Proc. Nati. Acad. Sci. USA
Vol. 77, No. 1,
p. 485-489, January 1980
Induction of a basement membrane glycoprotein in embryonic
kidney: Possible role of laminin in morphogenesis
(cell aggregation/epithelial cell differentiation/organ culture/immunofluorescence/tunicamycin)
PETER EKBLOM*, KARI ALITALOt, ANTTI VAHERIt, RUPERT TIMPLt, AND LAURI SAXEN*
*Department of Pathology and tDepartment of Virology, University of Helsinki, SF-00290 Helsinki 29, Finland; and tMax-Planck-Institut furBiochemie,
Martinsried bei Miunchen, D-8033 West Germany
Communicated by Clifford Grobstein, October 19, 1979
in the basement membranes of adult tissues, not in the mesen-
chymal stroma. We studied the appearance and distribution of
laminin during the early formation of kidney tubules in mouse
embryos and in an in vitro transfilter model system. In immu-
nofluorescence using affinity-purified antibodies, the distri-
bution of laminin showed a clear correlation, both spatially and
temporally, to the early stages of tubule formation. In vivo,
laminin was first detected in a punctate pattern in areas where
the pretubular aggregates form; later, it became confined to the
basement membranes of the tubules. In experiments in vitro,
the nephrogenic mesenchyme was found to form tubules after
12-24 hr of transfilter contact with the inductor. The first lam-
inin spots were found after 12 hr of culture, 24 hr before overt
morphogenesis. As the mesenchymal cells began to aggregate
and elongate (at 36 hr), laminin was detected in those cells
destined to become epithelial, and at 48 hr it was not found in
cells remaining in the stroma. In more mature tubules (at 72 hr),
laminin was seen as a sharp band in the basement membranes.
It is suggested that laminin is involved in the increased cell
adhesiveness during the early aggregation of the nephrogenic
The glycoprotein laminin is found exclusively
Inductive tissue interactions governing cell determination and
morphogenesis are mediated by hypothetical signal substances.
The target sites of the signals and the molecular response to
induction have not been clarified. In the embryonic kidney, a
relatively short interactive period triggers a whole sequence of
developmental events (1-3). These events can be followed and
timed by using the transfilter technique (4). The metanephric
mesenchyme is first separated from its natural inductor, the
ureter bud, and then combined with a heterotypic inductor
through a porous filter. This transfilter induction is completed
in 12-24 hr (5). Thereafter, the mesenchyme differentiates
autonomously, but there have been no means to distinguish the
determined cells from the uninduced ones prior to that. After
the morphogenetically silent period the mesenchymal cells are
soon aggregated into pretubular condensates. The trapping of
cells into aggregates and their subsequent behavior seen in
electron microscopy and time-lapse motion pictures suggest
increased adhesiveness of the cells as a response to the induction
(6). The molecular basis of the early aggregation phase, how-
ever, is still unknown. Kidney-specific antigenic and enzymatic
properties of the cells have been detected only considerably
later (3, 7).
The aggregation of the induced cells is followed by their
gradual conversion into epithelial elements surrounded by a
basement membrane. Thus, this model system provides a
unique opportunity for studying the synthesis and distribution
of basement membrane components during development.
Basement membranes have often been suggested to be involved
in morphogenesis (8-10), and they conceivably play a role in
the anchorage and positioning of the epithelial cells (11).
Basement membranes are known to consist of various collage-
nous and noncollagenous glycoproteins that have not yet been
thoroughly characterized (12), glycosaminoglycans (13), and
fibronectin (14). Recently, a large glycoprotein called "laminin"
was isolated from a mouse tumor producing a matrix of base-
ment membrane (15), and it has been characterized by chem-
ical and immunological methods (16, 17). This glycoprotein has
a molecular weight of about 850,000, is detected in all authentic
basal laminae studied, and is produced by various cultured cells
including epithelial cells (17). The expression of this glyco-
protein during kidney development will be described here.
MATERIALS AND METHODS
Kidneys. We made use of hybrid mouse embryos BALB/c
X CBA. The formation of aggregates in the kidney mesen-
chyme was examined in kidneys removed every 6 hr from day
11 to day 13 of gestation. The subsequent tubulogenetic phase
was followed in kidneys removed each day from day 13 to day
21 of gestation. Day 0 was the day of appearance of the vaginal
Experimental Model System. Mouse spinal cords, dissected
free from their basement membranes, of 11-day embryos were
used as heterologous inductors for lissue cultures (4). The un-
differentiated 11-day kidney mesenchyme was surgically
separated from the ureter bud and its basement membrane, and
one to five mesenchymes were placed on a filter (Nuclepore,
Pleasanton, CA). The nominal pore size of the filter used was
1.0 um. The tissues were cultured in Eagle's minimal essential
medium supplemented with 10% fetal calf serum and antibi-
otics. The Trowell-type culture technique was used as described
(18). In some of the experiments, tunicamycin at 50 ng/ml (19)
was added to the medium (a gift from R. Pratt, National In-
stitutes of Health, Bethesda, MD).
Fluorescence Microscopy. Laminin was purified from ex-
tracts of a tumor that produces an extracellular matrix of
basement membrane (16) and was used to raise antiserum in
rabbits. The antibodies were affinity purified and cross ab-
sorbed with type IV collagen. These antibodies did not cross-
react with fibronectin, basement membrane collagens, or in-
terstitial collagens in radioimrnunoassays. Their reaction with
authentic basement membrane could be blocked by absorption
with purified laminin. In the indirect immunofluorescence
assay the antibodies were used at a concentration of 15,g/ml.
Control experiments with antibodies(20,ug/ml) against oval-
bumin or fluoresceinated anti-rabbit immunoglobulin (Well-
come, Beckenham, UK) alone gave no staining.
The tissues were fixed and processed for immunofluorescence
according to the method of Sainte-Marie (20). Some tissue
sections were treated with hyaluronidase prior to staining (21)
with anti-laminin. After this treatment the fluorescence in-
The publication costs of this article were defrayed in part by page
charge payment. This article must therefore be her by marked "ad-
vertisement" in accordance with 18 U. S. C. §1734 solely to indicate
Developmental Biology: Ekblom et al.
tensity increased slightly, but there were no changes in the
staining patterns. For localization of nuclei, some sections were
stained for DNA (22).
We used a Zeiss Universal microscope with phase-contrast
and Nomarski differential interference contrast microscopy
optics together with an epi-illuminator. The microscope was
equipped with a high-pressure mercury lamp (HBO, 200 W)
and with a dichroic mirror (FT 510), a barrier filter (LP 520),
and excitation filters (BP 455-490) for fluorescein isothiocya-
nate-fluorescence or with a dichroic mirror (FT 420), a barrier
filter (LP 418), and an excitation filter (UG 1) for UV illumi-
The stages in the early morphogenesis of kidney tubules in the
metanephros are depicted in Fig. 1. By day 11 the ureter bud
has grown into loose mesenchyme, the mesenchymal cells ap-
pearing in a random pattern (Fig. 1A). After induction, dense
areas of mesenchymal cells become visible around the tip of the
ureter bud in 11.5- to 12-day kidneys (Fig. 1B). As the ureter
bud grows and branches, the condensed parts of the mesen-
chyme are left behind and they soon become S-shaped (Fig. 1
C and D). In 12.5- to 13-day kidneys, a lumen forms and a
peripheral basement membrane-like structure is observed (Fig.
1 D and E) and the tubule then connects with the branched
ureter (Fig. 1F).
In Vivo Findings. The undifferentiated nephrogenic mes-
enchyme of the 11-day embryo was devoid of laminin, as
judged by immunofluorescence. Fluorescence was observed
only as a discontinuous line in the basement membrane of the
ureter bud (Fig. 2a). In slightly older kidneys (11.5-12 days of
gestation), laminin fluorescence was regularly found distributed
as spots around the tips of the ureter at places known to ag-
gregate first. The layer with laminin spots was approximately
10 cells thick (Fig. 2b). The spots appeared prior to overt
aggregation. As the ureter bud was branching, spotty fluores-
cence was observed on both sides of the branching ureter
somewhat behind the tip (Fig. 2c).
In 12.5- to 13-day kidneys, laminin was found as spots in the
aggregates but also as discrete sharp staining in the basement
membrane of the newly formed S-shaped tubules (Fig. 2d).
There is a developmental gradient in older embryonic kid-
early pretubular condensate and S-shaped tubular Anlage. See text
for details. Drawings according to Rienhoff (23) and Saxon and
bryonic kidneys stained with anti-laminin antibodies and fluorescein
isothiocyanate-conjugated anti-immunoglobulin. (a) At 11 days (stage
A in Fig. 1). Only the ureter bud (U) shows laminin fluorescence. The
nephrogenic mesenchyme (M) does not stain. (X325.) (b) At 12 days
(stage B in Fig. 1). The first bright spots of laminin appear in the
mesenchyme. (X325.) (c) At 12.5 days (stage C of Fig. 1). The micro-
graph shows one branch of the uteric tree (U) and the surrounding
mesenchyme (M). Distinct spots of laminin are found on the lower
side of the branched ureter bud. Note that laminin staining is dis-
continuous in the basement membrane (BM) of the inducer (ureter
bud). Some staining is also seen on the inner side of the ureteric
basement membrane. (X325.) (d) At 13 days (stage D and E of Fig.
1). An early tubular Anlage that soon develops into the S-shaped body
is present. The area having an opened lumen (L) is shown to have
continuous basement membrane staining. In the other areas, laminin
is still found irregularly in punctate form. (X400.)
Immunofluorescence micrographs of 11- to 13-day em-
neys. More advanced tubules are found in deeper layers of the
metanephric blastema, and new aggregates appear in the pe-
riphery near the cortex. In deeper regions, laminin was detected
as long, thin strands in the basement membranes. In the pe-
riphery, punctate fluorescence still persisted within the young
aggregates and at places found also in morphologically undif-
ferentiated areas (Fig. 3 a and b). The luminal side of the tubule
did not display any staining. Spotty fluorescence was a regular
finding close to the forming basement membrane in the de-
veloping tubules. The mesenchymal cells and stroma sur-
rounding the tubules remained negative for laminin fluores-
cence. In older embryos (18-day to newborn), the spotty fluo-
rescence had disappeared, and laminin was always detected
in a linear continuous pattern associated with the basement
membrane (Fig. 3c).
In Vitro Findings. The experimental model system enabled
more exact examination of temporal relationships between the
induction phenomenon and the expression of laminin. The
inductor tissue did not show any fluorescence when stained with
laminin antibodies. The appearance of laminin was dependent
on induction; mesenchymes cultured alone did not express
Mesenchymes examined after 6-hr transfilter culture did not
show fluorescence (Fig. 4a). Small amounts of laminin were
detected as distinct spots in the mesenchymes examined after
12 hr (Fig. 4b). By 18-24 hr, bright spots were present (Fig. 4c)
in the mesenchyme, whereas sections studied after stainingfor
DNA showed no signs of aggregation (Fig. 4d). In the standard
procedure in which the inductor had been removed after 24
Proc. Natl. Acad. Sci. USA 77(1980)
Proc. Natl. Acad. Sci. USA 77 (1980)
areas (C) are found on the upper part of the micrographs (a and b). In the upper cortical areas of 14-day kidneys (a), laminin spots are seen in
the early aggregates but also in areas showing no morphological changes. The more mature tubules (T) show laminin distributed linearly in
the 15-day kidney (b), whereas one developing tubule in the upper cortex shows laminin in the two different patterns: linear in the lower area
and spotty on the upper edge of the growing tubule. In 21-day embryos (close to newborn), very bright fluorescence is found in the basement
membranes, but there are no spots (c). (X150.)
Immunofluorescence micrographs of 14-, 15-, and 21-day embryonic kidneys stained with laminin antibodies. The upper cortical
hr and the mesenchyme was subcultured alone, subsequent
development took place autonomously. Cell aggregation was
first detected at 36 hr (Fig. 4 f and h). All these aggregates
showed bright laminin staining in a spotty pattern (Fig. 4 e and
g). By 48 hr, the formation of tubuli was evident. In the tubuli,
laminin was found in two patterns, either within the basement
membranes or in their vicinity in the punctate form (Fig. 4 i
and j). A day later, the spotty fluorescence was lost, the base-
ment membrane alone expressing laminin (Fig. 4 k and 1).
Inhibition of Induction by Tunicamycin. Tunicamycin,
an inhibitor of protein glycosylation, has been shown to prevent
tubulogenesis when applied during the induction period (24).
Therefore, the effect of tunicamycin on laminin expression was
examined. When tunicamycin was present during the induction
period at concentrations preventing induction (50 ng/ml), the
mesenchymes remained negative for laminin fluorescence (Fig.
5a). The same concentrations of tunicamycin did not inhibit
differentiation when applied during the morphogenetic period
(24-72 hr of culture), and bright fluorescence appeared in the
basement membranes of the tubules (Fig. 5b).
The present data show that, prior to morphogenesis, induced
mesenchymal cells in the embryonic kidney express a glyco-
protein of the basement membrane type. We thus demonstrate
a difference between two morphologically indistinguishable
cell populations; only the terminally determined cells express
laminin. In these cells, laminin occurs in a punctate pattern. The
induced state was previously not apparent by any means
available and could only be deduced (5), as in other develop-
mental events, from the fate of the cells (25). The linear staining
typical of the basement membrane of adult tissues is acquired
only later when the epithelialization of the mesenchymal cells
is taking place. Presumably, these changes are dependent on
continuous synthesis of laminin rather than on rearrangement
of the protein deposited in the early stages of tubule induc-
The appearance and redistribution of laminin in the devel-
oping kidney show a striking correlation, both spatially and
temporally, to the previously described early stages of tubule
formation (3, 6, 23). The morphogenesis of the kidney tubules
in vivo is triggered by the ingrowing ureter bud. It has been
suggested that the triggering leads to a gradual increase in ad-
hesiveness of the responding cells, and a morphogenetic field
of determined cells 710 cell layers deep is assumed to form
around the inductor tissue (6). So far, there are no molecular
data to support this hypothesis. It is tempting to suggest that
laminin could be involved in the increased adhesiveness of the
induced cells during the aggregation phase. Laminin was first
detected around the inductor tissue in the undifferentiated
mesenchyme, the cell layer expressing laminin appearing 8-10
cell layers deep. Later, laminin was always detected in the
newly formed aggregates and, subsequently, predominantly
in close association with the basement membrane of the tubules.
When new aggregates formed in the upper cortex, the devel-
opmental sequence was the same.
The expression of laminin in the developing kidney differs
from that of fibronectin, another glycoprotein of basement
membranes. Fibronectin is found in the undifferentiated and
uninduced mesenchyme, and the staining remains positive in
the stroma that surrounds the formed tubules. The change
observed in the expression of fibronectin during tubulogenesis
is the loss of staining in the lumen of the developing tubules and
the appearance of linear staining in the basement membrane
(26). In the kidney, the expression of fibronectin can thus be
considered to be a property of the mesenchymal cells (27),
whereas the appearance of laminin coincides with the terminal
determination of mesenchymal cells into epithelial struc-
The transfilter in vitro experiments enabled a more detailed
analysis of the developmental time sequence and established
that only laminin-positive cells were incorporated into the tu-
bules. The first laminin spots were seen very early, at about 12
hr of culture. This is when the first mesenchyinal cells are in-
duced (5). When aggregates formed and the first cells acquired
an elongated shape, laminin was located prominently in these
areas, and the tissue around the aggregates became negative.
Immunofluorescence did not show the laminin spots in mes-
enchymes cultured alone or in explants treated with tunicam-
ycin during the induction period. Thus, the appearance of
laminin is dependent on unimpaired differentiation. Tuni-
camycin may interfere with tubule formation by inhibiting the
Developmental Biology:Ekblom et al.
Developmental Biology: Ekblom et al.
for staining with laminin antibodies and fluorescein isothiocyanate-conjugated immunoglobulin. Immunofluorescence staining for laminin
in transfilter cultures. (a) At 6 hr no fluorescence is seen in the tissues (M, S). The fluorescence in the filter (F) is a constant finding in short-term
cultures, but it disappears subsequently, probably due to cytoplasmic invasion into the filter pores. (X125.) (b) At 12 hr. The figure shows only
the mesenchyme in which few spots oflaminin are seen. (X335.) (c) At 18 hr there is an increase oflaminin spots in the mesenchyme. The inductor
remains negative. Note that the filter is negative for fluorescence. (X125.) (d) The section illustrated in c stained for DNA. Both tissues are
positive for DNA, whereas laminin was found only in the mesenchyme. (e and f) Laminin fluorescence (e) and phase-contrast micrograph (f)
of a 36-hr culture from which the inductor was removed at 24 hr. Laminin staining is prominent at the sites of cell aggregates (A). (X425.) (g
and h) Higher magnifications of e and f, showing how laminin fluorescence is located around the elongated cells. (X1100.) (i andj) Laminin
fluorescence (i) and Nomarski optics micrograph (j) of a 48-hr culture. Note that the cell aggregates (A) to the left express laminin in a spotty
pattern, but the tubule to the right (T) has laminin located also in the developing basement membrane. (X250.) (k and 1) Laminin fluorescence
(k) and Nomarski optics micrograph (1) of a 72-hr culture. Only linear staining is seen, and it corresponds to the basement membrane (BM)
of tubuli formed. The filter is fluorescent only at the sides where no cytoplasmic invasion into the filter pores has occurred. (X250.)
Sections of transfilter explants of nephrogenic mesenchymes (M) and inductor (S) cultured for various periods of time and processed
Proc. Natl. Acad. Sci. USA 77(1980)
Proc. Natl. Acad. Sci. USA 77 (1980)
transfilter cultures. (X300.) (a) Tunicamycin was present during the
induction period (0-24 hr) and the culture was processed for laminin
fluorescence at 24 hr. Only a few spots of fluorescence are seen. (b)
Tunicamycin was present only during the 24-72 hr of culture, not
during the preceding induction period (0-24 hr). The culture was
processed for immunofluorescence at 72 hr. Linear fluorescence is seen
in the basement membrane.
Effect of tunicamycin on the appearance of laminin in
secretion of laminin. However, this appears to be unlikely be-
cause the formation of basement membranes in the later stages
cannot be inhibited by tunicamycin.
Basement membranes and their components have frequently
been considered of importance for epithelial cell differentiation
and for the maintenance of their organized state (8-13, 28). The
present results are consistent with the possibility that laminin,
a basement membrane glycoprotein, may be responsible for
the increase in adhesiveness during the aggregation phase.
Subsequently, laminin participates in the development of the
basement membrane. The appearance of other basement
membrane glycoproteins and collagens remains to be analyzed
until appropriate immunological tools become available.
We thank Mrs. Anja Tuomi, Mrs. Inkeri Huttunen, and Mrs. Liisa
Pitkinen for their expert technical assistance. This study was supported
by Finska Lakaresallskapet, the National Institutes of Health (Grant
CA 24605), Deutsche Forschungsgemeinschaft (Ti 95/4), and the Sigrid
Grobstein, C. (1955) J. Exp. Zool. 130, 319-340.
Grobstein, C. (1967) Natl. Cancer Inst. Monogr. 26, 279-299.
Saxen, L., Koskimies, O., Lahti, A., Miettinen, H., Rapola, J. &
Wartiovaara, J. (1968) Adv. Morphog. 7, 251-293.
Grobstein, C. (1957) Exp. Cell Res. 13, 575-587.
Saxen, L. & Lehtonen, E. (1978) J. Embryol. Exp. Morphol. 47,
Saxen, L. & Wartiovaara, J. (1966) Int. J. Cancer 1, 271-290.
Ekblom, P., Miettinen, A. & Saxen, L. Dev. Biol., in press.
Bernfield, M. R., Banerjee, S. D. & Cohn, R. H. (1972) J. Cell Biol.
Hay, E. (1978) in Biology and Chemistry of Basement Mem-
branes, ed. Kefalides, N. A. (Academic, New York), pp. 119-
Thesleff, I., Lehtonen, E. & Saxen, L. (1978) Differentiation 10,
Vracko, R. (1974) Am. J. Pathol. 77,314-350.
Kefalides, N. A., ed. (1978) Biologyand Chemistry ofBasement
Membranes (Academic, New York).
Cohn, R. H. Banerjee, S. D. & Bernfield, M. R. (1977) J. Cell Biol.
Vaheri, A. & Mosher, D. (1978) Biochim. Biophys. Acta 516,
Orkin, R. W., Gehron, P., McGoodwin, E. B., Martin, G. R.,
Valentine, T. & Swarm, R. (1977) J. Exp. Med. 145,204-220.
Timpl, R., Rohde, H., Gehron Robey, P., Rennard, S. R., Foidart,
J.-M. & Martin, G. R. (1979) J. Biol. Chem. 254,9933-9937.
Timpl, R., Rohde, H., Ott-Ulbricht, U., Risteli, L. & Bachinger,
H. P. (1979) in Glycoconjugates, Proceedings of the Fifth In-
ternational Symposium, Kiel, September 1979, eds. Schauer,
P., Boer, P., Buddecke, E., Kramer, M. F., Vliegenthart, J. F. G.
& Wiegandt, H. (George Thieme, Stuttgart), pp. 145-146.
Saxen, L. & Saksela, E. (1971) Exp. Cell Res. 66, 369-377.
Tkacz, J. S. & Lampen, J. 0. (1975) Biochem. Biophys. Res.
Commun. 65, 248-257.
Sainte-Marie, G. (1962) J. Histochem. Cytochem. 10, 250-
von der Mark, H., von der Mark, K. & Gay, S. (1976) Dev. Biol.
Russel, W. C., Newman, C. & Williamson, D. H. (1975) Nature
(London) 253, 461-462.
Rienhoff, W. F. (1922) Johns Hopkins Hosp. Bull. 33, 392-
Ekblom, P., Nordling, S., Saxen, L., Rasilo, M-L & Renkonen, 0.
(1979) Cell Differ. 8, 347-352.
Dawid, J. B. & Wahli, W. (1979) Dev. Biol. 69,305-328.
Wartiovaara, J., Stenman, S., & Vaheri, A. (1976) Differentiation
Linder, E., Vaheri, A., Ruoslahti, E. & Wartiovaara, J. (1975) J.
Exp. Med. 142, 41-49.
David, G. & Bernfield, M. R. (1979) Proc. Natl. Acad. Sci. USA
Developmental Biology: Ekblom et al.