- Email: [email protected]
Sensitivity of Fibroblasts and Their Cytoskeletons to Substratum Topographies: Topographic Guidance and Topographic Compensation by Micromachined Grooves of Different Dimensions
EXPERIMENTAL CELL RESEARCH ARTICLE NO.
234, 413–424 (1997)
EX973625
Sensitivity of Fibroblasts and Their Cytoskeletons to Substratum Topographies: Topographic Guidance and Topographic Compensation by Micromachined Grooves of Different Dimensions Carol Oakley,* Nicolas A. F. Jaeger,† and Donald M. Brunette*,1 *Department of Oral Biology and †Department of Electrical Engineering, University of British Columbia, 2199 Wesbrook Mall, Vancouver, British Columbia V6T 1Z3, Canada
out the use of drugs such as cytochalasin and colcemid. Fibroblasts alter their shape, orientation, and direction of movement to align with the direction of micromachined grooves, exhibiting a phenomenon termed topographic guidance. In this study we examined the ability of the microtubule and actin microfilament bundle systems, either in combination with or independently from each other, to affect alignment of human gingival fibroblasts on sets of micromachined grooves of different dimensions. To assess specifically the role of microtubules and actin microfilament bundles, we examined cell alignment, over time, in the presence or absence of specific inhibitors of microtubules (colcemid) and actin microfilament bundles (cytochalasin B). Using time-lapse videomicroscopy, computer-assisted morphometry and confocal microscopy of the cytoskeleton we found that the dimensions of the grooves influenced the kinetics of cell alignment irrespective of whether cytoskeletons were intact or disturbed. Either an intact microtubule or an intact actin microfilament-bundle system could produce cell alignment with an appropriate substratum. Cells with intact microtubules aligned to smaller topographic features than cells deficient in microtubules. Moreover, cells deficient in microtubules required significantly more time to become aligned. An unexpected finding was that very narrow 0.5-mm-wide and 0.5-mmdeep grooves aligned cells deficient in actin microfilament bundles (cytochalasin B-treated) better than untreated control cells but failed to align cells deficient in microtubules yet containing microfilament bundles (colcemid treated). Thus, the microtubule system appeared to be the principal but not sole cytoskeletal substratum-response mechanism affecting topographic guidance of human gingival fibroblasts. This study also demonstrated that micromachined substrata can be useful in dissecting the role of microtubules and actin microfilament bundles in cell behaviors such as contact guidance and cell migration with-
1 To whom reprint requests should be addressed. Fax: (604) 8226698. E-mail: [email protected].
q 1997 Academic Press
INTRODUCTION
Fibroblasts can respond to physical cues in their environment such as the topography of their substratum. In a process termed contact [1] or topographic [2] guidance, cells alter their shape, orientation, and polarity of movement to align with features of the substratum topography such as the grooves and ridges of micromachined substrata. Most studies have been limited to cells that were already oriented with their substratum topography, and current theories of contact guidance focus on the cytoskeleton, namely stress fibers or actin MFBs2 [3, 4] and focal contacts [5]. One approach to investigating the role of cytoskeletal events in cell alignment is to determine which element aligns first with the surface features. When we examined the sequence of cytoskeletal alignment in HGFs spreading on micromachined grooves, MTs aligned first, at about 20 min in most cells, followed by MFBs, focal contacts, and even the cells themselves [6]. A direct test of the role of MTs as the primary determinant of HGF alignment was to observe cell spreading and alignment in the presence of colcemid, a specific inhibitor of MTs. Surprisingly, MT-deficient HGFs aligned, polarized, and migrated on micromachined grooves, a process we termed topographic compensation [7], but not on smooth surfaces. However, alignment, polarization, and directed migration on the grooved substrata were significantly delayed in the presence of colcemid, suggesting that although MTs were not essential for these processes, they still influenced these behaviors. More recently, Wojciak-Stothard et al. [8] studied the influence of cytoskeletal elements on topographic guid2 Abbreviations used: BHK cells, baby hamster kidney cells; CLSM, confocal scanning microscope; CB, cytochalasin B; FITC, fluorescein isothiocyanate; HGFs, human gingival fibroblasts; MFBs, microfilament bundles; MTs, microtubules.
413
AID
ECR 3625
/
6i24$$$161
07-15-97 08:12:51
0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
eca
414
OAKLEY, JAEGER, AND BRUNETTE
ance on micromachined surfaces at shorter times, using confocal microscopy and inhibitors of cytoskeletal polymerization in BHK fibroblasts. Although aligned MFBs were not reported at early times, the aggregation of actin along groove/edge boundaries was observed as early as 5 min and they concluded that actin condensations were the primary driving event of topographic guidance. Oakley and Brunette [6] also observed actin condensations along the ridge/groove interface up to 2 h after plating, at which time only 40% of cells demonstrated aligned MFBs, yet 80% of cells demonstrated aligned MTs. Another approach to investigating the role of the cytoskeleton in cell alignment is to vary the size of the topographic features in order to determine whether there are conditions under which MTs or MFBs are affected differentially and whether cell alignment can be obtained under such conditions. As the response of either MTs or MFBs could be affected by the presence of the other, we observed cell spreading and alignment in the absence (controls) or presence of specific inhibitors of MTs (colcemid) and MFBs (cytochalasin B) on smooth control surfaces and on three different sets of grooves that were identical in shape but differed in size. Thus, behaviors of cells with an intact cytoskeleton (controls) could be compared to MT-deficient/MFBcontaining colcemid-treated cells and MT-containing/ MFB-deficient cytochalasin B-treated cells. Using confocal microscopy and computer-assisted morphometry to examine HGF response at different time intervals, we determined that groove spacing significantly influenced the kinetics of cell alignment, irrespective of whether cytoskeletons were intact or perturbed. On the narrowest 0.5-mm-wide grooves, alignment of cytochalasin B-treated (MT-containing/MFB-deficient) cells was superior to alignment of cells with an intact cytoskeleton (controls), and colcemid-treated (MT-deficient/MFB-containing) cells failed to align altogether, behaving as if on a smooth surface. These findings suggest that in HGFs, the MT system may be the principal but not sole cytoskeletal component affecting topographic guidance and that substrata topographies can be produced that differentially affect the MT and MFB systems. MATERIALS AND METHODS Materials and methods are similar to those described previously [6, 7] and are reviewed briefly below. Cell culture, treatment conditions, and time-lapse observations. Fibroblasts between the 4th and 12th subcultures, isolated from human gingival explants [9] were cultured in a-minimal essential medium (MEM) (Stemcell, Vancouver, BC) supplemented with antibiotics [100 mg/ml penicillin G (Sigma, St. Louis, MO), 50 mg/ml gentamicin (Sigma), 3 mg/ml amphotericin B (Fungizone, Gibco, Grand Island, NY)] and 15% bovine serum (Calf Supreme, Gibco) at 377C in a humidified atmosphere with 5% CO2 .
AID
ECR 3625
/
6i24$$$162
07-15-97 08:12:51
HGF cells were removed from their growth surface by a trypsin solution [0.25% trypsin (Gibco) and 0.1% glucose dissolved in citrate– saline (pH 7.8)]. Cells were resuspended in either control medium containing less than 0.1% dimethyl sulfoxide (DMSO, Fisher Scientific, Ottawa, Canada) or medium containing cytoskeleton-perturbing agents (colcemid or cytochalasin B [CB]) solubilized in DMSO (õ0.1%) (Technical Information Bulletin AL-126, Aldrich Chemical Co., Inc., Milwaukee, WI). A stock solution of 100 mg/ml colcemid (Sigma), solubilized in DMSO, was prepared in medium so that the final concentration of DMSO was less than 0.1%. An appropriate volume of the colcemid stock solution was added to the medium used for resuspension of the cells so that a final concentration of 1.0 mg/ml colcemid was obtained. This concentration of colcemid was based on the work of Hollenbeck et al. [10] and is the same concentration used previously [7]. A stock solution of 1.0 mg/ml CB (Sigma), solubilized in DMSO, was prepared in medium and an appropriate volume of CB stock solution was added to the medium used for resuspension of the trypsinized cells so that a final concentration of 2.0 mg/ml CB was obtained. This concentration of CB was used in the present study because titration experiments indicated that this was the lowest CB concentration at which the vast majority of HGFs plated and maintained in the presence of CB displayed arborized MT-containing processes and no longer displayed MFBs [e.g., 11]. These titration experiments are discussed in detail in the Ph.D. thesis of Carol Oakley (1995, University of British Columbia, Canada). CB was used for time-lapse observations and for cytoskeletal observations in order to compare results to earlier reports of fibroblasts in the presence of CB [11–13]. However, cytochalasin D was used as an additional control (data not shown) to verify that the cytoskeletal behaviors elicited in the presence of CB were due to their action upon the actin system and were not related to the interference with sugar transport associated with CB [14]. The resuspended cells were seeded onto smooth and grooved titanium substrata at different cell populations in order to obtain cells which were spatially well isolated at the different time periods. For control conditions, population densities of 1 1 105 cells/ml were used for experiments lasting up to 2 h, 1 1 104 cells/ml for 6 h, and 1 1 103 cells/ml for 24 h. For drug conditions, cell densities of 1 1 105 cells/ml were used for 2-h experiments and 1 1 104 cells/ml were used for 24-h experiments. Cell suspensions of 1 1 105 cells/ml were used for all time-lapse observations, regardless of culture conditions or length of observation. For these observations, cell suspensions were injected into a Pentz chamber (Bachofer, Reutlingen, Germany) in which the substrata had been mounted. Cells were observed through a microscope (Reichert, Austria) equipped with reflected Nomarski differential interference contrast (DIC) optics and an Epiplan 8X or 16X objective (Zeiss, Oberkochen, Germany). Cell behavior was recorded using a television camera and control system (Hamamatsu model C2400; Japan) and a time-lapse video recorder (Panasonic Model 8050). Micromachined substrata. Titanium-coated micromachined silicon wafers were produced as described previously by Brunette [15]. A computer-generated pattern was produced on a photomask and transferred onto silicon wafers by photolithography and anisotropic etching, producing a series of grooves. In this study three different groove patterns were used. The grooves all had the same shape but differed in pitch and depth. In cross section the grooves had a truncated V shape and the interior corner of the ridge and wall of the groove formed an angle of 1257. The first groove pattern consisted of a series of 3-mm-deep grooves with a 30-mm pitch comprising a 15mm-wide groove and 15-mm-wide ridge. This pattern had been used in previous studies [6, 7, 16] and was termed the reference (R) groove pattern. The second pattern consisted of a series of 3-mm-deep grooves with a pitch that ranged from 6 to 9 mm in different areas of the silicon wafer. This pattern was designated the narrow (N) groove pattern. The third pattern consisted of 0.5-mm-deep grooves
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION with a 1.0-mm pitch and was termed the very narrow (VN) groove pattern. Both the grooved and the smooth silicon wafers were sputter coated with 50 nm of titanium, a procedure that renders the surface chemically homogeneous, as noted by Singvhi et al. [17]. The surfaces were cleaned by ultrasonication in a detergent formulated for tissue culture (71, ICN Biomedicals, Inc., Costa Mesa, CA) and then rinsed copiously and ultrasonicated in filtered deionized water and air-dried in a laminar-flow hood. Final cleaning and sterilization were accomplished by argon-gas glow-discharge treatment (18) just prior to seeding with the cell suspension. Immunofluorescence. At 2, 6, and 24 h and after all time-lapse observations, samples were rinsed with cytoskeletal stabilizing (CS) buffer (19), fixed for 10 min in 3.7% formaldehyde (BDH Inc., Toronto, Canada) in phosphate-buffered saline (PBS), and washed in CS buffer. To visualize cell morphology, some samples were prepared for incubation with fluorescent hydrazides to stain the cell membrane. Following oxidation in 4.2 mM sodium periodate (BDH Inc.), these samples were exposed to fluorescein-5-thiosemicarbazide (Molecular Probes Inc., Eugene, OR). Subsequently, preparation for staining of cytoskeletal elements was continued in all samples by permeabilization in 0.5% Triton X-100 (Sigma), quenching with 0.05% sodium borohydride (BDH Inc.), blocking with 1% bovine serum albumin (BSA, Sigma) and exposure to one or more of the following antibodies or fluorescent stains: Bodipy fluorescein phallacidin or rhodamine phalloidin (Molecular Probes Inc.) or monoclonal anti-b-tubulin (Boehringer–Mannheim GmbH, Mannheim, Germany) followed by sheep anti-mouse IgG conjugated to Texas red (Molecular Probes Inc.) or fluorescein isothiocyanate (FITC; Boehringer–Mannheim GmbH). All antibodies were diluted with PBS containing 0.1% BSA. Epifluorescence and confocal microscopy. Cells stained with the hydrazide were examined through a microscope equipped with epifluorescent illumination (confocal laser scanning microscope (CLSM), Zeiss) and photographed (Olympus PM4T camera, Kodak Tmax 400). Fields were selected randomly and the only criterion for cell inclusion in the sample was the absence of contact of a cell with other cells. Subsequently, images of the photographed cells were projected and cell outlines as well as corresponding groove–ridge patterns, where applicable, were traced onto paper. The tracings were retraced on the digitizing tablet of a morphometrics system (Videoplan, Zeiss) which calculated a form factor, Form Ell, that is the ratio of the minor to major axes. The longest diameter between two edges within the cell was considered the major axis and the minor axis was considered the longest axis perpendicular to the major axis. A form Ell value of 1.0 corresponds to a circular shape, and values less than 1 indicate elliptical or spindle shapes. The angle formed by the major axis to either the grooves or, for smooth surfaces, a line with randomly selected orientation was termed the orientation angle. For measurement of cell orientation no attempt was made to distinguish cell polarity and therefore the maximum angle possible was 907, with a value of 07 indicating a cell that was perfectly aligned with the direction of the grooves. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves [20]. Cell alignment was not contingent upon cell polarity and cell polarity did not imply alignment. Cell polarity was also used to describe cell shape, as in the approach of Glasgow and Daniele [21]. Cells were termed monopolar when there was a clearly identifiable leading lamella. Bipolar cells contained two distinct lamellae and apolar cells were round. Cytoskeletal elements were examined with both epifluorescence and confocal microscopy using a 63X oil-immersion plan-apochromat objective, NA 1.4 (Zeiss) on the CLSM (Zeiss). An argon laser (lmax Å 488 nm) was used for Bodipy fluorescein and FITC-labeled cells in confocal microscopy; a helium–neon laser (lmax Å 543 nm) was used for Texas Red-labeled cells. For double-labeled specimens, the Bodipy or FITC signal was examined first, before the Texas Red signal. Digital gray-scale images obtained by the CLSM were printed
AID
ECR 3625
/
6i24$$$162
07-15-97 08:12:51
415
using a video printer (UP-5000, Sony Canada, Richmond, British Columbia). Descriptions of cytoskeletal patterns and criteria used to determine cytoskeletal alignment followed those outlined in [6]. Statistics. Morphometric measurements were based on a mean number of 50 cells at 2 and 6 h and 83 cells at 24 h for each of the smooth and grooved surfaces, under control, colcemid, and CB conditions. Data were entered into a mainframe computer (IBM) and analyzed using SPSS-X (SPSS Inc., Chicago, IL). Analysis of variance (ANOVA) and Student–Newman–Keuls (SNK) were used to analyze the effects of surface, time, and treatment conditions on Form Ell and orientation angle. The proportions of aligned cells were expressed as a percentage and a nonparametric test, Kruskal–Wallis one-way analysis of variance, was used to analyze the percentages of aligned cells, as these data were not normally distributed.
RESULTS
This study, admittedly complex, includes observations of HGFs under 12 conditions over time; that is, observations were made on each of four substrata: smooth, R, N and VN, and under each of three culture conditions: control, colcemid, and cytochalasin B. To facilitate presentation of the data, observations are categorized by treatment condition (i.e., control, CB, colcemid) in terms of the MT or MFB content. Thus control or CB-treated cells would contain MTs and be designated MT/, whereas colcemid-treated cells would be designated MT0. The three groups were thus MT/MFB/ (controls), MT/MFB0 (CB-treated cells), and MT0MFB/ (colcemid-treated cells). Within each treatment group, the behaviors of cells on smooth, R-, N-, and VN-grooved surfaces are compared. Figure 1 presents an overview of the results, the appearance of cells cultured for 24 h on smooth, R, N, or VN grooves under control, CB, or colcemid conditions. In summary, all grooved surfaces oriented cells under all conditions, with the exception that VN grooves failed to orient cells when MTs were absent (i.e., colcemid-treated cells). These qualitative results are supported by quantitative data in Fig. 2, the percentage of aligned cells on each substratum as a function of time for each treatment group. In brief, under control conditions, alignment occurred in the order N ú R ú VN, whereas in CB-treated (MFB-deficient) cells, N and VN grooves oriented cells more quickly than R grooves, and surprisingly, produced greater orientation than that observed under control conditions. Colcemid-treated MT-deficient cells aligned best on N grooves followed by R grooves, but MT-deficient cells did not align on VN grooves. Moreover, in order to facilitate comparisons of cell alignment and cell shape among treatment groups, orientation angle (Fig. 3a) and Form Ell (Fig. 3b) are presented for all treatment groups together at 24 h, the longest time of observation. In addition, a qualitative comparison of directed motility by cells within each treatment group on each substratum is presented in Table 1. Finally, the typical appearance
eca
416
OAKLEY, JAEGER, AND BRUNETTE
FIG. 1. Light micrographs of HGFs spread 24 h on smooth (a, e, i), R- (b, f, j), N- (c, g, k), or VN-grooved (d, h, l) substrata under control (a–d), CB (e–h), or colcemid (i–l) conditions. Cells were stained with fluorescent hydrazide to stain the cell membrane. Thin lamellar and tail regions stained less brightly than the body of the cells. On smooth surfaces, cells showed no particular orientation, whereas cells on R and N grooves aligned to the direction of the grooves (double-headed arrows). On VN grooves, only controls and CB-treated cells aligned with the grooves.
of the actin and MT systems within each treatment group on smooth, R, N, and VN grooves is presented in Figs. 4 – 6. As the focus of this study is the topographic responses of entire cells, Figs. 4 – 6 are intended to provide a composite overview of cytoskeletal behavior under the various culture and substrata conditions rather than a detailed analysis of cytoskel-
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
etal interaction with the grooves. Detailed cytoskeletal patterns and behaviors of HGFs on smooth and R grooves under control [6] and colcemid [7] conditions have been published previously. A more detailed description of the behavior of control and treated cells as a function of time on the four surfaces is given below.
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
417
On grooved surfaces, HGFs generally spread radially across the ridges and grooves and gradually elongated in a direction parallel to the grooves but some differences between the types of grooves were seen in the kinetics of alignment and in the relationship of cells to the grooves. On R grooves, cells generally became aligned within one groove, whereas on N and VN grooves cells bridged across several grooves (Figs. 1b–
FIG. 2. Percentage of aligned HGFs plotted against time for fibroblasts plated 2, 6, and 24 h on smooth and R, N, and VN grooves under control (a), CB (b), or colcemid (c) conditions. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves or, for smooth surfaces, to a line of randomly selected orientation. The proportion of aligned cells at each time interval was expressed as a percentage. (a) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VNgrooved substrata under control conditions, in the absence of CB, or colcemid. Alignment was fastest and greatest on N grooves, followed by R grooves and VN grooves. (b) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VN-grooved substrata under CB conditions. Alignment was fastest and greatest on VN and N grooves, followed by R grooves. (c) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VN-grooved substrata under colcemid conditions. Alignment was fastest on N grooves, followed by R grooves but alignment on N and R grooves was similar at 24 h. The small proportion of cells classified as aligned on VN grooves is similar to the proportion of cells classified as aligned on smooth surfaces.
Control Cells (MT/MFB/ Cells) Under control conditions, in the absence of colcemid or CB, the actin MFB and MT systems were intact (Fig. 4). On the smooth surface, HGFs became monopolar and migrated across the surface. Cell orientation was essentially random, not differing significantly from the value of 457C expected if no orienting influences were present (Fig. 3a).
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
FIG. 3. (a) Orientation angle of HGFs 24 h after plating on smooth and R-, N-, or VN-grooved substrata plotted as a function of treatment (control, CB, or colcemid). Error bars { SEM. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves or, for smooth surfaces, to a line of randomly selected orientation. In the absence of any orientating influence, the orientation angle of cells on smooth surfaces did not differ greatly from the expected value of 457. On R and N grooves, cells under all treatment conditions were significantly (P õ 0.05) more aligned than on smooth surfaces. On VN grooves, only control and CB-treated cells were aligned but controls on VN grooves were significantly (P õ 0.05) less aligned than controls on R or N grooves or CB-treated cells on R, N, or VN grooves. On VN grooves, colcemidtreated cells were poorly aligned. (b) Form Ell of HGFs 24 h after plating on smooth or R-, N-, or VN-grooved substrata plotted as a function of treatment (control, CB, or colcemid). Error bars { SEM. Form Ell is a ratio of the minor and major axes of a cell and is an indicator of shape. A value of 1.0 corresponds to a circular shape and values less than 1.0 indicate elliptical or spindle shapes. On smooth surfaces, controls were significantly (P õ 0.05) more elliptical than colcemid- and CB-treated cells. On R and N grooves, cells were significantly (P õ 0.05) more elliptical than cells on smooth surfaces under the same treatment condition. On VN grooves, controls were the most elliptical (P õ 0.05), followed by CB-treated cells; colcemidtreated cells were round.
eca
418
OAKLEY, JAEGER, AND BRUNETTE
TABLE 1 Qualitative Ranking of Persistent and Directed Cell Migration
Smooth 30 mm (R) 6–9 mm (N) 1 mm (VN)
Control
CB
Colcemid
/ ///// ///// //
NM NM NM NM
NDM //// /// NDM
Note. On smooth surfaces and VN (1.0-mm pitch) grooves, only control cells were motile and the persistence of their directed motility along the grooves was less than that observed on R (30-mm pitch) or N (6- to 9-mm pitch) grooves. No motility (NM) was observed for CBtreated cells on any surface and no directional motility (NDM) was observed for colcemid-treated cells on smooth or VN surfaces. In contrast, migration along the grooves was observed on R and N grooves for controls and colcemid-treated cells although both continuity and persistence of forward locomotion differed with treatment and surface.
1d). Although cell alignment was fastest on N grooves (Fig. 2a), at 24 h alignment on N and R grooves was comparable (Figs. 2a and 3a) and significantly (P õ 0.05) greater than alignment on VN grooves (Fig. 3a). At 24 h, the shape of cells on grooved surfaces was typically elliptical (Fig. 3b) and monopolar but, again, some differences in cell morphology and alignment were seen between the types of grooves. On R and N grooves, leading edges, cell bodies, and tails of cells were typically aligned with the grooves, producing a very elongated cell shape (Figs. 1b and 1c). On VN grooves, cell shape and alignment were more variable than on N or R grooves; orientation of leading edges was frequently along the direction of the VN grooves, whereas alignment of cell bodies and tails was less uniform (Fig. 1d). On grooves, cells exhibited topographic guidance but otherwise traveled in a manner similar to that of cells moving on smooth surfaces. One difference in cell behavior dependent on topography was that cells on VN grooves showed less directional persistence than cells on R or N grooves (Table 1). Cell migration on R and N grooves also resulted in a larger displacement over a given time, relative to cells on VN grooves (Table 1) or on the smooth surface [22]. CB-Treated Cells (MT/MFB0Cells) CB (2.0 mg/ml) was used to inhibit the formation of actin MFBs (Fig. 5). At this concentration of CB, HGFs plated and maintained in the presence of CB attached and spread and the vast majority of HGFs extended cell processes. Actin MFBs were not observed but small foci of actin-positive material conformed to the outlines of cell processes that contained bundles of MTs (Fig. 5). On smooth surfaces (Figs. 1e, 6a and 6b), the cell
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
processes which contained bundles of MTs (Fig. 5) demonstrated no preferred orientation (Figs. 2b and 3a) but on grooves, the processes aligned with the direction of the grooves (Figs. 1f–1h, 5). CB-treated cells were not motile. Alignment of CB-treated cells on N and VN grooves was faster (Fig. 2b) and superior (Figs. 2b and 3a) to alignment on R grooves. Surprisingly, alignment of CBtreated cells on N and VN grooves exceeded alignment of controls at 2 h (Figs. 2a and 2b). Furthermore, at 24 h the overall alignment of CB-treated cells on R, N, and VN grooves was significantly (P õ 0.05) better than alignment of controls on VN grooves (Fig. 3a). Groove depth appeared to be a less important factor in determining alignment of CB-treated cells than groove/ ridge spacing. Shallow 0.5-mm-deep VN grooves produced alignment that occurred earlier but was comparable to that on the 3.0-mm-deep N grooves and both VN and N grooves produced better alignment than 3.0mm-deep, wider R grooves (Figs. 2b and 3a). Thus elimination of the MFBs by CB apparently increased the orientation response of cells to the N and VN grooves which produced faster and greater alignment than R grooves. Colcemid-Treated Cells (MT0MFB/ Cells) Colcemid was used to eliminate the MTs (Fig. 6). As previously reported, R grooves were able to effect topographic compensation, i.e., alignment, polarization, and directed migration of MT-deficient cells (7). Topographic compensation was also effected by N grooves; however, a smaller proportion of colcemidtreated cells was monopolar and directionally motile on N grooves as compared to R grooves. In addition, the directed migration of some colcemid-treated cells along N grooves appeared to be more intermittent than locomotion on R grooves (Table 1). By 24 h, alignment of colcemid-treated cells on R and N grooves (Figs. 2c and 3a) was comparable, although cells were significantly (P õ 0.05) less elliptical on N grooves than on R grooves (Fig. 3b). As expected from our previous study, colcemid-treated cells required significantly (P õ 0.05) more time to align on R (7) and N grooves than did controls (Figs. 2a and 2c). There were also differences between the grooved surfaces as colcemid-treated cells aligned more quickly on N grooves than on R grooves (Figs. 2a and 2c). Thus, at early times of plating, the N-grooved substratum was the most efficacious in promoting cell alignment in the presence or absence of MTs. On VN grooves (Fig. 1l), colcemid-treated cells behaved like colcemid-treated cells on smooth surfaces. That is, alignment (Figs. 2c and 3a), elliptical cell shapes (Fig. 3b), and directed cell migration (Table 1) did not develop. Although a few cells demonstrated
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
419
FIG. 4. CLSM-generated images of HGFs 24 h after plating on smooth (a, b), R- (c, d), N- (e, f), or VN- (g, h) grooved substrata under control conditions, in the absence of colcemid or CB. Groove directions indicated by double-headed arrows. Cells stained both for actin (a, c, e, g) and tubulin (b, d, f, h). All cells were monopolar and both actin–MFB and tubulin distributions reflected the overall shape of cells on smooth (a, b) and overall shape and alignment of cells on grooves (c–h). On R grooves (c, d) cells were typically aligned within one groove, whereas on N (e, f) and VN (g, h) grooves cells bridged laterally over several grooves and ridges. On VN grooves, MTs (h) were typically better aligned with the grooves than actin–MFBs. Some MFBs (arrowheads in g) appeared to be oriented across the grooves.
AID
ECR 3625
/
6i24$$3625
07-15-97 08:12:51
eca
FIG. 5. CLSM-generated images of HGF cells 24 h after plating on smooth (a, b), R- (c, d), N- (e, f), or VN- (g, h) grooved substrata under CB conditions. Groove directions indicated by double-headed arrows. Cells stained for both actin (a, c, e, g) and tubulin (b, d, f, h). Actin–MFBs were not evident but small foci of actin-positive material conformed to the outlines of cell processes that contained bundles of MTs. MT bundles radiated from a central source and on smooth surfaces MT bundles did not appear to have a preferred orientation (b). In contrast, on grooves the bundles aligned with the direction of the grooves.
07-15-97 08:12:51
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
421
FIG. 6. CLSM-generated images of HGF cells 24 h after plating on smooth (a, b), R (c, d), N (e, f), or VN (g, h) grooves under colcemid conditions. Groove directions indicated by double-headed arrows. Cells stained for both actin (a, c, e, g) and tubulin (b, d, f, h). A fibrillar network of tubulin staining was not observed. The distribution of actin MFBs reflected the shape of the cells and was similar to MFB distributions observed in controls with similar shapes on both smooth and grooved surfaces. On R (c) and N (e) grooves, actin MFBs reflected the alignment of the cells and both cells and MFBs were aligned to the direction of the grooves. In contrast, alignment of cells and their actin MFBs on VN grooves (g) was similar to that observed on smooth (a) surfaces.
AID
ECR 3625
/
6i24$$3625
07-15-97 08:12:51
eca
422
OAKLEY, JAEGER, AND BRUNETTE
aligned shapes, cell shapes were typically bipolar or apolar and changed frequently. Thus, using the criterion of alignment of MFBs to the grooves, the actin MFB system was not consistently influenced by VN grooves. DISCUSSION
The results of this study enabled us to distinguish between the roles played by the MT and actin–MFB systems in some aspects of topographic guidance and topographic compensation [7]. Whereas the actin system is essential for powering cell locomotion, it has been accepted that on smooth surfaces MTs are required for stabilizing cell shape and effecting polarization [e.g., 23]. In contrast, on some anisotropic substrata (i.e., R or N grooves) the presence of either an intact MT or an intact actin–MFB system appears to be sufficient to elicit alignment of cells to the substratum topography. However, the MT and actin–MFB systems differed in their responses to the substratum, both in the rates of effecting cell alignment and in their sensitivity to dimensions of the substratum topography. In general, MT-containing cells aligned faster with grooves of smaller dimensions (such as the VN grooves) than did MT-deficient but MFB-containing (i.e., colcemid-treated) cells. Thus, the MT system appeared to be the principal but not the sole cytoskeletal substratum-response mechanism effecting cell alignment on anisotropic substrata. However, given enough time, colcemid-treated cells lacking MTs aligned, polarized, and migrated on R and N grooves, a process termed topographic compensation [7]. Significantly, not all surfaces that can affect contact guidance in cells with an intact cytoskeleton can effect topographic compensation [7, 24]. That is, in contrast to their behavior on R and N surfaces, colcemid-treated (MT0MFB/) HGFs did not exhibit topographic compensation on VN grooves. In fact, the behavior of these MT-deficient cells on VN-grooved surfaces was similar to that on smooth surfaces. Thus, topographic compensation appeared to be associated with aligned MFBs that were observed in MT-deficient cells on the R- and N-grooved surfaces, but were only rarely observed in cells on the VN surfaces. By comparison, CB-treated cells which were MFB deficient could align on R, N, and VN grooves, although they could not move on the smooth or any of the grooved surfaces. Therefore, both the microtubular and the actin systems are required to obtain a full topographic guidance response from HGFs over the range of dimensions used in this study. The cellular response to substratum topography appears to involve some interaction between the MT and actin–MFB systems. This interaction is particularly evident on VN-grooved surfaces where the removal of MFBs by CB causes an increase in alignment of cells
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
and their processes with the grooves. Under control and colcemid (MT0MFB/) conditions, these VN grooved substrata did not significantly affect the MFB–actin system as assessed by the paucity of MFBs aligned with the substratum grooves. It appears that under control conditions, when the actin–MFB system is intact and presumably exerting its own effects on cell shape and translocation, the actin–MFB system diminishes the effects produced by the substratum topography on the MT system. Such an interaction of the MT and MFB systems might be expected because the intracellular course or placement of MTs is influenced by structural interactions with actin [25] as well as by tension generated by the actin network [25, 26]. Recently, Wojciak-Stothard et al. [8] used BHK fibroblasts to investigate the role of cytoskeletal systems in topographic reactions. They concluded that the aggregation of actin along groove/ridge boundaries was the primary driving event in determining BHK fibroblast orientation on microgrooved substrata. There are three germane differences between our study and that of Wojciak-Stothard et al. [8] that may explain the difference in conclusions. The first difference between the studies was that in our study, HGFs resuspended and maintained in the presence of cytochalasin developed arborized cell processes within 2 h after exposure to CB. Numerous studies have shown that exposure of cells to cytochalasin results in a decrease or cessation of ruffling behavior and cell motility; the cell body rounds up and numerous processes or branches are extended, giving the cells a stellate or arborized shape [e.g., 12–14, 27, 28]. In fibroblasts, cytochalasins do not depolymerize actin filaments but, rather, disrupt the supramolecular organization of actin filaments, and the filaments persist, albeit in dense focal accumulations [27]. As the cortical actin–MFB network is destroyed by cytochalasins, the restraining influence of the cortex is reduced and MTs are able to push out the pliable plasma membrane and traverse straight and uninterrupted to the distal ends of the arborized processes [11, 29]. In contrast, the CD-treated BHK fibroblasts in the Wojciak-Stothard et al. [8] study do not appear to be arborized at 24 h (see, for example, Fig. 1d of their paper which differs from Figs. 1f–1h of this study). However, BHK cells did adopt a more arborized appearance when exposed to cytochalasin for periods longer than 24 h (Curtis, personal communication, 1995). Thus, there are significant differences in the dynamics of the cytoskeleton in the cell types used in the two studies. There are several levels of supramolecular actin organization. Phallotoxins stain only filamentous actin and in both of these light microscopic studies only bundles of microfilaments were visualized; therefore, we cannot comment on the possible role of individual actin filaments in cell alignment. Nevertheless, as arboriza-
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
423
tion occurred in our CB-treated cells it appears that VN grooves. If a substratum such as R or N grooves the subcortical actin matrix was disrupted. Moreover, elicited alignment of both MT and actin–MFB systems, prominent MFBs were not present. In such circum- overall alignment was facilitated and enhanced. Constances the ability of the MT system to effect cell align- versely, if a substratum such as VN grooves failed to ment on R, N, and VN grooves independently of any direct alignment of the actin–MFB system, the influinfluence of functionally organized MFBs could be dem- ence of the MT system upon cell alignment was reduced onstrated. but not eliminated. A second difference between the work of WojciakA major finding of this study is that substrata can Stothard et al. [8] and this study lies in the chemical be designed to affect the MT and actin–MFB systems composition of the substrata. In this study, micromach- differentially. Such substrata may be useful in disined substrata were sputter coated with titanium secting the role of the MT and actin–MFB systems in [15, 30, 31], a process which, as noted by Singhvi et al. cell behaviors such as topographic guidance and cell [17], provides a completely homogeneous surface and migration without the use of drugs such as CB and avoids the possible problem of chemical heterogeneity colcemid with their attendant possibilities of nonspethat can be produced by etching methods [17]. Tita- cific effects. Understanding the principles of cytoskelenium was selected because of its biocompatability [32]. tal interaction with the substratum topography may In contrast, Wojciak-Stothard et al. [8] used a two-stage permit exploitation of certain behaviors to uncouple or dry-etching procedure on fused silica so that all of the amplify MT and actin-system responses and permit the surface was exposed to etching and thus could be con- design of substrata that specifically retard or enhance sidered as having a uniform surface chemistry. Never- cell alignment and migration. One practical application theless, the surface chemistries of the substrata doubt- of such substratum-mediated control of cell behaviors less differ between the studies. may be in the design of implanted devices where control A third difference between our study and that of Woj- of cell migration would be expected to improve device ciak-Stothard et al. [8] is the range of groove dimen- performance [31]. Ultimately, the substrata of biomatesions examined. Groove dimensions must be considered rials used for prosthetic devices and implants could be in relation to the typical size of the cells being investi- engineered to create specific cytoarchitectures which, gated [17, 33]. Our R and N grooves are comparable to in turn, could produce specific cell responses. the 10- and 5-mm-wide grooves, respectively, used by Wojciak-Stothard et al. [8]. Even though the cell types The authors thank Mr. Hiroshi Kato for technical assistance with differed, very similar orientation angles were obtained, substrata preparation. The authors are grateful to Dr. W. Vogl (Deon both our N grooves and their 5-mm-wide grooves partment of Anatomy, University of British Columbia), for advice under control, colcemid, or CB conditions. However, we and comments on this work. The authors are also indebted to Ms. Jackie McDiarmid for her diligence in tracing and digitizing the cells also employed substrata with much smaller grooves and thank Mrs. Lesley Weston and Mr. Andre Wong for technical (0.5-mm-wide VN grooves) and, significantly, it was on assistance. This work was supported by the Medical Research Counthese ultrafine grooves that differences in the threshold cil of Canada (MRC Dental Fellowship Grant 5-59108 to C. Oakley, sensitivities of the MT and actin–MFB systems were Operating Grant 5-97617 to D. M. Brunette). readily apparent. In summary, the MT system was not required for REFERENCES alignment, polarization, or directed migration if col1. Weiss, P. (1934) J. Exp. Zool. 68, 393–448. cemid-treated (MT-deficient) HGFs were provided with enough time and an appropriate substratum such as R 2. Curtis, A. S. G., and Clark, P. (1990) Crit. Rev. Biocompat. 5, 343–362. (7) or N grooves. However, on the VN-grooved sub3. Dunn, G. A., and Heath, J. P. (1976) Exp. Cell Res. 101, 1–14. strata, the presence of the MT system under CB condi4. Dunn, G. A., and Brown, A. F. (1986) J. Cell Sci. 83, 313–340. tions produced cell alignment even though MFBs were 5. Ohara, P. T., and Buck, R. C. (1979) Exp. Cell Res. 121, 235– absent. In contrast, under colcemid conditions on VN 249. grooves the presence of the actin-MFB system did not 6. Oakley, C., and Brunette, D. M. (1993) J. Cell Sci. 106, 343– produce cell alignment. Significantly, the MT system 354. responded faster and to smaller topographic features 7. Oakley, C., and Brunette, D. M. (1995) Cell Motil. Cytoskel. 31, than the actin–MFB system, and therefore we consider 45–58. that the MT system appeared to be the principal, but 8. Wojciak-Stothard, B., Curtis, A. S. G., Monaghan, W., McGrath, not the sole cytoskeletal substratum-response mechaM., Sommer, I., and Wilkinson, C. D. W. (1995) Cell Motili. Cynism in HGFs. Moreover, there is an interaction betoskel. 31, 147–158. tween actin–MFB and MT systems that affects their 9. Brunette, D. M., Melcher, A. H., and Moe, H. K. (1976) Arch. respective kinetics and sensitivities to the substratum. Oral Biol. 21, 393–400. This interaction is apparent in untreated control cells 10. Hollenbeck, P. J., Bershadsky, A. D., Pletjushkina, O. Y., Tint, I. S., and Vasiliev, J. M. (1989) J. Cell Sci. 92, 621–631. which aligned well on R and N grooves, but poorly on
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
eca
424
OAKLEY, JAEGER, AND BRUNETTE
11. Weber, K., Rathke, P. C., Osborn, M., and Franke, W. W. (1976) Exp. Cell Res. 102, 285–297. 12. Bliokh, A. L., Domnina, L. V., Ivanova, O. Y., Pletjushkina, O. Y., Svitkina, T. M., Smolyninov, V. A., Valisiev, J. M., and Gelfand, I. M. (1980) Proc. Natl. Acad. Sci. USA 77, 5919–5922. 13. Domnina, L. V., Gelfand, V. J., Ivanova, O. Y., Leonova, E. V., Pletjushkina, O. Y., Vasiliev, J. M., and Gelfand, I. M. (1982) Proc. Natl. Acad. Sci. USA 79, 7754–7757. 14. Yahara, I., Harada, F., Skeita, S., Yoshihara, K., and Natori, S. (1982) J. Cell Biol. 92, 69–72. 15. Brunette, D. M. (1986) Exp. Cell Res. 164, 11–26. 16. Oakley, C., and Brunette, D. M. (1995) Biochem. Cell Biol. 73, 473–489. 17. Singhvi, R., Stephanopoulos, G., and Wang, D. I. C. (1994) Biotechnol. Bioeng. 43, 764–771. 18. Baier, B. E., and Meyer, A. E. (1988) Int. J. Oral Maxillofac. Implants 3, 9–20. 19. Opas, M. (1989) Dev. Biol. 131, 281–293. 20. Clark, P., Connolly, P., Curtis, A. S. G., Dow, J. A. T., and Wilkinson, C. D. W. (1990) Development 108, 635–644. 21. Glasgow, J. E., and Daniele, R. P. (1994) Cell Motil. Cytoskel. 27, 88–96. 22. Damji, A., Weston, L., and Brunette, D. M. (1996) Exp. Cell Res. 228, 114–124.
23. Vasiliev, J. M. (1982) Phil. Trans. R. Soc. London 299, 159– 167. 24. Vasiliev, J. M., Gelfand, I. M., Domnina, L. V., Ivanova, O. Y., Komm, S. G., and Olshevskaja, L. V. (1970) J. Embryol. Exp. Morphol. 4, 625–640. 25. Schliwa, M., and van Blerkom, J. (1981) J. Cell Biol. 90, 222– 235. 26. Holifield, B. F., and Heath, J. P. (1992) Mol. Biol. Cell 3, 279a. 27. Schliwa, M. (1982) J. Cell Biol. 92, 79–91. 28. Sanger, J. W., and Holtzer, H. (1972) Proc. Natl. Acad. Sci. USA 69, 253–257. 29. Edson, K., Weishaar, B., and Matus, A. (1993) Development 117, 689–700. 30. Gould, T. R. L., Westbury, L., and Brunette, D. M. (1984) J. Prosthet. Dent. 52, 418–420. 31. Brunette, D. M., Kenner, G. S., and Gould, T. R. L. (1983) J. Dent. Res. 62, 1045–1048. 32. Bra˚nemark, P. I., Hansson, B. O., Adell, R., Breine, U., Lind¨ hman, A. (1977) in Osseointegrated Implants stro¨m, H., and O in the Treatment of the Edentulous Jaw: Experience for a 10Year Period, Almqvist Wiksell International, Stockholm. 33. Green, A. M., Jansen, J. A., van der Waerden, J. P. C., and von Recum, F. F. (1994) J. Biomed. Mater. Res. 28, 647–653.
Received January 30, 1997 Revised version received May 6, 1997
AID
ECR 3625
/
6i24$$$164
07-15-97 08:12:51
eca
234, 413–424 (1997)
EX973625
Sensitivity of Fibroblasts and Their Cytoskeletons to Substratum Topographies: Topographic Guidance and Topographic Compensation by Micromachined Grooves of Different Dimensions Carol Oakley,* Nicolas A. F. Jaeger,† and Donald M. Brunette*,1 *Department of Oral Biology and †Department of Electrical Engineering, University of British Columbia, 2199 Wesbrook Mall, Vancouver, British Columbia V6T 1Z3, Canada
out the use of drugs such as cytochalasin and colcemid. Fibroblasts alter their shape, orientation, and direction of movement to align with the direction of micromachined grooves, exhibiting a phenomenon termed topographic guidance. In this study we examined the ability of the microtubule and actin microfilament bundle systems, either in combination with or independently from each other, to affect alignment of human gingival fibroblasts on sets of micromachined grooves of different dimensions. To assess specifically the role of microtubules and actin microfilament bundles, we examined cell alignment, over time, in the presence or absence of specific inhibitors of microtubules (colcemid) and actin microfilament bundles (cytochalasin B). Using time-lapse videomicroscopy, computer-assisted morphometry and confocal microscopy of the cytoskeleton we found that the dimensions of the grooves influenced the kinetics of cell alignment irrespective of whether cytoskeletons were intact or disturbed. Either an intact microtubule or an intact actin microfilament-bundle system could produce cell alignment with an appropriate substratum. Cells with intact microtubules aligned to smaller topographic features than cells deficient in microtubules. Moreover, cells deficient in microtubules required significantly more time to become aligned. An unexpected finding was that very narrow 0.5-mm-wide and 0.5-mmdeep grooves aligned cells deficient in actin microfilament bundles (cytochalasin B-treated) better than untreated control cells but failed to align cells deficient in microtubules yet containing microfilament bundles (colcemid treated). Thus, the microtubule system appeared to be the principal but not sole cytoskeletal substratum-response mechanism affecting topographic guidance of human gingival fibroblasts. This study also demonstrated that micromachined substrata can be useful in dissecting the role of microtubules and actin microfilament bundles in cell behaviors such as contact guidance and cell migration with-
1 To whom reprint requests should be addressed. Fax: (604) 8226698. E-mail: [email protected].
q 1997 Academic Press
INTRODUCTION
Fibroblasts can respond to physical cues in their environment such as the topography of their substratum. In a process termed contact [1] or topographic [2] guidance, cells alter their shape, orientation, and polarity of movement to align with features of the substratum topography such as the grooves and ridges of micromachined substrata. Most studies have been limited to cells that were already oriented with their substratum topography, and current theories of contact guidance focus on the cytoskeleton, namely stress fibers or actin MFBs2 [3, 4] and focal contacts [5]. One approach to investigating the role of cytoskeletal events in cell alignment is to determine which element aligns first with the surface features. When we examined the sequence of cytoskeletal alignment in HGFs spreading on micromachined grooves, MTs aligned first, at about 20 min in most cells, followed by MFBs, focal contacts, and even the cells themselves [6]. A direct test of the role of MTs as the primary determinant of HGF alignment was to observe cell spreading and alignment in the presence of colcemid, a specific inhibitor of MTs. Surprisingly, MT-deficient HGFs aligned, polarized, and migrated on micromachined grooves, a process we termed topographic compensation [7], but not on smooth surfaces. However, alignment, polarization, and directed migration on the grooved substrata were significantly delayed in the presence of colcemid, suggesting that although MTs were not essential for these processes, they still influenced these behaviors. More recently, Wojciak-Stothard et al. [8] studied the influence of cytoskeletal elements on topographic guid2 Abbreviations used: BHK cells, baby hamster kidney cells; CLSM, confocal scanning microscope; CB, cytochalasin B; FITC, fluorescein isothiocyanate; HGFs, human gingival fibroblasts; MFBs, microfilament bundles; MTs, microtubules.
413
AID
ECR 3625
/
6i24$$$161
07-15-97 08:12:51
0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.
eca
414
OAKLEY, JAEGER, AND BRUNETTE
ance on micromachined surfaces at shorter times, using confocal microscopy and inhibitors of cytoskeletal polymerization in BHK fibroblasts. Although aligned MFBs were not reported at early times, the aggregation of actin along groove/edge boundaries was observed as early as 5 min and they concluded that actin condensations were the primary driving event of topographic guidance. Oakley and Brunette [6] also observed actin condensations along the ridge/groove interface up to 2 h after plating, at which time only 40% of cells demonstrated aligned MFBs, yet 80% of cells demonstrated aligned MTs. Another approach to investigating the role of the cytoskeleton in cell alignment is to vary the size of the topographic features in order to determine whether there are conditions under which MTs or MFBs are affected differentially and whether cell alignment can be obtained under such conditions. As the response of either MTs or MFBs could be affected by the presence of the other, we observed cell spreading and alignment in the absence (controls) or presence of specific inhibitors of MTs (colcemid) and MFBs (cytochalasin B) on smooth control surfaces and on three different sets of grooves that were identical in shape but differed in size. Thus, behaviors of cells with an intact cytoskeleton (controls) could be compared to MT-deficient/MFBcontaining colcemid-treated cells and MT-containing/ MFB-deficient cytochalasin B-treated cells. Using confocal microscopy and computer-assisted morphometry to examine HGF response at different time intervals, we determined that groove spacing significantly influenced the kinetics of cell alignment, irrespective of whether cytoskeletons were intact or perturbed. On the narrowest 0.5-mm-wide grooves, alignment of cytochalasin B-treated (MT-containing/MFB-deficient) cells was superior to alignment of cells with an intact cytoskeleton (controls), and colcemid-treated (MT-deficient/MFB-containing) cells failed to align altogether, behaving as if on a smooth surface. These findings suggest that in HGFs, the MT system may be the principal but not sole cytoskeletal component affecting topographic guidance and that substrata topographies can be produced that differentially affect the MT and MFB systems. MATERIALS AND METHODS Materials and methods are similar to those described previously [6, 7] and are reviewed briefly below. Cell culture, treatment conditions, and time-lapse observations. Fibroblasts between the 4th and 12th subcultures, isolated from human gingival explants [9] were cultured in a-minimal essential medium (MEM) (Stemcell, Vancouver, BC) supplemented with antibiotics [100 mg/ml penicillin G (Sigma, St. Louis, MO), 50 mg/ml gentamicin (Sigma), 3 mg/ml amphotericin B (Fungizone, Gibco, Grand Island, NY)] and 15% bovine serum (Calf Supreme, Gibco) at 377C in a humidified atmosphere with 5% CO2 .
AID
ECR 3625
/
6i24$$$162
07-15-97 08:12:51
HGF cells were removed from their growth surface by a trypsin solution [0.25% trypsin (Gibco) and 0.1% glucose dissolved in citrate– saline (pH 7.8)]. Cells were resuspended in either control medium containing less than 0.1% dimethyl sulfoxide (DMSO, Fisher Scientific, Ottawa, Canada) or medium containing cytoskeleton-perturbing agents (colcemid or cytochalasin B [CB]) solubilized in DMSO (õ0.1%) (Technical Information Bulletin AL-126, Aldrich Chemical Co., Inc., Milwaukee, WI). A stock solution of 100 mg/ml colcemid (Sigma), solubilized in DMSO, was prepared in medium so that the final concentration of DMSO was less than 0.1%. An appropriate volume of the colcemid stock solution was added to the medium used for resuspension of the cells so that a final concentration of 1.0 mg/ml colcemid was obtained. This concentration of colcemid was based on the work of Hollenbeck et al. [10] and is the same concentration used previously [7]. A stock solution of 1.0 mg/ml CB (Sigma), solubilized in DMSO, was prepared in medium and an appropriate volume of CB stock solution was added to the medium used for resuspension of the trypsinized cells so that a final concentration of 2.0 mg/ml CB was obtained. This concentration of CB was used in the present study because titration experiments indicated that this was the lowest CB concentration at which the vast majority of HGFs plated and maintained in the presence of CB displayed arborized MT-containing processes and no longer displayed MFBs [e.g., 11]. These titration experiments are discussed in detail in the Ph.D. thesis of Carol Oakley (1995, University of British Columbia, Canada). CB was used for time-lapse observations and for cytoskeletal observations in order to compare results to earlier reports of fibroblasts in the presence of CB [11–13]. However, cytochalasin D was used as an additional control (data not shown) to verify that the cytoskeletal behaviors elicited in the presence of CB were due to their action upon the actin system and were not related to the interference with sugar transport associated with CB [14]. The resuspended cells were seeded onto smooth and grooved titanium substrata at different cell populations in order to obtain cells which were spatially well isolated at the different time periods. For control conditions, population densities of 1 1 105 cells/ml were used for experiments lasting up to 2 h, 1 1 104 cells/ml for 6 h, and 1 1 103 cells/ml for 24 h. For drug conditions, cell densities of 1 1 105 cells/ml were used for 2-h experiments and 1 1 104 cells/ml were used for 24-h experiments. Cell suspensions of 1 1 105 cells/ml were used for all time-lapse observations, regardless of culture conditions or length of observation. For these observations, cell suspensions were injected into a Pentz chamber (Bachofer, Reutlingen, Germany) in which the substrata had been mounted. Cells were observed through a microscope (Reichert, Austria) equipped with reflected Nomarski differential interference contrast (DIC) optics and an Epiplan 8X or 16X objective (Zeiss, Oberkochen, Germany). Cell behavior was recorded using a television camera and control system (Hamamatsu model C2400; Japan) and a time-lapse video recorder (Panasonic Model 8050). Micromachined substrata. Titanium-coated micromachined silicon wafers were produced as described previously by Brunette [15]. A computer-generated pattern was produced on a photomask and transferred onto silicon wafers by photolithography and anisotropic etching, producing a series of grooves. In this study three different groove patterns were used. The grooves all had the same shape but differed in pitch and depth. In cross section the grooves had a truncated V shape and the interior corner of the ridge and wall of the groove formed an angle of 1257. The first groove pattern consisted of a series of 3-mm-deep grooves with a 30-mm pitch comprising a 15mm-wide groove and 15-mm-wide ridge. This pattern had been used in previous studies [6, 7, 16] and was termed the reference (R) groove pattern. The second pattern consisted of a series of 3-mm-deep grooves with a pitch that ranged from 6 to 9 mm in different areas of the silicon wafer. This pattern was designated the narrow (N) groove pattern. The third pattern consisted of 0.5-mm-deep grooves
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION with a 1.0-mm pitch and was termed the very narrow (VN) groove pattern. Both the grooved and the smooth silicon wafers were sputter coated with 50 nm of titanium, a procedure that renders the surface chemically homogeneous, as noted by Singvhi et al. [17]. The surfaces were cleaned by ultrasonication in a detergent formulated for tissue culture (71, ICN Biomedicals, Inc., Costa Mesa, CA) and then rinsed copiously and ultrasonicated in filtered deionized water and air-dried in a laminar-flow hood. Final cleaning and sterilization were accomplished by argon-gas glow-discharge treatment (18) just prior to seeding with the cell suspension. Immunofluorescence. At 2, 6, and 24 h and after all time-lapse observations, samples were rinsed with cytoskeletal stabilizing (CS) buffer (19), fixed for 10 min in 3.7% formaldehyde (BDH Inc., Toronto, Canada) in phosphate-buffered saline (PBS), and washed in CS buffer. To visualize cell morphology, some samples were prepared for incubation with fluorescent hydrazides to stain the cell membrane. Following oxidation in 4.2 mM sodium periodate (BDH Inc.), these samples were exposed to fluorescein-5-thiosemicarbazide (Molecular Probes Inc., Eugene, OR). Subsequently, preparation for staining of cytoskeletal elements was continued in all samples by permeabilization in 0.5% Triton X-100 (Sigma), quenching with 0.05% sodium borohydride (BDH Inc.), blocking with 1% bovine serum albumin (BSA, Sigma) and exposure to one or more of the following antibodies or fluorescent stains: Bodipy fluorescein phallacidin or rhodamine phalloidin (Molecular Probes Inc.) or monoclonal anti-b-tubulin (Boehringer–Mannheim GmbH, Mannheim, Germany) followed by sheep anti-mouse IgG conjugated to Texas red (Molecular Probes Inc.) or fluorescein isothiocyanate (FITC; Boehringer–Mannheim GmbH). All antibodies were diluted with PBS containing 0.1% BSA. Epifluorescence and confocal microscopy. Cells stained with the hydrazide were examined through a microscope equipped with epifluorescent illumination (confocal laser scanning microscope (CLSM), Zeiss) and photographed (Olympus PM4T camera, Kodak Tmax 400). Fields were selected randomly and the only criterion for cell inclusion in the sample was the absence of contact of a cell with other cells. Subsequently, images of the photographed cells were projected and cell outlines as well as corresponding groove–ridge patterns, where applicable, were traced onto paper. The tracings were retraced on the digitizing tablet of a morphometrics system (Videoplan, Zeiss) which calculated a form factor, Form Ell, that is the ratio of the minor to major axes. The longest diameter between two edges within the cell was considered the major axis and the minor axis was considered the longest axis perpendicular to the major axis. A form Ell value of 1.0 corresponds to a circular shape, and values less than 1 indicate elliptical or spindle shapes. The angle formed by the major axis to either the grooves or, for smooth surfaces, a line with randomly selected orientation was termed the orientation angle. For measurement of cell orientation no attempt was made to distinguish cell polarity and therefore the maximum angle possible was 907, with a value of 07 indicating a cell that was perfectly aligned with the direction of the grooves. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves [20]. Cell alignment was not contingent upon cell polarity and cell polarity did not imply alignment. Cell polarity was also used to describe cell shape, as in the approach of Glasgow and Daniele [21]. Cells were termed monopolar when there was a clearly identifiable leading lamella. Bipolar cells contained two distinct lamellae and apolar cells were round. Cytoskeletal elements were examined with both epifluorescence and confocal microscopy using a 63X oil-immersion plan-apochromat objective, NA 1.4 (Zeiss) on the CLSM (Zeiss). An argon laser (lmax Å 488 nm) was used for Bodipy fluorescein and FITC-labeled cells in confocal microscopy; a helium–neon laser (lmax Å 543 nm) was used for Texas Red-labeled cells. For double-labeled specimens, the Bodipy or FITC signal was examined first, before the Texas Red signal. Digital gray-scale images obtained by the CLSM were printed
AID
ECR 3625
/
6i24$$$162
07-15-97 08:12:51
415
using a video printer (UP-5000, Sony Canada, Richmond, British Columbia). Descriptions of cytoskeletal patterns and criteria used to determine cytoskeletal alignment followed those outlined in [6]. Statistics. Morphometric measurements were based on a mean number of 50 cells at 2 and 6 h and 83 cells at 24 h for each of the smooth and grooved surfaces, under control, colcemid, and CB conditions. Data were entered into a mainframe computer (IBM) and analyzed using SPSS-X (SPSS Inc., Chicago, IL). Analysis of variance (ANOVA) and Student–Newman–Keuls (SNK) were used to analyze the effects of surface, time, and treatment conditions on Form Ell and orientation angle. The proportions of aligned cells were expressed as a percentage and a nonparametric test, Kruskal–Wallis one-way analysis of variance, was used to analyze the percentages of aligned cells, as these data were not normally distributed.
RESULTS
This study, admittedly complex, includes observations of HGFs under 12 conditions over time; that is, observations were made on each of four substrata: smooth, R, N and VN, and under each of three culture conditions: control, colcemid, and cytochalasin B. To facilitate presentation of the data, observations are categorized by treatment condition (i.e., control, CB, colcemid) in terms of the MT or MFB content. Thus control or CB-treated cells would contain MTs and be designated MT/, whereas colcemid-treated cells would be designated MT0. The three groups were thus MT/MFB/ (controls), MT/MFB0 (CB-treated cells), and MT0MFB/ (colcemid-treated cells). Within each treatment group, the behaviors of cells on smooth, R-, N-, and VN-grooved surfaces are compared. Figure 1 presents an overview of the results, the appearance of cells cultured for 24 h on smooth, R, N, or VN grooves under control, CB, or colcemid conditions. In summary, all grooved surfaces oriented cells under all conditions, with the exception that VN grooves failed to orient cells when MTs were absent (i.e., colcemid-treated cells). These qualitative results are supported by quantitative data in Fig. 2, the percentage of aligned cells on each substratum as a function of time for each treatment group. In brief, under control conditions, alignment occurred in the order N ú R ú VN, whereas in CB-treated (MFB-deficient) cells, N and VN grooves oriented cells more quickly than R grooves, and surprisingly, produced greater orientation than that observed under control conditions. Colcemid-treated MT-deficient cells aligned best on N grooves followed by R grooves, but MT-deficient cells did not align on VN grooves. Moreover, in order to facilitate comparisons of cell alignment and cell shape among treatment groups, orientation angle (Fig. 3a) and Form Ell (Fig. 3b) are presented for all treatment groups together at 24 h, the longest time of observation. In addition, a qualitative comparison of directed motility by cells within each treatment group on each substratum is presented in Table 1. Finally, the typical appearance
eca
416
OAKLEY, JAEGER, AND BRUNETTE
FIG. 1. Light micrographs of HGFs spread 24 h on smooth (a, e, i), R- (b, f, j), N- (c, g, k), or VN-grooved (d, h, l) substrata under control (a–d), CB (e–h), or colcemid (i–l) conditions. Cells were stained with fluorescent hydrazide to stain the cell membrane. Thin lamellar and tail regions stained less brightly than the body of the cells. On smooth surfaces, cells showed no particular orientation, whereas cells on R and N grooves aligned to the direction of the grooves (double-headed arrows). On VN grooves, only controls and CB-treated cells aligned with the grooves.
of the actin and MT systems within each treatment group on smooth, R, N, and VN grooves is presented in Figs. 4 – 6. As the focus of this study is the topographic responses of entire cells, Figs. 4 – 6 are intended to provide a composite overview of cytoskeletal behavior under the various culture and substrata conditions rather than a detailed analysis of cytoskel-
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
etal interaction with the grooves. Detailed cytoskeletal patterns and behaviors of HGFs on smooth and R grooves under control [6] and colcemid [7] conditions have been published previously. A more detailed description of the behavior of control and treated cells as a function of time on the four surfaces is given below.
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
417
On grooved surfaces, HGFs generally spread radially across the ridges and grooves and gradually elongated in a direction parallel to the grooves but some differences between the types of grooves were seen in the kinetics of alignment and in the relationship of cells to the grooves. On R grooves, cells generally became aligned within one groove, whereas on N and VN grooves cells bridged across several grooves (Figs. 1b–
FIG. 2. Percentage of aligned HGFs plotted against time for fibroblasts plated 2, 6, and 24 h on smooth and R, N, and VN grooves under control (a), CB (b), or colcemid (c) conditions. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves or, for smooth surfaces, to a line of randomly selected orientation. The proportion of aligned cells at each time interval was expressed as a percentage. (a) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VNgrooved substrata under control conditions, in the absence of CB, or colcemid. Alignment was fastest and greatest on N grooves, followed by R grooves and VN grooves. (b) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VN-grooved substrata under CB conditions. Alignment was fastest and greatest on VN and N grooves, followed by R grooves. (c) The percentage of aligned cells 2, 6, and 24 h after plating on smooth, R-, N-, or VN-grooved substrata under colcemid conditions. Alignment was fastest on N grooves, followed by R grooves but alignment on N and R grooves was similar at 24 h. The small proportion of cells classified as aligned on VN grooves is similar to the proportion of cells classified as aligned on smooth surfaces.
Control Cells (MT/MFB/ Cells) Under control conditions, in the absence of colcemid or CB, the actin MFB and MT systems were intact (Fig. 4). On the smooth surface, HGFs became monopolar and migrated across the surface. Cell orientation was essentially random, not differing significantly from the value of 457C expected if no orienting influences were present (Fig. 3a).
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
FIG. 3. (a) Orientation angle of HGFs 24 h after plating on smooth and R-, N-, or VN-grooved substrata plotted as a function of treatment (control, CB, or colcemid). Error bars { SEM. Cells were considered aligned when their orientation angle measured less than 107 to the direction of the grooves or, for smooth surfaces, to a line of randomly selected orientation. In the absence of any orientating influence, the orientation angle of cells on smooth surfaces did not differ greatly from the expected value of 457. On R and N grooves, cells under all treatment conditions were significantly (P õ 0.05) more aligned than on smooth surfaces. On VN grooves, only control and CB-treated cells were aligned but controls on VN grooves were significantly (P õ 0.05) less aligned than controls on R or N grooves or CB-treated cells on R, N, or VN grooves. On VN grooves, colcemidtreated cells were poorly aligned. (b) Form Ell of HGFs 24 h after plating on smooth or R-, N-, or VN-grooved substrata plotted as a function of treatment (control, CB, or colcemid). Error bars { SEM. Form Ell is a ratio of the minor and major axes of a cell and is an indicator of shape. A value of 1.0 corresponds to a circular shape and values less than 1.0 indicate elliptical or spindle shapes. On smooth surfaces, controls were significantly (P õ 0.05) more elliptical than colcemid- and CB-treated cells. On R and N grooves, cells were significantly (P õ 0.05) more elliptical than cells on smooth surfaces under the same treatment condition. On VN grooves, controls were the most elliptical (P õ 0.05), followed by CB-treated cells; colcemidtreated cells were round.
eca
418
OAKLEY, JAEGER, AND BRUNETTE
TABLE 1 Qualitative Ranking of Persistent and Directed Cell Migration
Smooth 30 mm (R) 6–9 mm (N) 1 mm (VN)
Control
CB
Colcemid
/ ///// ///// //
NM NM NM NM
NDM //// /// NDM
Note. On smooth surfaces and VN (1.0-mm pitch) grooves, only control cells were motile and the persistence of their directed motility along the grooves was less than that observed on R (30-mm pitch) or N (6- to 9-mm pitch) grooves. No motility (NM) was observed for CBtreated cells on any surface and no directional motility (NDM) was observed for colcemid-treated cells on smooth or VN surfaces. In contrast, migration along the grooves was observed on R and N grooves for controls and colcemid-treated cells although both continuity and persistence of forward locomotion differed with treatment and surface.
1d). Although cell alignment was fastest on N grooves (Fig. 2a), at 24 h alignment on N and R grooves was comparable (Figs. 2a and 3a) and significantly (P õ 0.05) greater than alignment on VN grooves (Fig. 3a). At 24 h, the shape of cells on grooved surfaces was typically elliptical (Fig. 3b) and monopolar but, again, some differences in cell morphology and alignment were seen between the types of grooves. On R and N grooves, leading edges, cell bodies, and tails of cells were typically aligned with the grooves, producing a very elongated cell shape (Figs. 1b and 1c). On VN grooves, cell shape and alignment were more variable than on N or R grooves; orientation of leading edges was frequently along the direction of the VN grooves, whereas alignment of cell bodies and tails was less uniform (Fig. 1d). On grooves, cells exhibited topographic guidance but otherwise traveled in a manner similar to that of cells moving on smooth surfaces. One difference in cell behavior dependent on topography was that cells on VN grooves showed less directional persistence than cells on R or N grooves (Table 1). Cell migration on R and N grooves also resulted in a larger displacement over a given time, relative to cells on VN grooves (Table 1) or on the smooth surface [22]. CB-Treated Cells (MT/MFB0Cells) CB (2.0 mg/ml) was used to inhibit the formation of actin MFBs (Fig. 5). At this concentration of CB, HGFs plated and maintained in the presence of CB attached and spread and the vast majority of HGFs extended cell processes. Actin MFBs were not observed but small foci of actin-positive material conformed to the outlines of cell processes that contained bundles of MTs (Fig. 5). On smooth surfaces (Figs. 1e, 6a and 6b), the cell
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
processes which contained bundles of MTs (Fig. 5) demonstrated no preferred orientation (Figs. 2b and 3a) but on grooves, the processes aligned with the direction of the grooves (Figs. 1f–1h, 5). CB-treated cells were not motile. Alignment of CB-treated cells on N and VN grooves was faster (Fig. 2b) and superior (Figs. 2b and 3a) to alignment on R grooves. Surprisingly, alignment of CBtreated cells on N and VN grooves exceeded alignment of controls at 2 h (Figs. 2a and 2b). Furthermore, at 24 h the overall alignment of CB-treated cells on R, N, and VN grooves was significantly (P õ 0.05) better than alignment of controls on VN grooves (Fig. 3a). Groove depth appeared to be a less important factor in determining alignment of CB-treated cells than groove/ ridge spacing. Shallow 0.5-mm-deep VN grooves produced alignment that occurred earlier but was comparable to that on the 3.0-mm-deep N grooves and both VN and N grooves produced better alignment than 3.0mm-deep, wider R grooves (Figs. 2b and 3a). Thus elimination of the MFBs by CB apparently increased the orientation response of cells to the N and VN grooves which produced faster and greater alignment than R grooves. Colcemid-Treated Cells (MT0MFB/ Cells) Colcemid was used to eliminate the MTs (Fig. 6). As previously reported, R grooves were able to effect topographic compensation, i.e., alignment, polarization, and directed migration of MT-deficient cells (7). Topographic compensation was also effected by N grooves; however, a smaller proportion of colcemidtreated cells was monopolar and directionally motile on N grooves as compared to R grooves. In addition, the directed migration of some colcemid-treated cells along N grooves appeared to be more intermittent than locomotion on R grooves (Table 1). By 24 h, alignment of colcemid-treated cells on R and N grooves (Figs. 2c and 3a) was comparable, although cells were significantly (P õ 0.05) less elliptical on N grooves than on R grooves (Fig. 3b). As expected from our previous study, colcemid-treated cells required significantly (P õ 0.05) more time to align on R (7) and N grooves than did controls (Figs. 2a and 2c). There were also differences between the grooved surfaces as colcemid-treated cells aligned more quickly on N grooves than on R grooves (Figs. 2a and 2c). Thus, at early times of plating, the N-grooved substratum was the most efficacious in promoting cell alignment in the presence or absence of MTs. On VN grooves (Fig. 1l), colcemid-treated cells behaved like colcemid-treated cells on smooth surfaces. That is, alignment (Figs. 2c and 3a), elliptical cell shapes (Fig. 3b), and directed cell migration (Table 1) did not develop. Although a few cells demonstrated
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
419
FIG. 4. CLSM-generated images of HGFs 24 h after plating on smooth (a, b), R- (c, d), N- (e, f), or VN- (g, h) grooved substrata under control conditions, in the absence of colcemid or CB. Groove directions indicated by double-headed arrows. Cells stained both for actin (a, c, e, g) and tubulin (b, d, f, h). All cells were monopolar and both actin–MFB and tubulin distributions reflected the overall shape of cells on smooth (a, b) and overall shape and alignment of cells on grooves (c–h). On R grooves (c, d) cells were typically aligned within one groove, whereas on N (e, f) and VN (g, h) grooves cells bridged laterally over several grooves and ridges. On VN grooves, MTs (h) were typically better aligned with the grooves than actin–MFBs. Some MFBs (arrowheads in g) appeared to be oriented across the grooves.
AID
ECR 3625
/
6i24$$3625
07-15-97 08:12:51
eca
FIG. 5. CLSM-generated images of HGF cells 24 h after plating on smooth (a, b), R- (c, d), N- (e, f), or VN- (g, h) grooved substrata under CB conditions. Groove directions indicated by double-headed arrows. Cells stained for both actin (a, c, e, g) and tubulin (b, d, f, h). Actin–MFBs were not evident but small foci of actin-positive material conformed to the outlines of cell processes that contained bundles of MTs. MT bundles radiated from a central source and on smooth surfaces MT bundles did not appear to have a preferred orientation (b). In contrast, on grooves the bundles aligned with the direction of the grooves.
07-15-97 08:12:51
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
421
FIG. 6. CLSM-generated images of HGF cells 24 h after plating on smooth (a, b), R (c, d), N (e, f), or VN (g, h) grooves under colcemid conditions. Groove directions indicated by double-headed arrows. Cells stained for both actin (a, c, e, g) and tubulin (b, d, f, h). A fibrillar network of tubulin staining was not observed. The distribution of actin MFBs reflected the shape of the cells and was similar to MFB distributions observed in controls with similar shapes on both smooth and grooved surfaces. On R (c) and N (e) grooves, actin MFBs reflected the alignment of the cells and both cells and MFBs were aligned to the direction of the grooves. In contrast, alignment of cells and their actin MFBs on VN grooves (g) was similar to that observed on smooth (a) surfaces.
AID
ECR 3625
/
6i24$$3625
07-15-97 08:12:51
eca
422
OAKLEY, JAEGER, AND BRUNETTE
aligned shapes, cell shapes were typically bipolar or apolar and changed frequently. Thus, using the criterion of alignment of MFBs to the grooves, the actin MFB system was not consistently influenced by VN grooves. DISCUSSION
The results of this study enabled us to distinguish between the roles played by the MT and actin–MFB systems in some aspects of topographic guidance and topographic compensation [7]. Whereas the actin system is essential for powering cell locomotion, it has been accepted that on smooth surfaces MTs are required for stabilizing cell shape and effecting polarization [e.g., 23]. In contrast, on some anisotropic substrata (i.e., R or N grooves) the presence of either an intact MT or an intact actin–MFB system appears to be sufficient to elicit alignment of cells to the substratum topography. However, the MT and actin–MFB systems differed in their responses to the substratum, both in the rates of effecting cell alignment and in their sensitivity to dimensions of the substratum topography. In general, MT-containing cells aligned faster with grooves of smaller dimensions (such as the VN grooves) than did MT-deficient but MFB-containing (i.e., colcemid-treated) cells. Thus, the MT system appeared to be the principal but not the sole cytoskeletal substratum-response mechanism effecting cell alignment on anisotropic substrata. However, given enough time, colcemid-treated cells lacking MTs aligned, polarized, and migrated on R and N grooves, a process termed topographic compensation [7]. Significantly, not all surfaces that can affect contact guidance in cells with an intact cytoskeleton can effect topographic compensation [7, 24]. That is, in contrast to their behavior on R and N surfaces, colcemid-treated (MT0MFB/) HGFs did not exhibit topographic compensation on VN grooves. In fact, the behavior of these MT-deficient cells on VN-grooved surfaces was similar to that on smooth surfaces. Thus, topographic compensation appeared to be associated with aligned MFBs that were observed in MT-deficient cells on the R- and N-grooved surfaces, but were only rarely observed in cells on the VN surfaces. By comparison, CB-treated cells which were MFB deficient could align on R, N, and VN grooves, although they could not move on the smooth or any of the grooved surfaces. Therefore, both the microtubular and the actin systems are required to obtain a full topographic guidance response from HGFs over the range of dimensions used in this study. The cellular response to substratum topography appears to involve some interaction between the MT and actin–MFB systems. This interaction is particularly evident on VN-grooved surfaces where the removal of MFBs by CB causes an increase in alignment of cells
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
and their processes with the grooves. Under control and colcemid (MT0MFB/) conditions, these VN grooved substrata did not significantly affect the MFB–actin system as assessed by the paucity of MFBs aligned with the substratum grooves. It appears that under control conditions, when the actin–MFB system is intact and presumably exerting its own effects on cell shape and translocation, the actin–MFB system diminishes the effects produced by the substratum topography on the MT system. Such an interaction of the MT and MFB systems might be expected because the intracellular course or placement of MTs is influenced by structural interactions with actin [25] as well as by tension generated by the actin network [25, 26]. Recently, Wojciak-Stothard et al. [8] used BHK fibroblasts to investigate the role of cytoskeletal systems in topographic reactions. They concluded that the aggregation of actin along groove/ridge boundaries was the primary driving event in determining BHK fibroblast orientation on microgrooved substrata. There are three germane differences between our study and that of Wojciak-Stothard et al. [8] that may explain the difference in conclusions. The first difference between the studies was that in our study, HGFs resuspended and maintained in the presence of cytochalasin developed arborized cell processes within 2 h after exposure to CB. Numerous studies have shown that exposure of cells to cytochalasin results in a decrease or cessation of ruffling behavior and cell motility; the cell body rounds up and numerous processes or branches are extended, giving the cells a stellate or arborized shape [e.g., 12–14, 27, 28]. In fibroblasts, cytochalasins do not depolymerize actin filaments but, rather, disrupt the supramolecular organization of actin filaments, and the filaments persist, albeit in dense focal accumulations [27]. As the cortical actin–MFB network is destroyed by cytochalasins, the restraining influence of the cortex is reduced and MTs are able to push out the pliable plasma membrane and traverse straight and uninterrupted to the distal ends of the arborized processes [11, 29]. In contrast, the CD-treated BHK fibroblasts in the Wojciak-Stothard et al. [8] study do not appear to be arborized at 24 h (see, for example, Fig. 1d of their paper which differs from Figs. 1f–1h of this study). However, BHK cells did adopt a more arborized appearance when exposed to cytochalasin for periods longer than 24 h (Curtis, personal communication, 1995). Thus, there are significant differences in the dynamics of the cytoskeleton in the cell types used in the two studies. There are several levels of supramolecular actin organization. Phallotoxins stain only filamentous actin and in both of these light microscopic studies only bundles of microfilaments were visualized; therefore, we cannot comment on the possible role of individual actin filaments in cell alignment. Nevertheless, as arboriza-
eca
SENSITIVITY TO TOPOGRAPHIC GUIDANCE AND COMPENSATION
423
tion occurred in our CB-treated cells it appears that VN grooves. If a substratum such as R or N grooves the subcortical actin matrix was disrupted. Moreover, elicited alignment of both MT and actin–MFB systems, prominent MFBs were not present. In such circum- overall alignment was facilitated and enhanced. Constances the ability of the MT system to effect cell align- versely, if a substratum such as VN grooves failed to ment on R, N, and VN grooves independently of any direct alignment of the actin–MFB system, the influinfluence of functionally organized MFBs could be dem- ence of the MT system upon cell alignment was reduced onstrated. but not eliminated. A second difference between the work of WojciakA major finding of this study is that substrata can Stothard et al. [8] and this study lies in the chemical be designed to affect the MT and actin–MFB systems composition of the substrata. In this study, micromach- differentially. Such substrata may be useful in disined substrata were sputter coated with titanium secting the role of the MT and actin–MFB systems in [15, 30, 31], a process which, as noted by Singhvi et al. cell behaviors such as topographic guidance and cell [17], provides a completely homogeneous surface and migration without the use of drugs such as CB and avoids the possible problem of chemical heterogeneity colcemid with their attendant possibilities of nonspethat can be produced by etching methods [17]. Tita- cific effects. Understanding the principles of cytoskelenium was selected because of its biocompatability [32]. tal interaction with the substratum topography may In contrast, Wojciak-Stothard et al. [8] used a two-stage permit exploitation of certain behaviors to uncouple or dry-etching procedure on fused silica so that all of the amplify MT and actin-system responses and permit the surface was exposed to etching and thus could be con- design of substrata that specifically retard or enhance sidered as having a uniform surface chemistry. Never- cell alignment and migration. One practical application theless, the surface chemistries of the substrata doubt- of such substratum-mediated control of cell behaviors less differ between the studies. may be in the design of implanted devices where control A third difference between our study and that of Woj- of cell migration would be expected to improve device ciak-Stothard et al. [8] is the range of groove dimen- performance [31]. Ultimately, the substrata of biomatesions examined. Groove dimensions must be considered rials used for prosthetic devices and implants could be in relation to the typical size of the cells being investi- engineered to create specific cytoarchitectures which, gated [17, 33]. Our R and N grooves are comparable to in turn, could produce specific cell responses. the 10- and 5-mm-wide grooves, respectively, used by Wojciak-Stothard et al. [8]. Even though the cell types The authors thank Mr. Hiroshi Kato for technical assistance with differed, very similar orientation angles were obtained, substrata preparation. The authors are grateful to Dr. W. Vogl (Deon both our N grooves and their 5-mm-wide grooves partment of Anatomy, University of British Columbia), for advice under control, colcemid, or CB conditions. However, we and comments on this work. The authors are also indebted to Ms. Jackie McDiarmid for her diligence in tracing and digitizing the cells also employed substrata with much smaller grooves and thank Mrs. Lesley Weston and Mr. Andre Wong for technical (0.5-mm-wide VN grooves) and, significantly, it was on assistance. This work was supported by the Medical Research Counthese ultrafine grooves that differences in the threshold cil of Canada (MRC Dental Fellowship Grant 5-59108 to C. Oakley, sensitivities of the MT and actin–MFB systems were Operating Grant 5-97617 to D. M. Brunette). readily apparent. In summary, the MT system was not required for REFERENCES alignment, polarization, or directed migration if col1. Weiss, P. (1934) J. Exp. Zool. 68, 393–448. cemid-treated (MT-deficient) HGFs were provided with enough time and an appropriate substratum such as R 2. Curtis, A. S. G., and Clark, P. (1990) Crit. Rev. Biocompat. 5, 343–362. (7) or N grooves. However, on the VN-grooved sub3. Dunn, G. A., and Heath, J. P. (1976) Exp. Cell Res. 101, 1–14. strata, the presence of the MT system under CB condi4. Dunn, G. A., and Brown, A. F. (1986) J. Cell Sci. 83, 313–340. tions produced cell alignment even though MFBs were 5. Ohara, P. T., and Buck, R. C. (1979) Exp. Cell Res. 121, 235– absent. In contrast, under colcemid conditions on VN 249. grooves the presence of the actin-MFB system did not 6. Oakley, C., and Brunette, D. M. (1993) J. Cell Sci. 106, 343– produce cell alignment. Significantly, the MT system 354. responded faster and to smaller topographic features 7. Oakley, C., and Brunette, D. M. (1995) Cell Motil. Cytoskel. 31, than the actin–MFB system, and therefore we consider 45–58. that the MT system appeared to be the principal, but 8. Wojciak-Stothard, B., Curtis, A. S. G., Monaghan, W., McGrath, not the sole cytoskeletal substratum-response mechaM., Sommer, I., and Wilkinson, C. D. W. (1995) Cell Motili. Cynism in HGFs. Moreover, there is an interaction betoskel. 31, 147–158. tween actin–MFB and MT systems that affects their 9. Brunette, D. M., Melcher, A. H., and Moe, H. K. (1976) Arch. respective kinetics and sensitivities to the substratum. Oral Biol. 21, 393–400. This interaction is apparent in untreated control cells 10. Hollenbeck, P. J., Bershadsky, A. D., Pletjushkina, O. Y., Tint, I. S., and Vasiliev, J. M. (1989) J. Cell Sci. 92, 621–631. which aligned well on R and N grooves, but poorly on
AID
ECR 3625
/
6i24$$$163
07-15-97 08:12:51
eca
424
OAKLEY, JAEGER, AND BRUNETTE
11. Weber, K., Rathke, P. C., Osborn, M., and Franke, W. W. (1976) Exp. Cell Res. 102, 285–297. 12. Bliokh, A. L., Domnina, L. V., Ivanova, O. Y., Pletjushkina, O. Y., Svitkina, T. M., Smolyninov, V. A., Valisiev, J. M., and Gelfand, I. M. (1980) Proc. Natl. Acad. Sci. USA 77, 5919–5922. 13. Domnina, L. V., Gelfand, V. J., Ivanova, O. Y., Leonova, E. V., Pletjushkina, O. Y., Vasiliev, J. M., and Gelfand, I. M. (1982) Proc. Natl. Acad. Sci. USA 79, 7754–7757. 14. Yahara, I., Harada, F., Skeita, S., Yoshihara, K., and Natori, S. (1982) J. Cell Biol. 92, 69–72. 15. Brunette, D. M. (1986) Exp. Cell Res. 164, 11–26. 16. Oakley, C., and Brunette, D. M. (1995) Biochem. Cell Biol. 73, 473–489. 17. Singhvi, R., Stephanopoulos, G., and Wang, D. I. C. (1994) Biotechnol. Bioeng. 43, 764–771. 18. Baier, B. E., and Meyer, A. E. (1988) Int. J. Oral Maxillofac. Implants 3, 9–20. 19. Opas, M. (1989) Dev. Biol. 131, 281–293. 20. Clark, P., Connolly, P., Curtis, A. S. G., Dow, J. A. T., and Wilkinson, C. D. W. (1990) Development 108, 635–644. 21. Glasgow, J. E., and Daniele, R. P. (1994) Cell Motil. Cytoskel. 27, 88–96. 22. Damji, A., Weston, L., and Brunette, D. M. (1996) Exp. Cell Res. 228, 114–124.
23. Vasiliev, J. M. (1982) Phil. Trans. R. Soc. London 299, 159– 167. 24. Vasiliev, J. M., Gelfand, I. M., Domnina, L. V., Ivanova, O. Y., Komm, S. G., and Olshevskaja, L. V. (1970) J. Embryol. Exp. Morphol. 4, 625–640. 25. Schliwa, M., and van Blerkom, J. (1981) J. Cell Biol. 90, 222– 235. 26. Holifield, B. F., and Heath, J. P. (1992) Mol. Biol. Cell 3, 279a. 27. Schliwa, M. (1982) J. Cell Biol. 92, 79–91. 28. Sanger, J. W., and Holtzer, H. (1972) Proc. Natl. Acad. Sci. USA 69, 253–257. 29. Edson, K., Weishaar, B., and Matus, A. (1993) Development 117, 689–700. 30. Gould, T. R. L., Westbury, L., and Brunette, D. M. (1984) J. Prosthet. Dent. 52, 418–420. 31. Brunette, D. M., Kenner, G. S., and Gould, T. R. L. (1983) J. Dent. Res. 62, 1045–1048. 32. Bra˚nemark, P. I., Hansson, B. O., Adell, R., Breine, U., Lind¨ hman, A. (1977) in Osseointegrated Implants stro¨m, H., and O in the Treatment of the Edentulous Jaw: Experience for a 10Year Period, Almqvist Wiksell International, Stockholm. 33. Green, A. M., Jansen, J. A., van der Waerden, J. P. C., and von Recum, F. F. (1994) J. Biomed. Mater. Res. 28, 647–653.
Received January 30, 1997 Revised version received May 6, 1997
AID
ECR 3625
/
6i24$$$164
07-15-97 08:12:51
eca