- Email: [email protected]
Dermal fibroblasts from Down's syndrome patients share a cycloheximide-induced deficiency in collagen adhesion responses with normal aging cells
EXPERIMENTALCELL
RESEARCH
189,189-201(1990)
Dermal Fibroblasts from Down’s Syndrome Patients Share a Cycloheximide-Induced Deficiency in Collagen Adhesion Responses with Normal Aging Cells KELLY S. FLICKINGER’ lkportment
of
AND LLOYD A. CULP
Molecular Biology and Microbiology, Case Western Reserve University, School of Medicine, Clmelund, Ohio 44106
INTRODUCTION Human skin fibroblasts from three different Down’s syndrome patients (trisomy 21) of very different ages have been tested for their adhesion responses on tissue culture substrata coated with type I collagen, fibronectin (FN), and their combination after or during treatment of cells with cycloheximide to evaluate limitations in specific responses. It was shown previously that in vitro-aged papillary and reticular dermal fibroblasts from normal individuals do not generate F-a&in stress fibers when pretreated with cycloheximide on collagen substrata but do so on FN substrata, a deficiency linked to limiting amounts/function of collagen-specific receptors in aging cells. In these studies, all three Down’s fibroblast populations demonstrated a similar deficiency in stress fiber formation, evaluated by rhodaminephalloidin staining, upon cycloheximide treatment at all passage levels. They remained competent for stress fiber formation on FN substrata and for reorganization of microtubule and intermediate filament networks on all substrata, demonstrating the specificity for the collagen matrix and for the F-actin cytoskeleton in this deficiency. The cycloheximide-induced deficiency could be readily reversed in all three cell populations by further incubation of cells in drug-free medium and, in some cases, by prior growth of cells in ascorbate-supplemented medium to stimulate collagen and possibly collagen receptor production. However, several pieces of evidence indicate that reduced amounts of FN and collagen synthesized by fibroblasts do not contribute to the cycloheximide-induced deficiency, including the inability to reverse the effect by treatment of cells with TGF& Several conclusions are suggested from these studies: (a) Down’s dermal fibroblasts become deficient in collagen-specific receptor(s) upon cycloheximide treatment, which leads to altered transmembrane signaling and inability to reorganize F-actin into stress fibers; (b) Down’s dermal fibroblasts at all passage levels have matrix adhesive phenotypes similar to those of aging flbroblasts from normal individuals; and (c) these studies provide further support for cells from Down’s patients as a genetic model of aging in normal populations. 0 1990 Academic Press, Inc.
The molecular mechanisms by which dermal fibroblasts from human skin interact with and are regulated by their extracellular matrices are being analyzed with a variety of experimental approaches [l, 21. Fibronectin (FN) and type I collagen in these matrices are essential factors in fibroblast adhesion processes [3-S]. Much less is known about the similarities and differences in these regulatory mechanisms for the papillary and reticular fibroblast subsets 19] of human skin, as well as alteration of processes during natural and premature aging processesof skin. Dermal fibroblasts, including both papillary and reticular subsets, require at least two binding activities from FN in order to extend cytoplasmic processes in natural patterns and to reorganize their F-actin into microfilament stress fibers during adhesion on FN substrata in culture [lo-121. One of these activities is a cell-binding domain of FN, containing the Arg-Gly-Asp-Ser (RGDS) sequence [13], that recognizes a member of the integrin glycoprotein family on cell surfaces [4-6, 141. The second essential activity of FN recognizes a heparan sulfate proteoglycan on the cell surface in order to generate complete cytoskeletal reorganization in cells [15]. These two activities are insufficient to promote a complete adhesion response when presented to cells on separate proteolytic fragments of plasma or cellular FNs but become sufficient when they occur on a common fragment [ll, 121, suggesting some topological requirements of multiple receptor molecules in order to effect transmembrane signaling for stress fiber formation and focal contact formation as well [ll]. During analyses of substratum adhesion mechanisms of dermal fibroblasts aged in uiuo or in vitro (the latter commonly referred to as senescence), it was discovered that in vitro-aged papillary and reticular fibroblasts became deficient in collagen-specific receptors when they were pretreated with the protein synthesis inhibitor, cycloheximide [ 121. Their cytoplasmic spreading and 1 To whom
189
reprint
requests
should
be addressed.
0014~4S27/90
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
190
FLICKINGER
microtubule organizations were normal under these cycloheximide treatment conditions but stress fibers were completely absent from the cytoplasm of wellspread cells. This effect was not observed with in uiuoaged cells nor on FN substrata, revealing the specificity of the effect for both aging and substratum ligand contexts [ 121. The cycloheximide-induced deficiency could be reversed by prior growth of cells in medium containing ascorbate in order to stimulate their endogenous collagen production/maturation and was reversed upon cycloheximide withdrawal [ 121. These results raised the interesting question whether fibroblasts from other human aging model systems would demonstrate similar or different deficiencies in their abilities to reorganize their cytoskeletal networks and to facilitate other adhesion processes on defined substrata. For some time, Down’s syndrome patients have been considered to display some aspects of premature aging (reviewed in Ref. [16]). Therefore, dermal fibroblasts from three different Down’s patients, who died at very different ages,were tested in this adhesion experimental paradigm. All three fibroblast populations were shown to share sqme, but not all, of the aging-related deficiencies in collagen adhesion responses of fibroblasts from normal individuals. MATERIALS
AND METHODS
Cells andgrowth conditions. Human papillary and reticular dermal fibroblasts as separate populations were obtained from the inner, upper aspect of the arm of two individuals-a newborn male Caucasian infant (patient 5 and abbreviated PAP, and RET,, respectively) and a 78-year-old Caucasian male (patient 8 and abbreviated PAPs and RET*)-and have been described previously 111, 121. Full-thickness dermal fibroblasts from three separate Down’s syndrome patients (all Caucasian males with trisomy 21; cells obtained from the thorax region at autopsy) were obtained from the National Institute of Aging Cell Repository in Camden, New Jersey, as follows: AGO7096 (died at 5 months/8 days); AGO4823 (died at 5 years); and AGO8942 (died at 20 years/4 months/l9 days). All cells were free of Mycop.hnu contamination and were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% newborn calf serum, penicillin, and streptomycin in 10% CO*:humidifiedair at 37°C. In order to maximize collagen production/maturation [ 17,181 in some cultures as indicated, cells were supplemented twice a week with 50 fig/ml sodium ascorbate and were grown under these conditions at least two passages prior to experimental use. In other select cases, cells were pretreated for 24 or 48 h with 5 rig/ml TGFfi prior to cycloheximide treatments. Stock cultures were subcultured at split ratios of 1:3 and became senescent at prescribed population doubling levels as described previously [ll, 121.
Cycloheximide treatments. In all cases, the concentration of cycloheximide during the two treatment conditions described previously [12] was 2 pg/ml. In the first protocol, cells were only treated with drug during their 4-h reattachment period on new substrata and then evaluated for cytoskeletal reorganization. In the second protocol, cells were treated for 18 h in stock cultures containing complete medium, detached by EGTA treatment as described above, and then treated with the drug for 4 additional h during reattachment on new substrata [12]. Cells maintained their viability during these treatment conditions, as well as upon removal of the drug from the medium.
AND
CULP
Adhesion and cytoskeletal assays. Glass coverslips in 24-well cluster dishes were adsorbed with one of the following proteins whose preparation and purification have been described previously [12]: (a) 100 rg/ml rat tail collagen (type I; Vitrogen 100) with drying overnight at roomtemperature; (b) collagen for 1 h, followed by three PBS rinses and adsorption of 20 pg/ml human pFN for 1 h at 37°C; (c) human pFN alone for 1 h at 37°C; and (d) 250 pg/ml heat-treated bovine serum albumin (BSA) in DMEM (adhesion medium) for 1 h at 37°C. All coated substrata were then rinsed three times with PBS and adhesion medium added for 1 h at 37°C to guarantee coverage of substratum sites with BSA prior to cell inoculation. Meanwhile, stock cultures of dermal fibroblasts (with or without cycloheximide treatment as indicated) were rinsed three times with PBS and cells at 60-80% confluence detached with 0.5 mM EGTA in PBS by gentle shaking at 37°C for 30 min and pipetting of the cell suspension over the substratum to guarantee detachment of all cells. Cells were rinsed twice with adhesion medium by centrifugation/resuspension and finally suspended in adhesion medium prior to inoculation of 2.5 X 10’ cells into wells containing coated coverslips and 250 pl adhesion medium. In some cases to quantitate attachment of cells, stock cultures were radiolabeled with [methyl-3H]thymidine incorporation and radiolabeled adherent cells assayed as described previously [ 111. After 4 h when cell spreading had maximized, cells on coverslips were fixed with 3.7% (v/v) paraformaldehyde in PBS (with 1 mM CaCl, and MgCl& for 20 min at room temperature, permeabilized with 1 ml of 0.2% (v/v) Triton X-100 in PBS with divalent cations, and stained for cytoskeletal components as follows. F-actin reorganization was evaluated by staining with rhodamine-phalloidin (250 ~1 of 2.0 units/ml) in PBS for 20 min at room temperature [ 121. For microtubule staining, a rabbit polyclonal antiserum to chicken tubulin crossreactive with human tubulin was used at 1:60 dilution, and detection was with a rhodamine-conjugated goat anti-rabbit IgG serum. For intermediate filament staining, a goat polyclonal antiserum to mouse fibroblast vimentin cross-reactive with human vimentin was used at 1:50 dilution, and detection was with a rhodamine-conjugated rabbit anti-goat IgG serum. After staining, all coverslips were rinsed three times with PBS and inverted on glass slides in 50% glycerol:50% PBS. Fluorescent cells were observed and photographed on a Nikon Diaphot microscope equipped with epifluorescence illumination using Kodak 2475 recording film. Materials. Materials were purchased from the following sources: 24-well cluster dishes from Costar (Cambridge, MA); DMEM from GIBCO (Grand Island, NY); newborn calf serum from Biologos (Naperville, IL); la-mm-diameter glass coverslips from Dynalab (Rochester, NY); rat tail type I collagen as Vitrogen 100 from Collagen Corp. (Palo Alto, CA); cycloheximide, sodium ascorbate, bovine serum albumin, Triton X-100, and EGTA from Sigma Chemical Co. (St. Louis, MO); paraformaldehyde from Fisher Scientific Co. (Fairlawn, NJ); rhodamine-phalloidin from Molecular Probes, Inc. (Junction City, OR); anti-tubulin and anti-vimentin antisera and their indirect antibody conjugates from ICN ImmunoBiologicals (Lisle, IL); transforming growth factor-p from Calbiochem (LaJolla, CA); and [methyZ-3H]thymidine from New England Nuclear Corp. (Boston, MA).
RESULTS The fibroblasts from the youngest Down’s patient (the 5-month-old) at an early in vitro passage level were initially tested on a collagen-only substratum for any perturbation of their ability to reorganize F-actin stress fibers under untreated or experimental conditions. As shown in Fig. lA, when these cells were not treated with cycloheximide and when inoculated into wells coated with type I collagen only (without exogenous FN), they
DOWNS
CELLS
AND
COLLAGEN
ADHESION
Down’s Early-pass.
Normal Early-pass.
Normal Late-pass.
FIG. 1. F-actin reorganization in normal and Down’s dermal fibroblasts with or without cycloheximide treatment on collagen substrata. Early-passage Down’s fibroblasts from the &month-old patient (A and B), early-passage PAPS cells from a normal individual (C and D), and late-passage PAPS cells (E and F) were evaluated for F-actin reorganization with rhodamine-phalloidin staining on collagen-only substrata as described under Materials and Methods. Some cells were not treated with cycloheximide (Cycle-) while other cells were treated with the inhibitor (Cycle+) as follows: for 18 h in stock cultures, then EGTA-detached and rinsed with PBS, and finally inoculated into collagen-coated wells containing adhesion medium and cycloheximide. Cells were fixed and stained at 4 h as described under Materials and Methods. Typical F-actin stress fibers are shown at the solid black arrows and star-shaped clusters of F-actin in some cells with the curved black arrows. Magnification, X270.
attached and spread effectively within the 4-h time period and reorganized extensive arrays of stress fibers (e.g., at the arrow). However, when these cells were pretreated overnight for 18 h in stock culture and then during the 4-h reattachment period on the collagen-only
substratum with cycloheximide, cell spreading continued to be excellent but the cytoplasm was devoid of microfilament stress fibers (Fig. 1B). The cytoplasm was now filled with star-shaped clusters of F-actin (curved arrow in Fig. 1B) that may be some intermediate com-
192
FLICKINGER
AND
CULP
FIG. 2. Defective stress fiber formation in two other Down’s fibroblast populations on collagen substrata. Early-passage Down’s fibroblasts from the B-year-old patient (A) and the 20-year-old patient (B) were treated with cycloheximide, inoculated into collagen-only wells containing adhesion medium and cycloheximide, and evaluated for F-actin reorganization with rhodamine-phalloidin staining as described under Materials and Methods and in the legend to Fig. 1. Primitive linear bundles of F-actin are shown with open arrows and star-shaped clusters of F-actin with black arrowheads. Magnification, X380.
plex in formation and/or dissolution of microfilament stress fibers. This continued to be the case at all population doubling levels of these cells over their in vitro lifespan, pertaining to >85% of the cells in the population. This was not the case under another treatment condition (data not shown)-when the cycloheximide treatment was applied only during the 4-h reattachment period, indicating that a particularly labile set of proteins did not require resynthesis during the reattachment period as a consequence of the EGTA-mediated subculturing. When the collagen substratum was postadsorbed
with human plasma FN and cells were treated with cycloheximide for 22 h as in Fig. lB, they persisted in their inability to generate stress fibers, indicating that a FN deficiency in the collagen matrix cannot explain this effect. This collagen-specific deficiency in stress fiber reorganization is similar to previous observations of normal dermal fibroblasts in aging populations [121. For example, early-passage PAP5 cells from a normal newborn infant challenged with a collagen-only substratum make excellent stress fibers under untreated (Fig. 1C) or cyclo-
DOWNS
CELLS
AND
COLLAGEN
ADHESION
193
FIG. 3. Stress fiber formation in Down’s fibroblasts on FN substrata. Early-passage fibroblasts from the 5-month-old Down’s patient were left untreated (A) or were treated with cycloheximide (B) for 18 h prior to EGTA-mediated detachment and then for 4 h during reattachment on substrata as described under Materials and Methods and the legend to Fig. 1; however, cells were plated on substrata coated with human plasma FN only. Rhodamine-phalloidin-stainable stress fibers are prevalent in both classes of cells (solid black arrows). Magnification, X380.
heximide treatment (Fig. 1D) conditions. In contrast, late-passage PAP5 cells entering their senescent period displayed the same discriminating behavior as the Down’s fibroblasts did-untreated cells made excellent stress fibers on a collagen-only substratum (arrow in Fig. 1E) while cycloheximide-treated cells spread well and were completely devoid of stress fibers with a cytoplasm containing star-shaped clusters of F-actin (curved arrow in Fig. 1F). The cell in Fig. 1F is representative of >85% of the cells in the entire population under these experimental conditions. The same was true for RETs cells from the same individual (data not shown), indicating
that both major cell classes of the human dermis display this effect [12]. These results indicate that dermal fibroblasts from the 5-month-old Down’s patient at all population doubling values behaved like aging skin fibroblasts from normal individuals. In order to test the generality of this effect, dermal fibroblasts from two other Down’s patients were examined, these patients had died at 5 years and 20 years of age (see Materials and Methods). As shown in Fig. 2A, early-passage cells from the 5-year-old Down’s patient displayed the same incapacity for stress fiber formation after cycloheximide treatment, as did early-passage cells
194
DOWNS
CELLS
AND
from the 20-year-old Down’s patient (Fig. 2B), with only a few primitive F-actin bundles in the cytoplasm (open arrows) and star-shaped clusters of F-actin prevalent throughout the cytoplasm (solid arrowheads). The untreated cells from these two individuals reorganized stress fibers effectively (data not shown). Therefore, this observation is not a peculiarity of the dermal fibroblasts from one particular patient but applies to dermal fibroblasts harvested from three different Down’s patients who died at very different ages. The substratum specificity of the effect was evaluated as well. A substratum coated with human plasma FN was tested for comparison with the results above on a collagen-only substratum. Untreated cells from the 5 month-old Down’s patient generated stress fibers as expected (e.g., at the arrow in Fig. 3A). Furthermore, cycloheximide-treated cells generated stress fibers effectively (e.g., at the arrow in Fig. 3B). The same was true for the Down’s fibroblasts from the other two patients (data not shown). Therefore, the cycloheximide-induced deficiency is collagen-specific (and unrelated to fibronectin receptor activity) and cannot be explained by a deficiency in a cytoskeletal component per se, since the same cycloheximide-treated cells do very well on a FN-only substratum but are incompetent on a collagen-only substratum. Furthermore, supplementing the collagen layer with exogenous FN does not overcome the deficiency (see above), indicating a dominant effect by collagenspecific receptors at the surface of these cells in these phenotypic changes. In normal aging dermal fibroblasts, the cycloheximide-induced deficiency in stress fiber formation could be reversed by several experimental conditions [ 121. Potential reversal of the inhibition was therefore tested with the Downs’ cells as well. Removal of the drug was evaluated initially. Cells were pretreated for 18 h with the drug, EGTA-detached, inoculated into collagen-only wells in medium containing the drug for 4 more h, and then rinsed well with drug-free medium for an additional 16 h incubation in the absence of drug before evaluating F-actin organization. As shown in Fig. 4, cells from the &month-old Down’s patient (Fig. 4A), the &year-old patient (Fig. 4B), and the 20-year-old (Fig. 4C) regained competence for forming stress fibers throughout their cytoplasm (arrows). It should also be noted that the density of stress fibers and their lengths increased considerably during these reversal conditions. Therefore, cycloheximide treatment does not lead to irreversible damage to Down’s fibroblasts such that they are no longer competent for reorganizing their microfilament bundles.
COLLAGEN
ADHESION
195
A second approach for reversal was also tested. Down’s cells were pregrown for two passages in medium supplemented with sodium ascorbate in order to stimulate collagen production/maturation processes [ 17,181. In previous experiments [ 121, this approach has successfully reversed cycloheximide inhibition in the aging normal fibroblasts. In this case with fibroblasts from the Down’s patients, two different results were obtained. With cells from the 20-year-old Down’s patient, ascorbate supplementation was incapable of replenishing the receptor moieties required for cells to generate stress fibers on collagen-only substrata under cycloheximide treatment (Figs. 5A and 5C). In contrast, cells from the 5-year-old after ascorbate supplementation became competent in forming stress fibers (Figs. 5B and 5D). These results may indicate any one of several possibilities-(a) there are two different molecular mechanisms of deficiency operating here; (b) Down’s cells are differentially susceptible to stimulation with ascorbate; (c) there are different stoichiometric levels of receptor activity being depleted in these two populations; or (d) Down’s cells display an aging-related competence for ascorbate reversal since cells from the oldest patient could not be reversed while those from the younger patient could be reversed. A third approach was tested based on recent evidence that fibroblasts treated with transforming growth factor-@ (TGFB) produce much more collagen, FN, and FNspecific integrin-an up-regulation that occurs at the transcriptional level of the respective genes [19-211. When Down’s cells were treated with TGF@ for 24 or 48 h prior to their receiving cycloheximide, there was no reversal of the cycloheximide-induced deficiency upon stress fiber formation (data not shown). This result indicates a mechanistic difference of this treatment with the ascorbate reversal described above and with the TGFP stimulation of FN and collagen production per se. Therefore, this difference may also be consistent with stoichiometric differences in collagen-specific receptor levels under these treatment conditions and not in the matrix ligands themselves. The inability of TGF/3 pretreatment of cells to reverse the deficiency also indicates that collagenolytic destruction of the collagen substratum, as shown recently for aging fibroblasts [ 311, may not be increased upon cycloheximide treatment. To further evaluate this possibility, Down’s cells were treated with cycloheximide by the usual 22-h protocol but their adhesion responses were evaluated on collagen-only substrata in medium that contained 50 pg/ml human plasma FN which could re-
FIG. 4. Cycloheximide removal and regeneration of stress fibers. Early-passage Down’s fibroblasta from the 5-month-old (A), the B-yearold (B), and the 20-year-old (C) patients were pretreated with cycloheximide for 18 h as usual, EGTA-detached and rinsed, inoculated into collagen-only wells for 4 h in adhesion medium containing cycloheximide, and then fed medium without cyclo/zeximide for 16 h when cells were fixed for rhodamine-phalloidin staining. Dense and extremely long stress fibers are observed throughout all cells (solid black arrows). Magnification, X365.
FIG. 5. Differential recovery from the cycloheximide-induced inhibition by ascorbate-supplemented cells. Early-passage fibroblasts from the 20year-old (A and C) and &year-old (B and D) Down’s patients were grown for two passages in complete medium supplemented with sodium ascorbate as described under Materials and Methods. Cells were then left untreated (A and B) or treated with cycloheximide (C and D) as described in the legend to Fig. 1 in medium containing ascorbate. After 4 h on collagen-only substrata in medium with cycloheximide, cells were fixed and stained with rhodamine-phalloidin. Extensive stress fibers are observed in three cases (A, B, and D) as pointed out with solid black arrows, while in case C the cytoplasm remained devoid of stress fibers and contained star-shaped clusters of F-actin (solid white arrow). Magnification, X410.
DOWN’S
CELLS
AND
place collagen lost from the substratum. However, stress fibers were still not observed. It is not likely that increased production of collagenase in cycloheximidetreated cells, a stimulation difficult to visualize because of the general shutdown of protein synthesis in drugtreated cells, can explain the adhesion deficiency. However, these experiments do not rule out the possibility for increased proteolytic turnover of collagen-specific receptors in treated cells for any one of a number of reasons. The excellent spreading of these dermal fibroblasts (either normal aging or Down’s populations upon cycloheximide treatment) in the complete absence of microfilament stress fiber formation is a remarkable finding and may indicate that the cell’s other cytoskeletal networks may be functioning at maximal capacity in these spreading responses. This was tested with antisera to tubulin for evaluation of microtubule networks and to vimentin for evaluation of intermediate filament networks. As shown in Fig. 6A (white arrowhead), untreated late-passage RET5 cells generated an extensive microtubule network throughout their cytoplasm. This was also true for the same cells under cycloheximide treatment (Fig. 6C; white arrowhead) when they were completely devoid of F-actin stress fibers. The same case applied for vimentin-containing intermediate filaments (Figs. 6B and 6D; open arrows) and for papillary cells from the same individual (data not shown). Therefore, while aging normal fibroblasts become incompetent for forming stress fibers, they remain competent for forming two other cytoskeletd networks that may not require cell surface receptor mediation [S]. These data provide further support for depletion of a collagen-specific receptor at the cell surface upon cycloheximide treatment as being responsible for the lack of stress fiber formation. Down’s fibroblast networks were also evaluated for microtubule and intermediate filament networks. As shown in Figs. 7A and 7B for the 20-year-old patient, microtubule networks formed normally in both cell classes (white arrowheads). The same was true for the other two Down’s patients’ cells and for intermediate filament networks (data not shown). Therefore, fibroblasts from Down’s patients behave identically to fibroblasts from normal patients undergoing in vitro aging in many regards. DISCUSSION
These results indicate that dermal fibroblasts from Down’s patients share a deficiency, inducible with cycloheximide treatment, in collagen-specific adhesion responses with aging dermal fibroblasts from normal individuals, an alteration that is not observed in early-passage normal cell populations [12]. Furthermore, this deficiency applies to Down’s cell populations from three
COLLAGEN
ADHESION
197
individuals who died at very different ages. This evidence supports the argument that Down’s patients experience premature aging processes in skin, as shown by the studies here, and in some other tissues as well [16, 22-25, 301. It also implicates a genetic basis for this alteration in the Down’s populations and contrasts with the aging-related changes at the phenotypic level observed in dermal fibroblast populations from normal individuals. The adhesion deficiency is specific for F-actin stress fiber formation in cells that continue to spread very effectively, displaying normal arrays of microtubule and intermediate filament networks. This deficiency in stress fiber formation cannot be explained by limiting amounts of cellular FN or collagen synthesized by cells, since cycloheximide did not alter stress fiber formation on human plasma FN-only substrata. Also, treatment of cells with TGFj3, known to stimulate collagen and cellular FN production at the transcriptional level [19-211, could not reverse the effect. Additional evidence in this study points to a stoichiometric limitation in collagen-specific receptors during the 22-h cycloheximide treatment in Down’s fibroblasts as the molecular basis for the deficiency. There are no limitations on the amounts of cytoskeletal proteins required for stress fiber formation per se in Down’s cells since fibers form perfectly well on FN-only substrata during drug treatment. In support of this latter argument, late-passage human lung fibroblasts have increased amounts of F-actin in their cytoskeleton when compared with early-passage cells [36]. Cell-synthesized collagen and FN do not appear to be limiting since plasma FN supplementation of the collagen substratum or the medium failed to prevent the deficiency and TGF/3 was without effect. However, countering this argument is the reversal of the inhibition in some, but not all Down’s cells, by ascorbate supplementation of cells prior to and during the cycloheximide treatment. Ascorbate is known to stimulate collagen production in human derma1 fibroblasts [ 10,17,18,41] and in parallel may stimulate production of collagen-specific receptors, although this possibility has not been tested to date. In this regard, fibroblasts from dermal keloid tumors of certain individuals have been shown to up-regulate synthesis of both cellular FN and the FN integrin receptor in a coordinate fashion [ 341. Ascorbate treatment of human foreskin fibroblasts has been shown in two-dimensional cultures to generate highly aligned collagen fibrils which align in register with cell surface attachment regions [41]. Also, limitation of collagen-specific receptors in aging and in Down’s cell populations may have consequences on the abilities of these cells to contract a threedimensional collagen matrix [36-401, for example the matrix surrounding these cells in situ. Another possibility for the deficiency in stress fiber formation is the destruction of the collagen substratum
198
FLICKINGER
.
E ? .I
AND
CULP
DOWN’S
CELLS
AND
COLLAGEN
ADHESION
199
FIG. 7. Microtubule networks in cycloheximide-treated Down’s fibroblasts. Early-passage Down’s fibroblasts from the 20-year-old patient were treated as described in the legend to Fig. 6 without (A) or with (B) cycloheximide. After 4 h on collagen-only substrata in media with cycloheximide, cells were fixed and stained with antisera to tubulin as described under Materials and Methods. Microtubule networks are shown throughout the cytoplasm (white arrowheads) in both cases. Magnification, X380.
by collagenase secreted from aging cells as described recently [31]. However, several experiments run counter to this argument. First, collagenase production should be down-regulated, and not up-regulated, by cycloheximide treatment of these cells. If collagenase was generat-
ing an incompetent substratum, this effect should have also been observed in the noncycloheximide-treated cells. This was never observed in this and previous studies [X2]. Second, TGF/3 treatment, known to stimulate endogenous collagen production, could not reverse the
FIG. 6. Microtubule and intermediate filament networks in cycloheximide-treated normal normal RET5 population were left untreated (A and B) or were treated with cycloheximide (C under Materials and Methods. After 4 h on collagen-only substrata in adhesion medium with or antisera to tubulin (Anti-Tub.) or antisera to vimentin (Anti-Vim.) as described under Materials C are evident throughout the cytoplasm (white arrowheads). Similarly, extensive intermediate cytoplasm in B and D (open arrows). Magnification, X410.
fibroblasts. Late-passage fibroblasts from the and D) as described in the legend to Fig. 1 and without drug, cells were fixed and stained with and Methods. Microtubule networks in A and filament networks are evident throughout the
200
FLICKINGER
effect. Third, supplementation of the medium of cycloheximide-treated cells with high concentrations of plasma FN (in the absence of any bovine albumin) would be expected to overcome this deficiency; this was not observed. Fourth, anti-collagen I staining of the substratum throughout these experiments revealed a continuous collagen matrix which maintained its integrity, inconsistent with this possibility. In further consideration of this possibility, transcription of both stromelysin and collagenase genes in many cell types is coordinately upregulated [42] and in some cases the regulation is mediated by FN receptor activity [43]. In the studies reported here, collagen receptor activity and not FN receptor activity mediated the cycloheximide-induced deficiency in stress fiber formation, providing further evidence for a mechanism independent of collagenase action. Detailed molecular characterization of the collagen-specific receptors of these cell populations under these treatment conditions should ultimately resolve this issue. With regard to potential receptor activities of interest, two molecular classes require attention. The first is the integrin glycoprotein complex of receptors on the surface of cells that recognize many extracellular matrix molecules [4-7,141 and that organize into focal contacts in substratum-specific patterns [44]. This family in human fibroblasts is composed of several members, some of which are specific for the target matrix ligand and some of which have overlapping specificities [26,27,35]. Several members of the integrin family are found in various subpopulations of human keratinocytes as well [35] and some integrins are shared in common with dermal fibroblasts [26,27]. The a2/31 integrin appears to be collagen-specific [26,27,35,45,46] and may be the candidate receptor for the cycloheximide-induced deficiency observed in these and previous studies [ 121. The second class of potential receptors are the cell surface heparan sulfate proteoglycans which mediate adhesion of dermal fibroblasts to FN and possibly collagen matrices [lo-12, 151 and of epithelial cells to collagen matrices [28, 291. Analysis of the stoichiometric amounts, synthesis, and turnover properties of these receptor classes can now be initiated by several experimental approaches. The results of this study on molecular deficiencies in a matrix adhesion response can be compared to other studies of phenotypic changes in cells from various tissues of Down’s patients. The trisomy of chromosome 21 would be expected to increase the dosage of gene products coded on this chromosome in such cells. This was one explanation for the increased intercellular adhesion and aggregation (divalent cation-dependent) of lung or cardiac fibroblasts from Down’s patients when compared to those of normal controls [32]. However, in these same studies dermal fibroblasts demonstrated decreased adhesion, consistent with a tissue-specific reduction in some receptor activity as suggested from our own studies [32]. Lymphoblastoid cell lines generated from circulat-
AND
CULP
ing cells of Down’s patients and immortalized with Epstein-Barr virus infection demonstrated increased aggregation, relative to normal patient control cells [33]. This aggregation was dependent upon elevated levels of one of the integrin family members in these cells [33]. Therefore, the trisomy 21 condition of Down’s syndrome may lead to differing phenotypic changes in various cell types that cannot be explained by simple gene dosage effects. The authors acknowledge support for these studies from National Institutes of Health research Grant AGO2921 and the technical assistance of Victor Guinto with the photomicrography.
REFERENCES 1.
Uitto,
2.
Smith,
3. 4. 5. 6. 7. 8.
J. (1986)
Dermutol. Clin. 4,433.
L. T., Holbrook, Dermatol. 105,57.
K. A., and Byers,
P. H. (1982)
J. Invest.
Annu. Rev. Biochem. 62,761. Biophys. A 907,261. Ruoslahti, E. (1988) Annu. Rev. Biochem. 57,375. Buck, C., and Horwitz, A. F. (1987) Anna Rev. Cell Biol. 3,179. McDonald, J. A. (1988) Annu. Rev. Cell Biol. 4,183. Yamada,
K. M. (1983)
Juliano,
Burridge, (1988)
R. L. (1987)
Biochim.
K., Fath,
K., Kelly,
T., Nuckolls,
G., and Turner,
C.
Annu. Rev. Cell Biol. 4,487.
9.
Harper,
10.
Beyth,
R. A., and Grove,
11.
Hall, M. D., Flickinger, L. A. (1988) Erp. Cell
R. J., and Culp,
G. (1979)
Science
204,526.
Exp. Cell Res. 155,537.
L. A. (1984)
K. S., Cutolo,
M.,
Zardi,
L., and Culp,
Res. 179,115.
Flickinger,
K. S., and Culp,
13.
Ruoslahti,
E., and Pierschbacher,
14.
Hynes,
15.
Culp, L. A., Laterra, J., Lark, M. W., Beyth, R. J., and Tobey, S. L. (1986) in Functions of the Proteoglycans (Evered, D., and Whelan, J., Eds.), CIBA Symposium No. 124, pp. 158-183, Wiley, London.
16.
Brown, W. T. (1989) Kligman, A. M., Eds.),
17.
Pinnell,
18.
Booth,
R. 0. (1987)
S. R. (1979) B. A., Polak,
L. A. (1990)
Exp. Cell Res. 186,158.
12.
M. D. (1987)
Science
238,491.
Cell 48,549.
in Aging pp. 93-120,
and the Skin (Balin, Raven Press, New
A. K., and York.
J. Invest. Dermntol. 73,485. J. (1980) Biochim. Biophys.
K. L., and Uitto,
A 607,145. 19.
Dean,
D. C., Newby,
S. (1988)
R. F., and Bourgeois,
J. CeU Biol.
106,2159. 20.
Roberts, C. J., Birkenmeier, S. K., Yamada, K. M., Chen,
21.
Keski-Oja, tlethwaite,
C&em.
J., Raghow, A. E., Kang, 263,311l.
22.
Uitto,
J., Fazio,
T. M., McQuillan, J. J., Akiyama, W.-T., and McDonald, J. A. (1988)
J. Biol. Chem. 263,4586. R., Sawdey, M., Loskutoff, D. J., PosA. H., and Moses, H. L. (1988) J. Bid.
M. J., and Olsen,
D. R. (1988)
J. Cutaneous Aging
Cosmetic Dermatol. 1,13. 23.
Imayama, 1019.
S., andBraverman,
J. Neurol. Sci. 79,91.
24.
Plioplys,
A. V. (1987)
25.
Plioplys, 233. Wayner,
A. V., and Bedford,
26.
Amer. J. Puthol. 134,
I. M. (1989)
E. A., and Carter,
H. M. (1989) W. G. (1987)
Brain Res. Bull. 22,
J. Cell Biol. 105,1873.
DOWN'S CELLS AND COLLAGEN ADHESION 27.
Wayner, E. A., Carter, T. J. (1988) J. Cell&X.
28.
Koda,
29.
LeBaron, R. G., Hook, A., Esko, (1989) J. Biol. Chem. 264,795O.
30.
Martin,
31.
Sottile, J., Mann, D. M., J. Cell. Physiol. 138,281.
32.
Wright, T. C., Orkin, R. W., Destrempes, M., (1984) Proc. Natl. Ad. Sci. USA 81,2426.
33.
Taylor, G. M., Haigh, Immunology 64,451.
34.
Babu, M., 9.1642.
35.
Carter, (1990)
36.
Gibson, J. M., Milam, Dev. 48,101.
Received
W. G., Piotrowicz, 10’7,188l.
J. E., and Bernfield,
G. M. (1977)
Diegelmann,
M. (1984)
Birth
Defects Diemer,
H., D’Souza,
J. Biol.
1,199O
Chem.
J. D., Gay,
Hook,
A. J. T. (1989)
and Kurnit,
N. (1989)
Mol.
D. M.
42. 43.
Cell. Bid.
44.
Kaur,
45.
Mech.
Age@
P.
Schafer, I. A., Shapiro, A., Kovach, M., Lang, C., and Fratianne, R. B. (1989) Exp. Cell Res. 8,259. Danowski, B. A., and Harris, A. K. (1988) Exp. Cell Res. 17'7,47. Montesano, R., and Orci, L. (1988) Proc. Natl. Acad. Sci. USA 85,4894. Nakagawa, S., Pawelek, P., and Grinnell, F. (1989) Elcp. Cell Res.
182,572. 41.
R. (1988)
T. S., and
R. J. (1989)
39. 40.
S. W., and Harris,
S. B., and Klebe,
38. M.
Ser. 14,5.
V., and Millis,
R., and Oliver,
37.
259,11763.
S., and
Orig. Artic.
W. G., Wayner, E. A., Bouchard, J. Cell Biol. 110,1387.
March
R. S., and Kunicki,
201
46.
Grinnell, F., Fukamizu, H., Pawelek, P., and Nakagawa, S. (1989) Exp. Cell Res. 181,483. Frisch, S. M., Clark, E. J., and Werb, Z. (1987) Proc. Natl. Acad. Sci. USA 84,260O. Werb, Z., Tremble, P. M., Behrendtsen, O., Crowley, E., and Damsky, C. H. (1989) J. Cell Biol. 109,877. Fath, K. R., Edgell, C.-J. S., and Burridge, K. (1989) J. Cell Sci. 92,67. Ogle, R. C., Potts, J., Yacoe, M., and Little, C. D. (1989) J. Cell Sci. 94,361. Gullberg, D., Terracio, L., Borg, T. K., and Rubin, K. (1989) J. Biol. Chem. 264,12,686.
RESEARCH
189,189-201(1990)
Dermal Fibroblasts from Down’s Syndrome Patients Share a Cycloheximide-Induced Deficiency in Collagen Adhesion Responses with Normal Aging Cells KELLY S. FLICKINGER’ lkportment
of
AND LLOYD A. CULP
Molecular Biology and Microbiology, Case Western Reserve University, School of Medicine, Clmelund, Ohio 44106
INTRODUCTION Human skin fibroblasts from three different Down’s syndrome patients (trisomy 21) of very different ages have been tested for their adhesion responses on tissue culture substrata coated with type I collagen, fibronectin (FN), and their combination after or during treatment of cells with cycloheximide to evaluate limitations in specific responses. It was shown previously that in vitro-aged papillary and reticular dermal fibroblasts from normal individuals do not generate F-a&in stress fibers when pretreated with cycloheximide on collagen substrata but do so on FN substrata, a deficiency linked to limiting amounts/function of collagen-specific receptors in aging cells. In these studies, all three Down’s fibroblast populations demonstrated a similar deficiency in stress fiber formation, evaluated by rhodaminephalloidin staining, upon cycloheximide treatment at all passage levels. They remained competent for stress fiber formation on FN substrata and for reorganization of microtubule and intermediate filament networks on all substrata, demonstrating the specificity for the collagen matrix and for the F-actin cytoskeleton in this deficiency. The cycloheximide-induced deficiency could be readily reversed in all three cell populations by further incubation of cells in drug-free medium and, in some cases, by prior growth of cells in ascorbate-supplemented medium to stimulate collagen and possibly collagen receptor production. However, several pieces of evidence indicate that reduced amounts of FN and collagen synthesized by fibroblasts do not contribute to the cycloheximide-induced deficiency, including the inability to reverse the effect by treatment of cells with TGF& Several conclusions are suggested from these studies: (a) Down’s dermal fibroblasts become deficient in collagen-specific receptor(s) upon cycloheximide treatment, which leads to altered transmembrane signaling and inability to reorganize F-actin into stress fibers; (b) Down’s dermal fibroblasts at all passage levels have matrix adhesive phenotypes similar to those of aging flbroblasts from normal individuals; and (c) these studies provide further support for cells from Down’s patients as a genetic model of aging in normal populations. 0 1990 Academic Press, Inc.
The molecular mechanisms by which dermal fibroblasts from human skin interact with and are regulated by their extracellular matrices are being analyzed with a variety of experimental approaches [l, 21. Fibronectin (FN) and type I collagen in these matrices are essential factors in fibroblast adhesion processes [3-S]. Much less is known about the similarities and differences in these regulatory mechanisms for the papillary and reticular fibroblast subsets 19] of human skin, as well as alteration of processes during natural and premature aging processesof skin. Dermal fibroblasts, including both papillary and reticular subsets, require at least two binding activities from FN in order to extend cytoplasmic processes in natural patterns and to reorganize their F-actin into microfilament stress fibers during adhesion on FN substrata in culture [lo-121. One of these activities is a cell-binding domain of FN, containing the Arg-Gly-Asp-Ser (RGDS) sequence [13], that recognizes a member of the integrin glycoprotein family on cell surfaces [4-6, 141. The second essential activity of FN recognizes a heparan sulfate proteoglycan on the cell surface in order to generate complete cytoskeletal reorganization in cells [15]. These two activities are insufficient to promote a complete adhesion response when presented to cells on separate proteolytic fragments of plasma or cellular FNs but become sufficient when they occur on a common fragment [ll, 121, suggesting some topological requirements of multiple receptor molecules in order to effect transmembrane signaling for stress fiber formation and focal contact formation as well [ll]. During analyses of substratum adhesion mechanisms of dermal fibroblasts aged in uiuo or in vitro (the latter commonly referred to as senescence), it was discovered that in vitro-aged papillary and reticular fibroblasts became deficient in collagen-specific receptors when they were pretreated with the protein synthesis inhibitor, cycloheximide [ 121. Their cytoplasmic spreading and 1 To whom
189
reprint
requests
should
be addressed.
0014~4S27/90
$3.00
Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
190
FLICKINGER
microtubule organizations were normal under these cycloheximide treatment conditions but stress fibers were completely absent from the cytoplasm of wellspread cells. This effect was not observed with in uiuoaged cells nor on FN substrata, revealing the specificity of the effect for both aging and substratum ligand contexts [ 121. The cycloheximide-induced deficiency could be reversed by prior growth of cells in medium containing ascorbate in order to stimulate their endogenous collagen production/maturation and was reversed upon cycloheximide withdrawal [ 121. These results raised the interesting question whether fibroblasts from other human aging model systems would demonstrate similar or different deficiencies in their abilities to reorganize their cytoskeletal networks and to facilitate other adhesion processes on defined substrata. For some time, Down’s syndrome patients have been considered to display some aspects of premature aging (reviewed in Ref. [16]). Therefore, dermal fibroblasts from three different Down’s patients, who died at very different ages,were tested in this adhesion experimental paradigm. All three fibroblast populations were shown to share sqme, but not all, of the aging-related deficiencies in collagen adhesion responses of fibroblasts from normal individuals. MATERIALS
AND METHODS
Cells andgrowth conditions. Human papillary and reticular dermal fibroblasts as separate populations were obtained from the inner, upper aspect of the arm of two individuals-a newborn male Caucasian infant (patient 5 and abbreviated PAP, and RET,, respectively) and a 78-year-old Caucasian male (patient 8 and abbreviated PAPs and RET*)-and have been described previously 111, 121. Full-thickness dermal fibroblasts from three separate Down’s syndrome patients (all Caucasian males with trisomy 21; cells obtained from the thorax region at autopsy) were obtained from the National Institute of Aging Cell Repository in Camden, New Jersey, as follows: AGO7096 (died at 5 months/8 days); AGO4823 (died at 5 years); and AGO8942 (died at 20 years/4 months/l9 days). All cells were free of Mycop.hnu contamination and were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% newborn calf serum, penicillin, and streptomycin in 10% CO*:humidifiedair at 37°C. In order to maximize collagen production/maturation [ 17,181 in some cultures as indicated, cells were supplemented twice a week with 50 fig/ml sodium ascorbate and were grown under these conditions at least two passages prior to experimental use. In other select cases, cells were pretreated for 24 or 48 h with 5 rig/ml TGFfi prior to cycloheximide treatments. Stock cultures were subcultured at split ratios of 1:3 and became senescent at prescribed population doubling levels as described previously [ll, 121.
Cycloheximide treatments. In all cases, the concentration of cycloheximide during the two treatment conditions described previously [12] was 2 pg/ml. In the first protocol, cells were only treated with drug during their 4-h reattachment period on new substrata and then evaluated for cytoskeletal reorganization. In the second protocol, cells were treated for 18 h in stock cultures containing complete medium, detached by EGTA treatment as described above, and then treated with the drug for 4 additional h during reattachment on new substrata [12]. Cells maintained their viability during these treatment conditions, as well as upon removal of the drug from the medium.
AND
CULP
Adhesion and cytoskeletal assays. Glass coverslips in 24-well cluster dishes were adsorbed with one of the following proteins whose preparation and purification have been described previously [12]: (a) 100 rg/ml rat tail collagen (type I; Vitrogen 100) with drying overnight at roomtemperature; (b) collagen for 1 h, followed by three PBS rinses and adsorption of 20 pg/ml human pFN for 1 h at 37°C; (c) human pFN alone for 1 h at 37°C; and (d) 250 pg/ml heat-treated bovine serum albumin (BSA) in DMEM (adhesion medium) for 1 h at 37°C. All coated substrata were then rinsed three times with PBS and adhesion medium added for 1 h at 37°C to guarantee coverage of substratum sites with BSA prior to cell inoculation. Meanwhile, stock cultures of dermal fibroblasts (with or without cycloheximide treatment as indicated) were rinsed three times with PBS and cells at 60-80% confluence detached with 0.5 mM EGTA in PBS by gentle shaking at 37°C for 30 min and pipetting of the cell suspension over the substratum to guarantee detachment of all cells. Cells were rinsed twice with adhesion medium by centrifugation/resuspension and finally suspended in adhesion medium prior to inoculation of 2.5 X 10’ cells into wells containing coated coverslips and 250 pl adhesion medium. In some cases to quantitate attachment of cells, stock cultures were radiolabeled with [methyl-3H]thymidine incorporation and radiolabeled adherent cells assayed as described previously [ 111. After 4 h when cell spreading had maximized, cells on coverslips were fixed with 3.7% (v/v) paraformaldehyde in PBS (with 1 mM CaCl, and MgCl& for 20 min at room temperature, permeabilized with 1 ml of 0.2% (v/v) Triton X-100 in PBS with divalent cations, and stained for cytoskeletal components as follows. F-actin reorganization was evaluated by staining with rhodamine-phalloidin (250 ~1 of 2.0 units/ml) in PBS for 20 min at room temperature [ 121. For microtubule staining, a rabbit polyclonal antiserum to chicken tubulin crossreactive with human tubulin was used at 1:60 dilution, and detection was with a rhodamine-conjugated goat anti-rabbit IgG serum. For intermediate filament staining, a goat polyclonal antiserum to mouse fibroblast vimentin cross-reactive with human vimentin was used at 1:50 dilution, and detection was with a rhodamine-conjugated rabbit anti-goat IgG serum. After staining, all coverslips were rinsed three times with PBS and inverted on glass slides in 50% glycerol:50% PBS. Fluorescent cells were observed and photographed on a Nikon Diaphot microscope equipped with epifluorescence illumination using Kodak 2475 recording film. Materials. Materials were purchased from the following sources: 24-well cluster dishes from Costar (Cambridge, MA); DMEM from GIBCO (Grand Island, NY); newborn calf serum from Biologos (Naperville, IL); la-mm-diameter glass coverslips from Dynalab (Rochester, NY); rat tail type I collagen as Vitrogen 100 from Collagen Corp. (Palo Alto, CA); cycloheximide, sodium ascorbate, bovine serum albumin, Triton X-100, and EGTA from Sigma Chemical Co. (St. Louis, MO); paraformaldehyde from Fisher Scientific Co. (Fairlawn, NJ); rhodamine-phalloidin from Molecular Probes, Inc. (Junction City, OR); anti-tubulin and anti-vimentin antisera and their indirect antibody conjugates from ICN ImmunoBiologicals (Lisle, IL); transforming growth factor-p from Calbiochem (LaJolla, CA); and [methyZ-3H]thymidine from New England Nuclear Corp. (Boston, MA).
RESULTS The fibroblasts from the youngest Down’s patient (the 5-month-old) at an early in vitro passage level were initially tested on a collagen-only substratum for any perturbation of their ability to reorganize F-actin stress fibers under untreated or experimental conditions. As shown in Fig. lA, when these cells were not treated with cycloheximide and when inoculated into wells coated with type I collagen only (without exogenous FN), they
DOWNS
CELLS
AND
COLLAGEN
ADHESION
Down’s Early-pass.
Normal Early-pass.
Normal Late-pass.
FIG. 1. F-actin reorganization in normal and Down’s dermal fibroblasts with or without cycloheximide treatment on collagen substrata. Early-passage Down’s fibroblasts from the &month-old patient (A and B), early-passage PAPS cells from a normal individual (C and D), and late-passage PAPS cells (E and F) were evaluated for F-actin reorganization with rhodamine-phalloidin staining on collagen-only substrata as described under Materials and Methods. Some cells were not treated with cycloheximide (Cycle-) while other cells were treated with the inhibitor (Cycle+) as follows: for 18 h in stock cultures, then EGTA-detached and rinsed with PBS, and finally inoculated into collagen-coated wells containing adhesion medium and cycloheximide. Cells were fixed and stained at 4 h as described under Materials and Methods. Typical F-actin stress fibers are shown at the solid black arrows and star-shaped clusters of F-actin in some cells with the curved black arrows. Magnification, X270.
attached and spread effectively within the 4-h time period and reorganized extensive arrays of stress fibers (e.g., at the arrow). However, when these cells were pretreated overnight for 18 h in stock culture and then during the 4-h reattachment period on the collagen-only
substratum with cycloheximide, cell spreading continued to be excellent but the cytoplasm was devoid of microfilament stress fibers (Fig. 1B). The cytoplasm was now filled with star-shaped clusters of F-actin (curved arrow in Fig. 1B) that may be some intermediate com-
192
FLICKINGER
AND
CULP
FIG. 2. Defective stress fiber formation in two other Down’s fibroblast populations on collagen substrata. Early-passage Down’s fibroblasts from the B-year-old patient (A) and the 20-year-old patient (B) were treated with cycloheximide, inoculated into collagen-only wells containing adhesion medium and cycloheximide, and evaluated for F-actin reorganization with rhodamine-phalloidin staining as described under Materials and Methods and in the legend to Fig. 1. Primitive linear bundles of F-actin are shown with open arrows and star-shaped clusters of F-actin with black arrowheads. Magnification, X380.
plex in formation and/or dissolution of microfilament stress fibers. This continued to be the case at all population doubling levels of these cells over their in vitro lifespan, pertaining to >85% of the cells in the population. This was not the case under another treatment condition (data not shown)-when the cycloheximide treatment was applied only during the 4-h reattachment period, indicating that a particularly labile set of proteins did not require resynthesis during the reattachment period as a consequence of the EGTA-mediated subculturing. When the collagen substratum was postadsorbed
with human plasma FN and cells were treated with cycloheximide for 22 h as in Fig. lB, they persisted in their inability to generate stress fibers, indicating that a FN deficiency in the collagen matrix cannot explain this effect. This collagen-specific deficiency in stress fiber reorganization is similar to previous observations of normal dermal fibroblasts in aging populations [121. For example, early-passage PAP5 cells from a normal newborn infant challenged with a collagen-only substratum make excellent stress fibers under untreated (Fig. 1C) or cyclo-
DOWNS
CELLS
AND
COLLAGEN
ADHESION
193
FIG. 3. Stress fiber formation in Down’s fibroblasts on FN substrata. Early-passage fibroblasts from the 5-month-old Down’s patient were left untreated (A) or were treated with cycloheximide (B) for 18 h prior to EGTA-mediated detachment and then for 4 h during reattachment on substrata as described under Materials and Methods and the legend to Fig. 1; however, cells were plated on substrata coated with human plasma FN only. Rhodamine-phalloidin-stainable stress fibers are prevalent in both classes of cells (solid black arrows). Magnification, X380.
heximide treatment (Fig. 1D) conditions. In contrast, late-passage PAP5 cells entering their senescent period displayed the same discriminating behavior as the Down’s fibroblasts did-untreated cells made excellent stress fibers on a collagen-only substratum (arrow in Fig. 1E) while cycloheximide-treated cells spread well and were completely devoid of stress fibers with a cytoplasm containing star-shaped clusters of F-actin (curved arrow in Fig. 1F). The cell in Fig. 1F is representative of >85% of the cells in the entire population under these experimental conditions. The same was true for RETs cells from the same individual (data not shown), indicating
that both major cell classes of the human dermis display this effect [12]. These results indicate that dermal fibroblasts from the 5-month-old Down’s patient at all population doubling values behaved like aging skin fibroblasts from normal individuals. In order to test the generality of this effect, dermal fibroblasts from two other Down’s patients were examined, these patients had died at 5 years and 20 years of age (see Materials and Methods). As shown in Fig. 2A, early-passage cells from the 5-year-old Down’s patient displayed the same incapacity for stress fiber formation after cycloheximide treatment, as did early-passage cells
194
DOWNS
CELLS
AND
from the 20-year-old Down’s patient (Fig. 2B), with only a few primitive F-actin bundles in the cytoplasm (open arrows) and star-shaped clusters of F-actin prevalent throughout the cytoplasm (solid arrowheads). The untreated cells from these two individuals reorganized stress fibers effectively (data not shown). Therefore, this observation is not a peculiarity of the dermal fibroblasts from one particular patient but applies to dermal fibroblasts harvested from three different Down’s patients who died at very different ages. The substratum specificity of the effect was evaluated as well. A substratum coated with human plasma FN was tested for comparison with the results above on a collagen-only substratum. Untreated cells from the 5 month-old Down’s patient generated stress fibers as expected (e.g., at the arrow in Fig. 3A). Furthermore, cycloheximide-treated cells generated stress fibers effectively (e.g., at the arrow in Fig. 3B). The same was true for the Down’s fibroblasts from the other two patients (data not shown). Therefore, the cycloheximide-induced deficiency is collagen-specific (and unrelated to fibronectin receptor activity) and cannot be explained by a deficiency in a cytoskeletal component per se, since the same cycloheximide-treated cells do very well on a FN-only substratum but are incompetent on a collagen-only substratum. Furthermore, supplementing the collagen layer with exogenous FN does not overcome the deficiency (see above), indicating a dominant effect by collagenspecific receptors at the surface of these cells in these phenotypic changes. In normal aging dermal fibroblasts, the cycloheximide-induced deficiency in stress fiber formation could be reversed by several experimental conditions [ 121. Potential reversal of the inhibition was therefore tested with the Downs’ cells as well. Removal of the drug was evaluated initially. Cells were pretreated for 18 h with the drug, EGTA-detached, inoculated into collagen-only wells in medium containing the drug for 4 more h, and then rinsed well with drug-free medium for an additional 16 h incubation in the absence of drug before evaluating F-actin organization. As shown in Fig. 4, cells from the &month-old Down’s patient (Fig. 4A), the &year-old patient (Fig. 4B), and the 20-year-old (Fig. 4C) regained competence for forming stress fibers throughout their cytoplasm (arrows). It should also be noted that the density of stress fibers and their lengths increased considerably during these reversal conditions. Therefore, cycloheximide treatment does not lead to irreversible damage to Down’s fibroblasts such that they are no longer competent for reorganizing their microfilament bundles.
COLLAGEN
ADHESION
195
A second approach for reversal was also tested. Down’s cells were pregrown for two passages in medium supplemented with sodium ascorbate in order to stimulate collagen production/maturation processes [ 17,181. In previous experiments [ 121, this approach has successfully reversed cycloheximide inhibition in the aging normal fibroblasts. In this case with fibroblasts from the Down’s patients, two different results were obtained. With cells from the 20-year-old Down’s patient, ascorbate supplementation was incapable of replenishing the receptor moieties required for cells to generate stress fibers on collagen-only substrata under cycloheximide treatment (Figs. 5A and 5C). In contrast, cells from the 5-year-old after ascorbate supplementation became competent in forming stress fibers (Figs. 5B and 5D). These results may indicate any one of several possibilities-(a) there are two different molecular mechanisms of deficiency operating here; (b) Down’s cells are differentially susceptible to stimulation with ascorbate; (c) there are different stoichiometric levels of receptor activity being depleted in these two populations; or (d) Down’s cells display an aging-related competence for ascorbate reversal since cells from the oldest patient could not be reversed while those from the younger patient could be reversed. A third approach was tested based on recent evidence that fibroblasts treated with transforming growth factor-@ (TGFB) produce much more collagen, FN, and FNspecific integrin-an up-regulation that occurs at the transcriptional level of the respective genes [19-211. When Down’s cells were treated with TGF@ for 24 or 48 h prior to their receiving cycloheximide, there was no reversal of the cycloheximide-induced deficiency upon stress fiber formation (data not shown). This result indicates a mechanistic difference of this treatment with the ascorbate reversal described above and with the TGFP stimulation of FN and collagen production per se. Therefore, this difference may also be consistent with stoichiometric differences in collagen-specific receptor levels under these treatment conditions and not in the matrix ligands themselves. The inability of TGF/3 pretreatment of cells to reverse the deficiency also indicates that collagenolytic destruction of the collagen substratum, as shown recently for aging fibroblasts [ 311, may not be increased upon cycloheximide treatment. To further evaluate this possibility, Down’s cells were treated with cycloheximide by the usual 22-h protocol but their adhesion responses were evaluated on collagen-only substrata in medium that contained 50 pg/ml human plasma FN which could re-
FIG. 4. Cycloheximide removal and regeneration of stress fibers. Early-passage Down’s fibroblasta from the 5-month-old (A), the B-yearold (B), and the 20-year-old (C) patients were pretreated with cycloheximide for 18 h as usual, EGTA-detached and rinsed, inoculated into collagen-only wells for 4 h in adhesion medium containing cycloheximide, and then fed medium without cyclo/zeximide for 16 h when cells were fixed for rhodamine-phalloidin staining. Dense and extremely long stress fibers are observed throughout all cells (solid black arrows). Magnification, X365.
FIG. 5. Differential recovery from the cycloheximide-induced inhibition by ascorbate-supplemented cells. Early-passage fibroblasts from the 20year-old (A and C) and &year-old (B and D) Down’s patients were grown for two passages in complete medium supplemented with sodium ascorbate as described under Materials and Methods. Cells were then left untreated (A and B) or treated with cycloheximide (C and D) as described in the legend to Fig. 1 in medium containing ascorbate. After 4 h on collagen-only substrata in medium with cycloheximide, cells were fixed and stained with rhodamine-phalloidin. Extensive stress fibers are observed in three cases (A, B, and D) as pointed out with solid black arrows, while in case C the cytoplasm remained devoid of stress fibers and contained star-shaped clusters of F-actin (solid white arrow). Magnification, X410.
DOWN’S
CELLS
AND
place collagen lost from the substratum. However, stress fibers were still not observed. It is not likely that increased production of collagenase in cycloheximidetreated cells, a stimulation difficult to visualize because of the general shutdown of protein synthesis in drugtreated cells, can explain the adhesion deficiency. However, these experiments do not rule out the possibility for increased proteolytic turnover of collagen-specific receptors in treated cells for any one of a number of reasons. The excellent spreading of these dermal fibroblasts (either normal aging or Down’s populations upon cycloheximide treatment) in the complete absence of microfilament stress fiber formation is a remarkable finding and may indicate that the cell’s other cytoskeletal networks may be functioning at maximal capacity in these spreading responses. This was tested with antisera to tubulin for evaluation of microtubule networks and to vimentin for evaluation of intermediate filament networks. As shown in Fig. 6A (white arrowhead), untreated late-passage RET5 cells generated an extensive microtubule network throughout their cytoplasm. This was also true for the same cells under cycloheximide treatment (Fig. 6C; white arrowhead) when they were completely devoid of F-actin stress fibers. The same case applied for vimentin-containing intermediate filaments (Figs. 6B and 6D; open arrows) and for papillary cells from the same individual (data not shown). Therefore, while aging normal fibroblasts become incompetent for forming stress fibers, they remain competent for forming two other cytoskeletd networks that may not require cell surface receptor mediation [S]. These data provide further support for depletion of a collagen-specific receptor at the cell surface upon cycloheximide treatment as being responsible for the lack of stress fiber formation. Down’s fibroblast networks were also evaluated for microtubule and intermediate filament networks. As shown in Figs. 7A and 7B for the 20-year-old patient, microtubule networks formed normally in both cell classes (white arrowheads). The same was true for the other two Down’s patients’ cells and for intermediate filament networks (data not shown). Therefore, fibroblasts from Down’s patients behave identically to fibroblasts from normal patients undergoing in vitro aging in many regards. DISCUSSION
These results indicate that dermal fibroblasts from Down’s patients share a deficiency, inducible with cycloheximide treatment, in collagen-specific adhesion responses with aging dermal fibroblasts from normal individuals, an alteration that is not observed in early-passage normal cell populations [12]. Furthermore, this deficiency applies to Down’s cell populations from three
COLLAGEN
ADHESION
197
individuals who died at very different ages. This evidence supports the argument that Down’s patients experience premature aging processes in skin, as shown by the studies here, and in some other tissues as well [16, 22-25, 301. It also implicates a genetic basis for this alteration in the Down’s populations and contrasts with the aging-related changes at the phenotypic level observed in dermal fibroblast populations from normal individuals. The adhesion deficiency is specific for F-actin stress fiber formation in cells that continue to spread very effectively, displaying normal arrays of microtubule and intermediate filament networks. This deficiency in stress fiber formation cannot be explained by limiting amounts of cellular FN or collagen synthesized by cells, since cycloheximide did not alter stress fiber formation on human plasma FN-only substrata. Also, treatment of cells with TGFj3, known to stimulate collagen and cellular FN production at the transcriptional level [19-211, could not reverse the effect. Additional evidence in this study points to a stoichiometric limitation in collagen-specific receptors during the 22-h cycloheximide treatment in Down’s fibroblasts as the molecular basis for the deficiency. There are no limitations on the amounts of cytoskeletal proteins required for stress fiber formation per se in Down’s cells since fibers form perfectly well on FN-only substrata during drug treatment. In support of this latter argument, late-passage human lung fibroblasts have increased amounts of F-actin in their cytoskeleton when compared with early-passage cells [36]. Cell-synthesized collagen and FN do not appear to be limiting since plasma FN supplementation of the collagen substratum or the medium failed to prevent the deficiency and TGF/3 was without effect. However, countering this argument is the reversal of the inhibition in some, but not all Down’s cells, by ascorbate supplementation of cells prior to and during the cycloheximide treatment. Ascorbate is known to stimulate collagen production in human derma1 fibroblasts [ 10,17,18,41] and in parallel may stimulate production of collagen-specific receptors, although this possibility has not been tested to date. In this regard, fibroblasts from dermal keloid tumors of certain individuals have been shown to up-regulate synthesis of both cellular FN and the FN integrin receptor in a coordinate fashion [ 341. Ascorbate treatment of human foreskin fibroblasts has been shown in two-dimensional cultures to generate highly aligned collagen fibrils which align in register with cell surface attachment regions [41]. Also, limitation of collagen-specific receptors in aging and in Down’s cell populations may have consequences on the abilities of these cells to contract a threedimensional collagen matrix [36-401, for example the matrix surrounding these cells in situ. Another possibility for the deficiency in stress fiber formation is the destruction of the collagen substratum
198
FLICKINGER
.
E ? .I
AND
CULP
DOWN’S
CELLS
AND
COLLAGEN
ADHESION
199
FIG. 7. Microtubule networks in cycloheximide-treated Down’s fibroblasts. Early-passage Down’s fibroblasts from the 20-year-old patient were treated as described in the legend to Fig. 6 without (A) or with (B) cycloheximide. After 4 h on collagen-only substrata in media with cycloheximide, cells were fixed and stained with antisera to tubulin as described under Materials and Methods. Microtubule networks are shown throughout the cytoplasm (white arrowheads) in both cases. Magnification, X380.
by collagenase secreted from aging cells as described recently [31]. However, several experiments run counter to this argument. First, collagenase production should be down-regulated, and not up-regulated, by cycloheximide treatment of these cells. If collagenase was generat-
ing an incompetent substratum, this effect should have also been observed in the noncycloheximide-treated cells. This was never observed in this and previous studies [X2]. Second, TGF/3 treatment, known to stimulate endogenous collagen production, could not reverse the
FIG. 6. Microtubule and intermediate filament networks in cycloheximide-treated normal normal RET5 population were left untreated (A and B) or were treated with cycloheximide (C under Materials and Methods. After 4 h on collagen-only substrata in adhesion medium with or antisera to tubulin (Anti-Tub.) or antisera to vimentin (Anti-Vim.) as described under Materials C are evident throughout the cytoplasm (white arrowheads). Similarly, extensive intermediate cytoplasm in B and D (open arrows). Magnification, X410.
fibroblasts. Late-passage fibroblasts from the and D) as described in the legend to Fig. 1 and without drug, cells were fixed and stained with and Methods. Microtubule networks in A and filament networks are evident throughout the
200
FLICKINGER
effect. Third, supplementation of the medium of cycloheximide-treated cells with high concentrations of plasma FN (in the absence of any bovine albumin) would be expected to overcome this deficiency; this was not observed. Fourth, anti-collagen I staining of the substratum throughout these experiments revealed a continuous collagen matrix which maintained its integrity, inconsistent with this possibility. In further consideration of this possibility, transcription of both stromelysin and collagenase genes in many cell types is coordinately upregulated [42] and in some cases the regulation is mediated by FN receptor activity [43]. In the studies reported here, collagen receptor activity and not FN receptor activity mediated the cycloheximide-induced deficiency in stress fiber formation, providing further evidence for a mechanism independent of collagenase action. Detailed molecular characterization of the collagen-specific receptors of these cell populations under these treatment conditions should ultimately resolve this issue. With regard to potential receptor activities of interest, two molecular classes require attention. The first is the integrin glycoprotein complex of receptors on the surface of cells that recognize many extracellular matrix molecules [4-7,141 and that organize into focal contacts in substratum-specific patterns [44]. This family in human fibroblasts is composed of several members, some of which are specific for the target matrix ligand and some of which have overlapping specificities [26,27,35]. Several members of the integrin family are found in various subpopulations of human keratinocytes as well [35] and some integrins are shared in common with dermal fibroblasts [26,27]. The a2/31 integrin appears to be collagen-specific [26,27,35,45,46] and may be the candidate receptor for the cycloheximide-induced deficiency observed in these and previous studies [ 121. The second class of potential receptors are the cell surface heparan sulfate proteoglycans which mediate adhesion of dermal fibroblasts to FN and possibly collagen matrices [lo-12, 151 and of epithelial cells to collagen matrices [28, 291. Analysis of the stoichiometric amounts, synthesis, and turnover properties of these receptor classes can now be initiated by several experimental approaches. The results of this study on molecular deficiencies in a matrix adhesion response can be compared to other studies of phenotypic changes in cells from various tissues of Down’s patients. The trisomy of chromosome 21 would be expected to increase the dosage of gene products coded on this chromosome in such cells. This was one explanation for the increased intercellular adhesion and aggregation (divalent cation-dependent) of lung or cardiac fibroblasts from Down’s patients when compared to those of normal controls [32]. However, in these same studies dermal fibroblasts demonstrated decreased adhesion, consistent with a tissue-specific reduction in some receptor activity as suggested from our own studies [32]. Lymphoblastoid cell lines generated from circulat-
AND
CULP
ing cells of Down’s patients and immortalized with Epstein-Barr virus infection demonstrated increased aggregation, relative to normal patient control cells [33]. This aggregation was dependent upon elevated levels of one of the integrin family members in these cells [33]. Therefore, the trisomy 21 condition of Down’s syndrome may lead to differing phenotypic changes in various cell types that cannot be explained by simple gene dosage effects. The authors acknowledge support for these studies from National Institutes of Health research Grant AGO2921 and the technical assistance of Victor Guinto with the photomicrography.
REFERENCES 1.
Uitto,
2.
Smith,
3. 4. 5. 6. 7. 8.
J. (1986)
Dermutol. Clin. 4,433.
L. T., Holbrook, Dermatol. 105,57.
K. A., and Byers,
P. H. (1982)
J. Invest.
Annu. Rev. Biochem. 62,761. Biophys. A 907,261. Ruoslahti, E. (1988) Annu. Rev. Biochem. 57,375. Buck, C., and Horwitz, A. F. (1987) Anna Rev. Cell Biol. 3,179. McDonald, J. A. (1988) Annu. Rev. Cell Biol. 4,183. Yamada,
K. M. (1983)
Juliano,
Burridge, (1988)
R. L. (1987)
Biochim.
K., Fath,
K., Kelly,
T., Nuckolls,
G., and Turner,
C.
Annu. Rev. Cell Biol. 4,487.
9.
Harper,
10.
Beyth,
R. A., and Grove,
11.
Hall, M. D., Flickinger, L. A. (1988) Erp. Cell
R. J., and Culp,
G. (1979)
Science
204,526.
Exp. Cell Res. 155,537.
L. A. (1984)
K. S., Cutolo,
M.,
Zardi,
L., and Culp,
Res. 179,115.
Flickinger,
K. S., and Culp,
13.
Ruoslahti,
E., and Pierschbacher,
14.
Hynes,
15.
Culp, L. A., Laterra, J., Lark, M. W., Beyth, R. J., and Tobey, S. L. (1986) in Functions of the Proteoglycans (Evered, D., and Whelan, J., Eds.), CIBA Symposium No. 124, pp. 158-183, Wiley, London.
16.
Brown, W. T. (1989) Kligman, A. M., Eds.),
17.
Pinnell,
18.
Booth,
R. 0. (1987)
S. R. (1979) B. A., Polak,
L. A. (1990)
Exp. Cell Res. 186,158.
12.
M. D. (1987)
Science
238,491.
Cell 48,549.
in Aging pp. 93-120,
and the Skin (Balin, Raven Press, New
A. K., and York.
J. Invest. Dermntol. 73,485. J. (1980) Biochim. Biophys.
K. L., and Uitto,
A 607,145. 19.
Dean,
D. C., Newby,
S. (1988)
R. F., and Bourgeois,
J. CeU Biol.
106,2159. 20.
Roberts, C. J., Birkenmeier, S. K., Yamada, K. M., Chen,
21.
Keski-Oja, tlethwaite,
C&em.
J., Raghow, A. E., Kang, 263,311l.
22.
Uitto,
J., Fazio,
T. M., McQuillan, J. J., Akiyama, W.-T., and McDonald, J. A. (1988)
J. Biol. Chem. 263,4586. R., Sawdey, M., Loskutoff, D. J., PosA. H., and Moses, H. L. (1988) J. Bid.
M. J., and Olsen,
D. R. (1988)
J. Cutaneous Aging
Cosmetic Dermatol. 1,13. 23.
Imayama, 1019.
S., andBraverman,
J. Neurol. Sci. 79,91.
24.
Plioplys,
A. V. (1987)
25.
Plioplys, 233. Wayner,
A. V., and Bedford,
26.
Amer. J. Puthol. 134,
I. M. (1989)
E. A., and Carter,
H. M. (1989) W. G. (1987)
Brain Res. Bull. 22,
J. Cell Biol. 105,1873.
DOWN'S CELLS AND COLLAGEN ADHESION 27.
Wayner, E. A., Carter, T. J. (1988) J. Cell&X.
28.
Koda,
29.
LeBaron, R. G., Hook, A., Esko, (1989) J. Biol. Chem. 264,795O.
30.
Martin,
31.
Sottile, J., Mann, D. M., J. Cell. Physiol. 138,281.
32.
Wright, T. C., Orkin, R. W., Destrempes, M., (1984) Proc. Natl. Ad. Sci. USA 81,2426.
33.
Taylor, G. M., Haigh, Immunology 64,451.
34.
Babu, M., 9.1642.
35.
Carter, (1990)
36.
Gibson, J. M., Milam, Dev. 48,101.
Received
W. G., Piotrowicz, 10’7,188l.
J. E., and Bernfield,
G. M. (1977)
Diegelmann,
M. (1984)
Birth
Defects Diemer,
H., D’Souza,
J. Biol.
1,199O
Chem.
J. D., Gay,
Hook,
A. J. T. (1989)
and Kurnit,
N. (1989)
Mol.
D. M.
42. 43.
Cell. Bid.
44.
Kaur,
45.
Mech.
Age@
P.
Schafer, I. A., Shapiro, A., Kovach, M., Lang, C., and Fratianne, R. B. (1989) Exp. Cell Res. 8,259. Danowski, B. A., and Harris, A. K. (1988) Exp. Cell Res. 17'7,47. Montesano, R., and Orci, L. (1988) Proc. Natl. Acad. Sci. USA 85,4894. Nakagawa, S., Pawelek, P., and Grinnell, F. (1989) Elcp. Cell Res.
182,572. 41.
R. (1988)
T. S., and
R. J. (1989)
39. 40.
S. W., and Harris,
S. B., and Klebe,
38. M.
Ser. 14,5.
V., and Millis,
R., and Oliver,
37.
259,11763.
S., and
Orig. Artic.
W. G., Wayner, E. A., Bouchard, J. Cell Biol. 110,1387.
March
R. S., and Kunicki,
201
46.
Grinnell, F., Fukamizu, H., Pawelek, P., and Nakagawa, S. (1989) Exp. Cell Res. 181,483. Frisch, S. M., Clark, E. J., and Werb, Z. (1987) Proc. Natl. Acad. Sci. USA 84,260O. Werb, Z., Tremble, P. M., Behrendtsen, O., Crowley, E., and Damsky, C. H. (1989) J. Cell Biol. 109,877. Fath, K. R., Edgell, C.-J. S., and Burridge, K. (1989) J. Cell Sci. 92,67. Ogle, R. C., Potts, J., Yacoe, M., and Little, C. D. (1989) J. Cell Sci. 94,361. Gullberg, D., Terracio, L., Borg, T. K., and Rubin, K. (1989) J. Biol. Chem. 264,12,686.