Venoarterial communication mediates arterial wall shear stress-induced maternal uterine vascular remodeling during pregnancy

Abstract

Although expansive remodeling of the maternal uterine circulation during pregnancy is essential for maintaining uteroplacental perfusion and normal fetal growth, the underlying physiological mechanisms are not well understood. Using a rat model, surgical approaches were used to alter uterine hemodynamics and wall shear stress (WSS) to evaluate the effects of WSS and venoarterial communication (e.g., transfer of placentally derived growth signals from postplacental veins to preplacental arteries) on gestational uterine vascular remodeling. Changes in WSS secondary to ligation of the cervical but not the ovarian end of the main uterine artery and vein provoked significant expansive remodeling at the opposite end of both vessels, but only in pregnant animals. The ≈50% increase in lumen diameter (relative to the contralateral horn) was associated with an upregulation of total endothelial nitric oxide (NO) synthase expression and was abolished by in vivo NO synthase inhibition with N-nitro-l-arginine methyl ester. Complete removal of a venous segment adjacent to the uterine artery to eliminate local venous influences significantly attenuated the WSS-induced remodeling by about one-half (P < 0.05). These findings indicate that, during pregnancy, 1) increased WSS stimulates uterine artery growth via NO signaling and 2) the presence of an adjacent vein is required for arterial remodeling to fully occur.

NEW & NOTEWORTHY This study provides the first in vivo evidence for the importance of venous influences on arterial growth within the uteroplacental circulation.

INTRODUCTION

During pregnancy, increases in maternal blood volume combined with decreases in vascular resistance direct up to 20% of the total cardiac output to the uteroplacental circulation. As a result, uteroplacental blood flow (UPBF) increases more than 10-fold compared with the nonpregnant state in most mammalian species and approaches 1 l/min by term in humans (1, 2, 31). The progressive and substantial increase in UPBF is accomplished by the coordinated growth (expansive remodeling) of the entire uterine circulation (9, 10, 15, 26, 32, 33, 41, 50, 51). This process is essential for maintaining UPBF and assuring healthy pregnancy outcome, since aberrance in uterine vascular adaptation has been associated with gestational complications such as intrauterine growth restriction and preeclampsia (6, 27, 42, 43).

Earlier studies have shown that endothelial nitric oxide (NO) signaling is a key pathway in maternal gestational uterine vascular remodeling (36, 47, 53, 55, 56), although the actual physiological stimulus for increasing NO production is not known. A regulatory role of wall shear stress (WSS) on vascular diameter has been shown in the skeletal muscle microcirculation (20–22) and in myometrial arteries from pregnant women (24). Conversely, decreased WSS lowered NO production in a uterine blood flow occlusion model in pregnant sheep (19). Because endothelial NO synthase (eNOS) has been shown to be acutely and chronically regulated by WSS (34, 39, 55), we hypothesized that increased WSS is the principal physiological stimulus for both enhanced endothelial NO production and expansive remodeling in the uterine circulation during gestation.

Most studies have focused on arterial remodeling, and relatively little is known about venous remodeling and whether venous influences play a role in arterial remodeling during gestation. WSS is known to affect veins as well (4, 14, 21), and venous influences on adjacent arterial structure have been documented in several earlier studies (32, 50). Venoarterial communication has also been documented in the process of luteolysis (28), in which vasoactive molecules (PGF_2α) derived from the endometrium pass into uterine veins and are transferred to the ovarian artery. Its involvement in maternal uterine vascular remodeling during gestation remains hypothetical, however, and was evaluated using a combination of 1) arterial ligation at the cervical end of the main uterine artery (MUA) and 2) complete removal of a segment of main uterine vein (MUV) while leaving the artery intact.

WSS is difficult to assess in the mammalian uterine circulation because of its unique architecture. In rodents and humans, blood flow input to the uterus is normally bidirectional, with inflow from both the ovarian and uterine arteries (which anastomose to form a loop), with the latter normally being predominant (14, 45, 46). Earlier studies have used site-specific arterial ligation to study the pathophysiology of fetal growth retardation (4, 57) and to alter arterial remodeling (38). Surgical approaches such as unilateral oviduct ligation and vascular occlusion have also been used successfully to assess blood flow and WSS (13, 19, 42, 48).

To test the hypothesis that changes in shear stress secondary to placentation are the main physiological stimulus for uterine artery remodeling, we adapted the surgical approach to include unilateral ligation of the MUA and MUV at the cervical versus ovarian end to alter uterine hemodynamics in a way that restricts inflow and outflow to a single site (38, 46). The results indicate that WSS is an important physiological stimulus for uterine vascular remodeling, that the mechanism involves NO, and that normal arterial remodeling requires the presence of an intact and functional adjacent vein.

MATERIALS AND METHODS

Animals.

Nonpregnant (NP) and pregnant female Sprague-Dawley rats (12–14 wk old) were purchased from Charles River Laboratories (Kingston, NY, and Montreal, Quebec, Canada) and housed in the University of Vermont small animal care facility, which is accredited by the American Association for Accreditation of Laboratory Animal Care. On-site surgeries were performed on gestational day 10 in pregnant animals (n = 33, termed LP because most measurements were made late in pregnancy, on day 20/22 of gestation) and in age-matched NP rats (n = 16). Rats were singly housed after surgery, with food and water provided ad libitum. All experiments and procedures were approved by the Institutional Animal Care and Use Committee of the University of Vermont.

Reagents.

All chemicals were purchased from Fisher Scientific (Hampton, NH) unless otherwise specified. The composition of relaxing solution was HEPES physiological salt solution (containing 10 mM HEPES, 141.8 mM NaCl, 4.7 mM KCl, 1.7 mM MgSO₄, 0.5 mM EDTA, 1.2 mM KH₂PO₄, and 5 mM glucose, pH 7.4 without CaCl₂), with the addition of 100 μM papaverine (Sigma, St. Louis, MO) and 10 μM diltiazem (Sigma).

Vascular ligation surgery.

Surgeries were performed on day 10 of pregnancy and in age-matched NP rats. All animals were anesthetized with isoflurane, and a midline transverse incision was made to expose the uterus. The MUA and MUV were ligated with a 6-0 prolene suture (Ethicon, Somerville, NJ) at either the cervical or ovarian end of one randomly selected uterine horn. The opposite horn served as the sham-operated control, since it was exposed to the same surgical manipulation but with the suture removed.

MUA + MUV ligation was expected to alter uterine hemodynamics in the ligated horn in a way that restricted inflow to one site, thereby altering both the directionality and pattern of WSS within the vasculature (Fig. 1, A vs. B). The abdominal fascia was closed using a running suture with 5-0 vicryl thread (Ethicon), and the dermal layer was closed using a simple interrupted suture with 5-0 silk thread (Oasis, Mettawa, IL). Buprenorphine (0.06 mg/kg) was administered for analgesia and to minimize discomfort during recovery.

Fig. 1.

Vascular ligation model in rodents. A: rodents have a duplex uterus with normally bidirectional inflow of blood and considerable architectural redundancy in case of occlusion. The vessels in the mesometrium (arcuate and radial arteries) and uterine corpus are perfused by arterial blood coming from either the cervical or the ovarian end (with a predominance of cervical-end inflow, as shown by the red arrows). B: to create a unidirectional flow, surgical ligations of the main uterine artery (MUA) + main uterine vein (MUV) were placed at the ovarian end (to restrict inflow of blood to the cervical end, as shown in the horn on left) or at the cervical end (to allow blood flow from only the ovarian end, as shown in the horn on right). Black X’s show points of ligation.

MUV removal.

On gestational day 10, unilateral arterial ligations were performed on the MUA at the cervical end followed by the removal of the MUV flanking a section of the MUA at the ovarian end. This created a segment of the MUA that was no longer flanked by a vein and thereby removed any local venous influences on the artery. The contralateral uterine horn served as an internal control. After uterine excision on day 20, main uterine artery diameters along the vessel were measured under unstressed and pressurized conditions in relaxing solution.

Effects of NOS inhibition.

To determine if WSS-induced remodeling is NO mediated, as previously observed (39), a group of LP rats with cervical-end ligations (n = 4) was treated with N-nitro-l-arginine methyl ester (l-NAME; Sigma) added to drinking water (0.5 g/l) from the day after surgery (gestational day 11) until euthanasia on day 20 of pregnancy.

Arterial measurements.

All rats were euthanized with a high dose of isoflurane (3%) followed by decapitation in a small animal guillotine. Both uterine horns were excised and immersed in the relaxing solution for at least 20 min. Arterial and venous diameters at both the cervical and ovarian end were measured under a microscope to provide a measure of true remodeling, which can only be evaluated under unstressed conditions since pressurization may distort the extent of true remodeling secondary to altered distensibility.

The MUA operates at essentially systemic pressure; therefore, values at 90 mmHg should approximate the arterial diameter at mean arterial pressure (90–95 mmHg under normal conditions). Segments of ovarian- and cervical-end MUA were cannulated and pressurized to evaluate diameter as a function of transmural pressure from opening pressure (2–3 mmHg) up to 150 mmHg in relaxing solution. Wall thickness, wall-to-lumen ratio, and vessel cross-sectional area were calculated under both unstressed and pressurized conditions in relaxing solution to eliminate any residual tone. For the calculation of circumferential wall stress, intraluminal pressure was converted from units of mmHg to N/m² (1 mmHg = 1.334 × 10² N/m²). Circumferential stress (ε) was calculated as follows: ε = (D_i – D_o)/D_o, where D_o is the original diameter, defined as the internal diameter (D_i) at 3 mmHg. The stress-strain relation was fitted to an exponential curve [f(x) = ae^bx) for each MUA segment to derive the stiffness coefficient (β) as the slope of each single curve. The means of each calculated slope for vessels from the control versus ligation group were then compared.

Collagen and elastin analysis.

Segments of ovarian-end MUAs from both the control and ligated horn of rats with cervical-end ligation (n = 3) were collected, fixed in formalin (10% for 4 h), and transferred to 70% ethanol. Samples were processed by the Surgical Pathology Histology Laboratory at the University of Vermont Medical Center for paraffin embedding, sectioning (6 μm), and staining with either elastic Van Gieson (EVG) or picrosirius red (PSR) using standardized protocols to determine elastin or collagen, respectively. Images were captured by an Olympus BX50 light microscope at ×200 magnification with a QImaging Retiga 2000R digital camera. ImageJ software (National Institutes of Health, Bethesda, MD) was used to determine collagen and elastin proportions in the vessels. Vessel area was determined by tracing the outer edge and lumen of the vessel. Color threshold was set to highlight the positive staining for collagen (red for PSR) and elastin (purple/black for EVG) and set to remain the same for all subsequent images. Collagen and elastin areas (in %) were calculated using the following equation: percent collagen or percent elastin = positive threshold area within the traced region/total vessel area × 100%.

Western blot analysis.

Protein expression of eNOS and phosphorylated eNOS at Ser¹¹⁷⁷ (p-eNOS¹¹⁷⁷) were quantified by Western blot analysis. Segments of ovarian MUA were placed in Pierce RIPA buffer (ThermoFisher Scientific, Waltham, MA) supplemented with Halt protease and phosphatase inhibitor cocktail (ThermoFisher Scientific) in Lysis Matrix D tubes (MP Biomedicals, Solon, OH) and homogenized using two 30-s pulses on a FastPrep-24 instrument (MP Biomedicals). Total extracted protein was determined by a BCA protein assay kit (ThermoFisher Scientific). Ten micrograms from each sample were analyzed by SDS-PAGE and Western blot analysis using primary antibodies (Cell Signaling Technology, Danvers, MA), rabbit monoclonal eNOS (1:1,000), rabbit monoclonal p-eNOS¹¹⁷⁷ (1:1,000), and rabbit polyclonal β-tubulin (1:3,000) antibodies. Protein bands were revealed by binding to horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology), detected by SuperSignal West Pico chemiluminescent substrate (ThermoFisher Scientific), and analyzed by densitometry using ImageJ software.

Statistical analysis.

Data were analyzed by Student’s t-test or ANOVA with post hoc Tukey’s or Sidak’s multiple-comparisons tests to detect differences between treatment means. All analyses were performed with GraphPad Prism7 (GraphPad Software, La Jolla, CA). Values are presented as means ± SE. P values of ≤0.05 were considered statistically significant.

RESULTS

NP animals.

In NP animals, MUA lumen diameter did not change after ovarian-end ligation surgery (Table 1). Cervical-end ligation did not affect vessel diameter at the ovarian end, and although lumen diameters were somewhat reduced just above the ligation at the cervical end, the difference did not reach statistical significance (Table 1).

Table 1.

Effect of vascular ligation on the remodeling of cervical- and ovarian-end MUA lumen diameters in nonpregnant and late pregnant rats

		Cervical-End MUA Diameter, µm		Ovarian-End MUA Diameter, µm
Animal Group and Ligation Site	n	Control	Ligation	Control	Ligation
Nonpregnant
Ovarian end	8	143.1 ± 13.1	152.5 ± 20.1	153.8 ± 21.4	130.6 ± 14.3
Cervical end	8	140.6 ± 17.3	102.6 ± 11.5	174.0 ± 15.3	148.8 ± 14.0
Late pregnant
Ovarian end	6	244.4 ± 16.4	216.9 ± 15.8	186.5 ± 14.9	206.3 ± 8.6
Cervical end	14	240.0 ± 10.9	191.8 ± 16.8	203.6 ± 18.1	295.7 ± 20.6*

Values are means ± SE; n, number of rats. MUA, main uterine artery.

^*P < 0.05 vs. the control value.

LP animals.

In LP animals, ligation of the MUA at the ovarian end did not affect unstressed arterial diameters (Table 1), since cervical-end arterial enlargement occurred with or without ovarian ligation.

Vascular ligation at the cervical end reversed the spatial pattern of remodeling in both MUAs (Table 1) and MUVs (Fig. 2A) such that expansive remodeling was augmented at the ovarian end. Ovarian-end MUA diameter increased significantly in both unstressed and pressurized conditions (Fig. 2B). There was no significant change in MUA wall thickness (Fig. 2C) and wall-to-lumen ratio (Fig. 2D) at the ovarian end, but cross-sectional area were increased because of a larger MUA diameter (Fig. 2E). The effects of ligation on arterial elasticity were evaluated by plotting the stress-strain relationship (Fig. 2F); the stiffness coefficients (β) for MUAs from control versus ligated horns were not significantly different.

Fig. 2.

Cervical-end ligation of main uterine artery (MUA) + main uterine vein (MUV) induces significant expansive remodeling of both arteries and veins at the ovarian end in late pregnant (LP) rats (n = 14). A: unstressed MUV diameters at the ovarian versus cervical ends showed enlargement (outward remodeling) of the vein at the ovarian end and inward remodeling at the cervical end, where flow volumes would be maximal and minimal, respectively. B: ovarian-end MUA diameters in the ligated horn increased significantly in both unstressed and pressurized conditions. C and D: wall thickness and wall-to-lumen ratio of the ovarian-end MUA did not change after vascular ligation in either the unstressed or pressurized conditions. E: cross-sectional area was significantly increased in the unstressed (P < 0.05) but not pressurized (P = 0.057 at 90 mmHg) condition because of the larger lumen diameter and is indicative of outward (expansive) hypertrophic remodeling. F: effect of ligation on the circumferential wall stress vs. strain relationship in ovarian-end MUAs. G: stiffness coefficients (β) calculated for ovarian-end MUAs from control versus ligated horns were not different (P > 0.05). *P < 0.05.

Collagen and elastin analysis.

We next determined whether there were any changes in arterial wall collagen and elastin content secondary to ligation surgery (Fig. 3). Neither elastin nor collagen content was significantly altered during the remodeling process in vessels from ligated versus sham-operated contralateral horns.

Fig. 3.

Cervical-end vascular ligation surgery in late pregnant (LP) rats did not alter the relative amounts of elastin or collagen in the ovarian-end main uterine artery (MUA). A: ovarian-end MUA cross sections from control versus ligated horns stained with elastic Van Gieson (EVG) for elastin analysis (black). B: bar graph showing elastin as a percentage of total vessel area (n = 3; P > 0.05). C: ovarian-end MUA cross sections from control versus ligated horns stained with picrosirius red (PSR) for collagen analysis (red). D: bar graph showing collagen as a percentage of total vessel area (n = 3). Scale bars = 50 μm.

NO signaling mediates WSS-induced vascular remodeling.

NOS inhibition with l-NAME treatment completely prevented the increase in ovarian-end MUA lumen diameter after cervical-end ligation (Fig. 4A). Compared with the contralateral (control) horn, ligation significantly augmented the protein expression of total eNOS but not p-eNOS¹¹⁷⁷ in ovarian-end MUAs (Fig. 4, B–D).

Fig. 4.

Wall shear stress (WSS)-induced uterine vascular remodeling is mediated by nitric oxide (NO). A: nitric oxide synthase (NOS) inhibition with N-nitro-l-arginine methyl ester (l-NAME) added to drinking water (0.5 g/l; n = 4) abolished the increase in the ovarian-end main uterine artery (MUA) diameter by the cervical-end vascular ligation surgery in late pregnant (LP) rats (n = 14). B: Western blot analysis of endothelial nitric oxide synthase (eNOS) and phosphorylated eNOS at Ser¹¹⁷⁷ (p-eNOS¹¹⁷⁷) in the ovarian-end MUA from control versus ligated horns (n = 4). C and D: summary graphs of Western blot analysis indicating an increase in eNOS in the ovarian-end MUA on the ligated versus contralateral sham-operated control side but no significant change in p-eNOS¹¹⁷⁷ protein expression after cervical-end vascular ligation. *P < 0.05.

Effects of ligation on reproductive outcome.

Cervical-end ligation of the MUA + MUV did not significantly alter pup number, incidence of fetal resorption, pup weight, or placental weight (Fig. 5, A–D).

Fig. 5.

No change in the reproductive outcomes following cervical-end vascular ligation in late pregnant (LP) rats. A−D: numbers of pups per horn (A), numbers of fetal resorptions (B), pup weights (C), and placental weights (D) were similar in control versus ligated horns (n = 13).

Vein removal.

Figure 6A shows a photograph of a successful surgery with ligations, flanking the segment of vein that was removed. Postplacental venous blood flow was redirected to the main uterine vein via arcuate veins, and the weight of the corresponding pup was not affected (data not shown). After ligation at the cervical end, MUA (+ vein) diameters were significantly enlarged at the ovarian end compared with the control (sham-operated) horn, as before (Fig. 2). This difference was evident under both unstressed (Fig. 6B) and pressurized (Fig. 6C) conditions. Conversely, after vein removal, MUA (− vein) diameters were significantly smaller in both the unstressed and inflated state (Fig. 6, B vs. C) such that the extent of MUA (− vein) remodeling was approximately one-half of that measured in the neighboring MUA (+ vein) segment. The respective areas of measure are shown in Fig. 6A.

Fig. 6.

Vein removal combined with ligation surgery decreases wall shear stress (WSS)-induced main uterine artery (MUA) remodeling (n = 9). A: photograph of one uterine horn from a late pregnant (LP) animal showing three ligations flanking a segment of vein (completely removed). Blue arrowheads, points of ligation; black and gray bars, points of diameter measurements for the MUA (+ vein) and MUA (– vein) groups in the ligated horn, respectively. Both unstressed (B) and pressurized (C) ovarian-end MUA lumen diameters were larger in the cervical end-ligated horn, with intermediate values in MUA segments adjacent to the site of vein removal. *P < 0.05.

DISCUSSION

Maternal uterine vascular remodeling during pregnancy.

Growth of the maternal uterine vasculature during pregnancy sustains the progressively increasing demands of a growing fetus by maintaining adequate uteroplacental perfusion, which increases many fold in all mammalian species. We (39) and others (47) have found that endothelial NO is a key physiological mediator of this expansive remodeling process and that both eNOS and endothelial NO production are augmented during pregnancy (36, 56). It is therefore important to determine the actual stimulus responsible for augmenting NO signaling in the uterine circulation. Physical forces, particularly WSS (19, 38, 49, 55), stimulate eNOS expression and NO signaling, as do sex steroids (16, 44, 54) and growth factors (30, 37, 40), although how these various stimulatory inputs interact is not well understood.

Surgical approach for altering uteroplacental hemodynamics.

Rats were chosen as the animal model because rodents, like humans, exhibit a hemochorial type of placentation and a bilateral anatomical vascular arrangement to provide the uterus with a dual source of blood and redundancy in case of occlusion. We used surgical ligation of both the MUA and MUV as a means for creating a single point of uteroplacental inflow from either the cervical (uterine artery) or ovarian (ovarian artery) end. In contrast to the normal physiological situation, in which there is inflow and outflow at each end of the uterus, this intervention also created unidirectional flow within both the MUA and MUV. This ligation surgery creates a more linear WSS pattern within both the arterial and venous territories and likely increases WSS secondary to the enlarged perfusion territory resulting from a single point of inflow/outflow.

Ligation surgeries were performed relatively early in the pregnancy (gestational day 10) because earlier studies have shown that measurable increases in UPBF are first detectable on day 15 in rats and that most vascular remodeling occurs during the last week of gestation (12). Also, ligation at later gestational ages results in fetal resorption and severe growth restriction (4, 57). Sampling 2 days before term (day 20/22) was chosen to avoid preparturition changes that reportedly begin ~36 h before delivery (11).

Effects of ovarian- versus cervical-end vascular ligation.

Surprisingly, MUA diameters in NP rats did not change after either cervical- or ovarian-end ligation (Table 1), as might have been expected because of the increase in perfusion territory created by vascular ligation. This differs from the situation in the splanchnic circulation where vascular ligation to increase perfusion territory induces expansive arterial remodeling in NP rats (52). The difference in our findings versus those of Tarhouni and colleagues suggests that there may be regional adaptations in arterial responses to altered perfusion. Unfortunately, measuring blood flow was beyond the scope of this study, and we cannot state with certainty that the volume of flow is increased after cervical-end vascular ligation in the NP state, although there was no visible evidence of ischemia or necrosis that might result from severe underperfusion. The absence of any effect of ligation on reproductive outcomes (pup number, resorptions, or pup and placental weights; Fig. 5) supports this interpretation.

Ovarian-end ligation of the MUA and MUV in LP rats also did not affect MUA diameter at the cervical end (Table 1). This may be explained by previous observations that inflow from the cervical end is predominant during pregnancy in rats, with ~70% of the blood volume entering via the uterine arteries at the cervical end of the uterus and only 30% via the ovarian artery (14, 45, 46). A similar hemodynamic pattern has been reported in humans (8).

On the other hand, ligation of the uterine artery at the cervical end on day 10 of pregnancy resulted in an ~50% increase in unstressed MUA diameters at the ovarian end relative to the same vessel in the contralateral (sham-operated) horn 10 days later (Fig. 2), in agreement with an earlier study in which only the artery was surgically ligated (38). The combination of a larger lumen, increased cross-sectional area, and unchanged wall thickness under unstressed conditions indicated that the nature of the WSS-induced remodeling of the ovarian-end MUA was outward hypertrophic (35).

The MUA normally operates at, or close to, mean arterial pressure (33), and the pattern of MUA remodeling at 90 mmHg was similar to that of vessels under unstressed conditions. Stress-strain relationships and the stiffness coefficient (β) of MUAs from ligated versus control horns were not altered by the increase of shear stress, suggesting that there was no significant change in the composition of vascular wall, which was also confirmed by the histological analysis of collagen/elastin (Fig. 3).

Role of NO signaling.

The enlargement of ovarian MUA diameter after ligation was completely abolished in the l-NAME treatment group (Fig. 4), confirming the primary role of NO in the remodeling process. l-NAME is known to increase systemic blood pressure, which may affect UPBF (and the attendant change in WSS) in the uterine circulation. Furthermore, high doses of l-arginine analogs were shown to affect prostaglandin synthesis and decrease vessel tone (23); thus, the role of prostanoids cannot be completely excluded. Our data also demonstrated an upregulation of eNOS protein expression in the remodeled MUAs (Fig. 4), supporting the linkage between WSS, NO, and expansive remodeling. The increased NO signaling appears to be primarily because of increased enzyme content rather than activity, since there was no change in the phosphorylation of eNOS at Ser¹¹⁷⁷. Changes in other eNOS phosphorylation sites, and alterations in eNOS spatiotemporal dynamics in relation to mechanisms of WSS-induced vascular remodeling cannot be excluded, however, and future studies are warranted.

Venous structural changes and influences.

The extent of gestational expansive remodeling in the MUV was comparable to that of the MUA (Fig. 2A). Veins operate at a much lower intravascular pressure than arteries, but they also experience shear stress. Given the fact that arterial inflow must equal venous outflow, postplacental venous WSS would also have to be increased and may play a role in regulating venous structure. Several studies have shown that WSS regulates venous endothelial cell behaviors, e.g., transcriptional regulation of heparan sulfate proteoglycan expression (25), upregulation of transient receptor potential vanilloid channels (7), and release of vasoactive factors such as NO (5). Changes in shear stress are also thought to contribute to venous remodeling in vein grafts (18), although it is difficult to exclude other mechanical forces such as pressure and pulsatility, which are altered as well.

By removing a segment of vein adjacent to an MUA, we were able to create a localized area that was devoid of venoarterial communication. As shown in Fig. 6, vein removal reduced the extent of MUA widening by ~50%, which indicates that venous influences may indeed be a significant physiological mechanism for regulating uterine artery remodeling during pregnancy. This concept is substantiated by the fact that postplacental veins are quite permeable (6, 7) and that the placenta secretes an array of signals such as growth factors and hormones (e.g., placental growth factor and sex steroids) into the venous circulation, where their concentration is highest. Previous studies have also established that vasoactive factors present within [e.g., VEGF (see Refs. 6 and 7)] or secreted by (e.g., NO or prostaglandins) veins (4, 32) can modulate arterial tone and may be regulated by WSS via changes in endothelial signaling (8, 14, 19, 27, 50, 51). Although the situation with respect to remodeling is less clear, a linkage between altered tone and changes in vessel structure, with vasoconstriction leading to inward and vasodilation to outward remodeling, has been noted in several earlier papers (3, 17, 29).

In summary, these findings indicate that, during pregnancy, 1) increased WSS stimulates uterine artery growth via NO signaling and 2) the presence of an adjacent vein is required for arterial remodeling to fully occur. Our results are the first to demonstrate an active venous influence on maternal gestational uterine artery remodeling and encourage further examination of the underlying mechanisms, particularly of how placental signals carried within the venous effluent might alter the periarterial environment and/or level of tone, and thereby induce changes in arterial structure.

GRANTS

This work was supported by National Institutes of Health Grants R21-HD-080156 and RO1-HL-134371 (to G. Osol).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.L.K., M.M., and G.O. conceived and designed research; N.L.K., and M.M. performed experiments; N.L.K., L.J., A.G., and M.M. analyzed data; N.L.K., M.M., and G.O. interpreted results of experiments; N.L.K. prepared figures; N.L.K. drafted manuscript; N.L.K., M.M., and G.O. edited and revised manuscript; N.L.K., L.J., A.G., M.M., and G.O. approved final version of manuscript.