Author for correspondence: George Osol, Department of Obstetrics and Gynecology, University of Vermont College of Medicine, Burlington, VT, USA (fax +001 802 656 8871, e-mail George.Osol@uvm.edu).
Abstract: Sufficient uteroplacental blood flow is essential for normal pregnancy outcome and is accomplished by the coordinated growth and remodelling of the entire maternal uterine vasculature. The main focus of this MiniReview is to provide information on upstream (pre-placental) maternal uterine vascular remodelling that facilitates gestational increases in uterine blood flow. Consideration of the three-dimensional pattern of remodelling (circumferential enlargement versus axial elongation), changes in vessel biomechanical properties, and underlying mechanisms [shear stress, nitric oxide, vascular endothelial growth factor (VEGF)/placental growth factor (PlGF), the renin–angiotensin system] and pathways (local versus systemic; venoarterial exchange) are provided using the rat as the principal animal model, although findings from other species are incorporated wherever possible to provide a comparative perspective. The process of maternal gestational uterine vascular remodelling involves a number of cellular processes and mechanisms, including trophoblast invasion, hyperplasia and hypertrophy, and changes in extracellular matrix composition. In addition, changes in cellular function, e.g. the secretory and contractile properties of smooth muscle and an up-regulation of endothelial vasodilatory influences may contribute to uteroplacental blood flow increases through changes in tone as well as in structure. Future studies aimed at better understanding the inter-relationship between changes in vessel structure (remodelling) and function (reactivity) would likely generate new mechanistic insights into the fascinating process of maternal gestational uterine vascular adaptation and provide a more physiological perspective of the underlying cellular processes involved in its regulation.
Large increases in uteroplacental blood flow during gestation are essential for normal foetal growth and survival and occur in every mammalian species studied, including the human beings.
In women, and in experimental animals such as rodents and sheep, uteroplacental blood flow increases many-fold above the nonpregnant level, with the majority (usually >90%) being directed to the placenta by term [1–3]. Haemodynamic changes are principally effected by a profound decrease in uterine vascular resistance, which is accomplished through a combination of expansive remodelling (vessel growth) and enhanced vasodilation (functional changes in reactivity). Although this review is focused on the former, in vivo, uterine vascular resistance (and, therefore, blood flow) is ultimately determined by the combination of altered vessel structure and function. As considered below, there is also some evidence for a mechanistic linkage between the two processes.
The clinical importance of uterine vascular remodelling is underscored by the fact that insufficient growth and development of the uteroplacental circulation may lead to placental underperfusion and contribute to the development of significant gestational pathologies such as intrauterine growth restriction and preeclampsia [4,5].
Three-Dimensional Changes in Vascular Structure
Significant increases in both venous and arterial diameter and length occur during gestation in every mammalian species. Figure 1 shows the uterine horns of a nonpregnant (inset) versus late pregnant (day 20/22) rat to illustrate the magnitude of change.
Changes in Arterial Circumferential Dimensions
During pregnancy, the passive (fully dilated) diameter of the human uterine artery is approximately doubled , with similar changes reported in rodents, sheep, pigs and guinea pigs [7–11].
Enlargement in arterial calibre occurs most often with little or no thickening of the vascular wall [7,12], with the one exception being the mouse, in which media thickness increases significantly during the course of gestation . Even without any increase in wall thickness, however, an increase in lumen diameter represents outward hypertrophic remodelling because it results in increased cross-sectional area of the vascular wall and, therefore, wall mass hypertrophy. Measurement of cross-sectional area alone underestimates the actual increase in wall mass, because it ignores the several-fold increases in arterial length (discussed in the next section); when both dimensions are considered, the doubling in lumen diameter and the several-fold increases in arterial length combine to increase wall mass on the order of 400–1000%.
Spatial and Temporal Patterns of Remodelling Based on Vessel Type
While the pattern of remodelling in rodents is consistently outward hypertrophic for both arteries and veins, the spatial and temporal patterns of change in small versus larger vessels may differ. For example, growth of the smaller (e.g. radial) arteries precedes that of the larger vessels, suggesting that remodelling may initially begin in placental and pre-placental vessels and, as pregnancy continues, progresses to more proximal tissues so that changes in the main uterine artery are the last to occur .
There are also differences in the magnitude of change in vessels that feed the placenta versus those that perfuse the uterine corpus (i.e. pre-placental versus pre-myometrial arteries ), with significant additional widening observed in pre-placental vessels. The proximal portion of the vessel remodels by widening and elongating while preserving its functional (contractile) elements; conversely, the distal portion of the pre-placental vessel widens additionally but completely loses its contractile ability, presumably as a result of endovascular trophoblast invasion that destroys the intramural muscular elements (vascular smooth muscle) .
Cellular Hypertrophy and Hyperplasia
As the media occupies the greatest proportion of the wall, luminal enlargement could be accomplished most simply by an increase in vascular smooth muscle cell (VSMC) length (axial hypertrophy), along with a rearrangement of matrix elements. This does indeed appear to be the case. For example, unstressed smooth muscle cell length in arcuate vessels from pregnant versus virgin nonpregnant rats was increased by 20% in rats . In guinea pigs , smooth muscle cell length increased from 21 to 39 μm (86%), along with an increase in cellular thickness from 4.6 to 9.6 μm (108%). Elongation of VSMCs and an increase in protein content (including actin and myosin) have also been reported in other species such as the sheep .
In addition to cellular hypertrophy, there is also strong evidence for gestational activation of hyperplastic processes, because VSMC division is increased significantly in uterine arteries and veins from rats and guinea pigs [13,17–20]. While endothelial hyperplasia has been documented in the rat , changes in endothelial hypertrophy (increased cell area or mass) have not yet been detailed and await further study.
In rodents, significant axial growth of uterine arteries and veins is clearly evident, with a doubling in length having been measured in the main uterine artery and vein in several studies [21,22]. Rat and guinea pig mesometrial (arcuate and radial) vessels elongate as well, increasing three to five times in length [12,23].
A scaled drawing showing the relative extent of circumferential versus axial uterine radial artery remodelling in rodents is shown in fig. 2. Although axial length increases 4× on average, whereas diameter approximately doubles, the haemodynamic effects of widening should nevertheless predominate because the relationship between axial length and resistance is linear (i.e. a four times increase in length would quadruple flow resistance), while that of diameter (or radius) versus resistance is inverse and quadratic (i.e. doubling of diameter theoretically decreases resistance 16 times), as per Poiseiulle’s law (R = ηL/r4, where R = resistance, η = viscosity, L = length and r = inner radius).
In women, as the uterus grows and stretches to accommodate the fetoplacental unit, it is reasonable to expect that increases in arterial length occur, although these have not been documented thus far. Furthermore, it is not clear whether this is accomplished by actual longitudinal growth, stretching of elastic elements, progressive straightening of coiled vessels (especially in the larger uterine arteries that parallel the uterine corpus and are quite tortuous) or a combination thereof.
The mechanism by which arterial elongation occurs is completely unknown, although the approximate doubling of the thickness of VSMCs (such as has been measured in guinea pigs,  suggests that cellular rearrangement may take place, i.e. as cells grow, they may rearrange in a way that lengthens the vessel. Even in the absence of hyperplasia, which clearly occurs, a doubling of cellular width (or cross-sectional area) would – in the absence of any increase in wall thickness – lead to a doubling of axial length as well. Hyperplasia would augment the extent of elongation further. At this time, the specific factors that stimulate cellular hypertrophy or hyperplasia in this physiological setting have not been identified.
Remodelling of Uterine Veins
The process of expansive remodelling is not limited to arteries because the veins also enlarge substantially during pregnancy. For example, the fully distended diameter of the main uterine vein in pregnant versus nonpregnant rats averaged 1576 versus 956 μm, respectively (an increase of 65%), along with a doubling in unstressed length. Increased venous diameter was accompanied by significantly increased endothelial and VSMC mitotic indices, increased distensibility and a reduced elastin content – adaptations that would further enhance venous capacitance . Significant uterine venous enlargement during pregnancy was also reported in the mouse .
Based on these observations, it can be seen that uterine vascular remodelling during pregnancy is a complex adaptive process that is dependent on multiple coordinated and interactive influences. Compared to the changes in vessel dimensions and morphology, our understanding of the underlying mechanisms is much more limited.
Systemic versus Local Influences on the Remodelling Process
Data using different experimental approaches (induction of pseudo-pregnancy, hormone replacement in oophorectomized animals and surgical ligation) support the idea that steroids may initiate the process of uterine vascular adaptation during the earlier stages of pregnancy [9,25,26]. As gestation progresses, local (fetoplacental) influences become predominant, as is clearly shown by the results of rodent studies in which pregnancy was prevented in one of the two uterine horns [24,27]. At term, the nonpregnant horn is relatively small and shows only minor changes in vascular structure, while changes in its contralateral pregnant counterpart are striking (fig. 3). In addition to changes in vascular structure, overall fecundity is preserved such that the implanted horn is occupied by an average of 12–14 pups (twice the normal number in one horn).
Although the nature of local influences is not known, it seems reasonable to speculate that growth factors such as VEGF/PlGF and/or platelet-derived growth factor (PDGF) may be responsible. This raises several questions such as – what stimulates the release of growth factors and what is their origin?
The intercellular matrix is known to be a storage site for growth factors, and their release may be stimulated by steroids and/or haemodynamic forces such as shear stress. Finally, growth factor production or activation may be stimulated by placentally derived signals, and/or myometrial stretch by the fetoplacental unit. If so, venoarterial transfer (discussed below) is one putative pathway for their delivery to the arterial wall.
Oestrogen and Progesterone
An important role for female sex steroid hormones (progesterone and oestrogens) in uteroplacental vascular adaptation during pregnancy has been documented in a number of studies. For example, oestradiol 17β reduces uterine vascular resistance and promotes uterine artery vasodilation  through mechanisms linked to the nitric oxide (NO)/extracellular regulated kinase (ERK) pathway . Oestrogen also stimulates DNA synthesis in uterine radial arteries  and promotes mitotic activity within the uterine artery wall [17,19,31,32]; in one study, growth of VSMCs was shown to be protein kinase C dependent .
Oestrogen also stimulates endometrial secretion of a factor(s) that promotes VSMC migration as an early step in endometrial vessel remodelling  and plays a role in regulating invasion and remodelling of the uterine spiral arteries by extravillous trophoblasts. Low levels of oestrogen during early primate pregnancy are required to permit normal progression of trophoblast vascular invasion, while the surge in oestrogen that occurs during the second trimester has a physiological role in suppressing further arterial trophoblast invasion .
Less is known about the effects of oestrogen (and progesterone) on uterine artery expansive remodelling. The results of one study, in which pseudo-pregnancy (a state in which hormone levels rise without the presence of a foetus or placenta) was induced in the mouse , showed significant expansive remodelling of the main uterine artery 10 days after stimulation, although the extent was significantly less (about 25%) than the amount measured in pregnant animals. This result suggests the existence of synergism between local and systemic factors in the remodelling process.
A role for progesterone rather than oestrogen may also deserve consideration as it stimulates vessel maturation in the mouse endometrium during the menstrual cycle  although, surprisingly, its role in uterine vascular remodelling during pregnancy has not been evaluated.
Nitric Oxide and Shear Stress
Shear stress is a plausible mechanism for uterine arterial enlargement [36,37] during gestation in view of its well-established nature as a physiological mechanism for circumferential arterial growth. During pregnancy, the principal stimulus for shear stress elevation in the maternal uterine circulation derives from the reduction in downstream resistance secondary to hemochorial placentation, which leads to an acceleration of flow velocity and shear stress in upstream vessels . Vasodilation and/or vessel growth would allow the augmented flow to continue, but with a slower velocity, thereby normalizing shear stress in the process.
The molecular mechanism(s) by which shear stress leads to circumferential vessel growth appears to be endothelial in origin, and several recent studies suggest that nitric oxide (NO) plays a key role in this process [11,38–40]. This observation is pertinent to uterine gestational remodelling in view of the well-established up-regulation of endothelial nitric oxide synthase (eNOS) and NO signalling in the pregnant state [22,41] that arises from the confluence of increased oestrogen and growth factor (e.g. VEGF and PlGF) levels combined with elevated endothelial shear stress. Each of these influences augments eNOS activity and NO production in their own right and, during pregnancy, they occur in combination.
Evidence to support the role of NO in uterine gestational remodelling comes from several studies in rodents, e.g. those that showed attenuated remodelling in eNOS (NOS-3) knockout mice , and in rats given a NOS inhibitor during gestation, although axial growth was completely unaffected .
Another mechanism involved in the remodelling process is the renin–angiotensin system, as shown in a study where over-activation of the renin–angiotensin system by a low-sodium diet in pregnant rats significantly attenuated the normal increase in arcuate artery diameter . In some pre-eclamptic women, over-expression of an antibody that stimulates the angiotensin-1 (AT-1) receptor has been documented and is thought to be important in the genesis of hypertension characteristic of this disease, although its influence on uterine vascular remodelling has yet to be examined.
VEGF and PlGF
Members of the VEGF family, especially PlGF, are also appealing candidates for mediating this process because they induce vasodilation, stimulate endothelial mitosis and are associated with hypervascularization and enlargement of existing vessels [43–45]. Importantly, an attenuation of VEGF/PlGF signalling by placental secretion of an excess of their soluble receptor [soluble vascular endothelial growth factor receptor-1 or soluble Fms-related tyrosine kinase-1(sFlt-1)] is well documented in pre-eclamptic women, and over-expression of sFlt-1 in animals results in a gestational syndrome (hypertension, proteinuria and endothelial glomerulosis) that closely mimics pre-eclampsia .
The biomechanical properties of uterine vessels also change significantly during pregnancy with increased distensibility reported in sheep  and guinea pigs [17,46]. In rats, smaller arcuate and radial arteries are also more distensible than their nonpregnant counterparts [12,20] and in the rabbit, premyometrial radial arteries are significantly more distensible than pre-placental vessels .
Most commonly, changes in distensibility are attributed to alterations in extracellular matrix volume, composition and collagen and elastin fibre orientation. Some species differences exist in this regard. In main uterine arteries from sheep  and pigs , collagen content decreased significantly during gestation, with no change in elastin, while elastin content was reduced in spiral arteries of pregnant women  and in rat uterine arteries and veins .
Studies on matrix reorganization and altered composition have established a central role for metalloproteinases (MMPs) . In particular, the activities of MMP-2 and MMP-9 are elevated in vascular tissues during pregnancy . Notably, their activities may be reduced in some gestational pathologies such as pre-eclampsia, affording a mechanism to explain the attenuation in remodelling that has also been documented.
Although the mechanisms and pathways that govern the uterine vasculature remodelling during pregnancy are not well defined, the fact that the uterus and the placenta are highly synthetic organs that release a number of growth and vasoactive factors into the uterine venous effluent led us to suggest the existence of a local mechanism of venoarterial signalling whereby the uterus and placenta may regulate their own perfusion. Here, secreted placental and/or myometrial signals (e.g. growth factors) pass across the uterine venous wall and thereby influence the structure of adjacent uterine arteries.
An anatomical basis for venoarterial transfer does exist in the uterine circulation, because arteries and veins are arranged in close apposition in a number of species [50–52] and in human beings . Physiological support for the existence of venoarterial mechanisms comes from earlier studies on luteolysis, in which prostanoids secreted by the myometrium pass into the uterine veins and are then transferred to the ovarian artery, leading to vasoconstriction and ischaemia that result in the death of the corpus luteum. This venoarterial luteolytic mechanism has been demonstrated in a number of mammalian species . Thus, venoarterial exchange is plausible from both an anatomical and a physiological standpoint.
A recent study from our laboratory validated the existence of venoarterial communication in the rat uterine circulation in vitro by showing that small vasoactive molecules perfused within the uterine vein induce diameter changes in the adjacent uterine artery . Moreover, the venous permeability of isolated rat uterine veins is significantly increased, even to large molecular weight signals (70 kDa), enhanced during pregnancy and is regulated by physical forces such as wall tension and by molecular signals such as VEGF [51,52].
As growth factors such as VEGF and PlGF are also potent vasodilators, this mechanism may provide a short-loop pathway for regulating placental perfusion by inducing changes in arterial tone.
Nevertheless, experimental evidence that conclusively demonstrates (or refutes) the importance of venoarterial transfer as an in vivo mechanism for uterine vascular gestational remodelling is currently lacking, and the determination of whether this mechanism is physiologically important in the setting of maternal uterine vascular remodelling during gestation awaits further research.
As can be seen from this review of the literature, our knowledge about the process of maternal uterine vascular adaptation during pregnancy is heavily weighted towards its descriptive rather than mechanistic aspects. Some of the questions that future research will endeavour to answer include (i) What are the principal molecular signals involved in the remodelling process, and are they abrogated in pathological states such as pre-eclampsia? In this regard, little information is available about the role of prostanoids or leukotrienes in gestational remodelling. Alternatively, the problem may lie in altered tissue responsiveness, rather than signalling per se. (ii) Do systemic elevations in sex steroid concentrations modulate changes in vessel structure and reactivity, and are their actions synergistic with altered local physical forces (shear stress and transmural pressure) that arise from the process of placentation? (iii) Is venoarterial signalling an important short-loop mechanism for regulating placental perfusion via changes in arterial tone and structure? If so, how is the synthesis of vasoactive and growth-promoting signals regulated? (iv) Are changes in structure (remodelling) and function (reactivity) linked and, if so, does one precede the other? For example, by altering wall tension and inducing cell and matrix deformation, does the process of vasodilation initiate a cascade of events in vascular smooth muscle and endothelial cells that stimulates expansive remodelling? There is some basis for this in studies that have shown that vasoconstriction initiates inward remodelling . Is this mechanism bidirectional, i.e., does vasodilation also stimulate outward remodeling? Last, but not least, what are the mechanisms that stimulate vascular smooth muscle hypertrophy versus hyperplasia, and are these processes involved in axial versus circumferential remodelling? As already discussed, while NO and other mechanisms (shear stress and sex steroids) have been implicated in the latter, at this time absolutely nothing is known about the former.
The many-fold increase in uteroplacental blood flow required for normal pregnancy outcome is achieved by a coordinated process of uterine vascular adaptation that involves both vessel growth and vasodilation. A combination of local and systemic molecular mechanisms regulates this process; among them are shear- and steroid-induced up-regulation of endothelial nitric oxide, placental signalling vis-à-vis VEGF/PlGF and activation of the renin–angiotensin system. Although the pathways governing placental signalling of the arterial system are not well understood, venoarterial communication may provide a pathway for the placenta to modulate the structure and tone of afferent arteries directly and, in turn, influence its own perfusion.
The clinical relevance of maternal uterine vascular adaptation during pregnancy is underscored by the fact that its aberrance is associated with several common gestational pathologies, including intrauterine growth restriction, gestational diabetes and pre-eclampsia.
The authors would like to acknowledge RO1 support from the National Institutes of Health (HL79253 and HL73895); GO is an Established Investigator of the American Heart Association.