Cardiac fibrosis inhibitor CTPR390 prevents structural and morphological changes in human engineered cardiac connective tissue

Abstract

Cardiac fibrosis is a key characteristic of heart failure. CTPR390, an experimental anti-fibrotic inhibitor targeting Hsp90, has shown success in animal models, but remains unexplored in human cardiac models. This study evaluated an engineered cardiac connective tissue (ECCT) model, focusing on changes in the extracellular matrix and fibroblasts. Results showed that CTPR390 prevented architectural changes in TGFβ1-activated ECCT, preserving tissue perimeter, collagen fibers alignment while reducing structured areas and degree of collagen structuration. CTPR390 treatment reduced cell area of fibroblasts under tension, without changes in the internal rounded cells devoid of tension. Fibroblast recruitment to tension areas was diminished, showing biomechanical behavior similar to control ECCT. This treatment also lowered the gene and protein expression of key pro-fibrotic markers. Here, advanced biotechnology was employed to detect the detailed structure of tissue fibrosis reduction after administering CTPR390, representing a significant advancement toward clinical application for cardiac fibrosis treatment.

Keywords: Cell biology; Molecular biology.

Graphical abstract

Figure 1

Administering CTPR390 to TGFβ1-activated ECCT resulted in a reversal of its physical and biomechanical characteristics to those observed in control ECCT (A) Representative images (bright field image, en face view, side views, and 3D rendering) of an ECCT using PS-OCT technique that showed the morphology of the engineered tissue around the two flexible poles. The white scale bar indicated 1.3 mm. (B) Three panels showing the correlative behavior observed within a sigma area between the Stokes U and the attenuation coefficient (μm⁻¹), considering the quantified overlap in distributions in Control, TGFβ1, and TGFβ1-CTPR390 ECCT samples. n = 3 biological samples (ECCT) measuring 10⁸ pixels in triplicate distributed in layers of 3 3.6 μm/px thick, assuming a refractive index of n = 1.38, typical of biological samples, evaluated up to the first 100 micrometers. White scale bar indicated 1 mm. (C) Representative ECCT images of the angle of polarization (AoP) orientation aligned with the fibers within the ECCT using the same color-coding as in panel. n = 3 biological samples (ECCT) measuring 10⁸ pixels in triplicate distributed in layers of 3 3.6 μm/px thick, assuming a refractive index of n = 1.38, typical of biological samples, evaluated up to the first 100 micrometers in Control, TGFβ1 and TGFβ1-CTPR390 groups. The white scale bar indicated 1.3 mm. (D) Bar graph illustrating the quantification of areas with the same percentage of AoP aligned with ECCT fibers, using the PS-OCT technique. AoP color code: blue (0° – 60°), green (60° – 120°), and orange (120° – 180°). The process from data acquisition segmenting areas with the same AoP and assigning the color code for the AoP is detailed in Figure S2A. n = 2 biological samples (ECCT) measuring 10⁸ pixels in triplicate distributed in layers of 3.6 μm/px thick, assuming a refractive index of n = 1.38, typical of biological samples, evaluated up to the first 100 micrometers. (E) Evolution along depth (z) of the xy-averaged Stokes U parameter (a.u.) among the Control, TGFβ1, and TGFβ1-CTPR390 groups within the initial 100 μm from the surface of the ECCT. Control Y = (−554.8e-04 ± 8.0e-04) + (−129.8e-05 ± 5.7e-05) X. TGFβ1 Y = (−525.7e-04 ± 8.1e-04) + (−362.1e-05 ± 5.8e-05) X. TGFβ1-CTPR390 Y = (−431.9e-04 ± 8.0e-04) + (−138.3e-05 ± 5.7e-05) X. n = 3 biological samples (ECCT) measuring 10⁸ pixels in triplicate distributed in layers of 3 3.6 μm/px thick, assuming a refractive index of n = 1.38, typical of biological samples, evaluated up to the first 100 micrometers. (F) Detailed analysis of contraction at day 13 presented in a bar graph comparing Control and TGFβ1-CTPR390, to TGFβ1 group. Values presented as means ± SEM for n = 12 ECCT per group with triplicates of each biological sample (∗∗∗∗p < 0.0001) were assessed through two-way ANOVA with Tukey’s multiple comparison test. (G) ECCT contraction was evaluated over a 13-day period based on the percentage (%) of pole deflection in three groups (Control, TGFβ1, and TGFβ1-CTPR390), n = 12 ECCT per group with triplicates of each biological sample. (H) Bar graph illustrating ECCT stiffness (Young’s modulus) with significant differences observed in TGFβ1 compared to Control or TGFβ1-CTPR390 groups n = 8–9 ECCT per group with triplicates of each biological sample. (I) Bar graph depicting ECCT resilience with significant differences noted in TGFβ1 compared to Control or TGFβ1-CTPR390 groups, n = 7–9 ECCTs with triplicates of each biological sample. All datasets, except for (F), passed the Shapiro-Wilk test for normal distribution. p-values for datasets following a normal distribution (∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.0005, ∗∗∗∗p < 0.0001) were determined using one-way ANOVA with Tukey’s multiple comparison test. For (F), where the dataset did not follow a normal distribution, p-values (∗∗∗∗p < 0.0001) were determined using the non-parametric Mann-Whitney test. In all graphs, error bars represent the variability of the data as standard deviation.

Figure 2

Restoration of Control ECCT gene and protein expression following the introduction of CTPR390 to TGFβ1-activated ECCT (A) Bright-field and epifluorescence imaging of control representative images of ECCT, TGFβ1-activated ECCT, and TGFβ1-activated ECCT subjected to CTPR390 treatment (Control, TGFβ1 and TGFβ1-CTPR390). Both white scale bars in the two panels indicated 5 mm. (B) Diagram illustrating the detection of CTPR390-green fluorescence (resulting from Alexa 488 fluorophore conjugated to CTPR390) incorporated into ECCT. It included a close-up view of the fluorescent fibroblasts within the ECCT and the subsequent extraction of fibroblasts from the ECCT, transferred to a tube, and later seeded onto a culture plate, revealing both fluorescent and non-fluorescent fibroblasts. Diagram generated with Biorender. (C) CTPR390 fluorescence captured directly from the surface of the TGFβ1 ECCT following a 13-day treatment with CTPR390 (TGFβ1-CTPR390 group). The white scale bar indicated 10 μm. (D) Bar graphs presenting the percentage and total cells retaining CTPR390 over a 13-day period in TGFβ1-CTPR390 ECCT, n = 4 ECCT per group, with three technical replicates. (E) Visualization of bright-field, nuclei (Hoechst) and CTPR390 (Alexa 488), along with the merged representation of all three images, pertaining to purified fibroblasts derived from TGFβ1-CTPR390 group and cultured in 2D plates. The white scale bar indicated 50 μm. (F) Heatmap illustrating the prevalent pro-fibrotic genes, with darker red representing higher expression (2x over the median) and light reds indicating lower expression (0.5 lower the median) in Control, TGFβ1, and TGFβ1-CTPR390 ECCT groups; n = 6–10 ECCT per group with 3 technical replicates of each ECCT sample. (G and H) XY plots illustrating the comparison of expression for all pro-fibrotic genes tested between TGFβ1-CTPR390 and TGFβ1groups (G), as well as between TGFβ1 and Control groups (H). (I) Bar graphs illustrating significant gene expression differences of the most differentially expressed pro-fibrotic genes tested in Control, TGFβ1 and TGFβ1-CTPR390 groups; n = 6–10 ECCT per group with 3 technical replicates of each ECCT sample. All datasets passed the Shapiro-Wilk test for normal distribution. p-values for all graphs (∗p < 0.05, ∗∗p < 0.005, ∗∗∗p < 0.0005, ∗∗∗∗p < 0.0001) were determined using one-way ANOVA with Tukey’s multiple comparison test. (J) Representative confocal images depicting the localization of COL I A1, COL III A1, FN, and FAP proteins in cross-sections of ECCTs. A higher density of superficial cells (indicated by DAPI-stained nuclei), and COL IA1, was observed in TGFβ1-activated ECCT s compared to the Control and TGFβ1-CTPR390-treated groups. CTPR390 was visualized using Alexa 488 fluorescence. The white scar bars indicated 50 μm. In all graphs, error bars represent the variability of the data as standard deviation.

Figure 3

Attenuation of TGFβ1-activated ECCT pro-fibrotic cell features after CTPR390 treatment (A) Representative images of ECCT surface with fibroblasts stained with hematoxylin-eosin from cross-sections of Control, TGFβ1 and TGFβ1-CTPR390 ECCTs and representative image of inner fibroblast from an ECCT; the white scale bars indicated 25 μm. (B) Violin plots depicting significant differences in the number of cell surface layers of TGFβ1 ECCT compared to Control and TGFβ1-CTPR390 ECCT, n = 6 independent samples measured in triplicate. (C) Three panels featuring a depiction with a front view of fibroblasts on the surface of Control, TGFβ1 and TGFβ1-CTPR390 ECCTs by scan microscopy of 1 ECCT per group. Yellow arrows indicate surface blebs, and white arrows indicate filopodia in Control, TGFβ1 and TGFβ1-CTPR390 ECCTs; the white scale bars indicated 6 μm. (D and E) (D) Cell area count and, e) Cell perimeter count in the three groups of ECCT (Control, TGFβ1 and TGFβ1-CTPR390) performed in n = 12 ECCT per group, n of cells = 400–510 cells per group. (F) Confocal images showcasing the 3D reconstruction of cellular surfaces within 100 μm depth utilizing lysotracker (in red) as a cell dye for comparing cell surface density in Control, TGFβ1, and TGFβ1-CTPR390 ECCTs. The white scale bars in upper panels indicated 30 μm and the white scale bars in lower panels indicated 40 μm. (G) Violin graph presenting the analysis of significant differences in the density of cell area in Control, TGFβ1, and TGFβ1-CTPR390 ECCTs; n = 4 independent ECCT samples measured in triplicate. The dataset from (G) passed Shapiro-Wilk test for normal distribution. The p-values for (G) which followed a normal distribution (∗p < 0.05, ∗∗∗∗p < 0.0001), were determined using one-way ANOVA with Tukey’s multiple comparison test. p-values for the dataset in (B), (D), and (E) (∗p < 0.05, ∗∗p < 0.005, ∗∗∗∗p < 0.0001) were determined using the non-parametric Mann-Whitney test, as the data did not follow a normal distribution. (H) Representative confocal images of ECCT cross-sections visualizing human actin-F (phalloidin-TRICT staining), CTPR390 fluorescence (Alexa 488), and nuclei (DAPI staining) in the surface and inside areas of control, TGFβ1 and CTPR390 ECCT. The white scale bars indicated 10 μm. In all graphs, error bars represent the variability of the data as standard deviation.

Figure 4

Characterization of distinct fibroblast subtypes in the ECCT (A and C) Hematoxylin/eosin staining on day 1 (A) and day 13 (C) of cross-sections of Control, TGFβ1, and TGFβ1-CTPR390 groups including insets of the whole cross section (C). The white bars of all panels indicated 50 μm, including inset panels showing the whole ECCTs. (B and D) Quantification of total cell numbers on day 1 (B) and day 13 (D) of Control, TGFβ1, and TGFβ1-CTPR390 groups. Cell counts were performed in 3–5 biological samples per group including 3 technical replicates per sample. (E–G) Electron microscopy images of superficial fibroblasts showing representative ultrastructural features of Control (E), TGFβ1 (F), and TGFβ1-CTPR390 (G) ECCT on day 13. (F) The blue arrow marks a depression of the plasma membrane; the red arrow points to a sac of rough endoplasmic reticulum (RER); and the green arrow points to one of the vesicles with very electron-dense fibrous elements. (G) The red arrow points to a normal-shaped RER sac; and the green arrow indicates a vacuole. Black scale bars indicated 2 μm (E, G) and 3 μm (F). (H and I) Bar graphs showing the percentage of superficial and inner fibroblasts per area on day 1 (H) and day 13 (I); n = 3 biological samples per group in triplicate were measured. The p values (∗p < 0.05) were determined using a one-way ANOVA with Tukey’s multiple comparison test. All datasets, except for (B), passed the Shapiro-Wilk test for normal distribution. p-values for datasets following a normal distribution (∗p < 0.05) were determined using one-way ANOVA with Tukey’s multiple comparison test. The dataset for (B) did not show significant differences, and to assess significance, p-values were calculated using the non-parametric Mann-Whitney test, as the data did not follow a normal distribution. In all graphs, error bars represent the variability of the data as standard deviation.

Figure 5

Localization of organized collagen produced by human fibroblasts on the surface of TGFβ1-activated ECCT (A) Representative cross-sectional scanning microscopy images of Control, TGFβ1 and TGFβ1-CTPR390 long arms of the ECCTs. The white scale bars indicated 100 μm. (B) Representative magnified scanning microscopy images of inner-sections of Control, TGFβ1 and TGFβ1-CTPR390 ECCTs. The white scale bars indicated 10 μm. (C and D) Violin graph illustrating a significant reduction of the TGFβ1 ECCT cross section area (CSA in mm²) (C), and a reduction of the TGFβ1 ECCT perimeter (in mm) (D), indicating a higher degree of compaction in TGFβ1 samples in comparison to Control or TGFβ1-CTPR390 ECCTs (n = 15–33 independent ECCTs analyzed per group with 3 technical replicates per sample). (E) Positive and significant correlation (∗∗∗∗p < 0.0001) between CSA and perimeter of Control ECCTs. The other two groups (TGFβ1, and TGFβ1-CTPR390) did not show a correlation between the CSA and their respective perimeters (n = 25–33 ECCTs analyzed per group with 3 technical replicates per sample). (F) Representative images showing distinctively colored structured collagen at the periphery of the TGFβ1-activated ECCT (central panel) under polarized light. In contrast, there was very low detection of structured collagen in the other two groups under study. The white scale bars indicated 50 μm. (G) Bar graph showing the percentage of structured collagen assessed through polarized light; n = 4–5 ECCT samples per group with 3 technical replicates per sample. All datasets, except for (C), passed the Shapiro-Wilk test for normal distribution. p-values for datasets following a normal distribution (∗p < 0.05, ∗∗p < 0.005) were determined using one-way ANOVA with Tukey’s multiple comparison test. For (C), where the dataset did not follow a normal distribution, p-values (∗p < 0.05, ∗∗∗p < 0.0005) were determined using the non-parametric Mann-Whitney test. (H) Representative confocal imaging of the ECCT surface detecting human collagen (COL I A1 antibody) showed a clear increase in collagen detection in TGFβ1 samples compared to Control or TGFβ1-CTPR390 samples. The white scale bars indicated 10 μm. (I) Second Harmonic Generation (SHG) signal showing the clear accumulation of collagen in TGFβ1 ECCTs compared to Control or TGFβ1-CTPR390 ECCTs. Yellow arrows indicated the place of the augmentation. The white scale bars indicated 5 μm, except the fifth panel that indicated 10 μm. In all graphs, error bars represent the variability of the data as standard deviation.

Figure 6

Differences in organization of collagen produced by human fibroblasts on the surface of TGFβ1-activated ECCT (A–C) Maps of the preferential orientation and degree of orientation of collagen in control (A), TGFβ1(B), and TGFβ1-CTPR390 (C) ECCTs. The white scale bars indicated 200 μm. Preferential of orientation in each 20 × 20 μm² pixel is indicated by the direction of the arrow in each of the pixels of the figure and the color-coded discs. Degree of orientation in each 20 × 20 μm² pixel is indicated by the thickness of the arrows in each of the pixels of the figure. Maps from individual ECCTs are found expanded in Figures S5A–S5C. (D) Violin graph illustrating significant changes in the degree of orientation of collagen in TGFβ1 (increase) and TGFβ1-CTPR390 (decrease) ECCTs. Graph was obtained considering the values from all of the pixels from each ECCT analyzing one complete ECCT per group including 3000–4000 pixels analyzed per ECCT. The dataset for panel (D) did not pass the Shapiro-Wilk test for normality. p-values (∗∗∗∗p < 0.0001) were determined using the non-parametric Mann-Whitney test, as the data did not follow a normal distribution. N = 5191-2051 measurements of 1 ECCT per group. In all graphs, error bars represent the variability of the data as standard deviation.