Nanoparticles for Bioapplications : Study of the Cytotoxicity of Water Dispersible Quantum Dots

Semiconductor nanocrystals or quantum dots (QDs), have unique optical and physical properties that make them potential imaging tools in biological and medical applications. However, concerns such as the aqueous dispersivity, toxicity to cells and stability in biological environments may limit the use of QDs in bioapplications. Here, we report an investigation into the cytotoxicity of aqueously dispersed CdSe(S) and CdSe(S)/ZnO core/shell QDs in the presence of human colorectal carcinoma cells (HCT-116) and a human skin fibroblast cell line (WS-1). The cytotoxicity of the precursor solutions used in the synthesis of the CdSe(S) QDs was also determined in the presence of HCT-116 cells and compared to that of the heat-shock protein (Hsp90) inhibitor, 17-AAG. CdSe(S) QDs were found to have a low toxicity at concentrations up to 100 μg/ml, with a decreased cell viability at higher concentrations, indicating a highly dose-dependent response. Meanwhile, CdSe(S)/ZnO core/shell QDs exhibited lower toxicity than uncoated QDs at higher concentrations. Confocal microscopy images of HCT-116 cells after incubation with CdSe(S) and CdSe(S)/ZnO QDs showed that the cells were stable in aqueous concentrations of 100 μg of QDs per ml, with no sign of cell necrosis, confirming the cytotoxicity data.


Introduction
Semiconductor nanocrystals or quantum dots (QDs) have received a great deal of attention over the last decade due to their unique optical and physical properties.This has led to them being classified as a powerful new class of bio-imaging tools [1][2][3][4][5][6][7][8].Despite the desirability of using QDs as intense fluorescent nanoscale markers in biological applications, cytotoxicity is a serious constraint that limits potential uses [9][10][11].QDs can be considered as being potentially toxic due to both their nanoscale size and the presence of heavy metals [10,11].
Nanoscale particles can have a greater toxicity than bulk counterparts, with the small size enabling greater penetration into cells than corresponding bulk chemical materials [12].In addition, the generation of free radicals such as reactive oxygenated species (ROS) upon cell exposure to nanomaterials, is a significant cause of nanoparticle cytotoxicity [12][13][14][15].
Interactions of QDs with cell mitochondria and nuclei can result in the disruption of cell function, inhibition of cell proliferation and decreased cell viability due to ROS production, ultimately resulting in cell death, mutation, or induced immunotoxicity [13,[16][17][18].The longterm toxicity of QDs has also been attributed to bioaccumulation in organs, leading to organ damage and chronic illnesses [19].The high ratio of reactive surface area relative to the bulk generally increases the chemical activity of nanoparticles, which can result in an accumulation of nanoparticles in tissues and associated organ toxicity [20], particularly those having high blood flow such as spleen, kidneys, liver and lungs, along with the blood circulation system itself [21].QDs have therefore largely been classified as poisonous to both human and animal cells and as such, cytotoxicity assays of QDs are an essential requirement before applying QDs to cellular environments.
There are various methods to determine the toxicity of QDs in both in-vivo and in-vitro studies.In-vivo studies may involve introducing QDs to microorganisms [22] or the use of animal models [23][24][25][26][27], whereas in-vitro toxicity features the treatment of various cell types with QDs in order to investigate the cytotoxicity of nanocrystals in mammalian cell lines [28][29][30][31][32][33][34][35][36][37].A common feature in all reports of cadmium-based QDs, including CdTe, CdSe and CdS, is that Cd is a primary source of toxicity.Accordingly, core/shell QDs such as CdTe/CdS, CdTe/CdS/ZnS and CdSe/ZnS have been shown to exhibit less toxicity than core QDs alone [11,31].However, the cytotoxicity of QDs depends on a number of parameters including the surface modifications of the QDs, cell type, cellular morphology, cell growth and the interaction of QDs with cell membranes [36,30].
There are extensive reports detailing cytotoxicity assays of QDs synthesized in organic media [25,[29][30][31][32][33][34][35][36][37][38][39], but few papers have reported in detail the cytotoxicity of QDs directly obtained from aqueous solution [11,[40][41][42][43]. Bhatia and co-workers investigated the in-vitro cytotoxicity of organically synthesized CdSe QDs on liver cells and showed that the release of free cadmium ions from the QDs resulted in cell death [30].Meanwhile, Zhu et al. studied both the in-vivo and in-vitro cytotoxicity of aqueous CdSe and CdSe/CdS QDs and reported that the toxicity of QDs depends on both target cells and physicochemical properties of the QDs [42].Plank et al. investigated the cytotoxicity of CdSe and CdSe/ZnS QDs obtained in organic solvents, demonstrating that QD cytotoxicity is related to both particle size and the surface covering of functional groups such as amines and carboxylic acids used to disperse them in water [32].Fan et al. investigated the cytotoxicity of a wide range of QDs (CdTe, CdTe/CdS and CdTe/CdS/ZnS) synthesized in aqueous reactions, concluding that free cadmium ions are a major source of toxicity in CdTe-based QDs [11], also reporting that CdTe/CdS/ZnS QDs were slightly less toxic than CdTe in their experiments [11].The present work reports on the cytotoxicity of CdSe(S) and CdSe(S)/ZnO QDs, synthesized in wholly aqueous reactions, to human colorectal carcinoma cells (HCT-116) and human skin fibroblast cells (WS1).The water dispersible QDs were synthesized in a modified literature method [44] and the cytotoxicity of both the QDs and the precursor solutions were determined after incubation of HCT-116 cells with the QDs over a wide range of concentrations (25 to 500 µg/ml).Cytotoxicity of CdSe(S) QDs was benchmarked against the heat-shock protein (Hsp90) inhibitor, 17-AAG, and confocal images of HCT-116 cells after treatment with QDs were recorded.Cytotoxicity of QDs to WS1 cells was also studied to explore the action of QDs toward normal, resilient cell types.

Synthesis of QDs:
Water dispersible CdSe(S) and CdSe(S)/ZnO QDs were synthesized and characterized using optimized experimental parameters in a modified literature method [44], as detailed in Supplementary Data S1.

Characterization of QDs: UV-vis absorption spectra were measured with a Varian
Cary UV spectrometer.Photoluminescence spectra were measured on a Carry Eclipse fluorometer using an excitation wavelength of 350 nm.X-ray powder diffraction patterns (PXRD) were recorded on an X'pert PRO Multi-purpose X-ray diffraction system (MPD system) with a Cu  source ( = 0.154056 nm).X-ray photoelectron spectroscopy (XPS) was performed using an Escalab 250Xi spectrometer and a monochromated Al Kα X-ray source (hν = 1486.6eV) operated at 10 kV and 10 mA.High resolution transmission electron micrographs (HRTEM) were obtained using a Philips CM200 instrument.

Preparation of aqueous solutions of QDs:
A total of twenty different solutions of CdSe(S) and CdSe(S)/ZnO QDs, Cd-MPA (Sample 1, Supplementary Data S1) and Cd-Se-MPA (Sample 2, Supplementary Data S1) were prepared by diluting the aqueous solutions with ultra-pure water to achieve concentrations of 25, 50, 100, 250, 500 (µg/ml).Each sample was used in cytotoxicity assays without further purification.In order to remove excess cadmium ions and capping agent from the aqueous solution, a Slide-A-Lyzer dialysis cassette was used in the QD stock solution and the cytotoxicity of dialyzed CdSe(S) QDs in the presence of HCT-116 cell line was investigated in five samples at concentrations of 25, 50, 100, 250, 500 µg/ml.

Cytotoxicity assays:
Cell cultures of human HCT-116 and WS-1 cell lines were obtained according to the standard protocol [45], as detailed in Supplementary Data S2.The cells were then separately seeded in two 96-well plates (3000 cells/well) and allowed to adhere to the dish for 24 hours in a humidified incubator.As a control, both the cell media alone (3000 cell/well) and cell media in the presence of Hsp90 inhibitor 17-AAG (100 nM = 0.06 µg/ml) were also seeded.Finally, 10 µl of each aqueous solution was added to the plates.The plates were incubated in the presence of the QDs for 72 hours at 37ºC with 5% CO2, after which the samples were analysed to determine the cell proliferation using a Cell Counting Kit-8 assay.

Confocal microscopy studies: first two samples of either fixed or live HCT-116 cells-
QDs were prepared (Supplementary Data S3).Images of the cells in the presence of QDs were recorded under Leica TCS SP5 CW STED and Zias LSM 780 confocal microscopes, respectively.The live cells were stained by adding a Hoechst solution (5 µg /ml) 10 minutes before recording images, whilst the fixed cells were stained using DAPI.In addition, the emission profiles of QDs in both cell media and aqueous solution were recorded using a Zias LSM 780 confocal microscope.

Results and discussion
3.1 Synthesis of QDs: the method of synthesising QDs can play an important role in the overall toxicity, with literature reports indicating that QDs synthesized in organic solvents tend to be more toxic than those synthesized by aqueous pathways [46][47][48].However, the toxicity of Cd-based QDs has largely been attributed to the free Cd ions existing in equilibrium with the QDs in solution [11,30,49].Coating the QD cores with appropriate shell materials may prevent the core from oxidation, thereby reducing the number of free Cd ions released [42,[50][51][52][53][54].In this work, an aqueous hydrothermal method reported by Aldeek and co-workers [44] was used to synthesise CdSe(S) and CdSe(S)/ZnO QDs using optimized experimental parameters, leading to the formation of highly crystalline QDs.
PXRD patterns of the synthesized QDs showed that both CdSe(S) and CdSe(S)/ZnO QD cores had the cubic zinc blende structure (Figure 1), according to the standard CdSe cubic pattern [55].The diffraction peaks of the QDs were found to be consistent with standard patterns of the cubic phases of CdSe [55] and CdS [56] and are characteristic of alloyed CdSe-CdS, Table S1, Supplementary data.As ZnO is amorphous, there was no peak observed in PXRD that can be attributed to ZnO.Meanwhile, the cubic structure of CdSe(S) QDs remained intact after coating with an amorphous ZnO shell, in accord with a previous report [44].The size of the core was estimated using the Scherrer equation as 3.6±0.1 and 3.0±0.1 nm in CdSe(S) QDs and CdSe(S)/ZnO QDs, respectively.This indicates a potential decrease in the size of the QD core during the heating process in the formation of the zinc oxide shell.The decreased particle size of the CdSe(S) QD core after heating using reflux is   The obtained QDs were found to be highly luminescent, with narrow emission peaks at 560 and 550 nm for CdSe(S) and CdSe(S)/ZnO QDs, along with broad excitation bands starting at 530 nm for CdSe(S) and 520 nm for CdSe(S)/ZnO QDs, respectively (Figure 3).The optical properties of the QDs are related wholly to the CdSe(S) QD cores, with the ZnO present as an amorphous shell, exhibiting no optical properties in this region.The emission wavelength of the CdSe(S) QDs (λ = 560 nm) (particle size = 3.6±0.1 nm) was found to be slightly different to that of CdSe QDs (λ = 515 nm) (particle size = 2.5±0.5 nm) reported previously [44], consistent with both the larger particle size and modified Cd:MPA molar ratio used.Moreover, the emission wavelengths of CdSe(S)/ZnO QDs (λ = 550 nm) were found to shift to lower wavelengths relative to the cores CdSe(S) QDs (λ = 560 nm), in agreement with the estimated particle size (3.0±0.1 nm) but in contrast to a previous report [44].This is primarily due to the sensitivity of QDs on changes in environmental or experimental parameters.For example, Aldeek and co-workers [44] reported that the emission wavelengths of CdSe(S)/ZnO QDs shifted to lower wavelengths compared to uncoated CdSe(S) QDs after 4 minutes of UV exposure [44], indicating that even small changes in temperature or illumination can lead to shifts in emission and excitation wavelengths of the QDs.
The XPS spectra of both CdSe(S) and CdSe(S)/ZnO QDs (Figure 4), confirmed the existence of the peaks assigned to selenium, cadmium, sulfur and carbon, indicating that the shell growth does not influence the structure of CdSe(S) QD cores.The presence of sulfur was confirmed with the appearance of peaks characteristic of S2p at 161.8 eV in XPS data.It is also in agreement with previous reports in the literature where MPA releases sulfur at high temperature [58,59], resulting in sulfur from the 3-mercaptopropionic acid (MPA) participating the growth of CdSe(S) particles.Indeed, XPS data of CdSe(S)/ ZnO QDs confirmed the existence of both Zn and O in the core/shell CdSe(S)/ZnO QDs, with the observation of a new peak related to Zn2p at 1022 eV (Figure 4e), indicating the formation of a ZnO shell around the CdSe(S) cores; this is consistent with the standard X-ray photo electron spectrum of zinc oxide [60] and in accord with previous work [44].
XPS analysis was also used as quantitative method to determine elemental composition.The atomic percentages of as-prepared QDs, have been summarized in Table 1.This confirmed that CdSe(S) QDs were coated with a ZnO shell.The CdSe(S) QDs contained cadmium (13.7%), selenium (0.7%), sulfur (13.1%), oxygen (29.9%) and carbon (42.5%), with the ratio of atomic percentages C:O:S = 3.2:2.3:1,fully consistent with the molecular composition of the capping agent MPA, C3O2SH6.The atomic percentage of Cd was found to be approximately equal to the sum of the atomic percentages of the Se and S contributions, indicating that at the surface of the CdSe QDs, the MPA coordinates by substitution of S at the Se sites.The low intensity of the Se peak suggests a low atomic percentage of Se relative to that of S (1:18.2) and is due to the dominance of surface atoms in the XPS data, essentially MPA and the outer surface of the CdSe(S) QDs.Besides, according to quantitative analysis data (Table 1), the most abundant element in the CdSe(S)/ZnO QD spectrum was O (37.7 atomic %).Each MPA molecule accounts for two O and one S atom; the atomic % of S was found to be 12.3% and so MPA accounts for 24.6% of the O contribution, leaving 13.1%, which closely matches the 12.0% contribution of Zn.The Cd:(Se+S) ratio of 1:0.8 and Cd:Zn  Table 1.Elemental analysis of QDs.

Cytotoxicity assays:
The response of HCT-116 and WS1 cell lines upon exposure to the QDs was studied in order to investigate toxicity.Cancer cells such as HCT-116 cells have an irregular DNA pattern, making them more sensitive than healthy cells to free heavy metals, including cadmium.In contrast, normal skin cells (WS1) are known to be some of the most resistant cell types to free metal ions.In addition, 50% of cancer cells, including HCT-116, depend upon a heat-shock protein (such as Hsp90) to survive [61].The viability of these cancer cells decreases in the presence of Hsp90 inhibitor 17-AAG, which is a common antitumour compound [62,63].The viability of HCT-116 cells was therefore investigated both in the presence of 17-AAG and after treatment of the cells with QDs.

Cytotoxicity of CdSe(S) QDs toward HCT-116 cells:
The lethal concentration corresponding to the death of 50% of cells (LC50) was determined as 105 µg/ml for the HCT-116 cell line upon exposure to CdSe(S) QDs.CdSe(S) QDs were found to have a low toxicity at the highest dilutions studied (25, and 50 µg/ml), with 100.0±0.2%,91.5±2.6% cell viability respectively, but have a cell mortality rate of ≥50% at concentrations of ~100 µg/ml and beyond, with a dose-dependent cytotoxicity (Figure 5).AAG is an anti-tumour drug that inhibits the activity of Hsp90, a protein necessary for the growth of the cells [62,63].The cytotoxicity results showed that CdSe(S) QDs inhibit the cell viability by 64.4±0.5% at a concentration of 500 µg/ml; a similar value to that of 17-AAG (66.4±2.7%) at a concentration of only 0.06 µg/ml (Figure 5).Hence, CdSe(S) QDs are much less cytotoxic than Hsp90 inhibitor 17-AAG, presumably as a result of non-specific cell activity.

Cytotoxicity of precursor solutions:
The CdSe(S) QDs were synthesised by hydrothermal reactions of two precursor solutions containing Cd-MPA and Cd-Se-MPA.The viability of HCT-116 cells was measured to both of the precursor solutions (Figure 5), which were found to have no significant toxicity at concentrations of 25 and 50 µg/ml, but to be cytotoxic at a concentration 500 µg/ml, similar to that of CdSe(S) QDs.The Cd-MPA solution was found to be more toxic than the Cd-MPA-Se precursor solution.In addition, cells treated with the Cd-MPA precursor solution had a lower viability (67.3±9.8%) at 50 µg/ml than the Cd-Se-MPA precursor (97.0±8.3%),indicating that formation of the Cd-Se-MPA complex decreases the toxicity, presumably due to a decreased number of free cadmium ions.

Cytotoxicity of dialyzed CdSe(S) QDs:
A dialysis cassette was used to remove excess capping agent and cadmium ions from solution, however the results showed that the CdSe(S) QD solutions were in fact more toxic after dialysis than before (Figure 6).This is consistent with a report of CdTe QDs with thiol capping agents, in which a surface cadmium-thiol shell protected the QDs against oxidation, resulting in greater stability of the QDs [64].In aqueous solution, QDs are in an active equilibrium with excess cations and thiol in solution.During dialysis, excess thiol and cadmium ions are removed from solution, resulting in the destabilisation of the QD surface, thereby leaching more free cadmium ions into solution and so increasing toxicity.43,50,52,53,64].The toxicity of core/shell CdSe(S)/ZnO QDs towards HCT-116 cancer cells indicated that the core/shell QDs exhibited low toxicity at all concentrations studied.As shown in Figure 7, the viability of the cells was determined to lie between 72.5±1.0%and 56.9±1.0%across the concentration range of 25 to 500 µg/ml, indicating that the LC50 of the cells is not reached even at a concentration of 500 µg/ml, the highest concentration used in this series of experiments.Clearly the shell inhibits release of free cadmium ions, limiting the cell death.The emission profiles of QDs in both the cell media and water were recorded using a confocal microscope in order to determine any changes in photoluminescence spectra of QDs in cell media (Figure 10a).The photoluminescence emission of CdSe(S) QDs were found with no change in cell media, in contrast to literature reports which have indicated that water soluble MPA-capped CdTe QDs had an altered emission profile in cell growth media [65].
Emission spectra indicated that CdSe(S) QDs have a maximum emission at 548 nm in the cell media, with no significant change or shift in photoluminescence relative to that in water.The emission of CdSe(S)/ZnO QDs was found to have a lower intensity and shift to 546 nm in cell media from 556 nm in water (Figure 10 b).

Conclusions
We have shown that the cytotoxicity of QDs can be controlled with the aqueous synthesis of stable QDs.It was determined that coating of QDs with a ZnO shell protects the core against oxidation and the production of toxic free radicals, resulting in decreased cytotoxicity of QDs, even at high concentrations.Images of cells after incubation with QDs at a concentration of 100 µg/ml indicated that the cells remained viable, confirming the cytotoxicity data.The stability of the QD cores in the cell media was found to be related to the overall toxicity of QDs, which is largely governed by the free Cd-ion concentration.

S4. The comparison of PXRD of CdSe(S) QDs with cubic CdSe and CdS standard peaks
The diffraction peaks of CdSe(S) QDs were observed between standard cubic CdSe [3] and CdS [4], as shown in Table S1.
our previous studies on CdSe nanoparticles[57].However, the total size of CdSe(S)/ZnO QDs cannot be estimated by using the Scherrer equation because the ZnO shell is amorphous.

of 1 :
0.7 are fully consistent with a ZnO shell around the CdSe(S) core.In addition, the peaks of Cd and Se shift with the incorporation of ZnO, highlighting the modified bonding environments at the interfacial atoms.

Figure 5 .
Figure 5.The results of cytotoxicity assays of CdSe(S) QDs, Cd-MPA and Cd-Se-MPA precursors toward HCT-116 cell line.Error bars indicate standard error of the mean and the concentration of 17-AAG was 0.06 µg/ml

Figure 6 .
Figure 6.The results of cytotoxicity assays of dialyzed CdSe(S) QDs.Error bars indicate standard error of the mean and the concentration of 17-AAG was 0.06 µg/ml.

Figure 7 .
Figure 7.The results of cytotoxicity assays of CdSe(S)/ZnO core/shell QDs toward HCT-116 cell line.Error bars indicate standard error of the mean and the concentration of 17-AAG was 0.06 µg/ml

Figure 8 .Figure 9 .
Figure 8.The results of cytotoxicity assays of CdSe(S) and CdSe(S)/ZnO QDs toward WS1 cell line.Error bars indicate standard error of the mean and the concentration of 17-AAG was 0.06 µg/ml

Figure 10 .
Figure 10.Photoluminescence spectra of CdSe(S) QDs (A) and CdSe(S)/ZnO QDs (B): (a) in water and (b) in cell media.The excitation wavelength of the instrument was adjusted at 405 nm in time of measuring PL.