THE ANATOMICAL RECORD 290:1500–1507 (2007)
Vascular Arrangement and
Ultrastructure of the European Eelpout
Zoarces viviparus Ovary: Implications
for Maternal–Embryonic Exchange
PETER VILHELM SKOV,1* THOMAS FLARUP SORENSEN,2 HANS RAMLOV,2
1
AND JOHN FLENG STEFFENSEN
1
University of Copenhagen, Marine Biological Laboratory, Helsingør, Denmark
2
Roskilde University, Department of Life Sciences and Chemistry, Roskilde, Denmark
ABSTRACT
The structural basis for exchange between maternal serum and ovarian fluid in the viviparous teleost Zoarces viviparus was investigated.
Casts of the ovarian vasculature showed that blood supply to the ovary
is initially directed to the follicular appendages lining the ovarian wall
through thick-walled muscular arteries running along the ovary wall
and within the follicular appendages. The follicles had a rich capillary
network with diffusion distances between maternal blood and ovarian
fluid comparable to those found for gill epithelia, suggesting this is the
primary site of gas exchange between maternal plasma and ovarian
fluid. Follicular capillary beds were continuous with those in the ovary
wall and were eventually drained by the ovarian and intestinal venous
systems. The barrier between ovarian fluid and maternal blood consisted
of the endothelial cells of the maternal blood vessels and a layer of epithelial cells lining the ovarian lumen, with an intermittent layer of loose
connective fibers. Junctional complexes between cells were predominantly anchoring junctions with the occurrence of occasional occluding
junctions, supporting the possibility of paracellular transport from
maternal serum to ovarian fluid of small molecular weight compounds.
Heavy investment in keratin filaments suggests that follicles are tissues
of high structural integrity. Evidence for protein synthesis in the ovarian
lining was found in the form of Golgi apparatus and rough endoplasmic
reticulum. Although numerous cytoplasmic vacuoles and secretory granules were present in both epithelial and endothelial cells, the fate of synthesized protein remains to be determined. Anat Rec, 290:1500–1507,
2007. Ó 2007 Wiley-Liss, Inc.
Key words: matrotrophy; transfer; nutrient; diffusion; gas;
post-ovulatory follicles; calyces nutriciae
Viviparous reproduction has evolved in an estimated
2–3% of the known teleost species (Wourms and Lombardi, 1992). The evolution of a viviparous reproduction
strategy in teleosts included several adaptations: a shift
from external to internal fertilization, the ability to
retain embryos in the female reproduction system, the
utilization of the ovary as a site of gestation, and modifications of the endocrine reproductive control mechanisms as well as structural and functional modifications
of the embryo and the female reproductive system (see
Wourms and Lombardi, 1992). Viviparous teleosts generÓ 2007 WILEY-LISS, INC.
Grant sponsor: The Carlsberg Foundation; Grant sponsor:
The Danish Research Agency.
*Correspondence to: Peter Vilhelm Skov, University of Copenhagen, Marine Biological Laboratory, Strandpromenaden 5, DK3000 Helsingør, Denmark. Fax: 45-3532-1951.
E-mail: [email protected]
Received 12 December 2006; Accepted 8 August 2007
DOI 10.1002/ar.20605
Published online 30 October 2007 in Wiley InterScience (www.
interscience.wiley.com).
MATERNAL–EMBRYONIC EXCHANGE IN Z. VIVIPARUS
ally have a single fused ovary that undergoes significant
changes during development of the embryos in species
with intraluminal gestation. The ovary wall consists of a
layer of smooth musculature beneath the peritoneal covering, while the inner germinal epithelium often possesses folds, increasing the surface area, and becomes
hypervascularized during gestation (Soin, 1968; Schindler and Hamlett, 1993).
The European eelpout Zoarces viviparus is a common
inhabitant of the Eastern North Atlantic region ranging
from the coastal areas of the North Sea along the Norwegian coast to the White Sea as well as the Baltic Sea
(Andriashev, 1986). The reproductive strategy of
Z. viviparus can be characterized as viviparous with internal fertilization and subsequent intraluminal matrotrophic gestation of the embryos (Soin, 1968; Kristoffersson et al., 1973, Korsgaard, 1983; see Rasmussen et al.,
2006, for an excellent overview of embryo development).
During oocyte maturation and growth, the follicle surrounding the oocyte will grow from the ovary wall into
the ovary lumen to form richly vascularized structures
(Soin, 1968). After release of the ovulated eggs, the follicles do not undergo atresia but remain as epithelial
structures maintaining a highly vascularized network in
their walls and adding significantly to the inner surface
of the ovary wall (Kristoffersson et al., 1973). Fertilization is internal and takes place shortly after ovulation,
while the eggs lie freely in the ovary lumen. Hatching
takes place approximately 3 weeks after fertilization
(Rasmussen et al., 2006). The newly hatched embryos
obtain nutrition from the large yolk-sac which they
absorb during the first month (Kristoffersson et al.,
1973, Korsgaard, 1983). The remaining 3–4 months of
gestation are matrotrophic, where the embryos depend
on transfer of gasses, nutrients, electrolytes, and waste
removal by means of the ovarian fluid (Kristoffersson
et al., 1973, Korsgaard, 1983, 1994; Schindler and Hamlett, 1993). There appears to be no direct maternal–embryonic connection, nor is this transfer facilitated by anatomical modifications of the exterior anatomy of the
embryo. However, the embryos do develop an enlarged
hindgut and evidence for nutrient uptake by this route
was presented by Kristoffersson et al. (1973). The presence of a mobile gill and jaw apparatus at early embryonic stages might function to ensure circulation of ovarian fluid over the gills and into the gut by means of the
mouth (Soin, 1968; Kristoffersson et al., 1973).
Transfer of low molecular weight substances like inorganic ions, amino acids, sugars, and free fatty acids
from the maternal circulation to the ovarian cavity has
been reported in Z. viviparus (Kristoffersson et al., 1973;
Korsgaard, 1983, 1987; Korsgaard and Andersen, 1985).
Korsgaard (1987) concluded that the ovarian fluid is not
a static pool, but undergoes circulation and exchange of
metabolites with the maternal organism already in early
pregnancy. Several routes for transfer of nutrition from
the maternal organism to the ovarian fluid of viviparous
matrotrophic species have been suggested: through the
dense vascular network in follicles and the ovary wall,
through secretory cells in the ovarian wall, or through
enrichment from unfertilized eggs and the possible
resorption of unviable embryos (Turner, 1947; Soin,
1968; Kristoffersson et al., 1973; deVlaming et al., 1983;
Korsgaard, 1983). Analysis of maternal serum and ovarian fluid strongly indicate that inorganic ions and low
1501
molecular substances readily penetrate the ovarian wall
and that transfer of soluble nutrients to the ovarian
fluid probably occurs in a manner similar to the transfer
between blood and other extracellular body fluid compartments (Kristoffersson et al., 1973; Korsgaard, 1983).
In contrast, protein content in Z. viviparus ovarian fluid
has been reported to be very low, if present at all. Korsgaard (1983) found no evidence of protein in ovarian
fluid, while protein composition and content in follicular
fluid and maternal serum were nearly identical. Kristoffersson et al. (1973) found very low concentrations of
protein in the ovarian fluid, but suggested this observation was a result of influx of maternal blood due to mechanical stress on the ovarian epithelium exerted by
developing embryos.
Recent observations have demonstrated the presence
of small antifreeze proteins in ovarian fluid (Sørensen
and Ramlov, 2002). The antifreeze protein present in
Z. viviparus is a 7-kDa globular protein (Sørensen et al.,
2006) and is found in similar concentrations in serum of
the maternal organism and the embryos during winter
while lower concentrations are found in the ovarian fluid
(Sørensen and Ramlov, 2002). In addition, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) analysis of ovarian fluid showed the presence of a
few protein bands smaller than 15 kDa, with a different
protein composition than serum from both the maternal
organism and the embryos. This finding indicates that
the presence of proteins is not due to rupture of maternal
vessels, but rather a selective transport of proteins across
blood vessels, as has also been reported from other viviparous teleosts (Schindler and Hamlett, 1993).
The aim of the present study was to perform an anatomical and ultrastructural investigation of the ovarian
and follicular vasculature of Zoarces viviparus, based on
histological sections examined by light- and electron microscopy and vascular corrosion casts to evaluate the
structural basis for exchange between maternal plasma
and the ovarian fluid.
MATERIALS AND METHODS
Gravid Zoarces viviparus were caught using eel traps
set in the southern part of Roskilde Fjord (558400 N,
128040 E) in December and January. Fish were transported to the laboratory under aeration, where they
were processed for experiments on the day they were
caught. A total of 11 animals were used, ranging from
20–23 cm spinal length, and 75–105 g body mass.
Vascular Casting
Fish were anesthetized in an overdose of benzocaine
(0.12 g L21) until they were unresponsive to tactile stimulation. The heart was exposed by a small midline incision between the pectoral fins, and silk suture guided
underneath the junction between the ventricle and bulbus arteriosus. A length of flared polyethylene tubing
(o.d. 0.965 mm) connected to a reservoir of physiological
saline (0.9% NaCl) was inserted into the ventral aorta
by means of the ventricle, and the suture tightened.
Fish were flushed with 50–100 ml heparinized (10 IU
ml21) saline, and subsequently filled with 3–5 ml Mercox, diluted 4:1 with methacrylate containing 5 mg catalyst ml21, using moderate hand-pressure. Fish were
1502
SKOV ET AL.
stored overnight at room temperature to allow polymerization of the resin, after which tissues were digested in
one or more changes of 20% (w:v) KOH. Casts were
rinsed first in 5% HNO3 then distilled water, after which
they were allowed to air-dry. Relevant pieces of the casts
were dissected free under a dissection microscope (Leica
MZ125) and mounted on aluminum stubs using doublesided carbon tape for scanning electron microscopy
(SEM). Preparations were platinum coated, and viewed
in a JEOL JSM-6335F field emission scanning electron
microscope (JEOL, Tokyo, Japan).
Histology
For histological preparations, animals were cannulated through the heart following the procedure
described above. The circulatory system was flushed
with heparinized saline until effluent from the heart
was clear, and then perfusion fixed with 2% paraformaldehyde (PFA) in 0.1 M phosphate buffered saline (PBS).
After perfusion, the ovary was removed and post-fixed in
4% PFA in 0.1 M PBS overnight at 48C. Relevant sections were dissected free and dehydrated in alcohol (70%
1 hr, 95% 1 hr, and 100% 3 3 1 hr) and infiltrated in a
1:1 solution of 100% alcohol and histological resin (Technovit 7100 solution 1, Heraeus Kulzer GmbH and Co,
Wehrheim, Germany) overnight at 48C. After this process, tissues were infiltrated for several days in pure
Technovit solution 1, before being embedded in Technovit solution 2 at 48C. Blocks were mounted on Histoblocs
(Heraeus) and cut on a Leica 1516 microtome (Ernst
Leitz Wetzlar GmbH, Germany). Sections were put into
distilled water and collected on SuperFrost slides. Slides
were allowed to air-dry overnight at room temperature,
stained with toluidine blue for 1min, rinsed in running
tap water, air-dried, and cover-slipped using DPX.
Transmission Electron Microscopy
Fish were perfusion fixed as described above, using
2.5% glutaraldehyde in 0.1 M PBS. Ovaries were
removed and stored in 2.5% glutaraldehyde in 0.1 M
PBS at 48C until processed for transmission electron microscopy (TEM). After isolation of suitable specimen
blocks, samples were rinsed three times in 0.15 M sodium cacodylate buffer (pH 7.2) and subsequently postfixed in 1% OsO4 in 0.12 M sodium cacodylate buffer
(pH 7.2) for 2 hr. The specimens were dehydrated in
graded series of ethanol, transferred to propylene oxide,
and embedded in Epon according to standard procedures. Ultrathin sections were cut with a Reichert-Jung
Ultracut E microtome and collected on 200-mesh copper
grids with Formvar supporting membranes. The sections
were stained with uranyl acetate and lead citrate and
examined with a Philips CM 100 transmission electron
microscope, operated at an accelerating voltage of 80 kV
and equipped with a SIS MegaView2 camera. Digital
images were recorded with the analySIS software package.
Image Analysis
Determination of capillary area was performed from
scanning electron micrographs (n 5 14). Image brightness threshold was adjusted to 110 producing pure black
and white images with a sharp distinction between blood
vessels and background, and analyzed using computer
software (SigmaScan / Image v. 1.02, Jandel Scientific).
SDS-PAGE
Protein distribution in serum (n 5 2) and ovarian
fluid (n 5 5) was visualized by SDS-PAGE in accordance
with Laemmli (1970). A total of 10 ml of serum (103
dilution) and 7 mg of protein from ovarian fluid determined by the Bio-Rad Protein Assay were loaded onto a
Tris-glycine 12% acrylamide gel in a mini-Protean 3 system (Bio-Rad). The gel was run at fixed voltage (200 V)
for 60 min before being stained in Coomassie blue R-250
for 45 min and destained overnight in 10% acetic acid/
30% methanol. Molecular weights of the visualized proteins were estimated from Bio-Rad SDS-PAGE low-range
markers using a Bio-Rad GS-710 gel scanner.
RESULTS
Vascular casts of Z. viviparus produced three-dimensional self-supporting replicas of the vascular system of
the ovary and associated structures (Fig. 1A). The ovary
was heavily vascularized with primary blood supply
being derived from the aorta ovarica, which rapidly divided across the surface of the ovary wall (Fig. 1B), to
feed numerous postovulatory follicular appendages (subsequently termed follicle). These projected into the
lumen of the ovary (Fig. 1A,C) and were typically clubshaped, 1–2 mm wide and 8–10 mm in length, widening
toward the distal portion (Fig. 2A). The arterial vessels
that supplied the follicles were thick-walled and muscular running within the core of the organ (Fig. 2B). Toward the distal end of the follicle, repeated bifurcation
of these arteries gave rise to dense capillary beds (Fig.
3A). Distal views of the vascular casts showed the presence of a cone-shaped cavity at the tip of the follicle
(the location of the oocyte before ovulation), where
sheets of capillaries lay in parallel array (Fig. 3A). On
the exterior of the follicle, capillaries covered 69.0 6
3.4% (mean 6 SD; n 5 14) of the surface area and did
not appear to decrease in density toward the base (Fig.
3B). Here, vessels were continuous with the capillary
beds of the ovarian wall, and eventually drained into
the intestinal and postcardinal venous systems. Capillaries of the follicles had vessel diameters ranging
between 10 and 25 mm, with a diffusion distance of 6.5
6 1.6 mm (n 5 8). This size was slightly smaller than
those of the ovary wall that were typically 20–30 mm in
diameter, with a diffusion distance across the vessel
wall of 5.0 6 2.0 mm (n 5 12).
The tunica intima of the blood vessels from the follicles and ovary wall was formed by a continuous layer
of endothelial cells connected by junctional complexes.
Anchoring junctions displayed long and well-developed
desmosomes heavily invested with keratin filaments, but
only occasionally (10–20%) contained tight junctions
(Fig. 4B). Membrane bound vesicular profiles were present in abundance throughout endothelial cell bodies and
could frequently be seen in open connection with the
luminal or abluminal surface (Fig. 4A). Vesicles were
frequently fused to form vesicular–vacuolar–organelles
(VVO; Fig. 4A). Vesicles and VVOs frequently contained
electron dense ligands (Fig. 4B), presumably protein.
Endothelial cell cytoplasm contained mitochondria (Fig.
MATERNAL–EMBRYONIC EXCHANGE IN Z. VIVIPARUS
Fig. 1. A: Photograph showing the lateral view of a vascular cast
of the ovary (O). L, liver; G, gut; I, intestine; R, renal tissue. The postovulatory follicles (Calyces nutriciae) are denoted by an asterisk. B: Dorsal view of the ovary cast dissected free of surrounding organs show
the dense vascular network of the ovary wall as well as the aorta ovarica (ao) and vena ovarica (vo). C: Photograph of near term embryos
within the ovary. The ovary vasculature has been perfused with Evans
blue to visualize calyces nutriciae (*). Note the numerous differently
sized oocytes still embedded in the ovary wall. Scale bar 5 1 cm.
4C), Golgi apparatus with some budding vesicles (Fig.
4D), rough endoplasmic reticulum with free ribosomes,
multivesicular bodies containing vacuolar membranes,
and secretory granules (Fig. 4C,D). The endothelial cell
layer of the capillary vessels was separated from the
inner ovarian epithelium by a thin layer of loose connective tissue (Fig. 5A). The inner ovarian epithelium was
lined by a layer of epithelial cells connected by junctional complexes (Fig. 5B). Junctions between epithelial
cells of the ovary wall were similar to those of the endothelial cells, although tight junctions were less frequent,
but always contained a well-defined macula adherens
with desmosomes and large investments of keratin filament (Fig. 5B). Golgi complex, rough endoplasmic reticulum, and ribosomes, as well as mitochondria, secretory
granula, and vacuoles could be observed in epithelial
cells. For both cell types, mitochondria were concen-
1503
Fig. 2. A: Scanning electron microscope micrograph of a vascular
cast of a follicle. The capillary beds of the surface (c) are supplied by
several large bore central arteries (CA) running within the core of the
calyx. B: Light micrograph of a section through the basal region of a
follicle. The central arteries (CA) are thick-walled and muscular, presumably to prevent premature diffusion of gasses from the blood
before arrival at the capillaries (c). Scale bar 5 1 mm in A, 50 mm in B
(4-mm section, toluidine blue).
trated in regions around the cell nucleus, while Golgi
apparatus and associated structures were dispersed
throughout the cell body.
In the central region of the follicle, large interstitial
spaces containing the follicular fluid could be observed,
while blood vessels lining the ovary wall were closely
associated with the ovarian smooth muscle layer. The
large blood vessels destined for the calyces had a larger
diameter and were surrounded by several layers of
smooth muscle cells. The epithelial lining of the ovary
wall contained a moderate number of ciliated cells.
Total protein content in ovarian fluid from gravid
Z. viviparus varied from 0.9–1.6 mg ml21. The distribution of protein in ovarian fluid showed a different size
pattern as compared with maternal serum. Serum pro-
1504
SKOV ET AL.
Fig. 3. A: Scanning electron microscope (SEM) micrograph of the
distal portion of a vascular cast of a follicle. The location of the oocyte
before ovulation (*) is characterized by a cone-shaped depression
where capillary beds initially lie in parallel array. B: SEM micrograph
giving a lateral view of the mid- and distal portion of a vascular cast
of a follicle. Capillary organization is extremely dense throughout the
length of the follicle. Scale bar 5 100 mm in A, 1 mm in B.
teins in maternal serum were present throughout the
investigated range from 14–97 kDa, whereas the majority of protein from ovarian fluid was present in two distinct bands at approximately 14 and 32 kDa, respectively (Fig. 6).
DISCUSSION
The thick walls of the central arteries within the postovulatory follicles ensure that blood delivered to the tip
of the follicle remains well-oxygenated. The advantage of
such a vascular arrangement was already recognized by
Wourms et al. (1988). The frequent, repeated bifurcation
of blood vessels toward the tip of the follicle gives rise to
an intricate network of capillaries surrounding the follicles, which occupy the majority of their surface area.
With the presence of several hundred follicles on the
ovarian wall, each having an estimated surface area of
50 mm2, this correlates to an exchange surface of
35 cm2 in addition to the wall of the ovary. In comparison, the gill surface area of a similar sized Z. viviparus
is approximately 400 cm2 (Oliva, 1960). The most significant physiological barriers during the 4- to 5-month gestation period of Z. viviparus, are recognized as the
adequate transfer of oxygen and nutrients to the ovarian
fluid, a process that must be facilitated by maternal
structures (Wourms, 1994; Wourms et al., 1988). The
arrangement of the ovarian blood vessels provides
strong evidence that the postovulatory follicles are the
primary site of exchange between maternal plasma and
ovarian fluid in the eelpout.
Oxygen consumption in near-term eelpout embryos
has been measured at 4.2 mmol g21 h21 at 108C (Korsgaard and Andersen, 1985) and 3.8–4.4 mmol g21 h21 at
58C (Broberg and Kristoffersson, 1983). Measurements
of oxygen consumption in gravid females are not available, but adult nongravid eelpouts have a resting oxygen
consumption of 1.2 mmol g21 h21 at 118C (Van Dijk
et al., 1999; Zakhartsev et al., 2003). When correcting
for body mass, these values are nearly identical. From
the value reported by Korsgaard and Andersen (1985),
and assuming a brood of approximately 200 young
weighing 240 mg each, total oxygen consumption in
near-term embryos would account for 48 mmol h21. Hayakawa and Munehara (2003) reported dissolved oxygen
concentrations in the ovarian fluid of gravid Z. elongatus
at 108C ranging between 0.6 and 2.5 mmol l21 (0.02–0.08
mg l21) or less than 1% saturation, which is in striking
contrast to the observations by Hartvig and Weber
(1984) who reported 82.5 mmol l21 at 108C (38 torr). Considering that near-term embryos are suspended in only
a limited amount of ovarian fluid, which decreases during the gestation period, neither oxygen concentration
appears to allow eelpout embryos to maintain their metabolic rate. We hypothesize that a pH decreasing mechanism must be present, mediated either by CO2 uptake
from the ovarian fluid or intravascular production of lactic acid, which causes a displacement of the blood-oxygen binding curve. In combination with the vascular
arrangement of the follicles this would ensure a higher
rate of oxygen delivery to the ovarian fluid, but further
work on oxygen transfer to the ovarian fluid and the
underlying mechanisms in Z. viviparus are warranted.
The majority of junctional complexes between the endothelial cells of the ovary wall and postovulatory follicles of the eelpout were adherens type junctions. Thus,
there was no evidence to suggest that passage of low
molecular weight compounds from the maternal serum
to the ovarian fluid is hindered by barrier tissue. This
finding is in agreement with previous studies on viviparous teleosts, in particular the demonstration of matrotrophic transfer fluorescent microspheres (ø 40 nm) in
poeciliid fishes by DeMarais and Oldis (2005), and all
previous studies on live-bearing fishes have come to
the conclusion that proteins are transferred from maternal serum or synthesized in the ovarian epithelium
(Schindler et al., 1988; Schindler and Kujat, 1990). For
unknown reasons, protein is found in very low concentrations in the ovarian fluid of Z. viviparus. Korsgaard
(1983) analyzed the composition of maternal serum, follicular fluid and ovarian fluid from Z. viviparus and
found the latter to be free of proteins and, thus, concluded that Z. viviparus is incapable of maternal transfer to the ovarian fluid. Kristoffersson et al. (1973) was
able to detect protein concentrations of 0–7 mg ml21 in
Z. viviparus ovarian fluid, but suggested this to be a
result of bleeding from the ovary wall caused by mechanical stress exerted by the developing embryos,
rather than matrotrophic transfer. The low protein concentrations reported by Kristoffersson et al. (1973) is in
Fig. 4. Transmission electron microscope micrograph of follicle endothelial cells. A: Endothelial cell junction displaying desmosomes
(Des) and keratin filaments (KF), but no tight junction. Vacuoles (*) can
be seen throughout the cell. B: Occasionally, endothelial cells were
connected by tight junctions (TJ). Ribosomes were present (Rib).
Vacuoles frequently contained electron dense material (arrowheads).
C: Mitochondria (M) and secretory granula (SG) were numerous in
some regions of the endothelial cytoplasm. D: Golgi apparatus (Gol),
multivesicular body (MVB). Scale bars 5 500 nm.
1506
SKOV ET AL.
Fig. 6. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis
of maternal serum and ovarian fluid from gravid Z. viviparus. Lane 1–
2: Maternal serum diluted 1:10 in milliQ water. A total of 10 ml was
added in each lane. Lane 3: Bio-Rad low-range marker (97–14 kDa).
Lane 4–8: Ovarian fluid (n 5 5). A total of 7 mg of protein was loaded
in each lane.
Fig. 5. Transmission electron microscope micrograph from the
ovarian epithelial cells. A: Cytoplasm of the ovarian epithelial cells
contained several mitochondria (M) and secretory granules (SG). Ribosomes (Rib) were dispersed throughout the cytoplasm. Note the short
diffusion distance between the capillary lumen (CL) and the ovarian
lumen (OL). Nucleus (N), keratin filaments (KF), connective tissue (CT).
B: Adherens type cell junction between epithelial cells of the ovary
wall. Note also the presence of two punctum adherens. Epithelial cells
rest on a well-defined basement membrane (BM) overlying a layer of
loose connective tissue (CT). Desmosomes (Des). Scale bars 5 1 mm.
accordance with the findings in the present study and
the lack of protein found in the ovarian fluid by Korsgaard (1983) can most likely be accredited to the fact
that the lowest molecular marker used on the electrophoretic gel was 67 kDa. The SDS-PAGE gel in the present study did not demonstrate any significant presence
of proteins in the ovarian fluid beyond the 67-kDa
marker, but several bands were observed below. Of interest, the protein bands present in maternal serum and
ovarian fluid display differences in both the size classes
present as well as concentration (based on the intensity
of the bands). Vascular endothelial cell junctions do not
usually display keratin filaments (Schnittler, 1998). The
abundance of keratin filament between the endothelial
and epithelial cells, as well as the layer of connective tissue between the two, make the assumption by Kristoffersson et al. (1973) that protein origin in the ovarian
fluid is from blood vessel rupture less plausible. Rather,
this organization is considered adaptive to maintain
structural integrity against mechanical stress from the
movements of embryos within the ovary. The majority of
protein found in ovarian fluid has molecular masses less
than 32 kDa, so it is plausible that larger proteins are
restricted from transcapillary passage into the ovarian
fluid.
The presence of ribosomes associated with the rough
endoplasmic reticulum and a Golgi apparatus in the
ovarian vasculature in the present study suggest the
possibility that low molecular weight proteins can be
synthesized within the endothelial and epithelial cells.
Whether or not these proteins are exocytosed is
unknown. Although budding vesicles from the Golgi
were present, they were not found in an abundance to
suggest large production rates. There were differences
in protein size composition in the ovarian fluid compared
with serum suggesting that some protein is exocytosed
or that protein is modified during passage from the
maternal serum to the ovarian fluid. deVlaming et al.
(1983) reported differences in protein compositions in
the embiotocid species Cymatogaster aggregata, which
the authors concluded was the result of ovary wall protein synthesis. The high number of membrane bound
vesicles within the endothelial cells, demonstrate the
presence of a possible transcytotic pathway. A vacuolar
transport mechanism has been demonstrated in the rete
mirabile of the eel swim bladder (Wagner and Kachar,
1995; Bendayan and Rasio, 1996, 1997). Immunocytochemistry and perfusion studies of the eel swim bladder
have demonstrated that albumin and insulin are transported across the capillary wall by means of this vesicular system (Bendayan and Rasio, 1996), but similar
studies are required to further determine the nature of
the compounds observed within the vacuoles from the
present study.
In summary, the ovarian follicles are the first structures to receive blood from the ovarian arterial system,
and their intricate vascular arrangement present strong
evidence that these structures are the primary site of
gas exchange between the maternal blood and the ovarian fluid. The presence of any biochemical specializations that facilitate oxygen unloading from maternal
blood at this site deserves further attention. The endothelial cells of the follicular vasculature predominantly
MATERNAL–EMBRYONIC EXCHANGE IN Z. VIVIPARUS
display junctional complexes lacking a zonula occludens,
allowing for paracellular transport of small molecular
compounds. As Golgi apparatus and rough endoplasmic
reticulum are present, the prerequisites for protein synthesis are present, but most likely are used in both cell
turnover and exocytosed. This should be determined
using immunocytochemical techniques. The present findings from the SDS-PAGE gel showed the presence of a
protein size class not present in the maternal serum,
suggesting it was either modified during transcytotic
passage, produced in the ovarian lining, or secreted by
the developing embryos.
ACKNOWLEDGMENTS
The authors thank Dr. Klaus Qvortrup, Institute of
Medical Anatomy, University of Copenhagen for assistance with TEM work and interpretation. Dr. Moise
Bendayan, University of Montreal, is gratefully acknowledged for his comments on cell structures. P.V.S. was
funded by The Carlsberg Foundation, and J.F.S. was
funded by The Danish Research Agency.
LITERATURE CITED
Andriashev AP. 1986. Zoarcidae. In: Whitehead PJP, Bauchot M-L,
Hureau J-C, Nielsen J, Tortonese E, editors. Fishes of the northeastern Atlantic and the Mediterranean. Vol. 3. Paris: UNESCO.
p 1130–1150.
Bendayan M, Rasio EA. 1996. Transport of insulin and albumin by
the microvascular endothelium of the rete mirabile. J Cell Sci
109:1857–1864.
Bendayan M, Rasio EA. 1997. Evidence of a tubular system for
transendothelial transport in arterial capillaries of the rete mirabile. J Histochem Cytochem 45:1365–1378.
Broberg S, Kristoffersson R. 1983. Oxygen-consumption and lactate
accumulation in the intra-ovarian embryos and young of the
viviparous fish, Zoarces viviparus (L) in relation to decreasing
water O-2 concentration. Ann Zool Fenn 20:301–306.
DeMarais A, Oldis D. 2005. Matrotrophic transfer of fluorescent
microspheres in poeciliid fishes. Copeia 3:632–636.
deVlaming V, Baltz D, Anderson S, Fitzgerald R, Delahunty G,
Barkley M. 1983. Aspects of embryo nutrition and excretion
among viviparous embiotocid teleosts – potential endocrine
involvements. Comp Biochem Phys A 76:189–198.
Hartvig M, Weber RE. 1984. Blood adaptations for maternal-fetal
oxygen transfer in the viviparous teleost, Zoarces viviparus L. In:
Seymour R, editor. Respiration and metabolism of embryonic
vertebrates. Dordrecht: Dr W Junk Publishers. p 17–30.
Hayakawa Y, Munehara H. 2003. Comparison of ovarian functions
for keeping embryos by measurement of dissolved oxygen concentrations in ovaries of copulatory and non-copulatory oviparous fishes and viviparous fishes. J Exp Mar Biol Ecol 295:245–
255.
Korsgaard B. 1983. The chemical composition of follicular and ovarian fluids of the pregnant blenny (Zoarces viviparus (L.)). Can J
Zool 61:1101–1108.
1507
Korsgaard B. 1987. Assimilation and metabolism of exogenous glucose by yolk-sac embryos after intraovarian preincubation and in
vitro in Zoarces viviparus. Can J Zool 65:1201–1205.
Korsgaard B. 1994. Proteins and amino acids in maternal-embryonic trophic relationships in vivparous teleost fishes. Isr J Zool
40:417–429.
Korsgaard B, Andersen FO. 1985. Embryonic nutrition, growth and
energetics in Zoarces viviparus L as indication of a maternal-fetal
trophic relationship. J Comp Physiol B 155:437–444.
Kristoffersson R, Broberg S, Pekkarinen M. 1973. Histology and
physiology of embryotrophe formation, embryonic nutrition and
growth in the eel-pout, Zoarces viviparus (L.). Ann Zool Fenn
10:467–477.
Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 277:680–687.
Oliva O. 1960. The respiratory surface of the gills of the viviparous
blenny (Zoarces viviparus). Acta Biol Crac 3:71–89.
Rasmussen TH, Jespersen A, Korsgaard B. 2006. Gonadal morphogenesis and sex differentiation in intraovarian embryos of the
viviparous fish Zoarces viviparus (Teleostei, Perciformes, Zoarcidae): a histological and ultrastructural study. J Morphol 267:
1032–1047.
Schindler JF, Hamlett WC. 1993. Maternal embryonic relations in
viviparous teleosts. J Exp Zool 266:378–393.
Schindler JF, Kujat R. 1990. Structure and function of the trophotaenial placenta in the goodeid teleost, Skiffia bilineata. Z Mikrosk Anat Forsc 104:241–257.
Schindler JF, Kujat R, Devries U. 1988. Maternal-embryonic relationships in the goodeid teleost, Xenoophorus captivus – the internal ovarian epithelium and the embryotrophic liquid. Cell Tissue
Res 254:177–182.
Schnittler H-J. 1998. Structural and functional aspects of intercellular
junctions in vascular endothelium. Basic Res Cardiol 93:30–39.
Soin SG. 1968. Some features in the development of the blenny
[Zoarces viviparus (L.)] in relation to viviparity. Probl Ichth
8:222–229.
Sørensen TF, Ramlov H. 2002. Maternal-fetal relations in antifreeze
production in the eelpout Zoarces viviparus. Cryoletters 23:183–190.
Sørensen TF, Chen C-HC, Ramlov H. 2006. Isolation and some characterisation of antifeeze protein from the European eelpout Zoarces viviparus. Cryoletters 27:387–399.
Turner CL. 1947. Viviparity in teleost fishes. Sci Monthly 65:508–
518.
Van Dijk PLM, Tesch C, Hardewig I, Portner HO. 1999. Physiological disturbances at critically high temperatures: a comparison
between stenothermal Antarctic and eurythermal temperate eelpouts (Zoarcidae). J Exp Biol 202:3611–3621.
Wagner R, Kachar B. 1995. Linear gap and tight junctional assemblies between capillary endothelial-cells in the eel rete-mirabile.
Anat Rec 242:545–552.
Wourms JP. 1994. The challenges of piscine viviparity. Isr J Zool
40:551–568.
Wourms JP, Lombardi J. 1992. Reflections on the evolution of piscine viviparity. Am Zool 32:276–293.
Wourms JP, Grove BD, Lombardi J. 1988. The maternal-embryonic
relationship in viviparous fishes. In: Hoar WS, Randall DJ,
editors. The physiology of developing fish: viviparity and posthatching juveniles. New York: Academic Press. p 1–134.
Zakhartsev MV, De Wachter B, Sartoris FJ, Portner HO, Blust R.
2003. Thermal physiology of the common eelpout (Zoarces viviparus). J Comp Physiol B 173:365–378.