Cell Motility and the Cytoskeleton 40:13–21 (1998) Effect of Precise Mechanical Loading on Fibroblast Populated Collagen Lattices: Morphological Changes M. Eastwood,1* V.C. Mudera,2 D.A. McGrouther,2 and R.A. Brown2 1Centre for Tissue Engineering Research, Department of Technology and Design, University of Westminster, London, United Kingdom 2University College London, Division of Plastic Surgery Tissue Repair Unit, London, United Kingdom The contraction of a collagen lattice by resident fibroblasts causes strains to be developed within that lattice. These strains can be increased or decreased by altering the aspect ratio (ratio of length/width/thickness) of the fibroblast populated collagen lattice, as the cross-sectional area resisting the strain is changed and by the application of an external load. The fibroblasts align themselves with the direction of the maximum principle strain; in effect, these cells are ‘‘hiding’’ from the perceived strain. The direction of the maximum principle strain can be predetermined by the use of a computational finite element analysis. Using the tensioning-Culture Force Monitor to apply pre-determined loading patterns of known repeatable magnitudes, as calculated by the finite element analysis, we have succeeded in aligning fibroblasts into a deliberate predicted orientation. This study has shown that the resident fibroblast population will respond to changes in strain resulting from the most subtle of mechanical loads. This may be an important mechanism in development and repair of connective tissue. Cell Motil. Cytoskeleton 40:13–21, 1998. r 1998 Wiley-Liss, Inc. Key words: Culture Force Monitor; mechanical loading; fibroblast; collagen gel; cell morphology and alignment INTRODUCTION The biochemical effect of mechanical loading on fibroblast populations in collagen matrix has been studied previously. These studies have ranged from the stress effect on cyclic AMP pathway [He and Grinnell, 1994], the up-regulation of Tenascin C production [ChiquetEhrismann et al., 1994], to changes in free intercellular calcium ion concentration [Arora et al., 1994]. Margolis and Popov  reported that the application of a local force to mouse embryo fibroblasts induced the formation of cell processes. The effect of mechanical load on collagen expression was studied by Carver et al. , who found that the ratio of collagen type III to collagen type I was increased in mechanically stretched cells. Lambert et al.  found that fibroblast function could be effected by the application of mechanical stress in terms of collagenase activity and collagen synthesis. Butt r 1998 Wiley-Liss, Inc. et al.  reported that fibroblasts subjected to external mechanical stress resulted in increased cell growth and proliferation. All of these loading systems suffered from difficulties in defining the actual applied load, and how it would act on the resident cells. Methods of the application of mechanical stimulation differ. Sudden release of the loading across an attached lattice causes an effect known as strain recovery, or stress relaxation [Mochitate et al., 1991; He and Contract grant sponsor: Pearl Assurance Ltd; Contract grant sponsor: Phoenix Appeal; Contract grant sponsor: EEC; Contract grant number: BMH4-CT95-0396. *Correspondence to: Dr. Mark Eastwood, University College London, Division of Plastic Surgery Tissue Repair Unit, Charles Bell House, 67 Riding House St., London, W1P 7LD United Kingdom. Received 10 October 1997; accepted 26 January 1998 14 Eastwood et al. Grinnell, 1994]. In these experiments, the collagen lattice was allowed to remain attached to the culture well for a number of days. Fibroblasts become stellate in shape, contracted the gel thickness, and produced an internal strain. When the edges of the gel were released from the culture well, a rapid contraction of the gel was seen. Grinnell  has likened this to the transition from granulation tissue to dermis. Morphologically, the fibroblasts change during this strain recovery from large stellate or bipolar cells to rounded cells with the retraction of pseudopodia and the collapse of actin bundles [Tomasek et al., 1992; Mochitate et al., 1991; Grinnell, 1994]. The present study has used the tensioning-Culture Force Monitor (t-CFM) model [Eastwood, 1996]. This experimental model is a development of the Culture Force Monitor (CFM) [Eastwood, 1996; Eastwood et al., 1994, 1996] that enables precise uni-axial mechanical loads to be applied to fibroblast populated collagen lattices (FPCL) whilst simultaneously recording the total mechanical load across the FPCL (i.e., including the cellular contraction). This has allowed us to define accurately the magnitude and type of mechanical loading to which the matrix was subjected, and the resulting alignment of cells. In turn, this level of precision has made it possible to estimate the pattern of strain within the FPCL, through finite element analysis, and to manipulate this to test how cells respond to strain cues within that matrix. MATERIALS AND METHODS The tensioning-Culture Force Monitor (t-CFM) utilises the same force transducer, data collection system, and culture well as the Culture Force Monitor (CFM), as described previously [Eastwood et al., 1994]. Tension was applied to cultures via a microprocessor controlled microstepping motor (Micromech, Braintree, UK). The stepping motor drives a precision ground leadscrew with a pitch of 0.508 mm that is attached to the moveable table via a recirculating ball nut. The SX6 microprocessor controller (Micromech) enabled a resolution of up to 50,800 pulses per revolution of the microstepping motor. This, in conjunction with the precision ground leadscrew enabled a positional accuracy of 1 x 10-8 m to be achieved. The microstepping motor and table are mounted onto a base of dimensions 200 x 180 x 10 mm constructed of Perspex; a stainless steel post of 180 x 10 x 10 mm facilitates attachment of the force transducer in the same way as the CFM. The culture well is attached to the t-CFM by a cradle constructed from Perspex. A locking screw ‘‘nips’’ the petri dish and secures it in position. Initial positioning of the culture well during the experimental setup is through the moveable ‘‘X-Y’’ table mounted onto the motorised base. Control of the microstepping motor is achieved by the programming of the SX6 (Micromech) controller in ‘‘X-ware4’’ via the IBM computer. The code can be activated or deactivated by the RP240 instructional set (Micromech), which is run independently of the main computer. The t-CFM is shown in Figure 1. Calibration of the force transducer and data capture from the t-CFM was identical to the CFM and has been described previously [Eastwood et al., 1994, 1996; Eastwood, 1996]. Human dermal fibroblasts were cultured from explants of normal skin taken directly from the operating theatre [Eastwood et al., 1994; Burt and McGrouther, 1992] in Dulbecco’s Modified Eagle’s Medium (DMEM) Streptomycin/Penicillin (Gibco BRL, Paisley, Scotland) with 10% foetal calf serum (Advanced Protein Products [APP], West Midlands, UK). Fibroblast-seeded collagen gels were prepared by mixing 4 ml of 2.28 mg/ml solution of NATIVE acid soluble type I rat tail collagen (APP) with 0.5 ml of 10x strength, DMEM, and neutralised with 1M NaOH prior to the addition of fibroblasts (106 cells/ml of gel solution). The 5-ml collagen/cell suspension formed a gel within 5 min, at which stage the cell chamber was topped with a further 15 ml of culture medium, containing 2.5 mg/ml Amphotericin B (ICN, Paisley, Scotland). Rectangular FPCLs (75 x 25 mm) were restrained either by their opposing long or short edges (giving low or high aspect ratio configurations, respectively). Both configurations used identical culture wells. FPCLs were loaded with 120 Dyne force in a cyclic ramp after 8 h in culture. The loading cycle consisted of 15 min of loading in a linear progression at a rate of 480 dynes/h, followed by a 15-min resting phase, then unloading for 15 min at the same rate of loading prior to a further 15-min resting phase. The total cycle time being 1 h repeated for 16 h. Direct observation of collagen gels after 24 h in the t-CFM involved fixation (and subsequent processing, see below) by replacement of medium in the t-CFM culture chamber with 2.5% gluteraldehyde at 40C in 0.1M phosphate buffer, pH 7.5, at 40C, without the release of tension [Tomasek et al., 1992] for 1 h followed by washing in phosphate buffer, pH 7.5, at 4oC. The gels were stained with 1% toludine blue, (destained with distilled water) for routine light microscopy and stereomicroscopic examination on an Edge High Definition Stereo Light Microscope (Edge Scientific Instrument Corporation, Los Angeles, CA) [Greenberg and Boyde, 1993]. For scanning electron microscopy, fibroblast populated collagen gels were fixed as above, then rapidly cooled in liquid N2. After freezing, the gels were fractured down the long axis of the collagen gel, dehydrated, critical point dried, and sputter coated before examination on a JEOL (Peabody, MS) 35C electron microscope. Strain analysis of the collagen lattices was performed using the RASNA software suite of programs Mechanical Stimulation of Fibroblasts 15 Fig. 1. The tensioning-Culture Force Monitor. Indicated are the force transducer (f), the culture well (w), and the micro-stepping motor system (m). (RASNA Corporation, Los Angeles, CA). The collagen lattice was modelled in both the high aspect ratio of 3:1 (load and cell generated tension being measured along the short aspect) and the low aspect ratio of 0.33:1 (load and cell generated tension being measured along the long aspect). A load was applied that was representative of the force magnitude applied by the t-CFM. The model consisted of 3-dimensional elements connected through a P-Shell formulation, restraints were applied by the removal of the ux, uy, uz, fx fy and fz degrees of freedom along the grounded side of the model. The load was applied as uniform pressure acting out from the model edge. The analysis was computed on a Sun Sparkstation running a Unix operating system. RESULTS External mechanical loading by the t-CFM caused strains to be developed in the tethered collagen lattice, as did the force generated by the cells themselves. An assessment of the effect of external loading of the matrix was performed by Finite Element Analysis (FEA) to determine the strain gradients that were present in the FPCL in both the high and low aspect ratio gels. The strain gradients generated by the application of the mechanical loading are shown in Figure 2A,B. The FEA performed on the low aspect ratio gel (Fig. 2A) showed that there were no high strain gradients, and that the iso-strains were orientated parallel with the long axis, i.e., at 90o to the direction of the applied mechanical load. By contrast the strain gradients in the high aspect ratio gels (identical applied mechanical load) (Fig. 2B) were highly condensed, and aligned parallel with both the long axis and the direction of the applied load. Figure 2B also shows that all regions within the gel had high strain gradients except for the delta zone starting 4 mm behind the vyon bars. This zone had minimal strain due to the stiffness imparted into the collagen gel by its close proximity to the rigid vyon bar. The effect of applied mechanical load on cells within a low aspect ratio FPCL is shown in Figure 3A. After 16 h of cyclic loading, cells had a mixture of stellate and bipolar morphologies with no clear orientation, suggesting that there was no significant orientational cue in this configuration of the model. In contrast, application of the identical load in the high aspect ratio configuration of FPCLs produced a dramatic alignment parallel with the axis of the applied load and with the iso-strain lines Fig. 2. Principle strains developed within (A) low and (B) high aspect ratio collagen gels in response to the applied force from mechanical loading. Inserts: Direction of the applied force. Dotted, dashed, and solid lines show the contours of constant strain. Close lines indicate high strain gradients. Mechanical Stimulation of Fibroblasts predicted by the FEA. In addition shape control of the resident cells (Fig. 3B) was clearly produced in that most of the cells in these areas were elongate and approximately bipolar. This load produced a strain of only 0.2% of the total gel length. To test the idea that cells had aligned themselves parallel with the axis of the principle iso-strains, cell morphology was compared in adjacent areas of lowest and highest strain gradient, i.e., zone 2 (delta area) and zone 3 (high strain on the edge of the gel) as shown in Figure 4A. Figure 4B–D shows the corresponding micrographs of cells from zones 1–3 from those areas. Zone 2 (Fig. 4C) contained non-aligned cells with a mixture of stellate and bipolar morphologies. Aligned cells with bipolar morphologies, parallel with the axis of the principal iso-strain lines, were found in zones 1 (Fig. 4B) and 3 (Fig. 4D). Removal of all loads from the high aspect ratio FPCL experiment resulted in a rapid loss of cell alignment and return to a predominantly stellate morphology (Fig. 4E). Scanning EM of freeze fractured gel (Fig. 5A,B) confirmed the bipolar nature of the fibroblasts and their parallel alignment relative to each other and to the applied load. It is interesting to note that the cells are aligned in a predominantly ‘‘nose to tail’’ configuration (Fig. 5B), rather than in a staggered formation. This produced the appearance of strings of cells running in long lines through the matrix. Figure 5C shows a cell free collagen gel that has been loaded on the by the t-CFM for 24 h prior to being fixed under tension. After this loading cycle no alignment of the collagen fibrils is evident. DISCUSSION This series of experiments was performed to determine the effect on cell shape and alignment (morphology) in response to defined mechanical stimulation. Others [Butt et al., 1995] have reported that fibroblasts become aligned perpendicular to the direction of the applied force in a radially strained 2-dimensional system (Flexcell International Corp., McKeesport, PA). However, in this system the membrane is deformed by a vacuum; therefore, the force acts at a normal to the membrane surface. In addition, it has also been reported that fibroblasts in a 3-dimensional, ligament equivalent collagen matrix subjected to mechanical loading were aligned parallel with the direction of the applied load [Huang et al., 1993]. Data generated from this model are consistent with the proposal that fibroblasts become orientated parallel with the principle iso-strains. The response to mechanical load by other cell types has also been shown, notably in the observations of Julius Wolff  and in his work on bone remodelling, commonly known as Wolff’s law, where he proposed that bone grew in response to mechanical load in such a way as to resist the 17 prominent loads. This was later quantitated to the stress present in the system [Kummer, 1986]. It is predicted that orientation of the fibroblasts along the maximum principle strain will reduce the perceived strain across the cell. This will be further minimised by the long thin bipolar morphology. In effect, these cells appear to be hiding from the 3-dimensional strain, such that they only perceive the strain in one direction. Furthermore, the direction of the iso-strains in a 3-dimensional object will be dependent upon the shape of that object and can be independent of the direction of the applied load as in the case of the low aspect ratio configuration. Indeed, this was evident in this model, when identical loading of collagen matrices of different aspect ratios caused iso-strains to be aligned both parallel with, and at 900 to, the direction of the load. The FEA indicated that small shear and compressive force elements were also present at the sample zones and the possible contribution of these has been considered. However, the levels of shear strain were negligible, relative to the principal strains. Indeed, the centre of the gel (zone 1 in Fig. 4B) is devoid of all shear strain but is an area of maximum cell alignment. Low levels of shear strain were predicted in zone 3 (Fig. 4D) but the patterns did not match that of the observed cell alignment. Similarly, the levels of compressive strain are low and gave a poor match to the observed cell alignment. Cell processes can be induced by the action of a local force. Margolis and Popov  induced cell processes in rounded mouse embryo fibroblasts on coverslips by pulling the cells with two tungsten microelectrodes. This force caused cell processes to be formed within 20 sec of its application and parallel with the applied force. It was argued that applying a pulling, or tensile force, to the membrane reduced the load on the cytoskeletal components, namely the microtubules. The model of Margolis and Popov  also supports the hypothesis of cells being aligned along principal isostrains. The application of the tensile force caused strains to be generated. The cell processes formed were in the direction of the applied force, which would also be the direction of the maximum principal strain. The alignment of cells including fibroblasts has been attributed by others [Wojciak-Stothard et al., 1996; Ejim et al., 1993; Guido and Tranquillo, 1993; Dow et al., 1987] to contact guidance on substrates including collagen fibrils. This raises the possibility that cells here were orientated by contact guidance on collagen fibrils, which had been aligned due to the applied mechanical load. However, the SEM shown in Figure 5C shows that there is no alignment in the collagen gel after 24 h of mechanical loading. This tends to discount the idea of cells becoming aligned due to contact guidance during this early time course. Furthermore, it demonstrates that 18 Eastwood et al. Fig. 3. A: Effect of mechanical loading on a low aspect ratio FPCL; note the slight amount of cellular alignment. B: Effect on fibroblast morphology of mechanically loading a high aspect ratio FPCL. (Note all micrographs are the same magnification.) Mechanical Stimulation of Fibroblasts 19 Fig. 4. A: Diagram of a high aspect ratio FPCL. The zones for morphological studies are indicated as 1,2, or 3. B: Aligned fibroblasts from zone 1. C: Non-aligned fibroblasts from zone 2. D: Aligned fibroblasts from zone 3. E: Loss of cellular alignment and bipolarity from zone 1 with the removal of the mechanical load. The arrows indicate the direction of the long axis of the collagen gel. (Note all micrographs are the same magnification.) fibroblasts are highly responsive to the most subtle mechanical loads, and the resident environment. The principle that fibroblasts respond to minimise perceived strain within the matrix is perhaps not surpris- ing as such strains will produce cytoskeletal rearrangements consistent with observed shape changes [Eastwood et al., 1996; Eastwood, 1996]. However, the cell behaviour of hiding, or shielding, from strain does produce an 20 Eastwood et al. Fig. 5. Scanning electron micrographs of a high aspect ratio FPCL that has been freeze fractured along the major axis. A: Random orientation of collagen fibrils in between two aligned bipolar fibroblasts. B: Highly aligned fibroblasts, indicated with a small arrow is a cell, whilst the larger arrow indicates the general orientation of the cells and the direction of the major axis. C: Cell-free collagen gel, loaded in the t-CFM for 24 h and fixed under tension; note the random orientation of collagen fibrils. attractive mechanism to explain the intimate interrelationships between mechanical load and collagen fibre organisation in soft tissues. It is known that orientated fibroblasts tend to deposit collagen, which is aligned, in the same long axis [Trelstad and Birk,1985; Birk and Trelstad, 1985] (Brown et al., unpublished data). This new matrix will further tend to shield the resident cells from habitual strain in that axis, further minimising the perceived strain. Consequently, mechanical forces would be important in regulating the amount and the orientation of newly deposited collagen. Such a model would predict that increases in perceived strain would stimulate collagen production rates in fibroblasts, eventually falling again as the collagen matrix shielded the resident cells from the strain of applied loads. Disruption of the collagenous matrix in injury would then be expected to induce responses in fibroblasts consistent with those seen during soft tissue repair. Pathological failures of this feedback system may be seen in keloid and hypertrophic scarring where downregulation of matrix production is not elicited by a dense mechanically strong collagenous matrix. The dependence on strain and aspect ratio described here suggest that shape of a tissue will dictate the mechanical forces and, hence, the cellular activity. For example, a tendon has a high aspect ratio, and is subjected to high tensile loads. Consequently, the iso-strains will run parallel with the long axis, consistent with the orientation of the resident fibroblasts. Bipolar fibroblasts residing between collagen bundles of normal healthy tendon also have cell processes extending perpendicular to the main cell body [Squier and Magnes 1983]. These cell processes, in mature tissues, may act as parts of a strain sensing or ‘‘damage detection system.’’ Any breakage of the local collagen fibres would immediately place an unfamiliar strain on the fibroblasts attached to the surrounding intact collagen bundles, which its configuration is not adapted to minimise or shield. This would re-activate the cellular strain sensing system. From previous work we have determined that a load of 25–30% above the normal cell generated force will elicit a mechanical response from such cells [Eastwood et al., 1996] (Brown et al., unpublished data). The observation reported here of fibroblasts being aligned ‘‘nose to tail’’ rather than in less ordered (e.g., staggered) formation has analogies with slip planes, commonplace in engineering materials [Askeland, 1989]. Mechanical Stimulation of Fibroblasts Engineering materials fail along these slip planes as it represents the minimum energy requirement for failure to occur. When fibroblasts become aligned due to applied mechanical load they themselves will create local minimum potential energy wells in a similar manner to that seen in slip planes. It may be then that cells will be energetically drawn towards the same slip planes, further increasing the alignment in this head-to-tail manner. This cell arrangement is seen in the densest collagenous tissues such as tendons and ligaments where strain gradients are steepest. It is tempting to speculate that such energy wells and slip planes may have a developmental role in the formation of fascicle margins and epitendonous surfaces in vivo. In this study we have shown that fibroblasts are responsive to mechanical load, and mechanical stimulation can be a ‘‘cue’’ for the alignment of cells, necessary for many tissue engineering applications. The ability to control alignment of cells through strain, via mechanical load, and tissue shape provides potentially important insights into how soft tissue architecture is controlled during growth and repair. ACKNOWLEDGMENTS This work was supported by Pearl Assurance Ltd, the Phoenix Appeal, and EEC grant BMH4-CT95-0396, Bioartificial Organs and Tissues. We are grateful to Ms. Kirsty Smith for photography and electron microscopy. REFERENCES Arora, P.D. Bibby, K.J., and McCulloch, C.A. (1994): Slow oscillations of free intracellular calcium ion concentration in human fibroblasts responding to mechanical stretch. J. Cell. Physiol. 161:187–200. Askeland, D.R. (1989): In ‘‘The Science and Engineering of Materials.’’ Van Nostrand Reinhold, Chap. 3, pp. 74–80. Birk, D.E., and Trelstad, R.L. (1985): Fibroblasts create compartments in the extracellular space where collagen polymerizes into fibrils and fibrils associate into bundles. Ann. N. Y. Acad. Sci. 460:258–266. Burt, A. M., and McGrouther, D. A. (1992): Production and use of skin cell cultures in theraputic situation. In Spier, R.E. and Griffiths, J.B. (eds) ‘‘Animal Cell Biotechnology.’’ New York: Academic Press, pp. 150–168. Butt, R.P. Laurent, G.J., and Bishop, J.E. (1995): Mechanical load and polypeptide growth factors stimulate cardiac fibroblast activity. Ann. N.Y. Acad. Sci. 752:387–393. Carver, W., Nagpal, M.L., Nachtigal, M., Borg, T.K., and Terracio, L. (1991): Collagen expression in mechanically stimulated cardiac fibroblasts. Circ. Res. 69:116–122. Chiquet-Ehrismann, R., Tannheimer, M., Koch, M., Brunner, A., Spring, J., and Martin, D. (1994): Tenascin C expression by fibroblasts is elevated in stressed collagen gels. J. Cell. Biol. 127:2093–2101. Dow, J.A., Clark, P., Connolly, P., Curtis, A.S., and Wilkinson, C.D. 21 (1987): Novel methods for the guidance and monitoring of single cells and simple networks in culture. J. Cell Sci. 8(Suppl):55–79. Eastwood, M. (1996) Mechatronic devices for the investigation of fibroblast populated collagen lattices and the application of precise mechanical loads. Ph.D. thesis, University of London. Eastwood, M., McGrouther, D. A., and Brown, R. A. (1994): A culture force monitor for measurement of contraction forces generated in human dermal fibroblast cultures: Evidence for cell-matrix mechanical signalling. Biochim. Biophys. Acta. 1201:186–192. Eastwood, M., Porter, R., Khan, U., McGrouther, D.A., and Brown, R.A. (1996): Quantitative analysis of collagen gel contractile forces generated by dermal fibroblasts and the relationship to cell morphology. J. Cell. Physiol. 166:33–45. Ejim, O.S., Blunn, G.W., and Brown, R.A. (1993): Production of artificial-orientated mats and strands from plasma fibronectin: A morphological study. Biomaterials 14:743–748. Greenberg, G., and Boyde, A. (1993): Novel method for stereo imaging in light microscopy at high magnifications. Neuroimage 1:121– 128. Grinnell, F. (1994): Fibroblasts, myofibroblasts, and wound contraction. J. Cell. Biol. 124: 401–404. Guido, S., and Tranquillo, RT. (1993): A methodology for the systematic and quantitative study of cell contact quidance in orientated collagen gels. Correlation of fibroblast orientation and gel birefringence. J. Cell Sci. 105:317–331. He, Y., and Grinnell, F. (1994): Stress relaxation of fibroblasts activates a cyclic AMP signaling pathway. J. Cell. Biol. 126:457–464. Huang, D., Chang, T.R., Aggarwal, A., Lee, R.C., and Ehrlich, H.P (1993): Mechanisms and dynamics of mechanical strengthening in ligament equivilent fibroblast populated collagen matrices. Ann. Biomed. Eng. 21:289–305. Kummer B (1986): The musclo-skeletal system in man: Development, structure and function in dependance of gravity, and potential limitations for long term space flights. Adv. Space. Res. 6:323–330. Lambert, C. A., Soudant, E. P., Nusgens, B. V., and Lapiere, C. M. (1992): Pretranslational regulation of extracellular matrix molecules and collagenase expression in fibroblasts by mechanical forces. Lab. Invest. 66:444–451. Margolis, L.B., and Popov, S.V. (1991): Induction of cell processes by local force. J.Cell.Sci. 98:369–373. Mochitate, K., Pawelek, P., and Grinnell, F. (1991): Stress-relaxation of contracted collagen gels: Disruption of actin filament bundles, release of surface fibronectin, and down-regulation of DNA and protein synthesis. Exp. Cell. Res. 193:198–207. Squier, C.A., and Magnes, C. (1983): Spatial relationships between fibroblasts during the growth of rat-tail tendon. Cell Tissue Res. 234:17–29. Tomasek, J.J., Haaksma, C.J., Eddy, R.J., and Vaughan, M.B. (1992): Fibroblast contraction occurs on release of tension in attached collagen lattices: Dependency on an organised actin cytoskeleton and serum. Anat. Rec. 232:359–368. Trelstad, R.L., and Birk, D.E. (1985): The fibroblast in morphogenesis and fibrosis: Cell topography and surface related functions. Ciba. Found. Symp. 114:4–19. Wojciak-Stothard,B., Curtis, A., Monaghan, W., MacDonald, K., and Wilkinson, C. (1996): Guidance and activation of murine macrophages by nanometric scale topography. Exp. Cell Res. 223:426–435. Wolff, J. (1896): ‘‘The Law of Bone Remodelling.’’ Berlin: Springer (1986 translation of the original 1896 manuscript).