Binding of Soluble Fibronectin and Its Subsequent Incorporation into the Extracellular Matrix by Early and Late Passage Human Skin Fibroblasts*

The specific binding of soluble ‘251-labeled human plasma fibronectin (”‘I-HFN-P) to confluent cultures of early and late passage human skin fibroblasts was investigated. Previous studies of HFN-P bound to fibroblast cell layers indicated that HFN-P was present in the cultures in two separate pools, distinguishable on the basis of their solubility in 1% deoxycholate. Pool I contained deoxycholate-soluble fibronectin (cell-as-sociated), whereas Pool I1 contained deoxycholate-in- soluble fibronectin (matrix-associated). Time course studies indicated that HFN-P was initially incorporated into Pool I and then accumulated into Pool I1 (McKeown-Longo, P.J., and Mosher, D.F. (1983) J. Cell Biol. 97, 466-472). Examination of the kinetics of lZ5I-HFN-P binding to Pool I of early and late passage cultures revealed that both cultures required 2-4 h to approach steady-state conditions. Other kinetic studies showed that the rates of loss of 12’I-HFN-P from either Pool I or Pool I1 were similar for both cultures. However, the late passage cultures bound greater than twice as much fibronectin into Pool I, per cell, than the early passage cultures. This difference was not related to a difference in the level of endogenously produced fibronectins accumulating in the medium. Late passage cultures deoxycho-late-insoluble indicate that late passage exhibit and the they into

The specific binding of soluble '251-labeled human plasma fibronectin ("'I-HFN-P) to confluent cultures of early and late passage human skin fibroblasts was investigated. Previous studies of HFN-P bound to fibroblast cell layers indicated that HFN-P was present in the cultures in two separate pools, distinguishable on the basis of their solubility in 1% deoxycholate. Pool I contained deoxycholate-soluble fibronectin (cell-associated), whereas Pool I1 contained deoxycholate-insoluble fibronectin (matrix-associated). Time course studies indicated that HFN-P was initially incorporated into Pool I and then accumulated into Pool I1 Examination of the kinetics of lZ5I-HFN-P binding to Pool I of early and late passage cultures revealed that both cultures required 2-4 h to approach steady-state conditions. Other kinetic studies showed that the rates of loss of 12'I-HFN-P from either Pool I or Pool I1 were similar for both cultures. However, the late passage cultures bound greater than twice as much fibronectin into Pool I, per cell, than the early passage cultures. This difference was not related to a difference in the level of endogenously produced fibronectins accumulating in the medium. Late passage cultures incorporated "'I-HFN-P into the deoxycholate-insoluble Pool at an average rate 2.6 times greater than early passage cultures. The late passage cultures also chased a greater percent of their Pool I-bound fibronectin into Pool I1 and a lower percent into the chase medium. These results indicate that early and late passage cultures of human fibroblasts exhibit differences in the binding of soluble fibronectin and in the extent to which they incorporate soluble fibronectin into the extracellular matrix.
Fibronectin is a large multifunctional glycoprotein and is a major structural component of the extracellular matrix of mesenchymally derived cells. It functions in the processes of cell adhesion, migration, and differentiation (reviewed in * This work was supported by grants AG00697-09 from the Na- ll To whom all correspondence should be addressed.
Refs. 1-3). Peptide mapping and affinity chromatography experiments reveal the existence of protease-resistant domains which are believed to mediate interactions between the fibronectin molecule, the cell surface, and other components of the extracellular matrix (4 and reviewed in Ref.

5).
Recent studies indicate that fibronectin may interact with cells through two distinct regions of the molecule. One of these regions, termed the "cell attachment site," is located within the carboxyl-terminal half of the fibronectin molecule (6). This site binds to an integral membrane 140-kDa glycoprotein complex termed Integrin (7,8). In addition, the binding of soluble fibronectin to substrate-attached cells appears to involve the 70-kDa amino-terminal region of the molecule, which lacks the cell attachment site (9)(10)(11). It has been proposed that these two different modes of fibronectin-cell surface interaction are associated with unique cell surface molecules (9)(10)(11)(12).
Several reports have indicated that the association of fibronectin with the cell surface is altered in aged cultures (13-18). Indirect immunofluorescence studies have shown that the fibronectin-containing fibrils in the extracellular matrix of late passage human fibroblasts are thicker and less well organized than those found on early passage cells (19). Such observations suggest an age-related difference in the molecular machinery involved in the assembly of fibronectin into extracellular matrix fibrils.
To examine further the ability of early and late passage cells to organize fibronectin into the extracellular matrix, we have analyzed the interaction of soluble human plasma fibronectin with fibroblast cell layers. We report that late passage cultures accumulate more fibronectin into the extracellular matrix than early passage cultures. These results provide a quantitative comparison of the binding of soluble fibronectin and its subsequent incorporation into the extracellular matrix by early and late passage human diploid skin fibroblasts.

Fibronectin Binding to Early
to substratum-attached cell layers demonstrated that the fibronectin was retained in deoxycholate (1% deoxycholatesoluble (Pool I) and deoxycholate-insoluble (Pool 11) forms).
The kinetics of fibronectin incorporation into early and late passage cultures were compared and found to be similar, but not identical (Fig. 1). Binding into the Pool I fraction approached apparent steady-state conditions in 2-4 h. However, the total amount of fibronectin bound into Pool I was greater in the late passage cultures than in the early passage cultures. After correcting for nonspecific binding at 24 h, late passage cultures bound 2.4 times more fibronectin into Pool I than did early passage cells. Similarly, a t 24 h, the specific level of label in Pool I1 was 2.6 times greater for late passage cultures than for the early passage cultures.
In the absence of added unlabeled HFN-P, the late passage cultures bound more than twice the amount of labeled HFN-P as did the early passage cultures (Fig. 2 A ) . In both cell types, this label could be competed for by increasing the concentration of unlabeled HFN-P. These data were used to calculate saturation curves and to determine that saturation of Pool I binding sites was achieved somewhere between 25-75 pg/ml HFN-P (Fig. 2B). Half-maximal binding was obtained in the presence of approximately 10 pg/ml HFN-P.
These results suggested that late passage cell cultures should contain more cell layer fibronectin than early passage cell cultures. A dot-blot immunoassay was used to quantitate the levels of Pool I1 fibronectin in the cell layers. Pool I1 fibronectin was solubilized by treating the deoxycholate-insoluble pellet with buffer containing NaDodSO, and %mercaptoethanol. The solubilized materials were serially diluted, bound to nitrocellulose paper, and probed with antibodies to fibronectin. Table I shows that the amount of fibronectin extracted from late passage cells was 2.17-2.26 times greater than the amount extracted from the early passage cultures.
The level of fibronectin bound to isolated extracellular matrix was also examined in the two cultures. Table I1 shows the results of incubating iodinated fibronectin with early and late passage cultures in which the cells had been gently removed after lysis with deoxycholate. Approximately 26% of the total counts bound by early passage cell layers were bound to the isolated matrix material. For late passage cultures, the amount bound to isolated matrix was 13%. In both samples, the absolute number of counts bound was similar.
Role of Endogeneous Fibronectin-It was possible that the difference in radioactive HFN-P associated with the cell layers reflected a difference in the amount of endogenous The labeling medium was then removed, cell layers were washed, and radioactivity in Pool I was determined as described previously. The cell densities were 3.35 X lo5 cell/well for early passage cultures (H) and 3.14 X lo5 cell/well for late passage cultures (0). The competitive inhibition curves in A show that any Pool I-bound radioactivity at competitor concentrations greater than 150 pg/ml corresponds essentially to nonspecific binding. Therefore, radioactivity bound in the presence of 300 pg/ml unlabeled HFN-P was subtracted from all other values for each curve to generate the data shown in B. B shows the saturation curves for specific binding of fibronectin into Pool I.
The ordinate shows the total specific binding, expressed as ng/well, uersu the total concentration of HFN-P available (presented in pg/ ml). NaDodSO, and 5% 2-mercaptoethanol and boiled for 5 min. The extracts were serially diluted and transferred to nitrocellulose using a Bio-Rad "dot-blot" apparatus. Quantitation was achieved by comparison to known quantities of plasma fibronectin which had been similarly treated, diluted, and transferred to nitrocellulose. Immunoassays used either Y/3.116 (cross-reacts with human and bovine species fibronectins); or 0/4.41 as the primary antibody (specific for human fibronectins) (20 cellular fibronectins generated by the two cultures. To test that possibility, we measured the amount of fibronectin secreted into the culture medium by each culture and the binding of 9 -H F N -P t o cultures treated with cycloheximide. The amount of cellular fibronectin secreted into the medium by early and late passage cultures was measured during the 24-h labeling period (Fig. 3). A monoclonal antibody specific for human species fibronectins (0/4.41) (20) was used in an indirect enzyme-linked immunoabsorbent assay to quantitate the cellular fibronectin in the medium. Fig. 3 shows that the accumlation of endogenously produced cellular fibronectins was similar for both cultures. By 24 h, each conditioned medium contained approximately 0.1 pg of cellular fibronectin/10,000 cells, indicating that the difference in in- corporation of labeled fibronectin into the cell layers was not related to differences in the production of soluble cellular fibronectin present in the culture medium.

Fibronectin Binding to Early and Late Passage Fibroblasts
Cultures were treated with 20 pg/ml of cycloheximide and the level of '251-fibr~ne~tin binding measured at 0.5-24 h. Under those conditions, the late passage cultures bound 1.78 times more radioactivity into Pool I and 2.9 times more into Pool I1 than the early passage cultures (data not presented). These ratios are similar to those observed in the absence of cycloheximide (Fig. l), indicating that the difference between early and late passage cultures was not sensitive to cycloheximide.
Fate of Iodinated Fibronectin-Sensitivity to trypsinization was used to determine whether the incorporated fibronectin was retained at the cell surface or internalized. After a 6-h incubation with "'I-HFN-P, the cell layers were treated with trypsin (500 pg/ml) at 4 "C. Approximately 90% of the Pool I label and more than 97% of the Pool I1 label was trypsinsensitive (Table 111). Furthermore, extraction of the trypsintreated cell layers, with 1% deoxycholate after either a 2-or a 72-h incubation with '2'I-HFN-P revealed that approximately 10% of the total label incorporated into the cell layers was trypsin-resistant (data not shown). Both cultures retained most of the fibronectin in an enzyme-sensitive location. This finding is supported by recent electron microscopic localization studies which showed less than 1% of the exogenous fibronectin in intracellular locations (21).
The kinetics of the loss of "'I-HFN-P from Pool I was examined after a short pulse with radiolabeled fibronectin. The results, shown in Fig. 4, indicated that binding was reversible and that the kinetics of loss of label from Pool I were similar for both early and late passage cultures.
To follow the fate of the label leaving Pool I, the corresponding change in radioactivity in Pool I1 (Fig. 4) was assayed over a 5-h period after a short initial pulse labeling. The increase in Pool I1 radioactivity was greater in the late passage cultures as less Pool I radioactivity was lost to the culture medium. Fig. 5 shows the corresponding net increase in Pool 11-associated radioactivity during the 5-h chase period. Early passage cultures transferred approximately 10% of the label initially bound in Pool I into Pool 11, whereas late passage fibroblasts transferred approximately 30-35% of the fibronectin into Pool 11. Interestingly, skin fibroblasts cultured from donors suffering from diabetes, a condition associated with advancing age, in vivo, also showed increased transfer of fibronectin from a deoxycholate-soluble form to a deoxycholate-insoluble form (22,23).
The long-term turnover of labeled fibronectin incorporated into the deoxycholate-insoluble Pool was analyzed by measuring the kinetics of loss of label from Pool I1 over a period of 5 days. The loss of label from Pool I1 occurred with similar kinetics for both culture types (Fig. 6).

Fibronectin appears to interact
with cells through two separate sites on the polypeptide: a cell attachment sequence containing Arg-Gly-Asp located in the central portion of the molecule (7,8,(24)(25)(26) and a site located in the 70-kDa aminoterminal region of the molecule (9,27,28). Fibronectinderived peptides that lack the amino terminus but retain the sites for "cell attachment," collagen binding, and heparin binding fail to bind to attached cell layers (20). In the experiments presented here, soluble fibronectin was added to substrate-attached cell layers at concentrations too low to detect binding to the relatively low affinity fibronectin "RGD-receptor" (29-31). The levels of fibronectin bound to early and late passage cultures were compared. The evidence indicates that late passage cultures bind more fibronectin than early passage cultures. Furthermore, the late passage cultures transfer a higher percentage of the bound fibronectin from a deoxycholate-soluble Pool to a deoxycholate-insoluble form. The differences in the amounts bound do not reflect differences in soluble endogenous fibronectin, turnover from Pool I1 (Fig.  6), or proteolytic processing (data not presented). The stabilization of fibronectin into deoxycholate-insoluble fibrils is believed to result from increased disulfide bonding (27). Enhanced cross-linking of the extracellular matrix molecules collagen and elastin has been linked to in vivo aging (32).