Effect of Mechanical Forces on Extracellular Matrix Synthesis by Bovine Urethral Fibroblasts in Vitro

0022-534 7 /93/1502-0637$03 .00/0 THE JOURNAL OF UROLOGY Copyright © 1993 by AMERICAN UROL OGICAL ASSOCIATI O N , INC.

Vol. 1 50, 637-641 , August 1993

Printed in U. S. A .

EFFECT OF MECHANICAL FORCES ON EXTRACELLULAR MATRIX SYNTHESIS BY BOVINE URETHRAL FIBROBLASTS IN VITRO LAURENCE BASKIN,* PAMELA S. HOWARD AND EDWARD MACARAKt From the Connective Tissue Research Institute, University City Science Center and Division of Urology, Chiulren's Hospital of Phi/,adelphia, Phi/,adelphia, Pennsylvania

ABSTRACT

The role of mechanical forces in normal physiological processes is just beginning to be elucidated. Using a system developed in our laboratory, we can apply precise and reproducible mechanical deformations (biaxial strain) to cells. These deformations alter cell activities in a reproducible fashion and may mimic the physical environment found in portions of the urinary tract. At a low strain of 1.8% no change in the synthesis of types I and III collagen by urethral fibroblasts was found. However, at a high strain (4.9%) types I and III collagen showed a significant increase in synthesis compared to controls (type I, 1 .4 ± 0.25 µg. versus 0.9 ± 0.27 µg., p = 0.053; type III, 110 ± 7 ng. versus 88 ± 10 ng. , p = 0.036). In addition, fibronectin synthesis was increased at low and high strains when compared to controls (low strain 3.20 ± 1.03 µg. versus 1.46 ± 0.15 µg., p = 0.042; high strain 8.90 ± 1 .09 µg. versus 3.12 ± 0.69 µg. , p = 0.001). We have shown at the cellular level that mechanical force applied to fetal bovine urethral fibroblasts results in an increase in the amount of collagen synthesis and fibronectin synthesis. These findings suggest that alterations in the physical environment of cells found in the urethral wall can affect biochemical processes including those that govern the synthesis of structural macromolecules such as collagen. KEY WORDS: collagen, fibronectins, urethral stricture

Different tissues in the body and their component cells are subjected to tensional forces during normal function (such as forces experienced by vascular cells owing to the pumping action of the heart) and also as a result of pathological processes (wound healing, obstruction, tissue remodeling). The role of mechanical forces in normal physiological processes is just beginning to be elucidated. Urethral stricture disease is a common source of morbidity and frustration for the patient as well as the physician. 1 A urethral stricture can be defined as an abnormal constriction or loss of distensibility of the urethral lumen, 2 which results in pain and discomfort during micturition as well as a decrease in the urinary flow rate. If the urethral stricture progresses the end result will be an inability to void and urinary retention. Previously, we have shown an alteration in morphology as well as in the collagenous composition of urethral stricture tissue in comparison to the normal urethral spongiosum. 3 Urethral stric­ ture tissue consists of a relatively dense accumulation of types I and III collagen fibers with loss of the normal sinusoidal vascular network when compared to healthy urethral spon­ giosum. Quantitatively, urethral stricture tissue is characterized by an increase in the ratio of type 1:111 collagen in comparison to normal spongiosum. Histological analyses of the urethral stricture tissue suggest that the cell type responsible for the alteration in collagen synthesis is the urethral fibroblast. 3 Pre­ viously, we have isolated normal urethral fibroblasts in vitro using the fetal bovine as the cell source. 4 We characterized the extracellular matrix synthesis of these cells and have shown that urethral fibroblasts in vitro synthesize types I and III collagen and fibronectin. Clinical experience suggests that the strictured urethra has a pressure gradient across the lesion area. Since the cells within this region might be expected to be subjected to a different physical environment than they would under normal condi-

Supported in part by Public Health Service Grants HL34005 and DK45419. * Recipient of National Kidney Foundation Grant 1991-1992. t Requests for reprints: Division of Urology, Children's Hospital of Philadelphia, 34th St. and Civic Center Blvd., Philadelphia, Pennsyl­ vania 19104.

tions, we investigated the effects of mechanical deformation (stretch) on collagen and fibronectin synthesis by fetal bovine urethral fibroblasts. Using a system developed in our labora­ tory, we can apply precise and reproducible mechanical defor­ mations to cells. 5-7 MATERIALS AND METHODS

Cells. Fibroblast cell strains were developed from primary isolates of fetal bovine urethral spongiosum as previously de­ scribed. 4 The identification, proliferation characteristics and extracellular matrix synthesis of these fibroblasts have been previously reported in detail. 4 The fetal bovine urethras were obtained from mid to late gestational fetal calves (age deter­ mined by crown rump length). 8 Under sterile conditions, the calf skin overlying the urethra was removed. The urethra was then canalized with a 22 gauge needle, allowing the surrounding corpus cavernosum to be dissected from the urethral spon­ giosum. The spongiosum tissue was then minced and cultured in modified medium 199, supplemented with 10% newborn calf serum, 2.5 µg./ml. amphotericin B and 50 µg./ml. gentamicin, with additional modifications according to Lewis et al. 9 Cells were incubated at 37C in a humidified atmosphere of 5% carbon dioxide in air. In vitro collagen and fibronectin synthesis. To determine the in vitro secretion and accumulation of types I and III collagen and fibronectin in the absence of mechanical forces, equal numbers of urethral fibroblasts were seeded into 24 well plates. The cells were allowed to grow to a density of approximately 200,000 cells per well (confluence) and then the culture medium in each well was replaced with 1 ml. of serum-free medium. At periodic intervals during 160 hours the medium was removed and added to a protease inhibitor cocktail containing 4 mM. ethylenediaminetetraacetate, 0.9 mM. phenylmethyl sulfonyl fluoride and 10 mM. n-ethylmaleimide. The amounts of types I and III collagen and fibronectin in the fibroblast culture medium were then quantitated by enzyme linked immunosor­ bent assay (ELISA). ELISA. To quantitate the amount of types I and III collagen and fibronectin in culture medium, competitive ELISAs were

637

MECHANICAL FORCES ON EXTRACELLULAR MATRIX SYNTHESIS BY URETHRAL FIBROBLASTS

638

i FrG. 1. Chamber used to apply reproducible, quantifiable biaxial physical deformations (strain) to cells in vitro. Strain equals (l/lo)-1, where 1 equals stretched length and lo equals resting length.

A

B

2.0 1 .5

30

FIG. 3. Urethral fibroblast seeded at 300,000 cells per cm.2 attached to Tecoflex membrane after overnight incubation at 37C. Reduced from xlOO.

Fibronectin

20

ug/w�-� 10

0.5 0.0

Collagen Type Ill

· ······1'······ · �······ ····![ 0

50

1 00

Time ( h rs)

1 50

0

0

50

1 00

Time ( h rs)

1 50

FIG. 2. Quantitation of types I and III collagen (A ) and fibronectin (B) synthesized from urethral fibroblast (200,000 cells per well). Fibro­ nectin is predominant protein being secreted in vitro by urethral fibroblast. Its level is approximately 15 times that of type I collagen and 150 times that of type III collagen.

performed according to the method of Engvall.10 The ELISA for type I collagen was performed with rabbit anti-bovine type I collagen antibody at a dilution of 1:600. The experimental sample and antibody mixture were allowed to complex over­ night at 4C in bovine serum albumin coated wells. The complex was then transferred to the collagen coated wells (3.2 ng./µl. of bovine type I collagen) and incubated for 90 minutes. Anti­ rabbit alkaline phosphatase conjugate was added for 90 more minutes followed by phosphatase substrate for 30 minutes. All of these reactions occurred at room temperature. The reaction product optical density was measured at 405 nm. using a Biotek instrument connected to a computer equipped with application software that permitted storage of ELISA data and the gener­ ation of a standard curve by linear regression analysis. Concen­ trations of unknowns were calculated by reference to the stand­ ard curve. A standard curve was derived from known amounts of pure collagen I and III, which were dissolved in 0.5 M. acetic acid and diluted in phosphate buffered saline containing 0.5% Tween. The pH of the collagen standards was corrected to 7.2 before use. The ELISA for collagen type III was performed in a similar fashion using a rabbit anti-human type III collagen antibody at a 1:600 dilution. An ELISA for fibronectin was developed using an antibody that was made and characterized in our laboratory.11 This antibody was used at a 1:300 dilution. Results of the ELISA assays were expressed as mean plus or minus standard deviation in graph or bar form. A 2-tailed Student t test was used to evaluate statistical significance. Physical deformation (strain) experiments. To deform fibro­ blasts in vitro we used a system that was designed to apply precise and reproducible biaxial deformations [strain = (l/lo)1 where 1 equals stretched length and lo equals resting length] to a compliant membrane (fig. 1).6 Cells grown on this mem­ brane attach and are then subjected to deformation.5 -7 The membrane (Tecoflex) was previously stretched across the bot­ tom of a stainless steel cylinder (well) that was screwed into a base. The top of the well was covered by a cap that became the top of a small culture chamber.

For this series of experiments 2 different cell strains of bovine urethral fibroblasts were used, which had previously been sub­ cultivated between 6 and 10 times in cell culture. The fibro­ blasts were seeded onto the membrane at a density of about 150,000 to 300,000 cells per cm.2 and allowed to attach over­ night. The cells were gently washed with serum-free medium and 1 ml. of modified medium 199 without serum was added per well. The cells were then deformed (stretched) cyclically by injecting compressed air below the membrane. Using this ap­ paratus, the degree of deformation (strain), the rate of defor­ mation (peak strain rate) and the frequency of deformation (cycles per second) could be controlled. For controls, cells were cultured in the apparatus wells but not subjected to deforma­ tion. The fibroblasts were deformed at either 1.8 or 4.9% strain for 24 hours at a frequency of 1 Hz. Each experiment consisted of approximately 7 control wells and 7 wells that underwent mechanical deformation. At the completion of the experiments the serum-free medium containing newly synthesized proteins was removed and the protease inhibitor cocktail was added. The samples were then stored at 4C for analysis by ELISA of extracellular matrix protein synthesis. Cell number per well was estimated after removal by trypsinization and determining cell count in a particle counter. RESULTS

Types I and III collagen and fibronectin secretion. Fibroblast culture medium containing newly synthesized and secreted proteins was analyzed quantitatively by ELISA for types I and III collagen and fibronectin. Medium fractions were collected at sequential time points to determine the basal rate of secre­ tion and accumulation of these specific extracellular matrix proteins when cultured in a mechanically static environment (on typical tissue culture flasks). Type I collagen secretion increased by a slow but steady rate (fig. 2, A ). Type III collagen was detected at much lower concentrations compared to type I. In addition, no change was detected in the accumulation of type III collagen in the medium of these cultures (fig. 2, A ). In contrast, fibronectin levels increased linearly as a function of time (fig. 2, B). Furthermore, fibronectin was the predominant protein being secreted, since its concentration was approxi­ mately 15 times that of type I and 150 times that of type III collagen. Among different cell strains, the concentrations of these proteins that were synthesized remained approximately constant. Mechanical forces deformation (strain) experiments. Biocom­ patibility of the fibroblasts with the membrane was a necessary prerequisite for the initial studies. To investigate this aspect the cells were seeded on the membrane and allowed to attach

MECHANICAL FORCES O N EXTRACELLULAR MATRIX SYNTHESIS BY URETHRAL FIBROBLASTS

A

Collagen I (4.9% Strai n)

B

639

Col lagen I ( 1 .8%) Strai n 1 .2 �-----------,

P:0 .846

0.8

ug/we l l 0.6

ug/w e l l

0.4 0.2 0.0

C

Collagen I l l

(4.9% Stra in)

D

Collagen I l l ( 1 . 8% Stra i n ) 80

1 00

p:0.093

60

ng/we l l

75 n g / w e l l 40

50

20

25 0

FIG . 4 . Effect of mechanical forces on urethral fibroblast (300,000 cells per cm. 2 ) . Typ e I collagen synthesis at high (4.9% ) strain (A) and low (1.8% ) strain (B ) . Typ e III collagen synthesis at high (4.9 % ) strain ( C ) and low (1.8%) strain (D).

A

Fibronectin (4.9% Strain) 10 8

6 µg/w e l l

B

Fibronectin (1 .8% Strai n) 0

4

•

µg/we l l

2 0

0

FIG . 5. Effect of mechanical forces on urethral fibroblast (300,000 cells per cm. 2 ) . Fibronectin synthesis at high (4.9% ) strain (A) and low (1.8% ) strain (B ) .

overnight. Morphological analysis showed that the fibroblasts adhered well to the membrane in the absence of any specific coating agents, such as gelatin or fibronectin (fig. 3). Plating efficiency of the fibroblasts on the membrane after an overnight incubation showed that 88 ± 3% of the cells remained adherent. The effect of different strain levels on the adherence of the fibroblasts to the membrane was determined by measuring the number of cells found floating in the media after a 24-hour period of mechanical deformation. During this time approxi­ mately 8 ± 3% of the fibroblasts initially seeded lost attachment after experiencing 4.9% strain. The effect of strain levels on mitotic indexes of the fibroblasts was also determined for this time frame. During a 24-hour period the cell number increased approximately the same amount (10 ± 4 % ) for the controls and the experimental chambers. Experimental deformation did not seem to alter morphology (phase contrast light microscopy) or growth pattern of these cells (data not shown). To determine the effects of strain upon extracellular matrix synthesis we analyzed fibroblast cultures mechanically de­ formed for 24 hours at strain levels of 1.8% and 4.9% (fig. 4). At 1.8% strain no change in the synthesis of types I and III collagen was found. However, at a high strain (4.9%) there was

a significant increase in synthesis of types I and III collagens compared to nonstrained controls (collagen type I, 1.4 ± 0.25 µg. per experimental well versus 0.9 ± 0.27 µg. per well, p = 0.053; type III, 110 ± 7 ng. per well versus 88 ± 10 ng. per well, p = 0.036). Fibronectin synthesis was also analyzed at low and high strain levels. When fibroblasts were strained for 24 hours at low (1.8%) and high (4.9%) levels fibronectin synthesis was significantly increased compared to nonstrained controls (low strain 3.2 ± 1.03 versus 1.46 ± 0.15, p = 0.04; high strain 8.90 ± 1.06 versus 3.12 ± 0.69, p = 0.001, fig. 5). DISCUSSION

In this study we stretched fetal bovine urethral fibroblasts in vitro and determined the effect of different degrees of stretch or mechanical deformation on extracellular matrix synthesis by these cells. Our rationale for these experiments was based upon the fact that cells at the site of a urethral stricture may experience considerable mechanical deformation owing to the steep pressure decrease (pressure gradient) across this region. 1 2 For example, during normal voiding (without any obstruction) the pressure in the bladder is equal to the pressure in the posterior urethra which is equal to the pressure in the anterior urethra which, in turn, is equal to the pressure of the urine as it leaves the meatus (fig. 6). 2 When there is urinary tract obstruction, such as a urethral stricture or benign prostatic hypertrophy, the pressure during voiding is greater proximal than distal to the obstruction. A pressure gradient exists across the stricture. The extent of the pressure gradient (and likely the number of cells affected) is a function of severity of the obstruction (radius of the urethral stricture). The cells that will experience the greatest pressure gradient or stretch during each voiding cycle are likely to be at the site of the obstruction. Since no information is available on the response of these cells to cyclic biaxial mechanical deformation, we performed a series of experiments to examine this relationship. In the field of wound biology a critical factor in wound healing

640

MECHANICAL FORCES ON EXTRACELLULAR MATRIX SYNTHESIS BY URETHRAL FIBROBLASTS

Normal Voiding 10

During Voiding p, = P,

80 60 cm H 2 0

P,

40

30

20

15

0

0

mis

Uret hral Strict u re Voiding With Obstruction 10 80 60 cm H20

40

30

20

15

0

�-------'--,j

mis

0

FIG. 6 . Urodynamic diagram of normal and obstructed voiding. Note increased bladder pressure and decreased urinary flow rate in urethral stricture diagram. Also note increased pressure (Pl>P2) during voiding in obstructed state.

and subsequent scar formation is the surrounding tension the tissue experiences. 13• 14 In a classic abdominal incision the tissue heals under a relatively static condition with little scarring. In contrast, tissue that heals under tension tends to scar exten­ sively. In a urethral injury once a scar has formed the compli­ ance and, hence, the fluid dynamics along the urethra change. This result is reflected in a nonstatic condition in which a pressure gradient acts at the site of the urethral stricture. Unlike the abdominal scar, which soon (about 6 months) reaches a steady state, the cells within the scar in the urethra may be stimulated by a pressure gradient during each void. The steepness of the pressure gradient will depend on the extent to which the lumen has been narrowed. To begin to model a urethral injury (urethral scarring and subsequent progression to a clinically significant stricture) we made the assumption that a patient voids 8 times per day (likely greater in a patient with a urethral stricture). During a 20-year period this micturition rate would total 58,400 voids. For the purpose of this study, a second assumption was that the abnormal pressure increase and decrease during each uri­ nation are equivalent to 1 cycle of deformation at the cell level. Thus, cells subjected to 1 deformation or stretch every second for a 24-hour period would sustain 86,400 deformations, which is on the same order of magnitude. Our results show that for types I and III collagen a threshold strain must be reached before any significant difference in collagen synthesis occurs. In fact, at a low strain of 1.8% no change was seen in the synthesis of collagen types I and III. However, at the higher strains of 4.9% we were able to show that the synthesis of types I and III collagen was increased. For fibronectin, a glycoprotein known to be important in adhesion,

in matrix cell interactions and in wound healing, at low and high strains we were able to show an increase in synthesis when compared to controls. We can only speculate as to why fibronectin synthesis was increased at low and high experimental strains while types I and III collagen synthesis was increased only at a high strain. Fibronectin is known to form a framework or scaffold for cell reorganization at the site of a wound. 1 5 Fibronectin also has been shown previously to be increased in granulation tissue, hypertrophic scars, keloids, myocardial infarction and wounds. 16- 19 Further work has shown that the pattern of fibro­ nectin splicing during wound healing reverts back to an embry­ onic pattern. This suggests that alternative splicing (inclusion of the EIIIA and EIIIB regions) of the fibronectin gene may be used functionally during wound healing and tissue repair. 20 It may be that an increase in fibronectin synthesis sets the stage for an increase in collagen deposition. 2 1 Little is known about the complex regulatory mechanisms that govern collagen synthesis. Since collagen is a major con­ stituent of the urinary tract, its synthesis and turnover must be a tightly controlled process. Many different factors, such as inflammation and ischemia, undoubtedly affect collagen syn­ thesis. However, there is increasing evidence to suggest that the mechanical environment of a cell initiates biochemical changes within the cell. 5 • 7 Our studies support the idea that mechanical forces can alter cell behavior. We clearly have demonstrated a change in collagen synthesis as a result of mechanical deformation. In conclusion, we have shown that urethral fibroblasts sub­ jected to mechanical stretching respond by altering collagen and fibronectin synthesis. Presently, there is no good expla-

M E C HANICAL FORCES O N EXTRACELLULAR MATRIX SYNTHESIS BY URETHRAL FIBROBLASTS

nation for the progressive scarring observed in patients with stricture. We suggest that long-term and systemic mechanical perturbation of these cells may alter normal connective tissue synthetic rates. However, additional studies will be required to define the mechanical parameters that exist in vivo so that more exact in vitro studies can be performed. Nevertheless, our hypothesis is attractive, since it can explain the long-term buildup of urethral scar tissue observed clinically. If indeed, this hypothesis is true then we would expect that urethral stricture scar tissue would progress in a proximal fashion (toward the bladder) from the site of the initial injury. This proximal progression would be expected because the urethral fibroblasts that experience the greatest mechanical forces would be at the site of the greatest pressure gradient. This site would be the proximal edge of the urethral stricture scar tissue. Further validation of our hypothesis will require further investigation. Drs. John Duckett and Howard Snyder provided support and advice. REFERENCES

1 . Baskin, L. S . and McAninch, J. W.: Childhood urethral injuries: perspectives on outcome and treatment. J. Urol. , part 2, 1 4 7 : 275A, abstract 2 4 6 , 1992. 2. Kropp, K. A.: Strictures of the male urethra. In: Adult and Pediatric Urology, 2nd ed. Edited by J. Y. Gillenwater, J. T. Grayhack, S. S . Howards and J. W. Duckett. Chicago: Year Book Medical Publishers, Inc., pp. 1 297-1314, 199 1 . 3. Baskin, L. S . , Constantinescu, S . , Howard, P. S . , McAninch, J. W . , Ewalt, D . , Duckett, J. W . , Snyder, H. M. and Macarak, E. J . : Biochemical characterization and quantitation o f the collage­ nous components of urethral stricture tissue. J. Urol., part 2, 1 50: 642 , 1993. 4. Baskin, L. S., Macarak, E. J., Duckett, J. W., Snyder, H. S. and Howard, P. S.: Culture of urethral fibroblasts: cell morphology, proliferation and extracellular matrix synthesis. Unpublished data. 5. Gorfien, S. F., Winston, F. K., Thibault, L. E. and Macarak, E. J.: Effects of biaxial deformation on pulmonary artery endothelial cells. J. Cell. Physiol. , 1 3 9 : 492, 1989. 6. Winston, F . K., Macarak, E. J., Gorfien, S. F. and Thibault, L. E.: A system to reproduce and quantify the biochemical environment of the cell. J. Appl. Physiol., 6 7 : 397, 1989. 7. Gorfien, S . F., Howard, P . S., Myers, J. C. and Macarak, E. J.:

641

Cyclic biaxial strain of pulmonary artery endothelial cells causes an increase in cell layer-associated fibronectin. Amer. J. Resp. Cell Mol. Biol., 3 : 421, 1990. 8. Eckstein, P. and Kelly, W. A.: Implantation and development of the conceptus. In: Reproduction of Domestic Animals, 3rd ed. Edited by H. H. Cole and P. T. Cupps. New York: Academic Press, chapt. 12, p. 335, 1977. 9. Lewis, L. J., Hoak, J . C., Maca, R. D. and Fry, G. L.: Replication of human endothelial cells in culture. Science, 1 8 1 : 454, 1973. 10. Engvall, E.: Enzyme immunoassay ELISA and EMIT. In: Methods in Enzymology. lmmunochemical Techniques. Edited by H. Van Vunakis and J. J. Langone. New York: Academic Press, vol. 70, pp. 419-439, 1980. 1 1 . Macarak, E. J . and Howard, P . S.: Adhesion of endothelial cells to extracellular matrix proteins J. Cell. Physiol., 1 1 6: 76, 1983. 12. Barrett, D. M. and Wein, A. W.: Voiding dysfunction: diagnosis, classification and management. In: Adult and Pediatric Urology, 2nd ed. Edited by J. Y. Gillenwater, J. T. Grayhack, S. S. Howards and J. W. Duckett. Chicago: Year Book Medical Pub­ lishers, Inc., pp. 863-962, 199 1 . 13. Montandon, D., D'Andiran, G. and Gabbiani, G.: The mechanism of wound contraction and epithelialization. Clinical and experi­ mental studies. Clin. Plast. Surg., 4: 325, 1977. 14. Howes, E. L., Sooy, J. W. and Harvey, S. C.: Healing of wounds as determined by their tensile strength. J.A.M.A., 92: 42, 1929. 15. Kurkinen, M., Vaheri, A., Roberts, P . J. and Stenman, S.: Sequen­ tial appearance of fibronectin and collagen in experimental granulation tissue. Lab. Invest., 43: 4 7, 1980. 16. Grinnell, F., Billingham, R. E. and Burges, L.: Distribution of fibronectin during wound healing in vivo. J. Invest. Dermatol. , 76: 1 8 1 , 1981. 17. Kischer, C. W., Wagner, H. N., Jr. , Pindur, J., Holubec, H., Jones, M., Ulreich, J. B. and Scuderi, P.: Increased fibronectin produc­ tion by cell lines from hypertrophic scar and keloid. Conn. Tissue Res., 23: 279-288, 1989. 18. Grinnell, F.: Fibronectin and wound healing. J. Cell Biochem., 26: 107, 1984. 19. Knowlton, A. A., Connelly, C. M., Romo, G. M., Mamuya, W., Apstein, C . S. and Brecher, P . : Rapid expression of fibronectin in the rabbit heart after myocardial infarction with and without reperfusion. J. Clin. Invest. , 89: 1060, 1992. 20. Ffrench-Constant, C., Van de Water, L., Dvorak, H. F. and Hynes, R. 0.: Reappearance of an embryonic pattern of fibronectin splicing during wound healing in the adult rat. J. Cell. Biol., 109: 903, 1989. 2 1 . Hurme, T., Kalimo, H., Sandberg, M., Lehto, M. and Vuorio, E.: Localization of type I and III collagen and fibronectin production in injured gastrocnemius muscle. Lab. Invest. , 64: 76, 199 1 .