Open Access
Add to BibSonomy
Abstract
Yokukansan (YKS) is a traditional Japanese herbal medicine that is increasingly being studied for its effects on neurodegenerative diseases. In our study, we presented a novel methodology for a multimodal analysis of the effects of YKS on nerve cells. The measurements of 3D refractive index distribution and its changes performed by holographic tomography were supported with an investigation by Raman micro-spectroscopy and fluorescence microscopy to gather complementary morphological and chemical information about cells and YKS influence. It was shown that at the concentrations tested, YKS inhibits proliferation, possibly involving reactive oxygen species. Also substantial changes in the cell RI after few hours of YKS exposure were detected, followed by longer-term changes in cell lipid composition and chromatin state.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Traditional herbal medicines are naturally existing, plant-derived substances with minimal or no industrial processing that have been used to treat diseases as part of local or regional medicinal practices and are gaining significant attention in global health debates. Therapeutic herbs and their derived phytochemicals are increasingly being recognized as useful complementary cancer treatments [1] or as therapies for the long-term treatment of diseases associated with aging [2]. A number of clinical trials have demonstrated the beneficial effects of herbal medicines on survival, immune system modulation and quality of life in cancer patients when these herbal medicines are used in combination with conventional treatments [1]. Some herbal extracts that result in increased rates of tissue regeneration have also found applications in both stem cell therapy and tissue engineering [3]. Research into the clinical application of herbal extracts is still in progress due to the high variability and complexity of the bioactive components, so work on standardizing procedures for the preparation of herbal products will increase the possibility of their use in this context.
One of the most investigated herbal medicine in modern research is Yokukansan (YKS). This traditional herbal substance is considered a novel alternative treatment for many neurological diseases. The primary use is to treat diseases associated with aging, such as neurodegenerative disorders. YKS is used to alleviate behavioral and psychological symptoms in Alzheimer’s disease [4], and can also be used as a treatment for Parkinson’s disease [5]. YKS is considered to have neuroprotective [6], antidepressant and analgesic effects [7]. Modern studies have been carried out mainly on brain cells. It has been proven, that YKS maintains neuronal survival and function through a number of beneficial effects, including anti-apoptosis, antioxidation, endoplasmic reticulum stress and neurogenesis [8,9]. In research at cellular level multiple measurement techniques and imaging methods have been used to study Chinese medicine herbs. Immunohistochemical research is carried out using fluorescence or confocal microscopy [10,11]. Degradation tests of cerebral tissue and nerve cells are commonly performed using electron microscopy [12–14] and various Raman spectroscopy techniques have been applied for among others the fingerprinting [15]. However, most of these methods have rather qualitative nature. In order to overcome this limitation, quantitative phase imaging (QPI) techniques have also been applied to study Chinese herbal medicines [16]. Despite of all these efforts, there have been very few works that tried to combine the results obtained from different modalities.
Therefore in this paper we decided to conduct a first (to our best knowledge) multimodal analysis of YKS. However at this stage the three selected modalities (holographic tomography, fluorescence microsopy and Raman microspectroscopy) were implemented through separate (not combined through a common field of view) systems. The main measurement tool is holographic tomography (HT). Unlike the QPI methods that have been used so far in YKS analysis, the HT system provides quantitative and fully three-dimensional distribution of the refractive index. In most of the up-to-date research where HT is used, the central cross-section of HT 3D reconstructions has been analyzed only (with noteworthy exceptions including [17–20]). Here, we use the full 3D information to volumetrically analyze the influence of YKS on biological cells and their substructures. By combining high-quality HT reconstruction algorithms with state-of-the-art 3D segmentation solutions we obtain high accuracy of the calculated average refractive index. The second modality is Raman microspectroscopy (RmS), which is a vibrational spectroscopic technique, complementary to infrared spectroscopy, which provides bond-specific chemical information of the tested specimen in a non-invasive label-free manner. Finally, the results are correlated with fluorescence microscopy (FM) to determine the effect of YKS on cell proliferation. The cells under study are SH-SY5Y neuroblastoma cells. We have chosen these cells, as they are often used as in vitro models of neuronal function and differentiation [21,22]. They are adrenergic in phenotype but also express dopaminergic markers and, as such, have been used to study Parkinson’s disease, neurogenesis, and other characteristics of brain cells. The mulitimodality of our study gives a new perspective for the analysis of the effects of YKS on cells and enables identification of changes at the biophysical level.
In Section 2. we describe our methodology of the research, where we describe all systems. In addition, we have present extensively the procedure for preparing cell cultures and YKS herbal medicine which enables standardization of experiment preparation in the future. In Section 3. we show our data analysis methods for all modalities and in Section 4. we included the results and conclusions.
2. Methods
In this section, the complete scheme of multimodal methodology and workflow is introduced. Firstly, the workflow of the whole experiment is given. Then, details about measurement systems are provided. Finally, procedures for cell culturing and for preparation of the herbal compound are described.
2.1 Multimodal measurement methodology
In this paper, we investigate the effect of different doses of YKS on SH-SY5Y neuroblastoma cells. Specifically, we focus on the impact of the YKS on: (1) proliferation of the cells, (2) cellular stress which is expressed by the presence of lipid droplets in the cytoplasm, (3) RI changes within the cell and (4) chemical composition of the cell substructures.
Three microscopic systems are used to perform this experiment. The full workflow is shown in Fig. 1. In the first step, we use fluorescence microscopy together with Hoechst 33342 staining to examine the effects of multiple YKS doses on SH-SY5Y cells proliferation. Based on these results, we select 3 doses to be used in further analyses with holographic tomography and Raman micro-spectroscopy. For the second part of the research we used a HT system. It provides information about 3D morphology of cells based on the 3D distribution of the refractive index (with direct access to 2D RI distribution from any cross-section). From these results we retrieve information about volume of lipid droplets and about RI distribution of cellular substructures. Finally, the RmS is used, where the measurement is performed by scanning a sample with a focused laser light and recording the consecutive spectra. The YKS solution for all experiments was prepared using the same procedure.
Fig. 1. Workflow of the experiment. Fluorescence microscopy (FM) used for proliferation check, with quick measurements with Hoechst 33342 staining and for further dosage selection. Holographic tomography (HT) and Raman micro-spectroscopy (RmS) for characterisation of refractive index (RI) changes and biochemical changes for examination of YKS-induced cellular stress and nucleus and nucleolus changes under YKS influence.
2.2 Microscopic systems
The FM used in this study is DMi8 fluorescent wide-field inverted microscope (Fig. 2(A)) equipped with 10x/0.30 HC PL FLUOTAR objective, DFC360 FX CCD camera and a motorized stage (Leica Microsystems, Germany). Single-plane images of the whole well area were scanned in DAPI channel (EX350/50, EM460/50) and stitched together using Tile Scan module in LAS X software (Leica Microsystems, Germany). Theoretical lateral and axial resolution was 0.12 µm and 0.3 µm, respectively [23], and sampling varies between 0.135 µm and 0.155 µm, depending on the ROI selected at the beginning of the reconstruction process. The system was metrologically tested by measuring a dedicated phantom, imitating a cell, printed using the two-photon method [24,25].
Fig. 2. Measurement systems: A. fluorescence microscopy (EmF - emission filter; DBS - dichroic beam splitter; MO - microscope objective; S - sample; ExF - excitation filter); B. holographic tomography (TL1, TL2 - tube lenses; MO1, MO2 - microscope objectives; S - sample) with scanning based on galvanometer mirror (GM); C. Raman micro-spectroscopy (M1, M2 - mirrors; DM - dichroic mirror; ND neutral density filter; MS - motorized stage; S - sample).
For the three-dimensional analysis of RI distribution, a HT system based on an off-axis Mach-Zehnder interferometer was used (Fig. 2(B)). 90 optical projections of the sample were captured with 632 nm laser illumination. Detailed description of the system is given in [24]. The Gerchberg-Papoulis algorithm with finite object support regularization was used to reconstruct the 3D RI distribution of the analysed samples from the measured projections [20] (see Sec. 3.2.1).
For biochemical analysis a Raman micro-spectroscopy system in the mapping configuration (Fig. 2(C)) was utilized (inVia confocal Raman microscope). In this commercial setup consisting of bright-field microscope combined with a Raman spectrometer, one may choose the region of interest and plane of focus, for which the Raman spectroscopic measurement is then performed. The measured spectrum is obtained from a limited volume around the plane of focus. To highlight the methodology of Raman micro-spectroscopic measurement and for scheme clarity, the Fig. 2(C) represents the Raman measurement modality only. A Nd:YAG laser operating at 532nm was used as the excitation source. The system is equipped with a 60x 1NA water dipping microscope objective and a 2400 l/mm diffraction grating. The resulting spectra were recorded from 643 cm−1 to 1807 cm−1 with 1 cm−1/pixel resolution.
2.3 Cell culture and treatment
The SH-SY5Y cell line is the subline of the neuroblastoma SH-N-SH cell line. The cells were cultured in DMEM/F12 (1:1) medium (PAN Biotech, P04-41450) supplemented with $1\%$ Pen Strep (Gibco, 15140-122), $2$mM L-glutamine (Gibco, 25030-024), $10\%$ heat-inactivated Fetal Bovine Serum (Gibco, 10500-064). Cells were grown under standard conditions ($37^\circ C$, $5\%$ CO2, $90\%$ humidity). YKS stock solution ($40$mg/ml) was prepared in phosphate buffered saline (PBS) pH $7.4$, as follows. $1.6$g YKS powder was added to $40$ml PBS and incubated for 1h at 37 °C. The solution was centrifuged under $3000$RPM for $20$ minutes. The supernatant was subsequently filtered using 0.22 ml cellulose acetate syringe filter into sterile eppendorfs and the aliquots were stored at −20 °C. For the experiments, the cells were incubated with specified concentrations of YKS, by replacing part of the medium with YKS solution. For the control cells the equivalent volume of the medium was replaced with pure PBS. As we aim for multimodal measurements, the cell culture growing conditions remain the same, but preparation of the cells for measurement itself were different.
For proliferation assessment with FM, the cells were seeded onto 96-well TC-treated culture plate at 20 000 cells/well and incubated with and without YKS solution. After the incubation, cells were live-stained with Hoechst 33342 dye for 30 minutes at 5 µg/ml. For HT analysis, the cells were seeded onto ibidi 35mm Petri dish at 200 000 cells/dish density and measured directly in the nutrient medium with refractive index $n=1.338$. Finally, for RmS measurements, sterilized Raman grade CaF2 glass was placed into 35mm TC dish filled with medium and 200 000 cell were seeded into the dish. CaF2 glass is a commonly used material for Raman measurements as it does not introduce signal in the biologically-relevant spectral range.
3. Data analysis methods
3.1 Fluorescence microscopy data analysis
The data obtained through FM measurements were fully analyzed with LAS X software (Leica Microsystems, Germany). From the stitched images of Hoechst 33342 staining, full wells were segmented and the information generated in the software about the average intensity of the well was used. Hoechst 33342 dye stains the nuclei, thus by analyzing the intensity changes inside the whole well we are able to compare the average number of cells and draw conclusions which doses of YKS increase proliferation and which inhibit this process.
3.2 HT data analysis
3.2.1 HT reconstruction method
Captured off-axis holograms were processed with Fourier transform method to retrieve complex amplitudes. Then, Gerchberg-Papoulis algorithm with finite object support was used for the tomographic reconstruction. The full description of this technique can be found in [20]. The algorithm is a 2-step method. In the first step, a tomographic reconstruction with strong total-variation regularization is calculated. It results in a 3D reconstruction with geometry that is free from missing-cone artifacts, however with erroneous internal RI distribution. This result is then binarized, so that only information about external geometry is kept. In the second step, iterative Gerchberg-Papoulis method is applied in which the Fourier transform of the reconstruction and the inverse Fourier transform of the spectrum are calculated iteratively and 2 constraints are applied in each domain: nonnegativity constraint and finite object support which is the result of the first step of the method. The figure shows a 3D visualisation of the RI reconstruction (Fig. 3(A)), and a cross-section through the best focus plane (Fig. 3(B)).
Fig. 3. A. 3D tomographic reconstruction with magnified regions showing different structures withing cell volume; B. 2D cross-section from reconstruction with the best focus plane.
3.2.2 Segmentation and cells RI determination
In our work we used 3D watershed algorithm [26,27], which was applied to volume segmentation of 3D holographic tomography reconstructions of biological cells and for extracting 2D segmentation of the whole cells from chosen cross-section. Before segmentation, the tomographic data were preprocessed: reconstructions were thresholded, then morphological filtration (consecutive opening and closing operation) [28] was performed and artifacts were removed. After preprocessing, markers for background and biological objects were defined for the 3D watershed algorithm and then superimposed on the preprocessed reconstruction. The creation of background markers was performed in several following steps: binarization of the preprocessed reconstruction, dilation operation [28] and hole filling. In the next step, the watershed transform was calculated and a cell mask was created by assigning a value of 1 to all specified labels (example of 3D mask on Fig. 4(A)).
Fig. 4. A. Cross-section through 3D mask. B. Chosen manually 2D cross-section with the focus plane with masked nucleolus (yellow) and lipid droplets (blue) and borders of the whole cells (red).
The 2D segmentation for calculation of the mean refractive index of nucleolus was performed based on the 3D segmentation described above, by selecting the desired focus plane on the basis of the visibility and contrast of the cell nuclei (Fig. 4(B)). Cell nucleolus are highly homogeneous organelles compared to the whole cell, so use of one cross-section is reasonable. Further, by using the segmentation method from predefined Graph Cut function of the Matlab ImageSegmenter application the cells nucleolus were segmented (Fig. 4(B)). The lipid droplets (LDs) for analysis were detected in the 3D segmented images using a simple thresholding technique [29] and a dilation operation [28] and then the ratio of the volume of lipid droplets to the volume of cells in the field of view was calculated.
3.3 Raman micro-spectroscopy data analysis
3.3.1 Preprocessing
Firstly, the procedure called cosmic-spike removal was performed. Cosmic spikes are visible in the Raman spectrum as the abnormal spikes of very high intensity. The removal was performed by recognizing the spike and its position and replacing them with data from the neighbouring pixel. The second preprocessing step is baseline removal, whose goal is to suppress the fluorescence signal that is present in Raman measurements. This step is performed using the adaptive iteratively reweighted penalized least squares technique [30].
3.3.2 Cluster analysis
Clustering techniques are a group of unsupervised machine learning methods which are used for explorative analysis. One of these techniques is the additive hierarchical cluster analysis (HCA) [31], in which the pixels’ spectra are at first treated as single clusters, but are then connected in a step by step manner, based on their similarity. This method was applied to the masked hyperspectral data cubes of single cells. As a result, the dendrogram, which represents the differentiation of the single pixel’s spectra among defined distance measure, is created.
The next step is to visualise the results, by creating an image with the specified cluster number, which gives information about the spatial distribution of biochemical components in the measured cell. These components can be identified by the mean spectra of clusters, which include the peaks that are characteristic for specific chemical bonds.
3.3.3 Analysis of the chemical compartments of cells
First step in RmS analysis was calculating spectra of imaging media and YKS itself. Later for main analysis of the RmS data, we first focused on determining the relevant clusters that would correspond to the internal cellular structures: lipids, cytoplasm and nucleic acids. In order to find representative spectra for chosen structures, first the clusters obtained with one representative SH-SY5Y cell were compared in terms of line intensities. In Fig. 5 the lines characteristic for lipids were marked with red, nucleic acids were marked with green, proteins with blue, and carbohydrates with pink. Based on literature information about cells spectra [32] and knowing the visual structure of the cells, it can be concluded that the fist cluster (yellow) is connected to cytoplasm, second one (blue) is mainly connected to cell membrane, third cluster (pink) represents lipid droplets (=CH2 twist peak, C=O stretching peak), while the fourth one (green) indicates the positions in which high nucleic acids contribution is visible and its connected to nucleus.
Fig. 5. A. Clusters spectra for one reference SH-SY5Y cell. Lines characteristic for lipids were marked with red (717 cm−1 - CN+(CH3)3 str. - trimethylamine,1301 cm−1 - =CH2 twist, 1736 cm−1-1750 cm−1 - C=O - carbonyl group), nucleic acids were marked with green (788 cm−1 - O-P-O stretching DNA, 1578 cm−1 - guanine, adenine), proteins with blue (1005 cm−1 - Sym. Ring Phe - symmetric ring phenylalanine, 1033 cm−1- C-H in-plane Phe - C-H in place phenylalanine) and carbohydrates in pink (1420 cm−1-1480 cm−1 - CH def - CH deformation) - continuous line for points, dashed line for range [32]; B. cluster image of chosen cell.
4. Results and discussion
The summary of all measurements is shown in Table 1. Measurements taken with 3 systems for all YKS doses were accompanied by control measurements where no YKS was added. The difference in the number of analyzed cells depends on FOVs of the systems and difference in the duration of one measurement and potential ease of performing measurements in multiple wells at the same time. In FM measurements, there were 20000 cells per dose, because a measurement with stitching of multiple FOVs was used and the entire well was measured (stitched $FOV_{FM} = {60}\;\textrm{mm}^{2}$). In HT measurements, on average, 1-5 cells could be measured per seeded confluence in the field of view ($FOV_{HT} = \sim {100}\;\mathrm{\mu}\textrm{m} \times {100}\;\mathrm{\mu}\textrm{m}$), and in RmS due to long acquisition time, one cell per measurement was chosen. Additionally, due to the time-consuming nature of HT and RmS measurements, a limited number of cells and doses were analyzed. Only 3 doses were selected for HT, which were chosen as smaller representation of doses from FM measurements and then only one dose was selected for RmS, due to the very long measurement time (one cell measurement lasting $\sim 60$min). All results are available in Dataset 1 [33].
Table 1. Summary of all measurements taken with fluorescence microscope, holographic tomography system and Raman micro-spectroscopy system.
4.1 Influence of different doses of YKS on the proliferation of SH-SY5Y
We started with proliferation test, where we treated cells with multiple dose to recognize which dose has the strongest effect on the process and to chose doses for further experiments. For that reason, we conducted fluorescence microscopy studies. We treated cells with 9 doses (including control group) of YKS and measured the fluorescence signal at 4 time points. For a given dose and for a given time, there were 9 fluorescence measurements conducted (except for a few missing measurements due to technical errors in measurement). The averaged results are shown in Fig. 6. To determine if there is significant influence of YKS dose on the proliferation, we performed statistical analysis. First, we checked the normality of the fluorescence results calculated for a specific dose and a specific time with Shapiro-Wilk test [34]. The significance level was set to 5%. Out of all 37 datasets (9 doses x 4 time points + 1 additional control measurement without PBS), only in 4 cases there was enough evidence to accept alternative hypothesis that the distribution is not normal. Since the assumption of data normality is not severely violated, we performed a one-way repeated measures ANOVA test [35] with YKS dose as the independent variable. Obtained results showed that the YKS dose lead to statistically significant differences in proliferation ($F(8,32)=8.5088, p < 0.001$).
Fig. 6. Values of mean intensity of Hoechst 33342 staining in cell cultures with 0,5, 1, 2, 4, 5, 6, 7, 8 mg/ml doses of YKS measured after 0h, 24h, 48, 72h and 6days.
In Table 2 the proliferation doubling time is shown. It can be observed that, with increasing doses, cells divide more slowly, and at doses of 6-8mg/ml, the software calculated doubling time > 600h, which can be interpreted as complete lack of proliferation.
Table 2. Calculated proliferation doubling time for different YKS doses.
Doses $0,5, 1, 2$mg/ml inhibit proliferation to some extent and the doubling time is substantially longer comparing to dose $0$mg/ml, where only PBS was added. At doses of $6, 7, 8$mg/ml proliferation is fully inhibited and there are fewer cells after 6 days compared to measurements after 72h, which may suggest cell death.
Comparing the results to the study by Nakatani et al. and Tanaka et al. [9,36] where the effect of YKS on increasing the proliferation of neuroblastoma B65 was proven, our studies show that chosen doses inhibit proliferation from the 0.5mg/ml dose and it is also shown that doses above $5$mg/ml are already detrimental to proliferation. Based on these results we chose three doses for further research - $1$mg/ml, $4$mg/ml and $8$mg/ml dose.
4.2 Influence of different doses of YKS on the whole cell and nucleolus
Next, we analyzed the results of holographic measurements of cells under the influence of different doses of YKS herbal medicine ($1$mg/ml, $4$mg/ml and $8$mg/ml) and at different time points (2h, 24h, 48h after treatment). For a given dose and for a given time, there were approximately 28 RI measurements conducted. RI of the whole cell, and RI of the cell nucleolus were investigated separately. Figure 7 presents the averaged results. Analyzing the results from this Figure, one can observe whole cell RI changes as well as nucleolus RI changes in response to different YKS dose. Observation of the changes in whole cell RI shows (Fig. 7 (left)) the average RI increases with increasing YKS dose. Comparing to control measurement, RI decrease in case of 1mg/ml and 4mg/ml doses, and is at a similar level after 8mg/ml dose. Additionally, the mean RI values for doses of $4$mg/ml and $8$mg/ml are similar after 2h and 48h (RI for $4$mg/ml after 2h is 1.3529, after 48h is $1.3530$, RI for $8$mg/ml after 2h is $1.3544$, after 48h is $1.3546$). In the case of the $1$mg/ml dose, the refractive index after 48h decreases by $\Delta$RI= $0.0012$ in comparison to control measurement.
Fig. 7. Comparison of volume refractive index of the whole cell and nucleolus under the influence of different doses of YKS herbal medicine ($1$mg/ml, $4$mg/ml and $8$mg/ml) and at different time points (2h, 24h, 48h after treatment).
Observation of the mean RI of nucleolus shows similar tendencies of changes as observed in whole volume RI of the whole cells (Fig. 7 (right)), but the value of the nucleolus RI is higher (difference in control measurement is $\Delta$RI= 0.0106). For measurements after $4$mg/ml and $8$mg/ml values of the RI for all time points are similar. The values of RI after $1$mg/ml dose varies between time points (there is big increase jump 24h after treatment). Values after YKS are lower than control group in case of dose $1$mg/ml and $4$mg/ml. Decrease of refractive index was previously observed by another scientific group as a result of water accumulation in the cell, which suggests that adding YKS changes the percentage of intracellular water already as soon as 2 hours of incubation [37].
However, to see if these results are not due to randomness, we conducted a statistical analysis, similar to the one performed in Section 4.1. First, we analyzed if the sample distributions for a given time and a given dose are normal. For the measurements of the RI of the whole cell, out of 10 datasets (3 doses x 3 time points + 1 control measurement), there was enough evidence to accept alternative hypothesis that the distribution is not normal in 1 dataset only. In the case of RI measurement of the cell nucleoli, all 10 datasets can be assumed as coming from a normal distribution. Since the assumption of data normality is not severely violated in both cases, we performed a one-way repeated measures ANOVA tests with YKS dose as the independent variable. For RI of the whole cell, the influence of the YKS dose on the RI is not statistically significant for significance level $\alpha =5\%$ but is significant for $\alpha =10\%$ ($F(2,4)=5.8873, p=0.0643$). The same conclusion can be drawn from results obtained for RI of the cell nucleoli ($F(2,4)=4.3259, p=0.1000$). This result can be due to low number of time points and in the future, more time points should be taken into account.
Analyzing the averaged Raman spectra for the cluster associated with the nucleus (Fig. 8), we can see changes with nucleobases. In the measurement after 48h there are less of uracil, cytosine, thymine and guanine and adenine than in the control measurement and after 24h, what can be observed on Fig. 8. In addition at the same time observing changes in RI in the nucleolus, RI in these structures is approximately the same. It was previously shown that Raman signals associated with nucleic acids reflect the degree of chromatin compaction. Therefore, we suspect that YKS treatment promotes more open architecture of chromatin [38]. However, due to very small number of measurements, we cannot perform statistical analysis of the results, and thus, our conslusions should be treated as a hypothesis.
Fig. 8. Comparison of mean spectra for cluster connected to nucleus with characteristic for nucleic acids lines (green) (667 cm−1= - thymine, guanine; 729 cm−1 - adenine, 782 cm−1 - uracil, cytosine, thymine, 788 cm−1 - O-P-O stretching DNA, 811 cm−1- O-P-O stretching RNA, 1578 cm−1 - guanine, adenine)
4.3 Lipid droplets and stress granules as a stress symptom after treating cells with high doses of YKS
High-RI structures in a cell are usually either lipid droplets (LD) or stress granules (SG). They can usually be distinguished by their shape. LDs are uniformly round, SGs are irregular in shape [39]. Both structures are related to each other. Studies indicate that the pathways that induce lipid droplet accumulation, influence and mutually regulate each other through the formation of stress granules (SGs), which are a common adaptive response to various stresses. The inability to upregulate lipid droplets reduces the formation of stress granules. Stress granule formation, in turn, drives lipid droplet clustering and fatty acid accumulation [40].
Based on RI, we cannot clearly determine whether the high-RI structures in cells are only LDs or only SGs. For confirmation, other technique validation with higher specificity are required to distinguish these two structures [41]. Nevertheless based on HT measurements and RI analysis, we can observe LDs and SGs changes in volume in whole cells. With our results, based on 3D RI distribution, we focused on comparing the changes of average LD and SG volume per single cell (Fig. 9). As the dose increases, we notice an increase in the volume of LD and SG in a cell. The largest increase compared to a control measurement can be seen at a dose of $8$mg/ml and an increase in lipid structures in cells can be associated with cellular stress [39]. However, from performed statistical analysis we can conclude, that we do not have enough evidence to confirm these observations. After conducting Shapiro-Wilk tests for normal distribution of the samples (out of 10 datasets only 2 cannot be assumed normal), we performed one-way repeated measures ANOVA which resulted in $F(2,6)=3.3322, p=0.1063$. In order to prove the effect of YKS on lipid droplets, more measurements should be captured in the future.
Fig. 9. Comparison of mean volume of lipid droplets per single cell in all doses and times used for the experiment with holographic tomography.
In order to chemically confirm the presence of LDs and SGs in cells and the association of these structures with high-RI substructures in HT measurements, RmS measurements were performed, in which we observe that clusters corresponding to the location of high-RI structures in the spectrum are associated with lipids, because of changes in the lines characteristic only for them (Fig. 10 - lines 1736 cm−1-1750 cm−1 connected to carbonyl group). In addition, it can be observed that in the spectrum comparison in the cluster associated with LDs and SGs, there are more lipids with carbonyl group in the 24h and 48h measurements, but fewer lipids with CN+(CH2)3 compared to the control measurement. Also, in the cluster associated with cytoplasm area (Fig. 11), we observe significantly lower signal in tyrosinase, tryptophan proteins line in the measurement after 48h, and also with in line connected to carbonyl and alkene group. The LDs are present in most cells and can accompany with cell death. Under physiological conditions, they are few. In neurons, they accompany with neurodegenerative diseases and aging, but their role is not clear [42,43]. The observed increase of LDs volume in cells is proportional to increasing concentrations of YKS which is in line with the results of proliferation assay. It supports the thesis that exposure to the tested concentrations of YKS is toxic for SH-SY5Y cells. Analyzing the results from HT and RmS measurements, we can also conclude that YKS changes the lipid composition of LDs and SGs. There is an increase in the proportion of the carbonyl group (C=O), and this indicated that there are increased levels of ROS (reactive oxygen species) and oxidative stress in the cell. Increased levels of ROS in turn oxidise lipids (hence there are more oxygen groups) and RCS (reactive carbonyl species) may be formed [44,45]. In light of these results we speculate that the inhibition of SH-SY5Y cell growth by YKS exposure might involve excessive production of reactive oxygen species and oxidative stress.
Fig. 10. Comparison of mean spectra in cluster 3 connected to LDs and SGs with characteristic lines for lipids (717 cm−1 - CN+(CH3)3 str. - trimethylamine, 1301 cm−1 - =CH2 twist, 1736 cm−1-1750 cm−1 - carbonyl group).
Fig. 11. Comparison of mean spectra in cluster 1 connected to cytoplasm area with characteristic lines for proteins (blue - 1005 cm−1 - symmetric ring phenylalanine, 1617 cm−1 - tyrosinase, tryptophan) and lipids/carbohydrates (pink - 1060 cm−1, 1655 cm−1 - carbonyl group, alkene group stretching)
5. Conclusions
We demonstrated an approach to analyze effect of YKS on SH-SY5Y cells using 3 imaging modalities that provide complementary information. In terms of proliferation, we have shown and statistically proved that YKS dose has observable impact on this process. We also presented a methodology to analyze influence of YKS on RI of whole cells and cells’ nucleoli as well as influence on volume of the lipid droplets. In these cases, no statistically significant effect was observed, which may be due to relatively small number of measured samples. Nevertheless, we have shown that by combining results from Fluorescence Microscopy, Holographic Tomography and Raman micro-spectroscopy, we can draw unique conclusions about the analyzed biological samples. This multimodal approach allow for better interpretation of HT results, by complementing them with chemical information from RmS measurements as well as provide better understanding of YKS influence on cells. Additional studies, including wider range of concentrations could reveal YKS potential as neuroprotective and/or anti-proliferative drug. Also, the specific biochemical pathways involved and precise time-course of the process require further research.
Funding
Narodowe Centrum Badań i Rozwoju (PLTW/V/5/2018); Fundacja na rzecz Nauki Polskiej (POIR.04.04.00-00-16ED/18-00).
Acknowledgments
The research leading to the described results was carried out within the program of Foundation for Polish Science, co-financed by the European Union under the European Regional Development Fund and Polish-Taiwanese Joint Research Project (PLTW/V/5/2018) financed by the National Centre for Research and Development and supported project from Team-Net program (POIR.04.04.00-00-16ED/18-00) of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund. Fluorescence microscopy experiments were performed at the Laboratory of Imaging Tissue Structure and Function which serves as an imaging core facility at the Nencki Institute of Experimental Biology.
CRediT statement. MB: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft. MKr: Investigation, Data Curation, Formal analysis, Visualization, Writing – original draft. MM: Formal analysis, Data Curation, Visualization, Writing – original draft. NN: Investigation, Writing – Review & Editing, Resources. JS: Formal analysis, Validation, Resources, Writing – Review & Editing. WK: Formal analysis, Software, Visualization, Writing – Review & Editing. C-JC: Conceptualization, Resources. MKu: Conceptualization, Funding acquisition, Project administration, Supervision.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Detailed results presented in this paper are available in Dataset 1 cited in [33].
References
1. S.-Y. Yin, W.-C. Wei, F.-Y. Jian, and N.-S. Yang, “Therapeutic applications of herbal medicines for cancer patients,” J. Evidence-Based Complementary Altern. Med. 2013, 1 (2013). [CrossRef]
2. Y. Liu, W. Weng, G. Rui, and L. Yue, “New insights for cellular and molecular mechanisms of aging and aging-related diseases: Herbal medicine as potential therapeutic approach,” Oxid. Med. Cell. Longevity 2019, 1–25 (2019). [CrossRef]
3. V. L. Udalamaththa, C. D. Jayasinghe, and P. V. Udagama, “Potential role of herbal remedies in stem cell therapy: proliferation and differentiation of human mesenchymal stromal cells,” Stem Cell Res. & Ther. 7(1), 110 (2016). [CrossRef]
4. O. Hideki, Y. Masaomi, U. Keigo, H. Cheolsun, H. Yoshiro, and N. Takao, “Yokukan-san: a review of the evidence for use of this kampo herbal formula in dementia and psychiatric conditions,” Neuropsychiatr. Dis. Treat. 10, 1727–1742 (2014). [CrossRef]
5. J. Jang, K. Jung, J. Kim, I. Jung, H. Yoo, and C. Moon, “Potential application of Yokukansan as a remedy for Parkinson’s disease,” Front. Pharmacol. 2018, 1–19 (2018). [CrossRef]
6. Z. Kawakami, H. Kanno, T. Ueki, K. Terawaki, M. Tabuchi, Y. Ikarashi, and Y. Kase, “Neuroprotective effects of Yokukansan, a traditional Japanese medicine, on glutamate-mediated excitotoxicity in cultured cells,” Neuroscience 159(4), 1397–1407 (2009). [CrossRef]
7. Y. Suzuki, H. Mitrsuhata, M. Yuzurihara, and Y. Kase, “Antiallodynic effect of herbal medicine Yokukansan on peripheral neuropathy in rats with chronic constriction injury,” Evid.-based Complement. Altern. Med. 2012, 1–8 (2012). [CrossRef]
8. K. Mizoguchi and Y. Ikarashi, “Cellular pharmacological effects of the traditional Japanese kampo medicine Yokukansan on brain cells,” Front. Pharmacol. 8, 1 (2017). [CrossRef]
9. Y. Nakatani, T. Amano, H. Yamamoto, N. Sakai, M. Tsuji, and H. Takeda, “Yokukansan enhances the proliferation of b65 neuroblastoma,” J. Tradit. Complementary Medicine 7(1), 34–44 (2017). [CrossRef]
10. S. Shimizu, T. Tanaka, T. Takeda, M. Tohyama, and S. Miata, “The kampo medicine Yokukansan decreases microrna-18 expression and recovers glucocorticoid receptors protein expression in the hypothalamus of stressed mice,” BioMed Res. Int. 2015, 1–8 (2015). [CrossRef]
11. T. Ueki, Y. Ikarashi, Z. Kawakami, K. Mitsutoshi, and Y. Kase, “Promotive effects of Yokukansan, a traditional Japanese medicine, on proliferation and differentiation of cultured mouse cortical oligodendrocytes,” Pharmacol. & Pharm. 05(07), 670–680 (2014). [CrossRef]
12. Y. Ikarashi, S. Iizuka, S. Imamura, T. Yamaguchi, K. Sekiguchi, H. Kanno, Z. Kawakami, M. Yuzurihara, Y. Kase, and S. Takeda, “Effects of Yokukansan, a traditional Japanese medicine, on memory disturbance and behavioral and psychological symptoms of dementia in thiamine-deficient rats,” Biol. & Pharmaceutical Bulletin 32(10), 1701–1709 (2009). [CrossRef]
13. K. Azuma, T. Toyama, M. Katano, K. Kajimoto, S. Hayashi, A. Suzuki, H. Tsugane, M. Iinuma, and K.-y. Kubo, “Yokukansan ameliorates hippocampus-dependent learning impairment in senescence-accelerated mouse,” Biol. & Pharmaceutical Bulletin 41(10), 1593–1599 (2018). [CrossRef]
14. S. Iizuka, Z. Kawakami, S. Imamura, T. Yamaguchi, K. Sekiguchi, H. Kanno, T. Ueki, Y. Kase, and Y. Ikarashi, “Electron-microscopic examination of effects of Yokukansan, a traditional Japanese medicine, on degeneration of cerebral cells in thiamine-deficient rats,” Neuropathology 30(5), 524–536 (2010). [CrossRef]
15. C.-C. Huang, “Applications of Raman spectroscopy in herbal medicine,” Appl. Spectrosc. Rev. 51(1), 1–11 (2016). [CrossRef]
16. C.-H. Wu, X.-J. Lai, C.-J. Cheng, Y.-C. Yu, and C.-Y. Chang, “Applications of digital holographic microscopy in therapeutic evaluation of Chinese herbal medicines,” Appl. Opt. 53(27), G192–G197 (2014). [CrossRef]
17. X. Chen, M. E. Kandel, S. He, C. Hu, Y. J. Lee, K. Sullivan, G. Tracy, H. J. Chung, H. J. Kong, M. Anastasio, and G. Popescu, “Artificial confocal microscopy for deep label-free imaging,” Nature Photonics 17, 250–258 (2023). [CrossRef]
18. U. S. Kamilov, I. N. Papadopoulos, M. H. Shoreh, A. Goy, C. Vonesch, M. Unser, and D. Psaltis, “Learning approach to optical tomography,” Optica 2(6), 517–522 (2015). [CrossRef]
19. S. Chowdhury, M. Chen, R. Eckert, D. Ren, F. Wu, N. Repina, and L. Waller, “High-resolution 3D refractive index microscopy of multiple-scattering samples from intensity images,” Optica 6(9), 1211–1219 (2019). [CrossRef]
20. W. Krauze, “Optical diffraction tomography with finite object support for the minimization of missing cone artifacts,” Biomed. Opt. Express 11(4), 1919–1926 (2020). [CrossRef]
21. H. Xicoy, B. Wieringa, and G. J. M. Martens, “The sh-sy5y cell line in Parkinson’s disease research: a systematic review,” Mol. Neurodegener. 12(1), 10 (2017). [CrossRef]
22. J. Kovalevich and D. Langford, Considerations for the Use of SH-SY5Y Neuroblastoma Cells in Neurobiology (Humana Press, 2013), pp. 9–21.
23. V. Lauer, “New approach to optical diffraction tomography yielding a vector equation of diffraction tomography and a novel tomographic microscope,” J. Microsc. 205(2), 165–176 (2002). [CrossRef]
24. M. Ziemczonok, A. Kuś, P. Wasylczyk, and M. Kujawińska, “3D-printed biological cell phantom for testing 3D quantitative phase imaging systems,” Sci. Rep. 9(1), 18872 (2019). [CrossRef]
25. M. Ziemczonok, A. Kuś, and M. Kujawińska, “Optical diffraction tomography meets metrology—measurement accuracy on cellular and subcellular level,” Measurement 195, 111106 (2022). [CrossRef]
26. O. Wirjadi, “Survey of 3D image segmentation methods,” ITWM Rep. pp. 16–17 (2007).
27. P. S. Umesh Adiga and B. B. Chaudhuri, “An efficient method based on watershed and rule-based merging for segmentation of 3-D histo-pathological images,” Pattern Recognit. 34(7), 1449–1458 (2001). [CrossRef]
28. W. Burger and M. J. Burge, “Morphological Filters,” in Principles of Digital Image Processing: Fundamental Techniques, (Springer, 2009), Undergraduate Topics in Computer Science, pp. 157–184.
29. N. Otsu, “A Threshold Selection Method from Gray-Level Histograms,” IEEE Trans. Syst., Man, Cybern. 9(1), 62–66 (1979). [CrossRef]
30. Z.-M. Zhang, S. Chen, and Y.-Z. Liang, “Baseline correction using adaptive iteratively reweighted penalized least squares,” Analyst 135(5), 1138–1146 (2010). [CrossRef]
31. Y.-J. Liu, M. Kyne, C. Wang, and X.-Y. Yu, “Data mining in Raman imaging in a cellular biological system,” Comput. Struct. Biotechnol. J. 18, 2920–2930 (2020). Publisher: Elsevier. [CrossRef]
32. I. Notingher and L. L. Hench, “Raman microspectroscopy: a noninvasive tool for studies of individual living cells in vitro,” Expert Rev. Med. Devices 3(2), 215–234 (2006). [CrossRef]
33. M. Baczewska, M. Królikowska, M. Martyna, N. Nowak, J. Szymanski, and W. Krauze, “Influence of Yokukansan on the refractive index of neuroblastoma cells: dataset,” Zenodo (2022), https://zenodo.org/record/7331040#.Y_Sx4XbML5A.
34. S. Shapiro and M. Wilk, “An analysis of variance test for normality,” Biometrika 52(3-4), 591–611 (1965). [CrossRef]
35. A. Rutherford, ANOVA and ANCOVA: a GLM approach (John Wiley & Sons, 2011).
36. Y. Tanaka and K. Mizoguchi, “Influence of aging on chondroitin sulfate proteoglycan expression and neural stem/progenitor cells in rat brain and improving effects of a herbal medicine, Yokukansan,” Neuroscience 164(3), 1224–1234 (2009). [CrossRef]
37. M. A. Model and E. Schonbrun, “Optical determination of intracellular water in apoptotic cells,” The J. Physiol. 591(23), 5843–5849 (2013). [CrossRef]
38. F.-K. Lu, S. Basu, V. Igras, M. P. Hoang, M. Ji, D. Fu, G. R. Holtom, V. A. Neel, C. W. Freudiger, D. E. Fisher, and S. Xie, “Label-free DNA imaging in vivo with stimulated raman scattering microscopy,” Proc. Natl. Acad. Sci. 112(37), 11624–11629 (2015). [CrossRef]
39. T. Amen and D. Kaganovich, “Stress granules inhibit fatty acid oxidation by modulating mitochondrial permeability,” Cell Rep. 35(11), 109237 (2021). [CrossRef]
40. T. Amen and D. Kaganovich, “Small molecule screen reveals joint regulation of stress granule formation and lipid droplet biogenesis,” Front. Cell Dev. Biol. 8, 1 (2021). [CrossRef]
41. K. Kim, S. Lee, J. Yoon, J. Heo, C. Choi, and Y. Park, “Three-dimensional label-free imaging and quantification of lipid droplets in live hepatocytes,” Sci. Rep. 6(1), 36815 (2016). [CrossRef]
42. X. Han, J. Zhu, X. Zhang, Q. Song, J. Ding, M. Lu, S. Sun, and G. Hu, “Plin4-dependent lipid droplets hamper neuronal mitophagy in the MPTP/P-induced mouse model of Parkinson’s disease,” Front. Neurosci. 12, 1 (2018). [CrossRef]
43. V. Girard, F. Jollivet, O. Knittelfelder, M. Celle, J.-N. Arsac, G. Chatelain, D. M. Van den Brink, T. Baron, A. Shevchenko, R. P. Kühnlein, N. Davoust, and B. Mollereau, “Abnormal accumulation of lipid droplets in neurons induces the conversion of alpha-synuclein to proteolytic resistant forms in a drosophila model of Parkinson’s disease,” PLOS Genet 17(11), e1009921 (2021). [CrossRef]
44. C. A. Juan, J. M. Pérez de la Lastra, F. J. Plou, and E. Pérez-Lebe na, “The chemistry of reactive oxygen species (ROS) revisited: outlining their role in biological macromolecules (DNA, lipids and proteins) and induced pathologies,” Int. J. Mol. Sci. 22(9), 4642 (2021). [CrossRef]
45. H. M. Semchyshyn, “Reactive carbonyl species in vivo: generation and dual biological effects,” Sci. World J. 2014, 1–10 (2014). [CrossRef]
