Rosmarinic Acid as a Candidate in a Phenotypic Profiling Cardio-/Cytotoxicity Cell Model Induced by Doxorubicin

Figures

• 

  Figure 1. High-throughput imaging methodology. (A) Schematic representation of high-throughput imaging methodology. Cells are grown in flat-bottom plates and a compound is added to the desired wells. After incubation, cells are stained with the dye mix (Hoechst 33342, tetramethyl rhodamine methyl ester (TMRM), and Calcein-AM). Image acquisition is by a commercial instrument and four images were collected per channel per well. The image dataset is then subjected to image and data processing by CellProfiler. (B) Schematic representation of the image data analysis pipeline. Single-cell images were created for 302 cell properties. The cells treated with Dox were labeled as “positive” and the non-treatment cells were labeled as “negative.” A pattern recognition model was built by four algorithms and validated by Monte Carlo cross-validation. The sample scores were calculated by the classification model to represent the cell morphologies.

• 

  Figure 2. Pattern recognition of model and naïve cells. (A) Receiver operator characteristic (ROC) of four models. The area under the curve (AUC) of Partial Least Squares Linear Discriminant Analysis (PLS-LDA) equaled 0.999, which is the highest of all, and the AUC of LogReg equaled 0.719. (B) Visualization of the test samples according to the PLS-LDA scores vs. the first latent variable spaces. Data from individual images of model cells (models, red dots, n = 65) and naïve cells (naïve, green dots, n = 109) are separated along the horizontal axis. A score below 0 is classified as a model cell. (C) Variable importance in projection (VIP) values with cell properties. This represents the importance of each property in the trained PLS-LDA model.

• 

  Figure 3. Classification of the effect of natural compounds in model cells. (A) Heat map representing the morphologies of model cells, naïve cells, and model cells treated with the various natural compounds (10 μM, 24 h). (B) Chemical structures of the natural compounds whose scores were more than 0.5.

• 

  Figure 4. Phenotypes associated with model cells pretreated with RosA. The green fluorescence in the FITC channel is Calcein-AM, the red fluorescence in the Texas red channel is TMRM, and the blue fluorescence in the DAPI channel is Hoechst 33342.

• 

  Figure 5. Effect of RosA on Dox-induced AC16 cell apoptosis. (A) The cell viability of RosA-pretreated cells after Dox treatment for 24 h. (B) The green fluorescence is Annexin V-FITC staining positive cells, and the red fluorescence is Dox staining positive cells. The blue fluorescence is Hochest 33342 staining positive cells. Apoptotic cells were stained by green and blue fluorescence, while normal cells were stained only by blue fluorescence. (C) Cleaved Caspase-9 was measured by Western blot analysis in AC16 cells after treatment. GAPDH was used as an internal control. ** p < 0.01, *** p < 0.001, **** p < 0.0001, ns: no significant difference. 1 μM Dox-treated was the model group; 0.25% DMSO-treated was the negative control group, compared with the model group. Results are expressed as mean ± SD (n = 3).

• 

  Figure 6. The antioxidant effect of RosA in Dox-induced AC16 injury model. (A) The levels of intracellular ROS were measured by DCFH-DA fluorescence. (B) The mRNA levels of HO-1 were evaluated using real-time polymerase chain reaction (RT-PCR), with GAPDH as the internal control. (C) Levels of HO-1 expression with various concentrations of RosA for 24 h as measured by Western blot, with GAPDH as an internal control. ** p < 0.01, *** p < 0.001, 1 μM Dox-treated was the model group; 0.25% DMSO-treated was the negative control group, compared with the model group. Results are expressed as mean ± SD (n = 3).

• 

  Figure 7. The myocardial protective effect of RosA in Dox-induced dysfunctional cardiomyocytes. (A) Beating patterns, (B) beating rates, and (C) amplitude of dysfunctional cardiomyocytes treated by various concentrations of RosA at selected time points (0, 2, 6, 12, and 24 h). All data were normalized by 0 h. 1 μM Dox-treated was the model group; 0.25% DMSO-treated was the negative control group, compared with the model group. Results are expressed as mean ± SD (n = 3).

• 

  Figure 8. The myocardial protection of RosA in a Dox-induced hiPS-CMs injury model. (A) The levels of NT-proBNP and IL-6 were measured by ELISA. (B) The mRNA levels of NPPB and IL6 were evaluated using real-time PCR, with GAPDH as an internal control. (C) The levels of myocardial protection proteins, such as HDAC1, GATA4, and CTnI, were measured by Western blot, with β-tubulin as an internal control. *** p < 0.001, **** p < 0.0001, ns: no significant difference. 1 μM Dox-treated was the model group; 0.25% DMSO-treated was the negative control group, compared with the model group. Results are expressed as mean ± SD (n = 3).