Hemodynamic, Vascular, and Reproductive Impact of FMS-Like Tyrosine Kinase 1 (FLT1) Blockade on the Uteroplacental Circulation During Normal Mouse Pregnancy

ABSTRACT

To investigate the role of FMS-like tyrosine kinase 1 (FLT1, also known as VEGFR1) signaling during pregnancy, mice were injected with anti-FLT1 neutralizing antibody (Ab) beginning on Gestational Day 8 or 12 and every other day thereafter until Day 18; vehicle-only injected mice served as controls. Uterine artery blood flow was measured with ultrasound on Days 13 and 18, and morphometric measurements of the uterine arcade were carried out on Day 19 to provide a measure of gestational vascular remodeling; reproductive performance was evaluated by determining litter size, resorption rates, and pup and placental weights. Ab injections beginning on Day 8 or Day 12 resulted in significant reductions of uterine artery peak systolic and diastolic flows at Days 13 and 18. In addition, normal reproductive function was compromised, as evidenced by a significant reduction in average number of viable pups along with enhanced resorption rates. Reproductive performance was also significantly compromised in this group, although less severely. There was no evidence of a reduction in main uterine artery diameters, though arterial distensibility was reduced, and the diameter of the main uterine vein was significantly smaller in the Ab-injected mice. Significant reductions in main uterine artery and segmental artery length were also noted. Placental and pup weights were similar in all the groups. FLT1 inhibition during murine pregnancy impaired blood flow to the fetal-placental unit, compromised several indices of vascular remodeling, reduced fecundity, and increased fetal reabsorptions. The effects of FLT1 inhibition are most pronounced when targeted during early pregnancy.

Systemic inhibition of FLT1 signaling during normal murine pregnancy is associated with significant fetal loss, abnormal uteroplacental blood flow, and altered remodeling and biomechanical properties of the uterine circulation.

INTRODUCTION

The uterine vasculature undergoes considerable growth and remodeling during pregnancy, a process that leads to significant increases in uterine blood flow characteristic of mammalian gestation [1–6]. The purpose of this process is to accommodate the increasing demands of the developing fetus with respect to oxygen and nutrient delivery. As uterine vessels grow larger and become more distensible, both endothelial and vascular smooth muscle cell mitotic activity increases, and changes in the composition of the intercellular matrix alter biomechanical properties such as distensibility and compliance [1, 3, 5, 7–10]. The mechanisms underlying gestational uterine vascular remodeling are multi-factorial and carried out through a combination of both local and systemic influences [11].

Previous studies from our and other laboratories have suggested an important role for both vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) in placental angiogenesis, vascular remodeling, and vasodilation [12–15]. Much recent attention has been directed toward the role of soluble FLT1 (sFLT1) in the genesis of preeclampsia [13, 16]. The implication is that an excess of sFLT1 reduces the effectiveness of VEGF/PlGF signaling at the tissue level and results in an angiogenic imbalance and enhanced vasoconstriction. Comparatively less is known about the functions of the nonsoluble form of this receptor (VEGFR1, official symbol FLT1) in the vascular wall and about its physiological role in the maternal uteroplacental circulation during pregnancy. Because prior studies have shown that FLT1 activation by PlGF is a potent stimulus for uterine arterial and venous vasodilation [11, 15, 17], our hypothesis in this study was that increased PlGF production during pregnancy via FLT1 signaling may contribute to the profound increase in uterine blood flow noted during normal pregnancy. Vasodilatory effects of PlGF have also been noted in other types of vessels, e.g., those from the gut and kidney [13, 15].

Thus, the purpose of this study, which examined the in vivo effects of FLT1 inhibition during pregnancy, was to test the hypothesis that this receptor subtype plays an active role the hemodynamic and structural adaptations of the uterine circulation and that its inhibition would affect reproductive performance. Specifically, we first evaluated uterine artery blood flow noninvasively using Doppler-ultrasound and then performed morphometric measurements of the uterine arcade (unstressed diameter and length of the main uterine artery and vein, and segmental artery diameter and length) as a measure of gestational vascular remodeling prior to delivery. Antibody (Ab) effects on reproductive performance were evaluated by measurement of litter size, pup and placental weights, and resorption rate.

The results confirm a significant uteroplacental hemodynamic effect of FLT1 inhibition, along with an attenuation of arterial distensibility, venous growth, and deleterious actions on normal reproductive function as evidenced by a striking increase in the incidence of live pup resorptions that lead to significantly reduced litter size.

MATERIALS AND METHODS

Animals

Female CD1 mice were obtained from Charles River Laboratories (Wilmington, MA) and singly housed at the Animal Research Facility at the Beth Israel Deaconess Medical Center, which is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care, International. Feed and water were provided ad libitum. Prior to necropsy or tissue harvesting, animals were euthanized by exposure to CO₂ for 5 min. All experiments and procedures were approved by the Institutional Animal Care and Use Committee at the Beth Israel Deaconess Medical Center/Harvard Medical School.

Treatment Protocols

Pregnant mice received i.p injections with rat anti-mouse FLT1 (clone #MF1; Imclone systems, New York City, NY) [12] at 35 mg/kg every other day from Gestational Day 8 till Gestational Day 18. Animals injected with vehicle (PBS) on the same days served as controls. We also performed a second experiment using the same Ab where the animals were injected later during pregnancy, beginning at Gestational Day 12 and repeated every other day till Gestational Day 18. Animals injected with vehicle on same days served as controls.

Intravital Ultrasonographic Pulse-Wave Doppler Blood Flow Velocity Measurements

We measured uterine artery blood flow velocities and resistive indices in the mice at two time points during pregnancy (Gestational Days 13 and 18). Blood flow velocity measurements were obtained using Vevo 770 High Frequency Intravital Ultrasonography Platform (Visual Sonics, Toronto, ON). Briefly, mice were anesthetized using Isoflurane anesthesia and maintained in that state throughout the procedure. Electrocardiogram, respiratory rate, and body temperature were monitored and maintained to the optimal recommended values. Fur was removed from the ventral surface on supine-placed mouse by depilatory application. All measurements were obtained transabdominally. The abdominal aorta was identified, and followed by the internal iliac artery, as the uterine artery (UA) is a branch of the latter. Pulse-wave Doppler (PWD) mode was used to record blood flow pattern in real time in the right UA in all cases and was analyzed by a postacquisition processing software package (Visual Sonics). The following UA parameters were measured/calculated and recorded (an average of three measurements was used for each value; particular attention was paid to perform all the measurements during the same phase of the animal's breath cycle—expirium): 1) peak systolic velocity of blood flow (PSV) measured in millimeters per second, 2) late diastolic velocity of blood flow (LDV) measured in millimeters per second, and 3) resistive index, calculated according to the following formula: (PSV – LDV)/PSV measured in arbitrary units.

Main Uterine Artery, Main Uterine Vein, and Mesometrial Arcade Measurements

After animal euthanasia, the uterus and its vasculature was removed en bloc and shipped to Vermont in centrifuge tubes filled with physiological buffer packed in ice. Upon receipt, each uterus was positioned with pins in a silicone-coated Petri dish containing isotonic buffer to expose the course of the mesometrial uterine arcade. Care was taken to not stretch or compress the uterus and its vasculature during the pinning procedure. The length of the main uterine artery was measured along its curvature from the cervix to the ovarian end of the uterine horn. The unstressed lumen diameter of the main UA and uterine vein (UV) was measured at the approximate midpoint of the mesometrial arcade with a stereomicroscope (Zeiss, Germany) using a calibrated reticule while the vessel was bathed in a relaxing solution containing 100 μM papavarine at room temperature to insure full relaxation. Segmental artery length was determined at the midpoint of the uterine horn by measuring the distance from the main uterine artery to the nearest edge of the uterine corpus.

Pressurized Uterine Artery Preparation

Uterine arteries in the uterine horns were dissected free of connective tissue and placed in a relaxing solution (100 μM papaverine). Vessels were cannulated with buffer-filled hollow pulled-glass pipettes at room temperature in a custom-built arteriograph. Intralumenal pressure was measured and maintained with a pressure servo system (Living Systems Instrumentation, Burlington, VT) with an in-line transducer that was calibrated before each experiment. A thermostat controlled the superfusate temperature flowing through the vessel chamber of the arteriograph. After vessels were cannulated, they were rinsed free of any residual blood (≤10 mm Hg) and then maintained at 37° ± 0.2°C and 10 mmHg for 30 min, followed by a 30-min equilibration at 50 mm Hg. Transmural pressure was then reduced to the lowest pressure at which each vessel opened with a visible lumen (2–4 mm Hg) to allow accurate measurement of inner diameter and wall thickness in a minimally stressed state. Transmural pressure was then increased stepwise up to 100 mm Hg, and UA inner (lumen) diameter was recorded. Distensibility was expressed as the percent increase in lumen diameter over the diameter at the opening pressure (D_x/D_o × 100, where D_x = diameter at any particular pressure, and D_o = unstressed diameter @ 2–4 mm Hg).

Measures of Reproductive Performance

Fetuses were counted in each uterine horn, noting any resorption sites using a dissecting microscope in order to determine the number of degenerated fetuses per litter. Each fetus and placenta was then carefully dissected away from the uterus and weighed without membranes and umbilical cords to the nearest 0.01 g.

Chemicals

All chemicals were purchased from Fisher Scientific (Fair Lawn, NJ) unless otherwise specified. The composition of the isotonic buffer was HEPES (10 mM), NaCl (141.8 mM), KCl (4.7 mM), CaCl₂ (2.8 mM), MgSO₄ (1.7 mM), KH₂PO₄ (1.2 mM), Na₂-ethylenediaminetetraacetic acid (0.5 mM), and dextrose (5 mM). The pH was adjusted to 7.40 at 37.0°C by adding 10 M NaOH (Mallinckrodt Baker, Paris, KY). Relaxing solution for measurement of isolated vessels was prepared by adding 100 μM papavarine (Sigma, St. Louis, MO) to HEPES buffer.

Statistical Analyses

Data are expressed as mean ± SEM and were analyzed by two-way ANOVA followed by Dunn multiple comparisons test to evaluate differences between control and treatment means (P ≤ 0.05 considered significant).

RESULTS

Blood Flow Velocity Data

To study the effects of systemic FLT1 inhibition during pregnancy, we first evaluated UA blood flows using doppler-ultrasound at Gestational Days 13 and 18 as described in Materials and Methods. Representative UA flow abnormalities, including an image of waveform showing the PSV and end diastolic velocity (EDV), are shown in Figure 1, A and B. Mice exposed to anti-FLT1 neutralizing Ab from Gestational Day 8 demonstrated dramatically lower systolic and diastolic velocity and a higher resistive index as compared to controls at Gestational Day 13 (Fig. 2, A–C). These changes persisted at Gestational Day 18; however, there was no change in the resistive index between the two groups at this latter time point. When the experiment was repeated with the 4-day delay in the intitial Ab injection (Day 12 instead of 8), similar findings were noted except that alterations in blood flow were less dramatic (Fig. 3, A–C).

FIG. 1.

Representative ultrasound-doppler images of uterine circulation in pregnant mice. A) Image of pregnant mouse at Gestational Day 13 (E13) and Gestational Day 18 (E18). Upper panels are from control mice, and lower panels are obtained from anti-FLT1 Ab (MF1 clone)-injected mice. The uterine artery was localized using Doppler behind the bladder. A waveform was obtained using pulse Doppler, and the systolic and diastolic velocities were measured as shown. B) Image of waveform showing the PSV and EDV.

FIG. 2.

Uterine artery blood flow in the Day 8 (early) protocol. A) Change in the uterine artery peak systolic and diastolic velocities at Gestational Day 13 (E13) in the control and the anti-FLT1 (MF1 clone)-injected mice. B) Changes in the uterine artery peak systolic and diastolic velocities at E18 in the same group shown in A. C) Calculated resistive indices in the two Ab-injected groups at E13 and E18. There is an increase in EDV with advancing gestational age (P = 0.015). Values are presented as mean ± SD. *P < 0.05.

FIG. 3.

Uterine artery blood flow in the Day 12 (late) protocol. A) Change in the uterine artery peak systolic and diastolic velocities at Gestational Day 13 (E13) in the control and the anti-FLT1 Ab-injected mice in whom treatments began at Day 12 as described in the Materials and Methods. B) Changes in the uterine artery peak systolic and diastolic velocities at E18 in the same group shown in A. C) Calculated resistive indices in the two Ab-injected groups at E13 and E18. Values are presented as mean ± SD. *P < 0.05.

Reproductive and Fetal Data

Mice treated using the earlier (Day 8) injection protocol had a significantly smaller litter size and a higher rate of fetal resorption than controls (Fig. 4, A and B). Although this effect was present in the Day 12 injection group, some amelioration was evident since litter size increased, and the rate of resorption was reduced by approximately half. On the other hand, there was no significant difference between the Ab-treated and control groups with respect to either placental (Fig. 4C) or fetal (Fig. 4D) weights. Thus, while litter size was significantly reduced, the surviving pups appeared to be normal in size.

FIG. 4.

Reproductive performance data (different letters connote statistically significant differences). A) Anti-FLT1 Ab injections produced significant reductions in number of viable pups present on Day (d) 19 of pregnancy (litter size; P < 0.05). B) Anti-FLT1 Ab injections produced significant reductions in resorption rate (%; P < 0.05). C) Anti-FLT1 Ab injections did not affect Day 19 placental weights (g; P > 0.05). D) Anti-FLT1 Ab injections did not affect Day 19 pup weights (g; P > 0.05). Values are presented as mean ± SD.

Vascular Remodeling

The unstressed length of the main UA in Ab-treated animals was significantly shorter than that of controls, regardless of whether injections were started on Day 8 or Day 12 (Fig. 5A). Relative to controls, segmental artery length was also significantly reduced in the Ab-injected animals subjected to either protocol (Fig. 5B). When unstressed main UA inner diameter was assessed, no significant differences were observed between the three groups (Fig. 5C). Relative to controls, the passive distensibility of the main UA was significantly lower in Day 8; vessels from animals in the later (Day 12) injection protocol showed a similar trend, although this did not reach statistical significance (110% ± 10.6% vs. 126% ± 6.2% in control animals; P = 0.34; Fig. 5D). Assessment of main UV outer diameter (Fig. 5E) showed a significant reduction in unstressed diameter in veins taken from mice in early treatment group, while vessels from animals in later treatments showed a similar pattern, although the difference was not statistically significant (P = 0.15).

FIG. 5.

Uterine vascular data (different letters indicate statistically significant differences). A) Main uterine artery length (cm) was significantly reduced in Day (d) 8 and Day 12 anti-FLT1 Ab-injected mice relative to controls (P < 0.05). B) Segmental artery length (cm) was significantly reduced in Day 8 and Day 12 anti-FLT1 Ab-injected mice relative to controls (P < 0.05). C) Unstressed main uterine artery diameters (μm) were unchanged (P > 0.05) by anti-FLT1 Ab treatment. D) Distensibility (% change over opening diameter) was significantly (P < 0.05) reduced in Day 8 anti-FLT1 Ab-injected mice; values from the Day 12 group were intermediate (P > 0.05). E) Unstressed main uterine vein diameters (μm) were significantly reduced in Day 8 Ab-treated mice (P < 0.05); values from the Day 12 group were intermediate (P > 0.05). Values are presented as mean ± SD.

DISCUSSION

To our knowledge, this is first study to examine the effects of FLT1 inhibition on maternal uterine vascular and reproductive parameters during normal mouse pregnancy. There were three major findings. 1) Inhibition of systemic FLT1 signaling resulted in significant reductions in third semester (Day 13, Day 18) uterine artery peak systolic and diastolic blood flows and an increase in the resistive index. Here and elsewhere, Ab effects were more severe in the Day 8 vs. Day 12 injected groups. 2) A significant impact on uterine vascular remodeling was also noted; specifically, expansive venous remodeling was attenuated, as was axial remodeling of both large and small arteries. Although there was no measurable effect on main UA diameter, matrix composition may be affected based on the observed reductions in arterial distensibility in vessels from Ab-treated animals. 3) FLT1 inhibition resulted in a striking abrogation of the gestational process as evidenced by reduced fecundity and an increased rate of fetal absorption, suggesting that signaling via this receptor subtype is critical for normal placentation and early embryogenesis.

These findings are consistent with previous ex vivo studies on isolated uterine arteries that established the potency of FLT1 activation with PlGF in mediating both arterial and venous vasodilation [15, 17]. These findings also have implications for the development of therapeutic agents for preeclampsia, a disorder that is characterized by high-circulating concentrations of sFLT1 [13]. In particular, neutralizing antibodies against FLT1/sFLT1 may have deleterious consequences for the pregnancy, particularly when administered early in pregnancy.

The role of FLT1 in angiogenesis and vascular homeostasis is complex. VEGF-A induces much weaker tyrosine-kinase activity in FLT1 [18], most likely due to the presence of inhibitory sequence in its juxtamembrane domain [19]. The high affinity of FLT1 for VEGF-A led to the theory that FLT1 may act as a “molecular sink” or “decoy” receptor, with the primary objective being VEGF-A sequestration which, in turn, attenuates VEGFR-2 (Flk1)-related signaling and angiogenesis. Data showing that mouse mutants expressing the membrane-bound form of FLT1 but lacking the tyrosine-kinase domain (Flt^TK-/-) are viable and exhibit no vascular defects are consistent with this hypothesis [20]. On the other hand, more recent data demonstrating that FLT1 signaling in monocytes induces migration, and the aforementioned potent vasodilatory effects of PlGF—a FLT1 specific agonist—suggest that FLT1 may be an active receptor in both immune and vascular homeostatic functions [15, 21, 22]. The profound impairment in uteroplacental blood flow in FLT1 Ab-treated animals also supports the latter hypothesis; clearly, additional studies are warranted to explore these mechanisms in greater detail.

Regarding uterine vascular remodeling, we did not find any Ab effect on unstressed main uterine artery diameter, although distensibility was significantly reduced in the early Ab-treated (Day 8) group, which would predict increased blood flow resistance at physiological pressures. We also did not examine changes in UA tone and reactivity—if tone or vasoconstrictor effects were augmented by FLT1 inhibition, uterine vascular resistance would likewise be increased. Thus, we cannot exclude FLT1 effects on the uterine arterial circulation, especially in view of the in vivo hemodynamic data already discussed above. Another potential target for FLT1 inhibition is intraplacental angiogenesis. Although placental weights were normal, this is a fairly general parameter that does not afford any insights into possible changes in intraplacental architecture (such as intraplacental vessel density and connective tissue content, both of which could affect placental flow resistance). Future studies aimed at better understanding the role of FLT1 in placental angiogenesis would be useful in providing more direct insights into this concept.

The significant (>30%) reductions in venous diameters in both Ab-treated groups are interesting in this regard. The factors that regulate circumferential venous growth are not known; however, if shear stress is an important stimulus for venous expansive remodeling (as it clearly is in arteries [23, 24]), the significant reduction in venous diameters supports the hypothesis of increased placental resistance secondary to FLT1 inhibition leading to reduced venous remodeling and dovetails with the hemodynamic data. Even without any direct effects on the placental vasculature, reduced venous diameters would increase blood flow impedance and favor reduced uteroplacental blood flow.

Although the findings are provocative, our study has some limitations. We did not measure systemic blood pressure in these studies. Future studies should evaluate whether blood pressure is altered in response to uteroplacental abnormalities noted in animals exposed to systemic FLT1 inhibition. This is particularly relevant as we cannot rule out the possibility that the adverse effects of FLT1 Ab may also be related to inhibition of sFLT1 as the Ab cross-reacts with both receptor subtypes. At the same time, excess rather than reduced sFLT1 is associated with uterine vascular pathology and hypertension in preeclampsia [13, 16, 25]. The striking fetal losses in the Ab-injected animals and the fact that a 4-day delay in the initial injection ameliorated the resorption rate by approximately 50% suggest that FLT1 function is critical during the early embryonic period for reasons that we do not yet understand. We also cannot rule out whether the reduced fecundity was due to a direct abrogation of FLT1 signaling in the placental trophoblasts or secondary to impaired blood flow to the placenta. Prior data showing that a trophoblast-specific knock out of FLT1 leads to normal fetal growth [26] suggests that impaired blood flow to the placenta may be the etiology of reduced fecundity. We cannot explain why there were no significant changes in fetal weights in spite of the dramatic reductions in blood flows. One possibility is that the reduced fecundity allows for compensatory increases in blood flow to the remaining fetuses, thus maintaining normal growth in the surviving pups. Although reductions in main UA and segmental artery length were both measurable and significant, these effects may be indirect, i.e., secondary to reduced pup number, as a strong positive correlation between the number of pups per uterine horn and extent of arterial elongation was noted in an earlier study in rats [27]. A similar pattern was confirmed in this study, e.g., the coefficient of determination (r² value) between pup number and main UA length was 0.59; thus, by definition, approximately 60% of the variability in vessel length can be accounted for by differences in pup number.

In summary, this is the first study to show that systemic inhibition of FLT1 signaling during normal murine pregnancy results in a broad spectrum of reproductive, hemodynamic and uterine vascular consequences that is associated with significant fetal loss, abnormal uteroplacental blood flow, and altered remodeling and biomechanical properties of the uterine circulation.

DISCLOSURES

Dr. Karumanchi is a co-inventor of multiple patents related to angiogenic proteins for the diagnosis and therapy of preeclampsia. These patents have been licensed to multiple companies. Dr. Karumanchi reports having served as a consultant to Roche and Beckman Coulter and has financial interest in Aggamin LLC. The remaining authors report no conflicts.