SERS Quantification of Galunisertib Delivery in Colorectal Cancer Cells by Plasmonic-Assisted Diatomite Nanoparticles
Stefano Managò, Chiara Tramontano, Donatella Delle Cave, Giovanna Chianese, Gianluigi Zito, Luca De Stefano, Monica Terracciano, Enza Lonardo, Anna Chiara De Luca, and Ilaria Rea
1. Introduction
According to GLOBOCAN 2020 data, colorectal cancer (CRC) is the fourth most diagnosed cancer in the world. Despite the advances in diagnosis and surgical procedures, 20% of CRC patients manifest metastasis at the time of diagnosis, which of course affects the patient survival rate.[1] Most of the conventionaltherapies for CRC treatment, like sys- temic chemotherapy and immunotherapy, are expensive and invasive since they fail to target cancer cells selectively, causing severe toxicity on normal tissues besides side effects.[2–5] The epithelial-to-mesen- chymal transition (EMT) is a dynamic multistep process involved in several physiological and pathological conditions, including cancer.[6] Notably, EMT in CRC is associated with an invasive, metastatic, and chemoresistant phenotype.[7] In con- trast, the mesenchymal-to-epithelial tran- sition (MET), is the reverse program of EMT in metastases and is characterized by the upregulation of epithelial adhe- sive proteins such as E-CADHERIN, and downregulation of mesenchymal proteins, such as SNAIL-1 and TWIST-1. Slowing EMT or even reversing the process in metastases inducing a MET program not only is a powerful therapeutic approach alone, but can also assist in increasing the efficacy of other anticancer treat- ments. Among the molecular pathways promoting tumorigenesis in CRC, thetransforming growth factor-β (TGF-β) is described as a crucial factor in EMT induction.[8] Fighting the CRC with the TGF-β receptor I-specific inhibitor Galunisertib (LY2157299, LY) has been proposed as a powerful strategy to reduce tumor growth and the risk of relapse.[8] However, LY is cleared predominantly by cytochrome P450 3A4 (CYP3A4) oxidative metabolism in theliver, which leads to the formation of diverse toxic metabolites circulating in the plasma.[9] Thus, increasing the efficacy of LY delivery to CRC cells, promoting good results with reduced dosing, is the first crucial step to tackle side effects.
In this context, nanotechnology can play a key role. Indeed, biocompatible nanoparticles (NPs) able to carry therapeutic molecules inside cancer cells represent an excellent strategy to increase drug efficacy. In particular, the surface functionalization of NPs can enhance therapy selectivity and site-specific delivery of the drug. Targeted therapy can limit the mutual interactions between the drug and normal tissues, thus reducing systemic toxicity. In addition, NPs may overcome poor oral bioavail- ability of drugs and their instability in circulation. Several types of NPs, such as organic, inorganic, and biological NPs, as well as their combinations, have been explored in biomedicine.[10–19] We focused our attention on inorganic diatomite NPs (DNPs) obtained from diatoms and made of a natural biosilica. Indeed, DNPs have been explored as drug delivery systems thanks to their advantageous properties, including high surface-to-volume ratio, long half-life, and chemical stability in physiological con- ditions.[20–22] In addition, diatom’s amorphous silica is approved by the Food and Drug Administration (FDA), Generally Recog- nized as Safe for food and pharmaceutical production (GRAS, 21 CFR 182.90), and classified in the 3rd group of “Not classifiable as to its carcinogenicity to humans” by the International Agency for Research on Cancer (IARC).[23] DNPs can be easily preparedin the size range of 100–400 nm and possess nanopores able to load a wide range of therapeutic compounds, including aptamers, chemotherapeutic agents, and small molecules.[23] The easy-to-tune surface of DNPs offer the opportunity to per- form several surface functionalization strategies to enhance their biocompatibility, intracellular uptake, and improve drug loading and delivery efficiency.[24–27] In a previous study, we have already demonstrated the cell internalization of small-interfering RNA (siRNA)-modified DNPs via endocytic uptake. The DNPs penetrate cells within 18 h without causing toxicity and they are embedded in a lipid environment in the cytoplasmic region of cancer cells. Therefore, we consider DNPs as excellent candi- dates to promote enhanced LY delivery in CRC cells.[28]
To this end, in this work, we moved a crucial step forward devising a hybrid nanosystem constituted by a DNP decorated with AuNPs and capped with a gelatin (Gel) layer (Scheme 1) for LY delivery in CRC cells and concurrent quantification of LY release at the single-cell level with attogram-scale resolution via SERS imaging. The biocompatibility of the DNP-AuNPs com- plex has been preliminary demonstrated in HeLa cells during previous investigations.[29] The outer shell of gelatin is intro- duced to induce the sustained release of LY via degradation of the polymer chains. Gelatin-functionalized nanocarriers have already been employed for the treatment of CRC, where matrix metalloproteinase-2 (MMP-2) is overexpressed and can degrade the gelatin shell.[30] The realized nanosystem, denoted asDNP-AuNPs-LY@Gel, is a multifunctional platform serving as both a pH-triggered drug delivery system and SERS-nanosensor for intracellular release of LY. The release of LY from the nano- complex is both time and pH-dependent. At physiological pH (7.4) the gelatin layer is folded tightly and the drug is retained in the nanosystem. In contrast, due to the sufficient extension of gelatin chains and degradation of some gelatin molecules in the acidic environment, the release of LY is promoted at pH 5.5. This pH-dependent drug delivery behavior is useful to trigger the release of LY in CRC, where the accumulation of lactic acid in the close external microenvironment and the overexpression of MMP-2 make the physiological pH acidic.[31] The monitoring of the nanosystem internalization and LY release quantifica- tion is made possible by the presence of suitable AuNPs on the DNP providing a highly sensitive SERS readout of LY with non- invasive laser illumination exciting their plasmonic resonance. Indeed, the intracellular release of LY to CRC is measured by SERS as a function of the incubation time up to 48 h and cor- relatively quantified to provide the LY sensogram evolution with an unprecedented resolution of 7.5 1018 g. The quantifica- tion of the intracellular drug release from NP-based systems is essential for correlating the amount of internalized drug with the therapeutic effect induced in the cell. At present, many drug release monitoring systems are limited to qualitative determi- nation or they lack precise quantification due to imaging sen- sitivity and resolution. The urgency for monitoring the drug release in real-time can be satisfied by loading fluorescence payloads in NPs and quantifying the intracellular release by flu- orescence measurements.[32] This work describes an alternative, accurate, and label-free method based on SERS analysis for the intracellular quantification of LY release. We also demonstrate that the release of LY from the DNP-AuNPs-LY@Gel complex inside the CRC cells inhibits their proliferation and induces the reversion to a normal phenotype by MET with greater efficiency compared to the free drug.
2. Results and Discussion
2.1. DNP-AuNPs-LY@Gel Nanosystem Preparation and Characterization
The DNP-AuNPs-LY@Gel nanosystem was prepared following the fabrication procedure reported in Scheme 1. Amino- silanized DNP (reported in the text as “DNP”) resulted in a mean size of 400 50 nm and a surface ζ-potential of 20 5 mV; the positive surface charge was related to the presence of amino groups on the surface of the DNP. At this stage, the colloidal suspension of DNPs in de-ionized water appeared colorless. To decorate the surface of the DNP with metallic NPs, DNPs were dispersed in a chloroauric acid solution. The amine groups on the surface of the DNP allowed the electrostatic interac- tion between DNP-NH3 and AuCl4 gold precursor ions. The NaBH4 solution was used to reduce the gold precursor ions in gold NPs (AuNPs). The resulted hybrid system DNP-AuNPs had a mean size comparable to which of the bare DNP within the error (400 50 nm). Its ζ-potential decreased to 15 10 mV due to the presence of the carboxyl groups of dicarboxylic PEG embedding the AuNPs and used as a stabilizing agent in theAuNPs synthesis. [29,33] The presence of the AuNPs on the surface of the DNP was confirmed by the color change of the suspension that turned from transparent to light purple. The porosity structure of DNP-AuNPs and their surface area of about 6 m2 g1 (measured by gas porosimetry, data not shown here) was exploited to realize a drug delivery system with effi- cient drug loading and release capacity.[29] In this work, the small molecule LY used in clinical trials for the treatment of different types of cancer including CRC was loaded in the DNP- AuNPs nanocomplex. The loading of LY in the inner and outer surface of the DNP-AuNPs nanocomplex was performed in an acidic solution to promote the electrostatic interactions between the protonated isoform of the drug and dicarboxylic PEG mole- cules surrounding the DNP-AuNPs complex. Specifically, the elaborate chemical structure of LY can be protonated on various functional groups according to the multiple isoforms generated under different pH conditions. In this study, the acidic drug loading solution promoted the protonation of the quinoline- carboxamide group of LY and its attraction to the negative charges of the surface AuNPs. In addition, the chemisorption interactions (N-Au type) between the drug and AuNPs sup- ported the loading of LY in the complex DNP-AuNPs.[34,35] The nanosystem loaded with LY (DNP-AuNPs-LY) was finally capped with a layer of gelatin crosslinked by carbodiimide chemistry to prevent the burst release of the drug in aqueous solution. The DNPs-AuNPs-LY system was dispersed in an acidic solution of gelatin at pH 3.5 to preserve the electrostatic interactions between LY and the nanosystem. Finally, the gelatin absorbed on the surface of the DNP-AuNPs-LY was crosslinked and the DNP-AuNPs-LY@Gel final complex still purple in color was characterized. The characterizations of DNPs revealed an incre- ment of the mean size to 450 50 nm and a surface ζ-potential of 7 8 nm. The change of the ζ-potential to an almost neutral value (7 8 mV) was ascribed to the crosslinking phenomena occurring on the surface of the gelatin layer that reserved both carboxyl and amino groups, providing the nanosystem with a neutral surface. (Figure 1a,b).
The morphological changes of the nanocomplex were high-lighted by transmission electron microscopy (TEM) analysis (Figure 1c). The DNP 400 nm in size was characterized by an irregular shape and a porous morphology; pores with a size range between 15 and 35 nm were observed in TEM investi- gations. The pore size was calculated using a free-open-source image processing package (Image J). Previous studies dem- onstrated that the irregular shape of the DNP did not induce cellular toxicity. [20,28] The in situ synthesis of AuNPs promoted the growth of a dense coverage of AuNPs on the surface of the DNP. According to TEM analysis, the mean size of AuNPs on the biosilica surface was about 25 nm. The DNP-AuNPs-LY@ Gel nanosystem showed a morphology similar to that of the DNP before the drug loading and gelatin capping, revealing that the exposure to the chemical reagents used in the func- tionalization procedure (i.e., acetone, acidic solutions) did not alter the morphology of the nanosystem. The presence of the gelatin layer, visible as a thin and clear ring around the DNPs in TEM investigations, was further confirmed by UV-Vis anal- ysis. Figure 1d reports the absorbance spectra of bare DNPs, DNPs-AuNPs, and DNPs-AuNPs-LY@Gel suspended in de-ion- ized water. The suspension of DNPs did not show absorptionpeaks in the visible range under investigation (black curve in Figure 1d). The Localized Surface Plasmon Resonance (LSPR) of AuNPs on the biosilica surface was observed at 576 nm through absorbance spectroscopy (red curve in Figure 1d). The presence of LY loaded in the nanocomplex could not be detected through changes in the LSPR of the DNPs-AuNPs-LY suspension, due to the small size of LY (data not shown here). Conversely, the gelatin capping in the DNP-AuNPs-LY@Gel complex caused a 10 nm red-shift of the LSPR due to the increase of the effec- tive refractive index in the NP surroundings.[36] Figure 1e shows the FTIR spectra of the DNPs (black curve), DNPs-AuNPs(red curve), and DNPs-AuNPs-LY@Gel (green curve) suspen- sion. FTIR spectra of samples were characterized by two peaks at 1100 and 800 cm1 related to the asymmetric and symmetric stretching modes of siloxane (Si-O-Si), respectively;[37] the fea- tures at 3400 and 1640 cm1 were due to the O-H stretching and O-H bending of water physically absorbed on the silica surface.[38] The presence of the gelatin layer on the surface of the DNPs-AuNPs-LY@Gel dispersion was demonstrated by the peaks at 1530 and 1450 cm1 related to N-H bending (Amine II), and the features at 2910 and 2840 cm1 due to the C-H stretching vibrations.[39]
2.2. Evaluation of the Loading and Release Capacity of DNPs-AuNPs-LY@Gel
For the design of a sustained-release system, gelatin was selected as a capping element due to its advantageous fea- tures, including biodegradability, biocompatibility, and pH- responsive behavior. After the LY loading and gelatin capping, the DNP-AuNPs-LY@Gel nanosystem was immersed in an acidic solution of PBS and gently shaken for 48 h. The amount of drug-loaded in the nanocomplex (i.e., the loading capacity) was determined by analyzing the release solution by Reversed Phase-High Performance Liquid Chromatography (RP-HPLC) after 48 h. The loading capacity of LY in the nanosystem resulted to be 20 4 g mg1 of the DNP-AuNPs-LY@Gel com- plex. To investigate the release behavior of the nanosystem, we performed in vitro release tests under physiological (7.4) and more acidic (5.5) pH conditions, mimicking the tumor micro- environment. The “Warburg effect” is a well-accepted theory according to which cancer cells produce energy from glucose essentially through the glycolytic pathway. This effect leads to the production of huge amounts of lactate that decreases the tumor microenvironment pH value.[40] The DNP-AuNPs-LY@ Gel nanosystem exhibited a pH and time-dependent release behavior of LY in phosphate-buffered saline (PBS) at 37 C. As shown in Figure 2, the cumulative release of LY increased sig- nificantly with decreasing solution pH due to the gelatin matrix embedding the NPs. When the DNP-AuNPs-LY@Gel was dis- persed in PBS solution at pH 7.4, the amount of LY released to the medium was less than 10% of the total LY loaded in the nanocomplex. This phenomenon can be explained consid- ering that the attractive forces between the opposite charges of each gelatin molecule promoted the formation of intermo- lecular hydrogen bonding between the gelatin molecules at physiological pH value.[41] In the PBS solution 7.4, the gelatin matrix was folded tightly and almost all the drug was retained in the nanosystem within 48 h of incubation in the release solution. Since the electrostatic attraction between the gelatin molecules decreases in a more acidic environment, when the DNP-AuNPs-LY@Gel complex was dispersed in an acidic solu- tion the relaxation of the gelatin chains promoted the release of LY to the medium in a time-dependent manner (sigmoidal release profile). The release of LY from DNP-AuNPs-LY@Gel started after 10 h and gradually increased due to the extension of the gelatin chains that were disrupted after 48 h of incuba- tion in the acidic solution. When the gelatin matrix was com- pletely dissolved, 100% of the LY was released into the medium. To highlight the advantages of our system, we also carried out in vitro release tests of the DNP-AuNPs-LY complex in both acidic and physiological microenvironments (Figure S1, Sup- porting Information). The uncapped nanosystem exhibited a typical burst release profile, completely uncontrolled and without any sensitivity to pH value of the release solution. Indeed, the release profiles of the DNP-AuNPs-LY complex in PBS solutions 5.5 and 7.4 were comparable within the errors. Therefore, the gelatin coating in the DNP-AuNPs-LY@Gel com- plex not only delayed the release of LY up to 48 h, but it also prevented the uncontrolled burst release typically attributed to porous NP-based systems.[42] Moreover, the choice of the gelatin for the surface coating was based on the pathophysiologicalproperties of cancer and offered the opportunity of achieving stimuli-responsive drug release at tumor sites. In the cell, the upregulation of proteolytic enzymes degrading the gelatin chains, such as the MMP-2, can trigger the release of LY from the nanosystem, making the DNP-AuNPs-LY@Gel a stimuli- responsive drug delivery system.[30]
2.3. SERS Monitoring of Delivered LY in Living CRC Cells
The major shortcoming of existing diatomite-based nanovec- tors is their limited detection sensitivity that hinders the label- free monitoring and quantification of intracellular drug release. To overcome this limitation, hybrid systems made of diatoms and metal nanostructures have been proposed as SERS sub- strates.[43] The near-field optical amplification of metal nano- structures on diatoms increases their detection performance, allowing label-free sensing of biomolecules with excellent spec- ificity and sensitivity.[44,45] The combination of the drug-loading capacity of DNPs with the strong Raman enhancement of mole- cules close to AuNPs is an ideal strategy to combine therapeutic purposes with label-free intracellular drug monitoring. Indeed, the hybrid nanosystem can integrate multiple functionalities allowing bio-imaging and drug delivery goals simultaneously without the use of any fluorophore or external marker, avoiding fluorescence-quenching issues.
In this study, the incorporation of LY into the DNP-AuNPs system and the enhancement of the LY signal were prelimi- narily evaluated by carefully studying the Raman and SERS spectra. Figure 3a compares the SERS signals of the complex DNP-AuNPs-LY@Gel (blue line), DNP-AuNPs@Gel (red line), and the substrate DNP-AuNPs alone (green line). Each spec- trum was the average of 30 acquisitions. For reference, the Raman spectrum of the LY was shown (black line). The LY Raman spectrum was recorded on a drop of LY (5 mg mL1 inacetone) deposited on the coverslip and left to dry to allow the total evaporation of the solvent. The Raman spectra of LY and acetone were reported in Figure S2 (Supporting Information), showing that there was no overlap of the characteristic Raman bands of LY and acetone. The LY Raman spectrum exhibited strong bands in the spectral region between 1300–1600 cm1 ascribed to the main molecular vibrational bonds of the pyri- dine (basically CH, CN, CC stretching and bending modes). A detailed LY Raman band assignment was reported in Table S1 (Supporting Information). The SERS fingerprint of the LY-markedly similar to its Raman counterpart- was observed from the complex DNP-AuNPs-LY@Gel and the tentative assign- ment of the vibrational modes was reported in Table S1 (Sup- porting Information). Control experiments were carried out on the DNP-AuNPs@Gel and the substrate DNP-AuNPs alone. Notably, the DNP-AuNPs@Gel and DNP-AuNPs SERS signal did not show any significant spectral feature in the regions between 1300–1650 cm1, excluding major contributions to theSERS signal of the substrate itself and other organic compo- nents of the nanovector, such as the dicarboxylic PEG or gelatin capping. Our measurements confirmed that the LY molecules strongly interacted with the metallic NPs and hotspots. There- fore, the most intense SERS vibration, ascribed to a combined action of ring C–N stretching and ring bending, was found at 1360 cm1.[46–49] The intensity of this band was used for moni- toring the LY intracellular release from the developed platform. For SERS experiment, a drop (3 L) of the DNPs-AuNPs- LY@Gel dispersion (500 g mL1); LY loaded was 20 4 g per mg of DNPs-AuNPs-LY@Gel corresponding to 25 106 M of LY) was deposited on a CaF2 slide. The SERS spectrum of the complex DNP-AuNPs-LY@Gel revealed an enhancement (SERS Gain, G) of the LY signal of about 4.5 105, further confirmed by the analysis of the Enhancement Factor (EF) of the proposed metallic substrate reported in the Supporting Information. G was calculated as the ratio between the SERS and Raman signal of the LY bands at 1360 cm1, normalized to the power(1 mW vs 20 mW), integration time (1 s), and concentration (25 106 M vs 15 103 M) of the drug. G provided quantitative information on the signal gain expected from the DNP-AuNPs- LY@Gel complex with compared to the pure LY Raman meas- urement, assuming that all the experimental parameters, such as the objective, laser wavelength, and spectrometer were the same. This parameter had the advantage of being free from the overestimation error performed when calculating the probed molecules required for the evaluation of the EF reported in the Supporting Information. [50]
The SERS spectra of LY loaded in DNP-AuNPs-LY@Gel com- plex acquired within 30 s with an integration time of 1 s were reported in Figure S3a (Supporting Information). The spectra were well reproducible and the intensities on a selected nano- complex could vary up to about 18% (the standard deviations for the band at 1360, 1500, 1545, and 1580 cm1 were respec-tively 4%, 11%, 18%, and 8%). The SERS intensity for the band at 1360 cm1 was about 1500 counts s1 at the considered experi- mental parameters. A standard deviation of about 10% was reg- istered for signals acquired on different clusters of the DNPs, showing a good inter-sample reproducibility (see Figure S3b, Supporting Information).
Before investigating the release properties of the DNP- AuNPs-LY@Gel nanovector, we evaluated its internalization in the human LS-174T CRC cell line by Raman imaging. Figure 3b shows the Raman map of the LS-174T cell incubated with a dis- persion of 50 g mL1 of DNPs-AuNPs-LY@Gel for 18 and 24 h and the control sample at 0 h. Different colors are associated with the loadings/spectra (see Figure S4, Supporting Informa- tion) used to reconstruct the Raman map and are linked with a specific cell location (i.e., nucleus, cytoplasm) or with the nano- complex. By using the MCR approach, it was possible to recon- struct a false color Raman map of the cell showing that the nanovector was internalized and localized in few clusters dis- tributed through the cell cytoplasm after 18 h of incubation.[28]
We next investigated the LY release from the DNP-AuNPs- LY@Gel nanosystem in two different pH environments, 5.5 and 7.4, by SERS analysis. To this aim, a drop (3 L) of DNP- AuNPs-LY@Gel dispersion (500 g mL1) was deposited on a CaF2 slide, with a concentration of 250 ng mm2 on a circular area of 20 mm2. The total mass of DNPs was 1.5 g corre- sponding to an amount of loaded LY of 30 ng, with an LY mass for a single DNP of 0.25 fg (see section Quantification of the LY release). The SERS investigations under different pH condi- tions were performed in parallel (Figure 3c). In physiological conditions (PBS, pH 7.4), a very small amount of LY (about 4%) was released by the DNP-AuNPs-LY@Gel complex within 24 h. A slow release was observed after 30 h and about 60% of LY was delivered from DNP-AuNPs-LY@Gel after 48 h. At pH 5.5, mimicking the cancer cell microenvironment, the release of LY from the nanovector became faster and a great amount of LY (about 50%) was released after 24 h, reaching 90% of the released drug after 48 h. These results confirmed that the gel- atin coating provided the nanoplatform with a pH-responsive behavior and triggers the drug release in acidic conditions, as observed by HPLC analysis.[41]
To further demonstrate the accuracy of the nanosystem acting as SERS sensor, the DNP-AuNPs complex was incubated with four different concentrations of LY (25, 17.5, 12, and 9 106 m),capped with gelatin, and the SERS signals of the loaded LY was investigated. The mean LY SERS spectra acquired before (black) and after (red) the release were reported in Figure S5a–c (Supporting Information). Before the release, the intensity of the band at 1360 cm1 was 1500 counts s1, 1000 counts s1, 750 counts s1, and 600 counts s1 for the nanocomplex incu- bated with 25, 17.5, 12, and 9 106 M, respectively; after 48 h of release, the intensity was 150 counts s1 for all samples. The SERS intensity of the control sample, i.e., the DNP-AuNPs@ Gel complex without LY, was 30 counts s1 at 1360 cm1. These data reported in Figure S5d (Supporting Information) show that the SERS intensity was directly proportional to the LY con- centration with a high value of R2 (0.997). From this curve, it was also extrapolated that the LY SERS intensity measured after 48 h of release (150 counts s1) corresponded to about 2 106 M of LY. The SERS spectra of LY were well reproducible from 25 to 2 106 M and enabled to trace the release of LY in vitro.
The amount of LY released from the nanoplatform was quan- tified by SERS according to the nanoplatform loading capacity calculated by HPLC (Section 2.2). According to Figure 3c, about 50% of LY was released from 1.5 g of DNP-AuNPs-LY@Gel after 24 h in the acidic microenvironment, corresponding to about 0.125 fg per nanocomplex unit (about 15 ng overall of LY). These observations confirmed that the loaded anticancer drug was efficiently released in acidic conditions and that the release could be quantified per single nanocomplex unit thanks to the high sensitivity provided by the SERS technique.
We further investigated and quantified the time-dependent release of LY from the DNP-AuNPs-LY@Gel in living LS-174T cells by monitoring the LY SERS signal at various incuba- tion time intervals (0, 2, 18, 24, 42, and 48 h). A dispersion of DNPs-AuNPs-LY@Gel was incubated in 1.5 mL of cell medium at a final concentration of 50 g mL1, corresponding to a total mass of DNPs-AuNPs-LY@Gel of 75 g and a total mass of LY loaded on DNPs of 1.5 g. The mass of LY for DNP unit was 0.25 fg. After 2 h, the SERS signal was collected only from DNPs not internalized in the cells and the LY signal intensity was comparable to the control experiments without cells. At this point, the nanoplatform was not yet internalized and it was not possible to trace the SERS signal of LY in cells. Conversely, after 18 h the nanoplatform penetrated cancer cells and the LY SERS signal was detected inside cells, indi- cating a successful internalization. After 18 h the amount of LY released in LS-174T cells was about 30% of the total drug encapsulated, corresponding to 450 ng of LY (0.075 fg for DNP unit). The released LY was about 50% of the total encapsu- lated drug (750 ng overall, 0.125 fg for DNP unit) after 24 h. A smooth LY release was observed after 30 h and about 65% of LY was released (975 ng overall, 0.16 fg for DNP unit) after 48h. Interestingly, in this study, an efficient SERS intracellulartracing and imaging of LY in living CRC was demonstrated up to 48 h and quantified to provide LY sensing resolution down to 7.5 1018 g. For comparison, the LY release from the DNP- AuNPs-LY complex (without the gelatin coating) was inves- tigated in PBS pH 7.4 and pH 5.5 and in living CRC within 48 h by SERS analysis. A burst LY release of about 50% was observed after 60 s in all the release conditions (Figure S6, Supporting Information). The LY SERS signal completely dis- appeared after 10 min, in agreement with the in vitro releasetests investigated by HPLC (Figure S1, Supporting Informa- tion). These results further demonstrated that the nanosystem designed without the gelatin coating was not suitable for sus- tained drug delivery inside cancer cells.
2.4. Evaluation of MET Induced by DNPs-AuNPs-LY@Gel in CRC Cells
After assessing the drug release properties of the developed nanoplatform, we evaluated whether the controlled delivery of LY from the DNP-AuNPs-LY@Gel complex could increase the therapeutic efficacy of the drug, reducing tumor cell invasive- ness and metastasis.
For these studies, the LS-174T cell line was chosen due to its metastatic potential and mesenchymal phenotype.[51] The LS- 174T cells were also selected because they trigger an intact TGF- β signaling useful to test the inhibitory potential of LY delivered by our complex.[52] Firstly, the cellular uptake of DNP-AuNPs- LY@Gel labelled with Alexa Fluor 488 (DNP-AuNPs*-LY@ Gel) was investigated by confocal microscopy (Figure 4). To evaluate whether the gelatin layer affected the cellular uptake of the DNPs, the internalization of the labeled DNP-AuNPs-LY complex (DNP-AuNPs*-LY) was investigated as well. To this aim, 50 μg mL1 of DNP-AuNPs*-LY@Gel or DNP-AuNPs*-LY were incubated with the LS-174T cell line for 24 h; cell nuclei and membranes were stained with DAPI and WGA-Alexa Fluor 555, respectively. Untreated cells (first line of Figure 4) and cells treated with free dye Alexa Fluor 488 (last line of Figure 4) are also shown as controls.
The images revealed that the presence of the gelatin layer did not alter the cellular uptake of the DNPs since both the DNP- AuNPs*-LY@Gel and DNP-AuNPs*-LY complex were charac- terized by a comparable internalization; these results were also in agreement with previous studies performed on different cell lines.[20,25,28,29] The morphological analysis performed on the cells reported in Figure 4 allowed to evaluate the mesen- chymal to epithelial transition induced by the gelatin-coated and uncoated nanoplatform. The results, expressed in terms of elongated (mesenchymal) and rounded (epithelial) cells and reported as inset of the merge images, showed that only the DNPs-AuNPs*-LY@Gel complex was able to promote the rever- sion of the cell phenotype. Moreover, since the internalization of DNPs with and without the gelatin layer was comparable, the lower phenotype reversion observed with the DNP-AuNPs*-LY complex can be ascribed to the LY release from the nanosystem before its internalization in cells.
The biocompatibility of the DNPs-AuNPs@Gel suspen- sion was investigated by exposing the LS-174T cell line to the nanosystem at different concentrations (12.5, 25, 50, and 100 g mL1) for 24 and 48 h and monitoring the cell growth. The results reported in Figure S7a (Supporting Information) demonstrated that the multifunctional platform did not induce cell toxicity after 48 h up to a concentration of 50 g mL1, which was considered optimal for cell treatments. The cell growth of LS-174T was also investigated in presence of2.5 106 M of LY, 50 g mL1 of DNPs-AuNPs@Gel or DNPs- AuNPs, and 50 g mL1 of DNPs-AuNPs-LY@Gel or DNPs- AuNPs-LY (both containing 2.5 106 M of LY, according to theHPLC quantitative analysis reported in Section 2.2) (Figure 5a and Figure S7b: Supporting Information). We did not observe significant changes in cell growth after 48 h upon treatment with all the indicated nanoplatforms in the LS-174T cell line, demonstrating that also the nanoplatform loaded with LY can be considered safe.
For comparison, analogue studies were performed using the SW620 cell line, highly invasive colon cancer cells that display similar phenotype and genetic background of LS-174T cells.[52] In this case, we observed a slight increase in cell growth in the presence of DNPs-AuNPs@Gel, DNPs-AuNPs and DNPs- AuNPs-LY suspensions (Figure 5b and Figure S7c: Supporting Information). Overall, we observed that none of the compo- nents of the developed nanoplatform was toxic for the cells.
It has been shown that MET may be induced by blocking the activity of specific factors and signaling pathways that trigger EMT, including the TGF-β-mediated phosphorylation of SMAD proteins.[53] The TGF-β signaling pathway is a well- characterized factor involved in EMT and the development and progression of tumor metastasis. In CRC disease, the EMT is associated with an invasive and metastatic phenotype and the inhibition of the TGF-β signaling has been considered a valid strategy to halt EMT and enhance MET.[54] Therefore, we investigated whether the DNP-AuNPs-LY@Gel platform could directly reverse EMT by blocking the TGF-β1 receptor. The anti- metastatic effect of the developed hybrid multifunctional com- plex was assessed by investigating E-CADHERIN, SNAIL-1, and TWIST-1 levels in LS-174T and SW620 cells using quantitative PCR (qPCR) analysis. MET is characterized by loss of mes- enchymal markers (i.e., SNAIL-1 and TWIST-1) and, in turn, enhancement of epithelial markers (i.e., E-CADHERIN).[55] During the EMT, the cells lose their epithelial characteristics, like cell polarity and cell-cell contacts, and gain mesenchymal traits such as increased motility.[56] The transient nature of the EMT events allows mesenchymal cells to revert to an epithe- lial shape when signals driving the EMT decrease. In both cell lines, the DNP-AuNPs-LY@Gel complex significantly increased the expression of E-CADHERIN and suppressed the expres- sion of SNAIL-1 and TWIST-1 genes (Figure 5c,d). Of note, the modulation of the expression levels of E-CADHERIN, SNAIL-1, and TWIST-1 genes was much stronger upon treatment with DNP-AuNPs-LY@Gel than free LY at equal concentration. The encapsulation of LY in the hybrid nanoplatform (covered with gelatin) allowed lowering the LY dose necessary to induce MET in metastatic cells. Reduced doses can also help prevent adverse side effects and risks related to the circulation of toxic metab- olites in the plasma. Any gene modulation was not observed after treatment with DNPs-AuNPs@Gel, DNPs-AuNPs, and DNPs-AuNPs-LY suspensions (Figure 5c,d and Figure S7d,e: Supporting Information). These results confirmed that the gene modulation is observed only upon exposure to the nano- complex containing LY; the DNPs-AuNPs-LY complex realized without the gelatin layer did not induce any variation because it was not able to retain LY as demonstrated by HLPC (Figure S1, Supporting Information) and SERS (Figure S6, Supporting Information) analyses. Furthermore, since E-CADHERIN plays an essential role in maintaining epithelial integrity, its upregulation was also accompanied by morphological changes in both LS-174T and SW620 cells as observed in Figure 5e.
Specifically, cells treated with a suspension of 50 g mL1 of DNPs-AuNPs-LY@Gel showed an evident morphological shift from elongated and spindle-shape (mesenchymal) to rounded (epithelial) phenotype after 48 h compared to untreated or treated with DNPs-AuNPs@Gel or DNPs-AuNPs-LY cells (Figure 5e and Figure S7g, Supporting Information). The mor- phological changes were quantified recording the phenotype shapes over time (from 0 to 48 h) (Figure 5f). The transitionfrom elongated to rounded phenotype was already appreciable after 24 h of treatment with DNP-AuNPs-LY@Gel complex and it became stronger after 44 h since 60% of the cells lost mesen- chymal phenotype. The complete phenotype reversion occurred after 48 h of incubation in both cell lines, where 80% of cells acquired epithelial morphology (Figure 5f). The cell evolution to a rounded phenotype follows the same profile as the intracel- lular drug release evaluated by SERS analysis (section 2.3). Datacompared in Figure S8 (Supporting Information) confirmed the ability of the developed nanoplatform to provide a precise dose-response profile by correlating the amount of the internal- ized drug and cell outcome. Moreover, the observed morpho- logical changes confirmed that the metastatic process could be slackened by treating the CRC cells with the developed nano- platform which turns out to be more efficient if covered with gelatin.
Finally, we studied the effects of the LY delivery on normal human colon epithelial cells (CRL-1790) upon exposure to 50 g mL1 of the DNP-AuNPs-LY@Gel or DNP-AuNPs-@Gel complex.[57] The treatment with DNP-AuNPs-LY@Gel did not induce effects on the normal cell line since the expression levels of SNAIL-1 and TWIST-1 genes were unvaried (Figure S7f, Sup- porting Information). However, since the normal cells are gen- erally responsive to TGF-β signaling, we cannot exclude that some other genes, not discussed in this study, could be affected by the treatment with the DNP-AuNPs-LY@Gel complex.
In conclusion, even if the DNP-AuNPs-LY@Gel nanosystem exhibited the ability to revert the metastatic process in CRC cells, it is worth considering the administration of LY and a cancer-cytotoxic agents to block the metastasis and reduce the tumour growth simultaneously. [58,59]
3. Conclusions
Here, for the first time, we describe the production of a plas- monic-assisted nanoplatform for the real-time monitoring of Galunisertib release in colorectal cancer cells. The ability of gold NPs to enhance the SERS signal of Galunisertib loaded in the DNP was efficiently exploited to trace and quantify the release of the small molecule in living cells over days. Galunisertib was loaded in the DNP with a loading capacity of 20 g mg1 and the nanoplatform was capped by a thin layer of gelatin, over- coming the common burst release effect of porous drug car- riers. In vitro release studies showed that the DNP-AuNPs-LY@ Gel nanoplatform exhibited a pH-responsive release of Galuni- sertib starting at pH 6. The gelatin shell provided the pos- sibility to prevent burst release effects, enhance the LY delivery in cells and control the LY release in malignant cells, where the acidic microenvironment and the overexpression of MMP-2 enzymes promoted the gelatin degradation. The drug release profile investigated by either HPLC and SERS techniques revealed that SERS is a high-sensitive and reliable method for the investigation and quantification of drug release profiles. Significantly, the SERS enhancement of Galunisertib signals in the DNP-AuNPs-LY@Gel platform allowed extreme sensitivity down to sub-femtogram of the drug, enabling the label-free monitoring of drug release in living cells. The biocompatibility was demonstrated in vitro in both LS-174T and SW620 cancer cell lines, showing that the DNP-AuNPs-LY@Gel treatment was safe up to a concentration of 50 g mL1. The modula- tion of the expression levels of E- CADHERIN, SNAIL-1, and TWIST-1 genes upon treatment with the nanoplatform was significantly higher than free Galunisertib. Moreover, the anti- metastatic effect was further demonstrated by confocal micros- copy, revealing that the shape of LY2157299 cells efficiently turned from mesenchymal to epithelial-like phenotype in the presenceof the multifunctional platform. These studies showed that the encapsulation of Galunisertib in the developed nanoplatform can help lowering the amount of drug required to inhibit the metastatic process in cancer cells, reducing the formation of drug-related toxic metabolites.