Uteroplacental insufficiency programs regional vascular dysfunction and alters arterial stiffness in female offspring

Authors


Corresponding author M. Tare: Department of Physiology, Monash University, Clayton, Victoria 3800, Australia.  Email: Marianne.Tare@monash.edu

Abstract

Intrauterine growth restriction caused by uteroplacental insufficiency increases the risk of cardiovascular disease in adulthood. Vascular mechanisms in female offspring are poorly understood. The aim of this study was to investigate the effects of uteroplacental insufficiency on blood pressure, vascular reactivity and arterial stiffness in four vascular beds in female offspring born growth restricted. Uteroplacental insufficiency was induced on day 18 of gestation in Wistar Kyoto rats by bilateral uterine vessel ligation (Restricted) or sham surgery (Controls). Wire and pressure myography were used to test endothelial and smooth muscle function, and passive mechanical wall properties, respectively, in uterine, mesenteric, renal and femoral arteries of 18-month-old female offspring. Collagen and elastin fibres were quantified using circular crossed-polarized light microscopy and quantitative real time polymerase chain reaction. Restricted female offspring were born 10–15% smaller. Restricted females were normotensive, had plasma triglycerides 2-fold elevated and had uterine endothelial dysfunction, attributed to a 23% reduction in the maximal relaxation produced by endothelium-derived hyperpolarizing factor. Uterine artery stiffness was increased, with an augmented proportion of thick and decreased proportion of thin collagen fibres. Vascular reactivity and mechanical wall properties were preserved in mesenteric, renal and femoral arteries in growth restricted females. Female offspring born growth restricted have selective uterine artery endothelial dysfunction and increased wall stiffness. The preserved vascular function in other arteries may explain the lack of hypertension in these females. The uterine artery specific dysfunction has potential implications for impaired pregnancy adaptations and a compromised intrauterine environment of the next generation.

Abbreviations 
CCLP

circular crossed-polarized light

C t

cycle threshold

ECM

extracellular matrix

EDHF

endothelium-derived hyperpolarizing factor

eNOS

endothelial nitric oxide synthase

E max

maximal response

Indo

indomethacin

MMP-2

matrix metalloproteinase 2

MMP-9

matrix metalloproteinase 9

NO

nitric oxide

pD2

−log EC50

PE

phenylephrine

PSS

physiological saline solution

PGI2

prostacyclin

KCa2.3

small-conductance calcium-activated potassium channel

KCa3.1

intermediate-conductance calcium-activated potassium channel

SNP

sodium nitroprusside

TGF-β1

transforming growth factor β1

TIMP-1

tissue inhibitor of metalloproteinase 1

TIMP-2

tissue inhibitor of metalloproteinase 2

Introduction

Intrauterine growth restriction occurs in approximately 7–10% of human pregnancies, and is associated with an increased incidence of perinatal morbidity and mortality (McIntire et al. 1999). Epidemiological studies have shown a strong association between low birth weight, an indicator of intrauterine growth restriction, and the risk of higher blood pressure and cardiovascular disease later in life (Barker et al. 1989; Law & Shiell, 1996). The ‘developmental origins hypothesis’ proposes that suboptimal conditions in utero and during a period of rapid developmental plasticity produce adaptive changes in organ structure and physiological function to ensure survival of the fetus, but may be detrimental for long term postnatal health (Gluckman et al. 2008). These adaptations in early life, including those that occur in blood vessels, may lead to persistent alterations in arterial function that contribute to the manifestation of cardiovascular diseases in later life (Martyn & Greenwald, 2001).

Endothelial and smooth muscle function and arterial mechanical properties are important regulators of vascular function. Low birth weight humans tend to exhibit impaired endothelium-dependent vasodilatation that is already present in term infants (Martin et al. 2000a), and persists throughout childhood (Martin et al. 2000b) and into adulthood (Leeson et al. 2001). Individuals born small have premature stiffening of the carotid arteries (Martin et al. 2000b), increased aortic and coronary thickness (Skilton et al. 2005; Pesonen et al. 2006), reduced compliance in the conduit arteries of the trunk and legs (Martyn & Greenwald, 2001) and altered vascular dimensions, including smaller aortic diameter (Brodszki et al. 2005).

Intrauterine growth restriction, in many cases, occurs as a result of poor placental function leading to a decrease in uteroplacental perfusion and nutrients and oxygen to the fetus. Several animal models of fetal programming have been used to investigate the underlying vascular mechanisms and cardiovascular diseases later in life. In rats, maternal undernutrition during pregnancy results in reduced fetal growth, elevated blood pressure, vascular endothelial dysfunction (Franco et al. 2002; Brawley et al. 2003) and altered smooth muscle reactivity to vasoconstrictors and vasodilators in adult offspring (Ozaki et al. 2001).

In Western societies, uteroplacental insufficiency is the common cause of intrauterine growth restriction. There are only a few studies describing the effects of intrauterine growth restriction due to uteroplacental insufficiency on adult vascular function. Exposure to chronic maternal hypoxia is associated with impaired fetal growth and altered responsiveness to vasoconstrictors in postnatal rats (Williams et al. 2005) and endothelial dysfunction in adults (Morton et al. 2010). Uteroplacental insufficiency can also be induced in rats by clipping both the abdominal aorta and ovarian arteries in mid-to-late gestation (Alexander, 2003; Anderson et al. 2006). This model causes a rise in maternal blood pressure. The growth restricted male and female offspring have elevated blood pressure and enhanced vasoconstrictor responsiveness in mesenteric arteries (Anderson et al. 2006). Female offspring have altered endothelium-dependent and -independent relaxation in mesenteric arteries (Anderson et al. 2006), while male offspring have impaired endothelial vasodilator and/or altered smooth muscle reactivity in the aorta (Payne et al. 2003). However, the extent of vascular dysfunction across different regions of the vasculature is unknown. Furthermore, the effects of uteroplacental insufficiency on arterial wall stiffness and structural wall properties in female offspring have not been explored.

We have used a rat model of bilateral uterine vessel (artery and vein) ligation to induce uteroplacental insufficiency. Growth restricted male offspring develop hypertension and exhibit impaired endothelium-dependent and -independent smooth muscle relaxation and increased wall stiffness in mesenteric and femoral arteries in adulthood (Parkington et al. 2006). The males have a reduced nephron endowment and glomerular hypertrophy in adulthood (Wlodek et al. 2008). The same prenatal insult may result in markedly different disease outcomes for male and female offspring (Denton & Baylis, 2007). We have previously shown that growth restricted female offspring do not become hypertensive, despite having a nephron deficit and with ageing develop modest renal insufficiency and glomerular hypertrophy (Moritz et al. 2009).

The aim of this study was to investigate the effects of late gestation uteroplacental insufficiency on vascular smooth muscle and endothelial function, arterial stiffness and structural wall properties in four vascular beds in growth restricted female rat offspring. Ageing amplifies cardiovascular risk, particularly in females, and therefore we studied these rats at 18 months of age.

Methods

Ethics

All procedures were approved by The University of Melbourne Pharmacology, Physiology, Biochemistry & Molecular Biology and Bio21 Institute Animal Ethics Committee. All of the authors have read the article ‘Reporting ethical matters in The Journal of Physiology: standards and advice’ (Drummond, 2009) and our experiments comply with the policies and regulations.

Animals

Wistar Kyoto rats (9–13 weeks of age) were mated and on day 18 of gestation, pregnant rats were randomly divided into the Restricted (uteroplacental insufficiency) and Control (sham surgery) groups and surgery performed (Wlodek et al. 2007; O’Dowd et al. 2008). For surgery, all rats were anaesthetized with an intraperitoneal injection of a mixed solution containing ketamine (50 mg (kg body weight)−1) and ilium xylazil-20 (10 mg (kg body weight)−1). The Restricted group underwent bilateral uterine vessel (artery and vein) ligation to induce uteroplacental insufficiency. Sham surgery was performed in the Control group, as described previously (Wlodek et al. 2007; O’Dowd et al. 2008). All pregnant rats delivered naturally at term (22 days).

Control and Restricted (growth restricted) pups were cross-fostered the day after birth onto different Control and Restricted mothers (Wlodek et al. 2007; Siebel et al. 2008). One to three female offspring from each litter were studied per group at 18 months. The Control group consisted of nine females from five mothers and Restricted group had 11 females from five mothers.

Body weights and dimensions and mean arterial pressure

Female pups were weighed and measured (Siebel et al. 2008) on postnatal day 1, 3, 14 and 35 and at 6 and 18 months (postmortem). At 18 months conscious mean arterial pressure (MAP) and heart rate were measured using an indwelling tail artery catheter (Moritz et al. 2009).

Tissue and blood collection

Rats were used for experimentation during oestrus, as determined by vaginal smears. At least one and usually two or more complete oestrous cycles (4–5 days in length) were followed before a rat was used for an experiment. At postmortem, rats were anaesthetized with an intraperitoneal injection of a mixed solution containing ketamine (100 mg (kg body weight)−1) and ilium xylazil-20 (30 mg (kg body weight)−1), blood samples were taken and then the rats were killed by exsanguination. Organ and tissue weights were recorded. Blood samples were collected by cardiac puncture (non-fasted) or by tail vein (fasted, at 12 months) and plasma was frozen (−80°C) for later analysis.

Small mesenteric, renal, femoral and uterine arteries were isolated and used for functional studies. Remaining uterine artery was immediately fixed with 10% neutral buffered formalin for histological analysis or snap frozen and stored at −80°C for later molecular analysis.

Biochemical analyses

At 18 months, plasma (non-fasted) total cholesterol, triglycerides and free fatty acids were measured in duplicate by colorimetric enzymatic analysis on an automated centrifugal analyser (COBAS Mira; Roche Diagnostics Corp., Indianapolis, IN, USA). Plasma (non-fasted) insulin was measured using a radioimmunoassay (RIA) kit (Siebel et al. 2008). Plasma oestradiol concentrations were measured using an ultrasensitive RIA kit (Diagnostic Systems Laboratory, Webster, TX, USA) with detection limits of 0.06 nmol l−1 and 8.1 pmol l−1, respectively. Fasting blood glucose was measured at 12 months from tail vein blood sample using a glucometer (Advantage Blood Glucose Monitor, Roche Diagnostics Australia Pty Ltd, Castle Hill, NSW, Australia).

Vascular function

Rings of mesenteric (outside diameter (o.d.) ∼300 μm), renal (o.d. ∼400 μm), femoral (o.d. ∼400 μm), and uterine artery (o.d. ∼350 μm) ∼1–2 mm in length, were mounted on a four channel wire myograph (Model 610M, Danish Myo Technology, Aarhus, Denmark) for measurement of isometric tension, as previously described (Bubb et al. 2007). Arteries were bathed in warmed (36°C) physiological saline solution (PSS (mm): 120 NaCl, 5 KCl, 2.5 CaCl2, 25 NaHCO3, 11 glucose, 1 KH2PO4, 1.2 MgSO4) bubbled with 95% O2 and 5% CO2. Endothelial function was tested in arteries submaximally constricted with the α1-adrenoceptor agonist phenylephrine (PE). All vessels that relaxed completely to endothelial stimulation (maximal relaxation: uterine and mesenteric, 100%; renal, 90%; femoral, 80%) with acetylcholine (ACh) were used for experimentation.

Endothelium-dependent relaxation Arteries were submaximally constricted with PE (10−6–3 × 10−6m). The level of preconstriction was not different between arteries from Control and Restricted groups. The endothelium was stimulated with increasing concentrations of ACh (10−9–10−5m) added cumulatively for 2 min each. Responses were obtained before, and after blockade of endothelial nitric oxide synthase (eNOS) with Nω-nitro-l-arginine methyl ester (l-NAME; 2 × 10−4m) and cyclooxygenase with indomethacin (Indo; 10−6m) (Tare et al. 2000). Relaxations persisting in the presence of l-NAME and Indo were attributed to endothelium-derived hyperpolarizing factor (EDHF) (Tare et al. 2000).

Smooth muscle function Arteries were exposed to PE (10−9–10−4m) applied cumulatively. Contractions were expressed as a percentage of that evoked by high K+ PSS (isotonic replacement of Na+ with 100 mm K+). Endothelium-independent relaxation was tested using cumulative addition of the nitric oxide (NO) donor sodium nitroprusside (SNP, 10−9–10−5m).

Arterial passive mechanical wall properties

Leak free segments of artery were mounted on a pressure myograph (Living Systems Instrumentation, Burlington, VT, USA) with no lumenal flow (Wigg et al. 2004). Arteries were superfused at ∼15 ml min−1 in zero-Ca2+ PSS containing 1 mm EGTA at ∼36°C. Arteries were pressurized from 0 to 200 mmHg, in 10 mmHg increments and dimensions (length, outside diameter and wall thickness) measured at each increment. Wall stress and strain were derived as follows: wall stress (kPa) =[intraluminal pressure (Pa) × internal diameter (μm)]/[2 × wall thickness (μm)]; wall strain = (internal diameter − internal diameter extrapolated to 5 mmHg pressure)/internal diameter extrapolated to 5 mmHg pressure (Wigg et al. 2004; Bubb et al. 2007). For normalization of outside arterial diameters, values were expressed as a percentage of diameter at 5 mmHg.

Quantitative histological examination of collagen and elastin in uterine arteries

Segments of uterine artery were fixed in 10% neutral buffered formalin solution, deparaffinized and cut at 5 μm thickness. Collagen and elastin fibres were stained with picrosirius red (Sirius red F3B) or Gomori fuschin aldehyde, respectively (Junqueira et al. 1979). Artery sections were then mounted in dibutyl phthalate xylene (DPX) mounting medium and coverslipped. An equal number of Control and Restricted uterine artery samples were stained and photographed at the same time under identical conditions, using the same magnification and light intensity. Images were analysed by two independent examiners blinded to the study group. Sections were examined under a polarizing microscope (Olympus BX51 Pol) using a strain free dry objective lens (Olympus Uplan Apo FLP 40×/0.75NA). Two configurations were used: a conventional bright-field illumination and a circular crossed-polarized light (CCPL) set-up. The microscope was modified using two identical quarter wave plates (Olympus U-TP137) in addition to the conventional crossed polarizers to allow the CCPL observation (Bennet, 1961; Coats et al. 1997). Picrosirius red stain has birefringent properties that naturally enhances the collagen birefringence when viewed under CCPL, permitting the identification of thick (red to orange to yellow) and thin (green) collagen fibres (Junqueira et al. 1979; Whittaker et al. 1994).

Collagen and elastin content were calculated in the medial and adventitial areas separately. A high definition colour digital camera (Nikon DS-Fi1; Nikon, Tokyo, Japan) was used for image collection and further post-processing of images was performed using ImageJ software (v.1.41., National Institutes of Health, Bethesda, MD, USA).

Bright field data acquisition Original images were background subtracted, converted to an 8-bit greyscale and made binary (black and white). Quantification of total collagen and elastin content was calculated as a percentage of the area (expressed in mean fraction pixels) in the regions analysed.

Circular crossed-polarized light data acquisition The proportion of thick and thin collagen fibres influences mechanical wall stiffness (Jalil et al. 1989; Rowe et al. 2003). We assessed the proportion of different coloured fibres using published methods (Rich & Whittaker, 2005; Aikawa et al. 2006). To determine the proportion of different coloured collagen fibres, images were background subtracted and then resolved into hue, saturation and value components (using ImageJ software). A histogram of only the hue frequency was obtained from the resolved 8-bit hue images which contained 256 possible colours. The pixels were then separated out according to the hue definitions: red 2–9 and 230–256, orange 10–38, yellow 39–51, and green 52–128 (Rich & Whittaker, 2005). We determined the proportion of pixels within each hue range and expressed this as a percentage of the total number of birefringent collagen pixels which, in turn, was expressed as a percentage of the total number of birefringent collagen pixels in the regions analysed. Finally, results obtained for the number of red, orange and yellow containing pixels were added together to estimate the total number of thick fibres and this was compared with the total number of thin (green) collagen fibres.

Gene expression analysis in uterine arteries

Isolation of total RNA from pooled frozen uterine arteries was conducted using Tri Reagent. Frozen arteries were placed in pre-chilled capsules and pulverized with a steel ball using a Wig-L-Bug amalgamator (Crescent Digital Wig-L-Bug, Dentsply-Rinn, Elgin, IL, USA) as previously described (Novak et al. 2006). The quantity and quality of the total RNA was measured spectrophotometrically using a NanoDrop ND1000 (NanoDrop Technologies, Wilmington, DE, USA). One microgram of total RNA was reversed transcribed into first-strand cDNA using Invitrogen Superscript III RT. We performed quantitative real-time polymerase chain reaction (qRT-PCR) using a Rotorgene 3000 (Corbett Research, Mortlake, VIC, Australia) as previously described (Moritz et al. 2009). Fluorescence-based real-time PCR primers and TaqMan probes were designed using software and produced by Biosearch Technologies (Biosearch Technologies, Novato, CA, USA). The primer-probe design strategy was to situate primers/probes within the protein-coding region, exon spanning where possible to avoid genomic DNA contamination. The primer and probe sequences and concentrations used for qRT-PCR have been reported previously (Moritz et al. 2009) or as follows: Elastin forward primer (5′–3′) GGCATCGGTGGCTTAGGA, reverse primer (5′–3′) CTCCGACTCCAGCTCCAAGT, TaqMan Probe (5′–3′) (FAM)-TCAACAGGTGCTGTGGTGCCTC-(BHQ1); small-conductance calcium-activated potassium channel (KCa2.3) forward primer (5′–3′) GGGTGTCAAGATGGAACAAAGGAAG, reverse primer (5′–3′) GCTCCGTGATCAAGTCATACATG, TaqMan probe (5′–3′) (FAM)-TGAGTGACCAAGCCAACACCCTG-(BHQ1); intermediate-conductance calcium-activated potassium channel (KCa3.1) forward primer (5′–3′) CGCCAGGTACGGCTGAAAC; reverse primer (5′–3′) GCAGGTCGCACAGGATCAT, TaqMan probe (5′–3′) (FAM)-CGGAAGCTCCGGGAACAAGTGAA-(BHQ1). Optimal concentrations for all primers and probes were 300 and 100 nm, respectively. Relative quantification of gene expression was performed by the comparative CT (ΔΔCT) method with ribosomal 18S as the endogenous control. 18S expression in the uterine artery was not different between Control and Restricted groups as verified by the cycle threshold values (Ct) (P= 0.4).

Drugs and chemicals The following agents were used: acetylcholine, indomethacin, Nω-nitro-l-arginine methyl ester, phenylephrine (Sigma Aldrich, Castle Hill, NSW, Australia); ilium xylazil-20 (Troy Laboratories, Pty Ltd, Smithfield, NSW, Australia); ketamine (Parnell Laboratories, Pty Ltd, Alexandria, NSW, Australia); primers and probes (Biosearch Technologies, Novato, CA, USA); sodium nitroprusside (Ajax Chemicals, Auburn, SA, Australia); Sirius red F3B and DPX (BDH Laboratory Supplies, Poole, UK); Superscript III RT (Invitrogen, VIC, Australia); Tri Reagent (Ambion, Inc., Austin, TX, USA).

Statistical analysis

Data are expressed as means ± standard error of the mean (s.e.m.) with n representing the number of litters. Contraction and relaxation were analysed as previously described (Tare et al. 2000). Concentration–response curves were constructed using Prism (v.5.0; GraphPad Software, San Diego, CA, USA) and sigmoidal curves fitted to the data using the least squares method. Contractions to PE were expressed as a percentage of contraction to high K+ PSS. Relaxations evoked by ACh and SNP were expressed as a percentage of contraction evoked by PE. The concentration of drug required to evoke a half-maximal response (EC50) was determined and the pD2 (−log EC50) and maximum response (Emax) were compared between groups. In the renal and femoral arteries the higher concentrations of ACh tended to cause contraction, and thus the pD2 values (−log EC50) for relaxation were derived from sigmoidal curves fitted to the relaxation component of the dose response. Stress–strain relationships were analysed as previously described (Wigg et al. 2004). Data were analysed using a general linear model univariate ANOVA, ‘weighted by litter size’ (to account for measurement on multiple offspring from the same dam) using least squares regression (Festing, 2006), using SPSS (v18, SPSS Inc., Chicago, IL, USA) and differences were accepted as statistically significant when P < 0.05.

Results

Offspring characteristics and plasma analysis

Uteroplacental insufficiency resulted in a reduction in litter size (P < 0.05; Table 1). Restricted females were lighter than controls for the 6 days of postnatal life (days 1–6; by 10–17%, P < 0.05; Table 1). By postnatal day 14, both groups were of similar weight (Table 1) and were not different at 18 months (Table 1). At 18 months, body dimensions and organ weights were not different between the groups (data not shown). MAP and heart rate were not different between groups (Control 120 ± 2 mmHg, n= 5; Restricted 128 ± 4 mmHg, n= 5; Control 305 ± 8 beats per minute (bpm), n= 5; Restricted 305 ± 5 bpm, n= 5). Similarly, there were no differences in systolic (Control 137 ± 4 versus Restricted 143 ± 6 mmHg) or diastolic (Control 105 ± 2 versus Restricted 111 ± 4 mmHg) pressures. Non-fasted total plasma triglycerides were twofold higher in Restricted females compared with Controls at 18 months of age (P < 0.05; Table 1). Non-fasted plasma total cholesterol, free fatty acids, insulin and oestradiol were not different between groups (Table 1). Fasted blood glucose at 12 months was not different between groups (Control, 5.96 ± 0.14 vs. Restricted, 5.72 ± 0.17 mmol l−1, n= 5 litters per group).

Table 1.  Litter size, body weights and biochemical analyses
MeasurementsControlRestricted
  1. *P < 0.05 between Control and Restricted offspring. n= 5 litters per group.

Litter Size8.6 ± 0.66.2 ± 0.7*
Body Weight (g)
 Day 14.2 ± 0.13.4 ± 0.1*
 Day 35.6 ± 0.34.6 ± 0.3*
 Day 69.0 ± 0.67.4 ± 0.4*
 Day 1422.5 ± 1.019.3 ± 0.9
 Day 3575.4 ± 2.773.1 ± 2.1
 6 months247.8 ± 7.5241.6 ± 4.2
 18 months289.8 ± 8.6278.6 ± 2.8
Plasma parameters
 Total cholesterol (mmol l−1)3.67 ± 0.254.23 ± 0.27
 Triglycerides (mmol l−1)1.40 ± 0.232.73 ± 0.15*
 Free fatty acids (mEq l−1)2.14 ± 0.382.45 ± 0.20
 Insulin (ng ml−1)0.54 ± 0.120.51 ± 0.10
 Oestradiol (pg ml−1)10.0 ± 0.5013.1 ± 0.50

Vascular function

Endothelium-dependent relaxation Stimulation of the endothelium with ACh evoked concentration-dependent relaxation in uterine, mesenteric, renal and femoral arteries. In the absence of blockers, total endothelium-dependent relaxation was not different between Control and Restricted groups for any of the arteries tested (Table 2, Fig. 1). In the renal artery, the higher concentrations of ACh (10−6–10−5m) evoked contraction, with no differences between groups (Fig. 1C). The relaxation attributed to EDHF was revealed following the blockade of eNOS and cyclooxygenase with l-NAME and Indo, respectively. In the presence of l-NAME and Indo, concentration–relaxation curves to ACh were shifted to the right in the uterine, mesenteric, and renal arteries for both Control and Restricted groups (Fig. 1A, B and C). In the uterine arteries, maximal relaxation attributed to EDHF was significantly reduced (by 23%) in Restricted compared with Controls (P < 0.05; Table 2, Fig. 1A), with no differences in pD2 values (Table 2). EDHF-mediated relaxation was unchanged in mesenteric and renal arteries (Table 2, Fig. 1B and C). In femoral arteries, ACh failed to evoke relaxation in the presence of l-NAME and Indo, reflecting the absence of EDHF-mediated relaxation in this artery (Fig. 1D) (Wigg et al. 2001).

Table 2.  Vascular function in Control and Restricted female offspring
 ControlRestricted
 MesentericRenalFemoralUterineMesentericRenalFemoralUterine
  1. % K, percentage of contraction to 100 mmol l−1 K+ PSS. *P < 0.05 between Control and Restricted female offspring.

  2. — indicates no response. n= 5 litters per group.

Smooth muscle function
 100 mmol l−1 K (mN mm−1)8.0 ± 0.29.4 ± 0.410.2 ± 0.711.6 ± 0.57.8 ± 0.610.0 ± 0.310.3 ± 0.512.0 ± 0.2
 PE, Emax (%K)93 ± 3106 ± 3112 ± 395 ± 493 ± 5100 ± 2104 ± 298 ± 2
 SNP, Emax (%)97 ± 185 ± 354 ± 598 ± 198 ± 0.386 ± 458 ± 596 ± 0.4
Endothelium-dependent relaxation (Emax%)
 ACh, Emax100 ± 0.193 ± 185 ± 492 ± 398 ± 192 ± 388 ± 594 ± 2
 ACh+l-NAME+Indo, Emax96 ± 165 ± 454 ± 596 ± 163 ± 631 ± 3*
pD2
 PE5.30 ± 0.036.06 ± 0.026.50 ± 0.065.94 ± 0.075.24 ± 0.045.91 ± 0.046.55 ± 0.035.79 ± 0.05
 SNP7.71 ± 0.037.63 ± 0.057.20 ± 0.127.90 ± 0.067.71 ± 0.047.68 ± 0.067.34 ± 0.127.94 ± 0.04
 ACh8.40 ± 0.027.93 ± 0.087.36 ± 0.047.80 ± 0.088.40 ± 0.037.90 ± 0.057.42 ± 0.047.88 ± 0.09
 ACh+l-NAME+Indo7.26 ± 0.027.56 ± 0.147.46 ± 0.167.31 ± 0.057.37 ± 0.106.92 ± 0.18
Media thickness: lumen ratio (100 mmHg)
 0.05 ± 0.0040.06 ± 0.0030.08 ± 0.0040.04 ± 0.0040.04 ± 0.0010.05 ± 0.0070.07 ± 0.0040.04 ± 0.002
Figure 1.

Endothelium-dependent relaxation
Concentration–relaxation curves for ACh in the absence (continuous line) and presence (dotted line) of l-NAME and Indo in uterine (A), mesenteric (B), renal (C) and femoral (D) arteries of Restricted and Control female rats. n= 5 litters per group. *P < 0.05 Control vs. Restricted for maximal EDHF-mediated relaxation.

Smooth muscle function

PE (10−9–10−4m) evoked concentration-dependent contraction in all arteries that were not different between Control and Restricted groups (Table 2, Fig. 2AD). Absolute contraction (mN mm−1) to high K+ PSS was greatest for the uterine compared with mesenteric and renal arteries, whereas for the femoral artery, contraction was similar to renal artery but greater than in the mesenteric artery (P < 0.02, Table 2). However, there were no differences between Control and Restricted groups in the ability of all arteries to contract to high K+ PSS (Table 2). SNP elicited concentration-dependent relaxation in all arteries and responses were not different between Control and Restricted groups for any of the arteries tested (Table 2, Fig. 2EH). Maximal relaxation differed between arteries, but not between treatment groups (Fig. 2).

Figure 2.

Smooth muscle contraction and relaxation
AD, smooth muscle contraction: concentration–contraction curves for PE in uterine (A), mesenteric (B), renal (C) and femoral (D) arteries of Control and Restricted female rats at 18 months of age. EH, smooth muscle relaxation: concentration–relaxation curves for SNP in uterine (E), mesenteric (F), renal (G) and femoral (H) arteries of Restricted and Control female rats. n= 5 litters per group.

Arterial passive mechanical wall properties

The stress–strain relationship for uterine arteries from the Restricted group was shifted to the left compared with that for Controls (P= 0.004, Fig. 3A), indicating increased arterial wall stiffness. The stress–strain relationship for renal arteries from the Restricted group was modestly right shifted compared with Controls (P= 0.016, Fig. 3C), indicating a small decrease in arterial wall stiffness. There were no differences in the stress–strain relationships between groups for the mesenteric or femoral arteries (Fig. 3B and D).

Figure 3.

Passive mechanical wall properties
AD, stress–strain relationships for uterine (A), mesenteric (B), renal (C) and femoral (D) arteries of Restricted and Control rats. n= 5 litters per group. *P < 0.05 Control vs. Restricted (strain vs. treatment). EH, normalized outside diameter for uterine (E), mesenteric (F), renal (G) and femoral (H) arteries of Restricted and Control female rats. n= 5 litters per group. *P < 0.05 Control vs. Restricted (o.d. vs. treatment).

Over the pressurization range, o.d. (normalized to o.d. at 5 mmHg) of uterine arteries was reduced in Restricted compared with Controls (P < 0.0002, Fig. 3E). In femoral arteries, there was an increase in o.d. (P < 0.0001, Fig. 3H), whereas there were no differences between groups for mesenteric and renal arteries (Fig. 3F and G). The ratio of media thickness to lumen diameter at 100 mmHg was not different between Control and Restricted groups for any of the arteries (Table 2).

Quantitative histological analysis of collagen and elastin in uterine arteries

Given the increase in arterial stiffness as indicated by the stress–stain relationship in uterine arteries of Restricted offspring, we determined the relative contribution of collagen and elastin fibres which are involved in structural stability and distensibility of the vascular wall. In the uterine artery, there was an increase in elastin fibre content in the media layer of uterine arteries of the Restricted group compared to Controls (6%, P < 0.05, Fig. 4B). There were no differences in elastin content in the adventitia layer between groups (Fig. 4B). Collagen fibre contents were higher in the adventitia compared to the media, with no differences between groups (Fig. 4A). Using CCLP, in the media layer of uterine arteries, there were no differences in the proportion of red, orange, yellow and green birefringent collagen fibres between groups (Fig. 5A). Of the different fibres, the orange (intermediate-thick) birefringent collagen fibres were more abundant in both the media and adventitia of arteries from both groups (Fig. 5A and B). The proportion of orange (thick) fibres was significantly greater (5%) in the adventitia of Restricted compared with Controls (P < 0.01, Fig. 5B). There was a smaller proportion of yellow (2%) and green (4%) birefringent collagen fibres in Restricted compared with Controls (P < 0.05, Fig. 5B). When combining thick (red, orange, yellow) and thin (green) birefringent collagen fibre types, there were no differences in the proportion of thick and thin birefringent collagen fibres in the media layer between groups (Fig. 5C). However, in adventitia layer, there was an increase proportion of thick birefringent collagen fibres (4%) and decreased proportion of thin birefringent collagen fibres (4%) in Restricted compared with Controls (P < 0.01, Fig. 5D).

Figure 4.

Collagen and elastin content in uterine artery
Collagen (A) visualized by picrosirius red and elastin (B) visualized by Gomori aldehyde fuschin (purple) under bright field illumination in media and adventitia layer of uterine artery of Restricted and Control rats. n= 4 litters per group. *P < 0.05 Control vs. Restricted. Magnification, × 40. Bar = 100 μm.

Figure 5.

Thick and thin collagen fibres in uterine artery
Picrosirius red staining under circular crossed-polarized light to reveal birefringent collagen fibre thickness differentiated by colour; red (thick), orange (thick), yellow (thick) and green (thin) in media (A) and adventitia (B) layer of uterine artery of Restricted and Control rats. Combined proportion of thick (red, orange and yellow) and thin (green) birefringent collagen fibres in media (C) and adventitia (D) of Restricted and Control rats. n= 4 litters per group. *P < 0.05 Control vs. Restricted. Magnification, × 40. Bar = 100 μm.

Uterine artery gene expression

The mRNA expression of calcium-activated potassium channels (KCa2.3 and KCa3.1) involved in EDHF generation was investigated in the uterine artery. No differences were observed in the expression of these channels between groups (KCa2.3: 1.02 ± 0.11 vs. 1.16 ± 0.11; KCa3.1: 1.10 ± 0.28 vs. 0.91 ± 0.07, Control vs. Restricted, respectively, n= 4 litters per group). Expression of key structural extracellular matrix proteins involved in determining the mechanical wall properties was investigated and no differences were found in the relative mRNA expression of collagens type I (α1), III (α1) and IV (α1), elastin, matrix metalloproteinase (MMP) 2 and 9, tissue inhibitors of metalloproteinase (TIMP) 1 and 2, or transforming growth factor β1 (TGF-β1) in Restricted compared with Control uterine arteries (see the online data supplement).

Discussion

This study demonstrates that late gestation uteroplacental insufficiency leads to fetal growth restriction and region specific vascular dysfunction in 18 month female offspring in the absence of elevated blood pressure. In these females the uterine artery was targeted specifically in terms of an increase in arterial wall stiffness, reduced arterial diameter, altered thick and thin collagen and elastin fibres, and endothelial dysfunction, manifested by impairment in EDHF-mediated relaxation. Smooth muscle and endothelial function were preserved in mesenteric, renal and femoral arteries of these animals. These females born growth restricted were not obese and had normal plasma parameters apart from elevated triglycerides. Our findings indicate that late gestation uteroplacental insufficiency programs regional vascular dysfunction and altered mechanical wall properties in females born growth restricted.

In the Western world, the majority of intrauterine growth restriction is due to uteroplacental insufficiency. In this study late gestation uteroplacental insufficiency failed to program hypertension in growth restricted females. Despite the reduced EDHF-mediated relaxation in the uterine artery, smooth muscle and endothelial function was preserved in all other arteries tested. In contrast, uteroplacental insufficiency induced by chronic reduction of uteroplacental perfusion pressure earlier in pregnancy (day 14 of gestation) in rats resulted in female growth restricted offspring that were hypertensive as adults and had mesenteric artery dysfunction, including increased smooth muscle contraction and altered endothelium-dependent and -independent relaxation (Anderson et al. 2006). However, another group using the reduced uteroplacental perfusion pressure model found that females developed transient hypertension (at 4 and 8 weeks) that was absent at 12 weeks of age (Alexander, 2003). Of note is that the mothers with reduced uteroplacental perfusion pressure developed pregnancy-induced hypertension (Anderson et al. 2006) and this may explain the exacerbated vascular alterations seen in the offspring. In our model, the mothers do not become hypertensive and therefore the vascular defects in the offspring observed are not due to any maternal complications but reflect the effects of uteroplacental insufficiency. The later stage of pregnancy, the shorter duration and reduced severity of the uteroplacental insufficiency created in our model may account for the relative preservation of vascular reactivity in mesenteric arteries, and also the renal and femoral arteries, and for the discrepancy in vascular outcomes between models.

A few studies have shown sex specific differences in the programming of adult diseases (Denton & Baylis, 2007). Alterations in blood pressure or vascular reactivity tend to be greater in males than in females at the same age (Ozaki et al. 2001; Franco et al. 2002). Consistent with this paradigm, uteroplacental insufficiency in late gestation appears to program sex-specific differences in cardiovascular function, as our previous studies have demonstrated that males develop hypertension and exhibit more widespread vascular dysfunction (Parkington et al. 2006; Wlodek et al. 2007). Adult Restricted males exhibit impaired endothelium-dependent and -independent relaxation in mesenteric and femoral arteries and passive mechanical wall properties (Parkington et al. 2006). Furthermore, these males have reduced nephron endowment, and glomerular hypertrophy in adulthood (Wlodek et al. 2008). Thus, in our model, endothelial dysfunction and alterations in mechanical properties may, in part, be a cause or consequence of the elevated blood pressure seen in male growth restricted offspring. In contrast, the present study demonstrated that Restricted females have preserved smooth muscle and endothelial function in some of the major resistance beds including the mesenteric, femoral and renal arteries, which may, in part, explain why the Restricted females do not develop hypertension. This study highlights that intrauterine growth restriction due to uteroplacental insufficiency programs sexually dimorphic differences in vascular function and cardiovascular disease outcomes.

The vascular endothelium plays a key role in the regulation of vascular tone. Mechanisms of vascular dysfunction in fetal programming models have been attributed to impaired NO synthesis and/or release from endothelial cells (Franco et al. 2002; Payne et al. 2003; Anderson et al. 2006; Morton et al. 2010) and alterations in the NO–cGMP pathway in the smooth muscle (Payne et al. 2003; Brawley et al. 2003). In this study, sensitivity to sodium nitroprusside was not different between Control and Restricted groups for any of the arteries, indicating that the cGMP-mediated pathway of smooth muscle relaxation was unaltered. EDHF-mediated relaxation was reduced, but only in uterine arteries, whereas EDHF-mediated relaxation in mesenteric and renal arteries was unaltered. The fact that endothelium-dependent relaxation in the absence of blockers was not different in uterine arteries between Control and Restricted groups indicates that production and/or bioavailability of NO and/or vasodilator prostanoids may be upregulated in Restricted group to compensate for the reduced EDHF response. Upregulation of NO and/or vasodilator prostanoid contributions to endothelium-dependent relaxation have been reported in mesenteric and femoral arteries of rat offspring exposed to a high fat diet in pregnancy (Taylor et al. 2004). Upregulation of compensatory endothelium-dependent vasodilator pathways may also contribute to the absence of elevated blood pressure in Restricted females.

EDHF is an important vasodilator(s) in many small resistance arteries (Feletou & Vanhoutte, 2009). Diminished EDHF-mediated relaxation may contribute to the onset of some diseases including hypertension and pre-eclampsia (Coleman et al. 2004). EDHF generation is underpinned by the activation of endothelial KCa2.3 and KCa3.1, leading to hyperpolarization of the endothelium and relaxation of vascular smooth muscle (Coleman et al. 2004; Feletou & Vanhoutte, 2009). The reason for the uterine artery specific dysfunction of EDHF is unknown, but may reflect regional differences in the identity of EDHF (Feletou & Vanhoutte, 2009). EDHF in the uterine artery is modulated by oestrogen (Burger et al. 2009) and although oestradiol levels were low in these ageing females, the levels were not different between Control and Restricted groups. Although we found no differences in mRNA expression of KCa2.3 and KCa3.1 in uterine arteries of Restricted females, this does not preclude changes in protein levels of the channels. Alternatively, the downstream EDHF signalling pathways leading to relaxation may be impaired such as the function and/or the expression of myoendothelial gap junctions (Sandow et al. 2002; Morton et al. 2010), inward rectifier potassium channels, Na+,K+-ATPase, and other putative mechanisms (Feletou & Vanhoutte, 2009).

Arterial stiffening amplifies cardiovascular disease risk and is a hallmark of the ageing process (Zieman et al. 2005). The structural stability and distensibility of the vascular wall are dependent not only on the relative amount of collagen and elastin but, importantly, also the organization and relative thickness of collagen and elastin fibres (Zieman et al. 2005). Arterial stiffness and the deposition of collagen and elastin may be altered in intrauterine growth restricted individuals and this may contribute to the pathology of cardiovascular diseases (Khorram et al. 2007). Our findings demonstrate that the Restricted females had an increase in uterine artery stiffness and reduced arterial diameter when compared with Controls. Increased uterine artery wall stiffness is consistent with previous studies on intrauterine growth restricted individuals whereby they demonstrated altered vascular growth, smaller arterial diameters and increase arterial stiffness (Martyn & Greenwald, 2001; Brodszki et al. 2005). Although in our study arterial stiffening occurred in the uterine artery, it was absent in the mesenteric and femoral arteries. However, there was a subtle decrease in arterial stiffness in the renal arteries, possibly due to an adaptive response to the moderate renal insufficiency in the Restricted females (Moritz et al. 2009). The fact that there was not a generalized stiffening of the arterial vasculature, and reduced stiffness in the renal artery in the Restricted females may be another factor that contributes to the lack of hypertension in these animals. Our histological examination with bright field illumination revealed that Restricted females had an increase in elastin fibre content in the media layer of the uterine arteries compared with Controls. The synthesis of elastin in the media of developing arteries appears to be influenced by changes in fetal haemodynamics (Brodszki et al. 2005). Although increased elastin deposition may be counterintuitive in relation to increased stiffening, Arribas and colleagues (2008) showed that an increased deposition of abnormally organized elastin is associated with artery wall stiffening in blood vessels of spontaneous hypertensive rats even before the development of hypertension. Furthermore, our group has also shown that uteroplacental insufficiency was associated with increased internal elastic lamina defects in the aorta of male and female offspring (Pascoe et al. 2008). Thus, if elastin is disorganized, these alterations in elastin may contribute to the vascular stiffness and narrowing of uterine arteries in Restricted females.

The proportion, distribution and structural alignment of the collagen matrix are important determinants of wall stiffness (Jalil et al. 1989). Our study demonstrates that in the adventitia of uterine arteries of Restricted females, there is an increase in intermediate thickness (orange), stiff, collagen fibres and a decrease in thin (green) collagen fibres compared with Controls. These finding are consistent with the increased stiffness in the uterine arteries from the Restricted group. Thus, the increase in uterine arterial stiffness may be due to the complex rearrangement of collagen fibres. In a rat model of hypertrophied myocardium with fibrosis, there was an increase in the contribution of thick collagen fibres (high tensile strength) and fewer thin fibres (thought to provide resilience) (Jalil et al. 1989). Therefore, disequilibrium in the distribution of collagen fibres may lead to wall stiffening. To further decipher the mechanisms involved in increased uterine arterial stiffness we examined mRNA expression of key structural extracellular matrix proteins involved in determining the mechanical wall properties. Although we found no differences in key regulators of extracellular matrix proteins that may modulate the mechanical wall properties including collagen type I (α1), III (α1) and IV (α1), elastin, MMPs 2 and 9, TIMPs 1 and 2, and TGF-β1 at the mRNA level, protein levels may differ. Furthermore, other factors regulating arterial compliance including cross-linking, orientation and fibre alignment of the collagen network, glycosaminoglycans, and advanced glycation end products need to be pursued in future studies.

The growth restricted females had a twofold increase in plasma triglycerides but were of normal weight compared with Controls. There is evidence to suggest that plasma triglycerides concentrations are frequently elevated in experimental models with chronic renal insufficiency (Heuck et al. 1978). We have previously shown that these Restricted females also have a nephron deficit, and with ageing have elevated plasma creatinine, glomerular hypertrophy and modest renal insufficiency and may play a role in the observed increase in plasma triglycerides (Moritz et al. 2009).

Lastly, although we studied the effects of ageing on growth restricted females, it is likely that vascular dysfunction in the uterine arteries of Restricted females has been programmed and thus existed earlier in life. It is therefore of interest to consider their maternal uterine vascular adaptation to pregnancy. Pregnancy is associated with profound alterations of the uterine artery and inadequate adaptation of the uterine artery to pregnancy is a cause of maternal pregnancy and fetal complications. Indeed, there is evidence of altered vascular responses, including in the uterine artery, to pregnancy in offspring that experienced nutrient challenges during intrauterine life (Torrens et al. 2003; Hemmings et al. 2005). In our study, the Restricted females have selective impairment of uterine vascular function, particularly involving impaired EDHF-mediated relaxation, along with altered collagen and elastin, increase arterial stiffness and reduced arterial diameter. Thus, these Restricted females may exhibit impaired uterine artery adaptation to their pregnancy which may result in insufficient uteroplacental blood flow and pregnancy complications.

Summary

In conclusion, female offspring subjected to late gestation uteroplacental insufficiency have programmed EDHF-mediated endothelial dysfunction in the uterine artery. Furthermore, Restricted females have increased uterine artery wall stiffness associated with altered proportions of thick and thin collagen and elastin fibres. Vascular function was preserved in other regions of the body and may explain the absence of hypertension in growth restricted females. These finding differ from the male offspring subjected to the same prenatal insult, which had elevated blood pressure and widespread vascular dysfunction (Parkington et al. 2006; Wlodek et al. 2007). Intrauterine growth restriction has sex-dependent effects on vascular function in the offspring. The localization of uterine vascular dysfunction in Restricted females may have the potential for impacting on their pregnancy adaptations and the intrauterine environment of the next generation.

Appendix

Author contributions

M.E.W., M.T., H.C.P., N.M.D. and M.Q.M. were all involved in the conception and design of the experiments. M.Q.M. and M.T. performed the experiments and all authors were involved in the analysis, interpretation and writing of the manuscript. This study was conducted at the University of Melbourne and Monash University.

Acknowledgements

Authors were supported by the National Health and Medical Research Council of Australia. M.Q.M. was supported by a Kidney Health Australia Biomedical Scholarship and The University of Melbourne Fee Remission Scholarship. We thank Dr Andrew Siebel, Dr Laura Parry, Kerryn Westcott, Tania Romano, Amy Mibus, Dr Lenka Vodstrcil, Ian Boundy, Bruce Abaloz and Sandy Clarke (Statistical Consulting Centre, University of Melbourne) for their valuable contributions.

Ancillary