Human dermal fibroblasts (HDFs) in free-floating collagen matrices show minimal proliferation, although this may increase when the matrix is 'under tension'. We have investigated the detailed mechanics underlying one of the possible controls of this important cell behaviour, in particular the hypothesis that this is a response to substrate stiffness. Hyperhydrated collagen gels were plastic-compressed (PC) to give a predetermined collagen density and stiffness. Mechanical properties were tested using a dynamic mechanical analyser; cell number by Alamar blue assay. In the stiffest PC matrices, cell proliferation was rapid and seeding density-dependent, with a population doubling time of 2 days. In contrast, compliant attached matrices showed a 4 day lag period and a doubling time of 6 days. HDF growth was directly related to matrix stiffness, such that increasing stiffness using a range of compression levels (0-75% fluid removal) supported increasing proliferation rate, doubling times and matrix elastic modulus. HDF quiescence in compliant matrices was reversible, such that increasing stiffness in situ by compression at 1 and 5 days initiated proliferation. We conclude that collagen matrix stiffness regulates proliferation of fibroblasts (a duro-response), with important implications for understanding fibroblast-matrix feedback controls during wound healing and the design and regulation of engineered connective tissues based on collagen and other hydrogel-based scaffolds.

While matrix stiffness has been implicated in cell adhesion and migration, most studies have focused on the effects of substrate stiffness in 2D. The present work describes a novel continuous stiffness gradient model for studying such processes in 3D. Wedge-shaped collagen scaffolds were compressed to produce sheets of a desired (0.1 mm) uniform thickness, but with increasing collagen density along the length of the sheet. Dynamic mechanical analysis, carried out on 1 mm wide strips obtained from the two ends and the middle of each sheet, showed that the elastic modulus increased from 1057 +/- 487 kPa to 2305 +/- 693 kPa at the soft and stiff end respectively and was 1835 +/- 31 kPa in the middle. In constructs seeded with agarose marker beads prior to compression, mean agarose bead density rose from 10 +/- 1 to 71 +/- 12 at the soft and stiff end respectively and was 19 +/- 5 in the middle, indicating successful engineering of a density gradient corresponding to the measured stiffness gradient. Growth-arrested human dermal fibroblasts, initially seeded evenly within such constructs, accumulated preferentially towards the stiff part of the gradient after 3 and 6 days in culture. Durotactic migration was significant after 6 days. This model provides a new means for studying cellular mechanotaxis and patterning cells which is controllable, biomimetic and in 3D.

No abstract