Green synthesis of silver and iron oxide nanoparticles mediated photothermal effects on Blastocystis hominis

The spectral absorption of the explored metal nanoparticles was measured using a double-beam UV-Vis spectrophotometer (Cary 5000, Agilent Technologies, Santa Clara, CA, USA). Their morphology was imaged using a high-resolution transmission electron microscope (HRTEM, Tecnai, G20, FEI, Almelo, Netherlands) operating at an accelerating voltage of 200 kV. The drops of dilute prepared nanomaterial solutions were deposited on a carbon-coated copper grid and left to dry at room temperature before the examination by TEM. In addition, the electrokinetic potential (zeta potential) was measured with a zeta sizer analyzer (Nano ZS, Malvern Instruments, Malvern, UK) based on an electrophoretic light scattering technique.

Inoculum size was initially assessed by counting the number of live parasites using a Neubauer cell counting chamber after staining with a 0.4% Trypan blue solution (a viability indicator). A parasite inoculum of size 40×10⁴ parasites/mL from culture during the logarithm phase was considered suitable for assessment. After calculation of the inoculum volume required, the parasites were introduced into a set of culture tubes containing LE medium and three different concentrations of either Ag NPs or Fe₃O₄ NPs to evaluate the parasite sensitivity in the dark condition. Similarly, 2 mL of each parasitic suspension was seeded in each well of a 6-well plate with a flat bottom. The cells were photo-irradiated with broad-band visible light (400–700 nm) using an LED lamp as a light source (10 W, Wellmax, China) [41]. Culture tubes were used in triplicate for every concentration of each nanoparticle and the non-treated control. The summary of the groups used in the present work is displayed in Table 1.

The cell concentration and viability of Blastocystis cysts within each analyzed group were assessed by an acridine orange/propidium iodide (AO/PI) double staining mixture. This staining dye can quantify the morphological apoptotic changes in the treated cells and, hence, differentiate live cells from dead ones by utilizing a fluorescence microscope [42]. The AO dye can penetrate the cell membranes of viable and dead cells, staining their nuclear DNA and exhibiting green fluorescence under blue excitation [43], while the PI dye penetrates only decomposed cell membranes and stains the denatured DNA emitting red fluorescence under green excitation [43, 44]. After 24 h of incubation at 37 °C, counting the number of viable cells was done the same way as before. Subsequently, each group was washed three times with phosphate buffer (PBS), PH = 8, and then centrifuged. The supernatant was discarded, and the pellet was redispersed in 100 µL of PBS containing 10 µL of an equal volume AO/PI mixture. Then, the B. hominis cells were incubated with the mixture for 30 min in the dark. About 10 µL of each sample was examined by a ×63 confocal laser scanning microscope (CLSM, LSM 710, Carl Zeiss, Germany). The images were collected at excitation wavelengths of 458 nm for AO and 514 nm for PI. Under the excitation of 458 nm, the green fluorescence of AO emitted by dead cells was absorbed by PI generating weak green emission.

Automatic cell segmentation and clustering of the fluorescence images is a challenge. For example, morphological watershed-based algorithms were employed to separate different clustered nuclei [45, 46]. After weighting each pixel concerning its near boundaries’ curvature information, a distance transform and watershed method was used for the extraction of nuclei [47]. Nuclei segmentation from confocal images was performed employing gradient flow tracking and locally adaptive thresholding [48]. Besides, automatic threshold selection methods for the CLSM images of biofilms were established [49, 50].

In this study, the color features of the live and dead cells based on superpixels and k-means clustering were utilized to segment and classify the live and dead cells. Different superpixel approaches have been suggested and applied in computer vision algorithms as a key intermediate representation of an image to be broken into perceptually meaningful superpixel regions. As a result, segmentation and classification can be done over regions instead of the full image, providing high-quality results. An efficient superpixel approach should be fast and simple, maintain memory size and computation accuracy, and adapt well to image boundaries [51‐54].

Since the proposed method for the extraction of live and dead cells was based on their color features, the collected CLSM images were converted into the CIELAB colorspace. The presumed reason is that the CIELAB is a perceptually uniform colorspace that depicts all visible colors to the human eye [50, 51]. It comprises three components, \({l}^{*}\), \({a}^{*}\), and \({b}^{*}\). Whereas \({l}^{*}\) represents the luminance, while \({a}^{*}\) and \({b}^{*}\), respectively, define the amounts of red-green and yellow-blue tones. After such a preprocessing step, the fast superpixel approach called adaptive simple linear iterative clustering (ASLIC) [51] was applied to partition the image into superpixel regions employing the k-means clustering method for a fixed number of iterations. Then, we started with an initial estimate of the number of superpixels, k. Since the k-means method tries to minimize the average square distance between pixels in the same cluster [55], the distance measure, D, was defined as follows [51]:where \({(x}_{i},{y}_{i})\) and \(({x}_{j},{y}_{j})\) are the coordinate values of pixels \(i\) and \(j\), respectively. \({d}_{c}\) and \({d}_{s}\) are the CIELAB space and Euclidean distances, respectively. S is the average distance between two cluster centers, \(N\) is the total number of image pixels, and C is the compactness that weighs the relationship between the color and spatial distances. So by iteration, the ASLIC algorithm updates the values of S and C for each cluster, utilizing the maximum color and spatial distances from the preceding iteration [51]. Then, the mean color of the intensity distribution within each superpixel region was computed to quantify the superpixel regions and generate a label image. Further analysis by applying another k-means clustering to the label image was performed, by which the CLSM images were accurately segmented into three clusters, live cells, dead cells, and background.

Data were reported as mean ± standard deviation (SD) for quantitative variables of triplicate determinations, and comparisons between the different means of the studied groups were done using the Student t-test to assess the statistical significance. The values of p < 0.05 were considered statistically significant.

Silver nanoparticles were synthesized by reducing silver nitrate salts with non-toxic and biodegradable chitosan at given temperatures. The typical UV-Vis absorption spectrum of the resulting solutions is shown in Fig. 1a. The plot displays that the characteristic surface plasmon resonance (SPR) band is centered at about 425 nm. This band represents the formation of silver nanoparticles. The appearance of a yellowish-brown color also supported the formation of Ag nanoparticles. Such results are consistent with the previous reports on the fabrication of gold nanoparticles with chitosan as both a stabilizing and reducing agent [56]. The TEM result indicates the spherical shape of Ag NPs with a particle size in the range of 10 nm, distributed homogeneously in the Cs matrix (Fig. 1b). The average surface zeta potential of Ag coated with Cs was approximately 36.7 mV, as shown in Fig. 1c. This value can greatly influence their stability to form a stable solution in water using electrostatic repulsion between the particles. In addition, it facilitated its absorption by a cellular membrane and made them good candidates to be used for therapy and imaging while coated with Cs.

The successful synthesis of the magnetite nanoparticles was confirmed by the UV-Vis spectrum (Fig. 2a). The solution had a dark background color with a broad absorption band from UV to NIR. A representative TEM image of polymer-coated magnetite nanoparticles is shown in Fig. 2b. It displays the production of a large quantity of nearly uniform monodispersed nanospheres of iron oxide. The size of the magnetite was about 10 nm. The surface of magnetite nanoparticles was mostly negatively charged due to coating with PEG, as observed from zeta potential measurements (Fig. 2c). The average surface potential was −6 mV. This charge value enabled it to form a stable solution in the water. Such a finding was related to the small crystallite sizes of the magnetite particles. The superparamagnetic properties of iron oxide nanoparticles make them good candidates to be used in many biomedical applications [18, 40].

Although some studies reported the efficiency of PTT on pathogenic organisms, no previous studies represented the photothermal effect of LED combined with metal nanoparticles on B. hominis. Our design was prepared to obtain a synergetic, enhanced antiparasitic effect. This occurred by combining the intrinsic antiparasitic activity of metallic nanoparticles with their photothermal effect after irradiation with the LED (400–700 nm) to induce local hyperthermia that caused cell death. Consequently, the synergistic effect of metal nanoparticles (Ag NPs or Fe₃O₄ NPs) with the LED (400–700 nm) irradiation was able to induce an anti-blastocystis effect.

The accurate mechanisms postulated for the antipathogenic effects of metal NPs are still being studied. The mechanism suggested for the metal nanoparticles combined with photothermal activation might be related to the enhanced release of ions (Ag⁺ or Fe₃O₄⁺) from Ag NPs or Fe₃O₄ NPs, respectively, by rising temperature. Moreover, the permeability of the parasitic cell membrane might be disrupted by these ions. Furthermore, many studies have detected that NP force induces holes and gaps in the pathogenic organism’s cell membrane, which leads to cell fragmentation as well as cell death [58]. Otherwise, there was a possibility of oxidative stress occurring through the generation of reactive oxygen species (ROS) on the surface of NPs. Eventually, more damage occurs in the parasitic cell, leading to cell death [61].

To investigate the antiparasitic mechanism of the Ag NPs, Ag NPs, Fe₃O₄ NPs, and Fe₃O₄ NPs combined with LED, the B. hominis cyst viability was assessed using CLSM. The two fluorescent nucleic acid dyes, acridine orange and propidium iodide, were used to stain the B. hominis cells. The stained B. hominis cells were visualized under the CLSM to assess the viability and membrane integrity of the parasite both in the dark and after photothermal activation. Then, a method based on the color features of the live and dead cells employing superpixels and k-means clustering was proposed to segment and classify the live and dead cells (see the “Image processing for cell extraction” section). A stepwise example showing the segmentation process is presented in Fig. 4. The results of the extracted live cells and dead cells, as well as histograms stating the percentage of cells showing the proportion of apoptosis, are presented in Fig. 5.

It can be observed that the control group exhibited almost entirely green-colored cells stained with AC, which could be internalized in the live cells. On the other side, the cells treated in the dark with Ag NPs or Fe₃O₄ NPs exhibited the presence of yellow and red cells with higher percentages than the green ones. Interestingly, the cells treated with photothermally active Ag NPs or Fe₃O₄ NPs displayed more red cells, up to 68% and 98%, respectively, in comparison to the dark-treated cells group. The appearance of yellow or red colors is related to early and late apoptosis after staining with the PI dye. This reveals the presence of many dead cells with a damaged membrane that led to the entry of the PI stain into cells [43]. It is depicted that the results reveal the robustness of the proposed method. This robustness can be noticed in the detection of the cells in the very low-quality CLSM image captured from “Fe₃O₄ NPs.”