A prerequisite for the development of drugs to prevent or halt cardiac fibrosis is a better understanding of the involved mechanisms and the availability of suitable screening platforms. However, human cardiac fibroblasts (hCF) are difficult to obtain, especially from healthy donors. Therefore, hCF from commercial vendors are often used, e.g. to confirm data of complex animal models or as a basis for tissue engineering.

Thus, we investigated how similar or dissimilar commercial hCF from different donors behave, especially in our 3D engineered connective tissue (ECT) culture model. We used cells from 10 different donors in total (age 31-60 years, all Caucasian), including 2 from normal male hearts, 3 from normal female hearts, and 5 from male explanted hearts with end-stage heart failure (2 ICM, 2 DCM, 1 unknown). All cells were cultured in the same medium and usually expanded up to passage 4. Then, ECT were generated in molds with low (flexible model) or high (stiff model) mechanical constraints, and the biomechanical properties were analyzed after 5 days of culture. Key genes were analyzed by qPCR.

A first basal analysis demonstrated that hCF from healthy and diseased hearts appear in two clusters when plotted by their cell diameters in suspension versus their doubling times, indicating that cells from the diseased heart proliferate slower and possess a larger cell size. Next, the cells' ability to compact the collagen matrix in the ECT showed a strong linear correlation with the cell size, especially in the flexible model (flexible: R²=0.85, stiff R²=0.7). Surprisingly, the longitudinal tissue contraction was highly individual and ranged from 2-22.5% pole deflection. Although tissue compaction and contraction are both actin-dependent processes, they did not correlate. Ultimate tensile testing showed that all cells reacted to the higher mechanical constraint in the stiff model by producing stiffer tissues than in the flexible model. The fold difference in ECT stiffness ranged from 1.8 to 6.3 with a mean of 2.9-fold and was not related to any other investigated parameter. With respect to the absolute stiffness, the flexible ICM tissues were around 2-fold less stiff than their respective normal controls. By comparing the absolute stiffnesses of all tissues with their extensibility, a semi-logarithmic relation was found (R²=0.8), indicating that stiffer tissues are less strain-resistant. Finally, qPCR analysis demonstrated that almost all cells responded to the high mechanical constraint in the stiff model with an up-regulation of typical myofibroblast markers, like ACTA2, COL1A1, POSTN, and LUM. Only the two DCM cells and one female cell, which showed a rather high THY1 expression similar to one of both DCM cells, did not show a clear response in these genes, besides in LUM. This was largely based on an already increased expression level in the flexible model.