Chitosan overlaid Fe3O4/rGO nanocomposite for targeted drug delivery, imaging, and biomedical applications

A hybrid and straightforward nanosystem that can be used simultaneously for cancer-targeted fluorescence imaging and targeted drug delivery in vitro was reported in this study. A chitosan (CS) polymer coated with reduced graphene oxide (rGO) and implanted with Fe3O4 nanoparticles was fabricated. The fundamental physicochemical properties were confirmed via FT-IR, XRD, FE-SEM, HR-TEM, XPS, and VSM analysis. The in vivo toxicity study in zebrafish showed that the nanocomposite was not toxic. The in vitro drug loading amount was 0.448 mg/mL−1 for doxorubicin, an anticancer therapeutic, in the rGO/Fe3O4/CS nanocomposite. Furthermore, the pH-regulated release was observed using folic acid. Cellular uptake and multimodal imaging revealed the benefit of the folic acid-conjugated nanocomposite as a drug carrier, which remarkably improves the doxorubicin accumulation inside the cancer cells over-express folate receptors. The rGO/Fe3O4/CS nanocomposite showed enhanced antibiofilm and antioxidant properties compared to other materials. This study's outcomes support the use of the nanocomposite in targeted chemotherapy and the potential applications in the polymer, cosmetic, biomedical, and food industries.

and the epoxide (-O − ) groups results in the weak chemical interaction, such as H-bonding, or chemical interaction causing the transpire on the surface [8][9][10][11] . Even the entire areas of GO can be used for drug loading and noncovalent functionalization because of the possible π-π interactions on the basal plane. Thus, when designing an efficient drug delivery system, GO acts as a surfactant to stabilize hydrophobic molecules due to its amphiphilic nature 12,13 . However, Wang et al. 14 reported that pure GO is toxic to cells in vitro and animals, as the kidneys cannot empty it, thereby resulting in the granuloma formation in the lung or cell apoptosis lung granuloma formation even though apparent toxicity did not observe in the lower dose and moderate dose approximately 0.1 mg and 0.25 mg, respectively. Thus, GO must be coated with polymers, such as chitosan (CS), and polyethylene glycol, to improve its solubility. CS is made of two types of structural units: 2-amino-2-deoxy-d-glucose and N-acetyl-2-amino-2-deoxy-d-glucose linked by a β(1 → 4) bond, and CS is derived from chitin 15 . CS is a polycationic biopolymer, abundant in different crustacean shells like crab, shrimp and crawfish. CS is a low-cost material with biodegradable and biocompatible properties, and ease of chemical modification. CS posses chelation property and selectively binds with metal ions and biomolecules like proteins, cholesterol and tumor cells. Due to this property, CS is widely used in the food industry, pharmaceutical, and water management farms. CS possesses properties useful in the medical field to inhibit tumor cells, as antimicrobial agents, as a wound-healing agent and as immunostimulant 16 . CS-coated GO has improved solubility and drug loading capability, in addition to its increased potential to distribute the drug molecule at the tumor site, which has an acidic environment 7,9,15 .
When evaluating the designed drug delivery system's efficacy, the loaded drugs and ligands should be considered. Doxorubicin (DOX) is one of the best therapeutic agents for cancers. The cytotoxicity and DNA cleavage can be occurred through binding to DNA (intercalation mode) or reacting with topoisomerase II 16,17 . Folic acid (FA) is a highly selective ligand that binds to the folate receptor 18 . This ligand also has high target specificity to various types of cancer cells. Binding of the carrier molecules with FA improves the efficiency and reduces the adverse effects of the drug molecules 19,20 . The pH-responsive release behavior and anticancer activity of Dox-loaded, FA-conjugated, GO-Fe 3 O 4 nanocomposites were reported in the previous studies. However, The combination of rGO/Fe 3 O 4 /CS was a novel attempt. Such a facile protocol was resulted in Fe 3 O 4 with a size of less than 35 nm. This combination can be applied to the drug delivery system because it presented an appropriate cancer therapy result. Briefly, rGO can interact with DOX through a π-π bond and hence enhance the drug loading capacity, in which Fe 3 O 4 , as a magnetic agent, can readily move through the therapy without any external magnetic field. Besides, the biopolymeric backbone of nanocomposite (CS) can control the neutral pH's release behavior.
Furthermore, FA is a receptor and DOX is a model drug. Therefore, the current attempt was carried out as a novel study, in which there has been no report on trGO/Fe 3 O 4 /CS/FA/DOX combination. Accordingly, this study aimed to achieve efficient drug loading, release, and fluorescence imaging with chemo-photothermal combination therapy for cancer; herein, we report a biocompatible theragnostic nano platform, rGO/Fe 3 O 4 /CS, which was synthesized via a solvothermal method and studied as a targeted drug delivery system with DOX and FA (DOX/rGO/Fe 3 O 4 /CS/FA).

Results and discussion
Chemical bond analysis using FT-IR. The chemical properties and bonding nature of the synthesized rGO/Fe 3 O 4 (5%) and rGO/Fe 3 O 4 /CS (15%) nanocomposites were studied by infrared spectroscopy, and the obtained spectra are shown in Fig. 1. The nanocomposites' spectra were primarily compared with the FT-IR spectra of pure GO, CS, and Fe 3 O 4 NPs ( Fig. 1) to ensure the presence of the respective materials in both nanocomposites. The peaks present at 3563 cm −1 , 2036 cm −1 , 1410 cm −1 , and 959 cm −1 in the spectra of the pure GO corresponds to the O-H, C=O, C-OH, and C-O functional groups, respectively, (Fig. 1). In pure CS, peaks at 968 cm −1 , 2006 cm −1 , and 2975 cm −1 corresponded to the characteristic stretching vibrations of C-O, NH 2 , and C-H bonds, respectively ( Fig. 1) 21 . A common broad peak at 3378 cm −1 present in the FT-IR spectra of both GO and CS was credited to the hydroxyl groups present in the adsorbed water molecules present at the surface of both GO and CS. In the Fe 3 O 4 NP, the peak at 604 cm −1 represented the presence of a Fe-O bond in the material ( Fig. 1) 20,22 . Furthermore, the peak at 604 cm −1 attributed to the stretching vibration of the Fe-O bond. This band was shifted to 952 cm −1 , due to the bonding of Fe 3 O 4 NPs with the -COO-groups on the rGO surface 23  Crystal structural analysis using XRD. As shown in Fig. 2, the crystallographic properties of both rGO/ Fe 3 O 4 and rGO/Fe 3 O 4 /CS nanocomposites were obtained using X-ray powder diffraction (XRD) patterns. A sharp peak at 2θ = 12.7° and a broad peak at 2θ = 25.9° corresponding to the (001) and (002) planes, respectively, confirmed the formation of GO with an interlayer thickness of about 8.75 Å and weak amorphous peak at 2θ = 19.2° could be attributed to CS as shown in Fig. 2 25  FE-SEM analysis. The GO morphology was observed using field emission -scanning electron microscopy (FE-SEM), by which GO illustrated a network structure with a high specific surface area and minimum thickness formed by the GO nanosheets (Fig. 3). In pure CS (Fig. 3), smaller particles were agglomerated and formed large clusters with irregular shapes. The pure Fe 3 O 4 NPs had a spherical shape with a particle size distribution of 20-25 nm, and seemed to be well distinguishable and mostly agglomerate-free. These properties are essential to attain the superparamagnetic property. The morphological changes in both GO nanosheets and Fe 3 O 4 NPs during the formation of the rGO/Fe 3 O 4 (5%) nanocomposite were visible in their FE-SEM images, as shown in Fig. 3. The reduction of GO to rGO produced more thin and transparent sheets due to the additional exfoliation of the sheets during the removal of oxygen species.   (Fig. 4D). The crystalline plane of the Fe 3 O 4 system was detected by its diffraction dots and rings represented, which is in agreement with the XRD data.  20 (Fig. 5A). The presence of the Fe 2p peak in the survey spectra was from the Fe 3 O 4 NPs, whereas the N 1 s peak might have been from the CS group in the composite. The O 1 s peak was due to the oxygen species present in all of the three materials. Figure 5D illustrates the high-resolution XPS spectra of Fe 2p. The Fe 2p peak was further deconvoluted into 6 individual peaks, in which the peaks present at 710.8 and 723.9 eV emerged due to the 2p 3/2 and 2p 1/2 split orbitals of the Fe 2+ ions, respectively. The peaks at 712.8 and 725.7 eV were due to the 2p 3/2 and Magnetic measurements using VSM. Figure 6 illustrates the M-H loop of the rGO/Fe 3 O 4 /CS (15%) nanocomposite under a maximum induced magnetic field of ± 30 kOe at room temperature. The prepared nanocomposite was superparamagnetic with a saturation magnetization (MS) of 5.27 emu/g, attributed to Fe 3 O 4 NPs in the composition 8,20 . However, it should be noted that the obtained MS value was low compared to that of pure Fe 3 O 4 NPs according to the previous reports 28 . This significant decrease in the MS of rGO/Fe 3 O 4 /CS (15%) nanocomposite may have been due to the existence of a high amount of non-magnetic rGO and CS 29 . However, the prepared composite showed excellent response in the external magnetic field. The other magnetic parameters, such as remanent magnetization (MR) and coercive field (HC), were 1.23 emu/g and 0.020 kOe. Hence, it can be concluded that the presence of   www.nature.com/scientificreports/ Additionally, the nanocomposite's different material combinations did not impact the survival rate of zebrafish at various developmental stages (from 0 to 72 h), as shown in Fig. 7. This finding was related to the utilization of CS by the embryo as nourishment. Interestingly, our results did not exhibit any abnormalities compared with other materials. Earlier, toxicological studies on mice have presented that graphene strongly improves the dispersity in the physiological medium and enhances biocompatibility 32,33 . In the present study, rGO/Fe 3 O 4 /CS exhibited less toxicity and malformations in zebrafish.  Figure 8A shows the images of the rGO/Fe 3 O 4 /CS/FA nanocomposite loaded with DOX in necessary conditions (pH 7.4). A maximum of 95% of DOX molecules was loaded on the rGO/Fe3O4/ CS/FA nanocomposite under necessary conditions based on the studies. The saturated loading amount of DOX in the rGO/Fe 3 O 4 /CS/FA nanocomposite was 0.5 mg/mL −1 ; therefore, a maximum of 0.448 mg/mL −1 of DOX was loaded 34 . Sasikala et al. 34 claimed the highest loaded amount of drugs on the graphene. However, in this study, the loaded drug amount was higher than that of the reported value. This increased concentration of loaded DOX may be attributed to three significant phenomena: (i) π-π stacking and the H-bonding between GO and DOX (ii) the interaction between the phenolic OH and alkaline amino groups of the DOX molecule with the amino groups and hydroxy on the CS shell of the rGO/Fe 3 O 4 /CS/FA nanocomposite, which forms intermolecular complexes by hydrogen bonding; (iii) and the mucus nature of the CS shell, which provided physical adsorption, the rGO, which provided high surface volume binding with the OH group, and the superparamagnetic materials, which results in high drug loading efficiency of the synthesized nanocomposite. Although Fe 3 O 4 NPs occupied some surface area of GO, a comparable amount of DOX was loaded. This amount of loaded DOX was higher www.nature.com/scientificreports/ than that found in commonly used nanocarriers, such as polymer micelles, hydrogel microparticles, liposomes, and carbon nanohorns, where the loading amount is reported as below 1 mg/mg −135-38 . Additionally, the DOX molecules loaded under necessary conditions were highly efficient and reliable for target delivery because the free carboxylic acid groups can form hydrogen bonds, thus resulting in efficient pH-induced targeted delivery. Additionally, the prepared rGO/Fe 3 O 4 /CS/FA nanocomposite showed a homogenous dispersion in an aqueous medium due to the carboxylic and hydroxyl groups' existence graphene. Furthermore, the drug release at 37 °C from the composite was studied in simulated body fluid (phosphatebuffered saline; PBS, pH 7.4) and pH 5.5 (PBS). The release profile of DOX from the rGO/Fe 3 O 4 /CS/DOX/FA nanocomposite into the outer aqueous phase is shown in Fig. 8B. The faster pure DOX release behavior at acidic condition was detected, which was pH-dependent, in which the release value was 25% at pH 7.4 and 81% at pH 5.5. Such a pH-responsive property of MNDC can be practically utilized as a controlled release system, which can gently release doxorubicin in the cancer cell because the pH of intracellular is lower than healthy cells 39 . Under the two different pH conditions, the cumulative drug release prominently occurred at 10 h, and maximum (96.6% and 10%) drug release was achieved at pH 5.5 and at pH 7.4, respectively. This decrease in drug release at pH 7.4 may be due to DOX's decreased solubility at pH 7.4.

In vitro studies on drug loading and pH-regulated release properties of rGO/Fe
Additionally, the CS polymer was less soluble at pH 7.4, resulting in low permeability. The biodegradable combination of rGO with CS has presented stronger mechanical properties with a large drug delivery capacity, considered a pH-responsive and biodegradable drug delivery system. CS NPs conjugated with drug via pH-cleavable interaction tend to dissociate at the acid condition of endo-lysosomes, and then release the drug into the cytoplasm. The increment in the biodistribution level and drug concentration in the cancer cells is the CS's major anticancer mechanism. CS incorporated with mifepristone controlled the drug release in a constant release behavior and improved the drug's oral bioavailability and anticancer properties via pharmacokinetic study under in vivo analysis 40 . At pH 5.5, the release of DOX was favorable, as the CS polymer was water-soluble, resulting in increased water content and increased permeability of DOX. Because of this pH dependence, DOX can be specifically distributed around tumor tissues rather than non-selective release throughout the body. The pH-dependent release of anticancer drugs was high in acidic conditions. Thus, DOX molecules were protonated and the hydrophobic interactions between DOX molecules and the nanocomposites were decreased in an acidic environment, which accelerated the DOX release. The protonation of DOX molecules and reduced hydrophobic interactions between DOX molecules and the nanocomposites were observed at acidic conditions, which accelerated the DOX release. It demonstrates the potential use of the rGO/Fe 3 O 4 /CS/DOX/FA nanocomposite in cancer therapy 41 .   (Fig. 11A,B). The yellow fluorescence from the DOX emissions suggested the significant uptake of rGO/Fe 3 O 4 /CS/DOX/FA (DOX-loaded folate labeled particles) by the cells. The blue fluorescence, which represents DAPI, clearly indicated that rGO/Fe 3 O 4 /CS/DOX/FA was primarily localized in the nuclear region. Acridine orange (AO) was utilized to investigate DOX's localization in the perinuclear region and the nucleus. Under blue excitation, the green emission from DOX indicated its binding to the nucleus and cytoplasm 44 .
However, the drug-loaded nanocomposite was primarily internalized in the perinuclear region or in the nucleus, as shown in the image (Fig. 11). The AO fluorescent stain was used to observe the morphological changes in the apoptotic and necrotic cells. The green fluorescence and the yellow emission indicated that the nanocomposites entered into cells with broken membranes, such as apoptotic and dead cells, certain to targeted DNA fragments or apoptotic bodies. Quantitively, 80% of the cells were apoptotic and necrotic due to the DNA fragment injuries.
Furthermore, to corroborate the folate receptor-mediated drug delivery behavior, we compared the emission of rGO/  However, there were still quite a lot of biofilm structures remaining and the number of dead microbes significantly increased, indicating the biofilm ability of NPs. Results show that NPs induce the biofilm formation by P. aeruginosa, S. aureus, and C. albicans; however, the concentration of the NPs should be continuously increased for the biofilm to be destroyed. Therefore, the rGO/Fe 3 O 4 /CS nanocomposite destroyed a maturely formed biofilm structure and controlled the thickness by breaking the exopolysaccharides. Antioxidant activity. The 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay was utilized to investigate the nanomaterials' antioxidant activity in which ascorbic acid was considered positive control. DPPH is a stable compound that accepts hydrogens or electrons, and the reduction of DPPH is directly proportional to the samples' antioxidant nature. The free-radical scavenging ability of GO, CS, Fe 3 O 4 NPs, rGO/Fe 3 O 4 nanocomposite, and rGO/Fe 3 O 4 /CS nanocomposite as compared to the material combination as shown in (Fig. 13A,B). It reveals that both materials show free radical scavenging ability dependent on their materials. The free radical % scavenging potential of control, GO, CS, Fe 3 O 4 NPs, rGO/Fe 3 O 4 nanocomposite, and rGO/Fe 3 O 4 /CS nanocomposite were 0%, 48%, 58.6%, 66%, 76.6%, and 87.6%, respectively, based on the DPPH activity (Fig. 13A,B). The nanomaterials showed enhanced scavenging activity, increasing DPPH scavenging potential for CS, Fe 3 O 4 NPs, and rGO. Lower concentrations of the polymer-coated nanocomposite reflected its higher potency for free radical scavenging and total antioxidant capacity. Similar trends in the improvement of DPPH scavenging behavior by GO, Fe 3 O 4 NPs, and CS NPs have been previously reported [45][46][47][48] . www.nature.com/scientificreports/

Conclusions
We have developed a hybrid system by the solvothermal method that was highly biocompatible, water dispersible, fluorescent, superparamagnetic, and multifunctional. This hybrid possesses a large loading capacity of approximately 0.448 mg/mL −1 with the DOX anticancer drug. The release of DOX can be pH-triggered and controlled magnetically. Specifically, the surface modification of Fe 3 O 4 NPs with folic acid improved its uptake by cancer cells. The toxicological study revealed the cytocompatible nature of rGO/Fe 3 O 4 /CS nanocomposite in A549 and MCF7 cells. In vitro cellular imaging of rGO/Fe 3 O 4 /CS/DOX/FA in A549 and MCF7 cells showed significant co-localization to the cytoplasmic region. The toxicity studies showed high biocompatibility of the rGO/Fe 3 O 4 / CS to zebrafish without inducing significant abnormalities. Additionally, the zebrafish survival rate was not affected. Uniform distribution over the entire body was observed zebrafish using In vivo whole-animal imaging of rGO/Fe 3 O 4 /CS. Besides, rGO/Fe 3 O 4 /CS nanocomposite highly efficient in disrupting the biofilm formed by S. aureus, P. aureginosa and C. albicans and also exert productive antioxidant potential. Hence, rGO/Fe 3 O 4 /CS may serve as a biologically potent multipurpose material for future biomedical investigations and could be applied as a nano-deliver agent (nanocarrier) to the combination of photodynamic therapy with gene delivery or magnetically-guided drug.

Materials and methods
Chemicals and reagents. Graphite powder, ferric chloride (FeCl 3 .6H 2 O) (99%), ferrous sulfate (FeSO 4 .7H 2 O) (99.5%), sodium nitrate (NaNO 3 ) (99%), sodium hydroxide (NaOH) (97%), potassium permanganate (KMnO 4 ) (99%), sulfuric acid (H 2 SO 4 ) (95-97%), folic acid (FA) (97%), glutaraldehyde (98%), sucrose (99%), and ethylene glycol (EG) (99%) were purchased from Merck (India). Chitosan (CS) powder (medium molecular weight-190,000 to 310,000 Da) with a deacetylation degree of 85%, and doxorubicin hydrochloride (98% with 2 years of validation days) were purchased from Sigma-Aldrich (India). Microbial cultures (Strepto- Material synthesis. Preparation of graphene oxide. Modified Hummer's method was used to prepare to GO. The graphene oxide was prepared based on the Hummer's method with some modification. Briefly, graphite powder (2 g) and NaNO 3 (2 g) were mixed to 100 mL of concentrated sulfuric acid. After stirring the mixture in an ice bath for half an hour, 7 g of KMnO 4 was added to the solution. Afterward, the solution was further stirred for 1 h at ambient condition. Next, the solution was stirred overnight in order to oxidize the graphite. One hundred milliliters of Milli-Q water was poured into the mixture to neutralize. An additional 400 mL of Milli-Q water was added, and the mixture was stirred for 30 min. Next, 10 mL of 30% hydrogen peroxide was added to terminate the reaction. Finally, the solution was subjected to stirring, centrifugation, and then washing (with 5% HCl) until the pH became close to 7. The resultant was separated by filtration and extensively washed with H 2 O; the obtained GO was then freeze-dried.   Material characterization techniques. The infrared spectral studies were carried out using a Perkin-Elmer FT-IR spectrophotometer using KBr pellet. The crystalline structural analysis was performed using a PW3040/60 X'pert PRO X-ray diffractometer with Cu Kα radiation and a wavelength of 1.54060 Å. Microstructures of the pure and nanocomposite were examined by FE-SEM (Model: Hitachi S-4500). The chemical compositional analysis was studied using the EDX. HR-TEM (Tecnai) was used to study the particle size and nature. The chemical state analysis was performed with XPS (Carl Zeiss ) using Al Kα excitation at 250 W. The room temperature magnetization was measured by a vibrating sample magnetometer (VSM, Lake Shore, Model-7410, USA). The samples were also analyzed by UV-Vis spectroscopy (Shimadzu UV/Vis 1800 spectrophotometer), and the absorbance was measured at 620 nm using a Thermo Multiskan ELISA multiwell plate reader (EX, USA). The morphology of the A549 and MCF7 cells was studied using Nikon (Japan) bright-field inverted light microscopy at 40X magnification and fluorescence microscope (Nikon Eclipse, Inc, Japan) at 400X magnification with an excitation filter at 480 nm. Biofilm inhibition was inspected by light microscopy (Olympus cx21i) CLSM (NIKON ECLIPSETS 100) at a magnification of 20X.

Pharmacological Study. Microinjection of GO, CS, Fe 3 O 4 NPs, rGO/Fe 3 O 4 (5%), and rGO/Fe 3 O 4 /CS (15%)
into zebrafish embryos and microscopic measurements. Wild-type AB strains of Danio rerio (D. rerio-zebrafish) embryos were supplied from the zebrafish core facility center, National Tsing Hua University. The MNDC samples were ultrasonicated for 30 min in Milli-Q water and then microinjected (10 nL) in the fresh embryos placed on the microinjection embryo tray. The dark incubated (28 °C) control embryos (without MNDC) and MNDCloaded embryos were periodically observed using microscopic images 49,50  www.nature.com/scientificreports/ All of the measurements were done in triplicates. The control graph measured the actual releasing rate at various pHs, and the drug content loaded into the rGO/Fe 3 O 4 /CS/ FA/DOX was calculated by equation.
Maintenance of cell lines. Experimental cell lines include A549 and MCF7 were supplied from the National Centre for Cell Sciences (NCCS), Pune, India. The cell lines were sustained in Dulbecco's modified eagles medium (DMEM), holding 10% Fetal bovine serum (FBS). The cells were sustained in a humidified CO 2 incubator according to procedure 42 .
Cytotoxic analysis. The toxic activity of MNDC, MNDC/DOX and MNDC/DOX/FA against cancer cell lines were studied using MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] test 49 . Certain cancer cell lines (1 × 10 4 cells/well) were introduced into the well of 96-well plate at their IC 50 values. The medium was introduced with serially diluted MNDC, MNDC/DOX and MNDC/DOX/FA, and allowed to incubate for 48 h. After incubation, the medium was substituted with 100 µL of the MTT solution and incubated at 37 °C for 4 h. The plate was read at 620 nm in an ELISA multi-well plate reader (ThermoMultiskan EX, USA). The cell viability was calculated using the below formula, Cell viability (%) = Absorbance value of treated sample/ Absorbance value of control × 100.
Morphological analysis. Change in the morphology of the A549 and MCF-7 cancer cell lines (1 × 10 5 cells/ coverslip) concerning the release of drugs, which was probed using MNDC, MNDC/DOX and MNDC/DOX/ FA. The coverslips presented on the clean glass slides were detected using a Nikon (Japan) bright-field inverted light microscopy at 40X magnification.
Apoptotic analysis. The cell suspension (1 × 10 5 cells/mL) was stained using ethidium bromide (EthBr) (100 mg/mL) and AO and mildly spread on the slide to investigate apoptosis. The cancer cell lines were gently flooded with Na 2 HPO 4 -KH 2 PO 4 buffer solution (pH 7.2) before AO/EtBr staining. Afterward, the cells were rinsed with PBS twice and evaluate under a fluorescence microscope at 40X magnification with an excitation filter at 480 nm. The (4′,-6-diamidino-2-phenylindole) DAPI analysis was conducted based on the same protocol 52 .
Antibiofilm assessment. The samples' potential antibiofilm ability was investigated using S. pneumonia, P. aeruginosa, and C. albicans as Gram-positive, Gram-negative, and fungus, respectively. Briefly, The fungal and bacterial biofilm was grown on the sterile glass submerged potato dextrose broth and nutrient broth, respectively. Afterward, the nanomaterials were individually added to each well and incubated at 37 °C for 24 h. Consequently, the loosely attached cells on the glass were thoroughly removed using PBS, and the glass was stained nucleic acid fluorescent dye AO (0.4% www.nature.com/scientificreports/