Open Access
Add to BibSonomy
Abstract
The aim of this study was to clarify the dose- and time-dependent effect of the γ-secretase inhibitor (DAPT) combined with cisplatin on osteosarcoma (OS) cells, evaluated by confocal Raman microspectral imaging (CRMI) technology. The intracellular composition significantly changed after combined drug action compared with the sole cisplatin treatment, proving the synergistic effect of DAPT combined with cisplatin on OS cells. The principal component analysis-linear discriminant analysis revealed the main compositional variations by distinguishing spectral characteristics. K-means cluster and univariate imaging were used to visualize the changes in subcellular morphology and biochemical distribution. The results showed that the increase of the DAPT dose and cisplatin treatment time in the combination treatment induced the division of the nucleus in OS cells, and other organelles also showed significant physiological changes compared with the effect of sole cisplatin treatment. After understanding the cellular response to the combined drug treatment at a molecular level, the achieved results provide an experimental fact for developing suitable individualized tumor treatment protocols.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Osteosarcoma (OS) is an aggressive malignant tumor of the bones, mainly occurring in children and adolescents [1,2]. Current treatments for OS include surgical resection and systemic chemotherapy, and adjuvant chemotherapy and neoadjuvant chemotherapy significantly improved the long-term survival rate of patients [3–5]. However, drug toxicity and drug resistance are still the main factors limiting the cure of this disease [6].
Cisplatin is a well-known chemotherapeutic drug that crosslinks with purine DNA bases, interfering with DNA repair mechanisms, causing DNA damage, and inducing cancer cell apoptosis [7]. Although it has been widely used to treat many human cancers [8,9], high-dose cisplatin is accompanied by serious side effects despite its anti-cancer efficacies, and its long-term use can also induce intrinsic drug resistance [10,11]. Many research groups are trying to improve its efficacy and reduce its toxicity, for example changing its structure or administrating it as a combination therapy with other drugs [8,12].
Recent studies showed that Notch may be involved in the mechanism of cisplatin resistance [13,14]. Notch signaling is a highly conserved pathway in tumor biology, and shows important regulation on cell proliferation, metastasis and differentiation that is associated with tumorigenesis [14–16], such as osteosarcoma progression [14]. One of the main therapeutic targets of the Notch pathway is the Notch receptor, whose activity can be blocked using the γ-secretase inhibitors that inhibit its tumorigenic (intracellular) region [17]. The γ-secretase inhibitor N-[N-(3,5-difluorophenyl acetyl)-L-alanyl]-sphenylglycine butyl ester (DAPT) is a class of small molecules that suppress the cleavage of γ-secretase substrates, thus effectively blocking Notch signaling and consequently exerting a potent anti-tumor effect against several cancers [18,19]. The mechanism used by Notch to regulate the endogenous or acquired resistance of cisplatin is still not clear. Therefore, the clinical response of OS under cisplatin treatment after the inhibition of Notch activity by DAPT need to be investigated from the perspective of cell-drug interactions to evaluate the possibility of targeting Notch using a combined therapy to treat osteosarcoma.
Raman spectroscopy provides a non-invasive and label-free approach to obtain the biochemical fingerprint information of living cells and their response to drugs [20]. The content and conformation of nucleic acids, proteins and lipids in the cell are detected through the position, intensity and line-widths of different spectral bands [21]. Confocal Raman microspectral imaging (CRMI) is a Raman-based spectroscopic imaging technology that explores the subcellular biochemical structure and chemical stress responses caused by chemotherapeutics without changing the cell function [22]. Its potential in monitoring the mechanism of drug action has been confirmed in many clinical practices [23].
Therefore, the aim of this study was to reveal the molecular response of OS cells after cisplatin combined with DAPT treatment using the CRMI methodology. The spectral characteristics of the cisplatin control group (sole cisplatin treatment) and the experimental group (DAPT combined with cisplatin treatment) were compared, and the principal component analysis-linear discriminant analysis (PCA-LDA) was used to extract the molecular variations in OS cells induced by cisplatin and by its combination with DAPT. The multivariate spectral imaging method, K-means cluster analysis (KCA), was used to visualize the subcellular morphological and compositional characteristics under different conditions (changes in DAPT dose and cisplatin treatment time). The acquired spectral information revealed the cell changes induced by cisplatin and its combination with DAPT, and its dependency on the dose and treatment time of the combined drugs. The results obtained might provide an experimental basis for the establishment of new anti-cancer treatment strategies using combination therapies and Raman analysis to discover drug-cell interactions.
2. Materials and methods
2.1 Sample preparation
Samples were prepared according to the protocol published in our previous study [24,25]. The murine OS cell line K7M2 used in this work were purchased from the American Type Culture Collection (ATCC) and routinely cultured as suggested by ATCC. The osteosarcoma cells are cultured to the logarithmic growth phase, which can maximize the treatment effect of the drugs on the cells. Cisplatin was dissolved in PBS to obtain a stock solution of 50 mM, then diluted to a working dosage of 20 µM (called 20C). Considering that formalin may affect the internal proteins or nucleic acids of tumor cells [26], we did not use any material to fix the cells for eliminating interference from other materials as much as possible. OS cells were treated by 20 µM cisplatin and DAPT for 36 and 48 hours respectively as a control group, while the experimental groups were pretreated with DAPT at different concentrations (10, 20 and 40 µM, called 10D, 20D and 40D, respectively) for 24 hours, and subsequently treated with cisplatin at the same final dose of 20 µM for 12 and 24 hours. The detailed scheme to obtain the experimental and control group is shown in Fig. 1.
Fig. 1. Detailed experimental grouping scheme. 20C, OS Cells treated with 20 µM cisplatin; 10D, OS Cells treated with 10 µM DAPT; 20D, OS Cells treated with 20 µM DAPT; 40D, OS Cells treated with 40 µM DAPT.
2.2 Confocal Raman micro-spectroscopy
The confocal Raman microscopy used in this study is described in previous reports [27]. Briefly, a 532-nm fiber-coupled diode laser was collimated into a 63× water-immersion objective lens (NA=1.0, W Plan-APOCHROMAT, Zeiss, Germany) for Raman excitation and spectral measurements. To ensure an appropriate signal-to-noise ratio without damaging the samples, the laser power was controlled at about 10 mW. Cells were seeded in CaF2 slides that were placed on a multi-axis piezo scanning stage (P-524K081, PI GmbH, Germany) for multi-point spectral acquisition. The acquisition time of each Raman spectrum was 3×1 s, and the laser spot was located in different areas within the cell. All measurements were performed at room temperature after the calibration of the system wavelength and spectral intensity. After further screening for outliers, a total of 354 Raman spectra were collected and considered (Table S1). Five biological replicates were used.
2.3 Spectral data pre-processing and analysis
Spectral preprocessing was performed using the NWUSA Toolbox software according to the protocol previously reported in our articles [25,27,28]. Both point-acquired single spectrum and point-scanned spectral datasets were processed for further multivariate spectral analysis. The principal component analysis (PCA) is a classic unsupervised analysis that can explain the subtle differences between different types of samples. It extracts important feature information from a data set and express it as a new set of orthogonal variables called principal components (PCs) to reduce the dimensionality of the data set. Linear discriminant analysis (LDA) is used to identify an optimal projection direction that maximizes differences between samples in different groups and minimizes differences among samples within the same group, where the PCA scores are used as its input variable. After single spectral analysis, K-means cluster analysis (KCA) was used to automatically identify the cluster components and draw them into separate pseudo-color images. The average spectrum of each sub-cluster can be used to analyze the biochemical differences among different regions. Univariate spectral imaging is used to evaluate the integrated intensity of specific peaks found in cells to obtain the distribution of biochemical phenotypes.
3. Results
3.1 Spectral analysis
The mean spectra of OS cells treated with DAPT (20 µM, called 20D), cisplatin (20 µM, called 20C) and different doses of DAPT combined with cisplatin (10, 20, and 40 µM DAPT pretreatment, followed by 20 µM cisplatin treatment, which were called 10D+20C, 20D+20C, and 40D+20C, respectively) at both 36 and 48 hours is shown in Fig. 2. The spectra of all groups (Fig. 2 and Fig. S1) showed similar spectral characteristics based on DNA backbone, DNA ring bases, amino acids, lipids and proteins vibrations [21]. The tentative assignments for each peak are summarized in Table S2. Raman spectra of the cells treated with the combined drug showed evidential changes, in which the most significant one was the intensity decrease of the overall spectra compared with the cells treated with solo DAPT or cisplatin separately.
Fig. 2. Mean Raman spectra of OS cells treated with DAPT (20D group), cisplatin (20C group) and different doses of DAPT combined with cisplatin (10D+20C, 20D+20C, 40D+20C groups) at 36 h (A) and 48 h (B).
The anti-tumor effect of cisplatin was mediated by the formation of DNA adducts and inter- and intra-chain crosslinks [7]. The DNA characteristic peak at 787 cm−1 (ring breathing of cytosine/thymine/uracil and O–P–O symmetric stretch of the phosphodiester bond in DNA) and 1092 cm−1 band (Z-DNA, PO2− symmetric) [29] decreased in the DAPT combined with cisplatin treatment at 36 and 48 hours compared to sole cisplatin treatment, and a blue shift was observed (787 shifted to 793 cm−1, 1092 shifted to 1096 cm−1) with the increase of the DAPT dose, indicating that the DNA and the double-helical structure might be altered. The peak at 1580 cm−1 was attributed to the bases cytosine, adenine and guanine [30], and the base content gradually decreased with the increase of the DAPT dose at 36 and 48 hours, indicating that the DNA base pair might be damaged, thus affecting DNA replication (Fig. 2) [31,32]. In addition, the intensity of 941, 1172 and 1001 cm−1 Raman bands corresponding to proline [33], tyrosine [34] and phenylalanine [29] respectively, at 36 hours, decreased compared with sole cisplatin treatment, and the decrease trend became more significant with the increase of DAPT dose, suggesting the change of amino acid molecules in the cells [Fig. 2(A)]. The content of amino acid molecules in the combined high-dose DAPT and cisplatin (40D+20C group) group slightly increased at 48 hours [Fig. 2(B)]. The peak intensity of proteins and lipids in cells after the combination drug treatment decreased at 1246 [25], 1315 [33], 1450 [29], 1656 [33], 2885 [33] and 2934 [25] cm−1 compared with sole cisplatin treatment, and the difference in the decreasing intensity became more evidential with the increase of DAPT dose. Among them, the 1246 cm−1 spectral peak of the protein not only decreased in intensity, but a blue shift also appeared (1246 shifted to 1250 cm−1), indicating the changes in the protein content and spatial structure in the cells, and these changes were dependent on the dose of DAPT.
The Raman characteristic peak intensity of nucleic acids, proteins, and lipids at 48 hours showed different degrees of decrease compared with their intensity at 36 hours, as shown in Fig. 3. The intensity of nucleic acid peaks at 1092 and 1580 cm−1 in the 20C group decreased, and the blue shift occurred at 1092 cm−1 (1092 shifted to 1096 cm−1), indicating that the nucleic acid content decreased with the increase of the cisplatin treatment time, and the spiral structure also changed in the cell, as shown in Fig. 3(A). Furthermore, the band at 1338 cm−1 at 48 hours was attributed to the polynucleotide chain [21,33], and its intensity decreased, and its position disappeared, which might be due to the destruction of the double helix structure after the drug action. In addition, the characteristic peak of phenylalanine at 1001 cm−1 was red-shifted (1001 shifted to 997 cm−1), indicating that the phenylalanine structure in the cells was also changed. Moreover, a protein characteristic peak at 1315 cm−1 appeared after the long-term treatment with cisplatin. The intensity of the characteristic peaks at 787 and 828 cm−1 [33]of the nucleic acid in the 10D+20C group decreased with the increase of the cisplatin treatment time in the combined drugs, as shown in Fig. 3(B). The characteristic peak intensity of nucleic acids (787, 1092, 1315, 1338 and 1580 cm−1) proteins (1246, 1315 and 1656 cm−1) and lipids (1315, 2885 and 2934 cm−1) decreased in the 20D+20C group at 48 hours, as shown in Fig. 3(C). The peak intensity of nucleic acid (787, 1092 and 1580 cm−1) slightly increased in the 40D+20C group at 48 hours, as shown in Fig. 3(D), which might be due to the fragmentation of the cell nucleus due to the action of the drug, resulting in the separation of DNA and RNA and their package into different particles, eventually forming apoptotic bodies [35,36].
Fig. 3. Mean Raman spectra of OS cells treated with cisplatin (20C group (A)) and different doses of DAPT combined with cisplatin (10D+20C (B), 20D+20C (C), 40D+20C groups (D)) at 36 h and 48 h.
The bar chart in Fig. 4 was used to depict the relative spectral contributions of the experimental groups (20D, 20C, 10D+20C, 20D+20C, 40D+20C groups) at 36 and 48 hours. The spectral contributions of nucleic acids (787, 1092, 1338 and 1580 cm−1), phenylalanine (1001 cm−1), proteins (1246 and 1656 cm−1) and lipids (1450 and 2934 cm−1) (Table S2.) were significantly different among each group. The intensity of the 787 cm−1 band increased in the 10D+20C group, and gradually decreased in the 20D+20C group at 36 hours compared with the 20D and 20C groups. The trend of the 10D+20C group at 48 hours was the opposite to the one observed at 36 hours. Additionally, the intensity of 1092, 1338 and 1580 cm−1 bands in the 10D+20C and 20D+20C groups decreased with the increase of the DAPT dose and cisplatin treatment time in the combination drug compared with the 20D and 20C groups. However, the intensity of the 1092 and 1580 cm−1 peaks in the 40D+20C group at 48 hours was slightly higher than the intensity at 36 hours, which might be due to the nucleus rupture in the cell and apoptotic bodies formation under high-dose combination drugs and prolonged action [35,36]. The intensity of the characteristic peaks of protein and lipid at 1001, 1246, 1450, 1656 and 2934 cm−1 in the 20D, 20C, 10D + 20C, 20D + 20C and 40D + 20C groups gradually decreased with the increase of the DAPT dose and the treatment time of cisplatin. This might be due to the DAPT inhibition efficacy, which leads to the decrease in the expression of cell proliferation related proteins and the down-regulation of protein anabolism [21,36].
Fig. 4. Relative spectral contributions of biochemical components of the 20D, 20C, 10D+20C, 20D+20C, 40D+20C groups at 36 h and 48 h. (a), nucleic acid (787 cm−1); (b), phenylalanine (1001 cm−1); (c), nucleic acid (1092 cm−1); (d), protein (1246 cm−1); (e), nucleic acid (1338 cm−1); (f), lipid (1450 cm−1); (g), nucleic acid (1580 cm−1); (h), protein (1656 cm−1); (i), lipid (2934 cm−1).
3.2 PCA-LDA
PCA is used to extract the most important component information in the spectral dataset, and the highly specific biomolecule differences in each group is identified by the calculated spectral loadings and scores based on the acquired spectral features. The scatter points of the 20C and 10D+20C groups at 36 hours were distributed on the negative axis of PC1, while the scatter points of the 20D+20C and 40D+20C groups showed positive PC1 values, as shown in Fig. 5(A) and (B). The most discrete points of 20C, 10D+20C and 40D+20C groups were mainly concentrated on the positive axis of PC2, while the discrete points of the 20D+20C groups were mainly concentrated on the negative axis of PC2. The samples of the 10D+20C group were mainly concentrated on the PC3 positive axis, while most of the 20C group was on the PC3 negative axis, and the 20D+20C and 40D+20C groups were basically evenly distributed on the positive and negative PC3 axis, as shown in Fig. 5(C). Figure 5(D) shows three significant PCs loadings of the Raman spectral data set at 36 hours. The negative and positive characteristics of each PC loading were derived from the content changes in the biochemical molecules in the cells after drug treatment. In addition, the variance percentage of the first 10 PCs is plotted in Fig. S2 below. PC1 accounted for 82.46% of the total variance in the dataset. The evident peak positions in PC1 loading were at 1001, 1032, 1092, 1450, 1580, 1656 and 2934 cm−1, which could be attributed to phenylalanine, nucleic acid, and lipid components. PC2 explained the total variance of 9.08% of the data set. The negative peak corresponding to the loading could be attributed to the biochemical components such as nucleic acid at 1580 cm−1, protein Amide I at 1656 cm−1, and lipid at 2934 cm−1, while the positive peak could be attributed to protein at 1518 cm−1 [36] and phenylalanine at 1032 cm−1 [36]. PC3 represented the 2.88% of the total variance of the spectral data set. The loading of PC3 was distributed on both sides of the zero line. The positive features were mainly due to the nucleic acids (787 and 1092 cm−1), phenylalanine (1001 cm−1) and proteins (1032 and 1246 cm−1), while the negative characteristic peaks 1450 and 1656 cm−1 were attributed to the lipid C-H and protein amide I band.
Fig. 5. PCA analysis of 20C, 10D + 20C, 20D + 20C, 40D + 20C groups at 36 h. Score plot is plotted as PC1 vs PC2, (A); PC1 vs PC3, (B); PC2 vs PC3, (C); The corresponding loading spectra are plotted as PC1, PC2 and PC3, (D).
PC1, PC2, and PC3 at 48 hours explained 85.24%, 4.97%, and 3.73% of the total variance, respectively (The variance percentage of the first 10 PCs is plotted in Fig. S2). Figure 6(A) and (B) show that the 20C group was mainly distributed on the positive axis of PC1, and the 10D+20C, 20D+20C and 40D+20C groups were mainly distributed on the negative axis. The 10D+20C group was distributed on the positive axis of PC2, while the 20C, 20D+20C and 40D+20C groups were mainly distributed on the negative axis of PC2. The 20D+20C group was mainly distributed on the positive axis of PC3, as shown in Fig. 6(C). The loading of PC1 was mainly above the zero line, and significant peak positions at 787, 1001, 1092, 1246, 1450, 1580, 1656 and 2934 cm−1 could also be observed. The PCA scatter plot in Fig. 6(A) shows that the PC1 loading mainly represented the spectral characteristics of the 20C group. The positive characteristics of PC2 were located at 1338, 1580 and 2934 cm−1, while PC3 showed positive peaks at 1450, 1656 and 2934 cm−1, as shown in Fig. 6(D).
Fig. 6. PCA analysis of 20C, 10D+20C, 20D+20C, 40D+20C groups at 48 h. Score plot is plotted as PC1 vs PC2, (A); PC1 vs PC3, (B); PC2 vs PC3, (C); The corresponding loading spectra are plotted as PC1, PC2 and PC3, (D).
One-way ANOVA was used to identify the most diagnostically significant PCs (P < 0.01) in the dataset to better classify the different types of cell spectra. The results showed that PC1, PC2, PC3, and PC4 were the most diagnostically significant PCs at the 99% confidence level (Fig. S2). Therefore, the first four significant PC scores were used as input variables for the LDA algorithm to identify the significant difference among all groups. The LDA scores of the different groups calculated by PCA-LDA algorithm is shown in Fig. 7. The cell spectra of the 20C and 10D+20C groups at 36 hours were on the positive side of the first discriminant function, while the spectral distributions of the cells in the 20D+20C and 4D+20C groups were on the negative side. The zero line of the third discriminant function was able to clearly discriminate the 20C group from the 10D+20C group, as shown in Fig. 7(A). The 20C group at 48 hours was distributed on the negative axis of the first discriminant function, while the 10D+20C, 20D+20C and 40D+20C groups were mainly distributed on the positive axis, as shown in Fig. 7(B). The 40D+20C group was mainly distributed on the positive axis of the third discriminant function, while the 20D+20C group was mainly distributed on the negative axis of the second discriminant function.
Since two time points (36 and 48 hours) were investigated in this study, the first two PC scores were used as input variables for LDA to generate a “time-resolved” classification model. LDA projects the data onto a straight line in a two-dimensional graph. Figure 7(C-F) shows a scatter plot with the horizontal axis representing the number of spectra and the vertical axis representing the linear discriminant score, which revealed that the spectra of the cells in the two time points were separated. The distribution pattern of all the obtained data further confirmed that DAPT combined with cisplatin represented the effective combination therapy inhibiting OS cells in a DAPT dose and time-dependent manner.
3.3 Raman spectral imaging analysis
The effect of cisplatin combined with different doses of DAPT on the subcellular structure and biochemical components of the OS cells at 36 and 48 hours were investigated using the multivariate imaging method. The KCA algorithm was used to classify spectra with highly similar characteristics in both 700-1800cm−1 and 2800-3100 cm−1 spectral range. Then, the spectral dataset was divided into 4 different sub-clusters, and pseudo-hierarchical clustering trees in each group was constructed and presented in Fig. 8. The KCA results of the 20C, 10D+20C, 20D+20C, and 40D+20C groups were treated for 36 hours shown in Fig. 8(A) and 48 hours in Fig. 8(B). A microscopic image of each group of cells was added as a reference, as shown in Fig. 8(a). The regions of the subcellular structure of the nucleus, organelles, cytoplasm, and cell membrane are identified by the KCA pseudo-color image of each cluster [Fig. 8(b)]. Figure 8(c) shows the image spatially transformed between (a) and (b). Fig. S3 show the comparison of the average spectrum of the nuclear area (a), organelle area (b), cytoplasm area (c) and cell membrane area (d) of the 20C, 10D+20C, 20D+20C and 40D+20C groups at 36 and 48 hours, indicating the changes in biochemical characteristics of each subcellular structure after the treatment with different doses of DAPT combined with cisplatin.
Fig. 8. The KCA results of the 20C, 10D+20C, 20D+20C and 40D+20C groups at 36 h (A) and 48 h (B). For each sub-figure, image (a) shows the white light micrograph of the studied cell; image (b) shows the roots of a pseudo-hierarchical clustering tree of 20C, 10D+20C, 20D+20C, 40D+20C groups; image (c) displays the spatially transformed image between (a) and (b).
Cisplatin forms bivalent adducts with nucleophilic sites on the DNA purines, mainly producing intra-DNA chain cross-links between adjacent purines, leading to DNA damage and changes in the content and morphology of the nucleus, such as shrinkage of the nucleus and the cell, fragmentation of the cell nucleus, and the formation of apoptotic bodies [35], all signs of apoptosis. The KCA method was used to construct pseudo-color images of each group, and the results showed that cisplatin treatment did not result in significant changes in cell morphology, while DAPT pretreatment enhanced the effect of cisplatin in a dose- and time-dependent manner, inducing changes in cell morphology and mixing the cell regions, as shown in Fig. 8(A) and (B). The DNA damage in the nucleus (nuclear lysis and fragmentation) was significant with the increase of the dose of DAPT and the treatment time of cisplatin in the combined drug. An irregular distribution of the other cellular substructures was also observed, suggesting morphological and architectural modifications in the arrangements of the cellular components correlated with the different doses of DAPT combined with cisplatin, suggesting the promotion of apoptosis and the inhibition of proliferation of the tumor cells.
The role of DAPT combined with cisplatin in the biochemical environment of the cell was further evaluated by the analysis of the average Raman spectra extracted from each cluster in the intracellular region. The spectrum corresponding to the nuclear region was dominated by nucleic acid features, with bands displayed at 787, 1092, and 1580 cm−1, as shown in Fig. S3(A) and (B) (a) and Fig. S4. The nucleic acid concentration in the cells of the 10D+20C, 20D+20C and 40D+20C groups gradually decreased, and the nucleic acid content was gradually reduced with the increase of the cisplatin treatment time in the combination drug (from 36 to 48 hours), compared to the cells in 20C group. This might be because the DNA damage effect in cells caused by the combination drug was higher than that of cisplatin alone, and the degree of DNA damage gradually increased with the increase of DAPT dose and cisplatin treatment time. Changes in the concentration of cytochrome c (751 cm−1) [37], guanine (1315 and 1580 cm−1), adenine (1580 cm−1) and membrane lipids (1301 [33] and 2934 cm−1) were observed in the organelle region (Fig. S3(A) and (B) (b) and Fig. S5). The cytoplasmic region includes nucleic acid and protein related bands, as shown in Fig. S3(A) and (B) (c) and Fig. S6. The increase in nucleic acid (787 and 1092 cm−1) and protein (1246, 1450 and 1656 cm−1) content in the 10D+20C group was observed at 36 hours, which might be due to the early induction of cell apoptosis due to DAPT combined with cisplatin. In addition, the content of cytochrome c (751 cm−1) gradually increased in the 20C, 10D+20C, 20D+20C and 40D+20C groups at 36 and 48 hours with the increase of DAPT dose and cisplatin treatment time, which might be due to the release of cytochrome c caused by the action of drugs. The conformational changes in protein and lipid content in the cell membrane region might be caused by the shrinkage of the cell membrane in the early stage of apoptosis induced by the drugs. The spectrum of the cell membrane becomes noisy and many small peaks appeared under the action of high-dose DAPT and long-term cisplatin (40D+20C, 48 hours), shown in Fig. S3(B)(d) and Fig. S7, which might be due to the significant changes in the composition and conformation of the cell membrane [38,39].
Although KCA pseudo-color images could identify the changes of the subcellular structure under different drug conditions, it could not present the distribution changes of certain components in the cells. Therefore, the univariate spectral imaging was performed at the Raman bands of cytochrome c (751 cm−1), protein (1656 cm−1) and lipid molecules (2852 cm−1) in the cells for visualizing the drug-induced variations in the distribution of specific constituents. The cytochrome c (751 cm−1) was roughly located in the organelle and cytoplasmic area, the protein (1656 cm−1) was located almost in the entire cell, and the lipid molecules (2852 cm−1) were approximately located at the boundary between the cytoplasm and cell membrane (Fig. 9). The distribution of cytochrome c and lipid molecules in the cells treated with the combination of DAPT and cisplatin (20C+10D, 20C+20D, and 20C+40D groups) was significantly different compared with their distribution after the sole cisplatin treatment (20C group). Cytochrome c is a triggering factor in the activation of the caspase cascade, it can lead to the degradation of intracellular proteins, and its release from mitochondria is a sign of apoptosis [40,41]. The distribution of cytochrome c in the subcellular region was obtained through the univariate Raman spectrum image of cytochrome c (751 cm−1) combined with the KCA pseudo-color subcellular structure image (Fig. 8). The distribution of cytochrome c in the cytoplasm gradually increased with the increase of DAPT dose and cisplatin treatment time in the combination drug, which might be due to the gradual release and diffusion of cytochrome c from the mitochondria to the cytoplasm caused by the drug action. In the drug groups under different conditions (20C, 10D+20C, 20D+20C and 40D+20C groups), the 1656 cm−1 protein was almost all distributed in the whole cell region without significant changes. In addition, more lipid molecules (2852 cm−1) appeared in the cells with the increase of DAPT dose and cisplatin treatment time, compared with the 20C group, which might be related to the accumulation of intracellular vesicles during the apoptotic morphological stage [37,39].
Fig. 9. Raman images of 20C, 10D+20C, 20D+20C, 40D+20C groups at 36 h (A) and 48 h (B). Point-scanned Raman images of cytochrome c (751 cm−1), protein (1656 cm−1) and lipid molecules (2852 cm−1) in OS cells obtained by integration of labeled peaks.
4. Discussion
Although the cisplatin treatment is effective, the long-term treatment is still a challenge due to drug resistance and side effects [8,38]. It is of utmost importance to find chemotherapeutic drugs that can be combined to enhance the therapy efficacy and overcome its shortcomings. The response of cells to certain drugs can be better understanded without fixation, which avoid varying artefacts impacts on analyzation of the intracellular biochemical constitution nature [42]. The ideal situation would be to obtain information directly in living cells regarding the composition and distribution of biomolecules [42,43]. CRMI is a non-invasive method for obtaining cell information and establishing cell images without any chemical fixation or fluorescent markers. Since that, CRMI combined with multivariate statistical algorithms was used in this study to clarify the synergistic effect of cisplatin and DAPT in OS cells. After referencing to some biomedical studies on the cell responses to DAPT [8,44], cisplatin [45,46] and their combined treatment [47,48], and for achieving an experimental repeatability and reliability between our previous and current work [24,49,50], we decided to treat OS cells by 20 µM cisplatin and DAPT for 36 and 48 hours respectively in the control group, while 10, 20 and 40 µM DAPT was used to pretreat OS cells for 24 hours, and 20 µM cisplatin was used as the final dose for 12 and 24 hours.
OS cells treated with cisplatin combined with DAPT resulted in significant changes in the Raman spectrum of the cells. The most significant change was that the overall spectral intensity significantly decreased, and the changes became more significant with the increase of DAPT dose and treatment time (Fig. 2). In particular, the most sensitive molecular vibrations indicating cell death after drug treatment were the O-P-O stretching of DNA at 787 cm−1 and phenylalanine at 1001 cm−1, which gradually decreased with the increase of the DAPT dose and treatment time in the combined drug [51]. This observation could be related to the breakdown of the DNA strand following the breakage of the phosphodiester bond, and the longer treatment time led to the further DNA damage. In addition to the DNA backbone vibration, the ring vibration corresponding to cytosine and thymine at 1580 cm−1 was reduced [32,52]. The nucleic acid content decreased as the dose of DAPT increased in the combined treatment, which might be due to DAPT combined with cisplatin affecting DNA replication in a dose-dependent manner, potentially promoting G0/G1 phase arrest by the inhibition of G1-S phase progression, although this hypothesis needs further confirmation [8]. However, when the high DAPT dose is combined with cisplatin with the increase of the treatment time, the cellular content of nucleic acid (787, 1092 and 1580 cm−1) increased, which might be related to the formation of apoptotic bodies in the late stage of apoptosis [53,54]. Moreover, the protein and lipid content in the cell decreased with the increase in the dose of DAPT in the combination drug and the increase in the treatment time, mainly shown by the decrease of the protein Raman peak at 1246, 1301, 1450, 1656 and 2934 cm−1. This might be due to the increase of drug dosage and treatment time, leading to apoptosis of the OS cells, consequently causing a decrease in the content of the related proteins and lipids in the cells. Thus, DAPT combined with cisplatin could enhance the sensitivity of OS cells to cisplatin, inhibit cell proliferation, and significantly promote cell apoptosis in a dose- and time-dependent manner compared with sole cisplatin treatment.
The combination of cisplatin and DAPT induced complex changes in various biomolecules such as nucleic acids, proteins and lipids in OS cells. PCA analyzed the main differences in the spectral information of the investigated cells in 20C, 10D+20C, 20D+20C and 40D+20C groups, allowing to select the number of variables for subsequent LDA analysis to establish a discrimination model between all groups, as shown in Fig. 5 and 6. According to the results of PCA analysis, the spectra of most biochemical components in cells treated with cisplatin combined with DAPT were relatively reduced in a dose- and time-dependent manner compared with the spectra of cells treated with cisplatin, supporting the results of the single spectral study, as shown in Fig. 2. The first four significant PCs score were used in the LDA algorithm to develop an effective discriminant model for cell classification and identify the spectral changes of cells caused by different drug conditions.
5. Conclusions
The changes in the biochemical components of the mouse OS cell line K7M2 caused by the treatment with the γ-secretase inhibitor DAPT combined with cisplatin were analyzed using CRMI and multivariate analysis methods, and these changes were evaluated in a dose- and time-dependent manner. The results of the observed spectrum reflected the changes in the content and structure of the nucleic acid, lipid, and protein content of the intracellular biochemical components after the exposure to the combined drug. In addition, the obtained spectral information indicated that the synergistic pro-apoptotic effect of DAPT combined with cisplatin on cells might be related to the dose of DAPT and the treatment time of cisplatin in the combined drug. The multivariate analysis method PCA-LDA was further performed to distinguish the spectral characteristics of all the considered treatment groups, and the significant effects of different DAPT doses and different cisplatin treatment time in the combination drug were explained with high accuracy. The results of the KCA image clearly depicted the subcellular area and its characteristic spectrum, showing that the nuclear division of OS cells gradually increased, and the physiology of other organelles in the cell also underwent significant changes with the extension of the DAPT dose and cisplatin treatment time in the combination drug. These Raman spectroscopy analysis methods provided a unique and effective tool for studying the interaction of combined anticancer drugs with cancer cells and the cell heterogeneity under drug therapy. Thus, this approach might help the understanding of the drug resistance mechanism in tumors and the screen of drugs suitable for individualized tumor treatment.
Funding
National Natural Science Foundation of China (61911530695); Key Science and Technology Program of Shaanxi Province (2016ZDJC-15, 2018TD-018).
Disclosures
The authors declare that they have no conflict of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. S. Osasan, M. Zhang, F. Shen, P. J. Paul, S. Persad, and C. Sergi, “Osteogenic sarcoma: a 21st century review,” Anticancer Res. 36(9), 4391–4398 (2016). [CrossRef]
2. L. Mirabello, R. J. Troisi, and S. A. Savage, “International osteosarcoma incidence patterns in children and adolescents, middle ages and elderly persons,” Int. J. Cancer 125(1), 229–234 (2009). [CrossRef]
3. L. Wang, F. Jin, A. Qin, Y. Hao, Y. Dong, S. Ge, and K. Dai, “Targeting Notch1 signaling pathway positively affects the sensitivity of osteosarcoma to cisplatin by regulating the expression and/or activity of Caspase family,” Mol. Cancer 13(1), 139 (2014). [CrossRef]
4. G. Bacci, A. Longhi, F. Fagioli, A. Briccoli, M. Versari, and P. Picci, “Adjuvant and neoadjuvant chemotherapy for osteosarcoma of the extremities: 27 year experience at Rizzoli Institute, Italy,” Eur. J. Cancer 41(18), 2836–2845 (2005). [CrossRef]
5. G. Rosen, B. Caparros, A. G. Huvos, C. Kosloff, A. Nirenberg, A. C. acavio, R. C. Marcove, J. M. Lane, B. Mehta, and C. Urban, “Preoperative chemotherapy for osteogenic sarcoma: selection of postoperative adjuvant chemotherapy based on the response of the primary tumor to preoperative chemotherapy,” Cancer 49(6), 1221–1230 (1982). [CrossRef]
6. J. Du, Y. Zhou, Y. Li, J. Xia, Y. Chen, S. Chen, X. Wang, W. Sun, T. Wang, X. Ren, X. Wang, Y. An, K. Lu, W. Hu, S. Huang, J. Li, X. Tong, and Y. Wang, “Identification of Frataxin as a regulator of ferroptosis,” Redox Biol. 32, 101483 (2020). [CrossRef]
7. S. Dasari and P. B. Tchounwou, “Cisplatin in cancer therapy: molecular mechanisms of action,” Eur. J. Pharmacol. 740, 364–378 (2014). [CrossRef]
8. G. Dai, S. Deng, W. Guo, L. Yu, and T. Gao, “Notch pathway inhibition using DAPT, a γ-secretase inhibitor (GSI), enhances the antitumor effect of cisplatin in resistant osteosarcoma,” Mol. Carcinog. 58(1), 3–18 (2019). [CrossRef]
9. S. X. Chong, S. C. F. Au-Yeung, and K. K. W. To, “Monofunctional platinum (PtII) compounds-shifting the paradigm in designing new pt-based anticancer agents,” Curr. Med. Chem. 23(12), 1268–1285 (2016). [CrossRef]
10. K. A. Janeway and H. E. Grier, “Sequelae of osteosarcoma medical therapy: a review of rare acute toxicities and late effects,” Lancet Oncol. 11(7), 670–678 (2010). [CrossRef]
11. M. T. Kuo, S. Fu, N. Savaraj, and H. H. W. Chen, “Role of the human high-affinity copper transporter in copper homeostasis regulation and cisplatin sensitivity in cancer chemotherapy,” Cancer Res. 72(18), 4616–4621 (2012). [CrossRef]
12. M. Arita, S. Watanabe, N. Aoki, S. Kuwahara, R. Suzuki, S. Goto, Y. Abe, M. Takahashi, M. Sato, S. Hokari, A. Ohtsubo, S. Shoji, K. Nozaki, K. Ichikawa, R. Kondo, M. Hayashi, Y. Ohshima, H. Kabasawa, M. Hosojima, T. Koya, A. Saito, and T. Kikuchi, “Combination therapy of cisplatin with cilastatin enables an increased dose of cisplatin, enhancing its antitumor effect by suppression of nephrotoxicity,” Sci. Rep. 11(1), 750 (2021). [CrossRef]
13. Z. Wang, Y. Li, A. Ahmad, A. S. Azmi, S. Banerjee, D. Kong, and F. H. Sarkar, “Targeting Notch signaling pathway to overcome drug resistance for cancer therapy,” Curr. Drug Targets 1806(2), 258–267 (2010). [CrossRef]
14. L. Yu, K. Xia, T. Gao, J. Chen, Z. Zhang, X. Sun, B. M. Simões, R. Eyre, Z. Fan, W. Guo, and R. B. Clarke, “The Notch Pathway Promotes Osteosarcoma Progression through Activation of Ephrin Reverse Signaling,” Mol. Cancer Res. 17(12), 2383–2394 (2019). [CrossRef]
15. Z. Chen, Y. Zhu, X. Fan, Y. Liu, and Q. Feng, “Decreased expression of miR-184 restrains the growth and invasion of endometrial carcinoma cells through CDC25A-dependent Notch signaling pathway,” American Journal of Translational Research 11(2), 755–764 (2019).
16. R. A. Previs, R. L. Coleman, A. L. Harris, and A. K. Sood, “Molecular pathways: translational and therapeutic implications of the notch signaling pathway in cancer,” Clin. Cancer Res. 21(5), 955–961 (2015). [CrossRef]
17. I. M. Shih and T. L. Wang, “Notch signaling, γ-secretase inhibitors, and cancer therapy,” Cancer Res. 67(5), 1879–1882 (2007). [CrossRef]
18. D. Li, T. Li, Z. Shang, L. Zhao, and J. Zhou, “Combined inhibition of Notch and FLT3 produces synergistic cytotoxic effects in FLT3/ITD+ acute myeloid leukemia,” Signal Transduction Targeted Ther. 5(1), 21 (2020). [CrossRef]
19. P. Ranganathan, K. L. Weaver, and A. J. Capobianco, “Notch signalling in solid tumours: a little bit of everything but not all the time,” Nat. Rev. Cancer 11(5), 338–351 (2011). [CrossRef]
20. Z. Farhane, F. Bonnier, and H. J. Byrne, “An in vitro study of the interaction of the chemotherapeutic drug Actinomycin D with lung cancer cell lines using Raman micro-spectroscopy,” J. Biophotonics 11(1), e201700112 (2018). [CrossRef]
21. H. Huang, H. Shi, S. Feng, W. Chen, Y. Yu, D. Lin, and R. Chen, “Confocal Raman spectroscopic analysis of the cytotoxic response to cisplatin in nasopharyngeal carcinoma cells,” Anal. Methods 5(1), 260–266 (2013). [CrossRef]
22. F. A. Ryan Buckmaster, Myo Thein, Jian Xu, and M. Zhang, “Detection of drug-induced cellular changes using confocal Raman spectroscopy on patterned single-cell biosensors,” Analyst 134(7), 1440–1446 (2009). [CrossRef]
23. Z. Farhane, F. Bonnier, A. Casey, and H. J. Byrne, “Raman micro spectroscopy for in vitro drug screening: subcellular localisation and interactions of doxorubicin,” Analyst 140(12), 4212–4223 (2015). [CrossRef]
24. Q. Jie, W. Rui, B. Chenguang, W. Junxiang, D. Hui, W. Shuang, L. Yuhuan, Z. Yonglin, L. Jianjun, Y. Yiting, H. Xijing, and W. Dong, “Notch signaling regulates osteosarcoma proliferation and migration through Erk phosphorylation,” Tissue Cell 59, 51–61 (2019). [CrossRef]
25. J. Li, J. Qin, X. Zhang, R. Wang, Z. Liang, Q. He, Z. Wang, K. Wang, and S. Wang, “Label-free Raman imaging of live osteosarcoma cells with multivariate analysis,” Appl. Microbiol. Biotechnol. 103(16), 6759–6769 (2019). [CrossRef]
26. R. Smith, K. L. Wright, and L. Ashton, “Raman spectroscopy: an evolving technique for live cell studies,” Analyst 141(12), 3590–3600 (2016). [CrossRef]
27. J. Li, Z. Liang, S. Wang, Z. Wang, X. Zhang, X. Hu, K. Wang, Q. He, and J. Bai, “Study on the pathological and biomedical characteristics of spinal cord injury by confocal Raman microspectral imaging,” Spectrochim. Acta, Part A 210, 148–158 (2019). [CrossRef]
28. D. Song, Y. Chen, J. Li, H. Wang, T. Ning, and S. Wang, “A graphical user interface (NWUSA) for Raman spectral processing, analysis and feature recognition,” J. Biophotonics 14(5), e202000456 (2021). [CrossRef]
29. Z. Farhane, F. Bonnier, O. Howe, A. Casey, and H. J. Byrne, “Doxorubicin kinetics and effects on lung cancer cell lines using in vitro Raman micro-spectroscopy: binding signatures, drug resistance and DNA repair,” J. Biophotonics 11(1), e201700060 (2018). [CrossRef]
30. J. W. Chan, D. K. Lieu, T. Huser, and R. A. Li, “Label-free separation of human embryonic stem cells and their cardiac derivatives using Raman spectroscopy,” Anal. Chem. 81(4), 1324–1331 (2009). [CrossRef]
31. I. Schie, L. Alber, A. Gryshuk, and J. W. Chan, “Investigating drug induced changes in single, living lymphocytes based on Raman micro-spectroscopy,” Analyst 139(11), 2726–2733 (2014). [CrossRef]
32. I. Notingher, S. Verrier, S. Haque, J. M. Polak, and L. L. Hench, “Spectroscopic study of human lung epithelial cells (A549) in culture: living cells versus dead cells,” Biopolymers 72(4), 230–240 (2003). [CrossRef]
33. A. L. M. Batista de Carvalho, M. Pilling, P. Gardner, J. Doherty, G. Cinque, K. Wehbe, C. Kelley, L. A. E. Batista de Carvalho, and M. P. M. Marques, “Chemotherapeutic response to cisplatin-like drugs in human breast cancer cells probed by vibrational microspectroscopy,” Faraday Discuss. 187, 273–298 (2016). [CrossRef]
34. R. Deng, H. Qu, L. Liang, J. Zhang, B. Zhang, D. Huang, S. Xu, C. Liang, and W. Xu, “Tracing the therapeutic process of targeted aptamer/drug conjugate on cancer cells by surface-enhanced raman scattering spectroscopy,” Anal. Chem. 89(5), 2844–2851 (2017). [CrossRef]
35. N. Uzunbajakava, A. Lenferink, Y. Kraan, E. Volokhina, and C. Otto, “Nonresonant Confocal Raman imaging of DNA and protein distribution in apoptotic cells,” Biophys. J. 84(6), 3968–3981 (2003). [CrossRef]
36. H. D. Halicka, E. Bedner, and Z. Darzynkiewicz, “Segregation of RNA and separate packaging of DNA and RNA in apoptotic bodies during apoptosis,” Exp. Cell Res. 260(2), 248–256 (2000). [CrossRef]
37. M. Okada, N. I. Smith, A. Palonpon, H. Endo, S. Kawata, M. Sodeoka, and K. Fujita, “Label-free Raman observation of cytochrome c dynamics during apoptosis,” Proc. Natl. Acad. Sci. U. S. A. 109(1), 28–32 (2012). [CrossRef]
38. M. Apps, E. Choi, and N. Wheate, “The state-of-play and future of platin drugs,” Endocr.-Relat. Cancer 22(4), R219–R233 (2015). [CrossRef]
39. A. Zoladek, F. C. Pascut, P. Patel, and I. Notingher, “Non-invasive time-course imaging of apoptotic cells by confocal Raman micro-spectroscopy,” J. Raman Spectrosc. 42(3), 251–258 (2011). [CrossRef]
40. M. O. J. N. Hengartner, “The biochemistry of apoptosis,” Nature 407(6805), 770–776 (2000). [CrossRef]
41. H. Salehi, E. Middendorp, I. Panayotov, P. Y. C. Dutilleul, A. G. Vegh, S. Ramakrishnan, C. Gergely, and F. Cuisinier, “Confocal Raman data analysis enables identifying apoptosis of MCF-7 cells caused by anticancer drug paclitaxel,” J. Biomed. Opt. 18(5), 056010 (2013). [CrossRef]
42. A. J. Hobro and N. I. Smith, “An evaluation of fixation methods: Spatial and compositional cellular changes observed by Raman imaging,” Vib. Spectrosc. 91, 31–45 (2017). [CrossRef]
43. K. Klein, A. M. Gigler, T. Aschenbrenner, R. Monetti, W. Bunk, F. Jamitzky, G. Morfill, R. W. Stark, and J. Schlegel, “Label-free live-cell imaging with confocal Raman microscopy,” Biophys. J. 102(2), 360–368 (2012). [CrossRef]
44. J. X. Zhou, J. B. Han, S. M. Chen, Y. Xu, Y. G. Kong, B. K. Xiao, and Z. Z. Tao, “γ-secretase inhibition combined with cisplatin enhances apoptosis of nasopharyngeal carcinoma cells,” Exp. Ther. Med. 3(2), 357–361 (2012). [CrossRef]
45. H. Nawaz, F. Bonnier, P. Knief, O. Howe, F. M. Lyng, A. D. Meade, and H. J. Byrne, “Evaluation of the potential of Raman microspectroscopy for prediction of chemotherapeutic response to cisplatin in lung adenocarcinoma,” Analyst 135(12), 3070–3076 (2010). [CrossRef]
46. M. Kim, J. Y. Jung, S. Choi, H. Lee, L. D. Morales, J. T. Koh, S. H. Kim, Y. D. Choi, C. Choi, T. J. Slaga, W. J. Kim, and D. J. Kim, “GFRA1 promotes cisplatin-induced chemoresistance in osteosarcoma by inducing autophagy,” Autophagy 13(1), 149–168 (2017). [CrossRef]
47. L. I. Jun-Yang, L. I. Ru-Jun, and H. D. Wang, “γ-Secretase inhibitor DAPT sensitizes t-AUCB-induced apoptosis of human glioblastoma cells in vitro via blocking the p38 MAPK/MAPKAPK2/Hsp27 pathway,” Acta Pharmacol. Sin. 35(6), 825–831 (2014). [CrossRef]
48. T. Aleksic and S. M. Feller, “Gamma-secretase inhibition combined with platinum compounds enhances cell death in a large subset of colorectal cancer cells,” Cell Commun. Signal 6(1), 8 (2008). [CrossRef]
49. J. Li, R. Wang, J. Qin, H. Zeng, K. Wang, Q. He, D. Wang, and S. Wang, “Confocal Raman Spectral Imaging Study of DAPT, a γ-secretase Inhibitor, Induced Physiological and Biochemical Reponses in Osteosarcoma Cells,” Int. J. Med. Sci. 17(5), 577–590 (2020). [CrossRef]
50. J. Li, J. Qin, H. Zeng, J. Li, K. Wang, and S. Wang, “Unveiling dose-and time-dependent osteosarcoma cell responses to the γ-secretase inhibitor, DAPT, by confocal Raman microscopy,” J. Biophotonics 13(11), e202000238 (2020). [CrossRef]
51. S. Verrier, I. Notingher, J. M. Polak, and L. L Hench, “In situ monitoring of cell death using Raman microspectroscopy,” Biopolymers 74(1-2), 157–162 (2004). [CrossRef]
52. S. Ahlawat, A. Chowdhury, A. Uppal, N. Kumar, and P. K. Gupta, “Use of Raman optical tweezers for cell cycle analysis,” Analyst 141(4), 1339–1346 (2016). [CrossRef]
53. S. Hu, Y. Feng, D. Zhang, X. Lu, and L. Zhong, “Raman spectral changes of Artemisinin-induced Raji cells apoptosis,” Vib. Spectrosc. 81, 83–89 (2015). [CrossRef]
54. Z. Farhane, F. Bonnier, H. J. J. A. Byrne, and B. Chemistry, “Monitoring doxorubicin cellular uptake and trafficking using in vitro Raman microspectroscopy: short and long time exposure effects on lung cancer cell lines,” Anal. Bioanal. Chem. 409(5), 1333–1346 (2017). [CrossRef]
