MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT: EFFECTS OF THE LOCAL TISSUE ENVIRONMENT

0022-5347/01/1654-1283/0 THE JOURNAL OF UROLOGY® Copyright © 2001 by AMERICAN UROLOGICAL ASSOCIATION, INC.®

Vol. 165, 1283–1288, April 2001 Printed in U.S.A.

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT: EFFECTS OF THE LOCAL TISSUE ENVIRONMENT LAURENCE BASKIN, MICHAEL DISANDRO, YINGWU LI, WENHUI LI, SIMON HAYWARD GERALD CUNHA

AND

From the Departments of Urology, Anatomy and Pediatrics, University of California, San Francisco, California

ABSTRACT

Purpose: We have previously shown that mesenchymal-epithelial interactions are necessary for the development of bladder smooth muscle. Specifically without fetal or adult urothelium embryonic rat bladder mesenchyma does not differentiate into smooth muscle. The mechanism responsible for this interaction is not known, although it is postulated that diffusable growth factors have a role. Our hypothesis is that diffusable factors within adult rat bladders influence smooth muscle differentiation. Materials and Methods: Chimeric bladders were created by surgically implanting 14-day embryonic rat bladder mesenchyma before smooth muscle differentiation into the detrusor space of adult syngeneic hosts to test whether the host urothelium would induce smooth muscle differentiation without being in direct contact with fetal bladder mesenchymal tissue. Subdetrusor pockets were created between the serosa and smooth muscle layer, between the smooth muscle layer and lamina propria, and between the lamina propria and urothelium in direct contact with urothelium. Controls consisted of intact 14-day embryonic rat bladders with the urothelium not removed, and 14-day embryonic bladder mesenchyma recombined with urothelium (direct contact) placed within the sub-detrusor space of the bladder and under the renal capsule. Results: Immunohistochemical staining with antibodies directed against smooth muscle ␣-actin and urothelium (cytokeratin 7) revealed smooth muscle differentiation in intact embryonic bladders and bladder mesenchyma plus urothelium recombinants in contrast to bladder mesenchyma alone, which had no ␣-actin staining (morphometric smooth muscle analysis p ⫽ 0). There was no ␣-actin staining in chimeric bladders even when bladder mesenchymal grafts were placed directly in contact with host urothelium. In addition, bladder mesenchyma plus urothelial recombinants within the host bladder had less ␣-actin staining than their counterparts placed under the renal capsule (p ⫽ 0.001). Conclusions: A diffusable factor most likely exists within adult rat bladders that inhibits smooth muscle differentiation. KEY WORDS: bladder; muscle, smooth, epithelium; rats, inbred 344; mesoderm

The detrusor muscle of the bladder is one of the thickest smooth muscle layers in the body. It is responsible for the main function of the bladder, namely the storage and evacuation of urine. In certain pathological conditions, such as benign prostatic hyperplasia, posterior urethral valves, spina bifida and spinal cord injury, detrusor muscle organization and function are profoundly altered. These alterations lead to abnormal bladder compliance and subsequent high intravesical pressure, which if left untreated result in renal damage. The molecular mechanisms responsible for this altered state of detrusor growth and development are not well understood, although changes in epithelial-mesenchymal interactions most likely have an important role.1 We know from previous studies that epithelial-mesenchymal interactions are necessary for normal bladder smooth muscle development. In fact, fetal rat undifferentiated bladder mesenchyma does not differentiate into smooth muscle in the absence of bladder epithelium. However, when bladder mesenchyma is recombined with bladder epithelium in vivo and in vitro, the differentiation of bladder mesenchyma again occurs.2 The

smooth muscle inducing signal is not unique to fetal bladder epithelium, but also a property of adult urothelium.2 The mechanism whereby bladder epithelium signals bladder mesenchyma is currently not understood, although 3 possibilities exist. There is direct contact of epithelial and mesenchymal cells across a perforated basal lamina, short-range (20 to 40 ␮m.) interactions of mesenchymal cells with epithelial basal lamellae or a response of mesenchymal cells to a product produced by epithelial cells and diffusing several hundred microns into the mesenchyma. In some organ systems, such as bone, direct cellto-cell contact is necessary,3, 4 while in other organ systems, such as the eye, cell-to-cell contact is not necessary and differentiation occurs by a diffusable factor.5 Although previous recombination experiments answered the question of whether bladder epithelium does or does not induce bladder mesenchyma differentiation, these experiments shed no light on the mechanisms involved for several reasons. In recombination experiments bladder epithelium was isolated from bladder mesenchyma using trypsin, which nonspecifically digests protein and effectively removes the basement membrane from the epithelium. In other organ systems the basement membrane has an important role in mesenchymal-epithelial interactions as a promoter and in-

Accepted for publication October 13, 2000. Supported by National Institutes of Health Grants K08 DK0239704, RO1 DK51397-02, RO1 DK57246-01 and DK51101. 1283

1284

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT

hibitor of cell-to-cell communication.6 – 8 Also, the nature of the experiment requires the bladder epithelium and mesenchyma to be in direct contact and, thus, one cannot test whether differentiation occurs without direct contact, that is whether a diffusable factor is involved. Furthermore, the inductive potential of stroma and epithelium may be different in completely differentiated adult bladders, that is adult bladder epithelium may act differently in situ than when isolated from the normal bladder milieu. Cell-to-cell interactions have been shown to be active during and after tissue maturation in several other tissues,9 and there is no reason to believe that this finding may not also be true in the bladder. To overcome these shortcomings we devised an in situ recombination experiment using intact adult host bladder epithelium that was not separated from its surrounding mesenchyma. Embryonic rat bladder mesenchyma at 14 days of gestation before smooth muscle differentiation was surgically implanted into intact adult rat host bladders in direct contact with native bladder epithelium or at a defined distance from bladder epithelium within the detrusor space. Smooth muscle differentiation was then assessed in implanted bladder mesenchyma by immunohistochemical staining with antibodies directed against smooth muscle ␣-actin, while urothelial localization was defined by the expression of cytokeratin 7. MATERIALS AND METHODS

Tissue recombination experiments (fig. 1). Bladders from fetal Fisher 344 inbred rats (Simonsen, Gilroy, California) at 14 days of gestation based on plug date timing as day 0 were surgically isolated. Bladders of this gestational age do not express smooth muscle differentiation markers. Bladders were separated into mesenchyma and urothelium by tryptic digestion (0.1% trypsin in CMF Hanks on ice for 45 minutes), followed by manual separation of urothelium from mesenchyma by microscopic dissection. Embryonic bladder mesenchyma from 1 pregnant donor, consisting of mesenchyma from 5 to 14 bladders, was combined into 2 equal sized collections and each tissue aggregation was placed on solidified agar medium. Three types of specimens were then created, including intact bladder isolated from fetuses at 14 days of gestation by surgical methods only, bladder mesenchyma only and bladder mesenchyma plus bladder epithelium recombinants with

embryonic, newborn or adult bladder epithelium. Tissue recombinants were prepared by placing sheets of isolated embryonic, newborn or adult bladder epithelium directly onto the bladder mesenchyma aggregations. All specimens were cultured overnight at 37C on the agar medium to allow tissue layers to become adherent. A small particle of carbon was then placed on each tissue recombination experiment to help with subsequent localization of the implanted tissue. One tissue aggregation was placed in the sub-detrusor space of an adult Fisher 344 inbred rat host (fig. 1), while the other was placed under the renal capsule of the same rat host as a control. In each adult rat bladder tissue aggregations were placed between the serosa and smooth muscle layer, between the smooth muscle layer and lamina propria or between the lamina propria and urothelium directly in contact with urothelium. After 14 days of in vivo growth bladder and kidney grafts were assessed by immunohistochemical study for the expression of smooth muscle ␣-actin (smooth muscle fascicle formation). Cytokeratin 7 staining was done to assess the amount of epithelium in the recombined grafts. The amount of epithelium was graded as small—a few epithelial cells, moderate—several well-defined clumps of cells and large— epithelium forming a well-defined cystic structure resembling a normal bladder. Each experiment was repeated at least 5 times per group. Morphometrics. A DC330 digital camera (DAGE-MTI, Inc. Michigan City, Michigan) was used to capture images of the recombination experiments. Adobe PhotoShop software with Image Processing Tool Kit (Adobe, San Jose, California) was applied to analyze preferentially the image pixel values of the ␣-actin smooth muscle positive stained areas. After calibrating the image with its magnification bar in ␮m. the selected smooth muscle stained areas in the image were counted as ␮m.2 pixel values. Data were tabulated as the mean amount of smooth muscle ␣-actin staining per type of recombination experiment plus or minus standard deviation. The Student t test was done to calculate statistical significance among experimental groups. Immunohistochemistry. Immunohistochemical testing of tissue sections was performed with smooth muscle ␣-actin (Sigma Chemical Co., St. Louis, Missouri) on paraffin embedded specimens. Cytokeratin 7 was used as an urothelial marker.10 Briefly, avidin-biotin-peroxidase and alkaline phosphatase staining was performed with Vectastain ABC kits (Vector Laboratories, Burlingame, California), followed by cobalt intensification. All immunohistochemical studies were controlled with nonimmune or pre-immune serum, or IgG at equivalent dilutions. Biotinylated antimouse IgG was obtained from Amersham International (Arlington Heights, Illinois). Purified mouse IgG was obtained from Zymed Corporation (South San Francisco, California). Cytokeratin 7 was a gift from Dr. E. B. Lane, University of Dundee.11 Double staining with ␣-actin and cytokeratin was done on the same histological section to localize simultaneously epithelial and smooth muscle cells.12 For the double staining technique urothelial cells were labeled with a pink secondary color and smooth muscle ␣-actin was labeled brown. RESULTS

FIG. 1. Algorithm of 14 gestational day rat bladders consisting of intact fetal bladder, bladder mesenchyma (Mes) only or mesenchyma plus epithelial (BLE) recombinants grafted into detrusor space of host adult rat bladders or under kidney capsule in controls.

On gross examination implanted bladders were visualized within adult host chimeric bladders as a black stained mass (fig. 2). Immunohistochemical staining with antibodies directed against smooth muscle ␣-actin revealed that intact embryonic bladders had excellent smooth muscle differentiation, whether grown in the detrusor space or under the renal capsule (fig. 3). Bladder mesenchyma plus urothelium recombinants also showed smooth muscle differentiation in the detrusor space and under the renal capsule but not to the same degree. Bladder epithelium plus mesenchyma under the kidney capsule stained positive for smooth muscle ␣-actin

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT

FIG. 2. Gross specimen of host chimeric rat bladder with mesenchyma plus epithelial recombinant within detrusor space (arrow). Note carbon particles used for subsequent microscopic localization.

regardless of age, and embryonic, newborn and adult bladder epithelium induced smooth muscle differentiation (fig. 4). In contrast, bladder mesenchyma differentiated into smooth muscle in the detrusor space to a lesser degree overall (fig. 5). In fact, several bladder epithelium plus mesenchyma recombinants within the detrusor space showed minimal amounts of smooth muscle differentiation (fig. 5, D). Similar control recombination experiments in the kidney capsule did not show decreased smooth muscle differentiation (fig. 4). The table shows quantitative morphometric data. Intact bladder did not have a statistically different amount of smooth muscle formation at the kidney or detrusor space host site (p ⫽ 0.256). This finding was also true when comparing the amount of smooth muscle formation at the intact bladder kidney and detrusor graft sites, and the bladder mesenchyma plus bladder epithelium recombination experiment at the kidney host site (p ⫽ 0.217). In contrast, when comparing the bladder mesenchyma plus bladder epithelium recombination experiment at the detrusor space host site compared to the kidney capsule host site, a statistically significant less amount of smooth muscle had formed at the detrusor site (p ⫽ 0.001). Bladder mesenchyma alone had no ␣-actin staining when placed in the detrusor space (fig. 6) or under the renal capsule (p ⫽ 0, fig. 7).2, 12 There was no evidence of ␣-actin staining or smooth muscle differentiation in the bladder mesenchyma even when mesenchyma was placed directly in contact with host urothelium (fig. 6). In all cases the location of the recombinant within the detrusor space (subepithelial, lamina propria or muscle layer) had no bearing on the amount of positive smooth muscle ␣-actin staining. DISCUSSION

Bladder smooth muscle differentiates from embryonic mesenchyma in an orderly fashion under the direction of mesenchymal-epithelial interactions.2, 10 Bladder smooth muscle as well as uterine, lung and gastrointestinal tract smooth muscle does not develop in the absence of epithelium.13–16 Therefore, a signal must be transmitted from the epithelium to the mesenchyma, which instructs the mesenchyma to differentiate into smooth muscle. Although we know that the signal exists, the nature of the signal remains elusive.

1285

In our study bladder mesenchyma remained undifferentiated when grown in the detrusor space of adult rat hosts. In fact, implanted bladder mesenchyma remained undifferentiated even when it was in close contact with host bladder epithelium and separated only by the basement membrane (see table and fig. 6). Conversely recombinant bladder epithelium plus mesenchyma under the kidney capsule resulted in smooth muscle differentiation (fig. 4), whereas bladder epithelium plus mesenchyma grown in the detrusor space differentiated into smooth muscle but to a lesser degree (fig. 5). Why did native adult bladder epithelium not induce bladder mesenchyma to differentiate and why did the bladder epithelium plus mesenchyma recombination experiment show less smooth muscle differentiation in the host bladder environment? There are several possible explanations. Our results indicate there may be an inhibitory factor within the subdetrusor space. When isolated bladder epithelium and mesenchyma were recombined in the detrusor space of adult rat hosts, the amount of positive smooth muscle ␣-actin staining was much decreased compared to that in controls under the kidney capsule. In fact, in several examples there was no ␣-actin staining of bladder mesenchyma in the detrusor space despite recombination with amounts of bladder epithelium that always resulted in positive ␣-actin staining under the kidney capsule. This finding would be expected if one considered that maintaining the differentiated cell state depends on positive and negative feedback of interacting groups of cells.9, 17 Cell signals may be different in developing than in adult bladders. Peptide growth factors, which are the most likely mediators of mesenchymal-epithelial interactions, are expressed at various levels in bladder tissue depending on the age of the developing bladder.1 For example, transforming growth factor-␤2 is expressed at a high level during fetal periods of smooth muscle differentiation (14 days of gestation) but after birth transforming growth factor-␤2 is 10-fold lower. Transforming growth factor-␤ is an inhibitor of epithelial cell proliferation18, 19 and at the same time a stimulator of the transformation of fibroblasts to myofibroblasts.20, 21 Thus, transforming growth factor-␤ enables bladder muscle to differentiate but at the expense of epithelial proliferation. This situation is reversed in adult bladders, in which epithelium proliferates at the expense of further muscle differentiation. Such inhibitory factors may act over a distance and, thus, any tissue within the detrusor space, including implanted tissue, may be affected. To our knowledge no known growth factors to date have been shown to inhibit specifically smooth muscle differentiation. However, there are many examples of growth factors that inhibit epithelial proliferation, such as transforming growth factor-␤, tumor necrosis factor, interleukin-6 and urogenital sinus growth inhibitory factor.22 Recently the Notch gene has been identified that is downregulated in epithelial cells by mandibular mesenchymal cells, implying a direct inhibitory signal from mesenchyma to epithelium.23 There is no reason to believe that the reverse situation or an inhibitory signal from epithelium to mesenchyma may not also exist. Another explanation for the native bladder epithelium inability to induce smooth muscle differentiation in bladder mesenchyma may be that the basement membrane of native bladder epithelium somehow interferes with the transmitted signal. This explanation is based on the fact that the basement membrane is removed when epithelium and mesenchyma are separated by proteolytic enzymes such as trypsin, although it remains intact in the native bladder. However, most evidence indicates that the basement membrane is an integral part of cell-to-cell communication and, in contrast to being a hindrance, it is actually required for cell-to-cell signaling in some tissues.6, 7, 24 It has been shown that mandibular epithelium enables the mandibular mesenchyma in embryonic chicks to differentiate

1286

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT

FIG. 3. Whole 14-day gestation intact rat bladder A, grown in detrusor space. B, under kidney capsule of adult rat host for 2 weeks. Note excellent smooth muscle and urothelial differentiation confirmed by positive smooth muscle ␣-actin (brown) and cytokeratin 7 (pink) double immunohistochemical staining in both grafts. Inset is high power view of native host bladder urothelium and smooth muscle.

FIG. 5. Embryonic bladder mesenchyma at 14 days of gestation recombined with embryonic, newborn and adult bladder epithelium (BLE ⫹ BLM), and grown within detrusor space of adult host bladders. Double immunostaining technique with smooth muscle ␣-actin (brown) and urothelium stained with cytokeratin 7 (pink). Note black carbon particles localizing graft within host bladder and evidence of surrounding smooth muscle (arrows) in each recombination graft. A, fetal (f) recombination. B, newborn (n) recombination. C, adult (a) recombination. D, in adult recombination amount of smooth muscle differentiation is consistent with smaller amount of epithelium.

FIG. 4. Embryonic bladder mesenchyma recombined with embryonic, newborn and adult bladder epithelium (BLE ⫹ BLM), and grown under kidney capsule. Double immunostaining technique with smooth muscle ␣-actin (brown) and urothelium stained with cytokeratin 7 (pink). Note extensive smooth muscle and urothelial differentiation in all recombination grafts. A, fetal (f) recombination at 14 days. B, adult (a) recombination. C to F, newborn (n) recombinations.

into osteoblasts because of activity that resides within its basement membrane. In in vitro recombination experiments with epithelium mandibular mesenchyma isolated using proteolytic enzymes did not form bone but when it is isolated using a chelating agent such as ethylenediaminetetraacetic acid, which preserves the basement membrane, bone formed.7, 25, 26 Scanning electron micrography has revealed processes on each side of the basement membrane that intimately attach the basement membrane to the mesenchyma and epithelium.7 In our study the probable explanation of smooth muscle differentiation of bladder epithelium plus mesenchyma recombinants, despite being trypsinized, is that bladder epithelium and mesenchyma most likely act in concert to restore the basement membrane during recombination. This resto-

FIG. 6. Embryonic mesenchyma implanted into sub detrusor space (note carbon particles). A and B, mesenchyme implanted in direct contact with host urothelium. C and D, mesenchyme implanted within smooth muscle bundle of host bladder (C) or within lamina propria of host bladder (D). Note that each experiment shows no evidence of smooth muscle differentiation based on double immunostaining with antibodies directed against smooth muscle ␣-actin (brown) and cytokeratin 7 (pink). Contrast to figure 5.

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT Amount of smooth muscle staining in each type of recombination experiment Graft Site

Recombination Experiment

Representative Fig.

Mean ␣-Actin Pos. Staining ⫾SD ␮m.2

3, A 3, B 4

35,564.3 ⫾ 5,747 27,369.6 ⫾ 4,587 47,894.9 ⫾ 15,282

5

7,032.4 ⫾ 1,555*

Detrusor space Kidney capsule

Intact control bladder Intact control bladder Bladder mesenchyma, epithelium Bladder detruBladder mesenchyma, sor epithelium Bladder mesenchyma only Kidney capsule Bladder mesenchyma only * Statistically significant.

6

0*

7

0*

FIG. 7. Embryonic bladder mesenchyma (BLM) at 14 days of gestation implanted under renal capsule. Smooth muscle ␣-actin immunostaining revealed no evidence of smooth staining in bladder mesenchyma. Note positive control staining in renal blood vessels.

ration has been shown to occur in co-culture experiments using dental tissue and the restoration is rapid (less than 24 hours).6, 27 Several growth factors, including transforming growth factor-␤, fibroblast growth factor and platelet derived growth factor, have been shown to bind to basement membranes and these factors may be involved in basement membrane restoration.23, 28 Thus, basement membrane is thought to have an important role in the induction of cellular behavior.24 Therefore, it is unlikely that the basement membrane of native bladder epithelium hinders the cell-to-cell contacts necessary for bladder mesenchyma differentiation. Furthermore, differences in host graft sites in respect to readily available nutrients and/or trophic factors unique to the detrusor space versus the kidney capsule must be considered. Careful analysis of the implanted intact embryonic bladder showed no difference in the morphology or amount of smooth muscle and epithelium (fig. 3). This finding implies that sites may sustain intact control grafts in the same fashion. A major difference in host sites was the lack of mature smooth muscle in kidneys compared to adult bladders, representing indirect support for the inhibitory hypothesis described. CONCLUSIONS

Overall our results imply a diffusable inhibitory factor within adult bladders that does not allow further smooth muscle differentiation. It is conceivable that during periods of prolonged bladder obstruction, such as those in benign prostatic hyperplasia or posterior urethral valves, the blad-

1287

der reverts to a more embryonic state in which these inhibitory factors no longer exist. The resulting effect would be uninhibited smooth muscle growth and/or proliferation, leading to significant bladder dysfunction. The hope is that further understanding of the mechanisms of mesenchymalepithelial interactions in the bladder may lead to the development of new therapeutic approaches to bladder dysfunction.

REFERENCES

1. Baskin, L. S., Sutherland, R. S., Thomson, A. A. et al: Growth factors and receptors in bladder development and obstruction. Lab Investigation, 75: 157, 1996 2. Baskin, L. S., Hayward, S., Young, P. et al: Role of mesenchymalepithelial interactions in bladder development. J Urol, 156: 1820, 1996 3. Hall, B. K.: Matrices control the differentiation of cartilage and bone. In: Matrices and Cell Differentiation. Edited by R. B. Hinchcliffe and J. R. Kemp. New York: AR. Liss, pp. 147–169, 1984 4. Van Exan, R. J. and Hall, B. K.: Epithelial induction of osteogenesis in embryonic chick mandibular mesenchyme studied by transfilter tissue recombinations. J Embryol Exp Morphol, 79: 225, 1984 5. Chan, K. Y. and Haschke, R. H.: Epithelial-stromal interactions: specific stimulation of corneal epithelial cell growth in vitro by a factor(s) from cultured stromal fibroblasts. Exp Cell Res, 26: 231, 1983 6. Thesleff, I., Lehtonen, E. and Saxen, L.: Basement membrane formation in transfilter tooth culture and its relation to odontoblast differentiation. Differentiation, 10: 71, 1978 7. Hall, B. K. and MacSween, M. C.: An SEM analysis of the epithelial-mesenchymal interface in the mandible of the embryonic chick. J Craniofac Genet Dev Biol, 4: 59, 1984 8. Nakamura, M.: Regional ultrastructural and cytochemical comparisons of the epithelial-mesenchymal interface during rat incisor development. J Craniofac Genet Dev Biol, 4: 329, 1984 9. Cunha, G. R., Bigsby, R. M., Cooke, P. S. et al: Stromal-epithelial interactions in adult organs. Cell Diff, 17: 137, 1985 10. Baskin, L. S., Hayward, S. W., Young, P. et al: Ontogeny of the rat bladder: smooth muscle and epithelial differentiation. Acta Anatomica, 155: 163, 1996 11. Lane, E. B.: Monoclonal antibodies provide specific intramolecular markers for the study of epithelial tonofilament organization. J Cell Biol, 92: 665, 1982 12. DiSandro, M., Li, Y., Baskin, L. et al: Mesenchymal-epithelial interactions in bladder smooth muscle development: epithelial specificity. J Urology, 160: 1040, 1998 13. Taderera, J. V.: Control of lung differentiation in vitro. Develop Biol, 16: 489, 1961 14. Kedinger, M., Simon-Assmann, P., Bouziges, F. et al: Smooth muscle actin expression during rat gut development and induction in fetal skin fibroblastic cells associated with intestinal embryonic epithelium. Differentiation, 43: 87, 1990 15. Cunha, G. R., Battle, E., Young, P. et al: Role of epithelialmesenchymal interactions in the differentiation and spatial organization of visceral smooth muscle. Epithelial Cell Biol, 1: 105, 1992 16. Cunha, G. R., Young, P., Brody, J. R.: Role of uterine epithelium in the development of myometrial smooth muscle cells. Biol Reprod, 40: 861, 1989 17. Cunha, G. R. and Hom, Y. H.: Role of mesenchymal-epithelial interactions in mammary gland development. J Mammary Gland Biology and Neoplasia, 1: 21, 1996 18. Tucker, R. F., Shipley, G. D., Moses, H. L. et al: Growth inhibitor from BSC-1 cells closely related to platelet type beta transforming growth factor. Science, 226: 705, 1984 19. Martikainen, P., Kyprianou, N., Isaacs, J. T.: Effect of transforming growth factor-beta 1 on proliferation and death of rat prostatic cells. Endocrinology, 127: 2963, 1990 20. Pelton, R. W., Saxena, B., Jones, M. et al: Immunohistochemical localization of TGF beta 1, TGF beta 2, and TGF beta 3 in the mouse embryo: expression patterns suggest multiple roles during embryonic development. J Cell Biol, 115: 1091, 1991

1288

MESENCHYMAL-EPITHELIAL INTERACTIONS IN BLADDER SMOOTH MUSCLE DEVELOPMENT

21. Zhou, L., Dey, C. R., Wert, S. E. et al: Arrested lung morphogenesis in transgenic mice bearing an SP-C-TGF-beta 1 chimeric gene. Dev Biol, 175: 227, 1996 22. Rowley, D. R.: Characterization of a fetal urogenital sinus mesenchymal cell line U4F: secretion of a negative growth regulatory activity. In Vitro Cell Dev Biol, 28A: 29, 1992 23. Mitsiadis, T. A., Lardelli, M., Lendahl, U. et al: Expression of Notch 1, 2 and 3 is regulated by epithelial-mesenchymal interactions and retinoic acid in the developing mouse tooth and associated with determination of ameloblast cell fate. J Cell Biol, 130: 407, 1995 24. Paulsson, M.: Basement membrane proteins: structure, assembly, and cellular interactions. Crit Rev Biochem Mol Biol, 27: 92, 1992

25. Tyler, M. S. and Koch, W. E.: In vitro development of palatal tissues from embryonic mice. III: interactions between palatal epithelium and heterotypic oral mesenchyme. J Emb Exp Morph, 38: 37, 1977 26. Hall, B. K.: Initiation of osteogenesis by mandibular mesenchyme. Arch Oral Biol, 23: 1157, 1978 27. Ruch, J. V., Lesot, H., Karcher-Djuricic, V. et al: Extracellular matrix-mediated interactions during odontogenesis. In: Matrices and Cell Differentiation. Edited by J. R. Kemp and R. B. Hinchcliffe. New York: AR Liss, pp. 103–114, 1984 28. Adams, J. C. and Watt, F. M.: Regulation of development and differentiation by the extracellular matrix. Development, 117: 1183, 1993