3. Results and Discussions
3.1 Synthesis of fluorescent ZnO NPs and optimization of fluorescence: A facile chemical method was employed to synthesize the fluorescent zinc oxide NPs as represented in Figure 1 . A LSCM was employed to perform a detailed study on the optical properties of the synthesized fluorescent particles. First, we performed XYZʎ scanning in confocal microscopy, which suggested that these particles can be imaged using excitation at 405 and 488 nm with an emission wavelength in the range of 415-460 and 504-668 nm, respectively (Supplementary Figure S3 ). Since lower wavelength is not suitable for live imaging for longer period of time, 488 nm was used as the excitation wavelength for all the nanoparticle retention studies.
Figure 2(a) and (b) show that fluorescence intensity increases with Tween-80 concentration. Since we aim to retain fluorescence till 72hour, 20% (v/v) Tween-80, was selected for further investigations. Additionally, to retain the fluorescence in DMEM media and to obtain a slower release, the ZnO_T particles were coated by PLL. The dispersion of the ZnO_T and ZnO_T_PLL NPs in DMEM media along the course of 72 hours has been shown in Figure 3(a) and(b) . The time course of fluorescence intensities, as shown inFigure 3(c), demonstrated a significantly more retained fluorescence (p<0.0001) of the ZnO_T_PLL NPs over ZnO_T NPs. The result clearly shows that PLL encapsulation is able to retain the particle fluorescence level for a longer period of time. In order to justify the ability of the NPs to be stored and transported, the fluorescence of dry ZnO_T_PLL NPs was monitored over a period of one year, as shown in Supplementary Figure S4 . The particles were found to retain their fluorescence over three months, showing a potential for transportation and storage.
3.2 Characterization of NPs: FESEM images (Figure 4(a) and (b) ) and TEM images (Figure 4(c) and (d) ) show that the resultant NPs had a roughly spherical shape with size average size of 21 nm and 29 nm for ZnO_T and ZnO_T_PLL respectively. In order to obtain the size distribution, DLS analysis was performed. The results reveal the average hydrodynamic diameter of 73.05 nm and 104.82 nm for uncoated (ZnO_T) and coated (ZnO_T_PLL) NPs, respectively (Figure 4(e) and (f) ). EDX spectra (Figure 5(a) and (b) ) indicate the presence of Zn, O, and C both in ZnO_T and ZnO_T_PLL due to the presence of ZnO and Tween-80. The introduction of N in ZnO_T_PLL attests to the presence of PLL in the NP.
The FTIR analysis shown in Figures 5(c) and (d)confirms the peak at around 470 cm-1, which corresponds to the stretching vibration of the Zn-O bond present in all the NPs (Khan et al., 2010). The absorption peak at 1450 cm-1 from the spectra of ZnO_T is due to bending vibration of O-H, which may be accounted for by the interaction of ZnO with the C-H bond of Tween-80 ( ). A broad peak at 3400 cm-1 in ZnO_T suggests an abundance of O-H groups present on the surface of ZnO_T. The addition of a peak at 1590 cm-1 in the ZnO_T_PLL spectra is due to the introduction of amide groups present in PLL. The broad peak from 3100-3600 cm-1 is due to overlapping of amino and hydroxyl groups, the former introduced from PLL (Babic et al., 2008). The surface charge of ZnO_T and ZnO_T_PLL particles were measured by the zeta potential assay. Supplementary Figure S5 clearly depicts the variation of the surface zeta potential for a wide range of pH. The result shows that the surface charge of ZnO_T_PLL (4.6 mV) was found to be higher than ZnO_T (-0.5 mV) at pH 5.5, indicating the greater potential of ZnO_T_PLL to be attached to the cell surface.
3.3 NP internalization and retention study: In order to evaluate the efficacy of drug uptake and retention of ZnO_T_PLL NPs, they were tested on MCF-7 cells. The MCF-7 cells were chosen to mimic epithelial uptake in cells from the mammary gland, which is the cell type that is generally targeted for breast cancer treatment. Since LSCM coupled with an incubator can be employed for time-lapse 3D imaging, here we implement LSCM based imaging set-up for in-vitro testing of particle retention and ROS accumulation in tumor cells. The schematic diagram for the investigation on the particle internalization, cellular retention, and ROS generation is shown in Figure 1. In order to depict the interaction between the particles and cells, we performed time-lapse imaging for 5 hours using DIC mode after incubation for two days. This video demonstrates the dynamics of MCF-7 cells and consequent changes in cell morphology with time when treated with ZnO_T_PLL (Supplementary Video S1) . However, such a continuous time-lapse generation for three days was avoided as a long-exposure to laser may lead to cell apoptosis.
Although Supplementary video S1 was able to capture the dynamic changes in cells, such imaging cannot be used for quantification of retention of ZnO_T_PLL. In order to image the particle internalization along with the cell height, fluorescent intensity was captured for 38 stacks in order to cover the top-most and bottom-most planes, which varies in a range from 0 to 38µm. Specifically, 3D imaging was performed to obtain the integration of fluorescent signals from individual planes across the layers within a cell. The specific advantages of 3D imaging is shown in Figure 6 through a comparison of 2D and 3D images. It is evident from Figure 6(c) and Supplementary Video S2 that the internalized particles were present at multiple planes. The individual z-scans spanning a depth of 38µm were created by compression of individual 2D images presented in Supplementary Figure S2(i)and (ii).
Similarly, ROS generation at different z-planes are shown inSupplementary Figure S2(b) . The result shows how the 3D reconstruction provides accurate signals corresponding to the amount of internalized particles, and ROS generation, when compared to 2D imaging, thereby making the quantification more reliable (Figure 6(a) and (b) ). We also present the depth coding for all the cells with internalized particles in Figure 6(d) that clearly indicates that MCF-7 cells are present at different z levels, confirming the necessity of 3D imaging. To further attest to the internalization of the particles inside MCF-7 cells we combined fluorescence imaging and DIC imaging and the merged image in Supplementary Figure S6 shows that the ZnO_T_PLL particles are present inside the cells and not on the surface of the cells.
Since NP retention and drug-mediated cell death is known to be crucial in determining the efficacy of the NP formulation, first, we focus on the investigation of the cellular retention. MCF-7 cells were incubated with 40 µg/mL of ZnO_T_PLL NPs, and the imaging of particle internalization was acquired over 72 hours. Figure 7(a),(b)shows the representative images of particle retention at 0, 12, 24, 36, 48, and 72 hours along with intensity mapping. Since NPs are known to induce apoptosis through the generation of ROS in excess amounts, we focus on investigating the dynamics of the ROS produced in assessing this as the mechanism to induce toxicity in the presence of ZnO_T_PLL NPs. Figure 7(c) and (d) show the time-lapse images corresponding to intracellular ROS generation at various time intervals and corresponding intensity maps in a cell population.Figure 7(e) represents the merged images of cellular internalization and ROS formation.
Specifically, 40 µg/mL concentration ZnO_T_PLL was chosen for this study since the particle internalized was found to be significant at 48 hours for this concentration. The comparison of particle internalization in MCF-7 cells at 0, 12, 24, and 48 hours in the presence of 10 and 40 µg/mL of ZnO_T_PLL NPs and corresponding intensity maps are shown inSupplementary Figure S7 . The confocal images of the cells showed a dose-dependent fluorescence intensity with a significantly higher fluorescence signal at 40 µg/mL. The comparative study suggests the suitability of using 40 µg/mL as the concentration to be assessed. Moreover, we performed a study on differential cell toxicity and the preferential killing capability of the ZnO_T_PLL NPs for MCF-7 cells over healthy L929 cells. Supplementary Figure S8 shows that the cell viability of MCF-7 cells was significantly (p<0.0001) lower than normal L929 cells in the presence of ZnO_T_PLL NPs (p<0.0001). The viability of MCF-7 cells was found to be 43.53%, while L929 cells were 83.3% viable post 24 hours of incubation. The results clearly demonstrate the preferential killing of cancerous MCF-7 cells over healthy L929 cells. This study also showed that 40 µg/mL could be used as a subtoxic dose as the normal cells show significantly higher viability (p<0.0001) than cancer cells over a period of 72 hours (Supplementary Figure S8 ).
Next, we show how LSCM allowed internalization study in single cells through enabling 3D imaging at a higher resolution, which ensured the internalization of the particles in subcellular parts rather than particles being adsorbed on the cell surface. Figure 8 (a) and (b) show the internalization of ZnO_T_PLL in single cells.The result also shows the aggregation of ZnO_T_PLL NP clusters to be localized in the outer cell membrane of the MCF-7 cells up to 12 hours as found in Figure 8(ii) with no significant morphological changes in the cells. Detectable fluorescence at 12 hours post-treatment indicates the starting of ZnO_T_PLL internalization inside the cytoplasm of the MCF-7 cells (Figure 8(a),(b) ). It can be concluded that, early phases of ZnO_T_PLL internalization inside the cytoplasm of the cell is not associated with a prominent change in morphology (Figure 8(iii) and Supplementary Figure S9 ). The presence of higher levels of fluorescence at 48 hours, as evident from images of cellular uptake study, affirms the presence of the ZnO particles in the second day of treatment, which further confirms the cellular retention Figure 8(v) . The result shows that LSCM, being a potential tool to evaluate the NP internalization dynamics, can be employed to get an insight into the delivery of ZnO_T_PLL into MCF-7 cells for 72 hours.
Next, we present the simultaneous measurement of ROS in single cells at different times. Figure 8(c),(d) shows the intensity of MitoSOX dye and corresponding spatial intensity mapping in single cells. The results show that the ROS formation was rather sporadic up to 12 hours of treatment, owing to the lower amount of NPs present in the cells (Figure 8(c) and (d) ). A modest increase in ROS formation was observed after 24 hours of treatment. In contrast, at 36 hour, the abundance of red spots indicates a significant increase in ROS production likely due to accumulation of ZnO_T_PLL NPs in the cytoplasm (Figure 8(c) and (d) ). The result clearly shows that The NP internalization and ROS accumulation at 72 hours corresponds to complete destruction in cell morphology as a result of cell death(Figure 8(vi)). . From the above results, it can be concluded that the use of laser scanning helped us to gain a better insight not only in deciphering the internalization dynamics but also to identify the spatial distribution of particles inside single cells. Such spatial distribution study in single cells indicates a strong interdependence between ZnO_T_PLL particle retention and ROS generation in NPs.
3.4 Time course of cell viability for 72 hours: In order to assess the correlation between NP internalization, ROS formation, and cell death in a heterogeneous cell population, a live-dead imaging assay was conducted at 40 µg/mL concentration at the same time points, i.e., 0, 12, 24, 36 and 72 hours (Figure 9 ). As evident fromFigure 9 , early time points up to 12 h (cell viability of 78.41%) show fewer cells with PI uptake, attributed to the lower uptake of NPs. The results show that following 36 h, the percentage of cells with PI uptake increases and gets associated with the disruption of membrane integrity (viability=16.78%). Hence, it can be concluded that significant toxicity (cell viability= 5.02%) towards tumor cells can be induced at 48 h through the application of 40µg/mL ZnO_T_PLL particles (Figure 9 ). Such ROS-mediated cell death can be attributed due to the capability of ZnO NPs in inducing ROS like hydroxyl and superoxide radicals attributed to their semiconductor properties and to perturb electronic transfer processes in the cell (Ancona et al., 2018)).
3.5 Quantification of NP internalization, ROS generation, and cell viability: Since the imaging results indicated the interdependence between ZnO_T_PLL particle retention and ROS generation, we further performed quantification of NP retention, ROS generation, and cell viability dynamics. The time course of particle retention and percentage increase in ROS has been depicted in Figure 10 (a) and (b) . The percentage cell viability of MCF-7 cells with respect to time has been depicted in Figure 10(c) . The quantification of cellular internalization, ROS accumulation, and cell viability was performed based on numerous images taken simultaneously at a single time point. The result suggests a sigmoidal response for internalization of ZnO_T_PLL NPs with time. The results also indicate that it is possible to develop an in vitro assay for assessing the efficacy of the nanoformulation as an anticancer agent through the proposed imaging set up.
Additionally, we performed the fitting of various functions and found that a pseudo-first-order kinetic model can be fitted to the internalization of particles with time (Figure 10(a) ). A positive Pearson’s correlation value between percentage particle internalization and ROS generated (r= 0.942) shows that there is a strong correlation between ROS generation and particle internalization. On the other hand, negative Pearson’s correlation values between percentage particle internalization and cell viability (r = -0.9838) and ROS generation and cell viability (r = -0.98175) indicate the synergy between long term particle retention and toxicity towards tumor cells. The time-dependent analysis of intracellular events indicates a possible mechanism of apoptosis through mitochondrial ROS generation in cells when invaded by ZnO_T_PLL NPs. A schematic of these intracellular events has been shown in Figure 10(d) . Overall, it can be concluded that the confocal microscopy assisted investigation of mitochondrial ROS production inside the MCF-7 cells can also be used for deciphering mechanisms underlying the intracellular events.
4. Discussion
Fluorescent nanoformulations have been an attractive research topic in the field of cancer research both for imaging and therapeutic purposes. Specifically, ZnO NPs are one of the potential anticancer agents against MCF-7 cells and can be used for treatment of breast cancer (Sadhukhan et al., 2019; Hong et al., 2011; Ma et al., 2015; Sureshkumar et al., 2017; Gupta et al., 2015; Kavithaa et al., 2016; Salari et al., 2020; Boroumand Moghaddam et al., 2017; Farasat et al., 2020; Lestari et al., 2018; Wahab et al., 2014). However, quantitative imaging of spatial distribution of NPs in living cells over days was not investigated much. A detailed study on cell-NP interaction using time-lapse microscopy assumes importance in assessing the toxicity as well as therapeutic potential. In this paper, we performed optimization of the NP synthesis to achieve the fluorescence that can be used for 3D imaging for three days in cell culture medium. In order to achieve this, Tween-80 assisted fluorescent ZnO NPs were coated with PLL. The fluorescence in this case might be attributed to the chemically bonded oxygen molecules of the Tween-80. These act as scavengers of the photogenerated electrons and transfer them to deep traps. These electrons upon recombination with trapped holes may result in a recombination centre for visible emission (Khan et al., 2010). When characterized, there was a disparity in the aerodynamic and hydrodynamic diameter of the particles, which could be accounted to the fact that ZnO being an amphoteric oxide, undergoes hydrolysis in water. This results in the formation of a hydroxide coating, ultimately leading to physical adsorption of water molecules on the surface (Wang et al., 2017).
As evident from the increased hydrodynamic diameter in ZnO_T_PLL NPs, it can be concluded that PLL coating increased the hydrophilic nature of the NPs. Hanley et al. demonstrated the preferential uptake of ZnO NPs in cancerous cells compared to healthy T cells (Hanley et al., 2008). It has also been hypothesized that the hydrophilicity of the synthesized particles may remain beneficial for passive targeting owing to its ability to escape macrophage capture (Allahverdiyev et al., 2018). Although future experiments need to be conducted, it can be expected that ZnO_T_PLL NPs being hydrophilic may overcome the threat of macrophage capture generally found in lipid and polymeric particles.
There is limited information on hydrolysis of zinc-oxide inside lysosomes due to the acidic pH of cancer cells during long-term imaging ( ; Xia et al., 2008) . Although further imaging experiments need to be performed to investigate on this matter using lysosomal staining (Dong et al., 2018; Xia et al., 2008), in this work, we performed the measurement of zeta potential to check the stability of the synthesized zinc oxide material over a range of pH . Our results showed that the PLL coating leads to an improved range of zeta potential which provides a roadmap to synthesize a reasonably stable particle. At the same time the zeta potential remains in the favorable range for the attachment of the ZnO particles on the surface of MCF-7 cells. Poly-L-lysine being a cationic polymer, it exerts a layer of positive surface charge. However, the stability of the particle can be further improved by coating the fluorescent ZnO with a thicker PLL layer by altering the PLL concentration while coating the particles. This may lead to a tailored particle that is stable as well as suitable for faster attachment to the cell surface.
Our current study focuses on imaging of MCF-7 cells at 40 µg/mL of ZnO_T_PLL upto 72 hours. However, a future work is rather needed for a detailed study on retention dynamics at various concentrations of ZnO NPs. Since it has been found that the fluorescent ZnO nanostructures in the range 100-400 nm induce preferential killing in MCF-7 cells compared to L929 cells at 40 µg/mL, we chose to perform the 3D imaging study at this concentration. Additionally, a co-culture study was performed to show that normal T cells can be maintained at 85% viability in the range of 10-40 µg/mL concentration of FITC-tagged fluorescent ZnO particles, whereas cancerous T cells were found to be less than 10% viable at 40 µg/mL NP concentration (Hanley et al., 2008).
Also, there are multiple studies on non-fluorescent ZnO particles for which the viability of normal cells (HBL100 and MCF10A) were found to be significantly higher than MCF-7 cells in presence of ZnO nanostructure (Kavithaa et al., 2016; Farasat et al., 2020). It has been shown that 60% viability of HBL100 cells can be retained within 48 hour using ZnO nanorod of size range from 70 nm to 140 nm at a concentration of 10-20 µg/mL (Kavithaa et al., 2016). Based on the concentration range and particle size of the existing ZnO NP experiments (Table S1), here we choose 40 µg/mL to track the NP distribution and ROS generation while maintaining lower cytotoxicity to normal cells. Although, a comparison of cell viability between MCF-7 and L929 in presence of the synthesized NPs shows a preferential killing of the MCF-7 breast cancer cell line (Supplementary Figure S8 ), further optimization can be done for minimization of toxicity towards healthy cells.
Although generation of ROS is taken as a mechanism for inducing toxicity in cancer cells, there is limited data on spatiotemporal ROS dynamics induced by ZnO particles in MCF-7 cells (Gupta et al., 2015). Previously, Gupta et al. (Gupta et al., 2015) has shown the evidence of ROS generation with varying FITC tagged ZnO NP concentration at 24 hr, but the correlation between time course of particle internalization and ROS formation was not evident.
One of the limitations of the current study is that it focuses on monitoring of cell viability in individual cultures of MCF-7 and L929 cells. However, a comparison cell-NP interaction in a co-cultured MCF-7 and L929 cells will be required for gaining insight on the preferential internalization of these particles in the MCF-7 cells. Similarly, the formation of ROS generated as a result of oxidative phosphorylation can be monitored in normal cells as well as tumor cells inside the co-culture model. In order to understand the molecular interaction, a mechanistic model can be obtained for simultaneous prediction of internalization, ROS and viability.
Previously, it has been shown that fluorescent ZnO acts as a promising therapeutic agent for reducing the tumor volume from 1.2cc to 0.6cc in 15 days (breast tumor) in rat models (Hong et al., 2015). Very recently, non-fluorescent ZnO has been tested previously on rat models and has been seen to successfully reduce breast cancer tumor volume (Tanino et al., 2020; . Since the proposed work forms a solid basis for choosing ZnO particles as a potential candidate for treating breast cancer through in vitro imaging assay, we propose that the PLL coated ZnO NPs can further be tested in vivo in a rat model. Some of the recent investigations have revealed that ZnO can be effective as anti-proliferative agent and inhibition agent against various viral strains such as H1N1 and simplex viruses (Ghaffari et al., 2019; Tavakoli et al., 2018; Abdul et al., 2020; Faten & Ibrahim, 2018). Hence, future studies can also be conducted on the effectiveness of ZnO_T_PLL as an antiviral agent.