Decreased uterine vascularization and uterine arterial expansive remodeling with reduced matrix metalloproteinase-2 and -9 in hypertensive pregnancy

Abstract

Normal pregnancy involves extensive remodeling of uterine and spiral arteries and matrix metalloproteinases (MMPs)-mediated proteolysis of extracellular matrix (ECM). Preeclampsia is characterized by hypertension in pregnancy (HTN-Preg) and intrauterine growth restriction (IUGR) with unclear mechanisms. Initial faulty placentation and reduced uterine perfusion pressure (RUPP) could release cytoactive factors and trigger an incessant cycle of suppressed trophoblast invasion of spiral arteries, further RUPP, and progressive placental ischemia leading to HTN-Preg and IUGR; however, the extent and depth of uterine vascularization and the proteolytic enzymes and ECM proteins involved are unclear. We hypothesized that HTN-Preg involves decreased uterine vascularization and arterial remodeling by MMPs and accumulation of ECM collagen. Blood pressure (BP) and fetal parameters were measured in normal Preg rats and RUPP rat model, and the uteri were assessed for vascularity, MMP levels, and collagen deposition. On gestational day 19, BP was higher, and the uterus weight, litter size, and pup weight were reduced in RUPP vs. Preg rats. Histology of uterine tissue sections showed reduced number (5.75 ± 0.95 vs. 11.50 ± 0.87) and size (0.05 ± 0.01 vs. 0.12 ± 0.02 mm²) of uterine spiral arterioles in RUPP vs. Preg rats. Immunohistochemistry showed localization of endothelial cell marker cluster of differentiation 31 (CD31) and smooth muscle marker α-actin in uterine arteriolar wall and confirmed decreased number/size of uterine arterioles in RUPP rats. The cytotrophoblast marker cytokeratin-7 showed less staining and invasion of spiral arteries in the deep decidua of RUPP vs. Preg rats. Uterine arteries showed less expansion in response to increases in intraluminal pressure in RUPP vs. Preg rats. Western blot analysis, gelatin zymography, and immunohistochemistry showed decreases in MMP-2 and MMP-9 and increases in the MMP substrate collagen-IV in uterus and uterine arteries of RUPP vs. those in Preg rats. The results suggest decreased number, size and expansiveness of spiral and uterine arteries with decreased MMP-2 and MMP-9 and increased collagen-IV in HTN-Preg. Decreased uterine vascularization and uterine arterial expansive remodeling by MMPs could be contributing mechanisms to uteroplacental ischemia in HTN-Preg and preeclampsia.

NEW & NOTEWORTHY Preeclampsia is a pregnancy-related disorder in which initial inadequate placentation and RUPP cause the release of cytoactive factors and trigger a ceaseless cycle of suppressed trophoblast invasion of spiral arteries, further RUPP, and progressive placental ischemia leading to HTN-Preg and IUGR; however, the extent/depth of uterine vascularization and the driving proteolytic enzymes and ECM proteins are unclear. This study shows decreased number, size, and expansiveness of uterine spiral arteries, with decreased MMP-2 and MMP-9 and increased collagen-IV in HTN-Preg rats. The decreased uterine vascularization and uterine arterial expansive remodeling by MMPs could contribute to progressive uteroplacental ischemia in HTN-Preg and preeclampsia.

INTRODUCTION

Normal pregnancy is associated with maternal cardiovascular, uterine, and placental adaptations. Increases in plasma volume, heart rate, and cardiac output are associated with systemic vasodilation and decreased vascular resistance, thus maintaining blood flow to different organs with little changes in blood pressure (BP). The uterus expands progressively with gestational time to accommodate the growing fetus. The placenta undergoes marked development, and the extravillous trophoblasts invade the spiral arteries in the uterine decidua, gradually replacing the endothelium and smooth muscle and creating fully dilated low-resistance spiral arteries that maintain blood supply to the fetus (44, 56). These hemodynamic, uterine, and placental adaptations require extensive remodeling and structural and functional alterations in the maternal blood vessels, uterus, and placenta (45, 75).

Matrix metalloproteinases (MMPs) are zinc-dependent endopeptidases that play a role in tissue remodeling (41, 51) and promote proteolysis and degradation of different proteins in the extracellular matrix (ECM). MMPs are first produced as pro-MMPs, which are cleaved into active MMPs (63, 77). MMPs include collagenases, gelatinases, stromelysins, matrilysins, and membrane-type MMPs (63). Uterine trophoblasts and vascular cells are major sources of MMPs (16, 28, 39, 70), and collagen-IV is a major substrate of MMP-2 (gelatinase A) and MMP-9 (gelatinase B) (36, 55, 63, 77). MMP-2 and MMP-9 play a role in endometrial tissue remodeling during the estrous and menstrual cycles and during pregnancy (50, 73, 83). We have shown increases in rat aortic and uterine MMP-2 and MMP-9 during normal pregnancy, supporting their role in pregnancy-related vascular and uterine remodeling (15, 79).

In contrast with normal pregnancy, preeclampsia is a pregnancy-related disorder affecting 5–8% of pregnancies in the United States and ~8 million pregnancies worldwide (74). Preeclampsia is associated with new-onset hypertension in pregnancy (HTN-Preg), most often after 20 wk of gestation and frequently near term (35). Although often accompanied by new-onset proteinuria, HTN-Preg may present in some women in the absence of proteinuria (29, 35, 71). Preeclampsia could progress to eclampsia, with severe hypertension, cerebral edema, seizures (33, 58), and maternal death, accounting for 15–20% of maternal mortality (9, 61). Preeclampsia is also often associated with fetal intrauterine growth restriction (IUGR), accounting for 10–15% of preterm pregnancy and newborn low birth weight (66, 74). However, the pathophysiological mechanisms of preeclampsia are not clearly understood. Inadequate placentation and trophoblast invasion of the uterine wall could be initiating events. In contrast with the deeply invasive trophoblasts in normal pregnancy, in preeclampsia trophoblasts only partially invade the superficial decidual vessels, sparing the endothelial lining and musculoelastic tissue in the deeper spiral arteries, leading to partial dilation of the spiral arteries to only half their size during normal pregnancy, and resulting in placental ischemia/hypoxia (76). While delineating the mechanisms causing inadequate placentation in early human pregnancy has been challenging, studies in animal models have made substantial progress in understanding the events that follow the initial placental ischemia. We and others have shown that surgical induction of initial placental ischemia by reducing uterine perfusion pressure (RUPP) in pregnant rats and mice shows some of the characteristics of preeclampsia including HTN-Preg and IUGR (4, 25, 85). Also, rat models of gestational hypoxia produced by exposing Preg rats to extended periods of hypoxia show preeclampsia-like characteristics, supporting a role of hypoxia in HTN-Preg (84). Of note, RUPP is not a short-lived, one-point-in-time event but rather a progressive, self-feeding, incessant process that culminates in HTN-Preg. In support, surgical induction of RUPP does not cause immediate elevation of BP, suggesting the involvement of other intermediate biological factors and linking mechanisms in the fully manifested HTN-Preg pathology. Previous studies have suggested that initial faulty placentation and RUPP could release antiangiogenic and other cytoactive factors (3, 6, 13, 24, 26, 32, 37, 38, 71, 82, 85) that could trigger an incessant cycle of suppressed trophoblast invasion of the uterine wall and spiral arteries, further RUPP, and progressive placental ischemia/hypoxia that eventually lead to HTN-Preg and IUGR. Progressive placental ischemia/hypoxia could also lead to uteroplacental rarefaction and attrition, and trophoblast loss and deportation into the circulation, further contributing to the inflammatory response in preeclampsia (12, 67). Interestingly, the proteolytic enzymes MMPS have a hypoxia response element, and hypoxia could affect MMP expression/activity and MMP-mediated proteolysis of ECM (42, 69). However, the changes in uterine and spiral arteries and the driving proteolytic enzymes, ECM proteins, and collagen substrates underlying progressive placental ischemia and HTN-Preg have not been clearly identified. Specifically, the extent and depth of uterine vascularization and the MMPs and collagen substrates that could alter uterine structure and function and further interfere with trophoblast invasion of the uterine wall and the uterine arterial expansive remodeling in the setting of HTN-Preg are unclear.

The present study was designed to test the hypothesis that decreased uterine vascularization and uterine arterial expansive remodeling due to decreased MMPs levels/activity and accumulation of collagen are important mechanisms that contribute to uteroplacental ischemia in HTN-Preg. We used normal Preg rats and RUPP-initiated rat model of HTN-Preg to investigate whether 1) HTN-Preg is associated with decreased uterine vascularization and uterine arteriolar remodeling, 2) decreased uterine arterial remodeling affects the uterine artery distensibility and propensity to dilation and expansion in HTN-Preg, and 3) decreased uterine vascularization and uterine arterial expansive remodeling involve changes in uterine and uterine arterial MMP-2 and MMP-9 and the MMP substrate collagen.

METHODS

Animals.

Timed-pregnant (day 11) Sprague-Dawley rats (12 wk of age, Charles River, Wilmington, MA) were housed in the animal facility and maintained on ad libitum standard rat chow and tap water in 12-h:12-h light-dark cycle. Gestational day 1 was determined by the vendor by the detection of vaginal sperm plugs. On gestational day 14, rats underwent either sham operation (Preg) or surgical reduction of uterine perfusion pressure (RUPP) by banding the lower abdominal aorta above the iliac bifurcation and the main uterine branches of the ovarian arteries as previously described (17, 25, 85). Briefly, the abdominal aorta near the iliac bifurcation was dissected free of perivascular fat and separated from the vena cava. A blunt plastic rod [outer diameter (OD), 0.3 mm] was placed parallel to the aorta, and a 4-0 silk braided ligature was knotted twice around the aorta and adjacent plastic rod. Once taut, the rod was removed from the knotted ligature, thus creating a constrictive band [inner diameter (ID), 0.3 mm] and reducing blood flow through the aorta. This procedure reduces uterine perfusion pressure in the gravid rat by ∼40% (18). Since compensation of blood flow to the placenta occurs through adaptive increase in ovarian blood flow (54), a blunt plastic rod (OD, 0.1 mm) was used to place a silk ligature (ID, 0.1 mm) on the main uterine branches of both the right and left ovarian arteries, specifically avoiding the main ovarian artery. The RUPP banding procedure was successful. Out of 13 animals, one animal did not do very well and had to be euthanized, and another animal did not have any pups at the time of euthanasia, leading to an 85% success rate of the RUPP procedure. RUPP rats in which the banding procedure resulted in maternal death or complete resorption of the pups were excluded from data analyses. All procedures followed the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals and the guidelines of the American Physiological Society and were approved by the Institutional Animal Care and Use Committee at Brigham and Women’s Hospital.

Maternal and fetal parameters.

On gestational day 19, body weight and BP were measured. Rats were anesthetized with isoflurane, and a PE-50 catheter was inserted in the carotid artery and exteriorized at the back of the neck. Rats were allowed to recover from anesthesia for 1 h. The carotid arterial catheter was connected to a pressure transducer attached to an amplifier and pressure recorder (Living System Instrumentation, Burlington, VT), and BP in conscious rats was recorded (48). Rats were then euthanized by inhalation of CO₂, and the gravid uterus was excised and placed in Krebs-Henseleit buffer (Krebs). The gravid uterus was cut open, the pups and placentae were separated and gently blotted between filter papers, and the uterus weight (without pups or placentae), litter size, and individual pup and placenta weights were recorded.

Tissue preparation.

The uterine artery and arterioles were carefully dissected under microscopic visualization and cut into 4- to 5-mm segments for functional studies and biochemical assays. The uterus was preserved for tissue histology or cut into 3×3-mm segments to facilitate tissue homogenization and protein extraction for the biochemical studies.

Uterine artery expansion to intraluminal pressure.

Segments of small uterine artery (300–500 µm OD, 4 to 5 mm long) were transferred to a temperature-controlled perfusion chamber, mounted between two glass micropipettes (cannulas), and secured with 10-0 ophthalmic nylon monofilament (Living Systems Instrumentation) (10, 47). The uterine artery in the perfusion chamber was placed on an inverted microscope (TE300, Nikon, Melville, NY). The vessel was bathed in 5 ml normal Krebs solution bubbled with 95% O₂-5% CO₂ at 37°C and was continuously superfused with fresh Krebs at a rate of 1 ml/min using a peristaltic minipump (Master-Flex; Cole-Parmer, Vernon Hills, IL). A stopcock at the distal end of the vessel was closed, and the proximal end of the vessel was connected to a pressure transducer and pressure servo control system (Living Systems Instrumentation). The vessel was gradually pressurized to 60 mmHg and maintained at constant pressure with the pressure-servo control unit. The vessel was allowed to equilibrate for 60 min before testing its viability and functional responsiveness to high (51 mM) KCl solution. High KCl solution causes membrane depolarization and increases Ca²⁺ entry from the extracellular space (43, 47). Vessels were unacceptable if they showed leaks or failed to produce maintained constriction to KCl. The vessels were continuously observed using a video camera and a monitor, and the changes in arterial diameter in response to increases in intraluminal pressure were measured using automatic edge-detection system (Crescent Electronics, Sandy, UT) (10, 47). The vessel intraluminal pressure was first set to 10 mmHg, and a snap picture of the vessel diameter was taken using a digital camera (Cool-Snap, Photometrics, Tucson, AZ). The vessel intraluminal pressure was then raised in a 10-mmHg stepwise manner to a test 60 and 70 mmHg, and snap pictures of the vessel diameter were taken. The expansion in the vessel diameter in response to the test increases in intraluminal pressure was then measured relative to the vessel diameter at 10 mmHg. Because autoregulation and active myogenic response could affect arterial diameter in a Ca²⁺-dependent manner (8, 27), the uterine arteries were then equilibrated in Ca²⁺-free (2 mM ethylene glycol-bis-(2-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, EGTA) Krebs solution for 20 min, the vessel intraluminal pressure was set to 10 mmHg, and the passive changes in arterial diameter in response to a test 60- and 70-mmHg increases in intraluminal pressure were measured.

Tissue homogenate.

The uterus and uterine arteries were homogenized using a 2-ml tight-fitting homogenizer (Kontes Glass, Vineland, NJ) and a homogenization buffer containing 20 mM 3-[N-morpholino] propane sulfonic acid, 4% sodium dodecyl sulfate (SDS), 10% glycerol, 2.3 mg dithiothreitol, 1.2 mM ethylenediaminetetraacetic acid (EDTA), 0.02% bovine serum albumin (BSA), 5.5 μM leupeptin, 5.5 μM pepstatin, 2.15 μM aprotinin, and 20 μM 4-(2-aminoethyl)-benzenesulfonyl fluoride. The tissue homogenate was centrifuged at 10,000 g for 10 min. If the supernatant contained floating debris, centrifugation was repeated two times to obtain a clearer supernatant. Protein concentration in the supernatant was determined using a protein assay kit (Bio-Rad, Hercules, CA).

Western blot analysis.

To assess the role of MMPs and collagen in uterine vascularization and remodeling, MMP-2, MMP-9, and collagen-IV levels were measured using Western blot analysis. Protein extracts (20 μg) were combined with an equal volume of 2 × Laemmli loading buffer, boiled for 5 min, and size fractionated by electrophoresis on 8% SDS-polyacrylamide gels. Proteins were transferred from the gel to a nitrocellulose membrane by electroblotting. Membranes were incubated in 5% nonfat dry milk in phosphate-buffered saline (PBS)-Tween for 1 h and then overnight at 4°C with polyclonal rabbit antibody to MMP-2 (sc-10736, 1:1,000), MMP-9 (sc-10737, 1:1,000), or collagen-IV (sc-70246, 1:1,000) (Santa Cruz Biotechnology, Dallas, TX). The authenticity of the MMP and collagen-IV antibodies was validated by immunoreactivity with recombinant MMP and collagen-IV. Negative control experiments were performed with the omission of primary antibody or with heat-inactivated primary antibody (by heating at 75°C for 30 s and cooling at 4°C for 1 min, repeated 10 times) (34) and showed no detectable immunoreactive bands. Membranes were washed 5 × 15 min in PBS-Tween and then incubated with horseradish peroxidase-conjugated secondary antibody (1:3,000) for 2 h. Membranes were washed another 5 × 15 min in PBS-Tween, and the immunoreactive bands were detected using enhanced chemiluminescence (ECL) Western blotting detection reagent (GE Healthcare Bio-Sciences, Piscataway, NJ). The membranes were stripped in stripping buffer and subsequently reprobed with monoclonal β-actin antibody (A1978, 1:3,000, Sigma) as an internal control. Equal amounts of sample protein were loaded onto the gels, and equal exposure times were used for the processing and development of the Western blots. The reactive bands were analyzed by optical densitometry and ImageJ software (National Institutes of Health, Bethesda, MD). The densitometry values represented the pixel intensity normalized to β-actin to correct for loading (15, 40, 41).

Gelatin zymography.

To further assess the role of MMPs in uterine vascularization and remodeling, MMPs gelatinolytic activity was measured using gelatin zymography. The uterus and uterine artery homogenates were prepared in a homogenization buffer without dithiothreitol, centrifuged at 10,000 g for 10 min, and the protein concentration in the supernatant was determined using a protein assay kit (Bio-Rad). Our preliminary gelatin zymography analyses in uterine tissue homogenate from Preg or RUPP rats have shown concentration-dependent increases in the intensity of MMP-2 and MMP-9 bands at loading protein amount from 0.1, 0.2, to 0.5 μg, and clearly discernible bands at 1 and 2 μg protein. Further increases in loaded protein to 5 and 10 μg showed further increases in gelatinolytic activity, but the MMP-2 bands became almost saturated (15, 41), and therefore all gelatin zymography experiments for MMP-2 and MMP-9 were performed using 1 μg loading protein. To control for inter-gel variation, equal amounts of loading protein (1 μg) from tissues of Preg and RUPP rats were used in different gels.

Tissue homogenates were subjected to electrophoresis on 8% SDS polyacrylamide gel containing 0.1% gelatin (Sigma, St. Louis, MO). The gel was incubated in a zymogram renaturing buffer containing 2.5% Triton X-100 (Sigma) with gentle agitation for 30 min at room temperature. The gel was then incubated in a zymogram developing buffer (pH 6.7) containing 50 mM Tris-base, 0.2 M NaCl, 5 mM CaCl₂, 0.02% Brij35 (Fisher Scientific, Pittsburgh, PA), and 1 μM ZnCl₂ (Sigma) at room temperature (22°C) for 30 min and then at 37°C for 16 h. The gel was stained with 0.5% Coomassie blue R-250 (Sigma) for 30 min and then destained in a destaining solution (methanol:acetic acid:water = 50: 10: 40). Areas corresponding to MMP-2 and MMP-9 gelatinolytic activity appeared as clear bands against a dark-blue background. The clear bands were analyzed by optical densitometry and ImageJ software, and the integrated protease activity was measured as pixel intensity per squared millimeters, normalized to actin intensity to correct for loading (15, 41).

Histology and quantitative morphometry.

Uteri from Preg and RUPP rats were preserved in Tissue-Tek 4583 optimal cutting temperature compound (OCT, Fisher Scientific) and stored at −80°C. Peripheral sections of the uterus did not show the uterine lumen, making it difficult to identify the decidua and different regions of the uterus. Therefore, to be consistent in comparing uterine sections from Preg and RUPP rats, cross-sectional 6-μm-thick cryosections from the middle segment of the uterus were placed on glass slides and prepared for staining with hematoxylin and eosin (H&E), and the stained sections were coded and labeled in a blinded fashion. Images of the uterine sections of Preg and RUPP rats were acquired on a Nikon microscope using bright-field illumination and the same lens (×4 or ×40 objective), microscope magnification, and camera gain. Because the rat uterus is large, 16 to 20 picture frames of sequential parts of the same uterine tissue section were acquired using ×4 objective and then inserted as individual raw images in a PowerPoint slide presentation. The different image frames were aligned to produce a reconstructed image of the uterus; then, the individual image frames were grouped and saved as a composite JPG image of the whole reconstituted uterine tissue section as previously described (41, 46). Images of tissue sections were analyzed using ImageJ software. To determine the depth of uterine vascularization, images of the uterine decidua were divided by imaginary lines into shallow, middle, and deep regions, with the shallow region abutting the placenta attachment site and the deep region representing deep invasion of the uterine wall. The number, lumen size, and wall thickness of uterine spiral arteries in the whole uterus and different regions (shallow, middle, and deep) of the decidua were measured. To correct for any intraindividual/interindividual variability, experiments were conducted in a blinded fashion and data analysis of the slides/images was performed by two independent investigators who were blinded to the animal group.

Immunohistochemistry.

To further localize the spiral arteries in the uterine decidua, cryosections from the middle segment of the uterus were thawed and fixed in ice-cold acetone for 30 min. Endogenous peroxidase was quenched in 1.5% H₂O₂ solution for 30 min, and nonspecific binding was blocked in 10% horse serum for 30 min. To determine the extent of cytotrophoblasts invasion of spiral arteries and replacement of uterine arteriolar endothelium and smooth muscle, uterine tissue sections were incubated with polyclonal rabbit antibody to the endothelium marker cluster of differentiation 31 (CD31; 1:100, Ab29364, Abcam), monoclonal mouse antibody to the smooth muscle marker α-actin (1:100, A5228, Sigma), or polyclonal rabbit antibody to the epithelia and invading trophoblasts marker cytokeratin-7 (CK7; 1:100, sc-25721, Santa Cruz Biotechnology). Also, to further assess the role of MMPs and collagen in uterine vascularization and remodeling, the distribution of MMPs and collagen-IV in the uterus and uterine spiral arteries, tissue sections were incubated with polyclonal rabbit antibodies to MMP-2 (1:100, sc-10736), MMP-9 (1:100, sc-10737), or collagen-IV (1:1,000, sc-70246, Santa Cruz Biotechnology) for 30 min. After being rinsed with PBS, tissue sections were incubated with the appropriate biotinylated anti-mouse or anti-rabbit secondary antibody, and they were rinsed with PBS and then incubated with avidin-labeled peroxidase (VectaStain Elite ABC Kit, Vector, Burlingame, CA) for 30 min. Positive labeling was visualized using diaminobenzadine and appeared as brown spots. Negative control slides were run simultaneously with no primary antibody. Specimens were counterstained with hematoxylin for 40 s, rinsed with PBS, topped with cytoseal 60, and covered with slide coverslips.

Images were acquired on a Nikon microscope using the same light intensity, microscope magnification, and camera gain. Images of tissue sections were analyzed using ImageJ software. Outlines of the tissue exterior and interior were used to define the whole tissue area and lumen area, respectively, and the wall area was calculated as whole tissue area and lumen area. The total number of pixels in the tissue wall image was defined, and then the number of brown spots (pixels) corresponding to CD31, α-actin, CK7, MMP-2, MMP-9, or collagen-IV was counted and presented as percentage of total pixels. The number of pixels in the specific region of the decidua (shallow, middle, deep) was also defined and transformed into the area (in mm²) using a calibration bar. The number of brown spots (pixels) representing MMP-2, MMP-9, or collagen-IV in each region was then counted (presented as number of pixels/mm²) (15, 46, 72). Experiments were conducted in a blinded fashion, and data analysis of the slides/images was performed by two independent investigators who were blinded to the animal group.

Solutions and drugs.

Normal Krebs solution was used for tissue dissection and uterine artery expansion experiments and contained (in mM) 120 NaCl, 5.9 KCl, 25 NaHCO₃, 1.2 NaH₂PO₄, 11.5 dextrose, 2.5 CaCl₂, and 1.2 MgCl₂ and bubbled with 95% O₂-5% CO₂ (pH 7.4). High (51 mM) KCl depolarizing solution was used to test the uterine artery viability and was prepared as normal Krebs with equimolar substitution of NaCl with KCl. Ca²⁺-free (2 mM EGTA) Krebs was used to measure the passive changes in arterial diameter and was prepared as 0 Ca²⁺ Krebs supplemented with 2 mM EGTA. PBS contained (in mM) 137 NaCl, 2.7 KCl, 8 Na₂HPO₄, and 2 KH₂PO₄ (pH 7.4). All other chemicals were of reagent grade or better.

Statistical analysis.

Experiments were conducted on uterus and uterine arteries isolated from 8 to 11 rats per group, and cumulative data were presented as means ± SE, with n representing the number of rats. Data were analyzed using Prism (v. 5.01; GraphPad software, San Diego, CA). Data were first analyzed using two-way ANOVA. When a statistical difference was observed, data were further analyzed using Bonferroni’s post hoc correction for multiple comparisons. Student’s unpaired t-test was used for comparison of two means. Differences were statistically significant when P < 0.05.

RESULTS

Maternal and fetal parameters.

On gestational day 19, mean arterial BP was significantly higher, and maternal body weight was less in RUPP vs. Preg rats. The rats were euthanized, and the uterus with pups and placentae were isolated. The body weight without uterus, the uterus weight, litter size (number of pups), and average individual pup weight were reduced in RUPP vs. Preg rats (Fig. 1). Also, the individual placenta weight and the placenta/fetal weight as a measure of placental efficiency were reduced in RUPP vs. Preg rats as previously described (17).

Fig. 1.

Maternal and fetal parameters in reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Blood pressure (A) and body weight (B) were measured in Preg and RUPP rats. The rats were euthanized, the uterus was isolated, and the body weight without uterus (C), uterus weight (D), litter size (number of pups) (E), and average individual pup weight (F) were compared in RUPP vs. Preg rats. Bar graphs represent means ± SE; n = 8 Preg and 11 RUPP rats. Data were analyzed using 2-way ANOVA and Bonferroni’s post hoc correction. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

Extent and depth of uterine vascularization.

In uterine tissue sections stained with H&E, the number and lumen size of uterine spiral arterioles was reduced in RUPP vs. Preg rats. To confirm the H&E data, immunohistochemistry experiments also showed that the number and lumen size of uterine arterioles were reduced in uterine tissue sections isolated from RUPP vs. Preg rats and stained with endothelial cell marker CD31, smooth muscle marker α-actin, or invading trophoblasts marker CK7 (Fig. 2)

Fig. 2.

Number and lumen size of uterine arterioles in reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Uterine tissue sections from Preg and RUPP rats were stained with hematoxylin and eosin (H&E), endothelial cell marker CD31, smooth muscle marker α-actin, or invading trophoblasts marker cytokeratin-7 (CK7), and composite images of the whole uterine wall were reconstructed as described in the methods. The number of uterine spiral arterioles in the whole uterine wall was counted (A–D), and the average uterine arteriole lumen size (E–H) was determined. Total magnification, ×40. Bar graphs represent means ± SE; n =6 Preg and 5 RUPP rats. Data were analyzed using 2-way ANOVA and Bonferroni’s post hoc correction. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

To determine the depth of uterine vascularization, images of the uterine decidua were divided by imaginary lines into shallow, middle, and deep regions, with the shallow region abutting the placenta attachment site and the deep region representing deep invasion of the uterine wall (Fig. 3). In uterine tissue sections stained with H&E, endothelial marker CD31, smooth muscle marker α-actin, or trophoblast marker CK7, the number and lumen size of uterine spiral arteries were reduced particularly in the middle and deep regions of the decidua in RUPP vs. Preg rats (Fig. 3).

Fig. 3.

Number and lumen size of uterine spiral arterioles in different regions of the decidua in reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Uterine tissue sections from Preg and RUPP rats were stained with hematoxylin and eosin (H&E), endothelial cell marker CD31, smooth muscle marker α-actin, or invading trophoblasts marker cytokeratin-7 (CK7), and composite images of the decidua were reconstructed as described in the methods. Imaginary dashed lines were drawn to divide the decidua into 3 region: shallow, middle and deep. The number of uterine spiral arterioles in the shallow, middle, and deep decidua was counted (A–D), and the average uterine arteriole lumen size in the different regions (E–H) was determined in Preg and RUPP rats. Total magnification, ×40. Bar graphs represent means ± SE; n =6 Preg and 5 RUPP rats. Data were analyzed using 2-way ANOVA and Bonferroni’s post hoc correction. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

Trophoblast invasion of uterine spiral arteries.

To determine the extent of uterine vascular remodeling, the uterine arteriolar wall thickness and the trophoblast invasion and replacement of endothelial and smooth muscle cells in the uterine arteriolar wall were compared in higher magnification images of uterine tissue sections of RUPP and Preg rats. H&E staining and immunostaining appeared to be more condensed around the circumference of the uterine spiral arteries. Consistent with the data in Figs. 2 and 3, the lumen size appeared to be reduced in uterine arterioles of RUPP vs. Preg rats (Fig. 4). Of note, in tissue sections stained with H&E, the uterine arteriolar wall thickness was greater in RUPP vs. Preg rats. Also, endothelial cell marker CD31 and smooth muscle marker α-actin showed more intense staining in uterine arterioles of RUPP vs. Preg rats, suggesting inadequate trophoblast invasion and that the arteriolar endothelial and smooth muscle cells were spared from replacement by invading trophoblasts (Fig. 4). Notably, the trophoblast marker CK7 showed reduced staining in uterine arterioles of RUPP vs. Preg rats, supporting decreased trophoblast invasion of uterine arterioles in RUPP rats (Fig. 4)

Fig. 4.

Wall thickness and extent of trophoblast invasion of uterine spiral arterioles in reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Uterine tissue sections from Preg and RUPP rats were stained with hematoxylin and eosin (H&E), endothelial cell marker CD31, smooth muscle marker α-actin, or invading trophoblasts marker cytokeratin-7 (CK7). Images of uterine tissue sections were acquired, and the uterine arteriolar wall thickness (A), and immunostaining with CD31 (B), α-actin (C), and CK7 (D) were measured. The markers’ immunostaining was more condensed around the circumference of the spiral arterioles. Total magnification, ×400. Bar graphs represent means ± SE; n =6 Preg and 5 RUPP rats. Data were analyzed using 2-way ANOVA and Bonferroni’s post hoc correction. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

Expansive remodeling of uterine artery.

To further test the extent of uterine artery remodeling, we mounted uterine arteries on two cannulas and examined their distensibility and propensity to expand in response to increases in intraluminal pressure. Initial vascular viability tests confirmed responsiveness to high KCl depolarizing solution and suggested a significantly greater (P < 0.05) vasoconstriction to 51 mM KCl in RUPP (66.1 ± 3.1%) vs. Preg rats (41.4 ± 7.80%). The bathing solution was then changed to normal (Ca²⁺ containing) Krebs solution, and the uterine arteries were initially pressurized to small 10 mmHg intraluminal pressure; then, the changes in diameter in response to 10-mmHg stepwise increases in intraluminal pressure to 60 and 70 mmHg were recorded. The initial diameter at 10 mmHg was not significantly different (P = 0.601) in uterine arteries of Preg (394.7 ± 20.3 µm) and RUPP rats (379.1 ± 19.5 µm). In uterine arteries of Preg rats, increasing the intraluminal pressure to 60 and 70 mmHg was associated with marked uterine artery distensibility, expansion, and increases in diameter. In contrast, stepwise increases in intraluminal pressure to test 60 and 70 mmHg were associated with limited distensibility and expansion in uterine arteries of RUPP vs. Preg rats (Fig. 5). Because autoregulation and myogenic response could affect arterial diameter in a Ca²⁺-dependent fashion (8, 27), the bathing solution was changed to Ca²⁺-free (2 mM EGTA) Krebs for 20 min, and the passive changes in uterine artery diameter in response to increases in intraluminal pressure to 60 and 70 mmHg still showed less distensibility and expansion in RUPP vs. Preg rats (Fig. 5),

Fig. 5.

Uterine arterial expansion to increases in intraluminal pressure in reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Uterine arteries from Preg and RUPP were mounted on 2 cannulas, bathed in normal (Ca²⁺ containing) Krebs-Henseleit Buffer (Krebs), and initially pressurized to 10 mmHg, and changes in diameter in response to increases in intraluminal pressure were then recorded (A and B). The intraluminal pressure was increased gradually in 10-mmHg increments to 60 (C) and 70 (D) mmHg, and the percent increase in diameter relative to the initial diameter at 10 mmHg was compared in RUPP vs. Preg rats. To minimize the contribution of autoregulation and Ca²⁺-dependent myogenic response, the bathing solution was changed to Ca²⁺-free (2 mM EGTA) Krebs for 20 min, and the passive changes in arterial diameter in response to stepwise 10-mmHg increases in intraluminal pressure to 60 (E) and 70 (F) mmHg were compared in RUPP vs. Preg rats. Data represent means ± SE; n =5 Preg and 5 RUPP rats. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

MMP levels and distribution.

MMPs are produced by both uterine trophoblasts and vascular cells (16, 28, 39, 70). To test the potential role of MMPs in uterine and vascular remodeling, the levels of MMPs were measured in the uterus and uterine arteries of Preg and RUPP rats. Western blot analysis showed prominent immunoreactive bands corresponding to pro-MMP-2 (72 kDa), MMP-2 (63 kDa), pro-MMP-9 (92 kDa), and MMP-9 (82 kDa) in tissue homogenates of the uterus and uterine arteries of Preg rats. Western blot analysis revealed that the proMMP-2, MMP-2, proMMP-9, and MMP-9 immunoreactive bands were reduced in the uterus and uterine artery of RUPP vs. Preg rats (Fig. 6).

Fig. 6.

Matrix metalloproteinase (MMP) levels in the uterus and uterine artery of reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Uterine and uterine artery tissue homogenates from Preg and RUPP rats were prepared for Western blot analysis using antibodies to MMP-2 (1:500) and MMP-9 (1:500) (A and B). The intensity of immunoreactive bands corresponding to pro-MMP-2 and MMP-2 (C and D) and pro-MMP-9 and MMP-9 (E and F) were analyzed using optical densitometry and normalized to the house keeping protein β-actin to correct for loading. Data represent means ± SE; n =6 Preg and 4 to 5 RUPP rats. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

Gelatin zymography of tissue homogenates of the uterus and uterine artery of Preg rats revealed proteolytic bands corresponding to pro-MMP-2, MMP-2, pro-MMP-9, and MMP-9 (Fig. 7). Analysis of the gelatinolytic bands revealed that that pro-MMP-2, MMP-2, pro-MMP-9, and MMP-9 were reduced in uterus and uterine artery of RUPP vs. Preg rats (Fig. 7).

Fig. 7.

Matrix metalloproteinase (MMP) gelatinolytic activity in the uterus and uterine artery of reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Tissue homogenates of the uterus and uterine artery from Preg and RUPP rats were prepared, and equal protein amounts (1 μg) were loaded for gelatin zymography analysis (A and B). Gelatinolytic bands corresponding to pro-MMP-2 and MMP-2 (C and D) and pro-MMP-9 and MMP-9 (E and F) were measured (in pixel intensity × mm²) and normalized to actin to correct for loading. Data represent means ± SE; n =6 Preg and 4 to 5 RUPP rats. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

We tested whether the tissue distribution and localization of MMPs would parallel that of the observed changes in uterine vascularization and uterine arteriolar remodeling. Immunohistochemistry showed prominent MMP-2 and MMP-9 staining in the uterus and decidua of Preg rats. MMP-2 immunostaining was significantly reduced in the uterus, particularly in the shallow, middle, and deep regions of the decidua of RUPP vs. Preg rats. Also, MMP-9 immunostaining showed significant decreases particularly in the shallow and middle regions of the decidua of RUPP vs. Preg rats (Fig. 8). Both MMP-2 and MMP-9 showed prominent staining in the uterine arteriolar wall of Preg rats. The uterine arterioles appeared to be smaller, and showed significantly less MMP-2 and MMP-9 immunostaining in RUPP vs. Preg rats (Fig. 8).

Fig. 8.

Matrix metalloproteinase (MMP) distribution in uterus and uterine spiral arteries of reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Sections of the uterus of Preg and RUPP rats were prepared for immunohistochemical staining using MMP-2 and MMP-9 antibodies (1:100). Composite images of the whole uterine wall were reconstructed (×40 total magnification; A and B) as described in the methods, and images of uterine spiral arterioles were acquired (×400 total magnification; C and D). For MMP immunostaining of the uterine wall, imaginary dashed lines were drawn to divide the decidua into three regions: shallow, middle, and deep, and the number of pixels in each region was defined and transformed into the area (in mm²) using a calibration bar. The number of brown spots (pixels) representing MMP-2 (A) and MMP-9 (B) in each region was then counted and presented (in pixels/mm²). For MMP staining of spiral arteries, the number of pixels in the uterine arteriole wall was first defined, the number of brown spots (pixels) representing MMP-2 (C) and MMP-9 (D) in the vessel wall was then counted and presented as percentage of total pixels in the vessel wall. Cumulative data in bar graphs were presented directly below their respective representative microscopy images. Bar graphs represent means ± SE; n =6 Preg and 4 RUPP rats. Data were analyzed using 2-way ANOVA and Bonferroni’s post hoc correction. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

Collagen-IV levels and distribution.

To test for the role of MMP substrates in uterine and uterine arterial remodeling, the levels of collagen-IV were examined in uterus and uterine artery of Preg and RUPP rats. Western blot analysis showed detectable bands corresponding to collagen-IV in uterine tissue homogenates of Preg rats. The collagen-IV immunoreactive bands were enhanced in uterus and uterine artery of RUPP vs. Preg rats (Fig. 9). We tested whether the distribution and localization of collagen-IV would parallel the observed changes in uterine vascularization and uterine arteriolar expansive remodeling. Immunohistochemistry showed detectable collagen-IV in the uterus and decidua of Preg rats. Collagen-IV immunostaining was significantly greater in the uterus and uterine arteriolar wall of RUPP vs. Preg rats (Fig. 9).

Fig. 9.

Collagen-IV levels and distribution in uterus and uterine artery of reduced uterine perfusion pressure (RUPP) vs. pregnant (Preg) rats. Tissue homogenate of the uterus (A) and uterine artery (B) from Preg and RUPP rats were prepared for Western blots using antibodies to collagen-IV (1:500). The intensities of the immunoreactive bands corresponding to collagen-IV were analyzed using optical densitometry and normalized to the housekeeping protein β-actin. To determine collagen-IV distribution, cryosections of the uterus of Preg and RUPP rats were prepared for immunohistochemical staining using collagen-IV antibody (1:100). Composite images of the uterine wall were reconstructed (×40 total magnification; C) as described in methods, and images of uterine spiral arterioles were acquired (×400 total magnification; D). For collagen-IV immunostaining of the uterine wall, imaginary dashed lines were drawn to divide the decidua into 3 regions: shallow, middle, and deep, and the number of pixels in each region was defined and transformed into the area (in mm²) using a calibration bar. The number of brown spots (pixels) representing collagen-IV in each region was then counted and presented (in pixels/mm²). For collagen-IV staining of spiral arteries, the number of pixels in the uterine arteriole wall was first defined, the number of brown spots (pixels) representing collagen-IV in the vessel wall was then counted and presented (in %total pixels). Bar graphs represent means ± SE; n =6 Preg and 4 to 5 RUPP rats. Student’s t-test was used for comparison of 2 means. *P < 0.05, RUPP vs. Preg.

DISCUSSION

The main findings are as follows. First, uterine vascularization is reduced, and the uterine spiral arteries are decreased in number and size and are less deep in the decidua of RUPP vs. Preg rats. Second, uterine arteriolar remodeling is decreased, endothelial cells and smooth muscle are spared, and trophoblast invasion is decreased in RUPP vs. Preg rats. Third, the uterine spiral arteries are thicker, and the uterine arteries show reduced propensity to dilate and expand in response to intraluminal pressure in RUPP vs. Preg rats. Finally, the decreased uterine vascularization and uterine arterial expansive remodeling are associated with decreased MMP-2 and MMP-9 and increased collagen-IV deposition in the uterus and uterine arteries, highlighting their conjoint role in placental ischemia and HTN-Preg.

Decreased uterine vascularization in RUPP rat model of HTN-Preg.

Animal models of HTN-Preg have provided important information regarding the mechanisms of HTN-Preg and preeclampsia (33). Consistent with previous reports (4, 25, 85), the present study showed that BP was increased and the litter size and pup weight were decreased in RUPP vs. Preg rats, supporting that the RUPP rat is an appropriate model of HTN-Preg. The present study provided evidence of reduced uterine remodeling and vascularization in HTN-Preg because of the following. First, the uterine weight was reduced in RUPP vs. Preg rats, suggesting decreased uterine remodeling and expansion in HTN-Preg. Second, the uterine arterioles number and size as detected by H&E, the endothelial cell marker CD31, the smooth muscle marker α-actin, and the invading cytotrophoblast marker CK7 were decreased in RUPP vs. Preg rats, suggesting decreased vascularization of the uterine wall. The decreased uterine arterioles number and size in RUPP rats are consistent with reports that progressive placental ischemia could lead to placental rarefaction and attrition and trophoblast loss and deportation into the circulation in preeclampsia (12, 67). Finally, the uterine arterioles number and size were particularly decreased in the middle and deep regions of the decidua in RUPP vs. Preg rats. This is consistent with the concept that in preeclampsia, trophoblasts mainly invade the superficial decidua,and to a less extent the deeper decidua of the uterine wall (76).

Decreased trophoblast invasion of uterine arterioles in HTN-Preg.

During normal pregnancy the trophoblast are thought to invade the wall of the uterine arterioles, replacing endothelial and smooth muscle cells and transforming the vessels to thin-walled and dilated spiral arteries that allow sufficient blood supply to the developing fetus. The present study provided evidence of reduced trophoblast invasion of the spiral arteries in HTN-Preg because of the following. First, the endothelial cell marker CD31 was increased in uterine arterioles of RUPP vs. Preg rats, suggesting less trophoblast invasion and little replacement of the uterine vascular endothelial cell layer in HTN-Preg. We should note that the presence of endothelial cells in the uterine arterioles of RUPP rats may not translate into functional endothelium. In effect, we and others have shown decreased vascular relaxation and impaired endothelial cell function in the aorta and carotid, mesenteric, and renal arteries of RUPP vs. Preg rats (48, 85). Second, the smooth muscle marker α-actin was reduced in uterine arterioles of RUPP vs. Preg rats, suggesting decreased trophoblast invasion and little replacement of the uterine arteriolar vascular smooth muscle layer in HTN-Preg. Finally, in contrast with the increased endothelial cell and smooth muscle staining, the trophoblast marker CK7 was reduced in uterine arterioles of RUPP vs. Preg rats, supporting decreased trophoblast invasion of uterine arterioles in HTN-Preg.

Decreased uterine artery distensibility and expansion in HTN-Preg.

During normal pregnancy, the increased plasma volume and cardiac output are associated with systemic vasodilation and decreased vascular resistance. Pregnancy-induced vasodilation is particularly important in the uterine artery to allow adequate blood supply to the uterus, placenta, and the fetus. This is consistent with the present observations that the spiral arteries of Preg rats showed a large size and lumen and that the uterine artery of Preg rats showed marked distensibility and expansion in response to increases in intraluminal pressure. In contrast, in RUPP rats we observed an increase in wall thickness and a decrease in lumen size of uterine arterioles, which could be partly due to inadequate vascular remodeling and increased ECM deposition, and in turn increased uterine artery rigidity and decreased propensity to expand. This is consistent with the present observation that isolated pressurized uterine arteries showed decreased changes in diameter and limited distensibility and expansion in response to increased intraluminal pressure in RUPP vs. Preg rats. Because changes in intraluminal arterial pressure could also affect the myogenic response and autoregulation mechanisms (8, 27), we measured the passive changes in arterial diameter in Ca²⁺-free medium and still observed less distensibility and expansion in uterine arteries of RUPP vs. Preg rats, supporting structural changes in the uterine artery architecture in RUPP rats. The observed increase in wall thickness and decrease in lumen size of uterine arterioles in RUPP rats could also be partly due to increased vasoconstriction. Doppler screening at 23 wk of pregnancy and detection of early diastolic bilateral uterine artery notching have been suggested as a predictor of increased uterine artery constriction in women that develop preeclampsia (30). Also, experimental studies have shown increases in uterine artery vasoconstriction in pregnant sheep with gestational hypoxia (78). The present observation of increased α-actin smooth muscle staining in uterine arterioles of RUPP rats suggests that they are spared from invasion and replacement by trophoblasts, which would make them more responsive to vasoconstrictor agents in HTN-Preg. This is consistent with our previous reports that vascular contraction to phenylephrine and endothelin-1 is increased in the aorta and carotid, mesenteric, and renal arteries of RUPP vs. Preg rats (48, 85). Interestingly, our present initial viability tests of the isolated uterine arteries suggest that KCl (51 mM)-induced constriction is greater in uterine arteries of RUPP vs. Preg rats. Future studies should further examine the detailed changes in vascular contraction to various vasoconstrictor stimuli in uterine arteries during HTN-Preg.

Decreased uterine and uterine arterial MMPs in HTN-Preg.

MMPs are major regulators of tissue remodeling, and trophoblasts and vascular cells are major sources of MMPs (16, 28, 39, 70). The gelatinases MMP-2 and MMP-9 are thought to be involved in ECM degradation and facilitation of trophoblast invasion of the spiral arteries during normal pregnancy. MMP-2 and MMP-9 have been shown to play a role in endometrial tissue remodeling during the estrous and menstrual cycles and during pregnancy (50, 73, 83). MMP-2 is the main collagenolytic enzyme in umbilical cord artery (5), and serum MMP-9 levels are elevated in normal Preg women (51). Also, in cultured trophoblasts, suppression of MMP-9 expression inhibits their invasive capability, supporting a role of MMP-9 in modulating trophoblast invasion (81). We have previously shown increases in MMP-2 and MMP-9 in the aorta and uterus of mid- and late-Preg compared with nonpregnant rats (15), supporting that MMP-2 and MMP-9 play a role in the vascular and uterine remodeling during healthy pregnancy (15, 31, 68, 79). While some studies showed increases in circulating MMP-2 and MMP-9 (19), other studies showed a decrease in serum MMP-9 in preeclamptic vs. normal Preg women (51). The present study suggests that the decreased uterine vascularization and uterine arterial expansion in HTN-Preg could involve changes in MMP-2 and MMP-9 because of the following. First, Western blot analysis showed decreases in proMMP-2, MMP-2, proMMP-9, and MMP-9 levels in the uterus and uterine arteries of RUPP vs. Preg rats. Second, gelatin zymography showed decreases in proMMP-2, MMP-2, proMMP-9, and MMP-9 gelatinolytic activity in the uterus and uterine arteries of RUPP vs. Preg rats. Third, immunohistochemistry showed decreases in MMP-2 and MMP-9 immunostaining in the uterine decidua and uterine spiral arteries of RUPP vs. Preg rats. These observations are consistent with our previous report that MMP-2 and MMP-9 levels are reduced in the aorta, uterus, and placenta of RUPP vs. Preg rats (41). These observations are also in agreement with other reports that MMP levels are decreased in umbilical cord artery and microvascular endothelial cells of preeclamptic women (21, 22) and that pharmacological inhibition of MMPs is associated with decreased placenta weight and IUGR in both normal Preg and HTN-Preg rats and reduces trophoblast invasion and placental perfusion in HTN-Preg rats (57, 65, 80). Together these observations suggest that the decreased uterine vascularization and uterine arterial expansion are related to decreased uterine and uterine arterial MMP-2 and MMP-9. Of note, MMPs also promote angiogenesis by detaching pericytes from the vessels undergoing angiogenesis, releasing ECM-bound angiogenic growth factors, exposing cryptic proangiogenic integrin-binding sites in ECM, generating promigratory ECM component fragments, and cleaving endothelial cell-cell adhesions (60). Thus, the decrease in MMP-2 and MMP-9 could also decrease angiogenesis, trophoblast invasion of spiral arteries, and uteroplacental vascularization and thereby contribute to decreased fetoplacental growth and the observed IUGR in HTN-Preg rats.

In addition to their proteolytic effects on ECM, MMPs may affect vascular and uterine function and the mechanisms of smooth muscle contraction. We have previously shown that MMP-2 and MMP-9 cause relaxation of precontracted rat aorta (11) and uterus (79) and induce vasodilation in rat inferior vena cava via hyperpolarization and activation of K⁺ channels (62, 64). Other studies have shown that MMP-2 and MMP-9 increase the production of vasoconstrictor peptides such as endothelin-1 and decrease vasodilator peptides such as adrenomedullin (1, 20, 53), leading to endothelial dysfunction and imbalance between vasodilator and vasoconstrictor factors. We and other groups have shown that vascular contraction is enhanced in RUPP rats (14, 48, 85), and the observed decreases in vascular MMP-2 and MMP-9 could reduce uterine arteriolar expansion, enhance vasoconstriction, and aggravate placental ischemia, while changes in uterine MMPs could affect uterine contraction and trigger premature labor (41, 79).

Increased Collagen-IV in HTN-Preg.

In search for the MMP downstream targets that could influence uterine remodeling, we investigated possible changes in MMP substrates. MMPs degrade different substrates including gelatin, collagen, and other ECM proteins (36, 63, 77). We have previously shown an increase in total collagen content with no change in elastin in the aorta, uterus, and placenta of RUPP vs. Preg rats (41). However, collagen has 18 types and different subtypes (23). Gelatinases mainly degrade collagen-IV, although MMP-2 can also degrade collagen-I, II, III, V, VII, X, and XI and MMP-9 can degrade collagen-V, VII, X, and XIV (36, 63, 77). Thus, while MMP-2 is a gelatinase, it can also function as interstitial collagenase, acting much like MMP-1 but in a weaker manner (52). MMP-2 can degrade collagen in two stages: interstitial collagenase-like degradation followed by gelatinolysis promoted by the fibronectin-like domain (2). MMP-9 can also function as a collagenase where it binds the α₂-chains of collagen-IV with high affinity even when it is inactive and makes the substrate readily available, then as a gelatinase (55). The present study showed increases in collagen-IV level and distribution in the uterus and uterine arterial wall of RUPP vs. Preg rats. Because MMPs facilitate cell growth and migration by promoting proteolysis of ECM, the decreased MMP-2 and MMP-9 and the consequent increase in collagen deposition in RUPP tissues could impede cell growth, proliferation, and migration and in turn interfere with trophoblast invasion of the decidua and remodeling of spiral arteries. The increased collagen deposition in uterine vessels of RUPP rats could also be an important factor in the observed decrease in uterine arterial expansive remodeling.

Other observations/considerations.

Other observations and considerations include the following. First, MMP is a large family of at least 28 proteolytic enzymes (36, 63, 77). While we examined MMP-2 and MMP-9, other MMPs have been detected in the aorta, uterus, and placenta, and the changes in these MMPs in HTN-Preg need to be examined. For instance, trophoblast- and vascular smooth muscle-derived MMP-12 mediate proteolysis and spiral artery remodeling during pregnancy (28). Also, MMP activity could be influenced by other MMP activators and inhibitors. In effect, some MMPs may cleave other pro-MMPs, and membrane-type-1 MMP is a key activator of pro-MMP-2 (59, 63, 77). Also, tissue inhibitors of metalloproteases (TIMPs) are endogenous modulators of MMPs (51, 59, 77), and the changes in vascular and uteroplacental TIMPs in HTN-Preg need to be examined. Second, in the present study, we measured uterine vascularization and MMPs on gestational day 19, and one might argue that the changes in uterine vascularization and MMPs are solely due to the initial physical RUPP. This is unlikely as the changes in BP in the RUPP model do not occur immediately after banding of the vessels on gestational day 14 and the increased BP is only more robust on day 19 (26, 48). We propose that the initial RUPP is only an initiating event that triggers a cascade of biochemical and pathological events that lead to progressive placental ischemia and the different maternal and fetal manifestations associated with HTN-Preg and preeclampsia. The initial physically induced RUPP is thought to cause initial placental ischemia, which would increase the release of antiangiogenic and other cytoactive factors, leading to decreased MMPs, accumulation of collagen, further RUPP, progressive placental ischemia, and an incessant cycle that would lead to further decreases in angiogenesis and cause HTN-Preg and IUGR. Future time-course studies should determine the sequence of events between RUPP, release of antiangiogenic and cytoactive factors, changes in MMP levels/activity, collagen accumulation and impaired uterine vascularization, and uterine artery remodeling during the course of pregnancy, as well as the reversal of these events in the postpartum period. Third, while plasma levels of MMPs may show increase during pregnancy (49), MMPs are released from different maternal tissues, and plasma MMP levels reflect global changes in MMPs in multiple tissues. In the present study we showed decreases in uterine and uterine arterial MMP-2 and MMP-9 in HTN-Preg rats. The changes in MMPs expression/activity and their effects on the structure and function of other tissues, e.g., the small resistance vessels that control BP, need to be examined. Fourth, placental ischemia causes the release of antiangiogenic factors and other vasoactive factors including soluble fms-like tyrosine kinase-1 (sFlt-1) (32), soluble endoglin (sEng) (26), inflammatory cytokines such as tumor necrosis factor-α (TNFα) and interleukin-6 (IL-6) (37, 38, 82), reactive oxygen species (24), and hypoxia-inducible factor (3, 71). Interestingly, reactive oxygen species and hypoxia-inducible factors affect MMP expression/activity (7, 42) and their potential interaction with angiogenic factors in modulating uterine vascularization and uterine arteriolar remodeling, and MMP expression/activity in HTN-Preg should be examined in future studies. Finally, the present study in late pregnant rats with RUPP suggests that decreased uterine vascularization and uterine arterial expansive remodeling by MMP-2 and MMP-9 could be a factor in the progressive uteroplacental ischemia associated with HTN-Preg. Whether changes in MMPs participate in the inadequate placentation and the changes in trophoblast invasion that occur very early in human pregnancy before any apparent RUPP, and which are thought to precede and be the root cause of RUPP, are less clear and need to be further examined.

Perspective.

Understanding the role of uterine vascularization and the driving enzymes and ECM proteins should better help understand the factors contributing to placental ischemia during HTN-Preg and in turn help in the prediction, prevention, and management of preeclampsia. The present study demonstrated a decrease in uterine vascularization and uterine spiral arterioles number and size and increased BP and IUGR in the RUPP rat model of HTN-Preg. The study also showed decreased remodeling and trophoblast invasion, increased wall thickness of uterine arterioles, and decreased propensity of uterine arteries of RUPP rats to dilate and expand. The decreased uterine vascularization and uterine arterial expansive remodeling were associated with decreases in MMP-2 and MMP-9 levels and distribution and increases in the MMP substrate collagen-IV in the uterus and uterine arteries. The decreased uterine vascularization and uterine arterial expansive remodeling by MMP-2 and MMP-9 could be a factor in the progressive uteroplacental ischemia associated with HTN-Preg and preeclampsia. Progressive placental ischemia could also contribute to placental rarefaction and trophoblast deportation into the circulation in the pathogenesis of preeclampsia (12, 67). Advanced imaging techniques could help determine the extent of uterine vascularization and predict uteroplacental ischemia in early stages of HTN-Preg and preeclampsia. MMP modulators that increase expression/activity of MMP-2 and MMP-9 could be useful in improving uterine vascularization and uterine arteriolar remodeling and in turn BP and IUGR in the management of preeclampsia.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-65998, HL-111775, and R56-HL-147889. C. Lin was funded from Military Logistics Research Project Grant CLB18J035 and Natural Science Foundation (of Fujian Province) Grant 2017J01327.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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