Differentiation and vascularization of the metanephric kidney grafted on the chorioallantoic membrane

I)EVlCOPMENTAI,

BIOLOGY

96,

427-435 (1983)

Differentiation and Vascularization of the Metanephric Grafted on the Chorioallantoic Membrane HANNU SARIOLA,'~ETER

EKBLOM,EEROLEHTONEN,ANDLATJRI

Kidney

SAXON

The origin and development of mouse kidney vasculature were examined in chorioallantoic grafts of early kidney rudiments and of experimentally induced explants of separated metanephric mesenchymes. Whole kidney rudiments developed into advanced stages, expressed the segment-specific antigenic markers of tubules and the polyanionic coat of the glomeruli. In contrast to development in vitro, these grafts regularly showed glomeruli with an endothelial component and a basement membrane expressing type IV collagen and laminin. The glomerular endothelial cells in these grafts were shown to carry the nuclear structure of the host. This confirms the outside origin of these cells and the true hybrid nature of the glomeruli. When in vitro induced mesenchymes were grafted on chorioallantoic membranes, abundant vascular invasion was regularly found but properly vascularized glomeruli were exceptional. Uninduced, similarly grafted mesenchymal explants remained avascular as did the undifferentiated portions of partially induced mesenchgmal blastemas. It is concluded that the stimulation of the host cndothelial cells to invade into the differentiating mesenchyme requires the morphogenetic tissue interaction between the ureter bud and the mesenchgme. The induced metanephric cells presumably start to produce chemoattractants for endothelial cells at an early stage of differentiation. Kidney development thus seems to require an orderly, synchronized development of the three cell lineages: the branching ureter, the induced, tubule-forming mrsenchyme, and the invading endothelial cells of outside origin. INTRODUCTION

The early development of mammalian permanent kidney, the metanephros, involves a unique conversion of a mesenchyme into an epithelium. This conversion is initiated by an inductive tissue interaction between the nephrogenic mesenchyme and the ureter bud (Grobstein 1953,1955; Saxen et al., 1968). It has been recently demonstrated that the induced metanephric cells in vitro develop into advanced stages and express markers for each epithelial segment of the nephron (Ekblom et al., 1981a). On the other hand, the use of these markers also made it apparent that under these conditions, the glomeruli contained neither endothelial nor mesangial cells. This suggested that the endothelial cells of the kidney could be of extrinsic origin (Osathanondh and Potter, 1966; Ekblom, 1981b) rather than derivatives of the nephrogenic mesenchyme (Emura and Tanaka, 1972; Reeves et al., 1980). The origin of the glomerular endothelium has thus long been controversial (for reviews: Jokelainen, 1963; Potter, 1965; Kazimierzak, 1971; Ekblom, 1981a). Some investigators leave this question open, stating that glomerular capillaries grow from preexisting kidney vessels into the S-shaped body of the metanephric blastema (Huber, 1905; Kazimierzak, 19’71). To resolve this problem, early embryonic kidney rudiI To whom all correspondence should be addressed: Department Pathology, Haartmaninkatu 3, SF-00290 Helsinki 29, Finland.

of

ments with no evidence of vasculature within the blastema, were cultured on vascular areas of quail chorioallantoic membranes (CAM). The quail nuclear marker (Le Douarin, 1973) made it possible to demonstrate that the endothelial cells of the developing glomeruli were derived from the quail vasculature and not from the metanephric mesenchyme (Ekblom et al., 1982). The movement and proliferation of the capillary endothelial cells and their homing into the developing glomeruli suggest both a stimulation of the endothelium by the kidney explants and some guiding mechanism leading the vessels to their proper location in the kidney. To analyze these control mechanisms we have now grafted separated, uninduced as well as experimentally induced metanephric mesenchymes onto CAM of quail. We demonstrate that the uninduced mesenchymes remain avascular, but once induced to differentiate, they elicit a vascular response. The glomeruli which develop in these experimentally induced explants, however, mostly remain avascular. In contrast, both advanced tubules and well-developed glomeruli with a glomerular basement membrane form when whole kidneys are transplanted at the appropriate time of development. MATERIALS

AND

METHODS

Eleven- to fourteen-day whole murine (CBA X C57 Black) embryonic kidneys or mechanically separated ll-

428

DEVELOPMENTALBIOLOGY

day metanephric mesenchymes were grafted onto 7-day quail chorioallantoic membrane (CAM). Five-day quail (Coturnix coturnix juponica) kidney anlagen were similarly transplanted onto $--day chick (Gallusgallus) CAM. The transplantation was performed as follows: a window was cut into the egg, and the shell membrane was carefully scraped off. Then the graft was introduced by a pipet onto the CAM, near a branching large vessel, and the egg window was closed by a piece of Scotch tape. The eggs were incubated in humidified atmosphere at +37”C. Other 11-day mouse metanephric mesenthymes were induced by a piece of dorsal spinal cord (Grobstein, 1956) through a filter (Nuclepore, pore size 1.0 pm) for 24 or 48 hr prior to the grafting. This was done in Eagle’s minimum essential medium (MEM) supplemented with 10% horse serum and antibiotics. A total of 125 whole kidneys, 20 uninduced and 41 induced mesenchymes, were successfully grafted, incubated for 1 to 7 days on quail CAM and up to 10 days on chick CAM and analyzed histologically. This series does not include embryos that died during grafting (approximately one-third) and those in which the graft was lost (5% of the whole kidneys and approximately one-half of the isolated mesenchymes). The grafts were analyzed in light and electron microscopy and by immunohistochemistry. For histological studies, they were fixed in Carnoy solution, embedded in Tissue Prep, serially sectioned, and stained by the Feulgen procedure or with hematoxylin-eosin. For immunohistochemical analysis, the explants were fixed in cold alcohol (Sainte-Marie, 1962), embedded in Tissue Prep and cut into 7-pm sections. After deparaffinization the samples were incubated with antibodies against the brush border antigens of the proximal tubules (Ekblom et al., 1980b) or antibodies against Tamm-Horsfall glycoprotein (Ekblom et al., 1981b), a marker protein for distal tubules (Hoyer et al., 1974; Dawnay et al., 1980). They were then treated with fluorescein-isothiocyanate (FITC)-conjugated sheep anti-rabbit IgG, washed, and mounted in buffered glycerol. The details of the staining procedures have been described (Ekblom et al., 1980a,b). For lectin binding experiments, tetramethylrhodamine isothiocyanate-coupled wheat germ agglutinin (TRITCWGA) and fluorescein isothiocyanate-coupled peanut agglutinin (FITC-PNA) were used. After deparaffinization, the sections were incubated with TRITC-WGA or FITC-PNA. In some experiments, the deparaffinized sections were first treated with 0.01 U/ml of neuraminidase for 30 min in +37”C (from Vibrio cholerate, Behringwerke) (Ekblom et al., 1981b). For immunohistochemical analysis of extracellular matrix proteins, deparaffinized sections were reacted with antibodies to laminin, type IV collagen and type III procollagen (Timpl et al., 1979; Timpl, 1982) and their binding was analyzed

VOLUME 96,1983

by immunoperoxidase procedures. Briefly, deparaffinized sections were treated with HzOn, incubated with the antiserum, washed in PBS containing swine serum, and then incubated with swine anti-rabbit IgG serum. After washing, the preparations were treated with a rabbit anti-peroxidase anti-serum coupled to peroxidase (PAP, code No. Z 113, Dako, diluted 1:lOO in PBS) containing 2% swine serum. The chromogenic reaction was developed by incubation with 3,3’-diaminobenzidine-tetrahydrochloride and HzOB. Control stainings did not show brown staining (Ekblom et al., 1981a). For transmission electron microscopy the explants were fixed with 2.5% glutaraldehyde in 0.1 Mphosphate buffer, pH 7.4, for 3 hr at room temperature, postfixed with 1% Os04, in 0.1 M phosphate buffer, pH 7.4, for 1 hr at +4”C, and embedded in Epon 812. Thin sections were stained with uranyl acetate and lead citrate and examined in Jeol JEM-100CX transmission electron microscope. For scanning electron microscopy, the explants were fixed with 2.5% glutaraldehyde in phosphate buffer, pH 7.4, for 3 hr and embedded in paraffin. Sections were cut at 10 pm, deparaffinized with xylene, transferred to ethanol, and dried through CO2 in an Aminco critical point apparatus (American Instrument Co.). The samples were coated with a thin layer of gold in a Jeol JFC-1100 sputtering device and examined in a Jeol 35C scanning electron microscope (Department of Electron Microscopy, University of Helsinki). RESULTS

Development of Whole Kidneys on CAM When undifferentiated mouse kidneys were transplanted onto quail CAM and incubated up to 7 days, they showed all the main structures of the kidneys: glomeruli, tubules, collecting ducts, and abundant vasculature (Fig. 1A). The interstitial collagen was expressed in a fibrillar network in both the quail chorionallantoic tissue and in the interstitial tissue of kidney explants, as shown with antibodies to type III procollagen (Fig. 1B). Type IV collagen and laminin were expressed in the tubular and glomerular basement membrane (Fig. 1C) as well as in the endothelial basement membranes outside the glomeruli. However, antibodies against laminin reacted more intensely with the endothelial basement membrane than the antibodies against type IV collagen. In the grafts, tubular parts of the nephron expressed markers for proximal and distal segments after incubation of 5 to 7 days on CAM. This was demonstrated using antibodies to brush border antigens of the proximal tubules (Fig. 2A) and the antibody against TammHorsfall glycoprotein, a protein found in the distal tubules (Fig. ZB).

of Mouse

Kidmy

423

~~axulature

The glomeruli in the grafts showed two types of nuclei: mouse type in the podocytes and quail type in the endothelium (Fig. 3). The glomerular tuft formed normally (Fig. 4A), and its structure resembled the embryonic glomeruli with only a few podocytic folds. Staining of the glomerular podocytic surface was seen when sections were reacted with fluorescein-conjugated wheat germ agglutinin. When the sections were treated with neuraminidase and subsequently stained with fluorescein-conjugated peanut agglutinin, bright staining of the glomerular podocytic surface was seen (Fig. 4B). These observations indicate the presence of the glomerular polyanionic coat. Type III procollagen was seen in mesenchyme surrounding the developing glomerulus, but was not found within the glomerular tuft (Fig. 4C). Type IV collagen was expressed in the glomerular basement membrane, but only weakly in the mesangial area (Fig. 4D). The intensity was comparable to that seen for type IV collagen in the CAM and in the vessels outside the glomeruli. Laminin was found both in the glomerular basement membrane and in the mesangial area. The staining intensity for laminin was again similar to that of the vessel walls in the quail CAM. Older embryonic kidneys were similarly dissected and grafted onto quail CAM. They soon became necrotic in the center whereas the periphery remained alive. They showed only few glomeruli compared to the normal stage of the development in utero. Vnsculu rimt iov of Whole Kidneys

on CAM

When undifferentiated whole mouse (11-day-old) or quail (5-day-old) kidneys were grafted and incubated on quail or chick CAM, respectively, they were vascularized by chorioallantoic vessels in the vast majority of the explants (120 out of 125). The chorioallantoic endothelial cells invaded the lower crevice of the Sshaped body (Fig. 5A), which then formed the glomerulus of hybrid cells (Fig. 5B); the vascular part showed the host type of nuclear structure. Although this was the regular event, some exceptions could be seen. In four quail/chick and one mouse/quail interspecies grafts the kidney vasculature was partly of host and partly of graft type. When older, already vascularized mouse metanephric kidneys (from 12-day embryos) were cultured on quail CAM, they developed some immature glomerFIG. 1. Sections of a kidney graft after 7-day cultivation on quail chorioallantoic membrane. Ordinary light microscopy (A) illustrates advanced stages of nephron development and a rich vasculature. Antihodies against procollagen type III react exclusively with the interstitial nonepithelialized mesenchyme (BJ, while antihodies against collagen type IV react with the basement membrane of the developing nephrons and the vessel walls (C). Bound antibodies were detected with the unlabeled peroxidase-antiperoxidase technique. X70.

430

DEVELOPMENTAL BIOLOGY

FIG. 2. Immunofluorescence micrographs demonstrating the segmentation aCan chorioallantoic membrane for 7 days. Antibodies against the brush Horsfall glycoprotein (B) of the distal tubules were used. X400

ular cups and even an exceptional double glomerular tuft in the Bowman’s capsule. Thirteento fourteen-day mouse metanephric kidneys were also invaded by chorioallantoic vessels. While the central part of these older kidney grafts became necrotic, the cortical area was invaded by CAM vessels. Vuscularixation of Isolated Mesenchymes

on CAM

Separated, 11-day metanephric mesenchymes do not differentiate when cultured isolated in vitro. When such mesenchymal blastemas are cultured in contact with an inductor tissue, the mesenchyme converts into epithelial the kidney tubules. The inductor tissue can be removed before overt morphogenesis (Grobstein, 1967). In our transfilter cultures the induction is completed within et al., 1974; Lehtonen, 19’76; Sax&n 24 hr (Wartiovaara and Lehtonen, 1978). The induced and uninduced mesenchymes differ in extracellular matrix composition, although they are morphologically similar (Ekblom, 1981a). Here we compared the vascular response elicited by uninduced and induced mesenchymes. When 11-day, uninduced mouse metanephric mesenchymes were first separated and then grafted onto quail CAM, they remained as solid cell masses and contained only few, if any, vessels in the blastema. Yet, the mesenchymal cells survived; there were mitotic figures and very few pycnotic cells (Fig. 6A). In contrast, when mouse metanephric mesenchymes were induced i?z vitro by spinal cord prior to trans-

VOLUME 96, 1983

of the nephrons in an embryonic mouse kidney grafted on the border antigens of the proximal tubules (A) and for the Tamm-

plantation, they regularly vascularized during incubation of 1 to 7 days on quail CAM (Figs. 6B, C). All vessels in these grafts expressed the “quail marker” in the endothelial cells. The glomeruli remained mostly avascular. Based on the calculation of approximately 250 glomeruli in 41 explants, only l/10 of the glomeruli were vascularized, as shown by the expression of the “quail markers” inside the glomerular cup. One mesenchyme (out of 41) induced for 24 hr and subsequently incubated for 5 days on quail CAM also showed, however, endothelium of mouse type in the explant. The vascular response was dependent on the length prior to grafting. In those induced of in vitro induction for only 1 day a portion of the tissue seemed to remain as an undifferentiated blastema. In the mesenchymal areas only few vessels could be detected after incubation of 5 to 7 days on CAM. In contrast, no solid mesenchyma1 areas could be seen in 2-day-induced mesenchymal grafts, and vascularization was abundant in the transplants after similar incubation on CAM. Prolonged cultures up to 10 days could be performed with chick eggs, because they hatch 4 days later than quail eggs. When uninduced mesenchymes were grafted on chick CAM they were devoid of vessels which, again, were abundant in the experimentally induced ones. Here, like in the grafts on quail CAM, the majority of the glomeruli remained avascular. Since the mouse and chick cells cannot be easily distinguished from each other, no cytological analyses were performed in these experiments.

431

FIG. 3. Electron micrograph of the endothelial/podocyte interface of a hybrid quail/mouse glomerulus in a kidney graft on Day ‘7. The endothelial cell (E) shows nuclear features typical for quail, while the podocytes (P) are of mouse type. Basement membrane-like material is seen in the interface. Note the Weibel-Palade-like body and numerous pinocytic vesicles in the endothelial cell. X17,000.

DISCUSSION

Embryonic kidney rudiments grafted on the avian chorionallantoic membrane are known to develop well (Atterbury, 1923; Preminger et al., 1980). In the present study, we grafted embryonic kidney rudiments of different stages and followed this development by various criteria recently shown to characterize the kidney morphogenesis (Ekblom et al., 1981a,b). Examination of early kidney rudiments which were not yet vascularized when

grafted demonstrated that the vasculature in such experiments was derived from the host. Prior to grafting such undetermined and undifferentiated nephric blastemas expressed extracellular matrix proteins of an interstitial type. This pattern of the extracellular matrix rapidly changed into a basement membrane type during induction and early differentiation of the mesenchymal blastema. Thereafter the nephrons sequentially expressed their segment markers as during normal in V~VO development. Glomeruli were formed and, judged by

432

DEVELOPMENTAL

BIOLOGY

VOLUME

96, 1983

FIG. 4. Illustration of some features of hybrid mouse/quail glomeruli in kidneys grafted for ‘7 days. Scanning electron micrograph (A) demonstrating the developing podocytes. After neuraminidase treatment fluorescein-conjugated peanut agglutinin becomes attached to surface of the podocytes demonstrating their polyanionic surface (B). Immunoperoxidase stainings for type III procollagen (C) and type IV collagen (D) shows that only basement membrane collagen is found in the developing hybrid glomerulus. X600.

immunohistological and ultrastructural criteria, these showed an advanced stage of maturation and an endothelial component. Concerning the origin of the vasculature of the kidney and the endothelial glomerular loop, our results confirm earlier morphological observations (Osathanondh and Potter, 1966; Ekblom, 1981b) and direct experimental evidence which show that the vasculature of the metanephric kidney is derived from outside vessels (Ekblom et al., 1982) rather than from cells of the mesenchymal

blastema (Reeves et al., 1980). Our observations thus explain why otherwise well-differentiated glomeruli remain avascular during in vitro cultivation (Bernstein et al., 1981; Ekblom, 1981b) and why we have been unable to confirm an earlier report by Emura and Tanaka (1972) on a hematopoietic and endothelial bias of the metanephric mesenchyme (Ekblom, 1981b; Sariola et al., 1982). The exceptional cases where some of the vessels were of graft type might be explained by the variation of stages in grafted kidneys. Some “11-day kidney” ru-

SARIOLA

ET

AL.

FIG. 5. Two stages of vascularization of the glomeruli in a mouse kidney grafted on quail chorioallantoic membrane. At a S-shaped phase endothelial cells (E) carrying the quail marker invade the lower crevice of the body (A). A more advanced glomerulus (B) consists of abundant quail endothelial cells (E) and podocytes with mouse type nuclei (P). Feulgen stain. X400.

diments may have been slightly older and would thus already contain capillaries which then anastomose with chorionallantoic vessels. The 11-day metanephric blastema of the mouse is avascular when evaluated by morphological and immunohistological criteria (Ekblom, 1981b; Sariola et al., 1982). When grafted onto CAM, capillaries soon invade the mesenchyme and they seem to be directed towards th early condensates of the mesenchymal cells. When these aggregates become epithelialized and form the Sshaped anlage of the nephron, the endothelial cells are seen to invade the lower crevice of this body. Our immunohistological observations show, furthermore, that these invading host endothelial cells continuously deposit basement membrane components, laminin, and type IV collagen. Since they invade the glomerulus, they apparently also contribute to the formation of the glomerular basement membrane. Considering the forces stimulating and guiding the vascular development of the kidney anlage, two series of experiments reported here might become informative. We first demonstrated that both whole kidney anlagen and experimentally induced isolated mesenthymes became richly vascularized on CAM whereas uninduced mesenchymes grafted under identical conditions remained avascular. Similarly, in explants where induction was incomplete and only a portion of the mesenchyme was converted into epithelial structures, the undifferentiated part remained avascular while the ves-

sels were seeking the differentiated portion. Second, we have shown that while the formation of an endothelial glomerular loop was a rule in the grafted whole kidneys, it was a rare exception in the experimentally induced, similarly grafted mesenchymes, which otherwise showed good differentiation of the various epithelial components. The first set of observations suggests that vascularization of the nephric blastema was linked to its induction and early differentiation. Moreover, the invasion of the host capillaries was both temporally and spatially associated with the formation of pretubular mesenchymal aggregates. Hence it sounds feasible to search for angiogenic forces in these condensates and their metabolic products. Angiogenic factors have been demonstrated both in adult and neoplastic tissues (Folkman, 1974; Folkman and Cotran, 1974; Auerbach et al., 1976; Glaser et al., 1980) but their molecular nature and target sites are not yet understood (reviews: Auerbach, 1981; Gullino, 1981). Hence, one can only speculate upon such factors apparently produced by condensing mesenchymal cells after induction and recall the profound changes in their metabolism. This involves, among others, a rapid loss of the interstitial collagens (Ekblom et al., 1981a), and it has been suggested that split products of extracellular matrix proteins carry angiogenic activities (Gullino, 1981). The second set of our experiments demonstrated that the ultimate homing of the invading endothelial cells

434

DEVELOPMENTAL BIOLOGY

VOLUME 96, 1983

to the glomerular crevice was associated with the orderly development of the whole kidney rudiment and was confused in the experimentally induced mesenthymes. In both types of grafts, the mesenchyme was converted into epithelial elements of the glomerulus, as judged by lectin stainings demonstrating the glomerular anionic coat. Hence, the failure of the endothelial cells to invade the glomerulus might not be due to a lack of a proper epithelial “scaffold” but rather to an unsynchronized development of the epithelial and the endothelial cell populations. Finally, the possibility should be considered that the vascularization of the nephric blastema is a multistep process where the initial invasion and the directed migration of the endothelial cells are implemented through different mechanisms: First the induced mesenchyme starts producing angiogenetic factors leading to the vascular ingrowth. Subsequently the invading endothelial cells follow conductive pathways created by the orderly, gradual rearrangement of the molecular and cellular architecture of the induced mesenchyme. Such steps in angiogenesis have been previously emphasized (Gullino, 1981) but they remain to be explored in the vascularization of the metanephric kidney. Supported by grants from the Finska Lakaresallskapet, Finnish Cultural Foundation, and the Sigrid Juselius Foundation. We thank Drs. R. Timpl (Max-Planck Institut, Miinich, Federal Republic of German) and A. Miettinen (University of Helsinki, Helsinki, Finland) for their kind gifts of antibodies.

REFERENCES A~JERBACI~, R. (1981). Angiogenesis-inducing factors: A review. 1?l “Lymphokines” (E. Pick, ed.), Vol. 4, pp. 69-88. Academic Press, New York. AUERBACH, R., K~JBAI, L., and SIDKY, Y. (1976). Angiogenesis induction by tumors, embryonic tissues, and lymphocytes. Cancer Res. 34, 3435-3439. ATTERBURY, R. (1923). Development of the metanephric chick in allantoic grafts. Amer. J. Amt. 31, 409-431.

anlage

BERNSTEIN, J., CHENG, F., and ROSZKA, J. (1981). Glomerular entiation in metanephric culture. Lub. Zwuesf. 45, 183-190.

of the differ-

DAWNAY, A., MCLEAN, C., and CATTEL, W. R. (1980). The development

FIN;. 6. Sections of isolated induced and uninduccd metanephric mesenchyme (M) cultured on a quail chorioallantoic membrane for 7 days. (A) An uninduced mesenchymal blastema survives but remains undifferentiated, and no vascular invasion can be seen. (B) After a short inductive pulse ill. z&o nephric tubules (T) develop, but a portion of the mesenchyme still remains undifferentiated. Quail capillaries surrounding the graft invade the differentiated areas, but leave the undifferentiated mesenchyme (M) avascular. (C) Following an extended induction i?z &ro the entire graft shows tubular differentiation (T) with some glomerular bodies (G). The vasculature is abundant throughout the graft, but most of the glomerular bodies remain avascular. Feulgen stain. A and B, x200; C ~350

of radioimmunoassay B;och(~m.

for Tamm-Horsfall

glycoprotein

in serum.

J. 185, 679-687.

EKRLOM, P. (1981a). Determination and differentiation of the nephron. MNl. Bid. 59, 13%160. EKBI.OM, P. (1981h). Formation of basement membranes in the embryonic kidney. An immunohistological study. J. Cell Bid. 91, l-10. EKRLOM, P., ALITALO, K., VAHEKI, A., TIMPL, R., and SAX~~N,L. (1980a). Induction of a basement membrane glycoprotein in embryonic kidney: Possible role of laminin in morphogenesis. Proc. ik’ut. Acud. SC,. lLS.4 77, 485-489.

EKRLOM, P., MIETTINEN, A., and SAXON, L. (1980b). Induction of brush border antigens of proximal tubules in developing kidney. Dpc: Biol. 74, X-274.

EKULOM, P., LEHTONEN, E., SAXCN, L., and TIMPL, R. (1981a). Shift in collagen type as an early response to induction of the metanephric mesenchyme. J. Cell Bid. 89, 2’76-283. EKHLOM, P., MIETTIN~:N, A., VIRTANEN, I., WAHLSTKOM, T., DAWNAY, A., and SAXON, L. (1981b). In vitro segregation of the mctanephric nephron. @I,. Bid. 84, 88-95. E~IXOM, P., SAHIOLA, H., KARKINEN-J~BSKELRINEN, M., and SAXON, L. (1982). The origin of the glomerular cndothelium. Cell Diflerex 11, 35-39.

EMIX.~, M., and TAXAKA, T. (1972). Development of endothelial and rrythroid cells in mouse metanephrogenic mesenchyme cultured Lvith fetal liver. Up!,. Growth Difleren. 14, 237-246. FOLKMAN, J. (1974). Tumor angiogenesis. Ad~trn. C’u~ccr Res. 19, 33L 359.

735-753.

FOLKMAN, J., and COTRAN, R. (1976). Relation of vascular proliferation to tumor growth. Int. Rur~ E.tp Putho/. 16, 207-248. GL~SER, B. M., D’AMoRF,, P. A., SEPPK, H., SEPPA, S., and SCHIFFMANN, E. (1980). Adult tissues contain chemoattractants for vascular endothelial cells. Nuturc (Lomfvn) 288, 483-484. SROBSTEIN, C. (1953). Morphogenetic interaction between embryonic mouse tissues separated by a membrane filter. Nuture (Lo&n/) 172, 869-871. GROHSTEIN, C. (1955). Inductive interaction in the development of the mouse metanephros. J. EST). Zool. 130, 319-340. GROUSTEIX, C. (1956). Trans-filter induction of tubules in mouse metanephrogenic mesenchymc. &rp (%I1 Res. 10, 424-440. GRORSTEIP;,C. (1967). Mechanisms of organogenctic tissue interaction. Nut/.

JOKELAINEN, P. (1963). An electron microscope study of early development of the rat metanephric nephron. Actor Arm6 52, SuppI. 47, l-71. KXIMIERZAK, J. (1971). Development of the renal corpuscle and the juxtaglomerular apparatus: A light and electron microscopic study. Acttr Putho!. Microbid. Scrrntl. (A) (Suppl.), 218, 1-61. LX DOLIAKIN, N. (1973). A biological ccl1 labeling technique and its use in experimental embryology. UU~. Bid. 30, 217-222. LEIITONEN, E. (1976). Transmission of signals in embryonic induction. Md Bid. 54, 108-128. OSATHANONI)H, V., and POTTER, E. (1966). Development of human kidney as shown by microdissection. V. Development of vascular pattern of glomerulus. Arch. Pathol. 82, 403-411. POTTER, E. (1965). Development of the human glomerulus. Arch. Ptrthat. 80, 241-255. PREMINGER, G. M., KOCH, W. E., FRIED, F. A., and MANDELL, J. (1980). Utilization of the chick chorioallantoic membrane for in vitro growth of the embryonic murine kidney. Amer. J. Anut. 159, 17-24. REEVES, W. H., CAUI,FIF,LI), C. J. P., and FARQIJHAR, M. G. (1978). Differentiation of epithelial foot processes and filtration slits. Sequential appearance of occluding junctions, epithelial polyanion, and slit memhranes in developing glomeruli. Luh Ivwst. 39, 90100. REEVES, W. H., KANWAR, Y. P., and FAKQIJHAR, M. G. (1980). Assembly of the glomcrular filtration surface. Differentiation of anionic sites in glomerular capillaries of newborn rat kidney. J. Cell Bid. 85,

(‘rrrrwr

Inst. Mor~/gr.

26, 279-299.

GllLLINO, P. M. (1981). Angiogenesis factor(s). 1rr “Tissue Growth Factors” CR. Basrrga. ed.), pp. 427-449. Handbook of Experimental Pharmacology. Vol. 57, Springer-Verlag, New York/Berlin. IIOYF,K, J. R., RESNIU~, J. S., MI(XAEI,, A. D., and VERNIER, R. L. (1974). Ontogeny of Tamm-Horsfall urinary glycoprotein. Lab. Invest. 30, 757-761. HT,HER, G. G. (1905). On the development and shape of uriniferous tubules of certain of the higher mammals. Amer. .J. Aunt. Suppl. 4, 1-W

SAINTE-MARIE, G. (1962). A paraffin embedding technique for studies employing immunofluorescence. J. Histochym. C’ytochem. 10, 250256.

SARIOLA, H., EKBLOM, P., and SAXON, L. (1982). Restricted developmental options of the mctanephric mesenchyme. In “Embryonic Development, Part B: Cellular Aspects” (M. Burger and R. Weber, eds.), Progress in Clinical and Biological Research, Vol. 85b. pp. 425-431. Alan R. Liss, New York. SAS~N, L., KOSKIMIES, O., LAHTI, A., MIE:TTINF,N, H., RA~~OLA,J., and WARTIO~A-\K..\, J. (1968). Differentiation of kidney mesenchyme in an experimental model system. Adwn. Morphog 7, 251-293. SAX~~N,L., and LRHTONEN, E. (1978). Transfilter induction of kidney tubules as a function of the extent and duration of intercellular contacts. J. E~nh~& Ex~I. Morph 01. 47, 97- 108. TIMPL, R. (1982). Antibodies to collagens and procollagens. Advcuc. Enzymol. 82, 472-498, 1982. TIMPI,, R., ROIIDE, H., GEHRON ROBEY, P., RENNARI), S. I., FOIDART, J.-M., and MARTIX, G. R. (1979). Laminin-A glycoprotein from basement membranes. J. Biol. Chrm. 254, 9933-9937. WARTIOVAARA, J., NORDLING, S., LEHTONEN, E., and SAX~~N,L. (1974). Transfilter induction of kidney tubules: correlation with cytoplasmic penetration into Nuclepore filters. L Embryol. Ezp. Marphol. 31, 667-682.