Developmental programming of aortic and renal structure in offspring of rats fed fat-rich diets in pregnancy
- James A. Armitage1,
- Lorin Lakasing1,
- Paul D. Taylor1,
- Aswini A. Balachandran1,
- Runa I. Jensen1,
- Vasia Dekou1,
- Nick Ashton2,
- Jens R. Nyengaard3 and
- Lucilla Poston1
- 1Maternal and Fetal Research Unit, Division of Reproductive Health, Endocrinology and Development, King's College London, London, UK2Faculty of Life Sciences, University of Manchester, Manchester, UK3Stereological Research and Electron Microscopy Laboratory, University of Aarhus, Aarhus, Denmark
- Corresponding author J. A. Armitage: Maternal and Fetal Research Unit, Division of Reproductive Health, Endocrinology and Development, GKT Department of Women's Health, 10th Floor, North Wing, St Thomas' Hospital, Lambeth Palace Road, London SE1 7EH, UK. Email: james.armitage{at}kcl.ac.uk
Abstract
Evidence from human and animal studies suggests that maternal nutrition can induce developmental programming of adult hypertension in offspring. We have previously described a model of maternal dietary imbalance in Sprague-Dawley rats whereby administration of a maternal diet rich in animal lard programmes the development of increased blood pressure, insulin resistance, dyslipidaemia, obesity and mesenteric artery endothelial dysfunction in adult offspring. To further characterize the mechanism of hypertension in this model we have examined vascular and renal structure in adult offspring of Sprague-Dawley rats fed a control diet (OC) or lard-rich diet (OHF) during pregnancy and suckling followed by a control diet post-weaning. To gain further insight, we assessed aortic reactivity and elasticity in an organ bath preparation and renal renin and Na+,K+-ATPase activity. Plasma aldosterone concentration was also measured. Stereological examination of the aorta in OHF demonstrated reduced endothelial cell volume and smooth muscle cell number compared with OC. Adult OHF animals showed increased aortic stiffness and reduced endothelium-dependent relaxation. Renal stereology showed no differences in kidney weight, glomerular number or volume in OHF compared with OC, but renin and Na+,K+-ATPase activity were significantly reduced in OHF compared with controls. Programmed alterations to aortic structure and function are consistent with previous observations that exposure to maternal high fat diets produces systemic vascular changes in the offspring. Despite normal renal stereology, altered renal Na+,K+-ATPase and renin activity offers further insight into the mechanism underlying the increased blood pressure characteristic of this model.
Retrospective epidemiological studies of geographically disparate populations demonstrate an association between disproportionate fetal and postnatal growth and the development of adult cardiovascular risk factors including hypertension, dyslipidaemia, obesity, altered vascular endothelial function and altered glucose homeostasis. These are features of the metabolic syndrome, which represents an increasing health problem worldwide (Alexander et al. 2003).
It is suggested that in man, adult metabolic syndrome may, at least in part, have origins in fetal or early life (Gluckman & Hanson, 2004), and that maternal undernutrition leads to the development of several facets of this condition (Roseboom et al. 2001). This hypothesis is supported by several experimental animal models (Armitage et al. 2004).
Whilst maternal dietary restriction in pregnancy is a useful tool to study the mechanisms of developmental programming in the malnutritive state, it has limited relevance to dietary imbalance in the developed World. We have characterized a rat model of developmental programming based on a diet in pregnancy and suckling that is rich in animal fat and typical of a ‘Western’ diet (Khan et al. 2003; Taylor et al. 2003). Using radiotelemetry for remote recording of blood pressure in conscious unrestrained animals, we have reported raised systolic blood pressure in 6- and 12-month-old female offspring of fat-fed dams (OHF) compared with that observed in offspring of controls (OC) (Khan et al. 2003). However, both male and female OHF demonstrated markedly reduced endothelium-dependent relaxation in small mesenteric arteries of diameter ∼200 μm in vivo (Khan et al. 2003, 2004).
In the present study we aimed to further characterize this model of developmental programming induced by exposure to a lard-rich maternal diet during pregnancy and suckling. We have performed stereological assessment of aortic tissue to determine whether endothelial or smooth muscle structure was altered. These studies were carried out in offspring of control and lard-fed dams at 6 months of age. Additionally, we examined functional correlates: endothelium-dependent dilatation, vasoconstriction to pressor agonists and passive elasticity. Endothelium-dependent vasodilatation in conduit arteries is not a determinant of blood pressure per se but blunted endothelium-dependent dilatation is recognized to be a predictive indicator of atherosclerosis in man (Bisoendial et al. 2002). We assessed aortic endothelium-dependent dilatation in 6-month-old rats. We have also assessed vascular elasticity, as reduced vascular compliance is also a cardiovascular risk factor in man (Safar et al. 1998) and in hypertension-prone animal models, arterial stiffness precedes blood pressure changes (van Gorp et al. 2000). Whilst studies in man have suggested that individuals of low birthweight have reduced arterial compliance (Martyn et al. 1995), there are no reports of arterial compliance or stiffness in animal models of developmental programming. In the present study, passive length–tension curves, produced in organ bath experiments, have provided a useful indication of passive arterial stiffness. Whilst altered vascular structure and function could contribute to hypertension in this model, other investigators have highlighted a potential role of the kidney in developmental programming of cardiovascular disorders.
In view of several reports of reduced nephron number, associated with hypertension in other models of developmental programming (Vehaskari et al. 2001; Woods et al. 2001; Franco et al. 2002; Langley-Evans et al. 2003; Lisle et al. 2003; Sahajpal & Ashton, 2003), we also undertook a stereological assessment of the kidneys of offspring of lard-fed dams. Since perturbations of renal renin activity and the renin–angiotensin system (RAS) have also been demonstrated in other models of developmental programming (Woods et al. 2001; Sahajpal & Ashton, 2003; Vehaskari et al. 2004), we evaluated renal renin activity and the plasma aldosterone concentration in 6-month-old animals. Dysfunction of Na+,K+-ATPase has previously been implicated in certain forms of hypertension (Di Nicolantonio et al. 1993; Patel et al. 1996; de Wardener, 1997). Previous studies have highlighted the relationship between Na+,K+-ATPase activity and cell membrane phospholipid profile (Else et al. 1996) and we have demonstrated reduction of key fatty acids in the phospholipid membranes from offspring of lard-fed rats (Ghosh et al. 2001). We therefore assessed renal Na+,K+-ATPase activity. This was carried out in 1-year-old animals, littermates of those described elsewhere in this article.
Methods
Animal husbandry and experimental diets
All experiments conformed to national guidelines and local ethics requirements. Female Sprague-Dawley rats (100–120 days of age) were fed, for 10 days prior to mating and throughout pregnancy and lactation, either a control breeding diet of normal laboratory chow or an experimental diet of the standard chow (RM3) with 20% w/w animal lard and with 20% additional vitamins and minerals, protein, inositol and choline to correct for reduction in content of these micronutrients resulting from the addition of the lard (Table 1, Special Diet Services, Witham, Essex, UK). Composition of these diets was confirmed by independent analysis (Eclipse Scientific Group, Cambridge, UK). Before 48 h postpartum all litters were reduced to eight pups and where possible, to an equal number of males and females. After weaning, all offspring were fed ad libitum standard maintenance diet (RM1) and housed as four per cage. Food intake and animal weights were recorded once per week until the animals were fully grown. Rats were killed at 180 (or 360 for the Na+,K+-ATPase activity assay) days of age by CO2 inhalation. The studies described here were performed on male and female animals from a total of 20 control litters and 14 litters from lard-fed dams. These animals were bred in two sequential cohorts using identical protocols.
Offspring aortic studies
Preparation and sampling protocol for aortic tissue.
Abdominal aortas were dissected above the iliac bifurcation and isolated. Samples were then dehydrated at room temperature in ascending grades of ethanol (70%, 96%, 96%, 99%, 99%, 99% for 1 min each), and immersed in xylene. Aortas were embedded into plastic blocks (Technovit 7100 kit, Kulzer Histo-Technik, Heraeus Kulzer, Germany) and orientated so that transverse sectioning was possible. The blocks were sectioned exhaustively (section thickness 40 μm) using a Ralph glass knife mounted on a calibrated microtome (Supercut, Reichert-Jung, Germany). Systematic uniform random sampling of every 50th section was performed. Two adjacent thin sections (2 μm) were cut at random intervals for volume estimation as detailed below. Aortic sections were floated onto water, dried, dehydrated and mounted onto clean slides and incubated at 60°C overnight. The sections were stained with Mayers haematoxylin and counterstained with eosin, then dehydrated in absolute alcohol and cleared with xylene before coverslips were mounted using adhesive resin. Sections were then examined using light microscopy to check for artefactual changes that would invalidate the stereological results. Two aortas in the female control group were noted to have severely denuded or disrupted endothelium, presumably resulting from tissue processing, and were excluded from analysis of endothelial parameters.
Microscopy.
A modified Olympus BX50 microscope equipped with a MT12 microcator and ND281 readout (Heidenhain, Germany) was employed for structural analysis. Systematic uniformly random sampling of fields of view was performed using a motorized specimen stage (Merzhauser, Germany) which enabled movements in the x and y axes to be controlled. A CCD camcorder (KY-F 50E, JVC, Japan) connected to a PC was mounted on the microscope. The computer, fitted with a frame grabber (flashpoint 3D PRO, USA), was connected to a 21-inch monitor (EIZO 120-FlexScan, USA). CAST stereology software (v2.0, Visiopharm, Denmark) enabled 2D counting frames to be superimposed on the images. The aortic and renal tissues were investigated at a magnification of 3359 (Olympus objective 60×, S-Plan, NA 1.40) and 69.4 (Olympus objective 1.25×, S-Plan, NA 0.13), respectively.
A modification of the optical fractionator technique (Dorph-Petersen et al. 2001; Gundersen, 2002) was used in the analysis of 40 μm thick aorta sections. Sampling distances in the x and y axes were 175 μm. Areas of the 2D unbiased counting frames (aframe) were 3435 μm2 for endothelial cells and 343.5 μm2 for smooth muscle cells. The number of endothelial cells and smooth muscle cells in the defined volume was then determined by using a 60× (NA 1.40) oil immersion lens and counting the nucleoli and nuclei, respectively, that came into focus between a depth of 5–20 μm, i.e. a disector height (h) of 15 μm.
Structural analysis of aortic medial thickness.
The 2 μm sections were analysed using a test point system for estimation of mean profile area. A nine-test-point system was used for smooth muscle cells and 16 points for endothelial cells. Points overlying the whole tissue, and the cell type of interest, were counted separately. Medial thickness was determined by taking orthogonal readings (5 per vessel on 2 sections for each aorta) at randomly selected sampling points.
Estimation of aortic volume, cell number and mean cell volume.
The mean volume of the endothelial cells/smooth muscle cells,
, is given by:
VV,cell/vessel is the volume density of the vessel (media + intima) occupied by endothelial cells or smooth muscle cells; NV,cell/vessel is the numerical density of endothelial cells or smooth muscle cells; Ptcell and Ptvessel are the number of test points falling on endothelial or smooth muscle cells and on the vessel (media + intima) on thin sections;
ΣQ−cell is the number of cells sampled in the counting frames on thick sections; p is the number of test points per frame, here 4; aframe is the area of the counting frame, here 3435 μm2 and 343.5 μm2 for endothelial and smooth muscle cells, respectively; h is the disector height, here 15 μm and ΣPvessel is the number of points falling on the vessel (media + intima).
The volume of each vessel (media and intima), Vvessel, was estimated using the Cavalieri principle (Gundersen & Jensen, 1987).
F is the fraction of sections sampled, here 38; t is the section thickness, here 38.4 μm; a/p is the area per test point, here 859 and 85.9 μm2 for intima and media and ΣP is the total number of points overlying the area of interest.
The number of endothelial cells/smooth muscle cells, N(cell), in each vessel was estimated using:
SSF is the section sampling fraction, here 38; ASF is the area sampling fraction given by aframe/dxdy, dx= dy= 175 μm; HSF is the height sampling fraction estimated by h/t, where h is the height of the disector and t is the Q− weighted section thickness (Dorph-Petersen et al. 2001).
The error variance was estimated using the approximation formula (Gundersen et al. 1999). The average coefficient of error (CE) for total number of endothelial and smooth muscle cells were 0.06 and 0.12, respectively.
Functional examination.
Rats were killed at 180 days of age by CO2 inhalation and exsanguination. The thoracic aorta was excised from the aortic arch to the level of the diaphragm and rinsed in ice-cold physiological saline solution (PSS), and dissected free of connective tissue and fat. Two rings 2.5 mm in length from the middle portion of each vessel were mounted in an organ bath (Model 700MO, DanishMyo, Denmark) chamber containing PSS at 37°C and gassed with 95% O2–5% CO2. Vessels were equilibrated to a 5 mN force (equivalent to 1 g) for 30 min, then subjected to a run-up of three cumulative stretches to construct a length–tension curve (from basal internal diameter stretched by 500 μm, 200 μm then 50 μm), each lasting 2 min. After re-equilibration at 5 mN, three contractile responses to 125 mm K+ substituted PSS (KPSS) were performed. Vessels were then washed and equilibrated at a force of 5 mN and a cumulative dose–response curve to phenylephrine (3 × 10−9–10−5m) performed. After washout, vessels were submaximally constricted to phenylephrine (80% of maximum force) then dose–response curves constructed to assess endothelium-dependent vasodilatation to acetylcholine (ACh 3 × 10−9–10−5m), in the absence and presence of the NOS inhibitor Nω-nitro-l-arginine methyl ester (l-NAME) (10−4m). Smooth muscle sensitivity to the endothelium-independent vasodilator nitric oxide was assessed by an aqueous nitric oxide dose–response curve (10−7–3 × 10−5m). To measure arterial stiffness in the absence of active tension, PSS was replaced with warmed calcium-free PSS (with 10−4m EGTA) and the passive length–force relationship (a measure of passive stiffness) repeated over three cumulative stretches (500 μm, 200 μm, 50 μm) each lasting 2 min. Two rings from each aorta were studied and data averaged in subsequent analyses.
Renal studies
Sampling protocol for renal tissue.
Kidneys were dissected and halved coronally, dehydrated at room temperature in ascending grades of ethanol (70%, 96%, 96%, 99%, 99%, 99% for 1 min each), then immersed in xylene and embedded into plastic blocks (Technovit 7100 kit, Kulzer Histo-Technik, Heraeus Kulzer, Wehrheim, Germany) orientated to enable coronal sections. Blocks were sectioned exhaustively as described of the aortic tissue. Systematic uniform random sampling of every 25th section pair, sampling as well as look-up section of thickness 20 μm, was performed for glomerular number. At random intervals on two occasions, 10 2-μm-thick sections were cut instead of the 20-μm-thick section and used for glomerular volume estimation. A nine-test-point system was used with 225 points per frame for the glomeruli. All sections were floated, mounted, stained with PAS (periodic acid, Schiffs reagent and Mayers haematoxylin), dehydrated, cleared then mounted as per the aortic protocol.
Estimation of glomerular number and volume.
The physical fractionator technique (Sterio, 1984; Nyengaard & Bendtsen, 1990) was used to analyse the 20 μm thick renal section pairs for glomerular number and volume. Systematic uniformly random sampling of vision fields was performed in incremental movements of 8000 μm in both the x and y axes. The 2D unbiased counting frame had an area, aframe= 6.89 × 106μm2.
The physical fractionator was used to estimate the number of glomeruli, Nglom, independent of tissue deformation:
F is the fraction of sections sampled, here 1 in 25; dx and dy are the movements of the stage in x- and y-directions, here 8000 μm; aframe is the area of the counting frame, here 6.89 × 106μm; Ps is the number of test points hitting renal tissue; Pf is the number of test points hitting renal tissue in the five middle section pairs for estimating Q− and ΣQ− is the total number of glomeruli counted in the counting frame which is divided by 2 because we counted both ways.
The same principle was used to estimate the glomerular volume except that the disector height was equal to the section thickness. The error variance was estimated using the approximation formula (Gundersen et al. 1999). The average CE for glomerular volume and number were 0.10 and 0.08, respectively.
Renal tissue renin activity.
Rats were killed at 6 months of age as described above and kidneys rapidly dissected and snap-frozen in liquid nitrogen, then stored at −80°C. Tissue was homogenized on ice in Tris-HCl buffer, centrifuged (1000 g, 20 min, 4°C) and the supernatant diluted to a final concentration of 1: 4000 with 0.5 m phosphate buffer (pH 6.5). Following addition of 100 μl of renin-free plasma, renal renin activity was measured using a REN-CT2 kit (CIS UK Limited, UK) and was expressed as micrograms angiotensin I generation per gram kidney weight per hour (Gouldsborough et al. 2003).
Plasma aldosterone concentration.
Whole blood was obtained by cardiac puncture (anticoagulant K3 EDTA) from six-month-old rats after CO2 asphyxiation. Blood was centrifuged (1000 g, 10 min, 4°C) and plasma stored at −70°C until analysis. The plasma aldosterone concentration was measured using a commercial radioimmunoassay kit (Coat-a-Count, Diagnostic Products Corporation, Caernarfon, Gwynedd, UK) and was expressed in picomolar. Intra-assay variability was 3.2%.
Tissue Na+,K+-ATPase enzyme activity.
Following our observations with regard to renal renin content and plasma aldosterone concentration we measured Na+,K+-ATPase activity in kidneys from 1-year-old littermates of those animals studied at 6 months of age. The assay is based on Na+,K+-ATPase conversion of ATP to ADP with the liberation of a phosphate molecule and detected by colourimetric assay as described by Else et al. (1996).
Sample preparation. Rats were killed as previously described at 360 days of age and whole kidney rapidly dissected free and placed in ice-cold homogenization buffer, sliced coronally and homogenized mechanically in sucrose-based buffer. A sample was removed for determination of protein content according to manufacturer's instructions (Lowry Protein Assay, Bio-Rad DC protein assay, Bio-Rad, USA).
According to the procedure of Else et al. (1996), tissue homogenate was mixed 1: 1 with SDS (0.75 mg ml−1), and aliquots removed and mixed with incubation solution (150 mm histidine, 640 mm NaCl, 40 mm MgCl2 200 mm KCl). Non-specific phosphate production was assessed in duplicate samples after addition of 10 mm ouabain. After preincubation (10 min, 37°C) ATP was added and after 5 min the reaction was quenched with perchloric acid. Samples were centrifuged (1200 g, 15 min, 2°C) and equal volumes of supernatant mixed with distilled water and colour reagent (1 g (NH4)6MO7O24.4 H2O, 94.7 ml H2O, 3.3 ml concentrated sulphuric acid and 4 g FeSO4.7H2O). Absorbance of the final solution was read at 750 nm on an automated plate reader (Sunrise TS, Tecan Ltd, UK) with a phosphate standard curve (10–250 nm KH2PO4). Total Na+,K+-ATPase activity (μmol PO4 (mg protein)−1 h−1) was determined as the difference in inorganic phosphate liberated in the presence and absence of ouabain.
Statistical analysis
Stereological data were not normally distributed; thus data were transformed by a best power transformation (boxcox STATA v8.2, StataCorp, TX, USA) to achieve normality and then analysed by ANOVA. By convention, stereological data are presented as the mean and coefficient of variance (CV). Plasma aldosterone concentration data were not normally distributed, so data were also subject to a best power transformation before analysis. All other values were normally distributed and are given as means ±s.e.m. Dose–response curves to vasoactive agonists were analysed by repeated measures (RM) ANOVA (StatView v5, SAS Institute, NC, USA) between diet groups with increasing drug dose as the repeated measure. Additionally, dose–response curves were fitted to a three-parameter sigmoidal function, and the sensitivity coefficient, pEC50 determined (GraphPad Prism v3, GraphPad Software, CA, USA). Data for pEC50, Na+,K+-ATPase activity, renin activity and transformed stereological data were analysed by ANOVA. For all studies, sex was included in the initial ANOVA model; where sex was significant data are presented for male and female offspring separately. Where sex was not significant, male and female data were pooled. Significance was assumed where P < 0.05 and where appropriate, post hoc analyses incorporated correction to protect against the type I errors inherent to nested designs. One male and one female were studied from each litter, where n refers the number of litters studied.
Results
Growth curves of offspring from one of the cohorts described in this study have been published elsewhere; weights at birth did not differ significantly between groups (Khan et al. 2003). At 180 days there were no significant differences in bodyweights between offspring of control or lard-fed dams, though overall male animals were heavier than female animals (Male: OC 641.3 ± 18.6 g, n= 6, versus OHF 681.9 ± 16.8 g, n= 6; Female: OC 341.4 ± 8.4 g, n= 5, versus OHF 348.7 ± 9.3 g, n= 6).
Aortic studies
Aortic structure.
All 2-way ANOVA failed to find a significant main effect of sex, and therefore these were collapsed to a 1-way ANOVA on maternal diet (Table 2). Aortic stereology data are shown in Fig. 1 and Table 2 and presented here as means (CE).
Aortic endothelial cell layer volume (Fig. 1A) was significantly reduced in aortas from OHF compared with OC animals (OHF 2489.9 (0.35) μm3, n= 24 versus OC 3354.2 (0.40) μm3, n= 10, P < 0.02); however, endothelial cell number (Fig. 1B) was not affected by maternal diet (OHF 4.55 × 105 (0.25), n= 24 versus OC 4.33 × 105 (0.28), n= 9, P < 0.62).
Smooth muscle cell volume (Fig. 1C) was not affected by maternal diet (OHF 2742 (0.11) μm3versus OC 2751 (0.12) μm3, P < 0.93). However, OHF animals showed reduced numbers of smooth muscle cells (Fig. 1D) (OHF 5.71 × 106 (0.24), n= 18 versus OC 7.18 × 106 (0.25) n= 12, P < 0.007).
Medial thickness was not altered by maternal diet (OHF 102.3 (0.06) μm, n= 20, versus OC 100.4 (0.08) μm, n= 10, P < 0.47).
Aortic function.
There were no significant main effects of sex (Table 2) so all ANOVA models were collapsed to remove sex as an independent variable. In calcium-free PSS, OHF rats showed a significant increase in stretch-induced tone compared with OC, indicative of greater passive stiffness (Fig. 2A, Table 2). OHF produced significantly higher tension across all cumulative stretches than the OC (RM ANOVA F1,3= 4.58, P < 0.04, n= 11 (OHF), n= 13 (OC)). There was a significant interaction between maternal diet and passive stiffness, with the OHF showing a higher magnitude increase in tension for a given stretch, a further suggestion of increased aortic stiffness (RM ANOVA F1,3= 4.71, P < 0.005, n= 11 (OHF), n= 13 (OC)).
There was no evidence for alterations in contractile reactivity to the α1-adrenergic receptor agonist phenylephrine (Fig. 2B) as both the cumulative dose–response data (RM ANOVA F1,7= 0.60, n= 11 per group, P < 0.45) and pEC50 (ANOVA, OHF 7.42 ± 0.07 log m, n= 11 versus OC 7.28 ± 0.08 log m, n= 11, P < 0.08) were not significantly affected by maternal diet (Table 2).
OHF showed significantly blunted endothelium-dependent vasodilatation over the ACh cumulative dose–response curve (RM ANOVA F1,7= 11.19, P < 0.003, n= 11 (OHF), n= 15 (OC)) together with a significant interaction between dose and diet (Fig. 2C and Table 2). There was, however, no significant effect of maternal diet on offspring ACh sensitivity (pEC50 ANOVA, OHF 7.22 ± 0.14 versus OC 7.06 ± 0.20 log m, P < 0.55, n= 11 (OHF), n= 15 (OC)). Nitric oxide synthase blockade (by preincubation with 10−4ml-NAME) totally prevented ACh induced vasodilatation (data not shown). There was no significant effect of maternal diet on endothelium-independent vasodilatation (NO dose–response curve RM ANOVA F1,6= 1.00, P < 0.32 n= 10 (OHF), n= 14 (OC)) nor were any higher order interactions significant (Fig. 1D, Table 2). Similarly, there was no significant effect of maternal diet on offspring sensitivity to NO (pEC50 OHF 7.1 ± 0.18 versus OC 7.47 ± 0.21 log m, P < 0.54, n= 10 (OHF), n= 14 (OC)).
Renal studies
There was no effect of maternal diet upon kidney weight (P < 0.65); however, the effect of sex (P < 0.0001) was significant, with male kidneys being significantly heavier than female kidneys (Male OC 2.22 ± 0.04 g, n= 6 versus OHF 2.07 ± 0.18 g, n= 6, Female OC 1.23 ± 0.81 g, n= 5 versus OHF 1.19 ± 0.08 g, n= 6).
Renal structure.
There was no effect of maternal diet on kidney weights (P < 0.64); however, there was a significant effect of offspring sex as males had heavier kidneys than females (mean weight ±s.e.m. male OHF 2.07 ± 0.08 g, n= 6, male OC 2.22 ± 0.04 g, n= 6, female OHF 1.23 ± 0.08 g, n= 6, female OC 1.18 ± 0.08 g, n= 6, P < 0.0001). As is conventional, stereological data are given as the mean (CV). There were significant effects on sex for renal structures examined (Table 2); in all cases, males had higher glomerular number, an estimate of nephron number, (P < 0.04) and glomerular volume (P < 0.001) than females. There were, however, no significant effects of maternal diet on glomerular volume (OHF 1.20 (0.22) μm3, n= 12, OC 1.09 (0.18) μm3, n= 11, P < 0.12) (Fig. 3A) or number (OHF 30.35 (0.14) μm3, n= 12, OC 29.95 (0.17) μm3, n= 11, P < 0.75) (Fig. 3B). There were no significant interactions between maternal diet and offspring sex for these renal parameters studied (Table 2).
Renal renin activity.
Renal renin activity was significantly affected by maternal diet and offspring sex (Fig. 4A, Table 2). Overall, OHF had lower renin activity than OC and males had lower renin levels than females (μmol Ang I generation (g tissue)−1 h−1: male OHF 70.25 ± 9.45, n= 9, male OC 81.47 ± 8.82, n= 9, female OHF 90.25 ± 5.62, n= 10, female OC 117.16 ± 8.97, n= 11, P < 0.03 for diet, P < 0.002 for sex). There was, however, no significant interaction between maternal diet and offspring sex (Table 2).
Plasma aldosterone concentration.
Plasma aldosterone concentration was not affected by maternal diet (Fig. 4B, Table 2). However, there was a highly significant effect of offspring sex on the aldosterone concentration. Overall males had reduced aldosterone compared with females (male OHF 55.73 ± 18.77 pm, n= 7, male OC 56.57 ± 12.81 pm, n= 6, female OHF 357.78 ± 58.59 pm, n= 6, female OC 241.700 ± 92.12 pm, n= 5, P < 0.26 for diet, P < 0.0001 for sex). There was no significant interaction between maternal diet and offspring sex (Table 2).
Renal Na+,K+-ATPase activity.
Na+,K+-ATPase activity was affected by maternal diet, with OHF demonstrating a significant reduction in activity (Fig. 4C, Table 2). There was no effect of offspring sex and no significant interaction between maternal diet and offspring sex (Table 2). OHF demonstrated a 40% reduction in Na+,K+-ATPase activity compared with OC (OHF 181.09 ± 13.26 n= 20 versus OC 293.41 ± 33.45 μmol PO4 (mg protein)−1 h−1, n= 20, P < 0.002).
Discussion
To further assess hypertension in this model we have assessed renal structure and enzyme activity, and aortic structure and function. In the present study, both unbiased stereological and functional methods were used to estimate aortic contractile, relaxation and elastic function, together with endothelial and smooth muscle cell number and volume indices. Additionally, glomerular number and volume, renal renin activity, plasma aldosterone concentrations and renal Na+,K+-ATPase activity were evaluated in kidneys from offspring of dams fed a lard-rich diet during pregnancy. The exhaustive and technically rigorous approach to assessment of aortic and renal structure has not previously been used in any studies of developmental programming by maternal dietary manipulation, and although the technique used to estimate glomerular (and thus nephron) number has been used in other studies (Sterio, 1984), including developmental programming by prenatal dexamethasone exposure (Wintour et al. 2003), the methodology employed for assessing vascular structure was newly developed for the purposes of this investigation.
We have previously shown that by 6 months of age, offspring of dams fed a lard-rich diet in pregnancy and suckling develop a phenotype that is similar to that of human metabolic syndrome: hypertension, dyslipidaemia, insulin resistance obesity and blunted endothelium-dependent vasodilatation (Khan et al. 2003, 2004; Taylor et al. 2004b). The present study extended these observations and demonstrated that adult OHF animals exhibit reduced aortic endothelium-dependent vasodilatation and increased aortic stiffness. This was accompanied by reduced smooth muscle cell number and endothelial cell volume. There were no stereological or gross morphological differences observed in renal structure, but renal renin activity and renal Na+,K+-ATPase activity were compromised in OHF animals when compared with controls. Plasma aldosterone was not altered in OHF animals, relative to controls.
Our finding of increased aortic stiffness in OHF supports the hypothesis that disturbances to maternal nutrition programme altered aortic function in offspring. There is debate as to whether arterial compliance changes precede or follow hypertension, but the evidence implicating developmental programming of altered compliance is limited. In one study, adult humans of low birthweight were shown to have reduced conduit artery compliance in the trunk and legs, as assessed by pulse wave velocity (Martyn et al. 1995). Others have implicated early life influences of diet on compliance by observing that increased duration (< 10 months) of breast-feeding is associated with decreased adult brachial artery compliance in humans (Leeson et al. 2001). Potentially, this may result from increased exposure to fat in early postnatal life. Vascular remodelling, secondary to raised blood pressure, is presumed to occur in man since non-invasive imaging of the carotid arteries in untreated hypertensive subjects has suggested increasing arterial wall stiffening with increasing diastolic blood pressure (van den Berkmortel et al. 2001).
Given the observation of increased aortic stiffness in the offspring of lard-fed dams, we might anticipate increased medial thickness and smooth muscle hypertrophy, since increased vascular stiffness is often accompanied by alterations in the extracellular matrix collagen and elastin ratio and also changes in smooth muscle size and number (Wuyts et al. 1995). Surprisingly, in OHF animals we found reduced smooth muscle cell number, unchanged smooth muscle cell volume and unchanged medial thickness. This may be indicative of proliferation of the intracellular matrix, presumably increased collagen deposition, an interpretation that would be consistent with increased vessel stiffness if analogous to the human disease. It is possible that these animals might later develop increased medial thickness as a result of advancing hypertension with age. Indeed, animal studies have demonstrated increased arterial stiffness with advancing age in spontaneously hypertensive rats (Safar et al. 2001). Further stereological examination at a later time point would be helpful to determine whether aortic structural changes advance with age.
We also observed reduced endothelium-dependent vasodilatation in the aorta from OHF. It is proposed that endothelial dysfunction may initiate many of the abnormalities clustering in the metabolic syndrome (Bonora et al. 2003) and in the larger vessels this deficit is intimately involved in atherogenesis (Palinski & Napoli, 2002). Reduced nitric oxide bioavailability (implied in this model by reduced ACh-stimulated vasodilatation) is implicated in the genesis of atherosclerosis (Cooke, 2004). We cannot comment as to whether the defect observed is a primary or secondary event; nonetheless, results from the present study show clear aortic dysfunction at 180 days of age. The present data also showed a reduction in endothelial cell layer volume, evidence of altered endothelial cell morphology. The significance of the altered aortic endothelial morphology is difficult to interpret but it may be of relevance that altered endothelial morphology occurs in retinopathy in experimental animal models of diabetes (Dolgov et al. 1982). Endothelial apoptosis is observed in atherosclerosis (Stoneman & Bennett, 2004) and our observation of invariant endothelial cell number suggests that there is no significant apoptosis, and therefore minimal chance of widespread atherosclerosis in offspring of lard-fed dams.
There are conflicting reports as to whether the function of offspring aorta is affected in other animal models of developmental programming. Torrens et al. (2003) reported normal aortic function in female offspring of protein restricted dams, whereas Franco et al. (2002) report reduced aortic relaxation to ACh and increased constrictor function in aortas from male and female offspring of dams subject to global caloric restriction during pregnancy. Our present findings support the hypothesis that aortic endothelial dysfunction is manifest in offspring from lard-fed rats, and suggest that endothelial dysfunction is central to the phenotype observed in these animals.
In this study we have shown no change in glomerular number and volume, which contrasts with several previous investigations in animal models of developmental programming. In the low protein model, studies of renal structure using both basic light microscopy and stereological methods demonstrated a 13% (Langley-Evans et al. 1999) and 45% reduction (Woods et al. 2004), respectively, in glomerular count in offspring of low-protein-fed dams. Whilst the maternal fat-rich diet is associated with elevated systolic blood pressure in female offspring, fetal renal dysgenesis does not therefore appear to be a contributing mechanism in the high-fat model. Of course, chronic exposure to high blood pressure may cause structural damage to the kidney, leading to a loss of nephron function that could exacerbate the hypertension with ageing.
Previous studies in which we have employed radiotelemetric monitoring of blood pressure in this model have indicated progressive hypertension in female OHF (Khan et al. 2003). We have now shown this to be associated with low renal tissue renin activity. Low renin status is commonly associated with essential hypertension and has variously been attributed to raised aldosterone, adrenal super-sensitivity to angiotensin II, 11β-hydroxysteroid dehydrogenase (11β-HSD) type 2 deficiency and a reduction in glomerular number (Pratt, 2000). Since the plasma aldosterone concentration and glomerular number were normal we may exclude these possibilities. We have previously reported altered vascular sensitivity to angiotensin II in this model (Taylor et al. 2004a), but it is unknown whether this hyposensitivity also applies to the adrenal cortex. Decreased plasma renin activity has also been reported in the protein restricted model of prenatal programming of hypertension and insulin resistance (Vehaskari et al. 2001; Woods et al. 2001; Sahajpal & Ashton, 2004), but in these studies renin activity was reduced in young animals only and then normalized by approximately 4 weeks of age. Therefore, in the present study, the observation of low renin activity being maintained into adulthood differs from these studies of maternal protein restriction but may more faithfully mimic human low renin hypertension.
It has been proposed that the relationship between the RAS and aldosterone axis, of which a rennin: aldosterone ratio is a sentinel marker, may underlie essential hypertension (Pratt, 2000). Furthermore, in the low protein model of developmental programming, suppression of the renin–angiotensin system and alteration in renal function are observed (Woods et al. 2001). We did not find any evidence to support the hypothesis that maternal fat-rich dietary intake programmes altered plasma aldosterone concentration in offspring. Interestingly, the values for plasma aldosterone in the males were low compared with those observed in females. This result was not specific to either diet group, and unlikely to be due to measurement error as the intra-assay coefficient of variance was low (3.2%).
There is a reported association between low-renin hypertension and dysfunction of the Na+,K+-ATPase (Haddy & Pamnani, 1998). The activity of the renal Na+,K+-ATPase was significantly reduced by approximately 40% in OHF compared with OC animals at 360 days of age. It is not possible to attribute a causal role for the Na+,K+-ATPase dysfunction as these measurements were made on 360-day-old animals following our observation of altered renal renin activity. Thus the mechanism responsible for a reduction in the activity of the ATPase must remain speculative; however, there is strong evidence that Na+,K+-ATPase activity is modulated by the lipid environment, specifically the concentration of the long chain fatty acid docosahexaenoic acid (DHA), of the cell membrane within which the protein resides (Else et al. 1996; Wu et al. 2001; Turner et al. 2003). We have previously reported reduced DHA content in liver and aorta of OHF animals (Ghebremeskel et al. 1999; Ghosh et al. 2001). If renal DHA concentrations are similarly reduced this could provide an explanation. Alternatively, it is possible that a low Na+,K+-ATPase concentration, as opposed to a reduction in the activity of each functional unit, might explain the results. However, this is unlikely as a previous study of Na+,K+-ATPase concentration in the kidneys from several species found it to be invariant (Turner et al. 2004), despite differences in total activity.
It has been suggested that endogenous ouabain-like compounds are associated with essential hypertension and reduced Na+,K+-ATPase function in humans and experimental animals (de Wardener, 1997). However, it has also been proposed that the concentration of circulating endogenous ouabain-like substances is too low for them to act at renal sites (Di Nicolantonio et al. 1993), and instead they may activate neurones within the hypothalamic circumventricular organs (areas around the third ventricle not possessing a blood–brain barrier) (de Wardener, 1997). In addition to its important role in renal tissues, Na+,K+-ATPase activity is vital for myocardial function (Haddy & Pamnani, 1984, 1998) and Na+,K+-ATPase inhibition is associated with reduced vascular endothelial dilator function and increased contractile function (Bussemaker et al. 2002; Rossoni et al. 2002; dos Santos et al. 2003; Grbovic & Radenkovic, 2003). Hence, if this defect in Na+,K+-ATPase function were shown to be global, there may be some association with the aorta functional disturbances observed.
Conclusions
Abnormal mesenteric artery endothelial function and hypertension, previously reported in offspring of dams fed a lard-rich diet during pregnancy (Khan et al. 2003, 2004), are associated with abnormalities in aortic elasticity, smooth muscle morphology, endothelium-dependent vasodilatation and endothelial cell morphology. Reductions in renal Na+,K+-ATPase and renin activity were also observed, but plasma aldosterone concentration was not changed by maternal diet and stereological analysis revealed normal morphology in OHF kidneys.
This study highlights, for the first time, the possibility that altered conduit artery and renal function may exist in offspring of lard-fed dams, and provides a logical basis for further examination of renal function such as glomerular filtration rate, plasma electrolyte balance and a thorough investigation of the renin–angiotensin system.
Stereological analyses of aortic morphology were altered in OHF (filled symbols) compared with OC (open symbols) but not by offspring sex (male and female data are shown separately) All data are presented as the mean observation from each animal and horizontal lines show group means. A, endothelial cell volume was reduced in OHF animals compared with OC (*P < 0.03); however, there was no sex difference (P < 0.55). B, there was no effect of maternal diet (P < 0.62) nor offspring sex (P < 0.79) on endothelial cell number. C, there was no effect of maternal diet (P < 0.93) nor offspring sex (P < 0.18) on smooth muscle cell volume. D, there was a significant reduction in smooth muscle cell number in OHF animals compared with OC (†P < 0.008); however, there was no effect of offspring sex (P < 0.28). E, there was no effect of maternal diet (P < 0.44) nor sex (P < 0.19) on medial thickness.
Parameters of aortic reactivity were altered in OHF (filled symbols) compared with OC (open symbols) but not by offspring sex (male, circles; female, triangles) A, passive stiffness (measured in Ca2+ free media) was increased in OHF animals compared with OC (*P < 0.04). There was no effect of sex (P < 0.32). Data are shown as means ±s.e.m. (OC male n= 5, OHF male n= 6, OC female n= 8, OHF female n= 5). B, contractile responses to phenylephrine were not altered by maternal diet (P < 0.45) nor by offspring sex (P < 0.83). Data are shown as means ±s.e.m. (OC male n= 6, OHF male n= 6, OC female n= 5, OHF female n= 5). C, endothelium-dependent vasodilatation to ACh was blunted in OHF animals compared with OC (†P < 0.003); however, there was no effect of sex (P < 0.31). Data are shown as mean ±s.e.m. (OC male n= 7, OHF male n= 6, OC female n= 8, OHF female n= 7). D, endothelial independent vasodilatation to aqueous nitric oxide was not altered by maternal diet (P < 0.28) nor by offspring sex (P < 0.37). Data show means ±s.e.m. (OC male n= 6, OHF male n= 5, OC female n= 8, OHF female n= 5).
Stereological analyses of renal morphology were not altered in OHF (filled symbols) compared with OC (open symbols), but there was a significant effect of offspring sex (male, circles; female, triangles) All data are presented as the mean observation from each animal and horizontal lines show group means. A, there was no effect of maternal diet on glomerular volume (P < 0.12); however, there was a significant effect of offspring sex (*P < 0.001), with males having a greater glomerular volume. B, there was no effect of maternal diet on glomerular number (P < 0.75); however, there was a significant effect of offspring sex (†P < 0.04), with males having more glomeruli than females.
Activities of key renal enzymes were altered by maternal diet (OHF filled bars, OC unfilled bars) and offspring sex, but aldosterone was only affected by offspring sex A, kidney renin activity was reduced in OHF animals compared with OC (*P < 0.03) and was lower in males compared with females (†P < 0.002). Data represent means ±s.e.m. (OC male n= 9, OHF male n= 9, OC female n= 7, OHF female n= 10). B, plasma aldosterone concentration was not affected by maternal diet (P < 0.26); however, there was a significant reduction in concentration observed in males compared with females (‡P < 0.0001). Data represent means ±s.e.m. (OC male n= 7, OHF male n= 6, OC female n= 6, OHF female n= 5). C, kidney Na+,K+-ATPase activity was reduced in OHF animals compared with OC (§P < 0.003); however, for this parameter there was no effect of offspring sex (P < 0.33). Data represent means ±s.e.m. (OC male n= 10, OHF male n= 10, OC female n= 10, OHF female n= 11).
Composition and fatty acid content of control and lard-rich diets
Effect of maternal diet and sex on stereological functional and enzyme/hormone activity
Acknowledgements
This work was funded by The Guy's and St Thomas' Charitable Foundation (P.D.T., L.P.), Tommy's the Baby Charity (L.P.), The British Heart Foundation (L.P., P.D.T.) and The Wellcome Trust (J.A.A., A.A.B.). Thanks to Paul Seed for statistical assistance with regard to data transformation.
Footnotes
-
- Accepted March 17, 2005.
- Received February 11, 2005.
- Revision received March 10, 2005.
- The Physiological society 2005
