the SiO 2 surface. This could be attributed not only to the stronger absorbance but also to the narrower nanogap between Au NPs, leading to a highly amplified SERS signal. Transmission electron microscopy (TEM) images (Fig. 2) show that the nanogaps between Au NPs on the SiO 2 core gradually decreased as the concentration of Au NPs increased. The nanogap sizes were measured to be 4.16 ± 1.04, 3.76 ± 1.09, 3.68 ± 1.29, 1.98 ± 0.50, 1.17 ± 0.32, and 0.98 ± 0.19 nm for NPs treated with 50, 100, 200, 300, 400, and 500 μM Au precursor, respectively (Fig. 3e). The strongest Raman signal of SiO 2 @Au@Au NPs with 500 μM Au precursor during Au seed growth might be because the electromagnetic field was concentrated in 1-nm nanogaps. The seed-mediated growth method for SiO 2 @Au@Au NPs was validated as a powerful strategy to precisely control the nanogap size and maximize the SERS enhancement.
Considering that SiO 2 @Au@Au 500 NPs have a higher absorbance in the NIR region and show the strongest Raman enhancement ability among all prepared NPs, the following experiments were performed using SiO 2 @Au@Au 500 . Using 4-FBT molecule as an RLC, the SERS spectra of 20 single particles of SiO 2 @Au@Au 500 -4-FBT were measured, and the average EF value was estimated to be 3.8 × 10 6 with good uniformity (3.45% relative standard deviation on log scale) (Fig. 3f ). Compared to other noble metal-based NPs, the assembled structure had a lower EF value owing to the large surface area for RLC binding. However, the higher intensity and signal uniformity of each nanocomposite could be an advantageous feature of the assembled structures [44]. SiO 2 @Au@Au 500 exhibited higher EF values than those reported for other noble metal-assembled NPs (Table 1). Although the EF value is smaller than that of bumpy silver nanoshells, Au-assembled SiO 2 NPs are more stable than Ag-based NPs under biological conditions [45].

SERS imaging of HCT 116 cancer cell with SiO 2 @Au@ Au 500 -4-FBT
Before using SiO 2 @Au@Au for in vitro applications, a cytotoxicity test was conducted using HCT 116 cell line. SiO 2 @Au@Au 500 -4-FBT NPs were prepared at a concentration of 62.5 μg/mL (26.38 × 10 8 particles/ mL) and serially diluted for the cytotoxicity test (Additional file 1: Fig. S6). Cell viability was more than 90% at all concentrations of SiO 2 @Au@Au NPs within 24 h. Moreover, biocompatibility of SiO 2 @Au@Au 500 -4-FBT NPs at a concentration of 62.5 μg/mL or lower was confirmed.
To obtain images of HCT 116 cancer cells through SERS, the cells were incubated with SiO 2 @Au@Au 500 -4-FBT NPs for 24 h. NPs were either attached to the cell surface or entered the cell, whereas the remaining NPs were washed out. Additional file 1: Figure S7a shows the SERS mapping image at 1075 cm −1 . The overlay image of HCT 116 cells and the adsorbed NPs showed that SiO 2 @Au@Au 500 -4-FBT NPs were attached to the edge of the cell. We compared the Raman intensity at different locations on the cell and observed no Raman signal outside the cell (i), weak Raman signal at the cell surface (ii), and extremely strong Raman signal inside the cell (iii) (Additional file 1: Fig. S7b). This observation confirmed the possibility of SERS imaging of cancer cells using SiO 2 @Au@Au NPs. Bumpy silver nanoshell 2.2 × 10 7 [45] Gold (Au) Au/Ag hollow shellassembled silica nanosphere 2.8 × 10 5 [48] Au-assembled silica NP 3.8 × 10 6 Current study

Sensitivity of SERS signal of SiO 2 @Au@Au 500 -4-FBT
To investigate the SERS signal depth profile of SiO 2 @ Au@Au, we injected SiO 2 @Au@Au 500 -4-FBT NPs into the porcine tissue at different depths (1, 3, 5, 7, and 9 mm) and measured the SERS spectra (Fig. 4a). As the depth increased, the Raman intensity decreased (Fig. 4b). However, a measurable signal was detected up to a depth of 7 mm. For an accurate analysis, the Raman band intensities at 382, 620, and 1075 cm −1 were normalized to the signal intensity at a depth of 1 mm (Fig. 4c). The Raman intensity decreased as NPs were injected deeper inside the porcine tissue, and the Raman spectra distinct from that of the tissue without NP injection was observed until a depth of 7 mm. We conclude that the SiO 2 @Au@Au NPs generated a detectable SERS signal until the maximum depth of 7 mm in animal tissues. Thus, SERS detection using SiO 2 @Au@Au NPs was attempted through subcutaneous injection into animals.
For in vivo imaging, it is crucial to use small amounts of NPs to avoid side effects such as blood clots [46]. To determine the detectable concentration limit, various concentrations of SiO 2 @Au@Au 500 -4-FBT from 1000 to 4 μg/mL were subcutaneously injected into nude mice, and the SERS spectra were measured using a 785-nm laser (Fig. 5a, b). The Raman intensity decreased as the concentration of SiO 2 @Au@Au 500 -4-FBT NPs decreased; however, a sufficient signal was observed at a concentration of 16 μg/mL (Fig. 5c). To compare these results, the Raman bands intensities at 382, 620, and 1075 cm −1 were normalized to the Raman signal at 1000 μg/mL (Fig. 5d). The strong SERS signal of SiO 2 @Au@Au 500 -4-FBT allowed for the subcutaneous detection of particles even at a very low concentration (16 μg/mL), showing sufficient signal sensitivity. Fig. 4 a Schematic illustration for the depth profile evaluation of SiO 2 @Au@Au 500 -4-FBT using the porcine tissue. b Raman spectra of SiO 2 @Au@ Au 500 -4-FBT injected into the porcine tissue at different depths (1, 3, 5, 7, and 9 mm). c Correlation between normalized SERS intensities at 382, 620, and 1075 cm −1 for Raman spectra in b and the injection depth from the surface of the porcine tissue. The Raman intensity decreased as the injection depth of SiO 2 @Au@Au 500 -4-FBT increased, and was detectable up to the injection depth of 7 mm

In vivo multiplex imaging potential
To investigate the multiplex imaging potential of SiO 2 @ Au@Au, 14 different RLC-treated NPs (SiO 2 @Au@Au-RLC) were prepared and injected subcutaneously into nude mice (Fig. 6a). The Raman spectra from each location were measured using a 785-nm laser. Distinct Raman spectra were obtained for the 14 types of NPs (Fig. 6b), which showed unique bands for code (label) identification To the best of our knowledge, the present study used the highest number of labels for NIR-active nanoprobes; the previously reported maximum number of RLCs for multiplex imaging based on SERS was 10 [3]. Thus, our SiO 2 @Au@Au NPs with multiple hotspots and narrow nanogaps exhibited high stability, allowing attachment of 14 different RLCs.

Conclusions
SiO 2 @Au@Au NPs were prepared using the seed-mediated growth method; six SiO 2 @Au@Au NPs of different sizes were fabricated on the surface of SiO 2 NPs by controlling the concentration of Au precursor (50, 100, 200, 300, 400, and 500 μM). With increase in concentration of Au precursor, SiO 2 @Au@Au showed stronger absorbance, particularly in the NIR region. In addition, multiple hotspots and narrow nanogaps of approximately 1 nm were obtained by increasing the concentration of Au precursor during the growth process, enabling single particle-level detection. The SERS measurement revealed the Raman signal of high intensity after 785-nm laser photoexcitation. SiO 2 @Au@ Au NPs obtained using 500 μM Au precursor exhibited an average SERS EF value of 3.83 × 10 6 . SiO 2 @Au@Au NPs were successfully applied for the SERS imaging of HCT 116 cancer cells. In addition, owing to the advantages of NIR radiation and detection, the SERS signal could be measured even at a depth of 7 mm in the porcine tissue. The detectable concentration limit of NPs for subcutaneous injection was 16 μg/mL. Moreover, the multiplexing capability of the prepared SiO 2 @Au@Au was investigated by subcutaneously injecting 14 different SiO 2 @Au@Au-RLC NPs into nude mice. In this study, we fabricated highly sensitive NIR SERS nanoprobes with very strong SERS signals owing to their structure with uniformly synthesized multiple hotspots and narrow nanogaps. Along with the advantageous features of absorbing long-wavelength light and highly enhanced Raman signals, our SiO 2 @Au@Au structure can potentially be used for multiplex molecular imaging and in vivo applications.
Additional file 1: Fig S1. TEM image of SiO 2 NPs. Fig S2. TEM image of SiO 2 @Au used as a seed. Fig S3. TEM image of SiO 2 @Au@Au synthesized using 600 μM gold(III) chloride hydrate. Fig S4. TEM image of SiO 2 @Au synthesized by directly attaching large-sized Au NPs (10-15 nm) to aminated silica (not a growth method).  Fig. 6 a Photograph of mouse injected with 14 different SiO 2 @Au@Au-RLC with 14 different Raman labeling compounds (RLCs). b Comparison of the 14 normalized Raman spectra of SiO 2 @Au@ Au-RLC injected into nude mouse with spectra from location without NP injection at 785-nm photoexcitation light, 2.1-mW laser power, and 10-s acquisition time. Each showed distinct Raman spectra with unique bands for label identification