/. Embryo!. exp. Morph. Vol. 21, 3, pp. 517-37, June 1969
Printed in Great Britain
517
Differentiation of kidney antigens
in the human foetus
By EWERT U N D E R 1
Department of Serology and Bacteriology and Department of Pediatrics,
University of Helsinki
The appearance of new antigens in the embryo during differentiation has been
investigated by a number of authors. Among the proteins studied were myosin
(Holtzer, 1961; Ebert, 1962), Jens crystallin (Ten Cate & Van Doorenmaalen,
1950), chick embryo haemoglobin (Wilt, 1962), and keratin during feather
formation in chick embryo (Ben-Or & Bell, 1965). The development of liver
proteins in the chick embryo was studied by D'Amelio, Mutolo & Piazza (1963).
Okada & Sato (1963) and Okada (1965) studied the appearance of a 'kidneyspecific1 antigen in the developing mesonephros. Lahti & Saxen (1966) demonstrated the appearance of mouse kidney-specific tubule antigens during development both in vivo and in vitro.
'Kidney-specific' antigens are found in the metanephric proximal secreting
tubules of various mammals (Hill & Cruickshank, 1953; Weiler, 1956; Groupe
& Kaplan, 1967; Nairn, Ghose & Maxwell, 1967), including man (Nairn, Ghose,
Fothergill & McEntegart, 1962), and in the mesonephric tubules of birds and
fish (Nairn et al. 1967). In addition cross-reacting membrane antigens (Dumonde,
1966; Weinberger & Boss, 1966) have been demonstrated in the basement membranes and reticular fibres of many organs, including kidney (Cruickshank &
Hill, 1953). These are implicated in the glomerulonephritis induced by antikidney serum (Masugi, 1933). The appearance of kidney antigens in the human
foetus has not previously been investigated.
The purpose of the present investigation was to relate the appearance of new
antigens to the morphological differentiation of the metanephrogenic mesenchyme of the human kidney using immunodiffusion and immunofluorescence
techniques.
MATERIAL AND METHODS
Antigen
Human foetal and postnatal kidneys were used. Foetal kidneys from thirty-six
foetuses were obtained within 1 h after legal abortion performed by caesarian
section. The crown-rump length of the foetuses was 2-5—12 cm. Kidneys from
1
Author's address: Department of Serology and Bacteriology, University of Helsinki,
Haartmanink 3, Helsinki 25, Finland.
518
E. LINDER
children and adults ( = postnatal kidneys) were obtained at post-mortem within
12 h after death or occasionally as operation specimens.
Small tissue blocks were taken for immunohistology (see below). The main
bulk of the tissues was then roughly minced with scissors in saline to remove as
much blood as possible. The tissue specimens not used for immunohistology
were frozen ( - 2 0 °C) and thawed (room temperature) 2 or 3 times to rupture the
cell membranes. They were then homogenized in an equal amount of cold
distilled water in an ice bath with a Ultra Turrax homogenizer for a total of
5 min, with frequent intervals to avoid heating. The homogenates were centrifuged at 100 rev/min for 10 min to remove tissue debris. The supernatant was
used for immunization. Material for immunodiffusion was prepared from the
supernatant after centrifugation at 10000 g for 30 min. The supernatant was
lyophilized and stored in a desiccator at room temperature until used. The
pellets were stored at — 20 °C to be used as absorption material. Both pooled and
individual homogenates were used.
Tissue blocks for immunofluorescence studies were prepared by freezedrying or cryostat sectioning. Blocks about 2 x 3 mm were cut for freezedrying, and somewhat larger blocks for cryostat sectioning, then frozen in
liquid nitrogen. Freeze-dried blocks were embedded in paraffin wax (Gurr),
m.p. 45 °C, or in Polyester wax (B.D.H.) m.p. 37 °C and sectioned in a serial
microtome at 2-4 fi. Paraffin was removed from sections with xylene. The sections were then passed through ethanol and dried at room temperature.
After drying on the slide cryostat sections were rinsed in buffered saline
(PBS: 0-85 M-NaCl buffered at pH 7-2 with 001 M phosphate buffer) for 20 min
at 4 °C to remove material smeared at sectioning. The sections were then fixed
while still wet in 2 % formol in PBS, acetone, methanol or ethanol at 4 °C or in
a mixture of acetone, methanol and ethanol. After this the sections were rinsed
in PBS for 15 min at 4 °C, and used in immunofluorescence experiments.
Immunization and antisera
Adult rabbits of mixed stock were used. Material for immunization was prepared as described in the previous section, and mixed with an equal amount of
Freund's complete adjuvant (Difco). Fresh material was prepared for each
injection from pooled or individual kidneys. Two rabbits were immunized with
foetal kidney homogenate (one with 0-1 ml/injection, one with 0-5 ml/injection),
and four with postnatal kidney homogenate (two with 0-1 ml/injection, two
with 0-5 ml/injection). The rabbits were given monthly intracutaneous and
subcutaneous injections in the inguinal or scapular regions. The rabbits were
bled every second month, generally 10 days after the usual booster containing
adjuvant. The antibody response was tested by double diffusion. The number
of precipitin lines increased during the first year of immunization. Rabbits
receiving the larger dose of antigen at each injection produced potent antisera in
a somewhat shorter time. Some individual variations in the numbers of pre-
Foetal kidney antigens
519
cipitin lines produced by different antisera against the same antigenic material
were observed. This was not associated with differences in immunofluorescent
staining properties. Immunization was continued for 2-3 years.
A commercial preparation of sheep anti-rabbit globulin (Mann Research
Laboratories, New York) was used in immunofluorescence studies after
labelling (see below). The preparation of antiserum against a-foetoprotein has
been previously described (Linder & Seppala, 1968).
Absorption of immune sera
To increase the selectivity of the anti-kidney sera they were absorbed with
whole homogenate or occasionally with water-soluble or insoluble tissue
fractions or with pooled plasma or serum. Autopsy and tumour material was
prepared as described for kidney. Nephroblastoma tissue was obtained immediately after surgical removal of the tumour. The tumour usually used for
absorption of the antisera was T2. The tumour material is described in a separate
paper (Linder, 1969#). Plasma, serum and the soluble portion of the homogenate after high-speed centrifugation were lyophilized, while the sedimented
tissue fractions were stored at —20 °C as described previously. The minimum
amount of antigen necessary to obtain an effective absorption was determined by
adding increasing amounts of antigen (10, 20, 40, 80, 100, 150 and 200 mg/ml
antiserum). The mixtures were incubated at room temperature for 2 h and then
at 4 °C for 1-4 days. After incubation the mixture was centrifuged for 30 min at
5000 rev/min in the cold to remove the precipitate. Absorbed antisera were
tested by double diffusion against different dilutions of antigen, usually 40, 80
and 100 mg/ml. The tested supernatants were stored at 4 °C.
Immunodiffusion
Double diffusion in agar gel was carried out in a Petri dish in 1 % Ionagar
(Oxoid) or Agarose (lTndustrie Biologique Francaise) in PBS. A modfication of Wadsworth's (1957) microtechnique was used. The thickness of the
agar layer (0-25 mm) was regulated by inserting a nylon line between the matrix
and the supporting glass. The nylon line was removed when the agar had solidified, and the diffusion chamber was kept at 4 °C for at least 2 days before use.
After filling the wells with reactants the dish was kept for 2 h at room temperature
and subsequently transferred to 4 °C. The diffusion chambers were examined
daily. Precipitin lines were usually visible within 24 h, but new lines sometimes
appeared after 7 days. In addition to the conventional arrangement of wells filled
with reactants the cross-diffusion system of Abelev (1960) was used in some
experiments.
Immunofluorescence
Both the direct and indirect fluorescent antibody techniques were used
(Coons & Kaplan, 1950). Antiserum globulin was prepared either by sodium
520
E. LINDER
sulphate precipitation according to Kiraly & Jobbagy (1966), or by DEAESephadex A 50 chromatography (Dedmon, Holmes & Deinhardt, 1965).
Antiserum globulin from anti-kidney and anti-rabbit gamma globulin sera
were conjugated with fluorescein isothiocyanate (F1TC), crystallized and chromatographically pure (Baltimore Biological Co.) The procedure was that of
Lewis, Jones, Brooks & Cherry (1964), but using 0-2 mg F1TC per ml of phosphate buffer (pH 10-5), instead of 0-625 mg/ml, for adding to one ml of 1 % globulin solution. The labelling, for 5 h at room temperature, resulted in a conjugate with an FITC/protein molar ratio of 2-3 determined by the formula of
Mekler et al., quoted by Wagner (1967). Free fluorescein was removed using a
SephadexG-25 column equilibrated with PBS (Killander, Ponten & Roden, 1961).
Over-labelled protein molecules were removed by passage through a DEAE
Sephadex A-50 column, eluting with 0-01 M phosphate, pH 7-2 containing
0-21 M-NaCl (Dedmon et al. 1965). This absorption reduced the FITC/protein
ratio by only 0-2 or less. The eluted solution was passed through a bacterial
filter and then stored at 4 °C.
Procedure
Tissue sections were taken from the rinsing solution after fixation and incubated in a moist chamber with continuous agitation at room temperature. When
the direct method was used sections were incubated with fluorescent antiserum
for 2 h, then rinsed for 15 min in three changes of PBS. In the direct method the
sections were incubated with unlabelled antiserum or chromatographically
purified globulin from it for 2 h. This was followed by rinsing for 20 min in
three changes of PBS. Fluorescent anti-rabbit y-globulin serum, prepared as
described, was added for 45 min. The protein concentration of this antiserum
was about 0-3 %. After incubation the slides were rinsed in PBS and kept in this
solution until examined. The slides could be re-examined after 5 days without
loss of specific fluorescence.
Photomicrography
The slides were examined under a Wild fluorescence microscope equipped with
a high-pressure mercury vapour lamp and a bright-field quartz condenser using
Schott primary filter UG 1 (2 mm) and secondary filter GG 13. For photography
using Kodak Super XX film, Schott primary filter BG 12 (3 mm) and secondary
filter OG 1 gave better results. Immunodiffusion slides were photographed in
dark-field illumination on Agfa Agepe 35 mm document film.
RESULTS
Differentiation of kidney antigens was studied with absorbed antisera against
postnatal kidney and the distribution of foetal kidney antigens was studied
with absorbed antisera against foetal kidney.
The effect of absorptions was tested against the tissue material used in both
Foetal kidney antigens
521
immunodiffusion and immunofluorescence. The following controls for specificity of the immunofluorescence studies were performed:
A. Direct method. (1) Labelled globulin solution prepared from pre-immunization serum (0-serum). No fluorescence. (2) Blocking by pre-incubation of
slides with unlabelled immune serum; no fluorescence in the proximal tubule.
(3) Pre-incubation with O-serum; no effect on the tubule fluorescence.
B. Indirect method. (1) Fluorescent anti-rabbit y-globulin serum only; no
fluorescence. (2) O-serum in the first step instead of immune serum; occasional
staining of vessel media and a very faint collective tubule fluorescence (Fig. 4 c).
(3) Fluorescent anti-human y-globulin serum in the final step instead of fluorescent anti-rabbit y-globulin serum; no staining. (4) The y-globulin fraction of
immune serum instead of whole serum; no change in specific fluorescence; the
fluorescence of vessel media was not altered and there was a bright fluorescence
of elastic fibre. (5) y-globulin from O-serum. A bright elastic fibre fluorescence
was seen in addition to the fluorescence of vessel media.
W\
Fig. 1. Comparison of two antigen-antibody systems using cross-diffusion, (a)
Anti-kidney serum (AK) against pooled adult kidney (K), and anti-foetal kidney
serum (AFK) against pooled foetal kidney (FK). (b) AK against pooled newborn
kidney (A^) and AFK against/ 7 ^. Antigens present in/fare also found in FK and Kr.
However, a number of precipitin lines are formed against FK and Kx but not against
K. All activity against normal human serum (NHS) was absorbed from antisera.
.
i-rr .
Inmnmodijjusion
Foetal kidney antigens
J
b
Anti-foetal kidney serum gave precipitin lines against foetal kidney that were
not seen when it reacted against postnatal kidney (Fig. \d). These lines will be
called a and b. The foetal kidney antigens were also demonstrated in kidneys
from newborn infants (K^ (Fig. \b).
The antibodies in anti-foetal kidney serum that reacted with adult kidney
homogenate in double diffusion could be absorbed by 40 mg of this lyophilized
kidney tissue (soluble fraction) (Fig. 2). The absorbed antiserum gave lines a and b.
Absorption with 50 mg/ml pooled foetal organs (excluding kidney) or foetal
serum (50 mg/ml) prevented the formation of a lines. After these absorptions
522
E. LINDER
the anti-foetal kidney sera still formed one or two lines (marked b) when
reacted with foetal kidneys. However, anti-kidney serum also produced theline(s)
b when reacting with foetal kidney (Fig. 1 a and b). The antigen(s) responsible
for the production of line(s) b were demonstrated in some postnatal kidneys in
small amounts. This explains why anti-kidney serum produced line b. The
amount of antigen in pooled material was, however, not sufficient to absorb the
antibodies producing lines b from anti-foetal kidney serum. A reaction of
identity was seen between a lines and an antiserum against a-foetoprotein.
Fig. 2. Foetal kidney antigens. Reactions between FK and AFK absorbed with
NHS and increasing amounts of AK: 20 mg/ml (2), 40 mg/ml (3) or 80 mg/ml (4).
Additional absorption with foetal serum (20 mg/ml) abolished a number of precipitin lines (designated a), but did not affect two lines designated b (5). Absorption
with 100 mg/foetal serum per ml did not affect the formation of lines b (6). Abbreviations as in Fig. .1.
Fig. 3. Reaction between K and AK absorbed with tumour (T) and NHS. Eight
precipitin lines are formed against kidney, but none against the tumour used for
absorption. One line deviates slightly by another tumour (7^). Abbreviations as in
Fig. 1.
Immunofluorescence
Foetal kidney antigens were studied by immunofluorescence using antisera
against foetal kidneys absorbed with pooled adult kidneys and normal human
serum. The antiserum had an affinity for the undifferentiated foetal kidney
mesenchyme. The fluorescence had the same distribution as the interstitial
connective tissue (Fig. 5). The tubules and glomeruli were not stained. After
additional absorption with foetal serum the antiserum had no specific affinity for
foetal kidney. Thus the staining of undifferentiated kidney mesenchyme seems
to be due to antigens marked a.
Foetal kidney antigens
523
Antigens appearing during differentiation of the metanephrogenic mesenchyme
Immunodiffusion
Antisera against postnatal kidney absorbed with normal human serum still
gave about twenty precipitin lines in double diffusion against postnatal kidney.
Precipitin lines produced against postnatal kidney also formed against foetal
kidney. Antisera further absorbed with tumour gave 6-8 precipitin lines against
foetal or postnatal kidneys (Fig. 3). Antisera further absorbed with lung left six
precipitin lines, with liver four or five and with small intestine two lines against
kidney.
Immunofluorescence
Absorption of antisera against postnatal kidneys with whole homogenates of
a number of normal tissues and some nephroblastomas resulted in an antiserum
that stained different parts of the nephron but not the undifferentiated kidney
mesenchyme (Table 1). All of these absorptions removed the activity against
undifferentiated kidney mesenchyme. Normal tissues were more efficient than
tumour tissue as judged by double diffusion and by a more widespread fluorescence in the nephron.
Table 1. Immunohistochemical localization of kidney antigens
using various absorbed antisera
Anti-kidney serum absorbed with normal human serum and:
O
Interstitial CT & mes.
Glomerulus BM
Tubule BM
Glomerulus capill. BM
Glomerulus epithelial cells
Glomerulus endothelial cells
Proximal secreting tubules
Loops of Henle
Distal tubules
Collecting tubules
Vessel media
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
—
+
—
+
+
—
—
+
—
+
+
—
_
_
[
.
o
OS
o
Cu
PH
—
—
—
+
+
—
_
_
_
—
+
—
+
+
—
_
_
_
_
s
—
+
—
+
—
_
_
CT = connective tissue, mes. = undifferentiated mesenchyme, BM = basement membrane,
T = tumour homogenate, T' = soluble part of tumour homogenate, Tsed = sediment of
tumour homogenate, Spl = spleen homogenate, Plac = placenta homogenate, Lu = lung
homogenate, Li = liver homogenate, Gt = small intestine homogenate. + = presence of
fluorescence, - = absence of fluorescence, (+) = weak fluorescence.
524
E. LINDER
Antisera against postnatal kidneys absorbed with normal human serum and
the soluble fraction of tumour homogenate, localized diffusely in postnatal
kidney. It reacted with undifferentiated kidney mesenchyme but failed to stain
the condensed areas of mesenchyme in differentiating foetal kidney (Fig. 5).
Additional adsorption with the insoluble fraction of tumour homogenate
abolished the fluoresecence of the interstitial connective tissue and undifferentiated mesenchyme, but did not affect that of the different parts of the
nephron (Fig. 4b). Fluorescence was seen also in the basement membranes of
Fig. 4. Foetal kidney from 80 mm foetus (rump-head length), (a) Haematoxylin-eosin,
(b) treated with anti-adult kideny serum absorbed with NHS and tumour. Staining
present in all parts of the nephron. (c) control section treated with pre-immunization
serum.
the glomeruli and the tubules. (Figs. 6b, 7, 14, \6a). Staining of vessel media
was also present (Fig. 14). The effect of absorption with AB red blood cells
was tested in each new kidney studied. This absorption abolished staining of red
blood cells when present and weakened the collecting tubular fluorescence. If
an effect due to this absorption was noticed, absorption with red blood cells
was used in further experiments.
Foetal kidney antigens
525
Control pre-immunization sera sometimes stained the vessels, but no other
part of the sections. Unless otherwise stated the differentiation of kidney antigens was studied using antiserum against postnatal kidneys absorbed with
normal human serum, and whole homogenate of tumour. The sera were absorbed further with homogenates of various normal organs in order to distinguish between different antigens in the nephron (Table 1).
Fig. 5. Initiation of morphological differentiation of the metanephrogenic mesenchyme—the condensation stage. Fluorescence is seen in the undifferentiated mesenchyme (mes.) but not in condensed areas (c) at the tips of the collecting tubules (col.)
The antigens appearing during differentiation could be classified on the basis of
their distribution and their staining with antisera absorbed with different
normal tissues. There were five groups of antigens, (a) Basement membrane
antigens. The fluorescence was abolished by absorption with placenta, but only
partially abolished by absorption with spleen, (b) Glomerular epithelium
antigen(s). Stained after absorption with placenta, (c) Glomerular endothelium
antigen(s). Did not stain after placental absorption, (d) Proximal secreting
tubules contained at least two specific antigens. One of these was present also in
loops of Henle (but nowhere else in the kidney). The fluorescence was abolished
by absorbing antiserum with small intestine, (e) The proximal secreting tubule
fluorescence was still present after this absorption, but was abolished after
absorption with kidney.
526
E. LINDER
Foetal kidney antigens
527
Mesonephros
Mesonephric glomeruli and tubules were stained before metanephric development had led to the differentiation of the parts of the nephron (Fig. 6). At
early stages of morphogenesis the fluorescence in the metanephros was restricted to the basement membranes of the branching ureteric bud—the
prospective collecting tubules.
Metanephros
1. Basement membranes
Fluorescence of basement membranes was seen at epitheliomesenchymal
borders and in glomerular capillary tufts. The renal vesicle was surrounded by a
fluorescent limiting membrane (Fig. 7), which became more intensely stained
Fig. 7. Renal vesicle (rv) surrounded by a fluorescent limiting membrane. Note fusion
of this membrane with the basement membrane of the collecting tubule (col.).
Fig. 8. S-shaped body. Shows appearance of proximal secreting tubule antigens (pr).
The basement membrane is doubled next to the collecting tubule.
Fig. 9. Formation of glomerular epithelium antigens (ep) and proximal secreting
tubule antigens (pr) seen in the luminal part of tubule cells.
during the development of the S-shaped body. The membrane of the renal
vesicle bordering the tubule fused with the membrane of the collecting tubule
(Fig. 7). Concurrently the membrane fluorescence began to split as the S-shaped
body developed (Fig. 8). The cells of the future Bowman's capsule were surrounded by a less intensely stained membrane than the upper and medial segments. At the time of early cleft formation there was a marked increase in the
intensity of the basement membrane fluorescence of the medial and lower segments of the S-shaped body facing the cleft (Fig. 8).
Fig. 6. Section through 20 mm embryo showing early metanephros (mt) and degenerating mesonephros (ms). In addition to basement membrane staining, seen in
both mesonephros and metanephros, there is bright staining of mesonephric
tubules and glomeruli. In the mesonephric duct (d) there is only basement membrane
fluorescence. Note weak intestinal brush border staining at (g).
34
JEEM2I
528
E. LINDER
Foetal kidney antigens
529
The capillary basement membrane fluorescence was never clear and well
demarcated in the primitive glomerulus (Fig. 12). Instead the fluorescence of the
endothelial cells forming the primitive glomerular tuft was continuous with that
of the capillary walls. Gradually, during the maturation of the glomerulus, the
staining became more distinct and more like membrane staining.
Absorption with placenta homogenate abolished all the membrane fluorescence. After absorption with spleen there was still a faint fluorescence in some
basement membranes of the tubules, but the basement membranes of the
glomerulus were not stained.
2. Glomeruli
Antigens were demonstrated in both epithelial and endothelial cells of the
glomeruli.
Epithelial cells. Antigenic differentiation was seen when the single layers of
cells of the presumptive glomerulus epithelium began to proliferate, and to
form ingrowths between the cells filling the primitive glomerular tuft. At this
stage the future Bowman's capsule cells, which constitute the outer layer of the
bowl-like primitive glomerulus anlage, are flat and contain a reduced amount of
cytoplasm (Fig. 9).
No capillary structures could be seen in the cells constituting the glomerular
tuft (Figs. 10, 11, 14). In the primitive elongated epithelial cells the fluorescence
seemed to be associated with the cell membranes (Fig. 13). However, as proliferation proceeded the fluorescence was clearly seen to be associated with the cytoplasm (Fig. 15). At this stage the capillary lumina of the tuft had formed, and
the epithelial cells were arranged in septa between the primitive capillary loops.
As the glomerulus matured and the intercapillary spaces were reduced there
was still an irregular fluorescence associated with the membranes of glomerular
epithelium cells. Absorption with spleen or placenta homogenates did not affect
the epithelial fluorescence (Figs. 13,15). The fluorescence was however abolished
by absorption with lung homogenate.
Endothelial cells. At the stage when the glomerular epithelium antigen was
first demonstrated a new antigen was seen in cells of the primitive tuft bordering
Figs. 10, 11. Beginning of proliferation of glomerular epithelium cells towards the
primitive glomerular tuft. Endothelial and epithelial glomerulus antigens located at
the epithelio-endothelial border.
Fig. 12. Early lumen formation in primitive glomerular capillary tuft. Fluorescence
in both endothelial and epithelial cells of the glomerulus is seen.
Fig. 13. Glomerular epithelium fluorescence demonstrated with antiserum absorbed
with spleen. No fluoresence visible in endothelial cells and basement membranes.
Fig. 14. Fluorescence in the endothelium but not the epithelium of the primitive
glomerular tuft.
Fig. 15. Appearance of the glomerular epitheliumfluoresecenceafter absorption of
antiserum with spleen. The stage of development is that of Fig. 12.
530
E. LINDER
the epithelial layer. At about the time of formation of cavities in the cell mass
the endothelium antigen was seen in all the cells of the tuft (Fig. 14). The
fluorescence was homogeneous in the walls of the newly formed cavities (Fig. 12).
It was continuous with the capillary basement membrane fluorescence and formed
a continuous strand in the mature glomeruli.
Antibodies against the endothelial antigen(s) could be absorbed by spleen
hornogenate (Figs. 13, 15).
3. Tubules
Tubule fluorescence was demonstrated in all parts of the nephron (Fig. 4).
Fluorescence of the proximal secreting tubule was the most prominent, and
could also be demonstrated by the indirect fluorescent antibody technique. The
Fig. 16. In (a) the proximal tubule, glomerulus and basement membrane antigens
are stained. In (b) only proximal secreting tubulefluorescenceis seen after absorption
of antiserum with lung and placenta. Note the extension of proximal secreting tubule
cells to form a part of the parietal Bowman's capsule.
fluorescence was very intense in the brush border of the cells, which often filled
the whole lumen of the tubules. In addition to the brush border fluorescence
there was a distinct fluorescence in the cytoplasm of the cells. This was usually
more obvious in distal parts of the proximal secreting tubules. Both the cytoplasmic and the brush border fluorescence extended into the Bowman's space
(Fig. 1 b). During maturation of the kidney the fluorescence extending into the
glomerulus became gradually less prominent and the fluorescence was confined
to the brush border. Distally, where the proximal secreting tubule is continuous
with the loops of Henle, there was an abrupt change in the character of the fluorescence. The narrow loops of Henle had a uniform cytoplasmic staining. This was
most obvious in the renal papilla where the loops of Henle contrasted clearly
Foetal kidney antigens
531
with the weakly stained collecting tubules. In the cortical region the weakly
stained tubules were either collecting tubules or distal tubules. The collecting
tubule fluorescence was continuous with that of the epithelium of the renal
pelvis.
The antigens of the proximal secreting tubule in the medial segment of the
S-shaped body were demonstrated at about the same developmental stage as
glomerulus antigens (Fig. 9). Sometimes the proximal secreting tubule antigens
were seen a little earlier than glomerular epithelium antigens (Fig. 8).
Non-identity of glomerulus and tubule antigens was demonstrated by absorption experiments. Absorption with placenta and lung homogenates abolished the
staining of the glomeruli and basement membranes (Fig. 16), but not that of the
proximal secreting tubules and loops of Henle.
Absorption with small intestinal homogenate abolished the staining of loops
of Henle, but reduced the fluorescence of the cytoplasm of the proximal secreting tubules only slightly, and the fluorescence of the brush border not at all.
Absorption with whole kidney homogenate or washed sediment abolished all of
the staining, but absorption with the lyophilized soluble fraction did not affect
the staining.
CONCLUSION
Antigenic differentiation of the metanephrogenic mesenchyme
Condensation of the undifferentiated metanephrogenic mesenchyme at the
tips of the ureteric buds involved the first immunohistochemically demonstrated change—loss of connective tissue and foetal antigens. Development of the
renal vesicle was accompanied by surrounding fluorescence. This fluorescence
appeared to be associated with a basement membrane, although no basement
membrane can be seen in the renal vesicle at this stage with the light microscope. This fluorescence fused with that of the basement membrane of the
collecting tubule. As the S-shaped body was formed proximal secreting tubule
antigens appeared in the medial segment. Endothelium and epithelium antigens
became demonstrable at the interface between the cells of the tuft and those of
the lower segment of the S-shaped body after formation of the primitive glomerular tuft. As the primitive vascular spaces developed, the fluorescence in the
endothelium and in the vascular membrane basement became indistinguishable.
During maturation of the glomerulus the epithelial and endothelial cells of the
intercapillary spaces gradually degenerated and the antigens become closely
associated with the capillary basement membranes.
DISCUSSION
Connective tissue and foetal antigens were found in the undifferentiated
kidney mesenchyme. These antigens were lost during the initial step in differentiation—condensation of mesenchyme cells at the tips of the ureteric buds. It is
34-2
532
E. LINDER
not possible to decide whether these antigens are cellular or intercellular. If they
are intercellular their absence in the condensed areas may be the result of
increased cellular adhesion, leading to a gradual disappearance of the intercellular spaces (Saxen & Wartiovaara, 1966).
The reactions with foetal kidney antigens might be due to antibodies against
a-foetoprotein, since absorption with foetal serum removed the activity. This
interpretation is supported by the fact that antiserum against a-foetoprotein
became localized in the same areas as anti-foetal kidney serum (Linder &
Seppala, 1968).
Immunization with unfractionated kidney material resulted in antibodies
against many tissue antigens.
After absorption with nephroblastoma tissue a number of cross-reacting
antibodies were removed. This was seen as a decrease in the number of precipitin lines in double diffusion, and as a distinct localization of fluorescent antiserum in various parts of the nephron and vessel walls.
The antigens with which these antisera reacted were not detected by immunofluorescence in the undifferentiated foetal kidney mesenchyme. The absorbed
antiserum could therefore be used to study the appearance of a number of tissue
antigens during morphogenesis. Further absorptions with a number of normal
tissues made the anti-kidney sera more specific. Antisera could then distinguish
between common glomerular epithelium and endothelium antigens, tubule
antigens, cross-reacting and 'kidney-specific' proximal secreting tubule antigens.
Formation of the renal vesicle or pretubular structure was associated with a
fluorescence like that of basement membranes. No basement membrane could
be seen with the light microscope at this stage, but using the electron microscope
Kurtz (1958) and Jokelainen (1963) observed some amorphous material, resembling primitive basement membrane, surrounding the renal vesicle.
Wartiovaara (1966 a, b) used the electron microscope to study the early phase
of basement membrane formation during differentiation of mouse metanephrogenic mesenchyme in vitro. He demonstrated small fibrils embedded in a homogeneous matrix surrounding the pretubular structures, and provided evidence
suggesting that pretubular cells participated in the formation of this membrane.
The possibility that mesenchyme was also involved could not be excluded. This
agrees with the observation of Kallman & Grobstein (1965), who showed that
mesenchyme participates in membrane formation around salivary epithelium.
Absorption of antiserum with spleen abolished almost all of the basement
membrane fluorescence. This indicates that the fluorescence was principally due
to antigenic material of mesenchymal origin. A faint fluorescence remained
surrounding some tubules and may have been due to the presence of antigens of
epithelial origin (Pierce, 1966).
The cells of the glomerular tuft were shown to acquire antigenic properties
lacking in the cells of the stromal mesenchyme. The new antigen(s) seemed to
become associated with the walls of the early capillaries as development of the
Foetal kidney antigens
533
glomerular capillary tuft proceeds. This agrees with the observations of Jokelainen (1963). He showed that during the formation of the S-shaped body stromal
mesenchyme cells invade the cleft separating the lower and middle limbs, and
suggested that these cells are the progenitors of the endothelial cells of the
glomerular tuft. He demonstrated that the development of the lumens of the
glomerular capillaries is initiated by a gradual widening of the intercellular
spaces in the compact mass of endothelial cells. The cells containing the endothelial antigen(s) could not be seen to be directly derived from the mesenchyme
cells, and there may be an alternative to the postulated transformation of
mesenchymal cells to epithelial cells proposed by Jokelainen. It cannot be
excluded that a few cells of endothelial origin accompany the 'row of erythrocytes' observed by Jokelainen, even if he could not see any capillary wall surrounding these cells. The studies of Potter (1965) and Osathanodh & Potter
(1966) indicate that at least from a functional point of view there is a capillary
already present during early cleft formation and that this vessel, by gradual
broadening and fenestration, will give rise to the glomerular capillaries. This
developmental sequence seems to reflect rapid proliferation of endothelial cells.
Scott & Rowell (1967) observed similar fluorescence to that described here,
in the primitive capillary tuft of the developing rat glomerulus. They used an
antiserum against splenic reticulum or isolated glomeruli. This agrees with the
finding in the present paper that absorption of antiserum with spleen abolished
the endothelial fluorescence. Scott (1957) also observed fluorescence in the
glomerular epithelium. He showed that the antigen was not organ-specific by
obtaining identical fluorescence with antiserum against synovial membrane and
against glomeruli. The nature of the glomerular epithelium antigen(s) was not
investigated further in the present study. However, the epithelial basement membrane antigen (Pierce, 1966) does not seem to be responsible for the immunofluorescent staining of these cells, since absorption with placenta abolished the
fluorescence of basement membranes but not that of the glomerular epithelium.
The proximal secreting tubule antigens were demonstrated as a lumen fluorescence in the medial segment of the S-shaped body. It is interesting that there
was no demonstrable difference between the time of appearance of the 'kidneyspecific' and of the cross-reacting proximal secreting tubule antigens. Okada
(1965) studied the appearance of 'kidney-specific' proximal secreting tubule
antigens in the developing chick mesonephros. These antigens 'became demonstrable coinciding with an initiation of the epithelial architecture of the cell',
and there was a weak specific fluorescence in the cells of the condensed nephrogenic mesenchyme. Lahti & Saxen (1966) demonstrated 'kidney-specific'
antigens in differentiating mouse metanephrogenic mesenchyme in vitro at a
rather late stage in tubule formation (Wartiovaara, 19666).
The antigens of kidney tubules cross-react with those of lung and liver
(Okada, 1962; Okada & Sato, 1963). However, even after absorption with these
tissues the antisera are not necessarily kidney-specific, since Nairn et al. (1967)
34-3
534
E. LINDER
observed reactivity with seminal vesicle and Edgington et al. (1967) isolated
two proximal tubule antigens which cross-react with brush border antigens of
small intestinal mucosa. In addition there is some cross-reaction between the
antigens of proximal secreting tubules and those of the tubules of the epididymis
(Linder, 19696). Absorption with small intestine neutralized antibodies reacting
with epididymis, gut and seminal vesicle but not those reacting with the brush
border of the proximal tubules. Therefore these brush border antigens may be
kidney-specific.
SUMMARY
Differentiation of antigens in the metanephrogenic mesenchyme of the
human foetal kidney was studied by immunofluorescence and double diffusion
in agar gel using antisera against foetal and postnatal kidneys.
Absorption with nephroblastoma tissue removed antibodies against common
or cross-reacting antigens. The anti-kidney sera then did not react with undifferentiated kidney mesenchyme and could be used to detect kidney antigens
appearing during differentiation.
Foetal kidney antigens were demonstrated with anti-foetal kidney sera after
absorption with normal human serum and adult kidney. The kidney antigens
appearing during differentiation were classified in five groups on the basis of
their histological distribution: (1) basement membranes, (2) glomerular epithelium, (3) glomerular endothelium, (4) proximal secreting tubules, (5) loops of
Henle.
Antisera absorbed with various normal tissues could discriminate between
these groups, except that glomerular endothelial antigens could not be differentiated from capillary basement membrane antigens.
The antigens of the proximal secreting tubules were of two types: cytoplasmic
and brush border. The latter were both cross-reacting and kidney-specific.
During the condensation stage, which initiates the differentiation of the metanephrogenic mesenchyme, foetal and interstitial connective tissue antigens decreased and became undetectable. New antigen(s) could be seen in the limiting
membrane of the renal vesicle. Proximal secreting tubule antigens were demonstrated in the medial segment of the S-shaped body. Formation of the primitive
glomerular tuft was associated with the appearance of both epithelium and endothelium antigens in the glomerulus.
The present results show that tissue specificity is acquired during organogenesis in the human metanephros. The findings do not support the view that
initiation of histogenesis is preceded by the synthesis of new molecular species.
Foetal kidney antigens
535
RESUME
Differentiation d'antigenes du rein chez le foetus humain
La differentiation d'antigenes dans le mesenchyme metanephritique foetal
humain a ete etudie par immunofluorescence et double diffusion dans un gel
d'agar en utilisant des antiserums a l'egard des reins fcetaux et postnataux.
L'absorption par du tissu lisse de blasteme renal a enleve les anticorps vis a
vis des antigenes communs ou a reaction croisee. A la suite de cela, les serums
anti-rein n'ont pas reagi avec du mesenchyme renal indifferencie et n'a pas ou
etre utilise pour la detection d'antigenes renaux qui auraient pu apparaitre au
cours de la differentiation.
Des antigenes renaux fcetaux ont ete demontres en utilisant des serums antirein foetal absorbes avec du serum humain normal et du rein adulte. Les antigenes renaux apparaissant au cours de la differentiation ont ete classes en cinq
groupes sur la base de leur distribution histologique: (1) membranes basales,
(2) epithelium glomerulaire, (3) endothelium glomerulaire, (4) tubes proximaux,
(5) anses de Henle.
Des antiserums absorbes par des tissus normaux varies ont pu faire une
discrimination entre ces groupes, sauf que l'antigene glomerulaire endothelial
n'a pu etre distingue d'antigenes de la membrane basale des capillaires.
Les antigenes des tubes proximaux etsient de deux types: cytoplasmiques et de
bordure en brosse. Ces derniers montraient a la fois une reaction croisee et une
reaction renale specifique.
Pendant le stade de condensation qui marque le debut de la differentiation du
mesenchyme metanephritique, les taux d'antigenes des tissus conjonctifs foetal
et intersticiel diminuent au point que ces antigenes n'ont plus pu etre detectes.
De nouveaux antigenes (ou antigene) ont pu etre deceles dans la membrane
limitante des vesicules renales. Des antigenes de tubule secreteur proximal ont
ete demontres dans le segment moyen du corpuscule en S. La formation de la
houppe glomerulaire primitive s'est montree associee a l'apparition des deux
antigenes epithelial dans le glomerule.
Ces resultats demontrent que la specificite tissulaire est acquise pendant
l'organogenese du metanephros humain. Les faits constates ne sont pas en
faveur de Fidee que l'initiation de Fhistogenese puisse etre precedee de la
synthese de nouvelles especes de proteines.
REFERENCES
ABELEV, G.
I. (1960). Modification of the agar precipitation method for comparing two antigenantibody systems. Folia Biol. 6, 56-8.
BEN-OR, S. & BELL, E. (1965). Skin antigens in the chick embryo in relation to other developmental events. Devi Biol. 11, 184-201.
COONS, A. H. & KAPLAN, M. H. (1950). Localization of antigen in tissue cells. II. Improvements in a method for the detection of antigen by means of fluorescent antibody. /. exp.
Med. 91, 1-13.
536
E. LINDER
B. & HILL, A. G. S. (1953). The histochemical identification of a connective
tissue antigen in the rat. /. Path. Bad. 66, 283-9.
D'AMELIO, V., MUTOLO, V. & PIAZZA, E. (1963). A serological study of the cell fractions during
the embryonic development of liver in chick. Expl Cell Res. 31, 499-507.
DEDMON, R. E., HOLMES, A. W. & DEINHARDT, F. (1965). Preparation of fluorescein isothiocyanate-labelled globulin by dialysis, gel filtration, and ion exchange chromatography
in combination. J. Bad. 89, 734-9.
DUMONDE, D. C. (1966). Tissue-specific antigens. Advanc. Immunol. 5, 245-384.
EBERT, J. D. (1962). The acquisition of biological specificity. In The Cell (ed. J. Brachet and
E. E. Mirsky), vol. i, pp. 619-93. New York, London: Academic Press.
EDGINGTON, T. S., GLASSOCK, R. J., WATSON, J. I. & DIXON, F. J. (1967). Characterization
and isolation of specific renal tubular epithelial antigens. /. Immunol. 99, 1199-1210.
GROUPE, W. E. & KAPLAN, M. H. (1967). A proximal tubular antigen in the pathogenesis of
autoimmune nephrosis. Fedn Proc. Fedn Am. Socs exp. Biol. 26, 573.
HILL, A. G. S. & CRUICKSHANK, B. (1953). A study of antigenic components of kidney tissue.
Br. J. exp Path. 34, 27-34.
HOLTZER, H. (1961). Aspects of chondrogenesis and myogenesis. In Synthesis of molecular and
cellular structure, Growth Symp. no. 19 (ed. D. Rudnick), pp. 35-87. New York: Ronald
Press.
JOKELAINEN, P. (1963). An electron microscope study of the early development of the rat
metanephric nephron. Ada Anatomica (Suppl. 47 = 1 ad Vol. 52).
KALLMAN, F. & GROBSTEIN, C. (1965). Source of collagen at epithelio-mesenchymal interfaces during inductive interaction. Devi Biol. 11, 169-83.
KILLANDER, J., PONTEN, J. & RODEN, L. (1961). Rapid preparation offluorescentantibodies
using gelfiltration. Nature, Lond. 192, 182-3.
KIRALY, K. & JOBBAGY, A. (1966). Control of conjugation of fluorescent protein tracing.
Path. Microbiol. 29, 865-72.
KURTZ, S. M. (1958). The electron microscopy of the developing human renal glomerulus.
Expl Cell Res. 14, 355-67.
LAHTI, A. & SAXEN, L. (1966). Studies on kidney tubulogenesis. VIII. Appearance of kidneyspecific antigens during in vivo and in vitro development of secretory tubules. Expl Cell Res.
44, 563-71.
CRUICKSHANK,
LEWIS, W. J., JONES, W. L., BROOKS, J. B. & CHERRY, W. B. (1964). Technical considerations
in the preparation of conjugated fluorescent antibody. Appl. Microbiol. 12, 343-8.
E. (1969a). The antigen structure of nephroblastomas. Int. J. Cancer. 4, 232-47.
E. (19696). Cross-reacting antigens in kidney and other organs. Ann. Med. exp.
Fenn. 47. (In the Press.)
LINDER, E. & SEPPALA, M. (1968). Localization of a-foetoprotein in the human foetus and
placenta. Ada path, microbiol. scand. 73, 565-71.
MASUGI, M. (1933). Uber das Wesen der spezifischen Veranderungen der Niere und der
Leber durch das Nephrotoxin bzw. das Hepatotoxin. Zugleich ein Beitrag zur Pathogenese der Glumerulonephritis und der eklamptischen Leberkrankung. Beitr. path. Anat.
91,82-112.
LINDER,
LINDER,
NAIRN, R. C , GHOSE, T., FOTHERGILL, J. E. & MCENTEGART, M. G. (1962). Kidney specific
antigen and its species distribution. Nature, Lond. 196, 385-7.
R. C , GHOSE, T. & MAXWELL, A. (1967). Distribution of nephric antigens in Australian vertebrates. Nature, Lond. 215, 867-8.
OKADA, T. S. (1962). Tissue specificity in the soluble antigens in kidney microsomes. Nature,
Lond. 194, 306-7.
OKADA, T. S. (1965). Development of kidney-specific antigens: an immuno-histological study.
/. Embryol. exp. Morph. 13, 285-97.
OKADA, T. S. & SATO, A. G. (1963). Soluble antigens in microsomes of adult and embryonic
kidneys. Expl Cell Res. 31, 251-65.
OSATHANODH, V. & POTTER, E. L. (1966). Development of the human kidney as shown by
microdissection. V. Development of vascular pattern of glomerulus. ArchsPath. 82,403—11.
NAIRN,
Foetal kidney antigens
537
G. B. (1966). The development of basement membranes of the mouse embryo. Devi
Biol. 13, 231-49.
POTTER, E. L. (1965). Development of the human glomerulus. Archs Path. 80, 241-55.
SAXEN, L. & WARTIOVAARA, J. (1966). Cell contact and cell adhesion during tissue organization. //;/. /. Cancer. 1, 271-90.
SCOTT, D. G. (1957). A study of the antigenicity of basement membrane and reticulin. Br. J.
exp. Path. 38, 178-85.
SCOTT, D. G. & ROWELL, N. R. (1967). Alternations in the antigenic constitution of renal
glomerular capillaries accompanying the histological maturation of renal glomeruli in the
rat. Ann. rheum. Dis. 26, 10-17.
TEN CATE, G. & VAN DOORENMAALEN, W. J. (1950). Analysis of the development of the lens in
chicken and frog embryos by means of the precipitin reaction. Proc. K. ned. Akad. Wet.
53, 1-18.
WADSWORTH, C. (1957). A slide microtechnique for the analysis of immune precipitates in
gel. ////. Archs Allergy appl. Immun. 10, 350-60.
WAGNER, M. (1967). Floureszierende Antikorper und ihre Anwendung in der Mikrobiologie.
Injektionskrankheiten und ihre Erreger, Band 5. Jena: Gustav Fischer Verlag.
WARTIOVAARA, J. (1966a). Studies on kidney tubulogenesis. V. Electron-microscopy of
basement membrane formation in vitro. Ann. Med. exp. Fenn. 44, 140-50.
WARTIOVAARA, J. (19666). Cell contacts in relation to cytodifferentiation in metanephrogenic
mesenchyme in vitro. Ann. Med. exp. Fenn. 44, 469-503.
WEILER, E. (1956). Antigenic differences between normal hamster kidney and stilboestrolinduced kidney carcinoma. A histological demonstration by means of fluorescing antibodies.
Br. J. Cancer 10, 550-63.
WEINBERGER, N. J. & Boss, J. H. (1966). A comparative study of nephrotoxic serum antigens
of diverse rat organs. Path. Microbiol. 29, 324-40.
WILT, F. (1962). The ontogeny of chick embryo hemoglobin. Proc. natn. Acad. Sci. U.S.A.
48, 1582-90.
PIERCE,
{Manuscript received 2 December 1968)