Human Reproduction vol.12 no.10 pp.2297–2302, 1997
Remodelling of guinea-pig aorta during pregnancy:
selective alteration of endothelial cells
Sofija Jovanović1,3,4 and Aleksandar Jovanović2,3
1Department of Anatomy, Veterinary Faculty, Belgrade, Yugoslavia
and 2Department of Clinical Pharmacology, Pharmacology and
Toxicology, Medical School, Belgrade, Yugoslavia
3Present
address: Division of Cardiovascular Diseases and Clinical
Pharmacology Unit (G-7), Departments of Medicine and
Pharmacology, Mayo Clinic and Foundation, Rochester, MN 55905,
USA
4To
whom correspondence should be addressed
It is known that aortic blood flow is increased during
pregnancy, which may be due to a pregnancy-associated
decrease in aorta sensitivity to vasoconstrictors on one side,
and increased response to vasodilators on the other. Recent
studies have shown that alteration of blood flow or pressure
could remodel some arteries over a short time frame.
However, the possibility of remodelling of aorta during
pregnancy has not yet been examined. Therefore, the aim
of the present study was to assess the morphometric and
stereological characteristics of guinea-pig aorta during
different stages of pregnancy (non-pregnant, early-pregnant, mid-pregnant, late-pregnant, n 5 8–10 for each
group). The cross-sectional areas of different aortic layers
and of endothelial cells were measured using both light and
electron microscopy. The values of external and internal
diameters, wall thickness, total cross-sectional area and
cross-sectional area of media and adventitia were not
significantly different, regardless of the stage of pregnancy.
In contrast, the cross-sectional area of intima significantly
and progressively decreased during pregnancy (non-pregnant: 61 6 53104 µm2, late-pregnant: 38 6 3 µm2, P ,
0.01). The volume:surface density ratio of intima also
significantly and progressively decreased during pregnancy
(non-pregnant: 5.31 6 0.51, late-pregnant: 4.38 6 0.42,
P , 0.01). Electron microscopy revealed that the crosssectional area of endothelial cells was significantly
decreased during different stages of pregnancy (non-pregnant: 56.8 6 6.2 µm2, late-pregnant: 28.9 6 3.8 µm2, P ,
0.01). It is concluded that during pregnancy there is
selective thinning of intimal layer of guinea-pig aorta, which
probably reflects hypotrophy of aortic endothelial cells.
Key words: aorta/endothelium/pregnancy/remodelling/stereology
Introduction
Cardiac output is increased during pregnancy, which is associated with increased rates of perfusion to both reproductive
and non-reproductive organs (Rosenfeld, 1977). It has been
© European Society for Human Reproduction and Embryology
postulated that this vascular adaptation of the maternal organism to pregnancy may be due to a blunted vascular response
to vasoconstrictors on one side (Weiner et al., 1991), and
increased response to vasodilators on the other (Weiner et al.,
1989), although this concept has been recently contested
(Jovanović et al., 1994a, 1995a,b; Grbović and Jovanović,
1996). Although pregnancy is associated with adjustments in
cardiovascular function (Rosenfeld, 1977; Peeters et al., 1980),
the structure of the vascular system during pregnancy is
generally viewed as being very stable. However, some recent
studies have shown that the circulatory system is capable of
remodelling over a surprisingly short time frame (reviewed by
Angus, 1994). Indeed, such structural reorganization is manifest
whenever functional changes persist for more than a few days
(reviewed by Langille, 1993). It has been demonstrated that
aortic blood flow in both ewes and humans is increased during
pregnancy (Rosenfeld, 1977; Easterling et al., 1991), which
may be due to a pregnancy-associated decrease in aortic
sensitivity to vasoconstrictors (Jansakul and King, 1990;
St Louis and Sicotte, 1992). However, the possibility of
remodelling of the aorta during pregnancy has not been
previously examined.
Therefore, the aim of the present study was to assess
the morphological characteristics of guinea-pig aorta during
different stages of pregnancy. Guinea-pigs were specifically
used in this study, bearing in mind that functional changes in
the cardiovascular system during guinea-pig pregnancy are
apparently similar to those in humans (Jovanović et al.,
1994b,c, 1995b,c,d). Moreover, in order to analyse precisely
the morphology of the aorta at different stages of pregnancy,
we used morphometric and stereological methods. These
comprise the body of mathematical methods for obtaining
numerical information about anatomical structure in terms of
such parameters as volume, surface area, number of components, size of components etc. (Weibel, 1979).
Materials and methods
Non-pregnant female guinea-pigs and early-pregnant (22nd day of
pregnancy), mid-pregnant (44th day of pregnancy) and late-pregnant
(64–66th day of pregnancy) guinea-pigs were used in this study
(8–10 in each group). On the day of experiment, guinea-pigs were
anaesthetized with sodium pentobarbitone (40 mg/100 g, i.p. injection), the thoracic cavity was opened and the descending part of the
thoracic aorta was removed. For light microscopy, the vascular
segments (one segment per animal, 1 cm long) were fixed in 10%
buffered neutral formaldehyde (48 h) and cut into 2 mm long rings.
Subsequently, the specimens were dehydrated through a graded series
of ethanol solutions (70–100%) and embedded in paraffin wax.
Resultant blocks were cut on a Sorval JB-4 microtome at 3 µm.
2297
S.Jovanović and A.Jovanović
Sections were stained with haematoxylin and eosin. Each section was
examined under 310 magnification (Olympus Vanox microscope).
For electron microscopy, aortas from each animal were cut into rings
1 mm long (four rings per aorta), fixed in 3% glutaraldehyde for 24
h and, after rinsing in 0.1 M cacodylate buffer (pH 7.4), postfixed in
2% cacodylate solution of osmium tetroxide (1 h). After fixation and
postfixation, the specimens were dehydrated in increasing concentrations (70–100%) of ethanol, and then passed through propylene oxide
and embedded in Araldite (Lee et al., 1983). Ultrathin sections (50–
70 nm) were cut with a diamond knife. The sections were stained
with uranyl acetate and Reynold’s lead citrate and then examined and
photographed using a Philips EM 400-HMG electron microscope.
The investigation conformed with the ‘Guide for the Care and Use
of Laboratory Animals’ (NIH publication 85-23, revised 1985).
Table I. Values of external and internal diameters, and wall thickness of
guinea-pig aortas at different stages of pregnancy. Values are expressed as
mean 6 SEM (n 5 8–10). The range of values is given in parentheses
Stage of pregnancy
External
(102 µm)
Internal
(102 µm)
Wall thickness
(102 µm)
Non-pregnant
47.6 6 4.9
(38.5–54.8)
49.9 6 5.3
(38.4–56.8)
47.4 6 4.8
(36.6–56.3)
48.7 6 5.2
(37.4–57.8)
37.8 6 3.9
(30.2–44.4)
40.2 6 4.5
(31.1–45.2)
38.8 6 4.1
(29.8–45.2)
38.9 6 4.3
(30.0–44.8)
4.88 6 0.53
(4.10–5.18)
4.86 6 0.54
(3.64–5.92)
4.31 6 0.46
(3.42–5.56)
4.93 6 0.54
(3.68–6.46)
Early-pregnant
Mid-pregnant
Late-pregnant
Morphometric and stereological analysis
Light microscopy
Sections (3 µm thick) that included the entire circumference of each
ring were cut from five different blocks of each animal and viewed
with a light microscope equipped with an ocular micrometer accurate
to 0.01 mm. For each block, five sections were examined. Maximal
and minimal internal and external diameters were determined at
magnification 310, and these values were used to determine the
mean values of internal and external diameters. Wall thickness was
calculated by subtraction of the two diameters. To determine crosssectional areas of the different aortic layers including total crosssectional area, a standard point-counting system, Weibel M42, was
used (Weibel, 1979). In brief, this frame-square test grid is composed
of 21-line segments used to count intersects, and 42-line endpoints.
During point and intersect counting, each series of horizontal lines
was scanned to count all the points and intersects falling on the
various components of the vessels. The cross-sectional areas were
calculated using the following equation (Weibel, 1979; Lee et al.,
1983): Am 5 (Pi/Pt)3Ag, where Am 5 cross-sectional area, Pi 5
number of points falling on specific layer, Pt 5 total points of test
grid and Ag 5 area of the whole test grid. Correction for eccentricity
due to sectioning angle with reference to the long axis of the vessel
was used for all calculations: correction 5 d1/d2 where d1 5 minimal
radius of vessel, d2 5 maximal radius of vessel. Thus, the finite form
of the equation used was Am 5 (Pi/Pt)3Ag3(d1/d2).
To determine cellular hypertrophy and/or hyperplasia of aortic
intima, the volume surface density ratio (V/S) of intimal layer was
calculated using the formula: V/S 5 Pi3Z/43Ii, where Pi 5 total
number of points falling on intima, Ii 5 total number of intersects
with the borders of intima, Z 5 length of one test line from test grid
at 310 magnification.
Electron microscopy
Morphometric analysis was performed using 20 random electron
micrographs (five micrographs per block, four blocks per aorta)
obtained at 32800–4900 magnification. The size of endothelial cells,
defined as the cross-sectional area (Lee et al., 1983), was determined
using standard point-counting system B 100 (Weibel, 1979). This test
grid is composed of a coherent square lattice of lines, and total points
of 100. The intersections of the lines are considered as points for
point counting. During point counting, each series of points was
scanned to count all the points falling on the cross-section of
examined structure.
Cross-sectional area was calculated using the following equation
A 5 Pi3d9 (Weibel, 1979), where A 5 cell cross-sectional area;
Pi 5 number of points falling on cell cross sectional area; d 5
distinct between the nearest points of greed; d9 5 distance between
the nearest points of grid corrected with magnification: d9 5
d/magnification (Weibel, 1979; Lee et al., 1983).
2298
Statistical analysis
For all vessel parameters, separate data measurements of diameters,
wall thickness and cross-sectional profiles of vascular components,
from each animal were respectively pooled and an average value
recorded, so that in analysis each animal contributed only one value
for each parameter. All data were tested by the Kolmogorov Goodness
of Fit test and were found to be normally distributed. The results are
expressed as means 6 SEM with range of values in parentheses; n
refers to the number of animals. One-way analysis of variance
(ANOVA) was used when more than two groups were analysed. A
value of P , 0.05 was considered to be statistically significant.
Results
Light microscopy
The values of external and internal diameters and wall thickness
of aorta from guinea-pigs at different stages of pregnancy are
depicted in Table I. These values were not significantly
different (P . 0.05, n 5 8–10). In addition, the values of total
cross-sectional area and cross-sectional area of media and
adventitia were not significantly different, regardless of the
pregnancy status (Figure 1). In contrast, the cross-sectional
area of intima significantly and progressively decreased during
pregnancy (Figure 2). Since the cross-sectional area of intima
reflects the sum of the cross-sectional areas of endothelial
cells, we postulated that stereological intimal changes may be
due to either changes in endothelial cell number or cell size.
To investigate these possibilities, we calculated the volume/
surface density ratio (V/S) of intima. Similarly to the crosssectional area of intima, V/S values were significantly
decreased during pregnancy (Figure 3). Since both V/S ratio
and intimal cross-sectional area both decreased, it was possible
that thinning of the intimal layer may be due to a decrease in
size of endothelial cells (Lee et al., 1983).
Electron microscopy
To test the hypothesis that the decrease in intimal crosssectional area of guinea pig aorta during pregnancy was due
to endometrial cell hypertrophy, electron micrographs were
obtained at different stages of pregnancy (Figure 4). The values
of endothelial cell cross-sectional area are given in Figure 5.
The cross-sectional area of the endothelial cells significantly
and progressively decreased during pregnancy (non-pregnant:
Pregnancy and remodelling of aorta
Figure 1. The effect of pregnancy on total cross-sectional area and cross-sectional area of lumen, media and adventitia of guinea-pig aorta.
Bars represent mean 6 SEM (n 5 8–10).
Figure 2. The effect of pregnancy on cross-sectional area of intima
of guinea-pig aorta. Bars represent mean 6 SEM (n 5 8–10). *P ,
0.05, **P , 0.01 with respect to non-pregnant animals.
56.8 6 6.1
, 0.01).
µm2
versus late-pregnant: 28.9 6 3.8
µm2,
P
Discussion
Recently, it has been recognized that changes in blood flow
may be associated with alterations in morphology of the
affected blood vessels (Baumbach et al., 1991; reviewed by
Angus, 1994). In several different animal species, including
humans, it has been reported that, during pregnancy, blood
Figure 3. The effect of pregnancy on volume/surface density ratio
of intima of guinea-pig aorta. Bars represent mean 6 SEM (n 5 8–
10). *P , 0.05, **P , 0.01 with respect to non-pregnant animals.
flow through different organs is changed (Peeters et al., 1980;
Easterling et al., 1991; Magness et al., 1991). It has been
postulated that this cardiovascular adaptation may be due to
altered vascular reactivity toward neurotransmitters and humoral factors, leading to general maternal vasodilation (Conrad
et al., 1991; Grbović and Jovanović, 1996). In agreement with
this concept is the finding that the blood flow through the
2299
S.Jovanović and A.Jovanović
Figure 4. Photomicrographs of endothelial cells of non-pregnant (A), early-pregnant (B), mid-pregnant (C) and late-pregnant (D) guinea-pig
aorta. Horizontal bar represents 10 µm.
Figure 5. The effect of pregnancy on endothelial cells crosssectional area of guinea-pig aorta. Bars represent mean 6 SEM
(n 5 8–10). *P , 0.05, **P , 0.01 with respect to non-pregnant
animals.
aorta in ewes is gradually increased during pregnancy, reaching
a maximum value at the time of delivery (Rosenfeld, 1977).
It has been postulated that decreased aortic responsiveness to
vasoconstrictors during pregnancy may be the reason for the
increased blood flow through the aorta (Jansakul and King,
1990; St Louis and Sicotte, 1992). In principle, it has been
shown that changes in blood flow could lead to remodelling
of the aorta (Karr-Dullin et al., 1981; Limas et al., 1982).
However, a study of the effects of pregnancy on aortic structure
2300
failed to demonstrate differences between aortas taken from
non-pregnant and pregnant rats (Awal et al., 1995).
The results from the present study are partly in agreement
with this finding. We observed no changes in the diameters,
wall thickness, total cross-sectional area of aorta, and crosssectional areas of lumen, media and adventitia during pregnancy. However, in contrast to previous findings, we observed
that pregnancy is accompanied by a significant decrease in
aortic intimal cross-sectional area during pregnancy, reaching
a minimum at the time of delivery. This effect of pregnancy
was not previously recognized (Awal et al., 1995). Besides
species differences (rat and guinea pig) between these two
studies, this disagreement may result from the fact that intimal
cross-sectional area represents a minor component of total
cross-sectional area value (determined as 2–4% in the present
study), and the changes in intimal cross-sectional area do not
lead to a significant change in total cross-sectional area of the
aorta. This also could explain why, in the present study,
changes in intimal cross-sectional area did not affect total
cross-sectional area.
In order to determine the nature of intimal cross-sectional
area alterations during pregnancy, we calculated the volume/
surface density ratio. It is known that changes in cross-sectional
area of certain structures that are not accompanied by changes
in volume/surface density ratio suggest changes in the number
of units (cells) that form this structure. On the other hand, if
Pregnancy and remodelling of aorta
the volume/surface density ratio is changed in parallel with
the cross-sectional area of a structure, changes in cell volume
can be predicted (Lee et al., 1983). In the present study, the
intimal volume/surface ratio decreased in parallel with the
cross-sectional area. Therefore, it was likely that the pregnancyinduced decrease in the cross-sectional area of intima reflected
a decrease in the size of the cells that form the intima, i.e.
endothelial cells.
In order to examine this possibility, we determined the
parameters of endothelial cell size (cross-sectional area and
minimal and maximal diameter of endothelial cells). Electron
microscopy revealed that cross-sectional area and minimal
and maximal diameter of endothelial cells significantly and
gradually decreased during pregnancy, confirming our findings
obtained with light microscopy. This effect of pregnancy on
endothelial cells has not been previously reported.
It should be mentioned, however, that all the features
measured in the present study are profiles of cells or walls,
not the cells and walls themselves. In principle, since the
results were not obtained with native (unprepared) tissue, it is
possible that the presented results are not completely reliable,
but reflect the influence of sectioning, fixation or counting
(Weibel, 1979). However, the experiments were performed
according to the established principles of stereology (Weibel,
1979) and the results obtained with light and electron microscopy are in agreement, suggesting that this possibility is
rather unlikely.
At present, it is not possible to define the factors that are
responsible for this selective effect of pregnancy on endothelial
cells. In general, it is postulated that blood-flow remodelling
results primarily in a change in vessel diameter (reviewed by
Langille, 1993). Surprisingly, the diameter of the aorta was
not affected by pregnancy. Also, it has been reported that rat
uterine arterial smooth muscle undergoes hypertrophy and
hyperplasia in pregnancy (Cipolla and Osol, 1994). It has been
postulated that this effect of pregnancy may be a consequence
of oestrogen action (Leiberman et al., 1993). However, it is
unlikely that oestrogens mediate an effect of pregnancy on
aortic endothelial cells since the presence of oestrogen receptors
has not been demonstrated in aorta, in contrast to uterine artery
(Leiberman et al., 1993). Thus, at present, it is not possible
to define the mechanism of pregnancy-induced decrease in
endothelial cell size of the aorta. Nevertheless, this phenomenon suggests that, under certain conditions, endothelial cell
size may be altered, without changes to other vessel layers.
Despite the yet unknown mechanism(s) of endothelial cell
hypotrophy in pregnancy, it is likely that it has important
consequences in the process of vascular maternal adaptation.
It is well established that the vascular endothelium produces
a number of vasoactive substances that are important in
governing vascular tone (reviewed by Vane et al., 1990).
Therefore, it is tempting to speculate that hypotrophy of the
vascular endothelium may result in, or may be the result of,
changes in aortic smooth muscle reactivity, as has been
previously shown (Jansakul and King, 1990; St Louis and
Sicotte, 1992). At present, it is not possible to speculate on
the relation between functional and structural changes of aorta
during pregnancy, and further investigations are necessary to
clarify the mechanism(s) of this type of vascular adaptation.
We can conclude that during pregnancy there is selective
thinning of intimal layer of guinea-pig aorta, which probably
reflects hypotrophy of aortic endothelial cells.
Acknowledgements
We thank Dr Andre Terzic, Mayo Clinic and Foundation, for providing
osmium tetroxide and Araldite. A.J. is a recipient of a Merck Sharp
& Dohme International Fellowship in Clinical Pharmacology.
References
Angus, J.A. (1994) Arteriolar structure and its implications for function in
health and disease. Curr. Opin. Nephrol. Hypertens., 3, 99–106.
Awal, M.A., Matsumoto, M. and Nishinakagawa, H. (1995) Morphometrical
changes of the arterial walls of main arteries from heart to the abdomino–
inguinal mammary glands of rat from virgin through pregnancy, lactation
and post weaning. J. Vet. Med. Sci., 57, 251–256.
Baumbach, G.L., Siems, J.E. and Heistad, D.D. (1991) Effects of local
reduction in pressure on distensibility and composition of cerebral arterioles.
Circ. Res., 68, 338–351.
Cipolla, M. and Osol, G., (1994) Hypertrophic and hyperplastic effects of
pregnancy on the rat uterine arterial wall. Am. J. Obstet. Gynecol., 171,
805–811.
Conrad, K.P., Barrera, S.A., Friedman, P.A. and Schmidt, V.M. (1991) Evidence
for attenuation of myo-inositol uptake, phosphoinositide turnover and
inositol phosphate production in aortic vasculature of rats during pregnancy.
J. Clin. Invest., 87, 1700–1709.
Grbović, L. and Jovanović, A. (1996) Effect of the vascular endothelium on
contractions induced by prostaglandin F2α in isolated pregnant guinea pig
uterine artery. Hum. Reprod., 11, 2041–2047.
Easterling, T.R., Benedetti, T.J., Schmucker, B.C. et al. (1991) Maternal
hemodynamics and aortic diameter in normal and hypertensive pregnancies.
Obstet. Gynecol., 78, 1073–1077.
Jansakul, C. and King, R.G. (1990) Effects of pregnancy and endothelial cell
removal on alpha-adrenoceptor-mediated responses of rat thoracic aortae.
J. Auton. Pharmacol., 10, 353–362.
Jovanović, A., Grbović, L., Drekić, D. and Novaković, S. (1994a) Muscarinic
receptor function in the guinea-pig uterine artery is not altered during
pregnancy. Eur. J. Pharmacol., 258, 185–194.
Jovanović, A., Grbović, L. and Tulić, I. (1994b) Predominant role for nitric
oxide in the relaxation induced by acetylcholine in human uterine artery.
Hum. Reprod., 9, 387–393.
Jovanović, A., Grbović, L. and Tulić, I. (1994c) Endothelium-dependent
relaxation in human uterine artery in response to acetylcholine. Eur. J.
Pharmacol., 256, 131–139.
Jovanović, A., Grbović, L. and Jovanović, S. (1995a) Effect of the vascular
endothelium on noradrenaline-induced contractions in non-pregnant and
pregnant guinea-pig uterine arteries. Br. J. Pharmacol., 114, 805–815.
Jovanović, A., Grbović, L. and Jovanović, S. (1995b) K1 channel blockers
do not modify relaxation of guinea-pig uterine artery evoked by
acetylcholine. Eur. J. Pharmacol., 280, 95–100.
Jovanović, A., Grbović, L., Jovanović, S. and Z̆ikić, I. (1995c) Effect of
pregnancy on vasopressin-mediated responses in guinea-pig uterine arteries
with intact and denuded endothelium. Eur. J. Pharmacol., 280, 101–111.
Jovanović, A., Grbović, L., Z̆ikić, I. and Tulić, I. (1995d) Characterization of
arginine vasopressin actions in human uterine artery: lack of role of the
vascular endothelium. Br. J. Pharmacol., 115, 1295–1301.
Karr-Dullien, V., Bloomquist, I.E., Beringer, T. and El-Bermani, A.W.I. (1981)
Arterial morphometry in neonatal and infant spontaneously hypertensive
rats. Blood Vessels, 18, 253–262.
Langille, B.L. (1993) Remodeling of developing and mature arteries:
endothelium, smooth muscle, and matrix. J. Cardiovasc. Pharmacol., 21,
S11–S17.
Lee, R.M.K.W., Forrest, B.J., Garfield, E.R. and Daniel, E.E. (1983)
Ultrastructural changes in mesenteric arteries from spontaneously
hypertensive rats. Blood Vessels, 20, 72–91.
Leiberman, J.R., Wiznitzer, A., Glezerman, M. et al. (1993) Estrogen and
progesterone receptors in the uterine artery of rats during pregnancy. Eur.
J. Obstet. Gynecol. Reprod. Biol., 51, 35–40.
2301
S.Jovanović and A.Jovanović
Limas, C., Westrum, B., Iwai, J. and Limas, J.C. (1982) Aortic morphology
in salt dependent genetic hypertension. Am. J. Pathol., 107, 378–394.
Magness, R.R., Rosenfeld, C.R. and Carr, B.R. (1991) Protein kinase C in
uterine and systemic arteries during ovarian cycle and pregnancy. Am. J.
Physiol., 260, E464–E470.
Peeters, L.L.H., Gruters, G. and Martin, C.R. (1980) Distribution of cardiac
output in the unstressed pregnant guinea-pig. Am. J. Obstet. Gynecol., 138,
1177–1184.
Rosenfeld, R.C. (1977) Distribution of cardiac output in ovine pregnancy.
Am. J. Physiol., 232, H231–H235.
St Louis, J. and Sicotte, B. (1992) Prostaglandin or endothelium-mediated
vasodilation is not involved in the blunted responses of blood vessels to
vasoconstrictors in pregnant rats. Am. J. Obstet. Gynecol., 166, 684–692.
Vane, J.R., Anggerd, E.E. and Botting, R.M. (1990) Regulatory functions of
the vascular endothelium. N. Engl. Med. J., 32, 27–36.
Weibel, E.R. (1979) Stereological Methods. Vol. 1. Practical Methods for
Biological Morphometry. Academic Press, San Francisco.
Weiner, C., Martinez, E., Zhu, L.K. et al. (1989) In vitro release of endotheliumderived relaxing factor by acetylcholine is increased during the guinea pig
pregnancy. Am. J. Obstet. Gynecol., 161, 1599–1605.
Weiner, C., Zhu, L.K., Thompson, L. et al. (1991) Effect of pregnancy on
endothelium and smooth muscle: their role in reduced adrenergic sensitivity.
Am. J. Physiol., 261, H1275–H1283.
Received on April 10, 1997; accepted on July 11, 1997
2302